Characteristics of the spread of a wildlife rabies epidemic in Europe

Characteristics of the spread of a wildlife rabies epidemic in Europe K. BOGEL,1 H. MOEGLE,2 F. KNORPP,2 A. ARATA,3 K. DIETZ,4 & P. DIETHELM5 The control of rabies in wildlife by reducing the fox population has led in Europe to inconsistent results, since little was known of the dynamics of the fox population and the interaction between rabies epidemics, host populations, and control measures. As part of the WHO/FAO Coordinated Research Programme on Wildlife Rabies in Europe, data on epidemics and persisting reservoirs of the disease were processed by computer. The results led to a better understanding of the mechanism of spread of the epidemic and to proposals for the improvement of rabies control in animals and the protection of man. Within the WHO/FAO Coordinated Research Programme on Wildlife Rabies in Europe, intensive epidemiological and ecological studies have been car- ried out on the interaction between fox populations, rabies, and control measures under different topo- graphical conditions. Moreover, the role of other wild carnivores and the effect of rabies and fox control on these species, as well as on small game and rodents, have been investigated (5, 6, 7, 8). In some European countries, these research projects were supplemented by ecological studies on the composition and annual turnover of stable fox popu- lations and on the recovery rate of fox populations after reduction by rabies and control measures (1, 3, 9). Not all projects provided a basis for a numerical analysis of different phases of the wavelike spread of the disease. Country- or state-wide rabies records that could be evaluated for this purpose were not accurate enough and special surveys with the most intensive investigations were in areas too small for this analysis. The location of statistically reasonable numbers of rabies cases was therefore determined with ade- 1 Veterinary Public Health, World Health Organization, Geneva, Switzerland. ' Referat Veterinarwesen, Regierungsprasidium, Naukler- strasse 47, 74 Tubingen, Federal Republic of Germany. 3Chief, Vector Genetics and Bionomics, World Health Organization, Geneva, Switzerland. 4Health Statistical Methodology, World Health Organiz- ation, Geneva, Switzerland. Present address: Director, Institute of Medical Biometry, 74 Tubingen, Federal Repub- lic of Germany. I Data Processing, World Health Organization, Geneva Switzerland. quate precision for an area in the south of the Federa Republic of Germany. Preliminary results obtained by electronic data processing (EDP) have already been published (4) concerning the seasonal relation- ship between the spread of the epidemic, the fre- quency of rabies cases, and the density of fox popu- lations. These results are summarized in this report and supplemented by further analyses of the front- wave of rabies and its aftermath. DATA AND METHODS Area and period of investigation All available data on animal rabies cases recorded in the former administrative area of Sudwurttemberg- Hohenzollern and two adjacent districts (Stockach and Uberlingen) of the Land Baden-Wurttemberg, Federal Republic of Germany (Fig. 1) were stored for computer analysis. The size of the study area is 11 273 km2. The period under investigation was from January 1963, when rabies first spread into the area, to 31 March 1971 when the wave of rabies had spread from north to south over the whole area. As the frontwave of the epidemic moved, large-scale gassing of fox dens was carried out in each district until the disease disappeared. For special studies, the area was divided into the following three zones according to the different land- scapes, fox population densities, and the status of the rabies epidemic observed: A. Black Forest, comprising mostly unbroken coniferous woodland, with a relatively low fox density. 3538 - 433 BULL. WORLD HEALTH ORGAN., Vol. 54, 1976 K. BOGEL ET AL. SWITZERLAND Fig. 1. Study area and sub-divisions. Zone A=" Black Forest ", zone B=" Swabian Jura ", zone C=" pre- alpine area". B. Swabian Jura, including agricultural and mixed forested areas, part of which is characterized by steeply sloping valleys transecting the spine of the Jura and where wildlife rabies is difficult to control and tends to persist. C. Pre-alpine regions, comprising a moraine land- scape which, towards the south, merges with alpine areas. Woodland, agricultural, and pastoral areas are evenly represented. These areas produced the highest number of foxes shot per km2 and per year (HIPD, see below) before rabies and control oper- ations reduced the fox populations. Collection of data For each of the 2822 recorded and laboratory confirmed cases of rabies in animals (for species see Table 1), the following information was stored on magnetic tape: date; species; locality (longitude and latitude of community); code number of