THERE IS STRONG EPIDEMIOLOGICAL evidence indicating that human papillomavirus (HPV) plays a central role in the etiology of cervical cancer. Case-control studies using polymerase chain reaction (PCR)-based hybridization techniques have reported that women positive for HPV DNA have a risk of developing cervical cancer 15 to 50 times higher than those without HPV DNA.1–4 Similar strong associations have been reported in casecontrol and cohort studies on preneoplastic cervical lesions.5–9 The evidence linking HPV to cervical neoplasia has been recently reviewed by an interdisciplinary group who concluded that certain types of HPV are human carcinogens.10
Because most cases of cervical neoplasia arise in women infected with HPV,7 identifying risk factors for latent cervical HPV infection (defined as HPV infections causing no morphologic cervical abnormalities and identified only by HPV DNA detection11) might be a first step toward blocking the sequence of cervical cancer. With the advent of techniques for the detection of viral genomes, the natural history of genital HPV infection has just begun to be clarified. It has been shown that latent infection is at least three times as common as clinical infection.11 Different from other viral infections, diagnosis of latent genital HPV infection relies on detection of HPV DNA in cervical cells. Consequently, studies on HPV and cervical neoplasia have suffered from the use of HPV DNA detection assays with varying degrees of sensitivity and specificity. Among the several hybridization assays available, the polymerase chain reaction (PCR)-based methods have been considered to be the most accurate.12,13
Earlier reports on risk factors for latent HPV infection had several limitations, such as the use of insensitive hybridization methods and study populations limited to young women.14–28 Studies in older populations may be more valuable to identify risk factors for persistent latent infection since it has been suggested that most HPV infection could be transient in young women11 and persistent HPV infections might be more common in older women.29 In the multicenter case-control studies of cervical cancer coordinated by the International Agency for Research on Cancer (IARC), PCR-based techniques were applied to cervical smears of both cases and controls for the detection of HPV DNA. This study provided an opportunity to investigate the determinants of HPV DNA in middle-aged women in relation to demographic, sexual, and reproductive factors.
Materials and Methods
The study population consisted of 1,184 controls from five case-control studies in Spain, Colombia, and Brazil. Details in methods of recruiting cases and controls and data collection were described elsewhere.1,2,4 In brief, the field work was conducted in the nine provinces of Spain from June 1985 to December 1987 and in Cali, Colombia, from June 1985 to December 1988. Controls for invasive cases were randomly selected from the general population that generated the cases, stratified for age. Controls for CIN III were recruited from women who received cervical cytologic examination at the same clinics or public health centers as the corresponding cases. In Brazil, cases and hospital controls were recruited from five general hospitals in Sao Paulo City between 1990–1991. Women admitted for gynecologic conditions or diseases associated with known risk factors for cervical cancer were excluded from the control group.
Of 1,184 women, cervical smears from 895 women were tested for HPV DNA sequences by PCR. The laboratory methods were described in detail elsewhere.12,13 In brief, in Spain and Colombia, PCR amplification was applied to 70% of the specimens from women for whom sufficient material was available, using HPV L1 consensus primers. The amplification products were hybridized sequentially with probes to HPV 6, 11, 16, 18, 31, 33, and 35 and β-globin under high-stringency conditions. Subsequently, the filters were screened with a generic probe that contained a mixture of amplimers of HPV 16 and 18.12 Polymerase chain reaction-positive was defined as those samples positive with type-specific probes and/or generic probes. All specimens from Brazil were first screened to determine the overall presence of HPV using general primers (GP) 5/6 in the PCR, which permits the detection of the sequenced genital HPV types 6, 11, 16, 18, 31, and 33, but also detects still unsequenced genital HPV types at the subpicogram level. After low-stringency Southern blot analysis with probes of HPV-specific PCR products, the GP-PCR-positive scrapes were subjected to type-specific PCR to establish the specific type of HPV present. Mixtures of HPV 6-, 16-, 33-, and HPV 11-, 18-, 31-specific primer sets were used to detect the sequenced HPV genotypes.13 Eighty-five women whose samples were negative for both β-globin and HPV DNA were excluded from analyses. As a result, 810 women remained in the study. In addition, the presence of HPV DNA in cytologic specimens from the urethra and the penis was determined by PCR for 290 current husbands of control women.
