PURPOSE The risk of immune checkpoint inhibitor therapy-related GI adverse events in patients with cancer and inflammatory bowel disease (IBD) has not been well described. We characterized GI adverse events in patients with underlying IBD who received immune checkpoint inhibitors. PATIENTS AND METHODSWe performed a multicenter, retrospective study of patients with documented IBD who received immune checkpoint inhibitor therapy between January 2010 and February 2019. Backward selection and multivariate logistic regression were conducted to assess risk of GI adverse events. RESULTSOf the 102 included patients, 17 received therapy targeting cytotoxic T-lymphocyte antigen-4, and 85 received monotherapy targeting programmed cell death 1 or its ligand. Half of the patients had Crohn's disease, and half had ulcerative colitis. The median time from last active IBD episode to immunotherapy initiation was 5 years (interquartile range, 3-12 years). Forty-three patients were not receiving treatment of IBD. GI adverse events occurred in 42 patients (41%) after a median of 62 days (interquartile range, 33-123 days), a rate higher than that among similar patients without underlying IBD who were treated at centers participating in the study (11%; P , .001). GI events among patients with IBD included grade 3 or 4 diarrhea in 21 patients (21%). Four patients experienced colonic perforation, 2 of whom required surgery. No GI adverse event-related deaths were recorded. Anti-cytotoxic T-lymphocyte antigen-4 therapy was associated with increased risk of GI adverse events on univariable but not multivariable analysis (odds ratio, 3.19; 95% CI, 1.8 to 9.48; P = .037; and odds ratio, 4.72; 95% CI, 0.95 to 23.53; P = .058, respectively).CONCLUSION Preexisting IBD increases the risk of severe GI adverse events in patients treated with immune checkpoint inhibitors.

The influence of childhood pertussis infection and of purified pertussis toxin on histamine release from human basophil leucocytes was investigated. Three different stimuli, the peptide N-formyl-Met-Phe (NFMP), anti-IgE, and the calciumionophore A23187 were used to challenge the cells. When NFMP was the stimulus, histamine release in the control group (age 0.5-17 years) increased in an age-dependent fashion, whereas anti-IgE and A23187 stimulated release did not vary with age. During the convulsive state of pertussis infection there was a significant reduction of histamine release in response to 10 microM NFMP (from 9.5 +/- 1.4 [n = 21] to 6.7 +/- 1.5 [n = 19], P < 0.05) and in response to 800 and 80 U/ml anti-IgE (from 28.5 +/- 5 [n = 19] to 16.3 +/- 5 [n = 13], P < 0.05, and from 6.9 +/- 1.7 [n = 16] to 2 +/- 0.8 [n = 13], P < 0.01), whereas histamine release stimulated by A23187 was unchanged compared to release in control children. In vitro pretreatment of basophils from healthy children and adults with pertussis toxin also inhibited histamine release. When NFMP was the stimulus, release was completely blocked by pertussis toxin with an IC50 of about 11 ng/ml, whereas anti-IgE stimulated release was only inhibited by 20%-30% and release induced by A23187 was reduced to 40%-50% by toxin treatment. In conclusion we have demonstrated a functional impairment of histamine release during the convulsive state of pertussis and that this inhibition is likely to be mediated by pertussis toxin.