Skip to main content
Advertisement
  • Loading metrics

Clinical and epidemiologic characteristics associated with dengue during and outside the 2016 outbreak identified in health facility-based surveillance in Ouagadougou, Burkina Faso

  • Jacqueline K. Lim ,

    Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing

    kalim@ivi.int

    Affiliations Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Yaro Seydou,

    Roles Data curation, Project administration, Supervision, Writing – review & editing

    Affiliation Department of public health epidemiology, Centre MURAZ, Bobo-Dioulasso, Burkina Faso, Africa

  • Mabel Carabali,

    Roles Conceptualization, Data curation, Investigation, Methodology, Project administration, Resources, Writing – review & editing

    Affiliations Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea, Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, Canada

  • Ahmed Barro,

    Roles Data curation, Project administration

    Affiliation Action-Gouvernance-Integration-Renforcement (AGIR), Program Equité, Ouagadougou, Burkina Faso, Africa

  • Desire Lucien Dahourou,

    Roles Data curation, Investigation

    Affiliations Department of public health epidemiology, Centre MURAZ, Bobo-Dioulasso, Burkina Faso, Africa, Institut de Recherché en Sciences de la Santé, Ouagadougou, Burkina Faso, Africa

  • Kang Sung Lee,

    Roles Investigation, Resources, Software, Supervision

    Affiliation Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea

  • Teguewende Nikiema,

    Roles Data curation, Investigation, Project administration

    Affiliation Department of public health epidemiology, Centre MURAZ, Bobo-Dioulasso, Burkina Faso, Africa

  • Suk Namkung,

    Roles Data curation, Project administration

    Affiliation Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea

  • Jung-Seok Lee,

    Roles Conceptualization, Methodology, Writing – review & editing

    Affiliation Department Of Zoology, University of Oxford, Oxford, United Kingdom

  • Mee Young Shin,

    Roles Data curation

    Affiliation Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea

  • Emmanuel Bonnet,

    Roles Data curation, Visualization

    Affiliation Institute for Research on Sustainable Development (IRD), UMI Résilience, Bondy, France

  • Therese Kagone,

    Roles Data curation, Investigation

    Affiliation Department of public health epidemiology, Centre MURAZ, Bobo-Dioulasso, Burkina Faso, Africa

  • Losseni Kaba,

    Roles Data curation, Resources

    Affiliation Centre National de Transfusion Sanguine, Ouagadougou, Burkina Faso, Africa

  • Tansy Edwards,

    Roles Formal analysis, Writing – review & editing

    Affiliation Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • Paul-André Somé,

    Roles Data curation, Project administration

    Affiliation Action-Gouvernance-Integration-Renforcement (AGIR), Program Equité, Ouagadougou, Burkina Faso, Africa

  • Jae Seung Yang,

    Roles Data curation, Supervision, Writing – review & editing

    Affiliation Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea

  • Neal Alexander,

    Roles Methodology, Supervision, Writing – original draft, Writing – review & editing

    Affiliation Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom

  • In-Kyu Yoon,

    Roles Project administration, Resources, Supervision, Writing – review & editing

    Affiliation Global Dengue and Aedes-transmitted diseases Consortium, International Vaccine Institute, Seoul, Korea

  •  [ ... ],
  • Valéry Ridde

    Roles Conceptualization, Data curation, Investigation, Resources, Supervision, Writing – review & editing

    Affiliation Institute for Research on Sustainable Development (IRD), CEPED, Université de Paris, ERL INSERM SAGESUD, Paris, France

  • [ view all ]
  • [ view less ]

Abstract

Background

In Africa, the magnitude of dengue virus (DENV) transmission is largely unknown. In Burkina Faso, several outbreaks have been reported and data are often based on findings from outbreak investigations.

Methods

To better understand dengue epidemiology and clinical characteristics in Burkina Faso, a fever surveillance study was conducted among patients aged 1–55 years, who presented with non-malarial febrile illness at five primary healthcare facilities in Ouagadougou, Burkina Faso from December 2014 to February 2017, encompassing a 3-month dengue outbreak in September-November 2016. Acute and convalescent blood samples were collected within an interval of 10–21 days between visits. Acute samples were tested with dengue rapid diagnostic tests (RDT) and a selected subset with RT-PCR, and all acute/convalescent samples with IgM/IgG ELISA.

Results

Among 2929 non-malarial febrile patients, 740 (25%) were dengue–positive based on RT-PCR and/or IgM/IgG ELISA; 428 out of 777 patients (55%) and 312 out of 2152 (14%) were dengue-positive during outbreak and non-outbreak periods, respectively. There were 11% (316/2929) and 4% (129/2929) patients showing positive for NS1 and IgM, on the RDT, respectively. DENV 2 predominated during the outbreak, whereas DENV 3 predominated before the outbreak. Only 25% of dengue-positive cases were clinically diagnosed with suspected dengue. The odds of requiring observation for ≤3 days (versus routine outpatient care) were 11 times higher among dengue-positive cases than non-dengue cases. In adjusted analyses, dengue-positivity was associated with rash and retro-orbital pain (OR = 2.6 and 7.4, respectively) during the outbreak and with rash and nausea/vomiting (OR = 1.5 and 1.4, respectively) during the non-outbreak period.

Conclusion

Dengue virus is an important pathogen in Burkina Faso, accounting for a substantial proportion of non-malarial fevers both during and outside outbreak, but is only infrequently suspected by clinicians. Additional longitudinal data would help to further define characteristics of dengue for improved case detection and surveillance.

