The prevalence of human papillomavirus and its impact on cervical dysplasia in Northern Canada

Introduction Certain types of the Human Papillomavirus (HPV) are sexually transmitted and highly associated with development of cervical dysplasia and cervical cancer but the distribution of HPV infection in the North, particularly amongst First Nations, Metis, and Inuit peoples, is little known. The purposes of the study are to identify the prevalence of type-specific HPV infections and the association of different HPV types with cervical dysplasia among women in Northern Canada. Methods This was a cross-sectional study with attendants of the routine or scheduled Pap testing program in the Northwest Territories (NWT), Nunavut, Labrador and Yukon, Canada. Approximately half of each sample was used for Pap test and the remaining was used for HPV genotyping using a Luminex-based method. Pap test results, HPV types, and demographic information were linked for analyses. Results Results from 14,598 specimens showed that HPV infection was approximately 50% higher among the Aboriginal than the non-Aboriginal population (27.6% vs. 18.5%). Although the most common HPV type detected was HPV 16 across region, the prevalence of other high risk HPV types was different. The age-specific HPV prevalence among Aboriginal showed a ‘U’ shape which contrasted to non-Aboriginal. The association of HPV infection with cervical dysplasia was similar in both Aboriginal and non-Aboriginal populations. Conclusions The HPV prevalence was higher in Northern Canada than in other Areas in Canada. The prevalence showed a higher rate of other high risk HPV infections but no difference of HPV 16/18 infections among Aboriginal in comparison with non-Aboriginal women. This study provides baseline information on HPV prevalence that may assist in surveillance and evaluation systems to track and assess HPV vaccine programs.


Introduction
Cervical cancer is the second most common cancer of the women worldwide. In Canada invasive cervical cancer and associated mortality among Aboriginal women have been 2-4 times higher than among non-Aboriginal women [1][2][3][4]. From the Statistics Canada 2006 census Aboriginal groups make up more than 50% of the population in the Northwest Territories (NWT) [5] (First Nations, Metis, and Inuit), 85% of the population in Nunavut [6] (predominantly Inuit), 35% in Labrador [7] (Metis, Inuit, and First Nations) , and 25% in the Yukon [8] (predominantly First Nations). Increased risk of cervical cancer in Aboriginal women has been consistently observed in various regions across Northern Canada [2,9,10]. The gap in cervical cancer and associated mortality between the Aboriginal and non-Aboriginal groups has important implications for cancer control policy. Implementation and evaluation of cervical cancer prevention programs in Aboriginal women may require a different approach.
Human Papillomavirus (HPV) is a common sexually transmitted infection worldwide. Persistent infection with specific types of the virus is necessary for the development of invasive cervical cancer and its precursor lesions [11]. HPV is detected in almost 100% of women with invasive cervical cancer and types 16 and 18 are responsible for approximately 70% of all cases [12,13]. In Canada HPV 16/18 prevalence has been reported in 56.2% of high grade precursor lesions [14]. HPV 16/18 prevalence can vary by geography and ethnicity. Therefore, determination of type-specific HPV prevalence in a region may be one of the important steps towards cervical cancer prevention.
Only very limited regional data [2,3,15] provide HPV prevalence in Northern Canada which appears different with that of the general population. Population based research on HPV and related outcomes in the North could fill this information gap. For cervical cancer prevention and early detection of pre-cancerous lesions, routine Papanicolaou smear (Pap) test screening is offered in all regions of Northern Canada with good coverage of targeted population. In 2008, 88%, 74%, 83%, and 81% of women aged 18-69 in NWT, Nunavut, Labrador, and Yukon respectively, were reported as having had a Pap test in the previous three years [16]. It was therefore considered a practical and efficient approach to examine HPV prevalence as part of routine Pap smear screening.
The purpose of this study is to measure the prevalence of HPV infections by type and region and to determine their impacts on cervical dysplasia across Northern Canada.

