The American Journal of Orthopedics

ABSTRACT

The use of arthroscopy for purely diagnostic purposes has been largely supplanted by noninvasive technologies, such as magnetic resonance imaging (MRI). The mi-eye^+TM (Trice Medical) technology is a small-bore needle unit for in-office arthroscopy. We conducted a pilot study comparing the mi-eye^+TM unit with MRI, using surgical arthroscopy as a gold-standard reference. We hypothesized that the mi-eye^+TM needle arthroscope, which can be used in an office setting, would be equivalent to MRI for the diagnosis of intra-articular pathology of the knee.

This prospective, multicenter, observational study was approved by the Institutional Review Board. There were 106 patients (53 males, 53 females) in the study. MRIs were interpreted by musculoskeletally trained radiologists. The study was conducted in the operating room using the mi-eye^+TM device. The mi-eye^{+ TM} device findings were compared with the MRI findings within individual pathologies, and a “per-patient” analysis was performed to compare the arthroscopic findings with those of the mi-eye^+TM and the MRI. Additionally, we identified all mi-eye^+TM findings and MRI findings that exactly matched the surgical arthroscopy findings.

The mi-eye^+TM demonstrated complete accuracy of all pathologies for 97 (91.5%) of the 106 patients included in the study, whereas MRI demonstrated complete accuracy for 65 patients (61.3%) (P < .0001). All discrepancies between mi-eye^+TM and arthroscopy were false-negative mi-eye^+TM results, as the mi-eye^+TM did not reveal some aspect of the knee’s pathology for 9 patients. The mi-eye^+TM was more sensitive than MRI in identifying meniscal tears (92.6% vs 77.8%; P = .0035) and more specific in diagnosing these tears (100% vs 41.7%; P < .0001).

The mi-eye^+TM device proved to be more sensitive and specific than MRI for intra-articular findings at time of knee arthroscopy. Certainly there are contraindications to using the mi-eye^+TM, and our results do not obviate the need for MRI, but our study did demonstrate that the mi-eye^+TM needle arthroscope can safely provide excellent visualization of intra-articular knee pathology.

Continue to: Surgical arthroscopy is the gold standard...

Surgical arthroscopy is the gold standard for the diagnosis of intra-articular knee pathologies. Nevertheless, the use of arthroscopy for purely diagnostic purposes has been largely supplanted by noninvasive technologies, such as magnetic resonance imaging (MRI). Although MRI is considered the standard diagnostic tool for acute and chronic soft-tissue injuries of the knee, its use is not without contraindication and some potential inconveniences. Contraindications to MRI are well documented. In terms of inconvenience, MRI usually requires a separate visit followed by another visit to the prescribing physician. In addition, required interpretation by a radiologist may lead to a delay in care and increase in cost.

In the early 1990s, in-office needle arthroscopy was described as a viable means of diagnosing pathologies and obtaining synovial biopsies from the knee.^1-3 Initial results were good, and the procedures had very low complication rates. Nevertheless, in-office arthroscopy of the knee is not yet widely performed, likely given concerns about the technical difficulties of in-office arthroscopy, the potential for patient discomfort, and the cumbersomeness of in-office arthroscopy units. However, significant advances have been made in the resolution capability of small-bore needle arthroscopy, resulting in much less painful procedures. Additionally, the early hardware designs, which mimicked operating room setups using towers, fluid irrigation systems, and larger arthroscopes, have been replaced with small-needle arthroscopes that use syringes for irrigation and tablet computers for visualization (Figures 1A, 1B).  

The mi-eye^+TM technology (Trice Medical) is a small-bore needle unit for in-office arthroscopy with digital optics that does not need an irrigation tower. We conducted a pilot study of the sensitivity and specificity of the mi-eye^+TM unit in comparison with MRI, using surgical arthroscopy as a gold-standard reference. We hypothesized that the mi-eye^+TM needle arthroscope, which can be used in an office setting, would be equivalent to the standard of care (MRI) for the diagnosis of intra-articular pathology of the knee.

METHODS

Central regulatory approval for this prospective, multicenter, observational study was obtained from the Western Institutional Review Board for 3 of the sites, and 1 institution required and was granted internal Institutional Review Board approval.

The study was performed by 4 sports medicine orthopedic surgeons experienced in using the mi-eye^+TM in-office arthroscope. Patients were enrolled from December 2015 through June 2016. Inclusion criteria were an indication for an arthroscopic procedure of the knee based on history, physical examination, and MRI findings. Patients were excluded from the study if there were any contraindications to completing an MRI. Acute hemarthroses of the knee or active systemic infections were also excluded. Once a patient was identified as meeting the criteria for participation, informed consent was obtained. Of the 113 patients who enrolled, 7 did not have a complete study dataset available, leaving 106 patients (53 males, 53 females) in the study. Mean age was 47 years (range, 18-82 years).

Continue to: A test result form was used...

A test result form was used to record mi-eye^+TM, surgical arthroscopy, and MRI results. This form required a “positive” or “negative” result for all of several diagnoses: medial and lateral meniscal tears, intra-articular loose body, osteoarthritis (OA), osteochondritis dissecans (OCD), and tears of the anterior and posterior cruciate ligaments (ACL, PCL). MRI was performed at a variety of imaging facilities, but the images were interpreted by musculoskeletally trained radiologists.

The study was conducted in the operating room. After the patient was appropriately anesthetized, and the extremity prepared and draped, the mi-eye^+TM procedure was performed immediately prior to surgical arthroscopy. A tourniquet was not used. At surgeon discretion, medial, lateral, or both approaches were used with the mi-eye^+TM, and diagnostic arthroscopy was performed. During the procedure, the mi-eye^+TM was advanced into the knee. Once in the synovial compartment, the external 14-gauge needle was retracted, exposing the unit’s optics. Visualization was improved by injecting normal saline through the lure lock in the mi-eye^+TM needle arthroscope. An average of 20 mL of saline was used, though the amount varied with surgeon discretion. Subsequently, the surgeon visualized structures in the knee and documented all findings.

At the end of the mi-eye^+TM procedure, the scheduled surgical arthroscopy was performed. After the surgical procedure, if there were no issues or complications, the patient was discharged from the study. No follow-up was required for the study, as arthroscopic findings served as the conclusive diagnosis for each patient, and no interventions were being studied. There were no complications related to use of the mi-eye^+TM.

The mi-eye^+TM device findings were compared with the MRI findings within individual pathologies, and a “per-patient” analysis was performed to compare the arthroscopic findings with those of the mi-eye^+TM and the MRI. Additionally, we identified all mi-eye^+TM findings and MRI findings that exactly matched the surgical arthroscopy findings. When a test had no false-positive or false-negative findings in comparison with surgical arthroscopy, it was identified as having complete accuracy for all intra-articular knee pathologies. For these methods, the 95% confidence interval was determined based on binomial distribution.

RESULTS

The mi-eye^{+ TM} demonstrated complete accuracy of all pathologies for 97 (91.5%) of the 106 patients included in the study, whereas MRI demonstrated complete accuracy for 65 patients (61.3%) (P < .0001). All discrepancies between mi-eye^+TM and surgical arthroscopy were false-negative mi-eye^+TM results, as the mi-eye^+TM did not reveal some aspect of the knee’s pathology for 9 patients. On the other hand, MRI demonstrated both false-negative and false-positive results, failing to reveal some aspect of the knee’s pathology for 31 patients, and potentially overcalling some aspect of the knee’s pathology among 18 patients.

Continue to: The pathology most frequently...

The pathology most frequently identified in the study was a meniscal tear. The mi-eye^+TM was more sensitive than MRI in identifying meniscal tears (92.6% vs 77.8%; P = .0035) and more specific in diagnosing these tears (100% vs 87.5%; P < .0002). The difference in specificity resulted from the false MRI diagnosis of a meniscal tear among 24 patients, who were found to have no tear by both mi-eye^+TM and surgical arthroscopy.

Table 1. Raw Data of mi-eye^+TM and Magnetic Resonance Imaging Findings

The second most frequent pathology was an intra-articular loose body. The mi-eye^+TM was more sensitive than MRI in identifying loose bodies (86.7% vs 20%; P = .0007). The specificity of the mi-eye^+TM and the specificity of MRI were equivalent in diagnosing loose bodies (100%). Table 1 and Table 2 show the complete set of diagnoses and associated diagnostic profiles.

Table 2. Diagnostic Profiles: Sensitivity and Specificity of mi-eye^+TM and Magnetic Resonance Imaging

^aBold P values are significant. Abbreviations: CI, confidence interval; MRI, magnetic resonance imaging; N/A, not applicable.

DISCUSSION

The overall accuracy of the mi-eye^+TM was superior to that of MRI relative to the arthroscopic gold standard in this pilot study. Other studies have demonstrated the accuracy, feasibility, and cost-efficacy of in-office arthroscopy. However, likely because of the cumbersomeness of in-office arthroscopy equipment and the potential for patient discomfort, the technique is not yet standard in the field. Recent advances in small-bore technology, digital optics, and ergonomics have addressed the difficulties associated with in-office arthroscopy, facilitating a faster and more efficient procedure. Our goal in this study was to evaluate the diagnostic capability of the mi-eye^+TM in-office arthroscopy unit, which features a small bore, digital optics, and functionality without an irrigation tower.

This study of 106 patients demonstrated equivalent or better accuracy of the mi-eye^+TM relative to MRI when compared with the gold standard of surgical arthroscopy. This was not surprising given that both the mi-eye^+TM and surgical arthroscopy are based on direct visualization of intra-articular pathology. The mi-eye^+TM unit identified more meniscal tears, intra-articular loose bodies, ACL tears, and OCD lesions than MRI did, and with enough power to demonstrate statistically significant improved sensitivity for meniscal tears and loose bodies. Furthermore, MRI demonstrated false-positive meniscal tears, ACL tears, OCD lesions, and OA, whereas the mi-eye^+TM did not demonstrate any false-positive results in comparison with surgical arthroscopy. This study demonstrated statistically significant improved specificity of the mi-eye⁺ compared with MRI in the diagnosis of meniscal tears and OA.

There are several limitations to our study. We refer to it as a pilot study because it was performed in a standard operating room. Before taking the technology to an outpatient setting, we wanted to confirm efficacy and safety in an operating room. However, the techniques used in this study are readily transferable to the outpatient clinic setting and to date have been used in more than 2000 cases.

Continue to: The specificity of MRI...

The specificity of MRI for meniscal tears was unexpectedly low compared with previous studies, which may reflect the multi-institution, multi-surgeon, multi-radiologist involvement in MRI interpretation.^4-10 MRI was performed at a variety of institutions without a standardized protocol. This lack of standardization of image capture and interpretation may have contributed to the suboptimal performance of MRI, falsely decreasing the potential ideal specificity for meniscal tears. Although this study may have underestimated the specificity of MRI for meniscal tears, we think the mi-eye^+TM and MRI results reported here reflect the findings of standard practice, without the standardization usually applied in studies. For example, a study of 139 knee MRI reports at 14 different institutions confirmed arthroscopic findings and concluded that 37% of the operations supported by a significant MRI finding were unjustified.¹¹ The authors attributed the rate of false-positive MRI findings to the wide variety of places where patients had their MRIs performed, and the subsequent variation in quality of imaging and MRI reader skill level.¹¹

Before inserting the mi-eye^+TM needle arthroscope, the surgeons had a working diagnosis of the pathology based on their clinical examination and MRI results. Clearly, this introduced a bias. Further studies will be conducted in a prospective, blinded manner to address this limitation.

Although studies of in-office arthroscopy technology date to the 1990s, there is an overall lack of data comparing in-office arthroscopy with MRI. Halbrecht and Jackson² conducted a study of 20 knee patients with both MRI and in-office needle arthroscopy. Overall, MRI was poor in detecting cartilage defects, with sensitivity of 34.6%, using the in-office arthroscopy as the confirmatory diagnosis. Although the authors did not compare in-office diagnoses with surgical arthroscopic findings, they concluded that office arthroscopy is an accurate and cost-efficient alternative to MRI in diagnostic evaluation of knee patients. Xerogeanes and colleagues¹² studied 110 patients in a prospective, blinded, multicenter trial comparing a minimally invasive office-based arthroscopy with MRI, using surgical arthroscopy as the confirmatory diagnosis. They concluded that the office-based arthroscope was statistically equivalent to diagnostic surgical arthroscopy and that it outperformed MRI in helping make accurate diagnoses. The authors applied a cost analysis to their findings and determined that office-based arthroscopy could result in an annual potential savings of $177 million for the healthcare system.¹²

Modern imaging sequences on high-Tesla MRI machines provide excellent visualization. Nevertheless, a significant number of patients do not undergo MRI, owing to time constraints, contraindications, body habitus, or anxiety/claustrophobia. Our study results confirmed that doctors treating such patients now have a viable alternative to help diagnose pathology.

CONCLUSION

The mi-eye^+TM device proved to be more sensitive and specific than MRI for intra-articular findings at the time of knee arthroscopy. Certainly there are contraindications to using the mi-eye^+TM, and our results do not obviate the need for MRI; our study did demonstrate that the mi-eye^+TM needle arthroscope can safely provide excellent visualization of intra-articular knee pathology. More studies of the mi-eye^+TM device in a clinical setting are warranted.

ABSTRACT

The use of arthroscopy for purely diagnostic purposes has been largely supplanted by noninvasive technologies, such as magnetic resonance imaging (MRI). The mi-eye^+TM (Trice Medical) technology is a small-bore needle unit for in-office arthroscopy. We conducted a pilot study comparing the mi-eye^+TM unit with MRI, using surgical arthroscopy as a gold-standard reference. We hypothesized that the mi-eye^+TM needle arthroscope, which can be used in an office setting, would be equivalent to MRI for the diagnosis of intra-articular pathology of the knee.

This prospective, multicenter, observational study was approved by the Institutional Review Board. There were 106 patients (53 males, 53 females) in the study. MRIs were interpreted by musculoskeletally trained radiologists. The study was conducted in the operating room using the mi-eye^+TM device. The mi-eye^{+ TM} device findings were compared with the MRI findings within individual pathologies, and a “per-patient” analysis was performed to compare the arthroscopic findings with those of the mi-eye^+TM and the MRI. Additionally, we identified all mi-eye^+TM findings and MRI findings that exactly matched the surgical arthroscopy findings.

The mi-eye^+TM demonstrated complete accuracy of all pathologies for 97 (91.5%) of the 106 patients included in the study, whereas MRI demonstrated complete accuracy for 65 patients (61.3%) (P < .0001). All discrepancies between mi-eye^+TM and arthroscopy were false-negative mi-eye^+TM results, as the mi-eye^+TM did not reveal some aspect of the knee’s pathology for 9 patients. The mi-eye^+TM was more sensitive than MRI in identifying meniscal tears (92.6% vs 77.8%; P = .0035) and more specific in diagnosing these tears (100% vs 41.7%; P < .0001).

The mi-eye^+TM device proved to be more sensitive and specific than MRI for intra-articular findings at time of knee arthroscopy. Certainly there are contraindications to using the mi-eye^+TM, and our results do not obviate the need for MRI, but our study did demonstrate that the mi-eye^+TM needle arthroscope can safely provide excellent visualization of intra-articular knee pathology.

Continue to: Surgical arthroscopy is the gold standard...

Surgical arthroscopy is the gold standard for the diagnosis of intra-articular knee pathologies. Nevertheless, the use of arthroscopy for purely diagnostic purposes has been largely supplanted by noninvasive technologies, such as magnetic resonance imaging (MRI). Although MRI is considered the standard diagnostic tool for acute and chronic soft-tissue injuries of the knee, its use is not without contraindication and some potential inconveniences. Contraindications to MRI are well documented. In terms of inconvenience, MRI usually requires a separate visit followed by another visit to the prescribing physician. In addition, required interpretation by a radiologist may lead to a delay in care and increase in cost.

In the early 1990s, in-office needle arthroscopy was described as a viable means of diagnosing pathologies and obtaining synovial biopsies from the knee.^1-3 Initial results were good, and the procedures had very low complication rates. Nevertheless, in-office arthroscopy of the knee is not yet widely performed, likely given concerns about the technical difficulties of in-office arthroscopy, the potential for patient discomfort, and the cumbersomeness of in-office arthroscopy units. However, significant advances have been made in the resolution capability of small-bore needle arthroscopy, resulting in much less painful procedures. Additionally, the early hardware designs, which mimicked operating room setups using towers, fluid irrigation systems, and larger arthroscopes, have been replaced with small-needle arthroscopes that use syringes for irrigation and tablet computers for visualization (Figures 1A, 1B).  

The mi-eye^+TM technology (Trice Medical) is a small-bore needle unit for in-office arthroscopy with digital optics that does not need an irrigation tower. We conducted a pilot study of the sensitivity and specificity of the mi-eye^+TM unit in comparison with MRI, using surgical arthroscopy as a gold-standard reference. We hypothesized that the mi-eye^+TM needle arthroscope, which can be used in an office setting, would be equivalent to the standard of care (MRI) for the diagnosis of intra-articular pathology of the knee.

METHODS

Central regulatory approval for this prospective, multicenter, observational study was obtained from the Western Institutional Review Board for 3 of the sites, and 1 institution required and was granted internal Institutional Review Board approval.

The study was performed by 4 sports medicine orthopedic surgeons experienced in using the mi-eye^+TM in-office arthroscope. Patients were enrolled from December 2015 through June 2016. Inclusion criteria were an indication for an arthroscopic procedure of the knee based on history, physical examination, and MRI findings. Patients were excluded from the study if there were any contraindications to completing an MRI. Acute hemarthroses of the knee or active systemic infections were also excluded. Once a patient was identified as meeting the criteria for participation, informed consent was obtained. Of the 113 patients who enrolled, 7 did not have a complete study dataset available, leaving 106 patients (53 males, 53 females) in the study. Mean age was 47 years (range, 18-82 years).

Continue to: A test result form was used...

A test result form was used to record mi-eye^+TM, surgical arthroscopy, and MRI results. This form required a “positive” or “negative” result for all of several diagnoses: medial and lateral meniscal tears, intra-articular loose body, osteoarthritis (OA), osteochondritis dissecans (OCD), and tears of the anterior and posterior cruciate ligaments (ACL, PCL). MRI was performed at a variety of imaging facilities, but the images were interpreted by musculoskeletally trained radiologists.

The study was conducted in the operating room. After the patient was appropriately anesthetized, and the extremity prepared and draped, the mi-eye^+TM procedure was performed immediately prior to surgical arthroscopy. A tourniquet was not used. At surgeon discretion, medial, lateral, or both approaches were used with the mi-eye^+TM, and diagnostic arthroscopy was performed. During the procedure, the mi-eye^+TM was advanced into the knee. Once in the synovial compartment, the external 14-gauge needle was retracted, exposing the unit’s optics. Visualization was improved by injecting normal saline through the lure lock in the mi-eye^+TM needle arthroscope. An average of 20 mL of saline was used, though the amount varied with surgeon discretion. Subsequently, the surgeon visualized structures in the knee and documented all findings.

At the end of the mi-eye^+TM procedure, the scheduled surgical arthroscopy was performed. After the surgical procedure, if there were no issues or complications, the patient was discharged from the study. No follow-up was required for the study, as arthroscopic findings served as the conclusive diagnosis for each patient, and no interventions were being studied. There were no complications related to use of the mi-eye^+TM.

The mi-eye^+TM device findings were compared with the MRI findings within individual pathologies, and a “per-patient” analysis was performed to compare the arthroscopic findings with those of the mi-eye^+TM and the MRI. Additionally, we identified all mi-eye^+TM findings and MRI findings that exactly matched the surgical arthroscopy findings. When a test had no false-positive or false-negative findings in comparison with surgical arthroscopy, it was identified as having complete accuracy for all intra-articular knee pathologies. For these methods, the 95% confidence interval was determined based on binomial distribution.

RESULTS

The mi-eye^{+ TM} demonstrated complete accuracy of all pathologies for 97 (91.5%) of the 106 patients included in the study, whereas MRI demonstrated complete accuracy for 65 patients (61.3%) (P < .0001). All discrepancies between mi-eye^+TM and surgical arthroscopy were false-negative mi-eye^+TM results, as the mi-eye^+TM did not reveal some aspect of the knee’s pathology for 9 patients. On the other hand, MRI demonstrated both false-negative and false-positive results, failing to reveal some aspect of the knee’s pathology for 31 patients, and potentially overcalling some aspect of the knee’s pathology among 18 patients.

Continue to: The pathology most frequently...

The pathology most frequently identified in the study was a meniscal tear. The mi-eye^+TM was more sensitive than MRI in identifying meniscal tears (92.6% vs 77.8%; P = .0035) and more specific in diagnosing these tears (100% vs 87.5%; P < .0002). The difference in specificity resulted from the false MRI diagnosis of a meniscal tear among 24 patients, who were found to have no tear by both mi-eye^+TM and surgical arthroscopy.

Table 1. Raw Data of mi-eye^+TM and Magnetic Resonance Imaging Findings

The second most frequent pathology was an intra-articular loose body. The mi-eye^+TM was more sensitive than MRI in identifying loose bodies (86.7% vs 20%; P = .0007). The specificity of the mi-eye^+TM and the specificity of MRI were equivalent in diagnosing loose bodies (100%). Table 1 and Table 2 show the complete set of diagnoses and associated diagnostic profiles.

Table 2. Diagnostic Profiles: Sensitivity and Specificity of mi-eye^+TM and Magnetic Resonance Imaging

^aBold P values are significant. Abbreviations: CI, confidence interval; MRI, magnetic resonance imaging; N/A, not applicable.

DISCUSSION

The overall accuracy of the mi-eye^+TM was superior to that of MRI relative to the arthroscopic gold standard in this pilot study. Other studies have demonstrated the accuracy, feasibility, and cost-efficacy of in-office arthroscopy. However, likely because of the cumbersomeness of in-office arthroscopy equipment and the potential for patient discomfort, the technique is not yet standard in the field. Recent advances in small-bore technology, digital optics, and ergonomics have addressed the difficulties associated with in-office arthroscopy, facilitating a faster and more efficient procedure. Our goal in this study was to evaluate the diagnostic capability of the mi-eye^+TM in-office arthroscopy unit, which features a small bore, digital optics, and functionality without an irrigation tower.

This study of 106 patients demonstrated equivalent or better accuracy of the mi-eye^+TM relative to MRI when compared with the gold standard of surgical arthroscopy. This was not surprising given that both the mi-eye^+TM and surgical arthroscopy are based on direct visualization of intra-articular pathology. The mi-eye^+TM unit identified more meniscal tears, intra-articular loose bodies, ACL tears, and OCD lesions than MRI did, and with enough power to demonstrate statistically significant improved sensitivity for meniscal tears and loose bodies. Furthermore, MRI demonstrated false-positive meniscal tears, ACL tears, OCD lesions, and OA, whereas the mi-eye^+TM did not demonstrate any false-positive results in comparison with surgical arthroscopy. This study demonstrated statistically significant improved specificity of the mi-eye⁺ compared with MRI in the diagnosis of meniscal tears and OA.

There are several limitations to our study. We refer to it as a pilot study because it was performed in a standard operating room. Before taking the technology to an outpatient setting, we wanted to confirm efficacy and safety in an operating room. However, the techniques used in this study are readily transferable to the outpatient clinic setting and to date have been used in more than 2000 cases.

Continue to: The specificity of MRI...

The specificity of MRI for meniscal tears was unexpectedly low compared with previous studies, which may reflect the multi-institution, multi-surgeon, multi-radiologist involvement in MRI interpretation.^4-10 MRI was performed at a variety of institutions without a standardized protocol. This lack of standardization of image capture and interpretation may have contributed to the suboptimal performance of MRI, falsely decreasing the potential ideal specificity for meniscal tears. Although this study may have underestimated the specificity of MRI for meniscal tears, we think the mi-eye^+TM and MRI results reported here reflect the findings of standard practice, without the standardization usually applied in studies. For example, a study of 139 knee MRI reports at 14 different institutions confirmed arthroscopic findings and concluded that 37% of the operations supported by a significant MRI finding were unjustified.¹¹ The authors attributed the rate of false-positive MRI findings to the wide variety of places where patients had their MRIs performed, and the subsequent variation in quality of imaging and MRI reader skill level.¹¹

Before inserting the mi-eye^+TM needle arthroscope, the surgeons had a working diagnosis of the pathology based on their clinical examination and MRI results. Clearly, this introduced a bias. Further studies will be conducted in a prospective, blinded manner to address this limitation.

Although studies of in-office arthroscopy technology date to the 1990s, there is an overall lack of data comparing in-office arthroscopy with MRI. Halbrecht and Jackson² conducted a study of 20 knee patients with both MRI and in-office needle arthroscopy. Overall, MRI was poor in detecting cartilage defects, with sensitivity of 34.6%, using the in-office arthroscopy as the confirmatory diagnosis. Although the authors did not compare in-office diagnoses with surgical arthroscopic findings, they concluded that office arthroscopy is an accurate and cost-efficient alternative to MRI in diagnostic evaluation of knee patients. Xerogeanes and colleagues¹² studied 110 patients in a prospective, blinded, multicenter trial comparing a minimally invasive office-based arthroscopy with MRI, using surgical arthroscopy as the confirmatory diagnosis. They concluded that the office-based arthroscope was statistically equivalent to diagnostic surgical arthroscopy and that it outperformed MRI in helping make accurate diagnoses. The authors applied a cost analysis to their findings and determined that office-based arthroscopy could result in an annual potential savings of $177 million for the healthcare system.¹²

Modern imaging sequences on high-Tesla MRI machines provide excellent visualization. Nevertheless, a significant number of patients do not undergo MRI, owing to time constraints, contraindications, body habitus, or anxiety/claustrophobia. Our study results confirmed that doctors treating such patients now have a viable alternative to help diagnose pathology.

CONCLUSION

The mi-eye^+TM device proved to be more sensitive and specific than MRI for intra-articular findings at the time of knee arthroscopy. Certainly there are contraindications to using the mi-eye^+TM, and our results do not obviate the need for MRI; our study did demonstrate that the mi-eye^+TM needle arthroscope can safely provide excellent visualization of intra-articular knee pathology. More studies of the mi-eye^+TM device in a clinical setting are warranted.

ABSTRACT

The use of arthroscopy for purely diagnostic purposes has been largely supplanted by noninvasive technologies, such as magnetic resonance imaging (MRI). The mi-eye^+TM (Trice Medical) technology is a small-bore needle unit for in-office arthroscopy. We conducted a pilot study comparing the mi-eye^+TM unit with MRI, using surgical arthroscopy as a gold-standard reference. We hypothesized that the mi-eye^+TM needle arthroscope, which can be used in an office setting, would be equivalent to MRI for the diagnosis of intra-articular pathology of the knee.

This prospective, multicenter, observational study was approved by the Institutional Review Board. There were 106 patients (53 males, 53 females) in the study. MRIs were interpreted by musculoskeletally trained radiologists. The study was conducted in the operating room using the mi-eye^+TM device. The mi-eye^{+ TM} device findings were compared with the MRI findings within individual pathologies, and a “per-patient” analysis was performed to compare the arthroscopic findings with those of the mi-eye^+TM and the MRI. Additionally, we identified all mi-eye^+TM findings and MRI findings that exactly matched the surgical arthroscopy findings.

The mi-eye^+TM demonstrated complete accuracy of all pathologies for 97 (91.5%) of the 106 patients included in the study, whereas MRI demonstrated complete accuracy for 65 patients (61.3%) (P < .0001). All discrepancies between mi-eye^+TM and arthroscopy were false-negative mi-eye^+TM results, as the mi-eye^+TM did not reveal some aspect of the knee’s pathology for 9 patients. The mi-eye^+TM was more sensitive than MRI in identifying meniscal tears (92.6% vs 77.8%; P = .0035) and more specific in diagnosing these tears (100% vs 41.7%; P < .0001).

The mi-eye^+TM device proved to be more sensitive and specific than MRI for intra-articular findings at time of knee arthroscopy. Certainly there are contraindications to using the mi-eye^+TM, and our results do not obviate the need for MRI, but our study did demonstrate that the mi-eye^+TM needle arthroscope can safely provide excellent visualization of intra-articular knee pathology.

Continue to: Surgical arthroscopy is the gold standard...

Surgical arthroscopy is the gold standard for the diagnosis of intra-articular knee pathologies. Nevertheless, the use of arthroscopy for purely diagnostic purposes has been largely supplanted by noninvasive technologies, such as magnetic resonance imaging (MRI). Although MRI is considered the standard diagnostic tool for acute and chronic soft-tissue injuries of the knee, its use is not without contraindication and some potential inconveniences. Contraindications to MRI are well documented. In terms of inconvenience, MRI usually requires a separate visit followed by another visit to the prescribing physician. In addition, required interpretation by a radiologist may lead to a delay in care and increase in cost.

In the early 1990s, in-office needle arthroscopy was described as a viable means of diagnosing pathologies and obtaining synovial biopsies from the knee.^1-3 Initial results were good, and the procedures had very low complication rates. Nevertheless, in-office arthroscopy of the knee is not yet widely performed, likely given concerns about the technical difficulties of in-office arthroscopy, the potential for patient discomfort, and the cumbersomeness of in-office arthroscopy units. However, significant advances have been made in the resolution capability of small-bore needle arthroscopy, resulting in much less painful procedures. Additionally, the early hardware designs, which mimicked operating room setups using towers, fluid irrigation systems, and larger arthroscopes, have been replaced with small-needle arthroscopes that use syringes for irrigation and tablet computers for visualization (Figures 1A, 1B).  

The mi-eye^+TM technology (Trice Medical) is a small-bore needle unit for in-office arthroscopy with digital optics that does not need an irrigation tower. We conducted a pilot study of the sensitivity and specificity of the mi-eye^+TM unit in comparison with MRI, using surgical arthroscopy as a gold-standard reference. We hypothesized that the mi-eye^+TM needle arthroscope, which can be used in an office setting, would be equivalent to the standard of care (MRI) for the diagnosis of intra-articular pathology of the knee.

