Concurrent malaria and arbovirus infections in Kedougou, southeastern Senegal

The study was conducted in the Kedougou region, southeastern Senegal (Fig. 1) with an estimated population of 141,226 inhabitants, among whom 55 % are under 20 years old, and an average density of eight persons per sq km [19]. The Kedougou region borders Guinea, Mali and The Gambia, between isohyets 1200 mm and 1300 mm rainfall per year. The climate is Sudano-Guinean with a single rainy season from May to November [20]. The landscape consists of wooded grassland or woodland and dense gallery forest. The fauna is diverse with herbivores, insectivores, rodents, and monkeys.

Patients presenting with AFI were recruited from seven healthcare facilities of the Kedougou region, including Ninefesha rural hospital, Kedougou and Saraya health centres, Bandafassi and Khossanto health posts, the Kedougou military health post, and the Catholic Mission mobile team, which targets populations in remote areas (Fig. 1). AFI was defined as any patient older than 1 year of age with a fever (axillar temperature >38 °C) lasting for less than 2 weeks and exhibiting two or more of the following signs or symptoms: headache, myalgia, eye pain, arthralgia, cough, nausea/vomiting, diarrhoea, jaundice, bleeding, and neurological signs. Clinical manifestations and sociodemographic data were recorded on a standardized interview form. From each consented and eligible patient, health workers sampled 5 ml of venous blood in healthcare centres. The same day, samples were centrifuged and serum aliquoted and transported to the field where they were tested and stored at −20 °C until further use.

Human sera were systematically tested for malaria parasite and arbovirus infections. The malaria diagnosis procedure was performed using thick blood smear (TBS) and an RDT (Malaria Antigen P.f, Standard diagnostics, Ingbert, Germany) as previously described [21, 22], while IgM ELISA, real-time RT-PCR and other viral infection detection methods [23‐26] were used for detection of DENV, CHIKV, ZIKV, Crimean-Congo haemorrhagic fever (CCHFV) or West Nile virus (WNV), YFV, and RVFV infections. Negative and positive controls were used for serology and molecular methods and any samples with Ct value 35 was considered as positive for RT-PCR. Consequently, a case of confirmed arboviral infection was defined as any AFI that tested positive by any method used for detection of IgM and/or the genome of ZIKV, DENV, CHIKV, RVFV, YFV, CCHFV, or WNV, while a confirmed malaria case was defined as any AFI testing positive for TBS. Only TBS was used for the definition of confirmed malaria case as it is the gold standard for malaria diagnostics. Concurrent infection was defined as any AFI confirmed for both arboviral and malaria infections.

Data were analysed using STATA [27] and R software [28]. For small numbers of samples, exact methods, such as exact logistic regression and Fisher’s exact test or other non-parametric methods such as Kruskal–Wallis test, were used for qualitative and quantitative variables to investigate differences in measures between sub-groups. Significance was assigned at p < 0.05 and when pair-wise comparisons were performed among more than two groups, p values corrected for multiple testing by the Bonferroni method were considered.

The study objectives, benefits and risks were explained in the French language or local dialects to all participants before inclusion. Written informed consent was obtained from all adult participants and from parents or legal guardians of children. The study was examined and approved by the Senegalese National Health Research Committee.

From July 2009 to March 2013, 13,845 patients with AFI (representing 80 % of the patients with AFI attending the healthcare centres) were enrolled from the following seven health facilities: Ninefesha rural hospital (13.5 %), the Kedougou health centre (13.0 %), Saraya health centre (4.8 %), Bandafassi health post (9.7 %), the Military health post (53.8 %), Khossanto health post (3.6 %), and the Catholic mission mobile team (1.4 %). The median age was 16 years, ranged from 1 to 90 years, and the male/female sex ratio was 1:4. Among enrolled patients, 7387 (53.4 %) were malaria-confirmed cases (median age 13 years (range 1–90); male/female sex ratio 1:35), and 41 (0.3 %) were arboviral-infected individuals i.e., 12 YFV, nine ZIKV, one RVF, 16 CHIKV, and three DENV (median age 22 years [range 2–46]; male/female sex ratio 1:4). No patients tested positive for WNV or CCHF. The distribution of positivity for IgM and RT-PCR are summarized in Table 1.

As shown in Fig. 2, several major patterns can be evidenced from data collected during the 2009–2013 period: (i) the malaria parasite and arboviral infection peaks were coincidental; (ii) the yearly prevalence of malaria was stable with respect to dry and rainy seasons as shown by the red curve; and, (iii) different arboviruses circulated during malaria occurrence in this region of Africa. During rainy seasons, periods of high malaria prevalence, 15 positive IgM for CHIKV (green bars) and three for DENV (yellow bars) in 2009, six for YFV (red bars) in 2010, six for YFV (red bars) and nine for ZIKV (blue bars) in 2011, and one RVFV (dark grey bars) in 2012, were detected among samples. No arboviral infection was observed during dry seasons from 2009 to 2012 except one CHIKV acute illness IgM positive in 2010.

Among patients infected with arboviruses, 48.7 % (20/41) were concurrently harbouring positive testing for malaria parasites. Distinct differences in co-infection rates were observed among the tested arboviruses. Malaria co-infection was identified in 18.7 % (3/16) of CHIKV, 58.3 % (7/12) of YFV, 88.9 % (8/9) of ZIKV, 33.3 % (1/3) of DENV, and 100 % (1/1) RVF patients. These co-infection rates were not significantly different from one another except between CHIKV and ZIKV (Fisher exact p value = 0.021 after multiple tests correction). Given the prevalence of malaria and the circulation of distinct arboviruses in this region, these observed co-infection rates reached expected levels, with the exception of CHIKV where co-infection was significantly lower than expected (empirical p = 0.0089). Patients concurrently infected with malaria parasites and arboviruses were significantly younger, with median age of 14 years compared to 29 years in non-co-infected patients (p = 0.004, Kruskal–Wallis rank test). Among co-infected patients, 67 % presented high grade fever (temperature ≥40 °C), while only ten and 34 % of patients, respectively, were diagnosed with a single arboviral or malaria infection with temperature ≥40 °C (Tables 1 and 2). Therefore, co-infected patients had a significantly higher risk of exhibiting a high-grade fever than those infected with a single pathogen (OR = 19.5, CI = [3.38; 112.45], exact p value = 2 × 10–4). Also any other symptom was found significantly associated with co-infected patients (Table 3).