Incidence of Lyme Borreliosis in Europe: A Systematic Review (2005–2020)

Background: Lyme borreliosis (LB) is the most common tick-borne disease in Europe, but the burden of disease is incompletely described. Methods: We conducted a systematic review across PubMed, EMBASE, and CABI Direct (Global Health) databases, from January 1, 2005, to November 20, 2020, of epidemiological studies reporting incidence of LB in Europe (PROSPERO, CRD42021236906). Results: The systematic review yielded 61 unique articles describing LB incidence (national or subnational) in 25 European countries. Substantial heterogeneity in study designs, populations sampled, and case definitions restricted data comparability. The European Union Concerted Action on Lyme Borreliosis (EUCALB)–published standardized LB case definitions were used by only 13 (21%) of the 61 articles. There were 33 studies that provided national-level LB incidence estimates for 20 countries. Subnational LB incidence was available from an additional four countries (Italy, Lithuania, Norway, and Spain). The highest LB incidences (>100 cases per 100,000 population per year [PPY]) were reported in Belgium, Finland, the Netherlands, and Switzerland. Incidences were 20–40/100,000 PPY in the Czech Republic, Germany, Poland, and Scotland and <20/100,000 PPY in Belarus, Croatia, Denmark, France, Ireland, Portugal, Russia, Slovakia, Sweden, and the United Kingdom (England, Northern Ireland, and Wales); markedly higher incidences were observed at the subnational level (up to 464/100,000 PPY in specific local areas). Conclusions: Although countries in Northern (Finland) and Western (Belgium, the Netherlands, and Switzerland) Europe reported the highest LB incidences, high incidences also were reported in some Eastern European countries. There was substantial subnational variation in incidence, including high incidences in some areas of countries with low overall incidence. This review, complemented by the incidence surveillance article, provides a comprehensive view into LB disease burden across Europe that may guide future preventive and therapeutic strategies—including new strategies on the horizon.

