Antibiotic use and reduced effectiveness of second-line immunotherapy for lung cancer: all the time or just at the start of treatment?

It was with interest that we read the paper by Chalabi et al.1Chalabi M. Cardona A. Nagarkar D.R. et al.Efficacy of chemotherapy and atezolizumab in patients with non-small-cell lung cancer receiving antibiotics and proton pump inhibitors: pooled post hoc analyses of the OAK and POPLAR trials.Ann Oncol. 2020; 31: 525-531Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar in Annals of Oncology describing an unplanned analysis of two randomized clinical trials suggesting reduced overall survival when antibiotics and proton-pump inhibitors are used at the time of initiation of second-line immunotherapy for metastatic non-small-cell lung cancer. The rationale for their investigation was that these drugs modulate the microbiome which plays a key role in regulating the host innate and acquired immune response. That the association was less pronounced in the docetaxel group adds to the probability of causality. In the analysis, the use of antibiotics was defined as any use within a window of 30 days before and 30 days after the start of first study treatment. Considering that many patients with lung cancer are prescribed antibiotics during the course of their disease, this exposure definition does not answer the question do antibiotics have any impact when used while on immunotherapy (and not at start) and how long any impact may last. To explore this further we have conducted an additional analysis with the same source data as those used by Chalabi et al.1Chalabi M. Cardona A. Nagarkar D.R. et al.Efficacy of chemotherapy and atezolizumab in patients with non-small-cell lung cancer receiving antibiotics and proton pump inhibitors: pooled post hoc analyses of the OAK and POPLAR trials.Ann Oncol. 2020; 31: 525-531Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar In our analysis, we have explored the level of antibiotic use while on study treatment plus repeated the survival analysis with the so-called landmark method, wherein the study cohort was restricted to patients surviving up to at least 90 days and 180 days, respectively.2Anderson J.R. Cain K.C. Gelber R.D. Analysis of survival by tumor response.J Clin Oncol. 1983; 1: 710-719Crossref PubMed Scopus (898) Google Scholar Our additional analyses showed that the intensity of antibiotic use was much higher around the start of clinical trial study treatment compared with later on (Figure 1) and that at the two landmark points there was no association of antibiotics with overall survival after immunotherapy [hazard ratios: 1.05 (95% CI 0.79–1.39) and 0.83 (95% CI 0.56–1.23), respectively]. We consider our additional analyses relevant to further understand if the association of antibiotics with reduced immunotherapy effectiveness is likely to be causal. The high intensity of antibiotic use around the start of study treatment could be explained by symptoms of rapidly progressive disease at that time. Patients often receive antibiotics as an empirical way to relieve their symptoms or to treat pneumonia that occurs because of obstruction of the proximal airway.3Hsu-Kim C. Hoag J.B. Cheng G. Lund M.E. The microbiology of postobstructive pneumonia in lung cancer patients.J Bronchology Interv Pulmonol. 2013; 20: 266-270Crossref PubMed Scopus (12) Google Scholar Considering that patients with rapidly progressive disease have worse prognosis without early disease control, this might explain the temporal association of antibiotics with survival in the first months following start of immunotherapy. Some survival time is needed for an immune response to emerge, whereas early disease control is more probable with chemotherapy (i.e. docetaxel). Differential drop-out of the patients with worse prognosis in the first months can be the explanation why an association of antibiotics with survival cannot be detected later on. In our opinion, a next step to bring this important research question further would be to eliminate differences in early disease control from chemotherapy by replication of the analysis of Chalabi et al.1Chalabi M. Cardona A. Nagarkar D.R. et al.Efficacy of chemotherapy and atezolizumab in patients with non-small-cell lung cancer receiving antibiotics and proton pump inhibitors: pooled post hoc analyses of the OAK and POPLAR trials.Ann Oncol. 2020; 31: 525-531Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar with data from clinical trials comparing chemotherapy with the combination of chemotherapy and immunotherapy in a first-line setting. Data from the OAK (trial reference number: NCT02008227) and POPLAR trials (trial reference number: NCT01903993) were accessed according to Roche's policy and process for clinical study data sharing. Data analysis was supported by the foundation Personalized Healthcare Catalyst Alliance that aims to improve the personalized health care in The Netherlands. This work was not supported by funding.