Lyme disease, first recognized in 1976 because of an epidemic of oligoarthritis in children in Lyme, CT, is now the most common vector borne illness in both North America and Europe. The causative organism, Borrelia burgdorferi , was isolated from the tick vector, Ixodes scapularis , in 1982 and from patients in 1983. Erythema migrans (EM), a characteristic rash seen in the majority of cases, was reported in Sweden in 1909. The neurologic manifestations of Lyme disease, particularly peripheral neuropathy in association with meningitis, were described in Europe in the 1940s. Although Lyme disease in North America is almost always caused by B. burgdorferi , related species in Europe and Asia, Borrelia afzelli and Borrelia garinii, cause somewhat different clinical illnesses. Ticks capable of transmitting Lyme disease are part of the Ixodes ricinus complex, including I. scapularis in the northeastern and north central United States, I. pacificus in the western states, I. ricinus in Europe and Ixodes persulcatus in Asia.
In the United States, about 20,000 cases of Lyme disease are reported to the Centers for Disease Control and Prevention (CDC) annually ( Fig. 21-1 ). Most cases occur in the coastal northeastern states (Massachusetts to Virginia), the Midwest (Minnesota and Wisconsin), and the western states (California, Oregon and parts of Nevada) ( Fig. 21-2 ). Lyme disease is underreported to health authorities, so the actual number of cases in the United States may be five- to 10-fold higher than the number of reported cases. In association with a rapidly increasing incidence, there has been geographic expansion and widespread media interest, often sensationalizing Lyme disease.
Lyme disease has characteristic clinical features, and once the disease is established, confirmatory diagnostic tests. For the rheumatologist, Lyme arthritis can be recognized in the same way that gout or rheumatoid arthritis are diagnosed clinically, but confusion exists in the identification of Lyme disease, and it is often overdiagnosed and overtreated. In considering targeted therapies for Lyme disease, a well-grounded understanding of clinical manifestations will solve many treatment problems.
Lyme disease can be divided into stages :
Early localized infection.
Early disseminated infection.
Late Lyme disease.
Because Lyme disease is a tick-borne spirochetal infection, the earliest manifestations of the disease occur primarily during the late spring and early summer ( Fig. 21-3 ), when nymphal ixodes ticks are active ( Fig. 21-4 ). Early Lyme disease begins after an incubation period of days to several weeks following exposure to B. burgdorferi . Late disease occurs weeks to years later. Lyme arthritis, a late manifestation, often begins in the fall or early winter, 3 to 6 months after initial infection.
Overlap can occur between stages. The pattern of illness has changed somewhat since Lyme disease was described thirty years ago, primarily because the disease is more often recognized and treated at onset and such treatment is almost always curative. For this reason, fewer patients progress to early disseminated infection or to Lyme arthritis. For the same reason, Lyme arthritis patients are less likely to report antecedent EM. Recognition of EM leads to treatment, preventing progression of disease. Still, the concept of early and late disease is useful in recognizing patterns of illness. In considering whether a patient has Lyme disease, the clinician needs to understand how progression can occur and to approach the diagnosis based on suspicion of a specific stage of the illness.
Early Localized Infection
Early infection is most frequently recognized because of the EM rash seen in perhaps 80% of patients at the onset of Lyme disease. In early localized disease, spirochetal infection is confined to the skin. B. burgdorferi can be successfully cultured from EM lesions more frequently than from other sites later in the disease. There may be mild fever, malaise, and arthralgias, but other significant manifestations will be lacking. EM is an expanding erythematous rash, often with a well-demarcated outer border ( Figs. 21–5 through 21–8 ). The primary lesion occurs at the site of the tick bite, with centrifically expanding erythema. The lesion is often raised, sometimes indurated, warm, and itchy. It is not often painful. The rash is usually large (greater than 5 cm), and its size will usually distinguish EM from smaller, noninfectious, irritant skin reactions that occur at the bite site in many individuals. EM most typically develops 7 to 10 days after a deer tick bite, with a range of a few days to up to 30 days, but rarely occurs beyond this time. EM may expand within days or sometimes within hours. Appropriate antibiotic therapy promptly eliminates the rash and, even in untreated patients, EM disappears without scarring several days to several weeks after onset.
