Lyme Disease Testing in a High-Incidence State: Clinician Knowledge and Patterns

Lyme Disease Testing in a High-Incidence State: Clinician Knowledge and Patterns Abstract Objectives Lyme disease (LD) incidence is increasing, but data suggest some clinicians are not fully aware of recommended procedures for ordering and interpreting diagnostic tests. The study objective was to assess clinicians’ knowledge and practices regarding LD testing in a high-incidence region. Methods We distributed surveys to 1,142 clinicians in the University of Vermont Medical Center region, of which 144 were completed (12.6% response rate). We also examined LD laboratory test results and logs of calls to laboratory customer service over a period of 2.5 years and 6 months, respectively. Results Most clinicians demonstrated basic knowledge of diagnostic protocols, but many misinterpreted Western blot results. For example, 42.4% incorrectly interpreted a positive immunoglobulin M result as an overall positive test in a patient with longstanding symptoms. Many also reported receiving patient requests for unvalidated tests. Conclusions Additional education and modifications to LD test ordering and reporting systems would likely reduce errors and improve patient care. Lyme disease, Tickborne disease, Diagnostic testing, Test interpretation In the United States, Lyme disease is caused by Borrelia burgdorferi sensu stricto, which is transmitted by the bite of infected Ixodes scapularis and Ixodes pacificus ticks.1 A recently described bacterium, Borrelia mayonii, has also been shown to cause Lyme disease.2 Each year, at least 3.4 million Lyme disease tests are performed by commercial laboratories and an estimated 300,000 persons are diagnosed with Lyme disease.2,3 Most cases occur in the Northeast, mid-Atlantic states, and upper Midwest, with a smaller number of cases in Pacific Coast states. In Vermont, the incidence of Lyme disease is high (71 cases per 100,000 persons per year), and the number of reported cases has steadily increased since 2005.4 Patients with erythema migrans who live in or have traveled to areas endemic for Lyme disease can be diagnosed without laboratory testing. For all other patients, the Centers for Disease Control and Prevention (CDC) currently recommends a two-tiered serologic test that detects antibodies against B burgdorferiFigure 1.5 The first step uses an enzyme immunoassay (EIA) or immunofluorescent assay (IFA) to quantify total potential antibodies against B burgdorferi. If the first-step EIA or IFA is negative, no further testing of the specimen is necessary. If positive or equivocal, the second step, Western immunoblotting (WB), should be performed to assess for the presence of antibodies specific to B burgdorferi proteins. For patients who have had signs or symptoms for 30 days or less, both immunoglobulin M (IgM) and immunoglobulin G (IgG) antibodies should be assessed. For those who have been ill for more than 30 days, only WB IgG antibodies should be assessed, since this is typically sufficient to detect infection and avoids potential false-positive results from cross-reactive IgM antibodies. Results of two-tiered serology are considered positive only if the EIA is positive or equivocal and the WB is also positive (Figure 1).5,6 Figure 1 View largeDownload slide Two-tiered serologic testing algorithm for Lyme disease. Figure 1 View largeDownload slide Two-tiered serologic testing algorithm for Lyme disease. Because serology detects the body’s immune response to infection, it has low sensitivity during the first few weeks of infection while antibodies are still developing.7,8 After this time, sensitivity of two-tiered testing increases to 70% to 80% for early disseminated disease and nearly 100% for late disease. Because antibodies can persist for years after infection, serology should not be used as a test of cure.9 In addition to the recommended serologic tests, several other tests are available. Culture and polymerase chain reaction (PCR) assays of cerebrospinal fluid (CSF), skin biopsy specimens, or blood can occasionally be useful but are limited by low clinical sensitivity and are more commonly used in the research setting. Although PCR of synovial fluid is reasonably sensitive in patients who have not been treated, serologic testing remains the preferred diagnostic method for confirming Lyme disease in patients with arthritis.10 Some commercial laboratories claim to specialize in testing for tickborne diseases and offer their own laboratory-developed tests for Lyme disease. Tests offered include urine antigen assays, lymphocyte transformation tests, quantitative CD57 assays, and measurements of antibodies in synovial fluid.11–13 Such tests have not been sufficiently clinically validated. For example, one study by Klempner et al13 compared urine antigen testing to two-tiered serologic testing and found that it was unreliable and led to false-positive results. These tests are not cleared by the US Food and Drug Administration and not recommended for clinical use. Data suggest that some clinicians are unaware of proper testing procedures for Lyme disease. A 2011 nationwide survey of physicians and nurse practitioners found that in the previous year, 13.8% of respondents had ordered a Lyme disease test for a patient who was asymptomatic, which is not recommended, and 7.4% of respondents had incorrectly ordered a Lyme disease test as a test of cure (CDC, unpublished data). On a local level, anecdotal experiences from the University of Vermont Medical Center (UVMMC) laboratory indicate that some clinicians still find ordering and interpreting Lyme disease tests challenging and confusing. Inappropriate ordering or interpretation of test results can lead to misdiagnosis.14 Although national data indicate there is some confusion with testing, detailed and systematic information on laboratory testing knowledge and practices by clinicians is limited. This information would help tailor educational interventions and evaluate potential improvements to the laboratory ordering and reporting system. The objective of this study was to describe clinicians’ knowledge and practices regarding diagnostic testing for Lyme disease in Vermont. Materials and Methods This study was conducted at UVMMC, Vermont’s only academic medical center. Located in Burlington, the state’s largest city, it functions as a regional referral center for approximately 1 million people and a community hospital for 160,000 individuals. As a regional reference laboratory, UVMMC performs approximately 2.7 million tests annually. For Lyme disease testing, the UVMMC laboratory performs the recommended two-tiered serologic test in house. Two outlying hospitals perform their own Lyme EIA tests and, if positive or equivocal, send samples to UVMMC for follow-up WB testing. Otherwise, the WB as a standalone test is not offered by UVMMC, according to current guidelines. To order Lyme disease testing, clinicians order the “Lyme antibody” test. If the EIA is negative, the result is reported as “negative.” If the EIA is positive or equivocal, that result is reported as “Lyme antibody screen positive or equivocal. Western blot confirmation to follow.” For WB testing, the result is reported as positive or negative for IgM and IgG, as well as the specific bands identified. Comments are included in the report to aid in interpretation (see Discussion section for complete comments). UVMMC also offers CSF antibody testing; PCR on serum, CSF, and synovial fluid; and a serum tick-borne disease antibody panel as send-out tests through the Mayo Medical Laboratories. The serum tickborne disease antibody panel includes the two-tiered serologic testing for Lyme disease. We developed an online survey consisting of 13 questions regarding clinician characteristics and knowledge, attitudes, and practices related to diagnostic testing for Lyme disease (Supplementary Figure 1; all supplemental materials can be found at American Journal of Clinical Pathology online). A link to this anonymous survey was sent to 1,142 clinicians who had ordered any laboratory test from UVMMC within the past year. The survey collected information on respondent demographics, basic Lyme disease testing knowledge, alternative testing requests, and confusion with test interpretation. We also asked respondents to rate their satisfaction with the UVMMC laboratory’s Lyme disease test ordering system and the Lyme disease test result reporting system. A free-text option was available for some questions to allow respondents to elaborate