Risk of Overall and Cancer Mortality After Immunosuppression of Patients with Non-infectious Ocular Inflammatory Diseases.

Abstract

Purpose : To test whether use of immunosuppressive agents for ocular inflammatory diseases is associated with an altered risk of overall and cancer mortality.

Methods : Retrospective cohort study of all eligible patients seen at five ocular inflammation centers plus Tumor Necrosis Factor Inhibitor (TNF-I)-treated patients and matched controls at five other centers from inception through 31 December 2010. Patients with non-infectious ocular inflammation (uveitis, scleritis, pemphigoid with eye involvement, or others) were included unless they had HIV infection or (for this analysis) cancer. Overall and cancer-specific mortality was studied, obtained by linkage to the National Death Index (NDI) (1979-2013); death was attributed to cancer when the NDI indicated cancer was identified on the death certificate as cause of death. Exposures of primary interest were use of TNF-I, antimetabolites, calcineurin inhibitors, and alkylating agents. Exposure data were obtained by retrospective chart review, entering data on every patient at every visit. Relative hazard (RH) of overall or cancer mortality for the exposures of interest was assessed by survival analysis, adjusting for additional predictive factors, either including all patients or excluding those with systemic inflammatory diseases (who may have increased risk of death or cancer).

Results : Among 15,938 patients followed over 187,151 person-years, 1970 died (435 cancer deaths). Counts (Median follow-up time in years) of patients exposed to immunosuppression include: TNF-I: 1,333 (6.3); antimetabolites: 4,730 (8.2); calcineurin inhibitors: 1,558 (9.8); alkylating agents: 892 (9.4). RH’s are summarized in the forest plot images. Results excluding the first three to five years after exposure generally were similar; no cumulative dose or maximum dose thresholds with different outcome patterns were found.

Conclusions : The immunosuppressive agents commonly used for ocular inflammatory diseases, including TNF-I’s, were not associated with increased overall or cancer mortality. Alkylating agents were associated with higher overall mortality, but not after excluding patients with systemic inflammatory diseases (some with high mortality). Higher mortality in some tacrolimus and etanercept analyses might reflect selection factors (e.g., transplantation) and require further study.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.