BACKGROUND: Laboratory and epidemiological research suggests an association between human papillomavirus (HPV) and cervical intraepithelial neoplasia (CIN). We studied the natural history of incident cervical HPV infection and its relation to the development of CIN. METHODS: We recruited 2011 women aged 15-19 years who had recently become sexually active. We took a cervical smear every 6 months and stored samples for virological analysis. We immediately referred all women with any cytological abnormality for colposcopic assessment, but postponed treatment until there was histological evidence of progression to high-grade CIN. FINDINGS: In 1075 women who were cytologically normal and HPV negative at recruitment, the cumulative risk at 3 years of any HPV infection was 44% (95% CI 40-48): HPV 16 was the most common type. The cumulative risk at 3 years of detecting an HPV type not present in the first positive sample was 26% (20-32). 246 women had an abnormal smear during follow-up, of whom 28 progressed to high-grade CIN. The risk of high-grade CIN was greatest in women who tested positive for HPV 16 (risk ratio 8.5 [3.7-19.2]); this risk was maximum 6-12 months after first detection of HPV 16. All HPV types under consideration were associated with cytologically abnormal smears. Although abnormality was significantly less likely to be associated with low-viral-load samples, the cumulative risk at 3 years of a high-viral-load sample after a low-viral-load sample was 45% (95% CI 35-56). Five women who progressed to high-grade CIN consistently tested negative for HPV. INTERPRETATION: Our findings suggest that attempts to exploit the association between cervical neoplasia and HPV infection to improve effectiveness of cervical screening programmes might be undermined by the limited inferences that can be drawn from the characterisation of a woman's HPV status at a single point in time, and the short lead time gained by its detection.