Background
Lyme disease, caused by infection with the tick-borne spirochete
Borrelia burgdorferi, is the most common vector-borne disease in North America. In highly endemic regions of the United States, the annual incidence of infection may be as high as 1–3% with a cumulative prevalence as high as 7–15% [
1,
2]. In the majority of cases, the diagnosis of confirmed early Lyme disease is based on identification of the hallmark erythema migrans (EM) rash, which may occur in isolation or in conjunction with viral-like symptoms such as fever, malaise, fatigue, and generalized achiness [
3].
However, in up to 16% of early Lyme cases, patients do not present with a rash and the primary symptoms of their acute illness are viral-like [
4]. Historically, surveillance criteria requiring a high degree of specificity have excluded patients with only viral-like or subjective patient-reported symptoms in case definitions for Lyme disease [
5]. In 2008, however, the Centers for Disease Control (CDC) surveillance criteria for Lyme disease were modified to include patients with solely subjective symptoms and a positive confirmatory serology as probable Lyme disease [
6]. Based on the revised surveillance case definition, more than 30,000 new confirmed or probable cases of Lyme disease were reported in 2009 [
7], though studies have shown that the actual number of cases may exceed reported cases by a factor of 6 to 12 in endemic areas [
8,
9].
If left untreated, Lyme disease may progress to later stages involving the musculoskeletal, neurologic, or cardiovascular systems. The diagnosis of these late stages of Lyme disease is based on clinical diagnosis with serologic confirmation using CDC surveillance criteria [
10]. The CDC case definition for confirmed late Lyme disease relies on signs of specific organ damage such as inflammatory arthritis with synovitis and joint effusion, or objective neurologic disease, all confirmed by a positive IgG immunoblot (western blot) for antibodies to
B. burgdorferi[
6]. However, an initial longitudinal observation of untreated Lyme disease patients suggested that a significant number (18%) of late Lyme cases may only exhibit symptoms such as fatigue, arthralgias or myalgias, without development of classic physical signs of late Lyme arthritis or neurologic disease [
11]. The patient phenotype of IgG seropositivity and musculoskeletal pain, fatigue, and/or cognitive dysfunction without signs of organ inflammation or dysfunction corresponds with a late manifestation of the current surveillance category of probable Lyme disease. When present, this phenotype can be termed “probable late Lyme disease”.
The specific presentation of probable late Lyme disease, including untreated patients with a history of subjective symptoms and a positive IgG immunoblot, differs in two important ways from those meeting criteria for post-treatment Lyme disease syndrome (PTLDS), a disease category recently added to the Infectious Disease Society of America (IDSA) guidelines [
10]. First, patients with probable late Lyme have no history of a prior physician diagnosis of objective findings consistent with early or late Lyme disease, a requirement for patients with PTLDS. Second, patients with probable late Lyme have not been previously treated with an antibiotic regimen recommended for Lyme disease, also a requirement for PTLDS. Since patients with probable late Lyme disease have serologic evidence of remote exposure to
B. burgdorferi, they represent a distinct subset among patients with other chronic presentations that are often categorized as “medically unexplained” symptoms or syndromes such as chronic fatigue syndrome or fibromyalgia.
The existence of probable late Lyme disease, manifesting only as subjective symptoms with a concurrent positive IgG immunoblot serology, has remained controversial. This patient presentation was described in a recent review categorizing the spectrum of patients labelled as having “chronic Lyme disease” [
12]. The authors suggest that IgG seropositive patients with symptoms but no signs of illness have at most “equivocal evidence for infection with
B. burgdorferi and that any benefit from treatment would be unlikely” [
12]. Patients with probable late Lyme disease share clinical phenotypes which overlap with patients who have
PTLDS, fibromyalgia, chronic fatigue syndrome, and those whose symptoms remain medically unexplained after extensive medical evaluation. Thus, many have argued against serologic screening for Lyme disease among patients whose symptoms of fatigue, widespread pain, and subjective cognitive dysfunction exist in the absence of physical findings or laboratory abnormalities [
2,
12].
Patients with probable late Lyme disease, or those untreated patients with a positive IgG serologic test for Lyme disease and otherwise unexplained symptoms, represent an interesting subset of patients who have not been clinically characterized in the modern Lyme disease literature. This article offers a description of a sample of such patients, within the context of the wider spectrum of late and chronic Lyme patients seen for evaluation in community-based clinical practice.
Discussion
This article presents a sample of untreated patients with a history of persistent, subjective symptoms and IgG antibodies to B. burgdorferi, a group which has not been well-described. Patients with this presentation make up an unknown proportion of all patients presenting for evaluation of Lyme disease to a general practitioner or medical subspecialist. In our experience, this group accounted for 6% of all patients with symptoms greater than 12 weeks who presented for evaluation of Lyme disease in an endemic region, and 36% of patients with confirmed or probable late Lyme disease.
