Prevalence of malaria, prevention measures, and main clinical features in febrile children admitted to the Franceville Regional Hospital, Gabon

Recently, major progress has been made in controlling malaria in Africa. However, in Gabon, little information is available on the role of malaria in childhood febrile syndromes, the use and efficacy of preventive measures, and Plasmodium species distribution. Here, we characterized malaria in febrile children in Franceville, Gabon through a cross-sectional study at the pediatric unit of the Franceville Regional Hospital. We registered 940 febrile children. Their general condition was markedly altered in 11.7% of cases (n = 89/760); among them 19 (21.4%) had a severely altered condition. Malaria was the second most frequent etiology (22.0%; n = 162/738), after respiratory tract infections (37.3%; n = 275/738). Children with malaria (63 ± 39 months) were older than children without malaria (40 ± 37 months) (p = 0.0013). Hemoglobin, red blood cell, white blood cell, and platelet values were lower in children with malaria than in those without malaria (p < 0.0001). Anemia was the most common feature of severe malaria (70.6%; n = 12/17), followed by neurological involvement (23.5%; n = 4/17). The prevalence of malaria was significantly higher in children older than 60 months than in younger children (40% vs. 15.5%; p < 0.0001). Plasmodium falciparum accounted for 97.5% of cases (158/162), followed by Plasmodium malariae (2.5%; n = 4/162). Bed net use was high (74.4%; n = 697/936) and contributed to malaria prevention (p = 0.001). Good basic knowledge of malaria also had a preventive effect (p < 0.0001). The prevalence of malaria in children in Franceville did not decrease significantly from 2009 to 2012, remaining at about 20%, highlighting that preventive measures should be reinforced.


Introduction
Febrile syndromes are one of the main reasons for pediatric consultations in developing countries and may be due to various viral, bacterial, and parasitic infections. Malaria is the most widespread febrile disease, present in 99 countries and territories. One-third of the world population is at risk, and malaria causes about one-fifth of all childhood deaths worldwide [13]. In 2013, there were an estimated 198 million cases of malaria and 584,000 deaths [42].
New antimalarial policies implemented in recent years have mainly concerned P. falciparum, although P. vivax and P. knowlesi can also cause life-threatening malaria [14,19,34]. Plasmodium knowlesi can cause cerebral hemorrhage, lung heaviness, and cardiac hemorrhage [6,7]. Premature reticulocyte death due to P. vivax infection can cause severe anemia over a period of several months by blocking the formation of mature red blood cells [1].
Gabon is hyperendemic for malaria. Transmission is perennial because the equatorial climate favors mosquito proliferation and larval development. Changes in the national antimalarial policy, based on artemisinin combination therapy (ACT), distribution of impregnated bed nets, and intermittent preventive treatment during pregnancy, have led to a decline in the malaria burden in urban areas. Between 2005 and 2008, malaria prevalence dropped significantly from 31.2% to 18.3%, followed by a recrudescence in 2011 in Libreville (24.1%). In Franceville, malaria prevalence dropped from 69% (2004) to 19.5% (2009). Between 2010 (17.9%) and 2012 (21.45%), no significant change was observed [2,17,20]. Artemether-lumefantrine (AL) and artesunate-amodiaquine (ASAQ) are both drugs in the co-first lines of treatment. Only three human species (P. falciparum, P. malariae, and Plasmodium ovale) have been reported in Gabon [21,31]. The most prevalent is P. falciparum (94-99%), followed by P. malariae (0.5-5%) and P. ovale (0.5-2.4%). The near absence of human Plasmodium vivax infection in Central and Western Africa has been attributed to a low prevalence of the Duffy receptor in these populations [24]. However, cases of P. vivax infection have been described in Duffy (-) subjects in Uganda and Central Africa [8,23,29]. Surveillance of all malaria species will be necessary to eradicate the disease.
The aim of this study was to determine the place of malaria in febrile children consulting at the pediatric unit of the Franceville Regional Hospital, Gabon, as well as its characteristics and the efficacy of preventive measures.

