The Quality of Life Benefit Associated with Epoetin Treatment in Malignant Blood Disease Is Small. (Overview).

EPO (epoetin alfa or beta or darbepoetin alfa) is widely prescribed for cancer patients with anemia on the assumption that it will increase hemoglobin, reduce transfusion needs and improve quality of life (QOL). As EPO is expensive, its merits need to be critically examined. Statistically significant differences in QOL scores are not necessarily clinically meaningful. We evaluated the clinical significance of the QOL benefits reported in the three largest randomized trials, using two approaches. Effect sizes (ESs) (score difference divided by the standard deviation (SD)) were calculated, using the population derived SDs for the FACT-An questionnaire reported by Cella (2003). According to Cohen (1977), ESs of 0.2–0.5 are considered small, 0.5–0.8 moderate and above 0.8 large. We also used the results obtained by Osoba (1998), where breast cancer patients evaluated score changes (EORTC QLQ-C30 questionnaire) of 5–10 on the 0–100 scale as “a little” change, 10–20 as “moderate" change and above 20 as “very much” change. To facilitate comparison, the score differences reported in the three EPO trials with the FACT-An questionnaire and its subscales were transformed to a 0–100 scale. The table shows the maximum score differences between EPO and placebo groups, and their ESs. The ESs in Littlewood’s paper are small for the general part of the questionnaire (FACT-G) and somewhat larger, but still small for the fatigue and anemia subscales. The ESs in the Osterborg trial barely reach the level of “small” for the fatigue subscale while the score differences for the entire FACT-An, the FACT-G and the anemia subscale are unlikely to be clinically meaningful. In the Hedenus paper, only the results obtained with the fatigue subscale are reported; the ES is small. The score differences for the Fatigue and Anemia subscales in Littlewood’s paper are the only ones that would be considered meaningful (but small) according the subjective significance rating of Osoba. Whether these results reflect differences between the erythropoietin products is impossible to determine since the results have not been confirmed in independent trials and no comparisons have been performed. Although there are obvious caveats (e.g. the importance of score differences may vary across scales and between patient groups, the interpretation of ESs is a matter of debate, results obtained with the EORTC QLQ-C30 may not be transferable to the FACT-An), these data do suggest that the average QOL benefit obtained with EPO in hematological malignancies is small and of uncertain clinical significance. Fundamental issues as to who should be treated with EPO and how need to be resolved. FACT-An questionnaire FACT-G subscale Fatigue subscale Anemia subscale NDR= no data reported. *7 items only, no SD available. **calculated from Fig 3 of that paper, no numerical data reported Score diff (ES) Score diff (ES) Score diff (ES) Score diff (ES) Littlewood 2001 NDR 4.5 (0.27) 10.0 (0.50) 7.9 (0.45) Osterborg 2002 3.1 (0.19) 2.9 (0.18) 4.2 (0.21) 1.8* Hedenus 2003 NDR NDR 4.4 (0.22)** NDR