An epidemiologic investigation of an acupuncturist's practice in Rhode Island identified 35 patients who were infected with hepatitis B virus during 1984. Of 366 patients seen by the acupuncturist during 1984,316 (86%) completed questionnaires and submitted serum for hepatitis B serology. Use of tests for immunoglobulin M antibody to hepatitis B core antigen (IgM anti-HBc) identified 17 case-patients who otherwise may have gone undetected. Thirty-four of the 35 case-patients were treated in only one of the two clinics run by the acupuncturist. Patients who received a greater number of acupuncture needles during their treatment course were more likely to have been infected; the attack rate for patients who received <150 needles was 9%, compared with 33% for patients who received ≥450 needles (p < 0.001). Attack rates were higher during a one-month period when the index case-patient was more likely to have been viremic than during any other period in 1984 (relative risk = 4.1, 95% confidence interval = 2.3–7.3). While observing the acupuncturist's technique, the investigators noted several potential mechanisms for needle contamination. This study highlights the potential for transmission of hepatitis B in situations of repeated needle use.