Author summary

There is not much evidence on dengue in Africa, relative to the Asia-Pacific and Latin American regions. To estimate the proportion of dengue among patients with fever, and to identify clinical features of dengue during outbreak and non-outbreak periods, we studied 2929 patients with non-malarial fever, aged 1–55 years, who attended five primary healthcare centers in Ouagadougou, Burkina Faso. Patients were tested with a rapid test for dengue, and further tests were carried out on paired blood samples taken 10–21 days apart. Overall, a quarter of non-malarial febrile episodes identified between December 2014 and February 2017 were dengue-positive. Dengue-positive cases were 11 times more likely than non-dengue cases to require observation for ≤3 days. During the study period in 2016, there was a dengue outbreak where more than half of non-malarial febrile patients were identified to be dengue-positive. DENV 2 was the main serotype in circulation during the outbreak, whereas DENV 3 was the main serotype before the outbreak. Rash and retro-orbital pain were more frequently found among dengue-positive cases, compared to non-dengue cases, during the outbreak. During the non-outbreak period, rash and nausea/vomiting were more likely in dengue-positive versus non-dengue cases. There was a low level of clinical suspicion of dengue even during the 2016 outbreak. Therefore, a broader use of rapid diagnostic tests and more epidemiologic data would help to improve dengue case detection and surveillance in Burkina Faso.

Introduction

Dengue Fever (DF) is a mosquito-borne disease caused by four related but antigenically distinct dengue viruses (DENVs, serotypes 1–4). Approximately 50 to 100 million cases of DF and 500,000 severe dengue cases requiring hospitalization reportedly occur annually worldwide [13].

The Aedes mosquito vectors of DENV are widely distributed in Africa, and dengue cases have been reported in 34 African countries [46]. However, data are limited to retrospective testing of existing samples or outbreak investigations from a few countries [5, 79]. Several studies have identified DENV as a common cause of febrile illness in Africa, but there is a continued challenge to distinguish dengue from other causes of febrile illness given limited diagnostic capabilities [1012].

In Burkina Faso, several outbreaks have been reported since 1925 [5, 13, 14], including an outbreak declared in November 2013 by the Burkina Faso Ministry of Health (MoH) [11, 15]. Between 5 August and 12 November 2016, the Burkina Faso MoH conducted an outbreak investigation as part of emergency response in collaboration with World Health Organization (WHO) and 1266 suspected dengue cases were identified by the MoH, with 1061 cases positive by dengue rapid diagnostic test (RDT), and 15 deaths from all 12 districts of Ouagadougou [16, 17]. Most recently, an even larger outbreak occurred in September 2017, with 9029 suspected dengue cases, 5773 dengue RDT-positive cases, and 18 deaths throughout the country [18]. These repeated outbreaks suggest a considerable dengue burden in Burkina Faso.

Most African countries lack mandatory reporting or national surveillance systems for dengue [19]. Burkina Faso added dengue to its routine national surveillance system for diseases with epidemic-potential in 2016. Also, the MoH conducts outbreak investigations at several sentinel health centers [11].

To better understand the dengue problem in Burkina Faso, a passive facility-based fever surveillance study was conducted in Ouagadougou, from 2014–2017. During the study period, the 2016 dengue outbreak occurred, allowing for characterization of dengue epidemiology and comparison of clinical features during and outside the outbreak.

Methods

Study area and population

The study area was selected based on the existence of previous outbreaks and case reports, past seroprevalence and modelling studies, as well as the availability of research infrastructure [4, 20, 21]. Ouagadougou is the capital city of Burkina Faso in West Africa with most of its population residing in urban settings [22]. In March-May, temperatures may reach 43°C, and it is followed by the rainy season in May-September. Health services in Ouagadougou are provided by three university hospitals, five district hospitals, and 60 primary healthcare centers (CSPS, Centres de Santé et de Promotion Sociale), as well as private clinics [23].

The current study was implemented in five CSPSs (Pazani, CSPS22, CSPS25, Juvenat Fille, Zongo), serving a catchment population of 110,000 residents (Fig 1). The population in Ouagadougou is stable with an annual transmigration rate of 4.1% and >80% with home ownership [24].

thumbnail
Fig 1. A map of the study area in Ouagadougou, Burkina Faso.

The map shows the approximate location of the selected facilities of 5 CSPSs (Pazani, CSPS22, CSPS25, Juvenat Fille, Zongo), serving a catchment population of 110,000 residents of Ouagadougou, Burkina Faso [20].

https://doi.org/10.1371/journal.pntd.0007882.g001

Study design

Investigational methods can be found in previous publication [20]. The passive facility-based fever surveillance study enrolled outpatients and observation patients (for ≤3 days), as previously described [20], between December 2014 and February 2017 (27 months). Patients presenting with fever (body temperature ≥ 37.5º C) or history of (self-reported) fever for ≤7 days were tested for malaria using RDT (SD BIOLINE Malaria kit, Standard Diagnostics, Yongin-Si, Korea) as part of routine practice. Patients were eligible for study enrolment if they were malaria RDT-negative without localizing signs (i.e., no localized infection or known/confirmed non-dengue etiology), aged 1–55 years, resident of the catchment area covered by the study CSPSs, and provided informed consent, plus assent for individuals aged 8–17 years.

Malaria RDT-negative patients were tested using dengue RDTs. During the enrollment visit, an acute blood sample (7–10 ml) was collected (Fig 2). Then, a study physician/nurse conducted interviews and physical exams, and a surveillance case report form was completed capturing symptom history, medical history, treatment and laboratory results [20]. A convalescent blood sample was collected at the facility between 10–14 days after the initial visit, or if not possible within this timeframe, the patient was followed up at home within 21 days.

thumbnail
Fig 2. A chart of patient flow in passive fever surveillance.

A chart of patient flow in passive fever surveillance- The chart shows the study flow when a febrile patient presents at the CSPS from screening, enrollment, and lab testing.

https://doi.org/10.1371/journal.pntd.0007882.g002

Laboratory testing algorithm

Laboratory testing algorithm has been described in previous publication [20]. As described previously [20], acute samples were tested at enrollment at the CSPS using a commercial RDT for dengue nonstructural protein 1 (NS1) and Immunoglobulin type M and type G (IgM/IgG) (Dengue Duo, Standard Diagnostics, Yongin-Si, Korea). The acute and convalescent samples were transported in 4ºC ice boxes to Virology laboratory of CHU YO (University Hospital Center Yalgado Ouédraogo, in French: “Centre Hospitalier Universtaire Yalgado Ouédraogo”) where blood samples were centrifuged and separated into cryotubes in 0.5–1 ml serum aliquots under sterile conditions, labeled and stored at -70°C freezer. Subsequently, they were brought and tested in the Centre Muraz laboratory using dengue IgM/IgG Enzyme-Linked Immunosorbent Assay (ELISA) (SD Dengue IgM & IgG Capture ELISA, Standard Diagnostics, Yongin-Si, Korea). Furthermore, as described in previous publication [20], Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) for laboratory confirmation of dengue infection and serotyping [25] was performed at the International Vaccine Institute (IVI), on acute sera from patients who had: (i) NS1 or IgM positive by RDT in the acute sample; and/or (ii) sero-conversion between acute and convalescent samples by IgM and IgG capture ELISA. RT-PCR was also performed on a limited number of randomly selected acute sera that were: (iii) sero-positive in both acute and convalescent samples by IgM and IgG capture ELISA; or (iv) IgG positive by RDT in the acute sample; or (v) negative by RDT and ELISA on all samples.

