First cases of Angiostrongylus cantonensis infection reported in Martinique, 2002–2017

Neuroangiostrongyliasis is a parasitic disease caused by the accidental ingestion of the nematode Angiostrongylus cantonensis in its larval form. Human infection can lead to eosinophilic meningitis, sometimes complicated by life-threatening radiculomyelitis or encephalitis. Although some cases have been reported from other Caribbean Islands, no cases have been diagnosed in Martinique so far. Here, we report the first eight laboratory-confirmed cases of neuroangiostrongyliasis on the island of Martinique, French West Indies, between 1 January 2002 and 31 December 2017. One case was fatal and five resulted in neurological sequelae. The medical community should consider the risk of A. cantonensis infection in patients living in or returning from Martinique.


Introduction
Angiostrongylus cantonensis is a nematode parasite that is the leading cause of infectious eosinophilic meningitis in humans in tropical and sub-tropical regions [6]. The life cycle involves rats as definitive hosts (mainly Rattus spp.) [49], various gastropods as intermediate hosts, and crustaceans [23], fishes and various other species as paratenic hosts [6].
Human infection is accidental, by ingestion of stage 3 larvae (L3) in gastropods or in paratenic hosts [11]. Neuroangiostrongyliasis is commonly a self-limited meningitis syndrome, but a large spectrum of clinical manifestations is possible [5]. Clinical manifestations range from asymptomatic disease and mild headaches to radiculomyelitis and encephalitis that can lead to permanent neurological injury or even death [29]. Most human cases of neuroangiostrongyliasis have been recorded *Corresponding authors: dard.celine@gmail.com, nicole.desbois-nogard@chu-martinique.fr a These authors contributed equally to this work.
in East and Southeast Asia, and the Pacific Basin, but the disease appears to be emerging in Australia [1], South America [13,32,44], the United States [3,21], and some islands of the Caribbean (Cuba, Haiti, Dominican Republic, Jamaica, and Guadeloupe) [12,19,42]. Several patients infected with A. cantonensis have been diagnosed with neuroangiostrongyliasisbut not reported in the scientific literatureduring the last few years on the island of Martinique, a French overseas department in the Lesser Antilles with a population of 371,200inhabitants as of 1 January 2018 (INSEE census, French National Institute of Statistics and Economic Studies, https:// www.insee.fr). The objectives of this study were to estimate the incidence and assess the clinical and biological features of neuroangiostrongyliasis in Martinique.

Study design
A retrospective single-centre observational study was undertaken at the University Hospital of Martiniquethe main hospital on the islandspanning the period 1 January 2002 -31 December 2017.

Inclusion and exclusion criteria
The following clinical and biological inclusion criteria were used: (i) neurological symptoms requiring lumbar puncture; (ii) eosinophilic meningitis defined as the presence of more than 10 eosinophils/mm 3 in the cerebrospinal fluid (CSF) or 10% of the total CSF leukocyte count [5]; and (iii) detection of anti-A. cantonensis immunoglobulins in serum and/or CSF by indirect immunofluorescence assay (IIF) [22] or by western-blot (31-kDa antigen [34]). Patients with eosinophilia in the CSF due to a traumatic lumbar puncture or blood eosinophilia of another aetiology were excluded. Neuroangiostrongyliasis cases were defined as "confirmed" for patients who met the first two inclusion criteria, and with the detection of specific anti-A. cantonensis immunoglobulins in serum and/or CSF by western-blot. When A. cantonensis serology was found to be positive by IIF only, cases were defined as "probable" because of the low/lack of specificity of this method for nematode infections, in which some cross-reactions can be observed [47,48].

Data collection
Demographic data including exposure history, laboratory results (from blood and CSF), clinical presentation, imaging features (brain scan or MRI), and outcomes (recovery, sequelae, and death) were retrospectively collected from the medical charts, according to the legal and ethical guidelines of the French National Committee on Data Protection (CNIL).

