Frontiers in Nutrition · Oct 22, 2026

The obesity paradox in peritoneal dialysis: reassessing optimal BMI via longitudinal time in range analysis in a single-center cohort

By Wang M, Zeng X, Liu D et al.

Background: The optimal BMI target for mortality prevention in PD patients remains unclear, with existing evidence largely limited to single-time-point measurements that ignore cumulative exposure. Methods: This single-center retrospective cohort study enrolled 670 PD patients with repeated BMI measurements at baseline, 12 months, and 36 months. BMI time above target range (TAR) was calculated for each integer threshold from 18 to 32 kg/m ² using the Rosendaal linear interpolation method. Multivariable Cox proportional hazards models were used to identify the threshold with the strongest protective association with all-cause mortality. BMI time in target range (TTR) within the optimal range was then calculated, and its association with mortality was assessed using a landmark design (from 36 months onward), with adjustments for age, diabetes, serum albumin, estimated glomerular filtration rate, high-density lipoprotein cholesterol, 24-h urine volume, 3-year cumulative peritonitis episodes. Results: Over a median follow-up of 2.3 years, 75 deaths occurred. The TAR threshold scan revealed a non-linear protective pattern, with the nadir of mortality risk at 26 kg/m ² [HR per 10% TAR increment: 0.82, 95% CI: 0.73–0.92, P = 0.018 (Bonferroni-corrected)]. Notably, TTR of the 26–30 kg/m ² range was significantly associated with reduced mortality risk (HR per 10% increment: 0.77, 95% CI: 0.62–0.96, P = 0.02), whereas TTR of the 24–25.9 kg/m ² range did not reach statistical significance (HR = 0.98, 95% CI: 0.89–1.08, P = 0.77). Conclusions: Sustained BMI within 26–30 kg/m ² was associated with lower all-cause mortality in Chinese PD patients, suggesting a potentially higher optimal BMI target. However, these findings are exploratory and require confirmation in larger, multi-center studies.

doi.org/10.3389/fnut.2026.1982039Open accessPubMed 42990871Populations

AgingResearch.News