Screening for cervical cancer among HIV-positive and HIV-negative ...

Gynecologic Oncology 148 (2018) 118–125

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Screening for cervical cancer among HIV-positive and HIV-negative women in Cameroon using simultaneous co-testing with careHPV DNA testing and visual inspection enhanced by digital cervicography: Findings of initial screening and one-year follow-up Preetam Cholli a, Leslie Bradford a, Simon Manga b, Kathleen Nulah b, Edith Kiyang b, Florence Manjuh b, Geneva DeGregorio a, Rebecca K. Ogembo a, Enow Orock c, Yuxin Liu a,d, Richard G. Wamai e, Lisa Kennedy Sheldon f, Philimon N. Gona e, Zacharie Sando g, Thomas Welty b, Edith Welty b, Javier Gordon Ogembo a,h,⁎ a

University of Massachusetts Medical School, 55 N Lake Avenue, Worcester, MA 01655, United States Cameroon Baptist Convention Health Services, PO Box 1, Bamenda, North West Region, Cameroon c Regional Hospital, PO Box 32, Buea, Cameroon d Mount Sinai Health System, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Pl, New York, NY 10029, United States e Northeastern University, 360 Huntington Ave, Boston, MA 02115, United States f Oncology Nursing Society, 125 Enterprise Drive, Pittsburgh, PA 15275, United States g Yaoundé Gyneco-Obstetric and Pediatric Hospital, Yaoundé, PO Box 4362, Central Region, Cameroon h Beckman Research Institute of City of Hope, 1500 E Duarte Road, Duarte, CA 91010, United States b

H I G H L I G H T S • • • • •

Co-testing for cervical cancer and its precursors in resource-limited settings. Half of all VIA/VILI-DC positive women had no associated high risk HPV infection. VIA/VILI-DC false positives dropped from 50% overall to 14% in HIV-positive women Nearly half (44.3%) of all HPV-positive women spontaneously cleared infection in one year HPV co-testing with VIA/VILI-DC may identify false positives, reduce overtreatment

a r t i c l e

i n f o

Article history: Received 13 September 2017 Received in revised form 31 October 2017 Accepted 2 November 2017 Available online 16 November 2017 Keywords: Cervical cancer screening HPV DNA testing Visual inspection with acetic acid Women's health Sub-Saharan Africa

a b s t r a c t Objective. The World Health Organization (WHO)'s cervical cancer screening guidelines for limited-resource settings recommend sequential screening followed by same-day treatment under a “screen-and-treat” approach. We aimed to (1) assess feasibility and clinical outcomes of screening HIV-positive and HIV-negative Cameroonian women by pairing visual inspection with acetic acid and Lugol's iodine enhanced by digital cervicography (VIA/VILI-DC) with careHPV, a high-risk human papillomavirus (HR-HPV) nucleic acid test designed for low-resource settings; and (2) determine persistence of HR-HPV infection after one-year follow-up to inform optimal screening, treatment, and follow-up algorithms. Methods. We co-tested 913 previously unscreened women aged ≥30 years and applied WHO-recommended treatment for all VIA/VILI-DC-positive women. Baseline prevalence of HR-HPV and HIV were 24% and 42%, respectively. Results. On initial screen, 44 (5%) women were VIA/VILI-DC-positive, of whom 22 had HR-HPV infection, indicating 50% of women screened false-positive and would have been triaged for unnecessary same-day treatment. VIA/VILI-DC-positive women with HIV infection were three times more likely to be HR-HPV-positive than HIV-negative women (65% vs. 20%). All women positive for either VIA/VILI-DC or HR-HPV (n = 245) were invited for repeat co-testing after one year, of which 136 (56%) returned for follow-up. Of 122 women who were HR-HPV-positive on initial screen, 60 (49%) re-tested negative, of whom 6 had received treatment after initial screen, indicating that 44% of initially HR-HPV-positive women spontaneously cleared infection

⁎ Corresponding author at: Beckman Research Institute of City of Hope, 1500 E Duarte Road, Duarte, CA 91010, Fox South 2000D, United States. E-mail address: [email protected] (J.G. Ogembo). https://doi.org/10.1016/j.ygyno.2017.11.002 0090-8258/© 2017 Elsevier Inc. All rights reserved.

P. Cholli et al. / Gynecologic Oncology 148 (2018) 118–125

119

after one year without treatment. Women with HIV were more likely to remain HR-HPV-positive on follow-up than HIV-negative women (61% vs. 22%, p b 0.001). Treatment was offered to all VIA/VILI-DC positive women on initial screen, and to all women screening VIA/VILI-DC or HR-HPV positive on follow-up. Conclusions. We found careHPV co-testing with VIA/VILI-DC to be feasible and valuable in identifying falsepositives, but careHPV screening-to-result time was too long to inform same-day treatment. © 2017 Elsevier Inc. All rights reserved.

