Patterson JW

Pancreatic panniculitis represents a rare skin manifestation of underlying pancreatic pathology. The clinical presentation of the condition is remarkably consistent and privy to several unique clinical and histopathologic findings. We report a case of a 50-year-old white woman with pancreatic panniculitis and newly diagnosed pancreatic acinic cell adenocarcinoma. The clinical and histopathologic features, underlying causes, and treatments are reviewed.

Case Report
In June 2005, a 50-year-old white woman presented to the emergency department with a 6-month history of nausea, vomiting, abdominal pain, and weight loss, and a 3-week history of painful leg nodules that had been increasing in size and number in the days prior to admission. She currently was not taking any medications and was allergic to clindamycin and cefuroxime axetil. She smoked half a pack of cigarettes a day for the past 15 years and denied alcohol use. Her family history was notable for breast and colon cancer in her maternal grandmother and cervical cancer in her sister. Physical examination revealed multiple 2- to 4-cm, firm, tender, erythematous nodules on the anterior and anteromedial shins bilaterally (Figure 1). There also was mild tenderness on palpation of the abdomen in the epigastric region. The remainder of the physical examination was unremarkable. Pertinent laboratory findings included an elevated lipase level of 4000 U/L (reference range, 31–186 U/L) and a slightly elevated amylase level of 114 U/L (reference range, 27–131 U/L). A complete blood count and liver function panel were within reference range.

A 5-mm punch biopsy specimen obtained from one of the nodules revealed a predominantly septal panniculitis with some lacelike lobular infiltration of inflammatory cells (Figure 2). Lymphocytes and neutrophils were observed, and eosinophils were particularly prominent. In addition, there were small foci of lipocyte degeneration and calcification, with formation of ghost cells (Figure 3). Aggregates of granular basophilic material also were identified, particularly near the base of the specimen. Gram, Gomori methenamine-silver, and acid-fast bacilli stains were negative for organisms. A diagnosis of pancreatic panniculitis was made.

Further workup of the patient revealed a 5-cm ill-defined mass in the pancreatic head as well as a 2-cm liver mass. Biopsy specimens of the pancreatic and liver masses revealed pancreatic acinic cell adenocarcinoma with metastasis. The patient initially was started on octreotide acetate, gemcitabine hydrochloride, and nonsteroidal anti-inflammatory drugs. After 3 months of therapy, the tumor remained stable in size, but the leg nodules had begun to regress due to the octreotide acetate. Additional chemotherapeutic agents were added to her treatment, including streptozocin and doxorubicin hydrochloride liposome. In August 2005, the pancreatic carcinoma and liver metastasis had dramatically decreased in size and the panniculitis had resolved. The patient requested that the octreotide acetate and chemotherapy be discontinued. She presented again in December 2005 with the return of her panniculitis, this time involving her shins, arms, and hands. A few of the nodules on the shins were noted to express a brown-green oily fluid. Workup revealed an increase in size of her primary tumor and multiple liver masses. Octreotide acetate and chemotherapy were restarted. Two months later (February 2006), the patient's panniculitis had again regressed and her tumors slowly were decreasing in size.

Comment
Pancreatic panniculitis is a cutaneous finding marked by multiple subcutaneous, raised, firm, tender, edematous nodules varying from erythematous to violaceous to red-brown. These nodules most commonly present on the lower legs but also can involve the thighs, buttocks, trunk, and upper extremities.1-10 Individual nodules sometimes ulcerate and discharge a creamy, tan-brown, sterile, viscous substance made up of degenerated lipocytes. Lesions usually resolve with lipoatrophy and hypopigmented and/or hyperpigmented scars.1,2 Additional clinical findings can accompany the skin lesions and relate to lipocyte degeneration in other organs. Periarticular lipocyte degeneration results in a secondary acute arthritis that most frequently involves the ankles and may be migratory, intermittent, or persistent. Other joints subsequently or concurrently may be involved, including the knees, metacarpals, wrists, and elbows. Arthritis has been reported in 54% to 88% of cases.1,3 More rarely, submucosal lipocyte degeneration resulting in gastrointestinal tract bleeding can occur.1 Common laboratory abnormalities associated with pancreatic panniculitis include elevated sedimentation rates and lipase and trypsin levels (Table 1). Some cases are associated with eosinophilia and increased amylase.1-8 A differential diagnosis of panniculitides that may resemble pancreatic panniculitis could include erythema nodosum; sclerosing panniculitis (lipodermatosclerosis); α1-antitrypsin deficiency panniculitis; cutaneous polyarteritis nodosa; nodular vasculitis (erythema induratum); lupus panniculitis; and infective, traumatic, and factitial panniculitis (Table 2).2,5,11,12

The landmark article that first linked pancreatic disease with pancreatic panniculitis was published in 1883 by Chiari.13 Disease processes that resulted in pancreatic panniculitis included acute pancreatitis, chronic pancreatitis, pancreatic pseudocysts, pancreatic duct stenosis, abdominal trauma, and pancreatic carcinoma. A case of panniculitis associated with lupus pancreatitis also has been reported.14 Only 0.3% to 3.0% of patients with pancreatic disease develop associated panniculitis.2 Pancreatic carcinoma and pancreatitis are most intimately associated with pancreatic panniculitis.1 Specifically, acinic cell adenocarcinoma is responsible for more than 50% of all cases,4 though only 16% of acinic cell adenocarcinomas present with panniculitis.15 A small number of neuroendocrine carcinomas have been reported in the literature, as well as an isolated case of an intraductal carcinoid tumor in a pancreas divisum.2,9,10 Pancreatitis plays a role in the development of most of the remaining cases.1 Although pancreatic panniculitis only manifests in a small percentage of cases of pancreatic disease, its importance as a clinical sign should be recognized. As in our case, when panniculitis is observed, it is the presenting sign in 40% of cases of underlying pancreatic disease.16 The panniculitis usually precedes the diagnosis of pancreatic disease by an average of 13 weeks, with a reported range between 2 and 28 weeks.1 The characteristic histopathologic features of pancreatic panniculitis were first described by Szymanski and Bluefarb17 in 1961. Early lesions are nonspecific, marked by perivascular lymphocytic infiltrates that lack necrosis and may resemble erythema nodosum.4 In fact, Ball and colleagues18 have suggested that pancreatic panniculitis may begin as a septal panniculitis and only later develop lobular involvement. Biopsies performed on specimens from the nonulcerated, fully developed erythematous nodules reveal both lobular and septal panniculitis highlighted by focal areas of lipocyte degeneration populated by anucleate necrotic adipocytes surrounded by thickened acidophilic cell membranes, termed ghost cells. A unique feature, when present, is the deposition of granular or homogenous basophilic material resulting from the saponification of fat by calcium salts.12 A dense infiltration of lymphocytes, macrophages, neutrophils, and variable numbers of eosinophils exists at the periphery of the necrotic areas along with evidence of calcification. Resolution of the nodules is characterized by a granulomatous infiltrate that replaces the areas of necrotic tissue.1 The presence of numerous eosinophils was a striking feature in our case and has not been emphasized previously in the literature in this form of panniculitis. Although there is no universally accepted mechanism for the development of the skin lesions, a popular hypothesis states that a synergism exists between the elevated serum levels of lipase and trypsin. Trypsin alters the permeability of the tissue blood vessels, which allows lipase to hydrolyze lipids in the adipocyte cell membranes and interior, which leads to lipocyte degeneration of the tissue.16,19 Support for this hypothesis is garnered by the observations that more than 50% of patients with pancreatic portal fistulization develop panniculitis, and immunohistochemical analysis of the areas of lipocyte degeneration demonstrate pancreatic lipase.6,20 Potts and colleagues21 suggested a possible immunologic mechanism in a patient with pancreatic carcinoma and pancreatic panniculitis who was noted to have decreased complement levels and deposition of immunoglobulin G in the pleura. Successful treatment of pancreatic panniculitis usually requires diagnosis and treatment of the underlying pancreatic pathology. As the pancreatic enzyme levels decrease, the skin lesions usually tend to regress.3 There has been some success reported with the administration of octreotide acetate, a synthetic polypeptide that inhibits pancreatic enzyme production.1,2,4,6 In addition, general supportive measures, including rest, elevation of the legs, compression stockings, and nonsteroidal anti-inflammatory drugs, may be helpful.