Dengue infection status was categorized based on interpretation of laboratory results, following WHO diagnostic criteria [26]. Sero-conversion by dengue IgM and/or IgG between acute and convalescent samples and/or virus detection by RT-PCR in the acute sample were considered to be laboratory-confirmed dengue. Positive IgM by ELISA in a single acute sample or paired acute/convalescent samples, or NS1 and/or IgM positive by RDT were considered as probable dengue [26]. Confirmed and probable dengue cases were combined into a dengue-positive group for this analysis. Patients with negative RT-PCR and negative paired acute/convalescent IgM ELISA were classified as non-dengue.

Statistical analysis

There were 2 components in the analysis. First, a descriptive summary of clinical and laboratory characteristics is presented for dengue-positive and non-dengue cases. Elevated body temperature, as a dichotomous variable, was defined as body temperature ≥38.5°C, the 75th percentile of the body temperature measured at enrollment. Clinical diagnosis (i.e., made by clinician prior to laboratory confirmation) was grouped as suspected dengue, undifferentiated fever, and other illness. Our surveillance covered the entire outbreak from September to November 2016. Cases were also designated as outbreak or non-outbreak depending on date of occurrence, with outbreak cases considered as those occurring between September and November 2016, defined to be consistent with the outbreak period declared by Burkina Faso MoH/WHO [16, 17]. Yellow fever (YF) vaccination history was dichotomized between those who reported having been vaccinated versus those who did not remember or reported no vaccination. Categorical pair-wise comparisons were made across dengue infection status using χ2 or Fisher’s exact tests with significance level of 0.05 [27]. Continuous variables were compared using Student’s t-test or ANOVA [28].

Secondly, based on our a priori hypothesis that clinical presentation associated with dengue-positivity would be different between the outbreak and non-outbreak periods, logistic regression was used to build a multivariable model of clinical indicators associated with dengue-positive vs. non-dengue cases, to separately fit the outbreak and non-outbreak periods. The models contained age and gender as a priori confounders, possibly associated with exposure to Aedes vectors, and with some clinical features [29]. A backward stepwise process was used to select a final multivariable model for each outbreak status, with a significance level of 0.2 for entry and 0.1 for retention. Further variables investigated included: demographic and clinical variables such as YF vaccination history, requirement for observation, fever duration prior to enrollment, temperature at presentation, and clinical signs/symptoms. Some signs and symptoms were used only in the descriptive and univariate analyses, due to data sparsity. Clinical diagnosis of suspected dengue was considered to be closely related to the outcome of dengue-positivity and was not included.

Finally, a single set of variables was obtained as the union of the sets of variables from regression modelling in the outbreak and non-outbreak periods. Variables found to be significant in only one period were applied to both periods, producing a single list of variables. These variables were fitted to both outbreak and non-outbreak periods to give comparable results between them.

As part of sensitivity analysis, a descriptive summary of clinical and laboratory characteristics using three categories for dengue infection status—confirmed, probable, and non-dengue—is presented in supplementary S2 Table. Between dengue-confirmed and non-dengue groups, univariate logistic analyses were conducted for during and outside the outbreak (S3 and S4 tables). All analyses were performed using SAS version 9.4 (SAS Institute, Cary, North Carolina).

Ethical considerations

The study protocol received ethical approvals from the Institutional Review Boards (IRBs) of IVI (No. 2014–008), the London School of Hygiene and Tropical Medicine (Reference number: 17096), the National Ethical Committee for Health Research of Burkina Faso, and the Ethics Committee of the Centre Hospitalier de l'Université de Montréal (CRCHUM) at University of Montreal.

A written informed consent from (ICF) was obtained from each participant. For those aged between 8 and 17 years, an assent form was obtained, plus informed consent from at least one parent or legal guardian.

Results

Analysis was performed on 2929 out of 3012 enrolled patients with complete clinical and laboratory data; 83 withdrew consent or had incomplete laboratory data to determine dengue infection status (Fig 3). Although similar in terms of age, gender, requirement for observation, and days of illness before enrollment, these 83 patients were significantly different from the analysis sample in terms of residential neighborhood—the majority from Zongo (40%) and Pazani (28%)—and being mostly from non-outbreak periods (87%). In terms of missing data, only the patients requiring observation had information on the complete blood count (CBC) test and the results from CBC were not included in the analysis.

thumbnail
Fig 3. A chart of patient flow in passive fever surveillance.

The diagram shows how we reached the study population and the test results from collected samples, within the surveillance.

https://doi.org/10.1371/journal.pntd.0007882.g003

Clinical characteristics between dengue-positive and non-dengue cases

Table 1 describes demographic and clinical characteristics of dengue-positive vs. non-dengue cases. Of 2929 analyzed patients, 2189 (74.7%) were non-dengue and 740 (25.3%) were dengue-positive. Of the 740 dengue-positive patients, 540 (73.0%) were laboratory-confirmed and 200 (27.0%) were probable dengue. Of the dengue-positive cases, 42% (n = 317) were confirmed by RT-PCR and the remainder by paired ELISA (Fig 3). A small peak in dengue-positive cases was observed in October-December 2015. A much larger peak occurred in August-December 2016 (Fig 4). Both peaks occurred at the end or after the May-September rainy season. Of 777 fever cases from the outbreak, 55.1% (n = 428) were dengue-positive, with DENV2 predominating [181/258 (70%) of samples confirmed by RT-PCR] (Fig 4). Of 2152 non-outbreak fever cases, 14.5% (n = 312) were dengue-positive, mostly with DENV3 [28/43 (65%) of samples confirmed by RT-PCR] and a few DENV1 [5/43 (12%) of samples confirmed by RT-PCR].

thumbnail
Fig 4. Monthly distribution of febrile enrollees, dengue-positive and non-dengue cases & monthly distribution of dengue serotypes* in PCR-positive cases.

