Erythropoietin and painful leg ulcers: Thrombosis or vasculitis?
Recombinant human erythropoietin (rHuEPO) is used routinely in the management of anemia in end-stage renal failure. We report a case of thrombotic leg ulceration associated with rHuEPO use in a 69-year-old man with long-standing but inactive proteinase 3 antineutrophil cytoplasmic antibody–positive Wegener's granulomatosis, nondialysis-dependent renal failure, and secondary transfusion-dependent anemia.
The patient developed painful leg ulcers 12 weeks after starting rHuEPO. While receiving rHuEPO treatment the patient's hemoglobin gradually increased over 8 weeks from 70 gm/dl to 120 gm/dl. Other medical history included dilated cardiomyopathy with atrial fibrillation, hypertension, and late-onset diabetes mellitus, all of which were well controlled with treatment.
Histology of an initial biopsy sample from an ulcer crater and adjacent tissue suggested an acute necrotizing vasculitis. Because there were no clinical or laboratory signs of vasculitic disease related to Wegener's granulomatosis, a second biopsy sample from a more recent nonulcerated lesion was obtained. This second histology report read: “… Beneath the infarcted epidermis there is marked vascular congestion within the superficial dermis accompanied by fresh hemorrhage. Many of these capillaries are occluded by apparent thrombus. There are no significant associated inflammatory changes … This biopsy shows infarction of the skin secondary to occlusion of superficial capillaries by apparent thrombus …”
No coagulopathy or cryoglobulinemia was detected. The patient was started on warfarin therapy and his ulcers healed. He returned to working on his farm and has had no ulcer recurrence while continuing rHuEPO therapy.
Although the etiology of thrombotic ulceration in this patient was probably multifactorial, the prothrombotic effects of rHuEPO were likely to have contributed because his comorbid conditions were controlled. Rapid increases in hematocrit and direct effects on platelet aggregation have been postulated as potential prothrombotic mechanisms of rHuEPO (1–4) along with changes in protein concentrations including factor VIII, von Willebrand factor–associated coagulant activity (5), protein C, and protein S. In vitro, EPO induces endothelial secretion of soluble adhesion molecules, E-selectin, and vascular cell adhesion molecule, suggesting that EPO therapy may also activate vascular endothelium in vivo. In susceptible individuals these changes may combine to put the individual into a prothrombotic state.
Chronic ulceration of varying etiology is frequently associated with accumulation of acute and chronic inflammatory cells around vessels. It is important to not obtain biopsy specimens from within or immediately adjacent to an ulcer to minimize the risk of misinterpretation. On review, the first biopsy specimen had the same pattern of thrombosed vessels as seen on the second biopsy specimen, but this finding was partially obscured by a nonvasculitic inflammatory infiltrate. This aspect of the case also illustrates the importance of obtaining appropriate biopsy material to avoid confusion between primary thrombotic and primary vasculitic disease.