4 Laboratory Diagnostic Testing for Borrelia burgdorferi Infection1
Barbara J.B. Johnson
4.1 Introduction tests include culture of Borrelia from skin or blood and occasionally cerebrospinal fluid Serology is the only standardized type of (CSF), and detection of genetic material by laboratory testing available to support the PCR in skin, blood, synovial fl uid and CSF. clinical diagnosis of Lyme borreliosis (Lyme These tests have specialized roles in research disease) in the USA. It is also the only type of and in academic and reference laboratories diagnostic testing approved by the US Food but are not available for routine use. Culture and Drug Administration (FDA). Of the 77 and PCR each have distinct limitations that devices cleared by the FDA for in vitro will be noted in this chapter. diagnostic use for Lyme disease, all are Diagnostic tests are of clinical value designed to detect immune responses to only if they are used appropriately. This has antigens of Borrelia burgdorferi sensu stricto, become particularly important in the fi eld of particularly IgM and IgG (FDA, 2010). diagnostic testing for Lyme disease, as both Serological tests do not become positive until patients and doctors hear conflicting an infected individual has had time to information about the risk of Lyme disease develop antibodies. In Lyme disease, this in various environments. Furthermore, means that early acute disease characterized patients are sometimes given laboratory by an expanding rash (erythema migrans or diagnostic tests when they lack objective EM) at the site of a tick bite cannot be reliably signs of Lyme disease and a history of diagnosed by serology. Aft er a few weeks potential exposure to infected vector ticks. A of infection, however, immunocompetent healthcare provider must estimate the pre- people will have made enough antibodies test likelihood that a patient has Lyme that serology is useful for confirming disease in order to understand the positive exposure to B. burgdorferi in all subsequent and negative predictive values of tests for stages of Lyme disease. Antibody levels Lyme disease. Fortunately, there are remain elevated for months to years aft er the resources available to assist providers in infection is cured. making this judgement. A variety of direct tests for the agent of It is important to know that laboratories Lyme borreliosis have been developed. Direct may off er ‘in-house’ testing for Lyme disease
1 The fi ndings and conclusions in this article are those of the author and do not necessarily represent the views of the Centers for Disease Control and Prevention.
© CAB International 2011. Lyme Disease: An Evidence-based Approach (ed. J.J. Halperin) 73 74 B.J.B. Johnson
that does not require review and approval by 4.2 Two-tiered Serology: the Current the FDA. Because some in-house tests have Standard for Serodiagnosis in North not been rigorously developed and validated, America the Centers for Disease Control and Prevention (CDC) and FDA recommend that The public health agencies of the USA and these tests only be used when their accuracy Canada advocate a two-step process for and clinical usefulness have been documented measuring antibodies in blood when Lyme in peer-reviewed scientifi c literature (CDC, disease is suspected. The CDC recommends 2005). Unvalidated tests as of 2010 include two-tiered testing both for the evaluation of capture assays for antigens in urine, individual patients (CDC, 1995) and for immunofl uorescence staining or cell sorting epidemiological surveillance for Lyme of cell wall-defi cient or cystic forms of B. disease (CDC, 1997). This recommendation burgdorferi, lymphocyte transformation tests, was developed with the participation of the quantitative CD57 lymphocyte assays, relevant major agencies of the USA, including ‘reverse Western blots’ (Feder et al., 2007), the FDA, the National Institutes of Health, in-house criteria for interpretation of im- the Council of State and Territorial munoblots and measurements of antibodies Epidemiologists, the Association of Public in synovial fl uid. Health Laboratories and the Clinical This chapter considers the diagnostic Laboratory Standards Institute2 (ASTPHLD testing for B. burgdorferi sensu stricto and CDC, 1995). The Canadian Public Health infection, the only organism established to Laboratory Network (2007) guidelines also cause Lyme disease in North America. Lyme recommend two-tiered testing. The Infectious disease also results from infection by Borrelia Diseases Society of America (IDSA) has garinii or Borrelia afzelii in Europe and Asia, as endorsed two-tiered serology to support the well as by the recently described Borrelia diagnosis of Lyme disease in patients who spielmanii in Europe (Wang et al., 1999; Richter have manifestations other than acute EM et al., 2006; Fingerle et al., 2008). Borrelia (Wormser et al., 2006). valaisiana and Borrelia lusitaniae have been A schematic summarizing the features of associated anecdotally with Lyme disease in two-tiered serology is shown in Fig. 4.1. The some parts of Europe (Crowder et al., 2010), fi rst tier consists of a sensitive initial particularly B. lusitaniae in Portugal (Collares- serological test or tests that detect class- Pereira et al., 2004). Borrelia bisett ii has been specifi c antibodies (IgM and IgG, either cultured from a few patients in Europe (Strle together or separately). First-tier tests are et al., 1997), but has not been shown to cause enzyme immunoassays (EIAs) such as human disease in North America. Diagnostic ELISAs or, rarely today, indirect im- tests for B. burgdorferi sensu stricto will not munofl uorescence assays (IFAs) as they necessarily perform well for infections by require a skilled microscopist and cannot be other genospecies of Lyme disease bacteria, scored objectively. If the result of fi rst-tier although some do (e.g. assays based on the testing is negative, the serum is reported to be C6 peptide of the variable surface antigen negative for antibodies to B. burgdorferi and is (VlsE) or the whole VlsE protein). Guidelines not tested further. If the result is positive or for laboratory diagnosis of European Lyme indeterminate (a value that is sometimes borreliosis are available online in English called ‘equivocal’ or ‘borderline’), a second (Health Protection Agency of the UK, 2010; step should be performed. The indeterminate German Society for Hygiene and Micro- category is the range of test values that biology, 2000). overlaps between Lyme disease patients and
2 The latter two were known at the time as the Association of State and Territorial Public Health Laboratory Directors (ASTPHLD) and the National Committee for Clinical Laboratory Standards (NCCLS), respectively. Laboratory Diagnostic Testing for Borrelia burgdorferi Infection 75
