NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on Lyme Disease and Other Tick-Borne Diseases: The State of the Science. Critical Needs and Gaps in Understanding Prevention, Amelioration, and Resolution of Lyme and Other Tick-Borne Diseases: The Short-Term and Long-Term Outcomes: Workshop Report. Washington (DC): National Academies Press (US); 2011.

Cover of Critical Needs and Gaps in Understanding Prevention, Amelioration, and Resolution of Lyme and Other Tick-Borne Diseases

Critical Needs and Gaps in Understanding Prevention, Amelioration, and Resolution of Lyme and Other Tick-Borne Diseases: The Short-Term and Long-Term Outcomes: Workshop Report.

Show details

7Diagnostics and Diagnosis

Diagnostics and diagnosis, which are at the heart of the controversy surrounding tick-borne diseases (TBDs), have different connotations. Diagnostics provide a cluster of objective measures directed toward identifying the cause of a disease. After scientists discover the causative agent of an emerging infectious disease, such as Borrelia burgdorferi or Ehrlichia chaffeensis, they develop, evaluate, and refine diagnostic tests over time. Diagnosis, by contrast, rests on a patient’s history and symptoms and observed physical and laboratory findings. Ultimately, accurate diagnosis requires knowledge of the epidemiology, clinical manifestations, and diagnostic tests of a disease.

Lyme disease presents a significant challenge to this standard approach. The presentation of symptoms may not align directly with the diagnostic laboratory test results. Necessary and sufficient conditions for the diagnosis may not be met, and yet the constellation of findings might lead one to make a diagnosis. At the time of acute presentation to a health professional, serologies may not be definitive. Conversely, serology may be positive, but symptoms may not match the serological picture. This suggests opportunities to develop laboratory measures that are reliable, valid, and sensitive to change and that may help to define the phases/stages of Lyme disease, such as acute, post-acute, chronic, and recurrent.

In this chapter, three researchers explored the limitations of existing tests for Lyme borreliosis and other tick-borne diseases and suggested promising new approaches to diagnostics that can improve the diagnosis of those diseases, and four clinicians discussed the challenges and needs for improving diagnosis in the medical office.

DIAGNOSTICS FOR LYME DISEASE: KNOWLEDGE GAPS AND NEEDS

Maria Aguero-Rosenfeld, M.D., New York Medical College and Bellevue Hospital Center

Microbiologists share some of the concerns that patients have about the current diagnostic tests for Lyme borreliosis and other tick-borne diseases. The laboratory diagnostic challenges stem from the organism’s complex antigenic composition and its variation in expression depending on the environment where the organism is located. Borrelia burgdorferi has both linear and circular plasmids along with chromosomal DNA, and, in contrast to many other bacteria, a large portion of its genes are in plasmid DNA. The plasmid genes encode outer membrane components allowing the pathogen the flexibility of switching on and off antigens depending on the environment. This mechanism allows the pathogen to survive during the inactive tick stage (wintering) and to replicate during blood feeding on a suitable host.

Unlike other spirochetes, B. burgdorferi can be cultured in vitro. However, researchers are just beginning to understand the difference between immune responses to antigens expressed in vivo and antibodies detected using antigens from B. burgdorferi cultured in vitro. The Borrelia pathogen expresses some antigens as it first comes in contact with the host mammal leading to the early antibody response. Then, as more antigens are presented, the mammalian host develops the corresponding immune response in a sequential fashion. The intensity and type of antibodies developed depend on the duration of disease prior to antimicrobial treatment, the host immune system, and, perhaps, pathogenetic properties of the microorganism. Researchers have found that there are antigens expressed in vivo and others expressed in vitro. Therefore, the assays that are used to identify antibodies need to include those antigens expressed in vivo.

Two methods are available for directly detecting the presence of the pathogen in humans: culturing and polymerase chain reaction (PCR). Both of these have had mixed results in detecting B. burgdorferi (Table 7-1). Culturing spirochetes from a patient’s blood or synovial fluid has been difficult because the concentration of spirochetes is low. The key to this approach would be to optimize the culture methods that could allow scientists to detect these spirochetes efficiently. Microbiologists attain the best results from culture when using skin from a patient’s erythema migrans rash.

TABLE 7-1. Sensitivity of Direct Methods of Detection of Borrelia burgdorferi.

TABLE 7-1

Sensitivity of Direct Methods of Detection of Borrelia burgdorferi.

Modifications to the medium used to culture a patient’s blood have had mixed results. The sensitivity of the test depends on the volume of cultured blood and evidence of early disease dissemination, when the organism is most likely to be present in the bloodstream. However, only 40 percent of patients in the early stages of infection test positive for B. burgdorferi in their blood. European scientists can sometimes detect B. burgdorferi spirochetes in patients’ cerebrospinal fluid (CSF). The differences between being able to detect Borrelia in the blood or CSF may be a function of the Borrelia genotype. For example, the European Borrelia garinii is more neuroinvasive, which means that it is more often present in the cerebrospinal fluid, where it can be detected in culture.

The second technique for directly detecting Borrelia in patients is PCR, which amplifies specific sequences of spirochete DNA in samples of skin, synovial fluid, or blood. The efficiency of PCR depends on primers, number of sets used, sample type, and quality of the sample. In general, this detection method is more successful for B. burgdorferi when several different genetic sequences are amplified on DNA extracted from the skin of patients with erythema migrans and synovial fluid from the joints of patients with untreated Lyme disease arthritis. Obviously, PCR will have a higher yield on those samples with more spirochetes. Only a few U.S. scientists have used PCR to examine for spirochete DNA in cerebrospinal fluid where evidence of borreliae was seen in up to 60 percent of patients with early neuroborreliosis in one study. PCR is more sensitive in detecting infection in patients with untreated arthritis, as spirochete’s DNA is present in 78–96 percent of these patients. However, the high yield of PCR in CSF of patients with early neuroborreliosis and in synovial samples of patients with untreated arthritis was obtained when several PCR primer sets were used. Direct testing works well if applied to the best samples at the right time. Drawbacks are the unavailability of PCR in most clinical settings and the need to biopsy the skin or to perform a joint tap, which many primary care physicians do not do.

Because of the drawbacks of those two methods for directly detecting B. burgdorferi, most clinicians continue to rely on detecting antibodies to the pathogen when testing patients for Lyme disease. The first generation of assays in the 1980s did not use antigens that were effective in detecting antibodies in patients’ sera. These tests often failed to confirm that patients were infected with B. burgdorferi or gave positive results in patients who did not have Lyme disease.

In the mid-1990s, two-tier testing became the standard serological approach, and improvements were made on the antigen composition used in first- and second-tier tests, which produced better results. However, the two-tier approach is ineffective in detecting antibodies to B. burgdorferi during the acute phase of infection. Only 29 percent of patients later found to be infected with the pathogen have antibodies detected during the initial period. The sensitivity of the two-tier approach rises markedly during later stages of infection (Table 7-2).

TABLE 7-2. Performance of Two-Tier Testing.

TABLE 7-2

Performance of Two-Tier Testing.

The first step in the two-tier approach, the enzyme-linked immunosorbent assay (ELISA) test, has high sensitivity, but low specificity. That is, patients infected with B. burgdorferi are very likely to yield positive results on that test. However, ELISA also produces false positives, suggesting that some people are infected with the pathogen when they actually are not. For this reason, microbiologists apply the second step to confirm the results for those patients who test positive on the ELISA test. The Western immunoblot, by contrast, has high specificity: Most of the people who test positive for antibodies to B. burgdorferi are infected. The criteria for a positive result on this test are fairly stringent: Two of 3 specified bands on an IgM immunoblot or 5 of 10 specified bands on an IgG immunoblot must be detected for the specimens to be diagnostic. The IgM immunoblot should be used only within the first 4 weeks of illness, while the IgG immunoblot can be used at any stage in Lyme disease. If patients test positive on the ELISA but negative on the Western blot, they are considered not to have specific antibodies against B. burgdorferi.

One of the most immunodominant antigens in early disease is outer surface protein C (OspC). Antibodies to this antigen are among the first to appear after infection occurs. Another key result is evidence of antibodies to VlsE, which shows reactivity as early as 8 days after patients become ill. The addition of VlsE to both first- and second-tier tests has improved their performance. An increment in immunoreactive bands is observed in the IgG immunoblots of sera of patients with neuroborreliosis and Lyme disease arthritis.

