Urban Chikungunya in the Middle East and North Africa: A systematic review

Background The epidemiology of Chikungunya virus (CHIKV) in the Middle East and North Africa (MENA) is not well characterized despite increasing recognition of its expanding infection and disease burden in recent years. Methodology / Principal findings Following Cochrane Collaboration guidelines and reporting our findings following PRISMA guidelines, we systematically reviewed records describing the human prevalence and incidence, CHIKV prevalence/infection rates in vectors, outbreaks, and reported cases for CHIKV across the MENA region. We identified 29 human seroprevalence measures, one human incidence study, one study reporting CHIKV infection rates in Aedes, and nine outbreaks and case reports/series reported in the MENA from 1970–2015. Overall, anti-CHIKV antibody or reports of autochthonous transmission were identified from 10 of 23 countries in the MENA region (Djibouti, Egypt, Iraq, Iran, Kuwait, Pakistan, Saudi Arabia, Somalia, Sudan, and Yemen), with seroprevalence measures among general populations (median 1.0%, range 0–43%) and acute febrile illness populations (median 9.8%, range 0–30%). Sudan reported the highest number of studies (n = 11) and the highest seroprevalence among general populations (median 12%, range 0–43%) and undifferentiated acute febrile illness populations (median 18%, range 10–23%). CHIKV outbreaks were reported from Djibouti, Pakistan, Sudan, and Yemen. Conclusions / Significance Seroprevalence studies and outbreak reports suggest endemic transmission of urban cycle CHIKV in at least the Red Sea region and Pakistan. However, indications of seroprevalence despite a low quantity of CHIKV epidemiologic research from the region suggests that CHIKV transmission is currently underrecognized.


Introduction
Chikungunya virus (CHIKV) is an alphavirus whose recognized global distribution increasingly overlaps that of dengue and the distribution of their shared mosquito vectors, Aedes aegypti and Aedes albopictus [1].Clinical reports suggest that CHIKV may have been broadly distributed by the 1800s, though its similar clinical presentation to Dengue virus (DENV) makes its historic epidemiology uncertain [2].Since the first isolation of CHIKV in Tanzania in 1952-53 [3], large urban outbreaks have been detected in the Indian Ocean region and Latin America, yielding millions of suspected infections and the novel discovery of autochthonous CHIKV transmission in Mediterranean Europe in 2007 [4][5][6][7][8][9].The past decade has witnessed novel reports of CHIKV outbreaks in the Middle East and North Africa (MENA) as well.However, the present and historic epidemiology of CHIKV in this region remains poorly characterized.
Over the years, limited surveillance and diagnostic capacity have likely hindered the recognition of CHIKV in the MENA region.Although the earliest possible clinical description of CHIKV infection known to exist was first recorded in Cairo, Egypt in 1658 [2], it was not until 2011 that the presence of CHIKV was first confirmed (i.e. by viral culture or molecular detection) in the MENA during an outbreak in Yemen with over 15,000 suspected cases [10].The CHIKV lineage responsible for this epidemic was related to the enzoonotic East-Central-South African (ECSA) Indian Ocean Lineage and was isolated from A. aegypti, a finding that suggested urban cycle transmission [11,12].To date, the existence of a sylvatic transmission cycle has not been reported in the MENA region.Although debilitating and prolonged polyarthralgias can be a recognizable hallmark of CHIKV infection [1], the clinical syndrome can be difficult to distinguish from other mosquito-transmitted febrile illnesses including dengue fever [2,10,13], o'nyong'nyong fever [2], yellow fever [14], and malaria [14].Given the increasing global impact of Aedes-transmitted arboviruses and the limited knowledge of urban CHIKV in the MENA region, we performed a systematic review of the literature to describe the published evidence pertaining to the epidemiology of CHIKV in the MENA region.

Objectives
The objective of this study was to characterize the epidemiology of urban CHIKV in the MENA region through a systematic review of published human prevalence and incidence studies, human outbreaks and reported cases, and studies reporting CHIKV detections and prevalence/infection rates in Aedes mosquitoes.The original literature search was conducted in December 2015 and updated in May 2017.