controls and is specifi c to each test. Further (ASTPHLD and CDC, 1995; CDC, 1995). information is needed from a second test in Immunoblots and the recommended criteria order to call the specimen positive or for scoring them are shown in Fig. 4.2. These negative. scoring criteria have been validated for The second tier consists of standardized antibodies to B. burgdorferi sensu stricto, the immunoblott ing, either by using Western agent of Lyme disease in North America, but blots or blots striped with diagnostically not for immune responses to other geno- important purifi ed antigens. When an IgG species of Borrelia. immunoblot is scored as positive (Dressler et Two-tiered serology is considered al., 1993; CDC, 1995), two-tiered testing is positive only if the EIA (or IFA) and the reported as positive. When an IgM immunoblot are both positive. Skipping immunoblot is scored as positive (Engstrom either step increases the frequency of false- et al., 1995; CDC, 1995), Lyme disease serology positive results (see below). is reported as positive with the caveat that First-tier tests commonly use whole-cell this fi nding is clinically relevant only in early antigens of B. burgdorferi grown in vitro. The disease, that is, in the fi rst month of illness immunodominant antigen VlsE also has been
Two-tiered serology
(immunoblotting conditionally supplements EIAs)
Tier 1: IgG or IgM EIAs (combined or separate)
Positive or indeterminate Negative
Reported as negative; two-tiered protocol complete
Tier 2: IgG and/or IgM immunoblots (separate)
Either blot positive Negative
IgG-positive reported as positive Reported as negative IgM-positive reported as positive BUT clinically relevant only in early disease of less than 1 month duration
Fig. 4.1. Two-tiered serology for Lyme disease. 76 B.J.B. Johnson
test in 2001 (Liang et al., 1999; FDA, 2010). A diagnostic assay containing entire VlsE molecules expressed as recombinant proteins from both B. burgdorferi and B. garinii (Diasorin) became available as a fi rst-tier test in 2007 (Ledue et al., 2008; FDA, 2010). C6- and VlsE-based assays have the additional feature of detecting antibodies to Eurasian genospecies of Borrelia (i.e. B. garinii and B. azfelii) as well as B. burgdorferi sensu stricto. An extensive peer-reviewed scientifi c literature supports the rationale for and performance of two-tiered serological testing. This algorithm has been validated in both retrospective and prospective studies. The specifi city of two-tiered testing is high – 99% or greater in diagnostic reference centres. The sensitivity is also high aft er the acute phase of EM. Patients with Lyme arthritis or late neuroborreliosis are nearly always sero- positive (97–100%). The rate of seropositivity is lower in patients with acute-phase early neurological disease (80–100%, depending on the population studied). This stage of Lyme disease in particular is the subject of research to improve the sensitivity of serodiagnosis Fig. 4.2. Examples of conventional IgM (left panel) (Dressler et al., 1993; Engstrom et al., 1995; and IgG (right panel) immunoblots. Bands that are Johnson et al., 1996; Bacon et al., 2003; recommended for scoring are labelled. Two Peltomaa et al., 2004; Aguero-Rosenfeld et al., additional bands in the IgG blot are also labelled 2005; Steere et al., 2008; Branda et al., 2010; (OspA and OspB at 31 and 34 kDa, respectively; see text). Blots are considered to be positive if two Wormser et al., 2011; and others). of the three indicated IgM bands or fi ve of the ten Patients oft en inform themselves about indicated IgG bands (excluding OspA and OspB) diagnostic testing for Lyme disease before are present at an intensity equal to or greater than visiting a physician. Unfortunately, the the calibration control. Left panel: IgM blot profi les quality of information available on the for a patient with acute EM (lane 1) and for the Internet varies widely and some is not same patient at convalescence (lane 2). Note the evidence-based (Cooper and Feder, 2004). increase in the number and intensity of the bands Here are some questions that are commonly at convalescence. Right panel: IgG blot profi les for asked: eight patients with later manifestations of Lyme borreliosis (lanes 1–8). P, Positive-control serum; N, negative-control serum; C, calibration control 1. Aren’t ELISAs insensitive and therefore (weak positive control). The molecular mass is unsuitable as fi rst-tier tests? indicated (kDa). The calibration controls (weak 2. Aren’t immunoblots more sensitive than positive controls) have been digitally enhanced for ELISAs? Shouldn’t they be used instead of greater clarity in reproduction. two-tier testing? 3. Why do the recommended blot scoring criteria ignore outer-surface protein A (OspA) approved by the FDA. One small portion of and OspB? OspA was used as a vaccine, so VlsE, a 26 amino acid peptide called C6 that why isn’t it scored in serology? reproduces the sixth constant region of the 4. Why are you disregarding my IgM test protein, was authorized by the FDA for result just because I have had this illness for commercial use (Immunetics) as a fi rst-tier years? Laboratory Diagnostic Testing for Borrelia burgdorferi Infection 77
5. How sensitive is serology in late Lyme positive IgM ELISA test does not necessarily disease? How can you be sure, as support the diagnosis of a new B. burgdorferi seropositivity is part of the case defi nition for infection. Lyme disease, except in patients with EM? How did the misconception arise that ELISAs are insensitive in stages of Lyme Each of these questions will be addressed disease other than EM? Firstly, studies are with reference to the scientifi c literature. oft en cited that describe tests that are obsolete and no longer used. For example, a study conducted in 1992–1994, before two- 4.2.1 How sensitive are ELISAs? tiered testing was recommended as a national standard, is commonly quoted The sensitivity of fi rst-tier tests varies by (Bakken et al., 1997). Many early ELISAs stage of Lyme disease. The antibody response were designed to be stand-alone tests. Some to B. burgdorferi develops over the fi rst few tests were insensitive in order to achieve weeks aft er the spirochaete is introduced into bett er specifi city using whole-cell lysates. the body, in a fashion similar to other bacterial Despite this, false-positive results with some infections. Patients with EM are oft en serum samples from healthy donors seronegative at the time of presentation, as approached 55% (Bakken et al., 1997). Of EM can precede the development of a the 29 ELISAs approved by the FDA before measurable antibody response. The prob- 1993 (FDA, 2010), only three were used ability of seroreactivity increases with recently by a few laboratories (20/417) duration of EM and with the development of that participated in a College of American signs of disseminated disease (Aguero- Pathologists (CAP) profi ciency testing Rosenfeld et al., 1993, 1996; Johnson, 2006). programme (CAP, 2009). Most ELISAs in Although 60% or less of EM patients test current commercial use are suffi ciently positive by ELISA during acute disease, by sensitive to perform well in a two-tiered convalescence 80–90% of treated EM patients testing scheme aft er the EM stage of illness are seropositive (Aguero-Rosenfeld et al., (Aguero-Rosenfeld et al., 1993; Bacon et al., 1993, 1996; Engstrom et al., 1995; Bacon et al., 2003; Johnson et al., 2004; Johnson, 2006). 