Study population
This was a cross-sectional study including all women aged 14 years or older who presented for a regular Pap test in the four regions of Northern Canada: NWT (April, 2008  . Women with a cervical cancer history and those who chose to opt-out were excluded from the study. Only the first sample collected was kept for analysis for women with more than one Pap smear performed over the study period. An opt-in consent process was used in Yukon which was different with the other jurisdictions. Given the universal coverage of the health care system in Canada there are no other Pap test providers in these areas.

Authority and licensing
This research received ethical approval from the Health Canada Ethics Review Board, McGill University Ethics Review, and the local health authorities including NWT Department of Health and Social Services, Stanton Territorial Health Authority, the Yukon Department of Health and Social Services, and the Yukon Scientists and Explorers Act License. A Northern research licence was also obtained from Aurora Research Institute in the NWT. The project was facilitated by four data collection sites and the manuscript was reviewed by community stakeholders before being submitted for publication.

Data source
Pap test screening is performed in each Territory and the Labrador health region in Canada. It is offered to all women aged 18-69 and young women aged 14-18 who wish to participate. Pap smears are collected at community health centres, medical clinics, or hospitals whichever appropriate. Pap test registration form is used to assemble personal record and demographic information on site. Demographic details including Aboriginal status, age, and region information were abstracted from the routine Pap test registration form. Aboriginal status was only available among NWT, Nunavut, and Yukon datasets. Only area code (residence in Aboriginal area or non-Aboriginal area) was available in the Labrador dataset. Data were collected, organised, and owned by the local health authorities.

Specimen collection and preparation
Samples from NWT and Nunavut were collected in liquid based cytology (LBC) medium (SurePath W ) and were shipped to the DynaLife Medical Laboratory in Edmonton, where cytology was read and remaining half of the specimen was kept for HPV testing. Samples from Labrador were also collected in SurePath LBC medium and shipped to Newfoundland Public Health Laboratory for Pap test interpretation. In Yukon, conventional Pap smear slides were taken and shipped to the Cervical Cancer Screening Laboratory, BC Cancer Control Agency. A sample for HPV testing was prepared by taking a second cervical sample and by rinsing the brushes in LBC medium. Samples for HPV genotyping were shipped from all four areas to the National Microbiology Laboratory (NML) in Winnipeg, Canada. Cervical cytology and HPV testing were done independently to eliminate the potential for reporter bias.

Pap smear testing
Cytological reports were issued according to the Bethesda system [20,21]. The results were classified as: normal (including benign), atypical squamous cells of undetermined significance (ASC-US), low-grade squamous intraepithelial

lesion (LSIL), and high-grade squamous intraepithelial lesion (HSIL). Samples with atypical squamous cells (ASC-H),
where HSIL could not be excluded, were added to the HSIL category. To be consistent with reported data worldwide we classified LSIL and HSIL as abnormal categories for the correlation analysis and impact assessment. ASC-US cases were not included in this part of analysis.

Data preparation
Demographic information, HPV types and Pap test results were linked for HPV prevalence analysis. In the linkage dataset, 764 records (approximately 5% of collected data in the study) with one or more items of missing value were excluded from the analysis; 485 missing in age or age-group, 327 in ethnic status or ethnic area, and 399 in Pap test including 126 unsatisfactory cytology results. HPV_s and HR_s were defined as cases with only a single type of HPV infection and cases with only a single type of high risk HPV infection, respectively. Similarly, HPV_m and HR_m were defined as samples with two or more types of HPV or HR HPV infection. Samples with two or more HPV types were counted in for each type in type-specific analysis.