Materials and methods
The materials and methods used for this review are similar to those of a systematic review of DENV in the MENA region that we conducted in parallel to the current study [15].

Eligibility criteria
The eligibility criteria for this study follows similar criteria that was used for a review of DENV in the MENA region (Table 1) [15].In brief, reports containing primary human seroprevalence or incidence, outbreaks and reported cases, and CHIKV detections from Aedes mosquitoes in the MENA region published in any year were considered eligible for the systematic review.For incidence studies, those that reported the number of acute infections or seroconversions over any time interval, or overall attack rate if assessed during an outbreak, were eligible.Human CHIKV outbreaks, case series, and case reports in natives and returned travelers from the MENA region were also sought from the articles retrieved through the search databases using the original search criteria.Outbreak reports were included if at least some of the reported cases were laboratory-supported CHIKV infection; cases series and case reports were only included if they were laboratory-supported.We considered any report of CHIKV cases to constitute an outbreak if the author of the report qualified it as such.As with DENV, there is currently no consensus on how to define CHIKV outbreaks, so determining whether any number of cases represents a significant deviation from baseline transmission is often unclear [15].Finally, studies containing CHIKV prevalence in Aedes pools and single CHIKV isolates or vRNA detections from Aedes were included if they contained a measure of the estimated proportion of CHIKV-infected Ae. aegypti or Ae.albopictus at a given time and setting in the MENA region.Prevalence studies in animals were noted but excluded from the systematic review, as our study focused on urban cycle CHIKV mediated by human-mosquito transmission.Our review covered the 23 countries included in the MENA definitions of the WHO/ EMRO, World Bank, and the Joint United Nations Programme on HIV/AIDS (UNAIDS) for consistency with our systematic review of DENV in the MENA region as well as our earlier regional analyses of various infectious diseases such as HIV and other sexually transmitted infections and Hepatitis C virus [16][17][18][19].

Outcomes
For the systematic review, the primary outcomes were CHIKV human seroprevalence, human CHIKV incidence, human case reports/case series of CHIKV infection in MENA natives and returned travelers from the MENA region, CHIKV prevalence in Aedes, and single virus isolates of CHIKV in Aedes in the MENA region.

Data sources and search strategy
We conducted a systematic search for CHIKV in the MENA informed by the Cochrane Collaboration guidelines [20] and reported our findings using the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [21].The PRISMA checklist is found in S1 Table and our search criteria in S1 Fig.We searched PubMed (indexed since 1966 and selectively since 1865) and Embase (indexed since 1988) using text and MeSH/Emtree terms exploded to include all subheadings, as well as the World Health Organization (WHO) Index Medicus for the Eastern Mediterranean Region, WHO African Index Medicus (both indexed since 1984), and ProMED-MENA (indexed since 1994) using only the search term "chikungunya".

Study selection
The methodology for this section also follows our previous review of DENV [15].Titles and abstracts were imported into Endnote (Thompson Reuters, Philadelphia, PA, USA) and were screened by one author (JH) with potential eligibility determined by consensus with a second author (NC) when eligibility was unclear.Full texts of potentially relevant records were retrieved and assessed for eligibility, contacting the author of the report as necessary.Reference lists of all potentially eligible articles and reviews were also searched.For this study, 'report' refers to the document (paper, abstract, or public health record) containing an outcome measure of interest, while 'study' refers to the outcome measure(s) within that report.Hence, reports could contribute more than one study, though multiple reports of the same study were counted only once.

