Article Text

Download PDFPDF

Risk assessment of importation of dogs infected with Echinococcus multilocularis into the UK
  1. P. R. Torgerson, BA, VetMB, PhD, DipECVPH, DipEVPC, MRCVS1 and
  2. P. S. Craig, BSc, MSc, PhD2
  1. 1 Ross University Veterinary School, PO Box 334, St Kitts, West Indies
  2. 2 Cestode Zoonoses Research Group, School of Environment and Life Sciences, Salford University, Greater Manchester M5 4WT


A risk assessment was undertaken to assess the likelihood of importation of dogs infected with Echinococcus multilocularis should the requirement for treatment with praziquantel be relaxed under the Pet Travel Scheme. The analysis suggested that for every 10,000 dogs going on a short trip to Germany, the probability of at least one returning to the UK infected is approximately 98 per cent. If the dogs had been longer-term residents, this probability rises to over 99 per cent. This indicates that if praziquantel treatment of dogs on importation into the UK is abandoned, then it is almost inevitable that E multilocularis will be introduced.

View Full Text

Statistics from

Echinococcus multilocularis is a fox tapeworm, which is among the world's most lethal zoonoses (Craig 2003). Human infection results in alveolar echinococcosis following aberrant infection with the larval stage of the cestode parasite. The disease presents as slow-growing metacestodes, which behave like a malignant neoplasm in the liver, and is usually fatal if untreated. Although foxes are the usual definitive hosts, dogs are highly susceptible to infection (Kapel and others 2006) and infected dogs present a considerable hazard to their owners. Currently, the UK is believed to be free of E multilocularis: a large postmortem-based surveillance study of British foxes failed to find a single infected animal (Smith and others 2003). The parasite is highly endemic in many parts of Europe. Germany, France and Switzerland are key endemic areas where over 500 human alveolar echinococcosis cases were identified between 1982 and 2000 (Kern and others 2003). Furthermore, the parasite is believed to be extending its range in Europe (Romig and others 2006) (Fig 1). Because of the current parasite-free status of the UK, a derogation was obtained for the UK (also by Ireland, Sweden and Malta). This resulted in the Pet Travel Scheme (PETS), ensuring domestic dogs and cats are treated with praziquantel within 48 hours before embarkation. Praziquantel is highly effective in removing intestinal cestode infections including Echinococcus, Taenia and Dipylidium species. In addition, rabies vaccination and tick treatment are required. This derogation was due to expire in 2010 (Anon 2009a) although an extension to December 31, 2011 has since been proposed (Anon 2009b). If the derogation is allowed to expire, it will no longer be a requirement to treat pets with praziquantel before importation to the UK. This paper describes a simple risk assessment that was undertaken to estimate the likelihood of a dog travelling from the UK becoming infected with E multilocularis in an endemic European country, and the probability of infected dogs arriving in the UK.


Present known distribution of Echinococcus multilocularis in Europe (estimated from various sources from 2009)

Materials and methods

Mathematical model

A simple mathematical model was derived for the transmission of Echinococcus species in dogs (Budke and others 2005). An important aspect of this model was that, following data fits, there was little evidence of parasite-induced host immunity in canine E multilocularis infection and hence 'herd' immunity is unlikely to play a role in controlling the parasite biomass in the dog population (Torgerson 2006). Thus, a simple equation to model transmission of E multilocularis in dogs can be applied:

p(t)=(β/{β+μ}) * (1-exp{-(μ+β)t})

Where p(t) is the prevalence of E multilocularis infection at age t (in this report weeks are used, although time can be measured in years with appropriate conversions for the constants), β is the prevailing infection pressure (infections per unit time or incidence — also in weeks) and μ is the rate at which the host loses its parasites (or recovery rate).

