The Incidence of Lyme Borreliosis Among Children

Abstract

Background:

First recognized in children in the United States, Lyme borreliosis (LB) is the most prevalent tickborne illness in North America and Europe. However, the incidence of LB in children, including geographic variation and difference from that in adults, is incompletely described.

Methods:

We compiled surveillance data from public health agency websites reporting age-stratified LB case data, which was combined with census data to derive incidence estimates. Additional incidence estimates were obtained through a systematic literature review.

Results:

We identified 18 surveillance systems and 15 published studies for derivation of LB incidence in children. National incidence of >10 cases in children per 100,000 per year were estimated for the United States and parts of Eastern, Western and Northern Europe. However, there was substantial variation in incidence among countries in some European regions. National incidence estimates from the literature largely aligned with estimates from surveillance. Surveillance-reported pediatric incidence was lower than adult incidence in adults in 8 countries, similar to adult incidence in 3 countries, and higher than adult incidence in 1 country. Among all pediatric age strata, the 5–9 years of age stratum had the highest proportion of pediatric cases in most countries.

Conclusions:

As pediatric LB represents a substantial proportion of overall LB incidence across countries in Europe and North America, LB prevention and control efforts should target pediatric as well as adult populations. However, better data are needed to fully characterize the difference in incidence across geographic regions.

Lyme borreliosis (LB) was originally identified in the mid-1970s during the investigation of an outbreak of suspected juvenile arthritis reported among a cluster of children in Old Lyme, Connecticut.¹ Several subsequent studies in the Northeastern United States characterized the clinical presentation of LB,² discovered epidemiologic linkage to the ixodid tick vectors,³ and finally identified the etiologic bacterial agent Borrelia burgdorferi sensu lato.⁴ Today LB is recognized as the most prevalent tickborne illness in the United States and Europe.⁵

Clinical presentation of LB is similar in children and adults, typically occurring in 3 distinct stages with specific clinical manifestations: early localized disease manifests as erythema migrans (EM), early disseminated disease manifests as Lyme neuroborreliosis (LNB) and carditis and late disseminated disease primarily manifests as Lyme arthritis.⁶ In Europe, early localized disease in children may also present as borrelial lymphocytoma, while late disseminated disease may also include acrodermatitis chronica atrophicans; these manifestations are rare in North America due to differences in circulating Borrelia genospecies.⁷ LNB is of particular concern and most commonly manifests as facial nerve palsy in up to 56% of affected children,⁸ although meningitis may occur, potentially causing long-term neurologic deficits.^9–12 However, once diagnosed, all stages of LB are treatable with antibiotics with excellent long-term prognoses for children.^6,13

The incidence of LB can be monitored through public health surveillance systems, which report cases for both adult and pediatric populations. However, these systems vary widely in their methodologies including case definitions (all LB vs. specific manifestations), geographic scope (national vs. regional) and coverage (sentinel sites vs. comprehensive), reporting obligation (mandatory vs. voluntary) and reporting entity (general practitioners and/or hospital physicians and/or diagnostic laboratories and/or general public).¹⁴ Although these variations make direct comparison between surveillance systems challenging, data can still be used to identify areas of high pediatric LB incidence, defined by the US Centers for Disease Control and Prevention as ≥10 cases of 100,000 population,¹⁵ and analyze temporal trends. In addition, published epidemiologic studies can provide complementary information to that available through public health surveillance.

The global incidence of LB in children is incompletely described, with gaps in understanding how incidence in children compares among countries and regions, and how it compares to incidence in adults. This review seeks to provide a comprehensive summary of available surveillance data and published literature on LB in children across countries, to understand geographic variations and to characterize the incidence by age group.

MATERIALS AND METHODS

We reviewed government public health agency websites to identify surveillance systems with publicly available annual reports (International Prospective Register of Systematic Reviews CRD42021236906). Pediatric data were extracted from annual reports of age-stratified LB incidence from 2005 to 2020. Regional or state-level reports were pooled to provide country-level incidence.