The figure has two parts: the upper part shows monthly distribution of dengue-positive and non-dengue cases among the enrolled patients; and the lower part shows distribution of serotypes identified (numbers shown in the bars) by month. *number of identified serotypes shown in the bars.

https://doi.org/10.1371/journal.pntd.0007882.g004

thumbnail
Table 1. Demographic and clinical characteristics of dengue-positive and non-dengue cases in the facility-based fever surveillance established in Ouagadougou, Burkina Faso, between December 2014 and February 2017.

https://doi.org/10.1371/journal.pntd.0007882.t001

Overall, dengue-positive cases were older than non-dengue cases (Table 1). Among dengue-positive cases, those after the 2016 outbreak were younger than those before or during the outbreak (about 75% <30 years old, compared to before and during the outbreak with about 50% <30 years) (Fig 5); the age difference before, during and after the outbreak was statistically significant (ANOVA, p-value < .001). Differences in terms of presenting signs and symptoms are presented in Table 2.

thumbnail
Fig 5. Age distribution of dengue-positive cases before, during, and after the 2016 outbreak.

The figure shows age distribution of dengue-positive cases, compared to non-dengue cases, before, during, and after the 2016 outbreak.

https://doi.org/10.1371/journal.pntd.0007882.g005

thumbnail
Table 2. Signs and symptoms of dengue-positive and non-dengue cases in the facility-based fever surveillance established in Ouagadougou, Burkina Faso, between December 2014 and February 2017.

https://doi.org/10.1371/journal.pntd.0007882.t002

There were 180 patients requiring observation at the CSPS. Patients later determined to be dengue-positive were more likely, on presentation, to require observation: 18% of dengue-positive cases versus 2% of non-dengue cases (Table 1). A small but significant difference was observed in average time between fever onset and enrollment for dengue-positive versus non-dengue cases (2.9 days vs. 2.6 days, p < .001). Likewise, the entire duration of fever illness on average was significantly longer for dengue-positive cases (mean 4.7 versus 4.0 days, among the 2926 patients with such data, p < .001). Dengue-positive cases were half as likely to self-report that they had been vaccinated for YF (17%, versus 38% for non-dengue cases, p < .001).

Of 2929 available RDT results, 11% (316/2929) and 4% (129/2929) were positive for NS1 and IgM, on the RDT kit, respectively (Fig 1). There were 38 patients [28 (74%) during the outbreak and 10 (26%) during the non-outbreak periods] with positive results for both NS1 and IgM on the RDT. During the outbreak period, 86% (271/316) were NS-1 positive and 40% (52/129) were IgM positive (28 showing positive on both NS1 and IgM).

Only 25% of dengue-positive cases were clinically diagnosed with suspected dengue, prior to lab-confirmation, and more than 90% of non-dengue cases were clinically diagnosed with undifferentiated fever. During the outbreak, 31.3% (131/428) of dengue-positive cases were diagnosed with suspected dengue, while 17.0% (53/312) were diagnosed with suspected dengue during non-outbreak periods.

Clinical features associated with dengue during and outside the 2016 outbreak

Over the outbreak period, of the 740 dengue-positive patients, 357 patients (48%) were laboratory-confirmed. Of these dengue-confirmed cases, 258 (72%) were confirmed by RT-PCR; 55 (15%) by both IgM and IgG seroconversion; and 44 (12%) by IgM or IgG seroconversion on paired ELISA. Over the non-outbreak period, of the 740 dengue-positive patients, 183 (25%) patients were laboratory-confirmed. Of these dengue-confirmed cases, 59 (32%) were confirmed by RT-PCR; 10 (5%) by both IgM and IgG seroconversion; and 114 (62%) by IgM or IgG seroconversion.

Demographic and clinical associations with dengue-positivity are shown in Table 3 for the outbreak and in Table 4 for non-outbreak periods. During the outbreak, independently associated symptoms were: rash, retro-orbital pain, cough, headache, nausea/vomiting, and loss of appetite. During non-outbreak periods, retro-orbital pain, headache, nausea/vomiting, and constipation were independently associated. In addition to the symptoms, the multivariable model selected requirement for observation and lack of YF vaccination to be associated with dengue-positivity in both outbreak and non-outbreak periods. Age in non-outbreak periods and, gender, elevated temperature at enrollment, and fever duration prior to enrollment in the outbreak period were also selected. Age and gender were a priori confounders and were significantly associated with dengue-positivity. Enrolled CSPS may be a proxy for otherwise any unexplained variation across centers, but was not selected for either of the outbreak or non-outbreak periods. In the absence of observation of variation with respect to dengue-positivity, it was not entered in the models.

thumbnail
Table 3. Univariate logistic analyses showing significant indicators and their odds ratios of dengue-positivity during the outbreak period, from the facility-based fever surveillance established in Ouagadougou, Burkina Faso, between December 2014 and February 2017.

https://doi.org/10.1371/journal.pntd.0007882.t003

thumbnail
Table 4. Univariate logistic analyses showing significant indicators and their odds ratios of dengue-positivity during non-outbreak periods, from the facility-based fever surveillance established in Ouagadougou, Burkina Faso, between December 2014 and February 2017.

https://doi.org/10.1371/journal.pntd.0007882.t004

Table 5 shows the final set of variables. During both outbreak and non-outbreak periods, dengue-positive patients had increased odds of presenting with rash [outbreak: 2.6 (95%CI = 1.5–4.6); non-outbreak: 1.5 (95%CI = 1.0–2.4)] and retro-orbital pain [outbreak: 7.4 (95%CI = 3.7–14.7); non-outbreak: 1.4 (95%CI = 1.01–1.8)].

