SEARCH

SEARCH BY CITATION

Keywords:

  • female genital cancer;
  • risk factors;
  • HPV

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The etiology of vulvar and vaginal squamous cell carcinoma (VV-SCC) has received little attention. A total of 182 women with invasive VV-SCC (116 with VV-SCCvulva, 66 with VV-SCCvagina), 164 uterine corpus cancer controls and 518 population controls were interviewed in a population-based case–control study in Denmark, and 87 (48%) of the VV-SCC cases had tissue samples examined for human papillomavirus (HPV) DNA using the GP5+/6+ PCR-EIA assay and subsequent reverse line blotting for HPV typing. Logistic regression-derived odds ratios with 95% confidence intervals served as relative risks. Cervical cancer-associated high-risk HPVs (hrHPVs) were detectable in most (89%) examined cases of VV-SCCvagina and in half (50%) of cases of VV-SCCvulva (p < 0.001). In site-specific multivariate logistic regression analyses, statistically significant risk factors for both VV-SCCvulva and VV-SCCvagina included measures of hrHPV exposure (anogenital warts for VV-SCCvulva; cervical neoplasia and poor genital hygiene for VV-SCCvagina), tobacco smoking and alcohol consumption. Furthermore, socioeconomic variables (marital status and years at school) were associated with risk of VV-SCCvulva. Comparing hrHPV-positive and hrHPV-negative VV-SCCs in polytomous logistic regression analysis revealed that tobacco smoking and cervical neoplasia were significant risk factors only for hrHPV-positive VV-SCCs. Our study shows that VV-SCCvulva and VV-SCCvagina share measures of prior hrHPV exposure, tobacco smoking and alcohol consumption as statistically significant risk factors. HPV vaccination programs aimed at reducing the burden of cervical cancers are likely to also provide considerable protection against VV-SCCs. © 2008 Wiley-Liss, Inc.

Vulvar and vaginal squamous cell carcinomas (VV-SCC) are rare malignancies with a world-standardized annual incidence of around 1–3 pr. 100,000 women.1–3 The disease occurs mainly among women in their 60s or 70s,4, 5 the age-specific incidence increasing with age.1, 3, 4 The etiology of VV-SCC remains incompletely understood. Prior studies have reported infection with human papillomavirus (HPV) types associated with high risk of cervical cancer (hrHPVs) to be etiologically involved in VV-SCC.2, 6, 7 Accordingly, a number of traditional cervical cancer risk factors, including lifetime number of male sexual partners, genital warts, tobacco smoking and low socioeconomic status have been reported to be particularly relevant to VV-SCCs associated with hrHPV infection.4, 6–9 However, other, yet poorly defined etiologic factors that influence the risk of VV-SCC either without or in conjunction with hrHPV infection, must exist.10 We undertook a nationwide case–control study in Denmark to explore these issues further and to address whether VV-SCCs with vaginal involvement (VV-SCCvagina) differ etiologically from VV-SCCs restricted to the vulva (VV-SCCvulva).

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Participants

Three study groups were identified, including invasive VV-SCC patients diagnosed between 1993 and 1998 and 2 frequency-matched control groups consisting of female population controls and patients diagnosed with adenocarcinoma of the uterine corpus in the same time period as the VV-SCC patients. Information about newly diagnosed patients with invasive VV-SCC and uterine corpus cancer was obtained every 3rd month from the files of the Danish Cancer Registry and the Danish Pathology Registry, as well as through manual searches in the files of pathology departments throughout Denmark. Overall, we identified 796 women above the age of 18 years who had a primary invasive VV-SCC diagnosed between 1993 and 1998. Before inviting these women to the study, we contacted the hospital department or the private practitioner responsible for their treatment to ensure that each eligible participant had been fully informed about her cancer diagnosis, that she had no other health-related reason for nonparticipation known to the doctor (e.g., psychosis, hearing impairment), and that she was a Danish-speaking person. Of the original 796 VV-SCC patients, 435 patients (55%) had died, for 54 patients (7%) we received no reply from the hospital department or private practitioner responsible for the patient's treatment, and 4 patients (0.5%) were inaccessible due to emigration, residence in Greenland or unknown place of residence. Thus, of the total of 796 women with VV-SCC diagnosed between 1993 and 1998, we invited 303 (38%) patients who had been informed about their cancer diagnosis and were alive at the time we aimed to approach them between 1997 and 2002. Uterine corpus cancer controls were frequency-matched to VV-SCC patients on year of diagnosis and birth year (±5 years), and female population controls were identified through the Civil Registration System, a continuously updated nationwide demographic database11 and frequency-matched to VV-SCC patients on their year of birth. Eligible participants (303 VV-SCC patients, 238 uterine corpus cancer controls and 879 population controls) received a letter of invitation with information about the study and were asked to sign and return a consent form in a prepaid envelope to participate in the study.

All consenting participants underwent a structured telephone interview between July 1997 and July 2002. Interviews conducted by trained female medical students who were unaware of the study hypotheses covered a broad range of topics, including basic school attendance and post-school education, weight and height, tobacco and alcohol consumption, issues related to general and genital health, fertility, childbearing and contraception and details about sexual experiences and genital hygiene as well as venereal history. Patients with uterine corpus cancer were chosen as a supplementary control group because most examined risk factors for VV-SCC are not suspected to be risk factors for uterine corpus cancer. Consequently, information provided by these controls was anticipated to be comparable with the information obtained from population controls. We also anticipated the reliability of answers to be high among uterine corpus cancer controls, because these women are likely to be equally motivated as VV-SCC patients to provide honest answers with respect to sexual and venereal history, genital hygiene and other sensitive matters. Moreover, possible recall problems were expected to be rather similar for VV-SCC patients and uterine corpus cancer controls because both cancers originated in the female reproductive organs.

HPV status and anatomical localization

We identified paraffin-embedded archival tumour tissue in pathology departments throughout Denmark. These tumour tissues were examined for HPV DNA by the polymerase chain reaction technique (PCR) using general GP5+/6+ primers (GP5+/6+ PCR-EIA assay) as described in detail elsewhere.12 In brief, paraffin embedded tissues were cut for PCR analyses; outer sandwich sections were HE stained and analyzed for the presence of tumour cells, whereas inner sections were used for DNA extraction and subsequent PCR analyses. All DNAs were tested for B-globin PCR positivity after agarose gel analysis to check for the quality of DNA for PCR purposes.13 HPV was detected using the GP5+/6+ PCR-EIA using cocktail probes for 14 hrHPV types and 23 low-risk HPV (lrHPV) types as previously described.12 HPV-positive samples were subsequently analyzed for their HPV types using the reverse line blotting method.12 The following hrHPV types were detected and identified: HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66 and 68; lrHPVs detected were HPV 6, 11, 26, 34, 40, 42, 43, 44, 53, 54, 55, 57, 61, 70, 71 (CP8061), 72, 73, 81 (CP8304), 82/MM4, 82/IS39, 83 (MM7), 84 (MM8) and CP6108.

