Dapagliflozin and new-onset type 2 diabetes in patients with chronic kidney disease or heart failure: pooled analysis of the DAPA-CKD and DAPA-HF trials

BACKGROUND Chronic kidney disease and heart failure are insulin resistant states associated with a high incidence of diabetes. We assessed the effect of dapagliflozin on new-onset type 2 diabetes in a pooled analysis of patient-level data from the DAPA-CKD and DAPA-HF trials. METHODS This study is a pooled analysis of individual participant data from two phase 3, randomised, double-blind, placebo-controlled, multicentre, clinical trials. Participants with no history of diabetes and HbA1c less than 6·5% (48 mmol/mol) at baseline were included in this pooled analysis. New-onset type 2 diabetes was a prespecified exploratory endpoint in both DAPA-CKD and DAPA-HF trials and is the focus of this analysis. New-onset type 2 diabetes was identified by serial trial measurements of HbA1c (two consecutive values ≥6·5% [≥48 mmol/mol]) or following a clinical diagnosis of diabetes between trial visits. Time to new-onset type 2 diabetes was analysed in a Cox proportional Hazards model from random assignment to end of treatment. FINDINGS 4003 participants (1398 [34·9%] from the DADA-CKD trial and 2605 [65·1%] from the DAPA-HF trial) were included in our analysis: 1995 (49·8%) had received dapagliflozin and 2008 (50·2%) had received placebo. Over a median follow-up of 21·2 months (IQR 16·0 to 25·4), 126 (6·3%) of 2008 patients in the placebo group (event rate 3·9 per 100 patient-years) and 85 (4·3%) of 1995 patients in the dapagliflozin group (event rate 2·6 per 100 patient-years) developed type 2 diabetes (hazard ratio 0·67 [95% CI 0·51 to 0·88]; p=0·0040). There was no heterogeneity between studies (p interaction 0·77) and there was no clear evidence that the effect of dapagliflozin varied in prespecified subgroups including sex, age, glycaemic status, BMI, glomerular filtration rate, systolic blood pressure, and baseline cardiovascular medication use. More than 90% of the participants who developed type 2 diabetes had prediabetes at baseline (HbA1c 5·7% to 6·4% [39 to 46 mmol/mol]). Mean HbA1c remained unchanged (placebo-adjusted change in the dapagliflozin group of -0·01% [95% CI -0·03 to 0·01], -0·1 mmol/mol [95% CI -0·3 to 0·1] at 12 months). INTERPRETATION Treatment with dapagliflozin reduced the incidence of new-onset type 2 diabetes in participants with chronic kidney disease and HF without a reduction in HbA1c. FUNDING AstraZeneca.