METHODS

Central regulatory approval for this prospective, multicenter, observational study was obtained from the Western Institutional Review Board for 3 of the sites, and 1 institution required and was granted internal Institutional Review Board approval.

The study was performed by 4 sports medicine orthopedic surgeons experienced in using the mi-eye^+TM in-office arthroscope. Patients were enrolled from December 2015 through June 2016. Inclusion criteria were an indication for an arthroscopic procedure of the knee based on history, physical examination, and MRI findings. Patients were excluded from the study if there were any contraindications to completing an MRI. Acute hemarthroses of the knee or active systemic infections were also excluded. Once a patient was identified as meeting the criteria for participation, informed consent was obtained. Of the 113 patients who enrolled, 7 did not have a complete study dataset available, leaving 106 patients (53 males, 53 females) in the study. Mean age was 47 years (range, 18-82 years).

Continue to: A test result form was used...

A test result form was used to record mi-eye^+TM, surgical arthroscopy, and MRI results. This form required a “positive” or “negative” result for all of several diagnoses: medial and lateral meniscal tears, intra-articular loose body, osteoarthritis (OA), osteochondritis dissecans (OCD), and tears of the anterior and posterior cruciate ligaments (ACL, PCL). MRI was performed at a variety of imaging facilities, but the images were interpreted by musculoskeletally trained radiologists.

The study was conducted in the operating room. After the patient was appropriately anesthetized, and the extremity prepared and draped, the mi-eye^+TM procedure was performed immediately prior to surgical arthroscopy. A tourniquet was not used. At surgeon discretion, medial, lateral, or both approaches were used with the mi-eye^+TM, and diagnostic arthroscopy was performed. During the procedure, the mi-eye^+TM was advanced into the knee. Once in the synovial compartment, the external 14-gauge needle was retracted, exposing the unit’s optics. Visualization was improved by injecting normal saline through the lure lock in the mi-eye^+TM needle arthroscope. An average of 20 mL of saline was used, though the amount varied with surgeon discretion. Subsequently, the surgeon visualized structures in the knee and documented all findings.

At the end of the mi-eye^+TM procedure, the scheduled surgical arthroscopy was performed. After the surgical procedure, if there were no issues or complications, the patient was discharged from the study. No follow-up was required for the study, as arthroscopic findings served as the conclusive diagnosis for each patient, and no interventions were being studied. There were no complications related to use of the mi-eye^+TM.

The mi-eye^+TM device findings were compared with the MRI findings within individual pathologies, and a “per-patient” analysis was performed to compare the arthroscopic findings with those of the mi-eye^+TM and the MRI. Additionally, we identified all mi-eye^+TM findings and MRI findings that exactly matched the surgical arthroscopy findings. When a test had no false-positive or false-negative findings in comparison with surgical arthroscopy, it was identified as having complete accuracy for all intra-articular knee pathologies. For these methods, the 95% confidence interval was determined based on binomial distribution.

RESULTS

The mi-eye^{+ TM} demonstrated complete accuracy of all pathologies for 97 (91.5%) of the 106 patients included in the study, whereas MRI demonstrated complete accuracy for 65 patients (61.3%) (P < .0001). All discrepancies between mi-eye^+TM and surgical arthroscopy were false-negative mi-eye^+TM results, as the mi-eye^+TM did not reveal some aspect of the knee’s pathology for 9 patients. On the other hand, MRI demonstrated both false-negative and false-positive results, failing to reveal some aspect of the knee’s pathology for 31 patients, and potentially overcalling some aspect of the knee’s pathology among 18 patients.

Continue to: The pathology most frequently...

The pathology most frequently identified in the study was a meniscal tear. The mi-eye^+TM was more sensitive than MRI in identifying meniscal tears (92.6% vs 77.8%; P = .0035) and more specific in diagnosing these tears (100% vs 87.5%; P < .0002). The difference in specificity resulted from the false MRI diagnosis of a meniscal tear among 24 patients, who were found to have no tear by both mi-eye^+TM and surgical arthroscopy.

Table 1. Raw Data of mi-eye^+TM and Magnetic Resonance Imaging Findings

The second most frequent pathology was an intra-articular loose body. The mi-eye^+TM was more sensitive than MRI in identifying loose bodies (86.7% vs 20%; P = .0007). The specificity of the mi-eye^+TM and the specificity of MRI were equivalent in diagnosing loose bodies (100%). Table 1 and Table 2 show the complete set of diagnoses and associated diagnostic profiles.

Table 2. Diagnostic Profiles: Sensitivity and Specificity of mi-eye^+TM and Magnetic Resonance Imaging

^aBold P values are significant. Abbreviations: CI, confidence interval; MRI, magnetic resonance imaging; N/A, not applicable.

DISCUSSION

The overall accuracy of the mi-eye^+TM was superior to that of MRI relative to the arthroscopic gold standard in this pilot study. Other studies have demonstrated the accuracy, feasibility, and cost-efficacy of in-office arthroscopy. However, likely because of the cumbersomeness of in-office arthroscopy equipment and the potential for patient discomfort, the technique is not yet standard in the field. Recent advances in small-bore technology, digital optics, and ergonomics have addressed the difficulties associated with in-office arthroscopy, facilitating a faster and more efficient procedure. Our goal in this study was to evaluate the diagnostic capability of the mi-eye^+TM in-office arthroscopy unit, which features a small bore, digital optics, and functionality without an irrigation tower.

This study of 106 patients demonstrated equivalent or better accuracy of the mi-eye^+TM relative to MRI when compared with the gold standard of surgical arthroscopy. This was not surprising given that both the mi-eye^+TM and surgical arthroscopy are based on direct visualization of intra-articular pathology. The mi-eye^+TM unit identified more meniscal tears, intra-articular loose bodies, ACL tears, and OCD lesions than MRI did, and with enough power to demonstrate statistically significant improved sensitivity for meniscal tears and loose bodies. Furthermore, MRI demonstrated false-positive meniscal tears, ACL tears, OCD lesions, and OA, whereas the mi-eye^+TM did not demonstrate any false-positive results in comparison with surgical arthroscopy. This study demonstrated statistically significant improved specificity of the mi-eye⁺ compared with MRI in the diagnosis of meniscal tears and OA.

There are several limitations to our study. We refer to it as a pilot study because it was performed in a standard operating room. Before taking the technology to an outpatient setting, we wanted to confirm efficacy and safety in an operating room. However, the techniques used in this study are readily transferable to the outpatient clinic setting and to date have been used in more than 2000 cases.

Continue to: The specificity of MRI...

The specificity of MRI for meniscal tears was unexpectedly low compared with previous studies, which may reflect the multi-institution, multi-surgeon, multi-radiologist involvement in MRI interpretation.^4-10 MRI was performed at a variety of institutions without a standardized protocol. This lack of standardization of image capture and interpretation may have contributed to the suboptimal performance of MRI, falsely decreasing the potential ideal specificity for meniscal tears. Although this study may have underestimated the specificity of MRI for meniscal tears, we think the mi-eye^+TM and MRI results reported here reflect the findings of standard practice, without the standardization usually applied in studies. For example, a study of 139 knee MRI reports at 14 different institutions confirmed arthroscopic findings and concluded that 37% of the operations supported by a significant MRI finding were unjustified.¹¹ The authors attributed the rate of false-positive MRI findings to the wide variety of places where patients had their MRIs performed, and the subsequent variation in quality of imaging and MRI reader skill level.¹¹

Before inserting the mi-eye^+TM needle arthroscope, the surgeons had a working diagnosis of the pathology based on their clinical examination and MRI results. Clearly, this introduced a bias. Further studies will be conducted in a prospective, blinded manner to address this limitation.

Although studies of in-office arthroscopy technology date to the 1990s, there is an overall lack of data comparing in-office arthroscopy with MRI. Halbrecht and Jackson² conducted a study of 20 knee patients with both MRI and in-office needle arthroscopy. Overall, MRI was poor in detecting cartilage defects, with sensitivity of 34.6%, using the in-office arthroscopy as the confirmatory diagnosis. Although the authors did not compare in-office diagnoses with surgical arthroscopic findings, they concluded that office arthroscopy is an accurate and cost-efficient alternative to MRI in diagnostic evaluation of knee patients. Xerogeanes and colleagues¹² studied 110 patients in a prospective, blinded, multicenter trial comparing a minimally invasive office-based arthroscopy with MRI, using surgical arthroscopy as the confirmatory diagnosis. They concluded that the office-based arthroscope was statistically equivalent to diagnostic surgical arthroscopy and that it outperformed MRI in helping make accurate diagnoses. The authors applied a cost analysis to their findings and determined that office-based arthroscopy could result in an annual potential savings of $177 million for the healthcare system.¹²

Modern imaging sequences on high-Tesla MRI machines provide excellent visualization. Nevertheless, a significant number of patients do not undergo MRI, owing to time constraints, contraindications, body habitus, or anxiety/claustrophobia. Our study results confirmed that doctors treating such patients now have a viable alternative to help diagnose pathology.

CONCLUSION

The mi-eye^+TM device proved to be more sensitive and specific than MRI for intra-articular findings at the time of knee arthroscopy. Certainly there are contraindications to using the mi-eye^+TM, and our results do not obviate the need for MRI; our study did demonstrate that the mi-eye^+TM needle arthroscope can safely provide excellent visualization of intra-articular knee pathology. More studies of the mi-eye^+TM device in a clinical setting are warranted.

ABSTRACT

Proper reconstruction of proximal humeral anatomy is of primary importance to maximize patient outcomes after total shoulder arthroplasty. This article describes a new arthroplasty technique, where a fixed multiplanar bone resection is made and a novel implant, which is designed to precisely match the bone resection, is inserted. 

Continue to: The success of total shoulder arthroplasty...

The success of total shoulder arthroplasty (TSA) is largely dependent on how accurate the proximal humeral anatomy is reconstructed and the glenohumeral relationships are restored.^1-4 Numerous studies have demonstrated a relationship of worse clinical outcomes and implant failure with nonanatomic implant placement.^5-8 The majority of arthroplasty systems rely on surgeon-dependent decision-making to determine the location of the border of the articular surface and, ultimately, the amount and location of bone to be resected. Even in experienced hands, the ability to reproducibly restore the joint line is inconsistent.³

In contrast, the majority of total knee arthroplasty (TKA) systems have been designed with instrumentation that guides the surgeon precisely regarding where and how much femoral bone must be resected, and the corresponding implant is designed with the same thickness to preserve the location of the joint line. Cutting block instrumentation rather than freehand cuts enables reproducibility of TKA while being performed for an estimated 700,000 times annually in the US.⁹

To achieve similar high levels of reproducibility in shoulder arthroplasty, a new technique was developed based on the principle of providing instrumentation to assist the surgeon in accurately restoring the proximal humeral joint line. This technical article describes the technique of using a multiplanar instrumented cutting system and matching implants to perform TSA. The technique shown was previously studied and was found to allow surgeons to recreate the original anatomy of the humerus with very high precision.¹⁰

The humeral prosthesis described in this article has an articular surface that is slightly elliptical to more closely match the actual shape of the humerus bone.¹¹ Biomechanical studies have demonstrated that implants designed with a nonspherical shape have more similar motion and kinematics to those of the native humeral head.¹² The undersurface of the implant has a concave four-plane geometry that matches with the bone cuts created by the cutting guides (Figures 1, 2). 

This provides rotation stability, and the implant rests on the strong subchondral bone of the proximal humerus proximal to the anatomic neck rather than relying on metaphyseal bone or canal fixation, as recommended by Aldoiusti.¹³ It also allows optimal implant placement with complete freedom with respect to inclination, version, and medial/posterior offset from the humeral canal. 

Continue to: The implant respects the relationship...

The implant respects the relationship of the rotator cuff insertion and has a recessed superior margin to keep both the implant and the saw blade 3 mm to 5 mm away from the supraspinatus fibers to protect the rotator cuff from iatrogenic injury.

TECHNIQUE

The technique described in this article uses the Catalyst CSR Total Shoulder System (Catalyst OrthoScience), which was cleared to treat arthritis of the shoulder by the US Food and Drug Administration in May 2016.

A standard deltopectoral incision is made, and the surgeon dissects the interval between the pectoralis major medially and the deltoid laterally. The subscapularis can be incised by tenotomy; alternatively, the surgeon can perform a subscapularis peel or a lesser tuberosity osteotomy using this technique.

Once the glenohumeral joint is exposed, the surgeon delivers the humeral head anteriorly. A preferred method is to place a Darrach retractor between the humeral head and the glenoid, and a cobra or a second Darrach retractor behind the superolateral humeral head superficial to the supraspinatus tendon. By simultaneously pressing on both retractors and externally rotating the patient’s arm, the humeral head is delivered anteriorly. Osteophytes on the anterior and inferior edge of the humeral head are generously removed at this time using a rongeur.

Using a pin guide, the long 3.2-mm guidewire pin is drilled under power into the center of the articular surface. The pin guide is then removed, leaving the pin in the center of the humerus (Figure 3).

Continue to: Next, the surgeon...

Next, the surgeon slides the cannulated reamer over the long guidewire pin and under power removes a small portion of the humeral head subchondral bone until the surgeon feels and observes that the reamer is no longer removing bone (Figure 4). The patent-pending reamer design prevents the surgeon from removing more than a few millimeters of bone, after which point the reamer spins on the surface of the bone without resecting further.

The surgeon is aware that the reamer has achieved its desired depth when it is no longer creating new bone shavings, and the surgeon can hear and feel that the reamer is spinning and no longer cutting. Then the surgeon removes the reamer.

The surgeon places the first humeral cut guide over the long guidewire pin, oriented superiorly-inferiorly and secures the guide using 4 short pins, and the long pin is removed. The surgeon uses an oscillating saw to cut the anterior and posterior plane cuts through the saw captures in the cut guide (Figure 5). The humeral cut guide and short pins are removed (Figure 6).

The surgeon then applies the second humeral cut guide to the proximal humerus and secures it using 2 short pins. The surgeon then uses the 6-mm drill to drill the 4 holes for the pegs of the implant. The top portion of the guide is removed, and the surgeon makes the superior and inferior cuts along the top and bottom surfaces of the guide using an oscillating saw (Figure 7).

The surgeon then uses a rongeur to slightly round the edges of the 4 corners at the periphery of the humerus. The second humeral cut guide and short pins are removed (Figure 8).

Continue to: Next, the surgeon trials...

Next, the surgeon trials humeral implants to determine the correct implant size (Figure 9). Once the proper humeral size is chosen, the trial is removed and the humeral cover is placed over the prepared humeral head. The surgeon then proceeds to glenoid preparation (Figure 10), which is easily accessible and facilitated by angled planar cuts on the humeral head. Glenoid technique will be discussed in a subsequent article.

After glenoid preparation and insertion, the humerus is delivered anteriorly. The proximal humerus is washed and dried, and cement is applied to the peg holes in the humerus bone and the underside of the humeral implant. The implant is then inserted using the humeral impactor to apply pressure and assure that the implant is fully seated. Once the humeral cement is hardened, the glenohumeral joint is irrigated and closure begins. Postoperative radiograph is shown in Figure 11.

DISCUSSION

Numerous authors have demonstrated that accurate implant placement is crucial for restoring normal glenoid kinematics and motion,^1-4 while some authors have reported worsening clinical outcomes and higher rates of pain and implant loosening when the implants were not placed anatomically.^5-8 This is such an important concept that it essentially was the primary inspiration for creating this TSA system. In addition, the system utilizes a nonspherical, elliptical humeral head that more closely matches the anatomy of the proximal humerus,^14,15 and this type of shape has shown improved biomechanics in laboratory testing.¹²

Good results have been demonstrated in restoring the normal anatomy using stemmed devices on the radiographic analysis of cadavers.¹⁶ The creation of stemmed implants with variable inclination and offset has improved computer models¹⁷ compared with previous studies,¹⁸ with the exception of scenarios with extreme offset. 

In theory, resurfacing implants and implants without a canal stem should have a better implant placement than that with stemmed implants; however, the ability to restore the center of rotation was even worse for resurfacing prostheses, with 65% of all implants being measured as outliers postoperatively in one study.¹⁹ Most of the resurfacing implants and their instrumentation techniques offer little to help the surgeon control for implant height. The depth of the reaming is variable, not calibrated, and not correlated with the implant size, frequently leading to overstuffing after surgery. Second, the use of spherical prostheses forces the surgeon to choose between matching the superior-inferior humeral size, leading to overhang of the implant, or matching the anteroposterior, leading to frequent undersizing in the coronal plane. The nonspherical, elliptical head shape can potentially simplify implant selection.

In summary, new techniques have been developed in an attempt to achieve increased consistency and precision in TSA. By more accurately reproducing the proximal humeral anatomy, it is proposed that clinical outcomes in terms of the range of motion and patient satisfaction may also be improved through newer techniques. Cadaver studies have validated the anatomic precision of this technique.¹⁰ Clinical data comprising of patient-reported outcome measures and radiographic outcome studies are currently underway for this arthroplasty system.

ABSTRACT

Proper reconstruction of proximal humeral anatomy is of primary importance to maximize patient outcomes after total shoulder arthroplasty. This article describes a new arthroplasty technique, where a fixed multiplanar bone resection is made and a novel implant, which is designed to precisely match the bone resection, is inserted. 

Continue to: The success of total shoulder arthroplasty...

The success of total shoulder arthroplasty (TSA) is largely dependent on how accurate the proximal humeral anatomy is reconstructed and the glenohumeral relationships are restored.^1-4 Numerous studies have demonstrated a relationship of worse clinical outcomes and implant failure with nonanatomic implant placement.^5-8 The majority of arthroplasty systems rely on surgeon-dependent decision-making to determine the location of the border of the articular surface and, ultimately, the amount and location of bone to be resected. Even in experienced hands, the ability to reproducibly restore the joint line is inconsistent.³

In contrast, the majority of total knee arthroplasty (TKA) systems have been designed with instrumentation that guides the surgeon precisely regarding where and how much femoral bone must be resected, and the corresponding implant is designed with the same thickness to preserve the location of the joint line. Cutting block instrumentation rather than freehand cuts enables reproducibility of TKA while being performed for an estimated 700,000 times annually in the US.⁹

To achieve similar high levels of reproducibility in shoulder arthroplasty, a new technique was developed based on the principle of providing instrumentation to assist the surgeon in accurately restoring the proximal humeral joint line. This technical article describes the technique of using a multiplanar instrumented cutting system and matching implants to perform TSA. The technique shown was previously studied and was found to allow surgeons to recreate the original anatomy of the humerus with very high precision.¹⁰

The humeral prosthesis described in this article has an articular surface that is slightly elliptical to more closely match the actual shape of the humerus bone.¹¹ Biomechanical studies have demonstrated that implants designed with a nonspherical shape have more similar motion and kinematics to those of the native humeral head.¹² The undersurface of the implant has a concave four-plane geometry that matches with the bone cuts created by the cutting guides (Figures 1, 2). 

This provides rotation stability, and the implant rests on the strong subchondral bone of the proximal humerus proximal to the anatomic neck rather than relying on metaphyseal bone or canal fixation, as recommended by Aldoiusti.¹³ It also allows optimal implant placement with complete freedom with respect to inclination, version, and medial/posterior offset from the humeral canal. 

Continue to: The implant respects the relationship...

The implant respects the relationship of the rotator cuff insertion and has a recessed superior margin to keep both the implant and the saw blade 3 mm to 5 mm away from the supraspinatus fibers to protect the rotator cuff from iatrogenic injury.

TECHNIQUE

The technique described in this article uses the Catalyst CSR Total Shoulder System (Catalyst OrthoScience), which was cleared to treat arthritis of the shoulder by the US Food and Drug Administration in May 2016.

A standard deltopectoral incision is made, and the surgeon dissects the interval between the pectoralis major medially and the deltoid laterally. The subscapularis can be incised by tenotomy; alternatively, the surgeon can perform a subscapularis peel or a lesser tuberosity osteotomy using this technique.

Once the glenohumeral joint is exposed, the surgeon delivers the humeral head anteriorly. A preferred method is to place a Darrach retractor between the humeral head and the glenoid, and a cobra or a second Darrach retractor behind the superolateral humeral head superficial to the supraspinatus tendon. By simultaneously pressing on both retractors and externally rotating the patient’s arm, the humeral head is delivered anteriorly. Osteophytes on the anterior and inferior edge of the humeral head are generously removed at this time using a rongeur.

Using a pin guide, the long 3.2-mm guidewire pin is drilled under power into the center of the articular surface. The pin guide is then removed, leaving the pin in the center of the humerus (Figure 3).

Continue to: Next, the surgeon...

Next, the surgeon slides the cannulated reamer over the long guidewire pin and under power removes a small portion of the humeral head subchondral bone until the surgeon feels and observes that the reamer is no longer removing bone (Figure 4). The patent-pending reamer design prevents the surgeon from removing more than a few millimeters of bone, after which point the reamer spins on the surface of the bone without resecting further.

The surgeon is aware that the reamer has achieved its desired depth when it is no longer creating new bone shavings, and the surgeon can hear and feel that the reamer is spinning and no longer cutting. Then the surgeon removes the reamer.

The surgeon places the first humeral cut guide over the long guidewire pin, oriented superiorly-inferiorly and secures the guide using 4 short pins, and the long pin is removed. The surgeon uses an oscillating saw to cut the anterior and posterior plane cuts through the saw captures in the cut guide (Figure 5). The humeral cut guide and short pins are removed (Figure 6).

The surgeon then applies the second humeral cut guide to the proximal humerus and secures it using 2 short pins. The surgeon then uses the 6-mm drill to drill the 4 holes for the pegs of the implant. The top portion of the guide is removed, and the surgeon makes the superior and inferior cuts along the top and bottom surfaces of the guide using an oscillating saw (Figure 7).

The surgeon then uses a rongeur to slightly round the edges of the 4 corners at the periphery of the humerus. The second humeral cut guide and short pins are removed (Figure 8).

Continue to: Next, the surgeon trials...

Next, the surgeon trials humeral implants to determine the correct implant size (Figure 9). Once the proper humeral size is chosen, the trial is removed and the humeral cover is placed over the prepared humeral head. The surgeon then proceeds to glenoid preparation (Figure 10), which is easily accessible and facilitated by angled planar cuts on the humeral head. Glenoid technique will be discussed in a subsequent article.

After glenoid preparation and insertion, the humerus is delivered anteriorly. The proximal humerus is washed and dried, and cement is applied to the peg holes in the humerus bone and the underside of the humeral implant. The implant is then inserted using the humeral impactor to apply pressure and assure that the implant is fully seated. Once the humeral cement is hardened, the glenohumeral joint is irrigated and closure begins. Postoperative radiograph is shown in Figure 11.

DISCUSSION

Numerous authors have demonstrated that accurate implant placement is crucial for restoring normal glenoid kinematics and motion,^1-4 while some authors have reported worsening clinical outcomes and higher rates of pain and implant loosening when the implants were not placed anatomically.^5-8 This is such an important concept that it essentially was the primary inspiration for creating this TSA system. In addition, the system utilizes a nonspherical, elliptical humeral head that more closely matches the anatomy of the proximal humerus,^14,15 and this type of shape has shown improved biomechanics in laboratory testing.¹²

Good results have been demonstrated in restoring the normal anatomy using stemmed devices on the radiographic analysis of cadavers.¹⁶ The creation of stemmed implants with variable inclination and offset has improved computer models¹⁷ compared with previous studies,¹⁸ with the exception of scenarios with extreme offset. 

In theory, resurfacing implants and implants without a canal stem should have a better implant placement than that with stemmed implants; however, the ability to restore the center of rotation was even worse for resurfacing prostheses, with 65% of all implants being measured as outliers postoperatively in one study.¹⁹ Most of the resurfacing implants and their instrumentation techniques offer little to help the surgeon control for implant height. The depth of the reaming is variable, not calibrated, and not correlated with the implant size, frequently leading to overstuffing after surgery. Second, the use of spherical prostheses forces the surgeon to choose between matching the superior-inferior humeral size, leading to overhang of the implant, or matching the anteroposterior, leading to frequent undersizing in the coronal plane. The nonspherical, elliptical head shape can potentially simplify implant selection.

In summary, new techniques have been developed in an attempt to achieve increased consistency and precision in TSA. By more accurately reproducing the proximal humeral anatomy, it is proposed that clinical outcomes in terms of the range of motion and patient satisfaction may also be improved through newer techniques. Cadaver studies have validated the anatomic precision of this technique.¹⁰ Clinical data comprising of patient-reported outcome measures and radiographic outcome studies are currently underway for this arthroplasty system.

ABSTRACT

Proper reconstruction of proximal humeral anatomy is of primary importance to maximize patient outcomes after total shoulder arthroplasty. This article describes a new arthroplasty technique, where a fixed multiplanar bone resection is made and a novel implant, which is designed to precisely match the bone resection, is inserted. 

Continue to: The success of total shoulder arthroplasty...

The success of total shoulder arthroplasty (TSA) is largely dependent on how accurate the proximal humeral anatomy is reconstructed and the glenohumeral relationships are restored.^1-4 Numerous studies have demonstrated a relationship of worse clinical outcomes and implant failure with nonanatomic implant placement.^5-8 The majority of arthroplasty systems rely on surgeon-dependent decision-making to determine the location of the border of the articular surface and, ultimately, the amount and location of bone to be resected. Even in experienced hands, the ability to reproducibly restore the joint line is inconsistent.³

In contrast, the majority of total knee arthroplasty (TKA) systems have been designed with instrumentation that guides the surgeon precisely regarding where and how much femoral bone must be resected, and the corresponding implant is designed with the same thickness to preserve the location of the joint line. Cutting block instrumentation rather than freehand cuts enables reproducibility of TKA while being performed for an estimated 700,000 times annually in the US.⁹

To achieve similar high levels of reproducibility in shoulder arthroplasty, a new technique was developed based on the principle of providing instrumentation to assist the surgeon in accurately restoring the proximal humeral joint line. This technical article describes the technique of using a multiplanar instrumented cutting system and matching implants to perform TSA. The technique shown was previously studied and was found to allow surgeons to recreate the original anatomy of the humerus with very high precision.¹⁰

The humeral prosthesis described in this article has an articular surface that is slightly elliptical to more closely match the actual shape of the humerus bone.¹¹ Biomechanical studies have demonstrated that implants designed with a nonspherical shape have more similar motion and kinematics to those of the native humeral head.¹² The undersurface of the implant has a concave four-plane geometry that matches with the bone cuts created by the cutting guides (Figures 1, 2). 

This provides rotation stability, and the implant rests on the strong subchondral bone of the proximal humerus proximal to the anatomic neck rather than relying on metaphyseal bone or canal fixation, as recommended by Aldoiusti.¹³ It also allows optimal implant placement with complete freedom with respect to inclination, version, and medial/posterior offset from the humeral canal. 

Continue to: The implant respects the relationship...

The implant respects the relationship of the rotator cuff insertion and has a recessed superior margin to keep both the implant and the saw blade 3 mm to 5 mm away from the supraspinatus fibers to protect the rotator cuff from iatrogenic injury.

TECHNIQUE

The technique described in this article uses the Catalyst CSR Total Shoulder System (Catalyst OrthoScience), which was cleared to treat arthritis of the shoulder by the US Food and Drug Administration in May 2016.

A standard deltopectoral incision is made, and the surgeon dissects the interval between the pectoralis major medially and the deltoid laterally. The subscapularis can be incised by tenotomy; alternatively, the surgeon can perform a subscapularis peel or a lesser tuberosity osteotomy using this technique.

Once the glenohumeral joint is exposed, the surgeon delivers the humeral head anteriorly. A preferred method is to place a Darrach retractor between the humeral head and the glenoid, and a cobra or a second Darrach retractor behind the superolateral humeral head superficial to the supraspinatus tendon. By simultaneously pressing on both retractors and externally rotating the patient’s arm, the humeral head is delivered anteriorly. Osteophytes on the anterior and inferior edge of the humeral head are generously removed at this time using a rongeur.

Using a pin guide, the long 3.2-mm guidewire pin is drilled under power into the center of the articular surface. The pin guide is then removed, leaving the pin in the center of the humerus (Figure 3).

Continue to: Next, the surgeon...

Next, the surgeon slides the cannulated reamer over the long guidewire pin and under power removes a small portion of the humeral head subchondral bone until the surgeon feels and observes that the reamer is no longer removing bone (Figure 4). The patent-pending reamer design prevents the surgeon from removing more than a few millimeters of bone, after which point the reamer spins on the surface of the bone without resecting further.

The surgeon is aware that the reamer has achieved its desired depth when it is no longer creating new bone shavings, and the surgeon can hear and feel that the reamer is spinning and no longer cutting. Then the surgeon removes the reamer.

The surgeon places the first humeral cut guide over the long guidewire pin, oriented superiorly-inferiorly and secures the guide using 4 short pins, and the long pin is removed. The surgeon uses an oscillating saw to cut the anterior and posterior plane cuts through the saw captures in the cut guide (Figure 5). The humeral cut guide and short pins are removed (Figure 6).

The surgeon then applies the second humeral cut guide to the proximal humerus and secures it using 2 short pins. The surgeon then uses the 6-mm drill to drill the 4 holes for the pegs of the implant. The top portion of the guide is removed, and the surgeon makes the superior and inferior cuts along the top and bottom surfaces of the guide using an oscillating saw (Figure 7).

The surgeon then uses a rongeur to slightly round the edges of the 4 corners at the periphery of the humerus. The second humeral cut guide and short pins are removed (Figure 8).

Continue to: Next, the surgeon trials...

Next, the surgeon trials humeral implants to determine the correct implant size (Figure 9). Once the proper humeral size is chosen, the trial is removed and the humeral cover is placed over the prepared humeral head. The surgeon then proceeds to glenoid preparation (Figure 10), which is easily accessible and facilitated by angled planar cuts on the humeral head. Glenoid technique will be discussed in a subsequent article.

