To describe forms of unsafe protected sex (vaginal or anal intercourse where condoms are used unsafely) among a sample of drug users in London, data are drawn from a qualitative study of the sexual and drug taking lifestyles of opioid and stimulant users. Depth one-to-one interviews (n = 96) elicited detailed descriptions from interviewees of their sexual behaviour, including the last occasions they had protected and unprotected sex. Analysis of these accounts identified the phenomenon of 'unsafe protected sex' (UPS). Three forms of UPS were identified: (1) Condoms for ejaculation only. This is where a condom was used for penetrative sex, but only when ejaculation was imminent. In these situations, the perceived function of the condom related more to the prevention of unwanted pregnancy than the prevention of HIV/STDs. (2) Condoms after limited unprotected penetration. This is where sexual partners commenced unprotected penetration but used a condom soon after. Participants tended to see such unprotected penetration as a coerced or collaborative transgression from their usual safer sexual practices. (3) Condom failure. This is where condoms split or came off during penetration. This was sometimes only discovered after ejaculation and withdrawal, and was invariably perceived by participants to have been unsafe. The likelihood of condom failure may be increased in penetrative sex prolonged through the use of drugs. Findings point to the possibility that surveys of sexual risk behaviour underestimate levels of unprotected and unsafe sex. A broader and more sophisticated definition of 'sexual risk behaviour' is required with regard to condom use, one which incorporates UPS. If some forms of UPS are perceived to be 'safer sex', future interventions need to highlight the STD transmission risks associated with this activity. Also, some people may view UPS as a transgression towards unsafe behaviour, and this may be proffered as a rationalization for not using condoms at all.