Worldwide literature on epidemiology of human alveolar echinococcosis: a systematic review of research published in the twenty-first century

We performed a systematic literature search for worldwide relevant publications in the bibliographical databases MEDLINE (via the PubMed metasearch engine), EMBASE (via the OVID metasearch engine), the Chinese databases CNKI, VIP, and the Russian Scientific Electronic Library (via eLIBRARY.RU). These searches were supplemented with the Russian open access repository CyberLeninka, the Farsi language-based database Journals.research.ac.ir, the Arab language-based database of the University of Jordan E-Library and the web search engine Google Scholar.

The countries for our area-specific search strategy were selected after an initial screening of general reviews on E. multilocularis distribution, and its neighbouring nations. All the internationally relevant keywords for the disease were linked with the Boolean operator “OR”. The search key was designed to be as narrow as possible to ensure goal-oriented results but at the same time broad enough to capture all the relevant world literature. Search key optimisation was carried out by analysing the search details of each search term and subsequent pilot testing. The width of the search came primarily through automatic term mapping, the automatic generation of a more detailed search string, which also covered Medical Subject Headings (MeSH) terms (e.g. the MeSH term “Echinococcus multilocularis”). MeSH is the controlled vocabulary thesaurus generated in MEDLINE. A similar procedure was carried out with EMBASE (Emtree). The Boolean operator “AND” was used to add the country to be screened and the corresponding adjective, as well as any possible ethnonym, superordinate region (e.g. “Slavic” or “Baltic”) or subordinate region (e.g. “Alaska”, “Tibet”) to the end of the general search string. This resulted in PubMed search keys such as the one for France: Echinococcus multilocularis OR echinococcus alveolaris OR alveolar hydatid disease OR alveolar hydatid cyst OR alveolar hydatidosis OR alveococcosis AND (France OR French). Truncation, double quotes, and also the search term “alveolar echinococcosis” did not lead to a higher number of search results. Alongside the terms in Latin letters, synonyms in Chinese, Cyrillic and Arabic script were also searched through the suitable data bases.

The search was restricted to articles that were published from 2001 to 2018. There were neither restrictions in terms of language, place of publication, nor the time of the initial AE diagnosis; therefore, it can be dated before 2001. Overall case numbers included non-autochthonous cases, which were given in parenthesis in the final synopsis (e.g. a total of 65 cases, two of which were considered non-autochthonous, was given as “65 (2) cases”). With respect to the assumed infection locality, we used the information given by the authors. The searches were carried out between 01 June 2017 and 15 October 2017, February 2018, and between 14 January and 15 March 2019.

In order to be included in the final literature list, we established a quality assessment following previously defined including and excluding criteria for the collected data.

Inclusion criteria:

Exclusion criteria:

Qualification for the final literature list was carried out in two steps. First, we inspected all the articles found in the search results and applied the defined criteria. Then, we looked closely at all references that appeared relevant in each article. All the literature then discovered was inspected in the same way. We repeated the procedure until no more relevant information was generated (snowball method). The data obtained were stored in an Excel table (Microsoft Office 2017, version 15.30) and divided into the following categories: country, subordinate region (administrative unit), paper (ID, lead author, year of publication, title, journal, volume, pages), period of data collection, epidemiological data (case numbers, prevalence, incidence), case definition (serology, US, CT, MRI, histopathology, nucleic acid-based testing), and non-autochthonous cases.

A multilinguistic team of researchers screened the articles. Two researchers (SB, physician; JS, epidemiologist) independently inspected the literature according to the inclusion and exclusion criteria. In general, the full article was screened, unless the abstract was not clearly leading to an exclusion (e.g. the article was only about CE cases). All non-English articles were analysed in cooperation with native speakers. Literature in the Chinese language was independently screened by WXL, physician and RS, physician. Any uncertainties about the inclusion of an article were discussed, and if a consensus was not met, WK, physician, and HB, physician, were consulted to obtain it. An attempt was made to contact the corresponding author whenever there were any unresolved questions regarding the period of data collection or the case definition.

We used QGIS software (version 2.18.21) to generate the world map. Each country in which cases of AE had been reported in the literature between 2001 and 2018 was mapped. For the topographical colour shading of a country, the highest total number of cases in one reference within this period was the deciding factor.

This systematic overview follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure a transparent study [15]. The corresponding checklist is attached as supporting information. The search protocol was entered into the International Prospective Register of Systematic Reviews (PROSPERO) under Registration number CRD42017079097.

