This report represents a first attempt to estimate the global burden of AE although the global geographical distribution of E. multilocularis
has been reviewed previously (e.g. 
). Throughout much of its geographical range AE is sporadic in humans. In high income countries such as Germany and Switzerland the numbers of cases were the actual number reported (Switzerland) 
. Alternatively in Germanyreported figures were based on a reported capture recapture technique which modelled underreporting 
. These are believed to be accurate reports of the numbers of cases. Reviews of published data also gave estimates for a number of other upper income countries.
For some lower income studies there was limited official data (unpublished) reporting the total number of cases presenting for treatment. However in such countries a major underestimation of the numbers of cases is possible as only relatively wealthy individuals can pay for medical care and the majority of cases may not present for treatment and hence go undiagnosed. A criticism of our approach is that we did not use consistent inclusion criteria for data in different countries. What we did use was the best available data and balanced this inconsistency by using a stochastic approach to model uncertainty. In the countries where we believed the data was accurate a very narrow probability distribution was chosen for the Monte-Carlo routine. In contrast where there was poor data, a very wide distribution was used to model this uncertainty. Hence, the median incidence and estimated DALYs together with the 95% confidence limits give a good estimate of the burden of AE.
In certain districts of China several studies have indicated a very high prevalence of AE through mass screening studies. Surveys have consistently shown a high prevalence of AE using ultrasound studies across Sichuan, Gansu, Qinghai, and Ninxia (). This confirms that there are large numbers of AE cases in China and these represent at least 91% of the global incidence. In some communities the prevalence of AE is similar to that of tuberculosis 
. Because of the large population at risk and the consistent finding of high prevalences we believe the estimate of the number of cases in China is representative. However, the prevalence is not uniform with variations within these districts of between <1% to 12%. The calculations have tried to accommodate these variations in arriving at an overall incidence figure. In the Tibet Autonomous Region (TAR), the incidence could be much higher than the figures suggested. Tibetan communities in neighbouring districts of Sichuan for example have very high prevalences. The parasite is known to be endemic in TAR, but there is no published human surveillance data. We were only able to assume that prevalences in the eastern most part of TAR were similar to prevalences in neighbouring counties of Sichuan or Qinghai. It is possible, therefore, that the true incidence could be thousands rather than the hundreds that are suggested. There is a single case report of cerebral AE in a Tibetan monk who originated from Lhasa which is quite some distance west of the known highly endemic areas of Gansu, Qinghai, and Sichaun 
which is evidence that the parasite has a greater range than our conservative assumption. In Xingjiang the parasite is endemic, but human studies have mainly uncovered cases along the northwest of the province and hence the population at risk and actual case numbers are calculated accordingly.
Russia is a large endemic area for alveolar echinococcosis. In some districts, particularly in Siberia there are reports of a number of human cases whilst elsewhere the disease is sporadic. Detailed data is somewhat lacking despite intensive search of English and Russian language databases. Most data is reported in terms of hospital reports and is therefore estimated as an annual incidence. There are no mass ultrasound surveillance studies as in China although there are a few mass serological studies which tend to confirm the potential for large numbers of cases, especially in Siberia. It is clear, however, that the disease occurs sporadically across almost the entire country. Even in districts where there are no reported human cases there are reports of the parasite in animal hosts so transmission to humans is likely. The estimates for the whole of Russia initially relied on extrapolating data from districts where there were known reports or unpublished data. Bassonov wrote an extensive monograph 
detailing the epidemiology of echinococcosis throughout countries of the former Soviet Union including summaries of otherwise difficult to obtain material. This was also used as a basis for estimating the numbers of cases in Russia. In addition, we were able to access some local Russian language reports including articles in local newspapers and these largely confirmed our assumptions 
. Estimates based on the samples from case series (essentially a type of capture recapture technique) described in the text tended to confirm these estimates. We assumed a maximum ratio of 1
2 for AE
CE. However, one study of surgical cases from Omsk in western Siberia described 84 cases of AE and 44 cases of CE 
. As most of these cases were from Siberian districts such as Kemerova, Altai, Yakutia and Tomsk which could indicate higher numbers of AE than the ratios we used in our calculations.
The mountainous regions of Kazakhstan, Kyrgystan, Uzbekistan, and Tadjikistan are all endemic for E. multilocularis 
. Here the cycle of infection has been well described in terms of the animal hosts. The most accurate figures available are from Kyrgyzstan where approximately 35 cases per years are now being reported in the hospitals (unpublished figures from the Government Epidemiological Surveillance Unit, Bishkek). This figure is likely to be accurate in terms of numbers being treated. However, it may underestimate the actual numbers of cases occurring as the country is poor and expensive medical treatment is not available to a large part of the population.
Although ethnically distinct, in terms of geography and economy, Tadjikistan is very similar to Kyrgyzstan. Therefore, as it is also endemic for E. multilocularis, similar numbers of cases per year could be expected as there is likely a similar population size at risk. Kazakhstan and Uzebkistan are much larger countries in area and population but as much of their territory is in low or non-endemic areas the proportion of population at risk is smaller, but absolute numbers are similar. Turkmenistan is likely to have few cases of AE, as much of the territory is arid desert which is inimical to transmission. Mongolia has a number of reported cases. From unpublished data there were 9 cases of AE in both 2006 and 2007.