2003; Johnson et al., 2004; Johnson, 2006). The The excellent performance of ELISAs in well-known insensitivity of ELISAs in acute profi ciency tests can be reviewed by EM is the reason that the CDC and IDSA do subscribers to the surveys carried out by not advocate serological testing of these CAP (2009), although it must be kept in patients. It is appropriate to treat patients mind that only a small number of samples who have rashes compatible with EM with were used in each evaluation. antibiotics based on clinical presentation Secondly, the misconception that ELISAs alone. are insensitive in later Lyme disease is The controversies about serological supported by inappropriately applying data testing do not generally concern test from EM patients to people with later performance in patients with EM, of course. manifestations of this illness. Online Fortunately, aft er the fi rst weeks of illness, the statements such as ‘The test misses 35% of sensitivity of fi rst-tier serology is excellent. culture-proven Lyme disease (only 65% Numerous published studies indicate that sensitivity)’ (ILADS, 2010) fail to note that B. the sensitivity of whole-cell-lysate ELISAs is burgdorferi can be consistently cultured only essentially 100% aft er the EM stage of illness from patients with acute EM, and not from (e.g. Dressler et al., 1993; Bacon et al., 2003; patients with later disease (Aguero-Rosenfeld Johnson et al., 2004; Johnson, 2006). Antibody et al., 2005). It is incorrect to cite the levels remain elevated for months to years performance of a serological test with following antibiotic therapy (Engstrom et al., samples from patients with EM, for whom 1995; Aguero-Rosenfeld et al., 1996; Kalish et serological testing is not recommended, and al., 2001). Because IgM antibody levels may then claim that ELISAs are poor in diagnosing remain elevated aft er treatment, a single infections of longer duration. 78 B.J.B. Johnson
4.2.2 Why not skip the ELISA and go unclear because coinfection with B. burgdorferi directly to immunoblots? may be present (Wormser et al., 1997). Non- specifi c reactions due to polyclonal B-cell It is important to appreciate that fi rst- and activation may occur in conditions such as second-tier tests are not independent indi- Epstein–Barr virus infection or malaria cators of exposure to B. burgdorferi (Wormser (Magnarelli, 1995; Burkot et al., 1997). There et al., 2000). ELISAs and immunoblots are are reports of false-positive reactions in usually constructed with the same antigens – Helicobacter pylori infections and bacterial whole-cell antigens of bacteria grown in endocarditis, although this has not been well culture – but they are processed diff erently. studied (Kaell et al., 1993). In addition, non- There is no a priori reason for immunoblots infectious conditions within the diff erential to be more sensitive than ELISAs. ELISAs diagnosis of Lyme disease yield false-positive provide an estimate of the magnitude of the rates of around 10%, depending on the IgG/IgM humoral antibody response to all of patients studied. Cross-reactions are some- the antigens that are expressed under the times seen in serum from patients with anti- culture conditions used to produce the nuclear antibodies, rheumatoid factor, clinical whole-cell antigen or to the recombinant or rheumatoid arthritis or multiple sclerosis peptide antigens used. ELISA results are (Johnson et al., 2004; Johnson, 2006). objective and quantitative. They can be Omitt ing an ELISA as a fi rst-tier test and correlated with antibody titres. using immunoblot results alone decreases the Immunoblott ing techniques, in contrast, specifi city of serological testing. Decreased separate the many bacterial antigens spatially specifi city has been observed both with on a solid support so that the specifi city and serum samples from healthy blood donors complexity of the antibody responses are from non-endemic areas and with samples revealed. Immunoblots are qualitative or, at from patients with other illnesses within the best, semi-quantitative tests (Fig. 4.2). diff erential diagnosis of Lyme disease. For The rationale for determining IgM and donors, the decrease in specifi city was from IgG antibody profi les by immunoblott ing is 100% for two-tiered testing to 92% for blott ing to learn whether a patient’s antibodies alone in a study by Engstrom et al. (1995) and recognize proteins of B. burgdorferi that have from 100 to 98.5% in work by Johnson et al. been established to be more predictive of (1996). In patients with other illnesses, there Lyme disease than other components of the was a 4% decrease from 100% specifi city for bacteria (Dressler et al., 1993; Engstrom et al., two-tiered testing to 96% for blott ing alone 1995). Many antigens have similarities to (Johnson et al., 1996). those of other organisms, such as proteins Seemingly small changes in specifi city involved in motility (e.g. fl agellin) and have large public health impacts. The volume responses to stress (e.g. ‘heat-shock’ proteins). of laboratory diagnostic testing for Lyme Recognition of one or more antigens from this disease has recently been evaluated. In 2008, set by serum antibodies is not necessarily more than 3.4 million tests for Lyme disease indicative of exposure to B. burgdorferi, were performed in the USA (A. Hinckley, although these reactions contribute to the CDC, 2010, personal communication). Each signal strength measured in an ELISA. 1% decrease in testing specifi city would ELISAs for Lyme disease commonly may generate about 34,000 false-positive results give false-positive results (up to ~55%) in per year. To put this number in context, 38,468 patients with other spirochaetal diseases such cases of Lyme disease (confi rmed plus as tick-borne relapsing fever, syphilis or probable) were reported to the CDC as part of leptospirosis (Johnson et al., 2004: Johnson, the US national system for surveillance of 2006), and cross-reactivity with Treponema notifi able diseases in 2009 (Bacon et al., 2008; denticola in patients with periodontal disease CDC, 2011). has been reported anecdotally. False-positive Why does specifi city decrease if results also may occur in granulocytic immunoblott ing alone is used? The Clinical anaplasmosis, although the frequency is Laboratory Standards Institute identifi es one Laboratory Diagnostic Testing for Borrelia burgdorferi Infection 79
