Outbreak of Chikungunya in the French Caribbean Islands of Martinique and Guadeloupe: Findings from a Hospital-Based Surveillance System (2013–2015)

INTRODUCTION

Chikungunya fever is a vector-borne viral disease transmitted to humans by infected Aedes mosquitoes. Chikungunya virus (CHIKV) is a member of the Alphavirus genus of the Togaviridae family. From a preliminary historical hypothesis on suspicion of CHIKV cases described in the Caribbean in 1827–1828 through the first isolation of the virus in 1952–1953 in Tanzania until evidence of recent circulation in the Caribbean, knowledge on the natural history of the disease increased substantially. Simultaneously, enhanced surveillance systems were implemented across countries to detect and manage new health threats.^1,2 On December 3, 2013, two CHIKV laboratory-confirmed cases were reported in subjects living on the island of Saint-Martin.³ They were the first autochthonous CHIKV cases ever recorded in the Americas. Early microbiological investigations demonstrated that the circulating virus isolated in December 2013 belonged to the Asian genotype distinct from the ECSA genotype isolated during the 2005–2006 epidemics in the Indian Ocean.⁴ On December 9, 2013, the World Health Organization launched an epidemiological alert urging Caribbean countries and territories to enhance their surveillance system to improve their capacity of early detection of CHIKV cases.⁵ Because of intense travelers’ traffic within the area, every Caribbean country or territory was considered to be at high risk for CHIKV emergence.⁶ Countries implemented preparedness and response plans to monitor and prevent the risk of CHIKV dissemination through public health measures to fight against Aedes mosquitoes.^7–11

On December 18 and 25, 2013, health authorities reported the first confirmed CHIKV cases in Martinique and Guadeloupe, respectively.¹² Soon after, neighboring countries started to notify that they were detecting suspected and confirmed CHIKV cases.^13,14 Previous CHIKV outbreak reports highlighted the substantial impact of CHIKV introduction in naive immune populations. Although the course of the disease is usually benign, atypical, and severe cases, including cases with fatal outcomes, have been described.¹⁵ From December 2013 onward, the inhabitants of Martinique and Guadeloupe were affected by major CHIKV epidemics. After the end of the epidemics in November 2014 (Guadeloupe) and January 2015 (Martinique), an overall estimated 154,000 suspected clinical CHIKV cases had consulted general practitioners between December 2013 and January 2015, and we estimated that twice as many cases had not sought medical care. It has been estimated that around 308,000 suspected CHIKV cases arose, accounting for approximately 40% of the population of both islands.¹⁶

In the beginning of the outbreak, little was known about the virulence of the new CHIKV in a setting where the population had no immunity, and no specific treatment was available. Therefore, it was decided to implement a hospital-based surveillance system in both territories. The objectives were to estimate the incidence of hospitalized cases, to assess the burden of severe cases, and to describe features of hospitalized cases. Herein, we focus on the epidemiological and clinical description of hospitalized cases. We assessed the burden of severe cases and deaths related to CHIKV infection in patients admitted to public hospitals throughout the outbreak.

MATERIALS AND METHODS

Case definitions.

A case was defined as any patient cared for in the emergency room or acute ward of a public hospital in Martinique or Guadeloupe for at least 24 hours for acute CHIKV infection confirmed either by serological testing (Immunoglobulin M antibody titers) or by detecting the presence of CHIKV genome using a real-time reverse transcription polymerase chain reaction (RT-PCR) assays on blood or cerebrospinal fluid samples. Patients admitted to hospital more than 3 weeks after symptoms onset or for a reason not related to CHIKV infection were excluded from this study.

Neonatal CHIKV infections can be vector-borne or result from mother-to-child transmission (MTCT). A MTCT case was defined as a case developing within 10 days of life in a neonate whose mother had confirmed CHIKV infection (RT-PCR) around the time of delivery. We further sorted MTCT cases into two categories: a severe MTCT case had at least one organ failure (respiratory, cardiovascular, cerebral, hepatic, renal, and/or any other organ failure) while a nonsevere MTCT case had no organ failure. Vector-borne cases were assigned to one of the three following categories. A typical case was defined as a case aged 10 days or more and presenting with at least one of the following clinical manifestations: fever, stomach pain, diarrhea, rash, edema, arthralgia, headache, myalgia, itching, or tenosynovitis. An atypical case had at least one additional symptom among the following: severe pain (administration of class III analgesics); neurological manifestations (encephalitis, seizure, or any other neurological disorder); acute cardiovascular disorder (with or without pre-existing cardiac condition); dermatological features (other than rash and/or itching); respiratory, hepatic (hepatic cytolysis), renal or digestive (other than stomach pain and/or diarrhea) disorders; decompensation of a pre-existing comorbidity; and/or hemorrhagic manifestations. A severe case was defined as a case developing one or several organ failures in accordance with international consensus definitions: cardiovascular, cerebral, respiratory, renal, hepatic failure, or any other systemic or organ failure (Figure 1).¹⁷

View Full Size

Diagram of classification of clinical forms of chikungunya virus (CHIKV) infected hospitalized patients, Martinique and Guadeloupe - December 2013 to January 2015. This figure appears in color at www.ajtmh.org.