Angiostrongylus cantonensis serology by indirect immunofluorescence assays
From 2002 to 2010, samples provided for A. cantonensis serological testing were sent to the only laboratory performing neuroangiostrongyliasis infection diagnosis in France, including its overseas regions (Laboratory of the Centre Hospitalier de Gonesse, Gonesse, France). An indirect immunofluorescence (IIF) assay was used for the detection of antibodies against A. cantonensis antigens, as described in [18]. Due to logistical constraints and availability of certain reagents, patient serodiagnosis in this laboratory ceased in 2010 and no other laboratory then performed the test in France. Subsequent serological analyses were therefore performed in Thailand or Switzerland by western-blot.

Angiostrongylus cantonensis serology by western-blot
Western blot assays using antigens derived from A. cantonensis adult worms were performed to detect IgG against A. cantonensis in either the Department of Parasitology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand, or the Swiss Tropical and Public Health Institute, Basel, Switzerland. The detection of the 31-kDa band confirmed serum positivity as it shows high sensitivity and specificity (>99% for both) for the diagnosis of A. cantonensis infections [17].

Results
Descriptive results of the clinical presentations, including biological, imaging, and epidemiological features, are shown in Table 1 and statistically analyzed in Table 2. During the 16-year period of the study, four confirmed and four probable cases of neuroangiostrongyliasis were diagnosed in Martinique, among which three were children below 2 years of age, one was an 11-year-old boy, and four were adults aged from 37 to 64 years. The annual incidence rate was 0.14 cases/100,000 inhabitants/year (95% CI [0.04-0.23]) with six of the cases occurring during the rainy season from June to November. All patients were born and lived in Martinique and none reported recent travel. Contact with molluscs was reported in two cases. All five patients older than 2 years of age (Table 1) presented with pre-existing mental disorders (pica syndrome, bipolar disorder, intellectual disability, or autism), which may have promoted the accidental or even deliberate consumption of snails. All cases presented with acute neurological signs and/or symptoms requiring a lumbar puncture: dysfunction of the cranial nerves (highlighted by clinical neurological examination of the 12 pairs of cranial nerves), headaches, axial hypotonia, seizures, radiculalgia, and neck stiffness. Clinical examination was difficult for one patient because of autism. During hospitalisation, five patients had fever and three had digestive symptoms (vomiting, abdominal pain, loss of appetite, and/or diarrhoea). Brain imaging was performed for all patients (CT scan or MRI): five presented abnormalities, with abnormal enlargement of the cerebral ventricles or cortical atrophy, and three were normal. Blood eosinophilia at admission was inconsistent, with a median of 1.72 G/L (13% of the WBC count in blood) and a range of 0.49-6.43 G/L (5-31%). Median eosinophilia in the CSF at first lumbar puncture was 74.5 (25% of the WBC count in CSF), with a range of 0-1550/mm 3 (4-68%)  [35] and/or corticosteroids [18]. Clinical outcomes ranged from rapid recovery without sequelae for two patients to neurological sequelae manifested as strabismus and intellectual disability for five patients. One case was fatal for a 58-year-old man. Diagnosis was made by anti-A. cantonensis antibody detection in sera for seven patients and confirmed positive in CSF for five of them. One patient only showed detectable anti-A. cantonensis antibodies in the CSF. Serological analysis was negative for two patients nine days after admission, but positive 19-38 days after admission.