One scientific gap is the testing of cerebrospinal fluid for antibodies. Europeans measure intrathecal production of antibodies by measuring antibodies in CSF and comparing these results against the concentration of antibodies in the serum to produce a ratio. U.S. scientists have not had a sufficiently large population in which to evaluate the efficacy of this approach because fewer cases of neuroborreliosis are documented in the United States as compared to Europe and CSF sampling is not routinely done in patients with Lyme disease. The absence of this type of testing is a gap in diagnostics for neuroborreliosis caused by B. burgdorferi in the United States.

Besides the lack of sensitivity in detecting early Borrelia infection, the two-tier test cannot distinguish between active Lyme disease and past infection or reinfection. Promising new tests to address this problem are on the horizon. European researchers have advanced the use of recombinant antigens. Furthermore, a combination of immunodominant antigens in a bead format could be used instead of whole-cell lysates. For example, scientists at the Centers for Disease Control and Prevention evaluated the use of VlsE and pepC-10 (a synthetic peptide derived from OspC) in a kinetic ELISA. More recently C6, a synthetic peptide based on a component of VlsE, has been approved as a source of antigen in first-tier enzyme immunoassay (EIA). Three studies (Bacon et al., 2003; Steere et al., 2008; Wormser et al., unpub.) compared the C6 testing protocol with the standard two-tier method. The C6 testing protocol has performed comparably in accurately detecting the presence of antibodies to B. burgdorferi in sera of patients with acute EM, but it was slightly less effective in the case of neurological Lyme disease. Overall, the specificity of C6 testing protocol is lower than that of the two-tier testing protocol. False positives remain a significant concern as U.S. laboratories now perform more than 2 million tests for Lyme disease annually, with at least 1 percent of these tests generating false positives. Thus, current test protocols produce approximately 20,000 false positives each year, a problem that may increase with wider adoption of the C6 testing protocol.

In conclusion, education is crucial to the diagnostics for both the clinician and the community. The positive predictive value of a test relies on the test being applied to the appropriate patient. Furthermore, physicians need to be educated on the availability of the tests and their limitations. Clinicians sometimes order tests on patients with a low probability of infection, making the results difficult to interpret. In the diagnostic laboratory, education and training are also important for laboratorians as individual interpretation of the test often results in over-reading of the Western blots, in particular IgM immunoblots.

Knowledge Gaps and Research Opportunities

Aguero-Rosenfeld identified a number of key areas for future work:

  • Development of programs to educate practitioners on the appropriate use of laboratory tests for Lyme disease.
  • Improvement of direct methods for detecting B. burgdorferi in samples from patients. For example, scientists could improve and automate culture techniques and use PCR to target several gene sequences.
  • Development of immunoassays containing a combination of recombinant or peptide antigens of importance, such as VlsE and OspC, in a bead format or other comparable method that would allow measuring the quantity of antibodies to individual antigens.
  • Development and evaluation of assays on cerebrospinal fluid that can support a diagnosis of neuroborreliosis.
  • Development of an algorithm that enables laboratory tests for B. burgdorferi infection to determine the stage of disease or duration of infection.
  • Establishment of a repository of well-characterized samples from Lyme borreliosis patients for use in evaluating new assays.
  • Recombinant antigens and peptides such as C6, including variations in gene sequences, require further evaluation for sensitivity and specificity.

IMPROVED DIAGNOSTICS AND NOVEL APPROACHES TO TICK-BORNE DISEASES

Juan P. Olano, M.D., University of Texas Medical Branch

Diagnosis of infectious diseases is based on the same techniques for nearly all infectious agents: antibody detection, antigen capture, and culture and detection of nucleic acids with or without amplification. Antibody detection (serological techniques) is the most common diagnostic method used in infectious diseases, but as reiterated throughout the workshop, a serologic diagnosis is frequently rendered too late to be of clinical value for therapeutic decisions because the immune response requires time to develop so that pathogen-specific circulating antibodies can be reliably detected. More recently, great progress has been made following the development of PCR and other nucleic acid amplification techniques for detection of pathogens in blood and other tissues. Most types of diagnostic tests used to diagnose Lyme disease are also used to diagnose other tick-borne infections, including rickettsial diseases, ehrlichioses, and anaplasmosis. Rickettsia disease differs pathologically from those of Ehrlichia and Anaplasma in that it infects the microvascular endothelium, and therefore levels of circulating rickettsiae are usually low, posing a challenge for diagnosing rickettsioses during the acute phase of the disease. The indirect immunofluorescence assay (IFA) is considered the gold standard for diagnosis of rickettsial infections due to its high sensitivity and specificity when paired serum samples obtained 2–3 weeks apart are tested. However, its sensitivity is very low when single serum samples obtained in the acute phase are used. As with the diagnostic tests for Borrelia burgdorferi, the presence of IgM antibodies can be detected 5 to 7 days after the onset of symptoms, but the specificity of these tests is low. They produce false positives as other bacterial, viral, and parasitic infections can cross-react with the antigens. Similarly, IgG antibodies have rising titers after 7–10 days, but there is cross-reactivity within the spotted fever and typhus group rickettsiae. Diagnostic titers continue to rise, and, by day 30, approximately 100 percent of the patients have detectable circulating antibodies. As a result of the delayed diagnosis, appropriate treatment may be delayed, and case-fatality ratios are higher in the absence of specific antibiotic therapy.

There are a number of other diagnostic tests for rickettsial diseases. Western blots allow for early detection of IgM antibodies to lipopolysaccharide antigens, but they still have low specificity, and cross-reactions occur between the spotted fever and typhus group rickettsiae. Detection of diagnostic IgG bands parallels the IFA detection rates. Cross-adsorption studies can be used to distinguish between the various species, but the tests are cumbersome and expensive. Dot blot enzyme immunoassay has similar sensitivity as IFA but also allows for the use of multiple antigens.

The rash associated with rickettsial disease is caused by infection of the endothelial cells lining the microvessels in the skin. When the rash is present in the acute phase of the disease, detection of rickettsiae by immunohistochemistry of skin biopsies has a sensitivity of ~60–80 percent.

Nucleic acid amplification techniques (mostly DNA), including PCR and real-time PCR, are used in selected research laboratories around the world for detection of circulating rickettsiae (inside macrophages or circulating endothelial cells that have detached from their microvascular niche). These tests are not commercially available, and their sensitivity and specificity have not been evaluated systematically. The agents of human monocytic ehrlichiosis (HME), Ehrlichia ewingii ehrlichiosis (EEE), and human granulocytic anaplasmosis (HGA) infect mononuclear phagocytes, including circulating monocytes (HME) and neutrophils (EEE and HGA). Therefore, detection of these pathogens in blood is theoretically more sensitive compared to rickettsioses. However, the diagnosis of these diseases presents similar difficulties as described for rickettsiae. Direct observation of the pathogens in peripheral blood smears is usually insensitive (more so for HME than HGA) due to the lower abundance of circulating target cells for HME. Diagnosis relies primarily on IFA, and its sensitivity and specificity is similar to the rickettsioses. Other serological tests include Western immunoblotting using native or recombinant antigens. Antibodies against tandem repeat protein (TRP) 120 and TRP 42 provide diagnostic bands to differentiate Ehrlichia chaffeensis from other Ehrlichia spp., while the 42, 44, and 49 kDa proteins help in distinguishing HGA from other ehrlichiae.

Similar to Rickettsia, there are a number of other diagnostic approaches for ehrlichioses and anaplasmosis. Immunohistology is available, but not widely applied outside the research setting. Conventional and real-time PCR assays have been developed and evaluated in small series of cases. Sensitivity varies widely, from 50 to approximately 100 percent, depending on several factors including primers used, time of testing during the course of the disease, use of pre-test antibiotics, etc. Specificity for PCR is very high provided there is no amplification contamination. All amplification techniques are available at selected research laboratories and are not commercially available. Currently, several new technologies provide platforms for improving the performance of conventional serological assays and are based on antigen capture using pathogen-specific antibodies in microfluidic settings followed by different detection techniques, including electrochemiluminescence and microretroreflectors.

