Revisiting the Lyme Disease Serodiagnostic Algorithm: the Momentum Gathers

TEXT

Lyme disease, or Lyme borreliosis, is caused by the spirochete Borreliella (Borrelia) burgdorferi and is transmitted by the bite of infected ticks of the Ixodes ricinus complex. Lyme disease is the most common vector-borne illness in the United States and Europe. It is estimated that over 300,000 new cases of Lyme disease occur per year in the United States, with the majority of cases occurring in the Mid-Atlantic, Northeast, and upper Midwest regions. B. burgdorferi causes a complex, multisystemic infection in humans. For clinical purposes, Lyme disease is divided into early localized, early disseminated, and late stages. The disease typically begins with erythema migrans (EM), an expanding skin lesion that usually develops between 7 and 14 days at the site of the tick bite. If infection is untreated, spirochetes may disseminate from the site, and patients may present with additional skin lesions (multiple EM) and neurologic, cardiac, and/or rheumatologic manifestations (1). The vast majority of laboratory tests used to support the diagnosis of Lyme disease are based on the detection of the antibody responses against B. burgdorferi in serum, and this is the only type of diagnostic testing approved by the U.S. Food and Drug Administration (FDA).

Ten years ago, an editorial asked if it was time for a change in laboratory testing for Lyme disease (2). It accompanied a study comparing the two-tier algorithm for the serodiagnosis of Lyme disease with a peptide-based enzyme-linked immunoassay (EIA) test (3). Since then, there have been other studies that investigated this question, with the study by Pegalajar-Jurado et al. (4) being the most recent example. In order to fully appreciate how this study fits in the historical context and in the discussion regarding the need to change the currently recommended two-tier algorithm for serodiagnosis of Lyme disease, one needs to go back to the early years after the identification of the disease.

Lyme disease was first recognized in the United States in 1977, with the discovery of the pathogen, B. burgdorferi, in 1982. Because direct tests for B. burgdorferi are challenging and have low sensitivity in most presentations (5), the vast majority of laboratory tests used to support the diagnosis of Lyme disease are based on the detection of the antibody responses against B. burgdorferi in serum. The initial (first-generation) tests used antigen consisting of B. burgdorferi whole-cell sonicates (WCS) to detect immunoglobulin M (IgM) and/or immunoglobulin G (IgG) responses. Early studies showed very poor agreement between test and laboratory results (6) and demonstrated the potential of employing Western blots (WB) to supplement a primary EIA or indirect immunofluorescence assay (IFA) test (7, 8), as well as the need for established standardized interpretation criteria for the tests. In 1994, the Second National Conference on Serologic Diagnosis of Lyme Disease was held in Dearborn, MI, to address the problems of precision and accuracy in serodiagnosis of Lyme disease in the United States. The results led to the current standard two-tiered (STT) testing algorithm (Fig. 1), which was recommended by the Centers for Disease Control and Prevention (CDC) in 1995 (9). The first step of the algorithm uses a sensitive EIA (or, rarely, an IFA). If the initial test result is borderline or positive, the sample is retested using separate IgM and IgG WB as the second step. The WB is interpreted using standardized criteria, requiring at least 2 of 3 signature bands for a positive IgM WB and 5 of 10 signature bands for a positive IgG WB. The IgM WB results are used only for disease with duration of ≤30 days. This two-tier testing algorithm has been the basis for the national standardization of Lyme disease serologic testing methods in the United States and represented a major improvement in the field.

• Open in new tab
• Download powerpoint

FIG 1

Standard two-tiered (STT) testing algorithm and modified two-tiered (MTT) testing algorithm for serodiagnosis of Lyme disease. For patients with signs or symptoms consistent with Lyme disease for less than or equal to 30 days (steps marked with asterisks), the provider may treat the patient and follow up with testing of convalescent-phase serum. Patients with erythema migrans should receive treatment on the basis of the clinical diagnosis.

Since the introduction of the STT test, many studies have evaluated the algorithm in practice (3, 10–14). These studies have shown that, if used as recommended, the approach works relatively well, particularly for later manifestations of the infection. However, the current approach has low sensitivity during early infection because the sensitivity of the IgM WB component is half that of the first-tier component (3, 11, 12, 15, 16). Due to the subjective interpretation of immunoblot bands, there is still inter- and intralaboratory variability, and false-positive IgM WBs are common in commercial laboratories (17). Other issues include the high cost, substantial training requirement, and the labor-intensive process to perform the test, leading to an increased turnaround time for the WB. Finally, a major problem is the confusion by health care providers and patients regarding how to interpret the results from the tests. Laboratories will report the presence of individual IgG and IgM scored bands, plus the result (positive or negative) for the IgM and IgG WB, independent of the duration of the disease, leading to much confusion, misinterpretation, and uncertainty (18).

