Complicated malaria and other severe febrile illness in a pediatric ward in Libreville, Gabon

In Africa, one in six children dies before its fifth birthday from treatable conditions such as pneumonia, gastroenteritis and malaria [1]. In 2009, the World Health Organization (WHO) estimated that among the 761,000 deaths attributable to malaria, more than 80% occurred in African children less than five years old [2]. During recent decades, malaria has become a default diagnosis for acute febrile illness in many African settings though is presently declining [3]. Moreover, clinical manifestations of complicated forms of malaria, which include mainly cerebral malaria and severe anemia, usually overlap those of other severe febrile illness in endemic countries.

Presumptive treatment of fever was recommended by WHO for a long time. However, such a strategy is no longer satisfactory because of its poor specificity, leaving non-malarial febrile patients without an appropriate treatment. In fact, evidence suggests a higher mortality in children diagnosed and treated for malaria without microscopic confirmation compared to those who benefited from microscopically confirmed malaria [4]. The biological diagnosis of malaria is still not routinely done by health staff from many endemic countries, even when it is available [5]. In Gabon, where a substantial decline of malaria prevalence in Libreville, the capital city, has been observed, little is known about the suitability of treatment of patients in the pediatric wards in public hospitals |5]. Indeed, a threefold reduction of malaria cases was observed in the country [6, 7]. A shift toward a higher susceptibility of children older than five years was also noted for uncomplicated malaria, suggesting an epidemiological transition of the infection in the country. The frequency of complicated malaria should also be influenced by interventions. Therefore, the decline of malaria prevalence can also result in a parallel decrease of, or influence the frequency of cerebral malaria (CM) and severe malarial anaemia (SMA) as reported in north-eastern Tanzania and in rural Kenya [8, 9].

The aim of the present study was to assess the management of febrile illness and to determine the proportion of children with febrile illness hospitalized for primary diagnosis of malaria who had confirmed complicated malaria, at the Centre Hospitalier de Libreville, the main public health centre of Gabon, after implementation of new malaria control strategies.

Presumptive treatment of fever with antimalarial drugs is still frequent in public health settings of sub-Saharan Africa, even in countries where ACTs, bed nets, malaria diagnosis and IPT are freely provided to vulnerable population [8]. During the last years, the prevalence of fever due to malaria has been drastically reduced in several countries including Gabon [3, 6]. There is an urgent need to better manage febrile illness, especially in children, to avoid high mortality due to misdiagnosis of other etiologies, and overtreatment of non confirmed malaria cases which leads to impairment of patient outcomes, waste of money and time, and increased risk of drug resistance [4, 11]. Unfortunately, clinical diagnosis through the syndromic approach, according to the integrated management of children febrile illnesses, remains the main method for the diagnosis [11–13].

In Libreville, malaria declined from being the primary cause of hospitalization among febrile children admitted in the same pediatric ward in 2002, to representing the second cause six years later [14].

Hospital case management of seriously ill children is a challenge in middle and low income countries, mostly due to few resources for diagnosis and treatment, staff availability, and laboratory services [13, 15]. Respiratory tract infections, gastro-enteritis, urinary infections and meningitis are the main evocated etiologies of severe febrile illness in hospitalized children either in west, east, central Africa and Asia [11, 16–18]. Symptoms of these conditions always overlap with malaria, leading to difficult patient management, misuse of antibiotics and antimalarial and higher mortality rates [19, 20]. Less than one quarter of children had co-morbidities, mostly malaria and lower respiratory tract infection, as reported in Kenya and Uganda [12, 21]. Bacteremia was not confirmed when suspected in children included, although bacteria are the most frequently associated pathogens to P.falciparum infection, either through infection or co-incidental parasitaemia [20–22]. Beside bacteria, HIV infection and malnutrition are also frequent and responsible of life threatening conditions in malaria infected patients [22, 23], this is not the case in the present study. Likewise, between 2005 and 2007, an increase rate of non-malarial and non-bacteremia fevers has been observed, with appearance of Chikungunya and Dengue infections, among other viral infections in Libreville [24]. Approximately half of the patients treated with antimalarial were blood smear negative, suggesting another undiagnosed cause of fever and the necessity of an improvement of clinical and laboratory diagnosis as well as training and supervision of health workers for a proper case management.