A global systematic literature review on LB epidemiology was conducted for the period 2005–2020. The protocol was based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines,¹⁶ registered in International Prospective Register of Systematic Reviews (CRD42021236906) and has been described elsewhere.¹⁷ The literature was further surveyed from January 1, 2021, to December 31, 2021, to supplement the review. Studies identified in the literature utilizing data from public health surveillance systems were excluded from analysis to avoid duplication.

Non-English surveillance reports and journal articles were translated from their original published language into English using DeepL.¹⁸

Analysis

Surveillance data were extracted into Microsoft Excel (Microsoft Corporation, Redmond, WA). Annual incidence was the number of reported LB cases per 100,000 children per year (CPY). Where only incident LB case counts were reported, incidence proportions were calculated using available census data for the country population. For data presentation, Europe was divided into 4 regions per the World Health Organization Regional Classification scheme.¹⁹

Average LB incidence in children for each country was plotted over time for available years of surveillance data. For countries with annual LB incidence in children >5 of 100,000 CPY, the ratio of incidence for children versus adult cases was also calculated, with 95% confidence intervals calculated using a large sample normal approximation. In countries with >30 average annual surveillance-reported LB cases in children, we reported the proportion of cases in each age stratum. All plots were generated using R (The R Foundation, Indianapolis, IN).²⁰

RESULTS

We identified national LB public health surveillance systems in 28 countries, of which 19 (68%) reported age-stratified LB incidence from 2005 to 2020: 2 countries in Asia, 2 countries in North America, 4 countries in Eastern Europe, 6 countries in Northern Europe, 2 countries in Southern Europe and 3 countries in Western Europe. Additionally, our literature review identified 15 published studies from 2005 to 2021 that reported LB incidence in children. Among the 15 studies, 9 (60%) were conducted in countries that also had age-stratified LB public health surveillance; the remaining 6 (40%) were conducted in countries without age-stratified surveillance (Table 1).

TABLE 1.

Estimates of Incidence in Children of LB (Cases per 100,000 CPY) in From Surveillance Data (2005–2020) and Literature Published From 2005 to 2021

Region	Country	Surveillance Data		Literature Search Results
		Years Covered	Pediatric Age Strata	Author (Year)	Population	Manifestation	Age Strata	Incidence (Cases per 100,000 CPY)
Asia	Japan	2005–2020	0–4, 5–9, 10–14, 15–19
Asia	South Korea	2011–2020	0–9, 10–19
North America	Canada	2018–2019	0–4, 5–9, 10–14, 15–19
North America	United States	2008–2020	0–4, 5–9, 10–14, 15–19	Nelson et al21 (2015)	National, males	LB	0–4	30
							5–9	55
							10–14	45
							15–19	35
					National, females	LB	0–4	30
							5–9	45
							10–14	35
							15–19	30
				Schwartz et al22 (2021)	High-incidence states, males	LB	0–4	175
							5–9	300
							10–14	225
							15–19	200
					Neighboring states, males	LB	0–4	13
							5–9	23
							10–14	18
							15–19	14
					Low incidence states, males	LB	0–4	5
							5–9	10
							10–14	10
							15–19	8
					High-incidence states, females	LB	0–4	175
							5–9	225
							10–14	150
							15–19	175
					Neighboring states, females	LB	0–4	13
							5–9	20
							10–14	13
							15–19	19
					Low incidence states, females	LB	0–4	6
							5–9	9
							10–14	7
							15–19	14
Eastern Europe	Hungary	2005–2016	0, 1–2, 3–5, 6–9, 10–14, 15–19
Eastern Europe	Poland	2015–2019	0, 1, 2, 3, 4, 5–9, 10–14, 15–19	Toczylowski et al23 (2020)	Northeast Poland	LNB	<18	3
Eastern Europe	Romania	2011–2019	0–4, 5–9, 10–14, 15–19
Eastern Europe	Russia			Bogachkina et al24 (2011)	St. Petersburg	LB	<14	7
Eastern Europe	Slovakia			Svihrova et al25 (2011)	National	LB	0	1
							1–4	8
							5–9	14
							10–14	11
							15–19	9
Eastern Europe	Slovenia	2005–2012	0–1, 1–24
Northern Europe	Denmark	2015–2019	0–1, 1–4, 5–14	Tetens et al26 (2020)	National	LNB	0–15	4
				Dessau et al27 (2015)	National	LNB	0–15	4
Northern Europe	Estonia	2005–2018	0–1, 1–4, 5–9, 10–14, 15–19
Northern Europe	Finland	2015–2020	0–4, 5–9, 10–14, 15–19	Kanerva et al28 (2013)	Helsinki	Facial palsy	0–16	9
Northern Europe	Ireland	2016–2018	0–4, 5–9, 10–14, 15–19	Forde et al29 (2021)	National	LB	2–18	1
Northern Europe	Lithuania	2015–2019	0–6, 7–9, 10–14, 15–17
Northern Europe	Norway	2005–2020	0–9, 10–19	Oymar and Tveitnes11 (2009)	South-Rogaland District	LNB	<14	21
Northern Europe	Sweden			Bennet et al30 (2006)	Blekinge, male	EM	0	353
							1–4	460
							5–9	200
							10–14	133
					Blekinge, female	EM	0	339
							1–4	434
							5–9	232
							10–14	139
				Sodermark et al8 (2017)	Gothenburg Region	LNB	<15	28
Northern Europe	United Kingdom			Brellier et al31 (2022)	National	LB	<15	4
Southern Europe	Portugal	2013–2018	0–1, 1–4, 5–14
Southern Europe	Serbia	2013, 2016–2017	0, 1, 2, 3, 4, 5, 6, 0–6, 7–9, 10–14, 15–19
Southern Europe	Spain			Vázquez-López et al32 (2016)	Lugo	LNB	<15	5
Western Europe	Belgium	2005–2010	0–1, 1–4, 5–14
Western Europe	France	2010–2020	0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 0–4, 5–9, 10–14, 15–19
Western Europe	Germany	2016–2020	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 0–5, 5–10, 10–15, 15–20	Lohr et al33 (2015)	National	LB	6–9	29