Ethics statement
The study was approved by the Gabonese National Ethics Committee and the Ministry of Health (N°00370/MSP/ CABMD). Blood samples were collected with the parents' or guardians' informed consent.

Study location and population
We conducted a cross-sectional study between April 2011 and May 2012 in the pediatric unit of the Amissa Bongo Regional Hospital in Franceville, the capital of the Haut-Ogooué Province, Gabon. Children between 6 and 168 months of age presenting an axillary temperature >37.5°C or a >24-hour history of fever were recruited.
Global health alteration was classified as severe in the presence of anorexia, asthenia, and weight loss, and moderate in the presence of two of these symptoms. The alteration was defined as uncomplicated in the presence of only one of these three symptoms. Weight loss in the child was estimated by using the last weight recorded in the health book.

Recording of preventive measures
Data on the use of bed nets and insecticides, and some information on education and knowledge of malaria were collected with a questionnaire (Supplementary Material). The parents/guardians were asked whether their children slept under bed nets and whether they sprayed insecticide in their houses. To determine their level of education on malaria, we asked the following questions: ''Do you know malaria? What is malaria? Do you know the modes of transmission of malaria? How can you avoid malaria?'' Correct responses to these questions were considered a basic knowledge of malaria.

Diagnosis
The OptiMAL-IT Ò rapid diagnostic test (RDT) was used [27]. Parasite load was determined on blood smears using the Lambaréné method [30]. This method of counting slides is a variant of a method for counting thick films. Ten microliters of blood is evenly distributed on a 10-by 18-mm area of a microscope slide. Each high-power field (HPF) on this thick smear is 1/500th of a microliter (on a standard microscope at 1,000 magnification), and a count is made per 10 HPFs. The parasitemia per microliter is calculated by a precalibrated appropriate multiplication factor (500). Children with RDT or positive blood smears were considered to have malaria. Respiratory tract infections corresponding to colds and coughs, gastroenteritis corresponding to diarrhea and/or vomiting, dermatosis, ear, nose, and throat (ENT) infections, acute algetic syndromes, as well as clinical symptoms, such as pale conjunctivae and global health state alteration (anorexia, asthenia, and weight loss), were diagnosed by a physician according to clinical classification.

Hematological analyses
Routine hematological assays were performed with an automated blood cell counter (STKS Ò , Coulter Corporation, USA). Blood (5 mL) was collected in EDTA tubes. Plasma was stored at À20°C and blood pellets were used for DNA extraction. Moderate anemia was defined by a hemoglobin level between 5 and 10 g/dL, and severe anemia by a hemoglobin level 5 g/dL.

DNA extraction
DNA from all children was extracted with the Omega Bio-Tek E.Z.N.A.1 method (Omega Bio-Tek, USA) according to the manufacturer's protocol [16]. Briefly, 250 lL of blood ellets, 25 lL of OB protease (20 mg/mL), and 250 lL of lysis buffer were mixed and heated to 65°C for 30 min before adding 260 lL of isopropanol. The mixture was transferred to a column and centrifuged at 10,000 rpm for 1 min. The column was washed twice at 13,000 rpm for 2 min, and DNA was eluted with 90 lL of sterile water preheated to 65°C. DNA samples were kept at À20°C until use.

Identification of Plasmodium species by RFLP-PCR
Plasmodium speciation was based on nested PCR amplification of the cytochrome b gene, followed by enzymatic restriction, as previously described [39]. Five microliters of DNA was amplified with 1X Taq polymerase buffer (Invitrogen), 0.8 lM of each primer (Plas 1 and 2a for primary PCR and Plas 3 and 4 for nested PCR), 0.2 mM dNTP (Invitrogen), 2 mM MgCl 2 , and 0.024 U of Taq DNA polymerase (Invitrogen), using the following cycling program: 5 min at 94°C, followed by 35 cycles of 30 s at 94°C, 45 s at 58°C, 45 s at 72°C, and a final extension step for 7 min at 72°C. PCR products were detected by 2% agarose gel electrophoresis. In order to distinguish the four species of Plasmodium (falciparum, malariae, ovale, and vivax), restriction fragment length polymorphism (RFLP) analysis was performed using the restriction enzyme AluI (New England Biolabs, UK). PCR primers and Plasmodium species profiles after AluI enzyme digestion are shown in Table 1. DNA products were detected by electrophoresis on 2% agarose gel.