com- munity; and name of district. Moreover, the number of foxes shot per km2 and per year (Hunting Indi- cator of Population Density=HIPD, see ref. 5) was recorded for each district. The location of a case was defined by the coordi- nates officially given to a central point of the com- Table 1. The role of different animal species in the spread of rabies: study area in the southern part of the Federal Republic of Germany, 1963-1971 Observed cases Species ahead of frontline a behind frontline a No. % No. % Foxes 414 97.2 1912 79.8 Roedeer 6 1.4 175 7.3 Badgers 2 0.5 55 2.3 Martens - - 50 2.1 Other wildlife - - 4 0.2 Wildlife total 422 99.1 2196 91.7 Cats 3 0.7 54 2.3 Dogs 1 0.2 50 2.1 Other domestic animals - - 96 4.0 Grand total 426 2396 a Frontline determined for each preceding month. munity (Gauss-KrUiger's system of coordinates) in which the animal was shot or found. As the com- munities in the study area are rather small, the aver- age territory being 11 km2 in area, this definition appeared to be sufficiently exact for further com- putation. Considering that approximately 3000 cases were recorded, it was impossible to define the local- ities more precisely. Definition of the frontline Initially maps were produced by plotting cases monthly, both with and without the cases from the preceding months. These maps afforded an analysis of the movement of the frontwaves through the study area (Fig. 2). All attempts to draw frontlines and to determine the distances of new cases from the frontline by graphical and semi-graphical pro- cedures were found unsatisfactory. The " frontline " was therefore defined for the computer programme as follows: In view of the general trend of the epidemic to spread from north to south (Fig. 2), the study area or its zones A, B, and C were subdivided into north-south strips each about 2.3 km wide (exact width on west-east axis as defined by printing conditions). The frontline is composed of straight lines (latitude) that are drawn 434 WILDLIFE RABIES IN EUROPE 22.5 12 67.5 90 112.5 Kilometres Fig. 2. Spread of rabies through the study area. through the sites ofthe most southerly cases. The centres of each of these straight lines are called frontline points, and these points have been used for the calculations. In strips without cases, the frontline goes through the most northerly geographical point (community) in this particular strip. Distance of new cases from the frontline The frontline of the epidemic was determined at the end of each month. All cases observed south of the frontline until the end of the subsequent month were considered as " new cases south of the front- line ". The distance of a case from the frontline was taken to be the distance from the central point in the 2.3-km-wide strip, at the latitude of the new case, to the nearest frontline point in the same strip and the five strips west and east (Fig. 3). Initially we examined the extent to which the results may be influenced by fixing the frontline at 2-month intervals and by using different numbers of strips for the determination of the distance of new cases. Modification of these components of the computer programmes had very little influence on the results reported in this paper. The assumptions made were, therefore, technical aids for compu- tation rather than scientific determinants. The maximum and mean distances of new cases from the frontline are very important characteristics 0 25 50 _ an 0 ._ E 0 75 _ 100 F \ n * * * 4 * * <N a ""-VI1'' %M_ 001 30.4.65 ° . . .. . . G . U a3 * 0 i .0*S-. 0 .; Q GI ; . q C o ^t , , 0 0 G 31.10. 65 S . C6 \O . .eAEJr. L) /G\/'/b|--*- rO Cu G .' . . \u/.* .A.o a I C) \ ,. . ,; z <X.a 0 Dv Li/,,\~~~~~~~~~~~o/ yOoo.. ts9N-gFi - s mr/ e\~~~~~~~~~~~~~~4eo 125 F- 143 0 0 133 435 K. BOGEL ET AL. Fig. 3. Scheme for computing the distance of a new rabies case from the frontline. of an epidemic since they are measures of what can be expected to happen around a new case dur- ing the 30 days following its detection. Since these values refer only to new cases found south of the frontline and do not consider any stagnation of the spread of the epidemic, they should not be confused with the velocity of spread of the whole frontwave (see below). Movement offrontwave For the total of zones A, B, and C all observed cases were projected monthly on the y-axis (north- south direction). This frequency distribution based on latitude over time for the whole area and for the different zones is shown in Fig. 4-7. A tabulation of the cases was also made for further calculations and graphical analysis. In view of the results, a computer programme was developed to further characterize the movement of the centre of the frontwave and the seasonal appear- ance of index cases in the rabies-free areas towards the south. It proved most useful to project the cases on the y-axis for all months but including not only the cases