They consist of socio-demographic characteristic shown in Table 1. Women predominated at 58.46%, with a sex ratio (M/F) of 0.71. The average age was 27.04 ± 1.95 years, with extremes of 16 and 50 years, and a peak in the 20 - 30 age group represented at 47.69%. Christians peaked at 83.08%, as did singles at 64.62. The majority, 60.94%, came from monogamous families. Pupils and students were the most represented with 32.31%.

The presence of 58.46% was not motivated by a complaint and was part of their follow-up. For the remainder (41.54%), their presence was motivated by somatic complaints, of which 21.54% were pain-related, i.e. half of the latter.

The majority of patients (36.92%) had been under CUMAS care for less than a

year. Only one of the 65 patients had previously been under psychiatric care. Of the universal risk factors associated with suicide, alcohol consumption was the most common, accounting for 36.92%, and 87.5% of patients reported occasional consumption.

The scores obtained on the Ducher suicide risk assessment scale and their interpretations are presented in Table 2. According to this scale, 44.62% had a positive suicidal risk score, including 13.85% with a severe suicidal risk corresponding to an active desire to die, or even a defined short- or long-term suicide plan.

Among patients with a positive suicide risk, 44.83% mentioned pain as a factor motivating their suicidal thoughts or desires. Figure 1 illustrates this.

The different remedies used by patients are broken down in Figure 2. 46% of patients with a positive suicidal risk had shared their suicidal ideation or desire, confiding in either a family member (23%) or a close friend (14%), or both (9%). None of them had used a specialized health center on their own initiative, nor had they been referred to one by those to whom they had confided. They said they had received advice and encouragement.

In this study, as shown in Table 3, only the single-parent family appeared as a

factor significantly associated with suicidal risk in sickle-cell patients (OR = 8.03, CI95% [1.51 - 42.51]; p = 0.02). Gender (OR = 2.23, CI95% [0.79 - 6.17]; p = 0.12), ethnic group (OR = 1, p = 0.21) and monthly income (OR = 1, p = 0.40) were not significantly associated with suicidal risk.

The mean age in this study was 27.04 ± 1.95 years. This result is higher than those reported by some authors, who found 24.2 years, 24.75 years and 14 years respectively in their studies [9] [12] [13] . This observed difference could be explained by the age limit of 16 years in this study unlike theirs where no age limitation was made. The choice of this age limit is explained by the sensitivity of the subject studied and ethical considerations.

The most represented age group was between 20 and 30 (47.69%). Dodo et al. had also found the predominance of this age group in their study of sickle cell emergencies at the CNHU-HKM blood disease department [9] . This could be related to the relatively low life expectancy among sickle cell subjects in our context.

The female sex was predominantly represented (58.46%), thus agreeing with the results of Laghaf et al. in Mauritania and Dahmani et al. in Morocco, who also reported a female predominance [12] [13] . In contrast, Dodo et al. found a male predominance of 60.8% (2018) [9] . The female predominance could be explained by the female predominance (51.2%) of the Beninese population [14] and the fact that sickle cell disease is an autosomal recessive genetic disorder.

Overall, the most frequent reason was systematic follow-up (29.23%), but in terms of functional sign, the most frequent reason was pain (21.54%). Dodo et al. also found pain to be the most frequent reason for consultation [9] . This observation could be justified by the fact that vaso-occlusive crisis is the main complication of sickle cell disease [12] [15] [16] .

Based on aRSD scores, 44.62% of sickle cell patients in this study had a positive suicidal risk. Of these, 30.77% were at moderate risk, i.e. with thoughts of death or suicide, and 13.85% at severe risk, i.e. with an active wish to die, or even a defined short- or long-term suicide plan. We found no studies in the literature concerning suicidal risk in people with sickle cell disease. However, when comparing our results with similar studies on suicidal risk, they are similar to those of Trinanes et al. who reported 32.5% for suicidal ideation [17] . Tang et al. in their study of suicidal risk in chronic pain and Ducher et al. in their study of the descriptive epidemiology of suicidal risk in the French general medical system had respectively reported a suicidal risk of 20% and 24.3%, which are lower values than ours [4] [11] . This difference could be explained by the difference in the study population. In the study carried out by Ducher et al., the study population was represented by primary care consultants, regardless of their previous history, unlike ours, which concerned a specific population made up of people with sickle cell disease.

Our analysis showed that only single-parent family type was significantly associated with suicidal risk in sickle-cell subjects (p = 0.02; OR = 8.03; IC95% [1.51 - 42.51]). The World Health Organization, in its report published in 2014 on suicide prevention had also noted this factor as increasing suicidal risk [3] .

However, we note from our study that pain was an important factor. Indeed, 44.83% of the patients in our study with a positive suicidal risk had mentioned pain as the element justifying their suicidal ideas or desires. Tang et al. and Trinanes et al. also identified pain as a risk factor [4] [17] .

Age, gender and ethnic group were not significantly associated with suicidal risk in our study. This observation is also in line with the World Health Organization’s report on suicide prevention [3] . Ducher et al. also found no statistically significant association between age and suicidal risk. However, they did find a significant association between gender and suicidal risk; women were more at risk [11] .

Religion, marital status, occupational status, alcohol consumption and smoking were not significantly associated with suicidal risk in our study. In contrast, the World Health Organization and Ducher et al. found a significant association between marital status, occupational status, alcohol or tobacco consumption and suicidal risk. Indeed, divorce or separation, lack of employment, alcohol or tobacco consumption all increased suicidal risk according to their results [3] [11] . Van Praag, in his study on the role of religion in suicide prevention, concluded that religion was significantly associated with suicidal risk and was a protective factor against suicide [18] . This difference in results between our study and theirs could be explained by the difference in study population and also by the low statistical power associated with our sample size.

46% of patients with a positive suicidal risk had shared their suicidal idea or desire, and had confided in either a family member (23%), a close friend (14%) or both (9%). None of them had used a specialized or non-specialized health center on their own initiative, nor had they been referred to one by the person or persons to whom they had confided. They said they had received advice and encouragement.

These results are in line with those of the World Health Organization, which stipulated that the demand for help in the event of a suicidal crisis remains low, especially as regards recourse to specialized healthcare; populations due to the stigma attached to suicide or the absence of psychological or psychiatric services do not seek help [3] .

This study has limitations worth mentioning. As the study took place in a national referral center, the results obtained may not reflect the general population, due to the high frequency of severe cases and the limited number of people attending national referral hospitals. However, given that recruitment was carried out in Cotonou, the country’s cosmopolitan main city, these results can be used. Although suicide is a public health issue, it remains a taboo subject in some parts of Africa. Some patients do not readily admit to having suicidal thoughts. This may result in under-reporting or concealment of embarrassing information, which could have an impact on their score on the suicide risk assessment scale. Socio-cultural and religious beliefs could also reinforce this. The small size of our sample did not allow us to achieve sufficient statistical power. Similarly, as our study was cross-sectional, it had limitations in identifying risk factors; cohort and case-control studies being the most reliable.

Despite these limitations, the results of this work are interesting for the understanding of suicidal risk in our context on several levels.