Prevalence of cervical human papillomavirus infection in healthy women is related to sexual behaviours and educational level: a cross-sectional study

Abstract

This study reports the prevalence and risk factors of human papillomavirus (HPV) infection in healthy women in Singapore. Demography, education, sexual and reproductive history and cigarette smoking habits were obtained from a cross-sectional population of healthy women and girls aged above 12 years of age. Cervical or vaginal cytology samples were investigated for 37 known anogenital HPV subtypes using the linear array PCR method. Chi square statistics were used to test for associations of individual epidemiological factors with HPV infection. Independent risk factors were identified with binomial logistic regression analysis. Of 891 subjects, the prevalence of HPV infection was 9.31% (83/891 women) for any-type HPV and 5.05% (46/891 women) for the high-risk HPV (hrHPV). Of 30 HPV subtypes detected, the most prevalent genotypes in descending order of frequency were subtypes 51, 16, 52, 58 and 66 for hrHPV and subtypes 62, 61, 84, 72 and 53 for the low-risk HPV. This frequency distribution of HPV subtypes was different from reports from other countries within Asia. Forty-six virgins studied tested negative for HPV infection. Significant independent risk factors for any-type HPV infection were multiple sexual partners (adjusted OR 1.4) and low (≤6 years) educational level (adjusted OR 4.0). The distribution of HPV subtypes in healthy women varies between different countries within Asia. In Singapore, the prevalence of HPV infection was 9.31% and was related to penetrative sexual intercourse, multiple sexual partners and low educational level.

Keywords

Cigarette smoking human papillomavirus HPV sexually transmitted infection risk factors epidemiology sexual debut multiple sexual partners penetrative vaginal sexual intercourse

Introduction

Oncogenic or high-risk human papillomavirus (hrHPV) infection is the essential aetiological factor in invasive cervical cancer. Pooled data from 38 countries showed a remarkably consistent global trend that HPV types 16 and 18 collectively contribute to 71% of squamous cell carcinoma, and together with HPV 45, contribute to 94% of cervical adenocarcinoma.¹ In contrast, the prevalence of HPV subtype distribution in women without neoplasia shows a wide variation across geographic regions, ranging from 6.2% in South East Asia to 22% in Africa.² The accuracy of the estimated prevalence of HPV infection in South East Asia is questionable as data on epidemiology of HPV infection in the region is sparse. In fact, the available regional data contributed to a mere 0.48% of the sample size in the most recent review of HPV infection in more than 1 million normal women worldwide.³ We also noted that, compared to the plethora of reports on HPV prevalence, studies published on risk factors for HPV infection without cervical neoplasia have been few in number globally and none from South East Asia.⁴ We felt that the information gaps in these respects needed to be addressed urgently as the burden of invasive cervical cancer is particularly high in South East Asia.⁵

This study aimed to determine the general prevalence of genital HPV infection (any-type HPV) and individual HPV genotypes among women with normal cervices in Singapore. The study also examined the factors associated with detection of HPV infection.

Methods of study

Study population and data collection

This cross-sectional study was approved by the Singhealth Centralised Institutional Review Board for clinical research. Subjects agreed to participate in this research by signing an informed consent. Girls below 18 years of age had the consent endorsed by their mothers.

Participants for this study were recruited from the general community through an invitation published in local newspapers and on radio broadcasting. The inclusion criteria were females above the age of 12 years and women who had previously participated in cytological screening for cervical cancer unless they were virgins. Women were excluded if there was a history of hysterectomy, current or previous cervical neoplasia or abnormal cytology. Participants were interviewed for socio-demographic data (ethnic origin – Chinese, Malay, Indian and others – age, number of years in formal school education), sexual behaviours (ever had a history of penetrative vaginal intercourse, age at start of active sexual activity, lifetime number of sexual partners, use of contraception and method), reproductive history (number of pregnancies, termination of pregnancy and delivery) and smoking habits (participants and their spouse). A cervical smear was obtained in liquid-based Thinprep cell preservative for HPV DNA analysis. A vaginal lavage instead of cervical smear was obtained from the virgins.