Gray boxes indicate no data found. Incidence estimates rounded to nearest whole number to provide consistent precision between studies.

The pediatric age strata that were reported varied among LB surveillance systems and studies. Of the 19 surveillance systems reporting age-stratified LB incidence, 14 countries (74%) reported incidence in <20 years of age, 3 countries (16%) reported incidence in <15 years and 1 country each (5%) reported <17 and <25 years. Among the 15 studies, the most common age cutoff for children was <20 years of age (4 studies, 27%), followed by <15 years (3 studies, 20%), then <16 and <14 years (2 studies each, 13%); each of the remaining 4 studies use different age cutoffs (Table 1).

Incidence in Asia (Japan, South Korea)

Age-stratified LB surveillance data was available in Japan (2005–2020) and South Korea (2011–2020) (Table 1). Pediatric LB incidence was very low (<1/100,000 CPY) in both countries, across years of available data (Fig. 1). There were no pediatric data identified from elsewhere in Asia.

FIGURE 1.

Incidence of LB in children from surveillance systems (2005–2020). Age cutoff varies by country, based on available age-strata, indicated in parentheses. Denmark and Ireland only report LNB.

Incidence in North America (Canada, United States)

Age-stratified LB surveillance data was available in Canada (2018–2019) and the United States (2008–2020), and 2 studies reporting national estimates of LB incidence in children were available in the United States (Table 1). Surveillance data in Canada showed low national incidence in children <20 years over the 2 years of data collected (<10/100,000 CPY). National surveillance data from the United States in children <20 years showed higher incidence than in Canada, ranging from 5 to 12 of 100,000 CPY. LB incidence in children in the United States increased from 2012 to 2017, before decreasing from 2017 to 2020 to previous levels (Fig. 1). Nelson et al²¹ reported higher national US LB incidence in children in medical claims data than in surveillance data, ranging from 30 to 55 of 100,000 CPY. One study reported subnational estimates of pediatric LB incidence: these varied widely between high incidence states (range: 175–300/100,000 CPY) and low incidence states (5–17/100,000 CPY).²²

Incidence in Eastern Europe (Hungary, Poland, Romania, Russia, Slovakia, Slovenia)

Public health LB surveillance systems were identified in 8 countries in Eastern Europe; however, the Czech Republic, Russia and Slovakia did not report age-stratified estimates. Further, Bulgaria only reported age-stratified estimates for a subset of age groups, with no data available from ages 10–40. Thus, age-stratified surveillance data was available for analysis in Hungary (2005–2016), Poland (2015–2019), Romania (2011–2019) and Slovenia (2005–2012). In addition, 3 studies were published reporting pediatric LB incidence estimates in Eastern Europe: 1 reporting national estimates in Slovakia, 1 reporting estimates for St. Petersburg, Russia and 1 reporting LNB estimates for Northeast Poland (Table 1).

