The negative effect of concomitant medications on immunotherapy in non-small cell lung cancer: An umbrella review

•Previously inconsistent results were unified and new insights emerged.•OS was shortened using PPIs before or after the initiation of immunotherapy in NSCLC.•OS was shortened using steroids during the first course of immunotherapy in NSCLC.•The results of antibiotics weren’t supported by convincing evidence surprisingly.•The effect of opioids and NSAIDs on ICIs was similar to previous results. Conflicting results about the effect of concomitant medications on immunotherapy in non-small cell lung cancer (NSCLC) were reported by many meta-analyses (MAs), and the certainty of evidence linking concomitant medications with immunotherapy efficacy has not been quantified, which may cause some evidence to be misinterpreted. Four databases including Embase, Cochrane Library, PubMed, and Web of Science were searched from inception to January 2023 in English. Based on prospective or retrospective clinical controlled trials including immunotherapy with concomitant medications or not in NSCLC, quantitative MAs reporting the efficacy of immunotherapy with binary direct comparison and enough extractable data were collected. The methodological quality, reporting quality, and risk of bias of included MAs were evaluated respectively. New meta-analyses were conducted and their evidence certainty was classified as nonsignificant, weak, suggestive, highly suggestive, or convincing. Fifteen MAs with 5 medications were included. After being assessed by AMSTAR-2, PRISMA, and ROBIS, the major shortcomings were focused on the registration of protocol, literature retrieval or data extraction, implementation of sensitivity analysis or evidence certainty assessment, and incomplete reporting in the section of method and result. New pooled analyses indicated that antibiotics (HR = 1.545[1.318–1.811]), steroids (HR = 1.784[1.520–2.093]), proton pump inhibitors (PPIs) (HR = 1.303[1.048–1.621]) and opioids (HR = 1.910[1.213–3.006]) could shorten overall survival (OS) in patients with NSCLC receiving immunotherapy. Besides, antibiotics (HR = 1.285[1.129–1.462]) and steroids (HR = 1.613[1.315–1.979]) were harmful to progression-free survival (PFS) in these patients significantly. No negative effect was found in nonsteroidal anti-inflammatory drugs and the objective response rate of all medications. High-level evidence suggested that using PPIs before or after the initiation of immunotherapy and using steroids during the first-course immunotherapy could weaken the OS of patients with NSCLC. M