of the current month but also of the five preceding months. Thus, 6-month periods were plotted and tabulated at monthly intervals. For gra- phical evaluation, every third projection has been chosen (Fig. 8). Each quarterly projection on the y-axis in Fig. 8 therefore gives the smoothed distri- bution over the latitude of all cases recorded during the preceding 6-month period. Definition of the frontwave and its separation from a persisting reservoir Fig. 2 shows clearly the southward movement of a frontwave. However, from the 50th month onwards cases continued to appear a certain distance north of the frontwave. Practically all of these cases, which also appear in Fig. 6 and 7 for zones B and C, belong to a persisting rabies reservoir in the area of the Swabian Jura. For further computation it became necessary to separate these cases from the frontwave. A fairly accurate determination of the rear edge of the frontwave was made by drawing a line in Fig. 4 to reflect this continuing flow, although the longi- tudinal distribution of cases is missing in this computer printout. Between the frontwave thus determined and the reservoir defined by borderlines drawn on maps by hand, there were still 60 intermediate cases. How- ever, by further analysis these could be related to either the reservoir (29 cases) or the frontwave (31 cases); of the intermediate cases attributed to the frontwave, 27 were visually very closely linked with the frontwave as plotted in the right lower corner of Fig. 4. They were recorded in the districts of Wangen and Ravensburg, south of y-coordinate 5309, between May 1970 and March 1971 (see also Fig. 7 for zone C). Protrusions from the persisting reservoir to the south and east were observed for limited periods of time and accounted for 18 cases. An additional 11 cases were located very close to the borderlines of the reservoir. Thus, 29 of the 60 questionable cases could be attributed to the reservoir. Subdivision offrontwave and persisting reservoir In order to study the frequency of cases in dif- ferent animal species during the flow of the epidemic, all cases were categorized according to the scheme shown in Fig. 9. The frontwave was defined and its flow considered as described in the preceding sections. Also, the definition by EDP of new cases recorded monthly south ofthe frontline was described. The rear edge of the frontwave was defined by the most northerly case recorded per month in North East * Rabies cass in preceding months X Points used in determining the frontline at the end of the preceding month O New cas repored during the current month * Point of rew cax uxd for calctlation by cornputer ----Diaraces scanned by computer between point of new case and points of frontline ~--- istance betwen rew case and frontline ltssX{s*tisAs v 436 9 .0 e 16 'R2 1. 0 1, .5 1. 0 r- 0 LB s J: .1.2 WILDLIFE RABIES IN EUROPE 4.3/ Fig. 4. North-south distribution of rabies cases at monthly intervals over total study area (2822 cases). 1 - I -' ~~~~~~~~Linechosen to separate frontwavefrom persisting rabies reservoirI 5395- z~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 5375~~-''4. - E~~~~~~~~~~~~..4 === 5354~ ~ ~~4*4 o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .'-- ~ + - '3+--4-4-*''' -+ 440. ~ ~ - ' co 44-**4--''''--- 5314..44+.+--'' 4+4'-4--- 4-' +.'-4-144.''. '-,. 44- 53 3 1 .'.'.1 '44 '4. -4-4 '-- - 4/) I~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C Xi ---- ---- --- -'------ -- -- ---'.4 k.+t-- --- L-------- -'-- 'I_ 4 7~~~~~~~~~~~~~~4 2 + 5 8~~~~~~.U4U+.-4+44- + -- 4.-'~~~~~ .e= 3 = 9 438 K. BOGEL ET AL. Fig. 5. Progression of frontwave in zone A (479 cases). x x 5415t+5 - -' I '4- -' I S-.-. = I o~ + _ Z4- s Ss -.5375 0 o - - 0 0) _.- 4.4.~ I -. .s s v =e 53754 >_ =s Dec.-March -= .. + ._ I '4++ 4. I Co I ss +=.+4. o~~ ~~~~~ I..k.. -.- I. -.4. 8 + t @ , I J~~ *.4.- 5334- 4.'. + I -C --t+. - I I v >l to *'+--> -' M--' __X.++I l ~~~~~~~~~~~~~~~~~~~~~~~~~4.--.-. 4. Dec.-March I ~~~~~,~~ ~ 10s== _+ s+'' 5314 t- Oct.-March Southern border of area LL----------L---.----- -----.----- L------- L- Jan. 1963 Jan; 64 Jan. 65 Jan. 66 Jan. 67 Jan. 68 Jan. 69 Jan. 70 Jan. 71 Number of cases ~-1 ==4 7 *=10 -=2 +=5 K=8 1>10 += 3 *=6 R=9 zones A, B, and C (compare Fig. 5-7). We added direct, a separate printout of all case data and to these 157 cases 31 cases of the aftermath of the monthly maps permitted identification. epidemic which could not be related to the graphi- cally determined reservoir (see preceding section). RESULTS The animal cases attributed to the reservoir were mapped and thus determined. Moreover, these ani- Distance of new cases from the frontline mals accounted for the difference between all 2822 Of the new cases recorded ahead of the frontline cases and the cases of the frontwave. Where infor- within a month, 93.4%O were less than 10 km from mation on the species was not available from EDP the frontline. The maximum distance observed was WILDLIFE RABIES IN EUROPE 439 Fig. 6. Progression of frontwave in zone B (929 cases). x x 5415- 5395 L- -c , o , , eN =-.