|Home | About | Journals | Submit | Contact Us | Français|
There is no evidence of current or previous B. burgdorferi infection in most patients evaluated at University-based Lyme disease referral centers. Instead, psychological factors likely exacerbate the persistent diffuse symptoms or “Chronic Multisymptom Illness” incorrectly ascribed to an ongoing chronic infection with B. burgdorferi. The objective of this study was to assess the medical and psychiatric status of such patients and compare these findings to those from patients without CMI.
240 consecutive patients undergoing medical evaluation at an academic Lyme disease referral center in New Jersey were screened for clinical disorders (e.g. depression and anxiety) with diagnoses confirmed by structured clinical interviews. Personality disorders, catastrophizing, and negative and positive affect were also evaluated and all factors were compared between groups and to functional outcomes.
60.4% of our sample had symptoms that could not be explained by current Lyme disease or another medical condition other than CMI. After adjusting for age and gender, clinical disorders were more common in CMI than in the comparison group (p<.001, OR 3.54, 95% CI, 1.97 to 6.55), but personality disorders were not significantly more common. CMI patients had higher negative affect, lower positive affect and a greater tendency to catastrophize pain (p<.001) than did the comparison group. Except for personality disorders, all psychological factors were related to worse functioning. Our explanatory model based on these factors was confirmed.
Psychiatric comorbidity and other psychological factors are prominent in the presentation and outcome of some patients who inaccurately ascribe long-standing symptoms to “chronic Lyme disease.”
Lyme disease is a multi-system inflammatory disorder due to a symptomatic infection with the tick-borne organism Borrelia burgdorferi ( B. burgdorferi)1. Clinical features, including erythema migrans rash, musculoskeletal pain and/or joint inflammation, typically resolve with conventional antibiotic treatment2. Rarely do patients treated with adequate antibiotic therapy continue to manifest objective evidence of ongoing infection. Yet, in studies of post-therapy complaints, 4–40% of adult patients reported chronic physical, cognitive, and/or mood symptoms attributed to Lyme disease3–8.
Post Lyme Disease Syndrome refers to symptoms that continue 6 months after initial diagnosis and treatment2. Despite evidence from animal studies that viable B. burgdorferi can persist after antibiotic therapy9, there is no evidence of ongoing B. burgdorferi infection in humans10–11. Nonetheless, these patients are diagnosed as having “chronic Lyme disease” - assumed to be a persistent, perhaps incurable, infection12. Extended courses of antibiotic therapy are often prescribed for their “infection”2,13–16, even though double-blind, placebo controlled studies have shown that such therapies provide no benefit17. Frequently, symptoms are best explained by fibromyalgia4,14,16,18–19 and respond to treatment for this non-infectious disorder.
A second group of patients accounts for up to three quarters of patients seen in University-based Lyme disease referral centers14,19–20. These patients report multiple unexplained symptoms ascribed to “chronic Lyme disease,” but do not have an illness or objective evidence explicitly suggesting Lyme disease. Some patients are self-diagnosed, while others are misdiagnosed by physicians using non-validated laboratory tests, applying alternative criteria in the interpretation of validated laboratory tests, or eschewing laboratory tests entirely alleging fallibility and inaccuracy14,18,21. Many of these patients also inappropriately receive long-term and/or repeated antibiotic therapy2,14–16,19. Similar to Post Lyme Disease Syndrome patients, these patients experience numerous nonspecific complaints, e.g. joint and muscle pain, fatigue, headache, and cognitive impairment, that are often complicated by depression16,19, psychological stress19, and the presence of other stress-related syndromes, especially fibromyalgia14,16,19–20. Conditions including, but not limited to, fibromyalgia, chronic fatigue syndrome, and Gulf War syndrome may be considered collectively as Chronic Multisymptom Illness22. The case definition for Chronic Multisymptom Illness has included having at least one of more chronic symptoms from at least two of three categories of symptoms including musculoskeletal, fatigue and mood-cognition22.
Previous findings and our own clinical observations suggest that the complaints of some patients presenting to Lyme disease referral centers might have explanations rooted in psychological distress16,19,23. Not only does mood disturbance appear to be common amongst these patients19, but one study found a relationship between prior psychological trauma and/or treatment with psychotropic medication and poor outcome in Lyme disease patients24. However, psychological factors have not been adequately studied in the full range of patients presenting to Lyme disease centers. Of importance are both broad categories of psychiatric conditions: clinical disorders (e.g. depression, anxiety, somatization) and personality disorders (i.e., inflexible, maladaptive and enduring personality traits causing distress or functional impairment)25. In addition to psychiatric disorders, maladaptive emotional and cognitive/coping factors associated with poor outcomes should be evaluated.
