Article Text

Download PDFPDF
The importance of low-risk HPV infection for the risk of abnormal cervical cytology/histology in more than 40 000 Danish women
  1. Ann Nielsen1,
  2. Thomas Iftner2,
  3. Mia Nørgaard1,
  4. Christian Munk1,
  5. Jette Junge3,
  6. Susanne Krüger Kjaer1,4
  1. 1Unit of Virus, Lifestyle and Genes, Danish Cancer Society Research Center, Copenhagen, Denmark
  2. 2Medical Virology, University Hospital of Tuebingen, Tuebingen, Germany
  3. 3Department of Pathology, Hvidovre Hospital, University of Copenhagen, Denmark
  4. 4Juliane Marie Center, Gynecologic Clinic, Rigshospitalet, University of Copenhagen, Denmark
  1. Correspondence to Professor Susanne Krüger Kjaer, Unit of Virus, Lifestyle and Genes, Danish Cancer Society Research Center, DK-2100 Copenhagen, Denmark; susanne{at}cancer.dk

Abstract

Objectives To estimate the age and type-specific distribution of low-risk (LR) human papillomavirus (HPV) types in cervical samples from women in the general population and to assess the distribution of LR-HPV without the coexistence of high-risk HPV types in different cytology and histology categories.

Methods In a cross-sectional study, liquid-based cytology samples (SurePath) were collected over a 3-year period. The samples were HPV tested by Hybrid Capture II (HC2; Digene) and genotyped using a PCR-based assay (INNO-LiPAv2; Innogenetics Inc.). A total of 40 382 women (14–95 years of age) were included in the study. By linkage with the nationwide Pathology Data Bank, the HPV test results were directly linked to cytological diagnoses made from the same samples and to subsequent histology results.

Results Overall, 2790 women (6.9%) tested positive for LR-HPV types, with HPV6 and HPV70 being the most frequent types detected, whereas HPV11 was uncommon. The highest prevalence was observed in the youngest age group (≤19 years). The LR-HPV prevalence was 6.3% in women with normal cytology, 33.1% in atypical squamous cells of undetermined significance (ASCUS), 19.6% in low-grade squamous intraepithelial lesion and 12.7% in those with high-grade squamous intraepithelial lesion. When considering women with LR-HPV alone, the prevalence was 2.0% (normal cytology), 11.3% (ASCUS), 2.6% (low-grade squamous intraepithelial lesion) and 0.7% in women with high-grade squamous intraepithelial lesion, respectively. A similar pattern was observed in relation to the histological diagnoses with the majority of LR-HPV infections detected in CIN1 lesions (24.7%).

Conclusions LR-HPV types alone are relatively common in ASCUS, whereas LR-HPV types without coexisting high-risk HPV types are virtually never detected in severe cervical lesions.

  • Human papillomavirus
  • prevalence of low-risk types
  • association with cervical lesions

Statistics from Altmetric.com

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.

Introduction

Human papillomavirus (HPV) infection is the most common sexually transmitted infection among women worldwide with an estimated 291 million women with detectable HPV DNA.1 ,2 HPV infection is the necessary cause of invasive cervical cancer and its precursors.3 ,4 To date more than 100 HPV types have been identified of which around 40 are found to infect the anogenital region. Based on their oncogenic potential, they are grouped into high-risk (HR) and low-risk (LR) HPV types.5

In female populations, the prevalence of HR-HPV types is in general higher compared with LR-HPV types. In Europe, nearly 4% of the sexually active women aged 15–74 years are HR-HPV positive, whereas only around 2% of the women are LR-HPV positive.6 Invasive cervical cancer and its high-grade precursors are usually associated with the HR-HPV types, whereas benign lesions such as genital warts are more frequently associated with the LR-HPV types, notably HPV6 and HPV11. Interest has primarily been focused on HR-HPV types6–8 and the associated lesions, and there has been less focus on the LR-HPV types and lesions caused by these HPV types. With the introduction of the quadrivalent vaccine containing not only the HR-HPV types 16 and 18 but also the LR-HPV types 6 and 11, it has become important to examine the prevalence and the disease burden of these two LR-HPV types. In addition, it is important also to assess the prevalence and disease burden due to other LR-HPV genotypes and thereby establishing a baseline before the introduction of HPV vaccination.