Statistical analysis
Statistical analyses were carried out with Epi-info version 3.3.2 (2005, CDC, Atlanta, USA) and STATA version 10 (Stata Corp, College Station, USA). Age was expressed as the median and interquartile range (IQR), and parasite density as the geometric mean (GMPD). The Chi-square test was used to compare categorical variables, and the non-parametric Mann-Whitney U test, Pearson's test, and Fisher's exact test were used for group comparisons, as appropriate. p values <0.05 were considered to indicate statistical significance.  Table 2. Mean body temperature was 38.3 ± 1.1°C. Hematological values were normal.

Effect of malaria on hematological parameters and age distribution
The mean age of the Plasmodium-infected patients (63.4 ± 39.4 months) was significantly higher than that of the other children (40.3 ± 37.1 months) (p = 0.0013), ( Table 3). The body temperature of the Plasmodium-infected children (38.5 ± 1.1°C) was higher than that of the other children (38.2 ± 0.9°C) (p = 0.001). Hemoglobin, red blood cell, white blood cell, and platelet values were lower in Plasmodium-infected children than in uninfected children (p < 0.0001). The mean interval between symptom onset and consultation was 3 days in both Plasmodium-infected and uninfected children.

Distribution of Plasmodium species
Molecular diagnosis was based on amplification of an 816 bp fragment of the Cytb gene. Speciation was based on enzymatic digestion (RFLP) of PCR products. Plasmodium falciparum was characterized by two DNA fragments (640 and 159 bp), and P. malariae by three fragments (187, 249, and 381 bp). Only 162 PCR products could be digested, because of insufficient yield and the low sensitivity of the primer set used. We found that 158 (97.5%) infections were due to P. falciparum, and 4 (2.5%) to P. malariae. No mixed infections and no P. ovale, or P. vivax infections were found.

