Symptomatic Bilirubin Gallstones in Patients with PNH Treated with Eculizumab

Paroxysmal Nocturnal Hemoglobinuria is a stem cell disorder characterized by an acquired somatic mutation in the PIG-A gene. This results in a loss of GPI-linked complement inhibitors, CD55 and CD59, on the surface of red blood cells. Untreated patients experience chronic intravascular hemolysis and hemoglobinuria mediated by complement, as well as acute attacks. The lack of these complement inhibitors on platelets may lead to their activation and may explain the marked hypercoagulable state in this disorder. Eculizumab, a humanized monoclonal antibody targeting the C5 component of complement, prevents the formation of the C5b-9 complex and results in a major reduction in red cell transfusion requirements, a marked decrease in thrombotic events, and an improvement in quality of life. It was recognized soon after the introduction of this therapy that patients would often maintain an elevation in the reticulocyte count despite marked improvements in the LDH, and that some patients would have only a partial normalization of the hemoglobin--in some cases accompanied by an increase in bilirubin levels. It is now understood that in patients treated with C5 inhibition, C3 degradation products can accumulate on the red cell surface, resulting in opsonization and extravascular hemolysis. Based on this model, it would be expected that treated patients with PNH would now resemble those with other forms of extravascular hemolysis, with some of the same complications. Indeed, a rising ferritin has been reported in some patients, but until now, there have been only a few case reports of gallstone formation, and only one report commented on whether the stones were comprised of bilirubin. Here we review a set of 51 patients with PNH (31 women and 20 men) treated with eculizumab for a total of 280 patient years. There have been 7 cases of symptomatic cholelithiasis requiring surgery (5 women and 2 men), as well as one woman with a radiologically abnormal gallbladder removed at the time of a hemicolectomy for colon cancer. In four of these cases we have photographs of the surgical specimen: 3 patients had dark stones and/or debris, and one patient had stones with a mixed bilirubin/cholesterol appearance. For patients who had cholecystectomies at outside institutions, the pathology report indicated small bilirubin stones in one case, there was a verbal report of dark gallstones in another, and in two cases the nature of the stones is not known. Based on a Kaplan-Meier curve, the risk of gallstones leading to a cholecystectomy after 10 years of treatment was estimated to be 22.5% (95% CI 10%-46%) and did not differ based on gender. We then considered that markers of hemolysis might predict gallstone disease. The median bilirubin was only slightly higher in those with gallstones (2.1 vs 1.6 mg/dl, p = 0.06, one sided rank sum test), the median reticulocyte count was somewhat higher (303,500 versus 206,000 per μl, p=0.04, one sided rank sum test), and the median ferritin levels were approximately the same in both groups (361 versus 291, normal 10-143 ng/ml p=0.16). The median age at the time of the event or last follow-up was 46 years in both groups. All of the patients have recovered after their surgery. However, one of these patients had a complicated intraoperative course due to significant varices surrounding the gallbladder associated with cavernous transformation of the portal vein, due to prior thrombotic complications of PNH. This patient had presented with severe symptoms of cholelithiasis, including bacteremia from a biliary source. Another patient required a cholecystostomy tube due to severe gallbladder wall inflammation several weeks before the operation was performed. Another patient developed biochemical evidence of pancreatitis shortly before the procedure. We believe that bilirubin stones, a predictable consequence of extravascular hemolysis, represent, in addition to meningococcal infection and iron overload, a potentially significant side effect of inhibition of C5 in patients with PNH. This should be discussed with patients prior to initiation of therapy, and physicians should have a low threshold for obtaining appropriate imaging if a patient on a C5 inhibitor experiences complaints or laboratory evidence suggestive of gallstone disease.