After glenoid preparation and insertion, the humerus is delivered anteriorly. The proximal humerus is washed and dried, and cement is applied to the peg holes in the humerus bone and the underside of the humeral implant. The implant is then inserted using the humeral impactor to apply pressure and assure that the implant is fully seated. Once the humeral cement is hardened, the glenohumeral joint is irrigated and closure begins. Postoperative radiograph is shown in Figure 11.

DISCUSSION

Numerous authors have demonstrated that accurate implant placement is crucial for restoring normal glenoid kinematics and motion,^1-4 while some authors have reported worsening clinical outcomes and higher rates of pain and implant loosening when the implants were not placed anatomically.^5-8 This is such an important concept that it essentially was the primary inspiration for creating this TSA system. In addition, the system utilizes a nonspherical, elliptical humeral head that more closely matches the anatomy of the proximal humerus,^14,15 and this type of shape has shown improved biomechanics in laboratory testing.¹²

Good results have been demonstrated in restoring the normal anatomy using stemmed devices on the radiographic analysis of cadavers.¹⁶ The creation of stemmed implants with variable inclination and offset has improved computer models¹⁷ compared with previous studies,¹⁸ with the exception of scenarios with extreme offset. 

In theory, resurfacing implants and implants without a canal stem should have a better implant placement than that with stemmed implants; however, the ability to restore the center of rotation was even worse for resurfacing prostheses, with 65% of all implants being measured as outliers postoperatively in one study.¹⁹ Most of the resurfacing implants and their instrumentation techniques offer little to help the surgeon control for implant height. The depth of the reaming is variable, not calibrated, and not correlated with the implant size, frequently leading to overstuffing after surgery. Second, the use of spherical prostheses forces the surgeon to choose between matching the superior-inferior humeral size, leading to overhang of the implant, or matching the anteroposterior, leading to frequent undersizing in the coronal plane. The nonspherical, elliptical head shape can potentially simplify implant selection.

In summary, new techniques have been developed in an attempt to achieve increased consistency and precision in TSA. By more accurately reproducing the proximal humeral anatomy, it is proposed that clinical outcomes in terms of the range of motion and patient satisfaction may also be improved through newer techniques. Cadaver studies have validated the anatomic precision of this technique.¹⁰ Clinical data comprising of patient-reported outcome measures and radiographic outcome studies are currently underway for this arthroplasty system.

ABSTRACT

In a previous study, compared with throwing athletes with superior labral anterior posterior (SLAP) tears, those with concomitant SLAP tears and rotator cuff tears (RCTs) had significantly poorer outcome scores and return to play. Posterior shoulder instability also occurs in throwing athletes, but no studies currently exist regarding outcomes of these patients with concomitant RCTs.

The authors hypothesized that throwing athletes treated with arthroscopic capsulolabral repair for posterior shoulder instability with coexistent rotator cuff pathology would have poorer outcome scores and return to play.

Fifty-six consecutive throwing athletes with unidirectional posterior shoulder instability underwent arthroscopic capsulolabral repair. Preoperative and postoperative patient-centered outcomes of pain, stability, function, range of motion, strength, and American Shoulder and Elbow Surgeons Shoulder (ASES) scores, as well as return to play, were evaluated. Patients with and without rotator cuff pathology were compared.

Forty-three percent (24/56) of throwing athletes had rotator cuff pathology in addition to posterior capsulolabral pathology. All RCTs were débrided. At a mean of 3 years, there were no differences in preoperative and postoperative patient-centered outcomes between those with and without RCTs. Return-to-play rates showed no between-group differences; 92% (22/24) of athletes with concomitant RCTs returned to sport (P = .414) and 67% (16/24) returned to the same level (P = .430).

Arthroscopic capsulolabral reconstruction is successful in throwing athletes with RCTs treated with arthroscopic débridement. Unlike the previous study evaluating throwers outcomes after surgical treatment for concomitant SLAP tears and RCTs, the authors found no difference in patient-reported outcome measures or return to play for throwing athletes with concomitant posterior shoulder instability and RCTs. In throwing athletes with concomitant posterior instability and RCTs, arthroscopic posterior capsulolabral repair with rotator cuff débridement is successful.

Continue to: Posterior shoulder instability...

Posterior shoulder instability is an important and increasingly recognized pathology among throwers. Like the superior labrum, the posterior capsulolabral complex is also susceptible to injury during the throwing motion; the posterior labrum being most at risk during the late cocking and follow-through phases. Recent studies have found that arthroscopic capsulolabral reconstruction in posterior shoulder instability is successful in allowing athletes to return to their preinjury sports activities, with 2 studies detailing outcomes in throwing athletes.^1-4 However, superior labral anterior posterior (SLAP) tears are common in throwing athletes and have been treated with varying and limited success. Further, in a study of outcomes of arthroscopic repair of SLAP lesions, Neri and colleagues⁵ found that, compared with throwing athletes with SLAP tears, throwing athletes with concomitant SLAP tears and partial-thickness rotator cuff tears (RCTs) had significantly poorer outcomes and return-to-play rates after surgical repair.

The purpose of this study was to determine outcome scores and return to play of throwing athletes treated with arthroscopic capsulolabral repair for posterior shoulder instability with coexistent RCTs and to compare them with outcome scores as well as return to play of throwing athletes with isolated posterior shoulder instability. It was hypothesized that throwing athletes with a combination of posterior shoulder instability and RCT would have poorer outcomes and poorer return to play after surgery.⁵

METHODS

PATIENT SELECTION

After Institutional Review Board approval, informed consent was obtained, and consecutive throwing athletes who underwent arthroscopic posterior capsulolabral reconstruction for posterior shoulder instability were followed in the perioperative period. Inclusion criteria were throwing athletes participating in competitive sports at the high school, collegiate, or professional level, minimum 1-year follow-up, presence of unidirectional posterior instability, and absence of symptoms of instability in any direction other than posterior. Patients with inferior instability, SLAP pathology on examination and on magnetic resonance imaging, multidirectional instability, or habitual or psychogenic voluntary shoulder subluxations were excluded. Patients with diagnoses of both posterior shoulder instability and impingement treated with subacromial decompression and distal clavicle resection were also excluded.

After this cohort was identified, patient records were reviewed for pertinent operative data, such as procedure, complications, and evidence of RCT by operative report and arthroscopic photographs. A partial RCT was defined as a tear of 10% to 50%; those with rotator cuff fraying were determined not to be significant.

PATIENT EVALUATION

Surgeries were performed between January 1998 and December 2009 by the senior author (JPB). All patients were followed with clinical examinations, radiographs, and subjective grading scales. Recorded patient demographic data included age, sex, sport, position, competition level, and follow-up duration.

Continue to: All patients had...

All patients had symptomatic posterior shoulder instability, including posterior shoulder pain, clicking, a sensation of subluxation, or instability/apprehension with motion. Each athlete’s shoulder was palpated for tenderness and tested for impingement. Specific posterior glenohumeral instability tests, including the Kim test,⁶ the circumduction test, the jerk test,⁷ the posterior load-and-shift test,⁸ and the posterior stress test,⁹ were performed on all patients. Patients with multidirectional instability on the sulcus test, as well as provocative tests indicating SLAP pathology, such as the Crank test and the active compression test, were not included. Standard radiography and magnetic resonance arthrography (MRA) were performed to further narrow inclusion and exclusion criteria.

Both before surgery and at latest follow-up, patient outcomes were evaluated using the American Shoulder and Elbow Surgeons (ASES) score (range, 0-100) which combines a subjective functional scale measuring activities of daily living (0-3 for each of 10 tasks, with a total of 0-30) and a subjective pain scale (0-10, with 10 being worst pain). Values >80 were described as excellent, and failures were defined as scores <60 after surgery.¹⁰ A subjective stability scale (0-10, with 0 indicating completely stable and 10 completely unstable), strength scale (0-3, with 0 indicating none, 1 markedly decreased, 2 slightly decreased, and 3 normal), and ROM scale (0-3, with 0 indicating poor, 1 limited, 2 satisfactory, and 3 full) were evaluated both before surgery and at the latest follow-up. A stability score >5 after surgery was defined as a failure.^1,2,11 Patients were also asked if, based on their current state, they would undergo surgery again. Intraoperative findings and specific surgical procedures performed were correlated with the aforementioned subjective and objective outcome scores.

OPERATIVE TREATMENT

Throwing athletes who met inclusion criteria and failed nonoperative management underwent surgery by the senior author (JPB). Each patient was examined under anesthesia and, with the patient in the lateral decubitus position, a diagnostic arthroscopy was performed to identify posterior capsulolabral complex pathology, including a patulous capsule, capsular tears, labral fraying, and labral tears. A careful examination for rotator cuff pathology was also performed. Based on preoperative clinical examination, MRA, examination under anesthesia, pathologic findings at diagnostic arthroscopic surgery, and surgeon experience, capsulolabral plication was performed with or without suture anchors.^2,5 After capsulolabral repair, the capsule was evaluated for residual laxity, and additional plication sutures were placed, as indicated, with care to avoid overconstraint in these throwing athletes.¹ Posterior glenohumeral stability restoration was judged by removing traction and performing posterior load-and-shift and posterior stress tests. Any RCT with <50% thickness was débrided. Postoperative care and rehabilitation were carried out as previously described and were not altered by the presence or absence of a RCT.³

STATISTICAL ANALYSIS

Preoperative and latest follow-up ASES scores, stability scores, functional scores, and pain-level findings were compared using paired-samples Comparisons between groups, including throwing athletes with and without rotator cuff pathology, were done using the Student t test. Outcome comparisons between multiple groups, which included intraoperative findings and surgical fixation methods, were analyzed with c² modeling for nonparametric data. Statistical significance was set at P < .05. A power analysis found that this study was able to detect a meaningful difference of 10 ASES points.

RESULTS

PATIENT DEMOGRAPHIC CHARACTERISTICS

Of the 56 throwing athletes who met the inclusion criteria, 24 were found to have rotator cuff pathology in addition to posterior capsulolabral pathology, while 32 were found to have capsulolabral pathology alone. Demographic data are listed in Table 1. Mean age was 20.1 years for patients with rotator cuff pathology and 17.8 years for patients without RCTs. All 24 athletes with rotator cuff pathology were treated with arthroscopic débridement. Mean follow-up was 38.6 months (range, 16.5-63.6 months) for patients with RCTs and 39.1 months (range, 12-98.8 months) for patients without RCTs. No significant difference was found in age, sports level, or follow-up between groups.

Table 1. Demographic Data for Athletes With Posterior Instability With and Without Rotator Cuff Tears (N = 56 Shoulders)^a

^aThe majority of athletes were males in high school and their mean follow-up was 3 years.

Continue to: Outcomes

Table 2 lists the preoperative and postoperative scores for shoulder performance in throwing athletes with posterior shoulder instability, with and without RCTs.

Table 2. Preoperative and Postoperative Scores for Shoulder Performance in Throwing Athletes With Posterior Shoulder Instability With and Without Rotator Cuff Tears^a

^aThere was no difference in ASES, stability, pain, or functional scores between athletes with posterior instability alone compared with patients with concomitant rotator cuff tears.

Abbreviation: ASES, American Shoulder and Elbow Surgeons.

ASES Scores. Mean preoperative ASES scores for patients with RCTs improved significantly (t = –13.8, P < .001), as did those for patients without rotator cuff pathology (t = –8.9, P < .001). No significant differences in ASES score were found between patients with and without rotator cuff pathology before or after surgery (t = 1.9, P = .07; t = .58, P = .06). In addition, 70.8% (17/24) of throwing athletes with rotator cuff pathology had an excellent postoperative outcome (ASES score >80), and 29.2% (7/24) had a satisfactory outcome (ASES score, 60-80). Thus, 100% of those with concomitant posterior shoulder instability and RCTs had a good or excellent outcome after surgical intervention. In those without rotator cuff pathology, 78.1% (25/32) had an excellent outcome, 12.5% (4/32) had a satisfactory outcome, and 9.4% (3/32) had a poor outcome. Thus, 91% of those without rotator cuff pathology had a good or excellent outcome after surgery.

Stability. Preoperative stability scores improved significantly after surgery in both groups (t = 7.2, P < .001; t = 10.5, P < .001). There were no statistical differences between preoperative or postoperative stability scores in those with or without rotator cuff pathology (t = 1.7, P = .095; t = .03, P = .975). Of throwing athletes with RCTs, 54.2% (13/24) had an excellent outcome, 33.3% (8/24) a good outcome, and 12.5% (3/24) a satisfactory outcome. Thus, 87.5% (21/24) of those with RCTs had a good or excellent outcome in terms of stability. In those without rotator cuff pathology, 46.9% (15/32) had excellent stability, 46.9% (15/32) had good stability, and 3.1% (1/32) had satisfactory stability after surgery. Thus, 93.8% (30/32) of throwing athletes without rotator cuff pathology had good or excellent stability after surgery.

Pain. Mean preoperative pain scores for those with and without rotator cuff pathology improved significantly (t = 13.4, P < .001; t = 7.1, P < .001). There was no statistical difference in preoperative or postoperative pain scores between those with and without rotator cuff pathology (t = 1.99, P = .051; t = .49, P = .627).

Function. Mean preoperative function scores for both groups improved significantly (t = 7.7, P < .001; t = 8.0, P < .001). There was no difference in improvement in functional scores between the two groups before or after surgery (t = .36, P = .721; t = .5, P = .622).

Continue to: ROM

ROM. Of those with rotator cuff pathology, 54% (13/24) had normal ROM, 42% (10/24) had satisfactory ROM, and 4% (1/24) had limited ROM. In throwing athletes without rotator cuff pathology, 34% (11/32) had normal ROM, 53.1% (17/32) had satisfactory ROM, and 9% (3/32) had limited ROM after surgery. There was no significant difference in ROM between the groups (c² = 2.7, P = .260).

Strength. Of those with RCTs, 67% (16/24) reported normal strength, 29% (7/24) slightly decreased strength, and 4% (1/24) markedly decreased strength. Of those throwing athletes without rotator cuff pathology, 50% (16/32) had normal strength, 41% (13/32) had slightly decreased strength, and 9% (3/32) had markedly decreased strength. No statistical difference was noted between the two groups (c² = 1.7, P = .429).

Return to Sport. Of those with RCTs, 92% (22/24) returned to sport while 84% (27/32) of throwing athletes without RCTs returned to sport. There was no difference between the two groups (c² = .667, P = .414). Sixty-seven percent (16/24) of those with RCTs and 56% (18/32) of those without RCTs returned to the same level of sport. No statistical difference was found in return to play between throwing athletes with and without rotator cuff pathology (c² = .624, P = .430).

Failures. According to ASES scores, no throwers with RCTs failed, while 9.4% (3/32) with posterior instability alone failed. Regarding stability, 8.3% (2/24) of athletes with RCTs failed, while 6.3% (2/32) with posterior instability alone failed. 

SURGICAL FINDINGS AND PROCEDURES

Of the 24 throwing athletes with rotator cuff pathology, 92% (22/24) had labral tears, while 78% (25/32) of those without RCTs had labral tears. The majority of RCTs were in the posterior supraspinatus and anterior infraspinatus regions. This was not significantly different between groups (c² = 1.86, P = .172). All labral pathology was posterior-inferior, and all RCTs were <50% thickness, and therefore were débrided. Fifty-four percent (13/24) of those with RCTs had a patulous capsule and 63% (20/32) of throwing athletes without rotator cuff pathology had a patulous capsule. There was no significant difference between groups (c² = .393, P = .530). Of those with RCTs, 92% (22/24) had surgical fixation with anchors, while 78% (25/32) of those without rotator cuff pathology underwent repair with anchor fixation. There was no statistically significant difference in anchor use between groups (c² = 1.86, P = .172).

Continue to: Discussion

Throwing athletes with and without RCTs had similar rates of recovery and return to play after arthroscopic capsular labral repair, with rotator cuff débridement if a tear was present. The mean follow-up was 3.2 years. Further, there was no difference in return to play (92% vs 84%), ASES score, stability, pain, function, ROM, or strength between the 2 groups before or after surgery. In this cohort of 56 patients, 24 throwing athletes (43%) were found to have RCTs.

Return-to-play rates showed no between-group differences; 92% (22/24) of athletes with concomitant RCTs returned to sport, and 67% (16/24) returned to the same level. Eight percent of throwing athletes with RCTs were unable to return to sport after surgery. These return-to-play rates are an improvement over most previously reported rates in throwing athletes and in posterior shoulder instability in general.^1-4,11 When these athletes are compared with their counterparts with combined SLAP tears and RCTs, return-to-play rates are notably higher. There may be discrepancies in interpreting return-to-play between the two studies, but in the current study, 67% of those with concomitant RCTs achieved return to preinjury level of play. This is 10% higher than the rate reported in athletes with SLAP tears alone (57%) and even higher than in those with concomitant SLAP and RCTs. It is also essential to note that a number of this cohort’s athletes who did not return to play did so for factors (eg, graduation) unrelated to the shoulder. However, the study by Neri and colleagues⁵ included professional athletes who likely all attempted to return to play and, if unable to perform at the same level, likely were unable to continue their professional career.⁵

All patients with RCTs had a good or excellent outcome (ASES score), and 70.8% had an excellent outcome. Similarly, 97% of those without rotator cuff pathology had a good or excellent outcome, and 81.3% had an excellent outcome. There was no significant difference between the two groups. These results parallel those of Neri and colleagues’⁵ study of SLAP tears with RCTs, where 96% (22/23) of throwing athletes had a good or excellent outcome. Despite these high outcome scores in patients with SLAP tears, only 57% were able to return to elite pitching.⁵ In the current study, pain was slightly higher for those with rotator cuff pathology before surgery—a finding consistent with pain frequently being found in patients with isolated partial-thickness RCTs. Their postoperative pain scores were actually lower on average than those of patients without RCTs, which suggests simple débridement of undersurface tears adequately addressed the pathology. The authors theorize that the main pain generator in this population may be posterior instability, and that the rotator cuff has less of an influence. In the SLAP population, the main pain generator likely is the RCT.

Failures by ASES score or strength were fairly rare in this cohort. Many patients opted to have revision surgery because of continued instability, pain, decreased function, or reinjury. One potential cause of failure in this cohort is inadequate capsular shift. However, capsular plication in throwing athletes is difficult to address, as overtensioning the repair can lead to the inability to adequately perform overhead activites.^3,4 This cannot be overemphasized, particularly with pitchers.

Partial-thickness RCTs, particularly those on the articular side, are common in throwing athletes because of high tensile and compressive loads.¹² Despite the known risk of RCTs with posterior shoulder instability in throwing athletes, the authors are unaware of reports of the incidence or treatment of this pathology. RCTs in this posterior instability group likely represent a pathology other than internal impingement. The high proportion of throwing athletes with RCTs in this study (43%) indicates a need for close evaluation of rotator cuff pathology in young throwing athletes. Ide et al found that 75% of patients with SLAP tears had partial articular-sided RCTs.¹³ In the current study, all RCTs were small partial tears, and arthroscopic débridement was performed. It is unknown whether repair of these RCTs would impact return to play. However, rotator cuff repair in this population has been shown to have poor outcomes. Tear thickness typically is used to determine treatment, with débridement performed if <50% tendon thickness is affected. More recently, many have advocated having greater tendon involvement in throwers before repair, because of poor outcomes. Although studies are limited, tear size does seem to correlate with outcomes.¹⁴

Continue to: Study Limitations

Limitations of this study include its small number of professional throwing athletes, with the majority being high school athletes. Further, although ASES scores are consistently used in posterior shoulder instability studies, these scores are influenced highly by pain scores, and some argue that other scoring systems may provide more useful information. However, none of the more modern scoring systems have been studied extensively in posterior glenohumeral instability. Further, because the authors used the present scoring systems previously,^1-4 they were continued to be used for comparison and consistency. Outcomes such as ROM and strength may carry more weight if measured and documented by clinical examination. Further testing, such as clinical evaluation of the jerk test or the posterior load-and-shift test, and their comparison before and after surgery may provide more objective data.

CONCLUSION

Arthroscopic capsulolabral reconstruction is successful in throwing athletes with RCTs treated with arthroscopic débridement. Unlike a previous study of throwing athletes’ outcomes after surgery for concomitant SLAP tears and RCTs,⁵ this study of throwing athletes with concomitant posterior shoulder instability and RCTs found no difference in patient-reported outcome measures or return to play. In throwing athletes with posterior instability and RCTs, arthroscopic posterior capsulolabral repair with rotator cuff débridement is successful.

ABSTRACT

In a previous study, compared with throwing athletes with superior labral anterior posterior (SLAP) tears, those with concomitant SLAP tears and rotator cuff tears (RCTs) had significantly poorer outcome scores and return to play. Posterior shoulder instability also occurs in throwing athletes, but no studies currently exist regarding outcomes of these patients with concomitant RCTs.

The authors hypothesized that throwing athletes treated with arthroscopic capsulolabral repair for posterior shoulder instability with coexistent rotator cuff pathology would have poorer outcome scores and return to play.

Fifty-six consecutive throwing athletes with unidirectional posterior shoulder instability underwent arthroscopic capsulolabral repair. Preoperative and postoperative patient-centered outcomes of pain, stability, function, range of motion, strength, and American Shoulder and Elbow Surgeons Shoulder (ASES) scores, as well as return to play, were evaluated. Patients with and without rotator cuff pathology were compared.

Forty-three percent (24/56) of throwing athletes had rotator cuff pathology in addition to posterior capsulolabral pathology. All RCTs were débrided. At a mean of 3 years, there were no differences in preoperative and postoperative patient-centered outcomes between those with and without RCTs. Return-to-play rates showed no between-group differences; 92% (22/24) of athletes with concomitant RCTs returned to sport (P = .414) and 67% (16/24) returned to the same level (P = .430).

Arthroscopic capsulolabral reconstruction is successful in throwing athletes with RCTs treated with arthroscopic débridement. Unlike the previous study evaluating throwers outcomes after surgical treatment for concomitant SLAP tears and RCTs, the authors found no difference in patient-reported outcome measures or return to play for throwing athletes with concomitant posterior shoulder instability and RCTs. In throwing athletes with concomitant posterior instability and RCTs, arthroscopic posterior capsulolabral repair with rotator cuff débridement is successful.

Continue to: Posterior shoulder instability...

Posterior shoulder instability is an important and increasingly recognized pathology among throwers. Like the superior labrum, the posterior capsulolabral complex is also susceptible to injury during the throwing motion; the posterior labrum being most at risk during the late cocking and follow-through phases. Recent studies have found that arthroscopic capsulolabral reconstruction in posterior shoulder instability is successful in allowing athletes to return to their preinjury sports activities, with 2 studies detailing outcomes in throwing athletes.^1-4 However, superior labral anterior posterior (SLAP) tears are common in throwing athletes and have been treated with varying and limited success. Further, in a study of outcomes of arthroscopic repair of SLAP lesions, Neri and colleagues⁵ found that, compared with throwing athletes with SLAP tears, throwing athletes with concomitant SLAP tears and partial-thickness rotator cuff tears (RCTs) had significantly poorer outcomes and return-to-play rates after surgical repair.

The purpose of this study was to determine outcome scores and return to play of throwing athletes treated with arthroscopic capsulolabral repair for posterior shoulder instability with coexistent RCTs and to compare them with outcome scores as well as return to play of throwing athletes with isolated posterior shoulder instability. It was hypothesized that throwing athletes with a combination of posterior shoulder instability and RCT would have poorer outcomes and poorer return to play after surgery.⁵

METHODS

PATIENT SELECTION

After Institutional Review Board approval, informed consent was obtained, and consecutive throwing athletes who underwent arthroscopic posterior capsulolabral reconstruction for posterior shoulder instability were followed in the perioperative period. Inclusion criteria were throwing athletes participating in competitive sports at the high school, collegiate, or professional level, minimum 1-year follow-up, presence of unidirectional posterior instability, and absence of symptoms of instability in any direction other than posterior. Patients with inferior instability, SLAP pathology on examination and on magnetic resonance imaging, multidirectional instability, or habitual or psychogenic voluntary shoulder subluxations were excluded. Patients with diagnoses of both posterior shoulder instability and impingement treated with subacromial decompression and distal clavicle resection were also excluded.

After this cohort was identified, patient records were reviewed for pertinent operative data, such as procedure, complications, and evidence of RCT by operative report and arthroscopic photographs. A partial RCT was defined as a tear of 10% to 50%; those with rotator cuff fraying were determined not to be significant.

PATIENT EVALUATION

Surgeries were performed between January 1998 and December 2009 by the senior author (JPB). All patients were followed with clinical examinations, radiographs, and subjective grading scales. Recorded patient demographic data included age, sex, sport, position, competition level, and follow-up duration.

Continue to: All patients had...

All patients had symptomatic posterior shoulder instability, including posterior shoulder pain, clicking, a sensation of subluxation, or instability/apprehension with motion. Each athlete’s shoulder was palpated for tenderness and tested for impingement. Specific posterior glenohumeral instability tests, including the Kim test,⁶ the circumduction test, the jerk test,⁷ the posterior load-and-shift test,⁸ and the posterior stress test,⁹ were performed on all patients. Patients with multidirectional instability on the sulcus test, as well as provocative tests indicating SLAP pathology, such as the Crank test and the active compression test, were not included. Standard radiography and magnetic resonance arthrography (MRA) were performed to further narrow inclusion and exclusion criteria.

Both before surgery and at latest follow-up, patient outcomes were evaluated using the American Shoulder and Elbow Surgeons (ASES) score (range, 0-100) which combines a subjective functional scale measuring activities of daily living (0-3 for each of 10 tasks, with a total of 0-30) and a subjective pain scale (0-10, with 10 being worst pain). Values >80 were described as excellent, and failures were defined as scores <60 after surgery.¹⁰ A subjective stability scale (0-10, with 0 indicating completely stable and 10 completely unstable), strength scale (0-3, with 0 indicating none, 1 markedly decreased, 2 slightly decreased, and 3 normal), and ROM scale (0-3, with 0 indicating poor, 1 limited, 2 satisfactory, and 3 full) were evaluated both before surgery and at the latest follow-up. A stability score >5 after surgery was defined as a failure.^1,2,11 Patients were also asked if, based on their current state, they would undergo surgery again. Intraoperative findings and specific surgical procedures performed were correlated with the aforementioned subjective and objective outcome scores.

OPERATIVE TREATMENT

Throwing athletes who met inclusion criteria and failed nonoperative management underwent surgery by the senior author (JPB). Each patient was examined under anesthesia and, with the patient in the lateral decubitus position, a diagnostic arthroscopy was performed to identify posterior capsulolabral complex pathology, including a patulous capsule, capsular tears, labral fraying, and labral tears. A careful examination for rotator cuff pathology was also performed. Based on preoperative clinical examination, MRA, examination under anesthesia, pathologic findings at diagnostic arthroscopic surgery, and surgeon experience, capsulolabral plication was performed with or without suture anchors.^2,5 After capsulolabral repair, the capsule was evaluated for residual laxity, and additional plication sutures were placed, as indicated, with care to avoid overconstraint in these throwing athletes.¹ Posterior glenohumeral stability restoration was judged by removing traction and performing posterior load-and-shift and posterior stress tests. Any RCT with <50% thickness was débrided. Postoperative care and rehabilitation were carried out as previously described and were not altered by the presence or absence of a RCT.³

STATISTICAL ANALYSIS

Preoperative and latest follow-up ASES scores, stability scores, functional scores, and pain-level findings were compared using paired-samples Comparisons between groups, including throwing athletes with and without rotator cuff pathology, were done using the Student t test. Outcome comparisons between multiple groups, which included intraoperative findings and surgical fixation methods, were analyzed with c² modeling for nonparametric data. Statistical significance was set at P < .05. A power analysis found that this study was able to detect a meaningful difference of 10 ASES points.

RESULTS

PATIENT DEMOGRAPHIC CHARACTERISTICS

Of the 56 throwing athletes who met the inclusion criteria, 24 were found to have rotator cuff pathology in addition to posterior capsulolabral pathology, while 32 were found to have capsulolabral pathology alone. Demographic data are listed in Table 1. Mean age was 20.1 years for patients with rotator cuff pathology and 17.8 years for patients without RCTs. All 24 athletes with rotator cuff pathology were treated with arthroscopic débridement. Mean follow-up was 38.6 months (range, 16.5-63.6 months) for patients with RCTs and 39.1 months (range, 12-98.8 months) for patients without RCTs. No significant difference was found in age, sports level, or follow-up between groups.

Table 1. Demographic Data for Athletes With Posterior Instability With and Without Rotator Cuff Tears (N = 56 Shoulders)^a

^aThe majority of athletes were males in high school and their mean follow-up was 3 years.

Continue to: Outcomes

Table 2 lists the preoperative and postoperative scores for shoulder performance in throwing athletes with posterior shoulder instability, with and without RCTs.

Table 2. Preoperative and Postoperative Scores for Shoulder Performance in Throwing Athletes With Posterior Shoulder Instability With and Without Rotator Cuff Tears^a

^aThere was no difference in ASES, stability, pain, or functional scores between athletes with posterior instability alone compared with patients with concomitant rotator cuff tears.

Abbreviation: ASES, American Shoulder and Elbow Surgeons.

ASES Scores. Mean preoperative ASES scores for patients with RCTs improved significantly (t = –13.8, P < .001), as did those for patients without rotator cuff pathology (t = –8.9, P < .001). No significant differences in ASES score were found between patients with and without rotator cuff pathology before or after surgery (t = 1.9, P = .07; t = .58, P = .06). In addition, 70.8% (17/24) of throwing athletes with rotator cuff pathology had an excellent postoperative outcome (ASES score >80), and 29.2% (7/24) had a satisfactory outcome (ASES score, 60-80). Thus, 100% of those with concomitant posterior shoulder instability and RCTs had a good or excellent outcome after surgical intervention. In those without rotator cuff pathology, 78.1% (25/32) had an excellent outcome, 12.5% (4/32) had a satisfactory outcome, and 9.4% (3/32) had a poor outcome. Thus, 91% of those without rotator cuff pathology had a good or excellent outcome after surgery.

