Etiology of Acute Febrile Illnesses in Adults in the Defense Community in French Guiana

ABSTRACT.

In tropical countries, acute febrile illnesses represent a complex clinical problem for general practitioners. We describe the prevalence of different etiologies of acute febrile illnesses occurring among French service members and their families, excluding children, in general practice in French Guiana. From June 2017 to March 2020, patients with a fever ≥37.8°C with a duration of less than 15 days who sought medical care at the army medical centers in Cayenne and Kourou were prospectively enrolled. Based on clinical presentation, blood, urine, nasopharyngeal, and stool samples were collected for diagnostic testing for viruses, bacteria, and parasites (by direct examination, microscopic examination of blood smears, culture, serology, or polymerase chain reaction), and standardized biological tests were systematically performed. Among 175 patients retained for analysis, fever with nonspecific symptoms was predominant (46.9%), with 10 Plasmodium vivax malaria cases, 8 dengue infections, and 6 cases of Q fever. The second most frequent cause of acute febrile illness was upper respiratory tract infections (32.0%) due to influenza virus (n = 18) or human rhinovirus (n = 10). Among the causes of acute febrile illness in French Guiana, clinicians should first consider arboviruses and malaria, as well as Q fever in cases of elevated C-reactive protein with nonspecific symptoms and influenza in cases of signs and symptoms associated with upper respiratory tract infections. Despite an expanded microbiological search, the etiology of 51.4% of acute febrile illnesses remain unknown. Further investigations will be necessary to identify the etiology of acute febrile illnesses, including new pathogens, in French Guiana.

INTRODUCTION

Febrile diseases are one of the most frequent reasons for medical consultation in low- and middle-income countries and are often caused by infectious diseases, including malaria, various viral and bacterial infections, and zoonoses.^1–4 Febrile illnesses are also a complex medical problem for clinicians because of the broad differential diagnosis, nonspecific characteristics of most tropical febrile illnesses, and risk of severe infections.⁵

French Guiana (FG) is an overseas French territory located on the northeastern coast of South America, 90% of which is covered by Amazonian rainforest, and populated by 294,436 inhabitants in 2022.⁶ In FG, malaria is endemic, with transmission occurring in some hotspots along the river banks bordering Suriname to the west and Brazil to the east, as well as in remote forest areas where illegal gold miners constitute a hard to reach reservoir.^7–13 Other parasitic and fungal infections have been associated with acute febrile illnesses (AFIs) in FG, in particular toxoplasmosis due to atypical strains that could lead to severe forms even in immunocompetent persons and histoplasmosis, which can affect not only HIV-infected patients but also immunocompetent individuals.^14–18 Several arboviruses are endemic or have recently emerged in FG. Yellow fever is rare but is still reported because of incomplete vaccination coverage in at-risk populations.¹⁹ Dengue viruses (DENVs) serotypes 1 to 4 are endemo-epidemic, with regular outbreaks occurring on a 3- to 5-year cycle.²⁰ Chikungunya virus (CHIKV) and Zika virus (ZIKV) emerged in 2014–2015 but have not been reported in FG since 2016.^21,22 Mayaro virus (MAYV) and Tonate virus (TONV) infections have also been reported, with clinical pictures similar to those of CHIKV and DENV infections, respectively.^23,24

In addition to arboviruses, influenza, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), etiologic agent of coronavirus disease 2019 [COVID-19]), and other respiratory viruses are known to be common viral causes of AFIs in FG, where transmission occurs mainly during the rainy seasons (i.e., April to June and January to February).²⁵ Moreover, the incidence of Q fever is highest in the world (27.4 cases per 100,000 inhabitants per year).^26–29 About 24% of community-acquired pneumonia is due to Q fever, mostly attributable to Coxiella burnetii genotype MST17, which has only been described in FG so far.^26–29 A retrospective study on leptospirosis conducted from 2007 to 2014 suggested that only 20% of symptomatic patients were suspected of being infected by Leptospira spp. on hospital admission, probably leading to substantial underestimation of this disease.³⁰

