Sodium-glucose cotransporter 2 (SGLT2 inhibitors) are established therapies across the cardiovascular-renal continuum. While an increased risk of external genital infections with SGLT2 inhibition is well recognized, the relationship between SGLT2 inhibitors and urinary tract infection (UTI)—particularly clinically severe or complicated urinary infections—remains uncertain. The present trial-level systematic review and meta-analysis aimed to quantify the association between SGLT2 inhibitor treatment and two urinary safety outcomes: any UTI and serious/complicated urinary infection, across adults with type 2 diabetes (T2D), chronic kidney disease (CKD), or heart failure (HF).
The analysis included randomized, placebo-controlled trials of the following SGLT2 inhibitors: empagliflozin, dapagliflozin, canagliflozin, ertugliflozin, and bexagliflozin, enrolling adults with T2D, CKD, or HF. Co-primary urinary safety outcomes were defined as (1) any UTI and (2) serious/complicated urinary infection. The latter was defined using directly reported serious UTI endpoints when available or by selecting the closest reported severe urinary phenotype reported in trial reports. The investigators performed sensitivity analyses, subgroup analyses by individual agent and by clinical population, leave-one-out analyses, and exploratory meta-regression. The source abstract reports pooled effect estimates, participant and event counts, heterogeneity metrics, and statistical significance for the primary and secondary analyses.
Eleven trials contributed data to the primary any-UTI analysis, representing 62,542 participants and 4,685 UTI events. The pooled relative risk (RR) for any UTI with SGLT2 inhibitors versus placebo was 1.14 (95% CI 0.99–1.32; P = 0.073). This estimate did not reach conventional statistical significance. Heterogeneity across included studies was substantial, with I² = 76.0%, indicating variability in trial results or populations that influenced the pooled estimate.
A prespecified sensitivity analysis excluded reconstructed estimates from the CANVAS program. After exclusion, 10 studies comprising 52,400 participants and 4,019 events were analyzed. The pooled RR in this sensitivity set was 1.08 (95% CI 1.00–1.18; P = 0.060). Heterogeneity was reduced in this analysis (I² = 23.3%), suggesting that inclusion of the reconstructed CANVAS data contributed substantially to between-study variability. Despite the reduced heterogeneity, the point estimate remained close to unity and did not meet the conventional threshold for statistical significance.
For the clinically severe urinary phenotype, seven trials provided data, totaling 34,396 participants and 339 events. The pooled RR for serious/complicated urinary infection was 0.99 (95% CI 0.78–1.27; P = 0.960). There was no detectable heterogeneity for this outcome (I² = 0.0%), indicating consistent trial-level findings for serious urinary events. The confidence interval spans values that would be compatible with modestly lower or higher risk, but the pooled point estimate is essentially neutral.
According to the abstract, no significant subgroup differences were observed by individual SGLT2 inhibitor or by clinical population (type 2 diabetes, heart failure, or chronic kidney disease). The reported analyses therefore did not identify a specific agent or patient subgroup in which UTI risk differed meaningfully from the overall pooled estimates.
Across randomized placebo-controlled trials included in this meta-analysis, treatment with SGLT2 inhibitors was not associated with a statistically significant increase in either any UTI or serious/complicated urinary infection. The findings underscore a distinction between lower-grade urinary events captured as any UTI and clinically severe urinary infections: while point estimates for any UTI were slightly above unity, they did not achieve statistical significance and were sensitive to inclusion of specific trial estimates. For serious urinary infection—an outcome of greater clinical consequence—the pooled estimate was neutral with no heterogeneity.
These results support the view that urinary safety concerns after SGLT2 inhibitor therapy should not be equated with the established increased risk of external genital infections, and they provide trial-level evidence across T2D, HF, and CKD populations that SGLT2 inhibitors do not clearly increase risk of serious urinary complications.
All summary data and numerical results reported here derive from the abstract of the published meta-analysis. The abstract supplies participant counts, event counts, pooled relative risks, confidence intervals, P values, and heterogeneity statistics for the primary and co-primary outcomes, and it indicates the conduct of sensitivity, subgroup, and exploratory analyses. Detailed trial-level methods, risk-of-bias assessments, individual trial event definitions beyond the abstract description, and full meta-regression results were not provided in the source material available for this rewrite. When applying these findings in clinical practice, readers should consult the full published article for comprehensive methods, trial-level data, and supplementary analyses.
(Adapted from: Matteucci A et al. Are SGLT2 inhibitors really associated with increased urinary tract infection risk? A trial-level meta-analysis across type 2 diabetes, heart failure, and chronic kidney disease. Daru. 2026.)