Complement inhibitors (CI) are the mainstay therapy for classic complement-mediated disorders including paroxysmal nocturnal hemoglobinuria (PNH) and atypical hemolytic uremic syndrome (aHUS). Emerging recognition of the role of complement dysregulation in autoimmune neurologic disorders has also expanded CI use in diverse disease states. Recent evidence suggests complement may have both pro and anti-tumoral effects; thus, it is unclear if complement inhibition may increase cancer risk. We conducted a retrospective study of patients treated with CI at our institution to examine rates and trends in cancer incidence with long-term CI use.

A total of 190 adult patients who received CI in the past 20 years were identified. Patients who received only a single dose or limited course of ≤2 weeks (n=40) were excluded. Our overall cohort of 150 had 65 (43.3%) males and 85 (56.7%) females with a median age of 46 years. The most common indications for CIs included aHUS (32%), PNH (28.7%), myasthenia gravis (16%), neuromyelitis optica spectrum disorder (8%), and macular degeneration (4%). The remaining indications included transplant-associated complications, rare hemolytic conditions, and primary glomerular diseases.

Of the 118 patients (78.7%) who never developed cancer, the median time on CI was 40 months. Thirty-two (27.1%) had a family history of cancer,32 (27.1%) were current or former smokers, 4 (3.39%) had history of heavy alcohol use (>7 drinks/week in females and >14 in males), 2 (1.69%) had history of HIV, 3 (2.54%) had history of hepatitis B, 2 (1.69%) had history of hepatitis C, 8 (6.78%) had history of high-risk human papillomavirus (HPV), 87 (73.7%) had history of immunosuppressive drug use, and 38 (32.2%) had BMI ≥30 at CI initiation.

Thirty-two (21.3%) patients had a diagnosis of cancer with 12 (8%) diagnosed after CI initiation. Of these 12, median duration on CI was 59.5 months and median time from CI use to cancer diagnosis was 30 months; one had an unknown cancer diagnosis date and was excluded from the latter calculation. Six were treated with eculizumab and ravulizumab, 4 with eculizumab alone, 1 with ravulizumab alone, and 1 with eculizumab, ravulizumab, pegcetacoplan, and iptacopan at various points. Cancer types varied with breast, kidney, thyroid, prostate, lung, skin, kaposi sarcoma, bladder, and neuroendocrine cancers all seen. Breast cancer incidence was highest (n=4), but all patients were >40 years old at diagnosis. Six of the 12 had a family history of cancer, 3 were former smokers, 2 had a history of heavy alcohol use, 1 had a history of HIV, 1 had a history of high-risk HPV, 9 had a history of immunosuppressive drug use, and 3 had BMI ≥30 at CI initiation.

A separate, partially overlapping cohort analysis was performed on 89 patients with PNH, of whom 49 (55.1%) received CI and 40 (44.9%) did not. Four of the 49 (8.2%) on CI developed cancer after drug initiation while 6/40 (15%) not on CI developed cancer. The median duration on CI was 124.5 months and median time from CI to cancer diagnosis was 48 months. In the 10 PNH patients with cancer, the percent with cancer risk factors in the CI-treated group compared to those not treated with CI were as follows: family history of cancer (50 vs 66.7%), history of heavy alcohol use (25 vs 0%), any tobacco use (0 vs 16.7%), high-risk HPV (25 vs 0%), and BMI ≥30 (50 vs 0%). A cancer-free survival (CFS) analysis was performed, wherein a univariable Cox regression analysis compared time to cancer diagnosis between PNH patients who received CI and those who did not. Patients diagnosed with cancer before CI initiation or who received only a single dose or limited course of CI ≤2 weeks (n=6) were excluded from statistical analysis. Cox regression analysis evaluating CFS stratified by CI use in 83 patients with PNH produced a hazards ratio of 1.08 and p-value of 0.91.

In conclusion, CI use does not appear to be a significant predictor of cancer incidence in our cohort. For patients with PNH, CI use did not have a significant effect on cancer-free survival with patients receiving CI having 1.08 times the hazard of cancer incidence compared to those not receiving CI. Factors such as family history, smoking/alcohol use, and past medical history of HPV, hepatitis B, hepatitis C, and HIV remain established predictors for cancer incidence. Further analysis in a larger cohort is needed to validate these initial results.

This content is only available as a PDF.
2025
Sign in via your Institution