Dose of dialysis: what index?

Wider patient acceptance criteria in hemodialysis (HD) programs do not seem to completely explain the increasing mortality reported in the United States at a time characterized by reduced treatment time and dose. This raises the question of HD standard and adequacy. It stimulated us to analyze patient survival with unchanged 'old-times' methods. 445 unselected patients have been treated for 22 years by the same unchanged methods (24 m2/week, flat-plate dialyzers, cuprophane membrane, acetate buffer). Their survival data were compared to major HD registries and series. Survival was also evaluated as a function of mean arterial pressure (MAP), urea fractional clearance (Kt/V), and middle-molecule dialysis index (DI). Kaplan-Meier (with log-rank test) analysis and Cox proportional hazard model were used. Survival at short and long term was better in our series. This favorable survival difference was more obvious for older patients at the start of HD. It could not be accounted for by selection bias, but correlated with good MAP control without medication and with higher than usual Kt/V (1.67 +/- 0.41) and DI (1.47 +/- 0.38). Cox analysis including five covariates confirmed that survival was linked to MAP. It suggested that survival improvement might be expected from a DI increment of over 1.38 but not from a Kt/V increment of over 1.60. Adequate dialysis cannot be reduced to numbers; it should include both sufficient small- and middle-molecule diffusion and ultrafiltration with arterial pressure control without need for antihypertensive medication. The long-term satisfactory survival remains the best index of overall dialysis adequacy.