Self-collected cervical samples for detection of high-risk human papillomavirus: A community outreach approach in a remote region

Abstract

Background and objectives

Self collection of cervical samples is evolving as a potential screening approach to overcome cultural, geographical and logistical barriers that contribute to poor screening uptake. We conducted a study to compare performance of self and physician-collected cervical samples for high-risk human papillomavirus (hr-HPV) detection through community outreach approach in the hilly terrains of Sikkim.

Methods

A prospective longitudinal study was conducted for 2 y, sampling 210 sexually active women aged 18-65 y attending Central Referral Hospital and 3 district hospitals in Sikkim. Paired self and physician collected cervical samples were taken from consenting women for HPV DNA and pap test. Women who had total hysterectomy, diagnosed with cancer or were pregnant were excluded. Colposcopy was done among those positive for HPV infection.

Result

Prevalence of HPV (HPV DNA ± pap) was 17% (36/210) in this study of which HPV 16 was the most common variant (n=4,15%). Physician collected samples had the highest HPV detection rate (n=26,12%). Sensitivity and specificity were highest with physician-collected samples (71.4%, and 76.9%, respectively). Concordance between physician and self-sampling was good (Kappa 0.7). Women lost to follow up for colposcopy was 36% (n=13).

Interpretation and conclusions

Performance of physician-sampling was better than self-sampling and pap in the detection of hr-HPV. Physician-sampling should be used for HPV DNA detection when feasible.

Cervical cancer is the second most common cancer among women in India (17.7% as per GLOBOCAN 2022).¹ as well as in Sikkim (10.3%).² Cervical cancer has a long latent period from infection to development of cancer. Persistence of human papillomavirus (HPV) causes cervical epithelial changes leading to progression from cervical intraepithelial neoplasia (CIN) 1 to CIN2 to CIN3 and ultimately invasive cancer. Furthermore, most HPV infection (90%) gets resolved within two years. Thus, screening has a significant role to play in reducing the morbidity and mortality caused by cervical cancer.

Pap test is an easy and low/no-cost test made available for women in Sikkim and in India through the public health delivery systems. However, most women lack awareness and access to the test, and a large section of women living in the low- and middle-income countries (LMICs) refuse Pap testing as it requires pelvic examination for the collection of samples. The low screening uptake has led to the introduction of self-sampling method of sample collection by World Health Organization (WHO).³

Sikkim is a State in the northeastern Himalayan foothill region of India, which is culturally and socially divergent from the rest of the country. The geographical landscape of the hilly terrain further adds to local healthcare access challenges posed by seasonal interruptions such as landslides in monsoons and snow in winters. This study was designed with an objective to compare performance of self-collected versus physician collected cervical samples for high risk-HPV (hr-HPV) detection among women in Sikkim through community outreach approach.

Methods

This was a prospective longitudinal study conducted by the department of Biochemistry in collaboration with department of Obstetrics and Gynaecology and Pathology, Sikkim Manipal Institute of Medical Sciences, Sikkim Manipal University, Sikkim, India, over a period of 2 y (May 2022- May 2024). Institutional ethical clearance and permission to hold the outreach camps at the district hospitals were obtained from the Government of Sikkim. Accredited social health activists (ASHAs) were mobilised in the districts to disseminate the information about the screening camp per protocol.

Inclusion

Women patients and attendants, 18 to 65 y of age, with a history of sexual activity, visiting Central Referral Hospital (CRH), and the three district hospitals in Sikkim (West, South and North) were invited for screening of cervical cancer using Pap and HPV DNA hybrid capture 2 (HC2) tests on cervical samples. Digene HC2 is a qualitative test which is FDA (Food and Drug Administration) approved for detection of 13 types of hr-HPV (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68).

Exclusion

Women diagnosed with cancer or who had undergone total hysterectomy (uterus and cervix removed) or were pregnant were excluded from the study.

General demographic details of participants were collected after obtaining informed written consent. With the help of pictures and models, participants were taught to collect their self vaginal samples in specimen transport medium (STM) vials, digene® HC2 DNA collection device (Qiagen, USA) for the HPV DNA HC2 test. On the same day, the physician then collected two sets of cervical samples via a pelvic exam. The first cervical sample was smeared and fixed on a glass slide for the pap test and the second was collected with the help of a brush which was stored in STM vials for HC2 test. Glass slides were transferred to CRH for pap test while samples in STM vials (both self and physician collected) were stored at -20°C till transfer to the designated laboratory outside Sikkim for HPV DNA HC2 within 1-2 mo of sample collection, in batches. Relative light units/cut off (RLU/CO) ratios were used as per laboratory standard practice. An RLU/CO < 1 indicated low HPV DNA and negative for HPV DNA. The samples which were positive for hr-HPV DNA were further genotyped using real time polymerase chain reaction (PCR). Participants positive for HPV were advised for colposcopy at CRH to detect any CIN or malignant changes. Swede score was used to grade the disease on colposcopy, where score ‘0’ was taken as negative, score of 1-4 as CIN1, 5-6 as CIN2 (high grade/non–invasive) and score of 7–10 as CIN2+/3 (high grade/suspected invasive cancer).⁴ Women positive for hr-HPV who had CIN2 and CIN3 changes had large loop excision of transformation zone (LLETZ) for histopathology. All women who were positive for HPV infection were then followed up after 6-12 mo with Pap.

