Secondary Logo

Lyme Borreliosis in Children: A Tertiary Referral Hospital–Based Retrospective Analysis

Nassar-Sheikh Rashid, Amara, MD*; Boele van Hensbroek, Michael, MD, PhD; Kolader, Marion, MD; Hovius, Joppe, W., MD, PhD§; Pajkrt, Dasja, MD, PhD, MBA

The Pediatric Infectious Disease Journal: February 2018 - Volume 37 - Issue 2 - p e45–e47
doi: 10.1097/INF.0000000000001735
Brief Reports

Lyme borreliosis (LB) is an endemic disease in adults in Western countries. Although children may also be infected, pediatric studies on LB are scarce. This study aims to estimate the incidence of LB among children with a clinical suspicion for Lyme in a tertiary referral center in the Netherlands. Patient chart data on medical history, clinical signs and symptoms, diagnostic test results and diagnoses were collected using standardized case record forms. Patients were categorized based on clinical and laboratory findings using a modified, previously published classification system. We included 325 children, with a median age of 11.9 years, of whom 61.8% were female. LB was diagnosed in 38 of the referred children (11.7%). However, of the 85 patients who were specifically referred to the Lyme clinic, 28 (32.9%) were diagnosed with LB. Of the specifically referred Lyme-positive patients, 11 (39.3%) had a definitive LB diagnosis. Twelve children had a posttreatment LB syndrome. In line with previous reports in adults, only a small proportion of children referred with a suspicion of LB were diagnosed with definite or probable LB, which illustrates the difficulty in diagnosing LB by the general practitioner or pediatrician in a district hospital.

From the *Department of Paediatric Immunology, Rheumatology and Infectious Diseases, Emma Children’s Hospital/Academic Medical Center, Amsterdam, The Netherlands; Department of Paediatric Infectious Diseases, Emma Children’s Hospital and Department of Global Health, AMC, Amsterdam, The Netherlands; Department of Medical Microbiology, Amsterdam Multidisciplinary Lyme Center, AMC, Amsterdam, The Netherlands; and §Center for Experimental and Molecular Medicine, Department of Internal Medicine, Division of Infectious Diseases, Amsterdam Multidisciplinary Lyme Center, AMC, Amsterdam, The Netherlands.

Accepted for publication July 29, 2017.

The authors have no funding or conflicts of interest to disclose.

Address for correspondence: Amara Nassar-Sheikh Rashid, MD, Department of Paediatric Immunology, Rheumatology and Infectious Diseases, Emma Children’s Hospital/Academic Medical Center, H7-217, PO Box 22700, 1100 DD Amsterdam, The Netherlands. E-mail:

Lyme borreliosis (LB) is an endemic tick-transmitted infection caused by the Borrelia burgdorferi sensu lato (sl) spirochete.1 The incidence of LB in Western Europe has increased over the last decades. In The Netherlands, the incidence has been estimated at 140 per 100000 people,2 equal to approximately 25000 new LB cases each year. Clinical manifestations can vary by B. burgdorferi sl subspecies, although overlap occurs. Clinical presentations are commonly divided into three stages: early localized, early disseminated, and late (disseminated) LB. Late LB is due to a chronic B. burgdorferi sl infection and can occur in un- or inadequately treated patients.3 “Chronic Lyme” is a controversial term describing long-lasting nonspecific complaints attributed to LB.4

Identification and classification of LB is challenging due to the often nonspecific symptomatology and limitations of the available diagnostic tests. The sensitivity of serologic testing for LB can be low in the early stage but increases to approximately 95% by 8 weeks after the start of the infection.5 Polymerase chain reaction (PCR) of B. burgdorferi sl on skin, synovial fluid or cerebrospinal fluid (CSF) can be performed to support the diagnosis of LB, but it’s sensitivity varies, 25–90%, 46–88% and < 10–50% for skin, synovium and CSF, respectively6.

A recent retrospective study among adults described the characteristics of adult patients referred to the Amsterdam Multidisciplinary Lyme Center (AMLC), which is located within the Academic Medical Center (AMC) in Amsterdam found that of the 200 patients, 120 (60.0%) did not meet the criteria of LB. Thirty-one patients (15.5%) had localized or disseminated LB, of which 18 (58.1%) were classified as definite or probable and 13 (41.9%) as questionable.7 In this current study, we aim to describe the characteristics of referred children with a clinical suspicion of LB and categorize these children into groups with high and low probabilities of having LB.

