Background: Human papillomavirus infection with high-risk types (HR-HPV) is a necessary cause of cervical cancer, the most common malignancy among sub-Saharan African women. Little is known about prevalence of cervical HR-HPV infection in this region.
Methods: A cross-sectional study of 1528 women examined the determinants of HR-HPV infection among women in Kinshasa, Democratic Republic of Congo. Information was collected on sociodemographic, reproductive, lifestyle characteristics, and health-seeking behaviors. Cervical samples were tested for HPV-DNA by Hybrid Capture 2. Unconditional logistic regression identified predictors of HPV positivity.
Results: HR-HPV prevalence was 12.5% in all women and 8.7% in women with normal cytology. Prevalence was highest (18.3%) in individuals <35 years of age and gradually decreased with age. Excess HR-HPV infection risk was observed in women who were smokers (odds ratio [OR] = 1.60; 95% confidence interval [CI]: 1.11–2.31), divorced/separated (OR = 1.60; 95% CI: 1.11–2.32), in polygamous marriages (OR = 1.28; 95% CI: 0.90–1.82), using medical contraceptives (OR = 2.40; 95% CI: 1.20–4.80), and who preferred male physicians (OR = 1.90; 95% CI: 1.20–3.05). A statistically marginal increase was found in women whose partners had sex with prostitutes (OR = 2.40; 95% CI: 0.72–8.01). A higher standard of living was associated with reduced risk.
Conclusion: HR-HPV positivity was associated with behavioral and sexual characteristics thought to affect risk of new infections and immune function. However, HPV prevalence did not correlate with numbers of sex partners, possibly because of a high HPV infection rate per sexual contact or because subjects were older than 30 years. Our study should assist in designing strategies for control of cervical cancer in this low-resource, high cervical cancer risk setting in sub-Saharan Africa.
This community-based cross-sectional study analyzed high-risk types of human papillomavirus DNA infection prevalence and risk factors in an unscreened population in a low-resource and high-risk setting in Kinshasa, Democratic Republic of Congo.
From the *Department of Obstetrics and Gynaecology, University of Kinshasa, Kinshasa, Democratic Republic of Congo; †Faculté Saint-Jean, University of Alberta, Edmonton, Canada; ‡Division of Cancer Epidemiology, Department of Oncology, McGill University, Montreal, Canada; §Groupement de Recherche Cytologique, Lyon, France; ¶Ecole de Santé Publique de Kinshasa, Kinshasa, Democratic Republic of Congo; and ∥Wolfson Institute of Preventive Medicine, Barts and The London School of Medicine and Dentistry, London, England
Supported by PATH via the Alliance for Cervical Cancer Prevention, funded by The Melinda and Bill Gates Foundation, for fieldwork. Denis Vincent, MD, and Martine Roy, MD, financially supported the project “Santé ya ba Mama.” Additional funding support was also provided by the Canadian Institutes of Health Research (CIHR) via a grant to the CIHR Team in HPV infection and Associated Diseases.
The authors do not have commercial or other associations that might pose a conflict of interest.
Correspondence: Eduardo L. Franco, DrPH, Division of Cancer Epidemiology, Department of Oncology, McGill University, 546 Pine Ave West, Montreal, Quebec H2W1S6, Canada. E-mail: email@example.com.
Received for publication May 21, 2010, and accepted September 8, 2010.
Persistent infection with high risk types of human papillomavirus (HR-HPV) is a necessary cause of the majority of cervical cancers,1,2 the most common malignancy among women in sub-Saharan Africa.3 The overall prevalence of cervical HPV infection differs substantially from one population to another throughout the world. Age-adjusted prevalence rates for sexually active women of screening age range from 5.7% to 29.1% in most regions in the Americas, Europe, and Asia, whereas in African regions estimates range from 15.5% to 35.4%.4 However, little is known about the prevalence and risk factors of HPV in sub-Saharan Africa. HR-HPV infection prevalence ranged from 12% in a community-based study in South Africa5 to 46% in women attending antenatal and gynecological clinics in Gabon.6
Geographical differences in HPV prevalence could be explained by variations in exposure to risk factors. Among them are characteristics related to sexual activity, oral contraceptive use, high parity, lower socioeconomic status, lifestyle, male sexual partners, and immunosuppression.7,8 Understanding the role of determinants of acquisition of persistent HPV infection is an important first step in the development of effective strategies aimed at preventing cervical cancer. The objective of this study was to examine the prevalence and determinants of HR-HPV infection among women recruited in a population-based screening study conducted in Kinshasa, Democratic Republic of Congo (DRC).
