Detecting cervical precancer and reaching underscreened women by using HPV testing on self samples: updated meta-analysesBMJ 2018; 363 doi: https://doi.org/10.1136/bmj.k4823 (Published 05 December 2018) Cite this as: BMJ 2018;363:k4823
All rapid responses
We thank R Holman for her interest in our meta-analysis on high-risk HPV testing on self-samples . We agree that the editor-made summary of our meta-analyses in the printed BMJ issue was too short to describe all important aspects of the methods and results of our meta-analysis including standard terms for screen test performance and efficacy assessments. We accept the criticism in What this study adds and support the change of “accurate diagnostic methods” into “accurate tests”.
We want to clarify that the evaluated test is the application of a high-risk HPV (hrHPV) assay on a vaginal self-sample and on a clinician-taken cervical sample, and the disease outcome is the presence or absence of cervical intraepithelial neoplasia of grade 2 or worse (CIN2+) or grade 3 or worse (CIN3+) . The screen test should not be confused with the disease outcome. The test consists in a laboratory procedure in which a certain sequence of the viral genome (or their transcripts) from high-risk human papillomavirus types are identified. The diagnosis of the disease outcome consists in the histological examination of a punch or excision biopsy accepting negative satisfactory colposcopy without biopsy as sufficient assurance of absence of CIN2+. R Holman proposes describing the HPV test as a “triage of those at higher risk of having cervical pre-cancer”. We disagree with this terminology. Triage has a very specific meaning, which is an intermediate step in the screening-triage-diagnosis chain. A distinct triage is needed for women with a positive HPV screen test result and only those who are triage positive are referred for diagnostic work-up . Triage of women with a HPV-positive self-sample is not a study question but its importance is discussed. Triage of women with a hrHPV-positive self-sample by cytology requires an additional visit at a clinic for taking a Pap smear and this may result in significant losses to follow-up. A new systematic review on the accuracy of molecular triage tests feasible on self-samples is currently ongoing at the Belgian Cancer Centre.
The formulation of study questions was discussed and tailored by cervical cancer experts, guidelines specialists, clinical epidemiologists and colposcopists at the 2017 conference of the American Society for Colposcopy and Cervical Pathology (Orlando, USA). These questions were adequately phrased in the print summary. We do not agree with Holman’s suggested phrasing: “Is HrHPV testing an effective way to screen women for further investigation for CIN2+?“ This alternative question mixes the two separate investigations assessed in each of the meta-analyses: a) an accuracy question on HPV testing and b) an efficacy question whether the offer of a self-sample for HPV testing is more effective to reach under-screened or not-screened women. The efficacy meta-analysis comprised secondary objectives, such as the adherence to follow-up among screen-positive women and the detection of CIN2+ among invited and screened women; however most included trials were not empowered to assess these secondary outcomes.
The separation between HPV testing with a signal-amplification (SA) assay or a polymerase-chain reaction assay, was clearly expressed in the methods section and builds further on the conclusions of a previous meta-analysis [3;4]. Our updated meta-analysis confirmed that testing with a SA-based HPV assay is less sensitive for CIN2+on self- compared to clinician- samples.
M Arbyn1, P Castle2
1. Sciensano, Brussels, Belgium
2. Albert Einstein College of Medicine, New York, USA
(1) Arbyn M, Smith SB, Temin S, Sultana F, Castle PE, the Collaboration on Self-Sampling and HPV Testing. HPV testing on self-samples: accuracy to detect cervical precancer and efficacy to reach under-screened women. An update of meta-analyses. BMJ 2018; 363:k4823.
(2) Castle PE, Katki HA. Screening: A risk-based framework to decide who benefits from screening. Nat Rev Clin Oncol 2016; 13:531-2.
(3) Arbyn M, Verdoodt F, Snijders PJF et al. Accuracy of human papillomavirus testing on self-collected versus clinician-collected samples: a meta-analysis. Lancet Oncol 2014; 15:172-83.
(4) Arbyn M, Castle P. Offering self-sampling kits for HPV testing to reach women who do not attend in the regular cervical cancer screening program. Cancer Epidemiol Biomarkers Prev 2015; 24:769-72.
Competing interests: MA was supported by grants from Centers for Disease Control and Prevention (USA), the 7th Framework Programme of the European Directorate for Research and Innovation (through the COHEAHR network), the French High Authority for Health, the Dutch National Institute of Public Health and the Environment and VALGENT/VALHUDES. PC has received cervical screening tests and diagnostics from Roche, Becton Dickinson, Cepheid, and Arbor Vita Corporation at a reduced or no cost for research.
Re: Detecting cervical precancer and reaching underscreened women by using HPV testing on self samples: updated meta-analyses
I found the print summary of this article confusing
My main problem is with the statement in 'what this study adds' where it says hrHPV (high risk human papilloma virus) assays based on PCR (polymerase chain reaction) on self-samples are accurate diagnostic methods for detecting cervical precancer as this suggests an HrHPV test detects CIN (cervical intraepithelial neoplasia). An HPV test cannot diagnose CIN because CIN is a histological diagnosis. An HPV test triages those at higher risk of having cervical pre-cancer. I think it would be better to say that "Hr HPV tests are effective in detecting women needing further investigation".
The study question section and methods sections similarly suggests that high risk HPV tests directly detect pre-cancer. The methods say that Hr HPV assays are sensitive and specific to detect CIN2+. How can they do this if CIN is a histological diagnosis? The type of HPV test or the way it is submitted may lead to detecting more women who are subsequently found to have CIN but the HPV test itself is not a test for CIN, which is the impression given.
I also think the way the study question section is phrased implies that it is something about the clinician's taking of the sample that matters, whereas in the methods it is stated that it is comparing the take up rates of getting a letter for a sample compared to being sent a self sampling kit i.e. participation rates
I suggest the study questions should be:
Are participation rates higher when more women are offered self sampling compared to being invited to attend a clinician?
Is HrHPV testing an effective way to screen women for further investigation for CIN2+ ?
The paper article is also confusing because two types of HPV test are compared in the diagrams but this is not explained in the methods or the study question but suddenly appears in the study answer.
I think a better title would be "Role of HPV self sampling in reaching women underscreened for cervical pre-cancer".
Competing interests: No competing interests