Stability. Preoperative stability scores improved significantly after surgery in both groups (t = 7.2, P < .001; t = 10.5, P < .001). There were no statistical differences between preoperative or postoperative stability scores in those with or without rotator cuff pathology (t = 1.7, P = .095; t = .03, P = .975). Of throwing athletes with RCTs, 54.2% (13/24) had an excellent outcome, 33.3% (8/24) a good outcome, and 12.5% (3/24) a satisfactory outcome. Thus, 87.5% (21/24) of those with RCTs had a good or excellent outcome in terms of stability. In those without rotator cuff pathology, 46.9% (15/32) had excellent stability, 46.9% (15/32) had good stability, and 3.1% (1/32) had satisfactory stability after surgery. Thus, 93.8% (30/32) of throwing athletes without rotator cuff pathology had good or excellent stability after surgery.

Pain. Mean preoperative pain scores for those with and without rotator cuff pathology improved significantly (t = 13.4, P < .001; t = 7.1, P < .001). There was no statistical difference in preoperative or postoperative pain scores between those with and without rotator cuff pathology (t = 1.99, P = .051; t = .49, P = .627).

Function. Mean preoperative function scores for both groups improved significantly (t = 7.7, P < .001; t = 8.0, P < .001). There was no difference in improvement in functional scores between the two groups before or after surgery (t = .36, P = .721; t = .5, P = .622).

Continue to: ROM

ROM. Of those with rotator cuff pathology, 54% (13/24) had normal ROM, 42% (10/24) had satisfactory ROM, and 4% (1/24) had limited ROM. In throwing athletes without rotator cuff pathology, 34% (11/32) had normal ROM, 53.1% (17/32) had satisfactory ROM, and 9% (3/32) had limited ROM after surgery. There was no significant difference in ROM between the groups (c² = 2.7, P = .260).

Strength. Of those with RCTs, 67% (16/24) reported normal strength, 29% (7/24) slightly decreased strength, and 4% (1/24) markedly decreased strength. Of those throwing athletes without rotator cuff pathology, 50% (16/32) had normal strength, 41% (13/32) had slightly decreased strength, and 9% (3/32) had markedly decreased strength. No statistical difference was noted between the two groups (c² = 1.7, P = .429).

Return to Sport. Of those with RCTs, 92% (22/24) returned to sport while 84% (27/32) of throwing athletes without RCTs returned to sport. There was no difference between the two groups (c² = .667, P = .414). Sixty-seven percent (16/24) of those with RCTs and 56% (18/32) of those without RCTs returned to the same level of sport. No statistical difference was found in return to play between throwing athletes with and without rotator cuff pathology (c² = .624, P = .430).

Failures. According to ASES scores, no throwers with RCTs failed, while 9.4% (3/32) with posterior instability alone failed. Regarding stability, 8.3% (2/24) of athletes with RCTs failed, while 6.3% (2/32) with posterior instability alone failed. 

SURGICAL FINDINGS AND PROCEDURES

Of the 24 throwing athletes with rotator cuff pathology, 92% (22/24) had labral tears, while 78% (25/32) of those without RCTs had labral tears. The majority of RCTs were in the posterior supraspinatus and anterior infraspinatus regions. This was not significantly different between groups (c² = 1.86, P = .172). All labral pathology was posterior-inferior, and all RCTs were <50% thickness, and therefore were débrided. Fifty-four percent (13/24) of those with RCTs had a patulous capsule and 63% (20/32) of throwing athletes without rotator cuff pathology had a patulous capsule. There was no significant difference between groups (c² = .393, P = .530). Of those with RCTs, 92% (22/24) had surgical fixation with anchors, while 78% (25/32) of those without rotator cuff pathology underwent repair with anchor fixation. There was no statistically significant difference in anchor use between groups (c² = 1.86, P = .172).

Continue to: Discussion

Throwing athletes with and without RCTs had similar rates of recovery and return to play after arthroscopic capsular labral repair, with rotator cuff débridement if a tear was present. The mean follow-up was 3.2 years. Further, there was no difference in return to play (92% vs 84%), ASES score, stability, pain, function, ROM, or strength between the 2 groups before or after surgery. In this cohort of 56 patients, 24 throwing athletes (43%) were found to have RCTs.

Return-to-play rates showed no between-group differences; 92% (22/24) of athletes with concomitant RCTs returned to sport, and 67% (16/24) returned to the same level. Eight percent of throwing athletes with RCTs were unable to return to sport after surgery. These return-to-play rates are an improvement over most previously reported rates in throwing athletes and in posterior shoulder instability in general.^1-4,11 When these athletes are compared with their counterparts with combined SLAP tears and RCTs, return-to-play rates are notably higher. There may be discrepancies in interpreting return-to-play between the two studies, but in the current study, 67% of those with concomitant RCTs achieved return to preinjury level of play. This is 10% higher than the rate reported in athletes with SLAP tears alone (57%) and even higher than in those with concomitant SLAP and RCTs. It is also essential to note that a number of this cohort’s athletes who did not return to play did so for factors (eg, graduation) unrelated to the shoulder. However, the study by Neri and colleagues⁵ included professional athletes who likely all attempted to return to play and, if unable to perform at the same level, likely were unable to continue their professional career.⁵

All patients with RCTs had a good or excellent outcome (ASES score), and 70.8% had an excellent outcome. Similarly, 97% of those without rotator cuff pathology had a good or excellent outcome, and 81.3% had an excellent outcome. There was no significant difference between the two groups. These results parallel those of Neri and colleagues’⁵ study of SLAP tears with RCTs, where 96% (22/23) of throwing athletes had a good or excellent outcome. Despite these high outcome scores in patients with SLAP tears, only 57% were able to return to elite pitching.⁵ In the current study, pain was slightly higher for those with rotator cuff pathology before surgery—a finding consistent with pain frequently being found in patients with isolated partial-thickness RCTs. Their postoperative pain scores were actually lower on average than those of patients without RCTs, which suggests simple débridement of undersurface tears adequately addressed the pathology. The authors theorize that the main pain generator in this population may be posterior instability, and that the rotator cuff has less of an influence. In the SLAP population, the main pain generator likely is the RCT.

Failures by ASES score or strength were fairly rare in this cohort. Many patients opted to have revision surgery because of continued instability, pain, decreased function, or reinjury. One potential cause of failure in this cohort is inadequate capsular shift. However, capsular plication in throwing athletes is difficult to address, as overtensioning the repair can lead to the inability to adequately perform overhead activites.^3,4 This cannot be overemphasized, particularly with pitchers.

Partial-thickness RCTs, particularly those on the articular side, are common in throwing athletes because of high tensile and compressive loads.¹² Despite the known risk of RCTs with posterior shoulder instability in throwing athletes, the authors are unaware of reports of the incidence or treatment of this pathology. RCTs in this posterior instability group likely represent a pathology other than internal impingement. The high proportion of throwing athletes with RCTs in this study (43%) indicates a need for close evaluation of rotator cuff pathology in young throwing athletes. Ide et al found that 75% of patients with SLAP tears had partial articular-sided RCTs.¹³ In the current study, all RCTs were small partial tears, and arthroscopic débridement was performed. It is unknown whether repair of these RCTs would impact return to play. However, rotator cuff repair in this population has been shown to have poor outcomes. Tear thickness typically is used to determine treatment, with débridement performed if <50% tendon thickness is affected. More recently, many have advocated having greater tendon involvement in throwers before repair, because of poor outcomes. Although studies are limited, tear size does seem to correlate with outcomes.¹⁴

Continue to: Study Limitations

Limitations of this study include its small number of professional throwing athletes, with the majority being high school athletes. Further, although ASES scores are consistently used in posterior shoulder instability studies, these scores are influenced highly by pain scores, and some argue that other scoring systems may provide more useful information. However, none of the more modern scoring systems have been studied extensively in posterior glenohumeral instability. Further, because the authors used the present scoring systems previously,^1-4 they were continued to be used for comparison and consistency. Outcomes such as ROM and strength may carry more weight if measured and documented by clinical examination. Further testing, such as clinical evaluation of the jerk test or the posterior load-and-shift test, and their comparison before and after surgery may provide more objective data.

CONCLUSION

Arthroscopic capsulolabral reconstruction is successful in throwing athletes with RCTs treated with arthroscopic débridement. Unlike a previous study of throwing athletes’ outcomes after surgery for concomitant SLAP tears and RCTs,⁵ this study of throwing athletes with concomitant posterior shoulder instability and RCTs found no difference in patient-reported outcome measures or return to play. In throwing athletes with posterior instability and RCTs, arthroscopic posterior capsulolabral repair with rotator cuff débridement is successful.

ABSTRACT

In a previous study, compared with throwing athletes with superior labral anterior posterior (SLAP) tears, those with concomitant SLAP tears and rotator cuff tears (RCTs) had significantly poorer outcome scores and return to play. Posterior shoulder instability also occurs in throwing athletes, but no studies currently exist regarding outcomes of these patients with concomitant RCTs.

The authors hypothesized that throwing athletes treated with arthroscopic capsulolabral repair for posterior shoulder instability with coexistent rotator cuff pathology would have poorer outcome scores and return to play.

Fifty-six consecutive throwing athletes with unidirectional posterior shoulder instability underwent arthroscopic capsulolabral repair. Preoperative and postoperative patient-centered outcomes of pain, stability, function, range of motion, strength, and American Shoulder and Elbow Surgeons Shoulder (ASES) scores, as well as return to play, were evaluated. Patients with and without rotator cuff pathology were compared.

Forty-three percent (24/56) of throwing athletes had rotator cuff pathology in addition to posterior capsulolabral pathology. All RCTs were débrided. At a mean of 3 years, there were no differences in preoperative and postoperative patient-centered outcomes between those with and without RCTs. Return-to-play rates showed no between-group differences; 92% (22/24) of athletes with concomitant RCTs returned to sport (P = .414) and 67% (16/24) returned to the same level (P = .430).

Arthroscopic capsulolabral reconstruction is successful in throwing athletes with RCTs treated with arthroscopic débridement. Unlike the previous study evaluating throwers outcomes after surgical treatment for concomitant SLAP tears and RCTs, the authors found no difference in patient-reported outcome measures or return to play for throwing athletes with concomitant posterior shoulder instability and RCTs. In throwing athletes with concomitant posterior instability and RCTs, arthroscopic posterior capsulolabral repair with rotator cuff débridement is successful.

Continue to: Posterior shoulder instability...

Posterior shoulder instability is an important and increasingly recognized pathology among throwers. Like the superior labrum, the posterior capsulolabral complex is also susceptible to injury during the throwing motion; the posterior labrum being most at risk during the late cocking and follow-through phases. Recent studies have found that arthroscopic capsulolabral reconstruction in posterior shoulder instability is successful in allowing athletes to return to their preinjury sports activities, with 2 studies detailing outcomes in throwing athletes.^1-4 However, superior labral anterior posterior (SLAP) tears are common in throwing athletes and have been treated with varying and limited success. Further, in a study of outcomes of arthroscopic repair of SLAP lesions, Neri and colleagues⁵ found that, compared with throwing athletes with SLAP tears, throwing athletes with concomitant SLAP tears and partial-thickness rotator cuff tears (RCTs) had significantly poorer outcomes and return-to-play rates after surgical repair.

The purpose of this study was to determine outcome scores and return to play of throwing athletes treated with arthroscopic capsulolabral repair for posterior shoulder instability with coexistent RCTs and to compare them with outcome scores as well as return to play of throwing athletes with isolated posterior shoulder instability. It was hypothesized that throwing athletes with a combination of posterior shoulder instability and RCT would have poorer outcomes and poorer return to play after surgery.⁵

METHODS

PATIENT SELECTION

After Institutional Review Board approval, informed consent was obtained, and consecutive throwing athletes who underwent arthroscopic posterior capsulolabral reconstruction for posterior shoulder instability were followed in the perioperative period. Inclusion criteria were throwing athletes participating in competitive sports at the high school, collegiate, or professional level, minimum 1-year follow-up, presence of unidirectional posterior instability, and absence of symptoms of instability in any direction other than posterior. Patients with inferior instability, SLAP pathology on examination and on magnetic resonance imaging, multidirectional instability, or habitual or psychogenic voluntary shoulder subluxations were excluded. Patients with diagnoses of both posterior shoulder instability and impingement treated with subacromial decompression and distal clavicle resection were also excluded.

After this cohort was identified, patient records were reviewed for pertinent operative data, such as procedure, complications, and evidence of RCT by operative report and arthroscopic photographs. A partial RCT was defined as a tear of 10% to 50%; those with rotator cuff fraying were determined not to be significant.

PATIENT EVALUATION

Surgeries were performed between January 1998 and December 2009 by the senior author (JPB). All patients were followed with clinical examinations, radiographs, and subjective grading scales. Recorded patient demographic data included age, sex, sport, position, competition level, and follow-up duration.

Continue to: All patients had...

All patients had symptomatic posterior shoulder instability, including posterior shoulder pain, clicking, a sensation of subluxation, or instability/apprehension with motion. Each athlete’s shoulder was palpated for tenderness and tested for impingement. Specific posterior glenohumeral instability tests, including the Kim test,⁶ the circumduction test, the jerk test,⁷ the posterior load-and-shift test,⁸ and the posterior stress test,⁹ were performed on all patients. Patients with multidirectional instability on the sulcus test, as well as provocative tests indicating SLAP pathology, such as the Crank test and the active compression test, were not included. Standard radiography and magnetic resonance arthrography (MRA) were performed to further narrow inclusion and exclusion criteria.

Both before surgery and at latest follow-up, patient outcomes were evaluated using the American Shoulder and Elbow Surgeons (ASES) score (range, 0-100) which combines a subjective functional scale measuring activities of daily living (0-3 for each of 10 tasks, with a total of 0-30) and a subjective pain scale (0-10, with 10 being worst pain). Values >80 were described as excellent, and failures were defined as scores <60 after surgery.¹⁰ A subjective stability scale (0-10, with 0 indicating completely stable and 10 completely unstable), strength scale (0-3, with 0 indicating none, 1 markedly decreased, 2 slightly decreased, and 3 normal), and ROM scale (0-3, with 0 indicating poor, 1 limited, 2 satisfactory, and 3 full) were evaluated both before surgery and at the latest follow-up. A stability score >5 after surgery was defined as a failure.^1,2,11 Patients were also asked if, based on their current state, they would undergo surgery again. Intraoperative findings and specific surgical procedures performed were correlated with the aforementioned subjective and objective outcome scores.

OPERATIVE TREATMENT

Throwing athletes who met inclusion criteria and failed nonoperative management underwent surgery by the senior author (JPB). Each patient was examined under anesthesia and, with the patient in the lateral decubitus position, a diagnostic arthroscopy was performed to identify posterior capsulolabral complex pathology, including a patulous capsule, capsular tears, labral fraying, and labral tears. A careful examination for rotator cuff pathology was also performed. Based on preoperative clinical examination, MRA, examination under anesthesia, pathologic findings at diagnostic arthroscopic surgery, and surgeon experience, capsulolabral plication was performed with or without suture anchors.^2,5 After capsulolabral repair, the capsule was evaluated for residual laxity, and additional plication sutures were placed, as indicated, with care to avoid overconstraint in these throwing athletes.¹ Posterior glenohumeral stability restoration was judged by removing traction and performing posterior load-and-shift and posterior stress tests. Any RCT with <50% thickness was débrided. Postoperative care and rehabilitation were carried out as previously described and were not altered by the presence or absence of a RCT.³

STATISTICAL ANALYSIS

Preoperative and latest follow-up ASES scores, stability scores, functional scores, and pain-level findings were compared using paired-samples Comparisons between groups, including throwing athletes with and without rotator cuff pathology, were done using the Student t test. Outcome comparisons between multiple groups, which included intraoperative findings and surgical fixation methods, were analyzed with c² modeling for nonparametric data. Statistical significance was set at P < .05. A power analysis found that this study was able to detect a meaningful difference of 10 ASES points.

RESULTS

PATIENT DEMOGRAPHIC CHARACTERISTICS

Of the 56 throwing athletes who met the inclusion criteria, 24 were found to have rotator cuff pathology in addition to posterior capsulolabral pathology, while 32 were found to have capsulolabral pathology alone. Demographic data are listed in Table 1. Mean age was 20.1 years for patients with rotator cuff pathology and 17.8 years for patients without RCTs. All 24 athletes with rotator cuff pathology were treated with arthroscopic débridement. Mean follow-up was 38.6 months (range, 16.5-63.6 months) for patients with RCTs and 39.1 months (range, 12-98.8 months) for patients without RCTs. No significant difference was found in age, sports level, or follow-up between groups.

Table 1. Demographic Data for Athletes With Posterior Instability With and Without Rotator Cuff Tears (N = 56 Shoulders)^a

^aThe majority of athletes were males in high school and their mean follow-up was 3 years.

Continue to: Outcomes

Table 2 lists the preoperative and postoperative scores for shoulder performance in throwing athletes with posterior shoulder instability, with and without RCTs.

Table 2. Preoperative and Postoperative Scores for Shoulder Performance in Throwing Athletes With Posterior Shoulder Instability With and Without Rotator Cuff Tears^a

^aThere was no difference in ASES, stability, pain, or functional scores between athletes with posterior instability alone compared with patients with concomitant rotator cuff tears.

Abbreviation: ASES, American Shoulder and Elbow Surgeons.

ASES Scores. Mean preoperative ASES scores for patients with RCTs improved significantly (t = –13.8, P < .001), as did those for patients without rotator cuff pathology (t = –8.9, P < .001). No significant differences in ASES score were found between patients with and without rotator cuff pathology before or after surgery (t = 1.9, P = .07; t = .58, P = .06). In addition, 70.8% (17/24) of throwing athletes with rotator cuff pathology had an excellent postoperative outcome (ASES score >80), and 29.2% (7/24) had a satisfactory outcome (ASES score, 60-80). Thus, 100% of those with concomitant posterior shoulder instability and RCTs had a good or excellent outcome after surgical intervention. In those without rotator cuff pathology, 78.1% (25/32) had an excellent outcome, 12.5% (4/32) had a satisfactory outcome, and 9.4% (3/32) had a poor outcome. Thus, 91% of those without rotator cuff pathology had a good or excellent outcome after surgery.

Stability. Preoperative stability scores improved significantly after surgery in both groups (t = 7.2, P < .001; t = 10.5, P < .001). There were no statistical differences between preoperative or postoperative stability scores in those with or without rotator cuff pathology (t = 1.7, P = .095; t = .03, P = .975). Of throwing athletes with RCTs, 54.2% (13/24) had an excellent outcome, 33.3% (8/24) a good outcome, and 12.5% (3/24) a satisfactory outcome. Thus, 87.5% (21/24) of those with RCTs had a good or excellent outcome in terms of stability. In those without rotator cuff pathology, 46.9% (15/32) had excellent stability, 46.9% (15/32) had good stability, and 3.1% (1/32) had satisfactory stability after surgery. Thus, 93.8% (30/32) of throwing athletes without rotator cuff pathology had good or excellent stability after surgery.

Pain. Mean preoperative pain scores for those with and without rotator cuff pathology improved significantly (t = 13.4, P < .001; t = 7.1, P < .001). There was no statistical difference in preoperative or postoperative pain scores between those with and without rotator cuff pathology (t = 1.99, P = .051; t = .49, P = .627).

Function. Mean preoperative function scores for both groups improved significantly (t = 7.7, P < .001; t = 8.0, P < .001). There was no difference in improvement in functional scores between the two groups before or after surgery (t = .36, P = .721; t = .5, P = .622).

Continue to: ROM

ROM. Of those with rotator cuff pathology, 54% (13/24) had normal ROM, 42% (10/24) had satisfactory ROM, and 4% (1/24) had limited ROM. In throwing athletes without rotator cuff pathology, 34% (11/32) had normal ROM, 53.1% (17/32) had satisfactory ROM, and 9% (3/32) had limited ROM after surgery. There was no significant difference in ROM between the groups (c² = 2.7, P = .260).

Strength. Of those with RCTs, 67% (16/24) reported normal strength, 29% (7/24) slightly decreased strength, and 4% (1/24) markedly decreased strength. Of those throwing athletes without rotator cuff pathology, 50% (16/32) had normal strength, 41% (13/32) had slightly decreased strength, and 9% (3/32) had markedly decreased strength. No statistical difference was noted between the two groups (c² = 1.7, P = .429).

Return to Sport. Of those with RCTs, 92% (22/24) returned to sport while 84% (27/32) of throwing athletes without RCTs returned to sport. There was no difference between the two groups (c² = .667, P = .414). Sixty-seven percent (16/24) of those with RCTs and 56% (18/32) of those without RCTs returned to the same level of sport. No statistical difference was found in return to play between throwing athletes with and without rotator cuff pathology (c² = .624, P = .430).

Failures. According to ASES scores, no throwers with RCTs failed, while 9.4% (3/32) with posterior instability alone failed. Regarding stability, 8.3% (2/24) of athletes with RCTs failed, while 6.3% (2/32) with posterior instability alone failed. 

SURGICAL FINDINGS AND PROCEDURES

Of the 24 throwing athletes with rotator cuff pathology, 92% (22/24) had labral tears, while 78% (25/32) of those without RCTs had labral tears. The majority of RCTs were in the posterior supraspinatus and anterior infraspinatus regions. This was not significantly different between groups (c² = 1.86, P = .172). All labral pathology was posterior-inferior, and all RCTs were <50% thickness, and therefore were débrided. Fifty-four percent (13/24) of those with RCTs had a patulous capsule and 63% (20/32) of throwing athletes without rotator cuff pathology had a patulous capsule. There was no significant difference between groups (c² = .393, P = .530). Of those with RCTs, 92% (22/24) had surgical fixation with anchors, while 78% (25/32) of those without rotator cuff pathology underwent repair with anchor fixation. There was no statistically significant difference in anchor use between groups (c² = 1.86, P = .172).

Continue to: Discussion

Throwing athletes with and without RCTs had similar rates of recovery and return to play after arthroscopic capsular labral repair, with rotator cuff débridement if a tear was present. The mean follow-up was 3.2 years. Further, there was no difference in return to play (92% vs 84%), ASES score, stability, pain, function, ROM, or strength between the 2 groups before or after surgery. In this cohort of 56 patients, 24 throwing athletes (43%) were found to have RCTs.

Return-to-play rates showed no between-group differences; 92% (22/24) of athletes with concomitant RCTs returned to sport, and 67% (16/24) returned to the same level. Eight percent of throwing athletes with RCTs were unable to return to sport after surgery. These return-to-play rates are an improvement over most previously reported rates in throwing athletes and in posterior shoulder instability in general.^1-4,11 When these athletes are compared with their counterparts with combined SLAP tears and RCTs, return-to-play rates are notably higher. There may be discrepancies in interpreting return-to-play between the two studies, but in the current study, 67% of those with concomitant RCTs achieved return to preinjury level of play. This is 10% higher than the rate reported in athletes with SLAP tears alone (57%) and even higher than in those with concomitant SLAP and RCTs. It is also essential to note that a number of this cohort’s athletes who did not return to play did so for factors (eg, graduation) unrelated to the shoulder. However, the study by Neri and colleagues⁵ included professional athletes who likely all attempted to return to play and, if unable to perform at the same level, likely were unable to continue their professional career.⁵

All patients with RCTs had a good or excellent outcome (ASES score), and 70.8% had an excellent outcome. Similarly, 97% of those without rotator cuff pathology had a good or excellent outcome, and 81.3% had an excellent outcome. There was no significant difference between the two groups. These results parallel those of Neri and colleagues’⁵ study of SLAP tears with RCTs, where 96% (22/23) of throwing athletes had a good or excellent outcome. Despite these high outcome scores in patients with SLAP tears, only 57% were able to return to elite pitching.⁵ In the current study, pain was slightly higher for those with rotator cuff pathology before surgery—a finding consistent with pain frequently being found in patients with isolated partial-thickness RCTs. Their postoperative pain scores were actually lower on average than those of patients without RCTs, which suggests simple débridement of undersurface tears adequately addressed the pathology. The authors theorize that the main pain generator in this population may be posterior instability, and that the rotator cuff has less of an influence. In the SLAP population, the main pain generator likely is the RCT.

Failures by ASES score or strength were fairly rare in this cohort. Many patients opted to have revision surgery because of continued instability, pain, decreased function, or reinjury. One potential cause of failure in this cohort is inadequate capsular shift. However, capsular plication in throwing athletes is difficult to address, as overtensioning the repair can lead to the inability to adequately perform overhead activites.^3,4 This cannot be overemphasized, particularly with pitchers.

Partial-thickness RCTs, particularly those on the articular side, are common in throwing athletes because of high tensile and compressive loads.¹² Despite the known risk of RCTs with posterior shoulder instability in throwing athletes, the authors are unaware of reports of the incidence or treatment of this pathology. RCTs in this posterior instability group likely represent a pathology other than internal impingement. The high proportion of throwing athletes with RCTs in this study (43%) indicates a need for close evaluation of rotator cuff pathology in young throwing athletes. Ide et al found that 75% of patients with SLAP tears had partial articular-sided RCTs.¹³ In the current study, all RCTs were small partial tears, and arthroscopic débridement was performed. It is unknown whether repair of these RCTs would impact return to play. However, rotator cuff repair in this population has been shown to have poor outcomes. Tear thickness typically is used to determine treatment, with débridement performed if <50% tendon thickness is affected. More recently, many have advocated having greater tendon involvement in throwers before repair, because of poor outcomes. Although studies are limited, tear size does seem to correlate with outcomes.¹⁴

Continue to: Study Limitations

Limitations of this study include its small number of professional throwing athletes, with the majority being high school athletes. Further, although ASES scores are consistently used in posterior shoulder instability studies, these scores are influenced highly by pain scores, and some argue that other scoring systems may provide more useful information. However, none of the more modern scoring systems have been studied extensively in posterior glenohumeral instability. Further, because the authors used the present scoring systems previously,^1-4 they were continued to be used for comparison and consistency. Outcomes such as ROM and strength may carry more weight if measured and documented by clinical examination. Further testing, such as clinical evaluation of the jerk test or the posterior load-and-shift test, and their comparison before and after surgery may provide more objective data.

CONCLUSION

Arthroscopic capsulolabral reconstruction is successful in throwing athletes with RCTs treated with arthroscopic débridement. Unlike a previous study of throwing athletes’ outcomes after surgery for concomitant SLAP tears and RCTs,⁵ this study of throwing athletes with concomitant posterior shoulder instability and RCTs found no difference in patient-reported outcome measures or return to play. In throwing athletes with posterior instability and RCTs, arthroscopic posterior capsulolabral repair with rotator cuff débridement is successful.

Over the past few decades, there has been a dramatic increase in the number of shoulder arthroplasties performed around the world. This increase is the result of an aging and increasingly more active population, better implant technology, and the advent of reverse shoulder arthroplasty (RSA) for rotator cuff arthropathy. Additionally, as the indications for RSA have expanded to include pathologies such as rotator cuff insufficiency, chronic instabilities, trauma, and tumors, the number of arthroplasties will continue to increase. Although the results of most arthroplasties are good and predictable, any glenoid and/or humeral bone deficiencies can have detrimental effects on the clinical outcomes of these procedures. Bone loss becomes more of a problem in revision cases, and, as the number of primary arthroplasties increases, it follows that the number of revision procedures will also increase.

Many of the disease- or procedure-specific processes indicated for shoulder arthroplasty have predictable patterns of bone loss, especially on the glenoid side. Walch and colleagues¹ and Bercik and colleagues² made us aware that many patients with primary osteoarthritis have significant glenoid bone deformity. Similarly, there have been a number of first- and second-generation classification systems for delineating glenoid deformity in rotator cuff tear arthropathy and in revision settings. In revision settings, both glenoid and humeral bone deficiencies can occur as a result of implant removal, iatrogenic fracture, and even infection. Each of these bone loss patterns must be recognized and treated appropriately for the best surgical outcome.

The articles in this month of The American Journal of Orthopedics address the most up-to-date concepts and solutions regarding both humeral and glenoid bone loss in shoulder arthroplasty of all types.

HUMERAL BONE LOSS

Humeral bone loss is typically encountered in proximal humerus fractures, in revision surgery necessitating humeral component removal, and, less commonly, in tumors and infection.

In many displaced proximal humeral fractures indicated for shoulder arthroplasty, the bone is comminuted with displacement of the lesser and greater tuberosities. In these situations, failure of tuberosity healing may result in loss of rotator cuff function with loss of elevation, rotation, and even instability. Humeral shortening can also occur as a result of bone loss and can compromise deltoid function by loss of proper muscle tension, leading to instability, dysfunction, or both. In addition to possible instability, humeral shortening with metaphyseal bone loss can adversely affect long-term fixation of the humeral component, leading to stem loosening or failure. Cuff and colleagues³ showed significantly more rotational micromotion in cases lacking metaphyseal support, leading to aseptic loosening of the humeral stem.

Humeral bone loss can also result from humeral stem component removal in revision shoulder arthroplasty for infection, component failure or loosening, and even periprosthetic fracture resulting from surgery or trauma.

For the surgeon, humeral bone loss can create a complex set of circumstances related to rotator cuff attachment failure, soft-tissue balancing effects, and component fixation issues. Any such issue must be recognized and addressed for best outcomes. Best results can be obtained with preoperative imaging, planning, use of bone graft techniques, proximal humeral allografts, and, more recently, modular and patient-specific implants. All of these issues are discussed comprehensively in the articles this month.

Continue to: GLENOID BONE LOSS

Proper glenoid component placement with durable fixation is crucial for success in anatomical total shoulder arthroplasty and RSA. Glenoid bone deformity and loss can result from intrinsic deformity characteristics seen in primary osteoarthritis, cuff tear arthropathy, or glenoid component removal in revision situations and infection. These bone deformity complications can be extremely difficult to treat and in some cases lead to catastrophic failure of the index arthroplasty.

We are now aware that one key to success in the face of moderate to severe deformity is proper recognition. Newer imaging techniques, including 2-dimensional (2-D) computed tomography (CT) and 3-dimensional (3-D) modeling and surgical planning software tools, which are outlined in an upcoming article, have given surgeons important new instruments that can help in treating these difficult cases.

