Current issues of ACP Journal Club are published in Annals of Internal Medicine


Restrictive and liberal blood transfusion strategies had similar mortality rates in critically ill adults

ACP J Club. 1999 July-Aug;131:9. doi:10.7326/ACPJC-1999-131-1-009

Source Citation

Hébert PC, Wells G, Blajchman MA, et al., and the Transfusion Requirements in Critical Care Investigators for the Canadian Critical Care Group. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med. 1999 Feb 11;340:409-17.



In critically ill adults in the intensive care unit (ICU), is a restrictive red blood cell transfusion strategy as effective as a liberal strategy for preventing death and organ dysfunction?


Randomized, unblinded, controlled trial with 60-day follow-up.


22 tertiary care ICUs and 3 community ICUs in Canada.


838 patients (mean age 57.6 y, 63% men) who were admitted to the ICU with an expected stay of > 24 hours, had a hemoglobin (Hgb) level ≤ 9 g/dL within 72 hours of admission, and were considered to be euvolemic. Exclusion criteria were inability to receive blood products, current large blood loss, chronic anemia, pregnancy, brain death, imminent death, indecision about withholding or withdrawing treatment, or routine admission after cardiac surgery. Follow-up was 99%.


418 patients were allocated to a restrictive red blood cell transfusion strategy (maintenance of Hgb at 7 to 9 g/dL) and 420 patients to a liberal strategy (maintenance of Hgb levels at 10 to 12 g/dL). Transfusion was done 1 unit at a time, and Hgb levels were checked after each unit was transfused.

Main outcome measures

All-cause mortality at 30 days. Secondary outcomes were ICU, hospital, and 60-day mortality; organ failure; and length of stay in the hospital and ICU.

Main results

Fewer units of blood were used in the restrictive group than in the liberal group (2.6 vs 5.6 units, P < 0.01). The groups did not differ for 30-day mortality (Table). A trend toward lower hospital mortality was seen in the restrictive group (P = 0.056, Fisher exact test) (Table). Patients in the restrictive group had fewer cardiac events (13% vs 21%, P < 0.01), myocardial infarctions (1% vs 3%, P = 0.02), pulmonary edema (5% vs 11%, P < 0.01), and a a better mean adjusted multiple organ dysfunction score (P = 0.03). The groups did not differ for 60-day (22.7% vs 26.5%, P = 0.2) or ICU mortality (13.4% vs 16.2%, P = 0.3); number of organ failures (P = 0.5); or length of hospital or ICU stay (P > 0.5).


A restrictive red blood cell transfusion strategy in critically ill patients was at least as effective as a liberal policy with similar mortality rates, length of stay, and number of organ failures.

Sources of funding: Medical Research Council of Canada and Bayer.

For correspondence: Dr. P.C. Hébert, Department of Medicine, Ottawa General Hospital, 501 Smyth Road, Box 205, Ottawa, Ontario K1H 8L6, Canada. FAX 613-737-8141.

Table. Restrictive vs liberal transfusion strategy for critically ill patients*

Outcomes Restrictive Liberal RRR (95% CI) NNT
30-d mortality 18.7% 23.3% 20% (-4 to 39) Not significant
Hospital mortality 22.2% 28.1% 21% (-0.1 to 37) Not significant

*Abbreviations defined in Glossary; RRR and CI calculated from data in article.


Allogenic blood transfusion has many risks, but most are uncommon. Myocardial ischemia and infarction are more common after surgery if patients have low Hgb levels (1, 2). The circumstances under which transfusion should be done in critically ill patients remain controversial (3).

In the study by Hébert and colleagues, 41% of eligible patients gave consent, which is consistent with what is seen in similar studies. Enrolled patients were slightly younger and had less cardiovascular disease {the paper erroneously reports that patients had more cardiovascular disease}* although diagnoses and disease severity were similar to those of nonenrolled patients. The treatment groups were similar at baseline, and no difference was shown in 30-day mortality. Subgroup analyses showed that younger patients (age < 55 y) and patients with APACHE II (Acute Physiology and Chronic Health Evaluation) scores < 20 had a lower mortality rate with the restrictive transfusion practice, and the presence of cardiovascular disease did not affect mortality.

Although the authors did not show conclusive data on patients with acute myocardial infarction or ischemia, they caution clinicians about using the restrictive strategy in this patient population.

It may be prudent to individualize transfusion triggers in critically ill patients. A lower trigger may not adversely affect outcomes and may decrease blood transfusions and costs. Although subgroup analyses showed that mortality was not increased using the restrictive strategy in older patients who may have more risk factors for coronary artery disease, we feel that studies evaluating specific transfusion triggers should be done before these results are uniformly used in older patients.

Lakshmipathi Chelluri, MD, MPH
G. Daniel Martich, MDUniversity of Pittsburgh Medical CenterPittsburgh, Pennsylvania, USA


1. Hogue CW Jr, Goodnough LT, Monk TG. Perioperative myocardial ischemic episodes are related to hematocrit level in patients undergoing radical prostatectomy. Transfusion. 1998;38:924-31.

2. Carson JL, Duff A, Poses RM, et al. Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet. 1996;348:1055-60.

3. Valeri CR, Crowley JP, Loscalzo J. The red cell transfusion trigger: has a sin of commission now become a sin of omission? Transfusion. 1998;38:602-10.

*Erratum. N Engl J Med. 1999;340:1056.