To evaluate possible differences in risk factors between anatomical subgroups, we categorized VV-SCC patients according to their tumour localization as determined by the combined available information in clinical pathology reports (available for all participating VV-SCC patients) and examination of HE-stained tumour specimens (available for those VV-SCC patients whose archival tumour tissues were identified for HPV analysis). Specifically, we grouped VV-SCC patients in those whose carcinoma was entirely restricted to the vulva and those whose carcinoma involved the vagina with or without accompanying vulvar involvement, referred to as VV-SCCvulva and VV-SCCvagina, respectively.

Statistical analysis

To ensure comparable exposure ascertainment in all study participants, we assigned a pseudo-year of diagnosis to each population control according to the distribution of year of diagnosis among VV-SCC patients. Using this pseudo-year of diagnosis, we considered only exposures that preceded the pseudo-year of diagnosis to be relevant in population controls, similar to the situation in VV-SCC cases and corpus cancer controls, for whom only exposures that preceded the year of cancer diagnosis counted as exposures of potential etiologic relevance.

Univariate analyses

Before calculating odds ratios (ORs) for VV-SCC patients vs. all controls combined, we calculated univariate exposure ORs in logistic regression analyses for all VV-SCC patients vs. each of the 2 control groups, adjusted only for age at diagnosis in 10-year age groups (<40, 40-49, …, ≥70 years). ORs obtained with either of the 2 control groups were similar for most examined explanatory variables, so we subsequently tested the appropriateness of combining the 2 control groups by doing a logistic regression analysis restricted to uterine corpus cancer controls and population controls with adjustment for age in 10-year age groups. This analysis revealed that only one of the examined explanatory variables (labial herpes) was associated with statistically significantly increased risk of uterine corpus cancer, so to reduce complexity and gain statistical power we combined the 2 control groups in all subsequent analyses. Having confirmed the appropriateness of combining the 2 control groups, we next performed logistic regression analysis to calculate subsite-specific univariate exposure ORs separately for VV-SCCvulva and VV-SCCvaginavs. the combined group of control subjects, again with adjustment only for age at diagnosis in 10-year age groups. The results of these age-adjusted, site-specific analyses are referred toas univariate analyses. Likelihood ratio tests for homogeneity determined possible differences in risk between exposure categories.

Multivariate analyses

Univariate analyses suggested considerable overlap in risk factors for VV-SCCvulva and VV-SCCvagina but associations were not identical for all explanatory variables. We therefore performed a set of site-specific multivariate analyses, in which all explanatory variables for which the p-value in the test for homogeneity was <0.10 in the site-specific univariate analysis were entered in a logistic regression model for that particular site-specific VV-SCC subtype along with age in 10-year age groups. Using backward elimination to remove the least significant variables one by one until all remaining variables in the models were statistically significant (p-value in test for homogeneity <0.05), we arrived at 2 site-specific multivariate models, one for VV-SCCvulva and one for VV-SCCvagina. Subsequently, excluded explanatory variables were given an extra chance to enter the final models by means of forward inclusion in the models one by one, but none of the previously excluded variables reached statistical significance.

To examine for possible risk factor differences between hrHPV-positive and hrHPV-negative VV-SCCs, we finally did a polytomous logistic regression analysis, in which we calculated ORs for hrHPV-positive VV-SCCs and hrHPV-negative VV-SCCs, while simultaneously providing Wald tests for homogeneity in risk factor associations between the hrHPV-positive and the hrHPV-negative subsets. Chi-squared test evaluated proportional differences in hrHPV positivity between VV-SCC subsets defined by their anatomical localization (VV-SCCvulvavs. VV-SCCvagina). Throughout, 2-sided p-values <0.05 and 95% confidence intervals (CIs) excluding unity were considered as indicators of statistical significance. All logistic regression analyses were carried out using SAS Version 9.1 (SAS Institute, Cary, NC).

Ethics

The study was approved of by The Danish National Committee on Biomedical Research Ethics (approval no. C-1995-20) and The Danish Data Protection Agency (approval no. 1995-1200-187). Delayed reporting was due to shortage of funding.

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Participation rate

Of the 303 VV-SCC patients, 238 corpus cancer controls and 871 population controls who were invited to participate in the study, 182 VV-SCC patients (60.1%), 164 corpus cancer controls (68.9%) and 518 population controls (59.5%) agreed to participate in the telephone interview. The most common reasons for nonparticipation were unspecified unwillingness (25.4% of invited VV-SCC patients, 21.8% of invited uterine corpus cancer controls, and 29.2% of invited population controls), illness or death (7.3, 4.2 and 4.8%, respectively), or lack of telephone (7.3, 5.0 and 6.5%, respectively).

Characteristics of the study population

The median age at cancer diagnosis was 63 years among VV-SCC patients and 62 years among corpus cancer controls, while the corresponding median age at the pseudo-year of diagnosis was 57 years among population controls (Table I). We identified paraffin-embedded archival tumour tissue for 87 (48%) of the 182 VV-SCC cases. These tissues were examined by PCR for the presence of HPV DNA. Of the 87 examined tumour tissues, 54 (62%) were hrHPV-positive and one (1%) was positive to a low-risk HPV type (HPV 6). Tumours from patients with VV-SCCvagina were more often hrHPV-positive (89% of 27) than tumours from patients with VV-SCCvulva (50% of 60) (X2 = 11.96, df = 1, p < 0.001). None of a series of 17 tumour specimens from controls with uterine corpus cancer subjected to the same PCR analysis as the VV-SCC cases was found to be positive for any of the examined HPV types (Table I).