Geographic variations in LB disease burden reflect likelihood of exposure to infected ticks. Thus, the incidence of LB is higher in areas where there is an abundance of vertebrate animals that serve as reservoirs for infected ticks (Lindgren and Jaenson, 2006). LB is also more common in persons who undertake outdoor occupations or leisure activities that increase the risk of exposure to tick bites in Bbsl-endemic areas (Magnavita et al., 2022).
After infection of Bbsl, spirochetes burrow between tissues, bones, cells, joints, and nerves, and they can cross the blood-brain barrier into the central nervous system. There are various clinical manifestations of disease (Kullberg et al., 2020, Lantos et al., 2021, Marques et al., 2021, Stanek et al., 2011. Within days to weeks, Bbsl disseminates from the tick bite site to other body regions causing early localized infection and often-although not always-erythema migrans (EM) (Steere et al., 2016). Sometimes, Borrelial lymphocytoma also develops. After weeks or months, Bbsl uses a chemotaxis machinery system to disseminate into the host, causing early disseminated LB or late disseminated LB. Clinical manifestations of the disseminate form of the disease include: Lyme neuroborreliosis (LNB), Lyme carditis (LC), Lyme arthritis (LA), and acrodermatitis chronica atrophicans (ACA) (Aucott et al., 2009, Bernard et al., 2019, Stanek and Strle, 2018, Stanek et al., 2012, Steere et al., 2016, Verhaegh et al., 2017. Seroconversion can occur with or without clinical symptoms (Kullberg et al., 2020, Marques et al., 2021. A diagnosis of LB is first assessed clinically (Stanek and Strle, 2018, Stanek et al., 2012, Steere et al., 2016. After clinical evaluation, LB diagnosis is often supported through laboratory testing that can include serology, including detection of specific intrathecal antibodies, and microbiological examination of infected tissue in patients with suspected LNB. Serum antibody tests include enzyme immunoassays, immunofluorescence assays, enzyme-linked immunosorbent assays (ELISAs), or Western blots, and combinations of these. These tests vary in sensitivity and specificity, which can impact the likelihood of diagnosis of LB, and this has wide consequences for studies of LB epidemiology; antibody detection is not always the equivalent of disease (Kodym et al., 2018, Leeflang et al., 2016. The transient presentation of EM, the broad spectrum of clinical presentations observed in patients with early and late disseminated disease, and complexity of serological testing have impeded the development of generally accepted standard case definitions for LB, both in public health surveil-lance and in epidemiological studies Strle, 2018, Stanek et al., 2012). In the absence of widely accepted standardized case definitions for LB in Europe, the European Union Concerted Action on Lyme Borreliosis (EUCALB) published case definitions for manifestations of LB in 1996 and updated them in 2011 in an effort to encourage Europewide implementation (Stanek et al., 2011). To date, however, very few LB surveillance systems in Europe implement the EUCALB definitions (also see Nagarajan et al., in this edition).
The European Centre for Disease Prevention and Control (ECDC) has employed a reportable LNB case definition to detect and monitor LNB cases in Europe (European Centre for Disease Prevention and Control, 2018;European Commission, 2018;The Lancet, 2018), yet how often it is utilized across countries remains unknown. LNB is a clinical manifestation that represents the more severe, disseminated form of the disease involving systemic involvement (Radolf et al., 2021, Rauer et al., 2018, Trevisan et al., 2020. It is an appropriate indicator for surveillance given its high specificity in diagnosis and reproducibility in measurement (Stanek and Strle, 2018;The Lancet, 2018;Van den Wijngaard et al., 2017).
With EM as the leading indicator, and LNB to measure more severe forms of the disease, surveillance of other clinical manifestations periodically could also add value to surveillance and epidemiological studies to provide full insights of LB epidemiology in Europe ( Van den Wijngaard et al., 2017). Understanding the population-based incidence of LB is important for targeting and evaluating LB prevention strategies, which could potentially include vaccination. Because of the inherent limitations in LB surveillance (Stanek et al., 2011, Van den Wijngaard et al., 2017and Nagarajan et al., in this edition), more complete epidemiological studies may provide complementary data to understand LB incidence. Recent systematic reviews of LB have attempted to quantify LB incidence but have been limited to Western Europe (Sykes, 2014, Vandekerckhove et al., 2021. We conducted a comprehensive systematic review across all of Europe to understand the national-and subnational-level incidence of LB reported in the published literature over the past 15+ years.

Methods
The methodology, search strategy, and inclusion and exclusion criteria for the systematic review and analysis are included in a protocol developed by the Lyme Review Group, which included experts in Lyme vaccine development, clinical epidemiologists, and statisticians. The protocol was based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA, 2020) guidelines and was registered in PROSPERO (CRD42021236906). The protocol is for multi-objective review for a global study on LB. For the purposes of this article, we focus our scope on incidence estimates of LB in Europe.
All citations were merged into a database, and duplicates were removed. Titles and abstracts were screened independently by two reviewers for their relevance to the study objectives. Selected full-text articles were assessed based on predefined inclusion and exclusion criteria, by two reviewers. Full-text articles published in other European languages were translated into English using DeepL (DeepL SE, 2022). For articles that were not easily translated in DeepL, such as Finnish, we utilized P95/Pfizer colleagues who were native speakers or fluent in these languages to check and elaborate on translation accuracy. Relevant variables from selected articles were extracted into DistillerSR (Evidence Partners, 2021). A reviewer independently checked 20% of the articles and their extractions. All discrepancies identified during each phase were discussed and resolved.

Inclusion and exclusion criteria
We selected articles (obtained from our search) reporting LB incidence and/or LB cases for this study. Healtheconomic or cost studies, case studies, animal studies, as well as studies of biomedical mechanisms, modeling or simulations, or management or diagnostic guidelines, were excluded. Data only available in abstract form from conferences, letters, perspective or opinion papers, or commentaries were also excluded. Review articles were not included but were scanned for references. Articles reporting the results of national surveillance were excluded if the data were duplicated from the available public health surveillance reports, which have been analyzed and published separately (Burn et al., 2023, in this issue).