Based on reports from 13 Asiatic countries published since 2001, the epidemiological AE zone stretches across the north of the continent from Turkey to Japan, but with considerable gaps. The highest mean incidence of 7.1/10⁵/year was calculated in Oblast Nary, Kyrgyzstan (2010–2011) [16]. The highest prevalence of 9.43% was found in Banma County in the province of Qinghai, China, in 2014 [17]. The highest absolute number of cases, 3028 patients, came from an US screening in Shiqu County, Sichuan Province (2015–2017) in China [18].

From Turkey, there were five publications with data prior to 2005; in Southeastern Anatolia, the prevalence was calculated to be 0.4/10⁵ in the year 2000 [19]. In central Asia, we found literature from Kazakhstan, Kyrgyzstan and Tajikistan, however no original papers concerning Turkmenistan or Uzbekistan. Reporting the non-autochthonous case from Iran, Maddah and co-authors described the patient as being of Turkmen origin [20]. In the Kyrgyz Alay district, a study has been demonstrated a prevalence of 6.4% in 2012 [21]. From the west of Iran, we found a publication describing 18 cases. There was no literature describing possible Iraqi patients. Noticeably, a paper from the UK showed an Afghan patient who migrated from Pakistan; the authors discuss the infection originating in Afghanistan [22]. There was evidence of isolated cases in India and, as far as we know, the first case was reported in Pakistan [23‐30].

Apart from Kyrgyzstan, by far the highest prevalences of human AE were reported in China. Reports have been related exclusively to Western China, namely to the provinces of Qinghai, Gansu, Sichuan and the autonomous regions Xinjiang, Tibet and Ningxia. The highest prevalences were reported from counties of the eastern Tibetan Plateau, ranking up to 9.43% in Banma County (Qinghai, July–August 2014) [17] and 8.5% in Shiqu County (Sichuan, 2001–2002) including a prevalence of 15% in one of the 11 villages being studied [31]. In further administrative divisions of China, the highest prevalences were 6.36% in Xinjiang (2004) [32, 33], 5.2% in Tibet Autonomous Region (TAR) (2007) [34], 3.7% in Ningxia (1985–2001) [35] and 3.39% in Gansu (1996–1997, 2003) [36]. Up north in neighbouring Mongolia, four cases in the west of the country were confirmed by histopathology and molecular genetic testing; the rare Mongolian haplotype was identified in two of these cases [37]. In Japan, it is assumed that nearly all human infections occurred on the northern island of Hokkaido [38]; one paper calculated the mean incidence for the whole of the country to be 0.013/10⁵/year in the period from 2000 to 2005 [39]. The first case from South Korea was reported in a woman who had apparently never left the country [40]. The first AE reports in Taiwan and Thailand were considered to be non-autochthonous by the authors [41‐43].

In the twenty-first century, AE has been reported in 20 European countries, although the cases in the United Kingdom and Denmark were considered to be non-autochthonous by the authors. A (mean) incidence was calculated in six countries, the highest being in the Austrian Federal State of Vorarlberg in 2011 at 1.9/10⁵/year [44]. The incidence was up to 0.76/10⁵/year in France (in Doubs 1982–2009), up to 0.54/10⁵/year in Lithuania (2013), up to 0.26/10⁵/year in Switzerland (2001–2005), up to 0.20/10⁵/year in Poland (in Warmia-Masuria Province, 1990–2011) and 0.09/10⁵/year in Slovenia (2001–2005) [45‐50]. The prevalence was calculated only in a German and a Slovenian study, with the highest German figure of 2.18/10⁵ in the Federal State of Baden-Württemberg (1992–2016) and 0.45/10⁵ for whole Slovenia in the period from 2001 to 2005 [50, 51]. Highest case numbers were registered in the national databases in France (575 from 1982 to 2013) and Germany (523 in 1992–2016) [51, 52]. The first case thought to be autochthonously acquired in the Netherlands was reported in the province of Limburg [53, 54]. The first reported cases from neighbouring Belgium appeared nearly exclusively in Wallonia [55, 56]. Also for the first time in Hungary, a case of AE was thought to be autochthonous in the south-west of the country [57]. Evidence of the disease was found for the first time in Poland and in five patients in north-eastern Romania by molecular genetic testing [58, 59]. To the best of our knowledge, the first reported cases of human AE also originated in Slovakia, Lithuania, Latvia, and Slovenia [47, 50, 60, 61]. Since 2007, 20 cases were registered by the Czech National Reference Laboratory for Tissue Helminthoses [62]. Of the 26 confirmed AE cases in Slovakia between 2000 and 2013, Antolová et al. found that 23 of them occurred in the north-west of the country, in the Žilina and Prešov regions [63]. In Belarus, there were five case reports described in Grodno, a city in the border area to Poland and Lithuania [64], and one post-mortem diagnosis out of Gomel Oblast [65]. From Russia, the highest figure of 42 AE cases was reported in Tomsk Oblast until 2012 [66]. Literature focussing on human AE cases in Ukraine could not be found. In Southern Europe, single cases were documented in the north Spanish province of Navarre and in Greece [67, 68].