Patients with this clinical presentation have been considered to be most similar to other subsets of patients with the nonspecific label of “chronic Lyme disease” and thus thought not to have evidence of active, untreated infection with
B. burgdorferi[
12]. As a result, the etiology of their symptoms and any recommendations for treatment remain controversial. However, this group of narrowly defined patients with ongoing symptoms and a positive IgG immuno serology would meet CDC surveillance criteria for ‘probable’ Lyme disease. In the clinical context of months of ongoing symptoms, lack of an alternative diagnosis, lack of prior antibiotic therapy, and a positive IgG immunoblot blot serology, we propose that this group of patients has untreated, probable late Lyme disease.
A comparison of probable late Lyme and confirmed late Lyme patients in our sample show some similarities. Patients in both groups have significant immunoblot reactivity on IgG immunoblot blot analysis, exceeding the highly specific CDC criteria 5 band cut-off for positivity. This pattern strongly supports the exposure of both groups to
B. burgdorferi and is consistent with a diagnosis of late, untreated infection. The finding of a positive serology by itself is not diagnostic of active, untreated infection in probable late Lyme disease, as it is also present in the convalescent phase of resolved Lyme disease in the estimated 40% of individuals who are never treated during the acute phase and never develop late manifestations of Lyme disease. However in the context of otherwise unexplained symptoms it is a reasonable hypothesis that the patients’ symptoms are a result of previously untreated infection with
B. burgdorferi. By definition, probable late Lyme disease is characterized by patient-reported symptoms rather than objective physical exam signs of disease, distinguishing this group from those with untreated late Lyme arthritis or neurologic disease. However, the type and chronicity of symptoms in the probable group are very similar to a subset of those patients described in an early observational series of untreated Lyme disease in which 18% of patients developed periarticular or musculoskeletal pain for as long as 6 years, but never developed objective joint abnormalities [
11]. In the same study, other patient-reported symptoms including fatigue (41%), headache (16%), myalgias (9%), and abdominal pain (9%) were also found in addition to objective signs of Lyme arthritis.
These early descriptions suggest that the distinction between subjective, patient-reported symptoms and objective signs of late Lyme disease may not be absolute. For example, patient-reported symptoms of musculoskeletal pain may precede the onset of objective synovitis. In addition, the physical finding of joint swelling may be intermittent, so that on any given physician evaluation the only evidence of late Lyme disease may be the patient’s report of symptoms [
14].
Further, in probable late Lyme disease, the lack of physician-documented signs of early disease, including EM, is not unexpected and does not necessarily negate the possibility for development of late infection. It is known that during the acute phase of infection, EM may not occur, may not be seen, or may be misdiagnosed, allowing patients to progress to later stages of disease [
15,
16]. Series of patients with late Lyme arthritis report a history of EM in 23% of patients and “flu-like illness alone” in 16% of patients that preceded their diagnosis of late Lyme arthritis [
17]. Interestingly, approximately one third of patients in our series with probable late Lyme disease reported a history of a rash at illness onset that was never physician-documented or treated.
Until the causal relationship of symptoms to infection in probable late Lyme disease can be proven pathologically, the practicing physician must weigh the relative risk and benefit of antibiotic treatment in this group of patients. The finding that 8/12 patients with probable late Lyme eventually received antibiotic treatments for Lyme disease prior to their subsequent referral for evaluation demonstrates that current community practice is often to treat these patients. The treatment approach to this group of patients has not been clearly outlined in the medical literature, with one recent review suggesting that any benefit from treatment would be unlikely [
12]. The result that 75% of probable late Lyme patients reported clinically significant improvement with appropriate Lyme disease treatment is consistent with treatment outcomes in confirmed late Lyme arthritis, in which 90% of patients respond to one or more courses of antibiotic therapy. 10% of patients with definite late Lyme disease do not completely respond to antibiotics and are described as having antibiotic-refractory late Lyme arthritis. The pathophysiology of this syndrome is not well understood, although proposed to be an autoimmune based process [
17], the role of ongoing bacterial infection in the process remains under investigation [
18,
19]. The long term outcomes of patients with persistent symptoms after antibiotic treatment of probable late Lyme disease are unclear. In our small case series 4/9 (44%) of patients had relapse of symptoms after their initial antibiotic therapy suggesting that a subset of patients with probable late Lyme disease will go on to develop PTLDS. Future studies are needed to assess effective treatment modalities in a controlled fashion.
