Risk of bleeding and significant thrombocytopenia in cancer patients treated with immune checkpoint inhibitors.

Abstract

e15000 Background: Immune checkpoints inhibitors (ICIs) are associated with multiple side effects. Complications can affect any organ system including; endocrine, gastrointestinal, pulmonary, and less frequently hematological system.We evaluated the risk of bleeding and significant thrombocytopenia in cancer patients (pts) treated with ICIs. Methods: Cancer pts treated with ICIs between 01/2015 and 7/2019 at Cleveland Clinic Foundation were retrospectively reviewed. Platelets (plts) count and hemoglobin (hb) level at baseline, after one month and 3 months of treatment were reviewed. Thrombocytopenia was defined as plts count < 150,000. In pts with thrombocytopenia at baseline, more than 50% drop in the plts count was considered significant. Pts were followed-up for one year after ICIs treatment initiation for significant bleeding. Results: We identified 555 cancer pts treated with ICIs. Of these pts 65% were males. Among all included pts, 53% had lung cancer, 23% had genitourinary cancers, 17% had melanoma, 3% had gastrointestinal cancers, and 5% had other cancer types. Nivolumab, pembrolizumab, ipilimumab, and atezolizumab were used in 62%, 27%, 14%, and 12% respectively. Clinically significant bleeding, thrombocytopenia after one month and 3 months of treatment were identified in 21%, 6%, and 7% respectively. Of all pts who developed bleeding, 8% had gastrointestinal bleeding, 5% had intracerebral hemorrhage, and 3 % had hemoptysis. Only 23% of the pts who developed bleeding were receiving anticoagulation therapy. In multivariate analysis, breast cancer as primary malignancy (p = .03) and atezolizumab (p = 0.02) were predictors of clinically significant bleeding. Also, nivolumab was associated with increased risk of significant thrombocytopenia after one month of treatment (p = .02). There was no significant association between the use of specific medication and the development of significant anemia. Median overall survival time was not statistically different in patients who had bleeding compared to no bleeding group, 12.6 months (95% CI:10.8-14.5) and 12.4 months (95% CI:10.3-15) respectively (p = .21). Conclusions: In this cohort study, atezolizumab and breast cancer were significantly associated with increased risk of bleeding. The use of nivolumab was associated with a greater risk of thrombocytopenia after one month of treatment. Prospective studies are needed for better understanding of the hematological side effects of ICIs.