thumbnail
Table 5. Multivariate logistic analysis showing significant indicators and their odds ratios of dengue-positivity by outbreak or non-outbreak periods, in the facility-based fever surveillance established in Ouagadougou, Burkina Faso, between December 2014 and February 2017.

https://doi.org/10.1371/journal.pntd.0007882.t005

Discussion

Recent reports of dengue outbreaks in Burkina Faso suggest substantial DENV transmission in this region. However, existing evidence on epidemiological characterization of dengue in Burkina Faso was limited in scope prior to this study. The current study collected population-based epidemiologic data in Ouagadougou during a 27-month period from 2014–2017, including all three months of the 2016 dengue outbreak. Our data demonstrated that dengue infection is an important cause of febrile illnesses, accounting for one-quarter of non-malarial febrile illness in patients seeking care at CSPSs in the study. This proportion was very high (55%) during the outbreak itself, but even outside the outbreak, a considerable proportion (15%) of non-malarial febrile episodes was dengue-positive. Since then, Ouagadougou has experienced another, larger, dengue outbreak in 2017 [16, 18]. Recent outbreaks and the current study indicate that DENV transmission is likely to be underestimated and underdiagnosed in Burkina Faso [16, 18].

Nonetheless, Burkina Faso is one of the countries in West Africa with better defined dengue virus transmission and burden. Several other countries with some existing data are Nigeria, Senegal, Ghana, and Sierra Leone. In Nigeria, presence of antibodies to DENV 2 was documented in 45% of 1816 human samples [30]. In Senegal, there were reported outbreaks of DENV 3 in 2009 with 196 individuals affected and 5 cases of dengue haemorragic fever (DHF) [31]. In Ghana, 3.2% among 218 children were found to show dengue IgM in 2014 [32]. In Sierra Leone, presence of antibody to all four serotypes of dengue virus was documented, based on neutralization test results on the samples from patients with fever of unknown origin [33]. While these reports suggest dengue presence in several West African countries, some with even high rates of infections, there is continued lack of data on dengue epidemiology in the region and highlighted need for improved surveillance system.

Differences between outbreak and non-outbreak periods

The predominant DENV serotype identified from RT-PCR-positive outbreak cases in the study was DENV2 (Fig 4). This was consistent to the results of MoH/WHO investigation of the 2016 outbreak where DENV2 was the predominant serotype [16, 17]. DENV2 was also the dominant serotype detected in outbreaks in Burkina Faso in 1982 and 1983–1986 [9, 34]. The study found DENV3 to be predominant during the non-outbreak period preceding the 2016 outbreak. DENV3 was the dominant serotype in the 2013 outbreak in Burkina Faso [35]. A change in predominant DENV serotype may have fueled the outbreak in 2016. Although the current study did not determine DENV strain, DENV2 strains reported from ill French travelers returning from Burkina Faso in November 2016 were nearly identical to a DENV2 strain detected in Burkina Faso in 1983 [36]. This suggests that the 2016 outbreak may have been due to an endemic strain of DENV2 circulating in Burkina Faso for 30 years, perhaps maintained partly through a sylvatic cycle [36]. More detailed phylogenetic analysis of DENVs from the current study is planned.

Only a quarter of dengue-positive cases received a clinical diagnosis of suspected dengue in this study, with this proportion being only slightly higher during the 2016 outbreak (31% of dengue cases were suspected clinically) compared to outside the outbreak (17%). In the routine care system, clinicians in the CSPS refer to a guideline issued by the Burkina Faso MoH [37], primarily based on the 2009 WHO dengue guidelines. The dengue RDTs were made available at the CSPSs in the study, but the results of dengue RDT might not have contributed to the clinical assessment, if the results were not made available in time (dependent on patient volume and clinician availability). Dengue RDTs are typically unavailable for routine use in Africa; and many non-malaria febrile etiologies, including dengue, are likely to be under-diagnosed [12, 38]. Clinicians in Burkina Faso may need to consider dengue more frequently as a clinical diagnosis, with or without point-of-care assays.

Our multivariable analysis showed differing patterns of signs and symptoms associated with dengue-positivity during the outbreak period compared to non-outbreak periods. Rash was associated with dengue-positivity during both outbreak and non-outbreak periods. Rash is a common sign for dengue and listed in dengue classification in both 1997 and 2009 WHO dengue guidelines [3, 39, 40]. However, retro-orbital pain showed increased odds of dengue-positivity only during the outbreak. Retro-orbital pain, also listed in the 2009 WHO case definition, is another common sign associated with dengue-positivity [3, 39, 40]. Also, it was suggested that ocular symptoms, including retro-orbital pain, in dengue patients may possibly indicate thrombocytopenic state with increased likelihood of hemorrhage [41]. In our data, dengue-positive patients with retro-orbital pain were 5.8 times (95% C.I: 3.5–9.6, p < .001) more likely to require observation than dengue-positive patients without retro-orbital pain during the outbreak. During non-outbreak, it also showed a similar pattern with statistical significance, but with a wide confidence interval. Therefore, further information is needed for validation. While hemorrhagic signs were not commonly reported in our data, requiring observation may indicate severity of dengue illness and retro-orbital pain being associated with dengue-positive cases in the outbreak may indicate likely severity of dengue illness during the outbreak.

Our data showed a high proportion of individuals 15–40 years of age among dengue-positive cases in the outbreak period (a mean age of 26.8 years in dengue-positive patients). This was also found in the outbreak investigation by the Burkina Faso MoH with WHO where 70% of affected people were 25 years and older, with a mean age of 30 years [16]. It suggests that those in the labor force may be impacted, leading to significant economic and social burden [42]. Adjusted for age and gender, our model found higher odds that dengue-positive cases required observation, compared to non-dengue, during both outbreak (6.0 times) and non-outbreak (4.3 times) periods. Given the substantial proportion of dengue-positive cases among non-malarial febrile illnesses, this suggests that dengue may account for greater utilization of healthcare resources in CSPSs than other etiologies, during both outbreak and non-outbreak periods. As in many other parts of Africa, these primary healthcare centers have limited resources, such as beds [43], and could be especially overextended during outbreaks. Since the study only enrolled patients at CSPSs, the burden on the healthcare system due to dengue inpatients is unclear.

