Arboviruses as an unappreciated cause of non-malarial acute febrile illness in the Dschang Health District of western Cameroon

Acute febrile illness is a common problem managed by clinicians and health systems globally, particularly in the Tropics. In many regions, malaria is a leading and potentially deadly cause of fever; however, myriad alternative etiologies exist. Identifying the cause of fever allows optimal management, but this depends on many factors including thorough knowledge of circulating infections. Arboviruses such as dengue (DENV) cause fever and may be underdiagnosed in sub-Saharan Africa where malaria is a major focus. We examined cases of fever in western Cameroon that tested negative for malaria and found 13.5% (13/96) were due to DENV, with 75% (9/12) of these being DENV serotype 2 infections. Two complete DENV2 genomes were obtained and clustered closely to recent isolates from Senegal and Burkina Faso. The seroprevalence of DENV in this region was 24.8% (96/387). Neutralizing antibodies to DENV2 were detected in all (15/15) seropositive samples tested. Chikungunya (CHIKV) is an arthritogenic alphavirus that is transmitted by Aedes mosquitoes, the same principal vector as DENV. The seroprevalence for CHIKV was 15.7% (67/427); however, CHIKV did not cause a single case of fever in the 96 subjects tested. Of note, being seropositive for one arbovirus was associated with being seropositive for the other (Χ2 = 16.8, p<0.001). Taken together, these data indicate that Aedes-transmitted arboviruses are endemic in western Cameroon and are likely a common but underappreciated cause of febrile illness. This work supports the need for additional study of arboviruses in sub-Saharan Africa and efforts to improve diagnostic capacity, surveillance systems, and arbovirus prevention strategies.


Introduction
Acute febrile illnesses (AFI) commonly prompt presentation to medical attention in sub-Saharan Africa and represent challenging clinical problems. There are numerous possible etiologies, and the availability of diagnostic testing for non-malarial causes is generally limited. [1][2][3] Malaria remains the predominant concern in the febrile patient, and empiric antimalarials are administered in many areas for undifferentiated fever. While missing a diagnosis of malaria could have grave consequences, overuse of antimalarials could promote resistance and confer unnecessary side effects, as well as delay optimal management of the true cause of the presenting illness. Studies of AFI in sub-Saharan Africa (SSA) have identified some patterns, but what is most notable is the heterogeneity of reported etiologies. [2] Thus, further development of public health capacity, including surveillance systems, is needed to inform the best decisions for clinical management and health resource allocation at the local level. To address gaps in epidemiologic knowledge in western Cameroon, we asked whether arbovirus (ARBV) infection was a common cause of non-malaria acute febrile illness (nmAFI).
ARBVs transmitted by the anthropophilic vector Aedes aegypti [4,5] include dengue (DENV), Zika (ZIKV) and chikungunya (CHIKV) viruses and represent a major global health problem. [6,7] DENV is estimated to cause upwards of 400 million infections each year, [8,9] with 3.6 billion people at risk of acquiring a DENV infection. [10,11] DENV was responsible for about 41,000 deaths in 2017, [12] with an economic cost estimated at 2.1 billion USD between 2000-2007 for the Americas. [13] However, the precise global distribution and burden of DENV remains highly uncertain. [14] In SSA, DENV is endemic in at least 34 countries. There is currently no report of routine DENV diagnosis and treatment in health systems in Africa despite evidence of continuous, ongoing DENV transmission in East and West Africa, [1,15,16] and more recently in central Africa. [17][18][19] Consequently, febrile children and adults are managed as presumptive cases of malaria or bacterial infections, which limits the attention and awareness for infections caused by DENV or other viruses. Indeed, over treatment of malaria in Africa is well documented. Up to 70% of patients in Cameroon [20] and 88% in Ghana who received an artemisinin combination were negative for malaria by rapid diagnostic test. [21] The entrenched clinical paradigm of managing undiagnosed febrile patients with antimalarials, the lack of guidelines for ARBV control programs, and the accumulating record of DENV transmission among travellers to and residents of Cameroon and other places in SSA [2,3,[22][23][24][25][26] all point to the probability that DENV is an important public health problem in SSA. [27] In this study, we hypothesised that ARBV such as DENV represent an important but underappreciated etiology of febrile illness in SSA countries such as Cameroon and investigated the hospital prevalence of DENV infection in a cross-sectional acute febrile illness survey of patients suspected of having malaria infection in Dschang, West region of Cameroon.