Electrochemiluminescence is a highly sensitive technique that uses ruthenylated antibodies that in the presence of tripropylamine and an electrical current release photons. For rickettsial pathogens, its analytical sensitivity in vitro and in vivo using animal models is ~1,250 to 1,500 organisms/mL. Microretroreflector detection is based on reflection of light off a gold-coated surface. As rickettsiae are captured by antibody-coated magnetic beads or nanogold particles, their deposition on a reflective surface dims the amount of light bouncing off that surface. This variations are detected using conventional optics. Other antigen-capturing systems being evaluated with microfluidic systems include the use of microporous substances to increase the capturing surface area.

Improved serological assays using protein microarrays is another promising technology. In short, all open reading frames of the R. rickettsii genome have been cloned, expressed, and blotted onto microarray spots. The serological response is then analyzed and response patterns are delineated. This technique could improve both sensitivity and specificity for diagnosis of rickettsioses, ehrlichioses, and anaplasmosis. Promising preliminary results have been observed with rabbit and dog sera and a few human samples available for testing.

In conclusion, Olano noted that despite advances in antigen-capturing systems, detection technologies, automation, and nucleic acid amplification techniques, commercially available tests for the diagnosis of human rickettsioses, anaplasmosis, and ehrlichioses by nucleic acid amplification have not been developed. Aside from IFA, all these tests are available only at selected research laboratories around the United States and the world. Thus, the gap between the laboratory bench and the patient bedside in diagnosing tick-borne diseases remains wide. As better detection platforms are refined and become increasingly available, microfluidic technology, automation, nanotechnology, and point-of-care testing will result in accurate, fast, and inexpensive diagnosis of these diseases.

POTENTIAL BIOMARKER APPLICATIONS FOR LYME DISEASE: ALIGNING MULTIPLE SYMPTOMS WITH BIOLOGICAL MEASURES

Afton L. Hassett, Psy.D., University of Michigan Medical School

Biomarkers are cellular, biochemical, and molecular characteristics by which normal and abnormal processes can be recognized or monitored. Their identification has numerous clinical applications, such as improving diagnostic accuracy; assessing disease activity, prognosis, and efficacy of treatment; and tailoring treatment to the individual. However, studies that fail to properly characterize the patients studied (i.e., phenotyping) hamper progress. For example, a recent review of research on biomarkers for autoimmune diseases found that some investigators had failed to control for patients’ age or gender, while others had not controlled for medication use, other medical and psychological comorbidities, or the stage of disease (Tektonidou and Ward, 2010). These factors, and several others, must be accounted for as each can influence biological measures.

Currently, there are no credible biomarkers for post-Lyme disease despite ongoing efforts in this area. For example, in a series of studies, researchers using the CD57 marker for natural killer cells found that patients with post-Lyme disease had fewer natural killer cells, suggesting that the marker could be used to assess treatment outcomes (Stricker and Winger, 2001, 2003; Stricker et al., 2002). However, a more recent study using a combination of CD56 and CD16 surface markers in conjunction with CD3 markers showed that natural killer cell counts did not differ between post-Lyme disease patients and healthy controls or patients who had recovered from Lyme disease after treatment (Marques et al., 2009).

In another study, heightened anti-neural antibody reactivity was found in 49.4 percent of post-Lyme disease patients compared with control groups who had recovered from Lyme disease (18.5 percent) or were healthy (15 percent) (Chandra et al., 2010). The heightened reactivity was not greater compared to patients with systemic lupus erythematosus (73.3 percent). A similar reactivity was observed in an earlier study evaluating ongoing neurologic Lyme disease (Sigal and Williams, 1997). This antibody reactivity supports the hypothesis that a sustained immune response may contribute to persistent neurologic dysfunction in Lyme disease patients, even after the pathogens are eliminated (Sigal and Williams, 1997). Other ongoing and/or promising research includes

  • Immune system abnormalities (e.g., persistent activation, cytokine-induced sickness behavior);
  • Proteomics—proteins produced specific to post-Lyme disease;
  • Neuroendocrine dysfunction;
  • Neuroimaging (microglial activation, neural network differences); and
  • Genetic predisposition.

Charting a scientific inquiry into the nature of post-Lyme disease may rely on understanding the symptoms of the condition, which include arthralgias, musculoskeletal pain, radicular pain, paresthesia, fatigue, neurocognitive impairment, and mood disturbances. These symptoms are the very same symptoms observed in conditions such as fibromyalgia and chronic fatigue syndrome, which are currently thought to result from disturbances in the central nervous system processing of sensory information. These conditions are frequently referred to as “central sensitivity” syndromes. Thus, to understand the persistent symptoms of post-treatment Lyme disease, there is a need to understand the commonalities between these symptoms and those associated with central sensitivity syndromes. Moreover, aspects of the underlying pathophysiology of central sensitivity syndromes will likely inform biomarker research.

We begin with the most prominent symptom—pain. Findings from the past decade of neuroscience research suggest there are at least three types of pain. The first, peripheral pain, includes acute injury, osteoarthritis, rheumatoid arthritis, and cancer pain, and is “nociceptive,” meaning that a stimulus in the periphery (e.g., inflammation, mechanical malfunction, or tissue damage) is causing the pain. This type of pain responds to interventional procedures, non-steroidal anti-inflammatory drugs, and opioids. Second, neuropathic pain stems from damage to or dysfunction of peripheral nerves and can include diabetic neuropathic pain and post-herpetic neuralgia. Neuropathic pain responds to both peripheral and central interventions. Third, “central,” or non-nociceptive, pain results from disturbances in central nervous system processing and leads to diffuse hyperalgesia (increased response to painful stimuli) and allodynia (painful response to normal stimuli). Examples of central pain conditions include fibromyalgia, interstitial cystitis, and irritable bowel syndrome. Individuals can have a combination of these types of pain. For example, about 15 percent of patients with rheumatoid arthritis also have fibromyalgia: that is, they have both inflammatory pain and central pain. Even patients with well-controlled rheumatoid arthritis, as evidenced by a lack of inflammation, may continue to have persistent pain. In these cases, it appears that augmentation of central nervous system pain processing accounts for the persistent experience of pain.

Another common symptom in central sensitivity syndromes is fatigue, which is thought to have a peripheral or central origin. It has been proposed that peripheral fatigue is predominantly due to physical exhaustion and may be attributed to organ-system dysfunction (Silverman et al., 2010). This type of fatigue occurs commonly in patients with rheumatoid arthritis, cardio-respiratory diseases, and myasthenia gravis. By contrast, central fatigue is more cognitive in nature and is attributed to central nervous system dysfunction. Classic examples of central fatigue include chronic fatigue syndrome, fibromyalgia, and irritable bowel syndrome, although central fatigue can also accompany rheumatoid arthritis, lupus, and cancer. Importantly, the difficulty with memory and concentration reported by a myriad of patients with various systemic diseases may be a function of central fatigue.

As with many other medical conditions, a stress-diathesis model for the etiology of central sensitivity syndromes is widely accepted. Such a model purports that genetic and environmental factors likely contribute to central sensitivity syndromes in equal measure. A series of case-controlled studies suggest that in predisposed individuals these syndromes can be triggered by peripheral pain conditions (e.g., rheumatoid arthritis and lupus [Clauw and Katz, 1995], physical trauma [Buskila et al., 1997; McBeth, 2005; Miranda et al., 2010], or catastrophic events, such as war [Clauw et al., 2003]) and infections. Pertinent to Lyme disease, infections in general have been shown to trigger central sensitivity syndromes in approximately 10 percent of patients. More specifically, 5 to 30 percent of patients with enteric infections later manifest irritable bowel syndrome (Bayless and Harris, 1990; Saito et al., 2002; Thabane and Marshall, 2009). Similarly, urinary tract infections appear to later lead to interstitial cystitis and painful bladder syndrome (Warren et al., 2008). In the Dubbo population-based prospective cohort study of patients infected with three very different viruses—Epstein-Barr virus, Ross River virus, and Coxiella burnetii, the bacterium that causes Q fever—approximately 9 percent of infected patients continued to present with a chronic fatigue–like syndrome even after the agent was cleared (Nickie et al., 2006). More recently, there was a case report of a central sensitivity syndrome triggered by H1N1 influenza (Vallings, 2010).

