The relationship between hormonal contraception and cervical dysplasia/cancer controlling for human papillomavirus infection: A systematic review

Objective Studies on the effect of long-term use of combined oral contraceptives (COCs) on cervical dysplasia and/or cancer risk have been inconsistent. Less is known about the effects of other forms of hormonal contraception (HC). We examine whether HC use increases the risk of incident cervical intraepithelial neoplasia (CIN) 2, 3 and/or cancer after accounting for preexisting human papillomavirus (HPV) infection. Study Design Systematic review of prospective studies on HC use as risk factor for cervical dysplasia with HPV infection documented prior to outcome assessment including PubMed and EMBASE records between January 2000 and February 2020 (Prospero #CRD42019130725). Results Among nine eligible studies, seven described recency and type of HC use and therefore comprise the primary analysis; two studies limit comparisons to ever versus never use and are summarized separately. All seven studies explored the relationship between oral contraceptive (OC) use and cervical dysplasia/cancer incidence: two found increased risk (adjusted odds ratio, aOR = 1.5–2.7), one found no association but decreased risk when restricted to women with persistent HPV (adjusted hazard ratio = 0.5), and four found no association. None of the seven studies differentiated between COC and progestin-only pills (POPs) by use recency or duration. The only study that included injectable progestin-only contraception (DMPA) found increased CIN3 incidence among current versus never users (aOR = 1.6). The one study that included Norplant found no association. Two studies included intrauterine device (IUD) use, but did not differentiate between hormonal and copper IUDs, and found no association. Conclusion We found no consistent evidence that OC use is associated with increased risk for cervical dysplasia/cancer after controlling for HPV infection. There were too few studies of progestin-only injectables, implants or IUDs to assess their effect on cervical dysplasia/cancer risk. Implications Use of single self-reported HC measures and insufficient distinction by hormonal constituent cloud our understanding of whether some HCs increase risk for cervical cancer. Methodologically rigorous studies with distinct HCs measured as time-varying exposures are needed to inform cervical cancer prevention efforts and improve our understanding of cervical cancer etiology.