1. Introduction Although cervical cancer incidence and mortality have declined to historic lows in developed countries, less-developed regions continue to experience a high mortality burden, with N 230,000 women dying from the disease annually [1]. The introduction of cytological screening using the Papanicolaou (Pap) smear, allowing for early detection and treatment, is chiefly attributable for the steep decline in cervical cancer deaths in developed countries such as the United States, which has seen a N 70% reduction since its adoption [2,3]. Unfortunately, demands for materials, equipment, trained personnel, and infrastructure have prohibited cytology from being useful or cost-effective in resourcelimited settings [4,5], leading to the development of alternative, readily implementable screening techniques such as visual inspection with acetic acid (VIA) or Lugol's iodine (VILI) enhanced by digital cervicography (DC) (VIA/VILI-DC) [4,6]. Visual inspection methods are less resource-intensive, provide immediate results obtainable by midlevel providers (e.g. midwives, nurses), and demonstrate comparable or superior sensitivity to cytologic smears [4,7]. However, sensitivity of visualization methods is highly variable across studies and metaanalyses (VIA: 41%–79% [4] vs. 22%–74% [8] vs. 41%–92% [9]; VILI: 57%–98% [4]), depending on the quality of personnel training, experience, and supervision [4], as well as adjunctive quality assessment tools such as DC [10]. The utility of screening lies in connecting screen-positive women to treatment. To address the challenges and costs associated with loss to follow-up in resource-limited settings [11–13], the World Health Organization (WHO) outlined its “screen-and-treat” approach in 2006 whereby women testing-positive for cervical cancer or its precursors receive same-day treatment [14]. VIA/VILI's immediate results are advantageous in this context. However, the often-poor specificity and sensitivity of visual approaches may result in overtreatment or undertreatment [8,12]. Of the limited evidence available, complication rates of cryotherapy and loop electrosurgical excision procedure (LEEP) in low-resource settings are comparable to resource-rich settings, although adverse outcomes in low-resource settings are likely underreported [15]. While side effects of LEEP are generally mild, rare complications (e.g. severe bleeding, pelvic inflammatory disease) can pose a major risk in settings with limited access to medical care [16]. Additional limitations include the high prevalence of VIA/VILI-inadequate test results, in which the transformation zone (i.e. squamo-columnar junction) is not fully visualized (e.g. representing 12.2% and 22.0% of visual inspection results for programs in Nigeria and Cameroon [17,18], respectively). In light of these challenges, visual inspection in a singlevisit “screen-and-treat” protocol is not optimal [19], making co-testing with other modalities such as HPV DNA testing worth considering. The careHPV test, developed by QIAGEN (Germantown, MD), is a rapid test for 14 high-risk HPV (HR-HPV) genotypes designed to overcome logistical issues of traditional HPV DNA testing in the field, and may be an effective co-test alongside visual methods [20,21]. Invasive cervical cancer (ICC) is recognized as an AIDS-defining illness, with HIV-positive women at greater risk of infection with oncogenic HPV types and of developing cervical intraepithelial neoplasia (CIN) and ICC [22–24]. Few studies have explored the relationship between HIV infection and the performance of cervical cancer screening modalities, identified by the WHO as a research gap lacking quality

evidence [5,25]. This carries significant implications for sub-Saharan Africa, which in 2012 reported nearly 70% of all HIV infections worldwide [26]. Including HIV testing in screening algorithms for HIVendemic populations such as Cameroon may be a low-cost, readily implementable means of enhancing existing screening methods. Our team, comprised of the Cameroon Baptist Convention Health Services (CBCHS), University of Massachusetts Medical School (UMMS, Worcester), and University of Massachusetts College of Nursing and Health Sciences (UMass, Boston) evaluated the utility of VIA/ VILI-DC and careHPV co-testing among HIV-positive and HIV-negative women in Cameroon. Our goals were to 1) assess the applicability, feasibility, and outcomes of co-testing within a single-visit “screen-andtreat” approach, 2) determine 1-year HR-HPV persistence and clearance rates, and 3) explore how HIV status influences screening results, in order to inform optimal screening, treatment, and follow-up algorithms. 2. Materials and methods 2.1. Setting and study participants Institutional Review Boards of UMMS and CBCHS approved the study. careHPV kits and services for both initial and follow-up screening were provided through funding from the UMMS Office of Global Health, and additional follow-up careHPV tests were donated by QIAGEN. Initial screening was implemented between February–April 2015 at two CBCHS Women's Health Program (WHP) clinics, Baptist HospitalMutengene and Mboppi Baptist Hospital-Douala. Some participants were also recruited through mobile clinics near Mutengene. Recruitment was achieved through media announcements, word-of-mouth, church presentations, community gatherings, HIV care and treatment clinics, and outpatient clinics. Each patient received information about cervical cancer screening and treatment options for pre-cancer or cancer prior to enrollment. We targeted women aged ≥30 years (per careHPV DNA kit manufacturer's guidelines for FDA-approved intended use) who were not pregnant and had no history of past cervical cancer screening by CBCHS [21, 27]. For each consenting participant, a trained peer-educator or nurse administered a structured face-to-face interview and clinical questionnaire before screening, to collect medical history (including HIV status), sociodemographics (e.g. age, years of education, and occupation), and behavioral risk factors (e.g. age at sexual debut, number of lifetime sexual partners, and use of family planning methods). For all participants with negative or unknown HIV status, nurses reviewed CBCHS consultation records (kept and carried by patients) to confirm HIV testing in the previous six months. If their HIV status was unknown or their last HIVnegative test result had been N6 months prior, nurses counseled and referred women for HIV testing by a laboratory technician. To allow for active follow-up, participant contact information (cell phone numbers or permanent home address) was recorded at initial screening. 2.2. Cervical cancer co-testing protocol using both VIA/VILI-DC and careHPV Same-day co-testing with both careHPV and VIA/VILI-DC was performed alongside HIV testing (if indicated), a physical examination, and the four chief services of CBCHS's WHP: peer health education, family planning counseling, clinical breast exam, and management of