Ethics statement
The study was reviewed and approved by the institutional review board (IRB) of the Cameroon Baptist Convention Health Board (FWA00002077), Protocol IRB2019-40. Written informed consent was administered in French or English based on participant preference, via an independent translator (who also spoke the local language Yemba) as needed. For children, informed consent was obtained from a parent or guardian.

Study design and operations
Design. The analysis presented in this article was a secondary analysis of remnant samples collected during a parent study focused on malaria. The parent study was a prospective hospital based cross-sectional survey in the three main health facilities in the Dschang Health District in the West region of Cameroon: Dschang district hospital, St Vincent Catholic Hospital and Batsinglah Catholic Hospital.
Study population and site description. The parent study for this project was designed to study the species and genetics of parasites causing malaria in this region of Cameroon (S1 Text). In total, 431 patients were enrolled between June 12, 2020 -September 8, 2020. Selected demographic variables are reported in Table 1, and clinical symptoms are detailed in S1 Table. The Dschang health district with its 22 health areas has a surface area of about 1060 km 2 . It is a tropical, semi-urban environment at an elevation of 1380-1400m above sea level. The rainy season lasts mid-March-mid October.
Recruitment of study participants and biospecimens. Inclusion criteria of the parent study included fever (axillary temperature �37.5˚C or self-reported history of fever) in past 24 hours without signs and symptoms suggestive of severe malaria. Participants were excluded if they had used anti-malarial medicine in the past 14 days. Dried blood spots (DBS) were collected via finger prick. For serologic studies, plasma proteins were eluted from DBS as previously described. [23,28,29] DBS eluate was heat inactivated for 30 minutes at 56˚C and stored at 4˚C for up to 1 week or at -20˚C until use. One 6-mm hole punch of each DBS was reserved for molecular testing.
Laboratory assays and cartography (see S1 Text for greater detail). Samples were tested as extensively as allowed by specimen quantity and assay costs. Testing of 96 malaria-negative fever cases by real-time reverse transcriptase-polymerase chain reaction (rRT-PCR) was prioritized, followed by serologic testing of all samples with remaining material for DENV and CHIKV. Genomic sequencing could be accomplished from the same nucleic acid extraction as rRT-PCR testing.
rRT-PCR testing. Total nucleic acids were extracted from DBS on an EMAG instrument (bioMérieux, Durham, NC) by first placing 6mm punches into 400μL of Nuclisense lysis buffer and incubating overnight at room temperature on a tube rocker. Eluted nucleic acid (5μL) was immediately tested by multiplex rRT-PCR for ZIKV, CHIKV, and DENV, [30] and DENVpositive samples were subsequently tested in a DENV multiplex assay to determine the serotype (S1 Fig). [31,32] DENV genome sequencing and phylogenetic analysis. Seven samples underwent unbiased metagenomic sequencing followed by DENV-specific analysis. Reads underwent metagenomic classification and DENV genome assembly. For phylogenetic analysis, all complete DENV2 genomes were downloaded from NCBI on 7/3/22 (N = 3,229), and a maximum likelihood phylogenetic tree was constructed as described in S1 Text. After Cameroon sequences were confirmed to cluster with other DENV2 sequences of the Cosmopolitan genotype from West Africa (S2 Fig), a subset of 29 West African DENV2 Cosmopolitan sequences were used for further analysis (S2 Table). Sequence GQ398264 (Indonesia, 1976) was included as an outgroup. Maximum-likelihood and time-scaled Bayesian phylogenetic analyses were performed as described in the Supplementary Methods.
Antigen capture IgG ELISA. Binding IgG to DENV or CHIKV was measured by antigen capture ELISA as previously described. [34,35] Virus-reactive monoclonal antibody was used to capture virus antigen. Serum IgG binding to viral antigen was detected by an alkaline phosphatase-conjugated goat anti-human IgG secondary Ab and p-nitrophenyl phosphate Table 1. Sociodemographic characteristics of study population and association with DENV infection outcomes. Map creation. Data were summarized by study site (S3 Table) and were mapped to study site locations using pie charts. All maps were made using ArcGIS version 10.7.1 with Cameroon boundary shapefiles obtained from GADM.org.