reason: ‘The erroneous scoring of a faint band antibodies to OspA or OspB. When bands at is a common reason for false-positive 31 or 34 kDa are observed, they are virtually readings…’ (NCCLS, 2000). IgM results are always in the context of a robust IgG response more aff ected by this problem than IgG blots. to a large number of scored antigens. In general, IgM antibodies are more non- Patients may inquire specifi cally about specifi cally ‘sticky’ than IgG antibodies, in part why OspA is not scored when it was the basis because of their pentameric structure in serum for an eff ective vaccine (ILADS, 2010). People compared with monomeric IgG. In addition, naturally think of the usual way that vaccines only two of three specifi ed bands are required work, neutralizing infection in a mammalian for an IgM blot to be reported as positive, host, and expect a vaccine antigen to be a whereas fi ve of ten bands are necessary for an good diagnostic antigen. They may be IgG blot to be positive by the recommended unaware that the OspA vaccine works by blot interpretation criteria (CDC, 1995; Fig. killing B. burgdorferi in vector ticks as they 4.2). Consequently, a single erroneously scored feed (de Silva et al., 1996). OspA is well faint band will aff ect IgM results more readily expressed by B. burgdorferi in unfed ticks and than it will aff ect IgG results. is a suitable target for antibodies that enter a Faint bands, particularly in IgM blots, tick during a blood meal from an OspA- may not be diagnostically signifi cant even for vaccinated host. When ticks are exposed to a so-called ‘specifi c’ antigens. If healthcare blood meal and the body temperature of a providers adhere to the recommendation to mammal, B. burgdorferi stops expressing demonstrate that antibodies are present at a OspA (Schwan and Piesman, 2000). Another positive or indeterminate level by a fi rst-tier outer-surface protein, OspC, is expressed test before ordering an immunoblot, the risk instead. Reciprocal expression of these two of an erroneously positive serology based on Osps has been demonstrated at the level of scoring faint bands is reduced but not single cells (Srivastava and de Silva, 2008). It eliminated. is not surprising, therefore, that antibody responses to OspC are diagnostically useful in early Lyme disease, but responses to OspA 4.2.3 Why don’t the scoring criteria for are lacking. immunoblots include OspA and OspB? In later manifestations of Lyme disease, The bands at the 31 and 34 kDa positions of especially Lyme arthritis, some people immunoblots are produced by OspA and develop antibodies to OspA and/or OspB. OspB, respectively (Fig. 4.2). It has been OspA expression is upregulated in an recognized since the early 1990s that infl ammatory milieu such as an arthritic joint. antibodies to OspA and OspB are infrequently OspA expression can be artifi cially up- detected and when they are observed, it is regulated in a controlled in vivo environment usually in patients with longstanding Lyme by exposure to zymosan, a yeast cell-wall arthritis. Ma et al. (1992) wrote that ‘… extract that induces infl ammation (Kalish et antibodies against the 31- and 34-kDa al., 1993; Crowley and Huber, 2003). Thus, it proteins were rarely detected and, is no longer a paradox that B. burgdorferi consequently, became less signifi cant when expresses litt le or no OspA as it is transmitt ed compared with other protein bands in this to mammalian hosts, but that OspA can be study’. Steere’s laboratory reported in produced late in the course of untreated Dressler et al. (1993) that, although antibodies Lyme disease. to OspA and OspB were detectable in some Some claim that patients should be patients with Lyme arthritis or late judged seropositive based on fi nding neurological disease, the frequency of immunoblot bands solely at the 31 or 34 kDa antibody responses to these polypeptides positions, even when their serum is negative was not as high as to ten other antigens. Blot by an ELISA that uses whole-cell antigens. interpretation criteria that could best However, B. burgdorferi grown in culture discriminate Lyme disease patients from expresses OspA and OspB abundantly controls therefore did not include scoring (Crowley and Huber, 2003) and ELISAs made 80 B.J.B. Johnson
from cultured whole cells contain these by the serological fi ndings in patients with antigens. Thus, samples from patients who early neurological disease. In a study of have diagnostically signifi cant levels of patients with facial paralysis, 87% had antibodies to OspA or OspB will react in a diagnostic levels of IgM antibodies, 66% were whole-cell-lysate ELISA. When an ELISA is IgG positive and all were seropositive for at negative but an immunoblot of the same least one antibody class (Peltomaa et al., 2004). sample is scored positive, it is probable that This profi le of antibody reactivity by class faint immunoblot bands are being ‘over- (i.e. a greater frequency of positive IgM read’. responses than IgG, with many people seropositive for both classes) also is seen in patients with other manifestations of early 4.2.4 When is IgM testing clinically neurological disease, typically meningitis useful? and/or radiculoneuritis (Roux et al., 2007). In the event that a patient with a suspected IgM testing should be performed only in early manifestation of Lyme disease is sero- patients with early Lyme disease, defi ned by negative, CDC guidelines note that ‘serologic the CDC (1995) as within the fi rst month of evidence of infection is best obtained by infection. Some investigators have suggested testing of paired acute- and convalescent- recently that IgM responses may have phase samples’ (ASTPHLD and CDC, 1995) diagnostic utility for an additional 2 weeks obtained several weeks apart. (Branda et al., 2010, and personal com- By the time patients develop later munication). Whether the cut-off for IgM manifestations of Lyme disease, they are testing is best set at 4 or 6 weeks, IgM testing almost universally seropositive for IgG is appropriate only during a limited early (Dressler et al., 1993; Kannian et al., 2007). time window. Recall also that serological Numerous studies with robust sample sizes testing is not useful in patients with EM, the have been published about the immune earliest manifestation of Lyme disease, responses in Lyme arthritis. Patients with simply because antibodies have not yet had Lyme arthritis typically have high IgG titres, time to develop. This further restricts the higher than those seen in any of the other clinical utility of IgM testing. various manifestations of Lyme disease, and Some physicians use IgM serology to waning IgM responses. assess patients with longstanding illness Late neurological Lyme disease, pre- (many months to years). They point to the senting as encephalomyelitis, peripheral new IgM responses to OspB that have been neuropathy or encephalopathy, is rare observed to develop late in infection in (Wormser et al., 2006; Halperin et al., 2007). It patients with prolonged disease. This new has been speculated that late neuroborreliosis IgM response, however, occurs in the context has become rarer in recent years due to earlier of a robust IgG response to a large number of diagnosis and treatment, preventing pro- the antigens in the recommended IgG scoring gression to late-stage manifestations. Serum criteria (Kalish et al., 1993). The existence of a IgG antibodies have been found consistently new IgM response in Lyme arthritis patients in patients who have been available for study is not good evidence that IgM serology alone, (Dressler et al., 1993; Bacon et al., 2003). and especially not IgM immunoblott ing For these reasons, the CDC does not alone, can properly support the diagnosis of recommend the use of IgM responses in the late Lyme disease. absence of diagnostic levels of IgG antibodies During the month or so aft er initial to support the diagnosis of any manifestation infection, antibodies rise in titre, recognize an of Lyme disease aft er 1 month of illness. increasing number of borrelial antigens and Furthermore, as noted by Sivak et al. (1996), switch class from a predominantly IgM the predictive value of a positive IgM blot is response to IgG. The evolution of the immune ‘poor in patients with minimal clinical responses during early infection is illustrated evidence of Lyme disease’. Laboratory Diagnostic Testing for Borrelia burgdorferi Infection 81