Analysis
We synthesized data, with relevant descriptive and key outcome variables, into tables (Campbell et al., 2020). For the purposes of data presentation, we considered four European regions per the WHO Regional Classification scheme and organized national data accordingly (Table 1) (World Health Organization, 2022).
Forest plots were produced for national-level incidence estimates, organized by European region, by country, and study characteristics (study period, data source, and case definitions). The 95% confidence intervals (CIs) were displayed as applicable. When a study reported the cumulative number of cases over a study period without reporting the corresponding average number of cases, we calculated the mean number of cases by dividing the cumulative number of cases over the year period. In the case that the 95% CI of the LB incidence was not reported, we calculated it using the exact binomial method (Wilson, 1927), given the availability of the number of cases (and/or the corresponding sample sizes) and the LB incidence.
Plots were further stratified by clinical manifestations when reported in the study. Point estimates for each study are represented by a black box, and the magnitude of the black box represents the size of the study. The 95% CIs are represented by the horizontal lines for each plot. We measured the heterogeneity of the data using the I 2 statistic (Campbell et al., 2020, Deeks et al., 2022, Higgins et al., 2003. Metaanalyses were performed, but the results are not reported due to the considerable heterogeneity of the data. All analyses were performed using the statistical software R (RStudio, version 1.4.1103) (R Core Team, 2021;RStudio Team, 2020).

Search results
The systematic review captured 72 articles that reported LB incidence estimates, of which 11 duplicated data already provided by the national surveillance reports (provided in our companion incidence surveillance article for Europe, Burn et al., 2023, in this issue). This yielded 61 unique articles describing LB incidence (national or subnational) in 25 countries ( Fig. 1). No published data were obtained for Albania, Andorra, Austria, Bosnia and Herzegovina, Greece, Iceland, Liechtenstein, Luxembourg, Malta, Moldova, Monaco, Montenegro, North Macedonia, San Marino, or Ukraine.
Among the studies, there were countries with national LB incidence (n = 28); countries reported subnational LB incidence estimates (n = 11; with overlap of both national and subnational estimates from some studies) and subnational (n = 33) (Fig. 1). Estimates of national LB incidence are presented in Tables 2-6; subnational data are provided in Supplementary Tables S1-S6. Subnational variations in incidence values were substantial in many countries. We illustrate examples of this in the text below, and the reader is referred to Supplementary Tables S7-S11 for a more in-depth review.
The review yielded 131 articles with either LB incidence estimates (n = 61) and/or reported number of LB cases (n = 70). Two countries (Bulgaria and Hungary) reported the number of LB cases but not the incidence. The articles that only reported numbers of LB cases are summarized in Supplementary Tables S7-S11 and are intended to provide additional insights.

Study design
Study designs applied, case definitions utilized, and populations sampled varied markedly among the articles reviewed. Of the 61 articles with LB incidence estimates, 47 studies were retrospective, whereas 13 were prospective (Tables 2 -6 and Supplementary Tables S1-S6). Among the retrospective studies, most (43 of the 47 studies) used a cohort study design, 3 were observational, and 1 was an ecological study. Most of the 61 studies used population-based data, but some studies enrolled defined subpopulations, such as inpatients and outpatients presenting with manifestations of LB, persons with facial palsy, or persons consulting general practitioners (GPs). The types of the studies published also varied, and thus, there may be risk of bias in the reported LB incidence estimates. In brief, the considerable heterogeneity makes data interpretation and comparability more complex.
All studies that met the inclusion criteria were retained and were not excluded based on study design, sample size, populations sampled, and so on. Studies are summarized by key variables to allow for interpretation of incidence estimates and potential biases or limitations.

