Journal of Medical Virology 86:975–982 (2014)
Detection of Oncogenic Genital Human Papillomavirus (HPV) Among HPV Negative Older and Younger Women After 7 Years of Follow-Up Kim Agerholm Brogaard,1 Christian Munk,1 Thomas Iftner,2 Kirsten Frederiksen,1 and Susanne K. Kjaer1,3* 1
Unit of Virus, Lifestyle and Genes, Danish Cancer Society Research Center, Copenhagen, Denmark Department of Experimental Virology, Universitaetsklinikum Tuebingen, Tuebingen, Germany 3 Department of Gynaecology, The Juliane Marie Center, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark 2
The knowledge on risk factors of being human papillomavirus (HPV)-positive among older women is sparse. The aim was to determine the frequency of oncogenic HPV appearance after 7 years among initially HPV-negative women and to examine potential risk factors that influence the occurrence of HPV in older women using multiple logistic regression. For comparison, a younger cohort of women examined under identical study settings was included. This prospective cohort study comprised 1,577 older women (age 40–50 at enrolment) and 2,920 women aged 22–32. Participants were interviewed and underwent a gynecological examination at two time points (7 years apart). Cervical samples were tested for HPV using Hybrid Capture 2 (HC2) and only women who tested HC2-negative at baseline were included. The HPV prevalence among older and younger women was 6.4% and 10.7%, respectively, and there was no “second peak” observed among older women. Recent sexual partners were a strong determinant of HPV appearance irrespective of age. Lifetime number of sexual partners was a significant risk factor for HPV appearance among older women, even after adjustment for recent sexual behavior. In addition, menopause was associated with a non-significantly increased risk of HPV appearance at follow-up. In conclusion, appearance of HPV in previously HPV-negative older women may be due to both recent sexual behavior and previous exposure that is, reactivation of a latent HPV infection. J. Med. Virol. 86:975–982, 2014. # 2014 Wiley Periodicals, Inc.
KEY WORDS:
epidemiology; risk factors; sexual behavior; populationbased; natural history
C 2014 WILEY PERIODICALS, INC.
INTRODUCTION Human papillomavirus (HPV) is one of the most frequent sexually transmitted infections. Most infections are transient but persistent oncogenic or highrisk (HR) HPV infections are known to be the main etiological and necessary factor in cervical carcinogenesis [Koutsky et al., 1992; Ho et al., 1995; Remmink et al., 1995]. In most populations the prevalence of HPV is highest in younger age groups occurring after first sexual intercourse and decreasing gradually later in life. However, an increase in the prevalence of HPV infections in older women, the so-called “second peak” has been observed in some studies. This second peak in HPV prevalence coincides with an increase in the incidence of cervical cancer and has been observed most often in countries with a high incidence of cervical cancer [LazcanoPonce et al., 2001; Ferreccio et al., 2004; Munoz Grant sponsor: National Cancer Institute; Grant number: RO1 CA47812.; Grant sponsor: Sixth Research Framework Programme of the European Union, project INCA; Grant number: LSHC-CT-2005-018704.; Grant sponsor: Mermaid II project; Grant sponsor: Rigshospitalets Forskningspulje Disclosure of Potential Conflict of Interest: S. K. Kjaer has received lecture fees, advisory board fees, and unrestricted research grants through her institution from Merck and Sanofi Pasteur MSD. Christian Munk has received lecture fees and support for conference participation from Merck and Sanofi Pasteur MSD. Thomas Iftner has received lecture fees from Hologic and Beckton Dickinson, advisory board fees from GSK and Hologic, and research grants through his institution from Sanofi Pasteur MSD, GSK, and Hologic. Kirsten Frederiksen and Kim Brogaard report no potential conflicts of interest. Correspondence to: Susanne K. Kjaer, Unit of Virus, Lifestyle and Genes, Danish Cancer Society Research Center, Strandboulevarden 49, DK-2100 Copenhagen Ø, Denmark. E-mail: [email protected] Accepted 7 February 2014 DOI 10.1002/jmv.23914 Published online 7 March 2014 in Wiley Online Library (wileyonlinelibrary.com).
976
Brogaard et al.
