Dapagliflozin in Chronic Dialysis: DARE-ESKD-2 Trial

Key Takeaways

• The primary biomarker outcome and key symptom and exercise measures were broadly similar between study groups at 24 weeks.
• No clear between-group safety signal emerged, and symptomatic dehydration, euglycemic ketoacidosis, limb amputation, and drug-induced liver injury were not reported.
• The short follow-up and several design limitations temper broader conclusions from this randomized dialysis study.

In the DARE-ESKD-2 trial, adults receiving chronic dialysis were randomized to dapagliflozin 10 mg daily plus standard care or standard care alone for 24 weeks. After baseline adjustment, NT-proBNP did not differ significantly between groups. Symptom scores and 6-minute walk performance were also similar. The trial did not show demonstrable cardiovascular benefit over 24 weeks in this population.

This prospective randomized open-label, blinded end point trial was conducted at 3 dialysis clinics. Adults on chronic dialysis were randomized 1:1 to dapagliflozin plus standard care or standard care alone. Of 110 screened individuals, 80 were assigned, and 79 comprised the study population after one randomized participant lacked baseline assessment. The mean age was 56 years, 54% were male, 85% received hemodialysis, 54 had heart failure, and median dialysis duration was 3.5 years. The prespecified change from baseline in NT-proBNP at 24 weeks was the primary end point, with KCCQ scores and 6-minute walk distance as key secondary measures. Adherence was 93%, and mean follow-up was 171 ± 31 days. The study tested short-term surrogate, symptom, and functional outcomes in chronic dialysis.

After adjustment for baseline NT-proBNP, the between-group comparison was not statistically significant (P = 0.065). The unadjusted Hodges-Lehmann estimate of the median difference in change was -155 pg/ml, 95% CI -327 to -33. Median adjusted NT-proBNP change was 13 pg/ml in control and -109 pg/ml with dapagliflozin. KCCQ clinical summary and total summary least mean square differences were 7 points, with P values of 0.073 and 0.094. The 6-minute walk test difference was -0.9 m, with P = 0.956. Clinically meaningful responder analyses were also similar, including KCCQ clinical summary improvement in 18.8% versus 22.9% and walking improvement in 34.5% versus 40.0%. The primary and key secondary efficacy measures were similar between groups at 24 weeks.

Exploratory echocardiographic measures, including speckle-tracking assessments, and cardiac troponin I did not differ between groups. Prespecified interaction analyses covered age, sex, dialysis modality, dialysis duration, residual diuresis, NT-proBNP level, and heart-failure status, and no interaction signal was reported. Two deaths occurred in the control group and none with dapagliflozin, while hospitalization events were infrequent. Symptomatic dehydration, euglycemic ketoacidosis, limb amputation, and drug-induced liver injury were not reported. Hematocrit increased with dapagliflozin, whereas electrolytes, liver enzymes, and β-hydroxybutyrate did not increase. No clear between-group safety signal emerged during the study period.

The authors noted that sample-size assumptions were derived from trials that excluded dialysis populations. They also cited the open-label design as a possible source of bias for subjective end points. Relatively few peritoneal dialysis participants limited assessment of modality-specific effect modification, and secondary and exploratory outcomes were not adjusted for multiplicity. The investigators also stated that subgroup analyses had limited power and generalizability. The authors described dapagliflozin as safe in this setting but found no demonstrable cardiovascular benefit during 24 weeks of follow-up.