120


reproductive tract infections [28]. While WHO recommends co-testing with sequential screening, in which primary screening results inform decisions to follow-up with secondary screening, all women were cotested simultaneously to eliminate the possibility of verification bias. Nurses primarily performed VIA/VILI-DC in WHP clinics per clinical protocol, with results categorized and archived as previously described [29,30]. VIA-DC positive results were categorized as either cryotherapyeligible or LEEP-eligible, per WHO guidelines [31]. Concordance analysis between VIA-DC and VILI-DC interpretations was performed with Cohen's unweighted kappa and found significant agreement (k = 0.94). In rare discordance, VIA-DC results were used to guide treatment, follow-up, and analysis. The careHPV test detects 14 HR-HPV subtypes (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68) in cervical specimens, but does not provide type-specific results [21]. CBCHS staff experienced with cervical cancer screening procedures, including 8 nurses and 9 lab technicians, were trained in Cameroon by staff from QIAGEN on collecting and processing the careHPV test. Cervical swabs were collected by a nurse, labeled with a unique WHP ID number, frozen at − 20 °C, and either immediately tested or shipped to CBCHS Mutengene clinic for careHPV testing by laboratory technicians on-site. As it was not possible to provide same-day HPV results due to batching requirements (up to 90 samples needed to run a full plate), WHP nurses notified participants by phone or word of mouth to come to the clinic to receive their HPV test results as soon as they became available.

analysis. A total of 442 (48.4%) women received HIV testing at the time of screening. 3.2. Socio-demographics Table 1 summarizes socio-demographic characteristics of participating women. The mean age was 42 years (range: 30–80), with 707 (77%) women of child-bearing age and 28% between ages 30–35. Nearly half of the women sampled had b8 years of education; 20% had N 12 years of education. Women had a median of three lifetime sexual partners. About twice as many women did not use family planning methods as women who did (66% vs. 34%), with male condoms most commonly used (31% of all family planning users). Of all women screened, 42% (n = 384) were HIV-positive, of whom 87% were taking antiretroviral therapy. Only 194 of the women had CD4+ T-cells measured in the previous six months, of which 82% had CD4 cell counts N200 cells/ml. 3.3. VIA/VILI-DC and careHPV co-testing for HIV-positive and HIV-negative women

3. Results

Table 2 shows the prevalence of positive screen results for VIA/VILIDC and careHPV. Of all women screened, 4.8% (n = 44) were VIA/VILIDC-positive and 83.6% (n = 763) were VIA/VILI-DC-negative. An additional 10.8% (n = 99) of women screened VIA/VILI-DC-inadequate and 0.8% (n = 7) uncertain as previously defined [18]. The overall prevalence of HR-HPV infection was 24.4% (n = 223/913). Comparing test results by HIV status, 29 of the 384 HIV-positive women (8%) tested positive on the VIA/VILI-DC test, whereas 157 (41%) tested positive on the careHPV test (p b 0.0001). Excluding the 106 participants with VIA/VILI-DC-inadequate or uncertain results, the Cochrane-Armitage Trend Test showed no significant linear trend across age groups for either screening method (both p-values N 0.25). Of all VIA/VILI-DC-positive women, 50.0% were also HR-HPVpositive (n = 22/44), indicating that for HR-HPV subtypes detected by careHPV, half of the women with visible lesions lacked concomitant HR-HPV infection and would have been overtreated based on VIA-DC results alone. LEEP-eligible women were significantly more likely to be HR-HPV-positive (n = 6/7, 85.7%) than cryotherapy-eligible women (n = 16/37, 43.2%) (p = 0.0393). Prevalence of HR-HPV infection in VIA/VILI-DC-inadequate (n = 23/99, 23.2%) and VIA/VILI-DCnegative (n = 174/763, 22.8%) women were similar. Overall, HR-HPVpositive women were three times more likely to have VIA/VILI-DCpositive results than HR-HPV-negative women (n = 22/223, 9.9% versus n = 22/690, 3.2% of all VIA/VILI-DC positive results, p b 0.0001). Overall HIV prevalence was 42.1% (n = 384). VIA/VILI-DC-positive results were more than twice as prevalent in HIV-positive (n = 29/384, 7.6%) than HIV-negative (n = 15/530, 2.8%) women (p b 0.0001). Stratifying VIA/VILI-DC findings by HIV status, VIA/VILI-DCpositive women had a significantly higher prevalence of HIV infection (n = 29/44, 65.9%) than VIA/VILI-DC-negative (n = 321/763, 42.1%) and VIA/VILI-DC-inadequate women (n = 30/99, 30.3%), p b 0.0250). Overall, HIV-positive results were more than twice as prevalent in HRHPV-positive (n = 157/223, 70.4%) than HR-HPV-negative (n = 227/463, 32.9%) women (p b 0.0001). For each VIA/VILI-DC result category, HIV-positive women were at least 1.9 times more likely than HIVnegative women to test HR-HPV-positive, i.e., HR-HPV prevalence was 86.4% (n = 19/22) vs. 45.5% (n = 10/22) for VIA/VILI-DC-positive; 69.0% (n = 120/174) vs. 34.1% (n = 201/589) for VIA/VILI-DCnegative, 66.7% (n = 18/27) vs. 20.3% (n = 16/79) for VIA/VILI-DCinadequate or uncertain (Fig. 2).

3.1. Enrollment

3.4. Treatment uptake, referral, and follow-up screening

Enrollment (Fig. 1) was brisk, with approximately 900 of 1170 total women screened across two sites in the first week of study implementation. After all exclusions, a total of 913 women were eligible for

Due to careHPV batching limitations, HPV results were usually available more than a week after specimen collection, too late to inform same-day treatment. Therefore, same-day treatment and/or referral