About 4,500 French service members (FSMs) are present in FG. The FSMs live along the malaria-free coastal strip and are therefore exposed to the same health risks as the general population. They are also deployed along the rivers bordering Suriname and Brazil and to illegal gold mining sites in the rainforest where they are exposed to the risk of contracting malaria, helminthiasis, and enteritis, among others.³¹ In addition to individual and collective risks in the event of an epidemic, they can import pathogens from the forest to the coastal area or even be the source of the emergence of a new pathogen.³² To reduce these risks, soldiers receive specific vaccinations against hepatitis A, typhoid fever, Neisseria menigitidis serogroups A, C, W, and Y (meningococcal conjugate vaccine MenACWY), influenza, SARS-CoV-2, and yellow fever, in addition to the standard vaccinations included in the French national immunization schedule for adults (measles, mumps, rubella, diphtheria, tetanus, poliomyelitis, whooping cough, and hepatitis B). They are also instructed to take chemoprophylaxis against malaria (doxycycline) during their mission and for 4 weeks after returning from missions in endemic areas. Family members come in close contact with the returning military personnel and may be exposed to potential risks of communicable diseases. They receive the recommended vaccinations for travelers, including yellow fever vaccination.

French service members (n = 4,561), civilian employees of the French armed forces (n = 900), and their families (n = 2,700 adults) are at the interface between the forest and coastal areas of FG and thus serve as a sentinel population for infectious diseases in FG. This population represents about 8,000 individuals. In an earlier study conducted at the emergency department of Cayenne Hospital, it was found that 50% of AFIs were not confirmed by microbiological tests.³³ Furthermore, to date there is no description of the causes of AFIs in general practice in FG. The aim of the present study was to describe the etiologies of AFIs in the French defense community (FSMs and their families, excluding children) in the general practice setting in FG to improve diagnosis and management of AFIs.

MATERIALS AND METHODS

Study site.

The clinical research project on the etiologies of AFIs (2FAG study) was designed at the army medical centers (AMCs) in Cayenne and Kourou. The personnel of the AMCs consisted of general practitioners and paramedics providing healthcare to members of the defense community (FSMs, their families, and civilian employees of the French armed forces) residing in different places along the coastal areas of FG.

Inclusion criteria.

Adults aged 18 years and over were eligible for enrollment if they provided a written informed consent to participate in the study and met the following case definition of AFI: documented fever ≥37.8°C at the time of presentation to the AMC or a history of fever beginning less than 15 days before medical consultation.

Exclusion criteria.

Patients were excluded in cases of incomplete standardized data collection form; absence of first laboratory screening tests; and absence during the scheduled second visit for physical examination or lack of second biological screening in patients without a laboratory-confirmed diagnosis after the first round of screening (i.e., patients lost to follow-up).

Study protocol.

From June 2017 to March 2020, patients seeking medical care in the AMCs in Cayenne and Kourou who met the case definition of AFIs were prospectively enrolled in the 2FAG study (Figure 1). After physical examination, general practitioners explained the purpose of the clinical research project and, in cases of consent to participate, patients signed a written informed consent. Participants were interviewed to collect data that included sociodemographic and clinical information, signs and symptoms at onset, and dates of specimen collection. The information was recorded in a standardized data collection form. Blood samples were collected systematically to screen for the following viruses, bacteria, and parasites: DENV, ZIKV, CHIKV, TONV, MAYV, cytomegalovirus (CMV), Epstein-Barr virus (EBV), HIV; Leptospira spp. and C. burnetii; and Toxoplasma gondii and Plasmodium spp., respectively. Routine hematological and biochemical tests were performed, including complete blood count with differentials, glucose, blood urea nitrogen and creatinine, blood electrolytes, total bilirubinemia, total blood protein, inflammatory proteins (C-reactive protein [CRP], procalcitonin), erythrocyte sedimentation rate, liver function tests, and coagulation tests. A urine sample was collected for cytobacteriological testing, including culture. In addition, a rapid diagnostic test for group A Streptococcus was performed in patients with a sore throat (Streptatest; Biosynex, Illkirch-Graffenstaden, France).

Figure 1.

Flow chart of patients recruited in the 2FAG study, French Guiana 2017–2020.

If a patient presented with respiratory symptoms, serological tests for C. burnetii, Legionella pneumophila, Chlamydia pneumoniae, and Mycoplasma pneumoniae were performed, and a nasopharyngeal sample was collected to perform reverse transcriptase quantitative polymerase chain reaction (RT-qPCR) to test for influenza virus, respiratory syncytial virus (RSV), and human rhinovirus (HRV). If the patient had digestive symptoms, a stool sample was collected to evaluate the presence of gastrointestinal bacteria and parasites. At least 15 days after inclusion, another physical examination and a second series of standardized laboratory examinations, including serological tests, were performed. Clinical and laboratory data obtained on day 15 (D15) were used to evaluate the clinical evolution of the patient, and the results of the follow-up blood tests were used to assess the evolution of biological parameters. In patients whose diagnosis was based on serology, D15 serological tests also allowed the demonstration of IgM seroconversion.

