Background
Risk-based HPV screening strategies aim to optimize early detection of cervical cancer precursors while limiting unnecessary follow-up and healthcare resource use. In 2021, the Capital Region of Denmark implemented primary HPV screening with genotype-based risk stratification among women aged 30–59 years, using differentiated follow-up pathways for HPV positive women from immediate colposcopy to repeat testing at 12 and 24 months. Evidence from routine, population-based implementation is needed to inform screening policy and optimization.
Objective
To evaluate the population-level performance of HPV genotype-based risk stratification in routine cervical screening, focusing on timing of detection of high-grade lesions and implications for resource utilization.
Methods
All HPV-positive screening samples from women aged 30-59 years in the Capital Region of Denmark during 2021-2023 were included. HPV genotyping was available for all samples and follow-up was minimum two years. Individual-level linkage to the Danish Pathology Data Bank ensured complete follow-up with cytological and histological outcomes. Disease endpoints were classified as ≤CIN1, CIN2+, and CIN3+, and analysed by genotype-based risk group and by timing of detection (index screening, 12-month, and 24-month follow-up). Highest risk HPV genotypes (High-5) were HPV16, 18, 31, 33, 52, lowest risk HPV genotypes (Low-8) were HPV35, 39, 45, 51, 56, 59, 66, 68.
Results
Among 110,720 HPV screened women, 10,263 women tested HPV positive, and throughout the algorithm, 5,146 (50.1%) underwent colposcopy during follow-up. Colposcopy outcomes were ≤CIN1 in 58.1%, CIN2 in 21.2%, and ≥CIN3 in 20.7%.
Detection of ≥CIN3 occurred predominantly early in the screening pathway: 70.3% at index screening and 28.7% at 12-month follow-up, with only 1.0% detected at 24 months. In contrast, ≤CIN1 findings accumulated mainly at later follow-up visits. High-5 genotypes accounted for 87.1% of ≥CIN3 lesions across all time points, whereas Low-8 genotypes contributed disproportionately to ≤CIN1 outcomes, particularly during follow-up and surveillance.
Colposcopy yield declined substantially with increasing follow-up time, with an increase in the number needed to colposcope to detect one ≥CIN2 lesion from 1.74 at index screening to 3.01 at 12 months and 6.45 at 24 months.
Conclusion
In this real-world, large-scale population-based evaluation, HPV genotype-based risk stratification effectively concentrated detection of ≥CIN3 early in the screening pathway, whereas diagnostic yield of extended follow-up provided limited disease detection. Among women with Low-8 genotypes, high-grade disease was rare at any follow-up despite frequent low-grade findings, suggesting limited clinical benefit of intensive surveillance in this subgroup.
These findings support use of risk-based cervical screening strategies independent of setting. Learnings show that de-escalation of follow-up for selected lower-risk genotypes by extended surveillance periods may improve resource efficiency without compromising protection against high-grade disease.