2.3. Treatment, 1-year follow-up, and histology At initial screening, women were treated based on positive VIA/VILIDC results according to WHO guidelines [14]. Between June–July 2016, all women with VIA/VILI-DC-positive results and/or HR-HPV-positive results were invited for repeat VIA/VILI-DC and careHPV co-testing. If positive for any test on repeat screening, they were offered same-day treatment with cervical cryotherapy or thermal coagulation, or referral for LEEP. As same-day HR-HPV results were not available, women returned to receive them. All women with HR-HPV-positive results were advised to have appropriate treatment, even those screening VIA/VILI-DC-negative. Follow-up screening and treatment were free of charge. In both initial and follow-up screening, women with suspicious cervical lesions were biopsied and, if found to have ICC, subsequently referred to Mbingo Baptist Hospital for surgery, Douala General Hospital for radiotherapy, or palliative care. All biopsies and LEEP specimens underwent histologic review at Buea Regional Hospital (Dr. Orock), Cameroon, with additional blinded histologic review performed by Dr. Liu at the UMMS as a quality control measure. 2.4. Statistical analysis Descriptive statistics were calculated by HIV status. Continuous values were summarized using mean and standard deviation. Categorical variables were summarized using percentages. Prevalence of VIA/ VILI-DC-positive results, HR-HPV infection, and HIV co-infection were calculated. A Chi-squared test was used to compare proportions. Linear trends in prevalence across 10-year age groups were determined using the Cochrane-Armitage test for homogeneity. A two-sided p-value of b0.05 was considered statistically significant.


121

Fig. 1. Screening algorithm.

were offered based on VIA/VILI-DC results and careHPV results were reported to patients in a subsequent visit. Treatment was offered to all women who screened VIA/VILI-DC-positive. Of 44 VIA/VILI-DCpositive women, 37 were identified as eligible for cryotherapy or thermal coagulation and seven for LEEP. Of all VIA/VILI-DC-positive women, 40.9% (n = 18/44) received intervention, with 11 receiving same-day thermal coagulation and seven returning for follow-up LEEP (n = 5) or cryotherapy (n = 2). Of the 13 cryotherapy/thermal coagulation-eligible women who received treatment, eight (61.5%) were HR-HPV-positive, and of the remaining 25 with available data who were not treated, nine (36%) were HR-HPV-positive.

HIV-positive, and four (n = 4/5, 80%) were found to have a persistent HR-HPV infection. On follow-up screening, a total of 68 (n = 68/136, 50%) women qualified for treatment, consisting of those who tested positive for either VIA/VILI-DC or HR-HPV. Women who were VIA/VILI-DC-negative and HR-HPV-positive were advised to have cryotherapy or thermal coagulation. Of the 40 women who received treatment after re-screening (n = 40/68, 58.8%), five (n = 5/40, 12.5%) received same-day intervention (cryotherapy, biopsy), with the remaining women (n = 35/40, 87.5%) returning for cryotherapy, LEEP, or biopsy on a follow-up visit. Six (n = 6/40, 15%) of these treated women were VIA/VILI-DCpositive at baseline examination, but did not receive treatment until their follow-up.

3.5. Follow-up VIA/VILI-DC and HR-HPV results: changes from initial screening and treatment uptake 4. Discussion A total of 245 women initially screened positive for VIA/VILI-DC and/ or HR-HPV, and were therefore eligible for follow-up. Of these women, 136 returned for repeat co-testing at one year. Fig. 3 shows that of the 122 initially HR-HPV-positive women who returned for follow-up (including 15 who also screened VIA/VILI-DC positive), 62 re-tested HRHPV-positive (i.e. had persistent HR-HPV infection). Sixty of the 122 re-tested HR-HPV-negative (i.e. cleared HR-HPV infection), of whom six (10%) received treatment after initial screen, indicating 54 (44.3%) spontaneously cleared infection within one year. HR-HPV-persistent women were more likely to be HIV-positive than women who cleared HR-HPV infection (n = 55/62, 88.7% vs. n = 35/60, 58.3%; p b 0.0001). Of initially HR-HPV-positive women, those who were HIV-positive had a n HPV persistence rate nearly three times higher than HIV-negative women (61.1% vs. 21.9%; p b 0.001). Of all women who returned for screening due to an isolated HRHPV-positive result (i.e. did not screen VIA/VILI-DC-positive on initial screen), five (n = 5/107, 4.7%) converted from VIA/VILI-DC-negative to VIA/VILI-DC-positive within one year. All five of these women were

This longitudinal pilot study examined VIA/VILI-DC and careHPV cotesting results among HIV-positive and HIV-negative women. Our study highlights 1) the limits of utilizing visual inspection methods or HPV DNA testing alone to detect cervical cancer and its precursors, 2) how co-testing can effectively overcome the limitations of both tests (if both were to provide same day results), and 3) how adjunctive HIV testing may enhance the efficacy of both screening modalities. Perhaps most importantly, we demonstrated that co-testing with VIA/VILI-DC and careHPV can be effectively implemented and integrated with HIV testing in low-resource settings. Our results suggest relying exclusively on VIA/VILI-DC findings under a screen-and-treat model would result in a 50% overtreatment rate, subjecting these women to unnecessary procedures, associated personal and financial costs, and possible clinical risks. Pairing visual methods with HPV DNA co-testing can strengthen screening efforts by identifying false VIA/VILI-DC-positives, thereby reducing rates of overtreatment.