HPV DNA PCR and HPV genotyping

HPV detection and genotyping were performed using the Roche LA HPV Genotyping Test for 37 anogenital HPV subtypes: high-risk HPV (hrHPV) included types 16, 18, 31, 33, 39, 45, 51, 52, 56, 58, 59, 66, 68 and 73. The low-risk HPV included types 6, 11, 26, 35, 40, 42, 53, 54, 55, 61, 62, 64, 67, 69, 70, 71, 72, 81, 82, 83, 84, IS39 and CP6108. DNA extraction was carried out using the total nucleic acid isolation (TNAI) kit on the automated COBAS® Ampliprep Instrument (Roche Diagnostics, Germany) as previously described.⁶ PCR was carried out on the DNA extracts using biotinylated HPV primers that targeted the 450-base-pair fragment within the polymorphic L1 region of the HPV genome. An additional primer pair targeting a 268-base-pair fragment of the human β-globin gene was used to provide an internal control for adequacy of cells and DNA extraction and amplification. Hybridisation and detection were automated on the Auto-LiPA 48 instrument (Innogenetics N.V., Belgium). Results of the Linear Array HPV Genotyping Strips were read manually. One positive and one negative control were run with each PCR and hybridisation batch. Samples negative for both HPV and β-globin were considered inadequate for interpretation.

Data analysis

Exposure to HPV infection was defined by detection of HPV DNA of any subtype, denoted by HPV-(any-types). Data were analysed using SPSS (17.0) statistical package. Descriptive information on the prevalence of HPV-(any-types) and genotype-specific HPV subtypes for women in each age group was computed. Comparison of differences in the prevalence between 5-year age groupings, between women younger or older than 40 years and frequency difference in risk factors between HPV-positive and HPV-negative groups was evaluated with Chi square test, with Yate’s correction where the number in the group was less than 10. Binomial logistic regression analysis on data from socio-demography, sexual behaviours, reproductive history and cigarette smoking habits was carried out to determine the independent epidemiological risk factors for HPV infection.

Results

In all, 971 women between 13 and 82 years of age were recruited. The mean age was 41.8 (95% confidence interval, 22.30–61.28) years.

Technically valid HPV DNA results were obtained in 891 women who formed the subjects for this analysis. Eighty-three women had detectable HPV DNA from 30 HPV subtypes (Figure 1). The prevalence of HPV infection (HPV-any-type) and hrHPV infection was 9.31% and 5.05%, respectively. HPV infection involved a single genotype in 74.7% (62/83 cases), two genotypes in 18.1% (15/83 cases) and three genotypes in 6.0% (5/83 cases). The most prevalent HPV genotypes in descending order of frequency were HPV 51, 16, 52, 58, 66, 68 and 18 for hrHPV and HPV 62, 61, 84, 72 and 53 for the low-risk HPV subtypes (Figure 1). Of all HPV-positive cases, HPV-6 accounted for 2.7% and HPV-16 accounted for 7.3%.

Figure 1.

Relative frequency of the 30 HPV subtypes among HPV-positive women.

Impact of age

The age-specific prevalence of HPV-(any-type) and hrHPV infection showed a similar pattern with the highest frequency of 26.1% and 21.7%, respectively, for the 20–24 year-olds and the lowest frequency of 4.8% and 2.0%, respectively, for the 50–54 year-olds (Figure 2). Overall, the mean (95% CI) age for the HPV-positive group was 40.0 (18.3–61.7) years, compared to 42.0 (23.2–60.8) years for the HPV-negative group. This difference was not statistically significant. However, compared to women above 40 years of age, younger women had a significantly higher prevalence of any-type HPV infection (12.4% versus 8.0%, p = 0.036) and hrHPV infection (9.5% versus 2.4%, p < 0.001); see Table 1.

Figure 2.

Age-specific prevalence of HPV infection.

Table 1.