Analysis
Prevalence rates for type-specific HPV infection and cervical dysplasia were computed. Based on the 2008 census data among women in the NWT region [5] as the standard, the age-adjusted prevalence rates (AAPR) were calculated for area comparison. The overall prevalence rate in Northern Canada was also weighted by population size [5][6][7][8]

by region.
A multivariate logistic regression analysis was used to explore the associations between HPV infection and cervical dysplasia, adjusted for Aboriginal status, age, and region. Odd Ratios (ORs) were calculated by different level: (a) infection with any HPV type, single HR infection, multiple HR infection, single LR infection, multiple LR infection, and infection with both LR and HR types, relative to no HPV infection, and (b) infection with both HR and LR HPV types relative to a single LR HPV infection.
The proportion of cervical dysplasia caused by HPV infection in Northern Canada was measured by the exposure attributable risk fraction (ARF) and the population attributable risk fraction (PAR). In calculating the PAR, to avoid duplicate counting of those infected with both HR and LR, the exposed population was first stratified into those infected with only HR, only LR, and both. The PARs were summed across the strata. ORs are used in the calculation of both ARF and PAR: ARF = (OR-1)/ OR and PAR = (Number of exposed cases x ARF)/total cases in the population. Furthermore, PAR of HR HPV infection = (No. of HR_s cases x ARF reference to HPV negative + No. of only multiple HR cases x ARF reference to HPV negative + No. of both HR and LR infected cases x ARF reference to LR HPV) / total cases [22].
All data management and analyses were carried out using Statistical Analysis System 9.1. Two-tailed p value < 0.05 was the criteria to determine the statistical significant [23].

Results
A total of 14598 cytology samples were collected and tested from NWT (n = 6940), Nunavut (n = 4683), Labrador (n = 1370), and the Yukon (n = 1605). In NWT and Nunavut the study covered more than 95% of the Program targeted population, respectively; but only 77% and 27% in Labrador and the Yukon in the study period. In this study, approximately 42% of women were non-Aboriginal and 58% were Aboriginal. The overall mean and median ages of participants were 35 and 32, respectively.

HPV prevalence analysis
Type specific HPV-DNA was detected in 26% of subjects (Table 1). After weighting the four regional population sizes from census data, the overall prevalence rate in Northern Canada was estimated to be 25.2%. Although the specific prevalence of HPV types was dissimilar among regions, the highest prevalence HPV types detected were HPV-16  (Figure 1). Among HPV infected samples, 79% had an HR type, 33% had multi-types, and 25% had HPV 16/18. The overall prevalence of HPV_m, HR_s, HR_m, and LR_s were 8.4%, 13%, 4.0%, and 4.7%, respectively.

Age specific
The age specific HPV prevalence rate ranged between 13% and 48% ( Figure 2). The highest prevalence of HR HPV infection and HPV 16/18 was among women aged less than 20 years, with decreasing prevalence until age group 40-49. HPV prevalence was higher in the Aboriginal than the non-Aboriginal women in each age group; in particular among women aged 50 or over ( Figure 3). It is of note on the trend that the prevalence of HPV increased again after age 50 thus forming a U-shaped curve in the Aboriginal women, but following a monotonic decrease with age in the non-Aboriginal women.

Regional and ethnic differences in HPV prevalence
In

Impact of HPV on cervical abnormalities
After adjusting for Aboriginal status, age, and region, the logistic regression models showed that the ORs for cervical abnormalities associated with any type HPV, LR_s, HR_s, only multiple LR infections, only multiple HR infections and mixed HR and LR infections were 28.9, 5.3, 29.4, 6.2, 71.9, and 35.4, respectively, and that HPV infection accounted for 85.7% of the cervical abnormalities with 80.6% attributable to high risk HPV infection and 32.5% attributable to HPV 16/18 infection (Table 4).