et al., 2004; Castle et al., 2005; de Sanjose´ et al., 2007; Dartell et al., 2012]. It has been suggested that this peak may be due to newly acquired infections through an increased sexual activity among older women and their partners [Munoz et al., 2004; Trottier et al., 2010]. However, it has not been clarified if and to what extent both current and past sexual behaviors play a role. In addition, it is not known whether changes in hormonal and immunological factors with age can lead to reactivation of a latent HPV infection or impaired clearance [Garcia-Pineres et al., 2006; Gonzalez et al., 2010; Trottier et al., 2010; Gravitt et al., 2013]. Finally, some studies have indicated that the increase could be due to a cohort effect in HPV prevalence such as older women could have been infected later because sexual behaviors were not the same when they were young compared to young women today [Syrjanen et al., 2008; Gravitt, 2011; Gravitt et al., 2013]. Interestingly, animal and molecular studies have shown that HPV infections are able to persist in a latent state in the basal layer of the epithelium. This may support the hypothesis that HPV is able to remain in a latent state and subsequently to become re-activated under certain conditions, which could be related to a weakened immune response [Zhang et al., 1999; Gravitt, 2011; Maglennon et al., 2011]. Data on the incidence of genital HPV infections and risk factors for acquisition among older women are sparse. However, this is important knowledge for charting the natural history of genital HPV infections, and ultimately for planning future cervical cancer screening strategies in the era of HPV vaccination. In recent cohort studies from Costa Rica [Gonzalez et al., 2010] and the United States [Rositch et al., 2012], possible co-factors and determinants for incident HR HPV infections among older women were investigated including marital status, sexual behavior, smoking habits, reproductive history, contraceptive use, and menopausal status. It appears that no studies investigating risk factors for detecting incident HR HPV infections among older women in Europe have been conducted to date. The aim of this study was to assess the frequency of HPV appearance among HPV-negative older women after 7 years of follow-up, and to identify factors that influence the detection of HR HPV infection in older women in Denmark. For comparison, a younger cohort of women were included from whom data were collected simultaneously and in the same setting as the older cohort. MATERIALS AND METHODS Study Population and Data Collection The study population consisted of women from two Danish population-based prospective cohort studies conducted simultaneously on the natural history of HPV and cervical cancer, one study of women aged J. Med. Virol. DOI 10.1002/jmv
40–50 at enrolment (older cohort) and one study of women 20–30 years of age (younger cohort). The studies were approved by the Scientific Ethical Committee of Copenhagen and Frederiksberg Municipality and the Danish Data Protection Agency, and all study participants signed a written informed consent form. From October 1993 to January 1995 2,200 women 40–50 years of age were selected at random from the Copenhagen area. In Denmark, every citizen has a unique 10-digit personal identification number, which is used universally in all health registries. These identification numbers, which contain information on sex and date of birth, are registered in the computerized Danish Civil Registration System database. This register, which is updated on a daily basis, includes information on vital status, emigration, and current address. During the period October 1993 to January 1995 all the selected women were invited to visit a study clinic established especially for this purpose at the gynecological department at Rigshospitalet, Copenhagen, Denmark. The women were interviewed personally by female nurses trained specifically for the task. The structured interview at baseline included questions on socio-demographic factors, current and past sexual behavior, sexually transmitted diseases, contraception, reproductive history, smoking, and alcohol habits. In addition, the women had a gynecological examination with a Pap smear, and samples from endo- and ectocervix were obtained for HPV DNA testing. A total of 1,577 women (72%) were enrolled in the study. The younger cohort consisted of women included previously (2 years earlier) in an identical prospective study from the same study area and at the same study clinic. Of the 11,088 women enrolled originally [Kjaer et al., 2006], a total of 8,656 women (22–32 years of age) participated (78%) in the current study simultaneously with the women from the older cohort and at the same study clinic. The data collection (personal interview, gynecological examination, cervical samples) was identical to that of the older cohort. Approximately 7 years later (November 2001 to June 2002) all women in the older cohort (n ¼ 1,577) and a random sample of baseline HPV-negative women from the younger cohort (n ¼ 2,920) were considered eligible for the follow-up examination. To obtain information on the current address and vital status of all the women in the study, the cohort members were linked initially to the Danish Civil Registration System database by their unique personal identification number. Subsequently, the women were invited by mail to this study examination. For both cohorts data collection in the follow-up examination included another gynecological examination and a cervical sample to be tested for the presence of HR HPV DNA. The cervical samples from both examinations of the two cohorts were analyzed by Hybrid Capture 2 (HC2) to detect the presence of HR HPV types. In addition, the women were interviewed again
HPV Detection in Younger and Older Women