-= -.4 5375 _ _I S ~~~~~~+yt+t= ls E 3 - *++4 - -.+=-.= + -. s f _ I .- - _ -=+.sn:5354FFeb. ||t+' -+ - Co - o I -. s++=::.-.E44.4I 5314L en -.-.*K l++ --A 441 4-.+~~~+ iti X ++=fl.-..-.+C3 | Nov--March L ++ t.+ =-.++>-.=-MaI I 1 ->= + > v v t--I -.+.l| ~~~~~~II' * =+=+MF-15314v~ ~ ~ ii** + + 5 4IL I I= tt . +t- '.' Aug-Oct. IE++ -.+-.= .-.iSt+- 0 -..>--=*+=-l.+- - , ~I + 1 s *==>t"*fl+s+Oct-MarchI 5274 L j-= . I. xl7XF-----------i-----------L-----------I-----------L__----------L-----------L-----------L----------L--xL_Xk Jan. 1963 Jan. 64 Jan. 65 Jan. 66 Jan. 67 Jan. 68 Jan. 69 Jan. 70 Jan. 71 Number of cases -= 1 == 4 7 8= 10 _=2 =5 8 5> 10 + = 3 )(= 6 x = 9 440 K. BOGEL ET AL. Fig. 7. Progression of frontwave in zone C (1075 cases). 5375r'F=6' l s+~~~~+=+*+ + -C I _~~~~-+ ,4 - 5354ZI o+So+ss > r~I - - l Io S+S r ~~~~~Dec.-March + + 53541 1-++ + +-__ .- + 4 o I .4 -. i~~.I4.-'_ -.-=4= 2"2-+s++ +- I I5314 Is2 24._ - _~~~~~~~~~ I IuI+ =uK_ +4I U, I A -..=t=.- 4-_ _ I + +=> -.-.4. +~*4.- . -. 4. A Iri I ~~~~~~~~~~Oct.-Dec. .-.. 5294 r -.+ _ -8.4+++ -c If = Is+t -==-. os m +-%~~~~~~~~+k4 '. --+- 4-- 5274 F S pt.-March +K++-. += Jan. 1963 Jan. 64 Jan. 65 Jan. 66 Jan. 67 Jan. 68 Jan. 69 Jan. 70 Jan. 71 Number of cases - = 1 -= 2 + = 3 == 4 += 5 *= 6 K= 8 1=10 3 > 10 I= 9 20.5 km (Table 2 and Fig. 10). There were no sig- nificant differences among zones A, B, and C regard- ing the mean distance of new cases south of the frontline. These values were close to the 4.8 km per month determined for the entire study area. This mean distance appeared to be independent of the frequency of rabies cases and of the number of foxes shot per km2 and per year as an indicator of the population density (Table 3). This observation is further supported by analysis of a new epidemic in 1971-1972 in part of the area, starting from the district of Sigmaringen. The dis- ease spread from the focus in different directions, and the distance of 83 cases from the nearest cases that had occurred in preceding months was deter- mined on maps. Although the mean distance of 4.85 km calculated for the 83 cases was the same as that determined by EDP for the 1963-1971 epidemic, the frequency of rabies cases was extremely high in this new epidemic. In one reinfected district, WILDLIFE RABIES IN EUROPE Fig. 8. Movement of the epidemic wave of rabies through zone B of the study area. Cases for 6-month periods are shown at successive intervals of 3 months. The band width is determined by the printout conditions of the computer. The reference latitude is 5383.480 km according to Gauss-Kruger's coordinates. SOUTH 6 month period: May - Oct. 64 (7 cases) NORTH 20 Aug. 65 - Jan. 66 - (116 cases) 10 0 Aug. 64 - Jan. 65 (54 cases) Nov. 64 - Apr. 65 (158 cases) 20 Feb. - July 66 1 (91 cases) 10 o f\ 20 May - Oct. 66 - (57 cases) 10 _ O f { Aug. 66 - Jan. 67 (95 cases)) Distance from reference latitude 0.33 cases per km2 were recorded during the year with the highest frequency, representing a value 60% higher than the maximum frequency (about 0.2 cases per km2 and per year) observed in districts during the 1963-1971 endemic. In Fig. 11, the mean distance of new cases and the frequency of cases are compared for each month of the year; data were compiled for all Januaries, Distance from reference latitude Februaries, etc., of the 1963-1971 epidemic. In this seasonal analysis, mean distances and rabies fre- quency seem to parallel each other, except in Febru- ary, August, and September. Movement of the frontwave Maximum and mean distance of new cases from the frontline, as described in the preceding section, NORTH 20 10 _ SOUTH 4.- 3:E f.- 40 : 30 -o .20 0 10 . U X 30- 40 0 o 20 Z 10 - -E 441 20 10 n K. BOGEL ET AL. Frontwave identified graphically on a computer printout W projecting cases by month on north-south axis (Fig. 4) and supplemented by intermediary cases (group V) a, >0 Co m o Co a CE w OS, a, - .Z I E *ECO 1 E Persisting Rear edge of frontwave reservoir and aftermath 4) - X. 