The relationship between oral contraceptive use and the risk of invasive cervical cancer was investigated using data from a hospital-based case-control study conducted in the greater Milan area, Northern Italy. A total of 367 women under 60 years of age with a histologically confirmed diagnosis of invasive cervical cancer was compared with a group of 323 controls admitted for a spectrum of acute conditions, non-gynaecological, hormonal or neoplastic and apparently unrelated to oral contraceptive use. Cases had used oral contraceptives more frequently than controls, the age-adjusted relative risk (RR) being 1.53 (95% confidence interval 0.99-2.36). The risk increased with duration of use: compared with never users the age-adjusted RR was 1.48 for up to two years and 1.83 for more than two years (chi 2(1) = 5.28, p = 0.02). Allowing for major identified potential confounding factors, including sexual and reproductive habits, by means of multiple logistic regression, did not explain the association (multivariate RR 1.85 for ever use, 1.05 for up to two years and 2.47 for more than two years). When the interaction between oral contraceptive use and parity or sexual habits was analysed, the effects of various factors appeared independent: the point estimate for multiparous oral contraceptive users versus nulliparous never users was 8.01. There was no consistent influence on risk of invasive cervical cancer of age at first use, whereas the RRs were slightly greater for women who had first used oral contraceptives less than ten years before or had last used them less than five years before diagnosis: these findings, however, were far from significant.