We have previously reported only on the overall LR-HPV prevalence and type distribution in a smaller number of women (∼11 000 women).9 The aim of the present paper is to assess the occurrence of LR-HPV types (determined by hybrid Capture 2 (HC2) (Qiagen, Hilden, Germany)) and type-specific distribution of LR-HPV types (determined by genotyping (LiPAv2-test) (Innogenetics Inc., Gent, Belgium) of HC2-positive samples) in cervical samples from a large cohort of women (∼40 000) from the general Danish population. Furthermore, it is the aim to examine the frequency of cervical lesions in women exclusively being positive for LR-HPV types. This will provide potentially important information for the evaluation of the risk of squamous intraepithelial lesions (SIL) caused by LR-HPV types in women before introduction of HPV vaccination, and in addition, it will help to predict the potential effect of HPV vaccination programme against SIL caused by HPV6 and HPV11.

Materials and methods

From 2002 to 2005, we performed a population-based study of prevalent HPV infections in Copenhagen, Denmark, by collecting on random days consecutive liquid-based cytology (LBC) samples from one department of pathology (Department of Pathology, Copenhagen University Hospital, Hvidovre, Denmark), which covers all cervical cytology and subsequent histopathology in the Greater Copenhagen area. The women were included by either opportunistic screening or through the cervical screening programme that includes women between 23 and 65 years.10 All the samples were taken by general practitioners or gynaecologists using the SurePath LBC system. Within 2 days of collection, the samples were sent to the pathology department where they were marked with a microscopy number before they were analysed by certified cytotechnicians/cytopathologists and given a diagnosis. In the event of a subsequent biopsy or conization, the tissue was also sent to the same department for examination.

After the cytological examination, the remaining material was collected on a weekly basis by Unit of Virus, Lifestyle and Genes at the Danish Cancer Society, and within 2–3 weeks after the LBC was taken, the samples were sent to University of Tuebingen, Germany, for HPV testing. The study was approved by the Danish Data Protection Agency and the Scientific Ethical Committee (KF) 01-059/03.

In total, we collected 42 854 LBC samples. A total of 184 samples were excluded due to technical errors or missing/insufficient identification on the samples. We also excluded 2271 samples because they were duplicate samples from the same women (only the first one was included in the study). Finally, 17 samples were excluded due to missing information on HC2 result, HPV typing result or cytology. Consequently, a total of 40 382 women were included in the analysis.

The Pathology Data Bank is a Danish national pathology register, which operates through a real-time system. It contains information on all the cervical cytology (organised and opportunistic, normal and abnormal) and histology (biopsies, conizations and hysterectomies) samples collected in Denmark. With the use of the study number and the personal identification number, a unique identifier assigned to all Danish residents, we linked the cohort with the Pathology Data Bank and retrieved information on all the women in the study about the cytology result (normal atypical squamous cells of undetermined significance (ASCUS), low-grade SIL (LSIL), high-grade SIL (HSIL) of the index sample and on any subsequent histological examination (normal cervical intraepithelial lesion (CIN) of different grades) related to the index cytology.

All the samples were only marked with the study number. The specimens were then tested for the presence of HPV by the HC2 method detecting 13 HR (16,18,31,33,35,39,45,51,52,56,58,59,68) and 5 LR (6,11,42,43,44) HPV types, which are used to prepare a high (B) and low (A) probe mix that are used in two separate reactions. We used the Food and Drug Administration-recommended cut-off value for test-positive results 1.0 RLU/Co (equivalent to 1 pg HPV DNA per 1 ml of sampling buffer). When using this recommendation, the sensitivity of the HC2 test is above 95%.11 The samples positive for HPV were then analysed for specific genotypes by the LiPAv2 test which can identify 24 HPV genotypes (6,11,16,18,31,33,35,39,40,42,43,44,45,51,52,53,54,56,58,59,66,68,70,74). Those samples which were LiPA positive, but where no specific HPV type was identified, were classified as HPV of undetermined type.

We estimated the overall and the type-specific prevalence of LR-HPV types with 95% CI for all women and by age. Furthermore, we assessed the overall prevalence of LR-HPV infections according to cytological and histological diagnosis. We assessed both the prevalence of overall LR-HPV infections (with or without coexisting HR-HPV types) and the prevalence of HPV infections with only LR types. In women with LR-HPV types only, we estimated the prevalence of single-type infections, multiple-type infections as well as the type-specific prevalence of all LR-HPV types separately. Finally, we assessed the associations between LR-HPV types and prevalent cytological and histological abnormalities by means of age-adjusted logistic regression (SAS/STAT V.Windows 9.1). The associations were expressed as ORs with 95% CI. The HPV types 6, 11, 40, 42, 43, 44, 54, 70 and 74 were categorised as LR-HPV types.