Glenoid bone deformity in primary osteoarthritis was well delineated in the 1999 seminal study of CT changes by Walch and colleagues.¹ The Walch classification system, which characterized glenoid morphology based on 2-D CT findings, was recently upgraded, based on 3-D imaging technology, to include Walch B3 and D patterns (Figure 1).² Recognition of certain primary deformities in osteoarthritis has led to increased use of RSA in some cases of Walch B2, B3, and C deformities with substantial glenoid retroversion and/or humeral head subluxation.⁴

In cases of rotator cuff tear arthropathy, glenoid bone deformities are well described with several classification systems based on degree and dimension of bone insufficiency. The Hamada classification system defines the degree of medial glenoid erosion and superior bone loss, as well as acetabularization of the acromion in 5 grades; ⁵ Rispoli and colleagues⁶ defined and graded the degree of medicalization of the glenohumeral joint based on degree of subchondral plate erosion; and Visotsky and colleagues⁷ based their classification system on wear patterns of bone loss, alignment, and concomitant soft-tissue insufficiencies leading to instability and rotation loss.

In severe glenoid bone deficiency after glenoid component removal, Antuna and colleagues⁸ described the classic findings related to medial bone loss, anterior and posterior wall failure, and combinations thereof.

Continue to: All these classification systems...

All these classification systems are based on the 2-D appearance of the glenoid and should be considered cautiously. The glenoid is a complex 3-D structure that can be affected by any number of disease processes, trauma, and surgical intervention. Using more modern CT techniques and 3-D imaging, we now know that many deformities previously classified as unidirectional are, instead, complex and multidirectional.

Frankle and colleagues⁹ developed a classification based more 3-D CT models which has further classified severe glenoid vault deformities in relation to direction and degree of bone loss (Figures 2A-2E). Using this system, they were better able to determine degree and direction of deformity than in previous 2-D evaluations, and they were able to determine the amount of glenoid vault bone available for baseplate fixation. Scalise and colleagues¹⁰ further defined the influence of such 3-D planning in total shoulder arthroplasty.

With knowledge of these classification systems and use of contemporary imaging systems, shoulder arthroplasty in cases of severe glenoid deficiency can be more successful. Potentially, we can improve outcomes even more in the more severe cases of bone loss with use of patient-specific planning tools, including the guides and patient-specific implants that are now readily available with many implant systems.¹¹

Preoperative planning tools, bone-grafting techniques, augmented and specialized glenoid and humeral implants, and patient-specific implants are discussed this month to give our readers a comprehensive review of the latest concepts in shoulder arthroplasty in cases of significant bone loss or deformity.

This month of The American Journal of Orthopedics presents the most current and cutting-edge solutions for humeral and glenoid bone deformities and deficiencies in contemporary shoulder arthroplasties.

Over the past few decades, there has been a dramatic increase in the number of shoulder arthroplasties performed around the world. This increase is the result of an aging and increasingly more active population, better implant technology, and the advent of reverse shoulder arthroplasty (RSA) for rotator cuff arthropathy. Additionally, as the indications for RSA have expanded to include pathologies such as rotator cuff insufficiency, chronic instabilities, trauma, and tumors, the number of arthroplasties will continue to increase. Although the results of most arthroplasties are good and predictable, any glenoid and/or humeral bone deficiencies can have detrimental effects on the clinical outcomes of these procedures. Bone loss becomes more of a problem in revision cases, and, as the number of primary arthroplasties increases, it follows that the number of revision procedures will also increase.

Many of the disease- or procedure-specific processes indicated for shoulder arthroplasty have predictable patterns of bone loss, especially on the glenoid side. Walch and colleagues¹ and Bercik and colleagues² made us aware that many patients with primary osteoarthritis have significant glenoid bone deformity. Similarly, there have been a number of first- and second-generation classification systems for delineating glenoid deformity in rotator cuff tear arthropathy and in revision settings. In revision settings, both glenoid and humeral bone deficiencies can occur as a result of implant removal, iatrogenic fracture, and even infection. Each of these bone loss patterns must be recognized and treated appropriately for the best surgical outcome.

The articles in this month of The American Journal of Orthopedics address the most up-to-date concepts and solutions regarding both humeral and glenoid bone loss in shoulder arthroplasty of all types.

HUMERAL BONE LOSS

Humeral bone loss is typically encountered in proximal humerus fractures, in revision surgery necessitating humeral component removal, and, less commonly, in tumors and infection.

In many displaced proximal humeral fractures indicated for shoulder arthroplasty, the bone is comminuted with displacement of the lesser and greater tuberosities. In these situations, failure of tuberosity healing may result in loss of rotator cuff function with loss of elevation, rotation, and even instability. Humeral shortening can also occur as a result of bone loss and can compromise deltoid function by loss of proper muscle tension, leading to instability, dysfunction, or both. In addition to possible instability, humeral shortening with metaphyseal bone loss can adversely affect long-term fixation of the humeral component, leading to stem loosening or failure. Cuff and colleagues³ showed significantly more rotational micromotion in cases lacking metaphyseal support, leading to aseptic loosening of the humeral stem.

Humeral bone loss can also result from humeral stem component removal in revision shoulder arthroplasty for infection, component failure or loosening, and even periprosthetic fracture resulting from surgery or trauma.

For the surgeon, humeral bone loss can create a complex set of circumstances related to rotator cuff attachment failure, soft-tissue balancing effects, and component fixation issues. Any such issue must be recognized and addressed for best outcomes. Best results can be obtained with preoperative imaging, planning, use of bone graft techniques, proximal humeral allografts, and, more recently, modular and patient-specific implants. All of these issues are discussed comprehensively in the articles this month.

Continue to: GLENOID BONE LOSS

Proper glenoid component placement with durable fixation is crucial for success in anatomical total shoulder arthroplasty and RSA. Glenoid bone deformity and loss can result from intrinsic deformity characteristics seen in primary osteoarthritis, cuff tear arthropathy, or glenoid component removal in revision situations and infection. These bone deformity complications can be extremely difficult to treat and in some cases lead to catastrophic failure of the index arthroplasty.

We are now aware that one key to success in the face of moderate to severe deformity is proper recognition. Newer imaging techniques, including 2-dimensional (2-D) computed tomography (CT) and 3-dimensional (3-D) modeling and surgical planning software tools, which are outlined in an upcoming article, have given surgeons important new instruments that can help in treating these difficult cases.

Glenoid bone deformity in primary osteoarthritis was well delineated in the 1999 seminal study of CT changes by Walch and colleagues.¹ The Walch classification system, which characterized glenoid morphology based on 2-D CT findings, was recently upgraded, based on 3-D imaging technology, to include Walch B3 and D patterns (Figure 1).² Recognition of certain primary deformities in osteoarthritis has led to increased use of RSA in some cases of Walch B2, B3, and C deformities with substantial glenoid retroversion and/or humeral head subluxation.⁴

In cases of rotator cuff tear arthropathy, glenoid bone deformities are well described with several classification systems based on degree and dimension of bone insufficiency. The Hamada classification system defines the degree of medial glenoid erosion and superior bone loss, as well as acetabularization of the acromion in 5 grades; ⁵ Rispoli and colleagues⁶ defined and graded the degree of medicalization of the glenohumeral joint based on degree of subchondral plate erosion; and Visotsky and colleagues⁷ based their classification system on wear patterns of bone loss, alignment, and concomitant soft-tissue insufficiencies leading to instability and rotation loss.

In severe glenoid bone deficiency after glenoid component removal, Antuna and colleagues⁸ described the classic findings related to medial bone loss, anterior and posterior wall failure, and combinations thereof.

Continue to: All these classification systems...

All these classification systems are based on the 2-D appearance of the glenoid and should be considered cautiously. The glenoid is a complex 3-D structure that can be affected by any number of disease processes, trauma, and surgical intervention. Using more modern CT techniques and 3-D imaging, we now know that many deformities previously classified as unidirectional are, instead, complex and multidirectional.

Frankle and colleagues⁹ developed a classification based more 3-D CT models which has further classified severe glenoid vault deformities in relation to direction and degree of bone loss (Figures 2A-2E). Using this system, they were better able to determine degree and direction of deformity than in previous 2-D evaluations, and they were able to determine the amount of glenoid vault bone available for baseplate fixation. Scalise and colleagues¹⁰ further defined the influence of such 3-D planning in total shoulder arthroplasty.

With knowledge of these classification systems and use of contemporary imaging systems, shoulder arthroplasty in cases of severe glenoid deficiency can be more successful. Potentially, we can improve outcomes even more in the more severe cases of bone loss with use of patient-specific planning tools, including the guides and patient-specific implants that are now readily available with many implant systems.¹¹

Preoperative planning tools, bone-grafting techniques, augmented and specialized glenoid and humeral implants, and patient-specific implants are discussed this month to give our readers a comprehensive review of the latest concepts in shoulder arthroplasty in cases of significant bone loss or deformity.

This month of The American Journal of Orthopedics presents the most current and cutting-edge solutions for humeral and glenoid bone deformities and deficiencies in contemporary shoulder arthroplasties.

Over the past few decades, there has been a dramatic increase in the number of shoulder arthroplasties performed around the world. This increase is the result of an aging and increasingly more active population, better implant technology, and the advent of reverse shoulder arthroplasty (RSA) for rotator cuff arthropathy. Additionally, as the indications for RSA have expanded to include pathologies such as rotator cuff insufficiency, chronic instabilities, trauma, and tumors, the number of arthroplasties will continue to increase. Although the results of most arthroplasties are good and predictable, any glenoid and/or humeral bone deficiencies can have detrimental effects on the clinical outcomes of these procedures. Bone loss becomes more of a problem in revision cases, and, as the number of primary arthroplasties increases, it follows that the number of revision procedures will also increase.

Many of the disease- or procedure-specific processes indicated for shoulder arthroplasty have predictable patterns of bone loss, especially on the glenoid side. Walch and colleagues¹ and Bercik and colleagues² made us aware that many patients with primary osteoarthritis have significant glenoid bone deformity. Similarly, there have been a number of first- and second-generation classification systems for delineating glenoid deformity in rotator cuff tear arthropathy and in revision settings. In revision settings, both glenoid and humeral bone deficiencies can occur as a result of implant removal, iatrogenic fracture, and even infection. Each of these bone loss patterns must be recognized and treated appropriately for the best surgical outcome.

The articles in this month of The American Journal of Orthopedics address the most up-to-date concepts and solutions regarding both humeral and glenoid bone loss in shoulder arthroplasty of all types.

HUMERAL BONE LOSS

Humeral bone loss is typically encountered in proximal humerus fractures, in revision surgery necessitating humeral component removal, and, less commonly, in tumors and infection.

In many displaced proximal humeral fractures indicated for shoulder arthroplasty, the bone is comminuted with displacement of the lesser and greater tuberosities. In these situations, failure of tuberosity healing may result in loss of rotator cuff function with loss of elevation, rotation, and even instability. Humeral shortening can also occur as a result of bone loss and can compromise deltoid function by loss of proper muscle tension, leading to instability, dysfunction, or both. In addition to possible instability, humeral shortening with metaphyseal bone loss can adversely affect long-term fixation of the humeral component, leading to stem loosening or failure. Cuff and colleagues³ showed significantly more rotational micromotion in cases lacking metaphyseal support, leading to aseptic loosening of the humeral stem.

Humeral bone loss can also result from humeral stem component removal in revision shoulder arthroplasty for infection, component failure or loosening, and even periprosthetic fracture resulting from surgery or trauma.

For the surgeon, humeral bone loss can create a complex set of circumstances related to rotator cuff attachment failure, soft-tissue balancing effects, and component fixation issues. Any such issue must be recognized and addressed for best outcomes. Best results can be obtained with preoperative imaging, planning, use of bone graft techniques, proximal humeral allografts, and, more recently, modular and patient-specific implants. All of these issues are discussed comprehensively in the articles this month.

Continue to: GLENOID BONE LOSS

Proper glenoid component placement with durable fixation is crucial for success in anatomical total shoulder arthroplasty and RSA. Glenoid bone deformity and loss can result from intrinsic deformity characteristics seen in primary osteoarthritis, cuff tear arthropathy, or glenoid component removal in revision situations and infection. These bone deformity complications can be extremely difficult to treat and in some cases lead to catastrophic failure of the index arthroplasty.

We are now aware that one key to success in the face of moderate to severe deformity is proper recognition. Newer imaging techniques, including 2-dimensional (2-D) computed tomography (CT) and 3-dimensional (3-D) modeling and surgical planning software tools, which are outlined in an upcoming article, have given surgeons important new instruments that can help in treating these difficult cases.

Glenoid bone deformity in primary osteoarthritis was well delineated in the 1999 seminal study of CT changes by Walch and colleagues.¹ The Walch classification system, which characterized glenoid morphology based on 2-D CT findings, was recently upgraded, based on 3-D imaging technology, to include Walch B3 and D patterns (Figure 1).² Recognition of certain primary deformities in osteoarthritis has led to increased use of RSA in some cases of Walch B2, B3, and C deformities with substantial glenoid retroversion and/or humeral head subluxation.⁴

In cases of rotator cuff tear arthropathy, glenoid bone deformities are well described with several classification systems based on degree and dimension of bone insufficiency. The Hamada classification system defines the degree of medial glenoid erosion and superior bone loss, as well as acetabularization of the acromion in 5 grades; ⁵ Rispoli and colleagues⁶ defined and graded the degree of medicalization of the glenohumeral joint based on degree of subchondral plate erosion; and Visotsky and colleagues⁷ based their classification system on wear patterns of bone loss, alignment, and concomitant soft-tissue insufficiencies leading to instability and rotation loss.

In severe glenoid bone deficiency after glenoid component removal, Antuna and colleagues⁸ described the classic findings related to medial bone loss, anterior and posterior wall failure, and combinations thereof.

Continue to: All these classification systems...

All these classification systems are based on the 2-D appearance of the glenoid and should be considered cautiously. The glenoid is a complex 3-D structure that can be affected by any number of disease processes, trauma, and surgical intervention. Using more modern CT techniques and 3-D imaging, we now know that many deformities previously classified as unidirectional are, instead, complex and multidirectional.

Frankle and colleagues⁹ developed a classification based more 3-D CT models which has further classified severe glenoid vault deformities in relation to direction and degree of bone loss (Figures 2A-2E). Using this system, they were better able to determine degree and direction of deformity than in previous 2-D evaluations, and they were able to determine the amount of glenoid vault bone available for baseplate fixation. Scalise and colleagues¹⁰ further defined the influence of such 3-D planning in total shoulder arthroplasty.

With knowledge of these classification systems and use of contemporary imaging systems, shoulder arthroplasty in cases of severe glenoid deficiency can be more successful. Potentially, we can improve outcomes even more in the more severe cases of bone loss with use of patient-specific planning tools, including the guides and patient-specific implants that are now readily available with many implant systems.¹¹

Preoperative planning tools, bone-grafting techniques, augmented and specialized glenoid and humeral implants, and patient-specific implants are discussed this month to give our readers a comprehensive review of the latest concepts in shoulder arthroplasty in cases of significant bone loss or deformity.

This month of The American Journal of Orthopedics presents the most current and cutting-edge solutions for humeral and glenoid bone deformities and deficiencies in contemporary shoulder arthroplasties.

ABSTRACT

We conducted a study to determine the common characteristics of patients who developed radiculopathy symptoms and corresponding heterotopic ossification (HO) from transforaminal lumbar interbody fusions (TLIF) using recombinant human bone morphogenetic protein 2 (rhBMP-2). HO can arise from a disk space with rhBMP-2 use in TLIF. Formation of bone around nerve roots or the thecal sac can cause a radiculopathy with a consistent pattern of symptoms.

We identified 38 patients (26 males, 12 females) with a mean (SD) age of 50.8 (7.5) years who developed radiculopathy symptoms and corresponding HO from TLIF with rhBMP-2 in the disk space between 2002 and 2015. To document this complication and improve its recognition, we recorded common patterns of symptom development and radiologic findings: specifically, time from implantation of rhBMP-2 to symptom development, consistency with side of TLIF placement, and radiologic findings.

Radicular pain generally developed a mean (SD) of 3.8 (1.0) months after TLIF with rhBMP-2. Development of radiculopathy symptoms corresponded to consistent “pseudo-pedicle”-like HO. In all 38 patients, HO arising from the annulotomy site showed a distinct pseudo-pedicle pattern encompassing nerve roots and the thecal sac. In addition, development of radiculopathy symptoms and corresponding HO appear to be independent of amount of rhBMP-2. HO resulting from TLIF with rhBMP-2 in the disk space is a pain generator and a recognizable complication that can be diagnosed by assessment of symptoms and computed tomography characteristics.

Continue to: Bone morphogenetic proteins...

Bone morphogenetic proteins (BMPs), first isolated by Urist in 1964¹, are a family of growth factors that stimulate the cascade of bone formation. Recombinant human BMP (rhBMP), specifically rhBMP-2 and rhBMP-7 (also known as osteogenic protein 1 [OP-1]), was developed in the 1990s after the advent of gene splicing. Then, in 2002, the US Food and Drug Administration (FDA) approved use of rhBMP to stimulate fusion in the human spine. Specifically, rhBMP-2 (Medtronic) was approved for use in combination with a specific brand of interbody cage in 1-level anterior lumbar interbody fusion.² Over the past decade, off-label use of rhBMP-2 to achieve osseous union has increased dramatically, particularly in spinal surgery: transforaminal lumbar interbody fusion (TLIF), posterior lumbar interbody fusion, and posterolateral lumbar fusion.^3-9 However, this widespread off-label use for posterior spinal fusion began despite FDA data indicating that specific complications were underreported in the peer-reviewed literature.^10,11 Although rhBMP-2 is very effective in increasing osteoblast formation and improving osteogenesis and subsequent bone healing in spinal surgery,^12,13 its use in TLIF resulted in significant adverse side effects, including radiculopathy with and without neuroforaminal heterotopic ossification (HO); ^14-24 complications in the FDA studies; ^14,22,25-27 and osteolysis causing intervertebral cage subsidence, inflammatory radiculitis, genitourinary complications, infections, possible systemic effects, and significant HO complications.^10,28-30 Of these, HO complications involved rhBMP leakage through the annulotomy to the disk space that led to HO. Specifically, rhBMP leaked directly out of the disk space and formed a pillar of bone that encased the nerve roots and dura, which led to occlusion of the foramen and symptoms of radiculopathy.^10,28-30

Despite this frequent finding of HO in the intervertebral space outside the target fusion area, use of rhBMP-2 with intervertebral cages increased so rapidly that rhBMP-2 was used more often than autologous bone.^5,11,17,31 In this study, we reviewed the common characteristics of patients who developed HO and subsequent radiculopathy from TLIF with rhBMP.

METHODS

After this study received Institutional Review Board approval, we retrospectively reviewed cases of radiculopathy symptoms that developed after TLIF with rhBMP between January 2002 and January 2015. During this period, 38 patients (26 males, 12 females) with a mean (SD) age of 50.8 (7.5) years and radiculopathy symptoms arising from TLIF with rhBMP-2 were identified to determine commonalities and defining characteristics that will help facilitate diagnosis.

Inclusion criteria were computed tomography (CT)–documented HO arising from the TLIF annulotomy site in continuity with bone in the disk space or ectopic bone forming a distinctive shell with contouring around the thecal sac or nerve roots, as well as recurrence or initial occurrence of radiculopathy with signs and symptoms corresponding to the CT site of aberrant bone growth in terms of laterality and particular nerve root(s) involved. Exclusion criteria were malplacement of interbody cage or pedicle screws, disk herniation, systemic neuropathic disease, and new or unresolved radiculopathy immediately after index surgery.

To improve recognition of this complication, we also documented the amount of BMP used, common patterns of radiculopathy symptom development, and radiologic findings. Type and timing of radiculopathy symptom onset and consistency with side of TLIF placement were documented as well. Radiculopathy symptoms included shooting pain in the legs, incontinence, sexual dysfunction, and severe paralysis. Radiologic findings were specific to bone formation from the disk space (detected with CT).

Continue to: RESULTS

All 38 selected patients had radiculopathy symptoms from HO out of the intervertebral space. The Table lists the patients’ overall characteristics. The left side had the most radiculopathy symptoms (31/38 patients), followed by the right side (5/38) and both sides (2/38). Radiculopathy symptoms began a mean (SD) of 3.8 (1.0) months (range, 2-6 months) after index surgery. The 38 patients had 4 characteristics in common:

Table. Transforaminal Lumbar Interbody Fusion With Recombinant Human Bone Morphogenetic Protein 2: Onset Time for Radiculopathy Symptoms, Surgery Level, Side of Pseudo-Pedicle Bone Formation, and Subsequent Complications

1. Bone growing out of the annulotomy site for TLIF cage placement was present and in continuity with the disk space in 33 (87%) of the 38 cases. In the other 5 cases (13%), HO was present around the neural tissue, but not necessarily in continuity with the disk space. This bone appeared ectopic and not osteophytic and facet-related, as it formed a shell around either the nerve root or the thecal sac, contouring to the structure.

2. The common, novel finding on CT was a “pseudo-pedicle” (Figures 1A, 1B), which appeared as ectopic growth from the disk space—a solid piece of bone in the same direction as the anatomical pedicle. Confusing similarity to the anatomical pedicle is present on axial cuts and during surgery. The pseudo-pedicle varied in thickness and extent out of the disk space, but was always presented as a bar of bone arising from the annulotomy site. After arising from the disk space, the HO could disperse in any direction, further calcifying neural structures or the facet joints above or below. There was no apparent distinguishable repeating pattern, given the variable nature of arthritic facet changes, scoliotic deformities, size of annulotomies, amount of rhBMP used, and placement in cage and disk space or only in cage.

3. In 36 (95%) of the 38 cases, the initial interpretation of HO on magnetic resonance imaging (MRI) was of tissue other than bone, such as fibrous tissue, granulation tissue, recurrent disk herniation, or postoperative changes. However, this tissue was later determined to be bone from HO complications, which we confirmed with CT in all 38 cases. It is important to note that HO on MRI (Figures 2A, 2B) was initially interpreted by a radiologist as fibrous tissue, but same-level CT of the same case (Figures 3A, 3B) showed clear HO.

4. The radiculopathy symptoms caused by HO were independent of the amount of rhBMP-2 used in TLIF. Of the 38 patients, 19 had 1 rhBMP-2 sponge placed in the cage, 12 had a small kit sponge (1.05 mg), 5 had 1 sponge placed in the cage and 1 sponge placed directly in the disk space before cage placement (no notation of precise size or amount of rhBMP-2), and 2 had 1 sponge placed in the cage (no notation of rhBMP-2 amount). The data showed that HO can occur with even a small amount of rhBMP-2.

Continue to: Bone formation with rhBMP-2...

Bone formation with rhBMP-2 is robust and beneficial, but HO-related complications are significant, and identifiable on assessment of radiculopathy symptoms and CT characteristics.

DISCUSSION

We identified 38 patients with a recognizable and consistent pattern of complications of off-label use of rhBMP-2 in TLIF performed at our institution between 2002 and 2015. This pattern included consistent radiculopathy symptoms with corresponding HO at the annulotomy site in continuity with bone in the disk space or ectopic bone forming a distinctive shell around the thecal sac or nerve roots, as well as showing a distinct pseudo-pedicle pattern encompassing nerve roots and the thecal sac. Our finding differs from other findings of similar complication characteristics, but with much larger variations without consistency within the patient population.^19,20,22,24 Specifically, previous studies found an association between off-label rhBMP-2 use in the posterior spine and radiculopathy with and without neuroforaminal HO. However, our study found consistent radiculopathy symptoms with pseudo-pedicle-like HO complications in all its 38 patients a mean (SD) of 3.8 (1.0) months after surgery.

In this study, consistent radiculopathy symptoms with pseudo-pedicle-like HO complications were independent of the amount of rhBMP-2 used, as some complications occurred with use of small pack rhBMP-2 with TLIF. It is well understood that high doses of rhBMP-2 may be required to improve fusion rates, but to our knowledge an optimal dosing strategy for TLIF has not been reported, particularly with respect to potential complications.^8,20,31-33 For anterior lumbar interbody fusion surgery, the FDA-approved use of rhBMP-2 appears to have a significantly decreased risk of neuroforaminal HO complications. This may be attributable to the protective presence of the intact posterior annulus and longitudinal ligament for this procedure.^20,33 For TLIF, it has been suggested that rhBMP-2 should be placed only along the anterior annulus with a posterior strut and morselized bone allograft barricade,³³ and that fibrin glue should be used to limit BMP diffusion through the annulotomy site³¹ to prevent this complication.

Our study results suggest that radiculopathy symptoms with pseudo-pedicle-like HO complications appear to be caused by leakage of rhBMP-2 from the disk space through the annulotomy site. This was often initially interpreted incorrectly on MRI in the first year after surgery as being fibrous or granulation tissue, or even postoperative changes that the heterotopic tissue was bone was obvious only on CT. Even then the tissue may be incorrectly identified, as the encasing nerve roots in bone are similar to the scar tissue having no compressive effect. HO may compress, but it also has an inflammatory component that the scars lack. Additionally, the HO from the disk space, caused by leakage of the BMP placed in or around the fusion cage, can create a pseudo-pedicle of varying size and extent. This was present in all 38 of our cases.

This retrospective case series had its limitations. Its clinical and radiographic findings were not blinded. Confounding variables cannot be isolated for causal relationships, if any, to the complication in a case series such as this.

Bone formation with rhBMP-2 is robust and beneficial, but HO-related complications are significant, and identifiable on assessment of radiculopathy symptoms and CT characteristics.

ABSTRACT

We conducted a study to determine the common characteristics of patients who developed radiculopathy symptoms and corresponding heterotopic ossification (HO) from transforaminal lumbar interbody fusions (TLIF) using recombinant human bone morphogenetic protein 2 (rhBMP-2). HO can arise from a disk space with rhBMP-2 use in TLIF. Formation of bone around nerve roots or the thecal sac can cause a radiculopathy with a consistent pattern of symptoms.

We identified 38 patients (26 males, 12 females) with a mean (SD) age of 50.8 (7.5) years who developed radiculopathy symptoms and corresponding HO from TLIF with rhBMP-2 in the disk space between 2002 and 2015. To document this complication and improve its recognition, we recorded common patterns of symptom development and radiologic findings: specifically, time from implantation of rhBMP-2 to symptom development, consistency with side of TLIF placement, and radiologic findings.

Radicular pain generally developed a mean (SD) of 3.8 (1.0) months after TLIF with rhBMP-2. Development of radiculopathy symptoms corresponded to consistent “pseudo-pedicle”-like HO. In all 38 patients, HO arising from the annulotomy site showed a distinct pseudo-pedicle pattern encompassing nerve roots and the thecal sac. In addition, development of radiculopathy symptoms and corresponding HO appear to be independent of amount of rhBMP-2. HO resulting from TLIF with rhBMP-2 in the disk space is a pain generator and a recognizable complication that can be diagnosed by assessment of symptoms and computed tomography characteristics.

Continue to: Bone morphogenetic proteins...

Bone morphogenetic proteins (BMPs), first isolated by Urist in 1964¹, are a family of growth factors that stimulate the cascade of bone formation. Recombinant human BMP (rhBMP), specifically rhBMP-2 and rhBMP-7 (also known as osteogenic protein 1 [OP-1]), was developed in the 1990s after the advent of gene splicing. Then, in 2002, the US Food and Drug Administration (FDA) approved use of rhBMP to stimulate fusion in the human spine. Specifically, rhBMP-2 (Medtronic) was approved for use in combination with a specific brand of interbody cage in 1-level anterior lumbar interbody fusion.² Over the past decade, off-label use of rhBMP-2 to achieve osseous union has increased dramatically, particularly in spinal surgery: transforaminal lumbar interbody fusion (TLIF), posterior lumbar interbody fusion, and posterolateral lumbar fusion.^3-9 However, this widespread off-label use for posterior spinal fusion began despite FDA data indicating that specific complications were underreported in the peer-reviewed literature.^10,11 Although rhBMP-2 is very effective in increasing osteoblast formation and improving osteogenesis and subsequent bone healing in spinal surgery,^12,13 its use in TLIF resulted in significant adverse side effects, including radiculopathy with and without neuroforaminal heterotopic ossification (HO); ^14-24 complications in the FDA studies; ^14,22,25-27 and osteolysis causing intervertebral cage subsidence, inflammatory radiculitis, genitourinary complications, infections, possible systemic effects, and significant HO complications.^10,28-30 Of these, HO complications involved rhBMP leakage through the annulotomy to the disk space that led to HO. Specifically, rhBMP leaked directly out of the disk space and formed a pillar of bone that encased the nerve roots and dura, which led to occlusion of the foramen and symptoms of radiculopathy.^10,28-30

Despite this frequent finding of HO in the intervertebral space outside the target fusion area, use of rhBMP-2 with intervertebral cages increased so rapidly that rhBMP-2 was used more often than autologous bone.^5,11,17,31 In this study, we reviewed the common characteristics of patients who developed HO and subsequent radiculopathy from TLIF with rhBMP.

METHODS

After this study received Institutional Review Board approval, we retrospectively reviewed cases of radiculopathy symptoms that developed after TLIF with rhBMP between January 2002 and January 2015. During this period, 38 patients (26 males, 12 females) with a mean (SD) age of 50.8 (7.5) years and radiculopathy symptoms arising from TLIF with rhBMP-2 were identified to determine commonalities and defining characteristics that will help facilitate diagnosis.

Inclusion criteria were computed tomography (CT)–documented HO arising from the TLIF annulotomy site in continuity with bone in the disk space or ectopic bone forming a distinctive shell with contouring around the thecal sac or nerve roots, as well as recurrence or initial occurrence of radiculopathy with signs and symptoms corresponding to the CT site of aberrant bone growth in terms of laterality and particular nerve root(s) involved. Exclusion criteria were malplacement of interbody cage or pedicle screws, disk herniation, systemic neuropathic disease, and new or unresolved radiculopathy immediately after index surgery.