Sample size calculation

The sample size was calculated using women population of Sikkim as 287,507 (2011 census), prevalence of hr-HPV as 10.7% using regional data which ranged from 9%- 12.5.^5-8 Taking a confidence level of 95% and 5% margin of error the sample size was calculated as 139; however, assuming a 10% dropout and/or loss of biological samples, 210 women were enrolled. Proportionate sampling was used to enrol women from the four districts (North, South, East and West). The outcomes studied were prevalence of hr-HPV in Sikkim, association of HPV infection with demographic characteristics and the performance of the screening tests using colposcopy as the gold standard. Statistical analyses were performed using StataSE 17 (StataCorp LLC, College Station, TX, USA), IBM SPSS Statistics version 27 (IBM Corp., Armonk, NY, USA), and R version 4.5.1 (R Foundation for Statistical Computing, Vienna, Austria), using RStudio 2024.12.1+563 (Posit Software, PBC, Boston, MA, USA).

Results

The prevalence of HPV among 210 women recruited for the study was 17%. The detection of hr-HPV was higher on physician-collected samples compared to self-collected paired vaginal samples (12% vs. 9%, respectively). When HPV DNA and Pap test results were compared among the 26 HPV DNA positive women: 16 women tested positive on both self-collected and physician-collected samples. Eight women tested positive exclusively on physician-collected samples, and 2 women tested positive exclusively on self-collected samples. Among the 13 women who had positive Pap results: 2 were co-positive for HPV DNA on both collection methods and, 1 was co-positive for HPV DNA on the physician-collected sample only. Ten women tested positive for Pap alone, with no detectable HPV DNA ( Fig. 1).

Fig. 1.

Venn diagram to show positivity of HPV on paired cervical samples using different tests. It also represents the overlap of positivity in different tests and collection techniques.

General characteristics of the study participants who were positive for HPV DNA in terms of the age, district, ethnicity, BMI, age at first sexual intercourse, socioeconomic status and parity are described in Table I. All women who were positive for HPV infection were contacted to come for colposcopy at CRH but only 23 out of 36 positive women (64% of positive cases) attended the colposcopy clinic while the remaining 13 women (36%) were lost on follow up. The HPV DNA results from cervical smears collected by both physician and self-sampling were compared against colposcopy findings, which served as the gold standard.

Table I.

General characteristics of women with signs of HPV infection

Characteristics	Frequency (N=210); n (%)	Women with signs of HPV infection; (n=36) ; n (%)	P value
Age groups (y)
< 30	11(5)	2 (18)	0.925 (Chi square test)
≥ 30	199 (95)	34 (17)
Districts
East	101	23 (23)	0.20 (Chi square test)
West	49	6 (12)
North	14	1 (7)
South	46	6 (12)
Ethnicity
Bhutia/Sherpa/Tamang	34	4 (12)	0.712121 (Chi square test)
Lepcha	14	2 (14)
Nepali	150	27 (18)
Others	12	3 (25)
BMI
Underweight	10	2 (20)	0.07 (Chi square test)
Normal weight	92	10 (11)
Overweight	73	19 (26)
Obese	35	5 (14)
Age at first sexual intercourse (y)
13- 18	59	12 (20)	0.13 (Fisher’s exact test)
19 - 24	90	16 (18)
25 - 30	56	8 (14)
31 – 36	5	0
Socio-economic status
Class I	16	1 (6)	0.31 (Fisher’s exact test)
Class I and II	104	16 (15)
Class III and IV	90	19 (21)
Parity
0	14	2 (14)	0.656192 (Chi square test)
1 - 2	146	26 (18)
3-4	44	6 (14)
≥4	6	2 (33)

Results of analysis to assess the performance of pap test, physician- and self-collected samples with colposcopy as the reference standard are shown in Table II. The sensitivity and specificity of physician collected samples was better than self-collected samples and pap test. The sensitivity of pap and self-collected samples was comparable; however, the specificity of pap test (71.4%,) was better than HPV DNA test on self-collected samples (64.3% as shown in Table II).

Table II.