Pancreatic panniculitis represents a rare skin manifestation of underlying pancreatic pathology. The clinical presentation of the condition is remarkably consistent and privy to several unique clinical and histopathologic findings. We report a case of a 50-year-old white woman with pancreatic panniculitis and newly diagnosed pancreatic acinic cell adenocarcinoma. The clinical and histopathologic features, underlying causes, and treatments are reviewed.

Case Report
In June 2005, a 50-year-old white woman presented to the emergency department with a 6-month history of nausea, vomiting, abdominal pain, and weight loss, and a 3-week history of painful leg nodules that had been increasing in size and number in the days prior to admission. She currently was not taking any medications and was allergic to clindamycin and cefuroxime axetil. She smoked half a pack of cigarettes a day for the past 15 years and denied alcohol use. Her family history was notable for breast and colon cancer in her maternal grandmother and cervical cancer in her sister. Physical examination revealed multiple 2- to 4-cm, firm, tender, erythematous nodules on the anterior and anteromedial shins bilaterally (Figure 1). There also was mild tenderness on palpation of the abdomen in the epigastric region. The remainder of the physical examination was unremarkable. Pertinent laboratory findings included an elevated lipase level of 4000 U/L (reference range, 31–186 U/L) and a slightly elevated amylase level of 114 U/L (reference range, 27–131 U/L). A complete blood count and liver function panel were within reference range.

A 5-mm punch biopsy specimen obtained from one of the nodules revealed a predominantly septal panniculitis with some lacelike lobular infiltration of inflammatory cells (Figure 2). Lymphocytes and neutrophils were observed, and eosinophils were particularly prominent. In addition, there were small foci of lipocyte degeneration and calcification, with formation of ghost cells (Figure 3). Aggregates of granular basophilic material also were identified, particularly near the base of the specimen. Gram, Gomori methenamine-silver, and acid-fast bacilli stains were negative for organisms. A diagnosis of pancreatic panniculitis was made.

Further workup of the patient revealed a 5-cm ill-defined mass in the pancreatic head as well as a 2-cm liver mass. Biopsy specimens of the pancreatic and liver masses revealed pancreatic acinic cell adenocarcinoma with metastasis. The patient initially was started on octreotide acetate, gemcitabine hydrochloride, and nonsteroidal anti-inflammatory drugs. After 3 months of therapy, the tumor remained stable in size, but the leg nodules had begun to regress due to the octreotide acetate. Additional chemotherapeutic agents were added to her treatment, including streptozocin and doxorubicin hydrochloride liposome. In August 2005, the pancreatic carcinoma and liver metastasis had dramatically decreased in size and the panniculitis had resolved. The patient requested that the octreotide acetate and chemotherapy be discontinued. She presented again in December 2005 with the return of her panniculitis, this time involving her shins, arms, and hands. A few of the nodules on the shins were noted to express a brown-green oily fluid. Workup revealed an increase in size of her primary tumor and multiple liver masses. Octreotide acetate and chemotherapy were restarted. Two months later (February 2006), the patient's panniculitis had again regressed and her tumors slowly were decreasing in size.

Comment
Pancreatic panniculitis is a cutaneous finding marked by multiple subcutaneous, raised, firm, tender, edematous nodules varying from erythematous to violaceous to red-brown. These nodules most commonly present on the lower legs but also can involve the thighs, buttocks, trunk, and upper extremities.1-10 Individual nodules sometimes ulcerate and discharge a creamy, tan-brown, sterile, viscous substance made up of degenerated lipocytes. Lesions usually resolve with lipoatrophy and hypopigmented and/or hyperpigmented scars.1,2 Additional clinical findings can accompany the skin lesions and relate to lipocyte degeneration in other organs. Periarticular lipocyte degeneration results in a secondary acute arthritis that most frequently involves the ankles and may be migratory, intermittent, or persistent. Other joints subsequently or concurrently may be involved, including the knees, metacarpals, wrists, and elbows. Arthritis has been reported in 54% to 88% of cases.1,3 More rarely, submucosal lipocyte degeneration resulting in gastrointestinal tract bleeding can occur.1 Common laboratory abnormalities associated with pancreatic panniculitis include elevated sedimentation rates and lipase and trypsin levels (Table 1). Some cases are associated with eosinophilia and increased amylase.1-8 A differential diagnosis of panniculitides that may resemble pancreatic panniculitis could include erythema nodosum; sclerosing panniculitis (lipodermatosclerosis); α1-antitrypsin deficiency panniculitis; cutaneous polyarteritis nodosa; nodular vasculitis (erythema induratum); lupus panniculitis; and infective, traumatic, and factitial panniculitis (Table 2).2,5,11,12