'a - c M a, E0=1 0- 0 w .0 la E 0 SOUTH 8 Number , of cases Centre of Monthly cases frontwave ahead of frontline WHO 75912 Fig. 9. Categories of rabies cases in an epidemic moving from north to south. Table 2. Distance of new rabies cases from the frontline of the previous month Distance from the frontline in km New cases 05 a 5-10 10-15 15-20 >20 Number 238 115 20 5 3 Percentage 66.4 27.0 4.7 1.2 0.7 93.4 a Upper limit included. disregard the number of rabies cases recorded month- ly in previously rabies-free areas. The distribution of cases within the frontwave can, however, be analysed using computer printouts as shown in Fig. 4-7. For all three zones of the study area, certain trends become apparent by the monthly pro- jection of cases on the north-south axis. In zone A the disease moved southwards, mainly in December/ January (in 1965 even in October) and to some extent also in March, whereas the most southerly point appeared to become stationary from April to late autumn. Similarly, in zone B the major advances southwards occurred in November/December, and in 1965 in September/November. Zone C shows southward spread also during October/December, although a second step during March/April was more pronounced in this zone than in zones A and B. The graphical presentation of the frontwave of zone B illustrates better the periodic movement of the epidemic (Fig. 8). "Amoeboid " protrusions were produced as a first step in autumn and to a lesser extent in March/April, whereas the body of the wave seemed to move much more steadily (at a rather constant rate). According to Fig. 4-7, the centre of the frontwave moved at about 27 km per year. 442 cC 0. E £2 -0N_ NORTH [Pi WILDLIFE RABIES IN EUROPE Km 20 10 0 I I I I I I I I I I I J F M A M J J A S o N D Knm 7 5 9 xa4 c 3 2 C4 "O Fig. 10. Observed mean (solid line) and maximum (broken line) distances of new rabies cases from the frontline of the preceding month. J F M A M J J A S 0 N D WHO 761113 Fig. 11. Mean distance of rabies cases recorded ahead of the monthly determined frontline (solid line) and rabies cases as a percentage of all 2822 cases recorded from 1963 to 1971 (broken line). Role of different animal species during the epidemic As shown in Table 1, the fox has been almost solely responsible for carrying the disease in rabies- free areas. The differentiation of phases of a moving epidemic (Fig. 9), and the attribution of cases to these phases (Table 4), also allow analysis of the roles of different animal species towards the end of an epidemic. Because of the very high proportion of rabid foxes among the new cases ahead of the monthly determined frontlines, this initial phase of the spread of the epidemic should be excluded from further Table 3. Frequency and distance of rabies cases in zones of different indicator values for the density of fox populations A B c - Black Forest ' "Swabian Jura - - Pre-alpine area' HIPD a 0.7 1.1 1.5 Frequency of rabies b 0.044 0.051 0.065 Distance of new cases C 4.98 4.71 4.70 a HIPD = Hunting Indicator of Population Density: foxes shot per km2 per year. t Average number of recorded cases of rabies in animals per km2 and per year. c Mean distance in km of new cases ahead of the monthly determined frontline. 15 14 13 12 11 01 8 o6Z10 6g9 5 4 3 2 0 r grI 443 5 _- K. BOGEL ET AL. 00 0 OD O~ _- 00 0) ae N rlv CD CD CD N1 co at 0 co* ~CDN)ae|_- CD co D co co CN _ coED CD (D co cn .C 00 L_O cP LP N LO LOU) a) I I. 0 IrN 6; Uo WD IN aur_CD C -~~~~ E co~~ ~ ~ ~ ~ a) C .'C 0 CD E *0n o0 CL 0 0) _- C Co0 > iDt 41j°N i. pnoS N N1 co N1 N C- LO 0 La CD CN N 0 considerations. For example, the number of recorded rabid mustelids behind the frontline was 0.12 per fox, but in advance of the frontline it was only 0.005 per fox, a difference of 25 times. Other indi- cators, such as deer and domestic animals, showed a similar pattern of 0.22 per fox behind and 0.024 in advance of the frontline, a difference of 9 times. These differences may be due to differences in the awareness of rabies in wildlife and domestic animals and are therefore most difficult to interpret in detail, except that they show that the fox remains the best indicator of the presence of rabies. To study the conditions towards the end of an epidemic it seems more appropriate to compare the rear edge of the frontwave and its aftermath (which equals the final phase of the epidemic) with the centre of the frontwave (equalling the height of the epidemic). In Table 5, all cases observed during these two phases are given, as well as the percentages of these totals recorded during the final phase in foxes, mustelids, domestic animals, and deer. The latter two categories of animal are of no significance as transmitters of the disease and are well recorded by veterinary services. In particular, rabid domestic animals are excellent indicators for the purpose of comparison, since they are almost completely re- corded with little variation of public awareness dur- ing the height of the epidemic and its final phase. The proportion of cases found in the final phase of the epidemic varied significantly among the species (X2=9.084; P< 0.05; see Table 5). The main con- tribution to this value was given by mustelids (X2= 7.418). The other proportions were very similar. Mustelid cases also occurred in the persisting reservoir more frequently than expected, but the dif- ference between this and the centre of the frontwave is not statistically significant if the martens and bad- gers are considered together. However, out of the 42 rabid martens recorded in the centre of the epi- demic wave and the persisting reservoir, 13 (30.9%.) were observed in the reservoir. The proportion dif- fers significantly from the 12.5% of rabid domestic animals observed in the same persisting reservoir (Table 4). DISCUSSION The results will be discussed with respect to (a) the assessment of the risk of human exposure to rabid animals, (b) the improvement of wildlife rabies con- trol in different phases of the epidemic, and (c) the roles of different animal species towards the end of a rabies epidemic. 