Results

Of the 40 382 women included in the study, 9644 (23.9%; 95% CI 23.5 to 24.3) were HPV positive by HC2 (HR- and/or LR-HPV). Overall, 2790 women (6.9%; 95% CI 6.7 to 7.2) were HC2 positive for LR-HPV, of those were 1382 women (3.4%; 95% CI 3.2 to 3.6) positive only for the HC2 LR-HPV probe (table 1). The prevalence of LR-HPV was highest among the youngest women, 14–19 years old (26.3%; 95% CI 16.6 to 35.9). The prevalence then rapidly declined until the age group of 35–44 years (4.5%; 95% CI 4.1 to 4.9) and subsequently no significant change in prevalence was found, the prevalence among the oldest women (≥65 years) being 3.1% (95% CI 2.1 to 4.1).

Table 1

Prevalence of specific low-risk HPV types

In table 1, also the prevalence of the specific LR-HPV types is presented. In this study where genotyping was performed in women positive by HR-HC2 and/or LR-HC2, the most common LR-HPV types were HPV6 (1.5%; 95% CI 1.4 to 1.7) and HPV70 (1.5%; 95% CI 1.4 to 1.6) followed by HPV44 (1.4%; 95% CI 1.3 to 1.5). In contrast, HPV11 and HPV40 were uncommon with an overall prevalence of 0.4% (95% CI 0.3 to 0.4) and 0.3% (95% CI 0.3 to 0.4), respectively. For all the LR-HPV types, we observed a decreasing prevalence with increasing age (table 1).

In table 2, the prevalence of specific LR-HPV types in relation to the cytology result is shown. Among women with normal cytology, the overall prevalence of LR-HPV types was 6.3% (95% CI 6.1 to 6.6), whereas among women with a diagnosis of ASCUS, LSIL or HSIL, the prevalence was 33.1% (95% CI 25.8 to 40.4), 19.6% (95% CI 17.6 to 21.7) and 12.7% (95% CI 10.3 to 15.1), respectively. Similarly, the prevalence of infection with LR-HPV with coexisting HR-HPV types and infection with LR-HPV infections alone was decreasing with increasing severity of the lesion. In other words, LR-HPV types were significantly more prevalent in women with ASCUS than in women with more severe lesions. In women harbouring only LR-HPV types, single-type infections were significantly more frequent than multiple LR-HPV infections in women with normal cytology as well as in women with different grades of cytological abnormalities. Both in single-type infection and multiple LR-HPV types, the highest prevalence of LR-HPV infections was detected in ASCUS (10.6%; 95% CI 5.9 to 15.4 and 0.6; 95% CI 0.0 to 1.9), whereas virtually no LR-HPV types alone were detected in HSIL. HPV6 was detected as the most frequent LR-HPV type (10.6%; 95% CI 5.9 to 15.4) in women with ASCUS. The prevalence of HPV6 and/or HPV11 in women with ASCUS was 11.6% (95% CI 6.4 to 16.2) (data not shown).

Table 2

Low-risk HPV infection by cytological and histological diagnosis

Table 2 also displays the prevalence of LR-HPV types according to histological diagnosis. Both in women diagnosed with normal histology and CIN not otherwise specified (NOS), the overall prevalence of LR-HPV infection was 15.1%. In women with CIN1, CIN2, CIN3 and cervical cancer, the prevalence decreased from 24.7% (95% CI 19.8 to 29.7) in CIN1 to 7.4% (95% CI 0.0 to 17.3) in women diagnosed with cervical cancer. For LR-HPV infections with concurrent HR-HPV types, the prevalence according to the histological lesions showed the same pattern. This pattern was even more pronounced for LR-HPV infections only; the prevalence in CIN1 was 2.1 (95% CI 0.4 to 3.8), whereas practically no LR-HPV infections alone were detected in CIN2 (0.5; 95% CI 0.0 to 1.6) and CIN3 (0.4%; 95% CI 0.0 to 1.0). None of the women diagnosed with cervical cancer had LR-HPV types alone. In relation to CIN1 lesions, the most prevalent LR-HPV types were HPV6 followed by HPV70.

In table 3, the age-adjusted associations between infection with LR-HPV types (with and without HR-HPV coinfection) and different grades of cytological diagnoses are shown. Compared with HPV-negative women, a strongly increased risk for ASCUS was seen in both women infected with LR-HPV only (OR=33.4; 95% CI 17.7 to 63.3) and for those infected with LR-HPV with concurrent HR-HPV (OR=27.4; 95% CI 15.5 to 48.5). In contrast, the age-adjusted associations between LR-HPV alone and LSIL (OR=4.2; 95% CI 3.0 to 5.9) or HSIL (OR=2.7; 95% CI 1.1 to 6.8) were much less pronounced.