using a structured questionnaire covering questions as described under the baseline questionnaire. Baseline was defined as the examination in 1993– 1995 with follow-up in 2001–2002, giving data from two specific time points approximately 7 years apart. Among the 1,577 women in the older cohort, 212 women who had moved out of the study area, declined at baseline to participate in further examinations or died within the follow-up period were excluded. This left 1,365 women who were invited to participate in the follow-up examination, and a total of 1,007 women (73.8%) were included. Women with a missing or inadequate HC2 test at follow-up were excluded (n ¼ 47). In addition, women with a positive HC2 test at baseline were excluded (n ¼ 38). This left a total of 922 older women with a negative HC2 test at baseline for further analysis. From the younger cohort 2,920 women selected at random who were HC2 negative at baseline were invited. In all 2,339 younger women (80.1%) participated in the follow-up examination. Women with a missing or inadequate HC2 sample at follow-up (n ¼ 182) were excluded from the study, leaving a study population of 2,057 younger women for further analysis. HPV DNA Analysis The method of HR HPV DNA testing has been described previously [Iftner and Villa, 2003]. Laboratory personnel and scientists were blinded to the cytological sample and the clinical diagnosis, respectively. Cervical swabs and Pap smears for HR HPV testing were collected into a medium (PBS with 0.05% merthiolate) and stored at 80˚C until tested. Due to the fact that the cytological materials were collected into a medium, which is not recommended for HC2 test, a conversion protocol was carried out to allow HC2 testing (HC2; Digene Corporation, Gaithersburg, MD) [Kjaer et al., 2006]. The HPV16positive cervical carcinoma cell line SiHa was used as a positive control and the HPV negative C33A cells as a negative control. The analysis was carried out using the HR RNA probe to detect 13 HR types (HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68) [Iftner and Villa, 2003]. The threshold of 1.0 relative light unit coefficient (RLU/CO) was used as a positive HR HPV DNA test corresponding to the Food and Drug Administrationrecommended cutoff of 1.0 pg/ml viral load. In the younger cohort, the tests were performed using the robot platform device (RCS1), which processes four microtiter plates at the same time. Regarding the HPV testing on the older cohort, the same steps were carried out, but the HPV testing using the HR HPV probe was performed manually. Despite that, replicate assays were used on a number of plates in both cohorts with virtually the same results. Intra-laboratory reproducibility was determined by re-testing 1,000 samples on different days and resulted in an
977
overall high agreement (kappa 0.94; 95% CI: 0.92– 0.97). Statistical Analysis For the present analysis HPV correlates were chosen a priori based on the literature on acquisition of HR HPV infection and availability in the data set. The variables originated from the interviews conducted at baseline with the exception of data on recent sexual behavior, and alcohol habits, which were only available from the follow-up examination. Variables considered for inclusion in the statistical model constituted recent and lifetime sexual behavior, socio-demographic factors including age and educational level and lifestyle factors including alcohol and smoking habits. Parity, menopausal status, self-reported history of condom use at follow-up, oral contraceptive use and sexually transmitted diseases were also considered. Associations are presented as odds ratios (OR) with 95% confidence intervals (CI). The potential risk factors were examined individually in an age-adjusted logistic regression model assuming a linear association between age and HPV positivity. The association with lifetime number of sexual partners could be considered as increasing in a linear fashion in both younger and older women. Number of recent sexual partners were concentrated around values from 0 to 3 partners both in the younger and older cohort and showed significant deviations from linearity. As a consequence the categorized parameter estimates was reported for this variable. Variables were entered stepwise in the analysis and were retained in the final statistical model if they were statistically significant at a 10% level. Age was always included in the models. The variables in the final multivariate logistic regression model were number of recent sexual partners (i.e., within the last 12 months), lifetime number of sexual partners and age. An adjustment for the time interval between the two examinations was performed, but as this did not change the results, this variable was not included in the final statistical model. All analyses were conducted using SAS software (version 9.2, SAS Institute, Cary, NC) [Azevedo et al., 2011]. RESULTS The mean age at baseline was 45 years for the older women and 28 years for the younger women. The median lifetime number of sexual partners was nine partners among the older women and eight partners among the younger women. The majority of the older women had 10 or more years of schooling (72.2%), whereas a higher proportion of the younger women had 10 or more years of schooling (94.0%). Table I presents the distribution of sexual and reproductive factors in older and younger women. A higher proportion of the older women had no sexual partners the last year before follow-up (23.8%) J. Med. Virol. DOI 10.1002/jmv
978
Brogaard et al. TABLE I. The Distribution of Sexual and Reproductive Factors in Danish Women Stratified by Age
Baseline variables No. sexual partners the last yearb 0 1 2 3 Lifetime no. sexual partners 4 5–9 10–14 15 Condom useb Current use Non current use Use of oral contraceptives Never Ever No. of pregnancies 0 1 2
Younger women 22–32 years
Older women 40–50 years
Total ¼ 2,057
Total ¼ 922
n (%)a
n (%)a
120 1,648 176 106
(5.8) (80.4) (8.6) (5.2)
216 637 34 20
(23.8) (70.2) (3.8) (2.2)
590 574 366 525
(28.7) (27.9) (17.8) (25.6)
255 215 160 288
(27.8) (23.4) (17.4) (31.4)
651 (32.1) 1,376 (67.9)
134 (15.7) 717 (84.3)
369 (17.9) 1,688 (82.1)
133 (14.4) 789 (85.6)
948 (46.1) 537 (26.1) 572 (27.8)
109 (11.8) 164 (17.8) 649 (70.4)
a
The sum of one variable is not always equal to the total number of participants because of missing values for the specific variable. Measured at follow-up.
b
compared to the younger women (5.8%). Having had 15 of more lifetime number of sex partners was slightly more common among older women than among the younger (older women: 31.4%; younger women: 25.6%). Current use of condoms was rare among the older women (15.7%) compared to the younger women (32.1%) but a similar distribution of ever having used oral contraceptive was observed across the two age cohorts. A higher proportion of the older women reported two or more pregnancies (70.4%) compared to the younger women (27.8%).
Finally, only a small proportion of the older cohort reported being menopausal (6.2%). In all, 59 (6.4%) older women and 207 (10.7%) younger women had a positive HC2 test at follow-up. Figure 1 shows the detection (95% CI) of genital HR HPV among initially HPV negative women after 7 years of follow-up by age at follow-up. The highest proportion of women becoming HR HPV positive was observed among the younger women (aged 30 18.9%; 95% CI: 11.1–26.8) with a significantly decreasing prevalence with age (aged 39 6.0%; 95% CI: 2.8–9.1).
Fig. 1. Detection (95% CI) of genital high-risk (HR) HPV infections among initially HPV negative women after 7 years of follow-up by age at follow-up.
