OBJECTIVES: Zidovudine (ZDV) use has been shown to be associated with anemia. Our objective was to assess the impact of anemia on medical resource use and cost in patients treated with a ZDV-containing HAART regimens. METHODS: Case-control study using HIV Insight, which contains medical chart data from 21 US HIV care centers. ZDV-naïve patients who began a ZDV-containing HAART regimen on any date from January 2000 to June 2004 were assessed for baseline anemia (Hb < 11 g/dL). Those without anemia (N = 529) were followed through June, 2005 or until they had anemia or changed to a non-ZDV-containing regimen. Those with anemia (N = 69) were followed until anemia worsened (drop = 1.0 g/dL). Patients with an incident anemia event (initiating or worsening) with = 12 months of follow-up were selected as cases. Controls were matched on baseline anemia, CD4 < 350, gender and time period. Medical resource use was tabulated during the 1-year post anemia period. Costs were estimated for healthcare encounters using external data for prices expressed in 2006 $US. RESULTS: Cases included 70 patients with anemia (57 initiating, 13 worsening.) Cases had greater use of medical resources than controls: 9.2 vs 5.0 (P <0.001) office visits per year, 33% vs 3% hospitalized (P <0.001), 21% vs 3% treated with EPO (P = 0.001) and 7% vs 0% (P = 0.058) with a transfusion. Cases had higher average annual costs than controls; $37,124 vs $25,141 (P = 0.002). Cost components included EPO $1,369 vs $181 (P = 0.002), antiretrovirals $18,299 vs $19,837(P = 0.461), other drugs $12,488 vs $3,763 (P = 0.005) and hospital stays, $3,114 vs $169 (P <0.001). CONCLUSION: Patients treated with ZDV who experienced anemia required significantly more medical care and reported higher medical cost than patients without anemia.