No evidence of decline in malaria burden from 2006 to 2013 in a rural Province of Gabon: implications for public health policy

This is the first large retrospective survey conducted on malaria morbidity in North Eastern Gabon. Regular collection of data on malaria prevalence and at-risk populations is mandatory to establish and implement health policies and to plan programs to fight malaria [14]. The identification of the populations and regions the most at risk depends on the quality and completeness of the recorded data. In Sub-Saharan Africa, this information is often lacking, either because the source documents simply do not exist or because they are confidential and cannot be easily accessed [15]. Data published over the last decade nevertheless indicate a 70% reduction in malaria morbidity in Ethiopia and Eritrea, and similar patterns have been reported in Kenya [15-17]. In West Africa, surveillance of healthcare facilities indicates a 50-85% decline in the rate of positive smears in Gambia, and encouraging trends indicating a decline of malaria have also been recorded in Mozambique, South Africa, Tanzania, Zambia, and Zanzibar [18-22]. In contrast, data on Central Africa are particularly scarce and show little change in the malaria situation relative to historical figures [23]. In Congo, the proportion of pediatric hospitalizations due to malaria (estimated at 30%) has remained stable [24]. Two cross-sectional studies performed in Cameroon in 2000 and 2004 indicate a slight reduction in the pediatric prevalence of asymptomatic carriage, but found no change in malaria prevalence in patients with fever [25]. In this respect, the malaria situation in Gabon appears more favorable. There is now increasing evidence that the epidemiological situation of malaria is changing in urbanized areas. The prevalence of malaria in children suffering from fever who were admitted to the Libreville hospital has dropped by almost 70% over the past 10 years, from 45% in 2000 to 15% in 2008. This reduction in malaria morbidity was accompanied by an increase in the age of the most at-risk group [10]. A similar trend was reported in hospitals in Franceville, located in southeast Gabon [26]. In this region, the number of children suffering from fever who tested positive for malaria significantly decreased from 68% in 2004 to 18% in 2010 [11]. The overall reduction in the incidence of malaria in Gabon and the decline in mortality have been attributed to the implementation of priority control interventions such as the use of artemisinin-derived antimalarial drugs and the distribution of ITNs [27,28].

Urban centers account for about 70% of the population of Gabon, but they do not reflect the health situation in more isolated regions where the infrastructure is poor, and where malaria is endemic and its prevalence is certainly under-estimated [29]. Despite the decline in malaria burden in many endemic areas, our data demonstrate that it has remained mostly unchanged and is even deteriorating in North-Eastern Gabon. The increase in malaria indices in Makokou therefore contrasts with the recent changes in malaria epidemiology reported in Libreville and Franceville, and is consistent with earlier observations in Central Africa pointing to a patchy pattern of malaria transmission according to geographic location. Congo, some rural settings in Angola, Central African Republic, Sudan, and some other neighboring regions, are areas where the prevalence of malaria is the highest. These areas inevitably serve as pockets of transmission, and a similar situation is likely to occur in Gabon [30,31]. As a consequence, these reservoirs of malaria infections should now become the main focus of malaria-targeted interventions [32].

The increase in the number of both presumptive and confirmed malaria cases, the stable rate of the percentage of positive smears (approximately 50%), and the higher risk of infection in young children between 1 and 4 years of age are consistent with a sustained transmission of malaria in the rural area of Makokou. Hence, the decline in malaria observed in urban and semi-urban coastal areas cannot be extended to the country as a whole. The medical technicians who screen for malaria and the methods used for diagnosis have remained mostly unchanged in health facilities in Makokou. Thus, it is unlikely that the difference in malaria prevalence is due to the poor performance of medical staff and microscopists who are mostly well trained and experienced. Changes in malaria incidence are also unlikely to be caused by the replacement of microscopy with a more sensitive approach such as rapid diagnostic tests (RDT). Indeed, the implementation of alternative methods for biological diagnosis is a top priority in Gabon but the use of RDT has been hindered by many challenging constraints for shipping and long-term storage of RDT in the field. As a result, the use of RDT remained rather limited in Gabon, at least until 2010. While there is no systematic quality control of slide readings in Gabon, we have confidence in the data for malaria cases reported because comparative surveys conducted since our study at the Makokou health facilities in 2012 and 2013 revealed a concordance rate of 76% (unpublished data) between microscopic examinations and RDT readings.