J. Med. Virol. DOI 10.1002/jmv
HPV Detection in Younger and Older Women
Among the older women the proportion of women becoming HR HPV positive showed little variation across age groups (aged 47–48 7.1%; 95% CI: 2.9– 11.3) (aged 55–57 5.9%; 95% CI: 2.5–9.4). Number of Sexual Partners and the Association With Detection of HR HPV Table II presents some of the potential risk factors associated with detection of an HR HPV infection after 7 years among initially HR HPV negative women. Lifetime number of sexual partners was reported at baseline in contrast to recent sexual partners reported as number of partners the last 12 months before follow-up. Among the older women an increasing risk of HPV appearance after 7 years of follow-up was seen with increasing number of recent sexual partners. Older women who reported having had three sexual partners or more in the last year before follow-up had a higher risk of being HPV-positive than women who reported no sexual partners in the last 12 months (OR ¼ 4.0; 95% CI: 0.7–22.6). In the analysis regarding lifetime number of sexual partners among older women, the risk increased significantly by 3% per extra lifetime partner (P ¼ 0.017) even after number of recent partners was taken into account. When the analysis was restricted to women who reported only one partner during the last year, the effect of number of lifetime sexual partners was still present among the older women (OR ¼ 2.4; 95% CI: 1.3–4.4 for 10 or more partners vs. 9 or less partners). In the group of younger women number of recent sexual partners within the last 12 months was associated significantly with the risk of being HR HPV-positive (P < 0.001). Younger women who reported having had three sexual partners or more in the last year before follow-up had a significantly higher risk of being HR HPV-positive than women who reported no sexual partners in the last 12 months (OR ¼ 7.7; 95% CI: 3.2–18.4). In relation to lifetime number of sexual partners at baseline no statistically significantly increased risk per additional partner was seen (OR ¼ 1.00; 95% CI: 0.98–1.02) (P ¼ 0.984) when adjustment was performed for number of recent sexual partners. Other Factors and the Association With Detection of HR HPV The influence of other factors on the risk of becoming HR HPV-positive after 7 years of follow-up among younger and older women who were HPVnegative at baseline was also examined (Table II). All variables were obtained at baseline except for alcohol consumption and self-reported history of condom use only available at follow-up. Among the older women age at baseline was not related to the detection of an HR HPV infection. No increased risk of being HR HPV-positive by alcohol consumption was observed
979
(OR ¼ 1.01; 95% CI: 0.97–1.04 per unit per week). Smoking was not associated with the risk of becoming HR HPV-positive (OR ¼ 1.0; 95% CI: 0.5–1.9 current vs. non-smokers at baseline). Finally, menopausal women tended to be at increased risk of HR HPV, although the association did not reach statistical significance (OR 2.1; 95% CI: 0.8–5.6). Among the younger women the risk of detecting an HR HPV infection decreased significantly with increasing age (9% per year older; 95% CI: 0.86–0.96). Alcohol consumption did not increase the risk of detecting an HR HPV infection (OR ¼ 1.01; 95% CI: 0.98–1.04 per unit per week). However, smoking behavior tended to increase the risk of detecting an HR HPV infection (OR ¼ 1.4; 95% CI: 1.0–2.0 current vs. non-smokers at baseline). No association was found with educational level, parity, self-reported history of sexually transmitted diseases or oral contraceptive use at baseline or condom use reported at follow-up among younger or older women (data not shown).
DISCUSSION In the present study the detection of genital HR HPV among older women in Denmark was observed to be independent of age. The number of recent sexual partners was a strong determinant of HR HPV detection irrespective of age. Furthermore, lifetime number of sexual partners was associated significantly with the risk of being HR HPV-positive after 7 years of follow-up among older women, even after adjustment for recent sexual behavior. The detection of HR HPV at follow-up among initially HPV negative women remained stable from around age 36 and onwards, and an increasing HPV detection among older women was not observed, which has been shown in some other studies from Costa Rica, Tanzania, Chile, Mexico, and Columbia [Lazcano-Ponce et al., 2001; Ferreccio et al., 2004; Castle et al., 2005; de Sanjose´ et al., 2007; Munoz et al., 2009; Dartell et al., 2012]. One of the possible explanations for the absence of a “second peak” in older women in Denmark could be that increases in HPV and subsequent cervical abnormalities are detected and treated through the cervical cancer screening program. This could be one of the likely explanations why studies from countries without or with only limited screening report a “second peak.” The population from which the study is recruited is a highly screened population like many other populations from Western Europe and the U.S. [Gakidou et al., 2008; Saraiya et al., 2013]. Recently, a report from the U.S. has suggested that the absence of a second peak may be due to a cohort effect, primarily deriving from a difference in sexual behavior [Gravitt et al., 2013], that is, older women were infected later because sexual behaviors were not the same when they were young compared to younger women today. J. Med. Virol. DOI 10.1002/jmv