Similarly, various TBDs appear to trigger some central sensitivity syndromes. For example, approximately 39 percent of patients with human anaplasmosis developed chronic fatigue syndrome despite no serological evidence of persistent infection (Ramsey et al., 2002). In two prospective studies of patients with Lyme disease, 13 percent of patients eventually met the diagnostic criteria for fibromyalgia (Hassett et al., 2009), and 32 percent of patients eventually developed a fibromyalgia-like central sensitivity syndrome (not all met the criteria for fibromyalgia) (Hassett et al., 2010). Findings from studies conducted in Europe vary broadly, with anywhere from 2 to 48 percent of Lyme disease patients reporting symptoms after treatment consistent with central sensitivity syndromes (Cerar et al., 2010; Ljostaf and Mygland, 2010). The discrepancies in the European studies are likely due to differences in how the symptoms were measured. Biomarker research in central sensitivity syndromes could be highly pertinent to patients with persistent symptoms after treatment for Lyme disease. A recent study showed a decrease in natural killer cell cytotoxicity, and three different measures of CD26—an antigen located on cellular surfaces associated with immune regulation—accurately discriminated chronic fatigue syndrome patients from controls (Fletcher et al., 2010). Immune system biomarker research in irritable bowel syndrome has had mixed results, but there is some support for a dysfunctional mucosal immune response in these patients (Barbara and Stanghellini, 2009). For fibromyalgia, immune system research has suggested that IL-1β, IL-6, IL-8, and TNF-α contribute to central pain (Abbadie et al., 2003; Wang et al., 2008). The challenge, as noted earlier, is that many of the studies failed to control for variables such as obesity, autonomic nervous system dysfunction, and depression that are commonly observed in these populations and that have prominent effects on cytokine expression.

Studies involving twins have helped to reveal potential genetic biomarkers for central sensitivity syndromes. For example, one Swedish study (Kato et al., 2008) relied on a twin registry to investigate four somatic disorders (chronic widespread pain, chronic fatigue, irritable bowel syndrome, and recurrent headache) and two psychiatric disorders (major depressive disorder and generalized anxiety disorder). Multivariate twin analyses found a common genetic pathway for all six illnesses, but there were two distinct latent traits. One latent trait loaded heavily on the psychiatric disorders, while the other loaded on the somatic illnesses and not the psychiatric disorders. The somatic disorders were also affected by specific gene influences that were unique to each disorder. In family studies, first degree biological relatives of people with fibromyalgia had an eightfold higher risk of having fibromyalgia than did people without the familial link. There was also familial aggregation with mood disorders, but this was less pronounced (odds ratio of 2) (Arnold et al., 2004). Many of the genes thought to be involved in pain processing are also associated with mood disorders. This could help explain why pain and depression commonly co-occur. Psychiatric comorbidity is also common in patients with post-Lyme disease syndrome. For example, in a study using gold-standard assessment interviews, 45 percent of patients with post-Lyme disease syndrome were found to meet the criteria for major depressive disorder (Hassett et al., 2009). Moreover, that percentage was higher than that found in patients with fibromyalgia and those with medically unexplained symptoms. Causality could not be determined in that study because of its cross-sectional design, thus it is not clear if the depression predisposed patients to post-Lyme disease syndrome, or if post-Lyme disease patients are depressed because they have a difficult illness that has disrupted their lives.

In a prospective study of newly diagnosed Lyme disease patients, baseline depression, anxiety, and other medical and psychological factors were measured and the patients were followed over time. After antibiotic treatment, approximately 32 percent of the patients developed chronic symptoms ascribed to Lyme disease. Chronic symptoms were predicted by the severity of the Lyme disease symptoms at baseline (Hassett et al., 2010). Furthermore, positive affect was perhaps the best predictor of who would develop chronic symptoms, such that patients with high levels of positive affect when first diagnosed were less likely to have persistent symptoms. Level of positive affect at baseline was unrelated to the severity of their symptoms at baseline. That finding reinforces numerous studies (Horan and Dellinger, 1974; Adames et al., 1986; Bruel et al., 1993; Zautra et al., 2005; Connelly et al., 2007) linking positive affect to positive outcomes from surgeries and lower sensitivity to pain. The results from this prospective study suggest that the high rates of depression among patients with post-Lyme disease are more likely to be due to living with a chronic condition than to a predisposing factor for symptom chronicity. However, patients with chronic symptoms after treatment for Lyme disease may have an underlying genetic or immunologic vulnerability that predisposes to both chronic pain and other symptoms including depression and anxiety.

In research looking for fibromyalgia biomarkers, numerous investigators have explored stress-response system functioning. Dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis is frequently observed in central sensitivity syndromes. The nature of the reported abnormalities varies, but hyporesponsiveness of the HPA axis is found in 20 to 25 percent of the patients (Heim et al., 1998; Dedert et al., 2004; Van Den Eede et al., 2007; Wingenfeld et al., 2008). Further, in repeated-measure studies, a flat awakening cortisol level (McBeth et al., 2005; McLean et al., 2005; Weissbecker et al., 2006) and flat diurnal variation (Crofford et al., 2004; Dedert et al., 2004; Weissbecker et al., 2006) were observed in people with central sensitivity syndromes. In healthy individuals, cortisol peaks about 45 minutes after awakening as daily activities are initiated and then declines throughout the day. However, people with post-traumatic stress disorder and central sensitivity syndromes tend to have a low flat cortisol level, as if the HPA axis has lost its resiliency.

Although this pattern suggests an interesting biomarker for these patients, numerous cortisol studies have not accounted for depression, anxiety, childhood sexual or physical abuse, religiosity, positive affect, or other factors that can have a large effect on patients’ cortisol levels. For example, a Swedish study (Tjernberg et al., 2010) found a higher cortisol response on an adrenocorticotropic hormone stimulation test among post-Lyme disease patients. The investigators, however, did not assess subjects for depression, even though it has been associated with such a response, so the significance of the finding is unclear. These results underscore the need to control for these potential confounding factors with thorough patient phenotyping.

In addition to the HPA axis, the autonomic nervous system is an important part of the stress-response system. For individuals with fibromyalgia and other central sensitivity syndromes, another common pattern has emerged. Patients with central sensitivity syndromes tend to have high baseline sympathetic arousal, decreased parasympathetic activity, and an attenuated response to stressors (Adeyemi et al., 1999; Cain et al., 2007; Gockel et al., 2008).

In evaluating these findings on stress response among people with central sensitivity syndromes, it is important to note that even when studies reveal mean differences between patients and controls, there is substantial overlap between the two groups. The fact that cortisol levels are correlated with momentary pain in fibromyalgia patients also suggests that the pain may be causing the autonomic nervous system dysfunction, rather than vice versa. Finally, people with baseline hypo- or hyperactivity of these stress response systems may be more likely to develop central sensitivity syndromes after exposure to stressors, including even routine events such as cessation of exercise or restrictions on sleep.

Recent studies of potential neural biomarkers for central sensitivity syndromes have shown that a number of factors influence an individual’s sensitivity to pain. Neurotransmitters such as Substance P, nerve growth factor, and glutamate and other excitatory amino acids facilitate pain, while neural transmitters like norepinephrine, serotonin, dopamine, opioids, GABA (gamma-aminobutyric acid), adenosine, and cannabanoids inhibit pain. High levels of facilitatory neural transmitters and low levels of inhibitatory transmitters are linked to hyperalgesia, or higher sensitivity to pain. For fibromyalgia patients, there are elevations of neurotransmitters that facilitate pain and low levels of those that inhibit pain. Only the opioid levels appear to be appropriate in these patients, which may explain why opioids are not effective for relieving fibromyalgia pain.

Finally, neuroimaging has contributed significantly to our understanding of pain in fibromyalgia. In a 2002 study using functional magnetic resonance imaging (MRI), fibromyalgia patients showed activation of 12 cortical areas following pressure applied to the thumb bed. Controls need approximately twice the level of pressure applied to elicit the same pain response as demonstrated by a similar activation pattern as seen in the fibromyalgia group (Gracely et al., 2002). Finally, proton magnetic resonance spectroscopy imaging studies suggest that patients with fibromyalgia tend to have increased levels of glutamate in the insula and that these levels are associated with sensitivity to pain.

In summary, research suggests that

  • Although depression is common among people with post-Lyme disease, it does not explain the symptoms nor predict which patients with active Borrelia burgdorferi infection will later develop post-Lyme disease. These individuals could have a genetic vulnerability to central sensitivity syndromes, which might offer a productive approach to better identify those at risk.
  • Good biomarkers for post-Lyme disease are not yet available. However, conceptualizing the disease as having roots in central nervous system dysfunction could help chart the way toward identifying such biomarkers.
  • Promising areas to explore for biomarkers for post-Lyme disease include immune abnormalities, proteomics, genetics, neurotransmitter levels, stress-response system functioning, and neuroimaging.

