Benefits and potential harms of human papillomavirus (HPV)‐based cervical cancer screening: A real‐world comparison of HPV testing versus cytology

Abstract

Introduction: Human papillomavirus (HPV) testing as the primary cervical cancer screening method is implemented in several countries. We report data from the first round of a large Danish pilot implementation of HPV‐based screening. Our aim was to compare colposcopy referrals, detection of high‐grade cervical intraepithelial neoplasia (CIN) and cervical cancer, and positive predictive value (PPV) of colposcopy referral in HPV vs cytology‐based screening. Material and methods: From May 2017 to October 2018, women aged 30‐59 years attending cervical cancer screening in the uptake area of the Department of Pathology, Vejle Hospital, Region of Southern Denmark were screened by primary HPV testing (n = 16 067) or primary cytology (n = 23 981) depending on municipality of residence. In the HPV group, women with HPV16/18, or other high‐risk HPV types and abnormal cytology, were referred to immediate colposcopy. Women with other high‐risk HPV types and normal cytology were invited for repeat screening with HPV test and cytology after 12 months. From a nationwide pathology register, we obtained information on screening results and subsequent histological diagnoses during up to 2.9 years after the first screen. PPVs included diagnoses within 1 year after referral. Results: In the HPV group, 3.7% were referred to immediate colposcopy and 2.8% were referred at the 12‐month repeat screening. The total referral to colposcopy was higher in the HPV (6.6%) than cytology group (2.1%) (age‐adjusted relative referral = 3.05, 95% confidence interval [CI] 2.75‐3.38). The detection of CIN3+ was higher in the HPV (1.5%) than the cytology group (0.8%) (age‐adjusted relative detection = 1.88, 95% CI 1.56‐2.28). The probability of CIN3+ among women referred to colposcopy (= PPV) was lower in the HPV (21.1%; 95% CI 18.7%‐23.7%) than the cytology group (34.6%; 95% CI 30.7%‐38.9%). In the HPV group, the PPV was lower among women referred at repeat screening (12.1%) than among women referred immediately (27.8%). Conclusions: Compared with cytology‐based screening, HPV‐based screening provided a 90% increased CIN3+ detection at the cost of a threefold increase in colposcopy referrals, when considering complete data from the prevalence round. Our findings support implementation of HPV‐based screening in Denmark, but modifications of screening algorithms may be warranted to decrease unnecessary colposcopy referrals.