4.2.5 How can you study the sensitivity of EM (>80%) or sometimes early neurological tests for Lyme disease when Lyme disease (Wormser et al., 2006). Patients seropositivity is part of the defi nition with late neurological Lyme disease, a rare of a case? condition, generally have a history of other clinical manifestations of Lyme disease such The clinical signs and symptoms of Lyme as EM or Lyme arthritis. In a report by Bacon disease aft er the fi rst weeks of infection are et al. (2003), 100% of 11 late neurological not unique to this illness. Clinical fi ndings are Lyme disease patients were seropositive. All not specifi c enough to permit a confi dent of these patients had antecedent other clinical diagnosis without laboratory testing. As manifestations of Lyme disease that were the noted by Steere et al. (2008), ‘It is problematic basis for including the serum samples in the to determine the frequency of seroreactivity study. in patients with neurological, cardiac, or joint manifestations of Lyme disease, because serological confi rmation is a part of the case 4.3 Newer Serological Tests defi nition.’ These considerations raise the important question of how to properly select Two-tiered serology has good performance serum samples for studying the performance characteristics, that is, high sensitivity and of serological tests. To avoid circular specifi city aft er the fi rst weeks of B. burgdorferi reasoning, a previous positive serological infection. Experienced laboratories with result should not be the basis for inclusion of good-quality control and quality assurance a specimen in such a study. However, programmes obtain consistent results (e.g. independent assessment of infection status, Bacon et al., 2003; Kannian et al., 2007; CAP, for example by bacteriological culture, is 2009). Nevertheless, there are limitations to routinely successful only in early disease two-tiered testing that are being addressed (Aguero-Rosenfeld et al., 2005) and is by newer testing methods. As noted, two- generally performed only in research sett ings. tiered testing is insensitive in acute EM and To approach this problem, investigators may be negative in early neuroborreliosis. look to the natural history of untreated Lyme Other drawbacks are that the two-step disease. Patients with late disease frequently procedure is complex, technically demanding have a documented history of earlier signs and costly. Immunoblots are only semi- and symptoms of Lyme disease that support quantitative. Traditional blots are hard to the clinical diagnosis. For selection of ‘gold standardize, as reading them involves standard’ specimens for assessment of judgement about the signifi cance of weak serological test performance in later Lyme bands. Other diffi culties with two-tiered disease, serum from patients with antecedent serology are the need to know the date of clinical fi ndings compatible with earlier disease onset to appropriately request IgM Lyme disease are used. Supplementary testing and the inconvenience of sometimes research tests such as PCR add additional having to draw a second blood sample. The confi dence to the classifi cation of some latt er may occur if the second test is indicated specimens (Bradley et al., 1994; Nocton et al., and the fi rst test was performed by a 1994). laboratory that does not off er immunoblott ing. Patients with early neurological Lyme The research community is actively disease commonly have a history of recent addressing these limitations, and a number of EM. Lyme facial paralysis, for example, was new testing approaches have been developed. associated with EM in 72–87% of patients, The Public Health Service agencies have depending upon the study (Peltomaa et al., established the standard that new tests 2004). Patients with carditis, an uncommon should meet or exceed the performance of presentation of Lyme disease that occurs in two-tiered testing in order to be deemed the early weeks of infection and manifests suitable for clinical use (ASTPHLD and CDC, primarily as atrioventricular block, also 1995). New approaches are either improve- typically have either previous or concurrent ments in one of the steps of the two-step 82 B.J.B. Johnson
testing regime or a potential alternative to neuroborreliosis or Lyme arthritis. The two-tiered testing. specifi city of the C6 assay was slightly less Striped blots with defi ned, purifi ed than two-tiered testing (98.9 versus 99.5%, antigens were FDA-approved in 2009 and are P0.05), however, which will be a key now commercially available (FDA, 2010). consideration when the assay is reviewed for Viramed off ers these immunoblots (Virablots) approval as a stand-alone test. as an improvement over Western blots. Bands Various diagnostic testing approaches are striped at pre-defi ned positions so that will off er value to clinicians. The general calibration problems are avoided. They are practitioner may prefer a simple, objective, read with a scanning densitometer to provide less-costly one-step test. The specialist may an objective measure of whether each band prefer the added information that has suffi cient colour density to be scored as a immunoblots provide to diagnose atypical diagnostically signifi cant reaction. Branda et cases. The type and number of reactive bands al. (2010) have devised a two-tiered procedure off er insights about the stage of Lyme disease. consisting of whole-cell ELISAs and IgG Expanding profi les of reactivity with paired Virablots that include a new band of VlsE. samples may support suspicion of ongoing Only a VlsE band would be required for a infection. positive reaction in early Lyme disease and fi ve or more of 11 bands in the late disease (the bands in Fig. 4.2 plus VlsE). This 4.4 Direct Assays approach provides sensitivity comparable to or higher than standard two-tiered testing in Two types of direct assay have been important each stage of Lyme disease, while maintaining in Lyme disease research and are useful in the high specifi city. If adopted, it would render laboratory diagnosis of some patients. These IgM blots obsolete. The problems of false- assays are culture of B. burgdorferi and positive IgM blots due to over-reading of detection of DNA by molecular methods faint bands and the diffi culty of knowing (PCR or quantitative real-time PCR). Neither how long a patient has been infected would