Association between body mass and mortality in maintenance hemodialysis patients

Ther Apher Dial. 2010 Aug 1;14(4):400-8. doi: 10.1111/j.1744-9987.2010.00818.x.


A total of 959 Taiwanese patients undergoing maintenance hemodialysis-102 underweight (BMI < 18.5 kg/m(2)), 492 normal weight (BMI 18.5-22.9 kg/m(2)), 187 overweight (BMI 23.0-24.9 kg/m(2)), and 178 obese (BMI >or= 25 kg/m(2)) were recruited into this three-year, multicenter longitudinal study. It was found initially that the underweight group had more females, longer hemodialysis durations, less use of a biocompatible membrane (BCM) dialyzer, higher erythropoietin doses and Kt/V(urea), and lower white blood cell counts, hemoglobin, serum creatinine and phosphate, and high sensitivity C-reactive protein (hsCRP) than other groups (P < 0.001). Furthermore, a chi(2)-test demonstrated that underweight patients had poorer nutrition (P = 0.023), but less systemic inflammation (P < 0.001) than other groups. A stepwise multiple linear regression analysis established that age, sex, diabetes mellitus, hemodialysis duration, use of BCM dialyzer, Kt/V(urea), creatinine, high-density lipoprotein cholesterol, and hsCRP were significant risk factors associated with BMI (P < 0.001-0.002). After three years, 149 (15.5%) patients had died, including 22 of 102 (21.6%) underweight patients, 64 of 492 (13.0%) normal weight patients, 38 of 187 (20.3%) overweight patients, and 25 of 178 (15.5%) obese patients. The primary causes of mortality were cardiovascular (52.3%) and infection (39.6%). A multivariate Cox regression analysis revealed that age, diabetes mellitus, BMI, albumin, hsCRP, and cardiothoracic ratio were significant risk factors associated with all-cause mortality over three years (P < 0.001-0.022). Finally, Kaplan-Meier analysis confirmed that underweight patients suffer higher mortality than other groups (Log rank, P = 0.0392); therefore, the data have demonstrated a survival disadvantage of low BMI in Taiwanese patients undergoing maintenance hemodialysis.

Publication types

  • Multicenter Study

MeSH terms

  • Adult
  • Aged
  • Body Mass Index*
  • Cardiovascular Diseases / etiology
  • Cardiovascular Diseases / mortality
  • Female
  • Humans
  • Infections / etiology
  • Infections / mortality
  • Kaplan-Meier Estimate
  • Kidney Failure, Chronic / therapy*
  • Linear Models
  • Longitudinal Studies
  • Male
  • Middle Aged
  • Multivariate Analysis
  • Proportional Hazards Models
  • Prospective Studies
  • Renal Dialysis / mortality*
  • Risk Factors
  • Taiwan
  • Thinness / complications