Citation: The American Journal of Tropical Medicine and Hygiene 98, 6; 10.4269/ajtmh.16-0719

• Download Figure
• Download figure as PowerPoint slide

RESULTS

Overall, we collected 2,024 questionnaires corresponding to as many eligible patients, 1,836 of whom fulfilled the case definitions. Reasons for excluding the 188 remaining patients (9%) were length of hospital stay less than 24 hours (N = 26), admission to hospital more than 3 weeks after the onset of symptoms (N = 41) or admission for a reason not related to CHIKV infection (N = 74), incomplete medical records (N = 36), and miscellaneous (N = 11).

Between December 2013 and January 2015, we investigated an overall 1,836 hospitalized cases of which 1,196 cases were in Martinique, and 640 were in Guadeloupe. The number of hospitalized cases peaked in early May 2014 and early June 2014, respectively, in Martinique and Guadeloupe. Hospitalized case curves mirrored that of clinical CHIKV cases occurring in the general population (Figure 2).

View Full Size

Weekly number of hospitalized chikungunya virus (CHIKV) cases (left axis) and weekly estimated number of clinical CHIKV cases consulting general practitioners (right axis) in the population during the CHIKV epidemics in Martinique and Guadeloupe - December 2013 to January 2015. This figure appears in color at www.ajtmh.org.

Citation: The American Journal of Tropical Medicine and Hygiene 98, 6; 10.4269/ajtmh.16-0719

• Download Figure
• Download figure as PowerPoint slide

DISCUSSION

The emergence of chikungunya in the Caribbean in December 2013 was the start of the first reported epidemic caused by CHIKV in this part of the world. We set up an active hospital-based surveillance system in two French Caribbean territories to assess severity of the phenomenon and to detect atypical forms of the disease. Our main results show that a total of 1,836 CHIKV cases were hospitalized in Martinique and Guadeloupe: the estimated hospitalization rate was 60 per 10.000 CHIKV clinical cases. Severity rate accounted for approximately 12 per 10,000 CHIKV clinical cases and fatality rate for approximately 5 per 10,000 CHIKV clinical cases. Our results also emphasized MTCT occurrence during the epidemic with half of the MTCT cases being severe.

We targeted all major, acute, or emergency wards of public hospitals as severe cases would be mostly cared for in public hospitals. Official hospital data emphasize this hypothesis as public and private hospital activities indicate that 78% of the total number of hospital stays in medical departments rely on public hospitals (95% in Martinique and 62% in Guadeloupe). Therefore, it is unlikely that severity estimates based on the assumption that severe cases would mainly be treated in public hospitals have been biased by the setting of this observational study.

Severe clinical forms of the infection arose in Reunion Island during the massive CHIKV epidemic.^15,18 Although Reunion Island and Martinique/Guadeloupe have similar demographic sizes with, respectively, 785,000 and 785,076 inhabitants and similar medical infrastructures, the hospital surveillance system set up in the Caribbean captured more cases. In comparing the estimated population at risk for hospitalization in both regions during the epidemics, this population was larger in Martinique and Guadeloupe as an estimated additional 70,000 suspected CHIKV cases arose compared with that of Reunion Island. Indeed, in La Reunion, a total of 878 hospitalized cases were reported out of an estimated 244,000 suspected CHIKV cases (36 per 10,000 CHIKV cases) occurring in the background population whereas in the two French Caribbean territories, 1,836 cases were admitted to hospital out of an estimated 308,000 suspected CHIKV cases (60 per 10,000 CHIKV cases).¹⁹ Furthermore, sensitivity of the French Caribbean hospital surveillance was greater than that in Reunion island as any symptomatic biologically confirmed case was included in the first system, whereas patients with symptoms limited to fever and arthralgia were excluded from the second system.¹⁵ Other factors might have played a role in the difference, such as diagnostic and hospitalization practices differing across countries.