Malaria indices such as prevalence, incidence, and asymptomatic carriage are only slightly affected by seasonality in Gabon, and no major rainfall abnormalities were recorded in Gabon between 2000 and 2013 that could explain the malaria patterns observed [27,33]. Although annual variation in rainfall is modest in the Makokou region, the number of cases is twice as high in the peak seasons (May–June and November–December) than in off-peak seasons (February–April and July–September). This pattern is consistent with previous observations in regions where malaria is more seasonal: 1–2 months is required to re-establish larval habitats and for the re-initiation of the lifecycle of the parasite in host mosquitoes [34].

The epidemiological features and malaria trends observed in Makokou may be related to local factors. No demographic upsurge was apparent in the Ogooué-Ivindo Province during the study period. However, the construction of the hospital in 2011 possibly led to an increase in the frequency of medical consultations, and therefore in the number of malaria cases but, this element alone, cannot explain the differences in malaria trends we observed between Makokou and the Estuaire province as the number of cases increased from 1173 in 2006 to 3157 in 2012 in Makokou, corresponding to an increase by a factor of 5.3 over the period, whereas the population increased by a factor of only 1.2 over this period. Instead, the spread of resistance to conventional antimalarial drugs that continue to be used despite recommendations from the Ministry of Health, the uneven distribution of control and prevention methods across the country, and reluctance to seek treatment have undoubtedly contributed to the deterioration of the malaria situation over recent years. Artemisinin-based combination therapies (ACTs) for uncomplicated malaria were introduced in 2003 in Gabon, and the distribution of insecticide-treated mosquito nets started later on in 2005. The coverage of these interventions has however remained too limited to produce global and durable effects. In fact, ACTs were prescribed in only 24.5% of malaria cases in 2008 and in 52% of cases in 2010. Furthermore, the percentage of pregnant women receiving an intermittent preventive treatment (IPT) with sulfadoxine-pyrimethamine was largely unchanged between 2006 and 2010, remaining close to 60% for IPT1, and between 30% and 50% for IPT2. Between 2006 and 2010, only 323,586 ITNs were distributed for a national population of 1,556,222 in 2010. In 2010, the proportion of pregnant women who slept under an ITN was only 52% in the Ogouee-Ivindo Province, whereas in young children <5 years the proportion was 56% [9,28]. Incidentally, the decrease in malaria prevalence reported in Libreville and Port-Gentil between 2005 and 2008 was only observed later at Makokou, in 2008 and 2009 [7,10,11]. The delay in the decrease of malaria incidence in Makokou may presumably be a consequence of the late implementation of key health interventions. This finding is reminiscent of earlier studies showing that remote and underserved regions were the most affected by weakened efforts to combat malaria, due to inadequate funding, transport constraints and/or insufficient resources [35,36].

The analysis of malaria indicators is further clouded in Gabon by several confounding factors. Almost half of all cases are treated in private or informal sectors for which no statistics are available, suggesting that the incidence of falciparum malaria might be of about 200 per 1000 in the rural area of Makokou. Moreover, the current infrastructure is generally insufficient for the regular monitoring and follow-up of malaria patients. Many infections, although treated, are not confirmed by diagnostic tests. Finally, in the absence of a National Health Information System (NHIS) in Gabon, data are essentially collected from public hospitals in larger towns. The calculation of more accurate estimates of malaria burden in Gabon will require the implementation of screening programs, particularly in rural areas, to provide complete and representative values of various malaria indices, including incidence especially in children aged between 2 and 10 years [37]. The only national survey of this type carried out in Gabon between 2005 and 2011, reported a higher prevalence of infection in children <15 years old than in adults, and a higher risk of parasitic carriage in populations living in forest ecosystems [29]. Although these findings involve asymptomatic infections, they nonetheless illustrate the large heterogeneity of malaria endemicity in Gabon, and the divergent trends found in rural and urban areas.