IDEAL

Published

Cooper BA, et al. "A Randomized, Controlled Trial of Early versus Late Initiation of Dialysis". The New England Journal of Medicine. 2010. 363(7):609-619.
PubMed • Full text • PDF

Clinical Question

In patients with stage V CKD, does the initiation of early dialysis improve survival or outcomes?

Bottom Line

In patients with stage V CKD, there was no difference in survival or clinical outcomes between early or late initiation of dialysis.

Major Points

Previous non-randomized observational studies involving early initiation of dialysis demonstrated mixed results on survival. The Initiating Dialysis Early and Late (IDEAL) study is the only randomized trial that assessed time of dialysis initiation and mortality. This study randomized 828 patients with progressive CKD and GFR between 10-15 ml/min/1.73 m² to initiation of early dialysis when GFR was 10-14 or late dialysis when GFR was 5-7. At a median followup of 3.6 years, there was no significant difference in survival, cardiovascular events, infections, or dialysis complications between the two arms.

Of note, the design of the study allowed clinicians to start dialysis based upon symptoms secondary to uremia as well. In actuality, because of symptoms, 76% of patients assigned to the late dialysis arm started dialysis when GFR was ≥5-7 ml/min/1.73 m². Mean GFR at the start of dialysis was 12 ml/min/1.73 m² in the early dialysis arm vs. 9.8 in the late dialysis arm.

This trial supports current practice that dialysis should be initiated in symptomatic patients due to uremia or in asymptomatic patients with an extremely low GFR, approximately 10 ml/min/1.73 m².

Design

• Multicenter, parallel-group, randomized, controlled trial
• N=828 patients with stage V CKD
  • Early dialysis initiation (GFR 10-14 ml/min/1.73 m²) (n=404)
  • Late dialysis initiation (GFR 5-7 ml/min/1.73 m²) (n=424)
• Setting: 32 centers in Australia and New Zealand
• Enrollment: 2000-2008
• Mean follow-up: 3.6 years
• Analysis: Intention-to-treat
• Primary outcome: All-cause mortality
• Secondary outcomes: CV events, infectious events, complications of dialysis, and quality of life

Population

Inclusion Criteria

• Age ≥18
• Progressive CKD and estimated GFR between 10-15 ml/min/1.73 m² BSA (using Cockcroft-Gault)

Exclusion Criteria

• GFR <10 ml/min
• Plans to receive a living donor kidney transplant within next 12 mos
• Recently diagnosed malignancy likely to impact survival

Baseline Characteristics

Demographics:

• Mean age: 60.4 years
• Race or ethnic group
  • White: 71.5%
  • Asian: 8.8%
  • Maori: 6.2%
  • Pacific Islander: 5.8%
  • Aboriginal or Torres Strait Islander: 2.6%
  • Other: 5.1%

Health data:

• Weight: 82.1 kg
• BMI: 28.9
• BP: 142/79
• GFR: 13.1 ml/min/1.73 m² by Cockcroft-Gault vs. 9.9 ml/min/1.73 m² by MDRD

Medical history:

• Primary cause of ESRD
  • Diabetes: 34%
  • Glomerulonephritis: 16.7%
  • PCKD: 10.6%
  • HTN: 7.8%
  • Other: 30.9%
• Coexisting conditions:
  • Diabetes: 42.9%
  • HL: 60.8%
  • CV disease: 38.9%
  • Ischemic heart disease: 28.3%
  • Peripheral vascular disease: 17.9%
  • CHF: 5.5%
  • Stroke: 2.4%
• Current or former smokers: 60.2%
• Time since first seen by nephrologist: 30.9 mos
• Planned dialysis method:
  • Peritoneal dialysis: 56.3%
  • Hemodialysis: 43.7%

Baseline labs:

• Creatinine: 5.9 mg/dL
• Albumin: 3.84 g/mL
• Phosphate: 1.8 mmol/L
• Hemoglobin: 11.4 g/dL

Baseline Medications

• ACE inhibitors: 48.2%
• ARBs: 22.1%
• Statin: 56.2%
• EPO-stimulating agent: 40.8%

Interventions

• Early dialysis arm: initiation of dialysis when GFR 10-14 ml/min
• Late dialysis arm: initiation of dialysis when GFR 5-7 ml/min or when GFR >7 ml/min at physicians' discretion (uremic symptoms or difficult electrolyte abnormalities)
• Regimen of peritoneal dialysis vs. hemodialysis at discretion of patients and their physicians

Outcomes

Comparisons are early dialysis vs. late dialysis. Event rates are per 100 patient-years.

Primary Outcomes

All-cause mortality

10.2 vs. 9.8 (HR 1.04; 95% CI 0.83-1.30; P=0.75)

Secondary Outcomes

Composite CV events (CV death, nonfatal MI, nonfatal stroke, hospitalization with new-onset angina, TIA)

10.9 vs. 8.8 (HR 1.23; 95% CI 0.97-1.56; P=0.09)

Composite infectious events (death from infection or hospitalization for infection)

12.4 vs. 14.3 (HR 0.87; 95% CI 0.70-1.08; P=0.20)

Temporary catheter placement

10.0 vs. 9.7 (P=0.85)

Need for access revision

13.2 vs. 12.4 (P=0.54)

Access-site infection

3.4 vs. 3.5 (P=0.97)

Serious fluid or electrolyte disorder

13.2 vs. 15.0 (P=0.26)

Time to initiation of dialysis

1.80 vs. 7.40 months (HR 2.09; 95% CI 1.81-2.41; P<0.001)

GFR at the start of dialysis (ml/min/1.73 m²)

12.0 vs. 9.8 by Cockcroft-Gault (P<0.001)

9.0 vs. 7.2 by MDRD (P<0.001)

Subgroup Analyses

There were no differences for age, sex, diabetes, BMI, CV disease, or albumin levels.

Criticisms

• No standardized assessment of creatinine was used. The Cockcroft-Gault equation was used rather than MDRD.

Funding

• Grants from several public research agencies in Australia and New Zealand
• Three pharmaceutical and/or medical device companies
• Non-profit International Society for Peritoneal Dialysis

Further Reading

• Concurrent editorial
• NEJM Letters to the Editor