For all patients, the etiological diagnosis of AFIs was established by a medical staff consisting of general practitioners who examined the patients, a medical biologist, and two specialists in infectious and tropical diseases.

Laboratory testing.

Nasopharyngeal swabs were used to detect the presence of RSV (Supplemental Table 1), HRV, and influenza viruses by RT-qPCR assays at the National Reference Center for Respiratory Viruses at Pasteur Institute of French Guiana).^34,35 Serum samples were tested for DENV, ZIKV (Supplemental Table 1), CHIKV, TONV, and MAYV by specific RT-qPCR and virus-specific IgM antibody-capture ELISA at the National Reference Center for Arboviruses at the Pasteur Institute of FG.^36–40 Serological screening for toxoplasmosis, leptospirosis, Q fever, Legionella spp., Chlamydia spp., Mycoplasma spp., EBV, CMV, and HIV as well as standardized routine laboratory examinations were performed at two medical biology laboratories in Cayenne and Kourou. Polymerase chain reaction was also performed to detect the presence of Leptospira spp.

Definitions.

Acute fever was defined as body or axillary temperature ≥37.8°C at the time of consultation at the AMC or history of a fever less than 15 days before consultation.

A laboratory-confirmed case of DENV, ZIKV, CHIKV, TONV or MAYV infection was defined as a patient with DENV, ZIKV, CHIKV, TONV, or MAYV RNA detected in the serum by RT-qPCR.

Laboratory-confirmed cases infected with influenza virus, HRV, or RSV were defined as patients with a positive RT-qPCR assay.

Laboratory-confirmed cases with EBV, CMV, HIV, or toxoplasmosis were defined as follows: anti–Viral capsid antigen EBV IgM positive, anti-CMV IgM index >1, anti-HIV1/2 IgM positive, or anti-Toxoplasma IgM seroconversion (IgM index ≥0.6), respectively, determined by comparing titers in samples from acute and convalescent phases.

A laboratory-confirmed leptospirosis case was defined based on the following criteria: (1) positive PCR for Leptospira spp. or (2) IgM seroconversion for Leptospira spp. based on titers in acute and convalescent phase samples (IgM >19UA).

A laboratory-confirmed C. burnetii infection (Q fever) was defined as follows: phase II antibody titers that were consistent with acute Q fever (IgG II >200 and IgM II >50) or with increased phase I antibody titers (IgG I >800) and seroconversion with a 4-fold increase of phase II antibodies.

A positive test result for malaria was defined as the detection of malaria parasites by microscopic examination of thin and/or thick blood smears under oil immersion at a magnification of 1,000×.

White blood cells (WBC) counts were classified as leukopenia (total WBC <5,000/mm³) or leukocytosis (WBC > 11,000/mm³). The lymphocyte and neutrophil counts were classified as lymphopenia (<1,000/mm³) and lymphocytosis (>7,500/mm³), respectively. Platelet counts were classified as thrombocytopenia (total platelet count <150,000/mm³) or thrombocythemia (total platelet count >450,000/mm³). The intermediate values of WBC counts, lymphocyte counts, neutrophil counts, and platelet counts were within the range considered normal for adults. C-reactive protein was considered elevated if >5 mg/L. Anemia was defined as hemoglobin (Hb) <13 g/dL for men and Hb <12 g/dL for women.

Data collection.

All information recorded in the standardized data collection form and laboratory results were entered into an Excel spreadsheet with double input (Microsoft Office Excel 2016, Microsoft Corporation, Redmond, WA). Descriptive statistical analyses were likewise performed using Excel software.

RESULTS

A total of 262 patients were initially enrolled. Of 262 patients, 43 (16.4%) were excluded after the first visit and 44 (16.8%) were lost to follow-up. A total of 175 (66.8%) patients were included for analysis (Figure 1). The median age was 32 years (interquartile range [IQR], 26–39 years; range, 19–61 years). All patients but one were FSMs (99.4%), and one (0.6%) was a civilian employee of the French army. Most of the patients were men (148/175, 84.6%), with a sex ratio of 5.5. The median of delay between the onset of symptoms and the first clinical examination was 2 days (IQR, 1–3.75 days; range, 0–14 days).

