Forseter et al1 describe a population of 113 health care workers exposed to human immunodeficiency virus (HIV) after needlestick injury. Sixty of them (53%) took zidovudine. Only 21 (35%) of the 60 health care workers completed the full 42-day course of therapy, and most of them (73%) suffered many side effects. No seroconversion was detected in this cohort of untreated and treated subjects. We want to mention that seroconversion can occur during zidovudine treatment given after HIV exposure; several cases of clinical failure have been reported. We previously described a nurse in whom seroconversion developed after a deep needlestick injury with a needle that had been contaminated by a patient with acquired immunodeficiency syndrome.2 Preventive procedures were started immediately: the wound was bled and washed with bleach for 15 minutes, and oral zidovudine therapy (250 mg four times per day [1 g/d]) was initiated 90 minutes after the incident occurred. At the time of the incident, the nurse was HIV negative. She presented with fever, sweating, headache, and cervical lymphadenopathy 3 weeks after exposure. One week later, serologie tests were positive for HIV type 1 p24 antigen (170 pg/mL), and 7 weeks after exposure, further serologie tests (enzymelinked immunosorbent assay and Western blot) demon¬ strated seroconversion. No other risk factor was identi¬ fied. Zidovudine treatment was started within the time limits found to be crucial for a protective effect in animal studies.1 A decreased sensitivity of the HIV-1 strain to zi¬ dovudine is a possibility. However, zidovudine may not completely inhibit reverse transcription or block cell-tocell transmission of HIV-1.3 Some experiments with vari¬ ous animal models4 and cases of seroconversion that have occurred during zidovudine treatment3 seem to invali¬ date zidovudine's efficacy. These observations associated with the well-known side effects and potential long-term complications emphasize the need for strict adherence to recommended infection-control procedures. Zidovudine use may be justified, but studies are needed to estimate the real indications for, and the optimal dosage and du¬ ration of, prophylactic therapy. dovudine delays but does not prevent the transmission of MAIDS by LP-BM5 MuLv-infected macrophage-monocytes. J Acquir Immune Defic Syndr. 1992; 5:571-576. 4. Schinazi RF, Anderson DC, Fultz P, et al. Prophylaxis with antiretroviral agents in rhesus macaques inoculated with simian immunodeficiency virus. In: Program and abstracts of the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy; October 21-24, 1990; Atlanta, Ga. Abstract 962. 5. Looke DFM, Grove DI. Failed prophylactic zidovudine after needlestick injury. Lancet. 1990;335:1280. In replyWe appreciate the comments of Coutellier and colleagues, who point out that despite zidovudine therapy, human immunodeficiency virus (HIV) may still be transmitted to an exposed health care worker. This has happened on at least eight occasions,1-3 including the case cited by Coutellier et a...