The need to improve the testing algorithm has been recognized for a long time, including making testing simpler and independent from disease duration and increasing sensitivity in early disease while still maintaining a very high specificity. High specificity is an essential feature for any stand-alone test or test algorithms for serodiagnosis of Lyme disease. It is estimated that about 3.4 million Lyme serologic tests are done in the United States every year (19). Therefore, it is likely that most testing is done in situations with a low pretest probability of Lyme disease, which increases the likelihood of false-positive tests (20, 21). If one estimates a scenario where at least 3 million tests are being performed in cases with a low pretest predictive value, a 1% decrease in specificity can lead to an increase of almost 30,000 false-positive tests.

After the standardization afforded by the implementation of the two-tier testing, the second major advance in the serodiagnosis of B. burgdorferi infection was the discovery of the antibody response against VlsE, the variable surface antigen of B. burgdorferi, and to the C6 peptide, which is derived from invariable region 6 of VlsE (14). This response is immunodominant and highly conserved among B. burgdorferi strains. Because the plasmid containing VlsE is rapidly lost during in vitro passage and poorly expressed in culture, it was not a significant part of tests using WCS. These second-generation tests using recombinant VlsE or synthetic C6 peptide have sensitivities comparable to those of WCS-based EIAs, but with a much higher specificity, most markedly in patients with illnesses other than Lyme disease. The sensitivity of these assays was significantly higher than that of the STT algorithm in patients with earlier manifestations of the infection, while specificity of the assays was slightly lower (3, 10–12, 14, 16).

The use of an alternative (or modified) two-tiered (MTT) testing strategy (Fig. 1) was suggested in 2004 (22) in a letter commenting on a study comparing VlsE1, C6, and pepC10 (a 10-amino-acid segment of outer surface protein C of B. burgdorferi) to the STT approach using a WCS EIA as the initial test (10). A reanalysis of the data set using the proposed approach (WCS EIA first, followed by the two peptide EIAs if positive or equivocal) showed that the strategy performed quite well (23). Later, a follow-up study using a WCS EIA as the initial test and a multiplex assay with VlsE1-IgG and pepC10-IgM as the second-tier test showed again an advantage compared to WB (13). None of these tests were approved by the FDA. An approach using two commercially available EIAs was proposed in Europe, with a study published in 2005 showing that if the results of an IgG EIA with four recombinant antigens and the C6 peptide EIA were concordant, immunoblotting could be omitted (24).

Initial studies had added a VlsE band to the WB, and an alternative approach was proposed using the VlsE band alone as the second-tier test for patients with early or early disseminated disease, while requiring 5 of 11 bands for late disease (12). While this approach increased sensitivity compared to that of the two-tier test and removed the problems of IgM testing, it still required different interpretation criteria based on disease stage and the use of the WB.

Based on the favorable results of the initial studies of the two-EIA approach, this strategy was evaluated in the United States using two FDA-approved first-tier tests, a polyvalent WCS EIA followed by a C6 EIA, for subjects with a positive or equivocal first-tier result (16). The results showed that this two-EIA algorithm had sensitivity similar to that of the C6 EIA, while preserving the specificity of the STT testing. This strategy was shown to be cost-effective in a reanalysis of the data from a large study of the C6 EIA (25). Follow-up studies continue to show that the MTT strategy works well (26–29), confirming that MTT algorithms provided similar or greater sensitivity than STT tests, particularly for early disease, while retaining similar specificity. Noteworthy, in a recent study of patients with EM (26), an MTT approach using a Liaison B. burgdorferi chemiluminescent immunoassay (CLIA) that detects IgM and IgG antibodies against VlsE from B. burgdorferi B31 and B. garinii strain PBi, followed by the C6 EIA, had the highest sensitivity of the MTT tests for acute-phase disease (27).

