SEARCH

SEARCH BY CITATION

Keywords:

  • Cervical screening;
  • HPV self sampling;
  • HPV genotyping;
  • human papillomavirus;
  • participation rate

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

In France, about 40% of women aged 25–65 years do not participate in regular screening and thus are at high risk (HR) of cervical cancer. Human papillomavirus (HPV) vaginal self-sampling is a valuable alternative in this population. This study aimed to assess the prevalence of HR and LR (low-risk) HPV infection in 3767 women aged >35 years from mid-socioeconomic backgrounds who carried out HPV vaginal self-sampling at home. HPV vaginal self-sampling was better accepted than the Pap-test in women aged 35–69 years who were previously non-responders to individual invitation. From the 933 self-collected swabs studied (24.7%), 62 were HPV-infected (6.6%), and 73 HPV types were found. HPV 16 was the most frequently found (43.5%), followed by 53 (23.2%), 18 (12.3%), 66 (12.3%), 31 (6.8%), 33 (5.4%) and 58 (2.7%). Ten women (16.2%) were infected by multiple HR-HPV types. Median HPV 16 load was 104.000 copies/106 cells and median HPV 18 load was 833 copies/106 cells. Six women (9.3%) harboured LR-HPV types. The 12-month follow-up of 43 HR-HPV positive women (69.3%) revealed CIN2–3 lesions in three women (6.9%), all HPV 16 infected, and harbouring an HPV 16 load >5 log10 copies/106 cells. Women harbouring HR-HPV types other than HPV 16/18 were older than women harbouring HPV 16/18 types (55 years vs. 46.9 years, p 0.0008). The high frequency of HR-HPV types in women >50 years deserves further investigation to elucidate the mechanism involved (re-infection or reactivation).


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

High-risk oncogenic human papillomavirus (HR-HPV) infection is the major risk factor for the development of cervical cancer and cervical intraepithelial neoplasia (CIN) [1].

In France, 3000 new cases of cervical cancer are detected each year, leading to 1000 deaths per year [2]. Cervical cancer screening is performed on an individual basis and about 44% of the target population is not screened [3]. Cytological screening is often stopped at the age of 60 or 65 years, yet cancer incidence and mortality tend to increase from the age of 65 years [4]. According to recent studies, HPV DNA testing, in addition to cytology, could be useful for primary cervical cancer screening [5, 6]. Among women aged 35 years or older, primary HPV screening with cytology triage displayed a higher specificity than conventional Pap-smear screening [7]. In addition, HPV 16 viral load correlates with the severity of cervical lesions [8-10].

HPV vaginal self-sampling appears particularly useful for women with infrequent access to gynaecological healthcare [11-13] or refusing cytological screening because of cultural barriers or reluctance to have vaginal speculum examination.

In addition, more than 50% of all cases of cervical cancer are observed among women who do not respond to an invitation to have a Pap-test [14]. Moreover, the quality of self-sampling is usually satisfactory whatever the age. By contrast, in post-menopausal women the transformation zone is scarcely visible and the performances of cervical cytology are markedly decreased [15]. Recently, a preliminary pilot study using vaginal self-sampling for HPV testing was undertaken in two suburbs in the northern part of Marseille, where the socioeconomic level is low [13]. The rate of HPV infection was 23.3%, with 14.1% of HR-HPV types, including 25% of HPV-type 16. In that study, the age of HR-HPV-infected, LR-HPV-infected and multiple HPV-infected women was 42, 48 and 55 years, respectively.