on their responses. The UVMMC Laboratory Customer Service team collects basic information on phone calls received as part of standard practice. We collected additional data regarding phone calls received about Lyme disease testing prospectively during a 6-month period from May to October 2015. Customer service representatives filled out a standard data collection form each time they received a phone call regarding Lyme disease testing (Supplementary Figure 2). Responses were summarized using descriptive statistics. Inferential statistical calculations were performed with parametric methods using the SAS program, version 9.3 (SAS Institute, Cary, NC). Fisher exact test was used to compare the proportion of survey questions answered correctly among clinicians using the R program, version 3.3.1 (R Core Team, 2016).15 We compared the performance of those in primary care specialties, who are likely to be the first health care contacts of potential patients with Lyme disease, with those in other specialties; we excluded infectious disease specialists from this comparison since they likely have more detailed knowledge of testing protocols, and numbers were too few to perform separate statistical analyses. The study was reviewed and approved by the University of Vermont Institutional Review Board. Results Clinician Survey Of 1,142 survey links sent, 192 surveys were started and 147 were completed, yielding a 12.9% response rate. Of the 147 completed surveys, three were excluded because respondents were not physicians, nurse practitioners, physician assistants, or naturopathic providers. Thus, 144 total responses were included. Characteristics of survey respondents are listed in Table 1 . All respondents practiced within Vermont, primarily in the greater Burlington area. Table 1 Characteristics of Survey Respondents Characteristic   No. (%)  Total  144 (100)  Provider type   Attending physician  118 (81.9)   Physician assistant  11 (7.6)   Nurse practitioner  11 (7.6)   Physicians-in-training  2 (1.4)   Naturopathic provider  2 (1.4)  Specialty   Primary care    Family medicine  36 (25.0)    Internal medicine  27 (19.4)    Pediatrics  22 (15.3)   Subspecialists    Obstetrics and gynecology  8 (5.6)    Emergency medicine  7 (4.9)    Dermatology  6 (4.2)    Infectious diseases  4 (2.8)    Neurology  4 (2.8)    Orthopedics  4 (2.8)   Other  26 (18.1)  Practice setting   Outpatient  77 (53.5)   Inpatient  17 (11.8)   Combined  50 (34.7)  Characteristic   No. (%)  Total  144 (100)  Provider type   Attending physician  118 (81.9)   Physician assistant  11 (7.6)   Nurse practitioner  11 (7.6)   Physicians-in-training  2 (1.4)   Naturopathic provider  2 (1.4)  Specialty   Primary care    Family medicine  36 (25.0)    Internal medicine  27 (19.4)    Pediatrics  22 (15.3)   Subspecialists    Obstetrics and gynecology  8 (5.6)    Emergency medicine  7 (4.9)    Dermatology  6 (4.2)    Infectious diseases  4 (2.8)    Neurology  4 (2.8)    Orthopedics  4 (2.8)   Other  26 (18.1)  Practice setting   Outpatient  77 (53.5)   Inpatient  17 (11.8)   Combined  50 (34.7)  View Large Of three knowledge questions, the mean number of correct responses for all respondents was 2.1 with a standard deviation of 0.7 Table 2. Most clinicians obtained a score of 2 (73, 50.7%) or 3 (44, 30.6%). Using the Fisher exact test, we found no significant difference between the performance of clinicians working in primary care from those working in other noninfectious disease specialties (P = .84). Infectious disease specialists did have an overall higher score (2.8 correct answers for infectious disease vs 2.1 for primary care and 2.0 for others); however, statistical analyses were not performed given the low number of infectious disease specialists. Table 2 Clinician Responses to Survey Questions About Lyme Disease (LD) Serologya Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  1. A patient presents to your office with a rash resembling erythema migrans that began 3 days earlier. You order a serologic test for LD, which yields a negative result. This result rules out LD as a cause of the illness.  True  10 (6.9)  6 (7.1)  4 (7.3)  0 (0.0)  False  134 (93.1)  79 (92.9)  51 (92.7)  4 (100)  2. A 50-year-old female from southern Vermont presents with a swollen, erythematous knee for the past week. She does not remember a tick bite or rash, but she is active outdoors and went on a hiking trip two months ago. You suspect LD. What is the appropriate next step?  Treat with doxycycline (no testing needed)  13 (9.0)  8 (9.4)  5 (9.1)  0 (0.0)  Order WB  4 (2.8)  1 (1.2)  3 (5.5)  0 (0.0)  Order EIA followed by WB if positive  107 (74.3)  60 (70.6)  44 (80.0)  3 (75.0)  Aspirate joint, order PCR of the fluid  20 (13.9)  16 (18.8)  3 (5.5)  1 (25.0)  3. A 45-year-old woman from southern Vermont presents with fatigue and difficulty concentrating for the past two years. She does not remember a tick bite or rash, but occasionally gardens in her backyard. A LD test ordered by her PCP showed an equivocal EIA, positive IgM WB (2/3 bands), and negative IgG WB (1/10 bands). What is your interpretation of these results?  Patient has LD, treat with antibiotics  43 (29.9)  24 (28.2)  19 (34.5)  0 (0.0)  Not LD, IgM is false positive  61 (42.4)  39 (45.9)  18 (32.7)  4 (100)  The patient is in the “window period”  40 (27.8)  22 (25.9)  18 (32.7)  0 (0.0)  Mean (SD) number correct, range 0-3  2.10 (0.74)  2.09 (0.77)  2.05 (0.70)  2.75 (0.50)  Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  1. A patient presents to your office with a rash resembling erythema migrans that began 3 days earlier. You order a serologic test for LD, which yields a negative result. This result rules out LD as a cause of the illness.  True  10 (6.9)  6 (7.1)  4 (7.3)  0 (0.0)  False  134 (93.1)  79 (92.9)  51 (92.7)  4 (100)  2. A 50-year-old female from southern Vermont presents with a swollen, erythematous knee for the past week. She does not remember a tick bite or rash, but she is active outdoors and went on a hiking trip two months ago. You suspect LD. What is the appropriate next step?  Treat with doxycycline (no testing needed)  13 (9.0)  8 (9.4)  5 (9.1)  0 (0.0)  Order WB  4 (2.8)  1 (1.2)  3 (5.5)  0 (0.0)  Order EIA followed by WB if positive  107 (74.3)  60 (70.6)  44 (80.0)  3 (75.0)  Aspirate joint, order PCR of the fluid  20 (13.9)  16 (18.8)  3 (5.5)  1 (25.0)  3. A 45-year-old woman from southern Vermont presents with fatigue and difficulty concentrating for the past two years. She does not remember a tick bite or rash, but occasionally gardens in her backyard. A LD test ordered by her PCP showed an equivocal EIA, positive IgM WB (2/3 bands), and negative IgG WB (1/10 bands). What is your interpretation of these results?  Patient has LD, treat with antibiotics  43 (29.9)  24 (28.2)  19 (34.5)  0 (0.0)  Not LD, IgM is false positive  61 (42.4)  39 (45.9)  18 (32.7)  4 (100)  The patient is in the “window period”  40 (27.8)  22 (25.9)  18 (32.7)  0 (0.0)  Mean (SD) number correct, range 0-3  2.10 (0.74)  2.09 (0.77)  2.05 (0.70)  2.75 (0.50)  EIA, enzyme immunoassay; IgM, immunoglobulin M; PCP, primary care provider; PCR, polymerase chain reaction; SD, standard deviation; WB, Western blot. aCorrect answers are listed in bold text. View Large Most respondents correctly answered basic knowledge questions about Lyme disease testing (Table 2, questions 1 and 2). A clinical scenario regarding a woman with longstanding, nonspecific symptoms and an overall negative serologic test for Lyme disease, however, was answered correctly by only 45.9% of primary care providers and 32.7% of other noninfectious disease specialists (Table 2, question 3). Many respondents reported that patients had asked them for nonstandard serology (38.2%) or unvalidated tests from Lyme disease “specialty” laboratories (21.5%) (Table 3, questions 4 and 5). This was particularly true among primary care practitioners and infectious disease specialists. In addition to the nonstandard tests listed in Table 3, other free-text responses for tests requested included urine antigen testing and direct testing of ticks. Table 3 Clinician Responses to Survey Questions About Alternative Testing Requests Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  4. For diagnosis of Lyme disease, CDC recommends two-tier serologic testing consisting of an EIA test, which if positive or equivocal is followed by a reflex WB. Have you ever been pressured by a patient to order nonstandard serologic testing for Lyme disease (select all that apply)?  