A Comparative Study of Human Leptospirosis between Mayotte and Reunion Islands Highlights Distinct Clinical and Microbial Features Arising from Distinct Inter-Island Bacterial Ecology

ABSTRACT.

Leptospirosis is a neglected zoonosis for which investigations assessing host-pathogen interaction are scarce. The aim of this study was to compare the severity and bacterial species involved in human cases of leptospirosis on Reunion and Mayotte islands, territories located in the southwest Indian Ocean that have recorded high human leptospirosis incidence but display fairly distinct epidemiological situations. A retrospective multicentric study including all patients over 18 years of age from Mayotte or Reunion with proven leptospirosis was conducted from January 2018 to April 2020. This study collected demographic, geographical, clinical, and biological data. Overall, 490 patients were included, 222 on Mayotte and 268 on Reunion. More patients were hospitalized on Reunion (n = 215, 80%) compared with Mayotte (n = 102, 46%). Severe disease was more common on Reunion (n = 75, 28%) than on Mayotte (n = 22, 10%). The dominant Leptospira species on Reunion was Leptospira interrogans (79%) followed by Leptospira borgpetersenii (21%), contrasting with the epidemiological situation on Mayotte where L. interrogans was found in only a minority of patients (10%). The high frequency of severe cases on Reunion could be explained not only by higher comorbidities but also by the higher occurrence of L. interrogans infections compared with Mayotte. Finally, the distribution of cases linked to L. borgpetersenii was found almost exclusively on the west coast of Reunion, raising the potential role of a ruminant reservoir.

INTRODUCTION

It is estimated that leptospirosis affects more than one million people each year, causing nearly 60,000 deaths worldwide.¹ This neglected tropical disease, considered a priority public health issue by the WHO, is caused by helical-shaped pathogenic Leptospira belonging to the Spirocheta order.^2,3 Leptospira are characterized by an important genetic diversity and include environmental saprophytic species as well as intermediate/pathogenic species shed by animal reservoirs and involved in veterinary and human disease. The most common definitive hosts are small mammals, of which rodents are typically considered important reservoirs. Transmission to humans most often occurs through contact with contaminated fresh water or urine and is facilitated by warm and humid environments that might facilitate bacterial survival in the environment.² Symptoms are unspecific in benign forms and similar to those of dengue fever, but in some cases infection may progress to a severe form with multiorgan failure. The pathophysiology of progression to a severe form is not well understood but is probably linked to factors related to the host and infecting leptospiral species.^4,5

Mayotte and Reunion islands are two French overseas departments located in the southwestern Indian Ocean. Mayotte is an island of 374 km², part of the Comoros archipelago located northwest of Madagascar at the northern entrance of the Mozambique channel. Mayotte has a registered population of 256,500 inhabitants, with significant demographic growth (+3.8% per year).⁶ Reunion is located 942 km east of Madagascar and has an area of 2,512 km² and a population of 860,000 inhabitants. Both islands are tropical, with a climate alternating between a hot rainy season and a cooler dry season.^7,8 Leptospirosis has been present in the Indian Ocean for decades, and Mayotte displays one of the highest incidence rates in France.^9–13 Bacterial ecology differs across islands of this area with a dominance of the Leptospira interrogans serogroup Icterohaemorrhagiae within human cases on Reunion, whereas four distinct species have been reported on Mayotte, namely Leptospira borgpetersenii, Leptospira mayottensis, Leptospira Kirschneri, and L. interrogans.^14–16 When the epidemiology of human disease on both islands is compared, the scarce available data suggest a lower proportion of severe forms on Mayotte compared with Reunion.^5,17 The aim of the present study was to gather molecular and clinical data from human acute cases on Reunion and Mayotte to compare the diversity, geographic distribution, and associated symptoms in human cases of leptospirosis on both islands and also highlight distinctive features according to the infective genospecies.

MATERIALS AND METHODS

Study design and health facilities.