Figures from the Canadian Institute for Health Statistics were published in two articles from Canada, including a total of 12 cases in the southern states of British Columbia, Alberta, Saskatchewan, and Ontario for the period 2001–2014 [69, 70]. In the USA, molecular genetic analysis of the sample from a case reported in Minnesota in 1977 gave a 99.9% agreement in sequence homology with an E. multilocularis isolate of a fox in South Dakota and 99.4% agreement with a human sample from Japan [71].

In the literature from 2001 to 2018, we found only a single case report of a 54-year-old Moroccan man. This was the first reported case of AE in Morocco [72].

A huge increase in case numbers was seen in central Asia, especially in Kyrgyzstan, the only region in the world where prevalence dimensions were published which otherwise only could be found in China [21, 78]. While only 0–3 cases per annum were recorded in the period 1996–2003, numbers rose continuously from 2004 onwards, reaching 61 cases in 2011 [16]. And more than twice that number of patients was recorded in 2013 (148 cases) [79]. Omorov and Co-Workers collected 1179 AE cases (1996–2015) from nine different Kyrgyz institutions, though diagnostic criteria have been shown for 581 patients [80]. Possible reasons for the upsurge were the improved medical care and diagnostic investigation after the difficult economic period following the dissolution of the Soviet Union and the increasing spread of E. multilocularis-infected stray dogs [16, 81]. In Kazakhstan, one hospital in Almaty found that the recorded cases of AE more than doubled from 15 cases in 2004–2011 to 32 cases in 2012–2014 [82]. For the first time, AE cases were published from Tajikistan, where 22 patients were diagnosed from 2010 through 2013 [83]. Only a congress contribution, referring to 83 surgical AE patients missing diagnostic criteria, has been suggested the presence of the disease in Uzbekistan [84].

There is a distinct lack of studies from western Asia, with only estimates of a few AE cases per year existing for Armenia, Georgia and Azerbaijan [5]. Current epidemiological studies out of Turkey are missing. One publication reported 202 AE patients (1980–1998) for the whole country; no diagnostic criteria were given [182]. There are merely sporadic data or none at all from the countries that stretch across central Asia towards China (Iraq, Iran, Afghanistan, Pakistan, and India), so that we have to assume an under-representation [20, 30]. One case report from Iraq suggesting E. multilocularis as the causative agent has obviously not been presenting an AE patient [85], as the published CT figure showed the morphologic criteria of a CE lesion (WHO-type 3b) [Tilmann Graeter, personal communication].

By far the highest prevalences, apart from the discussed Kyrgyz numbers, are still to be found in China. Overall, the figures are unchanged and no recent spread, increase, or decrease of case numbers has been described by authors publishing Chinese data. The high prevalence in schoolchildren is particularly remarkable, as paediatric cases of disease are absolute rarities in other parts of the world [86‐88].

In recent years, there has been both a spread of AE reports across Europe and an increase in case numbers. Until the end of the 1980s, the disease was considered endemic only in the ‘historical’ AE area of western Europe (France, Germany, Switzerland, and Austria) [89], but human infection has already been reported in 20 European countries since 2001. Only the cases in Denmark and the United Kingdom were considered to be non-autochthonous in the respective publications [68, 90, 91]. Figures for all the affected western European countries have shown an increase. In France, according to the FrancEchino Register, the number of new cases per year in the period 2003–2012 (239 cases) almost doubled in comparison with the previous 10 years (122 cases); in particular, there was a significant increase in the incidence of AE in immunosuppressed patients [92, 93]. The mean incidence in Switzerland rose from 0.1/10⁵/year (1993–2000) to 0.26/10⁵/year (2001–2005) [48]. Austria had an average of 2.4 cases in 1991–2000, 2.8 in 2001–2010, and a sudden unexpected rise to 13 cases in 2011 alone [44]. Germany showed a progressive increase in the figures over five-year periods, with 97 cases (2002–2006), 107 cases (2007–2011), and 165 cases (2012–2016) [51]. In addition, AE seems to have spread to previously non-endemic areas such as Belgium and the Netherlands [54‐56].