Knowledge Gaps and Research Opportunities

Hassett noted that a state-of-the-art biorepository is essential to finding biomarkers for post-Lyme disease. To meet this research goal, the following features need to be considered:

  • A repository should include a wide range of samples, including serum, plasma, cerebrospinal fluid, organ tissue, heart rate variability, imaging studies, and genetic information.
  • The patients whose samples would be included in the biorepository must be carefully phenotyped. That detailed information should be collected and made available in regard to clinical characteristics of their Lyme disease, other medical comorbidities, psychiatric comorbidities, psychological factors, symptom profile and severity, and functional status.
  • The biorepository would have high-quality maintenance and provide open access to all researchers.
  • A conference or workshop convened specifically to hear all viewpoints regarding the attributes of such a biorepository would help make it a reality.

DISCUSSION

One participant noted that recent studies (Gomes-Solecki et al., 2007; Sillanpaa et al., 2007) suggest that the C6 testing paradigm is not as sensitive as it needs to be and was highly dependent on the particular Borrelia strain used during development. The participant questioned whether the test would have a higher sensitivity if it used the same B. burgdorferi strains that are in the patient’s geographic region. Aguero-Rosenfeld noted that the C6 assay that is commercially available in the United States as a Food and Drug Administration (FDA) approved test uses B. burgdorferi American genotype. While other researchers have looked into incorporating different sequences, the U.S. commercial test uses B. burgdorferi sensu stricto. What was not discussed in detail in the three studies (Bacon et al., 2003; Steere et al., 2008; Wormser et al., unpublished) described earlier is that two of the studies used in-house developed C6 assays. More recent, unpublished results suggest that the sensitivity of C6 is better than reported by Bacon et al. Aguero-Rosenfeld further noted the need for including other components, such as OspC and VlsE, to potentially improve these assays.

Another clinician noted that for research purposes, specificity of the testing is the goal, but many clinicians are more interested in sensitivity. They do not want to overtreat, but as noted during the workshop, complications can result when patients are not properly diagnosed. Aguero-Rosenfeld noted that the diagnostic test needs to combine the right antigen with less cross-reactivity. That would result in a test that has high sensitivity and high specificity. She further noted that the testing available today does not have the specificity needed and it can be difficult to find a balance between sensitivity and specificity.

Another participant questioned why IgM is only useful for diagnosis in early disease. Aguero-Rosenfeld noted that in very well-characterized Lyme disease patients, IgM positivity remained up to a year or more after a patient was well. She noted that IgM antibody-validated diagnostic criteria were restricted to acute Lyme disease in the first month of illness. Furthermore, IgM immunoblot reading and interpretation is prone to yield false-positive results when weak reactivities are scored and reported.

PANEL: CHALLENGES FOR CLINICIANS IN DIAGNOSIS AND MANAGEMENT OF CHRONIC ILLNESS MANIFESTATIONS: KNOWLEDGE GAPS

Sam T. Donta, M.D., Professor of Medicine (ret.), Infectious Diseases, Falmouth Hospital, MA

The clinical challenges involving Lyme disease are multifaceted, and research is needed to support the challenges in clinical diagnosis, diagnostic tools, and treatment. As most individuals will agree, the diagnosis of acute Lyme disease is relatively straightforward. The challenges arise in the diagnosis of “chronic Lyme disease” or “post-Lyme disease.” The criteria for chronic disease include fatigue, musculoskeletal symptoms, and neurocognitive impairments involving memory, concentration, and mood. Minor criteria include an array of nonspecific symptoms such as headaches, eye and/or ear symptoms, jaw/tooth pain, Bell’s palsy, disequilibrium, dyspnea, and others.

One difficulty that clinicians face is the absence of objective, measurable evidence for these symptoms. Unless a patient has an observable sign such as Bell’s palsy, a swollen joint, or a rash (EM), the clinician cannot easily attribute the patient’s symptoms to Lyme disease. A second difficulty arises in distinguishing the array of symptoms from those associated with other multisymptom illnesses such as chronic fatigue syndrome (CFS), fibromyalgia, or Gulf War syndrome. Clinicians are hard-pressed to say whether a patient’s complex of symptoms is caused by Lyme disease or some other etiology. For example, MRI findings in patients considered to have chronic Lyme disease can show signals that are confused or overlap with multiple sclerosis. Single-photon emission computerized tomography scans also can be positive. At the same time, patients press clinicians to provide an explanation, a diagnosis, for their symptoms.

Research is needed to define what is happening in terms of the pathogenesis of Lyme disease. It is important to understand “what about the organism is doing what,” as well as to understand how the host is responding. Greater focus needs to be placed on the former.

The bacteria are very difficult or impossible to find in the patient, even during the acute phase of the disease. Improvement in direct detection techniques may permit documentation of the bacteria in the future, but until that occurs, it is difficult to say definitively that a patient has Lyme disease or, conversely, that a patient no longer has Lyme disease following antibiotic therapy.

Another fertile area of research surrounds the mechanisms involved in Lyme disease (i.e., the ways in which the bacteria trigger the associated symptoms). One possibility is direct toxicity, in which a toxin produced by the bacteria perturbs the local nerve involved or the central nervous system, as is the case with tetanus and botulism. A toxin or other antigen might trigger the production of antibodies and an autoimmune reaction in the patient, although no evidence indicates this is the case in Lyme disease. Tetanus toxin, botulinum toxin, and other toxins bind to gangliosides in the nervous system, causing direct interference with nerve transmission, thus providing a precedent for other toxins to act in an analogous fashion. The fact that certain antibiotics can eliminate the symptoms patients experience supports the hypothesis that the continuing presence of the organism is responsible for the symptoms.

A third difficulty in diagnosing chronic Lyme disease is the lack of criteria for ELISA and Western blot results in those patients. The criteria that exist were developed for “late Lyme disease” and may not apply to patients farther removed from the initial infection. Our study of patients with chronic Lyme disease illustrates the difficulty in the diagnostic criteria as one third of the patients show neither a positive ELISA nor a positive Western blot. The number of positive Western blots increases when IgM is included. In addition, the IgM immunoblot tends to disappear following successful treatment with tetracycline (Donta, 1997), although not for all patients.

Considering the presence of IgM antibodies to be indicative of early disease, but to be considered a false positive in later disease, is illogical; perhaps IgM does persist as a sign of continuing, unresolved infection. A patient with IgM reactivity to the 23 kDa OspC protein and symptoms indicative of chronic Lyme disease definitely has been exposed to Borrelia, because there is no other identified cause of any cross-reactivity. The number of bands should not be what provides the laboratory support for the clinical diagnosis, but rather it is the specificity of the reaction that is important. The numbers were established for surveillance criteria, which are much more restrictive, and then translated into clinical criteria. Regardless, the current laboratory antibody-based tests are adjunctive to the clinical picture.

Neurologists routinely persist in the assertion that the cerebrospinal fluid of patients with neuro-Lyme disease will be positive for the bacteria. This misconception needs to change. The amount of IgG synthesis in the spinal fluid of those patients rarely is more than that found in the serum. All of these tests are fraught with difficulties.

In summary, symptom-based treatments are important, but they do not address the underlying pathology. A patient’s response to antibiotic treatment can be important as a diagnostic tool. More research involving clinical trials is needed for diagnostic tools, including direct antigen detection, and for vaccine development, as well as additional clinical antibiotic trials.

Brian A. Fallon, M.D., M.P.H., Columbia University Medical Center and the New York State Psychiatric Institute

Helping patients with chronic persistent symptoms requires understanding who the patient is, the chief complaint, the history of this person’s disease, and the history of treatment. Clinicians’ ability to help such patients depends on their experience with other similar patients and on what is known from the literature about the particular patient population. Before reviewing what is known about this patient group from the literature, it is important to emphasize that researchers and clinicians often have very different goals. Researchers aim to answer a specific question in a very tightly controlled setting, while clinicians aim to relieve suffering for an individual person and to help that person return to an active life. The clinical setting is filled with numerous confounding variables, and diagnostic and treatment decisions are often based on different levels of probability rather than on certainty. Treatment decisions also often reflect a cost–benefit analysis.

