The health district of Koumpentoum is located in the region of Tambacounda. In 2021, the population was 170,418 inhabitants on a surface area of 7652 Km², or a density of 22.27 inhabitants/Km². The district had 23 health care and service delivery points, including 1 reference health centre and 22 health posts. A little more than half, i.e. 51% of the population lived less than 5 km from a health facility, 35% between 5 and 15 km and 14% more than 15 km away.

This was a retrospective, descriptive and analytical study based on a database collected between 1^st January 2020 and 31^th December 2021 in all health care delivery points in the health district.

The study population consisted of women aged between 18 and 69 years targeted for screening at the district health centre or 22 health posts. The target population for screening was women aged between 30 and 69 years, 25,667 of whom 1907 were in urban areas and 23,760 in rural areas. The target for women aged between 18 and 69 was 37,870, of which 2813 were in urban areas and 35,057 in rural areas. The economic and social situation of the department is characterized by a low standard of living and a high rate of early marriage linked to socio-cultural realities; this justifies the practice of LPC screening from the age of 18 years [14].

All women aged between 18 and 65 years who were screened for cervical cancer at the centre and at the (22) district health posts and whose data were correctly and completely recorded were included.

Women screened whose data were not complete and correctly recorded were not included. Four women aged between 18 and 65 years with metastatic invasive cervical cancer at the time of screening were not included.

We conducted a comprehensive recruitment of women screened by VIAA/VIL from 1^st January 2020 to 31^th December 2021 and meeting the selection criteria. A total of 385 women aged between 18 and 65 years were included.

The dependent variable was the presence of PCLs at VIAA/VIL. The independent variables were socio-demographic data, medical history, macroscopic data after VIA/VILI and therapeutic aspects.

We used the LPC screening register to feed the electronic data recording platform.

The data were collected with the Excel workbook 97-2003. They were analyzed with Epi Info 7.2.5.0, SPSS 20 and R 4.0.5 softwares. In the descriptive part, the quantitative variables were described with the extremes, the means and their standard deviations while the qualitative variables were described with their frequencies surrounded by their confidence intervals (IC 95%). In the bivariate analysis, the Pearson’s chi2 statistical test and Fisher’s test under the conditions of applicability were used to verify the existence of a statistically significant relationship. All independent variables with a p < (0.25) in the bivariate analysis and those for which the literature review revealed a link with the dependent variable. The bottom-up stepwise approach allowed us to retain the independent variables with a statistically significant association with the presence of PCLs (p < 5%). The independent variables were then removed after a nested model comparison by the Aikake information criterion (AIC) [15] until no improvement was found by the maximum likelihood test. The Hosmer-Lemoshow test [16] was used to check the conformity of the final model. The strength of the relationship between the independent and dependent variables was assessed using the crude OR and the adjusted OR with their 95% CIs.

The research protocol had been submitted to the health authorities. The survey was explained in the local language and for each participant the free and informed consent was obtained. The data collected was kept anonymous and in compliance with medical secrecy.

In this study, 385 women aged between 18 and 65 years were interviewed. The minimum age was 18 years and the maximum age was 65 years. The average age was 30.07 ± 9.68 years and 89.8% were living in a couple. More than one third of the women (33.77%) were screened in urban areas and 73.25% of them were under 40 years old. Only 8.83% of the women were engaged in an income-generating activity (IGA) and 20.52% of them had attended school in French (Table 1).

The proportion of women aged between 30 and 69 years screened was 0.92% and varied according to their place of residence. In fact, it was 3.93% in urban areas

compared to 0.68% in rural areas. The same applies to the proportion of women aged between 18 and 69 years screened in health facilities, which was 1.01% and also varied according to place of residence, with 4.62% in urban areas and 0.72% in rural areas (Table 1). Most half (50.13%) of the women screened had more than five children and 29.35% of them had sexually transmitted infections (STIs) on speculum examination. At the VIA/VILI test, 5.45% of the women had precancerous cervical lesions. Of the women who tested positive, 90.48% received cryotherapy and 9.52% were referred for colposcopy and further management for suspected invasive cervical cancer (Table 2).