Various potential sources of exposure to infectious agents during the 30 days prior to medical consultation were reported by the patients. These included patients who had contact with another patient known to have contracted a viral, bacterial, or parasitic disease (16%) or contact with an animal (24.6%), insect bites (74.4%), dust exposure (71%), freshwater bathing (42.3%), drinking untreated water (26.9%), and eating food prepared or stored in unhygienic conditions (19.4%) or consuming wild game meat (8.6%).

Many patients reported high mobility during the 30 days prior to consultation. Many stayed near rivers and streams (32%), went into the forest (45.1%), searched for illegal gold mines as part of their mission (25.7%), returned to metropolitan France (12.6%), or visited Brazil (6.9%), Suriname (6.3%), or the French West Indies (4.6%). Among FSMs (N = 60, 34.3%) who traveled to a malaria endemic area, 60% did not take the complete chemoprophylactic scheme.

Ninety patients (51.4%) did not have a confirmed etiological diagnosis despite complete screening (Table 1). Fever with nonspecific symptoms was predominant, representing 46.9% (82/175) of the analyzed cases. An etiological diagnosis was established for 33 of them: 10 (5.7%) cases of Plasmodium vivax malaria, 8 (4.6%) DENV infections, 6 (3.5%) cases of Q fever, 2 (1.1%) patients with leptospirosis, 2 (1.1%) with acute Amazonian toxoplasmosis, and 1 case (0.6%) each of CMV, EBV, and TONV infections. Pericarditis associated with Mycobacterium tuberculosis and pulmonary embolism (a noninfectious cause of fever) were also diagnosed. The second most common cause of fever (56/175; 32%) was upper respiratory tract infections due to influenza virus (10.3%) and HRV (5.7%). Of 56 nasopharyngeal samples collected in patients with symptoms associated with upper respiratory tract infections, 19 were positive for influenza virus (11 influenza A H3N2, 4 influenza A H1N1, and 4 influenza B viruses [3 Victoria strains and 1 Yamagata strain]), and 11 were positive for HRV. Among these patients, one presented with influenza A H3N2–HRV coinfection.

Table 1.

Final diagnosis of the study population, 2FAG study, French Guiana, 2017–2020

Diagnosis	No. (%) of Cases
Fever with nonspecific symptoms	82 (46.9)
Unknown	49 (28)
Plasmodium vivax malaria	10 (5.7)
Dengue virus	8 (4.6)
Coxiella burnetii	6 (3.4)
Leptospira spp.	2 (1.1)
Toxoplasma gondii	2 (1.1)
Cytomegalovirus	1 (0.6)
Epstein-Barr virus	1 (0.6)
Tonate virus	1 (0.6)
Mycobacterium tuberculosis *	1 (0.6)
Pulmonary embolism	1 (0.6)
Upper respiratory tract infections	56 (32)
Unknown	23 (13.1)
Influenza virus	18 (10.3)
Rhinovirus	10 (5.7)
Streptococcus A	2 (1.1)
Klebsiella pneumoniae	1 (0.6)
Epstein-Barr virus	1 (0.6)
Influenza virus + rhinovirus	1 (0.6)
Gastroenteritis	25 (14.3)
Unknown	16 (9.2)
Shigella spp.	4 (2.3)
Salmonella spp.	2 (1.1)
Campylobacter spp.	2 (1.1)
Leptospira spp.	1 (0.6)
Urinary tract infections	6 (3.4)
Escherichia coli	4 (2.3)
Klebsiella pneumoniae	2 (1.1)
Lower respiratory tract infections	4 (2.3)
Unknown	2 (1.1)
C. burnetii	2 (1.2)
Skin and soft tissue infections†	2 (1.1)

^*Pericarditis due to Mycobacterium tuberculosis.

^† Erysipelas due to group A Streptococcus.

Gastroenteritis represented 14.3% (25/175) of the AFIs with three different bacteria found in stool samples: Shigella spp. (2.3%), Salmonella spp. (1.1%), and Campylobacter spp. (1.1%). Six cases of urinary tract infections were diagnosed on urine cultures: four associated with the presence of Escherichia coli (2.3%) and two with Klebsiella pneumoniae (1.1%). Four patients developed a lower respiratory tract infection (pneumonia), of whom two had Q fever. Two cases of erysipelas were also clinically diagnosed.