Incidence of LB in Northern Europe region: Baltic States
(Lithuania). The only Baltic country with published LB incidence estimates was Lithuania (Supplementary Table S2), with data reported in only one article and at the subnational level. This study in the Vilnius district reported an incidence of LB of 85.4/100,000 PPY in outpatients attending an ambulatory unit in 2014-2016. LB cases were defined on the basis of documented clinical characteristics, laboratory results, electrocardiograms, and skin biopsy findings (Petrulioniene et al., 2021).
Incidence of LB in Northern Europe: Nordic region (Denmark, Finland, Norway, Sweden). Four articles reported national incidence estimates of LB in Denmark, Finland, and Sweden (Table 3). National incidence estimates for LB in countries in Nordic region, considering any case definition, ranged from 1.9 to 7.3/100,000 PPY in studies reporting cases of LNB to 118/100,000 PPY in a study of clinically and microbiologically confirmed cases of LB (Table 3). Two articles reported incidences of LNB in Denmark, which ranged from 1.9 to 3.3/100,000 PPY between 1996 and 2015 , Tetens et al., 2020. The highest LB incidence in Nordic region (118/ 100,000 PPY) was reported in Finland, which conducts national surveillance for clinically diagnosed LB cases and laboratory-confirmed LB cases (Sajanti et al., 2017). One study in Sweden evaluated different national data sources, and incidence of LNB was 6.2/100,000 population in 2014, with a higher incidence in the south (Dahl et al., 2019).
Subnational variation. An additional 13 articles estimated LB incidence at the subnational level in Denmark, Finland, Norway, and Sweden (Supplementary Table S3). For example, at the subnational level, the incidence of LNB in Denmark was higher in outlying islands (>10/100,000 PPY) and in southern Denmark (5.1/100,000 PPY) than in other regions . In Finland, the highest LB incidence at the subnational level was observed in the Å land Islands, where incidence of clinically diagnosed EM was 884.6/100,000 PPY and laboratory-confirmed LB was 1597/100,000 PPY (Sajanti et al., 2017). The incidence of clinically diagnosed EM in other regions in Finland ranged from <5/100,000 PPY (Northern Ostrobothniam, Kainuu, and Lapland) to 161.8/100,000 PPY in South Karelia. The incidence of laboratory-confirmed LB followed a similar trend (Sajanti et al., 2017).

Incidence of LB in Northern Europe region: United Kingdom (England, Northern Ireland, Scotland, Wales) and
Ireland. Six articles reported national incidence estimates of LB in countries in the United Kingdom and Ireland (Table 4), and five articles reported subnational LB incidences (Supplementary Table S4). National incidence estimates of LB in the United Kingdom considering any definition were <10/100,000 PPY, except for one study in Scotland, where reported LB incidence was 37.3/100,000 PPY (Table 4) (Cairns et al., 2019).
Published LB incidence estimates for the United Kingdom have used a variety of data sources and include insights into incidence at the wider UK level, at individual country levels, and at subnational regional levels within countries. Within the published literature, there was evidence of an impact of case definition on reported LB incidence. For example, a study in the GP-based Clinical Practice Research Datalink (CPRD; 2001 -2012) found incidence of suspected and possible LB of 12.1/100,000 PPY for the United Kingdom and 37.3/100,000 PPY for Scotland (Cairns et al., 2019), whereas lower incidences were observed over a similar period (1998 -2016) in another study using a primary case database with diagnosis of LB, suspected LB, or related conditions based on Read codes (Tulloch et al., 2020). Both studies showed an increase in LB incidence over time.
Subnational variation. There were an additional five articles that reported subnational LB incidence estimates in specific localities in the United Kingdom or Ireland (Supplementary Table S4). CPRD data showed a higher incidence of (suspected and possible) LB in southwest England (23.4/ 100,000 PPY) and lower incidence in northern England and West Midlands (6.3/100,000 PPY) (Cairns et al., 2019). National LB testing laboratory in Scotland reported an incidence of laboratory-confirmed LB of 6.8/100,000 PPY between 2008 and 2013; highest incidence was observed in the Highlands (44.1/100,000 PPY) (Mavin et al., 2015).