Distribution frequency of epidemiologic characteristics between HPV-positive and HPV-negative women.

	n	Any-type HPV		High-risk HPV
	n	HPV−	HPV+	HPV−	HPV+
Age group (years)
≤39	347	304	43 (12.4)	314	33 (9.5)
≥40	498	458	40 (8.0)	486	12 (2.4)
p Value		0.036		<0.001
Ethnic group
Chinese	727	658	69 (9.1)	689	38 (5.2)
Malay	56	51	5 (8.6)	54	2 (3.6)
Indian	44	39	5 (10.2)	43	1 (2.3)
Eurasian	17	13	4 (8.3)	13	4 (23.5)
Others	1	1	0	1	0
p Value		0.391		0.016
Education (years)
≤6	32	25	7 (26.6)	32	0
7–12	323	290	33 (9.6)	308	15 (4.6)
≥13	448	414	34 (7.2)	425	23 (5.1)
Unknown	42	33	9 (20.9)	35	7 (16.7)
p Value		0.001^a		0.022^a
		0.045^b		0.67^b
Age of sexual debut^a
<20	92	78	14 (15.2)	86	6 (6.5)
20–29	569	521	48 (8.4)	540	29 (5.1)
30+	98	88	10 (10.2)	96	2 (2.0)
Unknown	86	75	11 (12.8)	78	8 (9.3)
p Value		0.253^a		0.313^a
		0.174^b		0.495^b
Number of sexual partners^a
1 only	631	586	45 (7.1)	609	22 (3.5)
2 or more	158	132	26 (16.5)	142	16 (10.1)
Unknown	56	44	12 (21.4)	49	7 (12.5)
p Value		<0.001^a		<0.001^a
		<0.001^b		0.001^b
Contraceptive method
None	361	324	37 (10.2)	345	16 (4.4)
Condom	259	234	25 (9.7)	242	17 (6.5)
Combined pills	66	59	7 (10.6)	64	2 (3.0)
Intra-uterine device	123	116	7 (5.7)	119	4 (3.3)
Tubal occlusion	13	12	1 (7.7)	12	1 (7.7)
Unknown	23	17	6 (2.6)	18	5 (2.2)
Yate’s p value		0.209^a		0.046^a
		0.758^b		0.751^b
Number of pregnancies
None	152	125	27 (17.8)	137	15 (9.9)
1–2	368	340	28 (7.6)	355	13 (3.5)
3 or more	303	281	22 (7.3)	291	12 (4.0)
Unknown	22	16	27 (17.8)	17	5 (2.3)
p Value		<0.001^a		<0.001^a
		<0.001^b		0.013^b
Number of abortions
None	655	591	64 (9.8)	623	32 (4.9)
1	119	111	8 (6.7)	115	4 (3.4)
2 or more	49	44	5 (10.2)	45	4 (8.2)
Unknown	22	16	6 (27.3)	17	5 (2.3)
p Value		0.031^a		0.011^a
		0.678^b		0.652^b
Number of deliveries
None	180	149	31 (17.2)	164	16 (8.9)
1–2	493	462	31 (6.3)	477	16 (3.2)
3 or more	148	134	14 (9.5)	141	7 (4.7)
Unknown	24	17	7 (29.2)	18	6 (25.0)
p Value		<0.001^a		<0.001^a
		<0.001^b		0.017^b
Women’s smoking status
Non-smoker	812	739	73 (9.0)	772	40 (4.9)
Smoker	19	14	5 (26.3)	18	1 (5.2)
Unknown	14	9	5 (35.7)	10	4 (28.6)
p Value		<0.001^a		0.004^a
		0.031^b		0.639^b
Spouse’s smoking status
Non-smoker	665	607	58 (8.7)	635	30 (4.5)
Smoker	163	143	20 (12.3)	152	11 (6.7)
Unknown	17	12	5 (29.4)	13	4 (23.5)
p Value		0.009^a		0.011^a
		0.215^b		0.328^b

HPV: human papillomavirus; number in ( ) denotes per cent.

p Value based on analysis including the ‘unknown’ category.

p Value based on analysis excluding the ‘unknown’ category.