Table 3

Age-adjusted association between HPV status and cytological and histological abnormalities

Table 3 also illustrates the association between LR-HPV infections and histological abnormalities. The risk for CIN1 or CIN NOS was increased in women with LR-HPV alone (OR=7.1; 95% CI 3.2 to 16.2) compared with HPV negative, but to a much lesser extent than when there was coinfection with HR-HPV types (OR=29.8%; 95% CI 19.4 to 45.8). Women infected with a single LR-HPV type tended to have a lower risk for CIN1/CIN NOS compared with women with multiple LR-HPV types. When looking at the more severe lesions (CIN2+), the association with LR-HPV types (alone) was somewhat weaker (OR=4.9; 95% CI 1.5 to 16.6), and the strongest association was found with LR-HPV infections with concurrent HR-HPV (OR=69.7; 95% CI 43.0 to 113.1)

Discussion

In this study, involving more than 40 000 Danish women from the general population of greater Copenhagen, the prevalence of LR-HPV infections was 6.9%. The LR-HPV types were relatively common in women diagnosed with ASCUS, whereas virtually no LR-HPV types were present without concomitant HR-HPV types in the severe cervical lesions.

Our result regarding the overall prevalence is largely in agreement with findings from some other published studies examining the prevalence of LR-HPV types,12–15 whereas the LR-HPV prevalence was higher in some other studies16–18 and in some studies from Asia, the LR-HPV prevalence was lower.19 ,20 However, comparisons may be problematic since the study populations differ and maybe more importantly the age distributions and the HPV testing method in the different studies also differ. Furthermore, in the present study, only HC2-positive samples were subsequently genotyped (LiPA). This has led to a lower LR-HPV prevalence than if all women had been tested with LiPA, as this test is more sensitive than HC2 test.21 Finally, there is a great variation in the number of HPV types included in the definition of LR HPV types in the different studies.

We found a peak prevalence of LR-HPV infections among the youngest age groups followed by a decline with increasing age. In contrast, we do not see a second peak in the older women as reported by Herrero et al.17 In the present study, the peak prevalence detected in the youngest women may be slightly overestimated as the youngest girls (≤23 years) are not part of the organised screening programme, and therefore, they may be more likely to have a specific reason for seeing a gynaecologist (eg, irregular bleeding or pain). Nevertheless, the high prevalence in the youngest women is concordant with the knowledge that most HPV infections occur soon after sexual debut.22

Regarding the distribution of the specific LR-HPV types in the study population, we found that HPV6 and HPV70 followed by HPV44 were the most frequent LR-HPV types. However, the prevalence data should be interpreted with some caution as we were only able to genotype the samples being positive to HC2, implying that we will underestimate the prevalence of LR-HPV, in particular the four LR-HPV types not in the LR-HC2 test (HPV40, 54, 70, 74), overall but especially for infections occurring alone.

Overall, HPV11 on the other hand was rare and only detected in <1% of the women. The higher frequency of HPV6 compared with HPV11 was also observed in women with cytological or histological abnormalities. Our findings related to HPV6 and HPV11 are in line with other studies among women in general13 ,15–18 ,23 ,24 and among women with abnormal cytology or histology.25–27

Similar to previous studies, we observe the highest prevalence of all LR-HPV infections among women with low-grade lesions (ASCUS, LSIL, CIN1), whereas in women with the more severe lesions, the prevalence of LR-HPV was substantially lower.12 ,17–19 25–28 This pattern was even more evident for LR-HPV infections without coinfection with HR-HPV types, where virtually no women with high-grade lesions had LR-HPV types alone.12 ,25–28 Compared with some other published data, the prevalence of LR-HPV types found in our study among women with cytological and histological abnormalities tended to be a little lower.12 ,18 ,25–27 One explanation could be that the specific LR-HPV types included in the LR-HPV group differ between studies. For example, some studies included HPV53 and HPV66 in the LR-HPV group.12 ,18 ,25–27 As these HPV types were rather prevalent, this may partly explain the higher prevalence of LR-HPV types found in those studies. HPV53 and HPV66 genotypes have now been classified as potentially carcinogenic,5 ,29 and based on this, we chose not to include them in the LR-HPV group in our study.