122


We found 44.3% of HR-HPV-positive women spontaneously cleared their infections in one year, indicating that treatment based solely on HR-HPV results at baseline would have led to rates of overtreatment

Table 1 Sociodemographic characteristics of women screened for cervical cancer and its precursors, HR-HPV, and HIV infection in CBCHS clinics. Characteristic

Number of Women (%)

Age, mean (stdev) Age brackets, n (%) 30–39 40–49 50–59 60+

41.84

(9.89)

442 265 156 50

(48.4) (29.0) (17.1) (5.5)

9.32

(4.04)

418 255 70 90 80

(45.8) (27.9) (7.7) (9.9) (8.8)

Years of education completed, mean (stdev) Highest education, n (%) b8 years 8–12 years 13–14 years 15+ years No data Occupation, n (%) Housewife Farmer Health care worker Secretary Trader Teacher Student Seamstress Domestic worker Hairdresser Other (business, CDC worker, police, banker, etc.) No data

135 100 44 15 295 89 6 51 12 38 96 32

(14.8) (11.0) (4.8) (1.6) (32.3) (9.7) (0.7) (5.6) (1.3) (4.2) (10.5) (3.5)

Age of first intercourse, median (range)

18

(16, 20)

Marital status, n (%) Married Single (never married) Widowed/separated/divorced Separated Divorced No data

527 183 142 28 24 9

(57.7) (20.0) (15.6) (3.1) (2.6) (1.0)

Husband has N1 wife, n (%) Yes Number additional wives, median (range) Number additional wives: 1, n (%) Number additional wives: 2, n (%) Number additional wives: 3, n (%) No No data

55 1 38 12 2 437 421

(6.0) (1-3) (73.1) (23.1) (3.9) (47.9) (46.1)

Partner has other partners, n (%) Yes No Unknown No current partner No data

90 293 306 136 88

(9.9) (32.1) (33.5) (14.9) (9.6)

Sexual partners… In lifetime, median (lower, upper quartile) In last 3 months, median (lower, upper quartile)

3 1

(2, 5) (1, 1)

Forced sex, n (%) Yes No No data

65 824 24

(7.1) (90.3) (2.6)

Sex for money or favors, n (%) Yes No Refused to answer No data

24 859 2 28

(2.6) (94.1) (0.2) (3.1)

Table 1 (continued) Characteristic

Number of Women (%)

Family planning methods Yes… n (%) …Method used, n (% of all methods) Abstinence Breastfeeding Withdrawal method Calendar method Male condoms Female condoms Depo Provera Birth control pills IUD Tubal ligation Implant Vasectomy Traditional methods Other No, n (%)

309

(33.8)

40 10 19 69 101 2 19 18 14 8 25 0 2 0 604

(12.2) (3.1) (5.8) (21.1) (30.9) (0.6) (5.8) (5.5) (4.3) (2.4) (7.6) (0.0) (0.6) (0.0) (66.2)

Gravida, median (lower, upper quartile) Para (number of births N 20 weeks)

4 3

(2, 6) (2, 5)

HIV results Positive Confirmed negative result in last 6 months

384 529

(42.1) (57.9)

CD4 count in last 6 months HIV positive, with available data CD4 b200 CD4 200–349 CD4 350–499 CD4 500+ HIV positive, no available data Mean CD4 count (stdev)

194 35 41 57 61 190 501.02

Currently taking antiretroviral therapy Yes No Unknown

334 15 35

(18) (21.1) (29.4) (31.4) (910.38)

(87) (3.9) (9.1)

which mirror overtreatment rates based on VIA/VILI-DC results alone. This high overtreatment rate echoes findings from another Cameroonian study that found HPV tests have a positive predictive value of 23.5% and a number needed to treat of 6 for positive atypical squamous cells of undetermined significance (indicating that among all women triaged for treatment based on HPV-positive results, only one of seven women would have a cytologic indication for treatment) [32]. The rate of overtreatment in our study differs considerably from overtreatment rates reported in WHO guidelines, which cite 15.7% overtreatment for HPV DNA testing and 12.7% for VIA screening [5,14,33]. However, it must be noted that definitions of false positives (and overtreatment) in the WHO guidelines are based on histologic standards (i.e. CIN2+), whereas Untiet et al. and this project used proxy variables for true infection to calculate overtreatment (i.e. cytology (32) vs. HR-HPV infection cleared in one year or positive visualization with negative HRHPV results, respectively). Our findings indicate that, due to their respective weaknesses, VIA/ VILI-DC and HPV DNA testing can be optimized for use in lowresource settings through a co-testing approach. Other studies support a sequential screening model, with HPV DNA testing as a primary screening tool to identify women at risk for cervical cancer or precancerous lesions, followed by VIA/VILI-DC as a point-of-care test to triage treatment [32,34]. In line with this, we found careHPV, which can be performed by trained technicians, to be less resource- and staffintensive in a mass-screening program than VIA/VILI-DC, which requires a dedicated diversion of nurses' time and resources to be performed on a large scale. As HPV prevalence was 24% among all women screened, using HPV DNA testing as a primary screening tool


123

Table 2 VIA-DC and HR-HPV results by age group on initial screen. 30–39 (n)

%

40–49 (n)

%

50–59 (n)

%

60+ (n)

%

Total (n)

%

VIA-DC results Positive Negative Inadequate or uncertain Total

26 403 13 442

5.9 91.2 2.9

12 223 30 265

4.5 84.2 42.0

5 109 42 156

3.2 69.9 26.9

1 28 21 50

2.0 56.0 42.0

44 763 106 913

4.8 83.6 11.6

HR-HPV results Positive Negative Total

115 327 442

26.0 74.0

61 204 265

23.0 77.0

38 118 156

24.4 75.6

9 41 50

18.0 82.0

223 690 913

24.4 75.6

could reduce the logistical burden of screening by about 75%, allowing for more efficient, targeted utilization of staff and resources. We found that co-testing accommodated high demand and rapid turnover, with 900 women successfully screened in the first week alone. However, despite its applicability in low-resource settings, careHPV's 90-sample batching requirement led to a prolonged screening-to-results time, precluding it from informing same-day treatment. To overcome this, HPV DNA self-sampling is another approach found to be comparable in efficacy to clinician-collected samples with high acceptability [32,34–36]. Bigoni et al. evaluated the efficacy of VIA as a secondary point-of-care test following HPV DNA self-sampling by comparing it to the histologic gold standard, demonstrating that such a sequential screening protocol is implementable [34]. Our study co-tested women with known HIV status, effectively “triple-screening” participants and revealing insights into both VIA/VILI-DC and HPV DNA testing. We found significant differences in VIA/VILI-DC overtreatment rates (i.e. positive visual lesion but HR-HPV-negative) among HIV-positive and HIV-negative women (34% vs. 80%, respectively). Of HIV-positive women who were VIA/VILI-DC positive, 66% were also HR-HPV-positive, compared to 20% among HIV-negative women, indicating acetowhite lesions detected by VIA/VILI-DC may be more likely related to HPV infection if a woman is HIV-positive. Therefore, HIV-positive women may have lower rates of overtreatment, if visual inspection is used as a lone modality under same-day screen-andtreat. While further studies are needed to assess the relationship between HIV status, VIA/VILI-DC, and HR-HPV infection, our findings