Self-report of not having received YF vaccination was associated with increased odds of dengue-positivity. A priori, one might have hypothesized cross-protection. However, the opposite phenomenon of a predisposition of YF vaccinated individuals to DHF has been suggested, with a possible explanation of cross-reactivity between antibodies from YF vaccination and dengue virus. [44]. Without much data on association of YF vaccination and dengue infection, self-reporting may be unreliable due to recall bias, and the study could not confirm YF vaccination using patient records.

Study limitations

DENV transmission can vary substantially over time and space. Hence, the generalizability of the current study is limited by enrollment from the five selected CSPSs in the capital during the 27-month study period. We would have missed those community residents with relevant symptoms seeking care elsewhere than study centers, including private providers. In addition, patients with severe illness would have not been enrolled since they would likely have sought care directly at inpatient facilities; and subclinical and mild DENV infections would also not have been detected.

The study surveillance excluded patients with malaria RDT positive results, localizing signs or known/confirmed diagnosis with other diseases, possibly omitting co-infections of dengue with another pathogen. In particular, given the prevalence of malaria in this region, dengue and malaria co-infection were not included in this study and may require further investigation. Nevertheless, the available information on co-infections suggests they are uncommon [9, 4548].

Performance of malaria RDTs, in terms of sensitivity, would depend on local conditions, especially the level of malaria transmission shown to be variable from reported incidence in Ouagadougou [49, 50]. There could have been misclassification among non-malarial patients (i.e. false negative results on malaria RDT included in the study being differently classified between dengue-positive and non-dengue groups). Also, this could vary by the level of dengue transmission (i.e. during and outside of the outbreak), leading to differential misclassification.

Our findings were based on outpatients and patients requiring observation, and clinical characteristics may be different for hospitalized patients and individuals with subclinical infections. Also, such findings may depend on other co-circulating pathogens endemic in the area, however our study did not confirm etiologies of non-dengue cases. Therefore, further information on the etiologies of non-dengue febrile cases may be needed to verify which signs are useful in distinguish non-dengue from dengue illnesses [51].

In our analysis, laboratory-confirmed and probable dengue cases were combined into the dengue-positive group. There may be some limitations with probable dengue being not as certain as lab-confirmed dengue. However, we performed analysis using 3 categories of dengue infection status (lab-confirmed-; probable-; and non-dengue) as part of sensitivity analysis and this yielded similar results (see S2S4 Tables).

Conclusion

Dengue is an important cause of non-malarial fever in Burkina Faso, both during and outside of outbreaks, despite being infrequently suspected by clinicians. Despite the many possible etiologies of febrile illness in this region, limited surveillance and diagnostic capacity will continue to pose challenges to dengue prevention and control. Additional longitudinal studies to better characterize dengue epidemiology and clinical presentation, including in inpatients and for subclinical/mild cases, along with encouraged use of dengue RDTs, would help to inform strategies to approach dengue countermeasures in this region.

Supporting information

S2 Table. Demographic and clinical characteristics of patients by dengue infection status from the health facility-based fever surveillance established in Ouagadougou, Burkina Faso.

https://doi.org/10.1371/journal.pntd.0007882.s002

(DOCX)

S3 Table. Univariate logistic regression analyses showing significant indicators and their odds ratios between dengue-confirmed and non-dengue cases during the period of outbreak in the health facility-based fever surveillance.

https://doi.org/10.1371/journal.pntd.0007882.s003

(DOCX)

S4 Table. Univariate logistic regression analyses showing significant indicators and their odds ratios between dengue-confirmed and non-dengue cases during the period of non-outbreak in the health facility-based fever surveillance.

https://doi.org/10.1371/journal.pntd.0007882.s004

(DOCX)

Acknowledgments

The authors thank the patients, doctors and staff of CSPSs of Ouagadougou. Also, we thank the laboratory staff of Centre Muraz. The authors would also like to thank the staff of Action-Gouvernance-Integration-Renforcement (AGIR), the health authorities of Ouagadougou and Burkina Faso, and particularly the Ministry of Health, for their support in the study execution. Lastly, we would like to thank the Dengue Vaccine Initiative (DVI) team and statisticians and administrative staff at the International Vaccine Institute for their helpful comments during the preparation of this manuscript and support during the studies.