A retrospective, multicenter comparative study was conducted on both islands from January 1, 2018 to April 15, 2020. The healthcare system on Reunion is divided into primary care facilities and four hospital sites: the Center Hospitalo-Universitaire (CHU) de la Réunion (North and South sites), the Center Hospitalier Ouest Réunionnais and the Groupe Hospitalier Est Réunionnais. Each center is equipped with an emergency and intensive care unit (ICU). Mayotte is composed of two inhabited islands. The majority of the population is located on Grande Terre, which hosts the reference hospital: the Center Hospitalier de Mayotte (CHM). The other healthcare structures are located mainly in the coastal areas and include dispensaries, maternity wards, general medical practices and outlying hospitals. The CHM is the only structure equipped with an ICU and carrying out microbiological analyses.

Case screening and case definition.

Potential cases were screened through laboratory database and medical records over the study period. On Reunion, molecular and serological diagnoses of leptospirosis are centralized at the CHU. On Mayotte, all diagnoses are performed at the unique hospital (CHM). A case of clinically suspected leptospirosis was considered confirmed if positive through polymerase chain reaction (PCR) on blood or urine samples (homemade PCR including primers directed to 23S ribosomal RNA on Reunion Island and directed to 16S ribosomal RNA and LIPL gene on Mayotte), or through ELISA serodiagnosis (Serion ELISA Classic Kit on Reunion, Wurzburg, Germany, and Mayotte or homemade serology if performed by the National Reference Center).^18–20 All confirmed adult cases of leptospirosis from January 2018 through April 2020 were included in the study. Patients <18 years old or having another more likely diagnosis after review of the records were excluded, especially when biological assessment was supported only by serodiagnosis. Demographic, clinical, and biological data were collected from the medical records of the hospitals. Only initial biological and demographic data could be collected from patients consulting general practitioners (i.e., outpatients) because of the inaccessibility of medical records. To identify a potential source of contamination, demographic data collected from patients’ medical records were used. A professional item was checked when the patient was involved in a high-risk activity (veterinarian, sewage or forestry worker, farming), a leisure item in cases of freshwater leisure activities (canyoning, kayaking, rafting, and swimming), and a housing environment item when the presence of rats around the home was reported.

Outcomes and definitions.

The primary endpoint was the proportion of severe forms in hospitalized patients. A patient was considered hospitalized when remaining at the hospital for more than 24 hours. Severity was defined in hospitalized patients according to a composite endpoint score including at least one of the following criteria: use of a vasoactive drug, mechanical ventilation, renal replacement therapy (RRT), transfusion, or death. Secondary endpoints were the proportion of severe forms in all patients (both outpatients and hospitalized cases), the percentage of hospitalized patients, the identification of the bacterial species by molecular biology, and the proportion of major organ failures according to the characterized species.

Bacterial genotyping.

Plasma remaining from PCR or serodiagnosis was collected for molecular analysis to identify Leptospira at the species level (177/268 for Reunion Island and 59/222 for Mayotte were available for the analyses). A total of 200 µL of nucleic acids (RNA/DNA) was extracted using a QIACUBE robot and Cador Pathogen kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. Reverse transcription was performed on 10 µL of nucleic acids using the ProtoScript II Reverse Transcriptase and Random Primer 6 (New England BioLabs, Ipswich, MA) under the following thermal conditions: 70°C for 5 minutes, 25°C for 10 minutes, 42°C for 50 minutes, and 65°C for 20 minutes. Leptospira were genotyped using multilocus sequence typing (MLST) scheme#3 (https://pubmlst.org/organisms/leptospira-spp/) and the six following loci: secY, rrs2, lipL32, adk, lipL41, and icdA. This genotyping scheme was selected as it was previously used in a number of studies carried out in the region and hence allows comparing the genotypes with previously characterized Leptospira from distinct southwestern Indian Ocean islands.^12,21–23 For clinical samples of Reunion Island, only secY and rrs2 loci were sequenced, as Leptospira diversity is much narrower than on Mayotte (two species on Reunion versus four on Mayotte) and hence can be assessed through the use of the two aforementioned loci. Two microliters of complementary DNA were used as a template in a PCR containing 12.5 µL of GoTaq G2 Hot Start Green Master Mix (Promega, Madison, WI), 0.5 µL (10 µM) of each primer (reverse and forward), and 9.5 µL of RNAse-free water. The PCR conditions were as follows: a first GoTaq activation step (95°C for 2 minutes) was followed by 45 cycles of denaturation (94°C for 30 seconds) and annealing (58°C for 30 seconds) and extension steps (72°C for 1 minute). Polymerase chain reaction thermal cycling was terminated by a final extension step (72°C for 5 minutes). Resulting amplicons were further Sanger sequenced (Genoscreen, Lille, France) on both strands. Chromatograms were manually aligned and edited using Geneious (9.1.8). Allelic numbers were determined using the PubMLST (scheme#3) database.

