Heterogeneity in Treatment Effects in the Reduction of Dietary Sodium to Less Than 100 mmol in Heart Failure (SODIUM-HF): A Secondary Post Hoc Analysis

Abstract

Background SODIUM-HF (Study of Dietary Intervention of Sodium Under100 mmol in Heart Failure) compared usual care with dietary sodium restriction in patients with heart failure (HF) and produced neutral results for the primary endpoint. Heterogeneity in treatment effects (HTE) analysis could enhance the original findings. Objective Explore the presence of HTE in the SODIUM-HF trial using a risk-effect–based approach. Methods HTE was assessed using a risk-based approach based on the Meta-Analysis Global Group in Chronic (MAGGIC) HF risk score. Interaction between MAGGIC quartiles and outcomes was assessed using a Bayesian regression model with neutral priors. The primary endpoint was the same for the original trial (composite of cardiovascular-related admission to hospital, cardiovascular-related emergency department visit, or all-cause death within 12 months in the intention-to-treat population); KCCQ was the secondary endpoint. Results Were included 806 patients. MAGGIC quartiles used for the risk-based analyses were 0.036 to 0.102 (low), '0.102 to 0.147 (medium-low), '0.147 to 0.209 (medium-high), and '0.209 to 0.591 (high). There was very strong evidence for the interaction between MAGGIC quartile and intervention (Bayes factor of 68). There was a strong suggestion of association between intervention arm and a lower occurrence of the primary endpoint for the medium-low MAGGIC quartile ('0.98 probability), and a suggestion that the intervention was associated with more frequent occurrence of the primary endpoint in the high MAGGIC quartile (probability of benefit of 0.06). Suggestion of HTE was also found for KCCQ with a strong suggestion of benefit for the intervention for those in the lowest MAGGIC quartiles. Conclusions HTE effects in the SODIUM-HF trial is probable. Further trials in sodium retention may benefit by incorporating this information.