Population and design
A repeated cross-sectional study was conducted in Abidjan, the economic capital of Côte d’Ivoire where the great majority of ICC were diagnosed in the country, during the 2009-2011 and 2018-2020 periods. During two 24-month periods (May 2009 to June 2011 and July 2018 to June 2020), clinical wards located in the urban area of Abidjan, known to manage women with gynaecologic malignancies were asked to include all adult women attending with a suspected or confirmed diagnosis of ICC. Cervical biopsies and histological examination by local pathology unit were systematically proposed and financially supported by the research project, when appropriate. During the 2009-2011 period, only wards from the three public referral hospitals were involved as previously reported in a first large case-referent study on Cancer and HIV conducted in West Africa [3]. The 2018-2020 period covered all wards potentially managing ICC from the public and private sector in the urban area of Abidjan.
Collected information
Women enrolled during the two time periods were administered a similar structured questionnaire to collect socio-demographic characteristics including age, formal education (categorized as no school, primary school, secondary school and over), personal monthly income, age at first sexual intercourse, parity, tobacco use (categorized s current or former tobacco use versus never users) and current hormonal contraceptive use. Cancer clinical stage at ICC diagnosis was assessed based on the International federation of gynaecology and obstetrics (FIGO) staging system [15]. Based on available information after the initial assessment of the tumour extension, clinical stage at diagnosis was reported by clinicians and dichotomised as early (stage I and II) or advanced (stage III and IV) disease.
Additional information was collected during the 2018-2020 period including the existence of any personal health insurance coverage. Pre-diagnosis history was also documented including date of first reported ICC-related symptoms, date of first consultation at an ICC referral centre, whether ICC was diagnosed following attendance to a systematic ICC screening without prior symptoms or not and attendance to a traditional healer or using any traditional treatment for ICC-related symptoms prior to diagnosis.
During these two periods, a rapid HIV test (Determine®, Abbott Diagnostics) was systematically performed by collecting capillary blood by a finger prick test at the time of interview. In case of positive result, a venous blood sample was collected for confirmation purposes, according to the national algorithm of Côte d’Ivoire. Participants with a previously known HIV infection were surveyed with regards to their HIV characteristics including their date of first HIV diagnosis, ART use, last known CD4 count (and last known HIV viral load measure, only for the 2018-2020 period). These HIV-related data were collected combining participants interview with specific data request to HIV programs following participants with a previously documented HIV infection.
The present research have been performed in accordance with the Declaration of Helsinki and must have been approved by the national ethic committee of Côte d’Ivoire [n°011-19/MSHP/CNESVS-kp]. All women enrolled in 2009-2011 and 2018-2020 provided their informed and written consent prior to participate.
Statistical analysis
Participants characteristics were compared according to the presence or absence of HIV infection and between studied periods using Pearson’s χ2 test or Fisher’s exact test when appropriate for categorical variables and Mann-Whitney test or Kruskall-Wallis test for continuous variables. An unconditional logistic regression model was used to estimate the association between HIV infection and participant characteristics in women diagnosed with ICC during the 2018-2020 period. Odds ratio (OR) estimates were reported with their 95% confidence interval (95% CI). A multivariate model was computed following a stepwise descending procedure. Available factors associated with a p-value<0.2 were systematically included in a full model. Additional relevant potential confounders known to be associated with HIV or ICC such as tobacco use, oral contraceptive use or socio-economic status (education, income) were also considered regardless of their statistical association and included in the initial multivariate model. The goodness of fit of the model was then assessed using the Akaike Information Criterion (AIC), a lower value of the AIC suggesting a better prediction of the model. Confounders that were not significantly associated with HIV infection and did not add any significant prediction to the model based on the AIC were sequentially removed. All statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc, Cary, NC, USA).