184
BURN ET AL.
Incidence of LB in Southern Europe region (Croatia, Italy, Portugal, Spain). Two articles reported national incidence estimates of LB for Portugal and Croatia (Table 5). The estimated national incidence of laboratory-confirmed LB (excluding EM) in Portugal was 0.4/100,000 PPY (de Carvalho and Núncio, 2006). The estimated national incidence of clinically diagnosed LB cases and laboratory-confirmed LB cases in Croatia was 6.6/100,000 PPY (Mulić et al., 2011).
Subnational variation. Five articles estimated LB incidence at the subnational level in specific localities in Croatia and Spain (Supplementary Table S5). In Croatia, for example, the highest LB incidence was reported in the northernmost provinces, ranging from 18.8 to 25.4/ 100,000 PPY ( Mulić et al., 2011). In a study among confirmed positive Western plots and patients diagnosed with possible LB in Spain, the annual incidence of LB increased over the study period from 2.6 to 11.6/100,000 PPY (Vazquez-Lopez et al., 2015). In Lombardy (northern Italy), where there is mandatory reporting of LB to the Rare Disease Registry using diagnostic codes, incidence was 0.1/100,000 PPY between 2000 and 2015 (Zanzani et al., 2019).

Incidence of LB in Western Europe region (Belgium,
France, Germany, Netherlands, Switzerland). Eleven articles reported national incidence estimates of LB in Western European countries ( Table 6). The national incidence of LB in countries in Western Europe ranged from 0.06 to 156/100,000 PPY. The highest LB incidence was observed in Switzerland using EUCALB definitions (156/ 100,000 PPY) (Altpeter et al., 2013). High LB incidences were also observed in Belgium and the Netherlands in studies that reported cases of EM or cases of EM and/or consultations for tick bites (106.1/100,000 PPY and 139.6/ 100,000 PPY, respectively) (Geebelen et al., 2019, Hofhuis et al., 2016. The LB incidences were <10/100,000 PPY in France and 17.8-41/100,000 PPY in Germany (Table 6). In the Netherlands, periodic national postal surveys were sent to GPs to ascertain the incidence of consultations for tick bites and EM between 1994 and 2014 (Hofhuis et al., 2016), which found an increase in LB incidence from 38.6/100,000 PPY in 1994 to 139.6/100,000 PPY in 2014.
Subnational variation. There were 15 articles that reported LB incidence at the subnational level in specific localities in Western European countries (Supplementary Table S6), including France and Germany, which had the lowest national LB incidences. Incidence of EM in Belgium ranged from 30.9/100,000 PPY in East Flanders to 390.9/100,000 PPY in Limburg (Geebelen et al., 2019). Incidence of LB cases diagnosed according to the EUCALB definition as reported from the French Sentinel GP network was 180/ 100,000 PPY in Alsace and ranged from 30 to 511/100,000 PPY (Schmitt et al., 2006). LB incidence in Germany also showed substantial subnational variation: 0.3/100,000 PPY (Thuringa) to 90.1/100,000 PPY (Brandenburg) (Adlhoch and Poggensee, 2010).
Incidence of LB across all four regions of Europe. The highest LB incidences (>100 cases per 100,000 PPY) were reported in Belgium, Finland, the Netherlands, and Switzerland. Incidences were 20 to 40/100,000 PPY in the Czech Republic,Germany,Poland,and Scotland and <20/100,000 PPY in Belarus,Croatia,Denmark,France,Ireland,Portugal,Russia,Slovakia,Sweden,and the United Kingdom (England,Northern Ireland,and Wales). Higher incidence was observed at the subnational level than at the national level in eight countries, including the Republic of Ireland (up to 43 per 100,000 PPY), Scotland (up to 56.4 per 100,000 PPY), England (up to 23.4 per 100,000 PPY), Bulgaria (up to 30.9 per 100,000), Poland (up to 200.9 per 100,000 PPY), Russia (up to 40.5 per 100,000 PPY), Slovak Republic (up to 52.1 per 100,000 PPY), and Sweden (up to 464 per 100,000 PPY). Local studies conducted in Lithuania and Norway reported LB incidence of 85.4 and 552 per 100,000 PPY, respectively.
Forest plots of incidence estimates, with corresponding 95% CIs (when available from studies) are displayed in Fig. 2 to better visualize the data and compare across countries and European regions, and to view the varying and wide range of incidences and identify outliers.
Due to the considerable heterogeneity of the data (I 2 > 80%) even after subanalyses by clinical manifestations and countrycase definition interaction, we did not pool data and a metaanalysis was not reported (Campbell et al., 2020). Maps showing variation of LB incidence estimates across countries were also not developed due to the heterogeneity of the data. The high degree of heterogeneity indicates that incidence results should be interpreted in light of the complexity of data due to variability in study design, study setting, populations sampled, study periods, sample size, and case definitions utilized.