444 0 I- 0 CD D 0 I 0 '7. 0 0 0at _0 cr en G) a) U) X 0a 0 -L C 0 El m en (A O 0 E ao-c a)0 en0)aC -C_ 0.CD o CD +0 tn 0 a) -o 0. 0 -0 -0E 0 0 WILDLIFE RABIES IN EUROPE Table 5. Distribution of species according to phase of the rabies epidemic Recorded rabies cases Species Proportion in Chi-square centre of final phase final phase values frontwave of frontwave Foxes 1531 140 0.084 0.798 Mustelids 72 17 0.191 7.418 Domestic animals 161 16 0.090 0.053 Deer 125 15 0.107 0.815 Total 1889 188 0.091 9.084 Characteristics of the spread of rabies The definition presented in this paper of a front- line of an epidemic wave and the calculation of distances of new cases from this line should be con- sidered as a first attempt to obtain some information on the distances between rabies cases. Although we did not observe significant differences among the mean distances in zones A, B, and C, it appears to be very important to study distances under other ecological conditions. Where the disease spreads into previously uninfected areas, this can be easily done by mapping the cases, as we have shown for re- infection of the study area in 1971. The spread of rabies is characterized by several variables. The mean distance between new front- line cases refers to points where new cases occurred within a one-month period and does not describe a movement of the frontline at its whole length. The mean distance between new cases is more infor- mative in conjunction with the total number of cases observed ahead of the frontline per month. Even more instructive is the ratio of new cases ahead of the frontline over the total number of cases observed per month. This ratio defines the progressivity of a rabies epidemic, according to Moegle et al. (4). The movement of the centre of the frontwave further characterizes the epidemic. Fig. 4-7 suggest that the centre moves at a rather constant rate of about 27 km per year. All these characteristics of the spread of rabies are not to be confused with frontwave protrusion defined by the geographically most advanced cases in each of the three zones studied (Fig. 5-8). Pro- trusions coincided with the periods of high rabies frequency in late autumn and in March. The ex- tremely advanced cases recorded in March and April did not seem to give rise to a further spread of the disease since generally about 6 months elapsed before further progression of the extreme points of infection were observed. Sites of extremely advanced cases were reached during the summer months by the more slowly moving centre of the epidemic wave, from which a further protrusion apparently originated in the autumn (Fig. 8), usually in late November or even in December (Fig. 5-7). Apart from this production of protrusions, rabies mobility increased in August, as shown by the sharp increase in the mean distance of new cases from the frontline (Fig. 11). This mobility must there- fore have occurred behind the most advanced cases in accordance with the movement of the centre of the frontwave (Fig. 8). These observations further support the conclusion that the geographically most advanced cases observed in March were of little or no significance for the further spread of the disease. This could be explained by the annual turnover of fox populations and the function of the new fox generations as a basis of the rabies reservoir. The pre-breeding population of adult foxes in spring represents only about one-third of the maximum population present after the birth of the new generation (3). Chains of infection should, therefore, tend to end in the adult fox population in March and April. Behavioural factors related to the period of pregnancy, the establishment of terri- tories, increasing food supplies, etc., may further lower the contact rate and aggressivity between foxes in this season. The probability that the epidemic spreads from extreme sites such as the protrusions may therefore be very low, whereas the high rabies frequency in the centre of the epidemic wave could provide suf- ficient chances for a transfer of the virus from the adult population to the new fox generation. Further studies must be carried out to show whether this break in the chain of infection in ex- 445 K. BOGEL ET AL. tremely advanced cases, and the shift of the reservoir from adult populations to include the new gener- ation in the centre of the epidemic, do indeed circum- scribe the critical paths of the epidemic. In particular, the