Cases/noncasesa
J. Med. Virol. DOI 10.1002/jmv
1.0 1.3 [0.6–2.8] 3.8 [1.6–8.9] 7.7 [3.2–18.4] 1.0 1.3 [0.9–2.0] 1.4 [0.9–2.3] 1.2 [0.8–1.9] 1.00 [0.98–1.02] 1.0 0.9 [0.6–1.2] 0.6 [0.4–0.8] 0.91 [0.86–0.96]
1.0 1.2 [0.8–1.7] 1.3 [0.7–2.6] 1.01 [0.98–1.04] 1.0 1.4 [0.9–2.2] 1.4 [1.0–2.0]
1.0 1.4 [0.9–2.1] 1.7 [1.1–2.6] 1.7 [1.1–2.5] 1.02 [1.00–1.03] 1.0 0.8 [0.6–1.2] 0.5 [0.4–0.8] 0.91 [0.86–0.95]
1.0 1.4 [1.0–2.0] 1.9 [1.0–3.7] 1.03 [1.00–1.07] 1.0 1.4 [0.9–2.1] 1.5 [1.1–2.1]
Multivariate ORc
1.0 1.3 [0.6–2.8] 3.8 [1.6–8.9] 7.7 [3.2–18.3]
OR 95% CIb
0.087
0.549 0.587
0.009 0.001
0.435 0.984
Detection of Oncogenic Genital Human Papillomavirus (HPV) Among HPV Negative Older and Younger Women After 7 Years of Follow-Up Kim Agerholm Brogaard,1 Christian Munk,1 Thomas Iftner,2 Kirsten Frederiksen,1 and Susanne K. Kjaer1,3* 1
Unit of Virus, Lifestyle and Genes, Danish Cancer Society Research Center, Copenhagen, Denmark Department of Experimental Virology, Universitaetsklinikum Tuebingen, Tuebingen, Germany 3 Department of Gynaecology, The Juliane Marie Center, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark 2
The knowledge on risk factors of being human papillomavirus (HPV)-positive among older women is sparse. The aim was to determine the frequency of oncogenic HPV appearance after 7 years among initially HPV-negative women and to examine potential risk factors that influence the occurrence of HPV in older women using multiple logistic regression. For comparison, a younger cohort of women examined under identical study settings was included. This prospective cohort study comprised 1,577 older women (age 40–50 at enrolment) and 2,920 women aged 22–32. Participants were interviewed and underwent a gynecological examination at two time points (7 years apart). Cervical samples were tested for HPV using Hybrid Capture 2 (HC2) and only women who tested HC2-negative at baseline were included. The HPV prevalence among older and younger women was 6.4% and 10.7%, respectively, and there was no “second peak” observed among older women. Recent sexual partners were a strong determinant of HPV appearance irrespective of age. Lifetime number of sexual partners was a significant risk factor for HPV appearance among older women, even after adjustment for recent sexual behavior. In addition, menopause was associated with a non-significantly increased risk of HPV appearance at follow-up. In conclusion, appearance of HPV in previously HPV-negative older women may be due to both recent sexual behavior and previous exposure that is, reactivation of a latent HPV infection. J. Med. Virol. 86:975–982, 2014. # 2014 Wiley Periodicals, Inc.
KEY WORDS:
epidemiology; risk factors; sexual behavior; populationbased; natural history
C 2014 WILEY PERIODICALS, INC.
INTRODUCTION Human papillomavirus (HPV) is one of the most frequent sexually transmitted infections. Most infections are transient but persistent oncogenic or highrisk (HR) HPV infections are known to be the main etiological and necessary factor in cervical carcinogenesis [Koutsky et al., 1992; Ho et al., 1995; Remmink et al., 1995]. In most populations the prevalence of HPV is highest in younger age groups occurring after first sexual intercourse and decreasing gradually later in life. However, an increase in the prevalence of HPV infections in older women, the so-called “second peak” has been observed in some studies. This second peak in HPV prevalence coincides with an increase in the incidence of cervical cancer and has been observed most often in countries with a high incidence of cervical cancer [LazcanoPonce et al., 2001; Ferreccio et al., 2004; Munoz Grant sponsor: National Cancer Institute; Grant number: RO1 CA47812.; Grant sponsor: Sixth Research Framework Programme of the European Union, project INCA; Grant number: LSHC-CT-2005-018704.; Grant sponsor: Mermaid II project; Grant sponsor: Rigshospitalets Forskningspulje Disclosure of Potential Conflict of Interest: S. K. Kjaer has received lecture fees, advisory board fees, and unrestricted research grants through her institution from Merck and Sanofi Pasteur MSD. Christian Munk has received lecture fees and support for conference participation from Merck and Sanofi Pasteur MSD. Thomas Iftner has received lecture fees from Hologic and Beckton Dickinson, advisory board fees from GSK and Hologic, and research grants through his institution from Sanofi Pasteur MSD, GSK, and Hologic. Kirsten Frederiksen and Kim Brogaard report no potential conflicts of interest. Correspondence to: Susanne K. Kjaer, Unit of Virus, Lifestyle and Genes, Danish Cancer Society Research Center, Strandboulevarden 49, DK-2100 Copenhagen Ø, Denmark. E-mail: [email protected] Accepted 7 February 2014 DOI 10.1002/jmv.23914 Published online 7 March 2014 in Wiley Online Library (wileyonlinelibrary.com).
976
Brogaard et al.