We also examined the age-adjusted association between LR-HPV infections and different grades of cytological and histological abnormalities. We found, not surprisingly, that women infected only with LR-HPV types had significantly lower risks compared with LR-HPV infections with coexisting HR-HPV for all cervical lesions except for ASCUS, where there was no difference whether the LR-HPV infection was with or without coinfection with HR-HPV. This supports the large body of evidence pointing to the importance of HR-HPV types for the development of cervical high-grade lesions.5 It also emphasises the importance of the LR-HPV types for the less severe lesions such as ASCUS. Although not statistically significant, multiple LR-HPV infections tended to be associated with a higher risk of histological abnormalities than single LR-HPV infections. This has also been found in other studies.17 However, as multiple HPV infections share the same sexual transmission route and as no HPV test is 100% sensitive, the tendency towards a higher risk in women with multiple LR-HPV types may be explained by the presence of HR-HPV types not yet detected.17

Attention has been focused on oncogenic HPV types and high-grade cervical lesions including cervical cancer; however, also the low-grade cervical lesions are important to consider as they outnumber the more severe lesions, and they imply that a large number of women are examined and followed in the triage of such lesions. In Denmark, every year approximately 15 000 women are diagnosed with ASCUS or LSIL.30 In our study population, a total of 4.4% and 1.1% of the women diagnosed with ASCUS and LSIL were infected with HPV6 and/or HPV11, respectively. Since 2009, a vaccination programme using the quadrivalent HPV vaccine (including HPV types 6, 11, 16 and 18) has been in place in Denmark. The programme includes vaccination of all 12-year old girls and had catch up vaccination of girls 13–15 years during the first 2 years, and it has been suggested that women up to the age of 26 years should be included in the vaccination programme in the near future. Given that the vaccination coverage is 100%, we will potentially be able to prevent around 6–700 cases of ASCUS/LSIL every year. Even though it is well known that many of these lesions will regress spontaneously, women with ASCUS/LSIL are referred for further examination/treatment, which leads to increased healthcare costs and also psychological stress for the woman involved and possible overtreatment. Furthermore, we have not in these calculations included the benefit which will be achieved by reduced treatments of genital warts which almost entirely are caused by HPV6 and HPV11.

The present study has several strengths. It is focused on LR-HPV infections and consists, to our knowledge, of one of the biggest sample sizes. In total, we included more than 40 000 Danish women from the general population of Copenhagen. Based on the large data material, we were able to make precise estimates. In addition, the selection of a screening population where women were included consecutively potentially makes our results generalisable to the general population. However, the Greater Copenhagen area may differ in terms of sexual habits and other behavioural factors, which may potentially have caused a higher HPV prevalence in this study compared with studies also including women from rural areas. It is a limitation of the study that we were only able to genotype the samples positive to HC2. This implies an underestimation of the prevalence especially for the four LR-HPV types not in the LR-HC2 test (HPV40, 54, 70, 74), both overall and for infections occurring alone.

In summary, we found HPV6 to be the most prevalent LR-HPV type in the general population of Danish women, whereas HPV11 was more rarely detected. This is consistent with findings in genital warts where HPV6 is found in the majority of the lesions and HPV11 is more rarely detected. We revealed that LR-HPV types are prevalent in women diagnosed with low-grade lesions, particularly ASCUS. Finally, we confirmed that practically no LR-HPV infections were present without concomitant HR-HPV types in the more severe lesions. In addition, it is possible that the few high-grade lesions (HSIL and CIN2+) with LR-HPV alone may truly have a HR-HPV which was not detected by the HC2 test. These data should inform cost-effectiveness analyses of using HR-HPV assays in cervical screening as this would prevent detection and treatment of benign abnormalities due to HPV 6/11.

Key messages

  • Low-risk HPV types are relatively prevalent in women diagnosed with ASCUS.

  • Practically no low-risk HPV types are present without concomitant high-risk HPV types in severe cervical lesions.

  • These data provide information valuable for modelling of the cost-effectiveness of prophylactic HPV6 and HPV11 vaccination.

Acknowledgments

We thank Betti Schopp and Barbara Holz at the University Hospital of Tübingen and the study staff at Hvidovre Hospital and the Danish Cancer Society for their technical support. We also thank Gabriele Breugelmans and Catherine Cohet at Sanofi Pasteur MSD Epidemiology department for their help in the initiation of the study.

References

Footnotes

  • Funding The study was funded through an institutional grant from Sanofi Pasteur MSD (Lyon, France) and from the Danish Cancer Society. Furthermore, the study was supported by MERMAID (Mermaid 2 project). The funders had no role in the conduction or interpretation of the study. The authors take full responsibility for the content of this contribution.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by the Danish Data Protection Agency and the Danish Scientific Ethical Committee.

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