The present study aimed to assess the prevalence of HR- and LR-HPV infection as well as HPV 16 or 18 load in women aged 35 years and older from two mid-socioeconomic cities around Marseille: Vitrolles and Marignane. These two cities have a relatively good healthcare system, but we tested women aged 35–65 years who did not attend regular cytological screenings for cervical cancer (referred to hereafter as non-attending). For this reason, self-collected vaginal swabs were considered valuable tools for detecting HPV infection, especially oncogenic HPV infection. Another objective was to evaluate the compliance with follow-up in this target population.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

Study population: the target population was composed of women aged 25–69 years living in two cities around Marseille (Vitrolles and Marignane) without a Pap-smear recorded in the National Insurance Registry for more than 2 years. A total of 17 330 women were invited to have a Pap-smear by individual mailing, free of charge in a local medical analysis laboratory. The number of women participating in self-sampling HPV tests was, however, limited by the financial support for the study (i.e. 4000 tests).Therefore, 9334 women aged 35–69 years without a Pap-test for more than 2 years (national insurance listing) who had not responded to a first invitation to undergo cytology were randomized into two groups: 4934 women receiving a second invitation to have a Pap-test and 4400 women receiving a proposal to perform HPV self-sampling at home.

The small unbalance between the two groups was due to the limited financial support for self-tests. After exclusion of 633 women for refusal or recent cytological screening, 3767 women remained in the HPV self-sampling arm. The flowchart of the study design is shown in Fig. 1.

image

Figure 1. Flowchart of the study design

Download figure to PowerPoint

Cervical cytology analysis

Among the 4934 women who received a second invitation to have a Pap-test, 4314 women were eligible for Pap-test (reasons for ineligibility were refusal, hysterectomy and pregnancy). Cervical cells were obtained using a cervical brush for conventional cytological slides.

Cytology was carried out at private cytology laboratories, and cytological diagnosis was carried out according to The French National Agency of Health consensus guidelines on the quality control of cervical smear screening [16] and formulated according to the 2001 Bethesda classification [17].Those carrying out cervical cytology analyses were blinded to the result of HPV genotyping.

Vaginal self-sampling and virological analysis

Flocked swabs (Copan Diagnostics, Brescia, Italy) were used as self-sampling devices. The women sampled their vaginal fluid using UTM tubes containing universal transport medium. The collected material was returned via mail in a prepaid envelope to the Department of Virology of the Timone Hospital (Marseille, France) and was stored at −20°C until use. After thawing, 250 μL of vaginal cell suspension was used for DNA purification using the QIAampDNA Mini Kit (QIAGEN, Courtaboeuf, France) modified as follows: samples were incubated at 56°C for 2 h in lysis solution containing proteinase K. The DNA was eluted in 100 μL of elution buffer and stored at −20°C until use. HPV genotyping was assessed by PCR (MY09/ MY11 primers), sequencing, phylogenetic analysis, and cloning if necessary, as described [13]. Quantitations of HPV 16 and 18 positive samples were performed using the quantitative duplex real-time PCR method, as reported [8]. HPV viral load was expressed as the number of HPV copies per million cells.

To follow-up abnormal tests, the virology and cytology laboratories communicated their results to the management centre ‘ARCADES’ in charge of sending them to the women with recommendations and to their referent physicians. A reminder regarding the management of abnormal results was also sent to these practitioners (gynaecologycal examination and colposcopy in the case of abnormal Pap-test and gynaecological examination and Pap-test if HPV-HR positive). The collection of follow-up data for abnormal tests was also monitored (mailing and telephone call) by the ARCADES centre after 3, 6 and 12 months.

Ethical approval

Ethical approval was given by the ethical committee of ‘Sud Mediterranée 2 and Marseille 2’ for all randomized studies on HPV testing in the Bouches du Rhône area performed by the ARCADES association.

Statistical analysis

Statistical analysis was performed using the spss (version 12.0) and epi info 6 softwares. A Yates's corrected chi-squared test was used to detect differences between groups. A descriptive analysis of the frequency and the distribution of the various HPV types and a comparison of median values using the Kruskal–Wallis test were performed.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

Participation in the re-invitation to have a Pap-smear and to carry out self-sampling in women aged 35–65 years

In the study group, the rate of attendance at the second Pap-smear screening invitation was 7.3%, compared with 25.1% who carried out self-sampling of vaginal fluid at home (p <0.001; OR = 4.64 (95% CI, 4.04–5.34). The rate of participation in Pap-smear screening did not decrease with age.

