West Nile virus in Europe: emergence, epidemiology, diagnosis, treatment, and prevention
Corresponding author: V. Sambri, Section of Microbiology, DIMES, Alma Mater Studiorum, University of Bologna, Via Massarenti, 9, 40138 Bologna, Italy
West Nile virus (WNV), a mosquito-borne flavivirus in the Japanese encephalitis antigenic group, has caused sporadic outbreaks in humans, horses and birds throughout many of the warmer regions of Europe for at least 20 years. Occasional cases of West Nile encephalitis have also been associated with infected blood transfusions and organ donations. Currently, WNV appears to be expanding its geographical range in Europe and causing increasing numbers of epidemics/outbreaks associated with human morbidity and mortality. This brief review reports on the current epidemic situation regarding WNV in Europe, highlighting the clinical, diagnostic and preventive measures available for controlling this apparently emerging human pathogen.
West Nile virus (WNV) is a mosquito-borne flavivirus in the Japanese encephalitis antigenic group. It was first isolated in 1937 from a febrile woman in Omogo in the West Nile district of Uganda . WNV has a natural transmission cycle in Culex spp. mosquitoes and wild and captive birds . In contrast, humans and horses are incidental dead-end hosts . The enzootic/epizootic cycle is strictly linked to the period of activity of the arthropod vectors, which, in Europe, results in a season of WNV activity that, depending on the different latitude, ranges from mid-June to mid-November [4, 5]. Since its first recognition in Uganda, the virus has gradually dispersed, via migratory birds, out of Africa to the more southerly regions of Europe, Asia, and Australasia . This review highlights some of the most relevant aspects of WNV human infections in Europe.
Epidemiology of WNV in EU Member Countries and Neighbouring Regions
Up to the mid-1990s, the dispersal of human WNV infections in Europe was mainly associated with sporadic cases . Initially, outbreaks involving significant numbers of cases presenting with neuroinvasive disease were reported in 1994 in Algeria and in 1996 in Romania . WNV was also introduced into the New York area of the USA in 1999, rapidly spreading to four north-eastern states and disseminating outwards in a wave across the entire country over subsequent years . Meanwhile, the virus continued to circulate in Europe and Asia/Australasia . The largest human outbreaks occurred in Bucharest in 1996 (393 hospitalized cases; 17 deaths) and Russia in 1999 (318 human cases; 40 deaths) [7, 8]. Three regions of the Russian Federation—Volgograd, Astrakchan, and Rostov—reported WNV fever/encephalitic activity. Between 1999 and 2010, 928 human cases were identified in the Volgograd area, and 322 and 40 in Astrakchan and Rostov, respectively. Peak epidemic activity was reported in Volgograd in 1999 (380 cases) and 2010 (413 cases) . Concerning other countries in the EU area, cases of WNV encephalitis were initially recorded in horses in Italy (1998) and southern France (2000). In Italy, the virus reappeared in association with human cases of neurological disease in 2008 [11, 12]. In addition, different seropositivity rates have been reported for horses, humans and birds in Spain [13, 14] and Portugal [15, 16], with the first clinical cases of WNV-related disease being reported in Spain for horses and humans in 2010 . Evidence of WNV circulation among birds has been reported in the UK since 2003, and in 2006 the use of sentinel chickens corroborated the epidemiological data about WNV in the UK by demonstrating actual seroconversion [18, 19]. As far as the remaining European countries are concerned, reports indicated that blood donors were seronegative for WNV in Tyrol, Austria  and in The Netherlands , whereas the prevalence of anti-WNV neutralizing antibodies was very low (0.03%) among blood donors in Germany. Nevertheless, a recent evaluation of immunoglobulin preparations produced from human plasma collected in Austria, Germany and the Czech Republic indicated increasing WNV antibody titres in preparations from 2009 and 2010, suggesting past WNV exposure of just under 1% of the population of these three central European countries .