Information was collected by interviews on demographic variables and smoking, sexual, and reproductive histories from study subjects. Serum antibodies against Chlamydia trachomatis and herpes simplex virus type 2 (HSV-2) were measured only in Spain and Colombia. Antibodies to HSV were determined by enzyme-linked immunoassay (ELISA) using type-specific antigens: glycoprotein C (gC) for HSV-1 and glycoprotein G (gG) for HSV-2. These two glycoproteins were purified and used in the ELISA as described elsewhere.30 Antibodies to C. trachomatis were measured by indirect immunofluorescence using a modification of the method originally described by Wang and Grayston.31 It was simplified (from 15 to 1 serotype) using an antigen derived from the L2 strain of C. trachomatis. The OR and 95% confidence intervals (95% CI) for HPV positivity by several demographic, sexual, and reproductive factors were estimated by unconditional logistic regression models,32 while adjusting for age, area and type of controls. Tests for linear trend were performed by the same models. Factors which were significantly associated with latent HPV infection in the univariate analyses were further analyzed in a final multivariate model.
The mean age of the study populations were: 41.7 years (Spain);, 42.8 years (Colombia); and 52.7 years (Brazil). Of 810 women, 85 were positive for HPV DNA by PCR (10.5%). The positive rate was highest in Brazil (17%), followed by Colombia (13%), and Spain (5%). Among these middle-aged women, there were no trends with age in Spain and Colombia, but the positivity rate was inversely associated with age in Brazil (Table 1).
The risk of detecting HPV DNA decreased with increasing levels of education and family income in all countries. The inverse trends was statistically significant for family income. No relation was found with cigarette smoking (Table 2).
Table 3 shows the relation of the HPV-positive rate to reproductive factors. No trends were observed for ages at menarche and at first birth and parity. The ORs for parity and age at first birth tended to fluctuate with exposure levels in all countries. The use of any kind of contraceptives reduced the risk of HPV infection. The ORs for contraceptives other than oral contraceptives and barrier methods (condom and diaphragm) and for oral contraceptives were statistically significant. However, the inverse trend with duration of oral contraceptives use was not statistically significant.
The relation between the HPV-positive rates and sexual histories is presented in Table 4. Sexually active women (defined as <1 year since last sexual intercourse) showed a statistically significantly decreased risk of HPV infection (OR = 0.52, 95% CI: 0.30–0.91). The number of lifetime sexual partners was significantly positively associated with the HPV-positive rates. Women who had six or more lifetime sexual partners had an about four times increased risk of being HPV DNA-positive. This association was observed in each country, but was most pronounced in Spain, followed by Colombia, and Brazil, where the proportion of nonmonogamous women was higher. There was also a progressive increase in risk of HPV infection with decreasing age at first sexual intercourse. The trend was similar in the three countries, but strongest in Brazil. The number of lifetime sexual partners and age at first sexual intercourse were correlated with each other. The simultaneous adjustment of both variables made the associations weaker, but the association with the number of lifetime sexual partners remained marginally statistically significant. The average frequency of sexual intercourse per week, a history of anal sex, and the presence of HPV DNA in the husbands were not related to the HPV DNA positivity rates in the women. The HPV positivity rate was twice as high in women with antibodies to C. trachomatis and to HSV-2 than those without, but the association was statistically significant only for C. trachomatis.
When all factors significantly related to latent HPV infection were included in a multivariate model (Table 5), the positive associations with lifetime number of sexual partners, low socioeconomic status and antibodies to C. trachomatis persisted.
The OR for women having six or more sexual partners was 3.82 (95% CI: 1.26–11.53) compared to those reporting only one. However, the test for trend became of borderline significance.