The landmark article that first linked pancreatic disease with pancreatic panniculitis was published in 1883 by Chiari.13 Disease processes that resulted in pancreatic panniculitis included acute pancreatitis, chronic pancreatitis, pancreatic pseudocysts, pancreatic duct stenosis, abdominal trauma, and pancreatic carcinoma. A case of panniculitis associated with lupus pancreatitis also has been reported.14 Only 0.3% to 3.0% of patients with pancreatic disease develop associated panniculitis.2 Pancreatic carcinoma and pancreatitis are most intimately associated with pancreatic panniculitis.1 Specifically, acinic cell adenocarcinoma is responsible for more than 50% of all cases,4 though only 16% of acinic cell adenocarcinomas present with panniculitis.15 A small number of neuroendocrine carcinomas have been reported in the literature, as well as an isolated case of an intraductal carcinoid tumor in a pancreas divisum.2,9,10 Pancreatitis plays a role in the development of most of the remaining cases.1 Although pancreatic panniculitis only manifests in a small percentage of cases of pancreatic disease, its importance as a clinical sign should be recognized. As in our case, when panniculitis is observed, it is the presenting sign in 40% of cases of underlying pancreatic disease.16 The panniculitis usually precedes the diagnosis of pancreatic disease by an average of 13 weeks, with a reported range between 2 and 28 weeks.1 The characteristic histopathologic features of pancreatic panniculitis were first described by Szymanski and Bluefarb17 in 1961. Early lesions are nonspecific, marked by perivascular lymphocytic infiltrates that lack necrosis and may resemble erythema nodosum.4 In fact, Ball and colleagues18 have suggested that pancreatic panniculitis may begin as a septal panniculitis and only later develop lobular involvement. Biopsies performed on specimens from the nonulcerated, fully developed erythematous nodules reveal both lobular and septal panniculitis highlighted by focal areas of lipocyte degeneration populated by anucleate necrotic adipocytes surrounded by thickened acidophilic cell membranes, termed ghost cells. A unique feature, when present, is the deposition of granular or homogenous basophilic material resulting from the saponification of fat by calcium salts.12 A dense infiltration of lymphocytes, macrophages, neutrophils, and variable numbers of eosinophils exists at the periphery of the necrotic areas along with evidence of calcification. Resolution of the nodules is characterized by a granulomatous infiltrate that replaces the areas of necrotic tissue.1 The presence of numerous eosinophils was a striking feature in our case and has not been emphasized previously in the literature in this form of panniculitis. Although there is no universally accepted mechanism for the development of the skin lesions, a popular hypothesis states that a synergism exists between the elevated serum levels of lipase and trypsin. Trypsin alters the permeability of the tissue blood vessels, which allows lipase to hydrolyze lipids in the adipocyte cell membranes and interior, which leads to lipocyte degeneration of the tissue.16,19 Support for this hypothesis is garnered by the observations that more than 50% of patients with pancreatic portal fistulization develop panniculitis, and immunohistochemical analysis of the areas of lipocyte degeneration demonstrate pancreatic lipase.6,20 Potts and colleagues21 suggested a possible immunologic mechanism in a patient with pancreatic carcinoma and pancreatic panniculitis who was noted to have decreased complement levels and deposition of immunoglobulin G in the pleura. Successful treatment of pancreatic panniculitis usually requires diagnosis and treatment of the underlying pancreatic pathology. As the pancreatic enzyme levels decrease, the skin lesions usually tend to regress.3 There has been some success reported with the administration of octreotide acetate, a synthetic polypeptide that inhibits pancreatic enzyme production.1,2,4,6 In addition, general supportive measures, including rest, elevation of the legs, compression stockings, and nonsteroidal anti-inflammatory drugs, may be helpful.

Pancreatic panniculitis represents a rare skin manifestation of underlying pancreatic pathology. The clinical presentation of the condition is remarkably consistent and privy to several unique clinical and histopathologic findings. We report a case of a 50-year-old white woman with pancreatic panniculitis and newly diagnosed pancreatic acinic cell adenocarcinoma. The clinical and histopathologic features, underlying causes, and treatments are reviewed.

Case Report
In June 2005, a 50-year-old white woman presented to the emergency department with a 6-month history of nausea, vomiting, abdominal pain, and weight loss, and a 3-week history of painful leg nodules that had been increasing in size and number in the days prior to admission. She currently was not taking any medications and was allergic to clindamycin and cefuroxime axetil. She smoked half a pack of cigarettes a day for the past 15 years and denied alcohol use. Her family history was notable for breast and colon cancer in her maternal grandmother and cervical cancer in her sister. Physical examination revealed multiple 2- to 4-cm, firm, tender, erythematous nodules on the anterior and anteromedial shins bilaterally (Figure 1). There also was mild tenderness on palpation of the abdomen in the epigastric region. The remainder of the physical examination was unremarkable. Pertinent laboratory findings included an elevated lipase level of 4000 U/L (reference range, 31–186 U/L) and a slightly elevated amylase level of 114 U/L (reference range, 27–131 U/L). A complete blood count and liver function panel were within reference range.

A 5-mm punch biopsy specimen obtained from one of the nodules revealed a predominantly septal panniculitis with some lacelike lobular infiltration of inflammatory cells (Figure 2). Lymphocytes and neutrophils were observed, and eosinophils were particularly prominent. In addition, there were small foci of lipocyte degeneration and calcification, with formation of ghost cells (Figure 3). Aggregates of granular basophilic material also were identified, particularly near the base of the specimen. Gram, Gomori methenamine-silver, and acid-fast bacilli stains were negative for organisms. A diagnosis of pancreatic panniculitis was made.

Further workup of the patient revealed a 5-cm ill-defined mass in the pancreatic head as well as a 2-cm liver mass. Biopsy specimens of the pancreatic and liver masses revealed pancreatic acinic cell adenocarcinoma with metastasis. The patient initially was started on octreotide acetate, gemcitabine hydrochloride, and nonsteroidal anti-inflammatory drugs. After 3 months of therapy, the tumor remained stable in size, but the leg nodules had begun to regress due to the octreotide acetate. Additional chemotherapeutic agents were added to her treatment, including streptozocin and doxorubicin hydrochloride liposome. In August 2005, the pancreatic carcinoma and liver metastasis had dramatically decreased in size and the panniculitis had resolved. The patient requested that the octreotide acetate and chemotherapy be discontinued. She presented again in December 2005 with the return of her panniculitis, this time involving her shins, arms, and hands. A few of the nodules on the shins were noted to express a brown-green oily fluid. Workup revealed an increase in size of her primary tumor and multiple liver masses. Octreotide acetate and chemotherapy were restarted. Two months later (February 2006), the patient's panniculitis had again regressed and her tumors slowly were decreasing in size.