Discussion
This review provides a unique and granular view of published estimates of LB incidence from epidemiological studies conducted in 25 countries in Europe from 2005 to 2020. Comparison of incidence across countries was limited by heterogeneity in study design and methods, including variations in the case definitions of LB. Four countries (Belgium, Finland, the Netherlands, and Switzerland) had national LB incidences >100/100,000 PPY, and three others (Czech Republic, Germany, and Poland) had national LB incidences of 20 to 40/100,000 PPY. National LB incidences were lower (<20/100,000 PPY) in Belarus, Croatia, Denmark, France, Ireland, Portugal, Russia, Slovakia, Sweden, and the United Kingdom (England, Northern Ireland, and Wales). Furthermore, several countries had high LB incidences at the subnational level in some locales, indicating that national estimates of LB incidence may not accurately reflect the incidence at local levels. Heterogeneity in disease incidence between subnational localities could reflect true differences in disease risk. Alternatively, the differences could reflect the use of inconsistent epidemiological methods to assess LB incidence, highlighting the potential advantage of using standardized case definitions, such as those published by EUCALB.
Our systematic review complements a study that evaluated incidence as reported in national surveillance systems in European countries (Burn et al., 2023, in this issue). As noted in the surveillance study, among 25 countries where national surveillance incidence data were published, incidence data from the literature were also identified in 15 of these countries (Belgium, Croatia, Czech Republic, England, Finland, France, Germany, Ireland, Portugal, Russia, Poland, Scotland, Slovakia, Switzerland, and Wales). We found additional published estimates of LB incidence in the literature for six countries currently without national surveillance in place: Belarus, Denmark, Italy, the Netherlands, Spain, and Sweden. Studies conducted in these countries serve as important complementary data to inform the use of LB prevention strategies in the future.
Two systematic reviews have previously estimated LB incidence in Europe, both of which were limited to Western Europe. The systematic review by Sykes (2014)  reported study years (1988 through 2011). The authors concluded that the LB incidence in Western Europe was 56.3/ 100,000 PPY, which equates to >200,000 cases per year (Sykes, 2014). A second systematic review across Western Europe by Vandekerckhove et al. (2021) evaluated 25 articles from 18 countries from database inception through 2018.
The authors found a similarly wide range of national LB incidence (0.001 to 632/100,000 PPY) across reported study years (1991 -2017). Limited LB incidence data available from some countries in Western Europe were noted, and it was concluded that the incidence of LB was increasing in some countries, mainly in the northern and central regions (Vandekerckhove et al., 2021). In contrast to these previous articles, our review included 61 articles from 25 countries (including European countries outside of Western Europe) and provides a more comprehensive overview of the contemporary incidence of LB in Europe.
There are limitations in interpreting the epidemiological studies included in our review. Studies used different case definitions: clinical, clinical and/or laboratory confirmed, or laboratory confirmed. Sometimes different sources of data with different case definitions were even used within articles, such as Sajanti et al. (2017). Other times, different case definitions were used for studies from the same countries. This can be illustrated by the studies included for the United Kingdom. The study by Cairns et al. (2019) included all patients tested and treated for LB regardless of test results, which may have resulted in an overestimation of incidence. However, the study by Tulloch et al. (2019b) used only laboratory-confirmed cases and thus may have missed cases, leading to underestimation.
While the use of different case definitions limits directly comparing incidence estimates of LB, it provides important insights to the countries' LB disease burden. Many of the studies included in our review also captured various clinical manifestations of LB, further complicating matters. While the majority of cases reported were clinical cases of EM, there were also cases of clinical EM and/or laboratory