In the last 2 years, WNV provoked large human epidemics in the Balkan area, including Greece, Romania, Croatia, the Former Yugoslav Republic of Macedonia, Kosovo, Montenegro, and Serbia. Additional European countries that were affected by WNV circulation with human involvement during the 2012 season were Italy and Hungary. During 2012, relevant WNV activity was also detected in the Russian Federation, Ukraine, Israel, Tunisia, Algeria, and the Occupied Palestinian Territories, as reported by the European Centres for Disease Control (ECDC) website on clinically apparent cases of WNV in Europe in 2012 (http://www.ecdc.europa.eu/en/healthtopics/west_nile_fever/West-Nile-fever-maps/pages/index.aspx; the most recent report was issued on 30 November 2012). During 2012, North America experienced the worst outbreak of WNV, the total number of human cases rising to 5387 and that of deaths to 243 in the latest set of national data released by the CDC on 11 December 2012. More than half of the cases (2734) were of the more serious neuroinvasive form (data from http://www.cdc.gov/ncidod/dvbid/westnile/surv&control.htm; last accessed on 26 February 2013).
Limited phylogenetic analyses are currently available for WNV isolates circulating in humans in Europe. In particular, WNV lineage 2, which had previously been detected only in birds and vectors in a range of European regions, was identified in encephalitic and WNV fever cases in Russia in 2004, 2007, and 2010 , in a Hungarian goshawk that died of encephalitis in 2004 , and in humans with encephalitic disease during the first outbreak in Greece in 2010 . A complete genome analysis performed in a WNV strain obtained during the Greek outbreak showed high (99.6%) nucleotide identity with the lineage 2 strain that emerged in Hungary in 2004 . Very recently, lineage 2 strains of WNV were also identified in two human cases in distinct Italian locations; these strains were closely related to each other and to those responsible for the outbreaks that occurred in Greece and Hungary in 2010 and 2004, respectively [26, 27]. The genetic diversity among the diverse viral strains identified in vectors and birds over the past 10 years in Europe is wide, including several different phylogenetic clades . Phylogenetic interpretation of these data has suggested two possible explanations for the dispersal patterns and survival strategies of WNV in this region of Europe. First, each year, the viruses circulating in Europe are introduced via migratory birds arriving each spring from Africa, and the introduced viruses do not survive at a significant level in Europe over the winter period. Second, the genetic variation may be the result of selective pressures exerted via innate and acquired immunological pathways in vertebrate hosts on currently circulating viruses. Whether or not either of these interpretations is correct will require further investigation. The emergence of an exotic WNV strain of lineage 2 in a Hungarian goshawk in 2004 emphasizes the role of migrating birds in introducing new viruses to Europe . As the goshawk is not a migratory species, the African WNV strain must have successfully adapted to local mosquito vectors before causing local transmission. As far as the genetic variation of human-derived WNV strains is concerned, three independent studies reported phylogenetic investigation of WNV human isolates in Europe. First, Rossini et al.  showed that seven strains derived from human patients in 2008 and 2009 in Italy clustered closely with two mosquito vector-derived isolates collected in 2008 in northern Italy. Second, Barzon et al. described a novel lineage, 1a (designated the Livenza strain), which correlated with the western Mediterranean subtype causing most of the human cases detected in north-eastern Italy in 2011 and that was divergent from the isolates reported 2 years earlier . More recently, another report showed that two different human isolates obtained in 2011 from Veneto and Sardinia were genetically diverse . These data support the hypothesis of seasonal introduction of WNV variants within Europe.
