Article Text

Original article
Vaginal high-risk human papillomavirus infection in a cross-sectional study among women of six different ethnicities in Amsterdam, the Netherlands: the HELIUS study
  1. C J Alberts1,2,
  2. R A Vos1,
  3. H Borgdorff2,3,
  4. W Vermeulen1,
  5. J van Bergen4,5,
  6. S M Bruisten1,2,
  7. S E Geerlings2,
  8. M B Snijder6,
  9. R van Houdt7,
  10. S A Morré8,9,
  11. H J C de Vries2,10,
  12. J H H M van de Wijgert2,3,11,
  13. M Prins1,2,
  14. M F Schim van der Loeff1,2
  1. 1Department of Infectious Diseases, Public Health Service of Amsterdam (GGD), Amsterdam, The Netherlands
  2. 2Division of Infectious Diseases, Department of Internal Medicine, Center for Infection and Immunity Amsterdam (CINIMA), Academic Medical Center (AMC), Amsterdam, The Netherlands
  3. 3Amsterdam Institute for Global Health and Development (AIGHD), Amsterdam, The Netherlands
  4. 4SOA Aids Nederland, Amsterdam, The Netherlands
  5. 5Department of General Practice, Academic Medical Center (AMC), Amsterdam, The Netherlands
  6. 6Department of Public Health, Academic Medical Center (AMC), Amsterdam, The Netherlands
  7. 7Department of Medical Microbiology and Infection Control, VU Medical Center, Amsterdam, The Netherlands
  8. 8Laboratory of Immunogenetics, Department of Medical Microbiology and Infection Control, Research School V-ICI, VU University Medical Center, Amsterdam, the Netherlands
  9. 9Department of Genetics and Cell Biology, Research Institute GROW (School for Oncology & Developmental Biology), Faculty of Health, Medicine & Life Sciences, Institute for Public Health Genomics, University of Maastricht, Maastricht, The Netherlands
  10. 10Department of Dermatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
  11. 11Institute of Infection and Global Health, University of Liverpool, Liverpool, UK
  1. Correspondence to Catharina J Alberts, Public Health Service, Department of Infectious Diseases, Nieuwe Achtergracht 100, Amsterdam 1018 WT, The Netherlands; nalberts{at}ggd.amsterdam.nl

Abstract

Objective In the Netherlands the incidence of cervical cancer is higher among ethnic minority populations compared with the general Dutch population. We investigated the prevalence of, and risk factors associated with, vaginal high-risk human papillomavirus (hrHPV) infection in women of six different ethnicities living in Amsterdam.

Methods For this cross-sectional study we selected women aged 18–34 years old of six ethnicities from the large-scale multiethnic HEalthy LIfe in an Urban Setting study. Self-collected vaginal swabs were tested for HPV DNA and genotyped using a highly sensitive PCR and reverse line blot assay (short PCR fragment (SPF)10-PCR DNA enzyme immunoassay/LiPA25-system version-1, delft diagnostic laboratory (DDL)). Participants completed a questionnaire regarding demographics and sexual behaviour. Logistic regression using generalised estimating equations was used to assess risk factors of hrHPV, and to investigate whether prevalence of hrHPV differed among ethnicities.

Results The study population consisted of 592 women with a median age of 27 (IQR: 23–31) years. Dutch and African Surinamese women reported the highest sexual risk behaviour. HrHPV prevalence was highest in the Dutch (40%) followed by the African Surinamese (32%), Turkish (29%), Ghanaian (26%), Moroccan (26%) and South-Asian Surinamese (18%). When correcting for sexual risk behaviour, the odds to be hrHPV-positive were similar for all non-Dutch groups when compared with that of the Dutch group.

Conclusions We found an overall higher hrHPV prevalence and higher sexual risk behaviour in the native Dutch population. Further research is needed to unravel the complex problem concerning cervical cancer disparities, such as differences in participation in the cervical cancer screening programme, or differences in clearance and persistence of hrHPV.

