Anaemia: a restrictive blood transfusion strategy reduced mortality in critically-ill patients aged < 55 or with low APACHE scores.

Clinical bottom line (level 1b)

  1. Normovolemic critically-ill patients with anaemia who had a restrictive blood transfusion policy compared with a liberal one were not clearly less likely to die or have complications.
  2. Patients aged < 55 who had a restrictive blood transfusion regimen were less likely to die (NNT = 14 at 30 days) .
  3. Patients with a low APACHE score ( = 20) who received a restrictive blood transfusion regimen were less likely to die (NNT = 14 at 30 days) .
Hebert et al: New England Journal of Medicine 1999; 340: 409-417
Expires November 2002

The study

Unblinded ?concealed randomised trial with intention-to-treat
Setting: 25 intensive care units, Canada

838 patients (aged mean 58, 63% male) critically ill with euvolaemic anaemia (Hb < 90 g/l) within 72 hours of admission

Excluded if
  • physicians deciding whether to withhold or withdraw therapy
  • admission after routine cardiac surgical procedure
  • pregnancy
  • active blood loss at time of enrolment (fall in Hb > 30 g/l in the last 12 hours or a requirement for at least 3 units of blood)
  • chronic anaemia (Hb < 90 g/l at least once more than a month before admission)
  • brain death or imminent death
  • stay < 24 hours
  • aged < 16
  • unable to receive blood products

 

Note:

  • Patients were stratified for study centre and disease severity (APACHE II score) before randomisation.
  • A priori subgroup analysis was performed based on age, APACHE score and cardiac disease.


Control Group: (n = 420, 420 analysed): liberal transfusion strategy: Hb maintained at 100 to 120 g/l with red cells transfused when Hb fell below 100 g/l
Experimental Group: (n = 418, 418 analysed): restrictive transfusion strategy: Hb maintained at 70 to 90 g/l, with red cells transfused if Hb fell below 70 g/l
Blood units were transfused a unit at a time - Hb was measured after each unit.
99% followed for 60 days
Outcome notes:
  • Death : at 30 days
  • Death : at 60 days
  • Any complication : cardiac, pulmonary, infection, GI, neurological, shock
  • Death : aged < 55
  • Death : APACHE score = 20

The evidence

Outcome Time to outcome CEREERRRR
(95% CI)		ARR
(95% CI)		NNT
(95% CI)
Death 30 days 98
(23.33%)		 78
(18.66%)		 20%
(-4% to 39%)		 4.67%
(-0.83% to 10.18%)		 21
(NNT = 10 to infinity;
NNH = 120 to infinity)
Death 60 days 111
(26.43%)		 95
(22.73%)		 14%
(-9% to 32%)		 3.70%
(-2.12% to 9.53%)		 27
(NNT = 10 to infinity;
NNH = 47 to infinity)
Any complication 30 days 228
(54.29%)		 205
(49.04%)		 10%
(-3% to 21%)		 4.67%
(-0.83% to 10.18%)		 19
(NNT = 8 to infinity;
NNH = 66 to infinity)

aged < 55
Outcome Time to outcome CEREERRRR
(95% CI)		ARR
(95% CI)		NNT
(95% CI)
Death 30 days 21
(13.0%)		 10
(5.78%)		 56%
(9% to 78%)		 7.26%
(1.01% to 13.52%)		 14
(7 to 99)

APACHE score = 20
Outcome Time to outcome CEREERRRR
(95% CI)		ARR
(95% CI)		NNT
(95% CI)
Death 30 days 35
(16.13%)		 18
(8.74%)		 46%
(7% to 68%)		 7.39%
(1.61% to 13.62%)		 14
(7 to 86)

Comments

  1. The study was too small to exclude potential benefit or harm for all patients.
  2. No significant difference in mortality was noted for a subgroup of patients with cardiac disease.
  3. No difference in length of intensive care or hospital stay was noted between the two groups.

Citation

  1. Hebert PC, Wells G, Blajchman MA, et al: a multicenter, randomized controlled clinical trial of transfusion requirements in critical care. New England Journal of Medicine 1999; 340: 409-417
Search Terms: transfus* in Cochrane
Contributor: Chris Ball and Musab Hayatli, October 1999
Reviewer: Arturo Marti-Carvajal

Clinical Question.
Patient critically-ill with euvolaemic anaemia
Intervention or Exposure restrictive blood transfusion
Comparison liberal blood transfusion
Outcome death