Given the parasite loss rate (the inverse of the life expectancy) and the age-stratified prevalences in the target population, an estimate of the infection pressure (infections per week) can be made. This infection pressure can then be used to calculate the probability of infection in the 90 days before importation into the UK — 90 days being the estimated life expectancy of the parasite in dogs (Kapel and others 2006), and hence, even if a dog has been resident in Europe for longer than 90 days, the probability of infection will not increase further. In cases where the dog has been resident in Europe for fewer than 90 days, the probability of infection will be lower due to the reduced time period of exposure at the prevailing infection pressure. Once the probability of infection of a single dog is known, it is a simple matter to calculate the probability of importation of at least one infected dog given the number of dogs imported over a period of time using the binomial probability distribution.


In contrast to that for red foxes (Vulpes vulpes) (Romig and others 2006), there are limited data from European countries on the prevalence of E multilocularis in dogs, but a large study from Germany suggested that 43 of 17,894 faecal samples from dogs (0·24 per cent) were positive for E multilocularis DNA (Dyachenko and others 2008). The results of that study indicated that E multilocularis is present throughout most of Germany. Likewise, in Switzerland it has been reported that two of 660 dogs (0·3 per cent) from the general dog population were shown to be infected with E multilocularis by coproantigen with PCR/postmortem confirmation (Deplazes and others 1999). Therefore, the German data were used to estimate the prevalence p(t) in the model. The estimated life expectancy from experimental data (Kapel and others 2006) was used to estimate μ of 0·078 per week. The data from the German study were not age stratified. Indeed, there was no data on the age of the dogs from which the faecal samples were examined. Therefore it was assumed that the samples were from a dog population of mid-age, that is, a mean age of five years (260 weeks), which would be a reasonable estimate of the average age of the dog population.

Data for the numbers of dogs entering the UK utilising the PETS passport programme were obtained from the Defra website (Defra 2009).


Given a mean age estimate of the German dog population of five years, the mean infection pressure β would be 1·87 x 10-4 infections per week. Thus, if 10,000 dogs from the UK made a two-week trip to Germany, the probability that at least one dog would return to the UK infected with E multilocularis is approximately 0·98, or 98 per cent (Fig 2). A long-term canine resident of Germany would have a probability of infection in the previous 90 days of 4·8 x 10-3. If 10,000 of these dogs were imported, the probability of at least one being infected would be over 99 per cent. In 2008, 93,719 dogs entered the UK through the PETS. If 30 per cent of these had originated from an endemic area, which includes much of the EU and Switzerland, and had been residing there for at least two weeks, the authors estimate that there would have been a probability of more than 99 per cent that at least one of these dogs would have been infected with E multilocularis.


Probability of at least one infection per 10,000 weeks of exposure. The solid arrow represents the estimated infection pressure to dogs in Germany. Thus for every 10,000 dogs there is a probability of approximately 0·85 that at least one dog is exposed per week, rising to 0·98 over a two-week period


This simple risk assessment indicates that there is a considerable threat of importation of dogs infected with Echinococcus multilocularis if the requirement for treatment with praziquantel is relaxed. The number of dogs being imported is rising each year and the total number that had participated in the PETS up to December 2008 was 423,219 (Defra 2009). Because of the large number of dogs and the likelihood that a substantial proportion spent a period of time in an area of Europe endemic for E multilocularis, the present study supports the claim that, in the absence of routine praziquantel treatment, introduction of E multilocularis into the UK would be inevitable. Even if the assumptions presented here substantially overestimate the prevalence of E multilocularis in dogs in endemic areas in Europe, the risk would still be considerable. However, copro-PCR methods for diagnosis, although highly specific, have shortcomings in sensitivity (Ziadinov and others 2008) and thus the German data used in this risk analysis are likely to underestimate the true prevalence and therefore the risk of infection in dogs. Furthermore, dogs in rural areas in Switzerland have a higher prevalence than that reported for the general population; in a rural area of western Switzerland, six of 86 dogs (7 per cent) tested positive for E multilocularis (Gottstein and others 2001). A prevalence of 2·8 per cent in rural dogs has been reported in Slovakia, which is a country where the disease is emergent (Antolová and others 2009). Thus, dogs imported into the UK from rural Europe could have a substantially increased risk. The conclusions from the risk assessment are also supported by recent observations in Japan. The northern island of Hokkaido is endemic for E multilocularis and a recent survey found that 18 of 4768 dogs were infected (0·38 per cent) (Nonaka and others 2009). This is similar in magnitude to the German data quoted in the present analysis. One of the 18 dogs had been transported from Hokkaido to Honshu, the main island of Japan, where E multilocularis is believed to be absent, and was found to have a patent infection after arrival in Honshu. As in many parts of Europe, the prevalence in foxes in Hokkaido is approximately 40 per cent (Nonaka and others 2009).