Prevalence of HR- and LR-HPV infection

Data from 946 women (946/3767; 25.1%) were recorded. To avoid false negative results, the samples from 13 women (1.3%) were excluded from the study because of paucicellularity. Finally, samples from 933 women (933/3767; 24.7%) were analysed.

Sixty-two samples from the 933 women were HR-HPV positive (6.6%), with 27 samples (43.5%) positive for HPV 16 (17 single infections and ten co-infections), nine samples (14.5%) positive for HPV 18 (six single infections and three co-infections) and 29 samples (46.8%) positive for other HR-HPV types (Table 1). As shown in Fig. 2(A), among the 73 HR-HPV strains, the HR-HPV types most frequently found were HPV 16 (n = 27, 43.5%), 53 (n = 17, 23.2%), 18 (n = 9, 12.3%), 66 (n = 9, 12.3%), 31 (n = 5, 6.8%), 33 (n = 4, 5.4%) and 58 (n = 2, 2.7%). Among the 62 HPV-infected women, six harboured ten LR-HPV types, including HPV type 6 (n = 3, 4.8%), 11, 70, 81 (n = 2, 3.2% each) and 23 (n = 1, 1.6%) (Fig. 2B), that were constantly associated with HPV 16 ± 18 types. Median HPV 16 viral load (104.000 copies/106 cells) was higher than median HPV 18 viral load (833 copies/106 cells) (p <0.001).

image

Figure 2. (A) HPV-HR type distribution among 62 HPV-positive samples. (B). HPV-LR type distribution among 62 HPV-positive samples.

Download figure to PowerPoint

The mean age of HPV-infected women was 51 ± 9.4 years, and the mean age of non-HPV-infected women was 52 years ± 8.5. Among the 62 HPV-infected women, 32 were ≥50 years, including 13 ≥ 60 years. No age-related differences were found between groups of HPV-uninfected, HPV-infected, HPV 16 or 18, multiple HPV-infected and LR-HPV-infected women (non-significant for all p values). Among the 29 women harbouring HR-HPV types other than HPV 16 and 18, 20 women were ≥50 years, compared with nine women ≤50 years (p 0.015). Women harbouring HPV 16 and 18 were significantly younger (mean age, 46.9 years) than women harbouring HR-HPV types (mean age, 55.1 years) (p 0.0008). No relationship was observed between HPV 16 or 18 viral load and age.

Table 1. Main virological characteristics of 62 HPV-infected women
HPV statusN (%)
HR-HPV-infected women62 (6.6)
Number of HPV 16 mono-infected women17
Number of HPV 18 mono-infected women6
Number of infected women with HR-HPV types other than HPV 16 or 18 (n)

29

HPV53 (14);HPV66 (8); HPV31 (5); HPV33 (1)

HPV58 (1)

Number of HPV16 + 18 dually infected women2
Number of HPV16 + 18 + other HR-HPV-infected women1
Number of HPV16 + other HR HPV-infected women (n)

7

HPV16 + 33 (2); HPV16 + 53 + 61 (1)

HPV16 + 53(2); HPV16 + 66 (1)

HPV16 + 58 (1)

Follow-up

After discovery of 62 women with positive DNA tests, at the end of follow-up (i.e. 12 months after stopping the study), 43 of these 62 women were referred to their physician (i.e. a follow-up rate of 67.7%) and were referred for cytology tests, whereas 14 women had colposcopy and 13 women had biopsy with the following data: three had CIN2+ (one CIN2, two CIN3), four had CIN1, and 36 had normal cytology findings (Table 2). Of note, all three CIN2+ lesions were related to HPV 16 infection and revealed an HPV 16 viral load >5 log10/106 cells in two CIN2+ cases and >6 log10/106 cells in one CIN3 case.

Table 2. Follow-up of 43 HPV-positive women at 12 months
HPV-positive women, n62
  1. a

    Including one HPV16+HR-HPV, three HR-HPV types.

  2. b

    Including one HPV16+.

  3. c

    Including two HPV16+.