Surveillance data showed highest LB incidence in children <24 years in Slovenia, ranging from 150 to 235 of 100,000 CPY, while incidence in children <20 years in Hungary, Poland and Romania did not exceed 25 of 100,000 CPY (Fig. 1). Similarly, study estimates of LB incidence in Slovakia and St. Petersburg ranged from 5 to 15 of 100,000 CPY.^24,25 Incidence of LNB reported in children <18 years in Poland was 3 of 100,000 CPY,²³ which was 4- to 8-fold lower than surveillance-reported incidence of all LB in children <20 years (ranging from 12 to 25/100,000 CPY) (Fig. 1).

Incidence in Northern Europe (Denmark, Estonia, Finland, Ireland, Lithuania, Norway, Sweden, United Kingdom)

In Northern Europe, 9 countries were found to have public health LB surveillance systems. Age-stratified incidence data was available in Denmark (2015–2019), Estonia (2005–2018), Finland (2015–2020), Ireland (2016–2018), Lithuania (2015–2019) and Norway (2005–2020); Latvia, Scotland and the United Kingdom did not report age-stratified estimates. Eight studies reporting incidence in children in Northern Europe were identified: 1 each with national estimates of LB in Ireland and the United Kingdom; 2 with national estimates of LNB in Denmark; 1 with estimates of LB associated with facial palsy in Helsinki, Finland; 1 with estimates of EM in Blekinge, Sweden; 1 each with estimates of LNB in South-Rogaland, Norway and Gothenburg, Sweden (Table 1).

Surveillance-reported LB incidence in children <20 years in Estonia was the highest in Northern Europe and appeared to be increasing overall from 2005 to 2018, but also showed the greatest range, from 12 to 82 of 100,000 CPY. Surveillance-reported LB incidence in children <20 years in Finland was relatively stable from 2015 to 2019, ranging from 40 to 57 of 100,000 CPY. Incidence in children <18 years in Lithuania from 2015 to 2019 showed a similar range, from 36 to 58 of 100,000 CPY (Fig. 1). Study estimates of LB-associated facial palsy in children <16 years in Helsinki was 8.6 of 100,000 CPY.²⁸ Surveillance-reported incidence of LB in children <20 years in Norway was relatively stable over time, ranging from 6 to 15 of 100,000 CPY from 2005 to 2020 (Fig. 1). However, a study in the South-Rogaland District reported LNB incidence of 21 of 100,000 CPY in children <14 years,¹¹ which suggests even higher overall LB incidence in this area. Surveillance data in Denmark and Ireland was only available for LNB, ranging from 4 to 6 of 100,000 CPY in children <15 years in Denmark and very low (<1/100,000 CPY) in children <20 in Ireland (Fig. 1). Study estimates of national LNB incidence in children <15 years in Denmark were 4.2 of 100,000 CPY²⁶ and 3.7 of 100,000 CPY.²⁷ An study estimated the national LB incidence in children 2–18 years of age in Ireland as 1.15 of 100,000 CPY,²⁹ similar to the very low incidence reported in LNB surveillance. Brellier et al³¹ estimated UK national LB incidence in children <15 years to be 3.71 of 100,000 CPY. The 2 studies available in Sweden reported high incidence estimates relative to the rest of Northern Europe: 133–460 cases of EM/100,000 CPY in children <20 years in Blekinge³⁰ and 14–42 cases of LNB/100,000 CPY in children <15 in the Gothenburg Region.⁸

Incidence in Southern Europe (Portugal, Serbia, Spain)

Public health LB surveillance systems were identified in 3 countries in Southern Europe. Portugal (2013–2018) and Serbia (2013, 2016–2017) reported age-stratified incidence, while Croatia did not. One study reporting incidence of LNB in Lugo, Spain was also identified (Table 1).