Comment
Pancreatic panniculitis is a cutaneous finding marked by multiple subcutaneous, raised, firm, tender, edematous nodules varying from erythematous to violaceous to red-brown. These nodules most commonly present on the lower legs but also can involve the thighs, buttocks, trunk, and upper extremities.1-10 Individual nodules sometimes ulcerate and discharge a creamy, tan-brown, sterile, viscous substance made up of degenerated lipocytes. Lesions usually resolve with lipoatrophy and hypopigmented and/or hyperpigmented scars.1,2 Additional clinical findings can accompany the skin lesions and relate to lipocyte degeneration in other organs. Periarticular lipocyte degeneration results in a secondary acute arthritis that most frequently involves the ankles and may be migratory, intermittent, or persistent. Other joints subsequently or concurrently may be involved, including the knees, metacarpals, wrists, and elbows. Arthritis has been reported in 54% to 88% of cases.1,3 More rarely, submucosal lipocyte degeneration resulting in gastrointestinal tract bleeding can occur.1 Common laboratory abnormalities associated with pancreatic panniculitis include elevated sedimentation rates and lipase and trypsin levels (Table 1). Some cases are associated with eosinophilia and increased amylase.1-8 A differential diagnosis of panniculitides that may resemble pancreatic panniculitis could include erythema nodosum; sclerosing panniculitis (lipodermatosclerosis); α1-antitrypsin deficiency panniculitis; cutaneous polyarteritis nodosa; nodular vasculitis (erythema induratum); lupus panniculitis; and infective, traumatic, and factitial panniculitis (Table 2).2,5,11,12

The landmark article that first linked pancreatic disease with pancreatic panniculitis was published in 1883 by Chiari.13 Disease processes that resulted in pancreatic panniculitis included acute pancreatitis, chronic pancreatitis, pancreatic pseudocysts, pancreatic duct stenosis, abdominal trauma, and pancreatic carcinoma. A case of panniculitis associated with lupus pancreatitis also has been reported.14 Only 0.3% to 3.0% of patients with pancreatic disease develop associated panniculitis.2 Pancreatic carcinoma and pancreatitis are most intimately associated with pancreatic panniculitis.1 Specifically, acinic cell adenocarcinoma is responsible for more than 50% of all cases,4 though only 16% of acinic cell adenocarcinomas present with panniculitis.15 A small number of neuroendocrine carcinomas have been reported in the literature, as well as an isolated case of an intraductal carcinoid tumor in a pancreas divisum.2,9,10 Pancreatitis plays a role in the development of most of the remaining cases.1 Although pancreatic panniculitis only manifests in a small percentage of cases of pancreatic disease, its importance as a clinical sign should be recognized. As in our case, when panniculitis is observed, it is the presenting sign in 40% of cases of underlying pancreatic disease.16 The panniculitis usually precedes the diagnosis of pancreatic disease by an average of 13 weeks, with a reported range between 2 and 28 weeks.1 The characteristic histopathologic features of pancreatic panniculitis were first described by Szymanski and Bluefarb17 in 1961. Early lesions are nonspecific, marked by perivascular lymphocytic infiltrates that lack necrosis and may resemble erythema nodosum.4 In fact, Ball and colleagues18 have suggested that pancreatic panniculitis may begin as a septal panniculitis and only later develop lobular involvement. Biopsies performed on specimens from the nonulcerated, fully developed erythematous nodules reveal both lobular and septal panniculitis highlighted by focal areas of lipocyte degeneration populated by anucleate necrotic adipocytes surrounded by thickened acidophilic cell membranes, termed ghost cells. A unique feature, when present, is the deposition of granular or homogenous basophilic material resulting from the saponification of fat by calcium salts.12 A dense infiltration of lymphocytes, macrophages, neutrophils, and variable numbers of eosinophils exists at the periphery of the necrotic areas along with evidence of calcification. Resolution of the nodules is characterized by a granulomatous infiltrate that replaces the areas of necrotic tissue.1 The presence of numerous eosinophils was a striking feature in our case and has not been emphasized previously in the literature in this form of panniculitis. Although there is no universally accepted mechanism for the development of the skin lesions, a popular hypothesis states that a synergism exists between the elevated serum levels of lipase and trypsin. Trypsin alters the permeability of the tissue blood vessels, which allows lipase to hydrolyze lipids in the adipocyte cell membranes and interior, which leads to lipocyte degeneration of the tissue.16,19 Support for this hypothesis is garnered by the observations that more than 50% of patients with pancreatic portal fistulization develop panniculitis, and immunohistochemical analysis of the areas of lipocyte degeneration demonstrate pancreatic lipase.6,20 Potts and colleagues21 suggested a possible immunologic mechanism in a patient with pancreatic carcinoma and pancreatic panniculitis who was noted to have decreased complement levels and deposition of immunoglobulin G in the pleura. Successful treatment of pancreatic panniculitis usually requires diagnosis and treatment of the underlying pancreatic pathology. As the pancreatic enzyme levels decrease, the skin lesions usually tend to regress.3 There has been some success reported with the administration of octreotide acetate, a synthetic polypeptide that inhibits pancreatic enzyme production.1,2,4,6 In addition, general supportive measures, including rest, elevation of the legs, compression stockings, and nonsteroidal anti-inflammatory drugs, may be helpful.

Scars may become pigmented for a variety of reasons, including the persistence and/or recurrence of an incompletely removed melanocytic nevus. However, development of an intensely hyperpigmented scar not long after a surgical procedure, in the absence of a clear explanation, would be a distinctly uncommon event. We recently encountered such a lesion in an otherwise healthy 20-year-old patient. In this case, the histopathologic findings led to further questioning of the patient and revealed a cause that had not been previously suspected. 

Case Report

A 20-year-old woman was seen for evaluation of a lesion on the left lower abdomen. Six weeks earlier, the lesion had been shave excised by an outside physician; pathology results were not initially available. The patient reported that the lesion had quadrupled in size and darkened considerably since the time of the excision. Her grandmother had died of malignant melanoma. She reported that her only medication was birth control pills. On physical examination, there was a 13X8-mm brown-black nodule with discrete but irregular borders (Figure 1). The clinical impression was recurrent nevus in a shave excision scar. However, because of the rapid growth, dark color, and family history of melanoma, there also was concern about the possibility of an atypical nevus or malignant melanoma. Therefore, an elliptical excision was performed. A report of the initial biopsy specimen was received, with the interpretation benign compound nevus.

Results of histopathologic evaluation of the reexcision specimen showed no residual melanocytic lesion. There was a prominent pigmented, cellular scar occupying the superficial to mid dermis in the central portion of the specimen. The pigmented material consisted of refractile, golden brown granules within macrophages and extracellularly, having a resemblance to hemosiderin (Figure 2). These granules stained positively with Perls stain for iron and with Fontana-Masson stain (Figure 3). Fontana-Masson staining was negative when performed after a bleaching procedure that employed potassium permanganate solution at a concentration of 3 g/L.