Turning to the literature pertaining to patients with chronic persistent symptoms, Fallon noted a number of areas need additional research. A European study compared patients with neurologic Lyme disease to those with erythema migrans, 3 years later, and found that 50 percent of those with neuroborreliosis experienced persistent symptoms versus 16 percent of the EM patients (Vrethem et al., 2002). These results suggest that rather than focusing solely on early EM, follow-up studies on chronic symptoms, should focus on the subpopulation of patients who present with neurologic or other disseminated symptoms.

Children are an at-risk population—another subpopulation that needs further research. A 2003 study found that 43 children with a history of cranial nerve palsy experienced more neck pain, behavioral changes, arthralgias, nerve sensations, and memory problems compared to age-matched controls (Vazquez et al., 2003). The study found no difference in functional impairments between the two groups, which raises the question of whether there is a difference between the effects on children and adults.

Finally, a study of Lyme disease encephalopathy at Columbia showed that most of the cognitive impairments the patients experienced were mild to moderate, primarily affecting verbal memory, working memory, and verbal fluency. Interestingly, the patients experienced only mild psychopathology; while some depression and anxiety were reported, these were not prominent (Fallon et al., 2008). Many patients (72 percent of Lyme disease patients versus 22 percent of controls), however, demonstrated sensory loss on their neurologic exams. On the rheumatologic exam, many patients had multiple-joint involvement compared to very few joints involved in the control group. Few of the Lyme disease patients exhibited multiple trigger points, suggesting that fibromyalgia was not a significant problem, at least if it is defined based on trigger points.

Future studies of patients with chronic symptoms should focus on the most prominent problems they report, which are pain, fatigue, and physical disability. When examining the patient’s treatment history, 70 percent of the patients screened for participation in the Lyme disease encephalopathy study had had at least 2 months of prior intravenous antibiotic therapy, and many of them had had significant oral treatment (Fallon et al., 2008). These facts suggest that many people continue to experience persistent symptoms despite having received a significant amount of antibiotic therapy.

Despite a large number of patients screened with persistent symptoms, few of them met the strict criteria for inclusion in the Columbia study. Only 1 percent had objective cognitive impairments, a positive IgG Western blot, prior intravenous therapy, and documentation of prior Lyme disease. We learned that a requirement of IgG Western blot positivity at time of enrollment for studies of patients with chronic symptoms following Lyme disease will exclude a large number of patients with good clinical histories, thus hampering enrollment and narrowing the generalizability of the findings.

Ten percent of patients in the Lyme disease encephalopathy sample had a history of co-infections. There was no difference in the incidence of human granulocytic anaplasmosis between the Lyme disease patients and the controls, but Babesia IgG was positive in 27 percent of the Lyme disease patients versus none of the controls. There was a high rate of Bartonella IgG positivity in the Lyme disease patients, but also in the controls, suggesting that exposure to Bartonella is common in the population.

Cerebrospinal fluid findings across four studies of patients with persistent symptoms following Lyme borreliosis showed elevated protein in 25.8 percent, 7.3 percent, and 12.1 percent of the patients, respectively (Klemper et al., 2001; Krupp et al., 2003; Fallon et al., 2008).

Brain imaging studies at the National Institutes of Health (NIH) reported that 55 percent of the patients with post-treatment Lyme disease exhibited hyperintensities on MRI (Morgen et al., 2001). Another study comparing the Lyme disease encephalopathy patients with controls well-matched for age, sex, and education showed no difference in the white matter hyperintensity density (DelaPaz et al., 2005). The study did find that the patients who had had Lyme disease were more likely to have blood flow deficits in their brains, as well as metabolic differences on positron emission tomography imaging, compared with the well-matched controls (Fallon et al., 2009). Prior work found the blood flow deficits to be reversible with intravenous ceftriaxone therapy (Logigian et al., 1997).

With respect to pathophysiology, Borrelia act directly and can invade neural cells in vitro (Livengood and Gilmore, 2006); there are also indirect actions, such as the induction of local cytotoxins or inflammatory mediators (reviewed in Fallon et al., 2010). European studies show that pro-inflammatory cytokines are increased, and chemokines, excitotoxin, and quinolinic acid are increased in patients with neuroborreliosis (Weller et al., 1991; Halperin and Heyes, 1992; Widhe et al., 2004; Rupprecht et al., 2005).

A rich field of research is psychoneuroimmunology, in which studies have suggested that individuals with risk factors for inflammatory disorders and for psychiatric disorders are more likely than those without them to experience chronic peripheral inflammation and chronic activation of the brain cytokine pathways following infections, leading to subjective health complaints similar to those in Lyme disease and a number of other disorders (Dantzer et al., 2008). Anti-neural antibody reactivity is increased in patients with persistent symptoms following treatment for Lyme disease (Alaedini and Fallon, 2010; Chandra et al., 2010). Approximately 50 percent of patients enrolled in two independent studies (Klemper et al., 2001; Fallon et al., 2008) of patients with chronic, persistent symptoms had elevated levels of anti-neural antibodies, indicating there is an abnormally activated immunological process at work in some of these patients with chronic symptoms.

A summary of the possible explanations for chronic, persistent symptoms in patients following treatment for Lyme disease includes: persistent infection in some patients; reinfection from a later tick bite; reactivation of a latent, dormant infection; widely distributed effects from a small amount of physiologically active but attenuated spirochetes; or post-infectious phenomena, such as spirochete-triggered immune abnormalities, neurotransmitter/receptor changes, or damage from prior infection. The symptoms also could be related to an unrecognized concurrent process, such as another TBD, another non-tick-borne infection, or another disease (e.g., depression or hypothyroidism).

There is a danger that persistent symptoms following treatment for Lyme disease will be labeled as somatoform. In part this may be due to clinicians’ assumptions that 2 to 4 weeks of antibiotic therapy is always curative and that any symptoms after minimal antibiotic treatment are due to other causes. It may also reflect a clinician’s failure to recognize that any infection may have a course of post-infectious symptoms that can continue for a year or more. In addition, some clinicians may experience hostility or frustration toward patients with chronic illnesses or may misinterpret the patient’s presentation with anxiety and multi-systemic, nonobjective symptoms as indicative of a psychiatric etiology.

Richard F. Jacobs, M.D., University of Arkansas for Medical Sciences and Arkansas Children’s Hospital Research Institute

One of the challenges that clinicians face in the diagnosis and management of tick-borne diseases (TBDs) in children is a poor understanding of the true incidence and geographic distribution of the diseases. The more information is known about the different variations of these organisms, the greater the realization will be that what has been taught about geographic distribution is not true. Another challenge is the similarity in the multisystem presentation among the TBDs. In addition, diagnostics are limited in acute illness, and the rates of chronic illness and morbidity are unknown. There is information about neuroborreliosis in adults, but aside from a few studies in children, there are no data on any of the TBDs that are sufficiently reliable to tell parents what the potential chronic neurologic or other sequelae may be.

Children are different; they are not little adults. They are still developing, and they have a very different central nervous system from adults, as well as a developing immune system. For this reason, it is important to recognize and study children as a distinct population. Furthermore, it is also important to keep in mind that the duration of any long-term effects of disease in this population will last 50 to 70 years.

Jacobs noted there is a need to provide enhanced educational information to clinicians—not only pediatricians, but also family physicians, advanced practice nurses, and physicians’ assistants—about the clinical manifestations of and other information regarding tick-borne illnesses. Conventional wisdom about spotted fever rickettsiosis (Rocky Mountain spotted fever) indicates that young children have a lower mortality rate, but a much higher infection rate, than do older adults (Dalton et al., 1995), although this generalization is not completely accurate. In addition, the fatality rate increases dramatically among cases in which treatment was not started until after the fifth day following the onset of symptoms (Dalton et al., 1995). Therefore, it is important to make treatment decisions presumptively and empirically based upon a patient’s clinical presentation.

Experience with Ehrlichia chaffeensis led to the recognition of a set of symptoms associated with ehrlichial infection. Fever and rash are common, but a host of other signs and symptoms occur as well: myalgia, headache, vomiting, diarrhea, and puffy eyes (Schutze and Jacobs, 1997). The physical presentation of ehrlichiosis has a large differential diagnosis, including a significant overlap with Rocky Mountain spotted fever (Buckingham et al., 2007), but the clinical laboratory triad of thrombocytopenia, leukopenia with lymphopenia, and elevated hepatic enzymes suggests human monocytic ehrlichiosis and warrants doxycycline therapy at admission (Schutze and Jacobs, 1997), given the importance of prompt treatment in reducing mortality.