We analyzed data from a community-based cross-sectional study of the efficacy of visual inspection, Papanicolaou cytology, and HR-HPV DNA testing as candidate screening methods for detecting cervical cancer and its precursor lesions in a primary healthcare setting. Between November 2003 and April 2004, the fieldwork was conducted in the community of Mbuku, Kinshasa, the DRC capital. The study methods have been presented elsewhere.9 In brief, all adult women were approached and invited to participate. The criteria for eligibility were age ≥30 years, had an intact uterus, and were not pregnant. Those who had been treated for cervical neoplastic disease in the past 6 months were excluded. The purpose and interventions of study were explained to the women before they consented to participate. The PATH Human Subjects Protection Committee and the Institutional Review Board of the Kinshasa School of Public Health approved the study protocol.
A questionnaire was designed to collect information through nurse-administered interviews on sociodemographic, reproductive, clinical, lifestyle, and sexual behavior variables; individual health-seeking behavior; standard of living indicators; and sexual behavior of long-term partners. A standard of living index was constructed from a set of 9 indicators, such as, ownership of house, availability of tap water, toilet, and electricity in the house, and ownership of radio, TV, car, refrigerator, and other goods. Each item was assigned a score ranging from 0 to 5, with the total score for the index ranging between 0 and 45. Cervical specimens for cervical cytology and HPV testing were collected by trained nurses. Conventional Papanicolau smears were read offsite (Lyon) and the remainder of the specimens shipped to Digene Corporation (later acquired by Qiagen, Gaithersburg, United States) where they were tested by one of the coauthors (A.L.) with the Hybrid Capture 2 (HC2) assay for 13 HR-HPV types (Types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) according to the manufacturer's recommendations. A relative light unit cut-off value >1 was used for defining HR-HPV positivity (the outcome variable) in most analyses, and an relative light unit >2 was used additionally for the multivariate analyses.
Unconditional logistic regression was used to compute age-adjusted odds ratios for putative risk factors for HR-HPV positivity; selection of independent variables into multivariate models was based on the likelihood ratio test, and a stepwise regression strategy was based on significant levels of 0.1 and 0.15 for entry and removal of variables. Multivariate models are presented for 2 outcomes: any HR-HPV infection and HR-HPV infection among cytology negative subjects. Models were adjusted for age, standard of living, use of contraception, gender preference of consultant, tobacco use, and preference of health care facility, number of wives, marital status, and sexual relations before the union.
Of the approximately 2000 women meeting the age eligibility criteria and who lived in Mbuku during the study, 1699 women were invited to participate, whereas only 1571 gave informed consent. Among the latter, 1528 women were enrolled in the study and were tested for HR-HPV DNA. After excluding 176 inadequate samples, the HR-HPV prevalence was 12.5% in all women and 8.7% in women with normal cytology. The age-specific HR-HPV prevalence rates for the latter 2 groups are shown in Figure 1. Prevalence was highest (18.3%) among women younger than 35 years of age and declined with age for those aged 40 to 64 years.
Table 1 shows the relationship between HR-HPV positivity and sociodemographic variables separately for all women and for those with negative Papanicolau results. As expected, there was a negative association with age in both sets of analysis. A high standard of living was associated with a significant reduction in infection risk of about 50%. Not being currently married or living with a male partner was associated with a significantly increased infection risk although the risk associated with sharing the house with other cowives was only marginal and nonsignificant. No other sociodemographic characteristics seemed to affect risk consistently.
Table 2 shows the relationship between HR-HPV positivity and reproductive and clinical variables. No association was found with parity, age at menarche, or sexually transmitted disease history. Infection risk was slightly higher for women who had become pregnant for the first time after the age of 18. Use of medical contraception was associated with a doubling in infection risk. Women who reported having generally consulted a gynecologist or physician had a slightly higher, albeit nonsignificant, risk than those who consulted nurses or traditional healers for their gynecologic ailments. Women who favored consulting with male health professionals were at higher risk compared with those who did not have a preference or favored consultations with female health providers. A few women reported having previously had a Pap smear. They were at a higher risk of infections but this was likely a reflection of their being investigated for HPV-associated Pap abnormalities.
