Reassessing the distribution of Burkholderia pseudomallei outside known endemic areas using animal serological screening combined with environmental surveys: The case of Les Saintes (Guadeloupe) and French Guiana

Abstract

Background: Melioidosis, an emerging infectious disease that affects both humans and animals, is caused by the soil-dwelling bacterium Burkholderia pseudomallei. It is endemic in South and Southeast Asia, and northern Australia, causing an estimated 165,000 human cases annually worldwide. Human cases have been reported in the French West Indies (Martinique and Guadeloupe) since the 1990s. Conversely, no human cases have been reported in French Guiana, a French territory in South America. Our study aimed to investigate whether B. pseudomallei is locally established in Guadeloupe and French Guiana using animals as a proxy.

Methodology/principal findings: Blood samples were collected from different animals from 56 farms in French Guiana (n = 670) and from two goat farms in Les Saintes (n = 31), part of the Guadeloupe archipelago and tested by enzyme-linked immunosorbent assay (ELISA). In Les Saintes, a serological follow-up was performed, and soil, water and goat rectal swabs were collected and analyzed by culture and PCR. The highest seroprevalence rates (39%) were observed in goats in Les Saintes, followed by horses (24%) and cattle (16%) in French Guiana. In the two goat farms, supplementary analyses detected B. pseudomallei from one goat rectal swab, and a B. pseudomallei strain was isolated from the soil.

Conclusions/significance: Our animal serological data suggest the presence of B. pseudomallei in Les Saintes and French Guiana. In Les Saintes, environmental surveys confirmed the endemicity of the bacteria, which is consistent with documented human cases of melioidosis on the island. We did not conduct an environmental survey in French Guiana. Nevertheless, our serological results call for local environmental surveys and a retrospective reassessment of human infections with melioidosis-like symptoms.

Fig 1. Map of sampling sites.

Blue dots represent farms where animals were tested in (A) Les Saintes (n = 2) and (B) French Guiana (n = 56). Les Saintes is part of the Guadeloupe archipelago, and consists of nine islands, of which only the two largest are inhabited, Terre-De-Haut and Terre-De-Bas; only Terre-De-Haut was sampled. The map layers (country boundaries) of South America, Martinique and French Guiana were downloaded from the geoBoundaries website https://www.geoboundaries.org/simplifiedDownloads.html).

Fig 2. Timeline of sampling at the two goat farms in Les Saintes (November 2021—March 2023).

November 2021 corresponds to the initial sampling. All other samplings correspond to the follow up study. The longitudinal studies, multiple sampling of the same individuals, occurred between August 2022 and March 2023 for farm A, and between August 2022 and December 2022 for farm B, after the animals were microchipped. The numbers in parentheses correspond to the number of samples collected from captured animals and the environment, on farms A and B.

Fig 3. Summary of GLANDA ELISA results per farm and species in French Guiana (A: cattle (n = 37), B: goats (n = 10), C: sheep (n = 15), and D: horses (n = 6)) and in Les Saintes (E: goats (n = 2)).

S/P%: Serological titer expressed as the ratio of the sample to the positive control. The x-axis of the graph represents farm identification numbers ranging from 1 to 56. Each farm is assigned a unique number, and if multiple species were present on the same farm, the same number is used. The total number of animals tested per farm is shown in parentheses below each farm number. The red dotted line represents the S/P% threshold (70%). The bar above and below the box plot indicates the minimum and maximum S/P%. The top and bottom of the box indicate the 1^st and 3^rd quartiles of S/P%. The bars inside the box indicate the median. Due to only one pig farm being tested, we have excluded these data from the figure.

Fig 4. Cultures of Burkholderia pseudomallei strain 22–10884_313#20.

Photographs of culture plates (A) on Ashdown agar and (B) on modified B. cepacia CHROMagar agar (supplemented with 4% glycerol, 500 mg/mL gentamicin and 130 mg/mL fosfomycin) after a 24, 48 and 120 h incubation at 37°C. The strain 22–10884_313#20 was isolated from soil 22–10884_313 collected from farm B in Les Saintes.

Fig 5. MLST-based phylogenetic tree of 24 B. pseudomallei strains from South America, including the strain 22–10884_313#20 isolated in this study.

The phylogenetic tree was constructed from 24 South American strains: 23 from the PubMLST database and the strain isolated in this study (22–10884_313#20) indicated by a red arrow, which was isolated from a goat farm in Les Saintes. The MLST tree was a maximum parsimony tree with bootstrapping (n = 100). The numbers on the branches correspond to the bootstrap values.