culture nor PCR are components of the be avoided. routine evaluation of patients with suspected A second approach, developed by Zeus Lyme disease and no nationally standardized Scientifi c, seeks to avoid immunoblott ing or FDA-approved tests are available. Both altogether by using defi ned peptides in a techniques have played important roles in multiplex microsphere assay on the Luminex understanding the pathogenesis of B. diagnostic platform. This assay, called the burgdorferi infections, however, and have AtheNA Multi-Lyte test system, has been assisted investigators in establishing serum FDA-approved as a fi rst-tier test and also banks from authenticated Lyme disease evaluated with favourable results as an patients. alternative to immunoblott ing when other Direct detection methods have been approved assays are used as the fi rst-tier test reviewed in detail by Aguero-Rosenfeld et al. (FDA, 2010; Porwancher et al., 2011). (2005) and have not changed signifi cantly Both the C6 peptide and whole VlsE since this work was published. B. burgdorferi assays have been approved as alternatives to can be recovered from skin biopsy samples of whole-cell ELISAs as fi rst-tier tests. In EM patients with 50% effi ciency. Effi ciency addition, the Immunetics C6 assay has of recovery is inversely correlated to the recently been evaluated as an assay that could duration of EM, indicating that spirochaetes be used in place of both steps of two-tiered are rapidly cleared from the region of skin testing, that is, as a simple ‘stand-alone’ test. inoculated by tick bite. In acute EM, The C6 ELISA as a single step is signifi cantly spirochaetes also can be grown from blood, more sensitive in patients with EM than two- especially high-volume plasma cultures, with tiered testing (66.5 versus 35.2%, P0.001; recovery rates of 40%. The period of Wormser et al., 2011). Furthermore, the C6 haematogenous dissemination of borreliae, assay performed comparably to two-tiered however, is brief (several weeks). In later testing in sera from patients with early stages of the disease, blood cultures are Laboratory Diagnostic Testing for Borrelia burgdorferi Infection 83
generally negative. There are only anecdotal test result truly has Lyme borreliosis. reports of B. burgdorferi cultured from Negative predictive value is the probability synovial fl uid, an apparently hostile that a patient who has a negative test result environment, and CSF. The low sensitivity of does not have Lyme borreliosis. An assay culture aft er the EM stage of illness (which with high diagnostic sensitivity improves can be treated based on the appearance of the negative predictive value; one with high rash) and the length of time necessary to diagnostic specifi city improves positive monitor cultures (3 weeks or longer, predictive value. depending on the protocol) greatly limit the Serological testing is recommended only clinical usefulness of bacteriological culture. for patients who have appropriate pre-test PCR is a sensitive method to detect B. probabilities of Lyme disease in order for the burgdoferi DNA in skin biopsy and synovial results to have useful predictive values. A fl uid specimens (Dumler, 2001). Aguero- position paper published by the American Rosenfeld et al. (2005) calculated median PCR College of Physicians (ACP) concluded that sensitivities of 64% in skin biopsy samples laboratory testing should be requested only from EM patients (four studies, range 59– for patients who have an estimated pre-test 67%) and 83% in synovial fl uid specimens probability of Lyme disease between 0.20 and (four studies, range 76–100%). PCR has 0.80 (Tugwell et al., 1997). The ACP panel been particularly useful diagnostically in members pointed out that patients who have evaluating patients with treatment-resistant only non-specifi c signs and symptoms of Lyme arthritis (Nocton et al., 1994). DNA illness such as headache, fatigue and muscle detection methods have been less helpful in or joint pains, even when they reside in a evaluating patients with neurological signs. geographical area endemic for Lyme disease, Reported PCR sensitivities in CSF have been have a pre-test probability of Lyme disease of low and highly variable. PCR tests were less than 0.20, usually much less. Patients positive in 38% of early and 25% of late US with non-specifi c fi ndings and no risk of neuroborreliosis patients (n = 60; Nocton et exposure to infected ticks will have an al., 1996). Urine is not a suitable sample for extremely low pre-test probability. PCR testing (Rauter et al., 2005). When the pre-test probability of Lyme disease is greater than 0.80, laboratory evaluation adds litt le useful information 4.5 Appropriate Use of (Tugwell et al., 1997). This situation only Diagnostic Tests occurs in patients presenting with EM in an endemic area, as all of the other clinical Laboratory testing of patients without manifestations of Lyme disease can be found objective signs of Lyme disease or a history in other conditions. of potential exposure to infected vector ticks The risk of Lyme disease is geographically is not clinically useful. Laboratory diagnostic focal. Of more than 300,000 cases reported to tests with excellent sensitivity and specifi city the CDC over the last 15 years, most occurred will not have helpful predictive values if they in ten states of the northeast and upper are used inappropriately (Sackett et al., 1991). midwest. Maps of reported cases of Lyme Predictive value is determined both by test disease by county and tables of incidence by characteristics (sensitivity and specifi city) state are updated annually by the CDC and and, importantly, by the population in which published online (CDC, 2011)c. The mapped it is used. The practice of testing patients density of host-seeking Ixodes scapularis with a low likelihood of Lyme disease can nymphs in the USA is consistent with the generate more false-positive results than patt ern of reported human cases (Diuk- true-positive results, resulting in mis- Wasser et al., 2006). A ‘Lyme disease tick map’ diagnosis and thereby harming ill people has recently become available as an iPhone (Seltzer and Shapiro, 1996; Tugwell et al., application through the Apple iTunes store 1997). (American Lyme Disease Foundation, 2010). The positive predictive value is the The concepts of positive and negative probability that a patient who has a positive predictive value are well established and 84 B.J.B. Johnson
Table 4.1. Effect of disease prevalence on predictive values of diagnostic tests a Prevalence = 1% Test positive Test negative Total Disease 10 0 10 No disease 20 970 990 Total 30 970 1000 Predictive value of a negative result = 970/970 = 100% Predictive value of a positive result = 10/30 = 33% (67% false positives) Prevalence = 40% Test positive Test negative Total Disease 392 8 400 No disease 12 588 600 Total 404 596 1000 Predictive value of a negative result = 588/596 = 99% Predictive value of a positive result = 392/404 = 97% (3% false positives)
aIllustration assumes that test sensitivity and specifi city are each 98%.