  • HPV
  • ETHNICITY
  • WOMEN
  • EPIDEMIOLOGY (GENERAL)

Statistics from Altmetric.com

Introduction

Cervical cancer is the fourth most common female cancer globally; the majority of this burden is in non-Western regions.1 When women from those regions immigrate to the west, they still show higher cervical cancer incidence and related mortality when compared with the native Western population,2 ,3 despite the fact that the same healthcare is offered to all women. In the Netherlands, these health disparities are also observed:4 ,5 women born in Morocco, Surinam, Aruba/Netherlands Antilles and Indonesia have a higher cervical cancer incidence compared with the Dutch population. Possible reasons for these differences, which are not mutually exclusive, are (A) differences in (sexual) risk behaviour;6–9 (B) differences in participation in the cervical cancer screening programme;10 ,11 (C) differences in high-risk human papillomavirus (hrHPV) prevalence,12 the distribution of hrHPV types13 ,14 or clearance of hrHPV;15 and/or (D) a reflection of the higher incidence of cervical cancer in their home country.5 ,16

In the Netherlands approximately 12% of the population is of non-Western origin and this group is expected to grow in the coming years.17 Data on prevalence and determinants of HPV infection among these groups are scarce and the few studies conducted lack the power to show associations.18–20 Our objective was to map the prevalence of hrHPV infection among ethnic minorities and Dutch women aged 18–34 years. In this way we hope to provide insight into why the incidence of cervical cancer is higher among ethnic minorities compared with the Dutch. Therefore, we selected from the multiethnic population-based HEalthy LIfe in an Urban Setting (HELIUS) study women with a Dutch, South-Asian Surinamese, African Surinamese, Ghanaian, Moroccan or Turkish origin living in Amsterdam, the Netherlands. By assessing the prevalence of vaginal hrHPV infections in these women and relating this to their sexual behaviour and other characteristics, we address in this cross-sectional study the following questions: (1) What is the reported sexual risk behaviour? (2) What is the prevalence of hrHPV? (3) Which hrHPV types are most common? (4) What are risk factors for being hrHPV-positive? (5) What is the trend of hrHPV prevalence by age? and (6) Does the prevalence of hrHPV infections differ between the six ethnic groups after adjustment for important risk factors?

Methods

Study population

Female participants were selected from the HELIUS study, a large-scale multiethnic cohort study being carried out in Amsterdam, the Netherlands.21 In brief, participants of Dutch, Surinamese, Ghanaian, Moroccan and Turkish origin aged 18 years and 70 years were randomly selected from the municipality registry of Amsterdam. Of the Surinamese immigrants in the Netherlands, approximately 80% are of either African origin (from West Africa) or South-Asian origin (from North India), and the remaining 20% are of Javanese, Chinese or other origin. Surinamese subgroups were classified according to self-reported ethnic origin. In this particular study, the South-Asian Surinamese group also included participants of Javanese origin. A participant was considered as of non-Dutch ethnic origin if she fulfilled one of the following criteria:22 (A) born outside the Netherlands and has at least one parent who was born outside the Netherlands (first generation) or (B) born in the Netherlands and at least one of the parents was born outside of the Netherlands (second generation).

At the time of this study, recruitment of participants into HELIUS was still ongoing. We selected participants recruited from January 2011 through December 2013. We randomly selected up to seven (if available) female participants per life year, in the age range of 18–34 years (=17 years), from each ethnic group, resulting in a maximum of 119 (=7×17) participants per ethnic group. Participants were eligible if they provided a vaginal sample and completed questions on their sexual behaviour. Participants who indicated to be vaccinated with a prophylactic HPV vaccine were excluded.

Data and sample collection

Detailed information regarding sociodemographic characteristics (age, education, civil status, religion, self-identified ethnicity in Surinamese participants), health-related characteristics (HPV vaccination status, ever used oral contraceptives and self-reported genital warts/ulcers/vesicles in the past), substance use (smoking and alcohol use) and sexual behaviour (ever had sex, currently in a steady relationship, age at sexual debut, number of lifetime male sex partners (LMSP), sexual activity in the preceding 6 months and condom use in the preceding 6 months during sex with a steady or casual partner) were obtained via standardised questionnaires. All female participants were asked to self-collect a vaginal swab (Copan Diagnostics, Murrieta, California, USA).