Ethics statement.

This retrospective study used previous data and adhered to the French regulation MR-004 on clinical research (N°2206739). Within this framework, there was no need for formal notification; however a study information leaflet was provided to all patients. The study received a favorable opinion from the ethics committee of the Société de Pathologie Infectieuse de Langue Française (SPILF), N° CER-MIT 2022-0505.

STATISTICAL ANALYSES

Nonparametric continuous variables were described as medians and interquartile ranges, and differences were tested using the Wilcoxon–Mann-Whitney test. Gaussian variables were described with means and SDs, and differences were tested using the Student’s t-test. Categorical variables were described as numbers and percentages. They were compared using the χ² test or Fisher’s exact test, as appropriate. Statistical analyses were performed using SPSS software (IBM SPSS 23.0; Chicago, IL). All tests were two-tailed, and a P-value below 0.05 was considered statistically significant. Incidence was calculated from the 2017 and 2019 population census on Mayotte and Reunion Island, respectively.

Mapping generation.

Incidence data corresponding to the annual number of confirmed cases of leptospirosis on Mayotte and Reunion Island were reported over the study period per 100,000 inhabitants by place of residence. Incidence maps per year and municipalities were created using QGIS 3.20 software (open-source geospatial foundation OSGeo^®) based on the residence locations recorded in the files. Species distribution maps were also based on the residence recorded for patients for whom a bacterial species could be identified. They were generated using My Maps software (freely available at https://www.google.com/maps/d/).

Data accessibility.

Sequence data from Reunion were deposited in GenBank under the following accession numbers: OQ179955–OQ180048 and OQ230476–OQ230548. Sequence data from Mayotte are available under accession numbers OP776245–OP776260, OP782690–OP782699, OQ230549–OQ230575, OQ248114–OQ248138, and OQ269484–OQ269510.

RESULTS

A total of 490 cases were retained for the final analysis: 222 from Mayotte and 268 from Reunion (Figure 1). Samples from Mayotte included in the present study are detailed in a recent investigation aiming at enlightening the epidemiology of leptospirosis on this island.²⁰

Figure 1.

Flowchart of the study selection process. PCR = polymerase chain reaction.

Patients’ baseline and epidemiological characteristics. The mean age of patients on Mayotte and Reunion was 35 ± 12.1 and 48.3 ± 15.18 years, respectively (P <0.001). Ninety two percent of human acute cases were males on Reunion (n = 246) versus 69% on Mayotte (n = 153; P <0.001). Consumption of alcohol and tobacco was more frequent on Reunion (17%, n = 39 and 28%, n = 64) than on Mayotte (2%, n = 2 and 11%, n = 15), whereas obesity was more frequently found on Mayotte (15%, n = 19; P = 0.009). Risk exposure was found in 90% of cases on Reunion and 68% of cases on Mayotte (P <0.001). The most frequently suspected contamination exposures were occupational activities on Reunion (47%, n = 86) and home-related on Mayotte (59%, n = 23) (Table 1). Reported cases of leptospirosis were mainly distributed on the coasts on both islands. Areas with the highest number of cases were those with the highest rainfall records on both islands (Supplemental Figure 1). Finally, on Reunion, 80% of patients (n = 215) were hospitalized compared with 46% on Mayotte (n = 102; P <0.001).

Table 1.

Study population

Variables	Reunion (N = 268)	Mayotte (N = 222)	P-Value*
Male Sex, n (%)	246 (92)	153 (69)	<0.001
Age in Years, Mean ± SD	48.3 ± 15.8	35.0 ± 12.1	<0.001
Comorbidities, N (% of known values)
Tobacco Use	64/232 (28)	15/135 (11)	<0.001
Alcohol Abuse	39/232 (17)	2/132 (2)	<0.001
Diabetes	32/232 (14)	12/136 (9)	0.148
Obesity	14/230 (6)	19/131 (15)	0.009
Suspected Contamination Reported, n (%)	193 (90)	40 (68)	–
Professional	86 (47)	14 (36)	0.213
Leisure	60 (33)	9 (24)	0.269
Housing Environment	53 (29)	23 (59)	<0.001
Hospitalization,† n (%)	215 (80)	102 (46)	<0.001

^*Significant differences are bolded.