All DENV infections were benign, with positive DENV RT-qPCR results in all cases. There were two DENV1 serotype and six DENV2 serotype. The main symptoms and laboratory findings in 175 patients are presented in Supplemental Tables 2 and 3.

DISCUSSION

The purpose of this study was to identify the main etiology of AFIs in general practice in the adult population of the defense community in FG. It highlights that the clinical presentation is most often nonspecific (46.9%), which represents a diagnostic difficulty for clinicians. In the present study, P. vivax malaria, dengue, and Q fever were most frequently diagnosed and microbiologically confirmed, accounting for 13.7% of all AFIs. The second most common clinical presentation (32%) was upper respiratory tract infections associated with influenza virus or HRV, representing 16% of all AFIs included in this study. Gastroenteritis was the third most frequent clinical presentation (14.3%), with the presence of bacterial infection in the stool samples of nine patients (5.1%).

Our results are in agreement with previous studies conducted in Central and South America, where DENV, CHIKV, and ZIKV were the most frequently identified etiologic agents of AFIs between 2012 and 2018.^3,33,41,42 Other arboviruses have been reported in western South America. For example, Venezuelan equine encephalitis virus (VEEV), MAYV, and Oropouche virus (OROV) were observed in 3% of AFIs.⁴³ In two hospitals in the Ecuadorean Amazon basin, Manock et al.⁴² identified pathogens responsible for fever in only 40.1% of cases: Leptospira spp. (13.2%), malaria (12.5%), Rickettsia spp. (5.9%), DENV (5.3%), Q fever (4.9%), Brucella spp. (1.3%), VEEV, OROV, and St. Louis encephalitis virus infections (<1%).

During our study period, malaria was still a significant cause of AFIs in FSMs owing to their mission to investigate illegal gold mining areas. However, the prevalence of malaria has considerably declined over the past decade in FG because of several interventions, such as active detection and treatment of asymptomatic parasite carriers and a specific program targeting illegal gold miners with the distribution of self-diagnosis kits and treatments.^44–46 In FG, DENV is responsible for large, cyclic, and prolonged epidemics in the coastal areas where FSMs reside.^20,22 During the study period, ZIKV and CHIKV were not circulating in the department, but a DENV outbreak occurred between 2019 and 2020, which explains our findings concerning eight DENV-infected patients.²² One of our patients was infected with TONV and presented with clinical signs and symptoms similar to a DENV infection.²⁴

Concerning eight cases of Q fever, only two of them presented with a community-acquired pneumonia. This contrasts sharply with the classical description of a high proportion (nearly 80%) of pulmonary involvement in Q fever in FG, probably owing to the use of doxycycline as antimalarial chemoprophylaxis in four to six patients without pneumopathy.²⁶ Although most individuals with Q fever recover without any treatment, doxycycline treatment is recommended.²⁸ Furthermore, in FG, it is likely that if a chest computed tomography (CT) scan were performed on all acute cases of Q fever, nearly 100% of parenchymal lung involvement would be expected. Unfortunately, CT scans were not performed in all of our cases presenting with respiratory symptoms, and the standard chest x-ray does not satisfactorily rule out moderate parenchymal involvement.

Compared with a previous study conducted in South America,⁴² fewer cases of leptospirosis were diagnosed in our cohort, and severe leptospiral infections were not observed. We can again hypothesize that doxycycline (100 mg per day for malaria chemoprophylaxis) also played a protective role against Leptospira spp., as Takafuji et al.⁴⁷ had shown a 95% reduction in leptospirosis in a randomized, double-blind, placebo-controlled field trial using doxycycline 200 mg per week.

Brucellosis and rickettsiosis were not observed in the present study, although they have been reported elsewhere in South America.⁴² The first three cases of Brucella spp. infections that have been reported in FG, including a new species of Brucella, involved illegal gold miners.⁴⁸ In FG, there is no intensive livestock farming, and we did not expect FSMs to be exposed to brucellosis. Recently, a survey of ticks was conducted in FG.⁴⁹ Of 819 ticks, 252 (30.8%) were positive for Rickettsia infection. Nineteen new genotypes were identified with no evidence of their pathogenicity, and two other species, R. amblyommatis and R. bellii, have never been found in vertebrate hosts, suggesting that they are nonpathogenic.

