From Wiki Journal Club
Jump to navigation Jump to search
Lowrie EG, et al. "Effect of the hemodialysis prescription of patient morbidity: report from the National Cooperative Dialysis Study". The New England Journal of Medicine. 1981. 305(20):1176-1181.
PubMedFull textPDF

Clinical Question

Among ESRD patients, how do HD duration and target BUN affect death and hospitalization rates?

Bottom Line

Among ESRD patients, death and hospitalization rates are decreased when dialysis prescriptions target a lower BUN compared to a higher BUN.

Major Points

The National Cooperative Dialysis Study (NCDS) was the first randomized trial of hemodialysis "dose". Prior to this study, there was no standardized target BUN or HD session duration among patients with ESRD.

The NCDS randomized 151 ESRD patients in a 2x2 factorial design to 1) long vs. short HD sessions and 2) high vs. low target BUN concentration. Patients were followed until withdrawal from study due to death, medical illness, or other reason. At one year of follow-up, the low BUN groups had fewer withdrawals due to death or medical illness compared to the high BUN groups (18% and 6% vs. 45% and 62%), regardless of HD session duration. Outcomes did not reach statistical significance between groups randomized to short vs. long dialysis sessions (P=0.06). There were no differences in mortality between the groups, although the trial was underpowered to detect mortality differences.

The trial was small, followed patients for only one year, and its use of surrogate outcomes (withdrawal from study due to death or medical illness) makes it difficult to translate these results into clinical practice. Nevertheless, the study's results led to the development of the Kt/Vurea,[1] which is currently used as the standard measurement of dialysis adequacy.


  • 2x2 factorial design, randomized, controlled trial
  • N=151 ESRD patients
    • Long session, low BUN (n=41)
    • Long session, high BUN (n=32)
    • Short session, low BUN (n=41)
    • Short session, high BUN (n=37)
  • Setting: 8 US centers
  • Enrollment: 1978
  • Follow-up: 1 year
  • Analysis: Intention-to-treat
  • Primary outcome: Time to withdrawal from trial, time to first hospitalization


Inclusion Criteria

  • Age 18-70 years
  • Receiving HD with residual CrCl ≤3 mL/min

Exclusion Criteria

  • Cancer
  • Diabetes
  • Uncontrolled HTN
  • Reversible cause of renal failure
  • Systemic disease
  • Unstable coronary or cardiovascular disease
  • Clinically important pulmonary or hepatic disease

Baseline Characteristics

Not presented in the article.


  • Randomized in 2x2 factorial design to:
    • Low target BUN (50 mg/dL) vs. high target BUN (100 mg/dL)
    • Short HD session (3h) vs. long HD session (4.5h) three times weekly


Comparisons are low BUN/long HD and high BUN/long HD vs. low BUN/short HD vs. high BUN/short HD groups.

Primary Outcomes

Removal from trial for death or medical reasons at 1 year
Low vs. high BUN groups: 18% and 6% vs. 45% and 62% (P<0.0001)
Short vs. long duration: 18% and 45% vs. 6% and 62% (P>0.5)

Secondary Outcomes

Low vs. high BUN groups: 2 vs. 1 patient (P=NS)
Short vs. long duration: 2 vs. 1 patient (P=NS)
Removal from trial for medical reasons
Low vs. high BUN groups: 4 vs. 28 patients
Short vs. long duration: 13 vs. 19 patients
Removal from trial for other reasons
Low vs. high BUN groups: 40 vs. 26 patients
Short vs. long duration: 34 vs. 32 patients
Not hospitalized at one year
Low vs. high BUN groups: 86% and 69% vs. 46% and 31%
Short vs. long duration: 86% and 46% vs. 69% and 31%

Additional Analyses

HD duration
Low vs. high BUN groups: 4.5 and 3.3h vs. 4.5 and 3.2h (P=NS)
Short vs. long duration: 3.3 and 3.2h vs. 4.5 and 4.5h (P<0.0001)
Average BUN
Low vs. high BUN groups: 51.3 and 54.1 mg/dL vs. 87.7 and 89.6 mg/dL (P<0.0001)
Short vs. long duration: 51.3 and 87.7 mg/dL vs. 54.1 and 89.6 mg/dL (P<0.05)


  • Small, unblinded study
  • Lacked baseline characteristics including baseline protein intake, etc.
  • Excluded diabetic patients and individuals with residual kidney function, which limits generalizability
  • The trial's primary outcomes are poor surrogates for clinically important outcomes
  • Underpowered to detect death; short follow-up time limits primary outcome observations


Supported by the National Institute of Arthritis, Metabolism, and Digestive Diseases.

Further Reading