Yes—WB only without preceding EIA  28 (19.4)  17 (20.0)  9 (16.4)  2 (50.0)  Yes—serologic tests that use nonstandard interpretation criteria (eg, fewer bands for positive)  34 (23.6)  24 (28.2)  8 (14.5)  2 (50.0)  Yes—other  12 (8.3)  10 (11.8)  2 (3.6)  0 (0.0)  No  89 (61.8)  46 (54.1)  41 (74.5)  2 (50.0)  5. Have you ever been pressured by a patient to order unvalidated tests from Lyme “specialty labs,” including tests that are not FDA approved for diagnosis of Lyme disease (select all that apply)?  Yes—CD57 assays  5 (3.5)  1 (1.2)  4 (7.3)  0 (0.0)  Yes—lymphocyte transformation tests  2 (1.4)  1 (1.2)  2 (3.6)  0 (0.0)  Yes—novel culture techniques  4 (2.8)  3 (3.5)  1 (1.8)  0 (0.0)  Yes—other  23 (16.0)  18 (21.2)  3 (5.5)  2 (50.0)  No  113 (78.5)  63 (74.1)  48 (87.3)  2 (50.0)  Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  4. For diagnosis of Lyme disease, CDC recommends two-tier serologic testing consisting of an EIA test, which if positive or equivocal is followed by a reflex WB. Have you ever been pressured by a patient to order nonstandard serologic testing for Lyme disease (select all that apply)?  Yes—WB only without preceding EIA  28 (19.4)  17 (20.0)  9 (16.4)  2 (50.0)  Yes—serologic tests that use nonstandard interpretation criteria (eg, fewer bands for positive)  34 (23.6)  24 (28.2)  8 (14.5)  2 (50.0)  Yes—other  12 (8.3)  10 (11.8)  2 (3.6)  0 (0.0)  No  89 (61.8)  46 (54.1)  41 (74.5)  2 (50.0)  5. Have you ever been pressured by a patient to order unvalidated tests from Lyme “specialty labs,” including tests that are not FDA approved for diagnosis of Lyme disease (select all that apply)?  Yes—CD57 assays  5 (3.5)  1 (1.2)  4 (7.3)  0 (0.0)  Yes—lymphocyte transformation tests  2 (1.4)  1 (1.2)  2 (3.6)  0 (0.0)  Yes—novel culture techniques  4 (2.8)  3 (3.5)  1 (1.8)  0 (0.0)  Yes—other  23 (16.0)  18 (21.2)  3 (5.5)  2 (50.0)  No  113 (78.5)  63 (74.1)  48 (87.3)  2 (50.0)  CDC, Centers for Disease Control and Prevention; EIA, enzyme immunoassay; FDA, US Food and Drug Administration; WB, Western blot. View Large When asked specifically about WB interpretation, many respondents (44.4%) indicated that a positive IgM WB result in a patient with longstanding symptoms lasting more than 30 days had created confusion for either themselves or their patients ( Table 4 , question 6). Similarly, many (46.5%) responded that a laboratory report that listed individually positive WB bands in the context of an overall negative test had also caused confusion for either themselves or their patients (Table 4, question 7). Again, primary care practitioners and infectious disease specialists reported the majority of these problems. Table 4 Clinician Responses to Survey Questions About Western Blot (WB) Interpretation Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  6. According to CDC guidelines, due to potential cross-reactivity with other conditions IgM WB should only be ordered or interpreted for patients with ≤30 days of symptoms. Has the presence of a false-positive IgM WB—but overall negative test—in a patient with longstanding symptoms ever caused confusion for you or your patient?  Yes, more than once in the past year  9 (6.3)  4 (4.8)  2 (3.6)  3 (75.0)  Yes, once in the past year  14 (9.7)  9 (10.6)  5 (9.1)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  30 (35.3)  11 (20.0)  0 (0.0)  No  80 (55.6)  42 (49.4)  37 (67.3)  1 (25.0)  7. According to CDC guidelines, IgM WB is considered positive if ≥2 out of 3 scored bands are present. IgG WB is considered positive if ≥5 out of 10 scored bands are present. Has the presence of individual bands on a WB—but an overall negative test—ever caused confusion for you or your patient?  Yes, more than once in the past year  14 (9.7)  8 (9.4)  3 (5.5)  3 (75.0)  Yes, once in the past year  12 (8.3)  9 (10.6)  3 (5.5)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  29 (34.1)  12 (21.9)  0 (0.0)  No  77 (53.5)  39 (45.9)  37 (67.3)  1 (25.0)  Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  6. According to CDC guidelines, due to potential cross-reactivity with other conditions IgM WB should only be ordered or interpreted for patients with ≤30 days of symptoms. Has the presence of a false-positive IgM WB—but overall negative test—in a patient with longstanding symptoms ever caused confusion for you or your patient?  Yes, more than once in the past year  9 (6.3)  4 (4.8)  2 (3.6)  3 (75.0)  Yes, once in the past year  14 (9.7)  9 (10.6)  5 (9.1)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  30 (35.3)  11 (20.0)  0 (0.0)  No  80 (55.6)  42 (49.4)  37 (67.3)  1 (25.0)  7. According to CDC guidelines, IgM WB is considered positive if ≥2 out of 3 scored bands are present. IgG WB is considered positive if ≥5 out of 10 scored bands are present. Has the presence of individual bands on a WB—but an overall negative test—ever caused confusion for you or your patient?  Yes, more than once in the past year  14 (9.7)  8 (9.4)  3 (5.5)  3 (75.0)  Yes, once in the past year  12 (8.3)  9 (10.6)  3 (5.5)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  29 (34.1)  12 (21.9)  0 (0.0)  No  77 (53.5)  39 (45.9)  37 (67.3)  1 (25.0)  CDC, Centers for Disease Control and Prevention; IgG, immunoglobulin G; IgM, immunoglobulin M. View Large Among the 142 respondents who answered a question about satisfaction with the UVMMC Lyme disease test ordering system, most indicated they were somewhat or very satisfied (85, 59.9%) or neutral (49, 34.5%). Most also indicated that they were somewhat or very satisfied (89, 62.7%) or neutral (47, 33.1%) with regard to the Lyme disease test reporting system. Some free-text comments indicated lack of knowledge or confusion about appropriate Lyme disease testing. For example, some clinicians responded that they would like to be able to order WB outside of the two-tiered testing algorithm, which is not recommended. One clinician also indicated that patients asking for nonstandard tests either have seen or would eventually see clinicians who have deemed themselves to be Lyme specialists. Lyme Disease Testing A total of 20,295 serum EIA tests were performed during the 2.5-year study period. The median age of patients tested was 49 years for female patients and 52 years for male patients; most EIA test samples were from female patients (55.3%). The largest number of EIA tests was ordered for women aged 50 to 59 years (Supplementary Figure 3). A total of 1,488 serum EIA tests (7.3%) were positive, 18,684 (92.1%) were negative, 99 were equivocal (0.5%), and 24 (0.1%) were unsuitable for analysis. Serum EIA testing peaked during the summer months, as did positive results of EIA testing (Supplementary Figure 4). During the same period, a total of 1,687 WB tests were performed, including 103 that were positive by EIA at an outside facility and sent for WB testing at UVMMC. Of those, 538 (31.9%) were positive by IgM only, 193 (11.4%) were positive by IgG only, 239 (14.2%) were positive by both, and 716 (42.4%) were negative by both; one was uninterpretable. Positive IgM with negative IgG peaked in the early summer months, which may correspond with early localized disease, while positive IgG results with or without a positive IgM were more sporadic throughout the year (Supplementary Figure 5). Patients with a positive IgG WB had a bimodal age distribution; median age was 51 years. Males aged 19 years or younger and 50 to 69 years had the highest number of positive IgG WB tests. While more EIA test samples were performed on female patients, more positive IgG WB results were from male patients (Supplementary Figure 6). Nine synovial fluid samples were sent for PCR testing, and all were negative. Of the 516 CSF samples sent for antibody testing, 14 (2.7%) were positive by EIA (Supplementary Figure 7). Of these 14 positive EIAs, three CSF samples were negative by WB, seven were positive for IgG but not IgM, and four were positive for both IgG and IgM (Supplementary Figure 7). Customer Service Calls A total of 13 telephone calls related to Lyme disease were logged by laboratory customer service during the study period. Clinicians called with questions regarding testing algorithm (seven calls), result interpretation (two calls), adding on tests to patient specimens (two calls), as well as requests for testing outside the recommended CDC guidelines (two calls, request for WB only and request for WB when EIA was negative). Discussion This study demonstrated that most clinicians surveyed in Vermont have basic knowledge of Lyme disease diagnostic testing protocols. In