Data management and statistical analysis
Clinical and sociodemographic data were collected at enrolment and were entered into the study database. Descriptive statistics were used to report prevalence of DENV cases and DENV and CHIKV seroprevalence. Associations between DENV infection outcomes and sociodemographic data were investigated by analysing cleaned data in Statistical Analysis Software (SAS). A logistic regression was performed to analyse the relationship between the parameters of interest and each of the infection outcomes. However, due to the highly dimensional data, the analysis model was prone to overfitting. Thus, a Chi-square test was conducted to compare the categorical variables to the outcome of interest. Fisher's exact test was conducted for variables with less than 5 observations when appropriate.

ARBV infection
Of 96 subjects with fever and negative malaria testing, 13 (13.5%) tested positive for DENV by ZCD assay [30] ( Table 2). DENV2 was identified in 9 samples (75%), and DENV3 was identified in 4, with a single sample testing positive for both serotypes. No samples were positive for ZIKV or CHIKV RNA, and all 96 samples yielded a positive signal for the RNase P target (S4 Table). One sample could not be serotyped.
Samples with the lowest Ct values for DENV2 were sequenced. Approximately 1 million unbiased metagenomic sequencing reads were generated per sample (range 850,949 to 1,097,048, S5 Table). Complete DENV genomes were assembled from samples DSG2020_345 and DSG2020_061, confirming serotype 2. No DENV reads were detected in the other five samples (S5 Table), which can be attributed to their relatively low DENV RNA content (C T values ranging from 34.95 to 42.90 in the ZCD assay), combined with our low-depth, unbiased sequencing approach; although a DENV-targeted approach likely would have yielded some DENV reads, it would be unlikely to allow full-genome sequencing and phylogenetic analysis for low-input samples such as these. Although one sample (DSG2020_346) contained GB virus C, no other potential pathogens were identified by metagenomic analysis. The two DENV2 genomes belonged to the Cosmopolitan genotype, and fell within a sub-lineage with sequences that have been circulating in West Africa for 20-35 years. [40,41] Interestingly, these sequences were distinct from and ancestral to sequences obtained from Burkina Faso, Cote d'Ivoire, and Senegal between 2016 and 2019 (Fig 1A). The Cameroon sequences, which were from samples collected in 2020, shared a common ancestor in approximately 2017 and diverged from other West African sequences in approximately 2013 (Fig 1B).
To better understand the extent of ARBV transmission in the region, we tested for IgG antibodies to DENV and CHIKV, which revealed a seroprevalence of 24.8% (96/387) for DENV and 15.7% (67/427) for CHIKV (Fig 2A-2D). The ELISA testing approach was validated by testing a small random subset of samples for neutralizing antibodies (Fig 2E and 2F). All 15 DENV IgG+ samples exhibited neutralizing antibodies to DENV2, the most prevalent DENV serotype identified among non-malarial acute fever cases tested by rRT-PCR. To assess for coherence in ARBV infection, we analyzed the proportion of samples positive for one virus  grouping by serostatus for the other virus (Fig 2C and 2D). There was a significant relationship between DENV and CHIKV seropositivity (χ 2 (1, n = 383) = 16.777, p�0.00005), consistent with the hypothesis that risk for multiple Aedes-borne viruses would be concentrated among some individuals.

Risk factors for ARBV infection
There was no clear geographic clustering by site for DENV cases or seroprevalence (Fig 3A  and 3B). A Chi-square independence test showed that there were no significant differences between the sociodemographic factors and the epidemiologic outcomes of DENV case or seropositivity (p> 0.05) ( Table 1). Analysis of a larger sample may reveal interesting differences in seroprevalence stratified by age, but no statistically significant relationships could be detected in the sample available for this study for age and DENV seroprevalence, χ 2 (3, N = 387) = 7.10, p = 0.0687.

