Angiostrongylus costaricensis infection in Martinique, Lesser Antilles, from 2000 to 2017

Human abdominal angiostrongyliasis (HAA) is a parasitic disease caused by the accidental ingestion of the nematode Angiostrongylus costaricensis in its larval form. Human infection can lead to severe ischemic and inflammatory intestinal lesions, sometimes complicated by life-threatening ileal perforations. Only one case had been reported in Martinique, an Island in the French Antilles, in 1988. We retrospectively reviewed the medical charts of patients diagnosed with abdominal angiostrongyliasis at the University Hospital of Martinique between 2000 and 2017. The objectives of this study were to evaluate the incidence and perform a descriptive analysis of the clinical, biological, radiological, and histopathological features of HAA in Martinique. Two confirmed cases and two probable cases were identified in patients aged from 1 to 21 years during the 18-year period, with an estimated incidence of 0.2 cases per year (0.003 case/year/100.000 inhabitants (IC95% = 0.00–0.05)). All patients presented with abdominal pain associated with high blood eosinophilia (median: 7.24 G/L [min 4.25; max 52.28 G/L]). Two developed ileal perforation and were managed by surgery, with diagnostic confirmation based on histopathological findings on surgical specimens. The other two cases were probable, with serum specimens reactive to Angiostrongylus sp. antigen in the absence of surgery. All cases improved without sequelae. The description of this case series highlights the need to increase awareness of this life-threatening disease in the medical community and to facilitate access to specific diagnostic tools in Martinique. Environmental and epidemiological studies are needed to broaden our knowledge of the burden of this disease.


Introduction
Human abdominal angiostrongyliasis (HAA) is a zoonotic disease caused by a nematode, Angiostrongylus costaricensis Morera & Céspedes, 1971 [44]. The definitive hosts are rodents of the Cricetidae, Heteromyidae, and Muridae families [19,40,60,62]. Adult nematodes reside in the mesenteric arterial system of wild rodents, in which females lay eggs that generate first-stage larvae (L1), which are shed in the rodents' feces. Larval maturation to the third-stage (L3) occurs in intermediate hosts, mainly slugs from the families Veronicellidae and Limacidae [11,21,46,62]. Human infection is accidental and occurs by ingesting third-stage larvae (L3) from mollusks or vegetables contaminated with their slime [43]. Once ingested, the larvae invade intestinal tissues, reach sexual maturity, and release eggs in the ileo-cecal mesenteric arteries, causing eosinophilic enteritis in humans [66].
A. costaricensis was first discovered in the mesenteric arteries of humans in Costa Rica in 1967 [7,45], followed by the description of adult worms in the rodent Sigmodon hispidus in 1971 [44]. A. costaricensis is now found from Texas [64] southward to Argentina [52], including Honduras [27,58], Venezuela [23,69], Mexico [70], Brazil [71], Colombia [35], Nicaragua [12], El Salvador [68], Ecuador [30], Guatemala [28], Panama [63], Peru, [60] and probably French Guiana [65]. The disease is a public health problem in South America, in particular in Costa Rica, where it affects 12/100,000 persons, with approximately 500 new cases each year [43]. Some sero-epidemiological studies in South America have shown strong seroprevalence rates in humans, i.e., from 29.8 to 66.0% in endemic areas of Southern Brazil. This implies numerous asymptomatic infections [17] and a far broader distribution of the parasite in the Americas than previously believed [34]. In contrast, HAA is rarely reported in the Antilles and only six sporadic cases have been described since 1974 (Table 2). Among them, one was reported in Martinique, an island in the French Antilles.
The main objective of this study was to evaluate the incidence of symptomatic HAA cases in Martinique. The secondary objective was to perform a descriptive analysis of the clinical, biological, radiological, and histopathological features of these cases.

Setting
Martinique is a French Overseas Department in the Antilles, with a population of 380,877 inhabitants as of January 1 st , 2015 (INSEE census, www.insee.fr). It has a tropical climate, with a rainy season from June to November and a dry season from December to May.

Study design
A retrospective monocentric observational study was performed in the University Hospital of Martinique between January 1, 2000 and December 31, 2017. Data were extracted from the hospital data information system (PMSI), in which classification is based on the International Classification of Diseases, Tenth Revision (ICD-10). Hospital data codes for HAA (B813) were selected from the PMSI databases. Demographic data, abdominal imaging, biological results, clinical features, and outcomes were anonymously and retrospectively collected from the medical charts according to the legal and ethical guidelines of the French National Committee on Data Protection (CNIL). Serological assays to detect IgG against Angiostrongylus sp. were performed at the Swiss Tropical and Public Health Institute, Basel, Switzerland. Sera were first tested using the ELISA helminth screening test (detecting Toxocara sp., Trichinella sp., Echinococcus sp., Fasciola sp., Filaria, Schistosoma sp., and Strongyloides sp.) followed by a western blot using antigens derived from A. cantonensis adult worms [13].