Impact of ethnicity and educational level

There was no difference in the prevalence of any-type HPV infection among the major ethnic groups in the country (Table 1). The significance of ethnicity on hrHPV infection was due to a disproportionately high prevalence of infection detected among a small group of Eurasians (23.5%). However, women’s duration of formal education was significantly related to any-type HPV infection (p = 0.001) and hrHPV infection (p = 0.002). The prevalence of HPV infection was 26.6% among women who had attended 6 years or less of formal schooling, 9.6% for women who received 7–12 years and 7.2% for those who attended more then 12 years of education.

Impact of sexual behaviours

The median (range) age among the 46 women who had never had vaginal sexual intercourse was 33 (13–57) years. None of these women was positive for any type of HPV infection. They were excluded from analysis of sexual behaviours as risk factors for HPV infection.

The mean age of sexual debut was 24.7 years, ranging from 14 to 45 years. The prevalence of HPV infection was not affected by age of sexual debut (Table 1). However, the prevalence of hrHPV infection was 6.5% for women whose age of sexual debut was below 20 years, 5.1% for those between 20 and 29 years and 2.0% for those above 30 years (p = 0.311). The difference did not reach statistically significance.

The mean number of sexual partners of these women was 1.3, ranging from 1 to 15. HPV exposure and the prevalence of hrHPV infection was 7.1% and 3.5%, respectively, for women who had only had one sexual partner, compared to 16.5% and 10.1%, respectively, for women who had multiple partners (Table 1). These differences between the 2 groups of women were statistically significant.

Impact of reproductive history

The prevalence of any-type HPV and hrHPV infection was similar for women employing different methods of contraception (Table 1). Women who omitted to report on their method of contraception showed a high prevalence of overall HPV infection (28.6%) and hrHPV infection (23.8%). In the remaining women, HPV infection detection rate was 10.2% (37/361 cases) among women who did not use any method of contraception, 9.7% (25/259 cases) among women whose partners used condoms and 7.4% (15/209 cases) among women who used all types of non-barrier methods pooled together.

The mean number of pregnancies among these women was 2.1 (range 0 to 9). The prevalence of any-type HPV infection was significantly higher among women who reported no history of pregnancy compared to those with history of pregnancy (17.8% versus 7.6%) and among those who reported no history of childbirth (17.2% versus 6.3%). The prevalence of hrHPV infection was significantly higher among women who reported no history of pregnancy compared to those with history of pregnancy (9.9% versus 3.2%) and among those who reported no history of childbirth (8.9% versus 6.3%).

Impact of cigarette smoking

In this cohort, merely 2.2% (19/845 cases) of women were active cigarette smokers. However, 19.2% (163/845 cases) of their partners were active smokers (Table 1). Smoking status of the women and the spouse had a significant impact on any-type HPV detection rate but not on the detection rate of hrHPV infection. After excluding women with unknown smoking habits, the prevalence of any-type HPV infection remained significantly higher among smokers than non-smokers (26.3% versus 9.0%, χ² = 4.674, Yate’s p = 0.031). Analysis of smoking history of spouse showed that, after excluding those with unknown smoking habits, the difference in the exposure to spouse’s smoking among women infected by any-type HPV infection was no longer significant statistically (p = 0.215).

Binomial logistic regression analysis

Factors related to any-type HPV infection were analysed by binomial logistic regression model. Table 2 summarised the results of the regression analysis. Compared to women without HPV infection, women detected with any-type HPV infection did not show a significantly different exposure rate to the following factors: women’s age, age at sexual debut, number of pregnancies and childbirths and smoking habits as well as the spouse’s smoking status. The significant independent risk factors for any-type HPV infection were low level of education (adjusted odd ratio 4.1, 95% CI 1.5–11.4, p = 0.007) and multiple sexual partners (adjusted odd ratio 1.4, 95% CI 1.2–1.8, p = 0.001).

Table 2.

Results of multinomial regression analysis showing risk for exposure to HPV infection of the uterine cervix.

