Pulse wave velocity demonstrated inferior prognostic power for all-cause mortality compared to clinical risk scores (C-index 73.7% vs 77.5% in the Manhes cohort) and worsened model calibration.
Cohort (n=533)
Sí
Does pulse wave velocity add meaningful prognostic information to simple clinical risk scores for predicting all-cause and cardiovascular mortality in patients with end-stage kidney disease?
In patients with end-stage kidney disease, pulse wave velocity has inferior prognostic power compared to simple clinical risk scores and does not meaningfully improve risk prediction when added to them.
High pulse wave velocity (PWV) is a hallmark of end-stage kidney disease (ESKD) where it is considered useful for risk stratification. We investigated whether PWV adds meaningful prognostic information to 2 simple, well-validated, clinical risk scores specific to ESKD (the Annualized Rate of Occurrence scores) for predicting all-cause and cardiovascular mortality by applying state-of-the-art prognostic tests including discrimination (Harrell C-index), risk reclassification (integrated discrimination improvement), and calibration. We performed these analyses in the 2 largest ESKD cohorts with available PWV data, the Manhes-Hospital cohort in Paris (n=287 patients) and the Quebec Research Center cohort in Canada (n=246 patients). The Harrell C-index of the 2 clinical risk scores was consistently higher than that by PWV both for all-cause (Manhes cohort, 77.5% versus 73.7%; Quebec cohort, 61.5% versus 58.9%) and cardiovascular mortality (Manhes cohort, 77.9% versus 77.2%; Quebec cohort, 63.8% versus 60.3%). Furthermore, PWV provided a very modest increase in discriminatory power over and above clinical risk scores by Harrell C-index (from 0.5% to 1.8%) and in risk reclassification by Integrated Discrimination Improvement (from 0.9% to 5.1%) and actually worsened models calibration. In patients with ESKD, PWV has a prognostic power for all-cause and cardiovascular mortality inferior to that by simple clinical risk scores and only modestly improves the risk discrimination and reclassification by the same risk scores and worsens models calibration. Clinicians may better rely on available clinical risk scores rather than on PWV for risk stratification in the ESKD population.
Tripepi et al. (Mon,) conducted a cohort in End-stage kidney disease (ESKD) (n=533). Pulse wave velocity (PWV) vs. Clinical risk scores (Annualized Rate of Occurrence scores) was evaluated on All-cause and cardiovascular mortality discrimination (Harrell C-index). Pulse wave velocity demonstrated inferior prognostic power for all-cause mortality compared to clinical risk scores (C-index 73.7% vs 77.5% in the Manhes cohort) and worsened model calibration.