Two-Tier Lyme Disease Serology in Children with Previous Lyme Disease

Introduction

Lyme disease serologic tests may remain positive indefinitely after successfully treated Lyme disease (Kalish et al. 2001, Glatz et al. 2006, Kannian et al. 2007). Protective immunity, however, is neither complete nor lifelong, and reinfection with Borrelia species after early Lyme disease is well described (Nadelman and Wormser 2007, Branda et al. 2018). When evaluating a patient for a suspected reinfection, the challenge for clinicians can be to determine whether a patient's antibody reactivity represents active versus previous infection.

We assembled an eight-center cohort of children undergoing emergency department evaluation for acute manifestations of Lyme disease. Our goals were to examine the proportion of children with a history of Lyme disease who had positive Lyme disease serology and to determine the relationship between seropositivity and the time from previous infection. We hypothesized that most children with a history of Lyme disease would have a positive two-tier Lyme disease serology and the proportion would decline with time from previous infection.

Materials and Methods

Results

Of the 6757 eligible children over the study period, study staff approached 3862 children of which 2745 (71.1% of approached) consented to study participation (Fig. 1). Of those enrolled, 2501 (91.1%) had both a research Lyme disease test and history recorded. Of these, 126 children (5.0%) had previously been diagnosed with Lyme disease (92 definite and 34 probable previous Lyme disease). The previous Lyme disease diagnosis had been made 1–3 months ago (n = 11), 4–12 months (n = 6), 13–24 months (n = 22), or >24 months (n = 67). Children with a history of Lyme disease had symptoms similar to those without Lyme disease history (early disease 1.7% previous vs. 2.3% no history, early disseminated 56.2% vs. 47.0%, late 42.1% vs. 50.7%, p = 0.14).

OPEN IN VIEWER

FIG. 1.

Patient enrollment.

Of the 126 children with a history of Lyme disease, the median patient age was 11 years (interquartile range 7–14 years), and 74 (58.7%) were male. Of these, 47 children (37.3%; 95% CI 29.1–45.7%) met the diagnostic criteria for Lyme disease at the time of enrollment. Of these, 2 (4.3%) had a single EM lesion and 45 had positive 2-tier Lyme disease serology (20 positive IgG alone, 8 positive IgM alone, and 17 both positive immunoblots), of whom 18 had early disseminated and 27 Lyme arthritis. A higher proportion of children with a Lyme disease history had Lyme disease than those without [47/126 (37.3%) vs. 476/2375 (20.0%); p < 0.001)]. The longer the time interval from the previous Lyme disease diagnosis, the less likely the patient had met diagnostic criteria for current Lyme disease (adjusted odds ratio 0.62 per time period; 95% CI 0.46–0.84).

Of the 79 children with a history of Lyme disease who did not meet the diagnostic criteria for acute Lyme disease, 4 had a positive and 1 an equivocal first-tier C6 EIA test. Of these, all five children had a negative supplemental immunoblot (Table 1).

Discussion

In our cohort of children undergoing emergency department evaluation for Lyme disease, approximately one-third of enrolled children with a history of Lyme disease met diagnostic criteria for current Lyme disease. The odds of testing positive for Lyme disease declined with time from prior diagnosis, likely reflecting recall bias, diagnostic confusion, and time-dependent seroreversion. In our cohort of children presenting to the emergency department for the evaluation of acute Lyme disease, the challenge for the treating clinician is to distinguish new from resolved infection (Nadelman and Wormser 2007).

Serologic testing for Lyme disease is indicated when clinicians suspect disseminated Lyme disease. Because serologic tests measure a host immune response rather than the presence of an infecting organism, the clinician must use clinical context to determine whether a positive test represents active or past infection. In the case of Lyme disease, prior infection does not guarantee protective immunity from a subsequent infection (Aguero-Rosenfeld et al. 1993). Thus, a challenge for clinicians is to determine whether positive serologic testing results after previous Lyme disease represent new infection or an immune response from prior exposure (Branda et al. 2013). In children with a history of Lyme disease, novel methods to directly detect the Borrelia pathogen will provide the most accurate way to distinguish between active and previous infection (Schutzer et al. 2019).

Our findings highlight an important obstacle clinicians may face when evaluating children for Lyme disease. Interestingly, two-thirds of children with a history of Lyme disease did not have positive two-tier serology. Some individuals who are treated for Lyme disease early in their infection may never become seropositive, whereas other treated individuals may have eventual (but unpredictable) decline and disappearance of the specific antibodies detected by these tests (Kalish et al. 2001, Kannian et al. 2007). In addition, many individuals who report prior Lyme disease lack definitive testing to support this history. In several treatment trials for symptoms persisting after Lyme disease, few screened subjects had adequate documentation of prior Lyme disease that the studies had difficulty meeting enrollment targets (Klempner et al. 2001, Krupp et al. 2003, Fallon et al. 2008). Thus, a stated history of Lyme disease should be critically examined, especially when previous medical records are not available for review. Despite these limitations, patient history is frequently the most accurate information available to treating clinicians when evaluating a patient for Lyme disease.

Diagnosing Lyme disease in a patient with prior seropositivity presents a diagnostic challenge. Serology should be interpreted in the context of a patient's prior probability of a diagnosis. In the case of Lyme disease, an intermediate prior probability can be “promoted” to a more confident diagnosis with a positive two-tier serologic test result (Tugwell et al. 1997). If one does not take past infection into account when evaluating a patient for Lyme disease, one may erroneously diagnose current Lyme disease based on an uninformative laboratory result. By contrast, if one discounts a positive serologic test as past infection, the diagnosis of active Lyme disease may be missed. This places a greater diagnostic burden on other sources of information, such as clinical evaluation to avoid both Lyme disease overdiagnosis and delays in making alternative diagnoses.

Our study has several important limitations. First, we obtained the history of Lyme disease from the patients and/or their caregivers and we were unable to confirm the previous diagnosis with either medical record or laboratory review. Although recall bias may have resulted in either under- or over-reporting of prior diagnoses, we utilized the best available clinical information available to the treating team evaluating a symptomatic child. In addition, we were unable to determine the adequacy of antibiotic treatment for the previous Lyme disease. Second, given the challenge for most parents to recall exactly when an event occurred, we collected the time from previous Lyme disease diagnosis in categories that limited our ability to quantify the time-based association between Lyme disease history and current criteria for Lyme disease. Last, as we did not follow children prospectively from the time of initial Lyme disease diagnosis, we cannot determine risk of reinfection.

Conclusions

Two-tier Lyme disease serology suffers from several well-recognized limitations, including the challenges of distinguishing past from current infections. Novel approaches to Lyme disease diagnosis including direct pathogen detection methods are needed to accurately discriminate between these clinical scenarios to avoid both under- and overdiagnosis of Lyme disease.