Women referred to colposcopy, n (%)14 (22.6)
Women who had biopsies, n (%)13 (30.2)
Women who had surgery, n (%)5 (11.6)
Histology results, n (%)
CIN14a (9.3)
CIN21b (2.3)
CIN32c (4.6)
Normal cervix36 (83.7)

Among the 4314 women eligible for the second invitation to have a Pap-test, 315 women actually had a Pap-test, four women had colposcopy and three women had a biopsy.

In the follow-up of women who responded to the second invitation to have Pap smears, two high-grade lesions were detected (both CIN3) among 17 abnormal cytological results.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

Since 2008, we have undertaken three campaigns including a re-invitation to undergo a Pap-test or HPV self-sampling. Two of these campaigns targeted low-income populations of the northern part of Marseille in 2008 [13] and specifically the poorest women living in the district of the Bouches du Rhône in 2010, who were recipients of universal health insurance (UHI) with free access to care. The present study targeted medium socio-economic level populations. If we consider the three campaigns as a whole, in all, 31 945 women were eligible for a second invitation to have a Pap-test, and 17 956 women received an invitation to carry out HPV self-sampling. This large sample size makes our research unique in France, because we were the first French team to undertake screening campaigns combining re-invitation to have Pap-smears and HPV-self sampling. Overall, in the three campaigns, the participation rates were 4.9% for the second invitation to have a Pap-test and 14.9% for the HPV test, respectively. This represents a three-fold increase in the participation rate for use of the HPV self-sampling device, which confirms the good acceptability of self-sampling in non-attendees, in agreement with previous studies [11, 18-20]. The highest participation rates in self-sampling were observed in the present study (25.1% vs. 2.3% in the pilot study, vs. 18.3% in the UHI study).

In the present study the targeted women had a medium socio-economic level, improved access to and use of the healthcare system, and a satisfactory public health setting. Of note, a dramatic increase in the participation rate was noted in ‘lower’ social groups of women from the first the pilot campaign in 2008 to the last UHI trial in 2010. The higher participation rate in self-sampling of the women from Vitrolles-Marignane and the UHI trials (25.1% and 18.3%) was optimized by the do-it-yourself at-home opportunity given by the device. In contrast, the low-income women from the northern part of Marseille had to go to a medical analysis laboratory to perform the self-sampling [13]. Therefore efforts are ongoing to extend these screening campaigns targeting women of low socioeconomic position. Indeed a meta-analysis of social inequality in cervical cancer showed an odds-ratio of 1.97 (95% CI, 1.80–2.15) for cervical cancer among women of low socioeconomic position compared with women of high socioeconomic position [21].

The overall frequency of HPV infection, including mostly HR-HPV types, was 6.6%, a frequency close to that found by Sanner et al. [22] in Swedish women aged 55–60 years, and by Gök et al. [18] in the Netherlands. It was lower than the rates of 14% previously found in under-privileged women of Marseille [13], or more recently in a large study in Mexico [23], and much lower than the rates found by Stenvall et al. [24] in Swedish women aged 35–50 years, and by Rossi et al. [19] in Italian women aged 35–64 years (26% and 21.8%, respectively).

Most HPV self-sampling studies are based upon the Hybrid capture II assay, which does not discriminate the different HR-HPV types. In the present study, the use of PGMY 09/11 PCR or HPV 16/18 real-time PCR allowed HR- and LR-HPV types to be accurately genotyped. Thus, HR-HPV types other than HPV 16 (including HPV 53, 31, 33, 66) were more frequently found than the HPV 16 type (37/62, 59.6% vs. 27/62, 43.5%), which is in line with other studies [13, 25]. In a recent study, the most prevalent HR-HPV types, 16 and 31, conferred the highest risk of HSIL in women with abnormal cytology results and referred to colposcopy [26]. Yet, in a previous French study, HPV 16, 18, 66, 31, 33 and 58 were the most prevalent HR-HPV types in women with HG CIN2+ lesions [27].

