OBJECTIVE: To analyse the data from Swiss nationwide voluntary reporting on non-occupational HIV-postexposure prophylaxis (HIV-PEP) by prescribing physicians. METHODS: One hundred and seventy-six persons, who received antiretroviral prophylaxis for community exposure to HIV between December 1997 and March 2000, were included in this prospective cohort study with standardised data collection. Information on the source, the exposed person, type of exposure, treatment, and outcome was reported by physicians on a voluntary basis to three co-ordinating centers. RESULTS: HIV-PEP was prescribed predominantly following sexual exposure (69%). Needle injury was the second most common type of exposure (19% of all exposures), mostly occurring in a non-healthcare related "professional" setting (i.e., housekeepers, concierges [caretakers], and policemen). Needle sharing accounted for only 4% of all cases of exposure. The HIV status of the source often remained unknown (56%). Most patients received a combination of three antiretroviral drugs (zidovudine/lamivudine/nelfinavir in 34.1%; zidovudine/lamivudine/indinavir in 22.8%; zidovudine/lamivudine/nevirapine in 18.6%; various triple combinations in 13.8%). Follow-up information was available for 86 patients. In this group 78 (91%) completed at least one week of prophylaxis. Side-effects were common (70.9%), particularly diarrhoea (29.6%) and nausea (20.9%). Two patients experienced severe side effects, nephrolithiasis with sepsis, and toxic hepatitis, respectively. CONCLUSIONS: In most of the cases where HIV-PEP was prescribed the indication was questionable, with the HIV status of the source unknown. The role of HIV-PEP as part of HIV prevention programs should be well defined in view of the cost and potential for causing severe side-effects.