Table I. Characteristics of 864 Participants in Case–Control Study of Risk Factors for Invasive Squamous Cell Carcinoma of the Vulva and Vagina, Denmark
 Cases (%)Controls (%)
VV-SCCvulva (n = 116)VV-SCCvagina (n = 66)CCC (n = 164)PC (n = 518)
  • VV-SCCvulva, squamous cell carcinoma restricted to the vulva; VV-SCCvagina, squamous cell carcinoma with vaginal involvement; CCC, corpus cancer controls; PC, population controls; hrHPV, high-risk human papillomavirus; lrHPV, low-risk human papillomavirus; NA, not applicable. Percentages may not sum to 100 because of rounding.

  • 1

    Age at diagnosis among population controls refers to the age in the pseudoyear of diagnosis attributed to each population control according to the distribution of year of diagnosis among VV-SCC cases.

  • 2

    None of 17 examined tumor tissues from controls with uterine corpus cancer was positive for any of the examined HPV types.

  • 3

    The 30 hrHPV-positive VV-SCCvulva patients were positive for the following specific HPV types; HPV 16 (n = 22), HPV 33 (n = 7), HPV 51 (n = 1), HPV 73 (n = 2). One hrHPV-positive VV-SCCvulva patient was positive for more than 1 hrHPV type; HPV 16/51/73 (n = 1). The 24 hrHPV-positive VV-SCCvagina patients were positive for the following specific HPV types; HPV 16 (n = 21), HPV 18 (n = 1), HPV 33 (n = 2), HPV 39 (n = 1) and HPV 45 (n = 1). Two hrHPV-positive VV-SCCvagina patients were positive to more than 1 hrHPV type; HPV 16/18 (n = 1) and HPV 16/33 (n = 1).

  • 4

    The lrHPV-positive VV-SCC patient was positive for HPV 6 (n = 1).

Age at diagnosis1    
 <40 years11 (9)6 (9)3 (2)66 (13)
 40–49 years9 (8)3 (5)22 (13)102 (20)
 50–59 years23 (20)9 (14)44 (27)113 (22)
 60–69 years28 (24)27 (41)52 (32)124 (24)
 ≥70 years45 (39)21 (32)43 (26)113 (22)
HPV statusVV-SCCvulva (n = 60)VV-SCCvagina (n = 27)CCC (n = 17)PC (n = 0)
 HPV-positive31 (52)24 (89)02 (0)NA
  hrHPV-positive330 (50)24 (89)0 (0)NA
  lrHPV-positive41 (2)0 (0)0 (0)NA
 HPV-negative29 (48)3 (11)17 (100)NA

Risk factors for VV-SCCvulvavs. VV-SCCvagina

Univariate analyses

A large number of variables were examined as possible risk factors for the 2 site-specific subsets of VV-SCC (Table II). For VV-SCCvulva, the following factors were found in univariate analysis to have a p-value less than 0.10: the lifetime number of male sexual partners (p = 0.05), history of anogenital warts (p < 0.001), history of preinvasive or invasive cervical cancer (p = 0.04), marital status (p = 0.05), the male partner's lifetime number of other sexual partners (p = 0.05), history of anogenital warts in the male partner (p = 0.02), smoker status (p < 0.001), cumulative alcohol consumption (p = 0.002), and years at school (p = 0.04). For VV-SCCvagina, risk factors with a p-value less than 0.10 in the univariate analysis were: the lifetime number of male sexual partners (p = 0.004), anal intercourse (p = 0.08), genital washing before and after intercourse (p = 0.05), history of anogenital warts (p = 0.03), history of preinvasive or invasive cervical cancer (p < 0.001), smoker status (p = 0.03) and cumulative alcohol consumption (p = 0.002) (Table II). These variables were selected for further evaluation in site-specific multivariate analyses.

Table II. Risk Factors for Invasive Squamous Cell Carcinoma of the Vulva and Vagina (Univariate Analyses)
 VV-SCCvulvaVV-SCCvaginaControlsUnivariate Odds Ratio (95% CI)1
VV-SCCvulva vs. controlsVV-SCCvagina vs. controls
  • VV-SCC, vulvovaginal squamous cell carcinoma. Numbers which do not add up are due to missing values. Percentages may not sum to 100 because of rounding. All p-values obtained in tests for homogeneity.

  • 1

    All univariate odds ratios are adjusted for age at the time of the diagnosis.

  • 2

    Numbers and percentages of respondents who reported that they had had the disease (or HIV test) in question. In the calculation of odds ratios for these variables, respondents reporting no such disease (or HIV test) served as the reference group.

  • 3

    Numbers and percentages of respondents who reported that their male partner had had the disease in question. In the calculation of univariate odds ratios for these variables, respondents reporting no such disease in their male partner served as the reference group.

  • 4

    Test for homogeneity only includes current smokers.

  • 5

    One pack-year equals 1 pack of cigarettes (20 cigarettes) per day for a year.

  • 6

    Test for homogeneity only includes ever smokers.

  • 7

    One consumption-year is equivalent to 2 alcoholic drinks per day for 1 year.

