Malaria is caused by parasites of the genus Plasmodium and transmitted by female mosquitoes belonging to the genus Anopheles. Although perfectly preventable and treatable, malaria continues to have a devastating impact on the health and livelihoods of people around the world [1] . It is therefore a public health priority in the world. Despite the efforts made in the global strategy to combat malaria, this disease remains a scourge in Africa, especially in the sub-Saharan region, where it is one of the main factors in child mortality. Globally, the number of malaria cases is estimated at 241 million in 2020 in 85 malaria-endemic countries, up from 227 million in 2000 [2] . The number of deaths related to this parasitosis was estimated at 627,000 in 2020, of which children under 05 years of age and pregnant women are the most vulnerable [3] . In most malaria-endemic countries, the disease disproportionately affects the poor and disadvantaged groups, who have limited access to health facilities and can barely afford the recommended treatment [1] . The WHO African Region bears a large and disproportionate share of the global malaria burden. In 2021, around 95% of malaria cases and 96% of malaria deaths were recorded on the African continent with 80% of all deaths among children under five years of age [4] . Malaria is also a priority in Senegal, especially in the south-east of the country with the regions of Kolda, Kédougou and Tambacounda, which accounts for 90% of malaria cases, 64% of severe cases, 58% of deaths of all ages and 84% of deaths under 5 years of age [5] . Despite the implementation of at least three-monthly visits each year of seasonal malaria chemoprevention (SMC) in the 3-to-120-month age group since 2014, malaria is one of the leading causes of infant mortality in the health districts of southeastern Senegal. Thus, the general objective of this study was to evaluate the management of children aged 0 to 120 months hospitalized for severe malaria at the Tambacounda health center from 2018 to 2021 according to the guidelines of the National Malaria Control Program (NMCP).

Descriptive part:

As a result of this work, 481 cases of children hospitalized for severe malaria were enrolled.

From 2018 to 2021, hospitalized cases of severe malaria in children were almost non-existent between January and June. The increase in cases starts from July for all four years. The peak is always reached between October and November, regardless of the year. The highest peak was recorded in November 2020 with 95 cases. The highest number of cumulative hospitalized cases was noted in 2018 with a total of 159 patients (33.05%). In 2019, only 60 children were hospitalized for severe malaria (12.47%). An increase in cases was seen in 2020 with 137 patients, or 28.48% of the cases in our series ( Figure 1 ).

The mean age was 65.64 months with a standard deviation of 29.28 months. The minimum age was 7 months and the maximum was 120 months. Children under 60 months accounted for 39.50%.

Males accounted for 53.43% with a sex ratio of 1.15. The Poular ethnic group was the most representative with 58.38% of cases, followed by the Mandinka (28.89%).

Among the seven health districts in the Tambacounda region, the majority of hospitalized children were from Tambacounda district (95.63%), followed by the border districts of Goudiry (3.12%), Maka Colibantang (0.83%) and Dianké Makhan (0.42%). The other three districts (Koumpentoum, Kidira and Bakel) were not represented. The majority of individuals were admitted from the outpatient clinic (57.79%). A total of 203 children (42.21%) were hospitalized for severe malaria at the Tambacounda Health Center following a referral from another peripheral health facility. Of the 27 health posts in the district, 22 referred cases of severe malaria to the Tambacounda health centre. The majority of referred cases came mainly from the health posts of Sinthiou Malem and Gouye (less than 15 km from the health centre) with 14.93% and 14.03% of referrals respectively.

The reasons for consultation varied ( Table 1 ), but fever and/or history of fever were in the foreground (97.51%), followed by vomiting (42.20%). However, all children had RDT and/or thick droplet (GE) positive (100%). Only 354 children in the series (74.01%) showed signs of malaria severity according to WHO criteria, 15.80% showed no signs of severity, and in 10.60% of hospitalized children, signs of severity were not mentioned in the record. Clinical signs of severity were observed in 67.36% of children and signs of biological severity were present in 17.67% of patients. In hospitalization, GE was positive in 93.80% of cases, negative in 5.20% of cases and results were unavailable (rupture of GE) in 1.00% of cases. In more than half of the series (50.94%), CBC was performed and 33.68% of them had anemia.