The objective of this cross-sectional study was to evaluate a large patient cohort representative of those seen in our Lyme disease referral center over the last 20 years. We assessed clinical and personality disorders, as well as a potential role for negative and positive affect (emotional state) and catastrophizing (maladaptive coping). Psychological factors were then compared between groups (those with and those without Chronic Multisymptom Illness) and to functional outcomes. An explanatory model for group inclusion consisting of clinical disorders, personality disorders, negative and positive affect and catastrophizing was tested.
We recruited 240 patients from the Lyme Disease Center at the University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School (UMDNJ-RWJMS). Participants were predominantly residents of New Jersey, New York, Connecticut and Pennsylvania - areas endemic for Lyme disease. Most were referred by physicians in their communities, while others were self-referred. The Institutional Review Board of UMDNJ-RWJMS approved this study.
All English-speaking patients ages 18–70 presenting to the Lyme Disease Center for an initial visit were invited to participate. Enrollment and tracking took place from September 2002 through March 2007. Patients arrived one hour before their appointment and completed consent and questionnaires in a private room. Structured clinical interviews were conducted when indicated by screening. Fewer than 17% (49 patients) declined to participate or did not complete at least 75% of the questionnaires. Participants were paid $10.
LHS conducted a standard medical evaluation for Lyme disease including a diagnostic interview, physical examination, and chart/records review. Special reference was given to the results and location of prior laboratory testing. Two-tiered serological testing (enzyme-linked immunosorbent assay [ELISA] and Western Blot) was ordered when current, but previously undiagnosed, Lyme disease was suspected. For patients diagnosed as having untreated Lyme disease antibiotic therapy was offered. These patients were reexamined at regular intervals and contacted by telephone six months post-treatment to evaluate the persistence of Lyme disease-related symptoms. Explanations other than Lyme disease were sought for all patients’ complaints, e.g. osteoarthritis, systemic lupus erythematosus, fibromyalgia, Parkinson disease, amyotrophic lateral sclerosis, anti-phospholipid antibody syndrome (all diagnoses made in the past in patients referred for evaluation of Lyme disease)14 and appropriate treatments and referrals were recommended. When diagnosis was unclear, patients were tracked and reviewed by LHS upon receiving test or referral results. Blind to questionnaire findings LHS assigned patients to one of four groups described below.
These patients met Centers for Disease Control (CDC) criteria for Lyme disease26 (Table 1); in most cases the two-tiered protocol for laboratory tests provided seroconfirmation. They received adequate prior antibiotic treatment defined as meeting or exceeding guidelines from the Infectious Diseases Society of America27, but continued to report persistent symptoms ascribed to Lyme disease. At evaluation there was: 1) no evidence of inflammatory disease that could be ascribed to active B. burgdorferi infection and 2) no serologic evidence of infection with B. burgdorferi that had not been treated previously with adequate antibiotic therapy.
These patients did not meet CDC criteria for Lyme disease at the onset of their symptoms26; nonetheless, symptoms were attributed to Lyme disease. These patients met the American College of Rheumatology criteria for fibromyalgia28 and no other medical condition.
These patients did not meet CDC criteria for Lyme disease at the onset of their symptoms26 and their multiple symptoms could not be accounted for by a medical condition. There was little or no evidence of fibromyalgia at evaluation.
Patients without Chronic Multisymptom Illness included those diagnosed by LHS as having previously untreated Lyme disease based on CDC criteria26. These patients were offered antibiotic therapy and contacted six months post-treatment for a telephone interview. Patients denying symptoms related to Lyme disease were considered recovered and assigned to the comparison group. Others without Chronic Multisymptom Illness had well-defined medical conditions, e.g., osteoarthritis, or were deemed “healthy” meaning that they had no chronic symptoms but were concerned about possible exposure, e.g. found tick on clothing.
Most studies assessing psychiatric comorbidity typically consider “depression” and/or “anxiety” using questionnaires that rarely result in a concrete diagnosis. Herein, participants were rigorously evaluated for the presence of clinical disorders like Major Depressive Disorder, Generalized Anxiety Disorder and Post-Traumatic Stress Disorder. The PRIME MD Patient Health Questionnaire was used to screen patients for the presence of clinical disorders which for the purposes of this study included diagnoses falling under the broader categories of mood, anxiety, somatoform, eating, and substance use disorders29. Next, to explore positive screening results corresponding modules from the Structured Clinical Interview for the DSM-IV30 were used to identify definitive diagnoses.