Table 3 shows the relationship between HPV positivity and lifestyle and sexual behavior of the participant herself. Tobacco use (smoking and chewing) was associated with 60% to 110% increased risk. There was no association between infection risk and age at first sexual intercourse (<15 years) or at regular intercourse. Having multiple lifetime sexual partners was associated with risk more noticeably among those with normal Pap smears.
Table 4 shows the relationship between participants' HPV positivity and characteristics of regular partners, as reported by participants. Having a partner of age ≥35 was associated with reduced risk of infection, albeit nonsignificantly. Knowledge that the partner visited commercial sex workers was associated with an increase in risk, although estimates lacked precision. In contrast, the woman's knowledge of whether her current partner or husband had had sexual relations with other unspecified female partners, either before or after the present union, did not seem to affect risk. Her knowledge of the partner's STI history also did not seem to influence risk.
Table 5 shows the results from multivariate models that selected all independent predictors of HR-HPV infections based on 2 separate thresholds for defining HPV positivity. Age, standard of living index, contraceptive method, gender of health provider, smoking, type of healthcare provider (general practitioner or gynecologist vs. nurse or traditional healer), polygamous marriage, marital status, and sexual relations before union were independently associated with risk in at least one of the models shown. Surprisingly, the history of partners' sexual relations was associated with a reduced risk of infection. Use of distinct positivity cutoff definitions did not yield materially different results.
In comparison with other geographical regions, few data are available on the epidemiology of HPV infection in Africa.10,11 To our knowledge, this population-based study on the prevalence and determinants of HPV infection is the first investigation of its kind in the DRC and in Central Africa. Our analyses extend existing knowledge on HPV epidemiology in Africa and are novel for the DRC. Data from this and similar analyses will assist in characterizing the burden of HPV infections and the identification of modifiable risk factors in Africa, thus informing strategies for cervical cancer prevention, including screening and HPV vaccination.
The rate of HR-HPV prevalence among all women in this study was 12.5%, whereas it was 8.7% in women with normal cytology. In general, comparing HPV prevalence across studies is complicated by factors such as, differences among study populations (e.g., age profiles, urban vs. rural setting, and clinic-based vs. general population), variation in defining HPV types to be detected (i.e., overall HPV positivity vs. HR-HPV types), HPV assays (i.e., target or signal amplification), and sampling (e.g., cervical vs. cervicovaginal specimens). Despite these sources of variability, HR-HPV prevalence in this study is consistent with that reported by most population-based studies in Africa5,10,12–14 which are between 12% in South Africa5 and 26% in Nigeria.13 However, prevalences as high as 36% and 46% have been reported for Mozambique15 and Gabon,6 respectively, through investigations of study populations that included younger and more self-selected women attending antenatal or gynecologic clinics, and tested for HPV DNA with polymerase chain reaction methods. The HR-HPV prevalence found in our study is also lower than the rate (30.8%) found among women attending planning family clinics in the urban setting of Nairobi in Kenya.16 Such high HPV prevalence is reported to be related to concomitant HIV infection. Indeed, the HIV prevalence in that study group was 11.5%. Although we did not collect blood samples for HIV testing, it is unlikely that a high prevalence of HIV-associated immunosuppression would have affected our results. A 2008 UNAIDS report indicates that HIV prevalence in Kinshasa is between 4% among antenatal clinic attendees and 12% among urban commercial sex workers.17 Our study location was in the suburb of Kinshasa, its population is mainly a rural population where 75% of the women are married and engaged in monogamous unions. This population may exhibit more conservative sexual behaviors than those of an urban population. Therefore, a lower prevalence of concomitant HIV can be expected for this suburban population than that of the inner part of Kinshasa.
HPV assay type is of major importance in cross-geographical comparisons. Our study was conducted as a cervical cancer screening investigation in which we compared cytology, HPV DNA testing, and visual inspection as modalities for detecting cervical lesions.9 We used the HC2 assay for HR-HPV DNA detection, an assay that probes simultaneously for 13 HR-HPVs and has a higher threshold of viral load than polymerase chain reaction protocols because it has been validated for clinical purposes. Therefore, it is likely that the prevalence estimates we observed are lower than what could have inferred if we had used more analytically sensitive protocols, such as polymerase chain reaction assays, which also test for HPV types other than those probed by the HC2 assay. However, the HR-HPV prevalence in Kinshasa can be considered high when compared with a sample of women attending screening in Montreal, a Canadian city. By applying to our Kinshasa sample, the age-specific HR-HPV prevalences in the age range of 30 to 69 years from a study we conducted in that city with the same assay (HC2)18 we would have expected to identify 105 HR-HPV positive cases of 1312 women in that age range. This estimate is more than one-third lower than what was effectively observed, that is, 167 HR-HPV positive cases.
