Real-world, large-scale studies on the association between immune-related adverse events (irAE) and immune checkpoint inhibitor therapy effectiveness are limited. We evaluated overall survival (OS) and progression-free survival based on the occurrence and grade of irAEs. We used data from Japanese patients with unresectable advanced or recurrent non-small cell lung cancer (NSCLC) who received atezolizumab and were enrolled in J-TAIL, a multicenter, prospective, single-arm observational study. Among the 1,002 patients, 190 (19.0%) developed irAEs. The most common irAEs were skin disorders (3.8%) of any grade and interstitial lung disease (1.5%) of grade ≥3. Patients who developed irAEs within 4 or 6 weeks of treatment initiation had higher baseline C-reactive protein levels than those without irAEs. OS was longer in patients with irAEs [HR, 0.66; 95% confidence interval (CI), 0.54-0.82], particularly in those with low-grade irAEs (HR, 0.45; 95% CI, 0.33-0.62), than in patients without irAEs. The HR (95% CI) for OS in patients with low-grade and high-grade skin or endocrine disorder-related irAEs was 0.42 (0.28-0.64) and 0.37 (0.15-0.88), respectively. The HR (95% CI) for OS in patients with low-grade and high-grade irAEs other than skin or endocrine disorders was 0.44 (0.30-0.65) and 1.27 (0.96-1.69), respectively. In patients with unresectable advanced or recurrent NSCLC treated with atezolizumab in real-world settings, irAEs are associated with a clinical benefit except in those with high-grade irAEs other than skin and endocrine disorders. Immune checkpoint inhibitors are useful for treating NSCLC but can cause life-threatening irAEs. This study had a large sample size and stratified the analysis by irAE type and grade. The results suggest that improved management of irAEs may improve the therapeutic effect of atezolizumab.