have been described carefully elsewhere (e.g. the approximate pre-test likelihood of Lyme Sackett et al., 1991; Seltzer and Shapiro, 1996; arthritis in patients with pronounced knee Tugwell et al., 1997). They are briefl y swelling who reside in an endemic area illustrated in Table 4.1 for two diff erent (Tugwell et al., 1997). clinical situations. In both cases, diagnostic Good tests have markedly diff erent tests with good performance characteristics predictive values depending on the sett ing of are assumed: 98% sensitivity and 98% use (Table 4.1). When the pre-test probability specifi city. In the fi rst situation, the true is 40%, the predictive values of both negative frequency of disease in the population to be and positive results are very high (99% and tested (prevalence) is only 1%. This represents 97%, respectively). However, when the pre- the pre-test likelihood of Lyme disease in a test probability is low, most positive test patient with non-specifi c symptoms and no results are false positives (67%). Clinicians objective physical signs of this illness who are currently ordering an extraordinary resides in an endemic area (CDC, 2011). For number of diagnostic tests for Lyme disease patients with no history of residence in or – more than 3.4 million tests annually, as travel to an endemic area, the prevalence of noted above. It is critically important to the Lyme disease is much less than 1%. In the well-being of patients that tests only be used second situation, the true frequency of Lyme when the predictive value of a positive result disease is 40%. This prevalence (or higher) is is high (Fig. 4.3).
Where disease is rare Positives mostly deceive Even with good tests
Paul Mead
Fig. 4.3. Haiku to diagnostic testing. Laboratory Diagnostic Testing for Borrelia burgdorferi Infection 85
References VlsE band as the second-tier test. Clinical Infectious Diseases 50, 20–26. Aguero-Rosenfeld, M.E., Nowakowski, J., Burkot, T.R., Schriefer, M.E. and Larsen, S.A. McKenna, D.F., Carbonaro, C.A. and Wormser, (1997) Cross-reactivity to Borrelia burgdorferi G.P. (1993) Serodiagnosis in early Lyme proteins in serum samples from residents of a disease. Journal of Clinical Microbiology 31, tropical country nonendemic for Lyme disease. 3090–3095. Journal of Infectious Diseases 175, 466–469. Aguero-Rosenfeld, M.E., Nowakowski, J., Bittker, CAP (2009) Tick-transmitted Disease Survey TTD- S., Cooper, D., Nadelman, R.B. and Wormser, B. College of American Pathologists, Northfield, G.P. (1996) Evolution of the serologic response IL. to Borrelia burgdorferi in treated patients with Canadian Public Health Laboratory Network (2007) culture-confi rmed erythema migrans. Journal of The laboratory diagnosis of Lyme borreliosis: Clinical Microbiology 34, 1–9. guidelines from the Canadian Public Health Laboratory Network. Canadian Journal of Aguero-Rosenfeld, M.E., Wang, G., Schwartz, I. Infectious Diseases and Medical Microbiology and Wormser, G.P. (2005) Diagnosis of Lyme 18, 145–148. borreliosis. Clinical Microbiology Reviews 18, CDC (1995) Recommendations for test 484–509. performance and interpretation from the second American Lyme Disease Foundation (2010) Lyme national conference on serologic diagnosis of disease tick map iPhone app
Diuk-Wasser, M.A., Gatewood, A.G., Cortinas,
immunodominant conserved region of Borrelia Sackett, D.H., Haynes, R.B., Guyatt, G.H., and burgdorferi VlsE. Journal of Clinical Microbiology Tugwell, P. (1991) Clinical Epidemiology: a 37, 3990–3996. Basic Science for Clinical Medicine. Little, Ma, B., Christen, B., Leung, D. and Vigo-Pelfrey, C. Brown and Co., Boston, MA. (1992) Serodiagnosis of Lyme borreliosis by Schwan, T.G. and Piesman, J. (2000) Temporal Western immunoblot: reactivity of various changes in outer surface proteins A and C of signifi cant antibodies against Borrelia the Lyme disease-associated spirochete, burgdorferi. Journal of Clinical Microbiology 30, Borrelia burgdorferi, during the chain of infection 370–376. in ticks and mice. Journal of Clinical Microbiology Magnarelli, L.A. (1995) Current status of laboratory 38, 382–388. diagnosis for Lyme disease. American Journal Seltzer, E.G. and Shapiro, E.D. (1996) Misdiagnosis of Medicine 98, S10–S12; discussion S12–S14. NCCLS (2000) Western blot assay for antibodies to of Lyme disease: when not to order serologic Borrelia burgdorferi; approved guideline. tests. Pediatric Infectious Disease Journal 15, National Committee for Clinical Laboratory 762–763. Standards document