^† >1 Day.

Clinical and biological characteristics at admittance.

The most frequently reported symptom at admittance on Reunion and Mayotte was fever, 82% (n = 186) and 92% (n = 123), respectively, followed by muscle pain, 73% (n = 167) and 66% (n = 85), respectively. There were more neurological symptoms including headache, retro-orbital pain, and dizziness on Mayotte (49%, n = 63; P <0.001). Regarding the main biological parameters at admittance, the platelet count was lower in human acute cases from Reunion (107 × 10⁹/L ± 84) than from Mayotte (136 × 10⁹/L ± 64; P <0.001). Mean creatininemia levels were 238 µmol/L ± 222 and 147 µmol/L ± 249; P <0.001), and aspartate-amino transferase levels were 157 UI/L ± 328 (n = 206) and 81 UI/L ± 145 (n = 145); P <0.001, in Reunion and Mayotte acute cases, respectively. There was no difference in the initial levels of leukocytes, creatinin kinase, and C-reactive proteins (Table 2).

Table 2.

Clinical and biological features of all patients at initial presentation

Variables	Reunion (N = 268)	Mayotte (N = 222)	P-Value*
Clinical Features, n (% of known values)
Fever	186 (82)	123 (92)	0.006
Asthenia	136 (60)	92 (74)	0.008
Muscle Pain	167 (73)	85 (66)	0.220
Gastro-Enteric Symptoms	98 (43)	72 (57)	0.013
Respiratory Symptoms	28 (12)	20 (16)	0.359
Neurological Symptoms	66 (29)	63 (49)	<0.001
Biological Features, Mean ± SD (number of known values)
Total Bilirubin (µmol/L)	87 ± 111 (171)	41 ± 55 (105)	<0.001
AST (IU/L)	157 ± 328 (206)	81 ± 157 (145)	<0.001
ALT (IU/L)	75 ± 73 (193)	62 ± 63 (146)	0.060
Creatininemia (µmol/L)	238 ± 222 (227)	147 ± 249 (153)	<0.001
Platelets (× 109/L)	107 ± 84 (228)	136 ± 64 (153)	<0.001
Hemoglobin (g/dL)	13.2 ± 2.1 (228)	12.7 ± 2.1 (155)	0.016
Neutrophils (× 109/L)	9.6 ± 4.6 (228)	7.6 ± 3.9 (138)	<0.001
Lymphocytes (× 109/L)	0.90 ± 1.18 (227)	0.94 ± 0.71 (138)	0.728
CK (IU/L)	3,120 ± 6,805 (189)	1,999 ± 8,991 (29)	0.431
CRP (mg/L)	233 ± 141 (199)	220 ± 115 (146)	0.352

ALT = alanine-amino transferase; AST = aspartate-amino transferase; CK = creatine kinase; CRP = C-reactive protein.

^*Significant differences are bolded.

Features recorded during follow-up of hospitalized patients.

When all patients were considered regardless of their hospitalization status, 28% on Reunion displayed a severe form (n = 75) versus 10% on Mayotte (n = 22; P <0.001). When only hospitalized patients were considered, again more severe forms were recorded on Reunion (35%; n = 75) than on Mayotte (22%; n = 22; P = 0.014) (Table 3).

Table 3.

Primary endpoint: Rate of severe forms

Variables	Reunion	Mayotte	P-Value*
All Patients, N	268	222	–
Severe Forms, n (%)	75 (28)	22 (10)	<0.001
Among Hospitalized† Patients, n	215	102	–
Severe Forms, n (%)	75 (35)	22 (22)	0.014

RRT = renal replacement therapy. Severe forms include use of vasoactive drug, and/or use of mechanical ventilation, and/or use of RRT, and/or use of transfusion, and/or death.

^*Significant differences are bolded.

^† >1 Day.

Cumulative symptoms and biological values of organ evaluation during hospital stay of patients are presented in Table 4. Overall, patients from Reunion presented biological parameters reflecting more severe damage to organs such as kidneys and liver. Overall, patients from Reunion presented worst values indicative of major organ involvement, such as the kidney, liver, and platelets.