extent to which this pattern depends on the density of fox populations in spring should be investigated. In the areas described in this report, the gassing of fox dens was in general done only after rabies had reached an administrative district. The process leading to epidemic protrusions could hardly have been influenced by control measures. However, the natural differences between the popu- lation densities of the three zones also had no sig- nificant influence on the patterns of spread (Table 3, Fig. 5-7). Suggestions for controlling the spread of the epidemic It may be essential in the future to include in routine surveillance an annual analysis of the front- wave. This could easily be done by mapping, by the identification of epidemic zones of about the size of zones A, B, and C, and by the projection of cases on an axis representing the major direction of spread, as shown in Fig. 8. Should the hypothesis of the transfer of rabies from the centre of the front- wave to the new fox generation, as presented in this report, be confirmed, it would be most useful to investigate the situation in February, March, and April so that decisions could be made on appropriate control operations. Experience must show whether it is in fact the area between the centre of the front- wave and the most advanced cases where, during the summer months, the transmission takes place and the further movement of the frontwave is prepared. Special efforts should be made to reduce the susceptible fox population rigorously in that area by the gassing of dens, intensified hunting, and any supplementary measures. This concentration of efforts on a selected area during a limited period could actually stop an epidemic wave. Insufficient control operations in areas further ahead of the frontwave could be dispensed with, and areas once infected could then be treated efficiently so that the disease could be eliminated. The results suggest that the speed of the spread of rabies is not influenced by a reduction in the fox population unless the popu- lation density falls below a critical level so that the epidemic wave can be completely stopped (5). Rabies control during the final phase of an epidemic The method described enabled us to follow the flow of the epidemic and to analyse continuously its final phase as it moved over the study area. The ratio of rabid foxes to rabid indicator animals remained rather constant from the centre of the frontwave (ratio 5.4: 1) to its rear edge and after- math (ratio 4.5: 1, see Table 4). Apparently, gassing operations and the further reduction of the fox popu- lation by rabies did not influence the main pattern of the epidemic in that the fox remained the prin- cipal reservoir animal and transmitter. Wild mustelids, although a minor factor in the rabies epidemic, showed a proportional increase from 3.8%Y of all cases in the centre of the frontwave to 9%O in the final phase of the epidemic. This may either be due to an actual increase in rabies incidence in these species, or it may only reflect a relative increase as the fox population is considerably reduced towards the end of an epidemic wave. Both actual and relative increases, however, would suggest that short chains of infection exist in mustelids and that such chains of infection become more apparent towards the end of an epidemic. Although martens and badgers do not seem to maintain the epidemic when rabies has disappeared from fox populations, their contribution to the total reservoir of rabies should not be neglected when the reservoir in foxes becomes exhausted. This should be considered particularly where rabies is difficult to control by the reduction of fox populations. An example is the persisting reservoir in the Swabian Jura. Special attention must be paid to martens under these conditions. Besides the secondary role of mustelids and the predominant role of the fox as the principal host and transmitter of wildlife rabies in Europe, the results do not indicate the involvement of other ani- mal species in the rabies reservoir. All the character- istics of the wavelike movement of the epidemic and its persistence in certain areas can easily be ex- plained with the fox as the principal host. This sup- ports earlier reports of many European investigators (e.g., 1, 2, 5, 7, 8). Public health significance of maximum distances between rabies cases Knowledge of maximum distances between rabies cases could be particularly useful when decisions are being taken on the risk of exposure and the necessity of post-exposure treatment in man. Knowledge of the maximum distances of cases within one-month periods would also be useful for certain control measures, e.g., the vaccination of dogs, cats, and cattle in those communities in the range of one or 446 WILDLIFE RABIES IN EUROPE 447 two " monthly maximum distances " from recorded rabies cases. Seasonal variations in these