Performance of Self Sampling compared to Physician Sampling and pap test using coloscopy as gold standard (N=36)

Statistics	Self-collected samples for HPV DNA (n=36)	Pap test (n=36)	Physician collected samples for HPV DNA (n=36)
Sensitivity (95% CI)	61.54% (31.5 - 86.1)	58.33% (27.6 - 84.8)	71.43% (41.9 -91.6)
Specificity (95% CI)	64.29% (35.1 - 87.2)	71.43% (41.9 - 91.6)	76.92% (46 - 94.9)
PPV (95% CI)	30.11% (15.9 - 49.5)	33.79% (16.4 - 57)	43.62% (21- 68.7)
NPV (95% CI)	86.99% (75.2 - 93.6)	87.27% (76.4 - 93.5)	91.50% (81.7- 96)
Accuracy (95% CI)	63.74% (43.1 - 81.2)	68.81% (47.7 - 85.3)	75.82% (55.6 - 90)

PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval; HPV, human papillomavirus

Cohen’s Kappa coefficient inter-rater agreement, was used to evaluate the agreement between the physician and self-collection methods which showed a good agreement between the two, with Kappa at 0.715. However, concordance between pap and HPV DNA tests on physician and self-collected samples was found to be low (Kappa <0.2).

Women who attended the colposcopy clinic were 23 out of 36 HPV positive cases (64%), of which 10 women did not show any changes on colposcopy. Among the remaining 13, eleven women had CIN1 changes while two women had CIN 2. Women who were positive on colposcopy had higher HPV DNA positivity rate as compared to the negative cases (10 vs. 6, respectively) and the pap smear also showed higher grade disease ( Fig. 2).

Fig. 2.

Study algorithm to classify and map the different screening test outcomes. Here, CIN, cervical intra-epithelial neoplasia; HSIL, high-grade squamous intraepithelial lesion; ASC-H, atypical squamous cells (cannot exclude a high-grade squamous intraepithelial lesion); ASC-US, atypical squamous cells of undetermined significance; LSIL, low grade squamous intraepithelial lesion; NILM, negative for Intraepithelial lesions or malignancy; PS, physician sample; SS, self-sample.

Women on scheduled visits have not developed malignancy till date. The samples which were positive for HPV DNA were further genotyped using the real time polymerase chain reaction (PCR). The most common hr-HPV detected in the study population was HPV 16 (n=4, 15%) followed by HPV 33, HPV 45, and HPV 58 (n=3, 12% each). High-risk HPV 18 accounted for only 8% (n=2) of positive samples.

Discussion

Prevalence of HPV infection by HPV-DNA (physician and self-collection methods) and Pap was 17% in this study. HPV DNA 16 was the most reported genotype among positives. Physician collected sample reported highest positivity, however the interrater agreement between physician and self-collected samples showed good concordance. The sensitivity and specificity of the physician collected sample was highest. Almost one-third of those positive were not available for colposcopy.

Laikangbam et al,⁷ 2007, reported prevalence of hr-HPV in Sikkim as 12.5% and a single institutional study conducted in Sikkim by Gupta et al⁹ reported 17.5% hr-HPV prevalence. Laikangbam et al⁷ also reported higher HPV prevalence in the age group of 30-50 y (20%), while the HPV prevalence amongst those aged ≤30 was 2.5%. In our study it was noted that hr-HPV prevalence in age group <30 y was 18% compared to 17% among women ≥ 30 y of age. Socioeconomic status, lifestyle and multiple child births have been found to be associated with HPV infection, however, in our study we found no association which may be due to the small sample size.^10-13 Though ethnicity had no significant association with HPV infection in our study, it was noted that women of non-ethnic origin had proportionately higher HPV infection than women of ethnic communities. The discrepancy between positive Pap results and lower HPV DNA detection rates may probably be attributed to the diagnostic scope of the hybrid capture platform, which is limited to testing only 13 HPV genotypes further supplemented by the fact that pap findings could also be due to infections by other HPV variants. Higher HPV DNA positivity on physician collected samples, as seen in our study, was comparable to findings reported by Sowjanya et al.¹⁴ However, another study that reported higher positivity for self-collected samples used alternative testing platforms.¹⁵ Physician-collected samples had better performance than self-collected samples and pap test in our study. A meta-analysis done by Tatara et al¹⁶ showed that sensitivity of self-collected samples HPV DNA test ranged from 74 to 86% while its specificity was more than 80% with regards to CIN2+ disease. However, another study which used PCR based platform for HPV DNA detection showed higher sensitivity but lower specificity for self-sampling compared to physician-sampling in the detection of CIN2+ disease.¹⁷ Since there were only two CIN2+ cases in our study, performance characteristics as per grade of disease was not evaluated. The accuracy of disease detection on colposcopy in the current study using physician-collected sample HPV DNA was the best in comparison to self-collected sample and pap test. Mayrand et al,¹⁸ showed that sensitivity of HPV test is higher than pap for high grade diseases (94.6% HPV DNA vs. 55.4% for Pap), however, the specificity for Pap was higher than HPV testing (96.8 vs. 94.1%, respectively). In our study physician-collected samples HPV DNA had higher sensitivity and specificity compared to pap test, however, specificity of pap was found to better than HPV DNA in self-collected samples.