We observed a decreasing rate of HR-HPV infection with increasing age. Although this pattern is similar to that found in some African countries,15,16,19 it is different from that described in Nigeria and some low-income Asian countries where HPV prevalence was similar across age groups.20 In other studies from Sub-Saharan countries (South Africa, Tanzania, Gambia, and Senegal), the HPV prevalence remained high, or even increased, in middle and old age.21
Relative to the youngest age group, HR-HPV infection risk was reduced by half for women aged 45 to 49 years, which could be associated with age-related changes in sexual behavior, mainly among men. Apart from the loss in precision, there were no noticeable differences between results for women younger than 45 years and those 45 and older when we examined associations with sexual activity variables for the woman and her partner (data not shown). As for couples in this region, males are generally much older than the women; it would be of interest to evaluate further how male-to-female transmission could have affected risk reduction. Some studies also attributed this age-specific risk reduction to the development of type-specific immune response to HPV infection.22 This immune protection may not be permanent, however, could explain the rise of HPV prevalence observed in postmenopausal women.1 Among other reasons for this increase could be the loss of specific immunity due to hormonal influences during the peri- and postmenopausal periods. It may also originate from a recent HPV reinfection.7
Risk reduction was consistently and independently associated with a high standard of living. This is in agreement with South African studies that found that earning opportunities reduced women's risk behaviors such as less involvement in unsafe sexual activity.23 In fact, low socioeconomic status appears to be a determinant of increased high risk sexual activity. This inverse association with socioeconomic level has also been shown in Latin-American studies.24,25
We found statistically marginal increases in HR-HPV risk for women who reported that their partners kept regular sexual contact with prostitutes. Previous studies have shown that a husband's sexual contact with prostitutes may be an even more important source of his wife's HPV infections than the number of sexual partners she had over her lifetime.26 These widespread sexual practices in combination with reporting biases limit the ability of cross-sectional studies to properly measure the association between partner's sexual behavior and HPV risk.
Tobacco smoking was also an independent predictor of infection risk, even after controlling for all other determinants, including those related to sexual behavior. This finding is consistent with data from epidemiologic studies.27,28 Residual confounding by unmeasured dimensions of sexual behavior seems unlikely because a much larger cross-sectional survey confirmed the association despite extensive adjustment for confounders.28 The credibility of the association with smoking hinges also on its biologic plausibility, as female smokers have reduced levels of immune effector cells in the cervix and thus could be more susceptible to acquiring new infections and delayed clearance of existing ones.29 Downstream (from HPV) mechanisms by which cigarette smoking may affect cervical cancer risk are also plausible, e.g., through a direct effect of smoking metabolites and reduced dietary antioxidants.30
HR-HPV infection risk was about 2-fold higher in women who reported using medically prescribed contraception methods. As suggested by others,31,32 unless detailed information is collected on the use of barrier methods of contraception, they may simply serve as a marker for “high-risk” sexual behaviors. Despite the establishment of the National Program on Reproductive Health in the DRC in 1983, which sought to promote contraceptive use, the proportion of women of reproductive age who regularly use modern contraceptive methods remains consistently <5%.33 Given the pronatalist culture that prevails in DRC, contraception use is not common among married women. The use of medical contraception in this population seems to be a marker of more liberal sexual behavior, which is not in accordance with local cultural values. Other factors that affect the use of modern contraceptive methods in Africa include socioeconomic status, being the head of the household, husband's approval, level of education, and urbanization and work status.33 Concerning nonbarrier methods, the mechanism by which the use of contraception influences infection involves the enhancement of estrogen effects on cervical ectopy, thereby permitting more ready exposure of the transformation zone to HPV and increased cell proliferation and HPV transcription induced by direct hormonal effects on cervical cells.34,35
There were no noticeable associations with the traditional cofactors of HR-HPV infection or cervical cancer risk (e.g., age at first intercourse, lifetime number of partners). It is possible that misclassification of sexual activity information may have played a role by dampening the magnitude of the associations. Although our interviewers were trained to avoid giving any impression that the questionnaire probed for items that implied social value judgment, it is conceivable that some women may have underreported number of partners, which is plausible in light of the male-centered structure of the DRC society. It is reasonable to assume that the effects of age at first intercourse and number of lifetime sexual partners are attenuated by the passage of time. In our population of women aged >30, more immediate secular factors may have prevailed on risk, such as profile and behavior of the current partner, leading to a preponderance of the male factor in HPV transmission. Although we collected information on partners' sexual behavior, it was through the women's perception of the extent of their extramarital activities, which is prone to substantial error. We attempted to examine the validity of the associations by restricting analyses to women with negative cytology and comparing 2 different thresholds for defining HR-HPV positivity. In general, findings were consistent for the multivariate analyses of independent risk predictors, which suggest that the associations, although moderate at best, are credibly because of direct or indirect causal pathways.
