Background and Purpose:

Novel antidiabetic agents have shown relative treatment benefits for renal outcomes in various patient populations compared to placebo; however, there is yet no comprehensive evaluation of absolute treatment effects – numbers needed to treat. We thus aimed to perform a meta-analysis of digitalized individual patient data from cardiovascular outcome trials. 

Methods:

Individual patient data on a composite renal outcome (e. g. sustained decline in renal function, need for dialysis or transplantation, death due to renal disease etc.) of Cardiovascular Outcome Trials (CVOTs) of sodium glucose transporter 2 (SGLT2) inhibitors or glucagon-like peptide-1 (GLP-1) receptor agonists against placebo were digitalized from Kaplan-Meier plots. Weibull regression models to calculate absolute treatment effects were fitted and random-effects meta-analysis was used to estimate Meta-NNTs with 95% confidence intervals (95% CI). Accuracy of data extraction and Weibull model fit were assessed using scatter plots to compare the originally reported hazard ratio (HR) to the Weibull HRs from extracted data. 

Results:

A total of 84,256 patients from eleven CVOTs (3 on GLP-1 receptor agonists, 8 on SGLT2 inhibitors) at high cardiovascular risk (mostly type 2 diabetes) were analysed; 5,415 (6.4%) experienced a renal event. Accuracy of data extraction was excellent, with an intra-class correlation of 99.6 % (95 % CI: 98.9 %; 100 %). Meta-NNTs at the estimated median follow-up time of 41 months were 71 (95% CI: 50; 121) for GLP-1 receptor agonists and 88 (95% CI: 68; 124) for SGLT2 inhibitors. 

Conclusion: 

In our meta-analysis of absolute treatment effects – numbers needed to treat – of GLP-1 receptor agonists and SGLT2 inhibitors for a composite renal outcome, we found modest treatment benefits for both drug classes in high cardiovascular risk patient populations.