Our meta-analysis of 40 randomized, controlled clinical trials found that dapagliflozin was associated with more frequent UTI compared with placebo and other active treatments. With respect to the effects of different dosage, a significantly increased risk of UTI was only identified with dapagliflozin 10 mg daily. We also confirmed the influence of different treatment durations of dapagliflozin on UTI occurrence at three levels of dosage in a subgroup analysis. The findings suggest that UTI is more common among patients treated with dapagliflozin 10 mg daily than among those receiving placebo and active therapy in treatment periods of more than 24 weeks. To our knowledge, this is the first meta-analysis to substantially expand on the data involving long-term treatment of dapagliflozin in patients with T2DM based on the current evidence.
Recent studies have still failed to reach a consensus on whether dapagliflozin exerts an influence on the risk of UTI3,4,7,8. According to its therapeutic mechanism, dapagliflozin induces glycosuria by inhibiting the urinary reabsorption of glucose, which is considered as a reasonable risk factor for UTI. Nonetheless, findings in patients with familial renal glucosuria are contrary to the hypothesis that it may not contribute to the increased incidence of UTI49. Furthermore, it has been verified that hyperglycaemia is significantly related to impaired immune system impairment50, which may increase the risk of diverse infectious events including UTI. Therefore, antihyperglycaemic therapy with dapagliflozin could theoretically reduce the prevalence of UTI. However, our findings reconfirms the results of some previous studies demonstrating that high-dose dapagliflozin is associated with an increased risk of UTI1,11,13,15. There are two possible reasons for these contradictory conclusions. First, several previous meta-analyses were limited by a relatively small number of RCTs with insufficient statistical power to detect the clinically important side effects. Second, the heterogeneity in the design of clinical trials among the included RCTs may also have affected the final conclusions. We augmented many updated RCTs and performed the subgroup analysis based on the elements of experimental design, which made the result more convincing.
Even though the relationships between two low-dose levels of dapagliflozin and the risk of UTI were not confirmed, our meta-analysis findings suggested that the risk still exhibited a dose-related trend (including the 10 mg daily dose). This is consistent with another meta-analysis conducted by Puckrin et al.1, which compared only two different dosing regimens. However, Musso et al.13 found that the risk of UTI was not dose-related. Several confounding variables may account for this, such as different sample sizes, publication bias originating from the diagnosis criteria of UTI and the influence of differences in the intervention measures in the control group.
Moreover, we realize that the combination therapy of dapagliflozin and other The importance of this aspect has not been highlighted in the limited published trials. In the study by Rosenstock et al.24, it was reported that UTI was more common in dual add-on therapy than in triple therapy, both of which included dapagliflozin. In contrast, another study demonstrated that the frequencies of UTI were slightly higher with triple therapy than with dual therapy in both men and women43. However, in addition to their conflicting results, neither of the studies compared the incidences of UTI between single-agent therapy of dapagliflozin and combination therapy. Based on a certain number of RCTs, we performed a stratified analysis to compare the risk of UTI between the abovementioned two treatment modes, which could generate relatively objective results. To date, there are no relevant clinical data that address this. Our findings demonstrate that administering dapagliflozin as an add-on therapy in the treatment of patients with T2DM calls for the careful consideration of the higher risk of UTI compared to its use as a monotherapy. The precise underlying mechanism is unknown. It may be related to the inadequacy of reaching glycaemic targets in those patients who required add-on therapy. This warrants further investigation in the future.
Based on prior studies, most UTI events were reported to be of mild or moderate intensity. They responded to standard antimicrobial treatment, and rarely led to an interruption of dapagliflozin therapy28,51,52. Only a small minority of infections develop into pyelonephritis or urosepsis1,28,53. Although a few studies have observed a greater rate of recurrence of UTI with dapagliflozin compared with placebo51,54, most of the infections were single episodes and rarely led to recurrence. It is also widely accepted that the clinical diagnosis of UTI occurs more frequently in women than in men. We did not conduct further research on the abovementioned clinical aspects, which were either established facts or lacked data.
As previously mentioned, our study has several advantages. First, the main strength of the review is that all the conclusions are based on a much larger number of RCTs with different durations of follow-up. In several published meta-analyses, the most common limitation was the small quantity of included trials with a short duration10,11,13. Second, although the systematic review by Puckrin et al.1 included a longer period of follow-up time, a difference in the risks of UTI between short-term and long-term treatment with dapagliflozin was not observed. Our study indicated that the risk of UTI may not increase even at a dose of 10 mg/d if the treatment duration is less than 24 weeks. Third, it was also observed that a combination therapy of dapagliflozin and another hypoglycaemia agent in T2DM patients may induce a higher risk of UTI, although few data are available. Moreover, the study was performed using a prespecified protocol in line with PRISMA guidelines, and the included studies were assessed with the Cochrane Collaboration’s tool.
The limitations of our study should be acknowledged. First, we decided to combine data that compared dapagliflozin with placebo or other active comparators in one set of analyses according to the heterogeneity in the clinical study design. The different effects on the risk of UTI among hypoglycaemic agents other than dapagliflozin cannot be completely ruled out. Second, we didn’t implement the subgroup analysis stratified by age because of the paucity of original data. The distinction of risk of UTI due to age among the RCTs may exist. Third, we only focused on the influences of three dose of dapagliflozin, and more unconventional doses including 2.5 mg/d and 20 mg/d were not analysed. Fourth, the diagnostic criteria of UTI varied from the presence of signs and symptoms in the reporting methods to specific laboratory testing. We combined the data in order to facilitate the analysis, which could be a potential confounder.