Data extraction and synthesis
Data were extracted by one of the authors (JH).Data from reports in English were extracted from the full texts, while reports in French (n = 1) and German (n = 1) were extracted from the English abstract and with the help of online language software and French and German language speakers [22].There were no records in other languages.Studies were compiled by country and organized by year.Prevalence studies were stratified as follows: 1) general prevalence studies assessing anti-CHIKV IgG prevalence (e.g.CHIKV exposure) among individuals not suspected to have acute CHIKV infection, including community members, blood donors, military, students, and hospitalized patients and outpatients receiving care for non-febrile illnesses, and 2) acute febrile illness studies assessing the prevalence of laboratory-supported CHIKV infection (i.e.any positive laboratory test suggesting CHIKV infection) in those with undifferentiated acute febrile illness (AFI) or suspected arbovirus infection (IgG prevalence measures obtained during the acute phase of illness in these studies are presumed to reflect an earlier infection).These stratifications were made because of the different study aims and pretest probabilities of having laboratory evidence of CHIKV infection in these populations.Reports were summarized by country and year, along with the sampling method, assay type/ make, sample size, and any data pertaining to alphavirus cross-reactions that was available.
For CHIKV outbreaks, we recorded the year, location, suspected or confirmed vector, and number of cases as provided in each report.We recorded similar information for case reports/ series and cases in returned travelelers as available.Finally, reports containing CHIKV prevalence, infection rates, and single virus isolations in Ae. aegypti and Ae.albopictus, both recognized urban cycle CHIKV vectors, were also sought using the original search criteria.We recorded the year, location, and methods of collection and laboratory methods, as well as additional bioecological aspects of the vectors as available.The geographic distribution of all included human prevalence studies, outbreak reports, and case reports/series were mapped according to the first-level administrative division (e.g.state, province) in which each event was recorded (Tableau Software, Seattle, WA, USA).

Risk of bias assessment
The risk of bias (ROB) was assessed for each seroprevalence study based on the Cochrane approach [20] and by evaluating the precision of the reported measures according to previously developed methodology [23,24].Each CHIKV seroprevalence measure was considered to have a low, high, or unclear ROB in two domains: sampling methodology and response rate.The latter was defined as the number of tested individuals divided by the number of persons invited to participate in the study [25].ROB was considered low if (1) sampling was probability-based (using some form of random selection), and (2) response rate was !80%.Studies with missing information for either of the domains were classified as having unclear ROB for that specific domain.We did not assess the ROB for the sampling methodology of populations with acute febrile illness, as these are defined populations presenting to a health facility with acute infection and no population-based sampling is needed to capture these populations.We also did not include the laboratory assay characteristics in our risk of bias assessment given that the epidemiology of antigenically similar alphaviruses (e.

g. O'nyong-nyong virus [ONNV],
Sindbis virus [SINV], and Semliki Forest virus [SFV]) in the MENA was also unclear, and nearly all identified studies utilized in-house assays.Studies were considered to have high precision if the number of individuals tested was !100.We considered this to be a reasonably sensitive cutoff for precision given the heterogeneous epidemiology of CHIKV across the region (e.g. a prevalence of 1% has a 95% CI of 0-3%).

Search results
The selection process based on PRISMA guidelines is illustrated in Fig 1 [21].Briefly, the search yielded 262 reports, 35 of which were eligible for inclusion in the study following screening process and after the addition of 8 reports identified from reference lists and reviews.
One animal seroprevalence study was excluded, in which anti-CHIKV antibody prevalence of 2.5% was reported in a sample of 157 rodents captured in Pakistan.In this study, all antibodies were cross-reactive with SINV [26].