All patients in the series received injectable artesunate as treatment (100%). According to the guidelines and the NMCP supervision grid, management was good in 46.55% of hospitalized children, acceptable in 42.61% of children, fair in 8.13% and poor in 2.71% of children in the series. A recovery rate of 81.29% was recorded, 10.19% of patients were referred to the regional hospital, 0.83% were discharged against medical advice. Deaths accounted for 4.99% of cases and in 2.70% of patients the evolution was not reported.

Analytical part:

Analytical analysis shows that not being referred to a health post, not having several signs of severity, not having a state of shock, not being in a coma, not having serious biological signs and not having serious hypoglycaemia were positively associated with recovery in children aged 0 - 120 months hospitalised for severe malaria ( Table 2 ).

Despite the implementation of malaria prevention campaigns in children such as seasonal malaria chemoprevention (SMC) in children under 10 years of age, and free distribution of long-lasting insecticide-treated nets (LLINs) in children under 5 years of age, children are still hospitalized in Tambacounda for severe malaria. In our work, among hospitalized children under 10 years of age, the 24 - 48 months age group was the most representative (28.92%). These results were comparable to those of Jacob et al. [6] in Côte d'Ivoire in 2017 and those of Keita et al. in Bamako [7] with 83.90% and 73.50% of malaria cases in children under 5 years of age, respectively. The predominance at this age could be explained by the fragility of this age group linked to the immaturity of the immune system. At this age, premunition is not yet acquired, and for effective protection against asexual forms of plasmodium in the blood in sub-Saharan areas, a minimum age of 5 years is very often required [8] . Male children were the most affected in this series (53.43%). These results were similar to those of Jacob (33) and Camara at the Albert Royer Children’s Hospital in Dakar (36), who also found a male predominance with sex ratios of 1.1 and 1.4 respectively. On the other hand, Coulibaly et al. in Mali [9] had regained a female predominance in her series (56.50%). However, neither of these studies nor ours was able to establish a statistically significant link between cure following antimalarial treatment and sex.

The seasonal nature of malaria was found in this work, with the beginning of hospitalization of children for severe malaria after the beginning of the winter. The highest numbers of children hospitalized for severe malaria were recorded in 2018 and 2020. This could be related to the fact that in 2018, some health care providers across the country observed a strike movement with a boycott of the strategy of seasonal malaria chemoprevention in children aged 3 to 120 months and the distribution of LLINs. In 2020, the COVID-19 pandemic handicapped the country’s health system, both in terms of attendance at health facilities for preventive purposes and in the quality of patient care. In its 2021 report, the World Malaria Report objectified these same findings with an increase in hospitalized malaria cases [2] . During these years, because the study health centre was the reference centre in the district, a large number of children hospitalized for severe malaria were admitted following a referral from a health post (42.20%). However, there was a statistically significant association between admission by referral and cure. Children hospitalized for severe malaria admitted for non-referral were almost twice as likely to be cured as children hospitalized for severe malaria and referred from a health post. The delay in adequate care through pre-transfer treatment, the delay in consultation at the rural level as shown by the study by Alassani et al. in Benin [10] , and especially the delay in decision-making from the referral from the health post to the health center as highlighted by Berthé et al. [11] (could be explanatory factors for this significant link between healing and referral. However, there is a discrepancy with the work of Mandoko et al. [12] in Kinshasa, which showed that the delay in referral did not have a negative influence on the evolution of patients. Clinically, fever and/or history of fever were present in almost all hospitalized children (97.50%). This result is consistent with the literature in which fever is the primary reason for consultation in malaria [13] .