Table 4.

Clinical and biological features of patients during their entire hospital stay

Variables	Reunion (N = 215)	Mayotte (N = 102)	P-Value*
Clinical Features, n (% of known values)
Fever	103 (51)	59 (67)	0.011
Muscle Pain	96 (45)	34 (42)	0.610
Arthralgia	37 (17)	24 (32)	0.011
Digestive Tract Symptoms	153 (74)	32 (40)	<0.001
Jaundice	118 (56)	23 (27)	<0.001
Cough	20 (9)	9 (11)	0.731
Hemoptysis	30 (14)	4 (5)	0.013
Alveolar Hemorrhage†	18 (9)	1 (1)	0.009
Bleedings	66 (31)	16 (24)	0.277
Headache	36 (17)	35 (46)	<0.001
Meningitis	2 (1)	3 (3)	0.143
Biological Features, Mean ± SD (number of known values)
Total Bilirubin Max (µmol/L)	215 ± 438 (198)	134 ± 235 (90)	0.155
AST Max (UI/L)	345 ± 2179 (209)	185 ± 366 (77)	0.522
ALT Max (UI/L)	147 ± 398 (207)	124 ± 149 (76)	0.627
Creatininemia Max (µmol/L)	347 ± 286 (212)	267 ± 374 (78)	0.053
Proteinuria, N‡ (%)	61/170 (36)	13/94 (14)	<0.001
Hematuria, N‡ (%)	107/170 (63)	48/94 (51)	0.081
Platelets Min (× 109/L)	68 ± 65 (212)	105 ± 71 (90)	<0.001
Hemoglobin Min (g/dL)	10.5 ± 2.5 (212)	10.5 ± 2.6 (90)	0.972
Neutrophils Max (× 109/L)	12.1 ± 6.5 (199)	11.2 ± 7.9 (71)	0.388
Lymphocytes Min (× 109/L)	0.74 ± 0.93 (199)	1.03 ± 1.24 (83)	0.032
PR Min (%)	77 ± 18 (206)	71 ± 16 (83)	0.017
CK Max (UI/L)	3,809 ± 6,193 (156)	3,277 ± 11,027 (19)	0.750
CRP Max (mg/L)	316 ± 880 (177)	279 ± 107 (78)	0.712

ALT = alanine-amino transferase; AST = aspartate-amino transferase; CK = creatine kinase; CRP = C-reactive protein; Max = maximal value during hospital stay; Min = minimal value during hospital stay; PR = prothrombin ratio. Proteinuria = positive urinary stick or proteinuria >10 cells/mm³. Hematuria = positive urinary stick or hematuria >10 cells/mm³. Among patients admitted to the hospital, admission to the ICU was higher on Reunion (59%, n = 127) than on Mayotte (18%, n = 18; P <0.001) (Table 5).

^*Significant differences are bolded.

^† Alveolar hemorrhage is defined by the association of hemoptysis and radiological interstitial lung disease.

^‡ Calculation based on all urine samples collected.

Table 5.

Secondary endpoints analysis among hospitalized patients

Variables	Reunion (N = 215)	Mayotte (N = 102)	P-Value*
Delay Symptoms–Hospitalization in Days, Mean ± SD (n)	4.7 ± 2.4 (202)	4.1 ± 4.7 (83)	<0.001
Delay Symptoms–Antibiotics in Days, Mean ± SD (n)	4.6 ± 2.2 (199)	4.0 ± 4.2 (77)	<0.001
Length of Hospitalization in Days, Mean ± SD (n)	8.8 ± 6.5 (211)	5.8 ± 4.9 (90)	<0.001
Organ Impairment,† N (%)
Kidney	123 (57)	29 (28)	<0.001
Hepatic	103 (48)	20 (20)	<0.001
Respiratory	59 (27)	7 (7)	<0.001
Hemodynamic	69 (32)	20 (20)	0.018
Hematology	35 (16)	9 (9)	0.063
Cardiac	21 (10)	3 (3)	0.020
Neurology	10 (5)	2 (2)	0.350
Total Organ Impairment, Median [IQR]	2 [0–3]	0 [0–2]	<0.001
KDIGO Score ≥2, n (%)	123 (57)	29 (28)	<0.001
ICU Admission, n (%)	127 (59)	18 (18)	<0.001
IGS II Score, Median [IQR]	33 [22–46]	35 [20–58]	0.207
Oxygen Use	66 (31)	12 (12)	<0.001
Mechanical Ventilation, n (%)	23 (11)	5 (5)	0.075
ARDS, n (%)	13 (6)	0	0.012
RRT, n (%)	49 (23)	9 (9)	0.002
Transfusion	36 (17)	12 (12)	0.239
Intravenous Fluid Therapy	75 (35)	41 (40)	0.361
Vasoactive Drug, n (%)	37 (17)	10 (10)	0.074
Fatality, n (%)	4 (1.9)	2 (2.0)	–