suggest VIA/VILI-DC sensitivity may be higher in HIV-positive women, and HIV screening may have clinical utility as an adjunct test alongside visual inspection—especially in settings where HPV DNA testing may be impractical as a primary screening test. We found potential value in considering HIV status when using HPV DNA testing to triage women for further screening or treatment. Treatment by positive HR-HPV results alone, irrespective of HIV status, would have led to an overtreatment rate of 44.3%, with overtreatment defined as same-day treatment for an isolated infection that spontaneously cleared within one year. Analyzing subsets of HIV-positive and HIV-negative women, HRHPV-positive women with concomitant HIV infection had a lower rate of overtreatment at 38.9%. In contrast, HR-HPV-positive, HIV-negative women were found to have more than twice (78.9%) the rate of overtreatment. There was also a strong correlation between HIV status and HR-HPV persistence after 1 year (61.1% for HIV-positive vs. 21.9% for HIV-negative women), with implications for surveillance. Our programmatic evaluation provides a longitudinal glimpse into trends of HR-HPV clearance and persistence. Few studies have longitudinally assessed screening results through follow-up co-testing or successfully collected treatment data after screening to analyze rates of same-day treatment uptake and referral. We achieved a 57.6% rescreening rate among women at higher risk for invasive cervical cancer and its precursors by virtue of a positive VIA/VILI-DC or HR-HPV result on initial screening. Reducing loss to follow-up is a significant consideration guiding the structure of screening programs in low-

Fig. 2. Association between VIA/VILI-DC, HR-HPV, and HIV status.

124


Fig. 3. Follow-up HPV and HIV* results.

resource settings [13]. It is noteworthy that our 1-year rescreening program recaptured 23 of 37 initially cryotherapy-eligible women (of whom 10 had received same-day treatment), including 52.0% of cryotherapy-eligible women who were initially untreated and unreferred, as well as six out of seven initially LEEP-eligible women (of whom four had received treatment). Our study is unique in that it was conducted within the financial and resource constraints of a mass screen-and-treat program in a lowresource setting, and therefore offers direct insight into the applicability and sustainability of implementing VIA/VILI-DC/HPV DNA co-testing with treatment in the field. For instance, although two previous studies in Cameroon utilized self-collected HPV DNA tests to evaluate sequential screening and triage in Cameroon, both shipped samples to Switzerland for analysis, leading to prolonged delays from sample collection to analysis (84 and 28 days, respectively) [32,34]. This resource-intensive process may not be feasible for mass-screening programs in low-resource settings without external funds, raising questions about practicality and sustainability. Moreover, material degradation in one study caused by specimen transport and storage contributed to high rates of inconclusive HPV results [34]. Delays and compromised results can have implications for loss to follow-up, treatment uptake, and community trust, especially in restrictive settings where effective outreach is paramount. We collected and analyzed HPV DNA samples alongside VIA/VILI-DC in the field, with no inconclusive results and timelier follow-up and treatment. Our study was an evaluation of a clinical-preventive program rather than a rigorously designed study, and showed limitations in project design. Participant ages were neither normally distributed nor sufficiently powered, and comparable populations within each age group were not available for comparison. Unlike other studies in Cameroon [32,34,37], we did not perform cytology or biopsies (the gold standard) to assess the sensitivity and specificity of VIA/VILI-DC and careHPV. Despite efforts to ensure treatment uptake, there was a low rate of same-day treatment among VIA/VILI-DC-positive women. WHP nurses reported the main barrier to acceptance of same-day treatment was women's desire to consult their spouses, with others preferring to wait for their HPV results before consenting to treatment. Although same-day treatment uptake was low, we were able to successfully record each woman's treatment, identify those who were not treated, and invite all women who were indicated for follow-up for re-screening after one year.