For AFIs without specific symptoms, laboratory findings and epidemiological and clinical context can help guide clinicians to establish the etiologic diagnosis. During or after travel to a malaria-endemic area, lymphopenia, thrombocytopenia, anemia, and elevated CRP may suggest malaria, but microscopic examination of blood smears (and/or rapid diagnosis test for malaria) is always necessary to formally exclude malaria infection.^50,51 Depending on the epidemiological situation, DENV infection may be suspected if there is lymphopenia and/or thrombocytopenia and a low CRP value.^41,51,52 An elevated CRP value (>50 mg/mL) may point to Q fever, leptospirosis, or any other bacterial infection.^53,54

In patients with upper respiratory tract infections, influenza virus (19/56, 33.9%) and HRV (11/56, 19.6%) were the most frequently detected pathogens in nasopharyngeal samples. In a previous study, Mahamat et al.²⁵ found a similar prevalence of influenza virus (40%), with a majority of influenza A (60%) (H3N2 and H1N1) and a minority of influenza B (40%) and with marked seasonality, as most of these viral infections occurred during both the short rainy season (January and February) and the long rainy season (from April to June). Influenza vaccination is administered every 2 years in FSMs (instead of the standard practice to administer it every year), with incomplete protection against influenza circulation in this population.

In another study conducted in Brazil, HRV was also one of the primary etiologic causes of upper respiratory tract infections, with rhinorrhea being the most prominent symptom.⁵⁵ In our study, there were no laboratory findings or symptoms that distinguished influenza virus infection from HRV infection. COVID-19 was not present in FG during the period study and therefore could not be the etiologic agent of an upper or lower respiratory tract infection. In our general practice, we have observed fewer cases of lower respiratory tract infections (2.3%) than in the hospital-based study conducted by de Lavaissiere et al.³³ (15.4%). This observation is probably due to the fact that serious infections, such as pneumonia, are usually managed in a hospital.

The third cause of AFIs in our study was gastroenteritis associated with bacterial infections in 36% of cases. We did not find any case of Salmonella enterica serovar Typhi. This finding is likely due to protection from typhoid fever vaccine that all soldiers receive before being stationed in FG, although the effectiveness of this vaccine is known to be incomplete.⁵⁶

Our study has several limitations. Because we included only adults, the circulation of common viruses leading to permanent immunity was underestimated, and the results could not be extrapolated to the pediatric population. Our study concerned essentially male adults without comorbidity factors, as FSMs are a selected healthy population. For this reason, we cannot extrapolate our results to the general populations in FG who present with other noninfectious diseases in a high proportion (e.g., diabetes or obesity).^57,58 For most of the diseases that were diagnosed in our study, clinical signs and symptoms and laboratory test results cannot be compared and correlated because of the small sample size. As for gastroenteritis, our laboratory facilities in FG do not have the capacity to detect viruses in stool samples. In a number of cases, the failure to identify the etiologic agent of AFIs could in part be related to the low sensitivity of some diagnostic tests because of the delay in seeking healthcare after several days of febrile episodes.

In general practice, nonspecific fever, upper respiratory tract infections, and gastroenteritis were the main clinical manifestations of AFIs among FSMs. Faced with an AFI without specific symptoms, clinicians would have to include malaria, dengue, Q fever, and leptospirosis in their differential diagnosis. In FG, serological tools do not allow early diagnosis of Q fever and leptospirosis. Therefore, the use of molecular tools would be needed to establish the diagnosis during the acute phase of these diseases to initiate an effective, appropriate treatment rapidly to avoid clinical evolution toward severe forms. Although microbiological diagnosis of upper respiratory tract infections has rarely been carried out in general practice until recently, this has become more common practice since the COVID-19 pandemic. Individual diagnosis will then have a double collective interest: the implementation of isolation measures and epidemiological surveillance of circulating strains of causal agents. In patients presenting with febrile gastroenteritis, stool samples are rarely examined in general practice. However, in human populations that consume food supplies of poor quality or poached meat, stool examination would be useful to diagnose and treat some bacterial infections. Finally, despite an expanded microbiological search, 51.4% of AFIs remained unexplained in our study. Further investigations based on the metagenomic approach will be necessary to identify new pathogens and better describe the etiology of AFIs in FG.