Psychiatric comorbidity also includes personality disorders; however, this other broad category is rarely explored. Personality disorders can co-occur with clinical disorders, but by definition represent an enduring pattern of maladaptive personality traits that are relatively stable over time and apparent in the absence of a clinical disorder such as Major Depressive Disorder25. The most extensively used written assessment instrument for personality disorders, the Millon Clinical Multiaxial Inventory-III31–32, was used to evaluate personality disorders. The single highest scale score was considered the best indicator of personality style. A cutoff score of 90 was used instead of the traditional cutoff score of 85 to reduce false-positives33.
The Coping Strategies Questionnaire’s catastrophizing subscale34 assessed catastrophizing, a dysfunctional cognitive process associated with poor medical outcomes in rheumatologic populations35. Catastrophizing is characterized by pessimistic beliefs where one anticipates the worst case scenario. The catastrophizing subscale has good construct validity36. The Positive and Negative Affect Scale evaluated emotional characteristics37. This reliable and valid self-report questionnaire consists of two mood scales with ten items each for the assessment of positive affect (e.g., inspired, enthusiastic) and negative affect (e.g., nervous, scared)37. Lastly, since fibromyalgia is common in this patient population14,18,20, level of functioning was determined by the global score from the Fibromyalgia Impact Questionnaire (FIQ). The FIQ is a 19-item self-report instrument with a mean of 50 and SD of 10 where higher scores indicate worse functioning38. A version of the FIQ modified for Lyme disease patients was used39.
In most cases, the comparison group was compared to the larger Chronic Multisymptom Illness group consisting of all three conditions. Linear regression was used to analyze continuous response variables (positive affect, negative affect, catastrophizing, and functioning), while logistic regression was used for binary responses (group membership, presence of clinical and personality disorders) and Poisson regression was used for the response of symptom number. Age and sex were used as covariates in all analyses, with the consequence that estimated differences and odds ratios are adjusted for age and sex. For secondary analyses, Tukey’s pairwise comparison method was used to compare subgroups. Variables in the predictive model were determined a priori, so no model selection was performed. Somers’ Dxy rank correlation40 was used to summarize the predictive capability of the model; 0.0 corresponds to no better than random predictions within discordant pairs, while 1.00 corresponds to perfect predictions. P-values and confidence intervals are not adjusted for multiple testing. The R statistical environment was used for analysis41, while the SPSS program (SPSS Inc., Chicago, IL) was used for recording data and generating tables.
Forty-six (19.2%) of the 240 patients evaluated had an active infection with B. burgdorferi at assessment. At follow up, 6 reported persistent symptoms and were classified as Post Lyme Disease Syndrome; another 25 of the 240 patients had been previously diagnosed and treated by other physicians and were also designated as Post Lyme Disease Syndrome (n=31). Seventy-two (30%) of the 240 patients presented primarily with musculoskeletal complaints related to fibromyalgia, while 42 (17.5%) had medically unexplained symptoms. Most patients (60.4%) were categorized as having Chronic Multisymptom Illness. Our comparison group consisted of the 40 patients who recovered from Lyme disease after treatment, 53 patients who had an identifiable medical condition other than fibromyalgia explaining the symptoms, and 2 completely healthy patients concerned about possible exposure (Table 2). Table 3 shows patient demographic and clinical characteristics by group.
In a secondary analysis, the two subgroups of the comparison group did not differ significantly (Tukey’s pairwise comparisons) on any of the primary outcomes. Similarly, the three CMI subgroups did not differ significantly among themselves on any of the primary outcomes. Patients in the Chronic Multisymptom Illness groups reported 47% more physical symptoms than did the comparison group (p<.001; 95% CI, 33% to 62%) with Chronic Multisymptom Illness patients commonly reporting pain (97.2%), fatigue (91.7%), poor concentration (74.5%) and sleep disturbance (75.9%). Patients in the Chronic Multisymptom Illness group had few observable clinical signs like joint inflammation (2.1%) that could be misinterpreted as indicative of Lyme disease. In Chronic Multisymptom Illness patients, equivocal test results, unreliable tests (e.g. Lyme urine antigen test) or unproven laboratories appeared to contribute heavily to previous misdiagnosis.