ARDS = acute respiratory distress syndrome; ICU = intensive care unit; IGS = simplified gravity index (calculated according to the worst values obtained during the first 24 hours of admission in the ICU); IQR = interquartile range; KDIGO = kidney disease improving global outcomes (scale 0–3); RRT = renal replacement therapy.

^*Significant differences are bolded.

^† Definition of organ impairment: Kidney: KDIGO score ≥2 or the need for renal replacement therapy; Hepatic: total bilirubin ≥100 µmol/L; Respiratory: need for oxygen therapy or mechanical ventilation; Hemodynamic: intravenous fluid therapy, use of vasoactive drug; Hematology: need for labile blood product transfusion; Cardiac: myocarditis, acute heart failure; Neurology: Glasgow score <15 or meningitis.

The median simplified gravity index (IGS) II score was similar on both islands: 33 [22, 46] and 35 [20, 58], P = 0.207, whereas cumulative organ failure was higher on Reunion than on Mayotte: 2 ± [0, 3] (n = 158) and 0 ± [0, 2] (n = 42), respectively; P <0.001. On Reunion, 23 patients (11%) underwent mechanical ventilation, 49 (23%) received RRT, and 37 (17%) received vasoactive drugs, whereas four deaths were recorded. On Mayotte, five patients (5%) required assisted ventilation, nine (9%) RRT, and 10 (10%) vasoactive support, with two deaths occurring in the time frame of the study.

Bacterial diversity and ecology according to the islands.

For patients whose plasma was available, bacterial species identification was performed using MLST. Overall, 102 samples could be analyzed on Reunion, whereas the typing of 40 patients from Mayotte was included in the study (Rajaonarivelo et al.²⁴) from an availability of 177 and 59 samples, respectively (Figure 1). The occurrence of four Leptospira species on Mayotte was confirmed, the dominant species being L. borgpetersenii (n = 17, 43%) followed by L. mayottensis (n = 11, 28%), L. kirschneri (n = 8, 20%), and L. interrogans (n = 4, 10%) (Figure 2). The sample size was too small to determine exposure factors or geographic distribution related to the genospecies. Only L. interrogans (n = 2) and L. borgpetersenii (n = 3) were associated with severe disease on Mayotte, whereas no severe cases were reported among the 11 L. mayottensis– or the eight L. kirschneri–infected patients (Supplemental Tables 1 and 2).

Figure 2.

Genetic and geographic distribution of Leptospira species from (A) Mayotte and (B) Reunion patients (n = 142). L. borgpetersenii = Leptospira borgpetersenii; L. interrogans = Leptospira interrogans; L. kirschneri = Leptospira kirschneri; L. mayottensis = Leptospira mayottensis.

On Reunion, L. interrogans and L. borgpetersenii represented 80% (n = 81) and 20% (n = 21) of patients, respectively (Figure 2). Occupational exposure was reported in 63% (n =10) of L. borgpetersenii cases but was not reported for L. interrogans. The latter was involved in 29 severe cases, among whom three died, and L. borgpetersenii in two severe cases (Supplemental Tables 1 and 2).

Analysis of severity according to genospecies.

For both islands combined, the genospecies most frequently involved in severe disease was L. interrogans (n = 31) followed by L. borgpetersenii (n = 5; P = 0.002) (Figure 3; Supplemental Table 3). Among these patients, four died: three related to L. interrogans on Reunion and one related to L. borgpetersenii on Mayotte.

Figure 3.

