OBJECTIVES: To assess whether a previously established low sexually transmitted infection/HIV risk in prostitutes in London has been sustained, and to measure other occupational risks, including mortality. DESIGN: 9 year prospective study in west London. SUBJECTS: 402 prostitutes recruited from 1985 to 1991, 320 were followed up for 675 person years to 1994. MAIN OUTCOME MEASURES: Condom use in commercial and non-commercial sex; viral and bacterial sexually transmitted infection at initial and follow up visits; death. RESULTS: Condom use increased significantly from 1986 to 1993 and protected the majority of commercial sexual contacts. Baseline prevalence: HIV 1.3%, hepatitis C 6.7%, hepatitis B 6.6%, syphilis 2.3%, HTLV-I/II 0.4%, gonorrhoea 3.0%, chlamydia 8.2%, genital herpes 16.8%. Incidence (per 100 person years): HIV 0.2, hepatitis C 0.3, gonorrhoea 5.6, chlamydia 12.6, genital herpes 6.5. Viral infections were associated with injecting drug use and non-British nationality; bacterial infections were associated with numbers of non-commercial partners but not with sexual contacts at work. Four women died during the course of the study; two had AIDS, two were murdered. This mortality of 5.93 per 1000 person years was 12 times the expected rate for women of a similar age. CONCLUSIONS: This study shows that it is possible to have a larger number of sexual partners and remain free from sexually transmitted infections provided that condoms are used consistently: there has been a sustained increase in condom use in the sex industry. None the less, prostitutes are at increased risk of sexually transmitted infections, primarily through non-commercial sexual partnerships. Infectious diseases are only one of the risks facing prostitutes, as illustrated by the mortality from violence as well as from HIV infection.