HPV testing

The self-collected vaginal swabs were kept at 2–8°C for up to 5–6 days, shipped to the Public Health Laboratory and then stored at −20°C. Selected swabs for this study were tested for Trichomonas vaginalis , Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) (Aptima Combo 2 for CT/NG assay and the Aptima Trichomonas vaginalis assay by PANTHER System, HOLOGIC (Bedford, USA)) and tested for HPV (SPF10-PCR-(DNA enzyme immunoassay)/DEIA/LiPA25 system version-1, DDL, Voorburg, the Netherlands) according to protocol and as described earlier.23 ,24 HPV types were categorised as indicated in earlier studies.24 ,25

Statistical analyses

Pearson's χ2 test was used to compare categorical data and Kruskal–Wallis test to compare continuous data across ethnic groups. The main outcome of interest was hrHPV positivity. Age and sexual debut were categorised based on quartiles and LMSP was categorised based on risk categories of interest. For logistic regression analyses, generalised estimating equation (GEE) models with exchangeable correlation structure were used to account for multiple HPV infections within the same participant;26 GEE models take account of clustered data as one participant is at risk for different HPV types at the same time. Statistical analyses were performed using Stata software V.13.1 (Stata Intercooled, College Station, Texas, USA).

Risk factor analyses and hrHPV by age

Bivariable logistic regression (using GEE) was used to investigate risk factors associated with hrHPV in each ethnic group. Subsequently, we assessed in an overall model whether these risk factors for hrHPV positivity differed between ethnicities by testing whether the interaction between ethnicity and the individual risk factors was statistically significant (p<0.05). We estimated the probability of hrHPV infections as a function of age using a restricted cubic spline with four knots that was allowed to differ by ethnic group.

Association between hrHPV and ethnicity

To study the association between hrHPV and ethnicity we conducted multivariable logistic regression analyses (using GEE). We controlled for risk factors by including a priori defined variables. We ran six different models. Models 1, 2 and 3 included all 592 female participants of the study population and models 4, 5 and 6 were restricted to 476 out of the 592 female participants who reported to ever have had sex with a man. The reason to perform separate analyses for those who reported to ever have had sex with a man is to enable adjusting for more than one sexual risk variable. Models 1 and 4 show the crude association between hrHPV and ethnicity; models 2 and 5 are adjusted for age, education and civil status; model 3 is adjusted for age, education, civil status and LMSP; while model 6 is adjusted for age, education, civil status, age of sexual debut, LMSP and sexual activity in the preceding 6 months. Risk factors were chosen to adjust for (behavioural) differences between ethnicities.

Results

Participants versus non-participants

During the period of January 2011 through December 2013, a total of 10 257 participants were enrolled in the HELIUS study; 1469 of these participants were women aged 18–34 years. In total 980 women aged 18–34 years agreed to provide a vaginal sample and 489 declined (see online supplementary figure S1). Dutch (94%), African Surinamese (87%), South-Asian Surinamese (75%) and Ghanaian (71%) women provided more often a vaginal swab compared with Moroccan (49%) and Turkish (37%) women (p<0.001). Turkish and Moroccan women who declined to provide a vaginal sample reported higher sexual risk behaviour than those who did provide a vaginal swab (data not shown).

Supplemental material

Study population

Of the 980 eligible female participants with a vaginal sample available, 610 were randomly selected. From those selected we excluded 10 participants because they indicated to have been vaccinated against HPV and 8 participants because of invalid HPV results (see online supplementary figure S1). Baseline characteristics of the 592 included participants stratified by ethnicity are shown in table 1. Overall, the median age was 27 years (IQR: 23–31) and was similar between ethnic groups (p=0.353). Most characteristics differed significantly between the ethnic groups. For example, women of Dutch origin more often had had a higher education (70%), were non-religious (92%), had ever used oral contraceptives (89%) and had used alcohol in the preceding year (94%) when compared with other ethnicities. Sexually transmitted infection (STI) diagnosed at visit did not differ significantly (p=0.070) between ethnicities, however the highest prevalence was found in African Surinamese women (8% STI positive, mainly CT). Almost all Dutch (95%) and African Surinamese (93%) reported to ever have had sex, whereas in the South-Asian Surinamese (82%), Ghanaian (69%), Moroccan (65%) and Turkish (77%) group, fewer indicated to ever have had sex with a man or a woman (p<0.001). Among those who reported to ever have had sex, sexual behaviour differed significantly between ethnicities: for example, Dutch and African Surinamese women reported a younger sexual debut, a higher number of LMSP, and more often having sex with a casual partner or with both a casual and a steady partner, in comparison to women of other ethnicities.