References

  1. 1. Gubler DJ, Clark GG. Dengue/dengue hemorrhagic fever: the emergence of a global health problem. Emerg Infect Dis. 1995;1(2):55–7. PubMed Central PMCID: PMC2626838. pmid:8903160
  2. 2. World Health Organization. Global strategy for dengue prevention and control 2012–2020. Geneva: World Health Organization; 2012. p. vi, 43p.
  3. 3. World Health Organization. Dengue and severe dengue Geneva: World Health Organization; 2019 [updated 2 February 2018 cited 2019 12 May]. Available from: http://www.who.int/news-room/fact-sheets/detail/dengue-and-severe-dengue.
  4. 4. Messina J, Brady O, Scott T, Zou C, Pigott D, Duda K, et al. Global spread of dengue virus types: mapping the 70 year history. Trends Microbiol. 2014;22(3):138–46. pmid:24468533
  5. 5. Amarasinghe A, Kuritsky J, Letson G, Margolis H. Dengue virus infection in Africa. Emerging Infectious Diseases. 2011;17(8):1349–54. Epub August 2011. PubMed Central PMCID: PMC3381573. pmid:21801609
  6. 6. Kraemer MU, Sinka ME, Duda K, Mylne AQ, Shearer FM, Barker CM, et al. The global distribution of the arbovirus vectors Aedes aegypti and Ae. albopictus. eLife. 2015;4:e08347. pmid:26126267
  7. 7. Baba M, Villinger J, Masiga DK. Repetitive dengue outbreaks in East Africa: A proposed phased mitigation approach may reduce its impact Reviews in Medical Virology. 2016;26(3):183–96. |https://doi.org/10.1002/rmv.1877.
  8. 8. Amoako N, Duodu S, Dennis FE, Bonney JHK, Asante KP, Ameh J, et al. Detection of dengue virus among children with suspected malaria, Accra, Ghana. Emerg Infect Dis 2018;24(8):1544–7. PubMed Central PMCID: PMCID: PMC6056106 pmid:30015610.
  9. 9. Ridde V, Agier I, Bonnet E, Carabali M, Dabiré K, Fournet F, et al. Presence of three dengue serotypes in Ouagadougou (Burkina Faso): research and public health implications. Infect Dis Poverty. 2016;5(5):23. PubMed Central PMCID: PMCpmid:27044528 PMCID: PMC4820922
  10. 10. Were F. The dengue situation in Africa. Paediatr Int Child Health. 2012;32 Suppl 1:18–21. Epub 2012/06/08. pmid:22668445; PubMed Central PMCID: PMC3381440.
  11. 11. Ridde V, Carabali M, Ly A, Druetz T, Kouanda S, Bonnet E, et al. The need for more research and public health interventions on dengue fever in Burkina Faso. PLoS Neglected Tropical Diseases. 2014;8(6):e2859. pmid:24945324
  12. 12. Brah S, Daou M, Salissou L, Mahaman S, Alhousseini D, Iroungou B A, et al. Fever of unknown origin in Africa: The causes are often determined! Health sciences and diseases. 2015;16(2).
  13. 13. Gonzalez J, Du Saussay C, Gautun J, McCormick J, Mouchet J. Dengue in Burkina Faso (ex-upper Volta): seasonal epidemics in the urban area of Ouagadougou. Bulletin de la Societe de Pathologie Exotique. 1985;78(1):7–14. PubMed Central PMCID: PMC3886182.
  14. 14. Robert V, Lhuillier M, Meunier D, Sarthou J, Monteny N, Digoutte J-P, et al. Yellow fever virus, dengue 2 and other arboviruses isolated from mosquitos, in Burkina Faso, from 1983 to 1986. Entomological and epidemiological considerations. Bulletin de la Societe de Pathologie Exotique. 1993;86(2):90–100. PubMed Central PMCID: PMC8102567. pmid:8102567
  15. 15. Ministère de la Santé. Rapport d'étape de l'investigation de cas suspects de Dengue dans la région sanitaire du Centre. Ouagadougou, Burkina Faso: Direction de la lutte contre la maladie, 2013.
  16. 16. World Health Organization. Dengue fever—Burkina Faso. Geneva, Switzerland 2016 [updated 2016 November 18; cited 2018 August 18]. Available from: http://www.who.int/csr/don/18-november-2016-dengue-burkina-faso/en/.
  17. 17. Tarnagda Z, Cissé A, Bicaba B, Diagbouga S, Sagna T, Ilboudo A, et al. Dengue fever in Burkina Faso, 2016. Emerg Infect Dis 2018;24(1):170–2. PubMed Central PMCID: PMC PMCID: PMC5749475 pmid:29260685
  18. 18. World Health Organization. Dengue fever—Burkina Faso. Geneva, Switzerland 2017 [updated 2017 November 6; cited 2018 August 18]. Available from: http://www.who.int/csr/don/6-november-2017-dengue-burkina-faso/en/.
  19. 19. Beatty M, Stone A, Fitzsimons D, Hanna J, Lam S, Vong S, et al. Best practices in dengue surveillance: a report from the Asia-Pacific and Americas Dengue Prevention Boards. PLoS Negl Trop Dis 2010;4(11):e890. pmid:21103381
  20. 20. Lim J, Carabali M, Lee J-S, et al. Evaluating dengue burden in Africa in passive fever surveillance and seroprevalence studies: protocol of field studies of the Dengue Vaccine Initiative. BMJ Open. 2018;2018(8):e017673. pmid:29358421
  21. 21. Brady OJ, Gething PW, Bhatt S, Messina JP, Brownstein JS, Hoen AG, et al. Refining the global spatial limits of dengue virus transmission by evidence-based consensus. PLoS Negl Trop Dis. 2012;6(8):e1760. pmid:22880140; PubMed Central PMCID: PMC3413714.
  22. 22. Secretariat General. Annuaire statistique 2016. In: sectorielles. Dgdéeds , editor. 03 BP 7009 Ouagadougou 03.: Ministere de la Sante, Burkina Faso; 2017.
  23. 23. Beogo I, Liu C-Y, Chou Y-J, Chen C-Y, Huang N. Health-care-seeking patterns in the emerging private sector in Burkina Faso: A population-based study of urban adult residents in Ouagadougou. PLoS ONE. 2014;9(5):e97521. pmid:24842536
  24. 24. Rossier C, Soura A, Baya B, Compaoré G, Dabiré B, Dos Santos S, et al. Profile: The Ouagadougou Health and Demographic Surveillance System. International Journal of Epidemiology. 2012;41(3):658–66. pmid:22685112
  25. 25. Alm E, Lindegren G, Falk KI, Lagerqvist N. One-step real-time RT-PCR assays for serotyping dengue virus in clinical samples. BMC Infectious Diseases. 2015;15(493). https://doi.org/10.1186/s12879-015-1226-z.