REVIEW OF LYME BORRELIOSIS INCIDENCE IN EUROPE
confirmed, and less frequently, specific clinical manifestations, such as LNB, LA, or LC. Serology is the most frequently used method of laboratory confirmation, which is supported by the results of our review. Nonetheless, many patients who present with EM, which is the feature most commonly used to diagnose Lyme disease, will have negative antibody test results. Compounding potential confusion is the fact that in patients with no clinical evidence of the disease, who have a low probability of infection, antibody assays for LB are likely to yield false-positive results (Shapiro, 2014).
The differences in incidence estimates and in clinical manifestations across countries in Europe were expected due to varying distributions of Bbsl across Europe, influenced by a range of factors-including geographical, environmental, and climate factors, compounded by human recreational and occupational risk factors, and so on (Van den Wijngaard et al., 2017). We did not exclude nor restrict our review to specific case definitions; rather, the variability in case reporting across countries complements incidence estimates from national surveillance systems in our companion article (Burn et al., 2023, in this issue). Given the wide range of incidence estimates of LB, heterogeneity in how varying case definitions were used in these measures sheds valuable insights into the disease across the region. Nonetheless, factors contributing to heterogeneity in varying study designs, case definitions utilized, data sources, and diagnostic methods should be considered when interpreting our results.
Most studies in the review are epidemiological studies. We did not perform a quality assessment, formally rank them on basis of quality of evidence, and consequently exclude certain articles and other data sources. While we realize that systematic reviews are susceptible to varying quality of studies that arise in any of the included primary studies, we sought to obtain estimates of incidence across Europe from published epidemiological studies. In an effort to be comprehensive, we did not want to exclude data that could give important insight into regions that may have LB burden that is not captured routinely from surveillance systems. Nevertheless, differences in quality of design and implementation of these epidemiological studies do exist. For example, in a study in our review conducted in Russia, the authors sought to evaluate effectiveness of areas treated with Baytex to eliminate ticks but not to specifically measure incidence (Bogachkina et al., 2011). The methodology is not clear, and the authors did not report case definition utilized, which are limitations in the interpretation of this incidence estimate for Russia; however, given this study met our inclusion criteria and reported incidence in their findings, we still report results here as part of our review.
Alternatively, a study conducted in Belgium among a representative sample of the national population clearly described their methodology, utilized EUCALB case definitions, and addressed possible information bias for EM (Vanthomme et al., 2012). Appropriate statistical analyses conducted to compare incidence rates using the Belgian population were clearly described, and the authors clearly stated the strengths and limitations of their study for full interpretation.
Our data may underestimate the future LB burden. The human health burden of LB is expected to increase in Europe as the range of tick populations expands (altitude and latitude), potentially as a result of the impact of climate change on tick life cycles, migratory animals, and human activities (Hussain et al., 2021). Furthermore, the LB incidence in countries in Europe is likely underestimated due to limitations of public health surveillance and epidemiological studies. For example, LB may be present in countries where data are currently absent, including at subnational levels.
New prevention methods for LB are on the horizon. A prophylactic monoclonal antibody for LB pre-exposure prophylaxis under development has demonstrated potential to offer protection (Schiller et al., 2021). Furthermore, a vaccine is currently in clinical development (Clinical-Trials.gov Identifier: NCT04801420). Recent studies have highlighted the value of a vaccine for the prevention of LB based on high acceptability and the limited capability of existing measures to prevent tick-borne diseases (Hook et al., 2021, Schwartz et al., 2022. Data presented in this review, and in the LB incidence in Europe from the National Public Health Surveillance Systems (Burn et al., 2023, in this issue) companion article, indicate that the incidence of LB disease in Europe is substantial but geographically heterogeneous, both among and within countries. Data reported at the national level can often mask subnational differences, particularly in areas with substantially higher incidence. Incidence data from the epidemiological studies included here can help identify subnational regions of high incidence. Therefore, data from this review serve as an important complement to incidence data from national surveillance systems, where subnational data may not always be available or reported in certain regional areas. Collectively, these data can be used to identify countries and localities with a high LB disease burden that may benefit from future preventive and therapeutic strategies, including a vaccine, to optimize reduction in LB disease burden.