Incidence of LB in children <15 years in Portugal was very low from 2013 to 2018 (<1/100,000 CPY). In Serbia, surveillance-reported incidence of LB in children <20 years was 11 of 100,000 CPY in 2013 and decreased to 6 of 100,000 CPY in 2016–2017 (Fig. 1). LNB incidence in children <15 years in Lugo, Spain was 5 of 100,000 CPY.³²

Incidence in Western Europe (Belgium, France, Germany)

Age-stratified LB surveillance data was available in Belgium (2005–2010), France (2010–2020) and Germany (2016–2020). Switzerland did not report age-stratified incidence in a format that could be used for analysis. One study reporting national LB incidence in Germany was also identified (Table 1).

The highest surveillance-reported pediatric incidence in Western Europe was in children <20 years in France, ranging from 9 to 52 of 100,000 CPY and increasing over time from 2012 to 2020. Incidence in children <20 years in Germany from 2016 to 2020 was lower than incidence in the same period in France, ranging from 21 to 31 of 100,000 CPY but increasing from 2016 to 2020 (Fig. 1). Lohr et al³³ also reported national incidence of 29 of 100,000 CPY in children 6–9 years old in Germany. Surveillance data showed incidence in children <15 years in Belgium to be much lower than in France and Germany, remaining below 12 of 100,000 CPY over the data collection period (Fig. 1).

Comparison of Incidence Across Age Groups

Surveillance-reported incidence of LB in children <20 years was not statistically significantly different from incidence in adults in the United States, Hungary and Estonia. In children <24 years in Slovenia, children <20 in Serbia and Germany, and children <15 in Belgium, incidence was moderately lower than in adults (incidence ratio between 0.6 and 0.8). In children <20 years in Poland, Finland and France and children <18 years in Lithuania, incidence was substantially lower than in adults (incidence ratio between 0.3 and 0.5). The only country in which incidence was higher in children than in adults was Norway (incidence ratio: 1.55 [95% confidence interval: 1.25–1.93] for children <20 vs. adults) (Fig. 2).

FIGURE 2.

Ratio of surveillance-reported incidence in children versus adults (95% CI). Incidence is averaged across years of available surveillance data. Countries with incidence in children <5 of 100,000 CPY not included (Canada, Denmark, Japan, Ireland, Portugal, Romania, South Korea). Age cutoff varies by country, based on available age strata, indicated in parentheses. Colors indicate WHO region. CI indicates confidence interval; WHO, World Health Organization.

The 5–9-year age stratum had the highest proportion of LB cases in children in surveillance reports from Canada, the United States, Poland, Romania, Estonia, Finland and France. The highest proportion of cases were reported in the 0–6-year stratum in Lithuania and Serbia, and in the 5–14-year stratum in Denmark and Belgium (Fig. 3).

FIGURE 3.

Age distribution of surveillane-reported pediatric LB cases. A: Proportion of surveillance-reported pediatric LB cases. B: Age stratum with highest proportion of pediatric LB cases, by country. Cases include all years of available data. Countries with <30 pediatric cases excluded (Japan, South Korea, Ireland, Portugal). Age strata vary by country.

DISCUSSION

We provide a comprehensive review of the global incidence of LB in children from 2005 to 2020, as identified through LB public health surveillance systems and published epidemiologic research. High incidence (>10/100,000 CPY) was reported in the United States and several countries in Eastern, Western and Northern Europe. The highest incidence was identified in Slovenia, followed by Estonia. There was substantial variability in LB incidence within European regions and countries, reinforcing that LB risk is highly localized, although these comparisons were complicated by differences in surveillance methodology. State-level variation in the United States was also reported and was consistent with categorization as “high-” and “low-incidence states” based on LB incidence in all ages. Gaps remain in fully characterizing the extent of regional variation, given the lack of age-stratified surveillance data for many countries.