Increasing figures were reported in the literature for all affected eastern European countries. In the five-year periods of the 1990s, the highest number of AE cases detected in Poland was 10, rising to more than 20 cases in 2000–2004, and then over 55 cases in 2005–2009, with a cluster in the north-east of the country bordering Lithuania [49]. Interestingly, five Belarusian case reports were described in a hospital, close to the border area of Poland and Lithuania [64]. In the region of Brest, there are partially unpublished data regarding eight registered AE patients since 1995 (Alla Korzan, personal communication). In Slovakia, there were 11 confirmed cases from 2000 to 2009, but already 15 cases in 2010–2013 [63]. No Czech cases were registered between 1998 and 2006, however 20 cases in the period of 2007–2014 [62]. We can assume the further spread of the disease from the first molecular genetic evidence for the existence of human AE in Romania, as well as the first autochthonous case from south-western Hungary [57, 59]. Despite some high estimated figures of more than 1000 new cases annually [5], there is a distinct lack of data from Russia. Besides numerous single-centre studies with preselected study groups undergoing surgery, epidemiological data could only be found from Moscow [94, 95], the south of the Volga and Ural Federal Districts (neighbouring Republic of Bashkortostan and Chelyabinsk Oblast) [96, 97], the southwest of Siberia Federal District (Tomsk Oblast, Altai Krai) [66, 98] and the Far East Sakha Republic and the peninsula of Kamchatka [99, 100]. Nevertheless, the official registry of the Russian federal agency Rospotrebnadzor reported higher figures in 2008 (41 cases) than in 2001 (30 cases) [13, 98].

Knapp and co-workers studied the genetics of the observed spread to eastern Europe. They used the EmsB microsatellite marker to analyse the genetic diversity of E. multilocularis in various European endemic areas. They found the lowest diversity in Slovakia and Poland and the highest in Switzerland and the Swabian Jura, arguing for the two latter regions being the oldest endemic areas in Europe in evolutionary terms and for a ‘mainland-island’ system governing pathogen transmission [101]. In northern Europe, the number of cases in Lithuania rose: the incidence increased from 0.03/10⁵/year in 2004 to 0.57/10⁵/year in 2009 and 0.74/10⁵/year in 2012, exceeding all the overall national incidences in Europe published since 2001 [47]. Six Swedish AE patients with assumed infection abroad were officially reported in the Public Health Agency in 2014 and 2017, with the first two diagnoses in 2012 [102, 103]. The European Food Safety Authority registered three cases in Estonia in 2013 (no information about site of infection or diagnostic criteria) [104]. In Finland, there are unpublished data about a native patient with PCR-confirmed AE and a travel history to endemic destinations in Europe (Antti Lavikainen, personal communication) [105].

In Southern Europe, Slovenian AE cases have been published for the first time. Although there have already been cases of the disease reported in Spain, E. multilocularis has never been demonstrated in the wildlife to date, so there is not sufficient evidence for the endemicity of the parasite [67, 106]. In Greece, a patient of Macedonian origin living in Thessaloniki was registered via the European Echinococcosis Registry (Petra Kern, personal communication) [68]. Since its reporting in the late 1990s, no new cases were documented out of this area. In summary, we can say that without exception, we found an increase in reported case numbers in all significant endemic European countries (i.e. countries from which there are at least 20 reported cases of AE). There are three relevant hypotheses to explain this rise in Europe.

First of all, there is an increase in the red fox population in Europe, which is also related to the elimination of rabies, together with higher infection rates with E. multilocularis [63, 106, 107]. One Swiss study showed a direct correlation between the growing fox population and the increase in human echinococcosis [43]. The observed increasing urbanisation of the fox habitat is also noteworthy [108‐110]. Secondly, improved awareness of AE in the general population and healthcare workers may also have contributed to more cases being diagnosed or fewer incorrect diagnoses being made [111]. Thirdly, the possibilities for diagnostic investigation have improved considerably in the last 20 years, particularly with respect to imaging and molecular genetics [8, 63].

If we compare our findings with the current literature on the distribution of E. multilocularis in European red foxes, there are several countries in which the parasite has been detected in red foxes in this century but not in humans (Serbia, Croatia, Italy, Luxembourg, and Ukraine) [106, 112]. Human cases in these countries probably have to be reckoned with in the future.