Regarding prevention policy, although a vaccine against HPV 16 and 18 could theoretically prevent about 50% of HSILs, genotypes not covered by the vaccine were frequent in our target population. However, note that the three CIN2+ lesions revealed by self-sampling HPV typing in this study were related to HPV 16, which confirms, as reported by others [27, 28], that HPV 16 is the most common type detected among women with abnormal cytology. Moreover, in these three CIN2+ lesions, HPV 16 viral load was elevated (at 104.000 copies/106 cells, 104.000 copies/106 cells and 1.000.000 copies/106 cells). Interestingly, these three HPV 16 load values are close to the cut-off of 3.0 × 106 copies per million cells predictive of CIN2+ lesions (91% specificity, 58.2% sensitivity), as pointed out recently [29].

A previous study had also defined an HPV 16 load cut-off value of 21 × 106 copies per million cells to be predictive of CIN2+ lesions [8]. Altogether, these data suggest that a high HPV 16 viral load plays a role in oncogenic HPV persistence and progression to cervical lesions.

In this study, about 16% of women harbouring HR-HPV types were infected with multiple HPV types. Co-infection in women with multiple HPV types is a common finding of many molecular studies, and according to some authors, this could increase the risk of HG-lesions and invasive cancers relative to women harbouring a single HPV type [30], although this is still debated [31].

Regarding the relationship between the age and the presence of HPV 16/18 and presence of HR-HPV types other than HPV 16, the present study revealed that HPV 16/18-infected women were younger (47.7 years) than HR-HPV-infected women (54.3 years).One of the major outcomes of this study was the high frequency of HR-HPV types other than HPV 16 in women >50 years. This finding is consistent with a second peak reported by others in HPV prevalence [26, 32] and the hypothesis of a potential re-infection of women >50 years with HPV, or reactivation of latent infection in the post-menopausal period due to decreasing specific immunity. To draw final conclusions about the role of HR-HPV types other than HPV 16 in the development of CIN2+, further studies will have to follow-up a higher percentage of women participating in self-sampling and harbouring HR-HPV types.

The main limitation of this study was that not all women with HR-HPV positive tests performed cytology or histology testing so that the three HG lesions detected further were probably underestimated. Interestingly, the strength of the recent large Mexican study was the systematic implementation of reference standard mobile colposcopy concomitant with positive HPV tests [23].

The present results, which revealed three CIN2+ cervical lesions (6.9%) among the 43 HR-HPV-infected women with a 1-year follow-up, should be compared with the results of the recent UHI trial conducted in free healthcare centres for low-income women in the Marseille area in 2010. In that study, 1604 self-collected cervicovaginal samples were obtained, including 283 HR-HPV-positive samples (17.5%). At 1 year of follow-up, 117 women (41%) were re-evaluated by cytology ± colposcopy and biopsy, which revealed a higher frequency of CIN2+ lesions, including two cancers in the more precarious low-income women (nine CIN1, nine CIN2, one in situ cancer and one invasive cancer (9.4%) (Sancho-Garnier H, Tamalet C, Halfort P, Leandri FX, Heid P, Le Retraite L, Piana L, unpublished data). Another limitation was the lack of second HPV testing 12 months later in most HPV-positive women to identify whether HR-HPV types detected at first were transient or persistent; in this latter case, a referral to immediate colposcopy should be the rule.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

This study confirms that self-sampling is an attractive alternative strategy to Pap-smear screening to increase the participation of unscreened women. The quality of self-sampling is highly satisfactory and allows detection of more lesions than by Pap-smear screening, especially in elderly women who are not adequately screened. However, compliance with follow-up remains to be improved to validate the increased coverage by a marked detection of CIN2+ lesions.

This study also points out the need to regularly survey the distribution of HPV types in target populations to adapt vaccines for protection against the prevailing HPV types. Indeed, HR-HPV types other than HPV 16 were more widespread than expected (59.6% vs. 43.5%), although the three HG lesions observed were only related to HPV 16 accompanied by an elevated HPV 16 viral load.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

The authors thank Ms Solange Riban for helping to collect the data and for involvement in the study.

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References

The authors declare no conflict of interests.

This study was supported by a grant from Institut National du Cancer (INCA).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Transparency Declaration
  10. References