Results
From 2011 through 2014, 1531 patients had elective general surgery at NMVAHCS. Twenty-two patients with preoperative anemia (hematocrit < 30%) and 1 patient who received a transfusion > 72 hours after the operation were excluded. Most elective operations (70%, n = 1075) were performed on an outpatient basis; none involved transfusion. Inguinal hernia repair was most common with 479 operations; 17 patients were treated inpatient of which 2 patients had routine postoperative hematocrit checks; (neither received transfusion). One patient with inguinal hernia surgery received transfusion without routine postoperative hematocrit monitoring.
Of 112 partial colon resections, 1 patient had a postoperative transfusion; and all but 3 received postoperative hematocrit monitoring. Nineteen patients undergoing partial colon resection had a clinical indication for postoperative hematocrit monitoring. None of the 5 patients with partial gastrectomy received a postoperative transfusion. Of 121 elective cholecystectomies, no patients had postoperative transfusion, whereas 34 had postoperative hematocrit monitoring; only 2 patients had a clinical reason for the hematocrit monitoring.
Of 430 elective inpatient operations, 12 received transfusions and 288 patients had ≥ 1 postoperative hematocrit test (67%). All hematocrit tests were requested by the attending surgeon, resident surgeon, or the surgical ICU team. Of the group that had postoperative hematocrit monitoring, there was an average of 4.4 postoperative hematocrit tests per patient (range, 1-44).
There were 12 transfusions for inpatients (2.8%), which is similar to the findings of a recent study of VA general surgery (2.3%).13 Five of the 12 patients received intraoperative transfusions while 7 were transfused within 72 hours postoperation. All but 1 patient receiving transfusion had EBL > 199 mL (range, 5-3000; mean, 950 mL; median, 500 mL) and/or signs or symptoms of anemia or other indications for measurement of the postoperative hematocrit. There were no statistically significant differences in patients’ age, sex, BMI, or race and ethnicity between groups receiving and not receiving transfusion (Table 1).
When comparing the transfusion vs the nontransfusion groups (after excluding those with clinical preoperative anemia) the risk factors for transfusion included: relatively low mean preoperative hematocrit (mean, 36.9% vs 42.7%, respectively; P = .003), low postoperative hematocrit (mean, 30.2% vs 37.1%, respectively; P < .001), high EBL (mean, 844 mL vs 109 mL, respectively; P = .005), large infusion of intraoperative fluids (mean, 4625 mL vs 2505 mL, respectively; P = .005), longer duration of operation (mean, 397 min vs 183 min, respectively; P < .001), and longer LOS (mean, 14.5 d vs 4.9 d, respectively; P < .001) (Table 2). Similarly, we found an increased risk for transfusion with high/intermediate cardiovascular risk (vs low), any wound not classified as clean, ICU stay, and postoperative symptoms of anemia.
We found no increased risk for transfusion with ethanol, tobacco, warfarin, or clopidogrel use; polycythemia; thrombocytopenia; preoperative INR; preoperative aPTT; preoperative albumin; Hemoglobin A1c; or diabetes mellitus; or for operations performed for malignancy. Ten patients in the ICU received transfusion (5.8%) compared with 2 patients (0.8%) not admitted to the ICU.
Operations were deemed high risk when ≥ 2 of patients having that operation received transfusions within 72 hours of their operation. There were 15 abdominoperineal resections; 3 of these received transfusions (20%). There were 7 total abdominal colectomies; 3 of these received transfusions (43%). We therefore had 22 high-risk operations, 6 of which were transfused (27%).