Table 1

Baseline characteristics of 592 female HELIUS participants aged between 18 years and 34 years who provided a vaginal swab and completed the questionnaire, by ethnicity, the HELIUS study Amsterdam, the Netherlands

HPV prevalence

HPV prevalence by ethnicity is presented in table 2 and a graphical representation of the distribution of the individual hrHPV types by ethnicity is presented in figure 1A. In total 280/592 (47%) of samples from women were PCR-DEIA positive, of which 42 were untypable. Any HPV (53%), hrHPV (42%) and low risk HPV (lrHPV) (31%) prevalence were highest among the Dutch. HPV type 16 (HPV-16) and HPV type 18 (HPV-18) were detected among all groups, but were most common among Turkish (HPV-16; 13%) and Dutch (HPV-18; 10%) women, respectively. The prevalence of individual hrHPV types did not differ significantly between ethnicities except for HPV-18 (p=0.006).

Table 2

HPV prevalence in 592 randomly selected female HELIUS participants aged between 18 years and 34 years, who provided a vaginal swab and completed the questionnaire, by ethnicity, the HELIUS study Amsterdam, the Netherlands

Figure 1

(A) Individual high-risk human papillomavirus (hrHPV) types as a percentage of the total number of detected high-risk types (indicated with n), by ethnicity. Samples were taken from self-collected vaginal swabs of 592 female HELIUS participants aged between 18 years and 34 years who had completed the questionnaire of the HELIUS study Amsterdam, the Netherlands. Note that one person can be infected with ≥1 hrHPV types. (B) Estimated probability of hrHPV infections as a function of age in years in women from six ethnicities: Dutch, South-Asian Surinamese, African Surinamese, Ghanaian, Moroccan and Turkish. Expected hrHPV prevalence is derived from a four-knot restricted cubic spline using default knot values (dashed line). The grey shading depicts the 95% CI for expected high-risk prevalence. Dots represent the observed prevalence of hrHPV for each age. The trend of hrHPV by age did not differ statistically significantly among ethnicities (p=0.796). HELIUS, HEalthy LIfe in Urban Setting.

Risk factor analyses and hrHPV by age

We did not find any sociodemographic characteristics, health-related characteristics or substance use characteristics that were significantly associated with hrHPV positivity in all ethnicities (for details see online supplementary tables S1 and S2). HrHPV positivity was significantly associated with a higher number LMSP in the Dutch (p=0.006), South-Asian Surinamese (p=0.005), Moroccan (p=0.001) and Turkish (p=0.002) groups. Odds for hrHPV positivity was the highest among those reporting sex with a casual only and a casual and a steady partner, but this was only statistically significant in the Dutch, South-Asian Surinamese and Turkish groups. No other sexual risk variables were found to be associated in almost all ethnic groups. No statistically significant interactions (all p>0.05) between ethnicity and any of the risk factors under investigation were found (data not shown). Figure 1B shows hrHPV prevalence by age for each ethnicity. Prevalence of hrHPV was highest around the age of 25–30 years and lower in higher ages; this trend did not differ significantly between ethnicities (p=0.796).

Association between hrHPV and ethnicity

Table 3 shows the association between ethnicity and hrHPV using GEE. In the full population (n=592) we observed that African Surinamese women did not significantly differ from Dutch women, whereas South-Asian Surinamese, Ghanaian and Moroccan women had a lower odds of being hrHPV-positive than Dutch women. When adjusting for non-sexual behavioural characteristics (ie, age, education and civil status) similar results were obtained, however the odds for African Surinamese women of being hrHPV-positive were now significantly lower when compared with the Dutch women (model 2). When we also adjusted for number of LMSP the differences between ethnicities were no longer significant (model 3). When restricting the analyses to female participants who reported to ever have had sex with a man, we observed similar results as in the full study population.