Sexual behaviour     
 Lifetime no. male sexual partners     
  0–126 (22)10 (15)210 (31)0.52 (0.31–0.86)0.33 (0.16–0.69)
  2–457 (49)34 (52)268 (40)1 (referent)1 (referent)
  5–919 (16)14 (21)109 (16)0.98 (0.55–1.77)1.28 (0.64–2.57)
  ≥1014 (12)8 (12)90 (13)1.03 (0.51–2.07)1.11 (0.46–2.69)
    p = 0.05p = 0.004
 No. male sexual partners before age 20 years     
  044 (39)24 (36)247 (37)1 (referent)1 (referent)
  133 (29)19 (29)226 (33)1.10 (0.66–1.83)1.17 (0.61–2.25)
  2 or 326 (23)14 (21)129 (19)1.68 (0.94–3.00)1.77 (0.85–3.70)
  >410 (9)9 (14)73 (11)1.35 (0.59–3.10)2.99 (1.15–7.77)
    p = 0.36p = 0.13
 Age at first sexual intercourse     
  ≤16 years23 (20)12 (18)126 (19)1.41 (0.78–2.57)1.07 (0.52–2.39)
  17 or 18 years34 (29)24 (36)238 (36)1 (referent)1 (referent)
  19 or 20 years31 (27)15 (23)143 (21)1.13 (0.65–1.97)0.82 (0.40–1.66)
  >20 years28 (24)15 (23)163 (24)0.81 (0.46–1.45)0.61 (0.30–1.24)
    p = 0.41p = 0.51
 Anal intercourse     
  Never103 (90)54 (82)566 (85)1 (referent)1 (referent)
  Ever12 (10)12 (18)101 (15)0.83 (0.43–1.63)1.95 (0.95–3.99)
    p = 0.59p = 0.08
 Genital washing before and after intercourse     
  Never (0 times out of 10)10 (9)9 (15)36 (6)1.63 (0.72–3.66)1.85 (0.77–4.49)
  Sometimes (1–5 times out of 10)34 (32)14 (23)197 (31)1.06 (0.62–1.79)0.58 (0.28–1.18)
  Often (6–9 times out of 10)32 (30)22 (36)210 (33)1 (referent)1 (referent)
  Always (10 times out of 10)31 (29)16 (26)196 (31)0.91 (0.53–1.57)0.60 (0.30–1.20)
    p = 0.58p = 0.05
Sexually transmitted diseases     
 Anogenital warts225 (22)9 (14)45 (7)4.98 (2.78–8.91)2.56 (1.14-5.76)
    p < 0.001p = 0.03
 Chlamydia, syphilis or gonorrhea28 (7)7 (11)61 (9)0.89 (0.41–1.95)1.50 (0.64–3.54)
    p = 0.77p = 0.37
 Genital herpes22 (2)1 (2)12 (2)0.99 (0.21–4.55)0.87 (0.11–6.91)
    p = 0.98p = 0.89
 Labial herpes257 (49)35 (53)316 (46)1.16 (0.78–1.73)1.34 (0.80–2.23)
    p = 0.48p = 0.27
 HIV test211 (9)4 (6)76 (11)1.18 (0.57–2.44)0.80 (0.26–2.46)
    p = 0.66p = 0.69
 Preinvasive or invasive cervical cancer220 (17)20 (30)76 (11)1.81 (1.04–3.13)3.84 (2.12–6.96)
    p = 0.04p < 0.001
Partner-related factors     
 Marital status     
  Unmarried without current male partner2 (2)3 (5)38 (6)0.25 (0.06–1.09)0.90 (0.26–3.19)
  Unmarried with current partner6 (5)3 (5)20 (3)2.40 (0.82–7.03)2.95 (0.68–12.8)
  Married62 (53)27 (41)375 (55)1 (referent)1 (referent)
  Widowed35 (30)22 (33)180 (26)0.65 (0.39–1.11)1.06 (0.55–2.05)
  Separated/divorced11 (9)11 (17)68 (10)0.92 (0.45–1.85)2.18 (1.02–4.68)
    p = 0.05p = 0.24
 Male partner's lifetime no. other sexual partners   
  013 (13)10 (16)114 (18)0.58 (0.29–1.16)0.70 (0.31–1.55)
  1 or 242 (41)16 (25)175 (28)1.32 (0.80–2.18)0.79 (0.39–1.57)
  ≥338 (37)23 (37)248 (40)1 (referent)1 (referent)
  No current male partner10 (10)14 (22)81 (13)0.66 (0.31–1.41)1.52 (0.73–3.16)
    p = 0.05p = 0.29
 Sexually transmitted diseases in male partner3  
  Anogenital warts in male partner5 (5)1 (2)10 (2)4.32 (1.37–13.6)1.18 (0.14–9.83)
    p = 0.02p = 0.88
 Chlamydia, syphilis or gonorrhea in male partner5 (4)0 (0)23 (3)1.47 (0.53–4.04)
    p = 0.47
Tobacco and alcohol consumption     
 Smoker status     
  Lifelong nonsmoker36 (31)27 (41)285 (42)1 (referent)1 (referent)
  Former smoker28 (24)12 (18)184 (27)1.25 (0.73–2.15)0.65 (0.32–1.33)
  Current smoker52 (45)27 (41)213 (31)2.69 (1.65–4.41)1.68 (0.93–3.03)
    p < 0.001p = 0.03
 Current tobacco consumption (g/d)     
  Current nonsmoker64 (55)39 (59)469 (69)1 (referent)1 (referent)
  <53 (3)3 (5)19 (3)1.38 (0.39–4.95)2.06 (0.56–7.54)
  5–911 (9)3 (5)24 (4)3.72 (1.70–8.15)1.51 (0.43–5.46)
  10–1412 (10)7 (11)45 (7)2.63 (1.29–5.40)2.15 (0.89–5.24)
  ≥1526 (22)14 (21)125 (18)2.24 (1.31–3.83)1.98 (1.00–3.90)
    p = 0.394p = 0.99
 Cumulative tobacco consumption (pack-years5)  
  Lifelong nonsmoker36 (31)27 (42)285 (42)1 (referent)1 (referent)
  <12.528 (24)12 (18)143 (21)1.82 (1.05–3.16)0.92 (0.45–1.89)
  12.5 to <2525 (22)14 (22)139 (21)1.89 (1.07–3.35)1.28 (0.64–2.57)
  ≥2527 (23)12 (18)107 (16)2.28 (1.30–4.01)1.22 (0.59–2.53)
    p = 0.656p = 0.67
 Cumulative alcohol consumption (consumption-years7)  
  016 (14)7 (11)167 (25)0.35 (0.19–0.63)0.25 (0.11–0.58)
  <1067 (59)40 (61)346 (52)1 (referent)1 (referent)
  10–2019 (17)12 (18)99 (15)1.03 (0.58–1.82)1.09 (0.54–2.20)
  >2012 (11)7 (11)57 (9)0.89 (0.45–1.78)0.86 (0.36–2.06)
    p = 0.002p = 0.002
Sociodemographic factors     
 Years at school     
  <1084 (72)42 (64)394 (58)1 (referent)1 (referent)
  ≥1032 (28)24 (36)286 (42)0.60 (0.37–0.98)1.11 (0.62–1.99)
    p = 0.04p = 0.72
 Post school education     
  None53 (46)26 (39)241 (36)1.12 (0.72–1.73)0.80 (0.45–1.41)
  Short (≤3 years)50 (43)31 (47)300 (44)1 (referent)1 (referent)
  Long (>3 years)12 (10)9 (14)135 (20)0.56 (0.29–1.10)0.67 (0.31–1.47)
    p = 0.11p = 0.54
 No. of children     
  0 (nulliparous)21 (18)8 (12)87 (13)1.41 (0.78–2.55)0.70 (0.31–1.60)
  117 (15)5 (8)109 (16)0.95 (0.51–1.77)0.35 (0.13–0.93)
  241 (35)32 (49)261 (38)1 (referent)1 (referent)
  320 (17)13 (20)140 (21)0.82 (0.46–1.47)0.65 (0.32–1.29)
  ≥417 (15)7 (11)85 (12)1.04 (0.55–1.96)0.46 (0.19–1.11)
    p = 0.65p = 0.13
Multivariate analyses

All variables with p-values < 0.10 in the univariate analyses were entered simultaneously in site-specific multivariate logistic regression models along with age in 10-year groups. After the procedure of backward elimination and subsequent forward inclusion, we identified the following statistically significant (p < 0.05) risk factors. For VV-SCCvulva: history of anogenital warts (p < 0.001), marital status (p = 0.03), smoker status (p < 0.001), cumulative alcohol consumption (p = 0.01) and years at school (p = 0.02). For VV-SCCvagina: genital washing before and after intercourse (p = 0.02), history of preinvasive or invasive cervical cancer (p < 0.001), smoker status (p = 0.02) and cumulative alcohol consumption (p = 0.003).

