Lower- Versus Conventional-Sodium Dialysate and Left Ventricular Mass in Patients Receiving Hemodialysis: Causal Analyses of a Randomized Clinical Trial

Rationale & objective In a recent trial, allocation to lower dialysate [Na+] reduced interdialytic weight gain (IDWG), B-type natriuretic peptide (BNP) levels, and possibly extracellular fluid volume based on bioimpedance analysis. It did not, however, reduce left ventricular mass index (LVMI) as assessed by cardiac magnetic resonance imaging. This study aims to elucidate the cause of this unexpected finding.Study design We defined a causal pathway between allocation and LVMI, quantifying effects between physiological variables within the model.Setting & participants Home and self-care satellite hemodialysis patients.Exposure Allocation to lower dialysate [Na+].Outcome LVMI.Analytic approach We used structural equations for path analysis to estimate simultaneous effects of the main exposure on intermediary outcomes and intermediary outcomes on the main outcome.Results The following latent constructs were identified as intermediary outcomes: "Fluid Volume" (indicated by N-terminal proBNP and BNP levels, overhydration index from bioimpedance analysis, and left ventricular end-diastolic volume from cardiac magnetic resonance imaging); "Dialytic BP" (indicated by pre-/postdialysis blood pressure [BP] measurements), "Interdialytic BP" (indicated by home/ambulatory BP measurements), antihypertensive dose, and IDWG. In the structural model, fluid volume had a large effect on LVMI (standardized path coefficient, 0.51; 95% CI, 0.35-0.67), and smaller ones on Dialytic BP and Interdialytic BP (respectively: 0.31; 95% CI, 0.15-0.47 and 0.22; 95% CI, 0.02-0.41). Dialytic BP and Interdialytic BP had small-to-medium-sized effects on LVMI (respectively: 0.23; 95% CI, 0.04-0.42 and 0.22; 95% CI, 0.07-0.37). Neither IDWG nor antihypertensive dose affected LVMI. Allocation to lower dialysate [Na+] had a large effect on IDWG (-0.37; 95% CI, -0.50 to -0.24), a marginal effect on antihypertensive dose (-0.20; 95% CI, -0.41 to 0.02), but none on fluid volume or BP constructs.Limitations Allocation rather than the intervention itself; residual confounding from unrecorded intermediary outcomes.Conclusions Allocation to lower dialysate [Na+] did not change LVMI because of insufficient effect on its main causes, namely fluid volume and BP. Although it reduced IDWG and antihypertensive dose, these factors did not causally affect LVMI.