Erythropoietin Therapy in the ICU: How Expensive?

Abstract & Commentary

By Leslie A. Hoffman, PhD, RN, Department of Acute/Tertiary Care, School of Nursing, University of Pittsburgh, PA, is Associate Editor for Critical Care Alert.

Dr. Hoffman reports no financial relationship to this field of study.

Synopsis: When study data were used to perform a formal analysis of costs associated with use of erythropoietin, the total cost to avoid one transfusion-related adverse event was $4.7 million.

Source: Shermock KM, et al. Number needed to treat and cost of recombinant human erythropoietin to avoid one transfusion-related adverse event in critically ill patients. Crit Care Med. 2005;33:497-503.

Using data from a recent study,1 Shermock and colleagues calculated the absolute risk reduction (ARR) of transfusion-related adverse events, the number of patients needed to treat (NNT), and the cost to avoid a transfusion-related adverse event by using recombinant human erythropoietin (EPO) in critically ill patients. The calculations were based on study data that indicated that the use of 40,000 units of EPO weekly reduced the likelihood of transfusion compared to placebo (50.5% vs 60.4%, respectively). There was a 19% decrease in total units of red blood cells transfused, resulting in an average of 3 transfusions in the placebo group and 2.4 in the experimental group. No difference was detected between the groups in morbidity or mortality (although the study was not powered to detect these). Shermock et al emphasized the benefit of fewer transfusions in terms of a decrease in transfusion-related adverse events.

Using published estimates of the risk and frequency of transfusion-related events, Shermock et al calculated the ARR, the NNT, and the cost to avoid one transfusion-related adverse event. The ARR is defined as the absolute arithmetic difference in rates of bad outcomes between experimental and control participants. The NNT is calculated by dividing the ARR into 1 (ie, 1/ARR). The NNT represents the number of persons who need to receive the experimental treatment to prevent one additional bad outcome.

Based on these calculations, routine use of EPO resulted in an ARR of 191 per million for all transfusion-related adverse events, 35 per million for serious transfusion-related adverse events, and 12 per million for likely fatal events. The NNT was 5,246 to avoid one transfusion-related adverse event (cost, $4.7 million), 28,785 to avoid one serious transfusion-related event (cost, $25.6 million) and 81,000 to avoid one likely fatal transfusion-related event (cost, $71.8 million). The magnitude of these results withstood extensive sensitivity analysis.


The main finding of this study was that 5,246 patients would have to be treated with EPO—at a cost approaching $5.0 million—to avoid one transfusion-related adverse event. The cost of this therapy escalated substantially if the focus was a serious or fatal transfusion-related event.

Anemia is a common problem in the ICU. The causes are multiple and include acute blood loss after trauma, hemorrhage, or surgery; prior treatment with chemotherapy; chronic medical illness; and frequent phlebotomy. Blunted erythropoiesis may also be a factor. There is evidence that critically ill patients may not produce appropriate levels of erythropoietin despite anemia and adequate iron stores.

In 1999 the Transfusion Requirements in Critical Care (TRICC) trial produced findings that indicated that 800 critically ill patients randomized to a liberal transfusion strategy (10 g/dL) had a mortality rate similar to those who received a more restrictive protocol (7 g/dL) 30-days post study enrollment. Moreover, the higher hemoglobin did not result in a shorter duration of mechanical ventilation.2

Findings of this and other studies led to questions about the potential benefits and risks of transfusions. By choosing transfusion reactions, which are rare, EPO was guaranteed to not be cost effective. Regardless, the magnitude of the cost inefficiency was impressive. Unless "zero risk" is the goal, economic considerations do not support routine use of EPO in critically ill patients. This therapy could be appropriate in some situations, eg, during blood shortages, or if the cost of transfusions continues to rise. There may also be some populations that benefit from this therapy, eg, long-stay ICU patients. Until benefits are better established, the preferred approach would appear to be to set and adhere to a strict transfusion threshold, as suggested by TRICC results, and to reduce blood waste and lab draws.


  1. Corwin HL, et al. Efficacy of recombinant human erythropoietin in critically ill patients: a randomized controlled trial. JAMA. 2002;288:2827-2835
  2. Hebert PC, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med. 1999; 340:409-417; Erratum in: N Engl J Med. 1999;340:1056.