Risk factor	Odd ratio	95% Confidence interval		p-value
Risk factor	Odd ratio	Lower Limit	Upper Limit	p-value
Age	0.998	0.965	1.032	0.919
Number of pregnancies	0.814	0.563	1.177	0.274
Number of childbirths	0.915	0.571	1.467	0.713
Age at sexual debut	0.993	0.930	1.059	0.823
Number of sexual partners	1.434	1.155	1.779	0.001
Years of education (≤6 years)	4.098	1.468	11.438	0.007
Women non-smoker	1.016	0.243	4.245	0.982
Spouse non-smoker	0.785	0.412	1.494	0.460

HPV: human papillomavirus.

Discussion

This study documented the prevalence of HPV infection and distribution of different HPV subtypes in women with normal cytology. This population provided a better representation of exposure to infection by different HPV subtypes than a population with abnormal cytology in whom hrHPV subtypes would be over-represented.

In this study population, the overall HPV exposure rate (9.31%) was higher than the estimated rate of 6.2% (95% CI 5.5%–7.0%) in a recent pooled analysis involving 4849 women in 11 reports from Vietnam, Thailand, the Philippines and Indonesia.³ The rate, however, was markedly lower than a report from southern Malaysia and Singapore using hybrid capture-II (HC-II) technology (25.6%), Taiwan (13.8%), the United States of America (26.8%) and Britain (29%).^7–10 The low exposure rate to HPV infection in the current study could be the effect of prior screening and low number of sexual partners in this population. Nonetheless, almost a tenth of healthy women had detectable HPV infection of the uterine cervix. It illustrated the ubiquitous nature of HPV infection in the female lower genital tract.

The most prevalent hrHPV subtypes found in this study were HPV 51 (1.0%), 16 (0.9%), 52 (0.9%), 58 (0.8%), 66 (0.4%), 68 (0.4%) and 18 (0.3%) compared to HPV 16 (2.5%), 18 (1.4%) and 52 (0.7%) in the pooled data for the Asian Continent and HPV 52 (1.9%), 16 (1.4%), 18 (1.4%) and 31 (0.9%) in Taiwan.^3,8 Although the high prevalence of HPV subtypes 16 and 18 in the pooled data and in Taiwan could possibly be explained by inclusion of women with cervical neoplasia in their studies, there was a distinct possibility that there was a genuine geographic variation in genotype-specific hrHPV distribution. Examination of the prevalence of low-risk HPV infection was useful as they were less often associated with cervical neoplasia as compared to HPV 16 and HPV18. In this respect, the most prevalent low-risk HPV subtypes were HPV 62 (1.0%), 61 (0.7%), 84 (0.7%), 72 (0.6%) and 53 (0.4%) in the current study and HPV 11 (1.8%), 53 (1%), 70 (0.8%), 82 (0.7%), 44 (0.7%) and 84 (0.7%) in Taiwan. These results supported geographic variations of HPV type-specific distribution, even within the same continent.

In contrast to a stark difference in geographic variation in the distribution of HPV subtypes among women with normal cytology, the five most common HPV subtypes contributing to more than 80% of invasive cervical cancer were remarkably similar across all the continents.¹ This observation could be explained by the specific oncogenic potential of HPV subtypes 16, 18, 45, 33 and 31. Therefore, the seemingly low prevalence of HPV subtypes 18 and 45 in women in this study should not negate the importance of the current HPV vaccines in prevention of cervical cancer.

Risk factors for HPV infection, compared to cervical cancer, are not well-studied and their variations across cultural divides between continents are even more obscure. In this respect, we believe that the current study added valuable information for the epidemiology of HPV infection. The current study showed that none of the 46 virgins across an age range from 13 to 57 years (median age 33 years old) was positive for HPV infection. This finding strongly supported the hitherto limited direct evidence that HPV infection of the cervix is related to penetrative vaginal sexual activity. One prospective longitudinal study of college girls in the USA reported that 1.7% of samples from 757 visits of virgins were HPV positive and, after controlling for penetrative vaginal sexual intercourse, other contacts such as penile-vulva or finger-vulva contacts were not significant risk factors for female HPV infection.¹¹