Our sample of probable late Lyme disease also showed some differences and some similarities when compared to patients with PTLDS. Patients with probable late Lyme disease had similar rates of patient reported symptoms of pain, fatigue, and cognitive complaints as those patients with PTLDS. In contrast to those with PTLDS, our group of probable late Lyme patients had not been previously diagnosed with Lyme disease, nor had they been previously treated with antibiotics for Lyme disease. The higher rates of IgG seroreactivity in patients with probable late Lyme disease is also in contrast to patients with PTLDS, who all have a history of physician-documented early Lyme disease and variable serologic reactivity. The lower rate of seroreactivity among those with PTLDS is likely explained by exposure to early, effective antibiotics during acute infection, which is known to potentially blunt serologic response to infection with
B. burgdorferi [
20]. It is also in contrast to patients with other, non-Lyme diagnoses and those with medically unexplained symptoms, all of whom lacked IgG immunoblot seropositivity at a level to meet the CDC surveillance criteria for Lyme disease. Until a gold standard with high sensitivity for exposure to
B. burgdorferi becomes available the percentage of patients with medically unexplained symptoms that are due to exposure to
B. burgdorferi infection will remain unknown. Because of the limitations of serology in documenting prior exposure to
B. burgdorferi with early antibiotic treatment Lyme disease, the case definition for PTLDS is clinically based. In an attempt to increase specificity the definition may lose sensitivity by excluding patients with onset of symptoms greater than 6 months after the antibiotic treatment or by excluding those with common pre-existing conditions such as mild depression. As many patients with no specific diagnosis and medically unexplained symptoms had poorly documented past medical histories, we cannot rule out that some may have had unrecognized or undocumented early Lyme disease that was treated, resulting in unrecognized PTLDS.
The observed demographic differences found across disease groups in our sample also warrant further research. The younger, female predominance in the medically unexplained group may reflect the inclusion of patients with syndromes such as fibromyalgia and chronic fatigue syndrome, which are known to have a female predominance. However, the possibility remains unexplored that certain group inclusion or exclusion criteria, or other factors such as patterns of interaction with the health care system, may be associated with specific demographic characteristics.
Previous recommendations have stated that the pre-test probability for Lyme disease in patients without a history of objective manifestations is too low to justify testing and treatment [
2]. Recommendations for Lyme disease testing in patients without objective physical findings have been based on assumptions that the incidence of disease in this population is low and not significantly higher than in the general population of a low-moderately endemic region (0.1–0.01% pre-test probability estimates). Our results suggest that the pre-test probability may be significantly higher than these estimates in patients with symptoms being evaluated in a Lyme endemic area; closer to the 6% assumption which has been used for patients with fibromyalgia-like symptoms from a very high incidence region. This assumption of higher disease prevalence leads to a post-test probability of approximately 25%, more than twice what has been reported in previous analyses [
2]. Among those with probable late Lyme disease, a patient-reported history of rash suggests that the pre-test probability of Lyme disease (and thus the predictive value of a positive serologic result in certain selected patients with very significant histories) may be even higher. We suggest that in patients from Lyme endemic regions, the possible diagnosis of probable late Lyme disease is reasonable to consider in the setting of an unexplained illness and a history highly suggestive of Lyme disease exposure. These patients may benefit from testing for IgG antibodies to confirm exposure to
B. burgdorferi and to suggest the possibility of late untreated infection.
There are several important limitations to this study and future research is warranted. First, the retrospective nature of the data relied largely on patient self-report as well as on serologic results from several commercial laboratories and past medical records from several physician offices. Because of this, results of testing for other tick-borne infections such as babesiosis, anaplasmosis, ehrlichiosis, bartonellosis, and rickettsiosis. were generally not available and were not included in this report. The insensitivity of serology for the early diagnosis of Lyme disease and for documenting remote exposure to B. burgdorferi infection may have led to unintended misclassification of an unknown number of cases currently defined as medically unexplained and led to an underestimation of the true numbers of cases of Lyme disease or PTLDS. Extrapolation of our findings in patients will probable Late Lyme disease to the much larger number of patients with “syndrome” diagnosis such as fibromyalgia, chronic fatigue syndrome, and medically unexplained symptoms with be the focus for future investigations when better biomarkers for Lyme disease and B. burgdorferi exposure becomes available.
In addition, the possibility of sex based differences in performance of serologic tests for Lyme disease may further complicate the ability to establish exposure to B. burgdorferi and to make the accurate diagnosis of Lyme disease. The opportunity exists for recall or other biases, particularly among patients eager to label previously unexplained symptoms. While the retrospective nature of the data is not ideal, we would argue that the opportunity for prospective studies capturing this subset of patients is challenging. Finally, because of the small sample size in our case series, additional studies with larger sample sizes are needed to see if our findings are replicable.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JA conceived of the study, participated in its design and coordination and helped draft the manuscript. AS assisted with the data analysis and helped draft the manuscript. AR participated in the study design, performed the statistical analysis and helped draft the manuscript. All authors read and approved the final manuscript.