The staining results suggested the possibility of minocycline-related hyperpigmentation. Subsequent questioning of the patient revealed that she had been taking minocycline 100 mg twice daily during the 2 years prior to her clinic visit. 

Comment

Pigmented scars can arise occasionally because of a number of factors. The sites of persistent and/or recurrent nevus are often pigmented. This pigment, confined to the scar, often shows irregular borders and may have a mottled appearance.1 Pigmented scars also are observed in spontaneously regressing malignant melanoma.2 In a related phenomenon called tumoral melanosis, sheets of melanophages may accompany either a regressed melanoma or epithelial neoplasm.3,4 Pigmentation of scars related to hemorrhage also could occur, eg, following postsurgical trauma or in association with clotting abnormalities, though it is difficult to find literature directly addressing this problem. Other reported associations with hyperpigmented scars include leishmaniasis,5 chickenpox,6 burns,7 Addison disease,8 and hemosiderin-related pigmentation in endometriosis arising in cesarean scars.9 Among other agents that cause cutaneous pigmentation and could potentially produce hyperpigmented scars are heavy metals (eg, gold) and drugs such as amiodarone, phenothiazines, and antimalarials.10,11 Biopsy results of oral hyperpigmentation due to long-term antimalarial therapy have shown macrophages that contain melanin and ferric iron,12 findings resembling those reported here. None of these causes was pertinent to our case.

Minocycline first became available for clinical use in 1967. An association between minocycline administration and black discoloration of thyroid gland follicles in animals was reported that same year.13,14 As early as 1972, Velasco et al15 reported a macular pigmentation of the legs in patients receiving minocycline for the treatment of venereal disease. Since that time, there have been a number of reports of minocycline-induced pigmentation of skin and mucous membranes. Journal articles and textbooks usually divide minocycline-related cutaneous pigmentation into 3 major types. The first, type I, is a blue-black pigmentation that develops in areas of inflammation and scar13,16-19; this is the type that we report here. The second, type II, is a blue-gray pigmentation that develops particularly over otherwise normal-appearing skin of the arms, legs, or face.18,20,21 The third, type III, is usually described as a diffuse or generalized "muddy brown" pigmentation,13,22-25 though in one report this type of pigmentation was actually described as dark blue-gray.24 The Table provides a summary of the clinical and histopathologic changes associated with the 3 major types of minocycline pigmentation. Pigmentation of the nails and nail beds also occurs19,26 and has coexisted with diffuse cutaneous and scleral pigmentation.25 A fourth type of pigmentation that is not specific to minocycline results from fixed drug eruption, as described by Chu et al27 and possibly also represented by the case of Tanzi and Hecker.28 Minocycline also has been associated with discoloration of teeth,23 pigmented conjunctival cysts,29 and black galactorrhea,30 as well as pigmentation of internal organs such as cardiac valves.31,32

The duration of treatment and total dose required for minocycline to produce cutaneous pigmentation is difficult to determine. Although data on duration and total dose are often provided in reports, these figures typically reflect the totals at the time the patients present to their physician, rather than the time of actual onset of pigmentation, which is much more difficult to determine. Localized pigmentation at a site of tissue injury does not appear to be directly related to the duration of treatment18 and has been reported to occur as rapidly as 1 to 3 months following the onset of minocycline therapy.16,19 The evidence suggests that the diffuse type of pigmentation is more dependent on total dose and duration of therapy; reported patients have been on minocycline for about 3 years, with total doses ranging from 130 to 144 g.24,25

As generally described, there are differences among the microscopic features of the 3 major types of minocycline pigmentation. In type I, the dermal pigment is present in macrophages and stains positively for iron in a manner similar to hemosiderin.13,16,17 Type II pigmentation stains for iron and also is reactive with Fontana-Masson.10,20,33 Type III pigmentation has shown an increase in basilar melanin and brown-black pigment in macrophages that stains positively with Fontana-Masson and negatively for iron.24 However, staining results are not always distinctive among the 3 types. For example, in our patient's scar and in the inflammatory lesions of Ozog et al19 (examples of type I pigmentation), there was dermal pigment that stained positively both for iron and with the Fontana-Masson method. Patients also may have more than one type of cutaneous minocycline pigmentation. In the case of Pepine et al,25 there were areas of blue-black pigmentation, as well as muddy brown discoloration in sun-exposed areas. Biopsy results showed black pigment deposition in perivascular and periadnexal areas, though it is not entirely clear whether these specimens were obtained from blue-black or muddy brown areas.25 Electron microscopy in cases with blue-gray or blue-black pigmentation has shown electron-dense particles in macrophages or extracellularly. Some intracytoplasmic granules are present within lysosomes, while others, including fine dustlike particles consistent with ferritin, are not bound by lysosomal membranes.10,17,20,25 Energy dispersive x-ray microanalysis has shown that the granules mostly contain iron, with lesser amounts of calcium.21,26

The Fontana-Masson staining method is routinely employed to demonstrate the presence of melanin in tissue sections. Therefore, positivity in instances of minocycline pigmentation has suggested to some that melanin is at least partly responsible for the changes. This idea has been supported by one ultrastructural study showing melanosome complexes in siderosomes in a case of minocycline-related hyperpigmentation.21 However, melanosomes have not been identified in other studies.10 It is reported that iron may give positive reactions with Fontana-Masson staining.20 Furthermore, the black staining of Fontana-Masson results from the action of a reducing substance on ammoniated silver nitrate; that reducing substance is not necessarily melanin.10 The failure of the pigment to bleach, in contrast to the case with melanin, has been used to support the idea that the pigment in question does not contain melanin.10 However, reported results with bleaching have been variable. Successful bleaching or partial bleaching has been observed in examples of cutaneous minocycline pigmentation,19 as well as minocycline pigmentation of the thyroid gland34 and heart valves.32 This also is true of our case, because Fontana-Masson staining became negative when preceded by a bleaching procedure. Because past studies have employed several bleaching agents—hydrogen peroxide and potassium permanganate—and because the concentrations used in bleaching and other technical details are rarely provided, in our view, one cannot rely on the results of bleaching alone as proof of the presence or absence of melanin.

