To evaluate the association between human papillomavirus (HPV) and cervical cancer, we performed a population-based case-control study in Columbia and Spain, the former country having an incidence rate of cervical cancer about 8 times higher than the latter. It included 436 cases of histologically confirmed invasive cervical cancer and 387 randomly selected population controls. Information on demographic variables, sexual behaviour and other risk factors was obtained by interview. HPV-DNA was measured in cervical-swab specimens with 3 hybridization assays: ViraPap, Southern hybridization (SH) and polymerase chain reaction (PCR). The presence of HPV-DNA and detection of types 16, 18, 31, 33 and 35 were strongly associated with cervical cancer in each country regardless of the assay used. For both countries combined the adjusted odds ratios and 95% confidence intervals were: ViraPap OR = 25.9 (10.0-66.7); SH OR = 6.8 (3.4-13.4); and PCR OR = 28.8 (15.7-52.6). HPV-16 was the most common type detected in both cases and controls. Our results indicate that there is a very strong association between HPV 16, 18, 31, 33 and 35 and invasive cervical cancer and that this association is probably causal.