Objectives To assess the performance of a 5-type human papillomavirus (HPV) messenger RNA (mRNA) test in primary screening within the framework of the Norwegian population-based screening programme.

Design Nationwide register-based cohort study.

Setting In 2003–2004, general practitioners and gynaecologists recruited 18 852 women for participation in a primary screening study with a 5-type HPV mRNA test.

Participants After excluding women with a history of abnormal smears and with cervical intraepithelial neoplasia grade 2 (CIN2+) before or until 3 months after screening, 11 220 women aged 25–69 years were eligible for study participation. The Norwegian Cancer Registry completed follow-up of CIN2+ through 31 December 2009.

Interventions Follow-up according to the algorithm for cytology outcomes in the population-based Norwegian Cervical Cancer Screening Programme.

Main outcome measures We estimated cumulative incidence of CIN grade 3 or worse (CIN3+) 72 months after the 5-type HPV mRNA test.

Results 3.6% of the women were HPV mRNA-positive at baseline. The overall cumulative rate of CIN3+ was 1.3% (95% CI 1.1% to 1.5%) through 72 months of follow-up, 2.3% for women aged 25–33 years (n=3277) and 0.9% for women aged 34–69 years (n=7943). Cumulative CIN3+ rates by baseline status for HPV mRNA-positive and mRNA-negative women aged 25–33 years were 22.2% (95% CI 14.5% to 29.8%) and 0.9% (95% CI 0.4% to 1.4%), respectively, and 16.6% (95% CI 10.7% to 22.5%) and 0.5% (95% CI 0.4% to 0.7%), respectively, in women aged 34–69 years.

Conclusions The present cumulative incidence of CIN3+ is similar to rates reported in screening studies via HPV DNA tests. Owing to differences in biological rationale and test characteristics, there is a trade-off between sensitivity and specificity that must be balanced when decisions on HPV tests in primary screening are taken. HPV mRNA testing may be used as primary screening for women aged 25–33 years and 34–69 years.