A retrospective cohort demonstrated that patients who are treated with antibiotics within 1 year before immune checkpoint inhibitor therapy may experience worse overall survival.
Findings from a population-level retrospective cohort study published in the Journal of Clinical Oncology demonstrated that patients with cancer who are treated with antibiotics within 1 year before immune checkpoint inhibitor (ICI) therapy may experience worse overall survival (OS) outcomes due to alteration in the gut microbiome. Antibiotic class analysis concluded that fluoroquinolones given within 1 year or 60 days before an ICI were associated with worse OS and a dose effect was observed; patients receiving anti–PD-1 ICIs and those with lung cancer or melanoma also experienced poorer OS.
Overall, the median OS was 302 days for patients (n = 2737) in the final cohort. Comparatively, patients who received antibiotics 1 year prior to beginning therapy with an ICI (n = 1627) had a worse OS with an adjusted HR (aHR) of 1.12 (95% CI, 1.01-1.24; P = .03). Investigators noted that “associations were significant for any antibiotic exposure within 1 year of starting ICIs, while a nonsignificant trend was seen for 60 days, and no significant associations were seen for exposure within 30 days.”
All patients included in the study were aged 65 years and older and 59% filled at least one antibiotic prescription in the year prior to starting ICI therapy with a median duration of 2 weeks (interquartile range [IQR], 1-3.4) and a median of 30 total doses (IQR, 18-60); 19% of patients filled one or more prescriptions 60 days prior to starting therapy (n = 532) with a median duration of 1.3 weeks (IQR, 1-2) and a median of 21 total doses (IQR, 14-36). Patients in the 2 groups received penicillin (29% vs 7%), cephalosporin (24% vs 5%), and fluoroquinolone (28% vs 7%), respectively.
Although investigators concluded that the antibiotics penicillin and cephalosporin were not associated with OS, patients who received fluoroquinolone had worse OS; those starting 1 year before ICIs (n = 771) had an aHR of 1.26 (95% CI, 1.13-1.40; P < .001) and patients starting 60 days prior (n = 188) had an aHR of 1.20 (95% CI, 0.99-1.45; P = .06). Dose effects were seen on the basis of total weeks of exposure for over 1 year with an aHR of 1.07 per week (95% CI, 1.03-1.11; P < .001) and for 60 days with an aHR of 1.12 per week (95% CI, 1.03-1.23; P = .01).
The study authors explained that “fluoroquinolones can alter many gut microbiota species including Faecalibacterium, Ruminococcus, Bifidobacteria, and Alistipes, which have all been found to affect ICI responses. Fluoroquinolones also have immunomodulatory activities that can influence cytokine release and the immune response, which in turn may affect response to ICI.”
Findings from a subgroup analysis showed that patients who were treated with the anti–PD-1 ICIs pembrolizumab and nivolumab (n = 2297) and received antibiotics 1 year before starting had a worse OS with an aHR of 1.14 (95% CI, 1.02-1.27; P = .02); a dose effect was recorded on the basis of the number of weeks of exposure to antibiotics 1 year before an ICI with an aHR of 1.01 per week (95% CI, 1.00-1.02; P = .007). Similarly, investigators noted that a nonsignificant trend was observed for weeks of exposure 60 days before starting ICI (aHR, 1.04 per week; 95% CI, 1.00-1.09; P = .06).
For patients who started an anti–PD-1 ICI, penicillin and cephalosporin were not associated with a poorer OS, but those who received fluoroquinolone within 1 year had an aHR of 1.28 (95% CI, 1.15-1.44; P < .001) and those in the 60-day group had an aHR of 1.19 (95% CI, 0.98-1.44; P = .08). Dose effects of an aHR of 1.08 per week (95% CI, 1.04-1.11; P < .001) and an aHR of 1.12 (95% CI, 1.02-1.23; P = .02) were observed for weeks of exposure, respectively.
An improvement in terms of OS was observed for patients who received the anti–CTLA-4 ICI ipilimumab (n = 362)and more weeks of antibiotics within 1 year before starting ICI therapy (aHR, 0.92; 95% CI, 0.86-0.99; P = .03). Furthermore, lung cancer and melanoma were disease sites that were associated with worse OS outcomes while there were no association with renal cancer and poorer OS.
Limitations of the study included that it was focused on antibiotic prescription claims from the Ontario Drug Benefit database, which may not reflect antibiotic adherence and use. Additionally, the claims were limited to patients receiving outpatient antibiotics, where the majority are oral antibiotics, but this was adjusted for.
Additionally, patients were 65 years old and above which may not reflect the overall population of patients with cancer, 1/3 of patients who received antibiotics 1 year before starting ICIs had potentially been exposed before their cancer diagnosis, and the study did not adjust for cancer stage. Further studies regarding gut microbiome manipulation may assist in understanding how antibiotic exposure impacts ICI outcomes.
Patients were mostly males (60%), had a median age of 73 years (IQR, 69-78) at the start of immunotherapy, a median of 357 days (IQR 147-861) from cancer diagnosis to start of immunotherapy, lived in rural areas (14%), had a median BMI of 26.1 (range, 15-56), a history of autoimmune disease (8%), a recent hospitalization within 1 year of starting immunotherapy (47%), and a recent hospitalization within 60 days of starting immunotherapy (20%).
The most common disease sites were lung (53%), melanoma (34%), renal (7%), bladder (3%), head and neck (2%), and other (2%). The most common first checkpoint inhibitors received were nivolumab (43%), pembrolizumab (41%), and ipilimumab (13%).
The study took place from June 2012 to October 2018 in Ontario, Canada, and used systemic therapy administration data from the New Drug Funding Program and the Cancer Activity Level Reporting databases to identify patients. The Ontario Drug Benefit database identified antibiotic prescriptions. The multivariable analysis conducted included sex, age, body mass index, if a hospitalization occurred within the previous year, history of autoimmune conditions, the facility level, and the John Hopkins Adjusted Clinical Groups score.
Eng L, Sutradhar R, Niu Y, et al. Impact of antibiotic exposure before immune checkpoint inhibitor treatment on overall survival in older adults with cancer: a population-based study. J Clin Oncol. Published online February 24, 2023. doi:10.1200/JCO.22.00074