Virulence of bacterial genospecies on both islands (n = 142). L. borgpetersenii = Leptospira borgpetersenii; L. interrogans = Leptospira interrogans; L. kirschneri = Leptospira kirschneri; L. mayottensis = Leptospira mayottensis.

Geographic structuration of pathogenic Leptospira.

A study carried out on Reunion in 2012–2013 reported L. borgpetersenii and L. interrogans within a cohort of 66 patients. As this study reported a limited diversity within L. interrogans, with ST02 and ST34 being the only sequences typing found among a cohort of 50 fully genotyped patients (see Table S2 in Guernier et al.²²), we used secY to address whether this low diversity was maintained over the 7 years separating both studies. secY1 and secY6 were used as markers of ST02 and ST34, respectively.²² We found secY1 and secY6 in 53 and three of the 102 typed patients, suggesting that ST02 and ST34 represented 65% and 4% of L. interrogans–infected patients (n = 81), respectively, within the study cohort. Interestingly, patients infected with L. interrogans came from different areas of the island, whereas 20 of the 21 patients infected with L. borgpetersenii were from the western or southern areas of the island (Figure 2).

DISCUSSION

This retrospective multicenter study, conducted over a 2-year period using samples collected in the same period and analyzed with common predefined criteria, confirms a higher severity of human leptospirosis on Reunion than on Mayotte in contrast to previous investigations.^5,17 Such a contrasted disease severity could be explained by a younger population on Mayotte, associated with reduced comorbidity. Indeed, the mean age of Mayotte’s inhabitants is almost 10 years lower than the mean age on Reunion (23 versus 33 years, respectively).^6,25 Distinct behaviors may also account for the presented pattern, as the majority of the population on Mayotte is Muslim; hence, alcohol consumption is less prevalent here than on Reunion. These three factors (age >46 years and alcohol and tobacco consumption) have been previously described as risk factors associated with severe forms of leptospirosis.^5,26,27 The ways of contamination also differed between islands. On Mayotte, exposure seemed to be related mainly to the immediate housing environment, which could be partly attributed to more precarious housing. The widespread use of wastewater or rainwater and less well-insulated homes in extreme weather events (flooding) may have contributed to the greater risk of contamination. On Reunion, the main exposure observed was occupational, mostly related to sugarcane farming. The differences in patient sex ratios across these two islands may also be explained by these differences in contamination pathways.

Clinical presentation on both islands did not differ from what was already known. Interestingly, headache was more frequent in the initial presentation of Mayotte patients, but was not associated with meningitis. In contrast, clinical and biological jaundice was more frequent among patients on Reunion. This may suggest selective tropisms of the species involved. Although L. interrogans is known to destroy hepatocyte junctions, neurological damage in leptospirosis remains unclear, whether as a direct bacterial lesion or a reactionary process.^28–30 Of the five cases of meningitis in our study, only one PCR was performed on cerebrospinal fluid, and the result was positive.

Our study also showed that a higher rate of leptospirosis patients were hospitalized on Reunion than on Mayotte (80% versus 46%, respectively). This could be explained by a higher number of severe forms recorded on Reunion but may also be related to Reunion’s greater downstream bed capacity. As of 2019, the number of downstream beds per inhabitant was 1.7 times higher on Reunion than on Mayotte.³¹ Similarly, as of January 1, 2020, Mayotte reported a mean of 7.2 ICU beds per 100,000 residents compared with 24.1 beds/100,000 residents on Reunion.³² These considerations could explain the similarities observed in the IGS II score in these two territories, reflecting the admission of patients in the most critical states on Mayotte. The admission rate to the ICU among hospitalized patients averaged 60% on Reunion, which was higher than previously described by Paganin et al.²⁶ (54.3%) and Delmas et al.⁵ (31.5%). This may reflect different practices between centers, particularly the easier use of continuous monitoring. On Mayotte, however, the admission rate to ICUs for all patients was relatively stable (8.1%) compared with the descriptions of Tantet et al.¹⁷ (5.5%).

