The impact of body mass index (BMI) variation on mortality of incident elderly patients on peritoneal dialysis: a joint model analysis

Abstract Introduction: Data on impact of high body mass index (BMI) on mortality of patients on peritoneal dialysis (PD), especially among elderly, are inconsistent. Objective: To evaluate impact of BMI on cohort of incident elderly PD patients over time. Methods: Prospective multicenter cohort study (December / 2004-October/2007) with 674 patients. Socio-demographic and clinical data evaluated with patients followed until death, transfer to hemodialysis (HD), recovery of renal function, loss of follow-up or transplant. Patients were divided into incident on renal replacement therapy (RRT) for PD (PD first: 230) and transferred from hemodialysis (HD first: 444). Analysis was performed comparing these two groups using chi-square or Kruskal Wallis. Similar analysis was used to compare patients on automated peritoneal dialysis (APD) vs. continuous ambulatory peritoneal dialysis (CAPD). Data were compared between patients according to BMI by ANOVA, Kruskal Wallis or chi-square. For analysis of survival, Kaplan Meier method was used and to adjust confounding variables, Cox regression proportional hazard. Joint model for longitudinal and time-dependent data was conducted, assessing impact that a longitudinal variable displays on time of survival. Results: Malnourished patients (76.79 ± 7.53 years) were older (p < 0.0001) with higher percentage of death (44.6%, p = 0.001); diabetes mellitus showed high prevalence in obese patients (68%, p < 0.0001); higher blood pressure levels (p = 0.002) were present in obese and overweight patients. Conclusions: Increased BMI variation over time proved to be a protective factor, with a decrease of about 1% in risk of death for every BMI unit earned