Nephrology

Dialysis transition linked to more hospital use than nondialytic care among veterans with advanced CKD

In a 1:1 propensity-matched cohort of 13,020 nondialytic and 13,020 dialysis patients, dialysis was associated with higher hospitalization rates versus nondialytic care (IRR 1.64). Dialysis was also linked to more days hospitalized (IRR 1.71) and a slightly longer mean LOS (8.6 vs 8.1 days). Earlier dialysis transition correlated with progressively higher hospitalization rates and longer LOS compared with nondialytic care. Findings derive from an observational cohort of US veterans and may reflect selection and residual confounding. 1

DKD-to-ESRD prediction models show good discrimination but pervasive bias and little external validation

Systematic review of 15 studies found pooled AUCs of 0.896 in development and 0.863 in validation datasets. All models were judged high risk of bias, often retrospective, single-center, poorly calibrated, and lacking external validation. Seven models used biopsy-proven DKD, limiting applicability to routine clinic populations. These models are promising but not ready for routine clinical decision-making without robust external validation. 2

High-volume online hemodiafiltration reduces mortality versus high-flux hemodialysis (CONVINCE era)

The CONVINCE randomized trial and pooled evidence show high-volume HDF reduced all-cause mortality by 23% versus high-flux HD. Systematic reviews cited support cardiovascular and survival benefits across trials. The main barrier now is real-world implementation, with structural, regulatory, and financial hurdles. Further work should prioritize pragmatic implementation studies, patient-reported outcomes, and mechanistic personalization of HDF. 3