Compounding the challenges of diagnosis and prompt treatment, different pediatric diseases within the differential for tick-borne illnesses, such as Kawasaki syndrome and meningococcemia, require very different therapies, but on a similarly urgent time line. The clinical challenge is that the diagnostics and clinicians’ understanding about the ecosystem and the organism do not allow them to separate these clinically.

Consider a child with a rash on her arms, legs, face, hands, and feet, along with fever, headache, and pleocytosis in her CSF. Rocky Mountain spotted fever and ehrlichiosis certainly are on the list of differential diagnoses. But the child had disseminated meningococcemia, which carries a 20 to 40 percent mortality rate had it not been treated with a third-generation cephalosporin on admission. This example highlights both the difficulty and the urgency of separating these diseases clinically.

The clinical challenge is compounded by the short incubation period of the TBDs. Antibody testing is not useful in the acute management of Rocky Mountain spotted fever or human monocytic ehrlichiosis. Both infections, as well as others, respond to treatment with doxycycline, however, so physicians are taught to administer doxycycline.

In terms of knowledge gaps, more data and understanding are needed on the genetics, predisposing factors, and epidemiology associated with tick-borne illnesses. In addition, more information is needed about the organisms, acute and persistent infections, diagnostics, spectrum of disease, and chronic manifestations and outcomes. In addition to the dearth of information in these areas for adults, there is virtually no knowledge of most of these infectious diseases as they relate to the unique attributes of children.

Lessons have been learned about the different impacts of acute disease from babies exposed to herpes simplex virus, different disease manifestations from children with tuberculosis, age-related immune responses to vaccinations, unknown central nervous system effects (autism-spectrum disorders), age-related exposures and adaptive immunity (Kawasaki’s disease), and central nervous system growth and development (use of folic acid to prevent neural tube defects). But much of the biology remains a mystery.

Jacobs noted the United States needs a study group to explore tick-associated and tick-borne infections in children. Two models currently exist. The NIH/National Institutes of Allergy and Infectious Diseases Collaborative Anti-Viral Study Group has operated for 30 years. Thirty-two sites now study rare diseases. The study of TBDs can follow the same model. The National Children’s Study, now NIH funded, involves randomly selected, geographically dispersed counties in the United States. that will follow pregnant women and their babies until the children are 20 years old. The study provides an opportunity to collect information in a repository, including biological samples and detailed historical information, and is charged to look at priority exposures and examples. Infectious agents are already listed as one of the study areas. The National Children’s Study provides a wonderful opportunity to use a currently NIH-funded, 20-year-long prospective study to focus some attention on TBD.

Matthew H. Liang, M.D., M.P.H., Harvard Medical School and Harvard School of Public Health

In unselected, non-specialty (primary care) practices, atypical manifestations of common illnesses are much more common than typical manifestations of uncommon illnesses. Primary care physicians build their practices on their patients’ trust in and access to them, both of which are necessary for fine-tuning a diagnosis or refining a treatment. The usual strategy is to identify treatable illnesses, make a working presumptive diagnosis, treat, and assess the outcome. If the treatment is not working, the physician may get more information, refine the diagnosis, and/or change the treatment. The process can be threatened, or lulled into complacency, by the primary care physician’s familiarity with the patient and the underlying probability that the illness is benign rather than serious. These factors sometimes make it difficult to keep an open mind about a patient with a persistent problem.

The presentation of Lyme disease in clinical practice is variable. Thirty-nine percent of patients ultimately considered to have Lyme disease do not meet the Centers for Disease Control and Prevention (CDC) criteria, and approximately 40 percent had negative Lyme disease serology and an acute viral-like illness without objective findings. Nearly one-third of the patients had a rash that did not meet the criteria for EM, and only 19 percent of those with EM exhibited the stereotypical bull’s-eye appearance (Aucott et al., 2009).

Given the variability, it is helpful when patients can provide the actual tick or a good description of a fed tick, as well as a time line to indicate how long the tick had been on them. A tick has to feed for 48 to 72 hours to transfer Borrelia to the host. The finding of a tick that is not well fed decreases the probability that it infected the patient. Clinicians also look for EM or forme-frustes, although EM can look like almost anything and is often mistaken for spider bites.

Generally clinicians would treat empirically because of the importance of early treatment and the assumption that there is little to lose because treatment can always be stopped. This presupposes the treatment has minimal negative effects. As mentioned previously, doxycycline is a common default treatment for suspected tick-borne illness; however, there are some downsides, including dental staining.

A prime area for intervention is educating people about prevention. Successful education requires understanding and reversing the thought barriers that prevent people from receiving and acting upon the message. There is, as already noted, a 48- to 72-hour time window within which to find a tick before it can infect the host, and a nightly shower provides ample opportunity to interrupt the life cycle. In theory, improved education will increase prevention and decrease the occurrence of Lyme disease.

A 5-year study randomized 29,000 people traveling by ferry to Nantucket (Daltroy and Phillips, 2007). The study group was exposed to an entertainment-based information session about Lyme disease and steps to prevent it. Participants also received a card with a Braille dot on it the size of a tick and a plastic shower card similar to those used for breast self-exam, tweezers, and a map of the island where ticks were prevalent. The study showed a reduction in Lyme disease in the group who received the message, both among year-round residents and among visitors, who constitute a high-risk population.

A population-based retrospective cohort study of 38 patients in a particular location in Ipswich, Massachusetts, who had been treated for Lyme disease showed 13 patients with ongoing symptoms of arthritis or recurrent arthralgias, neurocognitive impairment, neuropathy, or myelopathy (Shadick et al., 1994). The individuals with these sequelae tended to have higher IgG antibody levels to the spirochete and also to have received treatment later following infection. One of the 13, a 76-year-old woman, had been worked up by Lyme disease experts, had received two courses of ceftriaxone, and was negative for objective central nervous system findings. She died and at post mortem a Dieterle silver stain demonstrated two spirochetes, one in the cortex and another external to a leptomingeal brain vessel (Shadick et al., 1994).

Another study of approximately 6,000 year-round residents of Nantucket showed patients who had been diagnosed with Lyme disease and continued to be symptomatic following treatment, but the study presented few objective findings. This reinforces the challenges that clinicians face in the diagnosis and treatment of patients with chronic persistent symptoms following Lyme borreliosis.

In treating such patients, it is important to ensure that a thorough history and physical exam are conducted and that they have received sufficiently long courses of the appropriate doses and types of antibiotics. Beyond that, at a certain point the diagnosis matters less than treating the symptoms in an effort to maintain and improve function. In addition, it sometimes is necessary to assist patients in revising their expectations as well.

DISCUSSION

Much of the discussion focused on the challenges associated with the diagnosis of TBDs, both in patients with acute illness and in those experiencing persistent symptoms following an initial diagnosis of and treatment for Lyme disease. A second focus of discussion centered on the occurrence of Lyme borreliosis and other TBDs in children.

One participant questioned the reliability of screening tests, such as white blood cell count and standard neurological exam, to evaluate patients with chronic symptoms. Liang felt that these tests are not particularly useful and noted that there is tremendous variation among practitioners in terms of their approach to patients with persistent symptoms.

With regard to diagnosing TBDs in children, Jacobs noted that practitioners have come to rely on a clinical presentation of multisystem disease, in which the clinician must determine which systems are involved and create a differential diagnosis that is treatable. He reiterated the approach discussed by Krause, in which clinicians learn to look at particular presentations on screening tests (e.g., anemia, with thrombocytopenia) in acute infection to help make a diagnosis. The presence of multisystem disease and a specific picture of laboratory results together can generate a presumptive diagnosis and the initiation of treatment. Jacobs stressed, however, the need to develop better diagnostic testing that would permit more definitive diagnosis of TBDs.

Another participant questioned the practice of those clinicians who use the absence of direct markers of an infecting organism following treatment as evidence of the treatment’s success given the absence of, or the current inability to identify and test for, markers for the presence of infecting organisms prior to treatment as well.

With respect to cognitive dysfunction, one participant asked about the ability of brain SPECT scanning to distinguish hypoperfusion brain damage and cognitive dysfunction caused by TBDs from that caused by long-term excessive use of medication. Fallon agreed that certain medications can confuse the interpretation of SPECT scans. Cocaine use, although not a medication, can cause heterogenous hypoperfusion consistent with vasculitis, which appears similar to that seen in Lyme disease patients who have had SPECT scans. He also stated that despite the power of such imaging tools for research purposes, the use of SPECT scanning as a clinical tool is of questionable reliability because rarely are systematic methods used to evaluate the scans against healthy populations, making variation in readings across clinicians likely and interpretations of clinical significance difficult.