Characteristics of included studies
A total of 29 human seroprevalence studies for CHIKV were identified from eligible reports, 62% of which were conducted prior to 1990 (Table 2).Fig 2 illustrates the geographic distribution of all published CHIKV prevalence studies, outbreaks, and reported cases in the MENA region.Overall, anti-CHIKV antibodies were reported from 10 of 23 countries in the MENA: Djibouti, Egypt, Iran, Iraq, Kuwait, Pakistan, Saudi Arabia, Somalia, Sudan, and Yemen.The   median seroprevalence measures among general populations was 1.0% (range 0-43%), and 9.8% (range 0-30%) among populations with acute febrile illness.Sudan reported both the highest number of studies (n = 11) and the highest seroprevalence measures overall (median 12.9%, range 0-43%).Ninety-three percent of all studies utilized in-house assays; 82% of all studies that were conducted after 1990 (n = 11) utilized ELISA and/or RT-PCR (or RT-qPCR); 56% (10/18) of studies prior to 1990 utilized hemagglutinin inhibition (HI) assays.Serologic cross-reactions with related alphaviruses (e.g.ONNV, SINV, and SFV) or travel-acquired infections were observed could not be excluded in some studies from Djibouti [27], Iran [28], Kuwait [29,30], and Pakistan [26].Viral neutralization testing was performed in a total of five studies from Djibouti [31], Iraq [32,33], and Sudan [32][33][34].Among acute febrile illness studies, the prevalence of laboratory-supported CHIKV infection was also highest in Sudan, with anti-CHIKV antibody prevalence by HI or ELISA ranging from 1.8% to 30%.In Yemen, 0% of 222 cases of undifferentiated AFI were positive for CHIKV in the eastern coastal city of Al-Mukalla in June 2010 [35].Four months later, however, an outbreak of CHIKV was detected in the western coastal governorate of Al-Hudaydah, Yemen, and two studies demonstrated 9.8% and 28% ELISA IgM seroprevalence among subjects with undifferentiated AFI [10,36].Convalescent sera results were reported in two AFI studies, neither of which was positive as indicated by a four-fold rise in antibody titers [37,38].One study reporting detection of CHIKV in mosquito pools was identified in our search, in which CHIKV was detected by RT-qPCR in 26.6% of 11 pools of 30 Ae. aegypti mosquitoes that were collected at an Eritrean refugee camp in Al Hudaydah, Yemen during the aforementioned CHIKV outbreak [39].In this study, mosquitoes were collected by BG-sentinel TM traps, Knock-down pyrethroid spray, and indoor aspirations from houses of recently reported cases, and larvae were inspected in all possible containers per house or inhabited premises.The total container index of the sampled sites was 53.8, Breteau index was 100, and house index was 57.Ae. aegypti adult female minimum and maximum infection rates were 20% and 72%, respectively.There were no published reports of single CHIKV isolations from mosquitoes or population-based human incidence identified in our search.However, during the 2012 CHIKV outbreak in southern Yemen, one study reported an overall CHIKV attack rate of 7.5 per 1,000 people, ranging from 5.3 among children 0-4 years of age to 12.2 among adults !45 years [40].

Precision and risk of bias assessment results
The quality assessment for each prevalence study is found in Table 3. Overall, 66% (19/29) of studies contained high precision as defined by a sample size of !100 subjects.For the risk of bias assessment, response rates were !80% in 14% (4/29) of studies, <80% in 3% (1/29) of studies, and not reported in 83% (24/29) of studies.Among general seroprevalence studies, some form of probability sampling (i.e.low ROB) was reported in 31% (5/16) of studies, nonprobability sampling (e.g.high ROB) in 31% (5/16) of studies, and unclear sampling methods in 38% (6/16) of studies.
A total of four CHIKV outbreaks, three case reports/series, and one report of CHIKV in returned travelers, were identified through the search databases (Table 4) and were mapped along with the geographic distribution of prevalence studies (Fig 2).Outbreaks were reported from Djibouti, Sudan, Pakistan, and Yemen.In most cases, the vector was suspected to be Ae.aegypti given its known occurrence in the affected countries, but was only confirmed as such in the 2011-2012 outbreak in Yemen [10].