Any dog arriving in the UK infected with E multilocularis could have serious consequences. First, there is the direct hazard to the owner. Dog ownership is a risk factor for human alveolar echinococcosis in endemic areas such as Austria (Kreidl and others 1998). This is likely to be due to the close contact between the animal and its owner. Of greater concern is the possibility of introducing E multilocularis into the UK wildlife population. Infected dogs excrete viable eggs (Kapel and others 2006). Therefore, if infected dog faeces contaminate areas where there are concentrations of suitable intermediate hosts (principally rodents) then transmission is likely to occur eventually to red foxes (the main definitive reservoir host). There are a number of suitable intermediate hosts in the UK. The bank vole Myodes (Clethrionomys) glareolus is common and widespread through much of the UK (Harris and others 1995). Infection of this species with E multilocularis has been reported in Belgium (Hanosset and others 2008), so it would be likely to make a suitable intermediate host in the UK. The most common mammal in the UK is the field vole (Microtus agrestis), which has a population of 75 million (Harris and others 1995). Microtus species are known to be excellent intermediate hosts in many endemic areas, such as France, and represent a considerable proportion of the diet of foxes (Guislain and others 2008). Having had no previous exposure, foxes in the UK would be highly susceptible to infection and the cycle could be established rapidly. This contention is supported by the observation than when control of E multilocularis in foxes through an anthelmintic baiting programme is suspended, the parasite rapidly reestablishes in the host population at precontrol levels (Hegglin and Deplazes 2008). The long-term consequences of parasite introduction could be an eventual equilibrium of up to 40 per cent of foxes in the UK being infected, as is the case in many parts of Europe, including high infection rates in urban foxes (see, for example, Deplazes and others [2004]). In contrast to rabies, there is no vaccine for echino coccosis in canids. In Switzerland the numbers of human alveolar echinococcosis cases are increasing in parallel with increasing fox populations and now between 20 and 30 cases are seen annually (Schweiger and others 2007). If the same equilibrium prevalence in foxes in the UK was eventually reached, a similar alveolar echinococcosis incidence in the British population could be expected. As the UK has nearly 10 times the population of Switzerland, this could mean a substantial number of cases of alveolar echinococcosis per year. In Reuben Island (northern Japan), an island that was previously disease free, the first human alveolar echinococcosis cases were diagnosed within 12 years of the introduction in 1924/26 of 24 red foxes from Russia (Ito and others 2003). Without treatment, alveolar echinococcosis is fatal usually within 10 years of onset. Although modern treatment methods offer a good prognosis (Torgerson and others 2008), each case could possibly cost as much as £100,000 per patient (early 2009 exchange rates), assuming treatment costs would be similar to those in Switzerland.

A similar risk assessment for cats was not undertaken. Although there are reports of cats becoming infected (0·25 per cent in Germany [Dyachenko and others 2008]), there is no reliable data on the life expectancy of the parasite in cats. However, experimental studies (Kapel and others 2006) suggest that cats are poor hosts of E multilocularis and hence the risk of the parasite being imported with cats is likely to be considerably lower than with dogs.

To conclude, compulsory praziquantel treatment of all dogs (and cats) before importation into the UK should be continued under the PETS because of the high risk that a few dogs infected with E multilocularis will enter the UK, and the devastating zoonotic consequences if the infections were to establish in the UK wildlife. Such an introduction could result in considerable numbers of autochthonous human alveolar echinococcosis cases in the UK.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.