Globally, prostration, neurological symptoms like seizures and pallor are the main clinical forms of malaria among children with acute febrile illness as observed in other endemic countries [25, 26]. Thrombocytopenia and low Hb level, both predictors of malaria, have also been found to be associated to a greater risk of being infected [27, 28]. The increase of the frequency of Plasmodium infections with age observed since 2006 in outpatients was confirmed among the inpatients: children older than five years are 2.6 times more at risk for having complicated malaria. This suggests a change in the transmission profile of the disease as the prevalence of infection increases with age with decreasing transmission [26, 29, 30]. The median age of children with complicated malaria was 48 months compared to the non-infected (18 months). Moreover, cerebral malaria (CM), which has been shown to increase with age when the intensity of the transmission decreases, was observed in children aged more than 48 months old [30]. This was not the case six years earlier when 65% of CM was confined to children aged between 12 and 48 months. A recent meta-analysis showed that in areas with no seasonality, the median age of cases with CM increases with lowering transmission intensity (from 29 months in medium intensity areas to 49 months in low intensity areas), whereas the median age of cases of severe malarial anemia (SMA) did not significantly vary and remains concentrated among younger children [30]. In 2008, 75% of children with SMA were between 6 to 36 months old, and, in 2002, they were below 24 months [14]. The high burden of anemia (85%) and its multifactorial etiology are likely to be the reasons for the fact that SMA is the major clinical presentation of severe malaria (SM); it would therefore be difficult to use anaemia as marker for malaria transmission intensity characterization [11, 30, 31]. Indeed, despite the decline of complicated malaria prevalence observed in this study, the frequency of severe anemia has not drastically changed compared to 2001–2002 (9.4% in 2001–2002 and 10.5% in 2008), as well as that of SMA (14.3% in 2001–2002 and 14.7% in 2008) in Libreville [32]. This observation raises the problem of the complex aetiology of anaemia; moreover, it seems that P.falciparum infection often occurs in patients with chronic anaemia and increases its gravity [33].

Improvement of medical care is ongoing in Gabon, but most of the population cannot benefit from complementary examinations because of poor financial resources and lack of equipment at health centers. As an example, beside chest X-ray that is systematically prescribed in case of suspected respiratory tract infection, lactate levels cannot be measured at the CHL and material for blood and CSF cultures is always lacking. Therefore, it remains crucial to use clinical and biological signs to target malaria infected children for a rational use of antimalarial drugs in poor resource areas. Pallor, presence of neurological symptoms, BCS ≤ 2 and age >5 years, were predictors of malaria infection among children hospitalized for fever. These risk factors have also been found in children from other countries [16, 34, 35]. As always reported, Hb level below 8 g/dL and severe anemia are biological risk factors, thrombocytopenia also being a prognosis factor in children with parasitemia [28, 35]. Other factors associated with poor outcome in SM children are low BCS, co-infection, HIV, bacteremia, hyperlactatemia, malnutrition and respiratory distress [10, 20, 34, 36]. Children who died from SM had respiratory distress and CM, two clinical forms already reported as prognosis factors at Libreville in 2002.

Even if malaria diagnosis and hematological count were freely provided to patients, overdiagnosis (43%) and overtreatment with antimalarial drugs (46%) were common.

Although the low number of P.falciparum-infected patients, lethality was lower (1%) compared to that observed between 2001–2002 in the same paediatric ward (9%) and in 11 African countries (10.5%); probably due to i) the reduced cases of malaria, ii) a better case management because all malaria cases were microscopically confirmed and suspected co-morbidities treated with both antimalarial and antibiotics [37, 38].

After the deployment of new preventive measures, decreased frequency of malaria hospital admissions and its various clinical presentations were observed [9]. However, despite an adequate coverage, the reduction of the malaria burden observed among children or pregnant women, during the first years of the deployment, was followed with either stabilization or, conversely, an increase in the frequency of P.falciparum infection [39–41]. The protective effect of these control strategies, confirmed here with the lower prevalence of malaria among bednet users, is well recognized. Regular monitoring should absolutely be implemented in Gabon where ITNs, ACTs and diagnostic tools coverage is still weak or lacking in the health centers from remote and peripheral areas.

This pilot study allowed the inclusion of an enough number of patients to provide reliable data on hospitalized febrile children. A higher number of inclusions would have probably not given more indications on febrile illness management in the paediatric ward, neither more clinical data. Indeed, the main limit of this study concerns the lack of additional investigations necessary to finalize the diagnosis. They are most of the time not performed because of poor financial resources of the parents. Furthermore, there are some limitations on using health facility–based data including the lack of a knowledge on variations in access to, or use of the facilities in the community, extend of real coverage by various interventions.

Nevertheless, hospital based studies provide accurate data on children febrile illness etiologies, though they cannot replace epidemiological surveys in the community [42]. The present data can be used for the characterization of the malaria epidemiology over time in a specific area. However, local data are needed for WHO guidance application and are lacking in several endemic countries, including Gabon.

This study, and the one performed between 2001 and 2002, highlights the most likely causes of fever as well as a reduction in the frequency of P.falciparum infection in Gabon in 2008. These results can help for the design of an effective development of new guidelines for fever case management. Priority areas to achieve proper patient care are quality of diagnosis, treatment practices, availability of biological and radiographic diagnosis tools, introduction of malaria diagnosis RDTs to discriminate malarial to non malarial fevers and bring out rational use of drugs, health worker training and adherence to guidelines.