AE is extremely rare in North America, even though the infection rates recorded in animals are relatively high, e.g. 44.6% of foxes in northeast Nebraska and 35.3% of coyotes in Illinois [113, 114]. Despite these high figures, only two new cases have reliably been diagnosed in humans in the USA since more than four decades [198]. However, the literature was supplemented by two native individuals from the states of Alaska and Washington who were registered as deaths to AE on a death certificate in the National Center for Health Statistics (NCHS) and, therefore, convincing evidence is still lacking [115, Ben Bristow, personal communication]. From Alaska, where 54 human infections have been reported from 1947 to 1986 [reviewed by 116], no confirmed subsequent cases were found. In Canada, AE cases are thought to be (predominantly) non-autochthonous [69, 70]. Interestingly, a public health report of the government of the province of Alberta suggests some diagnosed autochthonous cases since 2013 [117]. Furthermore, the Canadian Institute for Health Information has documented at least three human infections in Ontario between 2014 and May 2018 [118] and a recent review of Wen and co-workers has mentioned unpublished case reports in Quebec and Manitoba [119]. Besides the possibility of misdiagnosis, one hypothesis to explain the discrepancy between the high infection rates in the wildlife and the extreme rarity of autochthonous cases in North America is the genetic expression of E. multilocularis in these areas combined with a human population of correspondingly low susceptibility. To the best of our knowledge, in fact, there has been only one proven case of human infection with the North American haplotype so far, detected in a DNA-based analysis of a sample of a patient diagnosed in Minnesota in 1977 [12, 120]. Remarkably, since 2009, multiple cases of animals infected with the European-type strain have been documented in Canada, including dogs infected as intermediate hosts. Unlike the North American strain, this strain is typically associated with human disease. We can, therefore, speculate that the European strain could have become established in the region and any human cases in Canada in the future may indeed be autochthonous in nature [121, 122, Janna Schurer, personal communication].

As with the case in Spain, despite the transparency of the diagnostic investigations in the case report from Morocco, E. multilocularis has not been confirmed in North African wildlife and there is not yet any concrete evidence of its presence [5, 8].

On the basis of the available literature, it is currently not possible to obtain a valid list of the worldwide prevalence of human AE without many gaps, as both the quantity and quality of the published data are insufficient for the purpose. The reasons for this are manifold. AE is a notifiable disease only in some places, e.g. in most European countries but, on the North American continent, only in the Northwest Territories and Ontario in Canada [105, 111, 118]. AE is frequently not distinguished from CE, even though the latter is a distinct disease entity with different transmission profiles, risk factors, and clinical manifestations, requiring quite different control and surveillance measures and treatment [1, 123]. Out of the echinococcosis cases officially notified to the European Union in 2013, 31.7% (253 cases) did not differentiate between AE and CE [124]. Furthermore, due to the initial asymptomatic period of 5–15 years, it is usually not possible to pinpoint the precise location where the parasite was ingested [75]. In addition, most entries in the registries do not have a case definition, i.e. are not based on firm diagnostic criteria, which makes it more difficult to compare the recorded data. If there was a case definition given by the authors, we could not verify its quality by reviewing the defining imaging data or other diagnostic criteria.

Because of the long incubation period, we have to suspect the existence of a large proportion of asymptomatic individuals, who have also not been recorded. There are probably also considerable numbers of symptomatic patients living in poor economic and/or remote areas who remain undiagnosed and are therefore not included in epidemiological registries or corresponding studies [8, 75].

One limitation of this article is the lack of comparability of the units in the data. The prevalence or incidence is only rarely calculated in the scientific papers, which makes the interpretation of case numbers more challenging.

Another limitation can be found in the fact that, although the literature was published in 2001–2018, the time frame of the diagnosed cases ranged from 1937 up to 2017, giving the data a temporal inhomogeneity. Even though an increase of reported cases as well as a spread in several mentioned areas has been confirmed by this work, a significant bias cannot be excluded; a statistical analysis did not seem reasonable due to the inhomogeneity of the collected data. Due to recent improvement of diagnostics, as discussed above, an increase in data quantity can be assumed [8, 63].

Following our inclusion and exclusion criteria, it can be supposed that grey literature of interest is missing in our synopsis. We discussed some of these sources above; adding those cases from Uzbekistan, Sweden, Finland and Estonia, AE could be assumed to have been reported in overall 40 countries in the twenty-first century. However, only in 36 of those nations, publications which meet some basic quality criteria were existent.

With respect to the mapping, the main limitation is the worldwide lack of valid data, insufficient even for topographical interpolation estimating the borders of endemic disease areas and the corresponding prevalences.

From the epidemiological point of view, a national obligation to report AE as a notifiable disease, including its differentiation from CE, would be desirable in endemic countries. The data should be entered into national AE registries, which should be standardised and coordinated on an international level to generate comparable datasets and ultimately ensure high validity [125]. In addition, every effort should be made to achieve a precisely defined uniform terminology relating to echinococcosis and its pathogenic agents.