Discussion
In this study, we demonstrated that DENV is clearly an important cause of AFI in SSA (at least in Cameroon). However, there may be substantial local and regional heterogeneity. Our study captured an outbreak driven primarily by DENV2 during 2020. The age distribution of DENV seroprevalence is consistent with endemic or intermittent transmission of DENV for at least several decades in the region. The seroprevalence of CHIKV further indicates that local ecology is amenable to human infection by multiple viruses transmitted by Aedes aegypti. Our two DENV2 genomes are important contributions to the scant phylogeographic data on DENV in SSA and suggest that a distinct lineage of the Cosmopolitan genotype of DENV2 may have been circulating unrecognized in western Cameroon for several years.
Our study adds to the limited existing evidence of DENV (including serotypes 1, 2 and 3) transmission in Cameroon in several regions. [3,17,[42][43][44][45][46][47][48][49][50][51] CHIKV has been less commonly reported in Cameroon, [48,[52][53][54] with exception of a notable outbreak in 2006. [53,54] The modest seroprevalence (~15%) of CHIKV but lack of cases in the AFI cohort illustrates a few important points. It is characteristic of many ARBV to exhibit unpredictable periodic outbreaks with intervening periods of low or absent transmission. The absence of detection of an ARBV like CHIKV in a single cohort study does not exclude this virus from being an important consideration in the etiology of AFI in this region. In fact, the seroprevalence data from our study and the sustained circulation of CHIKV in certain regions of Africa indicate the relevance of this pathogen. The second point is that diagnostic capacity is essential to support clinical practice and public health management of the myriad causes of fever in SSA. The association of DENV and CHIKV seropositivity strongly argues that common risk factors exist for Aedes-borne ARBV in this area, and there are several reasons why a comprehensive understanding of ARBV transmission dynamics is critical. As demonstrated by our study, ARBV such as DENV are important causes of fever, but attack rates can vary broadly from season to season or between relatively small distances. For example, 17.4% of 682 non-malaria febrile cases in a large health facility in Gabon were diagnosed as DENV, [18] and in Madagascar, DENV was the most common cause of AFI in one recent study. [55] Local epidemiology determines pre-test probability of a given diagnosis, which affects clinical decision making. Moreover, epidemiologic data permit priority setting and resource allocation by public health authorities. Detecting the true prevalence and incidence of ARBV infections can advocate for enhancing diagnostic capacity. Finally, the burden of ARBV infection represents a solvable global health problem. Vaccinology continues to advance, and it is likely that safe and effective vaccines could soon be available for ARBV including DENV, CHIKV and ZIKV. [56][57][58][59][60] There has also been reinvigoration of control efforts targeting mosquito vectors. [61][62][63][64][65] Effective implementation of these requires thorough knowledge of the local transmission ecology to design, monitor and maintain the benefit of novel interventions.
Limitations of our study include the relatively low sample amount, precluding more extensive serologic testing for other ARBV. For example, we did not test for IgM, which could have increased sensitivity for capturing acute ARBV infections. A longer longitudinal study would be necessary to capture cases of other ARBV, given the typical sporadic epidemic transmission patterns of these pathogens. Also, we were unable to collect follow up data on clinical outcomes of our DENV cases. Despite this constraint, we were able to generate a robust data set and provide new insight into ARBV transmission dynamics in SSA. DBS are an innovative tool that allows convenient sampling of large numbers of people in resource limiting setting. We and others have used DBS-based serology to study the epidemiology of arboviruses or other infectious diseases in both standard ELISA as well as multiplex platforms. [23,28,29,[66][67][68][69][70][71][72] We have recently generated data on the stability of ARBV genomic RNA on DBS for RT-PCR diagnostic testing as well as metagenomic sequencing, [73] which will expand the possibilities for clinical studies and surveillance activities.

Conclusion
In conclusion, this work highlights the need to consider DENV and other ARBV in the differential diagnosis of AFI in SSA. Our data indicate that western Cameroon is likely not hyperendemic for DENV, but sustained transmission of at least the DENV2 serotype in this region has likely been established for many years. Further study could also clarify if there is ongoing spill over from non-human mammalian hosts [48] or involvement of vector spp. other than Aedes aegypti and albopictus mosquitoes. [19,74] The methods used in this study provide an example of robust approaches to study ARBV. Optimal clinical management of febrile patients, along with efficiency and effectiveness of public health interventions such as vector control, would be strengthened were diagnostic capacity and surveillance systems improved and prioritized.
Supporting information S1