During 2005–2010 (time span needed to cover the study area) and during the months more favorable for infection (October–April [3]), foxes were either shot at night or trapped. The sampling size was chosen to collect ≈100 foxes from each department. Therefore, a grid of 5 km × 5 km to 10 km × 10 km, depending on the department size, was superimposed over the sampling area, and no more than 1 fox was collected in each square. The geographic district where the sample was taken was then noted, and each fox was randomly allocated geographic coordinates within the commune (a French administrative division of 10–100 km²).

Adult E. multilocularis worms were identified in departmental veterinary laboratories. Staff were trained by the Anses-Nancy laboratory (National Reference Laboratory for echinococcoses); that laboratory also confirmed any unrecognized specimens. For time- and cost-effectiveness during the analysis, we used the segmental and sedimentation counting technique (12).

We used the χ² test to compare E. multilocularis prevalence between departments. The distribution of E. multilocularis prevalence in foxes was modeled against geographic coordinates by using a generalized additive model with a logistic link function and a thin plate regression spline on 300 knots (13). Analyses and graphic displays were conducted by using ArcGIS 9.3, R 2.14.0 and the R packages maptools 0.8–10, mgcv 1.7–12, sp. 0.9–91, and splancs 2.01–29.

A total of 3,307 foxes were collected (Table 1). Eighty-five could not be assigned a commune code and were not kept for further analysis, except to compute E. multilocularis prevalence in departments. The mean number of foxes collected by department was 84.95 (± SD 25.76), which represents a mean of 1.56 foxes per 100 km² (± SD 0.57). For 4 departments, (36, 61, 67, and 69), full sampling could not be completed because of technical and/or administrative reasons. Urban areas, such as departments 93, 95, and 91, also were undersampled because of human population density and high urbanization, all factors preventing easy fox sampling.

Figure 2

Figure 2. . Model-predicted prevalence (A) and standard error (B) of Echinococcus multilocularis in foxes, France, 2005–2010. 1 = 100%.

We confirmed E. multilocularis in foxes in 35 departments (Figure 2). The prevalence varied widely among departments, from 0 (95% CI 0–5%) to 54% (95% CI 42%–64%) (Table 1) but was locally higher in some areas (Figure 2). The mean prevalence in the entire studied area was 17% (n = 3,307; 95% CI 16%–19%). The prevalence in the historically echinococcosis-endemic area was 41% (n = 789; 95% CI 37%–44%) and represented >55% of all infected foxes and <21% of the total area studied. Furthermore, in comparing our results with those of earlier similar studies during the same season with the same technique, we detected a significant increase of E. multilocularis prevalence in foxes over time in most of these departments (Table 2).

Our study confirms the presence of E. multilocularis in areas where it is known to be endemic and indicates its presence in 25 additional departments. However, we cannot discard the possibility that E. multilocularis was present but remained undetected during the 1980s–1990s. That E. multilocularis could have remained undetected if it were not already at a very low prevalence in general is doubtful. Isolated human cases recorded in the early 2000s outside areas to which it is known to be endemic corroborate this possibility (3). The same uncertainty applies in other parts of Europe (14). Taken as a whole, these findings indicate that the transmission intensity of E. multilocularis through fox populations in the occidental part of the European focus area is likely to have increased during the late 1990s and led to a much higher average prevalence than previously reported. Furthermore, infected foxes close to large-scale conurbations, such as Paris and its large suburban surrounding departments (93, 91, and 77) (Figure 1) amounting to 11,728,240 inhabitants, may create new conditions for human exposure similar to those already described in other highly urbanized cities, such as in Switzerland, Germany, and eastern France (Nancy), but on a much larger scale.

We believe that the public needs to be proactively informed and protected, including through awareness initiatives among urban residents and, in specific areas (15), more direct action toward the parasite may be considered. Monitoring the possible further extension of the parasite westward and southward and the evolution of prevalence in foxes in the historically and the newly echinococcosis-endemic areas also are essential.