Sexual behaviour

As seen in Table III, women with a lifetime number of only 0–1 male sexual partners were at reduced risk of VV-SCCvulva (OR = 0.59, 95% CI 0.34–1.02, compared to 2–4 male sexual partners) and VV-SCCvagina (OR = 0.37, 95% CI 0.16–0.84, compared to 2–4 male sexual partners), although the overall test for homogeneity was not formally significant at either site. Anal intercourse was not a significant risk factor for either subtype of VV-SCC, but the association almost reached statistical significance for VV-SCCvagina (OR = 2.18, 95% CI 0.97–4.90). Women who never washed their genitals in connection with sexual intercourse were at almost 3-fold increased risk of VV-SCCvagina compared to women who washed their genitals before and after intercourse 6–9 times out of 10 (OR = 2.76, 95% CI 1.07–7.08).

Table III. Risk Factors for Invasive Squamous Cell Carcinoma of the Vulva and Vagina (Multivariate Analyses)
 Multivariate Odds Ratio (95% CI)1
VV-SCCvulva vs. controlsVV-SCCvagina vs. controls
  • VV-SCC, vulvovaginal squamous cell carcinoma. All p-values obtained in tests for homogeneity.

  • 1

    Multivariate odds ratios for VV-SCCvulva are adjusted for age at the time of the diagnosis, anogenital warts, marital status, smoker status, cumulative alcohol consumption and years at school. Multivariate odds ratios for VV-SCCvagina are adjusted for age at the time of the diagnosis, genital washing before and after intercourse, preinvasive or invasive cervical cancer, smoker status and cumulative alcohol consumption.

  • 2

    Respondents reporting no such disease served as the reference group.

  • 3

    Respondents reporting no such disease in their male partner served as the reference group.

  • 4

    One consumption-year is equivalent to 2 alcoholic drinks per day for 1 year.

Sexual behaviour
 Lifetime no. male sexual partners
  0–10.59 (0.34–1.02)0.37 (0.16–0.84)
  2–41 (referent)1 (referent)
  5–90.79 (0.42–1.50)1.04 (0.47–2.26)
  ≥100.71 (0.31–1.65)1.02 (0.38–2.70)
 p = 0.29p = 0.07
 Anal intercourse
  Never1 (referent)1 (referent)
  Ever0.67 (0.31–1.44)2.18 (0.97–4.90)
 p = 0.29p = 0.07
 Genital washing before and after intercourse
  Never (0 times out of 10)1.63 (0.67–3.97)2.76 (1.07–7.08)
  Sometimes (1–5 times out of 10)1.15 (0.65–2.04)0.60 (0.29–1.27)
  Often (6–9 times out of 10)1 (referent)1 (referent)
  Always (10 times out of 10)0.82 (0.45–1.48)0.60 (0.29–1.26)
 p = 0.44p =0.02
Sexually transmitted diseases
 Anogenital warts25.77 (3.08–10.8)2.52 (1.00–6.31)
 p < 0.001p = 0.06
 Preinvasive or invasive cervical cancer21.76 (0.97–3.20)5.37 (2.79–10.3)
 p = 0.07p < 0.001
Partner-related factors
 Marital status
  Unmarried without current male partner0.20 (0.04–0.92)0.82 (0.16–4.19)
  Unmarried with current partner2.40 (0.72–8.02)2.27 (0.48–10.7)
  Married1 (referent)1 (referent)
  Widowed0.81 (0.38–1.73)1.12 (0.54–2.33)
  Separated/divorced0.64 (0.37–1.12)2.06 (0.88–4.85)
 p = 0.03p = 0.47
 Male partner's lifetime no. other sexual partners
  00.66 (0.31–1.41)0.84 (0.34–2.08)
  1 or 21.56 (0.90–2.71)1.04 (0.47–2.27)
  ≥31 (referent)1 (referent)
  No current male partner1.85 (0.42–8.06)2.19 (0.93–5.12)
 p = 0.08p = 0.24
 Sexually transmitted diseases in male partner
  Anogenital warts in male partner32.04 (0.56,7.48)1.26 (0.14–11.4)
 p = 0.32p = 0.85
Tobacco and alcohol consumption
 Smoker status
  Lifelong nonsmoker1 (referent)1 (referent)
  Former smoker1.21 (0.68–2.16)0.40 (0.18–0.89)
  Current smoker2.61 (1.53–4.46)1.17 (0.61–2.23)
 p < 0.001p = 0.02
 Cumulative alcohol consumption (consumption-years4)
  00.37 (0.20–0.70)0.21 (0.08–0.55)
  <101 (referent)1 (referent)
  10–201.02 (0.55–1.88)1.06 (0.48–2.30)
  >200.89 (0.43–1.86)0.84 (0.33–2.14)
 p = 0.01p = 0.003
Sociodemographic factors
 Years at school
  <101 (referent)1 (referent)
  ≥100.53 (0.31–0.90)0.98 (0.50–1.90)
 p = 0.02p = 0.94
Sexually transmitted diseases

Two proxy measures of hrHPV exposure emerged as independent risk factors: A history of anogenital warts was significantly associated with VV-SCCvulva (OR = 5.77, 95% CI 3.08–10.8), but less strongly associated with VV-SCCvagina (OR = 2.52, 95% CI 1.00–6.31), whereas a history of preinvasive or invasive cervical cancer was significantly linked with VV-SCCvagina (OR = 5.37, 95% CI 2.79–10.3), but less strongly associated with VV-SCCvulva (OR = 1.79, 95% CI 0.97–3.20).