Indirect markers for the role of sexual activity seemed to exert partial effects on risk. Women reporting one or more cowives were more likely to be infected than monogamous subjects. In the context of the local culture, the association with reported preference to consult a male doctor for a gynecological problem may reflect a level of emancipation that could be indicative of a more permissive sexual behavior.
Our findings underscore the need for enhanced knowledge about HPV determinants in this low-resource, high-risk setting to orient strategies to control the burden of cervical cancer. The observed prevalence of HR-HPV and risk associations in this unscreened population indicates a need for public health education to minimize the risks from sexual transmission and to formulate effective cervical cancer screening approaches.
1.Franco EL, Rohan TE, Villa LL. Epidemiologic evidence and human papillomavirus infection as a necessary cause of cervical cancer. J Natl Cancer Inst 1999; 91:506–511.
2.Bosch FX, Lorincz A, Munoz N, et al. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 2002; 55:244–265.
3.Parkin DM, Bray F, Ferlay J, et al. Global cancer statistics, 2002. CA Cancer J Clin 2005; 55:74–108.
4.de Sanjose S, Diaz M, Castellsague X, et al. Worldwide prevalence and genotype distribution of cervical human papillomavirus DNA in women with normal cytology: A meta-analysis. Lancet Infect Dis 2007; 7:453–459.
5.Denny L, Kuhn L, Pollack A, et al. Evaluation of alternative methods of cervical cancer screening for resource-poor settings. Cancer 2000; 89:826–833.
6.Si-Mohamed A, Ndjoyi-Mbiguino A, Cuschieri K, et al. High prevalence of high-risk oncogenic human papillomaviruses harboring atypical distribution in women of childbearing age living in Libreville, Gabon. J Med Virol 2005; 77:430–438.
7.Burchell AN, Winer RL, de Sanjosé S, et al. Chapter 6: Epidemiology and transmission dynamics of genital HPV infection. Vaccine 2006; 24:S52–S61.
8.Palefsky JM, Holly EA. Molecular virology and epidemiology of human papillomavirus and cervical cancer. Cancer Epidemiol Biomarkers Prev 1995; 4:415–428.
9.Sangwa-Lugoma G, Mahmud S, Nasr SH, et al. Visual inspection as a cervical cancer screening method in a primary health care setting in Africa. Int J Cancer 2006; 119:1389–1395.
10.Xi LF, Touré P, Critchlow CW, et al. Prevalence of specific types of human papillomavirus and cervical squamous intraepithelial lesions in consecutive, previously unscreened, west-African women over 35 years of age. Int J Cancer 2003; 103:803–809.
11.Smith JS, Melendy A, Rana RK, et al. Age-specific prevalence of infection with human papillomavirus in females: A global review. J Adolesc Health 2008; 43:S5.e1–S5.e62.
12.Wright TC Jr, Denny L, Kuhn L, et al. HPV DNA testing of self-collected vaginal samples compared with cytologic screening to detect cervical cancer. JAMA 2000; 283:81–86.
13.Thomas JO, Herrero R, Omigbodun AA, et al. Prevalence of papillomavirus infection in women in Ibadan, Nigeria: A population-based study. Br J Cancer 2004; 90:638–645.