To improve recognition of this complication, we also documented the amount of BMP used, common patterns of radiculopathy symptom development, and radiologic findings. Type and timing of radiculopathy symptom onset and consistency with side of TLIF placement were documented as well. Radiculopathy symptoms included shooting pain in the legs, incontinence, sexual dysfunction, and severe paralysis. Radiologic findings were specific to bone formation from the disk space (detected with CT).

Continue to: RESULTS

All 38 selected patients had radiculopathy symptoms from HO out of the intervertebral space. The Table lists the patients’ overall characteristics. The left side had the most radiculopathy symptoms (31/38 patients), followed by the right side (5/38) and both sides (2/38). Radiculopathy symptoms began a mean (SD) of 3.8 (1.0) months (range, 2-6 months) after index surgery. The 38 patients had 4 characteristics in common:

Table. Transforaminal Lumbar Interbody Fusion With Recombinant Human Bone Morphogenetic Protein 2: Onset Time for Radiculopathy Symptoms, Surgery Level, Side of Pseudo-Pedicle Bone Formation, and Subsequent Complications

1. Bone growing out of the annulotomy site for TLIF cage placement was present and in continuity with the disk space in 33 (87%) of the 38 cases. In the other 5 cases (13%), HO was present around the neural tissue, but not necessarily in continuity with the disk space. This bone appeared ectopic and not osteophytic and facet-related, as it formed a shell around either the nerve root or the thecal sac, contouring to the structure.

2. The common, novel finding on CT was a “pseudo-pedicle” (Figures 1A, 1B), which appeared as ectopic growth from the disk space—a solid piece of bone in the same direction as the anatomical pedicle. Confusing similarity to the anatomical pedicle is present on axial cuts and during surgery. The pseudo-pedicle varied in thickness and extent out of the disk space, but was always presented as a bar of bone arising from the annulotomy site. After arising from the disk space, the HO could disperse in any direction, further calcifying neural structures or the facet joints above or below. There was no apparent distinguishable repeating pattern, given the variable nature of arthritic facet changes, scoliotic deformities, size of annulotomies, amount of rhBMP used, and placement in cage and disk space or only in cage.

3. In 36 (95%) of the 38 cases, the initial interpretation of HO on magnetic resonance imaging (MRI) was of tissue other than bone, such as fibrous tissue, granulation tissue, recurrent disk herniation, or postoperative changes. However, this tissue was later determined to be bone from HO complications, which we confirmed with CT in all 38 cases. It is important to note that HO on MRI (Figures 2A, 2B) was initially interpreted by a radiologist as fibrous tissue, but same-level CT of the same case (Figures 3A, 3B) showed clear HO.

4. The radiculopathy symptoms caused by HO were independent of the amount of rhBMP-2 used in TLIF. Of the 38 patients, 19 had 1 rhBMP-2 sponge placed in the cage, 12 had a small kit sponge (1.05 mg), 5 had 1 sponge placed in the cage and 1 sponge placed directly in the disk space before cage placement (no notation of precise size or amount of rhBMP-2), and 2 had 1 sponge placed in the cage (no notation of rhBMP-2 amount). The data showed that HO can occur with even a small amount of rhBMP-2.

Continue to: Bone formation with rhBMP-2...

Bone formation with rhBMP-2 is robust and beneficial, but HO-related complications are significant, and identifiable on assessment of radiculopathy symptoms and CT characteristics.

DISCUSSION

We identified 38 patients with a recognizable and consistent pattern of complications of off-label use of rhBMP-2 in TLIF performed at our institution between 2002 and 2015. This pattern included consistent radiculopathy symptoms with corresponding HO at the annulotomy site in continuity with bone in the disk space or ectopic bone forming a distinctive shell around the thecal sac or nerve roots, as well as showing a distinct pseudo-pedicle pattern encompassing nerve roots and the thecal sac. Our finding differs from other findings of similar complication characteristics, but with much larger variations without consistency within the patient population.^19,20,22,24 Specifically, previous studies found an association between off-label rhBMP-2 use in the posterior spine and radiculopathy with and without neuroforaminal HO. However, our study found consistent radiculopathy symptoms with pseudo-pedicle-like HO complications in all its 38 patients a mean (SD) of 3.8 (1.0) months after surgery.

In this study, consistent radiculopathy symptoms with pseudo-pedicle-like HO complications were independent of the amount of rhBMP-2 used, as some complications occurred with use of small pack rhBMP-2 with TLIF. It is well understood that high doses of rhBMP-2 may be required to improve fusion rates, but to our knowledge an optimal dosing strategy for TLIF has not been reported, particularly with respect to potential complications.^8,20,31-33 For anterior lumbar interbody fusion surgery, the FDA-approved use of rhBMP-2 appears to have a significantly decreased risk of neuroforaminal HO complications. This may be attributable to the protective presence of the intact posterior annulus and longitudinal ligament for this procedure.^20,33 For TLIF, it has been suggested that rhBMP-2 should be placed only along the anterior annulus with a posterior strut and morselized bone allograft barricade,³³ and that fibrin glue should be used to limit BMP diffusion through the annulotomy site³¹ to prevent this complication.

Our study results suggest that radiculopathy symptoms with pseudo-pedicle-like HO complications appear to be caused by leakage of rhBMP-2 from the disk space through the annulotomy site. This was often initially interpreted incorrectly on MRI in the first year after surgery as being fibrous or granulation tissue, or even postoperative changes that the heterotopic tissue was bone was obvious only on CT. Even then the tissue may be incorrectly identified, as the encasing nerve roots in bone are similar to the scar tissue having no compressive effect. HO may compress, but it also has an inflammatory component that the scars lack. Additionally, the HO from the disk space, caused by leakage of the BMP placed in or around the fusion cage, can create a pseudo-pedicle of varying size and extent. This was present in all 38 of our cases.

This retrospective case series had its limitations. Its clinical and radiographic findings were not blinded. Confounding variables cannot be isolated for causal relationships, if any, to the complication in a case series such as this.

Bone formation with rhBMP-2 is robust and beneficial, but HO-related complications are significant, and identifiable on assessment of radiculopathy symptoms and CT characteristics.

ABSTRACT

We conducted a study to determine the common characteristics of patients who developed radiculopathy symptoms and corresponding heterotopic ossification (HO) from transforaminal lumbar interbody fusions (TLIF) using recombinant human bone morphogenetic protein 2 (rhBMP-2). HO can arise from a disk space with rhBMP-2 use in TLIF. Formation of bone around nerve roots or the thecal sac can cause a radiculopathy with a consistent pattern of symptoms.

We identified 38 patients (26 males, 12 females) with a mean (SD) age of 50.8 (7.5) years who developed radiculopathy symptoms and corresponding HO from TLIF with rhBMP-2 in the disk space between 2002 and 2015. To document this complication and improve its recognition, we recorded common patterns of symptom development and radiologic findings: specifically, time from implantation of rhBMP-2 to symptom development, consistency with side of TLIF placement, and radiologic findings.

Radicular pain generally developed a mean (SD) of 3.8 (1.0) months after TLIF with rhBMP-2. Development of radiculopathy symptoms corresponded to consistent “pseudo-pedicle”-like HO. In all 38 patients, HO arising from the annulotomy site showed a distinct pseudo-pedicle pattern encompassing nerve roots and the thecal sac. In addition, development of radiculopathy symptoms and corresponding HO appear to be independent of amount of rhBMP-2. HO resulting from TLIF with rhBMP-2 in the disk space is a pain generator and a recognizable complication that can be diagnosed by assessment of symptoms and computed tomography characteristics.

Continue to: Bone morphogenetic proteins...

Bone morphogenetic proteins (BMPs), first isolated by Urist in 1964¹, are a family of growth factors that stimulate the cascade of bone formation. Recombinant human BMP (rhBMP), specifically rhBMP-2 and rhBMP-7 (also known as osteogenic protein 1 [OP-1]), was developed in the 1990s after the advent of gene splicing. Then, in 2002, the US Food and Drug Administration (FDA) approved use of rhBMP to stimulate fusion in the human spine. Specifically, rhBMP-2 (Medtronic) was approved for use in combination with a specific brand of interbody cage in 1-level anterior lumbar interbody fusion.² Over the past decade, off-label use of rhBMP-2 to achieve osseous union has increased dramatically, particularly in spinal surgery: transforaminal lumbar interbody fusion (TLIF), posterior lumbar interbody fusion, and posterolateral lumbar fusion.^3-9 However, this widespread off-label use for posterior spinal fusion began despite FDA data indicating that specific complications were underreported in the peer-reviewed literature.^10,11 Although rhBMP-2 is very effective in increasing osteoblast formation and improving osteogenesis and subsequent bone healing in spinal surgery,^12,13 its use in TLIF resulted in significant adverse side effects, including radiculopathy with and without neuroforaminal heterotopic ossification (HO); ^14-24 complications in the FDA studies; ^14,22,25-27 and osteolysis causing intervertebral cage subsidence, inflammatory radiculitis, genitourinary complications, infections, possible systemic effects, and significant HO complications.^10,28-30 Of these, HO complications involved rhBMP leakage through the annulotomy to the disk space that led to HO. Specifically, rhBMP leaked directly out of the disk space and formed a pillar of bone that encased the nerve roots and dura, which led to occlusion of the foramen and symptoms of radiculopathy.^10,28-30

Despite this frequent finding of HO in the intervertebral space outside the target fusion area, use of rhBMP-2 with intervertebral cages increased so rapidly that rhBMP-2 was used more often than autologous bone.^5,11,17,31 In this study, we reviewed the common characteristics of patients who developed HO and subsequent radiculopathy from TLIF with rhBMP.

METHODS

After this study received Institutional Review Board approval, we retrospectively reviewed cases of radiculopathy symptoms that developed after TLIF with rhBMP between January 2002 and January 2015. During this period, 38 patients (26 males, 12 females) with a mean (SD) age of 50.8 (7.5) years and radiculopathy symptoms arising from TLIF with rhBMP-2 were identified to determine commonalities and defining characteristics that will help facilitate diagnosis.

Inclusion criteria were computed tomography (CT)–documented HO arising from the TLIF annulotomy site in continuity with bone in the disk space or ectopic bone forming a distinctive shell with contouring around the thecal sac or nerve roots, as well as recurrence or initial occurrence of radiculopathy with signs and symptoms corresponding to the CT site of aberrant bone growth in terms of laterality and particular nerve root(s) involved. Exclusion criteria were malplacement of interbody cage or pedicle screws, disk herniation, systemic neuropathic disease, and new or unresolved radiculopathy immediately after index surgery.

To improve recognition of this complication, we also documented the amount of BMP used, common patterns of radiculopathy symptom development, and radiologic findings. Type and timing of radiculopathy symptom onset and consistency with side of TLIF placement were documented as well. Radiculopathy symptoms included shooting pain in the legs, incontinence, sexual dysfunction, and severe paralysis. Radiologic findings were specific to bone formation from the disk space (detected with CT).

Continue to: RESULTS

All 38 selected patients had radiculopathy symptoms from HO out of the intervertebral space. The Table lists the patients’ overall characteristics. The left side had the most radiculopathy symptoms (31/38 patients), followed by the right side (5/38) and both sides (2/38). Radiculopathy symptoms began a mean (SD) of 3.8 (1.0) months (range, 2-6 months) after index surgery. The 38 patients had 4 characteristics in common:

Table. Transforaminal Lumbar Interbody Fusion With Recombinant Human Bone Morphogenetic Protein 2: Onset Time for Radiculopathy Symptoms, Surgery Level, Side of Pseudo-Pedicle Bone Formation, and Subsequent Complications

1. Bone growing out of the annulotomy site for TLIF cage placement was present and in continuity with the disk space in 33 (87%) of the 38 cases. In the other 5 cases (13%), HO was present around the neural tissue, but not necessarily in continuity with the disk space. This bone appeared ectopic and not osteophytic and facet-related, as it formed a shell around either the nerve root or the thecal sac, contouring to the structure.

2. The common, novel finding on CT was a “pseudo-pedicle” (Figures 1A, 1B), which appeared as ectopic growth from the disk space—a solid piece of bone in the same direction as the anatomical pedicle. Confusing similarity to the anatomical pedicle is present on axial cuts and during surgery. The pseudo-pedicle varied in thickness and extent out of the disk space, but was always presented as a bar of bone arising from the annulotomy site. After arising from the disk space, the HO could disperse in any direction, further calcifying neural structures or the facet joints above or below. There was no apparent distinguishable repeating pattern, given the variable nature of arthritic facet changes, scoliotic deformities, size of annulotomies, amount of rhBMP used, and placement in cage and disk space or only in cage.

3. In 36 (95%) of the 38 cases, the initial interpretation of HO on magnetic resonance imaging (MRI) was of tissue other than bone, such as fibrous tissue, granulation tissue, recurrent disk herniation, or postoperative changes. However, this tissue was later determined to be bone from HO complications, which we confirmed with CT in all 38 cases. It is important to note that HO on MRI (Figures 2A, 2B) was initially interpreted by a radiologist as fibrous tissue, but same-level CT of the same case (Figures 3A, 3B) showed clear HO.

4. The radiculopathy symptoms caused by HO were independent of the amount of rhBMP-2 used in TLIF. Of the 38 patients, 19 had 1 rhBMP-2 sponge placed in the cage, 12 had a small kit sponge (1.05 mg), 5 had 1 sponge placed in the cage and 1 sponge placed directly in the disk space before cage placement (no notation of precise size or amount of rhBMP-2), and 2 had 1 sponge placed in the cage (no notation of rhBMP-2 amount). The data showed that HO can occur with even a small amount of rhBMP-2.

Continue to: Bone formation with rhBMP-2...

Bone formation with rhBMP-2 is robust and beneficial, but HO-related complications are significant, and identifiable on assessment of radiculopathy symptoms and CT characteristics.

DISCUSSION

We identified 38 patients with a recognizable and consistent pattern of complications of off-label use of rhBMP-2 in TLIF performed at our institution between 2002 and 2015. This pattern included consistent radiculopathy symptoms with corresponding HO at the annulotomy site in continuity with bone in the disk space or ectopic bone forming a distinctive shell around the thecal sac or nerve roots, as well as showing a distinct pseudo-pedicle pattern encompassing nerve roots and the thecal sac. Our finding differs from other findings of similar complication characteristics, but with much larger variations without consistency within the patient population.^19,20,22,24 Specifically, previous studies found an association between off-label rhBMP-2 use in the posterior spine and radiculopathy with and without neuroforaminal HO. However, our study found consistent radiculopathy symptoms with pseudo-pedicle-like HO complications in all its 38 patients a mean (SD) of 3.8 (1.0) months after surgery.

In this study, consistent radiculopathy symptoms with pseudo-pedicle-like HO complications were independent of the amount of rhBMP-2 used, as some complications occurred with use of small pack rhBMP-2 with TLIF. It is well understood that high doses of rhBMP-2 may be required to improve fusion rates, but to our knowledge an optimal dosing strategy for TLIF has not been reported, particularly with respect to potential complications.^8,20,31-33 For anterior lumbar interbody fusion surgery, the FDA-approved use of rhBMP-2 appears to have a significantly decreased risk of neuroforaminal HO complications. This may be attributable to the protective presence of the intact posterior annulus and longitudinal ligament for this procedure.^20,33 For TLIF, it has been suggested that rhBMP-2 should be placed only along the anterior annulus with a posterior strut and morselized bone allograft barricade,³³ and that fibrin glue should be used to limit BMP diffusion through the annulotomy site³¹ to prevent this complication.

Our study results suggest that radiculopathy symptoms with pseudo-pedicle-like HO complications appear to be caused by leakage of rhBMP-2 from the disk space through the annulotomy site. This was often initially interpreted incorrectly on MRI in the first year after surgery as being fibrous or granulation tissue, or even postoperative changes that the heterotopic tissue was bone was obvious only on CT. Even then the tissue may be incorrectly identified, as the encasing nerve roots in bone are similar to the scar tissue having no compressive effect. HO may compress, but it also has an inflammatory component that the scars lack. Additionally, the HO from the disk space, caused by leakage of the BMP placed in or around the fusion cage, can create a pseudo-pedicle of varying size and extent. This was present in all 38 of our cases.

This retrospective case series had its limitations. Its clinical and radiographic findings were not blinded. Confounding variables cannot be isolated for causal relationships, if any, to the complication in a case series such as this.

Bone formation with rhBMP-2 is robust and beneficial, but HO-related complications are significant, and identifiable on assessment of radiculopathy symptoms and CT characteristics.

ABSTRACT

Anterior ankle impingement is a frequent cause of pain and disability in athletes with impingement of soft-tissue or osseous structures along the anterior margin of the tibiotalar joint during dorsiflexion.

In this study, we hypothesized that arthroscopic decompression of anterior ankle impingement would result in significant, reliable, and durable improvement in pain and range of motion (ROM), and would allow National Football League (NFL) players to return to their preoperative level of play.

We reviewed 29 arthroscopic ankle débridements performed by a single surgeon. Each NFL player underwent arthroscopic débridement of pathologic soft tissue and of tibial and talar osteophytes in the anterior ankle. Preoperative and postoperative visual analog scale (VAS) pain scores, American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scores, and ankle ROM were compared; time to return to play (RTP), events missed secondary to surgery, and complications were recorded.

All athletes returned to the same level of NFL play at a mean (SD) of 8.4 (4.1) weeks after surgery and continued playing for a mean (SD) of 3.43 (2.57) years after surgery. Mean (SD) VAS pain scores decreased significantly (P < .001), to 0.38 (0.89) from 4.21 (1.52). Mean (SD) active ankle dorsiflexion increased significantly (P < .001), to 18.86° (2.62°) from 8.28° (4.14°). Mean (SD) AOFAS hindfoot scores increased significantly (P < .001), to 97.45 (4.72) from 70.62 (10.39). Degree of arthritis (r = 0.305) and age (r = 0.106) were poorly correlated to time to RTP.

In all cases, arthroscopic débridement of anterior ankle impingement resulted in RTP at the same level at a mean of 2 months after surgery. There were significant improvements in VAS pain scores, AOFAS hindfoot scores, and ROM.

Arthroscopic débridement of anterior ankle impingement relieves pain, restores ROM and function, and results in reliable RTP in professional football players.

Continue to: Anterior ankle impingement...

Anterior ankle impingement is a frequent cause of disability in athletes.¹ This condition results from repetitive trauma over time, which leads to osseous and soft-tissue impingement, pain, and decreased ankle range of motion (ROM).

First termed footballer’s ankle, this condition is linked to repeated, forceful plantarflexion,² though later studies attributed the phenomenon to repeated dorsiflexion resulting in periosteal hemorrhage.³ Both osseous and soft-tissue structures can cause impingement at the tibiotalar joint, often with osteophytes anteromedially at the tibial talar joint. Soft-tissue structures, including hypertrophic synovium, meniscoid lesions, and a thickened anterior talofibular ligament, more often cause anterolateral impingement.^4-6 This process results in pain in extreme dorsiflexion, which comes into play in almost all football maneuvers, including sprinting, back-peddling, and offensive and defensive stances. Therefore, maintenance of pain-free dorsiflexion is required for high-level football. Decreased ROM can lead to decreased ability to perform these high-level athletic functions and can limit performance.

Arthroscopic débridement improves functional outcomes and functional motion in both athletes and nonathletes.^7,8 In addition, findings of a recent systematic review provide support for arthroscopic treatment of ankle impingement.⁹ Although arthroscopic treatment is effective in nonathletes and recreational athletes,¹⁰ there is a paucity of data on the efficacy of this procedure and on time to return to play (RTP) in professional football players.

We conducted a study to evaluate the outcomes (pain, ROM, RTP) of arthroscopic débridement for anterior ankle impingement in National Football League (NFL) players. We hypothesized that arthroscopic decompression of anterior ankle impingement would result in significant, reliable, and durable improvement in pain and ROM, and would allow NFL players to return to their preoperative level of play.

METHODS

After this study was granted Institutional Review Board approval, we retrospectively reviewed a consecutive series of arthroscopically treated anterior ankle impingement athletes by a single surgeon (JPB). Indications for surgery were anterior ankle impingement resulting in ankle pain and decreased ROM that interfered with sport. Active NFL players who underwent ankle arthroscopy for symptomatic anterior ankle impingement were included. Excluded were players who underwent surgery after retirement or who retired before returning to play for reasons unrelated to the ankle. Medical records, operative reports, and rehabilitation reports were reviewed.

Continue to: Preoperative and postoperative...

Preoperative and postoperative visual analog scale (VAS) pain scores, American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scores, and ankle ROM were compared; time to RTP, events missed secondary to surgery, and complications were recorded. These preoperative and postoperative variables were compared with paired Student 2-way t tests for continuous variables. Pearson correlation coefficients were calculated.

PROCEDURE

Ankle arthroscopy was performed with the patient supine after spinal or general anesthesia was induced. Prophylactic antibiotics were given in each case. Arthroscopy was performed with standard anterolateral and anteromedial portals. First, an incision was made through skin only, followed by blunt subcutaneous dissection down to the ankle capsule. A capsulotomy was then made bluntly. Care was taken to avoid all neurovascular structures. Posterior portals were not used. A 2.7-mm arthroscope was inserted and alternated between the anteromedial and anterolateral portals to maximally visualize the ankle joint. Diagnostic arthroscopy was performed to document synovitis, chondral injury, osseous, and soft-tissue impingement and any other noted pathology (Figures 1A-1C).

A full radius resector was then used to perform a synovectomy and débridement of impinging soft tissue from the anterior talofibular ligament or anterior inferior talofibular ligament. All patients underwent arthroscopic débridement of pathologic soft tissue and of tibial and talar osteophytes in the anterior ankle. A small burr was used to débride and remove the osteophytes on the talus and/or tibia. Soft-tissue and osseous structures were resected until the contours of the talus and tibia were normal. Any unstable articular defects were débrided and loose bodies were removed. Ankle ROM was checked to confirm complete resolution of impingement (Figures 2A-2D). Patients were not immobilized and were allowed progressive weight-bearing as tolerated. Crutches were used for assisted ambulation the first 3 to 5 postoperative days.

Physical therapy progressed through 3 phases: (1) inflammation control and ROM restoration, (2) initiation of ankle strengthening, including eversion and inversion, and (3) agility, proprioception, and functional rehabilitation.

RESULTS

Twenty-five NFL players (29 surgeries) were included in the study. Two players were excluded because they had retired at the end of the season before the surgery for reasons unrelated to the operative ankle. Mean (SD) age was 28.1 (2.9) years. Six included players had a history of ankle sprains, 1 had a history of ipsilateral ankle fracture, and 1 had a history of ipsilateral ankle dislocation. Table 1 lists the positions of players who underwent ankle arthroscopic decompression.

Table 1. Positions of National Football League Players Who Underwent Ankle Arthroscopic Decompression for Anterior Ankle Impingement

Continue to: During diagnostic arthroscopy...

During diagnostic arthroscopy, changes to the articular cartilage were noted: grade 0 in 38% of patients, grade 1 in 17%, grade 2 in 21%, grade 3 in 21%, and grade 4 in 3%. Four patients had an osteochondral lesion (<1 cm in each case), which was treated with chondroplasty without microfracture.

Each included patient returned to NFL play. Mean (SD) time to RTP without restrictions was 8.4 (4.1) weeks after surgery (range, 2-20 weeks). There was a poor correlation between degree of chondrosis and time to RTP (r = 0.305). In addition, there was a poor correlation between age and time to RTP (r = 0.106).

Dorsiflexion improved significantly (P < .001), patients had significantly less pain after surgery (P < .001), and AOFAS hindfoot scores improved significantly (P < .001) (Table 2).

Table 2. Preoperative and Postoperative Dorsiflexion, Pain, and AOFAS Score Before and After Arthroscopic Débridement of Anterior Ankle Impingement^a

^aAll values were significantly improved after surgery (P < .001).

Abbreviations: AOFAS, American Orthopaedic Foot and Ankle Society; VAS, visual analog scale.

The athletes played in the NFL for a mean (SD) of 3.43 (2.57) years after surgery (range, 1-10 seasons). These players included 6 who were still active at time of publication. No patient required revision surgery or additional surgery on the ipsilateral ankle. The one patient who was treated for superficial thrombophlebitis after surgery reported symptoms before surgery as well.

DISCUSSION

Arthroscopic decompression of anterior ankle impingement is safe and significantly improves pain and ROM in professional American football players. The procedure results in reliable RTP at an elite level, with durable results over the time remaining in their NFL careers.

Continue to: before the 1988 study by Hawkins...

Before the 1988 study by Hawkins,¹¹ ankle spurs were removed with open procedures. Hawkins¹¹ used arthroscopy for better and safer visualization of the ankle joint and used a burr for less painful removal of spurs from the tibia and the talus. In 2002, a series of 105 patients (median age, 35 years) had reduced pain and improved function a minimum of 2 years after arthroscopic débridement.¹² These patients had a mix of pathology, including soft-tissue impingement, bony impingement, chondral lesions, loose bodies, and osteoarthritis.

For many elite athletes, anterior ankle impingement can cause significant limitation. Reduced ankle dorsiflexion can alter all limb mechanics and predispose athletes to injury.¹³ In addition, because NFL players’ ankle ROM often approaches or exceeds normal physiologic limits,¹⁴ an ankle ROM limitation will often hinder their performance.

Miyamoto and colleagues¹⁵ studied a series of 9 professional athletes (6 soccer players, 1 baseball pitcher, 1 mixed martial artist, 1 golfer) who underwent decompression of both anterior and posterior impingement. With regard to anterior impingement, they found anterior osteophytes in all the ankles, as was seen in the present study. Furthermore, they noted that mean dorsiflexion improved from 10° before surgery to 15° after surgery and that their athletes returned to play 12 to 15 weeks after surgery. Their results are similar to ours, though we noted more improvement in dorsiflexion, from 8.28° before surgery to 18.86° after surgery.

One of the most important metrics in evaluating treatment options for professional athletes is time from surgery to RTP without restrictions. Mean time to full RTP was shorter in our study (8.4 weeks) than in the study by Miyamoto and colleagues¹⁵ (up to 20 weeks). However, many of their procedures were performed during the off-season, when there was no need to expeditiously clear patients for full sports participation. In addition, the patients in their study had both anterior and posterior pathology.

Faster return to high-level athletics was supported in a study of 11 elite ballet dancers,¹⁶ whose pain and dance performance improved after arthroscopic débridement. Of the 11 patients, 9 returned to dance at a mean of 7 weeks after surgery; the other 2 required reoperation. Although the pathology differed in their study of elite professional soccer players, Calder and colleagues¹⁷ found that mean time to RTP after ankle arthroscopy for posterior impingement was 5 weeks.

Continue to: For the NFL players in our study...

For the NFL players in our study, RTP at their elite level was 100% after arthroscopic débridement of anterior ankle impingement. In the literature, time to RTP varies. Table 3 lists RTP rates for recreational athletes in published studies.^18-27 In their recent systematic literature review, Zwiers and colleagues¹⁰ noted that 24% to 96.4% of recreational athletes returned to play after arthroscopic treatment for anterior ankle impingement. The percentage was significantly higher for the professional athletes in our study. Historical comparison supports an evolution in the indications and techniques for this procedure, with more recent literature suggesting a RTP rate much higher than earlier rates. In addition, compared with recreational athletes, professional athletes have strong financial incentives to return to their sports. Furthermore, our professional cohort was significantly younger than the recreational cohorts in those studies.

Table 3. Frequency of Recreational Athletes’ Return to Play After Arthroscopic Débridement of Anterior Ankle Impingement, as Reported in the Literature

Current recommendations for recreational athletes include initial conservative treatment with rest, ankle bracing, and avoidance of jumping and other repetitive dorsiflexing activities. Physical therapy should include joint mobilization and work along the entire kinetic chain. Night splints or a removable walking boot can be used temporarily, as can a single intra-articular corticosteroid injection to reduce inflammation and evaluate improvement in more refractory cases.²⁸ Commonly, conservative treatments fail if patients remain active, and soft tissue and/or osteophytes can be resected, though resection typically is reserved for recreational athletes for whom nonoperative treatments have been exhausted.^29,30

This study had several limitations, including its retrospective nature and lack of control group. In addition, follow-up was relatively short, and we did not use more recently described outcome measures, such as the Sports subscale of the Foot and Ankle Ability Measure, which may be more sensitive in describing function in elite athletes. However, many of the cases in our study predated these measures, but the rate of RTP at the NFL level requires a very high degree of postoperative ankle function, making this outcome the most meaningful. In the context of professional athletes, specifically the length of their careers, our study results provide valuable information regarding expectations about RTP and the durability of arthroscopic débridement of anterior ankle impingement in a high-demand setting.

CONCLUSION

For all the NFL players in this study, arthroscopic débridement of anterior ankle impingement resulted in return to preoperative level of play at a mean of 2 months after surgery. There were significant improvements in VAS pain scores, AOFAS hindfoot scores, and ROM. Arthroscopic débridement of anterior ankle impingement relieves pain, restores ROM and function, and results in reliable RTP in professional football players.

ABSTRACT

Anterior ankle impingement is a frequent cause of pain and disability in athletes with impingement of soft-tissue or osseous structures along the anterior margin of the tibiotalar joint during dorsiflexion.

In this study, we hypothesized that arthroscopic decompression of anterior ankle impingement would result in significant, reliable, and durable improvement in pain and range of motion (ROM), and would allow National Football League (NFL) players to return to their preoperative level of play.

We reviewed 29 arthroscopic ankle débridements performed by a single surgeon. Each NFL player underwent arthroscopic débridement of pathologic soft tissue and of tibial and talar osteophytes in the anterior ankle. Preoperative and postoperative visual analog scale (VAS) pain scores, American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scores, and ankle ROM were compared; time to return to play (RTP), events missed secondary to surgery, and complications were recorded.