Discussions
This study recruited most of the women aged 18-69 (and most of the incidental consultations for the 14-17 age group) who attended routine or scheduled pap screening over an approximately one year time period. The prevalence of overall HPV infection was 25.2% which was higher than in most countries [24][25][26][27] but similar to Denmark [28] and the US [29]. HPV prevalence rates in both the Aboriginal (31.6%) and non-Aboriginal women (18.9%) in this study were slightly higher than those in other studies in Aboriginal [2,15] and non-Aboriginal [30,31] populations in Canada. The proportion of HR types among the HPV-positive women was lower than in British Columbia and Ontario studies but higher than in a Quebec study [2,30,31].  Similar to most European and Canadian studies the two most common HPV types detected in Northern Canada were HPV-16 and 31 [28,[32][33][34][35][36]. The HPV 16/18 prevalence in the study was relatively lower than all similar studies in Canada [2,30,31,35], but comparable to studies in Europe and the US [25,26,28,29,37]. In this study HPV 16/18 comprised about 31% of all HR-HPV infections. Several HR HPV types other than HPV 16/18, such as HPV 31, 39, 45, 51, 59 and 66 were relatively common in Northern Canada.
The prevalence analysis shows that HPV prevalence was highest among women aged less than 20 years, which is consistent with other studies [27,30,31,38,39]. This finding confirms what is already known on the natural history of HPV infection, which is characterized by higher infection rates after sexual initiation. In agreement with published data [2,40], the age adjusted HPV prevalence rate in the Aboriginal group was higher than that in the non-Aboriginal group. Inconsistent trends in HPV prevalence by age were noted in older women, with a decrease or plateau of HPV prevalence among the non-Aboriginal group, whereas the Aboriginal group showed an increase of HPV prevalence (U-shaped age-specific HPV prevalence distribution). This has been observed in the other studies in Canada [2,15,30,31] and from Costa Rica and other Latin American countries [41,42], with the lowest prevalence among the age group 40-49. However, from most of the European literature, with the exception of the former Soviet Union [33], a decrease in HPV prevalence after age 20 and a levelling off after age 45, has been reported [25,26,28,42]. Further study needs to appreciate what role that social, cultural and lifestyle factors, as well as environmental determinants such as access to care may play in explaining this difference in age distribution.
Noteworthy is that, despite the difference in prevalence of HPV types and cytological abnormalities between Aboriginal and Non-Aboriginal women, the proportion of HPV 16/18 in high grade lesion is about 50% in both groups. This is very similar to the average proportion of HPV 16/18 in HSIL in North America and worldwide [13]. Since HSILs are the precursors of invasive cancer, these results suggest that the current HPV vaccines should be equally applicable to Aboriginal and Non-Aboriginal women.
Currently routine Pap testing is one of the pillars to prevent cervical cancer. Other studies have found similar Pap test abnormality rates [2,30]. HPV prevalence was higher in those women aged < 30 than in older women across the spectrum of abnormal Pap results. This is consistent with reports showing that HPV infection is common in young women and most often transient [43][44][45][46]. It is expected that women aged 30 years and older experience more persistent HR infections than younger women and that HPV-DNA testing would be able to triage this group for further follow-up.
Among participants with a normal pap test result, 26% of any type HPV infection among the Aboriginal group and 15% among the non-Aboriginal women (Table 3) translated into a 28 times higher risk of cervical abnormality than those with HPV negative (Table 4). By the same token, among 5.4% of the Aboriginal women and 3.4% of the non-Aboriginal women with both normal Pap test results and HPV 16/18 infection, they have a 50 times higher risk of cervical abnormality than those who are HPV negative. Interestingly, similar to our previous NWT report [47], results seem to indicate an additive risk model for multiple HPV infections. This information may be useful in considering HPV testing in cervical cancer screening programs in Northern Canada. There are potential limitations in the study. Firstly, no PCR protocol is ideal for all types of HPV detection. Any of the above comparisons in prevalence rates must be taken with caution as we have used the most recent IARC classification for HR HPV and a novel (Luminex) detection method as sensitive as Linear Array but able to detect more types than other previously reported technologies such as GP5+/6+, Elisa, or Hybrid Capture [17,18]. Secondly, this study represents HPV infection at various time points in the different regions, and does not represent the proportion of women at risk for disease or cancer related to HPV infection. Of note, there was no full regional coverage of the targeted populations during the study period. Different regions had different recruitment methods and therefore rates. Due to an "opt-in" process, the coverage for the Yukon sites lower than the other regions, and one thus should be cautious in interpreting the prevalence rates of the study population for Yukon. In addition, the Aboriginal status was not available in Labrador. To estimate HPV prevalence by ethnic group in the entire North in Canada, we used residence for Aboriginal or non-Aboriginal area instead as the purpose of the study was to understand the severity of HPV prevalence in Northern Canada and each northern region. Although we provided both crude rates and age adjusted rates for HPV infection, it still needs to be cautious to make comparison by region due to substantially different data collection processes. Finally, it's worth noting that routine screening is not recommended by standard guidelines for young women aged 14-18. In addition Pap screening is not mandatory for anyone; e.g., NWT has their own guidelines [48] and Yukon follows a guideline from British Columbia. Therefore, there may be a selection bias toward the young age group in data analysis. Nevertheless, the overall analysis was conducted with and without the 14-18 years age group and the results remained similar. We suspect also that there may be a potential bias affecting participation by Aboriginal versus non-Aboriginal people.

