HPV-test i screeningprogrammet mot livmorhalskreft for aldersgruppen 25-33 år: systematisk oversikt

Abstract

Introduction Prolonged infections with human papillomavirus (HPV) have a causative role in 99% of cervical cancer cases. The majority of HPV infections is nevertheless defeated by the immune system without causing cervical cancer. Since 1995, Norwegian women in the age group of 25 to 69 years have been invited to a national screening program for cervical cancer. The screening program has used conventional cytology to assess cell samples, but from 2015 primary HPV-testing has gradually been introduced for the age group of 34 to 69 years. This implementation was completed first quarter 2022. Also in 2022, the first cohort that was offered the HPV vaccine in the national vaccine program will be invited to participate in the national screening program for cervical cancer. Objective The objective of this systematic review was to investigate the effect of combinations of HPV-test and cytology in screening programs for cervical cancer for women aged 25-33 years. We intended to evaluate three interventions: (1) replace primary cytological evaluation with primary HPV-test, (2) use both HPV-test and cytology (cotesting), and (3) use HPV-test as quality assurance for cytology. Separate analyses and assessments were required for HPV-vaccinated and unvaccinated women. Method We conducted a systematic literature search in relevant databases. Two researchers assessed titles, abstracts, and full text of relevant references against the selection criteria. We used EPPI-Reviewer in the selection process, including various machine learning functionalities to streamline the selection process. One researcher extracted and analysed data from the included studies and another checked the data extraction. Two researchers assessed the methodological quality with validated checklists and tools and assessed our confidence in the results using GRADE. We also assessed the transferability to the Norwegian context. Results We included six randomized trials. Four trials investigated the effect of intervention 1 and two investigated the effect of intervention 2. We found no studies evaluating intervention 3. All trials were conducted in national or regional screening programs in high-income countries (in Australia, Sweden, Finland, Netherlands, Italy, and Canada). One trial investigated HPV-vaccinated women, and five trials investigated unvaccinated women. The trials investigated broad age groups, but also reported separate data for the age group that was relevant for this review (25-33 years) and/or for relevant age subgroups (25-29 years or 30-33 years). For HPV-vaccinated women, we found that primary HPV-test possibly results in more referrals to colposcopy and more findings of CIN2+ and CIN3+ than primary cytology in the first round of screening (Summary of findings 1). For unvaccinated women, we found that primary HPV-test compared to primary cytology probably results in more referrals to colposcopy and more findings of CIN2+ and CIN3+ in the first round of screening (Summary of findings 1), and probably less findings of CIN2+ and possibly less findings of CIN3+ in the second round of screening. The effects in the first round of screening were observed in both subgroups (25-29 years and 30-33 years) but were more pronounced for the youngest subgroup.

Summary of findings 1. Primary HPV-test compared to primary cytology for HPV-vaccinated and unvaccinated women aged 25-33 years

Outcome Anticipated absolute effects (95% CI) Relative effect (95% CI) Number of participants Certainty (GRADE) Primary cytology Primary HPV-test HPV-vaccinated women CIN2+ 5 per 1 000 26 per 1 000 (3 to 202) RR 5.55 (0.72 to 42.7) 629 ⨁⨁◯◯ Low a CIN3+ 5 per 1 000 22 per 1 000 (3 to 169) RR 4.54 (0.58 to 35.6) 629 ⨁⨁◯◯ Low a Colposcopy 47 per 1 000 82 per 1 000 (41 to 162) RR 1.72 (0.86 to 3.41) 629 ⨁⨁◯◯ Low a Unvaccinated women CIN2+ 21 per 1 000 25 per 1 000 (22 to 29) RR 1.18 (1.03 to 1.36) 35 234 ⨁⨁⨁◯ Moderate b CIN3+ 14 per 1 000 22 per 1 000 (14 to 34) RR 1.50 (0.96 to 2.35) 4 849 ⨁⨁⨁◯ Moderate c Colposcopy 40 per 1 000 58 per 1 000 (44 to 77) RR 1.45 (1.10 to 1.92) 35 234 ⨁⨁⨁◯ Moderate b a) Extremely broad confidence interval including no effect and both positive and negative effects. Less participants than the optimal information size. We have downregulated by two units. b) One of the studies has high risk of bias. We have downregulated by one unit. c) Broad confidence interval including large and small effects, and few participants. We have downregulated by one unit. The trials investigating intervention 2 evaluated unvaccinated women. We found that HPV-test and cytologi (cotesting) compared to cytology alone possibly results in more referrals to colposcopy and more findings of CIN2+ in the first round of screening (Summary of findings 2), and possibly less findings of CIN2+ in the second round of screening. The evidence for CIN3+ was highly uncertain for this comparison (Summary of findings 2). None of the included studies reported referrals to conization, adverse effects after conization, overtreatment, or quality of life. Summary of findings 2. HPV-test and cytology (cotesting) compared to cytology for unvaccinated women aged 25-33 years

Outcome Anticipated absolute effects (95% CI) Relative effect (95% CI) Number of participants Certainty (GRADE) Cytology Cotesting CIN2+ 14 per 1 000 20 per 1 000 (12 to 32) RR 1.43 (0.89 to 2.29) 18 677 ⨁⨁◯◯ Low a CIN3+ 9 per 1 000 10 per 1 000 (3 to 169) RR 1.07 (0.80 to 1.43) 18 677 ⨁◯◯◯ Very low a,b Colposcopy 41 per 1 000 48 per 1 000 (41 to 162) RR 1.17 (0.93 to 1.48) 6 267 ⨁⨁◯◯ Low a a) Broad confidence interval including no effect and both positive and negative effects. We have downregulated by two units. b) The included studies show opposite effects. We have downregulated by one unit.

Discussion The evidence in this review addresses intervention 1 and intervention 2, but studies evaluating intervention 3 were not found. All the included trials were conducted in national or regional screening programs in high-income countries and should be transferable to the Norwegian context. Most trials evaluating intervention 1 used the test that Norwegian women below 34 years are offered in the national screening program as comparator. This comparison is thus directly transferable to the Norwegian practice. The trials evaluating intervention 2 used cytology alone as comparator, and it is unclear to what extent this intervention would compare to the test women below 34 years are offered in the Norwegian national screening program. Only one trial investigated HPV-vaccinated women, and the number of participants in this trial was low. The evidence for HPV-vaccinated women is thus substantially weaker than for unvaccinated women. Novel studies investigating the effect of screening strategies for HPV-vaccinated women are thus needed. This need is urgent because the first cohort of women that were offered the HPV-vaccine in the Norwegian vaccine program are invited to the national screening program for cervical cancer in 2022. Conclusion Replacing primary cytology with primary HPV-test probably results in more referrals to colposcopy and earlier detection of CIN2+ and CIN3+ for unvaccinated women below 34 years. HPV-test and cytology (cotesting) possibly results in more referrals to colposcopy and earlier detection of CIN2+ than cytology alone. The evidence for HPV-vaccinated women is substantially weaker than for unvaccinated women. Primary HPV-test possibly results in more referrals to colposcopy and earlier detection of CIN2+ and CIN3+ than primary cytology, but novel studies investigating effects of screening strategies for HPV-vaccinated women are needed.