Most humans infected with WNV remain asymptomatic . Approximately 20–40% of infected humans develop symptoms, the vast majority of which range from a mild flu-like syndrome, West Nile fever (WNF), to severe West Nile encephalitic disease (WNED). This severe condition involves <1% of the infected patients. The neurological disease usually encompasses three different syndromes: meningitis, encephalitis, and acute flaccid paralysis . The clinical characteristics of the symptomatic cases of WNV infection reported in Europe correspond mainly with those indicated above. A recent update on the Greek epidemic reported >250 WNED cases in 2010 and 2011, with a fatality rate of 15% . Between 2008 and 2011, the case-fatality rate reported in Italy for WNED was 16% . The actual case-fatality rate reported by the CDC in the USA ranges from 3% to 15% (http://www.cdc.gov/ncidod/dvbid/westnile/qa/cases.htm; last accessed on 27 February 2013). In 2010, only seven clinically evident cases of WNV infection were described in Italy (four WNF and three WNED) , whereas Barzon et al. reported 13 cases, from the same area in northern Italy, early in 2012 (five WNED, three WNF, and five RNA-positive blood donors, mainly caused by the Livenza 1a autochthonous strain) . In the Volgograd region, 50 cases with neurological involvement were reported among 413 cases in 2010, including five fatalities (Fyodorova, personal communication). One of the most clinically relevant features of WNV infection in humans is the large percentage of infected asymptomatic subjects, which is usually reported being >80% of the total of individuals infected . However, the viral load present in these asymptomatic subjects is not known. Nevertheless, in at least some of these asymptomatic individuals, infectious virus is clearly present and able to be transmitted via blood transfusion and organ transplantation . During the recent past, several cases of WNV transmission related to transfusion of infected blood, solid organ transplants (SOTs) or haematopoietic stem cells have been reported in the USA . Moreover, a recent report demonstrated that WNV-specific IgM antibodies were present in the Italian SOT donor population in 2010, implying a likelihood of WNV transmission via organ donation in the absence of preventive screening procedures . In addition, during the past 4 years, transmission of WNV via infected transfused blood and/or organs has occurred in Europe, where at least two different episodes have been reported (involving six recipient patients) of WNV transmission generated by solid organ transplantion [39, 40]. The clinical diagnosis of WNED is mainly based on the presence of characteristic symptoms that suddenly appear in an epidemiologically exposed individual. On the other hand, most cases of WNF remain undiagnosed or can be categorized as so-called ‘summer fevers of unknown origin’, given the mild clinical presentation, which rarely leads to laboratory diagnosis .
Several different tools are presently available for the laboratory diagnosis of WNV infections . In the case of suspected WNED, the diagnosis is generally performed by the detection of viral RNA in serum or cerebrospinal fluid (CSF) samples with real time RT-PCR assays. Identification of the WNV genome in the CSF or serum during the acute stage of neurological involvement is generally considered to be a confirmatory diagnostic parameter . When viral RNA is not detectable in the CSF, the detection of a specific IgM immune response, either in the CSF or in serum, is accepted as a reliable diagnostic indicator of the early stage of an infection that can subsequently be confirmed by the detection of IgG seroconversion (indicated by a four-fold or greater increase in the level of IgG in the serum) during the convalescent phase of the infection. The detection of WNV-specific antibodies can be achieved by immunofluorescence assay and enzyme immunoassay . One important practical weakness of these techniques is the limited specificity resulting from the widespread immunological cross-reactions among flaviviruses . In order to overcome this lack of specificity, a confirmatory assay performed with the plaque reduction neutralization technique is generally advisable; however, the plaque reduction neutralization technique is complex to perform, requires viable virus isolates, and must be performed under BSL3 safety condition . As a consequence, only a few laboratories in Europe are capable of routinely performing this confirmatory test. The detection of a specific IgG response, mainly by enzyme immunoassay, has value in the context of the epidemiological studies that are warranted to monitor the potentially evolving circulation of WNV among humans in Europe. As an alternative, the isolation of WNV in cell cultures from CSF and serum samples could be performed, but it must be underlined that this is a scarcely sensitive assay, although it is the most specific method, and it requires the continuous availability of cell cultures under BSL3 conditions . For these reasons, the isolation techniques are generally performed only in research laboratories.
