Search tips
Search criteria 


Logo of bmjThis ArticleThe BMJ
BMJ. 2007 September 15; 335(7619): 537.
PMCID: PMC1976538
Shortcuts from Other Journals

Erythropoietin not recommended for most critically ill patients

Intensivists should not prescribe erythropoietin to critically ill patients unless they are part of a clinical trial says an editorial (p 1037), after a large randomised trial reported that weekly treatments made no difference to patients' transfusion requirements but increased their risk of thrombotic vascular events (hazard ratio 1.41, 95% CI 1.06 to 1.86).

The trial comprised 1460 medical, surgical, and trauma patients admitted to intensive care units in 115 different hospitals. Just over half received 40 000 units a week of epoietin alfa, a recombinant version of human erythropoietin. The rest had placebo injections. Both groups had up to three doses, starting at least two days after admission.

The results for mortality were inconclusive. In one analysis, epoietin alfa reduced overall mortality after 29 and 140 days. In another analysis, it didn't (0.79 (0.56 to 1.10) after 29 days and 0.86 (0.65 to 1.13) after 140 days). The study's authors believe and emphasise the results of the first analysis. The editorial's authors, who are independent of the trial's sponsors, believe and emphasise the results of the second analysis. The only subgroup to benefit consistently were patients admitted to the intensive care unit after trauma. Subgroup analyses are famously unreliable. This one may justify further trials, but not the wholesale treatment of critically ill trauma patients with epoietin alfa, says the editorial.


  • N Engl J Med 2007;357:965-76

Articles from The BMJ are provided here courtesy of BMJ Publishing Group