The mean number (1.3, range 1 to 15) of sexual partners of the women in this study was low. Nonetheless, the impact of multiple sexual partners was significant for both general exposure to HPV infection, hrHPV infection and infection by multiple HPV subtypes (Table 1) as reported in several longitudinal studies.^11–13

The peak prevalence of HPV infection was observed in the 20- to 24-year-old group with 26.1% for any-type HPV infection and 21.7% for hrHPV infection. The younger women also experienced a significantly higher prevalence of multiple HPV infection (3.2% versus 1.6%). The decline in age-specific prevalence of HPV infection for mid-adult women until 50 years of age observed in our study was consistent with other reports.^9,14 These data indicated that the observed age-related difference in prevalence of HPV infection in mid-adult women could be explained by natural clearance of HPV infection over time.

Our data showed that the prevalence of HPV infection was higher among women with no prior pregnancy or childbirth (17.8% for no-pregnancy group versus 7.6% for 1–2 pregnancy group for any-type HPV infection and 9.9% for no-pregnancy group versus 3.5% for 1–2 pregnancy group for hrHPV infection). These findings were similar to the report by Louvanto et al.¹⁵ However, the significance of pregnancy and childbirth diminished on binomial logistic regression analysis after adjusting for women’s age. Data in Figure 1 showed that the prevalence of HPV infection was the highest among young women who were more likely to experience either no or small number of pregnancy and childbirth.

The prevalence of active cigarette smoking among women in this study was low (2.2%). However, 19.2% of the partners were cigarette smokers. Both women’s own history and spouse’s history of smoking were significant risk factors for HPV infection in general (Table 1). Published reports were contradictory in the significance of cigarette smoking as an independent risk factor for HPV infection.^12,13,16,17 In our study, binomial logistic regression analysis showed that neither the woman nor the spouse’s cigarette smoking habits was significant independent risk factors for HPV infection. The impact of these factors was most likely confounded by the sexual behaviours.

The prevalence of HPV infection among women with 6 or less years of formal schooling was 26.6%, compared to 9.6% for women who received 7–12 years and 7.2% for those who attended more then 12 years of education (p = 0.02). This was an independent risk factor for HPV infection in our study (Table 2). This finding differed from an analysis of pooled data from 16 multicentre and multinational trials in which no significant association of lower education levels was found with higher prevalence of HPV infection.¹⁸ It is important to note a significant difference in the population between our study and the pooled data. In the pooled analysis, 66% of women had less than 5 years of formal schooling and a very small number had more than 10 years of schooling. In comparison, only 3.8% of women in our study had 6 or fewer years of formal study and a great majority (52.6%) had 12 or more years of formal education. Also, attention should be paid to the fact that the majority of women in the pooled analysis had no prior cervical cancer screening compared to our study.

There were several limitations in this study. The small sample size of the study limited the statistical power of identifying significance of distribution difference of variables between HPV-positive and HPV-negative women. Caution is needed in interpretation of exposure to the risk factors among women with HPV infection. The selection of study subjects were not based on a randomisation process in this study. The potential bias limited any conclusion to be drawn on the true prevalence of HPV infection in the population. We believe that the likelihood of an over-estimation on the prevalence of HPV infection introduced by bias on subject selection in this study was small as our results were consistent with results of similar studies elsewhere. The most important limitation of this study was the cross-sectional nature of the study which does not permit any conclusion on causality link of any of the risk factors identified in this study with HPV infection. Nevertheless, we believe that our data on the prevalence of HPV infection and distribution of HPV subtypes among women with normal cytology provide useful information, particularly in the understanding of geographic differences in the epidemiology of HPV infection.

Conclusion

Distribution of genital HPV genotypes varied between geographic regions, even within the same continent. In Singapore, despite normal cytology, almost one in ten women had detectable HPV DNA of one or more HPV subtypes, and 50% of these were high-risk HPVs. The independent risk factors associated with HPV infection were penetrative vaginal sexual intercourse, multiple sexual partners and low education level. Public health and education policies aiming at reducing HPV exposure could form an important additional means to screening and HPV vaccination in the efforts to curb the global burden of cervical cancer.