Concerning blood and organ donation, two different commercially licensed tests are available for the screening of WNV RNAs in donors [45, 46]. Both methods are fully automated and based on genome amplification techniques: the first uses real time RT-PCR, and the second is based on transcription-mediated amplification technology. These tests have a sensitivity in the range of 10–40 genome copies/mL for the standard WNV preparations derived from lineage 1 American isolates. A major concern relating to the sensitivity of these tests for the detection of WNV in Europe is the well-documented genetic variability of the European strains, which contrasts with American isolates, and the recent emergence of human infections caused by WNV lineage 2 in Europe [26, 29, 31, 47].
There are currently no specific therapeutic treatments for WNV infections , although high-throughput library and compound screening activities are being undertaken in a search for potential inhibitors of WNV infection under the 7th EU Framework grant agreement number 260644. WNED cases usually require intensive-care unit-related interventions and life support therapy, followed by comprehensive rehabilitation cycles, in order to achieve the best recovery . Experimental treatment of severe cases with large doses of either hyperimmune plasma or purified immunoglobulins with elevated titres of antibodies against WNV has been described in selected cases of infection acquired via SOTs . Although such treatments are currently rarely undertaken, there is a body of opinion in favour of encouraging the development of this type of procedure for treating WNED.
Within Europe, the quality of data concerning the circulation of WNV among vector birds and humans varies from country to country. Moreover, there are currently no common surveillance or health policies for the application of control measures in the event of disease outbreaks. Accordingly, as precise definition of viral circulation in vectors and vertebrate hosts, including humans, within defined geographical areas is essential for the definition of the risk of WNV transmission via mosquitoes, blood transfusion, and organ donations, the ECDC recently introduced a web-based publication of the WNV-affected areas (available at http://ecdc.europa.eu/en/healthtopics/west_nile_fever/west-nile-fever-maps/pages/index.aspx; last accessed on 1 December 2012). On the basis of these risk maps and the local surveillance data, each European country should now be able to define the areas and seasons for the implementation of vector control measures and laboratory screening of blood and organ donations, to reduce the risk of human-to-human transmission of WNV. As mentioned, two different commercially licensed tests are currently available for the screening of WNV RNA in donors [45, 46]. As far as the immune prophylaxis of human WNV disease is concerned, no licensed vaccination options for WNV are currently available, although some vaccine preparation candidates have recently been evaluated , and several licensed equine vaccines against WNV already exist [11, 48]. At present, most of the efforts to develop a licensed human vaccine have stalled. A commercial US vaccine preparation went successfully through a phase II trial some years ago, but, for reasons related to market uncertainty regarding the potential target population that should receive the vaccine, the manufacturer decided to stop the development process . The recent reports of the possible development of kidney damage among younger WNV-infected subjects  and the excretion of WNV from the urine of WNED patients years after their apparent recovery from the disease  could create a completely different set of priorities for a vaccine against WNV .
WNV appears to be expanding its geographical range in Europe and in the rest of the world, causing increasing numbers of outbreaks associated with human morbidity and mortality. Multiple de novo introduction of unrelated WNV strains has been demonstrated in Europe, raising concerns about the potential emergence of strains with increased virulence, and the limitations of current diagnostic tests in identifying novel and unexpected WNVs. Given this continuing unpredictability and the rapid development of epidemics, timely surveillance for WNV infection is needed on an EU-wide scale. This includes veterinary and entomological surveillance, as well as molecular surveillance of emerging strains.
V. Sambri has received research funds from Roche Diagnostics and Novartis Diagnostics. This review was supported, in part, by LABP3 from Regione Emilia Romagna and Ricerca Finalizzata RF-2009-1539631 to V. Sambri. Parts of this study were also supported under the Project PREDEMICS (7th Framework Program, Grant Agreement no. 278433). M. Fyodorova is partly supported by the Project ‘European West Nile Collaborative Research Project, EuroWestNile’, Grant Agreement no. 261391.