The evidence suggests that most examples of minocycline pigmentation—particularly types I and II—are due to cutaneous deposits of the drug or a metabolite thereof, chelated with iron.10,17,26,35 Clues to the mechanism of pigment deposition are provided by the studies of thyroid pigment by Enochs et al.36 Their in vitro modeling studies using electron paramagnetic resonance spectroscopy suggest that the pigment is a polymer caused by the in vivo oxidation of minocycline by thyroid peroxidase, which produces a melaninlike pigment.36 This pigment also contains significant amounts of iron, tightly bound in situ. A related phenomenon could well occur in the skin. Then, as suggested by Argenyi et al,10 the metabolite could act as a reducing substance, explaining the frequent positivity with the Fontana-Masson stain. It is possible that minocycline also may stimulate melanin production, accounting for the diffuse muddy brown type III pigmentation,17 but further studies are needed to clarify this point. The good news is that minocycline pigmentation resolves after cessation of therapy, though this may be a gradual process.17,19,25,37 

Conclusion

Minocycline therapy should be included in the differential diagnosis of hyperpigmented scars. Careful history taking and even repeated questioning may be necessary to elicit an accurate medication history. The pigmentation is most likely due to a minocycline metabolite, bound to iron; Fontana-Masson positivity may result from the action of reducing agents other than melanin. Slow resolution of the pigment can be expected following discontinuation of the drug. Nevertheless, biopsy is indicated when, as in this case, an atypical pigmented skin lesion raises concerns about malignant melanoma.

Scars may become pigmented for a variety of reasons, including the persistence and/or recurrence of an incompletely removed melanocytic nevus. However, development of an intensely hyperpigmented scar not long after a surgical procedure, in the absence of a clear explanation, would be a distinctly uncommon event. We recently encountered such a lesion in an otherwise healthy 20-year-old patient. In this case, the histopathologic findings led to further questioning of the patient and revealed a cause that had not been previously suspected. 

Case Report

A 20-year-old woman was seen for evaluation of a lesion on the left lower abdomen. Six weeks earlier, the lesion had been shave excised by an outside physician; pathology results were not initially available. The patient reported that the lesion had quadrupled in size and darkened considerably since the time of the excision. Her grandmother had died of malignant melanoma. She reported that her only medication was birth control pills. On physical examination, there was a 13X8-mm brown-black nodule with discrete but irregular borders (Figure 1). The clinical impression was recurrent nevus in a shave excision scar. However, because of the rapid growth, dark color, and family history of melanoma, there also was concern about the possibility of an atypical nevus or malignant melanoma. Therefore, an elliptical excision was performed. A report of the initial biopsy specimen was received, with the interpretation benign compound nevus.

Results of histopathologic evaluation of the reexcision specimen showed no residual melanocytic lesion. There was a prominent pigmented, cellular scar occupying the superficial to mid dermis in the central portion of the specimen. The pigmented material consisted of refractile, golden brown granules within macrophages and extracellularly, having a resemblance to hemosiderin (Figure 2). These granules stained positively with Perls stain for iron and with Fontana-Masson stain (Figure 3). Fontana-Masson staining was negative when performed after a bleaching procedure that employed potassium permanganate solution at a concentration of 3 g/L.

The staining results suggested the possibility of minocycline-related hyperpigmentation. Subsequent questioning of the patient revealed that she had been taking minocycline 100 mg twice daily during the 2 years prior to her clinic visit. 

Comment

Pigmented scars can arise occasionally because of a number of factors. The sites of persistent and/or recurrent nevus are often pigmented. This pigment, confined to the scar, often shows irregular borders and may have a mottled appearance.1 Pigmented scars also are observed in spontaneously regressing malignant melanoma.2 In a related phenomenon called tumoral melanosis, sheets of melanophages may accompany either a regressed melanoma or epithelial neoplasm.3,4 Pigmentation of scars related to hemorrhage also could occur, eg, following postsurgical trauma or in association with clotting abnormalities, though it is difficult to find literature directly addressing this problem. Other reported associations with hyperpigmented scars include leishmaniasis,5 chickenpox,6 burns,7 Addison disease,8 and hemosiderin-related pigmentation in endometriosis arising in cesarean scars.9 Among other agents that cause cutaneous pigmentation and could potentially produce hyperpigmented scars are heavy metals (eg, gold) and drugs such as amiodarone, phenothiazines, and antimalarials.10,11 Biopsy results of oral hyperpigmentation due to long-term antimalarial therapy have shown macrophages that contain melanin and ferric iron,12 findings resembling those reported here. None of these causes was pertinent to our case.

Minocycline first became available for clinical use in 1967. An association between minocycline administration and black discoloration of thyroid gland follicles in animals was reported that same year.13,14 As early as 1972, Velasco et al15 reported a macular pigmentation of the legs in patients receiving minocycline for the treatment of venereal disease. Since that time, there have been a number of reports of minocycline-induced pigmentation of skin and mucous membranes. Journal articles and textbooks usually divide minocycline-related cutaneous pigmentation into 3 major types. The first, type I, is a blue-black pigmentation that develops in areas of inflammation and scar13,16-19; this is the type that we report here. The second, type II, is a blue-gray pigmentation that develops particularly over otherwise normal-appearing skin of the arms, legs, or face.18,20,21 The third, type III, is usually described as a diffuse or generalized "muddy brown" pigmentation,13,22-25 though in one report this type of pigmentation was actually described as dark blue-gray.24 The Table provides a summary of the clinical and histopathologic changes associated with the 3 major types of minocycline pigmentation. Pigmentation of the nails and nail beds also occurs19,26 and has coexisted with diffuse cutaneous and scleral pigmentation.25 A fourth type of pigmentation that is not specific to minocycline results from fixed drug eruption, as described by Chu et al27 and possibly also represented by the case of Tanzi and Hecker.28 Minocycline also has been associated with discoloration of teeth,23 pigmented conjunctival cysts,29 and black galactorrhea,30 as well as pigmentation of internal organs such as cardiac valves.31,32

The duration of treatment and total dose required for minocycline to produce cutaneous pigmentation is difficult to determine. Although data on duration and total dose are often provided in reports, these figures typically reflect the totals at the time the patients present to their physician, rather than the time of actual onset of pigmentation, which is much more difficult to determine. Localized pigmentation at a site of tissue injury does not appear to be directly related to the duration of treatment18 and has been reported to occur as rapidly as 1 to 3 months following the onset of minocycline therapy.16,19 The evidence suggests that the diffuse type of pigmentation is more dependent on total dose and duration of therapy; reported patients have been on minocycline for about 3 years, with total doses ranging from 130 to 144 g.24,25