Table 3

Association between high-risk HPV and ethnicity using logistic regression with generalised estimating equations, among female HELIUS participants aged between 18 years and 34 years who provided a vaginal swab and completed the questionnaire, by ethnicity, the HELIUS study Amsterdam, the Netherlands

Discussion

In the Netherlands, incidence of cervical cancer is reported to be higher among ethnic minorities when compared with the general population.4 ,5 These differences might be caused, for example, by differences in prevalence of hrHPV infections or sexual behaviour. We found that among this cohort of young women aged 18–34 years, both hrHPV prevalence and sexual risk behaviour were overall highest in women with a Dutch origin. Furthermore, we did not find indications that the effects of risk factors for hrHPV differed significantly between ethnicities. To further understand the complex web of factors underlying the differences in cervical cancer incidence, we formulated six key questions that the data allowed us to address.

The first two questions were: (1) What is the sexual risk behaviour reported? and, (2) What is the prevalence of hrHPV? We found that sexual risk behaviour differed significantly between ethnicities with the Dutch and African Surinamese groups reporting the highest sexual risk behaviour. The hrHPV prevalence was highest in the Dutch group, followed by the non-Dutch groups. We did not expect that the groups at highest risk for cervical cancer would show the lowest overall hrHPV prevalence. There are several possible interpretations for these results. For instance, the prevalence of hrHPV in the current (younger) study population may be different from those in the women in which differences in cervical cancer incidence by ethnicity were observed in the past (older age group), and therefore perhaps may not explain the disparities observed among these ethnicities. Another explanation could be that other factors play a role in the development of cervical cancer, for example, differences in participation in cervical cancer screening, higher exposure at later age, or differences in clearance and persistence of hrHPV.

Next, we assessed which hrHPV types were most common in the six ethnic groups (question 3). We found that the prevalence of individual hrHPV types did not differ significantly between ethnicities, except for HPV-18. Although the prevalence of other hrHPV types was not significantly different, figure 1A suggests that among women with African ancestors (African Surinamese, Ghanaian and Moroccan) HPV-16 and HPV-18 do not comprise the largest fraction of all detected types, as suggested previously.13 ,14 It is therefore worthwhile considering whether the recently Food and Drugs Administration (FDA) approved 9-valent vaccine27 ,28 would be an asset in the efforts to control cervical cancer in Dutch ethnic minorities. However, a recent study from the USA found that only a small additional reduction in HPV-associated cancer cases is expected when using the new 9-valent vaccine across all ethnicities; nevertheless in some ethnic groups a larger additional reduction is expected for cervical cancer in situ and oropharyngeal cancers.29

What are the risk factors for being hrHPV-positive (question 4)? Our findings are in line with previous studies;6 ,7 ,30 sexual risk behavioural characteristics are associated with hrHPV positivity. However, in contrast to previous studies,8 ,31 we did not observe a clear association between hrHPV positivity and age of sexual debut in any of the ethnicities, which may be caused by lack of power. We also did not find indications that the effects of risk factors for hrHPV or the trend in hrHPV positivity by age (question 5) differed significantly between ethnicities.

The sixth and last question is whether the prevalence of hrHPV infections differed between the six ethnic groups after adjustment for important risk factors. We conclude that differences in hrHPV positivity are caused by differences in sexual risk behaviour between ethnicities. However, we also observe that among Moroccan and Turkish women the OR changed from <1 to >1. This change of direction is due to the fact that Moroccan and Turkish women with low sexual risk behaviour have higher hrHPV prevalence than Dutch women with similar reported sexual risk behaviour. There are three possible reasons for this higher prevalence: (1) Moroccan and Turkish women have a higher risk of becoming hrHPV-positive than Dutch women when practising the same sexual behaviour, (2) Moroccan and Turkish women more often have non-penetrative sex when reporting to never have had sex or (3) social desirability bias may differ between ethnicities.