The agreement between the physician and self-collected samples was good in our study, whilst the agreement with pap was found to be poor. Similar findings were reported by a study conducted at National Cancer Institute in Thailand¹⁹, Tranberg et al²⁰ and an Indian study by Bhatla et al.²¹ A study in 2009, showed that precursor lesion detection was highest with PCR testing on physician-collected samples, while it was lowest by HC2 on self-collected samples.¹⁴ The DNA detection rate in physician sample was higher than self-sample, (42% and 25%, respectively).¹⁴ The self-sampling method showed good acceptability in our society as per our earlier published report by Gibert et al²² wherein 91.7% of participants considered self-sampling as beneficial. There was high loss to follow up for colposcopy clinic. Despite counselling, and motivation during screening and perception change towards screening the high dropout could probably be due to, to the distance, geographical access, individual hesitation due to fear and prevalent culture of seeking second opinions in other centres. These factors need to be evaluated in future studies. Various studies have shown that non-attendance of colposcopy is high which ranges from 37–72%,^23,24 similar to what was seen in our study (34%). Since, the follow up of our study population for colposcopy was low, point-of-care HPV DNA testing of self-collected specimens and same-day thermal may would be of benefit to this segment of population.

Globally HPV-16 is the most prevalent, accounting for about 54% of cervical cancer cases, HPV16 was the most prevalent in our study followed by HPV 33, 45 and 58. Our study findings were similar to those reported by Laikangbam et al⁷ and Gupta et al.⁹ However, HPV 18 was not the second most common as was reported by other studies.

This is the only longitudinal study on HPV to be done across Sikkim. Despite geographical challenges, the study tried to compare sample collection methods and test parameters in remote locations. Certain insights like poor availability, and access to testing platforms at district and state level emerged as collateral findings. The limitation of the study was inability to conduct in-person follow up for colposcopy. The gold standard used to evaluate the performance of the tests on the various samples was colposcopy instead of biopsy due to anticipated anxiety among participants.

This current study in a remote State like Sikkim have demonstrated that self-sampling has comparable performance to physician sampling and may be a good approach to improve screening uptake in remote, accessed challenged, resource limited regions like Sikkim. However, wherever feasible, women should be given an option for sampling by the physician.

शोध-संदेश

बच्चेदानी की ग्रीवा के कैंसर का मुख्य कारण पैपिलोमा वायरस है जिसकी जाँच के लिए चिकित्सक द्वारा ग्रीवा से नमूने लिए जाते है। इस में समय लगता है एवं महिलाओं को जाँच के लिए चिकित्सक के पास जाना पड़ता है। महिलाओं दवारा स्वयं एकत्रित किये नमूने भी वायरस की स्क्रीनिंग बढ़ाने का एक उपयोगी और व्यावहारिक विकल्प हैं। सिक्किम के दुर्गम पर्वतीय क्षेत्रों में किए गए इस अध्ययन में पाया गया कि उच्च-जोखिम मानव पैपिलोमावायरस (High-risk Human Papillomavirus, hr-HPV) की पहचान के लिए चिकित्सक द्वारा लिए गए नमूने अधिक प्रभावी हैं। अध्ययन यह भी सुझाव देता है कि ऐसे दुर्गम क्षेत्रों में ‘स्क्रीन एंड ट्रीट’ (Screen and Treat) रणनीति अपनाने से समय पर जांच और उपचार सुनिश्चित करने में मदद मिल सकती है।

Author contributions

MLS, PGL: Conceptualisation, manuscript writing, formal analysis, funding acquisition; MS: Pap test analysis, manuscript writing; PGL, MLS: Methodology, data analysis, manuscript writing; AP, CG: Conceptualisation, project administration for camps, validation and visualisation. All authors have read and approved the final printed version of the manuscript.

Financial support and sponsorship

This project was funded by Sikkim Manipal Institute of Medical Sciences, Sikkim Manipal University, Sikkim under Dr. TMA PAI Seed Grant – Minor (SMIMS/ADRD/TMA PAI/2022/66 and SMIMS/RU/TMA PAI/2020-127).

Conflicts of Interest

None.

Use of Artificial Intelligence (AI)-Assisted Technology for manuscript preparation

The authors confirm that there was no use of AI-assisted technology for assisting in the writing of the manuscript and no images were manipulated using AI.