India, Nepal, Bhutan and Pakistan border on endemic zones and may have a few cases. The disease has been reported in India (Kashmir) 
. For Afghanistan, data is not available. However there is a case report of a patient originating from Afghanistan who was treated in the UK (which is non endemic) 
. Therefore, further cases are likely, especially in the north of the country. Iran is endemic for AE but there is little data. Between 1948 and 1993, 37 cases of AE were reported 
or less than 1 case per year. In view of the fact that neighbouring Turkey is known to be highly endemic this is likely underreported. There is a single case report of AE from northern Iraq 
Turkey is highly endemic for echinococcosis. In total, approximately 3000 cases are recorded annually. The number of cases of AE is uncertain. The proportion of cases of echinococcosis that are AE have been reported in a number of case series (). Assuming a similar ratio of AE to CE nationwide as found in the case report series this would suggest as many as 500 cases of AE per year. However, such selected case series may overestimate the true incidence of AE and these data contrast somewhat with the report of 206 cases of AE in Turkey for the period 1980–2000 or just 10 cases per year 
. AE has only recently been made reportable in Turkey 
and so more accurate figures are unavailable. Also of interest is a case finding study which specifically searched through the records of 47 neurosurgical units for cases of cerebral echinococcosis. This study found a total of 219 cases of intracranial echinococcosis in the five year period 1994–1999 
. Of these 16 were AE and 2 were AE with no extra CNS involvement. Another report found 4 CNS cases of AE in just 27 months at one centre 
. Intracranial AE is thought to be rare with CNS involvement usually a manifestation of metastases from a primary lesion. A large study in China suggested that 4% of AE cases had neurological involvement 
. Likewise, a large European study found 17 of 559 (3%) cases of AE had brain involvement 
. The incidence of diagnosed intracranial cases of AE therefore gives strong evidence there must be at least 100 cases of AE per year in Turkey. This type of approach may, nevertheless, substantially underestimate AE cases in resource poor endemic regions, as a smaller proportion of AE cases my receive hospital treatment then are actually present in the community. For example, in south Ningxi in China, CE represented 96% of hospital treated cases of echinococcosis, but ultrasound studies in the local community suggest that 56% of cases of echinococcosis are AE 
The USA and Canada are endemic for E. multilocularis
. However there is very little transmission to man. There have been reports from Native American communities of high incidence rates in Alaska 
but only locally and these have been eliminated by appropriate intervention programs. Otherwise, there have only been single cases reports in Minnesota 
and Manitoba 
. Thus there is very little evidence for autochthonous human cases presently in North America despite the active transmission in animal hosts.
Eastern Europe has highly variable data. The parasite is endemic in most of the former Soviet States. Lithuania has the best data and reports incidence rates 
and it is likely that the other Baltic States have similar incidences due to similarities in culture, geography, and population. From central and western Europe there is usually high quality data giving details of the numbers of cases (). The core endemic area is centred on Switzerland, southern Germany, and eastern France.
In North Africa, there have been two reports of autochthonous AE from Tunisia and a further case from Morocco diagnosed by histological examination of the lesions 
. However the parasite has never been recorded in animals from Africa and in the absence of molecular confirmation there is insufficient evidence to confirm any part of North Africa is presently endemic for E. multilocularis
AE is a serious disease and, although the prognosis is reasonably good when treatment is available 
, the prognosis is equally bleak in the absence of treatment. The overwhelming number of cases comes from an area of rural China which forms part of the Tibetan plateau. This population is remote and with few financial resources with an estimated annual income per head of less than US$500 
. Therefore it is reasonable to assume that most of these cases will be fatal and hence, the annual mortality due to AE is similar to the incidence. This is also likely to be true of most other cases outside of Europe. An annual mortality due to AE of approximately 18,000 is greater than one tenth of the total mortality of 177,000 as a result of the 10 diseases of the neglected tropical disease cluster (trypanosomiasis, Chagas disease, schistosomiasis, leishmaniasis, lymphatic filariasis, onchocerciasis, intestinal nematode infections, Japanese encephalitis, dengue, and leprosy ) 
. The disease burden can also be compared to that of rabies. Annual AE mortality is approximately one third of that due to rabies which has been estimated at approximately 55,000 
. Nearly all is from Africa and Asia. Unlike rabies, there is also no vaccine for canid echinococcosis and control requires repeated treatment of foxes or dogs with praziquantel. Thus, although AE is rare on a global scale it has a high burden in some highly endemic communities in China where it is likely to be one of the leading causes of death.
Likewise the global burden of disease, in terms of DALYs, is high. This is again due to the very high fatality rate resulting in a large number of YLLs but additionally due to the expected high disability weight that individuals have during the course of the disease. Diseases with similar magnitudes of DALYs include neglected tropical diseases such as onchocerciasis, and Chagas disease 
. An initial estimate of the global burden of cystic echinococcosis was approximately 1 million DALYs 
. However, this is likely to be an underestimate 
and a re-evaluation of the global burden of CE is ongoing.
Control of this disease depends on the risk factors for transmission. The wild life cycle can be disturbed through the treatment of foxes with praziquantel impregnated baits 
. However, the feasibility of applying such control measures over large parts of the Tibetan plateau, where the main burden of AE is found, would be questionable. Dog contact is a known risk factor for transmission to man 
and dogs are highly susceptible to infection with this parasite 
. However, it is not known if dogs participate in the cycle or are aberrant definitive hosts. If they are aberrant hosts, periodic treatment of dogs will not disturb the transmission cycle and will have much less effect on long term transmission rates to humans 
. Other means of reducing the disease burden would be through better control of food or water supplies which may be contaminated with parasite eggs. Such control would be dependent on the attributable fraction of disease burden due to these transmission pathways and the cost effectiveness of such intervention strategies.