In bivariate analysis, the factors statistically significantly associated with the presence of PCL were: age ≥40 years (OR = 7.84 (2.95 - 20.82); p < 0.001) and history of STI (OR = 8.8 (3.14 - 24.7); p < 0.001) (Table 3).

In multivariate analysis, the factors significantly associated with the presence of PCL were: age ≥40 years (ORa = 17.52 (4.51 - 67.14); p < 0.001) and history of STI (ORa = 10.15 (3.46 - 29.85); p < 0.001) (Table 4).

The practice of screening among women aged 30 to 69 is 0.92% in our sample with variations according to place of residence with 3.93% in urban areas against 0.68% in rural areas. Thiam found a similar proportion of 2% in Kédougou [17]. However, higher proportions of screening were found in Kaffrine by Faye and et al. with 6.5% [18] and by Guéye and et al. in Joal with 23.2% [19]. In Tanzania, Cunningham and al. found a screening rate of 6% (n = 576) [20], while Adrianampy HA et al. [21] found a low proportion of women screened in Fianarantsoa, Madagascar (0.47%). Overall, screening is higher in urban than in rural areas, probably due to the financial and geographical accessibility of health facilities and services.

The frequency of PCLs in our sample was 5.45%, which is in line with that found in Cameroon by Ngowa JDK and et al. with 6.4% (n = 421) [22], Tebeu PM and et al. with 5.1% (n = 2485) [23] and Ngekoum B et al. with 7.4% (n = 946) [24] and in Madagascar by Adrianampy HA et al. with 8.5% (n = 112) [21]. On the other hand, higher frequencies were observed in other African series, notably in that of Biaye B with 10.2% in Dakar (n = 899) [25], Makoussa D and al. in Pointe Noire with 15.36% (n = 1437) [26] and Keita M and et al. with 38.1% in Bamako (n = 3302) [27]. These differences could be explained by the experience of the provider, by the technical quality of the screening and by the different sample sizes.

In our sample, the frequency of STIs was 29.35% according to the physical examination with the speculum. This frequency is clearly higher than the national average obtained by survey 11.9% [28], but lower than the series by Ly et al. [29] with 37%. This high frequency of STIs could be explained by the syndromic approach used in primary health care, which is based on algorithms. Indeed, without the need for biological evidence, the syndromic approach makes it possible to make a presumptive diagnosis of an STI and to initiate antibiotic treatment [30]. This also illustrates a notable lack of awareness of STI prevention in the community.

Factors significantly associated with the presence of PCLs were: age ≥40 years (ORa = 17.52 (4.51 - 67.14); p < 0.001) and STI history (ORa = 10.15 (3.46 - 29.85); p < 0.001). These results are consistent with the scientific literature. Indeed, the persistence of HPV infection is an almost indispensable condition in the genesis of precancerous cervical lesions [30].

We did not find any link between parity and the presence of PCLs (ORa = 0.16 (0.13 - 1.3); p = 0.427), contrary to the scientific literature where trauma to the cervix during childbirth is considered to be a risk factor for cervical cancer [31] [32] [33]. Indeed, if chronic HPV infection is considered as an essential element in the genesis of PCL, there are many promoter factors necessary for the persistence or even the evolution of cellular abnormalities, notably the precocity of sexual intercourse before 18 years due to the immaturity of the squamocolumnar junction (SCJ), multiple sexual partners, age of the first child multi parity due to the numerous traumas of the cervix during childbirth, poor genital hygiene, immunodeficiency due to HIV or immunosuppressants, smoking, co-infection with HIV, co-infection with another STI, in this case herpes simplex virus and chlamydia trachomatis, prolonged use of oral contraceptives, illiteracy and the low socio-economic level of women [31] [32] [33].

Database was originally designed for administrative purposes only and did not include all relevant elements, including all risk factors for cervical cancers. This was a limitation in the comprehensive identification of factors associated with the presence of LPC.