In order to determine the prevalence of risk behaviour for, and antibodies to HIV and hepatitis B in clients of a needle-exchange scheme in central London we employed an anonymous, self-administered questionnaire along with salivary antibody testing by immunoglobulin (Ig) G antibody capture immunoassay. Two hundred and thirty-two subjects (193 men, 39 women; median age 32) participated; a response rate of 89%. Clients were long-term, frequent injectors. Lending used equipment at any time was reported by 55%, and borrowing by 52%. Of those who had shared needles and syringes during the last year, the majority had lent to, or borrowed from, one person only (53 and 55%, respectively). Younger clients (less than 29 years of age) reported more recent sharing than older clients (greater than 30 years of age). Five out of 211 (2.4%) samples tested for anti-HIV were positive. One hundred and eleven out of 199 (56%) samples were positive for anti-hepatitis B core (HBc). In this population of needle-exchange attenders there is no evidence of further spread of HIV, and a low prevalence of HIV infection appears to have been sustained. However, the high prevalence of anti-HBc provides evidence of previous risk behaviour and so constant vigilance is necessary if further viral spread is to be avoided. This study has established an acceptable method for the anonymous surveillance of current risk behaviour and salivary antibodies to HIV and hepatitis B virus (HBV) in a drug-using population.