et al., 2004; Castle et al., 2005; de Sanjose´ et al., 2007; Dartell et al., 2012]. It has been suggested that this peak may be due to newly acquired infections through an increased sexual activity among older women and their partners [Munoz et al., 2004; Trottier et al., 2010]. However, it has not been clarified if and to what extent both current and past sexual behaviors play a role. In addition, it is not known whether changes in hormonal and immunological factors with age can lead to reactivation of a latent HPV infection or impaired clearance [Garcia-Pineres et al., 2006; Gonzalez et al., 2010; Trottier et al., 2010; Gravitt et al., 2013]. Finally, some studies have indicated that the increase could be due to a cohort effect in HPV prevalence such as older women could have been infected later because sexual behaviors were not the same when they were young compared to young women today [Syrjanen et al., 2008; Gravitt, 2011; Gravitt et al., 2013]. Interestingly, animal and molecular studies have shown that HPV infections are able to persist in a latent state in the basal layer of the epithelium. This may support the hypothesis that HPV is able to remain in a latent state and subsequently to become re-activated under certain conditions, which could be related to a weakened immune response [Zhang et al., 1999; Gravitt, 2011; Maglennon et al., 2011]. Data on the incidence of genital HPV infections and risk factors for acquisition among older women are sparse. However, this is important knowledge for charting the natural history of genital HPV infections, and ultimately for planning future cervical cancer screening strategies in the era of HPV vaccination. In recent cohort studies from Costa Rica [Gonzalez et al., 2010] and the United States [Rositch et al., 2012], possible co-factors and determinants for incident HR HPV infections among older women were investigated including marital status, sexual behavior, smoking habits, reproductive history, contraceptive use, and menopausal status. It appears that no studies investigating risk factors for detecting incident HR HPV infections among older women in Europe have been conducted to date. The aim of this study was to assess the frequency of HPV appearance among HPV-negative older women after 7 years of follow-up, and to identify factors that influence the detection of HR HPV infection in older women in Denmark. For comparison, a younger cohort of women were included from whom data were collected simultaneously and in the same setting as the older cohort. MATERIALS AND METHODS Study Population and Data Collection The study population consisted of women from two Danish population-based prospective cohort studies conducted simultaneously on the natural history of HPV and cervical cancer, one study of women aged J. Med. Virol. DOI 10.1002/jmv
40–50 at enrolment (older cohort) and one study of women 20–30 years of age (younger cohort). The studies were approved by the Scientific Ethical Committee of Copenhagen and Frederiksberg Municipality and the Danish Data Protection Agency, and all study participants signed a written informed consent form. From October 1993 to January 1995 2,200 women 40–50 years of age were selected at random from the Copenhagen area. In Denmark, every citizen has a unique 10-digit personal identification number, which is used universally in all health registries. These identification numbers, which contain information on sex and date of birth, are registered in the computerized Danish Civil Registration System database. This register, which is updated on a daily basis, includes information on vital status, emigration, and current address. During the period October 1993 to January 1995 all the selected women were invited to visit a study clinic established especially for this purpose at the gynecological department at Rigshospitalet, Copenhagen, Denmark. The women were interviewed personally by female nurses trained specifically for the task. The structured interview at baseline included questions on socio-demographic factors, current and past sexual behavior, sexually transmitted diseases, contraception, reproductive history, smoking, and alcohol habits. In addition, the women had a gynecological examination with a Pap smear, and samples from endo- and ectocervix were obtained for HPV DNA testing. A total of 1,577 women (72%) were enrolled in the study. The younger cohort consisted of women included previously (2 years earlier) in an identical prospective study from the same study area and at the same study clinic. Of the 11,088 women enrolled originally [Kjaer et al., 2006], a total of 8,656 women (22–32 years of age) participated (78%) in the current study simultaneously with the women from the older cohort and at the same study clinic. The data collection (personal interview, gynecological examination, cervical samples) was identical to that of the older cohort. Approximately 7 years later (November 2001 to June 2002) all women in the older cohort (n ¼ 1,577) and a random sample of baseline HPV-negative women from the younger cohort (n ¼ 2,920) were considered eligible for the follow-up examination. To obtain information on the current address and vital status of all the women in the study, the cohort members were linked initially to the Danish Civil Registration System database by their unique personal identification number. Subsequently, the women were invited by mail to this study examination. For both cohorts data collection in the follow-up examination included another gynecological examination and a cervical sample to be tested for the presence of HR HPV DNA. The cervical samples from both examinations of the two cohorts were analyzed by Hybrid Capture 2 (HC2) to detect the presence of HR HPV types. In addition, the women were interviewed again
HPV Detection in Younger and Older Women