Back to Top | Article Outline


We included children (< 18 years) with a clinical suspicion of LB, referred to the AMC in The Netherlands between January 2011 and December 2015. Children were referred by general practitioners or (mostly general) pediatricians from all over The Netherlands. We included patients who were tested for LB as part of their patient care work-up. Tests were ordered by various medical specialists of the AMC, including pediatric rheumatologists, neurologists, dermatologists, and oncologists. Patients’ data were retrieved from electronic patient records and from the database from the Department of Medical Microbiology. The following groups of variables were collected in a standardized case record form: medical history, clinical symptoms, diagnostic test results and histopathologic results. Routine two-tier testing for LB consisted of B. burgdorferi sl C6-Lyme enzyme immunoassay (EIA) (immunoglobulin G (IgG)/immunoglobulin M (IgM), Immunetics®), and when indeterminate or positive a confirmation test with an immunoblot (IgG and upon indication IgM) (Mikrogen®) was performed. CSF analyses included cell counts, C6-Lyme EIA (IgG/IgM, Immunetics®) and/or B. burgdorferi sl qPCR. Skin biopsies were routinely performed in children with a suspicion of acrodermatitis chronica atrophicans (ACA). Patients were categorized based on clinical and laboratory findings by two independent doctors, using a modified classification system previously published (Table 1).7 Objective clinical findings compatible with LB were defined as “typical presentations of LB,” that is, erythema migrans (EM), cranial nerve palsy and mono or oligoarticular arthritis in larger joints. Patients with a “probable LB” were defined as having symptoms resembling LB, for instance oligoarthritis of smaller joints.



Back to Top | Article Outline


We included 325 children, 85 (26.2%) from the hospital patient databases and 240 (73.8%) more via the microbiology laboratory records. The median age was 11.9 years (interquartile age, 7.8–14.9), and 61.8% of the patients were female. The most reported complaints were fatigue/general malaise (18.2%) and arthralgia (17.2%). The duration of complaints varied widely. Only 8% of the included children had acute symptoms (with duration of less than 2 weeks), and more than 50% had symptoms for more than 6 months.

B. burgdorferi sl serologic tests were performed in 310 (95.4%) patients. Fifteen CSF analyses were performed in children suspected of a neuroborreliosis. Four patients had a pleocytosis, of which two tested positive for Borrelia antibody tests in CSF. One of the CSF analyses included B.burgdorferi sl PCR, which was negative. PCR on synovial fluid was not performed. In the one skin biopsy performed histopathologic changes suggestive of ACA were visible.

LB criteria were not met in 88% of the patients (Table 1). Fourteen children (4.3%) had a definite LB diagnosis. Four patients presented with EM and 10 with definite disseminated LB. Three patients were classified as having probable LB. One had a suspicion of neuroborreliosis with positive serologic tests, but negative CSF tests. Two patients had a peripheral facial palsy and an indeterminate B. burgdorferi sl IgG immunoblot result. One patient had persistent LB: Nine patients (2.8%) had a questionable LB. These patients presented with various complaints including polyarthritis, arthralgia, fatigue, chest pain and malaise with a rash. In two of them, a definitive alternative diagnosis, that is, Guillain Barre Syndrome and poststreptococcal reactive arthritis was established. Twelve patients (3.7%) met the definition of posttreatment LB syndrome (PTLBS) and presented with various complaints, including fatigue, general malaise and pain.

A total of 85 children were specifically referred to the AMLC. Of these, only 28 children (32.9%) had a diagnosis of LB. Eleven patients (12.9%) had a definite early localized or disseminated LB: EM (4), arthritis (4), neuroborreliosis (2) and ACA (1). Twelve patients presented with a PTLBS and four with a questionable LB.

One patient presented with a persistent LB, based on a persisting monoarthritis of the knee with positive B. burgdorferi sl serology in synovial fluid after being adequately treated with antibiotics. However, a PCR on synovial fluid was performed after we completed our data analysis, which was negative.