Discussion
Our review examines the epidemiology of urban CHIKV in the MENA region by summarizing published human prevalence and incidence studies, outbreaks, reported cases, and reports of CHIKV detected in Aedes mosquitoes.Serologic and outbreak evidence of CHIKV transmission has been identified in the countries surrounding the Red Sea (Djibouti, Egypt, Saudi Arabia, Somalia, Sudan, and Yemen) and Pakistan.This distribution overlaps the known distribution of DENV in the MENA region [15].No outbreaks, cases of autochthonous CHIKV antibodies in the populations tested.Given increasing reports of Ae. aegypti and Ae.albopictus occurrence and DENV transmission in the MENA over recent years, the possibility of unrecognized CHIKV transmission or cross-border spread of CHIKV to these countries merits consideration.
As the bulk of epidemiologic evidence for CHIKV is serologic, the genetic diversity of CHIKV circulating in the MENA is largely unknown.To date, the only published CHIKV strain from the MENA comes from the 2011 Yemen outbreak [10].This virus was found to be similar to the East-Central-South African (ECSA) Indian Ocean Lineage that was responsible for the 2004-2006 Indian Ocean epidemic [11, and subsequent urban outbreaks in Southeast Asia, Italy, and France [57].Still, CHIKV lineages may be diverse within North Africa and distinct from other regions in Africa.Distinct enzootic / epidemic ECSA lineages may be circulating in Sudan, for example, given the repeated detection of anti-CHIKV antibodies in Sudanese populations since the 1970s [32,33], Sudan's proximity to East Africa (where CHIKV has long been endemic), and the sylvatic habitats in Sudan's Nuba Mountain region where several yellow fever outbreaks may have emerged from in the past [14,32,[58][59][60].Contiguous spread of CHIKV from India to Pakistan is similarly plausible, as is import-related outbreaks from more remote regions, such as from South Asia to the Arabian Peninsula or from Mediterranean Europe to North Africa.
A variety of ecologic and social risk factors may be driving the spread of Aedes in the MENA, which we describe in detail in our recent publication on DENV in the MENA [15].These factors include increasing urbanization [61], use of open water storage that encourages Ae. aegypti breeding [39,62], armed conflict and poverty [63,64], and extensive inter-and intra-regional trade and migration [12,65].The 2011 CHIKV outbreak in Yemen illustrates several vulnerabilities to Aedes-borne arboviruses potentially faced by the MENA [10].Heavy travel and commerce through the port city of Al Hudaydah was thought to have led to CHIKV introduction, possibly from North African port cities along the Red Sea coast.A CHIKV outbreak was reported in Djibouti City [31] the same year of the Yemen outbreak, and the following year, autochthonous transmission of CHIKV was noted in Saudi Arabia [53].Similar transmission patterns have been cited as risk for DENV infection spread in the Red Sea subregion [10,12,27,66,67].High rainfall in Al Hudaydah months prior to the outbreak may have also amplified the Aedes population to a level capable of sustaining an outbreak.The outbreak was not detected until nearly 16 weeks after its suspected onset, in part attributable to a lack of surveillance and case detection, and initial misrecognition of the disease as DENV in a known dengue-endemic area [10].Prompter initiation of vector control measures may have forestalled the amplification of the outbreak over the following months.Serologic cross-reactions for alphaviruses like CHIKV are well documented and represent an important limitation of epidemiologic studies which rely on serology in the absence of confirmatory neutralization tests.Serologic evidence of CHIKV in the absence of neutralization tests must be interpreted with caution, as such results may represent cross-reactions between CHIKV and another alphavirus (e.g.SINV, ONNV, or SFV), co-cicrulation of both viruses, cross neutralization between both viriuses, or the presence of another antigenically-related virus that induces cross-neutralizing antibodies to both alphaviruses [68].These alphaviruses may have similar clinical presentations and geographic distributions in the MENA region, while the increase in international travel, intensification of trade, and the spread of potential vectors as a result of climate change may further drive overlapping endemicity [69].In our review, serologic cross-reactions were reported in 9 of 14 seroprevalence studies, most commonly due to SINV, another Old World alphavirus identified in the MENA whose epidemiologic impact is also poorly understood [70].Like West Nile Virus, SINV may be broadly distributed in the MENA given its enzoonotic transmission cycle involving migratory avian hosts and ubiquitous Culex mosquitoes [71][72][73].In a study by Ibrahim et al. in Kuwait, 78% (7/9) CHIKV positive samples were cross-reactive to SINV and SFV by HI testing, while crossreactivity between SINV and SFV occurred in an additional 9 of 21 SINV-reactive sera [39].In a second study from Kuwait, all CHIKV positive samples were cross-reactive to SINV [30].This evidence does not confirm CHIKV circulation in Kuwait.Antibodies against ONNV, another Old World alphavirus whose distribution is likely underreported, were only sought (and confirmed by VNT) in one study from Sudan conducted in 1973 [32].Yet although the distribution of ONNV in the MENA region requires further research, it is unlikely that the serologic studies reporting CHIKV are primarily detecting ONNV.A one-way antigenic cross-reactivity has been described between CHIKV and ONNV, in which antibodies against CHIKV neutralize ONNV but anti-ONNV antibodies do not recognize CHIKV [74,75].Further assurance of the detection of CHIKV antibodies is found in the study by Andayi et al. in Djibouti, in which a non-inactivated viral antigen was utilized for ELISA testing (allowing for selection of more specific anti-CHIKV antibodies) and results were confirmed by seroneutralization [31].
