A total of 19,090 individual blood culture bottles were collected from 1419 impatient patients as follows: normal profile (n = 18,042 bottles), AS profile (n = 677 culture bottles), SS profile (n = 362 bottles). The majority of blood culture orders is from pediatric services (70%) and account for 70% of the total number of prescriptions. Thus, 70%, 61% and 53.6% of the prescriptions made in normal profiles, AS and SS originate from pediatrics. Children under the age of 1 account for 54% of prescriptions issued. 67% of prescriptions issued in the SS are for patients under 10 years of age. The age range of 1 to 5 years includes the majority of hemoculture prescriptions in SS, accounting for 37% of prescriptions. The majority of the study population was quite young, with a mean age of 13.5 years [SD = 12.6 - 14.5]. Children under the age of 10 make up 58% of the study population. There was no significant age difference between normal and AS profiles (p = 0.997). Age differed significantly between the SS profile and the normal profile (p < 0.05). Emergencies and surgeries were more likely to have SS patients (Table 2).

Of 19,090 vials examined in this current study, 19.6% developed a positive blood culture. The normal and AS profiles have a similar positivity rate with 19.6% and 19.1% positive vials, respectively. The SC profile had the lowest positivity rate at 11.1%, while the highest values were in the SS with 24.3% positivity (Table 2).

The most commonly isolated bacteria were Staphylococci (41.1%), Enterobacteriaceae (36.7%), Bacillaceae (10.2%) unfermented (6.30%), Streptococci (5.01%) and a small proportion of yeast (0.75%). There is no significant difference in bacterial spectrum between the SS profile and the normal profile of subjects (p = 0.104) (Table 3).

Coagulase-negative staphylococci accounted for 32%, 24%, and 40% of species

isolated in the normal AS and SS profiles, respectively. Coagulase-negative staphylococci were the most frequently isolated germs in SS. Group E streptococci and sp-tagged streptococci each account for less than 2% of the isolated organisms in SS. Pneumococci were not found. Bacillus accounts for 25% of isolates in SS subjects compared to 9% in normal and AS subjects, respectively. Pseudomonas aeruginosa and Burkholderia cepacia then account for 10% of the isolates in the SS profile subjects as non-fermenters.

About 13% is found in subjects with AS profile, while in normal subjects and SC the frequency of Pseudomonas aeruginosa and B. cepacia does not exceed 5%. Klebsiella pneumoniae and E. coli are the main isolated Enterobacteriaceae, in normal (19% and 7%) and AS (11% and 8%) profiles, respectively. E. coli, P. mirabilis, and Serratia marcescens appear to be the most important enteric bacteria in SS individuals, accounting for 7%, 3.5%, and 3.5%, respectively. Salmonella was observed in less than 2% of cases, respectively, only in normal and SA-type subjects (Table 4, Figure 2).

The aim of our study was to determine the positive value of blood culture vials in sickle cell patients compared to normal people. Second, compare the spectrum of bacteria in SCD patients with normal individuals. Bacteremia is quite common in people with sickle cell disease, and in some studies, bacteremia can account for 10% - 32% of the causes of fever [16] . A total of 3741 blood culture

vials showed positive culture in our study on a total of 19,090 vials analyzed. Blood cultures were most frequently positive in homozygotes with a frequency of 24.3% compared to 19.6% in normal profile. Several studies in Africa have shown the higher frequency of 14% to 32% of bacteremia in patients with sickle cell disease. In Benin, 32.5% of febrile drepanocytes showed bacteremia and similar results of 32.2% were found in Nigeria in Ibadan [15] [16] . A similar study found a slightly lower yield in Saudi Arabia, which was 20.8% or 3546 positive vials from the 17,053 blood cultures tested [17] . In developed countries such as Canada [17] , infections are much less common, at only 1.3% among the 692 emergency patients admitted for bacteremia. Similar results were observed in the USA by Bansil [18] and by Morissey [19] . Overall, bacterial infections appear to be less common in developed countries, the USA [18] , London [19] , and Canada than in sub-Saharan Africa [20] . This disparity could be partly related to better living conditions and better prophylaxis, as well as greater vaccine efficacy [14] [15] .

