To assess the early effect of epoetin beta on survival and tumour progression in anaemic patients with cancer, data were pooled from nine randomised clinical trials comparing epoetin beta with placebo or standard care. Studies were not primarily designed to assess these end points. Follow-up was for treatment duration plus 4 weeks following therapy completion. All adverse events (AEs) were retrospectively reviewed blinded, for progression. Thromboembolic events were also assessed. Data analysis involved standard statistical tests. Overall, 1413 patients were included (epoetin beta, n=800; control, n=613; 56% haematological, and 44% solid). Median initial epoetin beta dose was 30 000 IU/week. Overall survival during months 0-6 was similar with epoetin beta and control (0.31 vs 0.32 deaths/patient-year). No increased mortality risk was seen with epoetin beta (relative risk (RR) 0.97, 95% CI: 0.69, 1.36; P=0.87). There was a significantly reduced risk of rapidly progressive disease for epoetin beta (RR 0.78, 95% CI: 0.62, 0.99; P=0.042). Epoetin beta was associated with a slightly higher frequency of thromboembolic events vs control (5.9% vs 4.2% of patients) but thromboembolic-related mortality was identical in both groups (1.1%). Epoetin beta provided a slight beneficial effect on tumour progression and did not impact on early survival or thromboembolic-related mortality.