Severe clinical forms of the infection arose in Reunion Island during the massive CHIKV epidemic. We also reported severe cases with fatal issues as reported in the Caribbean and Latin American countries affected by CHIKV epidemics during the same period.^20,21 Although CHIKV infection has commonly been considered a mild disease, the growing number of publications and reports from different countries underlines the potential severity of the infection. We emphasized that the incidence rate of severe cases in Reunion Island was similar to that in Martinique and Guadeloupe despite the fact that the genotypes circulating in the two regions (Indian Ocean and the Caribbean) were different. Incidence and fatality rate of severe hospitalized cases were respectively 10 and 2.9 per 10,000 suspected CHIKV cases in Reunion Island, compared with 11.5 and 2.4 per 10,000 suspected CHIKV cases in pooled cases from Martinique and Guadeloupe. Lahariya and Pradhan²² notified in their review on the reemergence of CHIKV in India that the infection is severe in infants, in elderly, and in immunocompromised patients. We have not identified a greater severity in the very young especially in children contaminated through vector transmission although two deaths were reported. Economopoulou and Dominguez emphasized that among other risk factors, older age (> 85 years) was associated with increased mortality.^15,18 Our results also suggested that deaths occurred mostly in elderly persons although five deaths were recorded in patients aged more than 1 year and less than 60 years of age. Comorbidity might be associated with severity as we reported higher proportion of comorbidity in severe cases compared with others. Investigations of risk factors in Reunion Island highlighted that cardio-vascular, respiratory disorders or hypertension were associated with increased risk of severe illness.^15,18 Furthermore, in our study diabetes was reported in 29% of severe clinical cases whereas prevalence of diabetes was estimated to be 5.6% and 6.9% in Martinique and Guadeloupe, respectively.²³ The observed discrepancy between diabetes prevalence in the general population and burden of diabetes in our study population suggests that diabetes could play a role in the clinical course of the infection or in hospitalization practices. The same finding was underlined in a previous publication on chikungunya.²⁴ Further investigations are needed to explore whether diabetes is a risk factor of severity especially in countries where diabetes is a major public health concern, such as in the Caribbean.

Regarding clinical signs, fever and arthralgia were the most frequently reported symptoms as widely described in the literature.^2,24–28 The atypical or severe clinical features affecting our patients were mainly neurological, cardiovascular, or renal disorders. Among cases with neurological manifestations, Guillain–Barré syndrome cases, seizure in children, and encephalitis in adults were reported. We hypothesized that Guillain–Barré syndrome might have occurred more frequently or have simply been more frequently reported compared with previous epidemics.^18,29 Whether the neurovirulence of the circulating Asian genotype led to greater neurological impact in comparison to that circulating earlier might be hypothesized. However, further clinical and epidemiological investigations are needed to explore this hypothesis. Although the disease was benign in most CHIKV cases occurring in the two French Caribbean islands of Martinique and Guadeloupe, severe and fatal outcomes were reported. In adults, occurrence of death was frequently related to the decompensation of an underlying disease and mostly indirectly caused by the CHIKV infection.

Another interesting finding is that 80% of hospitalized cases were nonsevere cases. This result suggests that given the context of an emerging infection, the debilitating nature of the disease, and uncertainty on clinical evolution including occurrence of deaths, clinicians unfamiliar with the infection were probably more likely to hospitalize CHIKV infected patients including those without severe clinical presentations.^26,30

Fifteen neonatal cases consecutive to MTCT occurred in Martinique and Guadeloupe, confirming vertical CHIKV transmission.^31,32 Almost half of reported MTCT cases in Martinique and Guadeloupe were severe cases. This result is concordant with study results on vertical transmission in Colombia, Salvador, and Dominican Republic emphasizing that severity also occurred in neonates born from CHIKV infected mothers. In Reunion Island where vertical transmission was first reported, a higher number of MTCT cases were confirmed (44 MTCT cases) despite the smaller magnitude of the epidemic.¹⁸ With enhanced health communication on the risk of vertical transmission already known during the 2013–2015 Caribbean epidemics, smaller birth cohorts in the French Caribbean (9,200 births in 2013) compared with those of Reunion Island (14,800 births in 2005) might explain the lower recorded neonatal cases consecutive to MTCT. Regarding the high sex ratio of MTCT cases with boys being overrepresented, we cannot provide any relevant explanation as gender difference practices in newborns are very uncommon in this setting.

Awareness on possible severity and deaths associated with the infection has grown over the past decade as chikungunya has been reemerging in different parts of the world where the competent vectors are endemic. Adequate measures and prevention could alleviate the health impact of this viral infection in population groups most at risk: infants, the elderly, persons with underlying health conditions, and pregnant women. Enhanced vector control strategies and vaccine research initiated with promising preliminary results should be strongly supported to empower public health stakeholders with a wider and more effective range of preventive tools.³³