RESUMO
PURPOSE: Recent reports suggest that cancer control may worsen if erythropoietin is administered. We investigated whether erythropoietin receptor expression on cancer cells may correlate with this unexpected finding. PATIENTS AND METHODS: Cancer tissue from patients with advanced carcinoma of the head and neck (T3, T4, or nodal involvement) and scheduled for radiotherapy was assayed retrospectively for erythropoietin receptor expression by immunohistochemistry. Patients were anemic and randomized to receive epoetin beta (300 U/kg) or placebo under double-blind conditions, given three times weekly starting 10 to 14 days before and continuing throughout radiotherapy. We administered 60 Gy following complete resection or 64 Gy subsequent to microscopically incomplete resection; 70 Gy were given following macroscopically incomplete resection or for definitive radiotherapy alone. We determined if the effect of epoetin beta on locoregional progression-free survival was correlated with the expression of erythropoietin receptors on cancer cells using a Cox proportional hazards regression model. RESULTS: We studied 154 of 157 randomly assigned patients; 104 samples were positive, and 50 were negative for receptor expression. Locoregional progression-free survival was substantially poorer if epoetin beta was administered to patients positive for receptor expression compared with placebo (adjusted relative risk, 2.07; 95% CI, 1.27 to 3.36; P < .01). In contrast, epoetin beta did not impair outcome in receptor-negative patients (adjusted relative risk, 0.94; 95% CI, 0.47 to 1.90; P = .86). The difference in treatment associated relative risks (2.07 v 0.94) was borderline statistically significant (P = .08). CONCLUSION: Erythropoietin might adversely affect prognosis of head and neck cancer patients if cancer cells express erythropoietin receptors.