As generally described, there are differences among the microscopic features of the 3 major types of minocycline pigmentation. In type I, the dermal pigment is present in macrophages and stains positively for iron in a manner similar to hemosiderin.13,16,17 Type II pigmentation stains for iron and also is reactive with Fontana-Masson.10,20,33 Type III pigmentation has shown an increase in basilar melanin and brown-black pigment in macrophages that stains positively with Fontana-Masson and negatively for iron.24 However, staining results are not always distinctive among the 3 types. For example, in our patient's scar and in the inflammatory lesions of Ozog et al19 (examples of type I pigmentation), there was dermal pigment that stained positively both for iron and with the Fontana-Masson method. Patients also may have more than one type of cutaneous minocycline pigmentation. In the case of Pepine et al,25 there were areas of blue-black pigmentation, as well as muddy brown discoloration in sun-exposed areas. Biopsy results showed black pigment deposition in perivascular and periadnexal areas, though it is not entirely clear whether these specimens were obtained from blue-black or muddy brown areas.25 Electron microscopy in cases with blue-gray or blue-black pigmentation has shown electron-dense particles in macrophages or extracellularly. Some intracytoplasmic granules are present within lysosomes, while others, including fine dustlike particles consistent with ferritin, are not bound by lysosomal membranes.10,17,20,25 Energy dispersive x-ray microanalysis has shown that the granules mostly contain iron, with lesser amounts of calcium.21,26

The Fontana-Masson staining method is routinely employed to demonstrate the presence of melanin in tissue sections. Therefore, positivity in instances of minocycline pigmentation has suggested to some that melanin is at least partly responsible for the changes. This idea has been supported by one ultrastructural study showing melanosome complexes in siderosomes in a case of minocycline-related hyperpigmentation.21 However, melanosomes have not been identified in other studies.10 It is reported that iron may give positive reactions with Fontana-Masson staining.20 Furthermore, the black staining of Fontana-Masson results from the action of a reducing substance on ammoniated silver nitrate; that reducing substance is not necessarily melanin.10 The failure of the pigment to bleach, in contrast to the case with melanin, has been used to support the idea that the pigment in question does not contain melanin.10 However, reported results with bleaching have been variable. Successful bleaching or partial bleaching has been observed in examples of cutaneous minocycline pigmentation,19 as well as minocycline pigmentation of the thyroid gland34 and heart valves.32 This also is true of our case, because Fontana-Masson staining became negative when preceded by a bleaching procedure. Because past studies have employed several bleaching agents—hydrogen peroxide and potassium permanganate—and because the concentrations used in bleaching and other technical details are rarely provided, in our view, one cannot rely on the results of bleaching alone as proof of the presence or absence of melanin.

The evidence suggests that most examples of minocycline pigmentation—particularly types I and II—are due to cutaneous deposits of the drug or a metabolite thereof, chelated with iron.10,17,26,35 Clues to the mechanism of pigment deposition are provided by the studies of thyroid pigment by Enochs et al.36 Their in vitro modeling studies using electron paramagnetic resonance spectroscopy suggest that the pigment is a polymer caused by the in vivo oxidation of minocycline by thyroid peroxidase, which produces a melaninlike pigment.36 This pigment also contains significant amounts of iron, tightly bound in situ. A related phenomenon could well occur in the skin. Then, as suggested by Argenyi et al,10 the metabolite could act as a reducing substance, explaining the frequent positivity with the Fontana-Masson stain. It is possible that minocycline also may stimulate melanin production, accounting for the diffuse muddy brown type III pigmentation,17 but further studies are needed to clarify this point. The good news is that minocycline pigmentation resolves after cessation of therapy, though this may be a gradual process.17,19,25,37 

Conclusion

Minocycline therapy should be included in the differential diagnosis of hyperpigmented scars. Careful history taking and even repeated questioning may be necessary to elicit an accurate medication history. The pigmentation is most likely due to a minocycline metabolite, bound to iron; Fontana-Masson positivity may result from the action of reducing agents other than melanin. Slow resolution of the pigment can be expected following discontinuation of the drug. Nevertheless, biopsy is indicated when, as in this case, an atypical pigmented skin lesion raises concerns about malignant melanoma.

Scars may become pigmented for a variety of reasons, including the persistence and/or recurrence of an incompletely removed melanocytic nevus. However, development of an intensely hyperpigmented scar not long after a surgical procedure, in the absence of a clear explanation, would be a distinctly uncommon event. We recently encountered such a lesion in an otherwise healthy 20-year-old patient. In this case, the histopathologic findings led to further questioning of the patient and revealed a cause that had not been previously suspected. 

Case Report

A 20-year-old woman was seen for evaluation of a lesion on the left lower abdomen. Six weeks earlier, the lesion had been shave excised by an outside physician; pathology results were not initially available. The patient reported that the lesion had quadrupled in size and darkened considerably since the time of the excision. Her grandmother had died of malignant melanoma. She reported that her only medication was birth control pills. On physical examination, there was a 13X8-mm brown-black nodule with discrete but irregular borders (Figure 1). The clinical impression was recurrent nevus in a shave excision scar. However, because of the rapid growth, dark color, and family history of melanoma, there also was concern about the possibility of an atypical nevus or malignant melanoma. Therefore, an elliptical excision was performed. A report of the initial biopsy specimen was received, with the interpretation benign compound nevus.

Results of histopathologic evaluation of the reexcision specimen showed no residual melanocytic lesion. There was a prominent pigmented, cellular scar occupying the superficial to mid dermis in the central portion of the specimen. The pigmented material consisted of refractile, golden brown granules within macrophages and extracellularly, having a resemblance to hemosiderin (Figure 2). These granules stained positively with Perls stain for iron and with Fontana-Masson stain (Figure 3). Fontana-Masson staining was negative when performed after a bleaching procedure that employed potassium permanganate solution at a concentration of 3 g/L.

The staining results suggested the possibility of minocycline-related hyperpigmentation. Subsequent questioning of the patient revealed that she had been taking minocycline 100 mg twice daily during the 2 years prior to her clinic visit. 

Comment

Pigmented scars can arise occasionally because of a number of factors. The sites of persistent and/or recurrent nevus are often pigmented. This pigment, confined to the scar, often shows irregular borders and may have a mottled appearance.1 Pigmented scars also are observed in spontaneously regressing malignant melanoma.2 In a related phenomenon called tumoral melanosis, sheets of melanophages may accompany either a regressed melanoma or epithelial neoplasm.3,4 Pigmentation of scars related to hemorrhage also could occur, eg, following postsurgical trauma or in association with clotting abnormalities, though it is difficult to find literature directly addressing this problem. Other reported associations with hyperpigmented scars include leishmaniasis,5 chickenpox,6 burns,7 Addison disease,8 and hemosiderin-related pigmentation in endometriosis arising in cesarean scars.9 Among other agents that cause cutaneous pigmentation and could potentially produce hyperpigmented scars are heavy metals (eg, gold) and drugs such as amiodarone, phenothiazines, and antimalarials.10,11 Biopsy results of oral hyperpigmentation due to long-term antimalarial therapy have shown macrophages that contain melanin and ferric iron,12 findings resembling those reported here. None of these causes was pertinent to our case.