The current study clearly showed a geographic variation in the prevalence of HPV, higher in Brazil and Colombia and lower in Spain. This pattern is consistent with that of the incidence of cervical cancer.33 Although study designs were different between Spain/Colombia and Brazil, it is unlikely that this would have affected the results. In Spain and Colombia where comparable study designs were used, the HPV-positive rates did not differ by type of controls (population and screening controls). In Brazil where hospital controls were used, too few patients had diseases that might have some degree of immunosuppression to affect the HPV DNA prevalence rate. The PCR-based assay used for the HPV DNA detection in Brazil was more sensitive than the early version of the PCR technique used in the Spain/Colombia studies.12,13 This may explain, at least in part, the higher prevalence of HPV in Brazil, and the lack of a decreasing trend with age in Spain and Colombia. Even low levels of misclassification of the HPV DNA status in the study populations from Spain and Colombia could lead to underestimation of HPV DNA prevalence and lack of association with age.34 Among these three countries, sexual behavior is also very different, especially between Spain and the two Latin American countries. Women tend to experience first sexual intercourse and first birth much earlier and to have more sexual partners and children in Colombia and Brazil than in Spain.4,35,36 Earlier commencement of sexual activity and more lifetime sexual partners in Brazil and Colombia probably contribute to the higher rates of HPV infection in these countries.4,35–37
Several studies have been reported on the prevalence of HPV DNA in women with normal cervical cytology.14–28 They vary greatly in hybridization assays used and size and nature of study populations, resulting in the broad range of HPV positive rates. Earlier studies based on insensitive hybridization methods, such as Southern blot hybridization,14–16 filter in situ hybridization (FISH),17–19 dot blot hybridization,20,21 and reverse DNA hybridization,22 generally yielded lower prevalence rates (0–32%). Prevalence rates based on PCR-based assays were around 35% to 45% in four populations,23,25–27 and ranged from 4.1% to 17.7% in other five populations.24,28
Age is another major determinant of the prevalence of latent HPV infection. In general, the younger the population, the higher the prevalence of DNA.24,25,27,28 This may explain the lower prevalence of HPV infection in the present study as compared with others. An inverse association with age was observed in Brazil, but not in Colombia and Spain. The number of study subjects in younger age groups may not have been enough to detect such an association, because the highest rates have been reported in the late teens and early 20s in an ongoing study in Colombia [unpublished data]. In addition, as discussed previously, misclassification of the HPV DNA status resulting from the use of a less sensitive PCR-based assay in Spain and Colombia may have affected the trend of the prevalence rates of HPV DNA in these countries. It has been suggested that most HPV infection in young women is transient, caused by sexual contacts with new partners, and that a persistent infection is established in a small proportion of women.11 If that is the case, older women should be a more appropriate study population to identify factors associated with persistent HPV infection, which may carry a higher risk for cervical neoplasia.38
Several studies have found associations between sexual behavioral factors and HPV infection,14,18–21,25–28 giving support to the sexual transmission of this virus. A positive association with number of sexual partners has most consistently been observed.14,20,21,25–28 An inverse association with age at first sexual intercourse has also reported in several studies.21,25,28 In the present study, both factors were associated with the risk of HPV infection, but the association was stronger for lifetime number of sexual partners. Particularly in low prevalence countries like Spain, multiple partners may be necessary to be exposed to a man who is positive for HPV.
An unexpected result was the higher prevalence of HPV DNA in sexually inactive women than in active women. This may indicate that recent sexual behavior in older women is not important to maintain the infection. These observations are also consistent with our finding that husbands' HPV DNA positivity is not related to their wives' HPV DNA positivity, although the low participation rates of husbands should be noted. In contrast, the situation appears to be different in young women. In a study in Washington, DC, it was reported that recent sexual behavior is a better predictor for HPV infection in young women than lifetime number of sexual partners.27
The presence of antibodies against C. trachomatis was associated with HPV DNA detection in Spain and Colombia. This does not necessarily indicate that the infection with C. trachomatis predisposes cervical tissue to subsequent HPV infection as the temporal relation between the two infections is unknown.
It has been hypothesized that estrogen or other hormones are capable either of reactivating latent HPV infection or of increasing viral gene expression.39,40 In support of this hypothesis, several studies found that number of pregnancies,27 current pregnancy,15,19,27 and use of oral contraceptives16,25,27 increased the detection rate of HPV DNA. The present study did not show any association with parity. Although oral contraceptive use tended to lower the risk of HPV infection, the relation is not likely to be causal. No clear dose-response was observed with duration of oral contraceptive use and the use of any type of contraceptives showed a similar degree of a decreased risk. Contraceptive users may generally differ from nonusers in other aspects than sexual behavior.
There was no association with cigarette smoking in the current study, in agreement with most of the previous studies.18,19,25–28
As observed for cervical cancer, an inverse association with socioeconomic status was found in the present study. A similar association with annual income was also reported in a study in Portland, Oregon.28 This inverse association may imply a summary of associations with both sexual (i.e., genital hygiene) and nonsexual factors (i.e., certain nutritional deficiencies), which were not assessed in our study.
In conclusion, the results of the present study support the sexual transmission route of HPV infection, and suggest that low socioeconomic status and antibodies to C. trachomatis are independent predictors of HPV DNA detection in middle-aged cytologically normal women.