Treatment uptake improved on follow-up screening, reaching a confirmed uptake of 52.9% for women screening either HR-HPV or VIA/ VILI-DC positive. To further increase treatment rates, the WHP is aiming to increase sensitization of communities regarding benefits of same-day treatment and is developing a cloud-based electronic medical record for use across all WHP clinics to more systematically monitor and follow up untreated women. Given the demonstrable mortality benefits of cervical cancer screening in low-resource settings, there is an urgent need to introduce effective, implementable, and sustainable screening methodologies [19]. While our study shows a promising link between HIV screening and VIA/VILI-DC and HPV DNA testing, further studies are needed to evaluate how HIV status can be used as an adjunct test alongside traditional screening methods to enhance triage, treatment, and follow-up. Given the impressive uptake of careHPV in our first week of screening, further technological development of low-cost, simple, rapid HPV tests that would enable immediate results (e.g. within 15 minutes) would overcome many of the limitations of batched HPV testing. Conflict of interest The authors have no conflicts of interest to disclose. References [1] J. Ferlay, I. Soerjomataram, M. Ervik, GLOBOCAN 2012: Cancer Incidence and Mortality Worldwide: IARC Cancer Base No. 10 [Internet], International Agency for Research on Cancer, 2012. [2] D. Saslow, C.D. Runowicz, D. Solomon, A.B. Moscicki, R.A. Smith, H.J. Eyre, C. Cohen, S. American Cancer, American Cancer Society guideline for the early detection of cervical neoplasia and cancer, CA Cancer J. Clin. 52 (2002) 342–362. [3] P.A. Wingo, C.J. Cardinez, S.H. Landis, R.T. Greenlee, L.A. Ries, R.N. Anderson, M.J. Thun, Long-term trends in cancer mortality in the United States, 1930–1998, Cancer 97 (2003) 3133–3275. [4] J. Sherris, S. Wittet, A. Kleine, J. Sellors, S. Luciani, R. Sankaranarayanan, M.A. Barone, Evidence-based, alternative cervical cancer screening approaches in low-resource settings, Int. Perspect. Sex. Reprod. Health (2009) 147–152. [5] WHO, Comprehensive Cervical Cancer Control: A Guide to Essential Practice, Second edition, 2014 (Australia). [6] L. Denny, L. Kuhn, M. De Souza, A.E. Pollack, W. Dupree, T.C. Wright, Screen-andtreat approaches for cervical cancer prevention in low-resource settings: a randomized controlled trial, JAMA 294 (2005) 2173–2181. [7] R. Sankaranarayanan, P. Basu, R.S. Wesley, C. Mahe, N. Keita, C.C. Mbalawa, R. Sharma, A. Dolo, S.S. Shastri, M. Nacoulma, M. Nayama, T. Somanathan, E. Lucas, R. Muwonge, L. Frappart, D.M. Parkin, Detection IMSGoCCE, Accuracy of visual


[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16] [17]

[18]

[19]

[20]

[21]

screening for cervical neoplasia: results from an IARC multicentre study in India and Africa, Int. J. Cancer 110 (2004) 907–913. J. Jeronimo, P. Bansil, J. Lim, R. Peck, P. Paul, J.J. Amador, F. Mirembe, J. Byamugisha, U.R. Poli, L. Satyanarayana, S. Asthana, Group S-US, A multicountry evaluation of careHPV testing, visual inspection with acetic acid, and papanicolaou testing for the detection of cervical cancer, Int. J. Gynecol. Cancer 24 (2014) 576–585. C. Sauvaget, J.M. Fayette, R. Muwonge, R. Wesley, R. Sankaranarayanan, Accuracy of visual inspection with acetic acid for cervical cancer screening, Int. J. Gynaecol. Obstet. 113 (2011) 14–24. C. Firnhaber, L. Mao, S. Levin, M. Faesen, D.A. Lewis, B.J. Goeieman, A.J. Swarts, N. Rakhombe, P.M. Michelow, S. Williams, J.S. Smith, Evaluation of a cervicographybased program to ensure quality of visual inspection of the cervix in HIV-infected women in Johannesburg, South Africa, J. Low. Genit. Tract Dis. 19 (2015) 7–11. J. Thinkhamrop, P. Lumbiganon, S. Jitpakdeebodin, Loss to follow-up of patients with abnormal pap smear: magnitude and reasons, J. Med. Assoc. Thail. 81 (1998) 862–865. R. Sankaranarayanan, R. Rajkumar, R. Theresa, P.O. Esmy, C. Mahe, K.R. Bagyalakshmi, S. Thara, L. Frappart, E. Lucas, R. Muwonge, S. Shanthakumari, D. Jeevan, T.M. Subbarao, D.M. Parkin, J. Cherian, Initial results from a randomized trial of cervical visual screening in rural south India, Int. J. Cancer 109 (2004) 461–467. J.D. Goldhaber-Fiebert, L.E. Denny, M. De Souza, T.C. Wright Jr., L. Kuhn, S.J. Goldie, The costs of reducing loss to follow-up in South African cervical cancer screening, Cost. Eff. Resour. Alloc. 3 (2005) 11. WHO, Guidelines for Screening and Treatment of Precancerous Lesions for Cervical Cancer Prevention: Supplemental Material: GRADE Evidence-to-recommendation Tables and Evidence Profiles for each Recommendation, WHO, Geneva, 2013. E. Chamot, S. Kristensen, J.S. Stringer, M.H. Mwanahamuntu, Are treatments for cervical precancerous lesions in less-developed countries safe enough to promote scaling-up of cervical screening programs? A systematic review, BMC Womens Health 10 (2010) 11. W. Castro, J. Gage, L. Gaffikin, C. Ferreccio, J. Sellors, Effectiveness Safety and Acceptability of Cryotherapy: A Systematic Literature Review, 2003. K.O. Ajenifuja, J.C. Gage, A.C. Adepiti, N. Wentzensen, C. Eklund, M. Reilly, M. Hutchinson, R.D. Burk, M. Schiffman, A population-based study of visual inspection with acetic acid (VIA) for cervical screening in rural Nigeria, Int. J. Gynecol. Cancer 23 (2013) 507–512. G.A. DeGregorio, L.S. Bradford, S. Manga, P.M. Tih, R. Wamai, R. Ogembo, Z. Sando, Y. Liu, C. Schwaiger, S.R. Rao, K. Kalmakis, L. Kennedy Sheldon, K. Nulah, E. Welty, T. Welty, J.G. Ogembo, Prevalence, predictors, and same day treatment of positive VIA enhanced by digital cervicography and histopathology results in a cervical cancer prevention program in Cameroon, PLoS One 11 (2016), e0157319. S.J. Goldie, L. Gaffikin, J.D. Goldhaber-Fiebert, A. Gordillo-Tobar, C. Levin, C. Mah, T.C. Wright, Cost-effectiveness of cervical-cancer screening in five developing countries, N. Engl. J. Med. 353 (2005) 2158–2168. W. Chen, J. Jeronimo, F.H. Zhao, Y.L. Qiao, M. Valdez, X. Zhang, L.N. Kang, P. Bansil, P. Paul, P. Bai, R. Peck, J. Li, F. Chen, M.H. Stoler, P.E. Castle, The concordance of HPV DNA detection by Hybrid Capture 2 and careHPV on clinician- and self-collected specimens, J. Clin. Virol. 61 (2014) 553–557. Y.L. Qiao, J.W. Sellors, P.S. Eder, Y.P. Bao, J.M. Lim, F.H. Zhao, B. Weigl, W.H. Zhang, R.B. Peck, L. Li, F. Chen, Q.J. Pan, A.T. Lorincz, A new HPV-DNA test for cervicalcancer screening in developing regions: a cross-sectional study of clinical accuracy in rural China, Lancet Oncol. 9 (2008) 929–936.