Discussion
Elimination or eradication of malaria is a new major challenge in some endemic areas. Herein, we assessed the epidemiology of malaria in febrile children in South-East Gabon (Franceville). More than three-fourth of febrile children reporting to the Franceville Regional Hospital during the study period had an altered global health state, probably linked to a high prevalence of anemia, as indicated by the large proportion of children with conjunctival pallor. Similar results have been obtained in Tanzania [25]. We confirm that children aged less than 5 years are the most likely to present fever [37], and that infectious diseases account for fever in 80% of these children. Low hemoglobin levels are due to parasite-induced hemolysis [3]. Indeed, P. falciparum antigens, such as RSP2/RAP2, could be transferred to the surface of uninfected and infected red blood cells, reducing their deformability, inducing their sequestration by the spleen, accelerating both complement-mediated lysis as well as macrophage uptake, leading to anemia [11]. Plasmodium falciparum infection also modulates the cytokine balance, leading to erythrocyte clearance [26]. However, malnutrition and digestive parasitic infections are also causes of anemia [15,44]. In addition, we confirm that malaria is the second leading cause of fever among children in Franceville, after respiratory tract infections [4,17]. Gastroenteritis was the third most common cause of fever, as also reported in Tanzania, where malaria was the second leading cause of fever [9]. Although we did not explore the pathogens responsible for respiratory tract infections, a recent report from Gabon (including Franceville) showed that the most prevalent are influenza-like viruses (adenoviruses, parainfluenza viruses, enteroviruses, respiratory syncytial viruses, and influenza viruses) [18]. The cause of fever remained undetermined in 10% of children, underlining the shortcomings of the Gabonese healthcare facilities.
Our findings confirm that malaria mainly affects older febrile children in Gabon [20]. However, our findings are based only on children in consultation at our study sites; we do not have information on malaria distribution in the overall population. Children under 5 years are more often protected by bed nets and other preventive measures than older children whose prevention measures are often unavailable. We found an overall high rate of bed net use. Insecticide-treated bed nets have been shown to significantly reduce the incidence, severe forms, and lethality of malaria, especially among children under five years old. As in several other African countries, the use of bed nets has contributed to the decline in the malaria burden in Gabon [40]. It should be noted that information on the use of bed nets in this study was based on the verbal statements of parents/guardians, and that bed net condition (presence of holes, insecticide impregnation) was not verified.
We found that children under seven years of age tended to have lower parasite loads. The same result has been reported after the increased coverage of insecticide-treated bed nets and the decrease of malaria transmission in some countries [38,43]. However, parasitemia tended to be lower after 108 months of age, a finding consistent with acquisition of semi-immunity.  The low frequency of malaria in children less than 12 months is consistent with maternal immune protection through breast-feeding. The higher prevalence of malaria among febrile children over 5 years is consistent with reports from Rwanda [32], where the risk was highest in the 5-to 15-year-old age group. A similar age distribution has been reported in asymptomatic Nigerian children [28]. This could suggest a delay in immunity acquisition, because of the use of preventive measures that limit the contact between children and the parasite. Children aged 5 to 15 years could thus represent a reservoir for Plasmodium. Surprisingly, the data indicate that malaria prevalence was highest in the short rainy seasons. Indeed, breeding sites in this season are maintained while they are cleaned in long rainy seasons.
We confirm the decrease in the malaria burden in Franceville in recent years, despite its stability since 2009 [17]. In contrast, however, a recrudescence has been reported in febrile children up to 24.1% in Libreville, capital of Gabon, after the decline observed between 2000 and 2009 [5,20]. This could be explained by different socioeconomic contexts and/or plasmodial transmission rates, and by a decrease in free distribution of impregnated bed nets and awareness campaigns on malaria. Since the year 2008, Gabon has lost a global grant.
Hematological values were lower in Plasmodium-infected children than in other febrile children. This is consistent with red blood cell lysis and platelet sequestration induced by P. falciparum, as recently reported in India [35]. We confirm that malaria is also associated with a loss of white blood cells [22]. Severe malaria anemia was found in 70.6% of our patients with severe malaria, consistent with data from Libreville, Gabon [12].
Plasmodium falciparum was responsible for the majority of cases of malaria in this study, as reported elsewhere in Gabon [21,31]. This suggests that, despite the decrease in the malaria burden in Gabon, the species distribution in febrile children has not changed. In contrast, among asymptomatic individuals living in rural areas of Gabon, it was recently reported that P. malariae and P. ovale accounted for, respectively, 47.6% and 9.9% of cases [10]. This could suggest selection of P. falciparum to the detriment of other species during acute  malaria. Another possible explanation is the use of different diagnostic methods. Delicat-Loembet et al used 454 sequencing to identify Plasmodium species, while we and other authors used less sensitive methods (PCR-RFLP, and blood smear with RDT, respectively). Despite the non-detection of P. ovale in our study, this species is occasionally found in Franceville, and we have diagnosed one case of P. ovale infection since the end of this study. The low sensitivity of the PCR can be explained by insufficient yield and the low sensitivity of the primer set used. We confirm that P. vivax does not circulate in humans in Franceville, and neither does P. knowlesi, a primate species that causes some episodes of severe malaria in humans [41].
We found that 66.7% of parents/guardians had a good knowledge of malaria, and confirm that this knowledge is associated with a lower risk of childhood malaria. Surprisingly, the use of window nets and insecticide sprays was not associated with a lower risk of malaria, possibly because few parents/guardians used these two measures.
We confirm the stability of the malaria burden in Franceville, Gabon, implying that stronger preventive measures are needed to reduce it further. Studies are also needed to evaluate the malaria burden throughout Gabon.