This study has some limitations. Although HELIUS participants were recruited from the general population of Amsterdam, and our study population was randomly selected from the HELIUS population, there may be selection bias in our study population. First, participants of the HELIUS study may be more concerned about their health, and second, less than half of the Turkish and Moroccan HELIUS participants agreed to provide a vaginal swab. Therefore, we compared our results on sexual behaviour with the national Dutch Public Health database32 and observed similar patterns of reported sexual risk behaviours. Another limitation of this study is the cross-sectional nature that prevents investigation of differences in incidence and persistence of hrHPV infection.

In our study in Amsterdam we found higher hrHPV prevalence than observed in previous studies in other parts of the Netherlands with a hrHPV prevalence of 12% (ages 18–29 years)30 and 11% (ages 18–34 years).33 A self-sampling study investigating hrHPV prevalence in various ethnic groups showed prevalence of 9% in ethnic-Dutch, 11% in the Surinamese, 8% in Moroccan and 8% in Turkish women.20 This suggests that our population concerns a higher-risk group.

The ethnic minority populations in the Netherlands, and especially in large cities such as Amsterdam, are increasing. Here we report that hrHPV prevalence and sexual risk behaviour were highest in the Dutch group, indicating that other factors must be responsible for the higher incidence of cervical cancer in ethnic minorities. These findings emphasise the need for further research to unravel the complex problem concerning cervical cancer disparities such as differences in participation in the cervical cancer screening programme or differences in clearance and persistence of hrHPV prevalence.

Key messages

  • We investigated the prevalence of, and risk factors associated with, vaginal high-risk human papillomavirus (hrHPV) infection in different ethnicities in Amsterdam.

  • hrHPV prevalence and sexual risk behaviour were overall lower among women of non-Dutch ethnicities compared with Dutch women.

  • Further research is needed to explain cervical cancer disparities among ethnic minorities.

Acknowledgments

The authors thank the participants of the HELIUS study and the management team, research nurses, interviewers, research assistants and other staff who have taken part in gathering the data of this study. The authors also thank the AMC Biobank for their support in biobank management and high-quality storage of collected samples. The authors also thank Ronald Geskus for his critical review on the statistical analyses.

References

Supplementary materials

  • Abstract in Dutch

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Handling editor Jackie A Cassell

  • Contributors MBS and MP conceived and designed the HELIUS study and obtained funding. WV performed laboratory analyses of the samples. WV, CJA, RAV and HB organised sample logistics. CJA and HB participated in data collection. All authors participated in designing the vaginal swab study within HELIUS and/or laboratory assay development. CJA and RAV performed data management and statistical analyses. CJA wrote the manuscript. MFSvdL (principal investigator) contributed to the design, statistical analyses and writing the manuscript. All authors contributed to data interpretation, reviewed successive drafts and approved the final version of the manuscript.

  • Funding The HELIUS study is conducted by the Academic Medical Center Amsterdam and the Public Health Service of Amsterdam. Both organisations provided core support for HELIUS. The HELIUS study is also funded by the Dutch Heart Foundation (2010T084), the Netherlands Organization for Health Research and Development (ZonMw) (200500003), and the European Union (FP-7) (278901). This study was funded by grant 7115 0001 and 204005002 (Academic Collaborative for Public Health) from ZonMw and an internal grant of the Public Health for Research & Development, the Netherlands.

  • Competing interests MFSvdL received research funding from Sanofi Pasteur MSD; he is a co-investigator in a Merck-funded investigator-initiated study on Gardasil; he is an investigator on a Sanofi Pasteur MSD sponsored HPV vaccine trial; he served on a vaccine advisory board of GSK; he received in-kind contribution for another study from Stichting Pathologie Onderzoek en Ontwikkeling (SPOO); his institution receives research funding from Janssen Infectious Diseases and Vaccines.

  • Ethics approval The study was approved by the Medical Ethics Committee of the Academic Medical Center Amsterdam (protocol number: 10/100; amendment 10/100# 10.17.1729; NL32251.018.10).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.