1. Muñoz N, Bosch FX, de Sanjosé S, et al. The causal link between human papillomavirus and invasive cervical cancer: A population-based case-control study in Colombia and Spain. Int J Cancer 1992; 52:743–749.
2. Bosch FX, Muñoz N, de Sanjosé S, et al. Human papillomavirus and CIN III: A case-control study in Spain and Colombia. Cancer Epidemiol Biomarker Prev 1993; 2:423–431.
3. Peng H, Liu S, Mann V, Rohan T, Rawls W. Human papillomavirus types 16 and 33, herpes simplex virus type 2 and other risk factors for cervical cancer in Sichuan Province, China. Int J Cancer 1991; 47:711–716.
4. Eluf-Neto J, Booth M, Muñoz N, Bosch FX, Meijer CJLM, Walboomers JMM. Human papillomavirus and invasive cervical cancer in Brazil. Br J Cancer 1994; 69:114–119.
5. Morrison EAB, Ho GYF, Vermund SH, et al. Human papillomavirus infection and other risk factors for cervical neoplasia: A case-control study. Int J Cancer 1991; 49:6–13.
6. Koutsky LA, Holmes KK, Critchlow CW, et al. A cohort study of the risk of cervical intraepithelial neoplasia grade 2 or 3 in relation to papilloma infection. N Engl J Med 1992; 327:1272–1278.
7. Schiffman MH, Bauer HM, Hoover RN, et al. Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia. J Natl Cancer Inst 1993; 85:958–964.
8. Becker TM, Wheeler CM, McGough NS, et al. Sexually transmitted diseases and other risk factors for cervical dysplasia among southwestern Hispanic and non-Hispanic white women. JAMA 1994; 271:1181–1188.
9. Olsen AO, Gjøfen K, Sauer T, et al. Human papillomavirus and cervical intraepithelial neoplasia grade II-III: A population-based case-control study. Int J Cancer 1995; 61:312–315.
10. IARC. Human papillomavirus. Monographs on the Evaluation of the Carcinogenic Risks to Humans. vol 64, Lyon: International Agency for Research on Cancer, 1995.
11. Schneider A, Koutsky LA. Natural history and epidemiological features of genital HPV infection. In: Muñoz N, Bosch FX, Shah KV, Meheus A, eds. The Epidemiology of Cervical Cancer and Human Papillomavirus. Lyon: International Agency for Research on Cancer, 1992:25–52.
12. Guerrero E, Daniel RW, Bosch FX, et al. Comparison of ViraPap, Southern hybridization and polymerase chain reaction methods for human papillomavirus identification in an epidemiological investigation of cervical cancer. J Clin Microbiol 1992; 30:2951–2959.
13. Walboomers JMM, Melkert PWJ, Van den Brule AJC, Snijders PJF, Meijer CJLM. The polymerase chain reaction for human papillomavirus screening in diagnostic cytopathology of the cervix. In: Herrington CS, McGee JOD, eds. Diagnostic Molecular Pathology. A Practical Approach. Oxford: IRL Press, 1992:153–172.
14. Fisher M, Rosenfeld WD, Burk RD. Cervicovaginal human papillomavirus infection in suburban adolescents and young adults. J Pediatr 1991; 119:821–825.
15. Schneider A, Hotz M, Gissmann L. Increased prevalence of human papillomaviruses in the lower genital tract of pregnant women. Int J Cancer 1987; 40:198–201.
16. Lorincz AT, Schiffman MH, Jaffurs WJ, Marlow J, Quinn AP, Temple GF. Temporal associations of human papillomavirus infection with cervical cytologic abnormalities. Am J Obstet Gynecol 1990; 162:645–651.
17. de Villiers E-M, Wagner D, Schneider A, et al. Human papillomavirus DNA in women without and with cytological abnormalities: results of 5-year follow-up study. Gynecol Oncol 1992; 44:33–39.
18. Kjaer SK, Engholm G, Teisen C, et al. Risk factors for cervical human papillomavirus and herpes simplex virus infections in Greenland and Denmark: a population-based study. Am J Epidemiol 1990; 131:669–682.
19. Villa LL, Franco ELF, Ludwig Institute for Cancer Research Human Papillomavirus Study Group. Epidemiologic correlates of cervical neoplasia and risk of human papillomavirus infection in asymptomatic women in Brazil. J Natl Cancer Inst 1989; 81:332–340.
20. Moscicki A-B, Palefsky J, Gonzales J, Schoolnik GK. Human papillomavirus infection in sexually active adolescent females: Prevalence and risk factors. Pediatr Res 1990; 28:507–513.