Patterns of EM are recognized (see Figs. 21–5 through 21–7 ), the most common being a diffuse, homogeneous rash (see Fig. 21-5 ). There may also be central clearing or a bull’s eye appearance (see Fig. 21-6 ). Vesiculation may occur (see Fig. 21-7 ). But beyond these patterns, EM may have atypical features, may be transient, lasting only hours, and does not occur in all patients. Characteristic EM rashes are observed on flat surfaces, such as the chest or in the groin, but if an individual is bitten on a contoured or irregular surface, for example, the ear or toe, there may only be erythematous swelling without more usual features. In these cases, the diagnosis will need to be considered in spite of an atypical rash because of other factors, including tick exposure, time of year, geographic risk of infection, and if present, laboratory confirmation.
Early Disseminated Infection
In some individuals, EM is associated with hematogenous dissemination of the organism to more widespread sites. These include the skin (secondary skin lesions), the liver and spleen (hepatitis), the peripheral nervous system and the brain (cranial and peripheral neuropathies, meningitis and encephalitis), the heart (myocarditis), and the joints (arthritis). Although hematogenous dissemination has been demonstrated to occur even in asymptomatic individuals, early disseminated infection is most often associated with high fever, headaches, stiff neck, significant arthralgias, and malaise. Secondary skin lesions may be seen (see Fig. 21-8 ) and can cover the entire body. They tend to be smaller than the primary lesion, wax and wane over time, and resolve independently from the primary lesion. Lyme disease causes mild hepatitis in about 5% of patients with early disseminated infection. Typically such individuals have mild to moderate transaminitis (aspartate transaminase less than 400 U/L) that resolves over several weeks.
Less than 5% of individuals with early Lyme disease will develop Lyme carditis, a well-characterized clinical syndrome that presents with fluctuating degree of atrioventricular nodal conduction disease. Complete heart block develops in approximately half of these individuals. Such heart block is almost always reversible. There may also be a mild acute myocarditis, but involvement of the heart valves does not occur, nor are there other cardiac arrhythmias such as atrial fibrillation or distal conduction system disease. Lyme carditis has been suggested as a cause of chronic cardiomyopathy in Europe, but there have been no well-documented cases in the United States. One elderly patient coinfected with both the B. burgdorferi and Babesia microtii died from cardiac complications of infection.
The neurologic manifestations of Lyme disease present the most difficult diagnostic problems, but the neurologic features of early Lyme disease are often characteristic. Patients may have disease of the peripheral nervous system; cranial neuropathies are common, particularly facial palsy, which may be unilateral or bilateral. There may be external ophthalmoplegia, and patients may initially seek medical care for diplopia. These manifestations can wax and wane, occur with other manifestations of early disseminated infection, and may be associated with meningitis or, less commonly, encephalitis. Many patients with early neurologic infection have headache, fever, and stiff neck. A significant percentage of these are found to have Lyme meningitis with a mild to moderate lymphocytic pleocytosis, elevated cerebral spinal fluid (CSF) protein levels and an increased CSF index (ratio of CSF to serum B. burgdorferi antibody titers greater than 1). Rare patients have a pattern consistent with meningoencephalitis and have acute cognitive difficulties, emotional lability, or other alterations in higher cortical function. These manifestations usually resolve even without antibiotic treatment, but untreated patients with Lyme meningoencephalitis may take months to recover.
A distinction is made between early and late stage neurologic disease, although this may be a continuum. Late Lyme disease can occasionally affect the central nervous system (CNS). Patients may have cognitive dysfunction, such as memory loss, fatigue syndromes, as well as disease of the peripheral nervous system. Upper motor neuron disease and demyelinating syndromes have been reported, but causation has not been established.
Patients with late neurologic Lyme disease usually have an antecedent history of earlier manifestations, such as EM, cranial nerve palsies or oligoarticular arthritis, and this is helpful in diagnosis. Diagnostic imaging tests of the CNS, such as brain magnetic resonance imaging or brain single-photon emission computerized tomography scanning, may help rule out other conditions, but are not useful in establishing the diagnosis of CNS Lyme disease. Lumbar puncture testing, as in early disease, is important and may show mild lymphocytic pleocytosis, increased CSF protein, and an elevated CSF index (ratio of CSF to serum B. burgdorferi antibodies by enzyme-linked immunosorbent assay [ELISA] greater than 1).