Discussion
This study reports the first eight laboratory-confirmed cases of neuroangiostrongyliasis in Martinique, which occurred between 2002 and 2017, thus extending the range of Caribbean islands with proven human cases of neuroangiostrongyliasis [12,20,44]. Given the potential lethality of neuroangiostrongyliasis, the medical community should therefore strongly consider the possibility of this infection in patients living in or returning from Martinique with eosinophilic meningitis.
In this study, neuroangiostrongyliasis cases occurred in two distinct epidemiological situations: three in infants less than 2 years old and five in patients over 11 years with mental diagnoses. The mode of transmission in infants was linked to poor living conditions with probable accidental consumption of contaminated snails or slugs when playing outside, as described on Mayotte, a French island in the Western Indian Ocean, in children under two years of age [18]. For adults, contamination was instead linked to intellectual disability associated with risky eating behaviour. In particular, pica syndrome may favour infection through the ingestion of gastropods usually not consumed as food in Martinique. No case was reported following consumption of raw paratenic hosts like shrimp, which is in contrast to the main source of infection in French Polynesia [35]. In our study, the diagnosis was initially unclear for all the patients and other kinds of helminthiases were initially suspected, as Martinique had never been reported as an endemic region for A. cantonensis. All but one of the patients were therefore treated with anti-helminthic drugs, although the efficacy and safety of albendazole or mebendazole for neuroangiostrongyliasis treatment remains controversial because of theoretical concerns that they may worsen the inflammatory response to dead and dying worms [5]. The four cases diagnosed after 2013 were also treated with corticosteroids, postulated to provide relief by reducing inflammation and thereby intracranial pressure and headache intensity [9,43].
Comparison of the incidence of human neuroangiostrongyliasis in Martinique with neighbouring Caribbean Islands is straightforward. Indeed, cases were mainly reported in travellers returning from the Caribbean and it is likely that numerous autochthonous cases have not been reported in the scientific literature. Most Caribbean cases were reported in Cuba with several dozen cases since the 1980s [2,36], mainly in the cities of Havana and Villa Clara [15,16,28,37], and one case in a Swiss traveller returning from Cuba [7]. In Guadeloupeanother French West Indies islandfour autochthonous cases were diagnosed (to our knowledge) between 1999 and 2017 in young children who may have been in contact with infected molluscs ( [12] and unpublished data), corresponding to an annual incidence rate of 0.053 cases per inhabitant per year (95% CI [0.001-0.105]), close to that of Martinique. For the Dominican Republic, two suspected cases were reported in travellers returning to Europe [24,41]. For Jamaica, twelve cases were diagnosed in adult travellers returning from Jamaica [38,42], seven in autochthonous young children [19,26] and one ocular case in a young woman [30]. No human cases have been reported in Grenada, the Bahamas, Haiti, and Puerto Rico, although A. cantonensis has been found in the environment on these islands [44]. It is noteworthy that the disease is also expanding in North and South America, in particular in the United States [27], Brazil [32], and some other South American countries [13,44]. In this study, most cases occurred during the rainy season, during which snails abound, particularly the giant African snail, Lissachatina fulica, which was introduced to Martinique in 1989 [31] and is known elsewhere to act as an intermediate host of A. cantonensis. In Guadeloupe, A. cantonensis infection in A. fulica was 32.4% in 2014 [12]. No doubt other snail and slug species could also act as hosts as there are close to 90 nonmarine mollusc species in Martinique [14,23]. Numerous rodents have also been reported as potential definitive hosts of A. cantonensis worldwide [49]. Two species of rats, Rattus norvegicus (brown rat) and Rattus rattus (black rat), have been present in the territory since the late 18th century and probably play the role of definitive host of A. cantonensis in Martinique as they are the only rodent species in Martinique other than the mouse Mus musculus. No study has evaluated infection of rats with A. cantonensis in Martinique. In neighbouring Grenada [8,10], Puerto Rico [4], Dominican Republic [45], Haiti [40], Jamaica [46], and Cuba [2], the proportion of infected Rattus spp. varies from 23.4% to 60.0% [2,8], while the parasite appears to be absent in rats in Barbados [25].
Neuroangiostrongyliasis cases in Martinique seem particularly severe relative to other case-series reported in China and South-East Asia, with a higher mortality and sequelae rate. However, given the low number of cases, this must be confirmed as the number of neuroangiostrongyliasis cases increases in Martinique.
All cases were diagnosed by anti-A. cantonensis immunoglobulin detection in serum and/or CSF. Diagnosis was not performed by PCR as it was not available in Martinique or metropolitan France at the time of initial diagnosis [33] and no remaining CSF samples were available in our biobanks. In fact, molecular detection of A. cantonensis in CSF was developed in the early 2010s [48] and was only recently validated for clinical use [39]. The recent availability of a specific A. cantonensis PCR test in the French departments of South America and the Caribbean should improve the diagnosis of this disease in this region and encourage local authorities to undertake epidemiological studies on the intermediate and paratenic hosts and reservoirs, which should broaden our understanding of disease transmission in Martinique.