  26. 26. World Health Organization. Handbook for clinical management of dengue. World Health Organization, Department of Control of Neglected Tropical Diseases (WHO/NTD), the Special Programme for Research and Training in Tropical Diseases (WHO/TDR), editors. Geneva, Switzerland: World Health Organization,; 2012.
  27. 27. Bewick V, Cheek L, Ball J. Statistics review 8: Qualitative data–tests of association. Crit Care 2004;8(1):46–53. Epub 2003 Dec 30. PubMed Central PMCID: PMCID: PMC420070 pmid:14975045.
  28. 28. Anders K. Resolution of Students t-tests, ANOVA and analysis of variance components from intermediary data. Biochem Med (Zagreb) 2017;27(2):253–8. Epub 2017 Jun 15. PubMed Central PMCID: PMCID: PMC5493167 pmid:28740445.
  29. 29. Ankera M, Arimab Y. Male-female differences in the number of reported incident dengue fever cases in six Asian countries. Western Pacific Surveillance and Response Journal. 2011;2(2):17–23. pmid:23908884
  30. 30. Fagbami A, Monath T, Fabiyi A. Dengue virus infections in Nigeria: a survey for antibodies in monkeys and humans. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1977;71(1):60–5. pmid:404737
  31. 31. Saluzzo J, Cornet M, Adam C, Eyraud M, Digoutte JP. [Dengue 2 in eastern Senegal: serologic survey in simian and human populations. 1974–85]. Bull Soc Pathol Exot Filiales. 1986;79(3):313–22. pmid:3769119
  32. 32. Stoler J, Delimini R, Bonney J, Oduro A, Owusu-Agyei S, Fobil J, et al. Evidence of recent dengue exposure among malaria parasite-positive children in three urban centers in Ghana. Am J Trop Med Hyg. 2015;92(3):497–500. pmid:25582693
  33. 33. de Araujo LJ, Mores C, Bausch D, Christofferson R. Short report: Serological evidence of under-reported dengue circulation in Sierra Leone. PLoS Negl Trop Dis. 2016;10(4).
  34. 34. Sang RC. Dengue in Africa Arbovirology/Viral Haemorrhagic Fever Laboratory, Centre for Virus Research: Kenya Medical Research Institute, PO Box 54628, Nairobi, Kenya; [cited 2019 January 21]. Available from: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.599.6007&rep=rep1&type=pdf.
  35. 35. Tarnagda Z, Congo M, Sagna T, Ouédraogo C, Nikiéma V, Cissé A, et al. Outbreak of dengue fever in Ouagadougou, Burkina Faso, 2013. International Journal of Microbiology and Immunology Research. 2014;2014(2):101–8.
  36. 36. Baronti C, Piorkowski G, Touret F, Charrel R, de Lamballerie X, Nougairede A. Complete coding sequences of two dengue virus type 2 strains isolated from an outbreak in Burkina Faso in 2016. Viruses. 2017. pmid:28450505
  37. 37. MoH of Burkina Faso. Directive national de prise en charge des cas de dengue au Burkina Faso. In: MoH, editor. Burkina Faso2014.
  38. 38. Zongo S, Carabali M, Munoz M, Ridde V. Dengue rapid diagnostic tests: Health professionals' practices and challenges in Burkina Faso. SAGE Open Medicine. 2018;6. Epub August 18, 2018 https://doi.org/10.1177/2050312118794589
  39. 39. World Health Organization. Dengue guidelines for diagnosis, treatment, prevention, and control. Geneva: World Health Organization, 2009.
  40. 40. Hadinegoro SRS. The revised WHO dengue case classification: does the system need to be modified? Paediatr Int Child Health 2012;32(S1):33–8. PubMed Central PMCID: PMCID: PMC3381438 pmid:22668448.
  41. 41. Chan DPL, Teoh SCB, Tan CSH, Nah GKM, Rajagopalan R, Prabhakaragupta MK, et al. Ophthalmic complications of dengue. Emerg Infect Dis 2006;12(2):285–9. PubMed Central PMCID: PMCID: PMC3373088 pmid:16494756.
  42. 42. Lee J, Mogasale V, Lim JK, Ly S, Lee K, Sorn S, et al. A multi-country study of the economic burden of dengue fever based on patient-specific field surveys in Burkina Faso, Kenya, and Cambodia. PLoS Negl Trop Dis 2019;13(2):e0007164. Epub 2019 Feb. PubMed Central PMCID: PMC PMCID: PMC6394908. pmid:30817776
  43. 43. Hospital bed density—World: indexmundi; 2018 [updated January 1, 2018; cited 2018 October 21]. The map displayed shows how Hospital bed density varies by country]. Available from: https://www.indexmundi.com/map/?t=0&v=2227&r=xx&l=en.
  44. 44. Guzman JR, Kron MA. Threat of dengue haemorrhagic fever after yellow fever vaccination. Lancet. 1997;349(9068):P1841. https://doi.org/10.1016/S0140-6736(05)61727-8.
  45. 45. Wiwanitkit V. Concurrent malaria and dengue infection: a brief summary and comment. Asian Pac J Trop Biomed 2011;1(4):326–7. PubMed Central PMCID: PMCID: PMC3614227 pmid:23569785.
  46. 46. Epelboin L, Hanf M, Dussart P, Ouar-Epelboin S, Djossou F, Nacher M, et al. Is dengue and malaria co-infection more severe than single infections? A retrospective matched-pair study in French Guiana. Malar J. 2012;11(142). Epub 2012 May 1. PubMed Central PMCID: PMCID: PMC3403992 pmid:22549018.
  47. 47. Magalhães BML, Siqueira AM, Alexandre MAA, Souza MS, Gimaque JB, Bastos MS, et al. P. vivax malaria and dengue fever co-infection: A cross-sectional study in the Brazilian Amazon. PLoS Negl Trop Dis. 2014. https://doi.org/10.1371/journal.pntd.0003239
  48. 48. Carme B, Matheus S, Donutil G, Raulin O, Nacher M, Morvan J. Concurrent dengue and malaria in Cayenne hospital, French Guiana. Emerg Infect Dis 2009;15(4):668–71. PubMed Central PMCID: PMCID: PMC2671414 pmid:19331769.
  49. 49. World Health Organization. Malaria rapid diagnostic test performance summary results of WHO product testing of malaria RDTs: round 1–8 (2008–2018). Geneva: 2018 Contract No.: Licence: CC BY-NC-SA 3.0 IGO.
  50. 50. Ouedraogo B, Inoue Y, Kambiré A, Sallah K, Dieng S, Tine R, et al. Spatio-temporal dynamic of malaria in Ouagadougou, Burkina Faso, 2011–2015. Malaria Journal. 2018;17(138). https://doi.org/10.1186/s12936-018-2280-y.
  51. 51. Yoon I-K, Srikiatkhachorn A, Hermann L, Buddhari D, Scott TW, Jarman RG, et al. Characteristics of mild dengue virus infection in Thai children. Am J Trop Med Hyg. 2013;89(6):1081–7. pmid:24127167