our survey, however, most clinicians struggled with recognizing false-positive IgM results in a patient with longstanding symptoms and interpreting WB results with individually positive bands but an overall negative test. Nearly half of survey respondents also indicated that situations such as these caused confusion in their clinical practices on at least one occasion. These data suggest there is a need for additional clinician education regarding Lyme disease testing. Furthermore, there may be a role for laboratories to modify how they report results to improve accuracy of interpretation by clinicians. At the time of the survey, the UVMMC laboratory reported both IgM and IgG results (whether positive or negative and which bands, if any, were identified), regardless of the time interval of symptoms. If only IgG is positive, the interpretation states, “Indicative of B. burgdorferi infection at some time in the past.” If only IgM is positive, the interpretation states, “Indicative of early B. burgdorferi infection. A new serum specimen should be analyzed in 14-21 days to demonstrate seroconversion of IgG. IgM Western blot is of diagnostic utility only during the first four weeks after the onset of disease. Specimens collected more than 1 month following the onset of disease with positive IgM and negative IgG results more likely represent a false positive.” If both IgM and IgG are positive, the interpretation states, “Indicative of active or previous B. burgdorferi infection. IgM Western blot is of diagnostic utility only during the first four weeks after the onset of disease.” If both are negative, the interpretation states, “Specific serologic response to B. burgdorferi infection is not detected. This may indicate lack of infection, lack of seroconversion, or low/undetectable antibody levels. If clinically indicated, a new serum specimen should be submitted in 7-14 days.” On all results, a comment is included that explains the CDC two-tiered approach and band interpretation criteria. IgM has high rates of cross-reactivity and false-positive results beyond the initial 30-day symptomatic period, and reporting IgM results for patients with symptoms lasting more than 30 days may lead to some confusion, despite comments added to all results. It is possible that some clinicians only read the “positive” or “negative” results without the accompanying interpretation or that there continues to be confusion despite the included comments. To address this issue, the UVMMC laboratory is currently exploring how best to modify the Lyme disease two-tiered test-ordering mechanism such that they are prompted to state whether or not the patient has had symptoms for more than 30 days. If yes, then only the IgG result would be reported; if no, then both the IgM and IgG results would be reported. This would be consistent with the current CDC-recommended guidelines and could minimize the number of false-positive results reported for patients with symptoms more than 30 days. Positive IgG results were identified during the summer months, as expected, and the groups with the highest proportion of positive tests were similar to the highest risk groups in CDC surveillance data. These results underscore that these particular age and sex groups are at higher risk and would benefit from targeted prevention efforts. Heightened clinician suspicion could enhance early recognition and treatment. This study was subject to limitations. Results were based on respondents’ willingness to complete the survey and therefore were subject to responder bias. However, the respondent characteristics seemed to be representative of the community and encompass many of the challenges in laboratory test ordering and interpretation typically identified. In addition, only 13 phone calls related to Lyme disease were logged by laboratory customer service during the study period, which is lower than expected. This is likely because the log sheet for this study was an additional form that customer service personnel needed to have on hand and complete. In a busy customer service call center that receives approximately 8,000 phone calls per month, it is possible that some of the phone calls related to Lyme disease were not recorded. Therefore, it is likely that the actual number of phone calls regarding Lyme disease testing and interpretation is higher than captured in this study. In conclusion, clinicians would benefit from targeted education regarding Lyme disease testing to improve test-ordering practices and accuracy of test interpretation. Furthermore, clinical laboratories could improve their Lyme disease test ordering and reporting system to tailor reporting of test results based on the patient’s duration of symptoms (ie, removing IgM WB results from the laboratory report for patients with longstanding symptoms). Last, education of patients regarding unvalidated Lyme disease tests could reduce alternative test-seeking behavior and decrease the burden of requests to clinicians for these tests. Acknowledgment: We thank Mark Delorey for guidance on statistical analysis. Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the official position of the US Centers for Disease Control and Prevention. References 1. Stanek G, Wormser GP, Gray Jet al.   Lyme borreliosis. Lancet . 2012; 379: 461- 473. Google Scholar CrossRef Search ADS PubMed  2. Pritt BS, Respicio-Kingry LB, Sloan LMet al.   Borrelia mayonii sp. nov., a member of the Borrelia burgdorferi sensu lato complex, detected in patients and ticks in the upper midwestern United States. Int J Syst Evol Microbiol . 2016; 66: 4878- 4880. Google Scholar CrossRef Search ADS PubMed  3. Nelson CA, Saha S, Kugeler KJet al.   Incidence of clinician-diagnosed Lyme disease, United States, 2005-2010. Emerg Infect Dis . 2015; 21: 1625- 1631. Google Scholar CrossRef Search ADS PubMed  4. Vermont Department of Health. Epidemiology of Lyme disease in Vermont. 2013. http://healthvermont.gov/prevent/lyme/documents/epidemiology.pdf. Accessed June 2017. 5. Centers for Disease Control and Prevention. Recommendations for test performance and interpretation from the Second National Conference on Serologic Diagnosis of Lyme Disease. MMWR Morb Mortal Wkly Rep . 1995; 44: 590- 591. PubMed  6. Moore A, Nelson C, Molins Cet al.   Current guidelines, common clinical pitfalls, and future directions for laboratory diagnosis of Lyme disease, United States. Emerg Infect Dis . 2016; 22:1169-1177. 7. Halperin JJ, Baker P, Wormser GP. Common misconceptions about Lyme disease. Am J Med . 2013; 126: 264.e1- 264.e7. Google Scholar CrossRef Search ADS   8. Steere AC, McHugh G, Damle Net al.   Prospective study of serologic tests for Lyme disease. Clin Infect Dis . 2008; 47: 188- 195. Google Scholar CrossRef Search ADS PubMed  9. Kalish RA, McHugh G, Granquist Jet al.   Persistence of immunoglobulin M or immunoglobulin G antibody responses to Borrelia burgdorferi 10-20 years after active Lyme disease. Clin Infect Dis . 2001; 33: 780- 785. Google Scholar CrossRef Search ADS PubMed  10. Aguero-Rosenfeld ME, Wang G, Schwartz Iet al.   Diagnosis of Lyme borreliosis. Clin Microbiol Rev . 2005; 18: 484- 509. Google Scholar CrossRef Search ADS PubMed  11. Nelson C, Hojvat S, Johnson Bet al.  ; Centers for Disease Control and Prevention (CDC). Concerns regarding a new culture method for Borrelia burgdorferi not approved for the diagnosis of Lyme disease. MMWR Morb Mortal Wkly Rep . 2014; 63: 333. Google Scholar PubMed  12. Dattwyler RJ, Arnaboldi PM. Comparison of Lyme disease serologic assays and Lyme specialty laboratories. Clin Infect Dis . 2014; 59: 1711- 1713. Google Scholar CrossRef Search ADS PubMed  13. Klempner MS, Schmid CH, Hu Let al.   Intralaboratory reliability of serologic and urine testing for Lyme disease. Am J Med . 2001; 110: 217- 219. Google Scholar CrossRef Search ADS PubMed  14. Nelson C, Elmendorf S, Mead P. Neoplasms misdiagnosed as “chronic Lyme disease.” JAMA Intern Med . 2015; 175: 132- 133. Google Scholar CrossRef Search ADS PubMed  15. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2016. © American Society for Clinical Pathology, 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Clinical Pathology Oxford University Press

Lyme Disease Testing in a High-Incidence State: Clinician Knowledge and Patterns

Loading next page...
 
/lp/ou_press/lyme-disease-testing-in-a-high-incidence-state-clinician-knowledge-and-VCrqgjeLFf
Publisher
Oxford University Press
Copyright
© American Society for Clinical Pathology, 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US.
