Proton Pump Inhibitors Linked to Reduced Durvalumab Benefit in Stage III NSCLC: Post-Hoc Analysis of PACIFIC Trial
Key Insights
A post-hoc analysis of the PACIFIC trial found that baseline proton pump inhibitor (PPI) exposure was associated with significantly shorter PFS and OS in durvalumab-treated patients with unresectable stage III NSCLC.
Median PFS was 9.4 months with PPI exposure versus 17.2 months without (HR 1.57, p<0.0001), and median OS was 33.0 versus 57.9 months (HR 1.66, p<0.0001).
Baseline antibiotic exposure was associated with shorter PFS (HR 1.50, p=0.016) but not a statistically significant difference in overall survival, and no significant treatment interaction was observed.
A new post-hoc analysis of the landmark PACIFIC trial has revealed that commonly used medications—proton pump inhibitors (search) (PPIs) and systemic antibiotics (search)—may be associated with diminished clinical benefit from consolidation durvalumab in patients with unresectable stage III non-small-cell lung cancer (search) (NSCLC). The findings, published by an international collaborative team led by Alessio Cortellini, raise important questions about how concomitant medications that alter the gut microbiome might influence immune checkpoint inhibitor (ICI) efficacy in the potentially curative setting.
The analysis included 660 of the 713 patients originally randomized in the PACIFIC trial, a phase 3 study that established durvalumab after concurrent platinum-based chemoradiotherapy as standard of care for unresectable stage III NSCLC. Baseline medication exposure was defined as oral or intravenous use within 30 days prior to randomization. Overall, 263 patients (40%) had baseline PPI exposure and 69 patients (10%) had baseline antibiotic exposure, with a median follow-up of 62.4 months.
PPI Exposure and Durvalumab Outcomes
Among patients treated with durvalumab, baseline PPI exposure was associated with significantly shorter progression-free survival (PFS). Median PFS was 9.4 months in patients with PPI exposure compared with 17.2 months in those without (HR 1.57; 95% CI, 1.28–1.93; p<0.0001). The impact on overall survival (OS) was similarly pronounced: median OS was 33.0 months with PPI exposure versus 57.9 months without (HR 1.66; 95% CI, 1.30–2.13; p<0.0001).
After adjustment for baseline clinical and pathological variables, PPI exposure remained associated with increased risk of both progression and death in durvalumab-treated patients. Critically, the same pattern was not observed in the placebo group. The treatment-by-PPI interaction was statistically significant for both PFS and OS, suggesting that the association is specifically relevant in the context of durvalumab rather than serving as a general marker of poorer prognosis.
Antibiotic Exposure: A Less Consistent Signal
Baseline antibiotic exposure was also associated with shorter PFS in the durvalumab group. Median PFS was 9.2 months with antibiotic exposure versus 15.6 months without (HR 1.50; 95% CI, 1.08–2.10; p=0.016). However, the OS difference did not reach statistical significance: median OS was 37.7 months in antibiotic-exposed patients versus 49.2 months in those without exposure (HR 1.33; 95% CI, 0.90–1.97; p=0.16).
In adjusted analyses, antibiotic exposure remained associated with shorter PFS in the primary multivariable model, but this association lost statistical significance in a sensitivity analysis incorporating additional variables with substantial missing data. Unlike PPIs, antibiotic exposure did not demonstrate a statistically significant treatment interaction, meaning the analysis could not establish that the association was specifically related to durvalumab treatment.
Microbiome Exploratory Analysis
The researchers also examined an ancillary cohort of 18 patients with stage III NSCLC treated with chemoradiotherapy followed by durvalumab in routine clinical practice. Stool samples were analyzed using a microbiome-based dysbiosis classification called Toposcore. Patients with a eubiotic microbiome profile (SIG2-positive) had longer PFS than those with a dysbiotic SIG1-positive profile (HR 3.47; 95% CI, 1.03–11.63; p=0.0314). Antibiotic exposure was numerically associated with a greater proportion of dysbiotic microbiome profiles, while PPI exposure alone did not show a clear association with dysbiosis in this small cohort.
These exploratory findings provide biological context but remain limited by the small sample size. The analysis was not designed to prove that medication-associated microbiome changes caused differences in durvalumab benefit.
Interpretation and Clinical Implications
The investigators emphasize that this was a post-hoc analysis, not a prospectively designed study. The reasons why patients received PPIs or antibiotics (search) were not available—PPI use could reflect reflux disease, esophagitis after chemoradiotherapy, or other gastrointestinal symptoms, while antibiotic use could reflect infection, hospital admission, or treatment complications. These underlying factors may themselves be associated with outcomes.
David Planchard, Thoracic Oncologist and Professor at the University Paris Saclay, Head of the Thoracic Cancer Group at Gustave Roussy (search), commented on the findings: "New data—Alessio Cortellini shows that using PPIs or ATBs was linked to worse outcomes for stage III NSCLC pts receiving durvalumab (PACIFIC trial) but not those on a placebo. This suggests these common drugs might actually weaken the benefits of ICI. Great collaborative work!"
The findings do not support stopping PPIs or withholding antibiotics (search) when clinically indicated. Patients receiving chemoradiotherapy and durvalumab may require these medications for important medical reasons, and avoiding appropriate treatment could cause harm. However, the study supports careful medication review—clinicians may consider whether long-term PPI therapy remains necessary, whether the indication is still present, and whether the lowest effective dose is being used. Antibiotic stewardship remains important, particularly when antibiotics are prescribed without a clear indication.
In the curative-intent setting of stage III NSCLC, even a modest reduction in the benefit of consolidation durvalumab could be clinically meaningful. The study reinforces a broader point in immuno-oncology: treatment outcomes may be influenced by factors beyond tumor biomarkers and drug selection, including the gut microbiome, concomitant medications, diet, comorbidities, and treatment-related complications. Prospective studies with standardized medication-exposure definitions and integrated microbiome profiling will be needed before medication use can be incorporated into routine immunotherapy decision-making.