Importance High body mass index (BMI) is independently associated with overall survival benefit from immune checkpoint inhibitor therapy in patients with melanoma, yet whether BMI is associated with outcomes in patients with advanced non-small cell lung cancer treated with atezolizumab remains unknown. Objective To examine whether BMI is associated with survival outcomes and adverse events in patients with non-small cell lung cancer (NSCLC) treated with atezolizumab. Design, Setting, and Participants A pooled analysis of individual patient-level data from 4 international, multicenter clinical trials was performed. Two were single-arm phase 2 trials (BIRCH [data cutoff of May 28, 2015] and FIR [data cutoff of January 7, 2015]), and 2 were 2-arm randomized clinical trials (POPLAR [phase 2; data cutoff of May 8, 2015] and OAK [phase 3; data cutoff of July 7, 2016]). Patients with advanced NSCLC previously untreated or treated with at least 1 line of systemic therapy, with measurable disease and good organ function and without contraindications for chemotherapy or immune checkpoint inhibitor therapy, were included in these trials. Data analyses were performed from February 28, 2019, to September 30, 2019. Interventions The control group was treated with docetaxel once every 3 weeks until disease progression or unacceptable toxic effects occurred in POPLAR and OAK. The experimental group was treated with atezolizumab once every 3 weeks until disease progression or unacceptable toxic effects occurred in all available trials. Main Outcomes and Measures Association between BMI and overall survival (OS), progression-free survival (PFS), and treatment-related adverse events. Intention-to-treat analysis was conducted. Results Adequate data were available for 2110 patients from a total pool of 2261 across 4 trials. Of those 2110, 1434 patients (median age, 64 years [range, 57-70 years]; 890 men [62%]) received atezolizumab and 676 patients (median age, 63 years[range, 57-69 years]; 419 men [62%]) received docetaxel. There was a linear association between increasing BMI and OS in patients treated with atezolizumab. Obesity (BMI >= 30 [calculated as weight in kilograms divided by height in meters squared]) was associated with significantly improved OS in patients treated with atezolizumab, but not in those who received docetaxel after adjusting for confounding variables. The association between BMI and OS/PFS was the strongest in the high PD-L1 expression subgroup. Overall survival for patients with the highest category of PD-L1 expression (>= 50% of tumor cells or >= 10% of tumor-infiltrating immune cells; n = 436) had hazard ratios of 0.36 (95% CI, 0.21-0.62) for the group with obesity and 0.69 (95% CI, 0.48-0.98) for the group with overweight. Patients with the highest category of PD-L1 expression had PFS hazard ratios of 0.68 (95% CI, 0.49-0.94) for the group with obesity and 0.72 (95% CI, 0.56-0.92) for the group with overweight. Treatment-related adverse events were not associated with BMI. Conclusions and Relevance High BMI appears to be independently associated with improved survival with atezolizumab in patients with NSCLC, raising the possibility that baseline BMI should be considered as a stratification factor in future immune checkpoint inhibitor therapy trials. This cohort study examines the association between body mass index and survival in patients with non-small cell lung cancer who receive immune checkpoint inhibitor therapy. Question Is high body mass index associated with survival outcomes with atezolizumab therapy, an immune checkpoint inhibitor, in patients with non-small cell lung cancer? Findings In this pooled analysis of 4 clinical trials that included more than 2261 patients with non-small cell lung cancer, those who had high body mass index had a significant reduction in mortality with atezolizumab, particularly in the presence of high expression of programmed cell death ligand 1. Meaning High body mass index appears to be associated with improved overall survival in atezolizumab-treated patients with advanced non-small cell lung cancer.