Although 169 (70.4%) patients at no time met criteria for Lyme disease, 114 (67.5%) of them had received at least one course of oral or intravenous antibiotics. Thirty-three (28.9%) of the 114 patients received multiple courses of oral antibiotics, while intravenous antibiotics were provided for 22 patients, with 12 receiving repeated intravenous antibiotic treatment often over a period of months or years. Nineteen (35.8%) of the 53 comparison group patients with medical conditions (not Lyme disease) received antibiotic treatment, with 9 receiving multiple courses of antimicrobials.
Twenty (21.1%) patients in the comparison group met criteria for a clinical disorder (Table 4). Conversely, clinical disorder rates were significantly higher for patients in the Chronic Multisymptom Illness groups (p<.001; Odds Ratio 3.54, 95% CI, 1.97 to 6.55); the highest rate was observed in the fibromyalgia group (52.8%). Current major depressive disorder and generalized anxiety disorder were the most commonly observed psychiatric disorders. Somatization disorder occurred at comparatively low rates across groups although the rate for patients with fibromyalgia (13.9%) was slightly elevated.
Personality disorders occurred at the highest rate in fibromyalgia group (38.9%); however, the difference between the Chronic Multisymptom Illness groups and the comparison group was not significant. Histrionic personality disorder was observed in 19.4% of fibromyalgia patients (Table 4).
Chronic Multisymptom Illness patients were more likely than comparison group patients to have lower levels of positive affect (p<.001, mean difference −4.3, 95% CI, −6.4 to −2.2), higher levels of negative affect (p<.001, mean difference +3.7, 95% CI, 1.7 to 5.7) and a greater tendency to catastrophize pain (p<.001, difference of 5.8, 95% CI, 3.0 to 7.8). The Chronic Multisymptom Illness group had worse functioning scores than did the comparison group (p<.001; increase of 15.1, 95% CI, 10.3 to 19.9) and these scores were related to catastrophizing (r=.522, p< 0.001), negative affect (r=.483, p< 0.001) and positive affect (r= −.342, p < 0.001). Lastly, clinical disorders were predictive of worse functioning scores for all patients (p<.001, difference of 21.6, 95% CI, 17.1 to 26.1).
The hypothesized prediction model consisting of age, gender, clinical disorders, personality disorders, catastrophizing, positive affect and negative affect adequately accounted for group inclusion (Table 5; p<.001 compared to a base model of age and gender only), with a Somers’ Dxy rank correlation of 0.53. Catastrophizing, positive affect and personality disorders all significantly contributed to the model after accounting for other effects, with odds ratios for the Chronic Multisymptom Illness groups of 2.09, 0.64, and 2.35, respectively, for an increase of 10 on the catastrophizing scale, an increase of 10 on the positive affect scale, or the presence of a personality disorder.
It is worth noting that although the clinical disorder rate differed significantly between groups (Chronic Multisymptom Illness and comparison) while the personality disorder rate did not, as discussed above, in the predictive model the opposite was true with regard to significance. As the tests in the predictive model can be thought of as adjusting for the other terms, these results reflect the relations among the terms in the predictive model and suggest the importance of considering several measures simultaneously.
Our data supported previous observations that the chronic symptoms of most patients presenting to academic Lyme disease referral centers cannot be attributed to ongoing infection with B. burgdoferi14,19–20. While some of our patients had another medical condition that explained the complaints attributed to Lyme disease, the physical, cognitive and emotional symptoms of most patients were more consistent with Chronic Multisymptom Illness and/or psychiatric disorders. Almost 49% of patients in our Chronic Multisymptom Illness group had clinical disorders like Major Depressive Disorder or Generalized Anxiety Disorder, in contrast to the comparison group with a rate of 21.1% which is closer rates observed in the general population, 26.2%42. In addition, psychological factors associated with poor medical outcomes including catastrophizing, high negative affect, and low positive affect were more pronounced in Chronic Multisymptom Illness than in the comparison group. A model based on psychiatric co-morbidity and these three psychological factors predicted the likelihood that a patient would be included in a Chronic Multisymptom Illness group. Moreover, the presence of clinical disorders, catastrophizing, high negative affect and low positive affect were highly related to poor functioning.