125

[22] S. de Sanjose, J. Palefsky, Cervical and anal HPV infections in HIV positive women and men, Virus Res. 89 (2002) 201–211. [23] M. Maiman, R.G. Fruchter, M. Clark, C.D. Arrastia, R. Matthews, E.J. Gates, Cervical cancer as an AIDS-defining illness, Obstet. Gynecol. 89 (1997) 76–80. [24] D. Serraino, P. Carrieri, C. Pradier, E. Bidoli, M. Dorrucci, E. Ghetti, A. Schiesari, R. Zucconi, P. Pezzotti, P. Dellamonica, S. Franceschi, G. Rezza, Risk of invasive cervical cancer among women with, or at risk for, HIV infection, Int. J. Cancer 82 (1999) 334–337. [25] J. Fokom-Domgue, C. Combescure, V. Fokom-Defo, P.M. Tebeu, P. Vassilakos, A.P. Kengne, P. Petignat, Performance of alternative strategies for primary cervical cancer screening in sub-Saharan Africa: systematic review and meta-analysis of diagnostic test accuracy studies, BMJ 351 (2015) h3084. [26] UNAIDS, Global Report 2012: UNAIDS Report on the Global AIDS Epidemic, ebookpartnership.com 2013. [27] L. Massad, M. Einstein, W. Huh, H. Katki, W. Kinney, M. Schiffman, D. Solomon, N. Wentzensen, H. Lawson, ASCCP Consensus Guidelines Conference. 2012 updated consensus guidelines for the management of abnormal cervical cancer screening tests and cancer precursors, J. Low. Genit. Tract Dis. 17 (2013) S1–27. [28] G. DeGregorio, S. Manga, E. Kiyang, F. Manjuh, L. Bradford, P. Cholli, R. Wamai, R. Ogembo, Z. Sando, Y. Liu, L.K. Sheldon, K. Nulah, T. Welty, E. Welty, J.G. Ogembo, Implementing a fee-for-service cervical cancer screening and treatment program in Cameroon: challenges and opportunities, Oncologist 22 (2017) 850–859. [29] S. Manga, G. Parham, N. Benjamin, K. Nulah, L.K. Sheldon, E. Welty, J.G. Ogembo, L. Bradford, Z. Sando, R. Shields, T. Welty, Cervical cancer screening in Cameroon: interobserver agreement on the interpretation of digital cervicography results, J. Low. Genit. Tract Dis. 19 (2015) 288–294. [30] G. DeGregorio, E. Welty, T. Welty, S. Manga, J.G. Ogembo, L.S. Bradford, Evaluation of see-and-treat cervical cancer screening in human immunodeficiency virus–positive and human immunodeficiency virus–negative women in Cameroon [332], Obstet. Gynecol. 125 (2015) 105S. [31] WHO Guidelines for Screening and Treatment of Precancerous Lesions for Cervical Cancer Prevention. , World Health Organization, 2013. [32] S. Untiet, P. Vassilakos, C. McCarey, P.M. Tebeu, G. Kengne-Fosso, P.A. Menoud, M. Boulvain, I. Navarria, P. Petignat, HPV self-sampling as primary screening test in sub-Saharan Africa: implication for a triaging strategy, Int. J. Cancer 135 (2014) 1911–1917. [33] WHO, Comprehensive Cervical Cancer Prevention and Control: A Healthier Future for Girls and Women, WHO, Geneva, 2013. [34] J. Bigoni, M. Gundar, P.M. Tebeu, A. Bongoe, S. Schafer, J. Fokom-Domgue, R. Catarino, E.F. Tincho, S. Bougel, P. Vassilakos, P. Petignat, Cervical cancer screening in sub-Saharan Africa: a randomized trial of VIA versus cytology for triage of HPVpositive women, Int. J. Cancer (2015)https://doi.org/10.1002/ijc.29353. [35] P. Petignat, D.L. Faltin, I. Bruchim, M.R. Tramer, E.L. Franco, F. Coutlee, Are selfcollected samples comparable to physician-collected cervical specimens for human papillomavirus DNA testing? A systematic review and meta-analysis, Gynecol. Oncol. 105 (2007) 530–535. [36] P.E. Gravitt, J.L. Belinson, J. Salmeron, K.V. Shah, Looking ahead: a case for human papillomavirus testing of self-sampled vaginal specimens as a cervical cancer screening strategy, Int. J. Cancer 129 (2011) 517–527. [37] P.M. Tebeu, J. Fokom-Domgue, V. Crofts, E. Flahaut, R. Catarino, S. Untiet, P. Vassilakos, P. Petignat, Effectiveness of a two-stage strategy with HPV testing followed by visual inspection with acetic acid for cervical cancer screening in a low-income setting, Int. J. Cancer 136 (2015) E743–750.