ISSN
0002-9173
eISSN
1943-7722
D.O.I.
10.1093/ajcp/aqx153
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives Lyme disease (LD) incidence is increasing, but data suggest some clinicians are not fully aware of recommended procedures for ordering and interpreting diagnostic tests. The study objective was to assess clinicians’ knowledge and practices regarding LD testing in a high-incidence region. Methods We distributed surveys to 1,142 clinicians in the University of Vermont Medical Center region, of which 144 were completed (12.6% response rate). We also examined LD laboratory test results and logs of calls to laboratory customer service over a period of 2.5 years and 6 months, respectively. Results Most clinicians demonstrated basic knowledge of diagnostic protocols, but many misinterpreted Western blot results. For example, 42.4% incorrectly interpreted a positive immunoglobulin M result as an overall positive test in a patient with longstanding symptoms. Many also reported receiving patient requests for unvalidated tests. Conclusions Additional education and modifications to LD test ordering and reporting systems would likely reduce errors and improve patient care. Lyme disease, Tickborne disease, Diagnostic testing, Test interpretation In the United States, Lyme disease is caused by Borrelia burgdorferi sensu stricto, which is transmitted by the bite of infected Ixodes scapularis and Ixodes pacificus ticks.1 A recently described bacterium, Borrelia mayonii, has also been shown to cause Lyme disease.2 Each year, at least 3.4 million Lyme disease tests are performed by commercial laboratories and an estimated 300,000 persons are diagnosed with Lyme disease.2,3 Most cases occur in the Northeast, mid-Atlantic states, and upper Midwest, with a smaller number of cases in Pacific Coast states. In Vermont, the incidence of Lyme disease is high (71 cases per 100,000 persons per year), and the number of reported cases has steadily increased since 2005.4 Patients with erythema migrans who live in or have traveled to areas endemic for Lyme disease can be diagnosed without laboratory testing. For all other patients, the Centers for Disease Control and Prevention (CDC) currently recommends a two-tiered serologic test that detects antibodies against B burgdorferiFigure 1.5 The first step uses an enzyme immunoassay (EIA) or immunofluorescent assay (IFA) to quantify total potential antibodies against B burgdorferi. If the first-step EIA or IFA is negative, no further testing of the specimen is necessary. If positive or equivocal, the second step, Western immunoblotting (WB), should be performed to assess for the presence of antibodies specific to B burgdorferi proteins. For patients who have had signs or symptoms for 30 days or less, both immunoglobulin M (IgM) and immunoglobulin G (IgG) antibodies should be assessed. For those who have been ill for more than 30 days, only WB IgG antibodies should be assessed, since this is typically sufficient to detect infection and avoids potential false-positive results from cross-reactive IgM antibodies. Results of two-tiered serology are considered positive only if the EIA is positive or equivocal and the WB is also positive (Figure 1).5,6 Figure 1 View largeDownload slide Two-tiered serologic testing algorithm for Lyme disease. Figure 1 View largeDownload slide Two-tiered serologic testing algorithm for Lyme disease. Because serology detects the body’s immune response to infection, it has low sensitivity during the first few weeks of infection while antibodies are still developing.7,8 After this time, sensitivity of two-tiered testing increases to 70% to 80% for early disseminated disease and nearly 100% for late disease. Because antibodies can persist for years after infection, serology should not be used as a test of cure.9 In addition to the recommended serologic tests, several other tests are available. Culture and polymerase chain reaction (PCR) assays of cerebrospinal fluid (CSF), skin biopsy specimens, or blood can occasionally be useful but are limited by low clinical sensitivity and are more commonly used in the research setting. Although PCR of synovial fluid is reasonably sensitive in patients who have not been treated, serologic testing remains the preferred diagnostic method for confirming Lyme disease in patients with arthritis.10 Some commercial laboratories claim to specialize in testing for tickborne diseases and offer their own laboratory-developed tests for Lyme disease. Tests offered include urine antigen assays, lymphocyte transformation tests, quantitative CD57 assays, and measurements of antibodies in synovial fluid.11–13 Such tests have not been sufficiently clinically validated. For example, one study by Klempner et al13 compared urine antigen testing to two-tiered serologic testing and found that it was unreliable and led to false-positive results. These tests are not cleared by the US Food and Drug Administration and not recommended for clinical use. Data suggest that some clinicians are unaware of proper testing procedures for Lyme disease. A 2011 nationwide survey of physicians and nurse practitioners found that in the previous year, 13.8% of respondents had ordered a Lyme disease test for a patient who was asymptomatic, which is not recommended, and 7.4% of respondents had incorrectly ordered a Lyme disease test as a test of cure (CDC, unpublished data). On a local level, anecdotal experiences from the University of Vermont Medical Center (UVMMC) laboratory indicate that some clinicians still find ordering and interpreting Lyme disease tests challenging and confusing. Inappropriate ordering or interpretation of test results can lead to misdiagnosis.14 Although national data indicate there is some confusion with testing, detailed and systematic information on laboratory testing knowledge and practices by clinicians is limited. This information would help tailor educational interventions and evaluate potential improvements to the laboratory ordering and reporting system. The objective of this study was to describe clinicians’ knowledge and practices regarding diagnostic testing for Lyme disease in Vermont. Materials and Methods This study was conducted at UVMMC, Vermont’s only academic medical center. Located in Burlington, the state’s largest city, it functions as a regional referral center for approximately 1 million people and a community hospital for 160,000 individuals. As a regional reference laboratory, UVMMC performs approximately 2.7 million tests annually. For Lyme disease testing, the UVMMC laboratory performs the recommended two-tiered serologic test in house. Two outlying hospitals perform their own Lyme EIA tests and, if positive or equivocal, send samples to UVMMC for follow-up WB testing. Otherwise, the WB as a standalone test is not offered by UVMMC, according to current guidelines. To order Lyme disease testing, clinicians order the “Lyme antibody” test. If the EIA is negative, the result is reported as “negative.” If the EIA is positive or equivocal, that result is reported as “Lyme antibody screen positive or equivocal. Western blot confirmation to follow.” For WB testing, the result is reported as positive or negative for IgM and IgG, as well as the specific bands identified. Comments are included in the report to aid in interpretation (see Discussion section for complete comments). UVMMC also offers CSF antibody testing; PCR on serum, CSF, and synovial fluid; and a serum tick-borne disease antibody panel as send-out tests through the Mayo Medical Laboratories. The serum tickborne disease antibody panel includes the two-tiered serologic testing for Lyme disease. We developed an online survey consisting of 13 questions regarding clinician characteristics and knowledge, attitudes, and practices related to diagnostic testing for Lyme disease (Supplementary Figure 1; all supplemental materials can be found at American Journal of Clinical Pathology online). A link to this anonymous survey was sent to 1,142 clinicians who had ordered any laboratory test from UVMMC within the past year. The survey collected information on respondent demographics, basic Lyme disease testing knowledge, alternative testing requests, and confusion with test interpretation. We also asked respondents to rate their satisfaction with the UVMMC laboratory’s Lyme disease test ordering system and the Lyme disease test result reporting system. A free-text option was available for some questions to allow respondents to elaborate on their