As of November 30, we had received samples from 129 horses and 2 donkeys clinically suspected of having WN virus infection by veterinary practitioners (neurologic signs such as ataxia, paresis, or paralysis, with or without fever >38.5°C). A confirmed case was defined as illness in an equine with clinical suspicion of WN virus infection and a positive WN virus IgM antibody test result; a probable case had a negative WN virus IgM test result and a positive IgG antibody test result. A total of 58 equines were defined as having confirmed cases (57 horses and 1 donkey) and 18 horses as probable cases. Twenty (34%) of the animals with confirmed cases and one (6%) of the probable cases died. Eight of the 58 confirmed cases had a negative IgG antibody test result; 4 of these 8 died. Of the probable cases, two had samples obtained 15 and 23 days after illness onset; the rest had samples obtained during the acute phase of illness. The clinical symptoms of the confirmed and probable cases were similar, as was the age distribution of the animals (mean 12.5 ± 5.3 vs. 12.0 ± 6.6 years for confirmed and probable cases, respectively).

Most positive samples (confirmed and probable) were reported in September (82.9%). The last case was reported on November 3. The clinical symptoms included mainly fever (>38.5°C), ataxia, paresis, and paralysis (Table).

Figure 3

Figure 3. . Age of horses with confirmed, probable, and fatal West Nile virus infection, France. ND = not determined.

Figure 2

Figure 2. . West Nile confirmed, probable, and fatal equine cases, by week of clinical onset, France.

Ages of confirmed and probable cases ranged from 3 to 30 years (mean 12 years, median 10 years). There were 4 stallions, 20 mares, and 49 geldings (no information for 3 horses). Age were not known for 4 of the 21 fatal cases. Of the 17 horses for which information was available, 41.2% and 29.6% were in the 6- to 10- and 16- to 20-year age categories, respectively (Figure 3). Most fatal cases (57.1%) were recorded before September 15 (Figure 2); among fatal cases, 18 (86%) were euthanized, including one donkey that had neurologic signs followed by a short period of remission and then severe hepatic failure.

All but three confirmed and probable cases were located in an area within a radius of 15 km, in a region in Hérault and Gard provinces called "la petite Camargue." Thirty-one (40.8%) of the horses were located within a 5-km radius of the first two reported cases (Lansargues). Three cases, all fatal, were located near this area, in Bouches du Rhone Province approximately 30 km from the first reported cases and 15 km outside the area where confirmed cases were reported. These animals, according to the owners, had not moved from this area during the 3 weeks preceding the onset of symptoms. However, because of the economic consequences of the restricted movement measures imposed by the Commission of European Communities, we assume some owners may not have observed the restrictions.

We also received 33 samples from other animals, some of them with neurologic signs, in the infected area during the outbreak: 16 cows, 8 goats, and 9 others (e.g., camel, dog, zebra). WN ELISA results were negative for all of them.

No human suspected of having WN infection has been laboratory-confirmed among 51 persons tested, including 33 hospitalized with signs of encephalitis or meningoencephalitis and 18 others with fever or living in close contact with horses. All these samples were obtained from persons living or traveling in the infected area during the outbreak. In contrast, WN IgG antibodies were detected in 3 of 33 gamekeepers working in this area. Two had WN neutralizing antibodies: one had no WN IgM antibodies; the other (who had no history of travel during recent years) had low but detectable IgM antibodies.

Virus Isolation and Molecular Characterization of Virus Isolates

WN virus was isolated after one passage into C6/36 and Vero E6 cells from the rachidian bulb of the first confirmed case and from cerebellum, cortex, and lumbar spinal cord of another horse that died on September 6. Viral RNA was extracted from culture supernatants and a reverse transcription-polymerase chain reaction (RT-PCR) was performed with primers located in the envelope gene fragments WN240 and WN132, as described (24). Nucleic acid sequences were obtained on an automated Applied Biosystems sequencer (PPE Biosystems, Foster City, CA). WN virus sequences were aligned by using the multiple sequence alignment software CLUSTAL.

Figure 4

Figure 4. . Phylogentic trees based on nucleic sequence data of E-glycoprotein gene fragments of 254 bp. GenBank accession numbers for the sequences included in the tree are indicated.

Phylogenetic analysis of an informative region of the E glycoprotein gene (Figure 4), using tree-view, showed that the WN France-2000 isolate belonged to lineage 1 and was closely related to both horse Morocco-1996 and Italy-1998 isolates. It is also closely related to mosquito isolates from Senegal-1993, Kenya-1998, and Romania 1996, as well as to the recent human isolate from Volgograd-1999. It is distinguishable from the group including both the New York-1999 and Israel-1998 isolates, as well as a WN virus recently isolated in our laboratory from the brain of a human fatal case that occurred during an outbreak in the governorates of Mahdia and Sfax on the Tunisian coast in 1997 (H. Triki, unpub. data).

