Predictors of hospitalization and death among pre-dialysis patients: a retrospective cohort study

Nephrol Dial Transplant. 2000 May;15(5):650-8. doi: 10.1093/ndt/15.5.650.


Background: Although there is abundant research describing predictors of patient morbidity and mortality among dialysis patients, predictors of adverse clinical outcomes among pre-dialysis patients are less well defined. The purpose of this study was to identify baseline predictors of first non-elective hospitalization among a retrospective cohort of 362 pre-dialysis patients.

Methods: Univariate and multivariate Cox proportional hazard models were used to identify predictors of hospitalization prior to dialysis initiation, adjusted for baseline creatinine level. Dialysis initiation, loss to follow-up, and study conclusion were censored events. Secondary outcomes included cause-specific hospitalization and death.

Results: Univariate analysis indicated that advanced age (RR 1.026, CI 1.016-1.037), number of prescribed anti-hypertensive medications (RR 1.149, CI 1.019-1.296), history of myocardial infarction (RR 1.979, CI 1.339-2.926), congestive heart failure (RR 2.299, CI 1.616-3.270), angina (RR 2.289, CI 1.695-3.091), peripheral vascular disease (RR 1.841, CI 1.282-2.644), renal failure secondary to nephrosclerosis (RR 1.413, CI 1.033-1.933) or renal artery stenosis (RR 1.587, CI 1.036-2.430), lower baseline haemoglobin level (RR 0.986, CI 0.979-0.992), and baseline creatinine greater than 300 micromol/l (RR 1.636, CI 1.233-2.171) were predictors of hospitalization. Gender, diabetes, diastolic blood pressure, mean arterial pressure, history of stroke, and hypoalbuminaemia did not predict outcome. Multivariate analysis, adjusted for baseline creatinine level, selected advanced age (RR 1. 017, CI 1.006-1.027), angina (RR 1.893, CI 1.371-2.613), peripheral vascular disease (RR 1.545, CI 1.054-2.266), and haemoglobin level (RR 0.987, CI 0.944-0.979) as independent predictors of hospitalization.

Conclusion: Advanced age, co-morbid cardiovascular illness and anaemia are independent predictors of non-elective hospitalization prior to dialysis initiation. Further study is needed to determine the extent to which aggressive pre-dialysis management of anaemia and cardiovascular disease can improve patient outcomes.

MeSH terms

  • Aged
  • Aging / physiology
  • Anemia / epidemiology
  • Cardiovascular Diseases / epidemiology
  • Cohort Studies
  • Comorbidity
  • Creatinine / blood
  • Death*
  • Female
  • Forecasting
  • Hospitalization*
  • Humans
  • Male
  • Middle Aged
  • Mortality
  • Multivariate Analysis
  • Proportional Hazards Models
  • Referral and Consultation*
  • Renal Replacement Therapy*
  • Retrospective Studies


  • Creatinine