Evaluation of primary HPV-based cervical screening among older women: Long-term follow-up of a randomized healthcare policy trial in Sweden

Background Evidence on invasive cervical cancer prevention among older women is limited, especially with the introduction of human papillomavirus (HPV)-based screening and longer interval. We conducted a long-term follow-up of the first phase of a randomized healthcare policy trial in cervical screening, targeting women aged 56 to 61 years old, to investigate the effectiveness of primary HPV-based screening in preventing invasive cervical cancer (ICC) and the safety of extending screening interval. Methods and findings The randomized healthcare policy trial of primary HPV-based cervical screening targeted women residing in Stockholm-Gotland region during 2012 to 2016, aged 30 to 64 years. The trial aimed to investigate the detection rate of cervical intraepithelial neoplasia grade 2 or worse (CIN2+) within 24 months and long-term protection against invasive cervical cancer, comparing primary HPV-based screening to primary cytology-based screening. The initial phase of the trial, which was the focus of this study, targeted women aged 56 to 61 years old in 2012 to 2014 who were randomized to primary cytology arm (n = 7,401) or primary HPV arm (n = 7,318). We used national registries to identify the subsequent cervical tests and all histopathological diagnoses including ICC before December 31, 2022. We calculated cumulative incidence, incidence rate (IR) and IR ratio (IRR) of ICC, by baseline test result. Furthermore, we calculated longitudinal sensitivity and specificity for detecting cervical intraepithelial neoplasia grade 2 or worse (CIN2+) by receipt of primary cytology or primary HPV test for the recommended screening intervals in this age group. We found that the IR of ICC among women in the primary HPV arm was 7.2/100,000 person-years (py) and 3.0 for women who tested HPV negative, compared to 18.4/100,000 py among women in the primary cytology arm and 18.8 for women who tested cytology negative. We further found that the overall point estimate for the risk of ICC over 10 years of follow-up among women in the primary HPV arm was 0.39 compared to women in the primary cytology arm, but this was not statistically significant (IRR: 0.39; 95% confidence interval, CI [0.14, 1.09]; p = 0.0726). However, among women with a negative test result at baseline, women in the primary HPV arm had an 84% lower risk of ICC compared to women in the primary cytology arm (IRR: 0.16; 95% CI [0.04, 0.72]; p = 0.0163). Moreover, primary HPV testing had a higher sensitivity for detecting CIN2+ within a 7-year interval than primary cytology testing within a 5-year interval (89.6% versus 50.9%, p < 0.0001). We were limited by a partial imbalance of invitations during the follow-up between the 2 arms which may have led to an underestimation of the effectiveness of primary HPV-based screening. Methods and findings The randomized healthcare policy trial of primary HPV-based cervical screening targeted women residing in Stockholm-Gotland region during 2012 to 2016, aged 30 to 64 years. The trial aimed to investigate the detection rate of cervical intraepithelial neoplasia grade 2 or worse (CIN2+) within 24 months and long-term protection against invasive cervical cancer, comparing primary HPV-based screening to primary cytology-based screening. The initial phase of the trial, which was the focus of this study, targeted women aged 56 to 61 years old in 2012 to 2014 who were randomized to primary cytology arm (n = 7,401) or primary HPV arm (n = 7,318). We used national registries to identify the subsequent cervical tests and all histopathological diagnoses including ICC before December 31, 2022. We calculated cumulative incidence, incidence rate (IR) and IR ratio (IRR) of ICC, by baseline test result. Furthermore, we calculated longitudinal sensitivity and specificity for detecting cervical intraepithelial neoplasia grade 2 or worse (CIN2+) by receipt of primary cytology or primary HPV test for the recommended screening intervals in this age group. We found that the IR of ICC among women in the primary HPV arm was 7.2/100,000 person-years (py) and 3.0 for women who tested HPV negative, compared to 18.4/100,000 py among women in the primary cytology arm and 18.8 for women who tested cytology negative. We further found that the overall point estimate for the risk of ICC over 10 years of follow-up among women in the primary HPV arm was 0.39 compared to women in the primary cytology arm, but this was not statistically significant (IRR: 0.39; 95% confidence interval, CI [0.14, 1.09]; p = 0.0726). However, among women with a negative test result at baseline, women in the primary HPV arm had an 84% lower risk of ICC compared to women in the primary cytology arm (IRR: 0.16; 95% CI [0.04, 0.72]; p = 0.0163). Moreover, primary HPV testing had a higher sensitivity for detecting CIN2+ within a 7-year interval than primary cytology testing within a 5-year interval (89.6% versus 50.9%, p < 0.0001). We were limited by a partial imbalance of invitations during the follow-up between the 2 arms which may have led to an underestimation of the effectiveness of primary HPV-based screening. Conclusions In this study, we observed that women over 55 years of age who received a primary negative HPV test result had substantially lower risk of CIN2+, and ICC, compared to women who received a primary negative cytology result. This should apply even if the screening interval were prolonged to 7 years.