responses. The UVMMC Laboratory Customer Service team collects basic information on phone calls received as part of standard practice. We collected additional data regarding phone calls received about Lyme disease testing prospectively during a 6-month period from May to October 2015. Customer service representatives filled out a standard data collection form each time they received a phone call regarding Lyme disease testing (Supplementary Figure 2). Responses were summarized using descriptive statistics. Inferential statistical calculations were performed with parametric methods using the SAS program, version 9.3 (SAS Institute, Cary, NC). Fisher exact test was used to compare the proportion of survey questions answered correctly among clinicians using the R program, version 3.3.1 (R Core Team, 2016).15 We compared the performance of those in primary care specialties, who are likely to be the first health care contacts of potential patients with Lyme disease, with those in other specialties; we excluded infectious disease specialists from this comparison since they likely have more detailed knowledge of testing protocols, and numbers were too few to perform separate statistical analyses. The study was reviewed and approved by the University of Vermont Institutional Review Board. Results Clinician Survey Of 1,142 survey links sent, 192 surveys were started and 147 were completed, yielding a 12.9% response rate. Of the 147 completed surveys, three were excluded because respondents were not physicians, nurse practitioners, physician assistants, or naturopathic providers. Thus, 144 total responses were included. Characteristics of survey respondents are listed in Table 1 . All respondents practiced within Vermont, primarily in the greater Burlington area. Table 1 Characteristics of Survey Respondents Characteristic   No. (%)  Total  144 (100)  Provider type   Attending physician  118 (81.9)   Physician assistant  11 (7.6)   Nurse practitioner  11 (7.6)   Physicians-in-training  2 (1.4)   Naturopathic provider  2 (1.4)  Specialty   Primary care    Family medicine  36 (25.0)    Internal medicine  27 (19.4)    Pediatrics  22 (15.3)   Subspecialists    Obstetrics and gynecology  8 (5.6)    Emergency medicine  7 (4.9)    Dermatology  6 (4.2)    Infectious diseases  4 (2.8)    Neurology  4 (2.8)    Orthopedics  4 (2.8)   Other  26 (18.1)  Practice setting   Outpatient  77 (53.5)   Inpatient  17 (11.8)   Combined  50 (34.7)  Characteristic   No. (%)  Total  144 (100)  Provider type   Attending physician  118 (81.9)   Physician assistant  11 (7.6)   Nurse practitioner  11 (7.6)   Physicians-in-training  2 (1.4)   Naturopathic provider  2 (1.4)  Specialty   Primary care    Family medicine  36 (25.0)    Internal medicine  27 (19.4)    Pediatrics  22 (15.3)   Subspecialists    Obstetrics and gynecology  8 (5.6)    Emergency medicine  7 (4.9)    Dermatology  6 (4.2)    Infectious diseases  4 (2.8)    Neurology  4 (2.8)    Orthopedics  4 (2.8)   Other  26 (18.1)  Practice setting   Outpatient  77 (53.5)   Inpatient  17 (11.8)   Combined  50 (34.7)  View Large Of three knowledge questions, the mean number of correct responses for all respondents was 2.1 with a standard deviation of 0.7 Table 2. Most clinicians obtained a score of 2 (73, 50.7%) or 3 (44, 30.6%). Using the Fisher exact test, we found no significant difference between the performance of clinicians working in primary care from those working in other noninfectious disease specialties (P = .84). Infectious disease specialists did have an overall higher score (2.8 correct answers for infectious disease vs 2.1 for primary care and 2.0 for others); however, statistical analyses were not performed given the low number of infectious disease specialists. Table 2 Clinician Responses to Survey Questions About Lyme Disease (LD) Serologya Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  1. A patient presents to your office with a rash resembling erythema migrans that began 3 days earlier. You order a serologic test for LD, which yields a negative result. This result rules out LD as a cause of the illness.  True  10 (6.9)  6 (7.1)  4 (7.3)  0 (0.0)  False  134 (93.1)  79 (92.9)  51 (92.7)  4 (100)  2. A 50-year-old female from southern Vermont presents with a swollen, erythematous knee for the past week. She does not remember a tick bite or rash, but she is active outdoors and went on a hiking trip two months ago. You suspect LD. What is the appropriate next step?  Treat with doxycycline (no testing needed)  13 (9.0)  8 (9.4)  5 (9.1)  0 (0.0)  Order WB  4 (2.8)  1 (1.2)  3 (5.5)  0 (0.0)  Order EIA followed by WB if positive  107 (74.3)  60 (70.6)  44 (80.0)  3 (75.0)  Aspirate joint, order PCR of the fluid  20 (13.9)  16 (18.8)  3 (5.5)  1 (25.0)  3. A 45-year-old woman from southern Vermont presents with fatigue and difficulty concentrating for the past two years. She does not remember a tick bite or rash, but occasionally gardens in her backyard. A LD test ordered by her PCP showed an equivocal EIA, positive IgM WB (2/3 bands), and negative IgG WB (1/10 bands). What is your interpretation of these results?  Patient has LD, treat with antibiotics  43 (29.9)  24 (28.2)  19 (34.5)  0 (0.0)  Not LD, IgM is false positive  61 (42.4)  39 (45.9)  18 (32.7)  4 (100)  The patient is in the “window period”  40 (27.8)  22 (25.9)  18 (32.7)  0 (0.0)  Mean (SD) number correct, range 0-3  2.10 (0.74)  2.09 (0.77)  2.05 (0.70)  2.75 (0.50)  Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  1. A patient presents to your office with a rash resembling erythema migrans that began 3 days earlier. You order a serologic test for LD, which yields a negative result. This result rules out LD as a cause of the illness.  True  10 (6.9)  6 (7.1)  4 (7.3)  0 (0.0)  False  134 (93.1)  79 (92.9)  51 (92.7)  4 (100)  2. A 50-year-old female from southern Vermont presents with a swollen, erythematous knee for the past week. She does not remember a tick bite or rash, but she is active outdoors and went on a hiking trip two months ago. You suspect LD. What is the appropriate next step?  Treat with doxycycline (no testing needed)  13 (9.0)  8 (9.4)  5 (9.1)  0 (0.0)  Order WB  4 (2.8)  1 (1.2)  3 (5.5)  0 (0.0)  Order EIA followed by WB if positive  107 (74.3)  60 (70.6)  44 (80.0)  3 (75.0)  Aspirate joint, order PCR of the fluid  20 (13.9)  16 (18.8)  3 (5.5)  1 (25.0)  3. A 45-year-old woman from southern Vermont presents with fatigue and difficulty concentrating for the past two years. She does not remember a tick bite or rash, but occasionally gardens in her backyard. A LD test ordered by her PCP showed an equivocal EIA, positive IgM WB (2/3 bands), and negative IgG WB (1/10 bands). What is your interpretation of these results?  Patient has LD, treat with antibiotics  43 (29.9)  24 (28.2)  19 (34.5)  0 (0.0)  Not LD, IgM is false positive  61 (42.4)  39 (45.9)  18 (32.7)  4 (100)  The patient is in the “window period”  40 (27.8)  22 (25.9)  18 (32.7)  0 (0.0)  Mean (SD) number correct, range 0-3  2.10 (0.74)  2.09 (0.77)  2.05 (0.70)  2.75 (0.50)  EIA, enzyme immunoassay; IgM, immunoglobulin M; PCP, primary care provider; PCR, polymerase chain reaction; SD, standard deviation; WB, Western blot. aCorrect answers are listed in bold text. View Large Most respondents correctly answered basic knowledge questions about Lyme disease testing (Table 2, questions 1 and 2). A clinical scenario regarding a woman with longstanding, nonspecific symptoms and an overall negative serologic test for Lyme disease, however, was answered correctly by only 45.9% of primary care providers and 32.7% of other noninfectious disease specialists (Table 2, question 3). Many respondents reported that patients had asked them for nonstandard serology (38.2%) or unvalidated tests from Lyme disease “specialty” laboratories (21.5%) (Table 3, questions 4 and 5). This was particularly true among primary care practitioners and infectious disease specialists. In addition to the nonstandard tests listed in Table 3, other free-text responses for tests requested included urine antigen testing and direct testing of ticks. Table 3 Clinician Responses to Survey Questions About Alternative Testing Requests Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  4. For diagnosis of Lyme disease, CDC recommends two-tier serologic testing consisting of an EIA test, which if positive or equivocal is followed by a reflex WB. Have you ever been pressured by a patient to order nonstandard serologic testing for Lyme disease (select all that apply)?  