We did not find personality disorders to be predictive of Chronic Multisymptom Illness group inclusion, although two of three Chronic Multisymptom Illness groups had elevated rates (Post Lyme Disease Syndrome = 29.0% and fibromyalgia = 38.9%) not seen in the comparison group (21.1%) or in the general population (~9%)43. Of interest, the presence of personality disorders did contribute significantly to our predictive model which emphasizes the importance of considering the relationships amongst several variables at once. However, our instrument is known to produce a high rate of false positives and despite our efforts to control for this, false positives could have been a problem herein. Others have found lower rates of personality disorders in fibromyalgia (8.7%)44. Despite these caveats, Lamberg noted that personality disorder is frequently present in some of our most “difficult” medical patients whether their symptoms are medically explained or not45. Conversely, half of our Chronic Multisymptom Illness patients did not meet criteria for psychiatric co-morbidity, which is consistent with other reports describing subgroups of psychologically healthy fibromyalgia patients who in one study reported lower levels of pain despite increased pain sensitivity46.
Chronic Multisymptom Illness patients had 47% more symptoms than the comparison patients and medicalizing these symptoms has adverse consequences for Chronic Multisymptom Illness patients. When attributing symptoms to an infectious disease, unnecessary antibiotic treatment is often given. Almost 68% of our patients with no evidence of Lyme disease received antibiotic treatment; 29% of them received multiple antimicrobials for months or even years. Baker recently argued that despite convincing results otherwise, some maintain that “chronic Lyme disease” is the result of a persistent infection with B. burgdorferi requiring several months of antibiotic therapy, which is an unprecedented treatment approach for a non-life threatening disease47. Antibiotics are unique in that their use affects the patient to whom they are prescribed, as well as future patients through the generation of new antibiotic-resistant strains of bacteria48–49.
Our study was limited by the cross-sectional design, thus it is inappropriate to infer causality. Depression and poor affect could be the result of living with chronic symptoms instead of being predisposing factors. Longitudinal studies assessing psychological factors in newly diagnosed Lyme disease patients followed over time could better address these questions. Our comparison group consisted of patients with a variety of medical conditions which likely result in varying levels of psychiatric comorbidity. Future studies would benefit from evaluating a more homogenous comparison group, e.g. only osteoarthritis patients. Also, the accurate assessment of the variables of interest was limited by our questionnaires despite selecting most for their “gold standard” status. Lastly, our Chronic Multisymptom Illness patients may not be representative of others who never ascribed symptoms to “chronic Lyme disease” and received antibiotic treatment.
In conclusion, psychiatric comorbidity and other psychological factors are prominent in the presentation and outcome of some patients who inaccurately ascribe long-standing symptoms to “chronic Lyme disease”. Less than 20% of patients presenting to our Lyme disease specialty center had a current infection with B. burgdorferi. Most patients had other medical conditions or Chronic Multisymptom Illness, but were being treated with antibiotics for Lyme disease. Depression and anxiety disorders were common in Chronic Multisymptom Illness patients as were other psychological factors, e.g., negative affect and catastrophizing, associated with poor functional outcomes. Our findings suggest that multidisciplinary treatment addressing the physical and often emotional suffering of such patients will be more effective than perpetuating the diagnosis of “chronic Lyme disease.”
This research was funded by a grant from the National Institute of Mental Health 1 K08MH65360–01.
NOTE: All authors had access to the data and a meaningful role in writing this manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Afton L. Hassett, Assistant Professor, Division of Rheumatology and Connective Tissue Research, Department of Medicine, University of Medicine and Dentistry of New Jersey- Robert Wood Johnson Medical School, New Brunswick, NJ.
Diane C. Radvanski, Senior Research Assistant, Division of Rheumatology and Connective Tissue Research, University of Medicine and Dentistry of New Jersey- Robert Wood Johnson Medical School, New Brunswick, NJ.
Steven Buyske, Associate Research Professor, Department of Statistics, Rutgers University, Piscataway, NJ.
Shantal V. Savage, Research Assistant, Division of Rheumatology and Connective Tissue Research, University of Medicine and Dentistry of New Jersey- Robert Wood Johnson Medical School, New Brunswick, NJ.
Leonard H. Sigal, Director, Pharmaceutical Research Institute, Bristol-Myers Squibb, Clinical Professor, Division of Rheumatology and Connective Tissue Research, Department of Medicine - Director, Lyme Disease Center, Adjunct Professor, Department of Molecular Genetics & Microbiology, University of Medicine and Dentistry of New Jersey- Robert Wood Johnson Medical School, New Brunswick, NJ.