using a structured questionnaire covering questions as described under the baseline questionnaire. Baseline was defined as the examination in 1993– 1995 with follow-up in 2001–2002, giving data from two specific time points approximately 7 years apart. Among the 1,577 women in the older cohort, 212 women who had moved out of the study area, declined at baseline to participate in further examinations or died within the follow-up period were excluded. This left 1,365 women who were invited to participate in the follow-up examination, and a total of 1,007 women (73.8%) were included. Women with a missing or inadequate HC2 test at follow-up were excluded (n ¼ 47). In addition, women with a positive HC2 test at baseline were excluded (n ¼ 38). This left a total of 922 older women with a negative HC2 test at baseline for further analysis. From the younger cohort 2,920 women selected at random who were HC2 negative at baseline were invited. In all 2,339 younger women (80.1%) participated in the follow-up examination. Women with a missing or inadequate HC2 sample at follow-up (n ¼ 182) were excluded from the study, leaving a study population of 2,057 younger women for further analysis. HPV DNA Analysis The method of HR HPV DNA testing has been described previously [Iftner and Villa, 2003]. Laboratory personnel and scientists were blinded to the cytological sample and the clinical diagnosis, respectively. Cervical swabs and Pap smears for HR HPV testing were collected into a medium (PBS with 0.05% merthiolate) and stored at 80˚C until tested. Due to the fact that the cytological materials were collected into a medium, which is not recommended for HC2 test, a conversion protocol was carried out to allow HC2 testing (HC2; Digene Corporation, Gaithersburg, MD) [Kjaer et al., 2006]. The HPV16positive cervical carcinoma cell line SiHa was used as a positive control and the HPV negative C33A cells as a negative control. The analysis was carried out using the HR RNA probe to detect 13 HR types (HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68) [Iftner and Villa, 2003]. The threshold of 1.0 relative light unit coefficient (RLU/CO) was used as a positive HR HPV DNA test corresponding to the Food and Drug Administrationrecommended cutoff of 1.0 pg/ml viral load. In the younger cohort, the tests were performed using the robot platform device (RCS1), which processes four microtiter plates at the same time. Regarding the HPV testing on the older cohort, the same steps were carried out, but the HPV testing using the HR HPV probe was performed manually. Despite that, replicate assays were used on a number of plates in both cohorts with virtually the same results. Intra-laboratory reproducibility was determined by re-testing 1,000 samples on different days and resulted in an
977
overall high agreement (kappa 0.94; 95% CI: 0.92– 0.97). Statistical Analysis For the present analysis HPV correlates were chosen a priori based on the literature on acquisition of HR HPV infection and availability in the data set. The variables originated from the interviews conducted at baseline with the exception of data on recent sexual behavior, and alcohol habits, which were only available from the follow-up examination. Variables considered for inclusion in the statistical model constituted recent and lifetime sexual behavior, socio-demographic factors including age and educational level and lifestyle factors including alcohol and smoking habits. Parity, menopausal status, self-reported history of condom use at follow-up, oral contraceptive use and sexually transmitted diseases were also considered. Associations are presented as odds ratios (OR) with 95% confidence intervals (CI). The potential risk factors were examined individually in an age-adjusted logistic regression model assuming a linear association between age and HPV positivity. The association with lifetime number of sexual partners could be considered as increasing in a linear fashion in both younger and older women. Number of recent sexual partners were concentrated around values from 0 to 3 partners both in the younger and older cohort and showed significant deviations from linearity. As a consequence the categorized parameter estimates was reported for this variable. Variables were entered stepwise in the analysis and were retained in the final statistical model if they were statistically significant at a 10% level. Age was always included in the models. The variables in the final multivariate logistic regression model were number of recent sexual partners (i.e., within the last 12 months), lifetime number of sexual partners and age. An adjustment for the time interval between the two examinations was performed, but as this did not change the results, this variable was not included in the final statistical model. All analyses were conducted using SAS software (version 9.2, SAS Institute, Cary, NC) [Azevedo et al., 2011]. RESULTS The mean age at baseline was 45 years for the older women and 28 years for the younger women. The median lifetime number of sexual partners was nine partners among the older women and eight partners among the younger women. The majority of the older women had 10 or more years of schooling (72.2%), whereas a higher proportion of the younger women had 10 or more years of schooling (94.0%). Table I presents the distribution of sexual and reproductive factors in older and younger women. A higher proportion of the older women had no sexual partners the last year before follow-up (23.8%) J. Med. Virol. DOI 10.1002/jmv
978
Brogaard et al. TABLE I. The Distribution of Sexual and Reproductive Factors in Danish Women Stratified by Age
Baseline variables No. sexual partners the last yearb 0 1 2 3 Lifetime no. sexual partners 4 5–9 10–14 15 Condom useb Current use Non current use Use of oral contraceptives Never Ever No. of pregnancies 0 1 2
Younger women 22–32 years
Older women 40–50 years
Total ¼ 2,057
Total ¼ 922
n (%)a
n (%)a
120 1,648 176 106
(5.8) (80.4) (8.6) (5.2)
216 637 34 20
(23.8) (70.2) (3.8) (2.2)
590 574 366 525
(28.7) (27.9) (17.8) (25.6)
255 215 160 288
(27.8) (23.4) (17.4) (31.4)
651 (32.1) 1,376 (67.9)
134 (15.7) 717 (84.3)
369 (17.9) 1,688 (82.1)
133 (14.4) 789 (85.6)
948 (46.1) 537 (26.1) 572 (27.8)
109 (11.8) 164 (17.8) 649 (70.4)
a
The sum of one variable is not always equal to the total number of participants because of missing values for the specific variable. Measured at follow-up.
b
compared to the younger women (5.8%). Having had 15 of more lifetime number of sex partners was slightly more common among older women than among the younger (older women: 31.4%; younger women: 25.6%). Current use of condoms was rare among the older women (15.7%) compared to the younger women (32.1%) but a similar distribution of ever having used oral contraceptive was observed across the two age cohorts. A higher proportion of the older women reported two or more pregnancies (70.4%) compared to the younger women (27.8%).
Finally, only a small proportion of the older cohort reported being menopausal (6.2%). In all, 59 (6.4%) older women and 207 (10.7%) younger women had a positive HC2 test at follow-up. Figure 1 shows the detection (95% CI) of genital HR HPV among initially HPV negative women after 7 years of follow-up by age at follow-up. The highest proportion of women becoming HR HPV positive was observed among the younger women (aged 30 18.9%; 95% CI: 11.1–26.8) with a significantly decreasing prevalence with age (aged 39 6.0%; 95% CI: 2.8–9.1).
Fig. 1. Detection (95% CI) of genital high-risk (HR) HPV infections among initially HPV negative women after 7 years of follow-up by age at follow-up.