Footnotes

Acknowledgements

Shilin Chen, Pei-Zhi Ong, Tiong-Han Yeo, and staff of the Molecular Laboratory, Department of Pathology, Singapore General Hospital for their technical assistance in this study.

Conflict of interest

The authors declare no conflict of interest.

Funding

Investigator initiated study supported by Merck Sharp & Dohme (Singapore).

References

De Sanjose

Quint

Alemany

. Human papillomavirus genotype attribution in invasive cervical cancer: a retrospective cross-sectional worldwide study. Lancet Oncol 2010; 11: 1048–1056.

de Sanjosé

Diaz

Castellsagué

. Worldwide prevalence and genotype distribution of cervical human papillomavirus DNA in women with normal cytology: a meta-analysis. Lancet Infect Dis 2007; 7: 453–459.

Bruni

Diaz

Castellsagué

. Cervical human papillomavirus prevalence in 5 continents: meta-analysis of 1 million women with normal cytological findings. J Infect Dis 2010; 202: 1789–1799.

Bosch

Qiao

Castellsagué

. The epidemiology of human papillomavirus infection and its association with cervical cancer. Int J Gynecol Obstet 2006; 9(4): S8–S21.

Jemal

Bray

Center

. Global cancer statistics. CA Cancer J Clin 2011; 61: 69–90.

Dobec

Bannwart

Kaeppeli

. Automation of the linear array HPV genotyping test and its application for routine typing of human papillomaviruses in cervical specimens of women without cytological abnormalities in Switzerland. J Clin Virol 2009; 45: 23–27.

Tay

. The prevalence and significance of high-risk human papillomavirus (HPV) infection in southern Malaysia and Singapore. Aust N Z J Obstet Gynaecol 2009; 49: 323–327.

Chen

You

Hsieh

. Prevalence of genotype-specific human papillomavirus infection and cervical neoplasia in Taiwan: a community-based survey of 10,602 women. Int J Cancer 2011; 128: 1192–1203.

Dunne

Unger

Sternberg

. Prevalence of HPV infection among females in the United States. JAMA 2007; 297: 813–819.

10.

Johnson

Mercer

. Epidemiology of, and behavioural risk factors for, sexually transmitted human papillomavirus infection in men and women in Britain. Sex Transm Infect 2012; 88: 212–217.

11.

Winer

Lee

Hughes

. Genital human papillomavirus infection: incidence and risk factors in a cohort of female university students. Am J Epidemiol 2003; 157: 218–226.

12.

Louvanto

Rintala

Syrjanen

. Incident cervical infections with high- and low-risk human papillomavirus (HPV) infections among mothers in the prospective Finnish Family HPV Study. BMC Infect Dis 2011; 11: 179–179.

13.

Sellors

Karwalajtys

Kaczorowski

. Survey of HPV in Ontario Women Group. Incidence, clearance and predictors of human papillomavirus infection in women. CMAJ 200; 168: 421–425.

14.

Baay

MFD

Smits

Tjalma

WAA

. Can cervical cancer screening be stopped at 50? The prevalence of HPV in elderly women. Int J Cancer 2004; 108: 258–261.

15.

Louvanto

Rintala

Syrjänen

. Genotype-specific persistence of genital human papillomavirus (HPV) infections in women followed for 6 years in the Finnish Family HPV Study. J Infect Dis 2010; 202: 436–444.

16.

Iftner

Eberle

Iftner

. Prevalence of low-risk and high-risk types of human papillomavirus and other risk factors for HPV infection in Germany within different age groups in women up to 30 years of age: an epidemiological observational study. J Med Virol 2010; 82: 1928–1939.

17.

Cercato

Mariani

Vocaturo

. Predictors of human papilloma virus (HPV) infection in Italian women. J Med Virol 2010; 82: 1921–1927.

18.

Franceschi

Plummer

Clifford

. Differences in the risk of cervical cancer and human papillomavirus infection by education level. Br J Cancer 2009; 101: 865–870.