Among cases in children, the 5–9 age stratum (or other strata spanning that age range) was consistently found to contribute the highest proportion of cases (Fig. 3). Although overall surveillance-reported incidence in children was only higher than adults in Norway (Fig. 2), it is possible that incidence in narrower pediatric age groups (eg, 5–9-year-olds) is higher than in adults; we were unable to perform this analysis due to inconsistencies in the available data.

Although most of the analyses relied on surveillance data, our literature search results suggest that these may underestimate incidence³⁴ and that alternative data sources could aid in estimation of true incidence. A recent study used seroprevalence data in children in Germany, along with assumptions on antibody duration and asymptomatic infection, to estimate incidence. They then compared this to surveillance data and estimated that, in 2005, there were 9 symptomatic LB cases for every surveillance-reported LB case.³⁴ Another study in Finland used a similar approach, although only in adults ≥30 years old, and estimated 2.7 symptomatic LB cases for every surveillance-reported case.³⁵ Further, 3 studies included in our results leveraged medical claims databases to estimate incidence: 1 in Germany and 2 in the United States.^21,22,33 Although the German study found comparable incidence in claims versus surveillance data, both studies in the United States estimated higher pediatric incidence than that reported through surveillance. However, these studies noted that there also limitations in solely relying on claims data to understand LB epidemiology. Thus, although surveillance data may underestimate incidence, the extent of this underestimation is unclear and likely varies between countries and surveillance systems.

On the other hand, there is also evidence of over-reporting and overdiagnosis of LB. Studies in endemic regions in the United States and some European countries have shown that a minority of individuals referred for presumptive LB are found to be cases.³⁶ This suggests that estimates of incidence using claims data may be inflated, although it is unclear how many of those misdiagnoses would be reflected in surveillance due to case definition requirements. In addition, a study conducted in a large health system covering both endemic and nonendemic regions of the United States found that the number of under-reported true positive cases to a surveillance system was exactly offset by the number of false positive cases.³⁷ Nevertheless, judicious testing of individuals with a low pretest probability is essential, especially given the potential harms of inappropriate antibiotic use.^38–40

A few studies reported incidence of specific manifestations of LB rather than LB overall. The majority of these focused on LNB, likely because of the potential for serious long-term complications.^9–12 In addition, the European Commission added LNB to the list of communicable diseases to be covered by epidemiologic surveillance in 2018, resulting in a common case definition for LNB in the European Union.⁴¹ As such, we can expect estimates of LNB incidence to be more robust and comparable across countries than LB incidence overall, for which case definitions vary widely.

The heterogeneity of LB incidence data may limit the validity of comparisons between countries and studies. Differences in methodology, especially case definitions, make it difficult to compare surveillance data between countries. For countries that only include certain LB clinical manifestations (eg, LNB or EM), there is currently no agreed upon methodology to extrapolate from syndrome-specific incidence to overall incidence. Further, surveillance systems may have differing levels of coverage, while epidemiologic studies often only focus on a small catchment area that may not be comparable to the national population. Nevertheless, the overall comparison of incidence among countries is still informative in identifying areas of higher or lower risk. Age stratification also varies widely between surveillance systems and between studies, requiring assumptions about the homogeneity of incidence between age strata in different countries that may not be valid. Although the data available is useful for the high-level comparisons done here, standardization of reported age strata as well as surveillance methodology and case definitions would allow more robust analyses.

There is a paucity of data on LB incidence in children in Asia. Although the 2 surveillance systems we identified suggest that there is very low incidence in Eastern Asia, studies in adult populations in China estimate seroprevalence ranging from 0% to 37%.⁴² This is consistent with the regional variability observed in Europe and North America and suggests additional incidence studies across Eastern Asia are needed to further characterize regional LB incidence. Further, it may motivate investigation of incidence in other Asian regions, although these should be limited to areas were Ixodes ticks are present.

Overall, this article provides a comprehensive summary of available evidence on the global incidence of LB in children. Our results show that pediatric LB is an important contributor to overall LB disease incidence and affirms the need for intervention strategies that target pediatric populations as well as adults. We also identify key data gaps in terms of geographic coverage and further characterization of manifestation-specific incidence in children.