J. Med. Virol. DOI 10.1002/jmv
HPV Detection in Younger and Older Women
Among the older women the proportion of women becoming HR HPV positive showed little variation across age groups (aged 47–48 7.1%; 95% CI: 2.9– 11.3) (aged 55–57 5.9%; 95% CI: 2.5–9.4). Number of Sexual Partners and the Association With Detection of HR HPV Table II presents some of the potential risk factors associated with detection of an HR HPV infection after 7 years among initially HR HPV negative women. Lifetime number of sexual partners was reported at baseline in contrast to recent sexual partners reported as number of partners the last 12 months before follow-up. Among the older women an increasing risk of HPV appearance after 7 years of follow-up was seen with increasing number of recent sexual partners. Older women who reported having had three sexual partners or more in the last year before follow-up had a higher risk of being HPV-positive than women who reported no sexual partners in the last 12 months (OR ¼ 4.0; 95% CI: 0.7–22.6). In the analysis regarding lifetime number of sexual partners among older women, the risk increased significantly by 3% per extra lifetime partner (P ¼ 0.017) even after number of recent partners was taken into account. When the analysis was restricted to women who reported only one partner during the last year, the effect of number of lifetime sexual partners was still present among the older women (OR ¼ 2.4; 95% CI: 1.3–4.4 for 10 or more partners vs. 9 or less partners). In the group of younger women number of recent sexual partners within the last 12 months was associated significantly with the risk of being HR HPV-positive (P < 0.001). Younger women who reported having had three sexual partners or more in the last year before follow-up had a significantly higher risk of being HR HPV-positive than women who reported no sexual partners in the last 12 months (OR ¼ 7.7; 95% CI: 3.2–18.4). In relation to lifetime number of sexual partners at baseline no statistically significantly increased risk per additional partner was seen (OR ¼ 1.00; 95% CI: 0.98–1.02) (P ¼ 0.984) when adjustment was performed for number of recent sexual partners. Other Factors and the Association With Detection of HR HPV The influence of other factors on the risk of becoming HR HPV-positive after 7 years of follow-up among younger and older women who were HPVnegative at baseline was also examined (Table II). All variables were obtained at baseline except for alcohol consumption and self-reported history of condom use only available at follow-up. Among the older women age at baseline was not related to the detection of an HR HPV infection. No increased risk of being HR HPV-positive by alcohol consumption was observed
979
(OR ¼ 1.01; 95% CI: 0.97–1.04 per unit per week). Smoking was not associated with the risk of becoming HR HPV-positive (OR ¼ 1.0; 95% CI: 0.5–1.9 current vs. non-smokers at baseline). Finally, menopausal women tended to be at increased risk of HR HPV, although the association did not reach statistical significance (OR 2.1; 95% CI: 0.8–5.6). Among the younger women the risk of detecting an HR HPV infection decreased significantly with increasing age (9% per year older; 95% CI: 0.86–0.96). Alcohol consumption did not increase the risk of detecting an HR HPV infection (OR ¼ 1.01; 95% CI: 0.98–1.04 per unit per week). However, smoking behavior tended to increase the risk of detecting an HR HPV infection (OR ¼ 1.4; 95% CI: 1.0–2.0 current vs. non-smokers at baseline). No association was found with educational level, parity, self-reported history of sexually transmitted diseases or oral contraceptive use at baseline or condom use reported at follow-up among younger or older women (data not shown).
DISCUSSION In the present study the detection of genital HR HPV among older women in Denmark was observed to be independent of age. The number of recent sexual partners was a strong determinant of HR HPV detection irrespective of age. Furthermore, lifetime number of sexual partners was associated significantly with the risk of being HR HPV-positive after 7 years of follow-up among older women, even after adjustment for recent sexual behavior. The detection of HR HPV at follow-up among initially HPV negative women remained stable from around age 36 and onwards, and an increasing HPV detection among older women was not observed, which has been shown in some other studies from Costa Rica, Tanzania, Chile, Mexico, and Columbia [Lazcano-Ponce et al., 2001; Ferreccio et al., 2004; Castle et al., 2005; de Sanjose´ et al., 2007; Munoz et al., 2009; Dartell et al., 2012]. One of the possible explanations for the absence of a “second peak” in older women in Denmark could be that increases in HPV and subsequent cervical abnormalities are detected and treated through the cervical cancer screening program. This could be one of the likely explanations why studies from countries without or with only limited screening report a “second peak.” The population from which the study is recruited is a highly screened population like many other populations from Western Europe and the U.S. [Gakidou et al., 2008; Saraiya et al., 2013]. Recently, a report from the U.S. has suggested that the absence of a second peak may be due to a cohort effect, primarily deriving from a difference in sexual behavior [Gravitt et al., 2013], that is, older women were infected later because sexual behaviors were not the same when they were young compared to younger women today. J. Med. Virol. DOI 10.1002/jmv
Cases/noncasesa
J. Med. Virol. DOI 10.1002/jmv
1.0 1.3 [0.6–2.8] 3.8 [1.6–8.9] 7.7 [3.2–18.4] 1.0 1.3 [0.9–2.0] 1.4 [0.9–2.3] 1.2 [0.8–1.9] 1.00 [0.98–1.02] 1.0 0.9 [0.6–1.2] 0.6 [0.4–0.8] 0.91 [0.86–0.96]
1.0 1.2 [0.8–1.7] 1.3 [0.7–2.6] 1.01 [0.98–1.04] 1.0 1.4 [0.9–2.2] 1.4 [1.0–2.0]
1.0 1.4 [0.9–2.1] 1.7 [1.1–2.6] 1.7 [1.1–2.5] 1.02 [1.00–1.03] 1.0 0.8 [0.6–1.2] 0.5 [0.4–0.8] 0.91 [0.86–0.95]
1.0 1.4 [1.0–2.0] 1.9 [1.0–3.7] 1.03 [1.00–1.07] 1.0 1.4 [0.9–2.1] 1.5 [1.1–2.1]
Multivariate ORc
1.0 1.3 [0.6–2.8] 3.8 [1.6–8.9] 7.7 [3.2–18.3]
OR 95% CIb
0.087
0.549 0.587
0.009 0.001
0.435 0.984