General Survey in Horses

To determine the number of infected horses and thus the number of asymptomatic infections, a serosurvey study has been undertaken, which includes all horses located within a 10-km radius of confirmed cases. A total of 5,133 sera were collected from September to November 2000 from the three provinces where cases were reported (Herault, Gard, and Bouches du Rhone). Preliminary results showed 428 (8.3%) horses with IgG antibodies; 248 had IgM antibodies. Analysis of these data is in progress, especially to determine rates of seropositivity for each commune. (A commune is the smallest French administrative subdivision, which approximately corresponds to an English parish).

Figure 5

Figure 5. . Geographic location of confirmed and probable clinical cases and serologically positive cases (according to serosurvey) of West Nile infection in equines, France. Data are grouped by commune, the boundaries of...

The geographic locations of the seropositive horses were compared with those of the clinically confirmed and probable cases (Figure 5).

In this report we described the 2000 WN outbreak in horses in southern France. Only a few reports of WN virus encephalomyelitis cases in equines have been published. In Egypt, a high prevalence of WN antibodies (54%) was observed during a serologic survey conducted from January to May 1959 in 436 equines (horses, donkeys, and mules). One suspected case was fatal and confirmed by viral isolation from the brain (25). In France during the 1962 WN outbreak in Camargue, several horses were suspected to be infected (6). In Morocco (provinces of Kenitra and Larache), 94 equines were affected from August to mid-October 1996; 42 died. The disease was reported in all age categories (9), and virus was isolated from a brain biopsy sample (10). In Italy, from August to beginning October 1998, 14 horses in Tuscany had laboratory-confirmed WN virus infection, and 6 animals died (12). Virus was isolated from a brain biopsy sample (V. Deubel, unpub. data). In Israel in 1998, 18 serum samples from horses with encephalomyelitis had WN-neutralizing antibodies, and virus was isolated from the brain of a stork (26). In 1999, thousands of geese were destroyed when WN virus was identified in commercial flocks (27). In the northeastern United States, 20 horses were infected by WN virus in 1999; 9 died (28). In 2000, 63 equine cases, approximately 35% fatal, were confirmed. The first case was identified in mid-August 2000 (29).

In France, the outbreak started in August 2000 and ended in November. Most positive samples (80%) were obtained before September 30. The death rate during this outbreak (28%) was lower than observed in Morocco (45%), Italy (43%), and North America (45% and 35%). Most fatal cases (57.1%) occurred at the beginning of the outbreak, before September 15. At that time, most veterinary practitioners thought that the disease could not be cured and did not apply symptomatic treatment.

Horses, as well as humans, are generally considered to be dead-end hosts of WN infection; however, little is known about the duration and magnitude of viremia. Experimental infections in horses and donkeys in Egypt (25) and in France (30,31) showed undetectable or low viremia of short duration. However, these experiments were conducted in different conditions with different WN strains; therefore, it is difficult to extrapolate from these results.

During the present outbreak, we were unable to detect WN virus (by RT-PCR and inoculation into cell culture) in the blood of a few animals tested, including animals with virus identified in the brain. In contrast, using intracerebral inoculation into mice, virus was isolated from blood samples during previous epidemics in Egypt, Israel, South Africa, and France.

During the French outbreak, WN IgM antibodies were not detected in 18 IgG-positive animals with neurologic signs from which blood samples were obtained. In the absence of a second blood sample and in the context of the outbreak, we concluded that these animals had a recent WN virus infection. This conclusion was supported by previous serosurveys conducted in 1999 and 2000 that demonstrated absence of Flavivirus antibodies in horses (unpub. data). These data suggest that in some cases IgM response was very low or nondetectable by commonly used techniques. This may have implications for the diagnosis of recent infections.

The outbreak appeared limited to a restricted area within a 15-km radius. This region, near Camargue National Park, where the 1962 outbreak occurred, is characterized by its original flora (wetlands, rice fields, garriguea [a geographical dry area, typical of the mediterranean basin]) and fauna (more than 300 migratory and resident bird species and large populations of mosquitoes). No abnormal deaths of birds were reported. An epidemiologic investigation of WN virus in birds, including five common nesting species, is in progress. The vector(s) involved in the present outbreak is (are) still unknown.

Thousands of horses live in this region. Thus, our large serosurvey should allow us to precisely determine the geographic distribution of the infection and the number of asymptomatic infections in the geographic area where the outbreak occurred.

Two gamekeepers had WN-neutralizing antibodies, and one of them also had IgM antibodies. Thus, human transmission occurred during this outbreak. However, in the absence of a serologic survey, evaluating the level of human infection among persons living in the infected area is not possible.

The main concern for 2001 is the possibility of persistence of virus transmission and thus the risk of human infections. Natural vertical transmission of WN virus in Culex mosquitoes (32) or survival in overwintering mosquitoes (3,33) could explain the persistence of the virus. However, human transmission likely depends on several factors, including environmental factors, vectors, and amplifying host conditions. Usually in most countries where WN outbreaks have been documented, a few cases are reported during subsequent years (34).

Several questions about WN virus infection are still unresolved, among them whether WN infection is a major health problem for humans and horses. The main problem during this outbreak was not the disease itself but the economic consequences from the restricted movement measures imposed by the Commission of the European Communities (cancellation of horse exhibitions, layoffs of employees in equestrian centers). Gaining more knowledge about the role of horses in virus transmission under natural conditions is important.