This brings us to the current study (4), in which the authors compared three MTT algorithms with three STT algorithms using the same assays as the initial tier test, followed by WB. The samples used in the study were from the CDC Lyme Serum Repository, and they have been described and analyzed in other studies (28, 29). The tests chosen included the Liaison VlsE CLIA and the Captia B. burgdorferi IgG/IgM EIA that detects antibodies against whole-cell lysate from B. burgdorferi B31. These were analyzed with results from testing using the C6 B. burgdorferi (Lyme) EIA and the ViraStripe IgM and IgG WBs (read using densitometry) that were performed on the same samples under previous studies (28). Again, the study showed that all MTT algorithms had higher sensitivities than STT algorithms due to the increase in sensitivity in early stages of the infection. There were no significant differences between the sensitivities of the three MTT algorithms in this study, with high agreement between the results. For patients with EM, VlsE/C6, WCS/C6, and WCS/VlsE MTT algorithm sensitivities were 50%, 55%, and 58% for acute-phase samples and 76%, 79%, and 76% for convalescent-phase samples. STT algorithm sensitivities for VlsE/ViraStripe, C6/ViraStripe, and WCS/ViraStripe were 43%, 43%, and 50% for acute-phase samples and 61%, 61%, and 63% for convalescent-phase samples. There were no differences for patients with later manifestations of Lyme disease. The three MTT algorithms also had high specificities, similar to the SST algorithms. There was improved specificity with the MTT VlsE/C6 compared to that of the STT WCS/ViraStripe in control subjects with other diseases.

The Pegalajar-Jurado et al. study results are similar to those of previous studies (16, 26–29) and again prove that MTT algorithms that included VlsE or C6 have improved sensitivity in early Lyme disease presenting with erythema migrans while maintaining high specificity. The need for the algorithm to be highly specific is the reason that two EIAs are necessary. While the WCS EIA and C6 EIA are not completely independent tests (30), the use of two assays increases the specificity compared to that of either test alone.

As shown here and in previous studies (4, 16, 26–29), MTT algorithms will be positive in only about 50% of acute-phase samples from patients with EM, while the STT algorithm will be positive in about 40% (Table 1). Therefore, patients with EM should receive treatment based on the clinical diagnosis (1). While a positive serological test is not necessary for the diagnosis of Lyme disease in patients with EM, this group is an important benchmark for test development, and changes that reduce the window period between time of infection and an accurate test result are welcome improvements. The sensitivity of antibody-based tests increases with the duration of the infection, and MTT and SST algorithms have similar sensitivities in patients with later manifestations of Lyme disease.

There are many advantages to the MTT approach compared to use of STT algorithms, and they are discussed in detail in a report from a conference on diagnostic tests for Lyme disease which was held in 2016 (31). Overall, the MTT approach is more sensitive than the STT algorithm for early disease, has similar sensitivity for later infection, and maintains similar specificity. Additionally, the MTT approach is much simpler than the STT algorithm in all aspects. The MTT algorithm gives an overall single result (positive or negative) for seropositivity in Lyme disease, facilitating the interpretation of the test. The tests are less expensive and less labor-intensive, can be performed using automated instruments, and have objective, quantitative results. The MTT approach can also be used in patients who acquired the infection in Europe when such cases are evaluated in the United States. Alternatively, the WB has the advantage of providing information regarding the expansion of the immune response, and the WB will still have a place in evaluating difficult cases. But for the vast majority of tests performed, this information is not necessary and only increases the confusion regarding interpretation of the results. Certain MTT algorithms may be positive in cases of Borrelia miyamotoi infection, as well as other relapsing-fever Borrelia infections (32).

Moreover, the MTT algorithms open the way for a sensitive and specific point-of-care diagnostic for Lyme disease. These tests would be particularly useful in evaluating patients with stage 2 manifestations of Lyme disease, like facial palsy or carditis, as well as in children with arthritis. At this point, if these patients do not have other manifestations of Lyme disease (erythema migrans) or a very suggestive history, the diagnosis may depend on serological test results. This can cause delay in appropriate therapy and lead to unnecessary treatment, including surgical procedures in pediatric patients with Lyme arthritis (33) and permanent pacemakers in Lyme carditis.

In summary, MTT approaches are a valid substitute for the STT algorithm for the majority of tests performed for the serodiagnosis of Lyme disease, and they are a welcome addition to the currently recommended algorithm for serodiagnosis of Lyme disease. The MTT approach has the potential for rapid diagnostics that substantially improve upon the time to result and point-of-care diagnostic testing that could be used in many clinical settings. These improvements can lead to better management of the individual patient. Hopefully, further advances in diagnostic technologies will continue to refine and expand first-line tests, possibly with the development of combination technologies to detect both antibodies and antigens, which could shorten the detection window period and the time to results, provide information about disease stage, and possibly serve as biomarkers for active infection.