Lyme disease was recognized in the United States in patients with Lyme arthritis, a late manifestation of the disease. Affected individuals are often active outdoors in locations endemic for Lyme disease and typically have a characteristic pattern of arthritis. In the past, most patients also had antecedent EM, but this is now less likely because recognition of EM leads to curative treatment. Such treated individuals will not develop further disease manifestations, including Lyme arthritis. On the other hand, if early Lyme disease is not treated, 60% of these individuals will subsequently develop Lyme arthritis, often 3 to 6 months after disease exposure ( Fig. 21-9 ). Some patients develop migratory arthralgias without frank arthritis, but most develop monoarthritis or, less frequently, an asymmetric oligoarthritis arthritis, usually including the knee ( Fig. 21-10 ). Lyme arthritis usually affects fewer than five joints, typically large joints ( Fig. 21-11 ).
Lyme arthritis tends to occur in intermittent attacks lasting from several days to several weeks, although in about 10% of untreated patients, joint swelling lasts for more than a year and can be considered chronic. Large joint effusions are common in the knees and elsewhere, and often recur after aspiration. Popliteal cysts occur and can rupture if tense synovial effusions are allowed to persist. The arthritis is often not painful and usually resolves with surprisingly little joint dysfunction, although chronic, unremitting arthritis sometimes causes permanent joint destruction.
Although Lyme arthritis is an infection, there are only two reports of successful cultivation of B. burgdorferi from synovial fluid. However, B. burgdorferi DNA is detectable in synovial fluid by polymerase chain reaction (PCR) in most untreated Lyme arthritis patients and synovial fluid PCR testing is usually negative after treatment. Antibiotic therapy clearly improves the natural history of Lyme arthritis, but not all patients respond immediately. It often requires several months for arthritis to resolve, even with successful therapy.
It is unclear why Lyme arthritis responds promptly to antibiotic therapy in some patients but not others. There is evidence that in some patients, persistent Lyme arthritis results from postinfectious, immune reactivity, rather than persistent infection. There may both bacterial and host determinants of arthritis propensity, severity and antibiotic responsiveness. Clinically, patients with Lyme arthritis can be divided into two groups: antibiotic responsive (successful antibiotic treatment in less than three months) and antibiotic refractory (treatment requires more than three months). It is not possible to prospectively identify which patients will be antibiotic refractory, but as discussed later, these patients may require nonantibiotic therapies to achieve resolution of arthritis.
Although Lyme arthritis is an infectious, inflammatory arthritis, its frequent presentation as an acute monoarticular knee effusion in an active individual is often confused with a mechanical internal derangement. Compared with other forms of bacterial arthritis, Lyme arthritis is only occasionally associated with fever or significant constitutional symptoms and is usually less painful. Lyme arthritis shares some clinical features with spondyloarthropathy, but not with most cases of rheumatoid arthritis, because it is not polyarticular or symmetric. There is no convincing evidence that Lyme disease causes osteoarthritis or an osteoarthritic predisposition.
It is possible to culture B. burgdorferi from EM skin biopsy specimens, but such testing is usually not necessary or routinely clinically available. B. burgdorferi has also been cultured, in a research setting, from blood samples from 93 of 213 (43.7%) patients with untreated EM. As noted, there are only two reports of successful culture of B. burgdorferi from synovial fluid in Lyme arthritis patients. B. burgdorferi has been detected by PCR from CSF and, most reliably, from synovial fluid, but such testing has not been validated for widespread use. Thus, the assessment of humoral immunity to B. burgdorferi by serologic testing, in spite of its inherent limitations, remains the primary diagnostic tool in testing for Lyme disease. Because such testing relies on the patient’s immune response, which can take up to 4 weeks to develop, a significant minority of patients with early localized infection will be seronegative for B. burgdorferi antibodies by ELISA at the onset of illness. Because the goal of treatment at this stage is not only to shorten the duration of EM but also prevent late stage manifestations of disease, it is appropriate to treat such individuals with antibiotics based on clinical suspicion, even in the absence of serologic confirmation. These patients will not only have complete recovery but may never make a detectable serologic response against B. burgdorferi.