A question was posed about the implications of multiple phenotypes or strains of Borrelia and other organisms for the development of new, improved diagnostics. It has to do with moving forward into the new diagnostics. Fallon reiterated that 17 different isolates of Borrelia in the United States have been sequenced and their antigens are now known. He mentioned the possibility of using that information to study a wide sample of patients and perhaps trying to correlate some of their clinical profiles, clinical histories, and/or treatment outcomes with these actual antigenic profiles. Doing so would require a very large study of many patients followed up with good bioinformatics over a long period of time, but the ability is there to do it.

With respect to the experience and impact of Lyme disease and chronic persistent symptoms in children, Jacobs emphasized how difficult it is to have to tell concerned parents that there is simply no solid information about the long-term effects and impact of the disease on the child.

A participant observed that it seems as if the numbers of children experiencing symptoms and being diagnosed with illnesses such as fibromyalgia or chronic fatigue syndrome have risen since the last generation and asked whether schools could be surveyed to obtain information on the school-age population and the kind of symptoms and difficulties they are experiencing. Jacobs discussed the current National Children’s Study as an example of a study looking at the complex interactions among the environment, infectious agents, and genetics. He expressed the need to tap into the new area of bioinformatics, which can provide detailed information on participants, and combine that information with access to environmental samples from the National Children’s Survey, as well as human samples in a biorepository. Such a data repository would provide a very rich source of information to probe once there are better diagnostics, better biomarkers, or a better understanding of which imaging system or type of testing to do.

A question was raised regarding concerns about privacy and confidentiality, among citizens and schools, as well as publicized knowledge of being located in a tick-endemic area. Jacobs acknowledged the possibility of such concerns, but that generally when members of a community are or become vested in a project, the schools and other community organizations follow. Another participant indicated that the interest level in participating in such a trial would be high in that community.

Related to questions about the long-term impact of Lyme disease and other TBDs on children as they develop and mature, a request for greater consideration of gender differences and issues specific to women in these diseases was made by one participant. She specifically mentioned the impact of hormonal fluctuations (e.g., during adolescence and puberty, pregnancy, and menopause) on symptoms. Questions were also raised about congenital transmission of the diseases as well as their impact on fertility. Donta noted that there are changes in the severity of symptoms experienced by women not only with Lyme disease but also with various chronic conditions, such as chronic fatigue syndrome, as hormone levels fluctuate. In addition there is a gender difference in Lyme disease, perhaps related to the presence of estrogen and progesterone receptors in glial and neural cells. Jacobs reiterated the need for a large-scale, long-term study, such as the National Children’s Study, involving bioinformatics, although he acknowledged that the results would not be available in time to help inform the parents of children currently experiencing symptoms. Nevertheless, such a study would provide hope that the present large knowledge gaps will be filled in the future.

CONCLUDING THOUGHTS ON DIAGNOSIS AND DIAGNOSTICS

Lynn Gerber, M.D., Center for the Study of Chronic Illness and Disability, George Mason University

David H. Walker, M.D., Department of Pathology, University of Texas Medical Branch at Galveston

Many participants in this session noted that diagnosis of tick-borne diseases remains problematic. This could be ameliorated using a three-pronged approach: (1) education of clinicians about which diagnostic tests to use, when to use them during the course of the disease, and how to interpret these results; (2) developing and applying new technology for serological assays to close the gap between bench and bedside using microfluidic technology, automation, and nanotechnology to achieve accurate, fast, and inexpensive diagnostic tests; and (3) consensus building to establish criteria for clinical phases of disease in children and adults, possibly describing necessary and sufficient criteria for arriving at common nomenclature, such as in systemic lupus erythematosus, chronic fatigue syndrome, and fibromyalgia, among others.

Clinicians need better education regarding the limitations of existing tests, and how to interpret the results. Some 10–20 percent of healthy people in some regions may already carry antibodies to a particular organism, such as Rickettsia rickettsii or Ehrlichia chaffeensis. The antibodies might stem from exposure to a related organism that caused a subclinical infection. In that situation, a clinician who does not realize that a patient with acute febrile illness has had antibodies for a long period might wrongly diagnose rickettsiosis or ehrlichiosis on the basis of a single acute serologic test. In fact, clinicians often fail to do a follow-up serologic test to determine whether the concentration of antibodies to a tick-borne disease in a patient is rising.

Clinicians need to understand that testing patients with a low likelihood of a tick-borne disease strongly undermines the test’s positive predictive value. When clinicians do test such patients, a substantial proportion will be false positives. Lengthy menus of tests also present a barrier to effective diagnosis of tick-borne illness by clinicians who are not familiar with the advantages and disadvantages of many laboratory assays.

Some clinicians also use serologic assays for IgM antibodies that have not been validated through well-documented series of cases. The soaring incidence of reported spotted fever rickettsial infections—few of which have been confirmed by methods specific to R. rickettsii—is one result.

Some presenters in this session emphasized that patients with tick-borne diseases do not develop antibodies to an infectious organism until some time after the onset of illness, because of the nature of the immune response. That means existing tests that may be highly reliable later in the disease are insensitive early on. It also means that diagnosing a tick-borne infection requires knowledge of a patient’s geographic and seasonal exposure to ticks as well as the clinical manifestations of tick-borne diseases.

Despite these diagnostic shortcomings, tick-borne diseases such as human ehrlichioses and anaplasmosis likely have undiagnosed incidence equal to that of Lyme borreliosis. And they and Rocky Mountain spotted fever carry the threat of a fatal outcome, which Lyme disease does not.

New methods that can determine which species of Rickettsia or Babesia a person has encountered are particularly important. Some participants noted that investigators also need to develop tests that can shed light on the etiology of southern tick associated rash illness (STARI), whose erythema migrans resembles that of Lyme disease but is not caused by B. burgdorferi.

Creating biorepositories and a network of clinical studies, such as those now supported by the National Institute of Child Health and Human Development, would greatly enhance the opportunity to improve diagnostics by providing wider access to stored specimens as well as clinical information. Sera from documented cases of tick-borne diseases—during both acute and convalescent phases—would enable scientists to validate new serologic tests. Such repositiories would aid in enabling investigators correlate symptoms and biological findings and help develop evaluative and treatment outcome criteria. It might help determine whether children, for example, have a different course of illness, given that their central nervous and immune systems are still developing.

Samples of whole blood and cerebrospinal fluid, tissue biopsies, and other specimens would allow scientists to validate the use of PCR for amplifying nucleic acids and to identify and validate novel methods of detecting tick-borne pathogens. The recently sequenced genomes of different strains of B. burgdorferi also promise to allow scientists to develop new diagnostics and ultimately, preventive measures. The community affected by Lyme borreliosis and other tick-borne diseases seeks guidance on prognosis and treatment, and that has not yet been achieved.

Throughout the workshop, during podium presentations and comments and questions from the floor, participants employed descriptive terminology pertaining to Lyme disease in different ways. This presented several challenges to discussants in that it was not always clear that the topic under discussion was addressing acute, chronic, recurrent phases of illness or other co-infections. Better descriptors would provide a uniform vocabulary for clinicians, researchers, and patients. They would also provide a basis for building and validating a comprehensive, sensitive battery of tools for evaluating both objective and patient-reported outcomes for tick-borne diseases. Improved descriptors should include both signs and symptoms: that is, information that is both objective and self-reported, and that includes physical findings, serological measures, and psychological measures, among others.

Further exploration of the stress response to tick-borne pathogens could help expand our understanding of the pathogenesis and natural history of Lyme disease. One approach is to investigate the role of the hypothalamic–pituitary–adrenal axis and the cortisol response in people with chronic Lyme disease symptoms that resemble those of other chronic fatigue or pain syndromes. Another would be to determine whether there is a genetic vulnerability to central sensitivity syndromes, which might offer a productive approach to better identify those at risk. Good biomarkers for post-Lyme disease are not yet available. However, conceptualizing the disease as having roots in central nervous system dysfunction could help chart the way.

Copyright © 2011, National Academy of Sciences.
Bookshelf ID: NBK57026
PubReader format: click here to try

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (15M)
  • Disable Glossary Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...