Organ involvement on both islands was similar to that described in previous studies.^5,17 Kidney failure remained the main cause of severe leptospirosis on both islands. The occurrence of acute respiratory distress syndrome (ARDS) in patients from Reunion only could result from a strong involvement of L. interrogans in respiratory damage: Seven of 13 patients with ARDS were infected with L. interrogans, whereas no samples were available for species determination of the six other cases. Our results confirmed that the use of mechanical ventilation remains infrequent on Mayotte, as previously reported, whereas the use of vasoactive drugs is comparable to use in previous reports.^5,17 Overall, the fatality rate was low on both islands: 1.9% on Reunion and 2% on Mayotte and appears slightly lower than previously reported (Delmas et al.⁵ and Tantet et al.¹⁷), although the study populations differed. This possibly resulted from a better knowledge of the disease and hence improved preparedness of the healthcare system. The delay for antibiotic initiation was shorter on Mayotte, with a mean difference of 0.6 day. Whether this difference could partly explain a lower rate of disease severity is unknown, especially because the results of studies exploring the importance of delays in antibiotic treatment have conflicted.^27,33

Lastly, this comparative work highlighted the possible strong involvement of the infecting leptospiral species in the outcome of the disease. The increased severity of the disease on Reunion could be explained by the widespread circulation of L. interrogans on this territory (80%) compared with Mayotte (10%). Similarly, most of the severe forms were related to L. interrogans on Mayotte, although one of the two deaths was caused by L. borgpetersenii (Supplemental Tables 1–3). Although it is classically described that all pathogenic leptospires are virulent, phylogenomic analyses allowed classifying them into four subgroups according to their genetic material, coding for virulence-related proteins. Subgroup 1, characterized by a large copy number of virulence genes, is mainly represented by L. interrogans serovar Icterohaemorrhagiae and Copenhageni and by L. kirschneri. Subgroup 2 genospecies, including L. borgpetersenii and L. mayottensis, has a shorter genome with possibly fewer virulence genes.³⁴ Studies in mice described L. interrogans as one of the most virulent species, with a high load found in renal biopsies.³⁵ Finally, few studies have investigated the association between species and severity in human cases. A study carried out in New Caledonia between 2008 and 2011 showed that infection with L. interrogans serovar Icterohaemorrhagiae, which is the most common serovar found on Reunion, was significantly associated with severe cases.²⁷ Similarly in New Zealand, L. interrogans serovars Copenhageni and Pomona were reported to generate more hospitalizations than L. borgpetersenii serovar Ballum.³⁶ Recent data from Rajaonarivelo et al.²⁴ highlight the important role played by the leptospiral genome, especially the involvement of L. interrogans in severe forms.

Finally, our data revealed a geographic structuration of L. borgpetersenii in the western and southern portions of Reunion, whereas L. interrogans appeared evenly distributed. Among the 21 cases of L. borgpetersenii, occupational exposure was found in 10 cases. These regions shelter most of the cattle production on the island, and L. borgpetersenii was previously detected from cattle.²² Although the distribution of L. interrogans genotypes ST02 and ST34 on Reunion appears to be stable over time (no statistical difference compared with data from Dietrich et al. study), we observed a significant decrease in the circulating L. interrogans-to-L. borgpetersenii ratio over the same period, thus suggesting that L. borgpetersenii is playing an increasing role in human leptospirosis on Reunion.²¹

To date and to our knowledge, this is the first study comparing human cases of leptospirosis between two island territories. We therefore took the opportunity to study some unique features of the epidemiology of leptospirosis in the French Islands of southwest Indian Ocean to address the importance of the infecting species in the development of severe symptoms in human disease. These French-administrated islands are particularly relevant to tackle this question as 1) they host four distinct species of pathogenic Leptospira (four on Mayotte and two on Reunion), and 2) they have displayed distinct leptospirosis epidemiological features while under the control of the same health surveillance authorities, hence limiting possible biases.

This survey showed more severe forms on Reunion, possibly related to an older, comorbid population. However, this difference can also be explained by the higher occurrence of L. interrogans on Reunion. The increasing role of L. borgpetersenii in the local epidemiology of human leptospirosis on Reunion as well as the sharp geographic structuration of human cases infected with this species underlined by our data strengthen the need for identifying the reservoir(s) of L. borgpetersenii, including through the exploration of ruminants. Presented data and proposed perspectives are bringing invaluable information toward a comprehensive understanding of human leptospirosis on these two islands, where leptospirosis is approximately 20 times more prevalent than on continental France and hence is considered a major medical concern.^22,37

Note: Supplemental materials appear at www.ajtmh.org.