All athletes returned to the same level of NFL play at a mean (SD) of 8.4 (4.1) weeks after surgery and continued playing for a mean (SD) of 3.43 (2.57) years after surgery. Mean (SD) VAS pain scores decreased significantly (P < .001), to 0.38 (0.89) from 4.21 (1.52). Mean (SD) active ankle dorsiflexion increased significantly (P < .001), to 18.86° (2.62°) from 8.28° (4.14°). Mean (SD) AOFAS hindfoot scores increased significantly (P < .001), to 97.45 (4.72) from 70.62 (10.39). Degree of arthritis (r = 0.305) and age (r = 0.106) were poorly correlated to time to RTP.

In all cases, arthroscopic débridement of anterior ankle impingement resulted in RTP at the same level at a mean of 2 months after surgery. There were significant improvements in VAS pain scores, AOFAS hindfoot scores, and ROM.

Arthroscopic débridement of anterior ankle impingement relieves pain, restores ROM and function, and results in reliable RTP in professional football players.

Continue to: Anterior ankle impingement...

Anterior ankle impingement is a frequent cause of disability in athletes.¹ This condition results from repetitive trauma over time, which leads to osseous and soft-tissue impingement, pain, and decreased ankle range of motion (ROM).

First termed footballer’s ankle, this condition is linked to repeated, forceful plantarflexion,² though later studies attributed the phenomenon to repeated dorsiflexion resulting in periosteal hemorrhage.³ Both osseous and soft-tissue structures can cause impingement at the tibiotalar joint, often with osteophytes anteromedially at the tibial talar joint. Soft-tissue structures, including hypertrophic synovium, meniscoid lesions, and a thickened anterior talofibular ligament, more often cause anterolateral impingement.^4-6 This process results in pain in extreme dorsiflexion, which comes into play in almost all football maneuvers, including sprinting, back-peddling, and offensive and defensive stances. Therefore, maintenance of pain-free dorsiflexion is required for high-level football. Decreased ROM can lead to decreased ability to perform these high-level athletic functions and can limit performance.

Arthroscopic débridement improves functional outcomes and functional motion in both athletes and nonathletes.^7,8 In addition, findings of a recent systematic review provide support for arthroscopic treatment of ankle impingement.⁹ Although arthroscopic treatment is effective in nonathletes and recreational athletes,¹⁰ there is a paucity of data on the efficacy of this procedure and on time to return to play (RTP) in professional football players.

We conducted a study to evaluate the outcomes (pain, ROM, RTP) of arthroscopic débridement for anterior ankle impingement in National Football League (NFL) players. We hypothesized that arthroscopic decompression of anterior ankle impingement would result in significant, reliable, and durable improvement in pain and ROM, and would allow NFL players to return to their preoperative level of play.

METHODS

After this study was granted Institutional Review Board approval, we retrospectively reviewed a consecutive series of arthroscopically treated anterior ankle impingement athletes by a single surgeon (JPB). Indications for surgery were anterior ankle impingement resulting in ankle pain and decreased ROM that interfered with sport. Active NFL players who underwent ankle arthroscopy for symptomatic anterior ankle impingement were included. Excluded were players who underwent surgery after retirement or who retired before returning to play for reasons unrelated to the ankle. Medical records, operative reports, and rehabilitation reports were reviewed.

Continue to: Preoperative and postoperative...

Preoperative and postoperative visual analog scale (VAS) pain scores, American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scores, and ankle ROM were compared; time to RTP, events missed secondary to surgery, and complications were recorded. These preoperative and postoperative variables were compared with paired Student 2-way t tests for continuous variables. Pearson correlation coefficients were calculated.

PROCEDURE

Ankle arthroscopy was performed with the patient supine after spinal or general anesthesia was induced. Prophylactic antibiotics were given in each case. Arthroscopy was performed with standard anterolateral and anteromedial portals. First, an incision was made through skin only, followed by blunt subcutaneous dissection down to the ankle capsule. A capsulotomy was then made bluntly. Care was taken to avoid all neurovascular structures. Posterior portals were not used. A 2.7-mm arthroscope was inserted and alternated between the anteromedial and anterolateral portals to maximally visualize the ankle joint. Diagnostic arthroscopy was performed to document synovitis, chondral injury, osseous, and soft-tissue impingement and any other noted pathology (Figures 1A-1C).

A full radius resector was then used to perform a synovectomy and débridement of impinging soft tissue from the anterior talofibular ligament or anterior inferior talofibular ligament. All patients underwent arthroscopic débridement of pathologic soft tissue and of tibial and talar osteophytes in the anterior ankle. A small burr was used to débride and remove the osteophytes on the talus and/or tibia. Soft-tissue and osseous structures were resected until the contours of the talus and tibia were normal. Any unstable articular defects were débrided and loose bodies were removed. Ankle ROM was checked to confirm complete resolution of impingement (Figures 2A-2D). Patients were not immobilized and were allowed progressive weight-bearing as tolerated. Crutches were used for assisted ambulation the first 3 to 5 postoperative days.

Physical therapy progressed through 3 phases: (1) inflammation control and ROM restoration, (2) initiation of ankle strengthening, including eversion and inversion, and (3) agility, proprioception, and functional rehabilitation.

RESULTS

Twenty-five NFL players (29 surgeries) were included in the study. Two players were excluded because they had retired at the end of the season before the surgery for reasons unrelated to the operative ankle. Mean (SD) age was 28.1 (2.9) years. Six included players had a history of ankle sprains, 1 had a history of ipsilateral ankle fracture, and 1 had a history of ipsilateral ankle dislocation. Table 1 lists the positions of players who underwent ankle arthroscopic decompression.

Table 1. Positions of National Football League Players Who Underwent Ankle Arthroscopic Decompression for Anterior Ankle Impingement

Continue to: During diagnostic arthroscopy...

During diagnostic arthroscopy, changes to the articular cartilage were noted: grade 0 in 38% of patients, grade 1 in 17%, grade 2 in 21%, grade 3 in 21%, and grade 4 in 3%. Four patients had an osteochondral lesion (<1 cm in each case), which was treated with chondroplasty without microfracture.

Each included patient returned to NFL play. Mean (SD) time to RTP without restrictions was 8.4 (4.1) weeks after surgery (range, 2-20 weeks). There was a poor correlation between degree of chondrosis and time to RTP (r = 0.305). In addition, there was a poor correlation between age and time to RTP (r = 0.106).

Dorsiflexion improved significantly (P < .001), patients had significantly less pain after surgery (P < .001), and AOFAS hindfoot scores improved significantly (P < .001) (Table 2).

Table 2. Preoperative and Postoperative Dorsiflexion, Pain, and AOFAS Score Before and After Arthroscopic Débridement of Anterior Ankle Impingement^a

^aAll values were significantly improved after surgery (P < .001).

Abbreviations: AOFAS, American Orthopaedic Foot and Ankle Society; VAS, visual analog scale.

The athletes played in the NFL for a mean (SD) of 3.43 (2.57) years after surgery (range, 1-10 seasons). These players included 6 who were still active at time of publication. No patient required revision surgery or additional surgery on the ipsilateral ankle. The one patient who was treated for superficial thrombophlebitis after surgery reported symptoms before surgery as well.

DISCUSSION

Arthroscopic decompression of anterior ankle impingement is safe and significantly improves pain and ROM in professional American football players. The procedure results in reliable RTP at an elite level, with durable results over the time remaining in their NFL careers.

Continue to: before the 1988 study by Hawkins...

Before the 1988 study by Hawkins,¹¹ ankle spurs were removed with open procedures. Hawkins¹¹ used arthroscopy for better and safer visualization of the ankle joint and used a burr for less painful removal of spurs from the tibia and the talus. In 2002, a series of 105 patients (median age, 35 years) had reduced pain and improved function a minimum of 2 years after arthroscopic débridement.¹² These patients had a mix of pathology, including soft-tissue impingement, bony impingement, chondral lesions, loose bodies, and osteoarthritis.

For many elite athletes, anterior ankle impingement can cause significant limitation. Reduced ankle dorsiflexion can alter all limb mechanics and predispose athletes to injury.¹³ In addition, because NFL players’ ankle ROM often approaches or exceeds normal physiologic limits,¹⁴ an ankle ROM limitation will often hinder their performance.

Miyamoto and colleagues¹⁵ studied a series of 9 professional athletes (6 soccer players, 1 baseball pitcher, 1 mixed martial artist, 1 golfer) who underwent decompression of both anterior and posterior impingement. With regard to anterior impingement, they found anterior osteophytes in all the ankles, as was seen in the present study. Furthermore, they noted that mean dorsiflexion improved from 10° before surgery to 15° after surgery and that their athletes returned to play 12 to 15 weeks after surgery. Their results are similar to ours, though we noted more improvement in dorsiflexion, from 8.28° before surgery to 18.86° after surgery.

One of the most important metrics in evaluating treatment options for professional athletes is time from surgery to RTP without restrictions. Mean time to full RTP was shorter in our study (8.4 weeks) than in the study by Miyamoto and colleagues¹⁵ (up to 20 weeks). However, many of their procedures were performed during the off-season, when there was no need to expeditiously clear patients for full sports participation. In addition, the patients in their study had both anterior and posterior pathology.

Faster return to high-level athletics was supported in a study of 11 elite ballet dancers,¹⁶ whose pain and dance performance improved after arthroscopic débridement. Of the 11 patients, 9 returned to dance at a mean of 7 weeks after surgery; the other 2 required reoperation. Although the pathology differed in their study of elite professional soccer players, Calder and colleagues¹⁷ found that mean time to RTP after ankle arthroscopy for posterior impingement was 5 weeks.

Continue to: For the NFL players in our study...

For the NFL players in our study, RTP at their elite level was 100% after arthroscopic débridement of anterior ankle impingement. In the literature, time to RTP varies. Table 3 lists RTP rates for recreational athletes in published studies.^18-27 In their recent systematic literature review, Zwiers and colleagues¹⁰ noted that 24% to 96.4% of recreational athletes returned to play after arthroscopic treatment for anterior ankle impingement. The percentage was significantly higher for the professional athletes in our study. Historical comparison supports an evolution in the indications and techniques for this procedure, with more recent literature suggesting a RTP rate much higher than earlier rates. In addition, compared with recreational athletes, professional athletes have strong financial incentives to return to their sports. Furthermore, our professional cohort was significantly younger than the recreational cohorts in those studies.

Table 3. Frequency of Recreational Athletes’ Return to Play After Arthroscopic Débridement of Anterior Ankle Impingement, as Reported in the Literature

Current recommendations for recreational athletes include initial conservative treatment with rest, ankle bracing, and avoidance of jumping and other repetitive dorsiflexing activities. Physical therapy should include joint mobilization and work along the entire kinetic chain. Night splints or a removable walking boot can be used temporarily, as can a single intra-articular corticosteroid injection to reduce inflammation and evaluate improvement in more refractory cases.²⁸ Commonly, conservative treatments fail if patients remain active, and soft tissue and/or osteophytes can be resected, though resection typically is reserved for recreational athletes for whom nonoperative treatments have been exhausted.^29,30

This study had several limitations, including its retrospective nature and lack of control group. In addition, follow-up was relatively short, and we did not use more recently described outcome measures, such as the Sports subscale of the Foot and Ankle Ability Measure, which may be more sensitive in describing function in elite athletes. However, many of the cases in our study predated these measures, but the rate of RTP at the NFL level requires a very high degree of postoperative ankle function, making this outcome the most meaningful. In the context of professional athletes, specifically the length of their careers, our study results provide valuable information regarding expectations about RTP and the durability of arthroscopic débridement of anterior ankle impingement in a high-demand setting.

CONCLUSION

For all the NFL players in this study, arthroscopic débridement of anterior ankle impingement resulted in return to preoperative level of play at a mean of 2 months after surgery. There were significant improvements in VAS pain scores, AOFAS hindfoot scores, and ROM. Arthroscopic débridement of anterior ankle impingement relieves pain, restores ROM and function, and results in reliable RTP in professional football players.

ABSTRACT

Anterior ankle impingement is a frequent cause of pain and disability in athletes with impingement of soft-tissue or osseous structures along the anterior margin of the tibiotalar joint during dorsiflexion.

In this study, we hypothesized that arthroscopic decompression of anterior ankle impingement would result in significant, reliable, and durable improvement in pain and range of motion (ROM), and would allow National Football League (NFL) players to return to their preoperative level of play.

We reviewed 29 arthroscopic ankle débridements performed by a single surgeon. Each NFL player underwent arthroscopic débridement of pathologic soft tissue and of tibial and talar osteophytes in the anterior ankle. Preoperative and postoperative visual analog scale (VAS) pain scores, American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scores, and ankle ROM were compared; time to return to play (RTP), events missed secondary to surgery, and complications were recorded.

All athletes returned to the same level of NFL play at a mean (SD) of 8.4 (4.1) weeks after surgery and continued playing for a mean (SD) of 3.43 (2.57) years after surgery. Mean (SD) VAS pain scores decreased significantly (P < .001), to 0.38 (0.89) from 4.21 (1.52). Mean (SD) active ankle dorsiflexion increased significantly (P < .001), to 18.86° (2.62°) from 8.28° (4.14°). Mean (SD) AOFAS hindfoot scores increased significantly (P < .001), to 97.45 (4.72) from 70.62 (10.39). Degree of arthritis (r = 0.305) and age (r = 0.106) were poorly correlated to time to RTP.

In all cases, arthroscopic débridement of anterior ankle impingement resulted in RTP at the same level at a mean of 2 months after surgery. There were significant improvements in VAS pain scores, AOFAS hindfoot scores, and ROM.

Arthroscopic débridement of anterior ankle impingement relieves pain, restores ROM and function, and results in reliable RTP in professional football players.

Continue to: Anterior ankle impingement...

Anterior ankle impingement is a frequent cause of disability in athletes.¹ This condition results from repetitive trauma over time, which leads to osseous and soft-tissue impingement, pain, and decreased ankle range of motion (ROM).

First termed footballer’s ankle, this condition is linked to repeated, forceful plantarflexion,² though later studies attributed the phenomenon to repeated dorsiflexion resulting in periosteal hemorrhage.³ Both osseous and soft-tissue structures can cause impingement at the tibiotalar joint, often with osteophytes anteromedially at the tibial talar joint. Soft-tissue structures, including hypertrophic synovium, meniscoid lesions, and a thickened anterior talofibular ligament, more often cause anterolateral impingement.^4-6 This process results in pain in extreme dorsiflexion, which comes into play in almost all football maneuvers, including sprinting, back-peddling, and offensive and defensive stances. Therefore, maintenance of pain-free dorsiflexion is required for high-level football. Decreased ROM can lead to decreased ability to perform these high-level athletic functions and can limit performance.

Arthroscopic débridement improves functional outcomes and functional motion in both athletes and nonathletes.^7,8 In addition, findings of a recent systematic review provide support for arthroscopic treatment of ankle impingement.⁹ Although arthroscopic treatment is effective in nonathletes and recreational athletes,¹⁰ there is a paucity of data on the efficacy of this procedure and on time to return to play (RTP) in professional football players.

We conducted a study to evaluate the outcomes (pain, ROM, RTP) of arthroscopic débridement for anterior ankle impingement in National Football League (NFL) players. We hypothesized that arthroscopic decompression of anterior ankle impingement would result in significant, reliable, and durable improvement in pain and ROM, and would allow NFL players to return to their preoperative level of play.

METHODS

After this study was granted Institutional Review Board approval, we retrospectively reviewed a consecutive series of arthroscopically treated anterior ankle impingement athletes by a single surgeon (JPB). Indications for surgery were anterior ankle impingement resulting in ankle pain and decreased ROM that interfered with sport. Active NFL players who underwent ankle arthroscopy for symptomatic anterior ankle impingement were included. Excluded were players who underwent surgery after retirement or who retired before returning to play for reasons unrelated to the ankle. Medical records, operative reports, and rehabilitation reports were reviewed.

Continue to: Preoperative and postoperative...

Preoperative and postoperative visual analog scale (VAS) pain scores, American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scores, and ankle ROM were compared; time to RTP, events missed secondary to surgery, and complications were recorded. These preoperative and postoperative variables were compared with paired Student 2-way t tests for continuous variables. Pearson correlation coefficients were calculated.

PROCEDURE

Ankle arthroscopy was performed with the patient supine after spinal or general anesthesia was induced. Prophylactic antibiotics were given in each case. Arthroscopy was performed with standard anterolateral and anteromedial portals. First, an incision was made through skin only, followed by blunt subcutaneous dissection down to the ankle capsule. A capsulotomy was then made bluntly. Care was taken to avoid all neurovascular structures. Posterior portals were not used. A 2.7-mm arthroscope was inserted and alternated between the anteromedial and anterolateral portals to maximally visualize the ankle joint. Diagnostic arthroscopy was performed to document synovitis, chondral injury, osseous, and soft-tissue impingement and any other noted pathology (Figures 1A-1C).

A full radius resector was then used to perform a synovectomy and débridement of impinging soft tissue from the anterior talofibular ligament or anterior inferior talofibular ligament. All patients underwent arthroscopic débridement of pathologic soft tissue and of tibial and talar osteophytes in the anterior ankle. A small burr was used to débride and remove the osteophytes on the talus and/or tibia. Soft-tissue and osseous structures were resected until the contours of the talus and tibia were normal. Any unstable articular defects were débrided and loose bodies were removed. Ankle ROM was checked to confirm complete resolution of impingement (Figures 2A-2D). Patients were not immobilized and were allowed progressive weight-bearing as tolerated. Crutches were used for assisted ambulation the first 3 to 5 postoperative days.

Physical therapy progressed through 3 phases: (1) inflammation control and ROM restoration, (2) initiation of ankle strengthening, including eversion and inversion, and (3) agility, proprioception, and functional rehabilitation.

RESULTS

Twenty-five NFL players (29 surgeries) were included in the study. Two players were excluded because they had retired at the end of the season before the surgery for reasons unrelated to the operative ankle. Mean (SD) age was 28.1 (2.9) years. Six included players had a history of ankle sprains, 1 had a history of ipsilateral ankle fracture, and 1 had a history of ipsilateral ankle dislocation. Table 1 lists the positions of players who underwent ankle arthroscopic decompression.

Table 1. Positions of National Football League Players Who Underwent Ankle Arthroscopic Decompression for Anterior Ankle Impingement

Continue to: During diagnostic arthroscopy...

During diagnostic arthroscopy, changes to the articular cartilage were noted: grade 0 in 38% of patients, grade 1 in 17%, grade 2 in 21%, grade 3 in 21%, and grade 4 in 3%. Four patients had an osteochondral lesion (<1 cm in each case), which was treated with chondroplasty without microfracture.

Each included patient returned to NFL play. Mean (SD) time to RTP without restrictions was 8.4 (4.1) weeks after surgery (range, 2-20 weeks). There was a poor correlation between degree of chondrosis and time to RTP (r = 0.305). In addition, there was a poor correlation between age and time to RTP (r = 0.106).

Dorsiflexion improved significantly (P < .001), patients had significantly less pain after surgery (P < .001), and AOFAS hindfoot scores improved significantly (P < .001) (Table 2).

Table 2. Preoperative and Postoperative Dorsiflexion, Pain, and AOFAS Score Before and After Arthroscopic Débridement of Anterior Ankle Impingement^a

^aAll values were significantly improved after surgery (P < .001).

Abbreviations: AOFAS, American Orthopaedic Foot and Ankle Society; VAS, visual analog scale.

The athletes played in the NFL for a mean (SD) of 3.43 (2.57) years after surgery (range, 1-10 seasons). These players included 6 who were still active at time of publication. No patient required revision surgery or additional surgery on the ipsilateral ankle. The one patient who was treated for superficial thrombophlebitis after surgery reported symptoms before surgery as well.

DISCUSSION

Arthroscopic decompression of anterior ankle impingement is safe and significantly improves pain and ROM in professional American football players. The procedure results in reliable RTP at an elite level, with durable results over the time remaining in their NFL careers.

Continue to: before the 1988 study by Hawkins...

Before the 1988 study by Hawkins,¹¹ ankle spurs were removed with open procedures. Hawkins¹¹ used arthroscopy for better and safer visualization of the ankle joint and used a burr for less painful removal of spurs from the tibia and the talus. In 2002, a series of 105 patients (median age, 35 years) had reduced pain and improved function a minimum of 2 years after arthroscopic débridement.¹² These patients had a mix of pathology, including soft-tissue impingement, bony impingement, chondral lesions, loose bodies, and osteoarthritis.

For many elite athletes, anterior ankle impingement can cause significant limitation. Reduced ankle dorsiflexion can alter all limb mechanics and predispose athletes to injury.¹³ In addition, because NFL players’ ankle ROM often approaches or exceeds normal physiologic limits,¹⁴ an ankle ROM limitation will often hinder their performance.

Miyamoto and colleagues¹⁵ studied a series of 9 professional athletes (6 soccer players, 1 baseball pitcher, 1 mixed martial artist, 1 golfer) who underwent decompression of both anterior and posterior impingement. With regard to anterior impingement, they found anterior osteophytes in all the ankles, as was seen in the present study. Furthermore, they noted that mean dorsiflexion improved from 10° before surgery to 15° after surgery and that their athletes returned to play 12 to 15 weeks after surgery. Their results are similar to ours, though we noted more improvement in dorsiflexion, from 8.28° before surgery to 18.86° after surgery.

One of the most important metrics in evaluating treatment options for professional athletes is time from surgery to RTP without restrictions. Mean time to full RTP was shorter in our study (8.4 weeks) than in the study by Miyamoto and colleagues¹⁵ (up to 20 weeks). However, many of their procedures were performed during the off-season, when there was no need to expeditiously clear patients for full sports participation. In addition, the patients in their study had both anterior and posterior pathology.

Faster return to high-level athletics was supported in a study of 11 elite ballet dancers,¹⁶ whose pain and dance performance improved after arthroscopic débridement. Of the 11 patients, 9 returned to dance at a mean of 7 weeks after surgery; the other 2 required reoperation. Although the pathology differed in their study of elite professional soccer players, Calder and colleagues¹⁷ found that mean time to RTP after ankle arthroscopy for posterior impingement was 5 weeks.

Continue to: For the NFL players in our study...

For the NFL players in our study, RTP at their elite level was 100% after arthroscopic débridement of anterior ankle impingement. In the literature, time to RTP varies. Table 3 lists RTP rates for recreational athletes in published studies.^18-27 In their recent systematic literature review, Zwiers and colleagues¹⁰ noted that 24% to 96.4% of recreational athletes returned to play after arthroscopic treatment for anterior ankle impingement. The percentage was significantly higher for the professional athletes in our study. Historical comparison supports an evolution in the indications and techniques for this procedure, with more recent literature suggesting a RTP rate much higher than earlier rates. In addition, compared with recreational athletes, professional athletes have strong financial incentives to return to their sports. Furthermore, our professional cohort was significantly younger than the recreational cohorts in those studies.

Table 3. Frequency of Recreational Athletes’ Return to Play After Arthroscopic Débridement of Anterior Ankle Impingement, as Reported in the Literature

Current recommendations for recreational athletes include initial conservative treatment with rest, ankle bracing, and avoidance of jumping and other repetitive dorsiflexing activities. Physical therapy should include joint mobilization and work along the entire kinetic chain. Night splints or a removable walking boot can be used temporarily, as can a single intra-articular corticosteroid injection to reduce inflammation and evaluate improvement in more refractory cases.²⁸ Commonly, conservative treatments fail if patients remain active, and soft tissue and/or osteophytes can be resected, though resection typically is reserved for recreational athletes for whom nonoperative treatments have been exhausted.^29,30

This study had several limitations, including its retrospective nature and lack of control group. In addition, follow-up was relatively short, and we did not use more recently described outcome measures, such as the Sports subscale of the Foot and Ankle Ability Measure, which may be more sensitive in describing function in elite athletes. However, many of the cases in our study predated these measures, but the rate of RTP at the NFL level requires a very high degree of postoperative ankle function, making this outcome the most meaningful. In the context of professional athletes, specifically the length of their careers, our study results provide valuable information regarding expectations about RTP and the durability of arthroscopic débridement of anterior ankle impingement in a high-demand setting.

CONCLUSION

For all the NFL players in this study, arthroscopic débridement of anterior ankle impingement resulted in return to preoperative level of play at a mean of 2 months after surgery. There were significant improvements in VAS pain scores, AOFAS hindfoot scores, and ROM. Arthroscopic débridement of anterior ankle impingement relieves pain, restores ROM and function, and results in reliable RTP in professional football players.

ABSTRACT

Quadriceps tendon ruptures disrupt the extensor mechanism of the knee and require urgent surgical management. Traditional repair techniques have had mixed biomechanical and clinical results risking weakness and extensor lag. We describe a novel technique using tape suture and knotless anchors, which has performed superiorly during biomechanical testing and yielded terrific early clinical results. 

Continue to: Quadriceps tendon rupture...

Quadriceps tendon rupture is an uncommon yet potentially devastating knee injury with an estimated incidence of 1.37 in 100,000.¹ It most often occurs in male, middle-aged or older patients with degenerative tendon changes and serious systemic diseases, such as chronic renal failure, diabetes mellitus, rheumatoid arthritis, and disorders requiring long-term steroid use (tissue quality is often compromised by patient age and comorbidities).^2-10 Whereas partial tears with an intact extensor mechanism may be managed nonoperatively, prompt operative intervention is indicated in cases of complete tear or an incompetent extensor mechanism to facilitate early range of motion (ROM) and return of knee function.^2-4,8,9

The standard of care is repair with a nonabsorbable suture passed through transosseous patellar tunnels, often with several weeks of postoperative immobilization to protect the repair.^3,4,7,10-12 Reported complications of this method include significant extension lag, decreased strength, and ROM compared with the contralateral knee, chronic pain, and iatrogenic patellar fracture.^8,13-18 Repair techniques using suture anchors have been proposed as viable alternatives, but biomechanical studies comparing them with standard transosseous repair have reported mixed results.^{7,10-12,18-20} Two studies found improved biomechanical characteristics with suture anchors,^10,21 but 2 others found the characteristics of suture anchor fixation equal to¹¹ or worse than¹² those of transosseous fixation. In light of the controversy regarding strength and clinical outcomes of suture anchor repair compared with transosseous repair, new and potentially superior surgical interventions should be considered.

We recently completed a cadaveric study comparing the biomechanical properties of a novel quadriceps tendon repair technique using 4.75-mm biocomposite knotless suture anchors with suture tape and the properties of conventional techniques using either transosseous or suture anchor repair alone.²² In the cadaveric model, compared with transosseous and fully threaded suture anchor techniques, repair of quadriceps tendon ruptures with this knotless suture anchor with suture tape technique was biomechanically superior in cyclic displacement, construct stiffness, and ultimate load to failure.²² Additionally, this method allows for less extensive dissection, shorter operative times, and the potential for earlier and more aggressive rehabilitation protocols.²² We propose this technique, presented in this article, as a superior alternative to traditional quadriceps tendon repair techniques.

TECHNIQUE

The patient is placed in supine position with a tourniquet placed on the proximal thigh. A midline incision is made from the proximal pole of the patella, proximally by 5 cm. A combination of sharp and blunt dissection is performed through skin and subcutaneous tissues down to the extensor mechanism, exposing the proximal pole of the patella and the torn quadriceps tendon.

The distal aspect of the quadriceps tendon is then débrided of any devitalized tissue and secured with an Allis clamp. A long tape suture (FiberTape; Arthrex) is then used to place a locking Krackow stitch in a distal-to-proximal and then proximal-to-distal direction for 5 throws in each direction within the quadriceps tendon, with the tails exiting distally at the tear site. Care is taken with each pass to ensure that there is no slack within the system.

Continue to: The proximal pole of the patella...

The proximal pole of the patella is then prepared by débriding any remaining soft tissue back to an area of exposed subcortical bone, which is débrided to a bleeding bony bed. Holes are drilled in the medial and lateral thirds of the patella at the proximal pole using the drill for 4.75-mm biocomposite knotless suture anchors (SwiveLock; Arthrex). The tap for the 4.75-mm anchors is then passed at each guide hole. In hard bone, double-tapping is recommended.

Next, the medial strand of tape suture is loaded within a 4.75-mm biocomposite knotless suture anchor eyelet and reduced to the patella. The medial anchor is malleted and screwed into place, while tension is kept on the lateral strand with the knee in full extension. The lateral strand is then placed into its 4.75-mm biocomposite knotless suture anchor, reduced to the patella, and then malleted and screwed into place in the lateral hole, thereby completing the core portion of the repair (Figures A-D). The core strands from the 4.75-mm biocomposite knotless suture anchors are then back-passed in mattress fashion and tied, and medial and lateral retinacular repairs are then performed using supersuture tape (SutureTape or FiberWire; Arthrex).

After surgery, the patient is placed in a knee brace locked in full extension and allowed to weight-bear as tolerated using crutches. During the first week, knee ROM is allowed up to 30°. During weeks 2 to 6 passive ROM is gradually increased to 90°, and use of crutches is tapered. At week 6 the brace is unlocked for ambulation; it may be discontinued after 7 to 8 weeks or when determined safe. Light activity is permitted from month 4 to month 6. A patient who achieves satisfactory strength, is clinically examined, and progresses through rehabilitation is allowed to return to fully unrestricted sport.

DISCUSSION

Quadriceps tendon rupture is an uncommon clinical entity that requires early surgical management.^1-5,12,17,19 The standard of care is passage of nonabsorbable sutures through transosseous patellar bone tunnels, but repair with suture anchors has been studied as an alternative that allows for less tissue trauma, decreased operative time, safe early initiation of rehabilitation protocols, and reduced risk of patella fracture or damage.^{3,7,10-12,18-20,21,23} Despite these potential advantages, biomechanical studies have yielded inconsistent results regarding the superiority of suture anchor repair over repair with transosseous tunnels.^{7,10-12,18-20} We propose quadriceps tendon repair using the 4.75-mm biocomposite knotless suture anchor with tape suture technique as a biomechanically superior alternative to either transosseous tunnels or suture anchor repair alone, with significant advantages both in and out of the operating room.

