The present study took place in the town of Parakou in north-eastern Benin. It was carried out on 23 prevention of mother-to-child transmission (PMTCT) sites and 7 care sites in the town, involved in the screening, care and/or follow-up of people living with HIV (PLHIV). Parakou is a cosmopolitan town located 415 km from the capital, Cotonou. The total population was estimated at 255,478 with a sex ratio of 0.99. Of this young population, 41.16% were children under the age of 15, while 51.97% were aged 18 and over. In 76% of cases, men were the head of household [5] . HIV seroprevalence was 0.4% in 2021 [6] .

This was a descriptive, analytic, cross-sectional study conducted over a 6-month period from July to December 2022. Participants were selected according to the following criteria: women or men aged 18 and over; living with HIV for at least 6 months; whose spouse’s status was known to health care services; in or having been in a stable union; received at one of the data collection sites during the study period and having given informed consent to the study. Patients physically or mentally unable to answer the questions and cases of withdrawal of consent were excluded from this study. We used non-probability sampling for convenience. The Schwartz minimum sample size was 345 subjects. Quantitative and qualitative data on participants’ lives were collected using a questionnaire, focusing on socio-demographic aspects, antecedents, behaviours, marital and psycho-social aspects presented by the participants. Clinical and therapeutic data were collected from patients’ personal records. The dependent variable, represented by serodiscordance within HIV+ couples, was defined as a couple in which one partner was HIV-infected and the other was not.

The data collected were analyzed using R 4.3.0 software. Quantitative variables were expressed as means with standard deviations, and qualitative variables as proportions. The Chi-square test was used to compare categorical variables with numbers > 5, and the Fisher test for variables with numbers ≤ 5. The Student’s T-test was used to compare means. The association between the two variables was established for a value of p < 0.05. A threshold of 20% was used to select variables for inclusion in a logistic regression model to identify potential predictors of HIV serodiscordance among infected couples in the city of Parakou.

The average age of respondents was 38.17 ± 10.25 years. Females were the most represented with a proportion of 79.62%, in other words, a sex ratio of 0.26. Most of the respondents (309 subjects, in other words. 74.11%) were married, and 83.69% of them were in an active relationship (married, cohabiting, couple). The average duration of the relationship was 130.75 ± 101.67 months. The minimum duration was 7 months and the maximum was 520 months. Monogamous unions were the most common, accounting for 67.87% of couples. The respondent was head of household among only 140 couples (33.57%). Nearly a third of the respondents in our study had a primary or secondary education (136 respondents, or 32.61% of the sample). The average income of the respondents was 38,349 F CFA ± 43,443 with extremes of 0 and 340,000F. The urban environment was the most represented (88.73% of cases, in other words. 370 respondents).

All subjects included in our study were infected with HIV1 (100% of cases). Liver serology was made on 357 respondents, in other words, 85.61% of cases. Of these, 37 (10.36%) were co-infected with HIV and viral hepatitis, including 9.52% with hepatitis B. Nearly one patient in eight (51, 12.23%) had a history of STI. During the study period, the last children of those surveyed were aged between 1 and 5 years in the majority of cases, 171 (44.65%). Their serological status, for those known, was negative in 95.29% of cases.

The vast majority of respondents (99.28%) were aware of at least one route of transmission and at least one means of HIV prevention. Among these routes of transmission, sexual and blood transmission were the most familiar to 410 (98.32%) and 312 (74.82%) respondents respectively. Vertical transmission was only known to 165 respondents, representing 39.57% of the total. As for known means of prevention, 387 (92.81%) knew about condoms, followed by abstinence (262 or 62.83%) and faithfulness (197 or 47.24%). Most of the PLHIV surveyed kept themselves informed about HIV, and 81.05% were actively researching the subject. The most common sources of information were health workers (84.61% of cases), television (40.53% of cases) and social networks (23.67% of cases).

The 417 subjects surveyed in our study included 299 who were involved in an HIV serodiscordant relationship, while 118 were in a couple with a serodiscordant positive partner. The prevalence of HIV serodiscordance among couples affected in the city of Parakou in 2022 was therefore 71.70% (IC95% [67.07 - 75.92]). It was 70.91% [65.54 - 76.28%] at PEC sites and 73.24% [65.95% - 80.52%] at PTME sites.

According to the results of our study, there is a statistically significant difference between HIV serodiscordance and:

● Socio-demographic variables: marital status and household head status (Table 1).

● Antecedents: duration of the couple’s relationship, serostatus of the last child and desire for (other) children (Table 2).

● None of the clinical or therapeutic variables studied in our study (Table 3).

● Behavioral variables: occasional sexual partnership, sharing of serostatus with spouse and spouse’s extra-marital relationships (Table 4).

● Social and marital variables: state of stress or anxiety and relationship outcome (Table 5).

a: sexually transmitted infections.

a: preventing mother-to-child transmission.

The multivariate analysis model of this study shows that the potential predictors of serodiscordance were: the status of head of household, the serological status of

the last child, the existence of multiple sexual partners, the existence of occasional sexual partners, the sharing of serological status with the spouse, the duration of the couple’s relationship, the spouse’s extra-marital relationships and strained relations with children (Table 6).

In our study, the prevalence of HIV serodiscordance was 71.70%. It was 73.24% at sites involved solely in monitoring pregnant women. This result corresponds to the model described by Chemaitelly H. [7] in 2012. According to the author, in countries with a low prevalence of HIV infection, around 75% of partnerships affected by HIV are discordant. Benin, with an HIV prevalence of around 1.2%, ranks among these low-prevalence countries. The prevalence described in our study is similar to that reported by Reed et al. [8] , in the African cohort study,

which was 78.52%. Similarly, in Benin City, Nigeria, the reported prevalence varied between 77% and 78.8% [9] [10] .

In contrast, Ujah I. [11] in Lagos and Akani [12] at the University Hospital of Port Harcourt, reported lower prevalences, respectively 66.8% and 52%. In Nweni, south-east Nigeria, Ezeama [13] and Nwachukwu [14] found respective prevalences of 54.7% and 38.9% between 2014 and 2021.

This difference can be explained by variations in regional HIV prevalence rates in these countries, which may or may not limit the risk for a subject in the general population to couple up with an HIV-negative partner. While HIV seroprevalence in Lagos was estimated at 1.2% in 2021 [15] , it was several times higher in other cities: 7.3% in Port-Harcourt [16] ; 9.6% in Nwéni [13] .

The multivariate analysis model of this study shows that the potential predictors of serodiscordance were the status of the head of household, the duration of the couple’s relationship, the serostatus of the last child, the existence of multiple sexual partners, the existence of occasional sexual partners, the sharing of serostatus with the spouse, the spouse’s extra-marital relationships and strained relations with children. In Congo, the explanatory factors for serodiscordance were: the number of partners, region of residence, number of children, length of union, age, level of education, ethnicity, protected sex, religion, standard of living and knowledge of HIV/AIDS transmission mechanisms, love and knowledge of infected partner’s serostatus [17] . In Mozambique, the factors found by Juga et al. [18] in 2020 included: HIV prevalence, number of unions for women, presence of STIs in men, condom use by women, wealth index, condom use by men and number of unions for men. In Nigeria, Akani [12] found the following associated factors: history of at least two ends of pregnancy (OR: 3.05; 95% CI: 2.91 - 3.89) and five or more sexual relationships in the course of a lifetime (OR: 2.3; 95% CI: 1.27 - 3.21).