Minocycline first became available for clinical use in 1967. An association between minocycline administration and black discoloration of thyroid gland follicles in animals was reported that same year.13,14 As early as 1972, Velasco et al15 reported a macular pigmentation of the legs in patients receiving minocycline for the treatment of venereal disease. Since that time, there have been a number of reports of minocycline-induced pigmentation of skin and mucous membranes. Journal articles and textbooks usually divide minocycline-related cutaneous pigmentation into 3 major types. The first, type I, is a blue-black pigmentation that develops in areas of inflammation and scar13,16-19; this is the type that we report here. The second, type II, is a blue-gray pigmentation that develops particularly over otherwise normal-appearing skin of the arms, legs, or face.18,20,21 The third, type III, is usually described as a diffuse or generalized "muddy brown" pigmentation,13,22-25 though in one report this type of pigmentation was actually described as dark blue-gray.24 The Table provides a summary of the clinical and histopathologic changes associated with the 3 major types of minocycline pigmentation. Pigmentation of the nails and nail beds also occurs19,26 and has coexisted with diffuse cutaneous and scleral pigmentation.25 A fourth type of pigmentation that is not specific to minocycline results from fixed drug eruption, as described by Chu et al27 and possibly also represented by the case of Tanzi and Hecker.28 Minocycline also has been associated with discoloration of teeth,23 pigmented conjunctival cysts,29 and black galactorrhea,30 as well as pigmentation of internal organs such as cardiac valves.31,32

The duration of treatment and total dose required for minocycline to produce cutaneous pigmentation is difficult to determine. Although data on duration and total dose are often provided in reports, these figures typically reflect the totals at the time the patients present to their physician, rather than the time of actual onset of pigmentation, which is much more difficult to determine. Localized pigmentation at a site of tissue injury does not appear to be directly related to the duration of treatment18 and has been reported to occur as rapidly as 1 to 3 months following the onset of minocycline therapy.16,19 The evidence suggests that the diffuse type of pigmentation is more dependent on total dose and duration of therapy; reported patients have been on minocycline for about 3 years, with total doses ranging from 130 to 144 g.24,25

As generally described, there are differences among the microscopic features of the 3 major types of minocycline pigmentation. In type I, the dermal pigment is present in macrophages and stains positively for iron in a manner similar to hemosiderin.13,16,17 Type II pigmentation stains for iron and also is reactive with Fontana-Masson.10,20,33 Type III pigmentation has shown an increase in basilar melanin and brown-black pigment in macrophages that stains positively with Fontana-Masson and negatively for iron.24 However, staining results are not always distinctive among the 3 types. For example, in our patient's scar and in the inflammatory lesions of Ozog et al19 (examples of type I pigmentation), there was dermal pigment that stained positively both for iron and with the Fontana-Masson method. Patients also may have more than one type of cutaneous minocycline pigmentation. In the case of Pepine et al,25 there were areas of blue-black pigmentation, as well as muddy brown discoloration in sun-exposed areas. Biopsy results showed black pigment deposition in perivascular and periadnexal areas, though it is not entirely clear whether these specimens were obtained from blue-black or muddy brown areas.25 Electron microscopy in cases with blue-gray or blue-black pigmentation has shown electron-dense particles in macrophages or extracellularly. Some intracytoplasmic granules are present within lysosomes, while others, including fine dustlike particles consistent with ferritin, are not bound by lysosomal membranes.10,17,20,25 Energy dispersive x-ray microanalysis has shown that the granules mostly contain iron, with lesser amounts of calcium.21,26

The Fontana-Masson staining method is routinely employed to demonstrate the presence of melanin in tissue sections. Therefore, positivity in instances of minocycline pigmentation has suggested to some that melanin is at least partly responsible for the changes. This idea has been supported by one ultrastructural study showing melanosome complexes in siderosomes in a case of minocycline-related hyperpigmentation.21 However, melanosomes have not been identified in other studies.10 It is reported that iron may give positive reactions with Fontana-Masson staining.20 Furthermore, the black staining of Fontana-Masson results from the action of a reducing substance on ammoniated silver nitrate; that reducing substance is not necessarily melanin.10 The failure of the pigment to bleach, in contrast to the case with melanin, has been used to support the idea that the pigment in question does not contain melanin.10 However, reported results with bleaching have been variable. Successful bleaching or partial bleaching has been observed in examples of cutaneous minocycline pigmentation,19 as well as minocycline pigmentation of the thyroid gland34 and heart valves.32 This also is true of our case, because Fontana-Masson staining became negative when preceded by a bleaching procedure. Because past studies have employed several bleaching agents—hydrogen peroxide and potassium permanganate—and because the concentrations used in bleaching and other technical details are rarely provided, in our view, one cannot rely on the results of bleaching alone as proof of the presence or absence of melanin.

The evidence suggests that most examples of minocycline pigmentation—particularly types I and II—are due to cutaneous deposits of the drug or a metabolite thereof, chelated with iron.10,17,26,35 Clues to the mechanism of pigment deposition are provided by the studies of thyroid pigment by Enochs et al.36 Their in vitro modeling studies using electron paramagnetic resonance spectroscopy suggest that the pigment is a polymer caused by the in vivo oxidation of minocycline by thyroid peroxidase, which produces a melaninlike pigment.36 This pigment also contains significant amounts of iron, tightly bound in situ. A related phenomenon could well occur in the skin. Then, as suggested by Argenyi et al,10 the metabolite could act as a reducing substance, explaining the frequent positivity with the Fontana-Masson stain. It is possible that minocycline also may stimulate melanin production, accounting for the diffuse muddy brown type III pigmentation,17 but further studies are needed to clarify this point. The good news is that minocycline pigmentation resolves after cessation of therapy, though this may be a gradual process.17,19,25,37 

Conclusion

Minocycline therapy should be included in the differential diagnosis of hyperpigmented scars. Careful history taking and even repeated questioning may be necessary to elicit an accurate medication history. The pigmentation is most likely due to a minocycline metabolite, bound to iron; Fontana-Masson positivity may result from the action of reducing agents other than melanin. Slow resolution of the pigment can be expected following discontinuation of the drug. Nevertheless, biopsy is indicated when, as in this case, an atypical pigmented skin lesion raises concerns about malignant melanoma.