Conclusions
In Northern Canada, the prevalence of HPV infection was higher than other Canadian areas and varied by region. Between the Aboriginal and the non-Aboriginal group, the age specific distribution rates of HPV infections appear to be different; however, the effect of HPV infection to cervical dysplasia was similar. The relatively high prevalence of HR-HPV other than HPV 16/18, in particular among the Aboriginal group, call for further study to understand the impact of specific predictors [49]. This study provides baseline information on HPV prevalence that may assist in surveillance and evaluation systems to track and assess HPV vaccine programs.
Abbreviations AAPR: Age-adjusted prevalence rate; ASC-H: Atypical squamous cells where HSIL could not be excluded; ASC-US: Atypical squamous cells of undetermined significance; ARF: Attributable risk fraction; CI: Confidence interval; HPV: Human papillomavirus; HPV_m: Two or more types of HPV infection; HPV_s: a Single type of HPV infection; HR: High risk; HR_m: Two or more types of high risk infection; HR_s: a Single type of high risk infection; HSIL: High-grade squamous intraepithelial lesion; LR: Low risk; LSIL: Low-grade squamous intraepithelial lesion; NML: National microbiology laboratory; NWT: Northwest territories; OR: Odds ratio; Pap: Papanicolaou test; PAR: Population attributable risk fraction; PCR: Polymerase chain reaction.

Competing interests
There is no conflict of interest related to this work from any author.
Authors' contributions YJ: linked the datasets, analyzed the data, drafted the manuscript. PB: participated in the data collection, data analysis and results interpretation. AS: carried out all HPV DNA typing tests, revised the manuscript and helped to finalize the manuscript. YM: designed the study and made the substantial contributions to interpretation of data. YAL: involved in the majority of the project management work for NWT and Yukon. JAL: involved in data interpretation and helped revise the manuscript. SC: carried out the data collection in NWT and involved in the results discussion. AC: carried out the study in NWT site and involved in the results discussion. KK: carried out the study in NWT site and involved in the data collection and linkage. BH: carried out the data collection in Yukon site and provided with the critical review of the results. IS: initiated and managed the data collection in Nunavut. MA-R: initiated and carried out the data collection in Labrador-Grenfell Health, and involved in the results discussion. GJ: carried out Pap test for samples from NWT and Nunavut. JL: carried out Pap test for samples from Yukon. SR: carried out Pap test for samples from Labrador and involved in the results discussion. TW: involved in the data collection in NWT, Nunavut, and Labrador. AD: participated in the analysis of data and the critical review of the results. GJ: participated in conducting HPV type-specific surveillance and linkages to cytology results in Nunavut, involved in data interpretation and helped revise the manuscript. ST: involved in data interpretation and helped revise the manuscript. HM: performed the study and revised the manuscript. All authors read and approved the final manuscript.