Back to Top | Article Outline


We conclude that in this study the large majority of pediatric patients who were referred to either a tertiary referral hospital or a specialized Lyme center did not meet the criteria of LB. This finding is new and illustrates the difficulty in diagnosing LB by the general practitioner or pediatrician in a district hospital.

Defining LB can be controversial. Our definitions were based on published LB guidelines6 and comparable with previously published criteria for adults.7 The small number of patients with definite LB might have reflected increasing societal concerns about LB diagnostics, treatment and (long-term) clinical outcomes. The diagnosis LB is further complicated by the incapacity of current serologic tests to discriminate between an active and past B. burgdorferi sl infection.

The low incidence in definite LB may further be explained by the fact that children with a clinical suspicion of LB were primarily diagnosed and treated in local hospitals, without prompt referral to the tertiary referral center or AMLC. As more than 50% of the children had a duration of symptoms of more than 6 months, this further points toward other disorders than LB eliciting the complaints of the studied patients. The incidence rate of LB cannot be translated to the total Dutch pediatric population, due to a potential referral bias. Prior antibiotic treatments (for other reasons than the treatment of LB) may further contribute to the low incidence rate of LB in this study as this may affect Borrelia antibody levels in both serum and CSF.

Despite being the largest study describing pediatric LB in a Dutch cohort presented at a tertiary referral hospital, our study has limitations. We included children who were evaluated by various medical specialists and routinely tested for LB. This might have influenced the number of negative serologic results in this group as they might have a low a-priori chance of having LB. We did not include children with symptoms resembling LB but who were not tested, and those who were not referred specifically to the AMC or AMLC. However, we believe that this number of patients is very low, due to the high awareness level in both the tertiary referral center and the specialized Lyme center.

There is increasing evidence that viable B. burgdorferi sl do not persist after conventional treatment with antimicrobials, which implies that persistent symptoms, after receiving conventional treatment for LB, should not be attributed to persistent active infection.8 Indeed, in our study, only one patient was diagnosed with antibiotic refractory Lyme arthritis.9

In summary, we report a low incidence of LB in referred children in The Netherlands. It is important that children with a suspicion of LB are evaluated by a pediatrician with affinity for LB. This helps to adequately establish or rule out the diagnosis of LB, potentially prevents antibiotic under- or overtreatment, and could help to downscale the societal concerns on LB.

Back to Top | Article Outline


1. Jahfari S, Krawczyk A, Coipan EC, et alEnzootic origins for clinical manifestations of Lyme borreliosis. Infect Genet Evol. 2017;49:48–54.
2. Hofhuis A, Bennema S, Harms M, et alDecrease in tick bite consultations and stabilization of early Lyme borreliosis in the Netherlands in 2014 after 15 years of continuous increase. BMC Public Health. 2016;16:425.
3. Hofhuis A, Harms M, Bennema S, et alPhysician reported incidence of early and late Lyme borreliosis. Parasit Vectors. 2015;8:161.
4. Ljøstad U, Mygland ÅThe phenomenon of ‘chronic Lyme’; an observational study. Eur J Neurol. 2012;19:1128–1135.
5. Leeflang MM, Ang CW, Berkhout J, et alThe diagnostic accuracy of serological tests for Lyme borreliosis in Europe: a systematic review and meta-analysis. BMC Infect Dis. 2016;16:140.
6. CBO Guideline Lymeziekte 2013. Available at Accessed October 2016.
7. Coumou J, Herkes EA, Brouwer MC, et alTicking the right boxes: classification of patients suspected of Lyme borreliosis at an academic referral center in the Netherlands. Clin Microbiol Infect. 2015;21:368.e11–368.e20.
8. Oliveira CR, Shapiro EDUpdate on persistent symptoms associated with Lyme disease. Curr Opin Pediatr. 2015;27(1):100–104.
9. Steere AC, Angelis SMTherapy for Lyme arthritis: strategies for the treatment of antibiotic-refractory arthritis. Arthritis Rheum. 2006;54:3079–3086.

Lyme borreliosis; children; incidence; clinical algorithm; diagnostic tests

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.