14.Wall SR, Scherf CF, Morison L, et al. Cervical human papillomavirus infection and squamous intraepithelial lesions in rural Gambia, West Africa: Viral sequence analysis and epidemiology. Br J Cancer 2005; 93:1068–1076.
15.Castellsagué X, Menéndez C, Loscertales M, et al. Human papillomavirus genotypes in rural Mozambique. Lancet 2001; 358:1429–1430.
16.De Vuyst H, Steyaert S, Van Renterghem L, et al. Distribution of human papillomavirus in a family planning population in Nairobi, Kenya. Sex Transm Dis 2003; 30:137–142.
18.Mayrand MH, Duarte-Franco E, Coutlee F, et al. Randomized controlled trial of human papillomavirus testing versus pap cytology in the primary screening for cervical cancer precursors: Design, methods and preliminary accrual results of the Canadian cervical cancer screening trial (CCCaST). Int J Cancer 2006; 119:615–623.
19.Serwadda D, Wawer MJ, Shah KV, et al. Use of a hybrid capture assay of self-collected vaginal swabs in rural Uganda for detection of human papillomavirus. J Infect Dis 1999; 180:1316–1319.
20.Franceschi S, Herrero R, Clifford GM, et al. Variations in the age-specific curves of human papillomavirus prevalence in women worldwide. Int J Cancer 2006; 119:2677–2684.
21.Clifford GM. HPV in sub-Saharan Africa [editorial]. Papillomavirus Rep 2005; 16:322–326.
22.Kjaer SK, Svare EI, Worm AM, et al. Human papillomavirus infection in Danish female sex workers: Decreasing prevalence with age despite continuously high sexual activity. Sex Transm Dis 2000; 27:438–445.
23.Jewkes RK, Levin JB, Penn-Kekana LA. Gender inequalities, intimate partner violence and HIV preventive practices: Findings of a South African cross-sectional study. Soc Sci Med 2003; 56:125–134.
24.Herrero R, Hildesheim A, Bratti C, et al. Population-based study of human papillomavirus infection and cervical neoplasia in rural Costa Rica. J Natl Cancer Inst 2000; 92:464–474.
25.Giuliano AR, Papenfuss MR, Denman CA, et al. Human papillomavirus prevalence at the USA-Mexico border among women 40 years of age and older. Int J STD AIDS 2005; 16:247–251.
26.de Sanjose S, Bosch FX, Munoz N, et al. Socioeconomic differences in cervical cancer: Two case-control studies in Colombia and Spain. Am J Public Health 1996; 86:1532–1538.
27.Castellsague X, Munoz N. Chapter 3: Cofactors in human papillomavirus carcinogenesis—role of parity, oral contraceptives, and tobacco smoking. J Natl Cancer Inst Monogr 2003:20–28.
28.Vaccarella S, Herrero R, Snijders PJ, et al. Smoking and human papillomavirus infection: Pooled analysis of the international agency for research on cancer HPV prevalence surveys. Int J Epidemiol 2008; 37:536–546.
29.Franco EL, Spence AR. Commentary: Smoking and human papillomavirus infection: The pursuit of credibility for an epidemiologic association. Int J Epidemiol 2008; 37:547–548.
30.Bosch FX, de Sanjose S. The epidemiology of human papillomavirus infection and cervical cancer. Dis Markers 2007; 23:213–227.
31.Kjaer SK, van den Brule AJ, Bock JE, et al. Determinants for genital human papillomavirus (HPV) infection in 1000 randomly chosen young Danish women with normal pap smear: Are there different risk profiles for oncogenic and nononcogenic HPV types? Cancer Epidemiol Biomarkers Prev 1997; 6:799–805.
32.Richardson H, Franco E, Pintos J, et al. Determinants of low-risk and high-risk cervical human papillomavirus infections in Montreal university students. Sex Transm Dis 2000; 27:79–86.
33.Kayembe PK, Fatuma AB, Mapatano MA, et al. Prevalence and determinants of the use of modern contraceptive methods in Kinshasa, Democratic Republic of Congo. Contraception 2006; 74:400–406.
34.Green J, Berrington de Gonzalez A, Smith JS, et al. Human papillomavirus infection and use of oral contraceptives. Br J Cancer 2003; 88:1713–1720.
35.de Villiers EM. Relationship between steroid hormone contraceptives and HPV, cervical intraepithelial neoplasia and cervical carcinoma. Int J Cancer 2003; 103:705–708.