Reply to: Comments on cumulative risk of cervical intraepithelial neoplasia for women with normal cytology but positive for human papillomavirus: Systematic review and meta‐analysis

Abstract

In their letter to the editor, Jayaraj et al have raised several points regarding our recent systematic review and meta-analysis on the risk of progression to precancer for human papillomavirus (HPV) positive but cytology normal women.1 We welcome the opportunity to further address these points, which may have been less clear in our original paper. We start with their most relevant point, and the last in their letter, about our adherence with PRISMA guidelines. Although we did not specifically mention the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines in our publication, we assure readers that we did use PRISMA reporting guidelines when drafting the article.2 Readers will find that all PRISMA items are reported in the publication and its appendix.1 Jayaraj et al stated that the categorisation of high-risk (HR) HPV types was not detailed for readers. This reflects the fact that we used each individual study's definition of what it considered to be HR-HPV types, which differed slightly between studies due to differences in the HPV assays used. However, the vast majority of studies in our review considered the same 13 to 14 HPV types as HR (HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66 and 68). These are the types detected by commercial DNA assays approved for cervical cancer screening by the Food and Drug Administration,3 and HPV-DNA tests have generally been found to have comparable sensitivity and specificity in a screening population.4 Therefore, our analysis for grouped HR-HPV types is an accurate reflection of the risk associated with these 14 HPV types at the time the studies were performed. It is likely that much of the heterogeneity in cervical precancer risk between populations is due to differences in HPV-type distribution between populations, rather than differences in types detectable between assays. This is an issue that affects not just our study, but all observational and clinical trials of HPV tests, and screening in general: the underlying predictive value of a positive test result for grouped HPV types will always depend on the underlying distribution of these types in a population. This is why our type-specific HPV results are likely to be very important going forward as HPV-16 and HPV-18 prevalence is declining in many populations due to HPV vaccination.5 We disagree with Jayaraj et al that HPV type-specific progression risks are not useful for informing risk-based cervical cancer screening guidelines, given the numerous studies which found that HPV type is a strong predictor of cervical precancer risk, and which we cited in our review. Recently updated cervical cancer screening guidelines for the United States6 and Australia7 integrate HPV genotyping as a tool to stratify HPV-positive women based on their risk of high-grade cervical lesions. We expect that more countries will follow suit as they adopt HPV testing and revise their approaches to cervical cancer screening. The current perspective espoused by many scientists is that extended HPV genotyping is a promising molecular biomarker for further risk stratification of women with positive HPV test results,8 especially in the context of HPV self-sampling.9 Finally, Jayaraj et al's inappropriate argument that only 5% of all cancers are caused by HPV is not relevant here; we examined risk for women who are HPV positive in the context of cervical cancer screening, where it is established that HPV is a necessary cause of this disease; virtually all cervical cancers worldwide are caused by infection with HPV.10 ELF has served as occasional consultant to Merck and GSK on HPV vaccines, to Roche and to BD on HPV diagnostics. His institution has received unrestricted grants from Merck. ELF holds a patent related to the discovery “DNA methylation markers for early detection of cervical cancer”, registered at the Office of Innovation and Partnerships, McGill University, Montreal, Quebec, Canada (October, 2018). A provisional utility patent application before the United States Patent & Trademark Office was also filed (November, 2018).