21. Azocar J, Abad SMJ, Acosta H, et al. Prevalence of cervical dysplasia and HPV infection according to sexual behavior. Int J Cancer 1990; 45:622–625.
22. Bornstein J, Ben-David Y, Pascal B, et al. Occult human papillomavirus cervical infection: Reverse DNA hybridization detection and review of the literature. Israel J Obstet Gynecol 1993; 2:58–65.
23. Kjaer SK, de Villiers E-M, Çaglayan H, et al. Human papillomavirus and Herpes simplex virus and other potential risk factors for cervical cancer in a high-risk area (Greenland) and a low-risk area (Denmark): A second look. Br J Cancer 1993; 67:830–837.
24. Melkert PWJ, Hopman E, van den Brule AJC, et al. Prevalence of HPV in cytomorphologically normal cervical smears, as determined by the polymerase chain reaction, is age-dependent. Int J Cancer 1993; 53:919–923.
25. Ley C, Bauer HM, Reinngold A, et al. Determinants of genital human papillomavirus infection in young women. J Natl Cancer Inst 1991; 83:997–1003.
26. Wheeler CM, Parmenter CA, Hunt WC, et al. Determinants of genital human papillomavirus infection among cytologically normal women attending the University of New Mexico Student Health Center. Sex Transm Dis 1993; 20:286–289.
27. Hildesheim A, Gravitt P, Schiffman MH, et al. Determinants of genital human papillomavirus infection in low-income women in Washington, DC. Sex Transm Dis 1993; 20:279–285.
28. Bauer HM, Hildesheim A, Schiffman MH, et al. Determinants of genital human papillomavirus infection in low-risk women in Portland, Oregon. Sex Transm Dis 1993; 20:274–278.
29. Hildesheim A, Schiffman MH, Gravitt PE, et al. Persistence of type-specific human papillomavirus infection among cytologically normal women. J Infect Dis 1994; 169:235–240.
30. Ho DWT, Field PR, Sjogren-Jansson E, Jeansson S, Cunningham AL. Indirect ELISA for the detection of HSV-2 specific IgG and IgM antibodies with glycoprotein G (gG-2). J Virol Methods 1992; 36:249–264.
31. Wang SP, Grayston JT. Human serology in TRIC-LGV (Chlamydia trachomatis
) infection with micro immunofluorescence. J Infect Dis 1974; 130:388–397.
32. Breslow NE, Day NE. Statistical Methods in Cancer Research. In: The Analysis of Case-Control Studies. vol 1. Lyon: International Agency for Research on Cancer, 1980.
33. Parkin DM, Muir CS, Whelan SL, Gao Y-T, Ferlay J, Powell J. Cancer Incidence in Five Continents. vol 6. Lyon: International Agency for Research on Cancer, 1992.
34. Franco EL. Measurement errors in epidemiological studies of human papillomavirus and cervical cancer. In: Muñoz N, Bosch FX, Shah KV, Meheus A, eds. The epidemiology of cervical cancer and human papillomavirus. Lyon: International Agency for Research on Cancer, 1992:181–197.
35. Bosch FX, Muñoz N, de Sanjosé S, et al. Risk factors for cervical cancer in Colombia and Spain. Int J Cancer 1992; 52:750–758
36. Muñoz N, Bosch, FX, de Sanjosé S, et al. Risk factors for cervical intraepithelial neoplasia grade III/carcinoma in situ in Spain and Colombia. Cancer Epidemiol Biomarkers Prev 1993; 2:423–431.
37. Bosch FX, Muñoz N, de Sanjosé S, et al. Importance of the human papillomavirus endemicity in the incidence of cervical cancer: An extension of the hypothesis on sexual behaviour. Cancer Epidemiol Biomarkers Prev 1994; 3:375–379.
38. Ho GY, Burk RD, Klein S, et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. J Natl Cancer Inst 1995; 87:1365–1371.
39. Mitrani-Rosenbaum S, Tsvieli R, Tur-Kaspa R. Oestrogen stimulates differential transcription of human papillomavirus type 16 in SiHa cervical carcinoma cells. J Gen Virol 1989; 70:2227–2232.
40. Mittal R, Tsutsumi K, Pater A, Pater MM. Human papillomavirus type 16 expression in cervical keratinocytes: role of progesterone and glucocorticoid hormones. Obstet Gynecol 1993; 81:5–12.