Yes—WB only without preceding EIA  28 (19.4)  17 (20.0)  9 (16.4)  2 (50.0)  Yes—serologic tests that use nonstandard interpretation criteria (eg, fewer bands for positive)  34 (23.6)  24 (28.2)  8 (14.5)  2 (50.0)  Yes—other  12 (8.3)  10 (11.8)  2 (3.6)  0 (0.0)  No  89 (61.8)  46 (54.1)  41 (74.5)  2 (50.0)  5. Have you ever been pressured by a patient to order unvalidated tests from Lyme “specialty labs,” including tests that are not FDA approved for diagnosis of Lyme disease (select all that apply)?  Yes—CD57 assays  5 (3.5)  1 (1.2)  4 (7.3)  0 (0.0)  Yes—lymphocyte transformation tests  2 (1.4)  1 (1.2)  2 (3.6)  0 (0.0)  Yes—novel culture techniques  4 (2.8)  3 (3.5)  1 (1.8)  0 (0.0)  Yes—other  23 (16.0)  18 (21.2)  3 (5.5)  2 (50.0)  No  113 (78.5)  63 (74.1)  48 (87.3)  2 (50.0)  Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  4. For diagnosis of Lyme disease, CDC recommends two-tier serologic testing consisting of an EIA test, which if positive or equivocal is followed by a reflex WB. Have you ever been pressured by a patient to order nonstandard serologic testing for Lyme disease (select all that apply)?  Yes—WB only without preceding EIA  28 (19.4)  17 (20.0)  9 (16.4)  2 (50.0)  Yes—serologic tests that use nonstandard interpretation criteria (eg, fewer bands for positive)  34 (23.6)  24 (28.2)  8 (14.5)  2 (50.0)  Yes—other  12 (8.3)  10 (11.8)  2 (3.6)  0 (0.0)  No  89 (61.8)  46 (54.1)  41 (74.5)  2 (50.0)  5. Have you ever been pressured by a patient to order unvalidated tests from Lyme “specialty labs,” including tests that are not FDA approved for diagnosis of Lyme disease (select all that apply)?  Yes—CD57 assays  5 (3.5)  1 (1.2)  4 (7.3)  0 (0.0)  Yes—lymphocyte transformation tests  2 (1.4)  1 (1.2)  2 (3.6)  0 (0.0)  Yes—novel culture techniques  4 (2.8)  3 (3.5)  1 (1.8)  0 (0.0)  Yes—other  23 (16.0)  18 (21.2)  3 (5.5)  2 (50.0)  No  113 (78.5)  63 (74.1)  48 (87.3)  2 (50.0)  CDC, Centers for Disease Control and Prevention; EIA, enzyme immunoassay; FDA, US Food and Drug Administration; WB, Western blot. View Large When asked specifically about WB interpretation, many respondents (44.4%) indicated that a positive IgM WB result in a patient with longstanding symptoms lasting more than 30 days had created confusion for either themselves or their patients ( Table 4 , question 6). Similarly, many (46.5%) responded that a laboratory report that listed individually positive WB bands in the context of an overall negative test had also caused confusion for either themselves or their patients (Table 4, question 7). Again, primary care practitioners and infectious disease specialists reported the majority of these problems. Table 4 Clinician Responses to Survey Questions About Western Blot (WB) Interpretation Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  6. According to CDC guidelines, due to potential cross-reactivity with other conditions IgM WB should only be ordered or interpreted for patients with ≤30 days of symptoms. Has the presence of a false-positive IgM WB—but overall negative test—in a patient with longstanding symptoms ever caused confusion for you or your patient?  Yes, more than once in the past year  9 (6.3)  4 (4.8)  2 (3.6)  3 (75.0)  Yes, once in the past year  14 (9.7)  9 (10.6)  5 (9.1)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  30 (35.3)  11 (20.0)  0 (0.0)  No  80 (55.6)  42 (49.4)  37 (67.3)  1 (25.0)  7. According to CDC guidelines, IgM WB is considered positive if ≥2 out of 3 scored bands are present. IgG WB is considered positive if ≥5 out of 10 scored bands are present. Has the presence of individual bands on a WB—but an overall negative test—ever caused confusion for you or your patient?  Yes, more than once in the past year  14 (9.7)  8 (9.4)  3 (5.5)  3 (75.0)  Yes, once in the past year  12 (8.3)  9 (10.6)  3 (5.5)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  29 (34.1)  12 (21.9)  0 (0.0)  No  77 (53.5)  39 (45.9)  37 (67.3)  1 (25.0)  Survey Question   All Respondents, No. (%)  Primary Care Providers, No. (%)  Subspecialists (Except Infectious Disease), No. (%)  Infectious Disease Specialists, No. (%)  6. According to CDC guidelines, due to potential cross-reactivity with other conditions IgM WB should only be ordered or interpreted for patients with ≤30 days of symptoms. Has the presence of a false-positive IgM WB—but overall negative test—in a patient with longstanding symptoms ever caused confusion for you or your patient?  Yes, more than once in the past year  9 (6.3)  4 (4.8)  2 (3.6)  3 (75.0)  Yes, once in the past year  14 (9.7)  9 (10.6)  5 (9.1)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  30 (35.3)  11 (20.0)  0 (0.0)  No  80 (55.6)  42 (49.4)  37 (67.3)  1 (25.0)  7. According to CDC guidelines, IgM WB is considered positive if ≥2 out of 3 scored bands are present. IgG WB is considered positive if ≥5 out of 10 scored bands are present. Has the presence of individual bands on a WB—but an overall negative test—ever caused confusion for you or your patient?  Yes, more than once in the past year  14 (9.7)  8 (9.4)  3 (5.5)  3 (75.0)  Yes, once in the past year  12 (8.3)  9 (10.6)  3 (5.5)  0 (0.0)  Yes, at some point during my practice but not in the past year  41 (28.5)  29 (34.1)  12 (21.9)  0 (0.0)  No  77 (53.5)  39 (45.9)  37 (67.3)  1 (25.0)  CDC, Centers for Disease Control and Prevention; IgG, immunoglobulin G; IgM, immunoglobulin M. View Large Among the 142 respondents who answered a question about satisfaction with the UVMMC Lyme disease test ordering system, most indicated they were somewhat or very satisfied (85, 59.9%) or neutral (49, 34.5%). Most also indicated that they were somewhat or very satisfied (89, 62.7%) or neutral (47, 33.1%) with regard to the Lyme disease test reporting system. Some free-text comments indicated lack of knowledge or confusion about appropriate Lyme disease testing. For example, some clinicians responded that they would like to be able to order WB outside of the two-tiered testing algorithm, which is not recommended. One clinician also indicated that patients asking for nonstandard tests either have seen or would eventually see clinicians who have deemed themselves to be Lyme specialists. Lyme Disease Testing A total of 20,295 serum EIA tests were performed during the 2.5-year study period. The median age of patients tested was 49 years for female patients and 52 years for male patients; most EIA test samples were from female patients (55.3%). The largest number of EIA tests was ordered for women aged 50 to 59 years (Supplementary Figure 3). A total of 1,488 serum EIA tests (7.3%) were positive, 18,684 (92.1%) were negative, 99 were equivocal (0.5%), and 24 (0.1%) were unsuitable for analysis. Serum EIA testing peaked during the summer months, as did positive results of EIA testing (Supplementary Figure 4). During the same period, a total of 1,687 WB tests were performed, including 103 that were positive by EIA at an outside facility and sent for WB testing at UVMMC. Of those, 538 (31.9%) were positive by IgM only, 193 (11.4%) were positive by IgG only, 239 (14.2%) were positive by both, and 716 (42.4%) were negative by both; one was uninterpretable. Positive IgM with negative IgG peaked in the early summer months, which may correspond with early localized disease, while positive IgG results with or without a positive IgM were more sporadic throughout the year (Supplementary Figure 5). Patients with a positive IgG WB had a bimodal age distribution; median age was 51 years. Males aged 19 years or younger and 50 to 69 years had the highest number of positive IgG WB tests. While more EIA test samples were performed on female patients, more positive IgG WB results were from male patients (Supplementary Figure 6). Nine synovial fluid samples were sent for PCR testing, and all were negative. Of the 516 CSF samples sent for antibody testing, 14 (2.7%) were positive by EIA (Supplementary Figure 7). Of these 14 positive EIAs, three CSF samples were negative by WB, seven were positive for IgG but not IgM, and four were positive for both IgG and IgM (Supplementary Figure 7). Customer Service Calls A total of 13 telephone calls related to Lyme disease were logged by laboratory customer service during the study period. Clinicians called with questions regarding testing algorithm (seven calls), result interpretation (two calls), adding on tests to patient specimens (two calls), as well as requests for testing outside the recommended CDC