Results of biomechanical studies comparing transosseous tunnel repair and standard suture anchor repair have been mixed, though the heterogeneity of their study methods and endpoints makes direct comparisons difficult.^{7,10-12,18-20} Petri and colleagues¹⁰ and Sherman and colleagues²¹ reported statistically significant higher load to failure¹⁰ and reduced gapping during cyclic loading^10,21 with suture anchor repair relative to transosseous repair. However, Hart and colleagues¹² found that repair with suture anchors had lower ultimate tensile load, and they concluded that transosseous repair is superior. Lighthart and colleagues¹¹ found no significant difference in displacement between the 2 repairs.

Continue to: In our cadaveric biomechanical study...

In our cadaveric biomechanical study, a novel 4.75-mm biocomposite knotless suture anchor with suture tape repair was compared with traditional 3-tunnel transosseous repair and with standard 2-anchor suture anchor repair.²² Statistically significant superiority was found across multiple parameters, including initial tendon displacement, stiffness, and ultimate load to failure (vs 5.5-mm biocomposite fully threaded suture anchor repair), as well as initial and late tendon displacement, stiffness, and ultimate load to failure (vs transosseous repair).²² Although definitive conclusions are difficult to draw on the basis of prior cadaveric studies comparing standard suture anchor repair and transosseous repair, our results decidedly favor the biomechanical characteristics of this 4.75-mm biocomposite knotless suture anchor with suture tape repair and make it a potentially superior repair technique based on biomechanics alone.²²

Similarly to standard repair with suture anchors, repair using a 4.75-mm biocomposite knotless suture anchor with tape suture eliminates the need to expose the distal pole of the patella.^7,10-12,21 This allows for a smaller surgical incision, less extensive dissection, and prevents possible interference with the patellar tendon.^7,10-12,21 Additionally, it eliminates the risk of iatrogenic patellar fracture and damage to the articular surface from drilling the transpatellar tunnels.^17,18 Both our own review of cases repaired with our 4.75-mm biocomposite knotless suture anchor with suture tape technique as well as studies of suture anchor repair have consistently found operative times of <1 hour.²¹ Shorter operative times and smaller surgical wounds are advantageous given that many of these patients have medical comorbidities that predispose them to intraoperative and wound-healing complications.^12,19-22

Optimal rehabilitation protocols for quadriceps tendon repair are a matter of controversy. Multiple studies of repair with transosseous patellar tunnels describe immobilization for 6 weeks after surgery, but there has been a recent push toward early motion.^7,13,23,24 Reported complications of extended immobilization include limited flexion, pain, weakness, decreased patellar mobility, and patella baja.¹⁴ Studies have suggested that, while excessive loading can cause gap formation and weaken the repair, some controlled motion is necessary to heal the tendon^23,25 and reduce the risks of stiffness and atrophy.¹⁴ The improved biomechanical characteristics of the 4.75-mm biocomposite knotless suture anchor with tape suture technique allow for safe early initiation of ROM exercises and accelerated rehabilitation protocols.

In our early experience with this technique, functional outcomes have been excellent. A formal 2-year outcome study of patients who have undergone quadriceps tendon repair with this 4.75-mm biocomposite knotless suture anchor with tape suture technique is under way.

ABSTRACT

Quadriceps tendon ruptures disrupt the extensor mechanism of the knee and require urgent surgical management. Traditional repair techniques have had mixed biomechanical and clinical results risking weakness and extensor lag. We describe a novel technique using tape suture and knotless anchors, which has performed superiorly during biomechanical testing and yielded terrific early clinical results. 

Continue to: Quadriceps tendon rupture...

Quadriceps tendon rupture is an uncommon yet potentially devastating knee injury with an estimated incidence of 1.37 in 100,000.¹ It most often occurs in male, middle-aged or older patients with degenerative tendon changes and serious systemic diseases, such as chronic renal failure, diabetes mellitus, rheumatoid arthritis, and disorders requiring long-term steroid use (tissue quality is often compromised by patient age and comorbidities).^2-10 Whereas partial tears with an intact extensor mechanism may be managed nonoperatively, prompt operative intervention is indicated in cases of complete tear or an incompetent extensor mechanism to facilitate early range of motion (ROM) and return of knee function.^2-4,8,9

The standard of care is repair with a nonabsorbable suture passed through transosseous patellar tunnels, often with several weeks of postoperative immobilization to protect the repair.^3,4,7,10-12 Reported complications of this method include significant extension lag, decreased strength, and ROM compared with the contralateral knee, chronic pain, and iatrogenic patellar fracture.^8,13-18 Repair techniques using suture anchors have been proposed as viable alternatives, but biomechanical studies comparing them with standard transosseous repair have reported mixed results.^{7,10-12,18-20} Two studies found improved biomechanical characteristics with suture anchors,^10,21 but 2 others found the characteristics of suture anchor fixation equal to¹¹ or worse than¹² those of transosseous fixation. In light of the controversy regarding strength and clinical outcomes of suture anchor repair compared with transosseous repair, new and potentially superior surgical interventions should be considered.

We recently completed a cadaveric study comparing the biomechanical properties of a novel quadriceps tendon repair technique using 4.75-mm biocomposite knotless suture anchors with suture tape and the properties of conventional techniques using either transosseous or suture anchor repair alone.²² In the cadaveric model, compared with transosseous and fully threaded suture anchor techniques, repair of quadriceps tendon ruptures with this knotless suture anchor with suture tape technique was biomechanically superior in cyclic displacement, construct stiffness, and ultimate load to failure.²² Additionally, this method allows for less extensive dissection, shorter operative times, and the potential for earlier and more aggressive rehabilitation protocols.²² We propose this technique, presented in this article, as a superior alternative to traditional quadriceps tendon repair techniques.

TECHNIQUE

The patient is placed in supine position with a tourniquet placed on the proximal thigh. A midline incision is made from the proximal pole of the patella, proximally by 5 cm. A combination of sharp and blunt dissection is performed through skin and subcutaneous tissues down to the extensor mechanism, exposing the proximal pole of the patella and the torn quadriceps tendon.

The distal aspect of the quadriceps tendon is then débrided of any devitalized tissue and secured with an Allis clamp. A long tape suture (FiberTape; Arthrex) is then used to place a locking Krackow stitch in a distal-to-proximal and then proximal-to-distal direction for 5 throws in each direction within the quadriceps tendon, with the tails exiting distally at the tear site. Care is taken with each pass to ensure that there is no slack within the system.

Continue to: The proximal pole of the patella...

The proximal pole of the patella is then prepared by débriding any remaining soft tissue back to an area of exposed subcortical bone, which is débrided to a bleeding bony bed. Holes are drilled in the medial and lateral thirds of the patella at the proximal pole using the drill for 4.75-mm biocomposite knotless suture anchors (SwiveLock; Arthrex). The tap for the 4.75-mm anchors is then passed at each guide hole. In hard bone, double-tapping is recommended.

Next, the medial strand of tape suture is loaded within a 4.75-mm biocomposite knotless suture anchor eyelet and reduced to the patella. The medial anchor is malleted and screwed into place, while tension is kept on the lateral strand with the knee in full extension. The lateral strand is then placed into its 4.75-mm biocomposite knotless suture anchor, reduced to the patella, and then malleted and screwed into place in the lateral hole, thereby completing the core portion of the repair (Figures A-D). The core strands from the 4.75-mm biocomposite knotless suture anchors are then back-passed in mattress fashion and tied, and medial and lateral retinacular repairs are then performed using supersuture tape (SutureTape or FiberWire; Arthrex).

After surgery, the patient is placed in a knee brace locked in full extension and allowed to weight-bear as tolerated using crutches. During the first week, knee ROM is allowed up to 30°. During weeks 2 to 6 passive ROM is gradually increased to 90°, and use of crutches is tapered. At week 6 the brace is unlocked for ambulation; it may be discontinued after 7 to 8 weeks or when determined safe. Light activity is permitted from month 4 to month 6. A patient who achieves satisfactory strength, is clinically examined, and progresses through rehabilitation is allowed to return to fully unrestricted sport.

DISCUSSION

Quadriceps tendon rupture is an uncommon clinical entity that requires early surgical management.^1-5,12,17,19 The standard of care is passage of nonabsorbable sutures through transosseous patellar bone tunnels, but repair with suture anchors has been studied as an alternative that allows for less tissue trauma, decreased operative time, safe early initiation of rehabilitation protocols, and reduced risk of patella fracture or damage.^{3,7,10-12,18-20,21,23} Despite these potential advantages, biomechanical studies have yielded inconsistent results regarding the superiority of suture anchor repair over repair with transosseous tunnels.^{7,10-12,18-20} We propose quadriceps tendon repair using the 4.75-mm biocomposite knotless suture anchor with tape suture technique as a biomechanically superior alternative to either transosseous tunnels or suture anchor repair alone, with significant advantages both in and out of the operating room.

Results of biomechanical studies comparing transosseous tunnel repair and standard suture anchor repair have been mixed, though the heterogeneity of their study methods and endpoints makes direct comparisons difficult.^{7,10-12,18-20} Petri and colleagues¹⁰ and Sherman and colleagues²¹ reported statistically significant higher load to failure¹⁰ and reduced gapping during cyclic loading^10,21 with suture anchor repair relative to transosseous repair. However, Hart and colleagues¹² found that repair with suture anchors had lower ultimate tensile load, and they concluded that transosseous repair is superior. Lighthart and colleagues¹¹ found no significant difference in displacement between the 2 repairs.

Continue to: In our cadaveric biomechanical study...

In our cadaveric biomechanical study, a novel 4.75-mm biocomposite knotless suture anchor with suture tape repair was compared with traditional 3-tunnel transosseous repair and with standard 2-anchor suture anchor repair.²² Statistically significant superiority was found across multiple parameters, including initial tendon displacement, stiffness, and ultimate load to failure (vs 5.5-mm biocomposite fully threaded suture anchor repair), as well as initial and late tendon displacement, stiffness, and ultimate load to failure (vs transosseous repair).²² Although definitive conclusions are difficult to draw on the basis of prior cadaveric studies comparing standard suture anchor repair and transosseous repair, our results decidedly favor the biomechanical characteristics of this 4.75-mm biocomposite knotless suture anchor with suture tape repair and make it a potentially superior repair technique based on biomechanics alone.²²

Similarly to standard repair with suture anchors, repair using a 4.75-mm biocomposite knotless suture anchor with tape suture eliminates the need to expose the distal pole of the patella.^7,10-12,21 This allows for a smaller surgical incision, less extensive dissection, and prevents possible interference with the patellar tendon.^7,10-12,21 Additionally, it eliminates the risk of iatrogenic patellar fracture and damage to the articular surface from drilling the transpatellar tunnels.^17,18 Both our own review of cases repaired with our 4.75-mm biocomposite knotless suture anchor with suture tape technique as well as studies of suture anchor repair have consistently found operative times of <1 hour.²¹ Shorter operative times and smaller surgical wounds are advantageous given that many of these patients have medical comorbidities that predispose them to intraoperative and wound-healing complications.^12,19-22

Optimal rehabilitation protocols for quadriceps tendon repair are a matter of controversy. Multiple studies of repair with transosseous patellar tunnels describe immobilization for 6 weeks after surgery, but there has been a recent push toward early motion.^7,13,23,24 Reported complications of extended immobilization include limited flexion, pain, weakness, decreased patellar mobility, and patella baja.¹⁴ Studies have suggested that, while excessive loading can cause gap formation and weaken the repair, some controlled motion is necessary to heal the tendon^23,25 and reduce the risks of stiffness and atrophy.¹⁴ The improved biomechanical characteristics of the 4.75-mm biocomposite knotless suture anchor with tape suture technique allow for safe early initiation of ROM exercises and accelerated rehabilitation protocols.

In our early experience with this technique, functional outcomes have been excellent. A formal 2-year outcome study of patients who have undergone quadriceps tendon repair with this 4.75-mm biocomposite knotless suture anchor with tape suture technique is under way.

ABSTRACT

Quadriceps tendon ruptures disrupt the extensor mechanism of the knee and require urgent surgical management. Traditional repair techniques have had mixed biomechanical and clinical results risking weakness and extensor lag. We describe a novel technique using tape suture and knotless anchors, which has performed superiorly during biomechanical testing and yielded terrific early clinical results. 

Continue to: Quadriceps tendon rupture...

Quadriceps tendon rupture is an uncommon yet potentially devastating knee injury with an estimated incidence of 1.37 in 100,000.¹ It most often occurs in male, middle-aged or older patients with degenerative tendon changes and serious systemic diseases, such as chronic renal failure, diabetes mellitus, rheumatoid arthritis, and disorders requiring long-term steroid use (tissue quality is often compromised by patient age and comorbidities).^2-10 Whereas partial tears with an intact extensor mechanism may be managed nonoperatively, prompt operative intervention is indicated in cases of complete tear or an incompetent extensor mechanism to facilitate early range of motion (ROM) and return of knee function.^2-4,8,9

The standard of care is repair with a nonabsorbable suture passed through transosseous patellar tunnels, often with several weeks of postoperative immobilization to protect the repair.^3,4,7,10-12 Reported complications of this method include significant extension lag, decreased strength, and ROM compared with the contralateral knee, chronic pain, and iatrogenic patellar fracture.^8,13-18 Repair techniques using suture anchors have been proposed as viable alternatives, but biomechanical studies comparing them with standard transosseous repair have reported mixed results.^{7,10-12,18-20} Two studies found improved biomechanical characteristics with suture anchors,^10,21 but 2 others found the characteristics of suture anchor fixation equal to¹¹ or worse than¹² those of transosseous fixation. In light of the controversy regarding strength and clinical outcomes of suture anchor repair compared with transosseous repair, new and potentially superior surgical interventions should be considered.

We recently completed a cadaveric study comparing the biomechanical properties of a novel quadriceps tendon repair technique using 4.75-mm biocomposite knotless suture anchors with suture tape and the properties of conventional techniques using either transosseous or suture anchor repair alone.²² In the cadaveric model, compared with transosseous and fully threaded suture anchor techniques, repair of quadriceps tendon ruptures with this knotless suture anchor with suture tape technique was biomechanically superior in cyclic displacement, construct stiffness, and ultimate load to failure.²² Additionally, this method allows for less extensive dissection, shorter operative times, and the potential for earlier and more aggressive rehabilitation protocols.²² We propose this technique, presented in this article, as a superior alternative to traditional quadriceps tendon repair techniques.

TECHNIQUE

The patient is placed in supine position with a tourniquet placed on the proximal thigh. A midline incision is made from the proximal pole of the patella, proximally by 5 cm. A combination of sharp and blunt dissection is performed through skin and subcutaneous tissues down to the extensor mechanism, exposing the proximal pole of the patella and the torn quadriceps tendon.

The distal aspect of the quadriceps tendon is then débrided of any devitalized tissue and secured with an Allis clamp. A long tape suture (FiberTape; Arthrex) is then used to place a locking Krackow stitch in a distal-to-proximal and then proximal-to-distal direction for 5 throws in each direction within the quadriceps tendon, with the tails exiting distally at the tear site. Care is taken with each pass to ensure that there is no slack within the system.

Continue to: The proximal pole of the patella...

The proximal pole of the patella is then prepared by débriding any remaining soft tissue back to an area of exposed subcortical bone, which is débrided to a bleeding bony bed. Holes are drilled in the medial and lateral thirds of the patella at the proximal pole using the drill for 4.75-mm biocomposite knotless suture anchors (SwiveLock; Arthrex). The tap for the 4.75-mm anchors is then passed at each guide hole. In hard bone, double-tapping is recommended.

Next, the medial strand of tape suture is loaded within a 4.75-mm biocomposite knotless suture anchor eyelet and reduced to the patella. The medial anchor is malleted and screwed into place, while tension is kept on the lateral strand with the knee in full extension. The lateral strand is then placed into its 4.75-mm biocomposite knotless suture anchor, reduced to the patella, and then malleted and screwed into place in the lateral hole, thereby completing the core portion of the repair (Figures A-D). The core strands from the 4.75-mm biocomposite knotless suture anchors are then back-passed in mattress fashion and tied, and medial and lateral retinacular repairs are then performed using supersuture tape (SutureTape or FiberWire; Arthrex).

After surgery, the patient is placed in a knee brace locked in full extension and allowed to weight-bear as tolerated using crutches. During the first week, knee ROM is allowed up to 30°. During weeks 2 to 6 passive ROM is gradually increased to 90°, and use of crutches is tapered. At week 6 the brace is unlocked for ambulation; it may be discontinued after 7 to 8 weeks or when determined safe. Light activity is permitted from month 4 to month 6. A patient who achieves satisfactory strength, is clinically examined, and progresses through rehabilitation is allowed to return to fully unrestricted sport.

DISCUSSION

Quadriceps tendon rupture is an uncommon clinical entity that requires early surgical management.^1-5,12,17,19 The standard of care is passage of nonabsorbable sutures through transosseous patellar bone tunnels, but repair with suture anchors has been studied as an alternative that allows for less tissue trauma, decreased operative time, safe early initiation of rehabilitation protocols, and reduced risk of patella fracture or damage.^{3,7,10-12,18-20,21,23} Despite these potential advantages, biomechanical studies have yielded inconsistent results regarding the superiority of suture anchor repair over repair with transosseous tunnels.^{7,10-12,18-20} We propose quadriceps tendon repair using the 4.75-mm biocomposite knotless suture anchor with tape suture technique as a biomechanically superior alternative to either transosseous tunnels or suture anchor repair alone, with significant advantages both in and out of the operating room.

Results of biomechanical studies comparing transosseous tunnel repair and standard suture anchor repair have been mixed, though the heterogeneity of their study methods and endpoints makes direct comparisons difficult.^{7,10-12,18-20} Petri and colleagues¹⁰ and Sherman and colleagues²¹ reported statistically significant higher load to failure¹⁰ and reduced gapping during cyclic loading^10,21 with suture anchor repair relative to transosseous repair. However, Hart and colleagues¹² found that repair with suture anchors had lower ultimate tensile load, and they concluded that transosseous repair is superior. Lighthart and colleagues¹¹ found no significant difference in displacement between the 2 repairs.

Continue to: In our cadaveric biomechanical study...

In our cadaveric biomechanical study, a novel 4.75-mm biocomposite knotless suture anchor with suture tape repair was compared with traditional 3-tunnel transosseous repair and with standard 2-anchor suture anchor repair.²² Statistically significant superiority was found across multiple parameters, including initial tendon displacement, stiffness, and ultimate load to failure (vs 5.5-mm biocomposite fully threaded suture anchor repair), as well as initial and late tendon displacement, stiffness, and ultimate load to failure (vs transosseous repair).²² Although definitive conclusions are difficult to draw on the basis of prior cadaveric studies comparing standard suture anchor repair and transosseous repair, our results decidedly favor the biomechanical characteristics of this 4.75-mm biocomposite knotless suture anchor with suture tape repair and make it a potentially superior repair technique based on biomechanics alone.²²

Similarly to standard repair with suture anchors, repair using a 4.75-mm biocomposite knotless suture anchor with tape suture eliminates the need to expose the distal pole of the patella.^7,10-12,21 This allows for a smaller surgical incision, less extensive dissection, and prevents possible interference with the patellar tendon.^7,10-12,21 Additionally, it eliminates the risk of iatrogenic patellar fracture and damage to the articular surface from drilling the transpatellar tunnels.^17,18 Both our own review of cases repaired with our 4.75-mm biocomposite knotless suture anchor with suture tape technique as well as studies of suture anchor repair have consistently found operative times of <1 hour.²¹ Shorter operative times and smaller surgical wounds are advantageous given that many of these patients have medical comorbidities that predispose them to intraoperative and wound-healing complications.^12,19-22

Optimal rehabilitation protocols for quadriceps tendon repair are a matter of controversy. Multiple studies of repair with transosseous patellar tunnels describe immobilization for 6 weeks after surgery, but there has been a recent push toward early motion.^7,13,23,24 Reported complications of extended immobilization include limited flexion, pain, weakness, decreased patellar mobility, and patella baja.¹⁴ Studies have suggested that, while excessive loading can cause gap formation and weaken the repair, some controlled motion is necessary to heal the tendon^23,25 and reduce the risks of stiffness and atrophy.¹⁴ The improved biomechanical characteristics of the 4.75-mm biocomposite knotless suture anchor with tape suture technique allow for safe early initiation of ROM exercises and accelerated rehabilitation protocols.

In our early experience with this technique, functional outcomes have been excellent. A formal 2-year outcome study of patients who have undergone quadriceps tendon repair with this 4.75-mm biocomposite knotless suture anchor with tape suture technique is under way.

Take-Home Points

• Suture augmentation improves construct strength for patella fixation.
• Krackow sutures may be placed in the quadriceps and patella tendons, then secured over the anterior patella (much like an anterior tension band).
• The Krackow technique described was superior to the suture cerclage technique based on mean load values, but did not reach statistical significance.
• The Krackow suture technique is a viable and easily applied technique for suture augmentation of patella fixation constructs.

Patella fractures are relatively uncommon, accounting for only 1% of skeletal injuries.¹ Restoration of the function of the patella and the extensor mechanism is vital for knee extension and gait. However, patella fractures have an inherently high rate of complications, making these injuries challenging to treat.^2-4 In patients with intact extensor function, displacement of <4 mm, and articular step-off of <3 mm, nonoperative management is extremely effective, with 99% of patients reporting favorable results.⁵ However, for fractures in which the extensor mechanism is disrupted, surgical intervention typically is indicated.⁶

Authors have reported various surgical interventions, one of the most commonly used being the anterior tension band (ATB) technique, first described by the AO (Arbeitsgemeinschaft für Osteosynthesefragen) group in the 1950s.⁷ By converting distractive anterior force during knee flexion to compressive force at the fracture site, the ATB technique provides a repair stronger than the previously used cerclage repair.⁸ Although initially considered standard of care, the ATB technique was soon found to be associated with implant failure and subcutaneous irritation prompting implant removal.⁹

To address these issues, Berg¹⁰ and Carpenter and colleagues¹¹ evaluated an ATB technique that used cannulated screws instead of Kirschner wires (K-wires). This variation on the ATB technique reduced the implant-related issues while maintaining the mechanical advantage of the tension band. The more rigid design also permitted earlier postoperative rehabilitation, which significantly reduced development of arthrofibrosis.^6,7,10 This modified ATB (MATB) technique has since been investigated for additional augments, mainly focusing on use of different tension band materials, including polyester suture and braided composite suture.^12-14

However, there is little research on augments that incorporate the surrounding soft tissue, specifically the quadriceps and patellar tendons. In a recent retrospective clinical study, Oh and colleagues¹⁵ found positive clinical results with use of Krackow sutures, though 2 or 3 vertically oriented stainless steel wires were used instead of cannulated screws.

We conducted a study to determine the biomechanical efficacy of using a cerclage suture augment and a Krackow suture augment coupled with and compared with conventional MATB repair. If effective, this technique may represent another strategy for increasing repair strength and thereby improve postoperative outcomes.

Specimen Preparation

Fresh-frozen cadaver extensor mechanisms (quadriceps tendon, patella, surrounding retinaculum, patellar tendon) were kept frozen at –4°C until preparation. Fifteen specimens were selected. Mean (SD) age at death was 68 (10) years (range, 51-85 years). One specimen was excluded for a short patella tendon, which precluded adequate attachment for testing. All specimens were free of overt osseous pathology.

After specimens were thawed overnight, the patellae were transversely osteotomized with an osteotome at the junction of the middle and distal thirds of the patella. Sharp dissection was performed to carry the division through the medial and lateral retinaculum at the same level. All 14 specimens were then repaired using the MATB technique. First, the transverse fracture was reduced with a reduction clamp. Then, two 4-mm cannulated screws (DePuy Synthes) were inserted parallel to each other and perpendicular to the fracture. An 18-gauge stainless steel wire was then passed through each screw, crossed anteriorly, and tightened to create a figure-of-8 ATB. The specimens were then randomly divided into 3 groups—MATB; MATB with cerclage suture augment; MATB with Krackow suture augment—while ensuring specimens from a single cadaver were placed in different groups to avoid confounding based on bone density differences.

Experimental Setup

Repaired specimens were secured with tissue clamps at the quadriceps and patellar tendons on an MTS Bionix 858 (MTS Systems) hydraulic arm.

Each patella was secured for cyclic testing. Initially it was placed under 10 N of tension. Then it underwent tensile loading from 10 N to 300 N at 50 N/s for 10 cycles. These parameters were based on previous biomechanical patella studies.^10,11 Load was measured with the MTS load cell and displacement with the displacement transducer. Fracture displacement associated with 300-N cyclic tension was recorded. Displacement was calculated as the difference between 10th cycle and 2nd cycle values, which accounted for any degree of initial tissue slippage. After cyclic testing, the patella was placed back in 10 N of tensile loading and subjected to maximum force loading to determine ultimate repair strength. For maximum loading, the patella was stretched progressively at 50 N/s until failure. Again, load and displacement were measured with MTS.

Statistical Analysis

After testing, fracture displacement and maximum load force data were compiled for analysis. One-way analysis of variance with Bonferroni correction was used to determine if there were significant differences between groups. Significance level was set at P < .05.

For cyclic testing, mean total displacement was measured over 10 cycles for each group. Again, displacement was determined by taking the difference between 10th cycle and 2nd cycle values, allowing for system stabilization.

Discussion

Our main objective was to compare the efficacy of a novel suture augment technique with that of other patella fracture repair techniques. Our hypothesis—that adding a Krackow suture augment would increase strength in both cyclic and maximum loading—was supported. Although testing results were not statistically significant because of the small sample size, we think this novel technique has clinically relevant descriptive significance and warrants further investigation.

Proper anatomical reduction and postoperative stabilization are of utmost importance in clinical approaches to patella fractures. In addition, regardless of which technical procedure is used, open reduction should also allow for early range of motion to prevent joint arthrofibrosis. Ever since the ATB technique was first described by the AO group, postoperative outcomes have improved significantly. In a retrospective study by Levack and colleagues,¹⁷ 30 of 64 patients with patella fractures underwent internal fixation. Mean follow-up was 6.2 years. By both objective and subjective measures, the best functional outcomes were associated with internal tension band fixation (vs cerclage repair). Lotke and Ecker¹⁸ also documented the efficacy of the tension band technique. Sixteen patients with patella fractures underwent anterior tension banding; those with a comminuted fracture also underwent cerclage repair for patella stabilization during tension banding. At 6-week follow-up, all patients had good range of motion (≥90° flexion), relatively few symptoms, and no implant failures. Results were similar to those of Levack and colleagues.¹⁷

Although it improves stability and functional outcomes over conventional patellectomy and cerclage wiring, the ATB technique has been associated with subcutaneous irritation caused by the K-wires used to secure the band. Hung and colleagues⁹ followed up 68 patients with patellar fractures. Five of these patients underwent tension banding. Although there was a high level of adequate functional outcomes, implant irritation was found to be “quite frequent.”

To address this issue, Carpenter and colleagues¹¹ evaluated an ATB technique that uses K-wires instead of cannulated screws. Biomechanical testing in a cadaver model revealed less fracture displacement and overall more repair strength through cyclic and maximum load testing. Clinically, these results were supported by Berg,¹⁰ who followed up 10 patients with transverse patella fractures repaired with the MATB technique. At a mean follow-up of 24 months, 7 of the 10 patients had good to excellent outcomes, and there were no implant failures.

Further investigation into patella repairs has mainly focused on improving the MATB technique and experimenting with different tension band materials. Rabalais and colleagues¹³ biomechanically tested high-strength polyethylene suture as a replacement for standard 18-gauge wire, and Bryant and colleagues¹⁴ tested a braided composite suture (FiberWire; Arthrex) as a replacement for standard 18-gauge stainless steel wire. Both found no significant difference with use of augmented tension band material, but Rabalais and colleagues¹³ did find more advantages with a parallel tension band construct than with a standard figure-of-8 arrangement.

In developing our novel technique, we considered that Krackow sutures are routinely used in both quadriceps tendon repair and patellar tendon repair, including partial patellectomy for distal patella fracture. With a suture placed in both tendons, the augment could be expected to resist longitudinal gapping and augment the tension band across the anterior patella. First described by Krackow and colleagues,¹⁹ the Krackow suture is widely used for tendon reconstruction. In an interlocking system of sutures, the Krackow suture provides a repair that is more stable than repair with conventional suture techniques, specifically in the context of tendon repair.²⁰ Given the sesamoidal nature of the patella, its repair shares the goal of gap prevention with other tendon repairs. In theory, anchoring the supporting structures that are above and below the patella provides support for the intervening patella and ultimately improves fracture fixation strength.

Oh and colleagues¹⁵ reported on the clinical efficacy of a Krackow augment in distal pole patella repairs. Similarly, we found a Krackow augment to be efficacious, supporting its potential in clinical approaches to patella repairs. Our results indicate this augment can be a useful clinical adjunct in biomechanical evaluation. 

Limitations of this study include its use of dissected extensor mechanisms, which may have less biofidelity than whole-knee specimens. In our model, specimens were secured at the patellar tendon and the quadriceps tendons, as opposed to the quadriceps tendon and the tibia distally. Use of this model could have led to an increase in early displacement during cyclic testing as a result of tissue slippage. Furthermore, our small sample size could have affected our ability to demonstrate a difference between these techniques.

Given its increased strength as demonstrated by mean displacement during cyclic loading and mean load to failure, as well as the early clinical data recently published, the Krackow suture augment represents a feasible technique for patella fixation. It likely will be most useful in cases in which conventional techniques are prone to failure or cannot be applied, such as severe distal comminution or poor bone density. Further biomechanical testing with a larger number of specimens may be required for statistical significance.

Conclusion

In patella fracture repair strategies, the Krackow suture augment increased strength when used with a MATB technique. Failure to reach statistical significance likely resulted from our small sample size. Further biomechanical testing and clinical studies are needed for more complete evaluation of this technique. We think it will be most useful in the setting of poor bone quality or severe comminution, which can limit fixation options. As increased repair strength allows earlier postoperative rehabilitation and maintains fracture reduction, patient outcomes should improve. This novel technique represents another strategy for managing challenging patella fractures.