Case definition
We defined a confirmed case as a patient with clinical symptoms and biological results consistent with HAA (fever, abdominal tenderness, and blood eosinophilia) and histopathological findings of HAA (identification of worms, eggs, or larvae in the intestinal wall). A probable  case was defined as a patient with clinical symptoms consistent with HAA and a serum specimen with IgG reactive to Angiostrongylus sp. antigen.

Ethics statement
The variables were secondarily anonymized and retrospectively collected from medical charts. The French National Committee on Data Protection (CNIL)authorizes the retrospective use of anonymous patient files on the site of patient care in a single hospital.

Results
During the 18-year period of the study, two confirmed and two probable cases of HAA were identified (male: female 50:50, median age: 7.5 years [min 1; max 21 years]). The annual incidence rate was 0.003 cases/100,000 habitants/year (95 CI% = 0.00-0.05). Most cases (75%) were diagnosed during the rainy season. All cases presented abdominal pain associated with high blood eosinophilia (median: 7.24 G/L [min 4.25; max 52.28 G/L]). The eosinophilia rate was not related to the severity of the disease. Cases 1 and 2, diagnosed in 12-month old children, were particularly severe and required surgical procedures with diagnostic confirmation by histological findings. These cases were characterized by anemia, a marked loss of weight and the presence of Charcot Leyden crystals in feces. Cases 3 and 4, probable, were diagnosed in a teenager and an adult with serum specimens reactive to Angiostrongylus sp. antigen. The length of hospitalization was variable (median 17.5 days [min 7; max 53 days]) and correlated with disease severity. All cases improved without sequelae. The clinical presentation along with the biological, imaging, histopathological, and epidemiological features are described in Table 1.

