Association Between Prior Antibiotic Treatment and Immune Checkpoint Inhibitor Therapy Outcomes

Get Permission

In a study presented in a brief report in JAMA Oncology, David J. Pinato, MD, PhD, and colleagues found that past but not concurrent use of broad-spectrum antibiotic therapy was associated with poorer treatment outcomes in patients receiving immune checkpoint inhibitors for various cancers in routine clinical practice.

David J. Pinato, MD, PhD

David J. Pinato, MD, PhD

As stated by the investigators, “Gut dysbiosis impairs response to immune checkpoint inhibitors and can be caused by broad-spectrum antibiotic therapy.”

Study Details

The study involved 196 patients with non–small cell lung cancer (NSCLC, n = 119), melanoma (n = 38), and other tumor types (n = 39) consecutively treated with immune checkpoint inhibitor therapy in routine clinical practice at two academic medical centers between January 2015 and January 2018. Past broad-spectrum antibiotic use was defined as use up to 30 days before start of immune checkpoint inhibitor treatment. Disease refractory to immune checkpoint inhibitor therapy was defined as progressive disease 6 to 8 weeks after first immune checkpoint inhibitor dose without evidence of pseudoprogression. Most patients (84%) had disseminated disease at the start of immune checkpoint inhibitor therapy, a performance status of 0 to 1 (81%), and received anti­–programmed cell death protein 1/programmed cell death ligand 1 (PD-1/PD-L1) immune checkpoint inhibitors (96%) as first-line therapy (62%).

Overall, 29 patients had past broad-spectrum antibiotic treatment (26 for ≤ 7 days) and 68 received antibiotic treatment concurrent with immune checkpoint inhibitor treatment (39 for ≤ 7 days). Most patients received β-lactam-based antibiotics (20 in the past treatment group and 46 in the concurrent treatment group). Past antibiotic use was not associated with corticosteroid use up to 30 days prior to or during immune checkpoint inhibitor therapy.

Antibiotic Use and Outcomes


  • Past antibiotic therapy but not concurrent antibiotic therapy was associated with poorer response and overall survival.
  • Poorer overall survival with past antibiotic use was observed in patients with NSCLC, melanoma, and other tumor types.

Prior antibiotic therapy (hazard ratio [HR] = 7.4, P < .001), but not concurrent antibiotic therapy (HR = 0.9, P = .76), was associated with significantly poorer overall survival vs no antibiotic therapy. Median overall survival was 2 months vs 26 months with prior antibiotic treatment vs no prior antibiotic treatment among all patients. Median overall survival was poorer with vs without prior antibiotic treatment among patients with NSCLC (2.5 vs 26 months, P < .001), melanoma (3.9 vs 14 months, P < .001), and other tumor types (1.1 vs 11, P < .001).

Among evaluable patients, poor response to immune checkpoint inhibitor treatment was observed in 81% of the past antibiotic treatment group vs 44% of the no past antibiotic treatment group (P < .001). Objective response vs disease progression was observed in 5 vs 21 patients with past antibiotic use and in 33 vs 33 patients with concurrent antibiotic use.

On multivariate analysis, past antibiotic treatment (HR = 3.4, P < .001) and no response to immune checkpoint inhibitor therapy (HR = 8.2, P < .001) were associated with poorer overall survival, independent of tumor site, disease burden, and performance status.

The investigators concluded, “Despite being limited by sample size, geographic origin, and the lack of correlative analyses on patients’ gut microbiota, this study suggests that [past antibiotic] therapy but not [concurrent antibiotic] therapy is associated with a worse treatment response and [overall survival] in unselected patients treated with immune checkpoint inhibitors in routine clinical practice. Mechanistic studies are urgently required to investigate [antibiotic]-mediated alterations of gut microbiota as a determinant of poorer outcome following immune checkpoint inhibitor treatment.”

Dr. Pinato, of Imperial College London, is the corresponding author for the JAMA Oncology article.

Disclosure: The study was supported by the Imperial College National Institute for Health Research Biomedical Research Centre, Imperial College Tissue Bank, and Imperial Cancer Research UK Centre. For full disclosures of the study authors, visit

The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.