The treatment efficacy of erythropoietin (EPO) in end-stage renal disease (ESRD) can be limited by deficiencies of iron, folate, or vitamin B12, by hyperparathyroidism, or by aluminum intoxication. Since EPO costs are significant, this study attempted to determine the cost-effectiveness of performing a panel of screening tests for anemia before starting EPO. Anemia screening was performed prospectively in 48 new-onset ESRD patients at the Ralph H. Johnson Veterans Affairs Medical Center before EPO treatment was started. Serum iron, transferrin, folate, vitamin B12, parathyroid hormone, and aluminum levels were determined, and transferrin saturation (Tfsat) was calculated at the first dialysis session. At presentation for dialysis, the mean hematocrit was 0.264 +/- 0.036 and the mean blood urea nitrogen was 32 +/- 2 mmol/L. Eighteen patients (37.5%) had a serum iron level lower than 7 micromol/L, suggesting iron deficiency. Twenty-five patients (52%) had Tfsat less than 0.20, consistent with overt iron deficiency. No patient was found to be vitamin B12 deficient, to be aluminum intoxicated, or to have significant hyperparathyroidism. One patient had folate deficiency. A cost-effectiveness analysis was performed assuming that (1) EPO would be given at an average starting dose of 6,000 U/wk at a cost of $14/2,000 U of EPO; (2) that without screening 1 month would elapse before a poor response was identified; and (3) that the failure to treat aluminum intoxication and hyperparathyroidism or to replete iron, vitamin B12, or folate deficiency would significantly impair the response to EPO. The Tfsat screen had a cost-effectiveness ratio of 0.2019, saving approximately $5.00 in EPO use for each dollar of test administration. All other screens had cost-effectiveness ratios greater than 1.0, indicating that their testing costs exceeded dollar savings in EPO use. In conclusion, iron deficiency is common in anemic patients starting dialysis, but other causes of anemia are not. It is imperative that current clinical practices be influenced by cost-effectiveness considerations. Given the cost of laboratory screens, and the relative ineffectiveness of the other screens examined here to identify factors known to impair the response to EPO, anemia screening before initiating EPO therapy should be limited to tests to identify iron deficiency.