Background <p>An increased risk of infection has been suggested for patients with triple-negative breast cancer (TNBC) treated with atezolizumab plus nab-paclitaxel based on the IMpassion130 trial.</p> Methods <p>This Japanese postmarketing study aimed to compare the incidence of severe infections in patients with TNBC treated with atezolizumab plus nab-paclitaxel versus nab-paclitaxel alone. Anonymized patient records regarding medical claims and laboratory tests were extracted from the Japanese Medical Data Vision database. Patients with a drug index date of November 27, 2019,–May 31, 2022, were included in the analysis. Based on the published literature, severe infections were declared when a patient had a confirmed diagnosis of infection, a hospitalization record, infection as the primary reason for hospitalization, and a record of immunological infection tests. In the sensitivity analysis, the definition of severe infections was modified as a combination of confirmed infection diagnosis and intravenous antibacterial drug prescription records.</p> Results <p>Overall, 321 and 319 patients were included in the exposure (atezolizumab plus nab-paclitaxel) and control (nab-paclitaxel alone) groups, respectively. After adjusting for standardized mortality/morbidity ratio weighting, the baseline characteristics were balanced between the groups. The incidence rate ratio (exposure/control) of severe infections was estimated at 3.29 (95% confidence interval [CI]: 0.93–13.53) originally and 1.05 (95% CI: 0.56–1.84) in the sensitivity analysis. Additional analyses supported the appropriateness of the revised definition of severe infections.</p> Conclusions <p>Overall, our results did not indicate a significant increase in the risk of severe infections with atezolizumab plus nab-paclitaxel in daily clinical practice. Further research is required.</p>