Partner-related factors

With the exception of marital status, which exhibited statistically significant heterogeneity in relation to VV-SCCvulva, there remained no significant association between partner-related factors (male partner's lifetime number of other sexual partners and male partner's history of anogenital warts) and the risk of VV-SCCvulva or VV-SCCvagina, after adjustment for age and other confounders.

Tobacco and alcohol consumption

Smoker status and cumulative alcohol consumption remained statistically significant risk factors in multivariate analyses for both site-specific subtypes of VV-SCC. Specifically, being a current smoker was associated with more than double risk of VV-SCCvulva (OR = 2.61, 95% CI 1.53–4.46) compared to being a lifelong nonsmoker and alcohol abstinence remained significantly associated with reduced risk of VV-SCCvulva (OR = 0.37, 95% CI 0.20–0.70, for 0 vs. <10 consumption-years) and VV-SCCvagina (OR = 0.21, 95% CI 0.08–0.55, for 0 vs. <10 consumption-years).

Sociodemographic factors

Having completed ≥10 years at school was significantly associated with reduced risk of VV-SCCvulva (OR = 0.53, 95% CI 0.31–0.90) compared to having less than 10 years of school education.

Risk factors for hrHPV-positive VV-SCC and hrHPV-negative VV-SCC

All exposure variables of interest as potential risk factors (p < 0.10 in Table II) were included in a polytomous logistic regression analysis (Table IV) aimed to explore possible risk factor differences between hrHPV-positive and hrHPV-negative VV-SCCs, regardless of their specific anatomical localization in the vulva or the vagina. All variables that were statistically significant in at least 1 of the 2 final site-specific multivariate models were included as potential confounders. Most variables were associated with hrHPV-positive VV-SCC in ways that were not statistically significantly different from those applying for hrHPV-negative VV-SCC, as judged from the tests for homogeneity in the right column of Table IV. The only exceptions were that marital status was a significant variable only for hrHPV-negative VV-SCC, and smoker status, which was only statistically significantly associated with the risk of hrHPV-positive VV-SCC. History of preinvasive or invasive cervical cancer was also only significantly associated with risk of hrHPV-positive VV-SCC, but the test for homogeneity between hrHPV-positive and hrHPV-negative VV-SCC for this variable was not formally significant (p = 0.08).

Table IV. Risk Factors for High-Risk Human Papillomavirus Positive and Negative Invasive Squamous Cell Carcinoma of the Vulva and Vagina
 hrHPV-positive VV-SCChrHPV-negative VV-SCCControlsMultivariate Odds Ratio (95% CI)1hrHPV-positive VV-SCC vs. hrHPV-negative VV-SCC2
hrHPV-positive VV-SCC vs. controlshrHPV-negative VV-SCC vs. controls
  • VV-SCC, vulvovaginal squamous cell carcinoma; hrHPV, high-risk human papillomavirus. Controls include corpus cancer controls and population controls. Numbers which do not add up are due to missing values. Percentages may not sum to 100 because of rounding. All p-values obtained in tests for homogeneity.

  • 1

    Multivariate odds ratios are adjusted for age at the time of diagnosis, genital washing before and after intercourse, preinvasive or invasive cervical cancer, anogenital warts, marital status, smoker status, cumulative alcohol consumption and years at school.

  • 2

    Test for homogeneity of associations between hrHPV-positive and hrHPV-negative VV-SCC.

  • 3

    Because of no hrHPV-negative vulvovaginal cases with male partner having had anogenital warts the analysis was performed as an ordinary logistic regression analysis.

  • 4

    One consumption year is equivalent to 2 alcoholic drinks per day for 1 year.

Sexual behaviour
 Lifetime no. of male sexual partners
  0–14 (7)5 (15)210 (31)0.23 (0.07–0.76)0.32 (0.11–0.95)p = 0.46
  2–428 (52)21 (64)268 (40)1 (referent)1 (referent)
  5–98 (15)5 (15)109 (16)0.57 (0.22–1.49)1.16 (0.37–3.66)
  ≥1014 (26)2 (6)90 (13)1.56 (0.56–4.39)0.63 (0.12–3.34)
 p = 0.03p = 0.18
 Anal intercourse
  Never46 (85)31 (97)566 (85)1 (referent)1 (referent)p = 0.46
  Ever8 (15)1 (3)101 (15)0.72 (0.26–2.02)0.30 (0.04–2.47)
 p = 0.53p = 0.26
 Genital washing before and after intercourse
  Never (0 times out of 10)6 (12)3 (10)36 (6)3.38 (1.05–10.87)1.47 (0.35–6.20)p = 0.70
  Sometimes (1–5 times out of 10)15 (29)7 (23)197 (31)1.46 (0.61–3.53)0.73 (0.25–2.15)
  Often (6–9 times out of 10)13 (25)9 (30)210 (33)1 (referent)1 (referent)
  Always (10 times out of 10)17 (33)11 (37)196 (31)1.47 (0.62–3.49)0.97 (0.36–2.61)
 p = 0.24p = 0.83
Sexually transmitted diseases
 Anogenital warts
  Yes18 (33)6 (18)45 (7)8.29 (3.57–19.22)6.05 (2.06–17.74)p = 0.62
  No36 (67)27 (82)636 (93)1 (referent)1 (referent)
 p < 0.001p = 0.001
 Preinvasive or invasive cervical cancer
  Yes19 (35)5 (15)76 (11)4.91 (2.34–10.29)1.65 (0.57–4.74)p = 0.08
  No35 (65)28 (85)605 (89)1 (referent)1 (referent)
 p < 0.001p = 0.36
Partner-related factors
 Marital status
  Unmarried6 (11)3 (9)58 (9)1.00 (0.28–3.50)1.37 (0.33–5.81)p = 0.02
  Married26 (48)20 (61)375 (55)1 (referent)1 (referent)
  Widowed, seperated or divorced22 (41)10 (30)248 (36)1.29 (0.61–2.75)0.26 (0.10–0.70)
 p = 0.79p = 0.02
 Male partner's lifetime no. other sexual partners
  03 (6)3 (10)114 (18)0.41 (0.11–1.62)0.53 (0.13–2.28)p = 0.59
  1 or 217 (33)14 (47)175 (28)1.41 (0.62–3.24)1.72 (0.64–4.61)
  ≥323 (45)9 (30)248 (40)1 (referent)1 (referent)
  No current male partner8 (16)4 (13)81 (13)1.12 (0.35–3.55)0.24 (0.02–2.42)
 p = 0.35p = 0.13
 Sexually transmitted diseases in male partner
  Anogenital warts in male partner3
   Yes3 (7)0 (0)10 (2)2.16 (0.37–12.5) 
   No41 (93)30 (100)589 (98)1 (referent)
 p = 0.41 
Tobacco and alcohol consumption
 Smoker status
  Lifelong nonsmoker17 (31)14 (42)285 (42)1 (referent)1 (referent)p = 0.006
  Former smoker5 (9)12 (36)184 (27)0.36 (0.11–1.18)1.50 (0.61–3.70)
  Current smoker32 (59)7 (21)213 (31)2.79 (1.30–5.99)1.03 (0.36–2.94)
 p < 0.001p = 0.65
 Cumulative alcohol consumption (consumption-years4)
  05 (9)7 (22)167 (25)0.13 (0.03–0.45)0.42 (0.14–1.22)p = 0.45
  <1035 (65)17 (53)346 (52)1 (referent)1 (referent)
  10–2010 (19)6 (19)99 (15)1.31 (0.53–3.21)0.96 (0.32–2.86)
  >204 (7)2 (6)57 (9)0.51 (0.16–1.67)0.50 (0.10–2.42)
 p =0.01p = 0.68
Sociodemographic factors
 Years at school
  <1038 (70)26 (79)394 (58)1 (referent)1 (referent)p = 0.12
  ≥1016 (30)7 (21)286 (42)0.38 (0.16–0.91)0.42 (0.15–1.19)
 p = 0.004p = 0.27