The WHO criteria recommend the presence of at least one sign of severity for hospitalization of a patient for severe malaria [14] . However, in this work, 74.01% of the children showed at least one sign of severity. This was mostly related to the fact that providers hospitalized children for severe malaria on the basis of 2 to several episodes of vomiting only. However, it should be noted that the presence of at least 2 signs of severity had a statistically significant association with non-recovery. Children who showed two or more signs of severity were 1.89 times more likely not to be cured of malaria. These results were also similar to those of Faye in Dakar [15] and Moyen [16] who also showed that the presence of several signs of clinical severity had a strong influence on the adverse course of malaria. Among the signs of severity, seizures were most objectified in children (45.90%) and there was a statistically significant association between recovery and the existence of seizures. Children with seizures were 1.51 times more likely to be not cured of malaria than those without malaria. These results were in agreement with those of Matumbo et al. in the Democratic Republic of Congo in 2018 [17] and Keita in Mali [7] who had established that most of the adverse malaria outcomes were related to cerebral malaria with the presence of seizures. Coma was also found in 3.70% of children and had a statistically significant association on recovery (p < 0.001). Children hospitalized with severe malaria who were comatose were 7.69 times more likely to not recover than those who did not present with a coma. In 2016, Okoko et al. in Congo Brazzaville [18] confirmed in their work that coma had a strong influence on malaria lethality. The existence of prostration or shock also had a negative influence on the recovery of hospitalized children. This was consistent with the findings of Boushab et al. in Mauritania [19] who had shown that prostration was a factor in poor prognosis for patients hospitalized with malaria. The work of Fall et al. [20] and Camara et al. [15] also objectified that shock had a negative association with the recovery of patients with malaria. On the other hand, respiratory distress did not have a statistically significant negative influence on recovery (p = 0.543). These results were consistent with those of Camara et al. [21] in Dakar but were inconsistent with the results of Gbadoé et al. in Togo [22] and those of Okoko et al [18] . Indeed, in the studies of the latter, acute respiratory distress had a strong influence on the case fatality rate of malaria.

There was a statistically significant association between the existence of signs of biological seriousness and non-cure. In this series, children with signs of biological gravity were 3.70 times more likely to not be cured of malaria than those without malaria. Mansour et al. in Niger [23] in 2019 and Saissy et al [24] . In 2003, it was also shown that signs of biological severity were factors in poor prognosis for malaria. Severe anemia was at the forefront of the signs of biological severity in this work as well as in the study by Bobossi et al. (51). Although there was no statistically significant association between anaemia and malaria cure, hypoglycemia had a strong negative influence on recovery. Children hospitalized for severe malaria with hypoglycemia were 5.88 times more likely to be not cured of malaria than those who did not. These results were consistent with those of Camara et al. [16] , Saissy et al. [24] and Adademy et al. [25] who showed that hypoglycemia was a factor associated with death in severe malaria. Among the biological signs, renal failure was also representative in this series (7.79%) but did not have a statistically significant association with cure. This was contrary to the work of Gbadoé et al. [22] who showed that although acute kidney injury was a rare complication, it had a strong influence on lethality. There was no statistically significant association with cure (p = 0.459) for health care providers’ compliance with malaria severity criteria. This result was similar to that of Yaméogo et al. [26] in Bobo-Dioulasso in 2015 who had shown that compliance with the diagnosis of malaria severity with guidelines did not have a statistically significant association with the death rate (p = 0.78). However, more than 3/4 of the children in the series received adequate care, even though all of the children had benefited from injectable artesunate thanks to its availability and free of charge. In his work in Bobo-Dioulasso, Yamégo et al. [26] also showed that management was adequate in 57.40% of the cases in his series. In our work, the result of the treatment goes hand in hand with the cure rate which was 81.24%. The case fatality rate was much lower in our study (4.99%) compared to the case fatality in the work of Camara et al. (11.10%) [15] and Jacob et al. (9.83%) [6] .

A limitation of our work is that potential confounding factors are not taken into account using multiple logistic regression.

Despite the implementation of proven effective malaria control strategies, malaria remains a public health priority in Senegal, particularly in Tambacounda, where it is one of the main factors in child mortality. The evaluation of the care of children under 10 years of age hospitalized at the Tambacounda Health Center from 2018 to 2021, with a search for factors associated with recovery, showed essential points on which action will have to be taken to reduce malaria-related infant mortality. Thus, in addition to the routine malaria prevention and management strategies implemented in Tambacounda, the early referral of cases of severe malaria from health posts to the health center, children hospitalized for severe malaria with certain symptoms such as coma, prostration, cardiovascular shock, convulsive seizures and/or hypoglycemia deserve special attention and could even be systematically referred to the regional hospital to increase their chance of recovery. In the future, the uptake of SMC drugs and their effectiveness will have to be established in these children hospitalized for severe malaria.

Thanks go to the members of the management team of the Tambacounda health district, the major of the Tambacounda referral health center, not to mention the members of the district health development committee who financed the reproduction of the tools.

This research was funded by the Ministry of Health and Social Action of Senegal through the Tambacounda health district.