The incidence and relative risk of PD-1/PD-L1 inhibitors-related colitis in non-small cell lung cancer: A meta-analysis of randomized controlled trials

The programmed cell death-1 (PD-1)/programmed cell death ligand-1 (PD-L1) inhibitors have shown encouraging merits in non-small cell lung cancer (NSCLC) patients, however, they are often related to potentially fatal immune-related adverse events (irAEs) including colitis. Considering the incidence and characteristics of immune-related colitis may have significant implications for the appropriate utilization of PD-1/PD-L1 inhibitors in clinical practice, we conduct this meta to systematically analyze the correlation between PD-1/PD-L1 inhibitors for the treatment of NSCLC and the incidence of immune-associated colitis.Electronic databases including PubMed, Embase, Cochrane Library and ClinicalTrials.gov (http://clinicaltrials.gov/) were searched up to May 2019, clinical trials reporting all grade (1-5), higher grade (3-5) colitis and grade 3-5 diarrhea were included, data were expressed as relative risk (RR), incidence, corresponding p value and 95% confidence intervals (CIs).9 randomized controlled trials (RCTs) were identified (7 with PD-1 inhibitors [n = 4526]) and 2 with PD-L1 inhibitors [n = 1464]). The overall incidence of PD-1/PD-L1 target agents was 1.40% for all grade colitis, 0.89% for severe colitis, 11.62% for all grade diarrhea and 1.36% for severe diarrhea. Compared with chemotherapy group, the PD-1/PD-L1 inhibitors had a significantly higher risk of all grade (RR: 3.68, p < 0.001) and high-grade (RR: 2.97, p = 0.01) colitis. Additional analysis of relative risk of diarrhea revealed that PD-1/PD-L1 treatment moderately reduce the risk of all grade diarrhea (RR: 0.64, p = 0.03), while the difference was not statistically significant in the risk of grade 3-5 diarrhea (RR: 0.83, p = 0.64). Subgroup analyses showed that the RR of all grade and higher grade colitis in PD-1 inhibitors was more significant (RR: 3.56, p = 0.001 vs RR: 2.98, p = 0.02 respectively). However, there was no appreciable difference in PD-L1 inhibitors (RR: 4.75, p = 0.15 vs RR: 2.85, p = 0.52 respectively). When compared with first-line therapy, second-line therapy associated with a higher risk of all grade colitis than first-line therapy (RR: 3.29, p = 0.006; RR: 4.69, p = 0.026).Our meta-analysis indicates when compared with control group, the PD-1/PD-L1 inhibitors may lead to a higher risk of all grade and high grade immune-mediated colitis, but may result in a reduction in all grade diarrhea. PD-1 inhibitors in NSCLC patients, but not PD-L1 inhibitors, increase the risk of all- and high grade colitis. These results suggest that clinicians shall pay more attention to this rare but life-threatening toxic effect.