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Our findings corroborate the view that cervical cancer-associated types of HPV are etiologically linked to VV-SCC risk. Earlier reports have suggested a direct association between the lifetime number of male sexual partners and risk of VV-SCC.2, 6–8, 14 While women with 0 or 1 male sexual partner were at reduced risk in the present Danish setting we observed no difference in risk between women in the reference category reporting a lifetime number of 2–4 male sexual partners and those reporting 10 or more partners. Also, unlike a prior report,2 we found no strong association between age at first sexual intercourse and risk of VV-SCC. However, we confirm strong positive associations with histories of anogenital warts and cervical neoplasia. Unlike some previous studies6, 7 our findings fail to support a role of herpes simplex virus in the etiology of VV-SCC. Taken together, our findings and those of previous studies argue that sexually transmitted hrHPVs play a central role in the etiology of a considerable proportion of VV-SCCs, notably those with vaginal involvement. This view is in accordance with recent prospective evidence that women receiving prophylactic vaccination against HPV types 16 and 18 are effectively protected from developing the VV-SCC precursor lesions VIN3 and VAIN3.15

Abstinence from alcohol consumption was associated with low risk for both VV-SCCvagina and VV-SCCvulva in our study. In a Swedish study,14 women with a heavy alcohol consumption were at elevated risk for VV-SCCvagina but not for VV-SCCvulva, but these associations were not adjusted for tobacco consumption. Our finding that VV-SCCvulva patients had significantly less education than controls accords well with prior US-based observations. Three studies reported that VV-SCC patients tend to have shorter education, lower household incomes and more often belong to lower social strata of society than controls.2, 8, 16, 17 Our findings regarding tobacco consumption are in agreement with prior reports suggesting that tobacco smoking, particularly current smoking, is linked to significantly increased risk of VV-SCC.1, 2, 4, 16 Of note, however, our findings suggest that the role of tobacco smoking is limited to those VV-SCCs that harbour hrHPV DNA in the tumour tissue, reflecting possibly that tobacco smoking might somehow exert its function through biologic interaction with viral proteins.

In spite of the well-established etiologic link between hrHPVs and cervical cancer and the suspected common etiology between cervical cancer and VV-SCC,17, 18 the evidence regarding the role of hrHPV in invasive VV-SCC has been conflicting.8, 19, 20 Tumour tissues from patients with VV-SCC are less often hrHPV-positive than those of patients with cervical carcinoma, which has led to the suggestion that vulvar carcinoma may be composed of 2 different subtypes, one related and one unrelated to infection with hrHPV.3 Earlier studies have reported between 36% and 86% of VV-SCCs to be hrHPV-positive.2, 3, 21 In our material, 62% of 87 PCR-examined VV-SCCs were hrHPV-positive, with HPV types 16 and 18 being present in 80% of all HPV-positive cases. Of note, these 2 hrHPV types are the ones targeted in the 2 currently available prophylactic HPV vaccines aimed at preventing cervical carcinoma and its precursors.15 Consequently, in the longer run these vaccines seem to carry the potential for reducing the burden also of noncervical HPV 16 and HPV 18 associated anogenital cancers, including large proportions of VV-SCCs as well as anal and penile carcinomas.

The risk factor associations we observed for VV-SCCvulva are generally in accordance with findings by Brinton et al.,8 who reported a high number of lifetime sexual partners, an early age at first sexual intercourse, low socioeconomic status, genital warts, previous abnormal pap smear and current tobacco smoking as risk factors for vulvar carcinoma. The uterine cervix is anatomically closer to the vagina than to the vulva, so our findings that DNA from cervical cancer-associated hrHPVs was present in 89% of the VV-SCCvagina cases vs. in only 50% of the VV-SCCvulva cases, and that a history of preinvasive or invasive cervical cancer was more strongly linked to VV-SCCvagina than to VV-SCCvulva both appear plausible. The strong link we observed between anogenital warts and risk of VV-SCCvulva confirms prior reports,8 but our finding of a similar, though somewhat weaker, association between anogenital warts and VV-SCCvagina was not seen in a recent Seattle-based case–control study of vaginal carcinoma.2 Prior results favouring the association of anogenital warts with risk of VV-SCCvulva rather than VV-SCCvagina may partly be due to the limited power of vaginal cancer studies and possibly reflect that self-reported anogenital warts are dominated by visible lesions of the vulva.

Our study has some limitations that need to be considered. Participation rates were modest (around 60%) among invited VV-SCC patients and population-controls and only slightly higher (69%) among cancer controls. We share these difficulties in recruiting VV-SCC patients with other case–control studies whose participation rates among invited VV-SCC patients ranged between 48 and 61%.2, 7, 8, 22 Theoretically, as in the other studies, the ∼40% of invited VV-SCC patients and population controls who declined our invitation to participate could somehow represent socioeconomic or behavioural segments of the population that would make generalization of our findings problematic. Another limitation is the small number of VV-SCC patients, notably patients with VV-SCCvagina, which made some of our statistical analyses unstable. However, unlike other case–control studies, our study benefited from having 2 independent control groups. Only after (a) having seen that risk factor associations were similar in initial logistic regression analyses obtained with the 2 control groups separately and (b) having formally examined the appropriateness of combining the 2 control groups in a logistic regression analysis comparing the 2 control groups, did we combine them to gain statistical power, thus adding robustness and credibility to the reported risk factor associations.