However, it should be noted that 24.6% and 40.4% of isolates in homozygotes are coagulase negative Staphylococcus and Bacillus species, respectively, which are mostly considered non-pathogenic. According to some authors, coagulase-negative staphylococci can account for up to 80% of the contaminations in their series [20] [21] . A similar study conducted in Bahrain in a cohort with sickle cell disease showed that Staphylococcus epidermidis was the most common germ, isolated in 42.86% of cases. Pathogenicity depends on the expression of virulence factors that are unexplorable in resource-constrained settings. For this reason, if only one bottle is positive for negative staph, contamination is considered. After isolation of negative staphylococci twice, the result is considered positive and an antibiotic susceptibility test is initiated. The greater frequency of contamination in homozygous sickle cell disease may be related to more severe symptomatology, often requiring more blood cultures to find pathogens.

Since the introduction of the pneumococcal vaccine, predicting the bacteria involved in septicemia has become increasingly challenging. In many cases, it lies outside the classic pathogen spectrum of sickle cell infection [21] . Our results show that among the proven pathogens, enterobacteria occupy the first place with 29% and 16% in subjects with AS and SS profiles, respectively. This is followed by non-fermenters with 16% and 12% in the AS and SS profile subjects, respectively. Gram-positive cooci rank third position with 5% and 11% respectively in the SS and AS profile subjects. This distribution is quite distant from the classical distribution of the causative agent of sickle cell anemia, in which gram-positive cocci with pneumococci and staphylococci in the head predominate. We did not find isolates of Streptococcus pneumoniae in either AS or SS individuals. The first studies on the pathogen spectrum of bacteremia carried out in Lagos in the 1980s already showed a dominance of gram-negative rods with Klebsiella spp. (38%), E. coli (23%), Pseudomonas spp. (23%) [16] [22] . In 1993, Okuonghae in Nigeria found in 166 repertory infection episodes in 162 drepanocytes that gram-negative germs predominated, accounting for 70.4% of the isolates. Salmonella and Klebsiella were the predominant germs, each with 25.9% of all isolates [15] . Similar results were found for Ibadan, where 59% of isolates were Gram-negative and for Staphylococcus aureus, the major Gram-positive cocci [15] [22] . It appears that the etiological agents found in sickle cell patients better reflect the microbial ecology of African countries characterized by a dominance of Gram negatives and Staphylococcus aureus [23] . First studies carried out in Fann CHU by GAKi-Zerbo between 1985 and 1987 have already shown that 792 blood culture or 80.16% of the total, were gram-negative bacilli [24] . For some authors, the Gram-positive Gram-negative shift begins in the early 1990s with the advent of the first prophylaxis that altered germ ecology and resistance [25] .

The effectiveness of pneumococcal prophylaxis in developed countries has been supported by several studies, including those conducted in the United States between 1997-2006 and 1998-2018. These two studies both showed a reduction in pneumococcal infections by more than 90% under the effect of vaccination [26] [27] . In our study Streptococcus pneumoniae was not found among blood culture isolates. For Akinyanju of the children’s emergency hospital, gram-positive cocci were not only a minority, but no more S pneumoniae was found. This observation has already prompted the author to discuss the suitability of European guidelines based mainly on Streptococcus pneumoniae prophylaxis in the Nigerian context [22] [28] . Similar results by Alsaif in the vaccine era between 2005 and 2015 showed an important reduction in severe episodes of infection, affecting less than 10% of the febrile sickle cell population. On the other hand, none of these infections was related to Streptococcus pneumoniae [29] . For some authors, this could be the result of almost systematic prophylaxis against infections with penicillin and vaccination results in sickle cell disease subjects. Especially since our study population consisted of subjects who benefited from hemoglobin electrophoresis and thus had made clinical arguments likely to initiate pneumococcal prophylaxis. In fact, severe forms of pneumococcal infections continue to be described in the general population, such as at the Albert Royer Center where 218 cases were recorded between 2008 and 2013. However, meningeal and pulmonary forms remained predominant and only 27/218 had bacteremia. Sickle cell disease accounted for only 2.3% of the total study population, suggesting that pneumococcal bacteremia is rare in Senegalese sickle cell patients [30] . According to some authors, the evolution of the affected sickle cells subject in the malaria endemic zone allows maintaining of a certain spleen activity sufficient to fight pneumococcal infections [31] .

For other authors, the fetal hemoglobin of the affected sickle cells would play the same protective role in prophylaxis against microorganisms [32] . For other authors, the abuse of antibiotics in prophylaxis could explain the rarity of pneumococci infections in our populations [11] . It is clear that pneumococcus diagnosis can be affected by sub-optimal growing conditions or a low level of pathogens. Therefore, the new diagnostic techniques such as PCR (Polymerase chain reaction) and NGS (new generation sequencing) often make it possible to catch up on false negative blood culture results [33] .