distances could be taken into consideration (Fig. 10). Although such control measures will not influence rabies in wildlife, they may considerably reduce the number of persons receiving post-exposure treatment in areas at risk. Moreover, knowledge of maximum and mean distances could be of great importance where, for legal or economic reasons, control operations should be limited to " infected" fox populations. RtSUMt CARACTERISTIQUES DE LA PROPAGATION D'UNE EPIDEMIE DE RAGE DES ANIMAUX SAUVAGES EN EUROPE On a etudie les caracteristiques d'une vague epid& mique de rage des animaux sauvages dans une region du sud de la Republique federale d'Allemagne. Le traite- ment des donnees sur ordinateur a permis de suivre la progression de cette vague et d'y distinguer a tout moment les cas observes en avant du front mensuel, ceux du centre de la vague et ceux qui venaient en queue ou en arriere- garde. En ce qui concerne les nouveaux cas, leur distance maximale en avant du front determine tous les mois a ete de 20,5 km; leurs distances moyennes dans des zones differentes quant a la topographie et la densite des popu- lations vulpines etaient tres similaires, a savoir a peu pres 4,8 km. Sur la totalite des cas, 93,4% se sont produits a moins de 10 km du front. Quant au centre de la vague epidemique, il se deplarait it une vitesse a peu pres constante de 27 km par an, independamment de la topographie des zones. On obser- vait une periodicite pour les cas initiaux les plus avances, correspondant a des saillants de la vague a la fin de l'automne et, dans une certaine mesure, en mars egale- ment. Cette progression, rapportee a une zone entiere, coincide avec Jes periodes oui les cas de rage ont une frequence elevee. I1 y avait cependant une rupture dans ce parallelisme en aofit, car a ce moment la distance moyenne des nouveaux cas par rapport au front aug- mentait brusquement. Pour mieux lutter contre la pro- pagation de la rage des animaux sauvages par la reduc- tion selective des populations vulpines, il faut verifier l'hypothese selon laquelle la rage passerait pendant l'ete des adultes aux nouvelles generations de la population vulpine au centre de la vague epidemique. A cette periode, cette nouvelle generation represente au moins les deux tiers de la population totale. On estime que les cas qui se produisent tres loin en avant du centre de la vague, par exemple ceux qui apparaissent en mars dans la popu- lation adulte, sont d'une importance mineure en ce qui concerne la propagation ulterieure de la maladie. Les renards representent 97% de la totalite des ani- maux enrages trouves en avant du front. Au cours de la phase finale de l'epidemie et dans un reservoir persis- tant, la proportion de Mustelides rabiques, en particulier ceux du genre Martes, etait nettement plus 6levee qu'au centre de la vague epidemique. En revanche, les cas se produisant parmi d'autres especes animales, y compris les animaux domestiques et les chevreuils qui constituent des (( indicateurs #, ont ete distribues tres egalement pen- dant ces phases de l'epidemie. Bien que le renard soit le principal h6te du virus rabique dans toutes les phases epidemiques, il faut tenir compte du role des Martes et des blaireaux dans les programmes de lutte vers la fin d'une epidemie. En particulier la chasse aux fouines doit etre intensifiee. REFERENCES 1. BOGEL, K. ET AL. Recovery of reduced fox popu- lations in rabies control. Zentralblatt fur Veterindr- medizin, B, 21: 401-412 (1974). 2. KAUKER, E. Vorkommen und Verbreitung der Tollwut in Europa von 1966-1974. Berlin, Heidelberg, New York, Springer, 1975. 3. LLOYD, H. G. ET AL. Annual turnover of fox popu- lations in Europe. Zentralblattfur Veterinarmedizin, B, 23: 580-589 (1976). 4. MOEGLE, H. ET AL. Zur Epidemiologie der Wildtier- tollwut: Untersuchungen im siudlichen Teil der Bundesrepublik Deutschland. Zentralblatt fur Veteri- narmedizin, B, 21: 647-659 (1974). 5. MOEGLE, H. ET AL. Einfluss der Begasung der Fuchs- baue auf die Fuchsdichte und die Wildtollwut in Baden-Wurttemberg. Berliner und Munchener tier- arztliche Wochenschrift, 84: 437-441 (1971) (English summary in Weekly epidemiological record, 47: 49-52 (1972)). 6. MULLER, J. The effect of fox reduction on the occur- rence of rabies. Observations from two outbreaks of rabies in Denmark. Bulletin de l'Office international des Epizooties, 75: 763-776 (1971). 7. WANDELER, A. ET AL. Rabies in wild carnivores in Central Europe. I. Epidemiological studies. Zentral- blatt fur Veterinarmedizin, B, 21: 735-756 (1974). 8. WANDELER, A. ET AL. Rabies in wild carnivores in Central Europe. II. Virological and serological exam- inations. Zentralblatt fur Veterindrmedizin, B, 21: 757-764 (1974). 9. WANDELER, A. ET AL. Rabies in wild carnivores in Central Europe. III. Ecology and biology of thr fox in relation to control operations. Zentralblatt fur Veterindrmedizin, B, 21: 765-773 (1974).