In conclusion, our population-based case–control study provides epidemiological and molecular biological evidence that most cases of VV-SCC, notably those with vaginal involvement, are etiologically linked to sexually transmitted infections with hrHPV types known to cause cervical cancer. Combined with recent evidence that vaccination against primary HPV 16 and HPV 18 infection is effective in preventing VV-SCC precursor lesions,15 it seems that HPV vaccination programs aimed at reducing the burden of cervical cancers will also provide considerable protection against VV-SCC.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank Ms. Murielle Verkuyten, Department of Pathology, VU Medical Center, Amsterdam, for excellent technical assistance in HPV detection and typing. Dr. Frisch has served as a member of a national advisory board for Glaxo Smith Kline, the manufacturer of a vaccine against HPV 16 and HPV 18.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    Ansink AC,Heintz AP. Epidemiology and etiology of squamous cell carcinoma of the vulva. Eur J Obstet Gynecol Reprod Biol 1993; 48: 11115.
  • 2
    Daling JR,Madeleine MM,Schwartz SM,Shera KA,Carter JJ,McKnight B,Porter PL,Galloway DA,McDougall JK,Tamimi H. A population-based study of squamous cell vaginal cancer: HPV and cofactors. Gynecol Oncol 2002; 84: 26370.
  • 3
    Madeleine MM,Daling JR. Cancers of the vulva and vagina. In: SchottenfeldD,FraumeniJF. Cancer epidemiology and prevention. New York: Oxford University Press, 2006. 106874.
  • 4
    Mabuchi K,Bross DS,Kessler II. Epidemiology of cancer of the vulva. A case-control study. Cancer. 1985; 55: 18438.
  • 5
    Merino MJ. Vaginal cancer: the role of infectious and environmental factors. Am J Obstet Gynecol 1991; 165: 125562.
  • 6
    Hildesheim A,Han CL,Brinton LA,Kurman RJ,Schiller JT. Human papillomavirus type 16 and risk of preinvasive and invasive vulvar cancer: results from a seroepidemiological case-control study. Obstet Gynecol 1997; 90: 74854.
  • 7
    Madeleine MM,Daling JR,Carter JJ,Wipf GC,Schwartz SM,McKnight B,Kurman RJ,Beckmann AM,Hagensee ME,Galloway DA. Cofactors with human papillomavirus in a population-based study of vulvar cancer. J Natl Cancer Inst 1997; 89: 151623.
  • 8
    Brinton LA,Nasca PC,Mallin K,Baptiste MS,Wilbanks GD,Richart RM. Case-control study of cancer of the vulva. Obstet Gynecol 1990; 75: 85966.
  • 9
    Sherman KJ,Daling JR,Chu J,Weiss NS,Ashley RL,Corey L. Genital warts, other sexually transmitted diseases, and vulvar cancer. Epidemiology 1991; 2: 25762.
  • 10
    zur Hausen H. Papillomaviruses in anogenital cancer as a model to understand the role of viruses in human cancers. Cancer Res 1989; 49: 467781.
  • 11
    Pedersen CB,Gotzsche H,Moller JO,Mortensen PB. The Danish Civil Registration System. A cohort of eight million persons. Dan Med Bull 2006; 53: 4419.
  • 12
    van den Brule AJ,Pol R,Fransen-Daalmeijer N,Schouls LM,Meijer CJ,Snijders PJ. GP5+/6+ PCR followed by reverse line blot analysis enables rapid and high-throughput identification of human papillomavirus genotypes. J Clin Microbiol 2002; 40: 77987.
  • 13
    Walboomers JM,Jacobs MV,Manos MM,Bosch FX,Kummer JA,Shah KV,Snijders PJ,Peto J,Meijer CJ,Munoz N. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999; 189: 1219.
  • 14
    Weiderpass E,Ye W,Tamimi R,Trichopolous D,Nyren O,Vainio H,Adami HO. Alcoholism and risk for cancer of the cervix uteri, vagina, and vulva. Cancer Epidemiol Biomarkers Prev 2001; 10: 899901.
  • 15
    Joura EA,Leodolter S,Hernandez-Avila M,Wheeler CM,Perez G,Koutsky LA,Garland SM,Harper DM,Tang GW,Ferris DG,Steben M,Jones RW, et al. Efficacy of a quadrivalent prophylactic human papillomavirus (types 6, 11, 16, and 18) L1 virus-like-particle vaccine against high-grade vulval and vaginal lesions: a combined analysis of three randomised clinical trials. Lancet 2007; 369: 1693702.
  • 16
    Daling JR,Sherman KJ,Hislop TG,Maden C,Mandelson MT,Beckmann AM,Weiss NS. Cigarette smoking and the risk of anogenital cancer. Am J Epidemiol 1992; 135: 1809.
  • 17
    Peters RK,Mack TM,Bernstein L. Parallels in the epidemiology of selected anogenital carcinomas. J Natl Cancer Inst 1984; 72: 60915.
  • 18
    Stern BD,Kaplan L. Multicentric foci of carcinomas arising in structures of cloacal origin. Am J Obstet Gynecol 1969; 104: 25566.
  • 19
    Crum CP. Carcinoma of the vulva: epidemiology and pathogenesis. Obstet Gynecol 1992; 79: 44854.
  • 20
    Sturgeon SR,Brinton LA,Devesa SS,Kurman RJ. In situ and invasive vulvar cancer incidence trends (1973 to 1987). Am J Obstet Gynecol 1992; 166: 14825.
  • 21
    Iwasawa A,Nieminen P,Lehtinen M,Paavonen J. Human papillomavirus in squamous cell carcinoma of the vulva by polymerase chain reaction. Obstet Gynecol 1997; 89: 814.
  • 22
    Brinton LA,Nasca PC,Mallin K,Schairer C,Rosenthal J,Rothenberg R,Yordan E,Jr,Richart RM. Case-control study of in situ and invasive carcinoma of the vagina. Gynecol Oncol 1990; 38: 4954.