As B. burgdorferi infection becomes more established, however, most patients with early disseminated Lyme disease will have B. burgdorferi– specific antibodies by ELISA, and virtually all patients with late stage Lyme disease will have such seroreactivity. Criteria have been established to determine positive tests for B. burgdorferi by ELISA and Western blot (WB). According to validated standards, established by the CDC/Association of State and Territorial Public Health Laboratory directors, the criteria for immunoglobulin G (IgG) reactivity require the presence of five or more significant bands. These criteria were developed to maximize sensitivity and specificity in the diagnosis of all stages of Lyme disease. Patients with Lyme arthritis usually widely exceed this WB IgG reactivity and often have IgM reactivity as well ( Table 21-1 ).
|Lyme Disease IgG and IgM, WB|
|Lyme Disease (IgG), WB||POSITIVE||ABN||NEGATIVE|
|18 kD (IgG) band||REACTIVE||ABN|
|23 kD (IgG) band||REACTIVE||ABN|
|28 kD (IgG) band||REACTIVE||ABN|
|30 kD (IgG) band||REACTIVE||ABN|
|39 kD (IgG) band||REACTIVE||ABN|
|41 kD (IgG) band||REACTIVE||ABN|
|45 kD (IgG) band||REACTIVE||ABN|
|58 kD (IgG) band||REACTIVE||ABN|
|66 kD (IgG) band||REACTIVE||ABN|
|93 kD (IgG) band||REACTIVE||ABN|
|NOTE: IgG Western Blot strips which have 5 (or more) of the 10 significant bands are considered positive for specific antibody to B. burgdorferi . The diagnosis of Lyme disease must include careful clinical evaluation and should not be based only on detection of antibodies to B. burgdorferi . A negative western blot interpretation does not exclude the possibility of infection with B. burgdorferi . (From Dressler F, Whalen JA, Reinhardt BN, Steere AC. Western blotting in the serodiagnosis of Lyme disease. J Infect Dis 1993;167:392–400 and the Recommendations of the Second Conference of Lyme Disease, Dearborn, Michigan, 1994.)|
|Lyme Disease (IgM), WB||POSITIVE||ABN||NEGATIVE|
|23 kD (IgM) band||REACTIVE||ABN|
|39 kD (IgM) band||NON-REACTIVE|
|41 kD (IgM) band||REACTIVE||ABN|
|NOTE: IgM Western Blot strips which have a minimum of 2 of the 3 significant bands are considered positive for specific antibody to B. burgdorferi . The diagnosis of Lyme disease must include careful clinical evaluation and should not be based only on detection of antibodies to B. burgdorferi . A negative western blot interpretation does not exclude the possibility of infection with B. burgdorferi . (From Dressler F, Whalen JA, Reinhardt BN, Steere AC. Western blotting in the serodiagnosis of Lyme disease. J Infect Dis 1993; 167:392–400 and the Recommendations of the Second Conference of Lyme Disease, Dearborn, Michigan, 1994.)|
|The State of CT asks us to remind all physicians to report all cases of Lyme disease. Please call them at 860–509–7994 for reporting forms.|
Lyme disease test results must be interpreted in the clinical context of the patient’s history. Not only are negative serologic responses often seen during the first several weeks of infection but persistent positive responses are seen in many patients for months after successful antibiotic therapy for early Lyme disease. Among individuals with Lyme arthritis, the immune response against B. burgdorferi is even more robust. For this reason, most patients with Lyme arthritis remain seropositive for years after successful antibiotic therapy. Such seropositivity, in the absence of other evidence for infection, is not an indication for retreatment and may confer protective immunity. In my experience, patients with a history of Lyme arthritis are unlikely, for a period of several years, to be reinfected with Lyme disease.