Discussion
Here, we report two confirmed and two probable cases of HAA in Martinique, thus bringing the total number of HAA cases to 10 in the entire Antilles. Indeed, only six sporadic cases of HAA have been reported in the Antilles over the last two decades. Two cases were diagnosed in travelers returning from the Greater Antilles, one from Puerto Rico [47], and the other from the Dominican Republic [59]. In the Lesser Antilles, one case was reported in Martinique in 1988 in a 16-month-old boy [24], followed by two cases in Guadeloupe in 1987 and 1989 in a 20month-old and a five year-old, respectively [26], and a presumed case in the Commonwealth of Dominica in a North-American student in 1997 [50]. Clinical and biological features of these cases are summarized in Table 2.
In our case series, the diagnosis of angiostrongyliasis was considered after admission to hospital because of the nonspecific clinical presentation of the disease [33]. Symptoms usually include abdominal pain in the right iliac fossa along with fever, anorexia, vomiting, and persistent eosinophilia (> 2 G/L). The disease is generally mild and self-limiting, but some cases can be complicated by intestinal infarction, pseudo-tumor, acute appendicitis, or digestive perforation, requiring emergency laparotomy and surgical care with an unpredictable prognosis [33].
Typically, diagnosis occurs unexpectedly when an exploratory laparotomy or laparoscopy is required with histological examination of unhealthy tissues. Definitive diagnosis is established when histological examination of resected specimens shows eggs, larvae, or adult parasitic forms in mesenteric arteries [19]. In the absence of parasites, histopathological findings can help the diagnosis when they show granulomatous reactions with massive eosinophilic and giant cell infiltration in the intestinal wall and regional lymph nodes and/or eosinophilic vasculitis of arteries, veins, and lymph vessels [19]. In subclinical forms not requiring laparotomy or surgery, diagnosis may be established when IgG anti-crude adult worm antigens are found by ELISA-based serological analysis, but such analyses are available in only a few laboratories worldwide [1,3,15,48,67]. Serodiagnosis of A. costaricensis is somewhat unsatisfactory because of cross-reactions with A. cantonensis, Strongyloides stercoralis, and Gnathostoma spinigerum [47]. Better specificity is observed when antigens are derived from A. costaricensis eggs or the reproductive organs of females [4,15]. New tools are now being used to improve the diagnosis in countries in which the disease is endemic, such as Brazil, particularly PCR on paraffin-embedded biopsy tissue or sera, which can lead to  [8,53]. Unfortunately, such biological tools are not yet available in the French territories. In our study, the pediatric cases (cases 1 and 2) illustrate the severe and chronic form of the disease, with necrotizing intestinal inflammation, requiring laparotomy and partial intestinal resection. These two cases were characterized by weight loss, anemia, and a long hospital stay (Table 1), consistent with the results observed in the three reported pediatric cases in Martinique and Guadeloupe in the 1980's [24,26]. In both of our pediatric cases, examination of the ileo-cecal surgical specimen unexpectedly led to the diagnosis of HAA through microscopic identification of A. costaricensis eggs and larvae in the context of a typical, intense ileo-cecal inflammatory, eosinophilic, and granulomatous reaction. Case 1 was particularly intriguing due to the presence of degenerated helminth eggs in the feces after surgery (Fig. 2). We could not confirm them as A. costaricensis eggs based solely on morphological observation and molecular investigation was not performed. Indeed, detection of A. costaricensis eggs in feces has rarely been described, since their elimination is prevented by the inflammatory reaction in the intestinal wall. However, in this case, surgery may have liberated the eggs in the digestive tract. Cases 3 and 4 illustrate the presumptive and probable diagnoses of less severe forms of HAA, based solely on abdominal symptoms and marked eosinophilia. The absence of histopathological examination of digestive specimens and specific A. costaricensis serological and PCR tests in the French territory hampered a definitive diagnosis. The main elements supporting the diagnosis of HAA were the positive results to A. cantonensis serological tests (which are often cross-reactive A. costaricensis antigens), combined with negative results for serological tests for other parasites. Several negative parasitological examinations of feces and the absence of headaches and neurological symptoms, respectively ruled out a possible differential diagnosis of strongyloidiasis and angiostrongyliasis due to A. cantonensis. All patients recovered without sequelae.
There is no consensus concerning the treatment of HAA [38]. It is mainly supportive, focusing on analgesia, hydration, and nutrition. Surgery can solve ischemiarelated intestinal damage and perforation. Anthelmintic treatment using benzimidazole-derived compounds is debatable because their larvicidal effect aggravates the inflammatory response, leading to more severe lesions, and may favor the erratic migration of adult parasites and larvae [39]. Recent studies in mouse models showed that prophylactic enoxaparin treatment does not prevent tissue damage and mortality related to abdominal angiostrongyliasis [54,55]. The four patients in our case series were treated with an anthelmintic as standard treatment for cases of high eosinophilia before diagnostic confirmation.
The mode of transmission of HAA varies depending on the geographical area, generally through the slime of mollusks (i.e., mollusks mouthed by young children or in poorly washed vegetables or aromatic plants) or the consumption of raw mollusks (i.e., during atypical medicinal practices) [28]. The mode of transmission for the two confirmed cases is unclear, as no evident contact with mollusks was reported for either patient. However, environmental investigation in one case found the frequent presence of slugs (und. species) near the house, sometimes reaching the bathroom, and the frequent presence of Achatina fulica snails in their favorite strolling zone in the Morne-Rouge district. In the other case, the parents did not exclude contact between their child and mollusks, but no specific event was reported. The modest family house was in a district infested with slugs and snails during the rainy season, including Limicolaria aurora and A. fulica, and surrounded by brush and sugar cane fields with many rodents. These mollusks were not examined to ascertain the presence of A. costaricensis. Aside from the adult case diagnosed during the dry season in February 2017, all diagnoses in children were made during the rainy season, when slugs and snails are abundant. Environmental studies are needed to better understand the routes of HAA transmission and evaluate the infection rate and dissemination in mollusks and rodents in Martinique.
The only environmental investigation in the French Antilles was conducted on Rattus rattus and Rattus norvegicus in Guadeloupe (an island close to Martinique) in 1992 and showed that 7.5% of rats tested were naturally infected by A. costaricensis [25]. These rat species are also found in Martinique and may be the main definitive hosts there [49] (Table 3). Among the most common definitive hosts in South America, the rodent families Cricetidae and Heteromyidae are absent in Martinique [52,61,63]. Slugs, acting as intermediate hosts for A. costaricensis in South America, are also found in Martinique, including the Veronicellidae family (Sarasinula plebeia, Diplosolenodes occidentalis) and limacid slugs [10,62,63] (Table 3). The aquatic snails Biomphalaria glabrata and B. straminea could have been a potential intermediate host, but are now considered to have been eradicated in Martinique, following a control program on intestinal parasitosis initiated in 1978 [10,22] (Table 3). Finally, A. fulica snails are not considered to be major intermediate hosts in the wild, although they are capable of hosting A. costaricensis larvae in laboratory models [6]. This invasive species, first described in 1989 in Martinique, is responsible for the emergence of central nervous system angiostrongyliasis due to Angiostrongylus cantonensis in the Lesser Antilles [9].
HAA is an emerging parasitic disease in the neotropics, which is not critical in most cases, but nonetheless potentially life-threatening. In Martinique, this zoonosis is sporadic and rare, with an estimated incidence of 0.003 cases/100,000 habitants/year in this study. However, HAA can be misdiagnosed due to its nonspecific clinical presentation, paucisymptomatic cases, and the lack of awareness and information in the medical community concerning this disease [18]. Eosinophilia of undetermined origin is often treated using empirical anthelmintic  treatment (generally a combination of albendazole or flubendazole and ivermectin) to cover a broad range of parasitic disease etiologies known in Martinique, including ascariasis, enterobiasis, strongyloidiasis, trichuriasis, and ankylostomiasis [14]. The combination of abdominal pain and hypereosinophilia should suggest potential HAA disease as for other well-known intestinal helminthiases, and clinicians should then seek histological or biological Panama [63]