CHIKV may continue to spread in the MENA region as it has in other regions.Surveillance programs and clinicians should maintain alertness for CHIKV, particularly in dengueendemic areas, given the overlapping clinical presentation and geographic distribution of both pathogens.Addressing this challenge will require multiplexed serologic and molecular diagnostics able to simultaneously detect and discriminate between CHIKV and other pathogens with overlapping clinical manifestations, such as DENV, malaria, ONNV, Yellow fever virus, and Zika virus [2,10,13,14,48,[76][77][78][79][80].Such diagnostics could also circumvent the issue of cross-reactivity and the frequent unavailability of convalescent serology.However, molecular diagnostics are limited by the presence of viral RNA or antigen for a limited duration during acute infection.Studies to understand the distribution of Aedes and infection rates in Aedes in the MENA are also important, as they can inform our understanding of transmission dynamics which is useful for vector control strategies and predicting future transmission risk [71][72][73].Our review identified only one published study of CHIKV isolated from Aedes in the region [39], depite the published occurrence of Ae. aegypti and Ae.albopictus in 11 and 7 MENA countries, respectively [15].
Our study was limited by its use of select databases of peer-reviewed literature with the possible exclusion of some grey literature which may have provided additional data (the African Index Medicus and Index Medicus for the Eastern Mediterranean Region databases do index grey literature).Reports from public health laboratories in each country may also have provided additional data, though we chose to focus our review on published literature that would be available to the broader academic community in the region.In addition, although we did not impose any publication year restrictions in our search, Embase and the regional databases began indexing articles after 1980, so earlier publications may have been overlooked during our search.PubMed began indexing articles prior to the discovery of CHIKV in 1952-53, however.The urban cycle focus of CHIKV epidemiology must also be underscored in our manuscript, as we did not include seroprevalence studies in animals (i.e.sylvatic cycle) in our review criteria.Still, only one seroprevalence study in animals was identified in our search.It is likely that our study generally reflects the available seroprevalence literature concerning CHIKV in the MENA, as we identified no mention of sylvatic cycle CHIKV in the MENA over the course of our review.CHIKV may primarily exist in a human-peridomestic mosquito cycle in the MENA, as it existis in Asia [6], though sylvatic cycle CHIKV is known to exist elsewhere in Africa [81].Non-publication of studies with small or zero effect size may also have biased the distribution of studies identified in our review.The absence of confirmatory neutralization tests in many studies is an additional limitation, along with the lack of convalescent serum tests to demonstrate antibody titer kinetics in acute infections, both of which would reduce the likelihood of cross-reaction.Indeed, serologic cross-reactions may have overestimated CHIKV seroprevalence in multiple studies.Finally, given the small number of studies and sparse distribution and content of available studies, we did not perform a meta-analysis or explore bias in overall outcome measures through a funnel plot or Egger test nor did we exclude studies based on the risk of bias assessment.This resulted in the inclusion of studies of varying methodological rigor and scientific quality.

Conclusions
In the MENA region, published reports suggest endemic circulation of CHIKV in the Red Sea Region and Pakistan, similar to the known geographic distribution of DENV.Published epidemiologic studies are lacking in many sub-regions, including some Aedes and DENV endemic areas, further suggesting underrecognition of CHIKV in the MENA.These findings articulate the need for further research to understand the epidemiology of CHIKV in the MENA.

Table 1 . Criteria for study inclusion or exclusion. Study type Inclusion Criteria Exclusion Criteria
Any reports of single CHIKV isolates or vRNA detections from Ae. aegypti or Ae.albopictus obtained by any laboratory method Mosquito captured in non-MENA country https://doi.org/10.1371/journal.pntd.0005707.t001

Geographic distribution of human prevalence studies and reported outbreaks and cases for Chikungunya virus in the Middle East and North Africa.
https://doi.org/10.1371/journal.pntd.0005707.g002