Abstract
KEY POINTS: Use of sodium-glucose transport protein-2 inhibitors (SGLT2i) compared with glucagon-like peptide-1 receptor agonists portends a greater risk of candida urinary tract infection (UTI), but no increased risk of noncandida UTI. Abnormal genitourinary pathology is a risk factor for noncandida UTI, but no additive risk was observed in those taking SGLT2i. SGLT2i initiation is associated with an additive larger risk of candida UTI in those with genitourinary tract abnormalities. BACKGROUND: Although sodium-glucose transport protein-2 inhibitors (SGLT2i) possess multiple beneficial effects, the drugs are associated with genitourinary (GU) infections. We sought to define the precise relationship between SGLT2i exposure, types of urinary tract infections (UTIs), and clinical risk factors. METHODS: We used an incident user design with active comparator analysis to derive SGLT2i and glucagon-like peptide-1 receptor agonist (GLP1ra) user cohorts from US nationwide insurance claim data. We used both covariate-adjusted Cox models and Cox models with inverse probability of treatment weighting to investigate the risk of noncandida UTI and candida UTI after drug exposure. We compared the risk between SGLT2i and GLP1ra exposure in the general population and subpopulations with GU abnormalities. RESULTS: SGLT2i exposure compared with GLP1ra exposure was associated with a greater risk of candida UTI (all hazard ratios [HRs] ≥2.42 and all P values < 0.001), but a lower risk of noncandida UTI (all HRs ≤0.91 and all P values < 0.001). Prior GU abnormalities such as prior UTI, prior genital infection, GU malignancy, indwelling Foley, or other GU pathology were associated with greater risk of noncandida and/or candida UTI (all adjusted HRs ≥1.26 and all P values ≤ 0.002). However, no difference in comparative risk of SGLT2i to GLP1ra exposure for noncandida UTI was observed in these subpopulations. By contrast, an additive effect between SGLT2i exposure and several GU abnormalities was observed for candida UTI (all adjusted HRs ≥2.37 and all P values < 0.001). CONCLUSIONS: SGLT2i exposure was associated with greater risk of candida UTI, but not noncandida UTI. SGLT2i to GLP1ra comparative risk of noncandida UTI did not differ in individuals with abnormal GU pathology compared with those without.