M34-A, Vol. 20, No. 20. Sivak, S.L., Aguero-Rosenfeld, M.E., Nowakowski, Nocton, J.J., Dressler, F., Rutledge, B.J., Rys, P.N., J., Nadelman, R.B. and Wormser, G.P. (1996) Persing, D.H. and Steere, A.C. (1994) Detection Accuracy of IgM immunoblotting to confi rm the of Borrelia burgdorferi DNA by polymerase clinical diagnosis of early Lyme disease. chain reaction in synovial fl uid from patients Archives of Internal Medicine 156, 2105–2109. with Lyme arthritis. New England Journal of Srivastava, S.Y. and de Silva, A.M. (2008) Medicine 330, 229–234. Reciprocal expression of OspA and OspC in Nocton, J.J., Bloom, B.J., Rutledge, B.J., Persing, single cells of Borrelia burgdorferi. Journal of D.H., Logigian, E.L., Schmid, C.H. and Steere, Bacteriology 190, 3429–3433. A.C. (1996) Detection of Borrelia burgdorferi Steere, A.C., McHugh, G., Damle, N. and Sikand, DNA by polymerase chain reaction in V.K. (2008) Prospective study of serologic tests cerebrospinal fl uid in Lyme neuroborreliosis. for Lyme disease. Clinical Infectious Diseases Journal of Infectious Diseases 174, 623–627. 47, 188–195. Peltomaa, M., Mchugh, G. and Steere, A.C. (2004) Strle, F., Picken, R.N., Cheng, Y., Cimperman, J., The VlsE (IR6) peptide ELISA in the serodiagnosis of Lyme facial paralysis. Otology Maraspin, V., Lotric-Furlan, S., Ruzic-Sablijic, E. and Neurotology 25, 838–841. and Picken, M.M. (1997) Clinical fi ndings for Porwancher, R.H., Hagerty, C.G., Fan, J., patients with Lyme borreliosis caused by Landsberg, L., Johnson, B.J.B., Kopnitsky, M., Borrelia burgdorferi sensu lato with genomic Kulas, K., and Wong, S.J. (2011) Multiplex and phenotypic similarities of strain 25015. immunoassay for Lyme disease using VlsE1- Clinical Infectious Diseases 25, 273–280. IgG and pepC10-IgM antibodies: improving test Tugwell, P., Dennis, D.T., Weinstein, A., Wells, G., performance through bioinformatics. Clinical Shea, B., Nichol, G., Hayward, R., Lightfoot, R., and Vaccine Immunology 18, 851–859. Baker, P. and Steere, A.C. (1997) Laboratory Rauter, C., Mueller, M., Diterich, I., Zeller, S., evaluation in the diagnosis of Lyme disease. Hassler, D., Meergans, T. and Hartung, T. (2005) Annals of Internal Medicine 127, 1109–1123. Critical evaluation of urine-based PCR assay for Wang, G., Van Dam, A.P., Schwartz, I. and Dankert, diagnosis of Lyme borreliosis. Clinical and J. (1999) Molecular typing of Borrelia burgdorferi Diagnostic Laboratory Immunology 12, 910–917. sensu lato: taxonomic, epidemiological, and Richter, D., Postic, D., Sertour, N., Livey, I., clinical implications. Clinical Microbiology Matuschka, F.R. and Baranton, G. (2006) Reviews 12, 633–653. Delineation of Borrelia burgdorferi sensu lato Wormser, G.P., Horowitz, H.W., Nowakowski, J., species by multilocus sequence analysis and confi rmation of the delineation of Borrelia Mckenna, D., Dumler, J.S., Varde, S., Schwartz, spielmanii sp. nov. International Journal of I., Carbonaro, C. and Aguero-Rosenfeld, M. Systematic and Evolutionary Microbiology 56, (1997) Positive Lyme disease serology in 873–881. patients with clinical and laboratory evidence of Roux, F., Boyer, E., Jaulhac, B., Dernis, E., Closs- human granulocytic ehrlichiosis. American Prophette, F. and Puechal, X. (2007) Lyme Journal of Clinical Pathology 107, 142–147. meningoradiculitis: prospective evaluation of Wormser, G.P., Carbonaro, C., Miller, S., biological diagnosis methods. European Journal Nowakowski, J., Nadelman, R.B., Sivak, S. and of Clinical Microbiology and Infectious Diseases Aguero-Rosenfeld, M.E. (2000) A limitation of 26, 685–693. 2-stage serological testing for Lyme disease: 88 B.J.B. Johnson
enzyme immunoassay and immunoblot assay babesiosis: clinical practice guidelines by the are not independent tests. Clinical Infectious Infectious Diseases Society of America. Clinical Diseases 30, 545–548. Infectious Diseases 43, 1089–1134. Wormser, G.P., Dattwyler, R.J., Shapiro, E.D., Wormser, G., Schriefer, M., Aguero-Rosenfeld, M., Halperin, J.J., Steere, A.C., Klempner, M.S., Levin, A., Steere, A., Nadelman, R., Krause, P.J., Bakken, J.S., Strle, F., Stanek, G., Nowakowski, M., Marques, A., Johnson, B. and Bockenstedt, L., Fish, D., Dumler, J.S. and Dumler, J. (2011) Single-tier testing with the C6 Nadelman, R.B. (2006) The clinical assessment, peptide ELISA kit compared with two-tiered treatment, and prevention of Lyme disease, testing for Lyme disease. Journal of the human granulocytic anaplasmosis, and American Medical Association (in press).