guidelines (two calls, request for WB only and request for WB when EIA was negative). Discussion This study demonstrated that most clinicians surveyed in Vermont have basic knowledge of Lyme disease diagnostic testing protocols. In our survey, however, most clinicians struggled with recognizing false-positive IgM results in a patient with longstanding symptoms and interpreting WB results with individually positive bands but an overall negative test. Nearly half of survey respondents also indicated that situations such as these caused confusion in their clinical practices on at least one occasion. These data suggest there is a need for additional clinician education regarding Lyme disease testing. Furthermore, there may be a role for laboratories to modify how they report results to improve accuracy of interpretation by clinicians. At the time of the survey, the UVMMC laboratory reported both IgM and IgG results (whether positive or negative and which bands, if any, were identified), regardless of the time interval of symptoms. If only IgG is positive, the interpretation states, “Indicative of B. burgdorferi infection at some time in the past.” If only IgM is positive, the interpretation states, “Indicative of early B. burgdorferi infection. A new serum specimen should be analyzed in 14-21 days to demonstrate seroconversion of IgG. IgM Western blot is of diagnostic utility only during the first four weeks after the onset of disease. Specimens collected more than 1 month following the onset of disease with positive IgM and negative IgG results more likely represent a false positive.” If both IgM and IgG are positive, the interpretation states, “Indicative of active or previous B. burgdorferi infection. IgM Western blot is of diagnostic utility only during the first four weeks after the onset of disease.” If both are negative, the interpretation states, “Specific serologic response to B. burgdorferi infection is not detected. This may indicate lack of infection, lack of seroconversion, or low/undetectable antibody levels. If clinically indicated, a new serum specimen should be submitted in 7-14 days.” On all results, a comment is included that explains the CDC two-tiered approach and band interpretation criteria. IgM has high rates of cross-reactivity and false-positive results beyond the initial 30-day symptomatic period, and reporting IgM results for patients with symptoms lasting more than 30 days may lead to some confusion, despite comments added to all results. It is possible that some clinicians only read the “positive” or “negative” results without the accompanying interpretation or that there continues to be confusion despite the included comments. To address this issue, the UVMMC laboratory is currently exploring how best to modify the Lyme disease two-tiered test-ordering mechanism such that they are prompted to state whether or not the patient has had symptoms for more than 30 days. If yes, then only the IgG result would be reported; if no, then both the IgM and IgG results would be reported. This would be consistent with the current CDC-recommended guidelines and could minimize the number of false-positive results reported for patients with symptoms more than 30 days. Positive IgG results were identified during the summer months, as expected, and the groups with the highest proportion of positive tests were similar to the highest risk groups in CDC surveillance data. These results underscore that these particular age and sex groups are at higher risk and would benefit from targeted prevention efforts. Heightened clinician suspicion could enhance early recognition and treatment. This study was subject to limitations. Results were based on respondents’ willingness to complete the survey and therefore were subject to responder bias. However, the respondent characteristics seemed to be representative of the community and encompass many of the challenges in laboratory test ordering and interpretation typically identified. In addition, only 13 phone calls related to Lyme disease were logged by laboratory customer service during the study period, which is lower than expected. This is likely because the log sheet for this study was an additional form that customer service personnel needed to have on hand and complete. In a busy customer service call center that receives approximately 8,000 phone calls per month, it is possible that some of the phone calls related to Lyme disease were not recorded. Therefore, it is likely that the actual number of phone calls regarding Lyme disease testing and interpretation is higher than captured in this study. In conclusion, clinicians would benefit from targeted education regarding Lyme disease testing to improve test-ordering practices and accuracy of test interpretation. Furthermore, clinical laboratories could improve their Lyme disease test ordering and reporting system to tailor reporting of test results based on the patient’s duration of symptoms (ie, removing IgM WB results from the laboratory report for patients with longstanding symptoms). Last, education of patients regarding unvalidated Lyme disease tests could reduce alternative test-seeking behavior and decrease the burden of requests to clinicians for these tests. Acknowledgment: We thank Mark Delorey for guidance on statistical analysis. Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the official position of the US Centers for Disease Control and Prevention. References 1. Stanek G, Wormser GP, Gray Jet al.   Lyme borreliosis. Lancet . 2012; 379: 461- 473. Google Scholar CrossRef Search ADS PubMed  2. Pritt BS, Respicio-Kingry LB, Sloan LMet al.   Borrelia mayonii sp. nov., a member of the Borrelia burgdorferi sensu lato complex, detected in patients and ticks in the upper midwestern United States. Int J Syst Evol Microbiol . 2016; 66: 4878- 4880. Google Scholar CrossRef Search ADS PubMed  3. Nelson CA, Saha S, Kugeler KJet al.   Incidence of clinician-diagnosed Lyme disease, United States, 2005-2010. Emerg Infect Dis . 2015; 21: 1625- 1631. Google Scholar CrossRef Search ADS PubMed  4. Vermont Department of Health. Epidemiology of Lyme disease in Vermont. 2013. http://healthvermont.gov/prevent/lyme/documents/epidemiology.pdf. Accessed June 2017. 5. Centers for Disease Control and Prevention. Recommendations for test performance and interpretation from the Second National Conference on Serologic Diagnosis of Lyme Disease. MMWR Morb Mortal Wkly Rep . 1995; 44: 590- 591. PubMed  6. Moore A, Nelson C, Molins Cet al.   Current guidelines, common clinical pitfalls, and future directions for laboratory diagnosis of Lyme disease, United States. Emerg Infect Dis . 2016; 22:1169-1177. 7. Halperin JJ, Baker P, Wormser GP. Common misconceptions about Lyme disease. Am J Med . 2013; 126: 264.e1- 264.e7. Google Scholar CrossRef Search ADS   8. Steere AC, McHugh G, Damle Net al.   Prospective study of serologic tests for Lyme disease. Clin Infect Dis . 2008; 47: 188- 195. Google Scholar CrossRef Search ADS PubMed  9. Kalish RA, McHugh G, Granquist Jet al.   Persistence of immunoglobulin M or immunoglobulin G antibody responses to Borrelia burgdorferi 10-20 years after active Lyme disease. Clin Infect Dis . 2001; 33: 780- 785. Google Scholar CrossRef Search ADS PubMed  10. Aguero-Rosenfeld ME, Wang G, Schwartz Iet al.   Diagnosis of Lyme borreliosis. Clin Microbiol Rev . 2005; 18: 484- 509. Google Scholar CrossRef Search ADS PubMed  11. Nelson C, Hojvat S, Johnson Bet al.  ; Centers for Disease Control and Prevention (CDC). Concerns regarding a new culture method for Borrelia burgdorferi not approved for the diagnosis of Lyme disease. MMWR Morb Mortal Wkly Rep . 2014; 63: 333. Google Scholar PubMed  12. Dattwyler RJ, Arnaboldi PM. Comparison of Lyme disease serologic assays and Lyme specialty laboratories. Clin Infect Dis . 2014; 59: 1711- 1713. Google Scholar CrossRef Search ADS PubMed  13. Klempner MS, Schmid CH, Hu Let al.   Intralaboratory reliability of serologic and urine testing for Lyme disease. Am J Med . 2001; 110: 217- 219. Google Scholar CrossRef Search ADS PubMed  14. Nelson C, Elmendorf S, Mead P. Neoplasms misdiagnosed as “chronic Lyme disease.” JAMA Intern Med . 2015; 175: 132- 133. Google Scholar CrossRef Search ADS PubMed  15. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2016. © American Society for Clinical Pathology, 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Journal

American Journal of Clinical PathologyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off