For the rheumatologist, evaluating patients with chronic musculoskeletal complaints and questions about late Lyme disease (“Do I have Lyme arthritis?”), serologic testing is a useful adjunct in establishing or excluding the diagnosis, but it is important to determine whether signs and symptoms fit the clinical syndrome. Lyme disease testing should not be indiscriminant because the predictive value of such testing will be greatly diminished if the prevalence of disease in the population being tested is low. Although false-negative results are possible at disease onset, among individuals with a low likelihood of Lyme disease, false-positive results are the bigger problem. At the present time, hundreds of thousands of Lyme disease serologies are being performed to find hundreds of cases. As a result, the predictive value of Lyme disease serologies has been diminished, and false-positive tests are a common clinical problem.
A 56-year-old construction foreman was turkey hunting in Orange, CT, in May. To hunt turkey, he sat on the ground in the woods for several hours. Following this, he noted a minimally engorged deer tick in the right groin. He believed that he had been bitten one and a half hours before he removed the tick. Two days later he developed an erythematous rash in the right groin that was 3 inches in diameter. He otherwise felt well but suspected Lyme disease and visited his primary care physician. He was diagnosed with EM and given doxycycline 100-mg two pills in one dose. He was told to return if he developed any new symptoms. Over 4 days the rash resolved, but 2 weeks later, he developed fever, achiness, and headache. He was seen again by his primary doctor and given doxycycline 50 mg bid to be taken for 21 days. By day 8 of this therapy, he was improved but still had significant fatigue and was unable to function full time at the construction job site. His employer suggested that he get a second opinion about treatment.
A second physician evaluated him in mid June. On examination, the skin was clear. The right groin revealed a bite site, but no rash. There was no significant inguinal lymphadenopathy. The remainder of the examination was unremarkable. The past medical history was one of generally good health. The patient reported having had three previous episodes of EM, 3 years, 5 years, and 8 years previously. Each of these was treated with antibiotic therapy with prompt resolution of illness.
The patient was given doxycycline 100 mg PO bid for an additional 10 days. At the completion of this treatment, his energy had returned to normal and he was entirely asymptomatic 1 month later in follow-up.
Discussion of Case 1
Lyme disease risk is a function of the time of year, geographic exposure, and vocational or recreational activities that increase the likelihood of deer tick bites. This individual, a turkey hunter, spent prolonged periods sitting on the ground during hunting season in the late spring in Connecticut. Perhaps it is not surprising that this was his fourth EM episode. When Lyme disease is recognized and treated at the onset, most patients develop little protective immunity, become seronegative within weeks or months, and may develop the illness again if re-exposed.
This patient was aware of a deer tick bite and removed the tick, by his estimate, one and a half hours after exposure. He saved the tick. It was not engorged. In spite of this, he developed EM. This is not typical. Generally, the duration of tick attachment predicts the likelihood of disease development. The transmission of B. burgdorferi from the mid gut of I. scapularis to a host is relatively inefficient. In experimental animal models and in humans, 24 to 48 hours of tick attachment is required for infection to become reliably established. Therefore, prompt tick removal is one protective strategy to prevent Lyme disease. This patient noted the tick bite and the EM rash in the groin. Warm, moist sites such as the groin or the axilla are common locations for tick bites (see Fig. 21-5A ). It is also common for an individual engaging in high-risk activity, such as turkey hunting, to sustain multiple bites, and it is conceivable that the patient was unaware of the actual tick bite that caused this infection.
The patient recognized that he had early Lyme disease. He presented to his primary care physician, who confirmed this diagnosis based on his risk for Lyme disease and the typical appearance of the EM rash. However, the physician prescribed only two doxycycline 100 mg pills. This is an acceptable strategy to prevent Lyme disease in asymptomatic individuals with engorged deer tick bites. Another appropriate strategy is simply to tell the patient to watch for signs and symptoms of Lyme disease and withhold any treatment, but it is not an appropriate treatment of early Lyme disease. EM indicates that a localized skin infection is already present. In this patient, single dose doxycycline was insufficient to eradicate infection, and he went on, over several weeks, to have further symptoms of early Lyme disease. When seen in follow-up, the patient’s physician again correctly recognized the persistence of Lyme disease but gave less than a full dose of doxycycline therapy. The patient improved but had persistent fatigue that prevented him from returning to work. He felt better after doxycycline was increased to 100 mg bid. Full-dose doxycycline was continued for 10 days. As expected, he was cured of the illness and did well.