Lyme Disease: Recommendations for Diagnosis and Treatment Daniel W. Rahn, MD; and Stephen E. Malawista, MD

Transcription

1 DIAGNOSIS AND TREATMENT Lyme Disease: Recommendations for Diagnosis and Treatment Daniel W. Rahn, MD; and Stephen E. Malawista, MD The incidence and the endemic range of Lyme disease in the United States have increased steadily since the disease was originally recognized in Lyme, Connecticut, in Because of the varied clinical manifestations of this illness and the use of unstandardized serologic testing methods, diagnosis is often uncertain and treatment outcomes are often difficult to evaluate. The antibiotic regimens that are commonly used in clinical practice have changed rapidly. They show much regional variation with little critical comparison of treatment results. The clinical diagnosis and the literature on the treatment of the various stages of Lyme disease are reviewed. The reported data are supplemented with recommendations based on IS years of clinical experience with this illness. Annals of Internal Medicine. 1991;114: From the Yale University School of Medicine, New Haven, Connecticut. For current author addresses, see end of text. Lyme disease is a systemic, tick-borne illness with protean clinical manifestations. The onset of the disease is usually heralded by the appearance of a pathognomonic skin lesion, known as erythema chronicum migrans (1), but not all patients manifest this unique marker (2). Unless treatment is initiated early, the disease usually disseminates, often resulting in cardiac, neurologic, or joint manifestations (2, 3). Lyme disease is of increasing importance to clinicians because of its rising incidence and expanding clinical spectrum and geographic range. Cases have been reported from 43 states in the continental United States, but the distribution of cases is still regional; through 1989, 97% of the 7402 reported cases were reported from 9 states in the northeast, upper midwest, and pacific coastal regions (4). Clinicians have had to make patient management decisions with inherent uncertainty in a milieu dominated by both confusing media reports and panicked patients. We therefore developed a set of treatment recommendations to provide a framework for such decision making. These recommendations address treating patients in various stages of this illness. Data on treatment are still emerging, and large areas of uncertainty remain; these recommendations, developed through a review of the literature on the treatment of Lyme disease as well as through our own 15 years of experience with this illness, therefore must be viewed as preliminary and as subject to modification as more information becomes available. Definition Lyme disease is typically defined by clinical evidence supported by serologic test results. Infection can rarely be proved, because of the scarcity of organisms seen on histologic staining of affected tissues and the technical difficulty associated with culturing the causative organism from clinical specimens (5). Assessments of disease incidence and treatment outcomes have been hampered by the absence of a uniform set of diagnostic criteria. The Centers for Disease Control (CDC), in association with state health departments, has recently developed a national surveillance case definition for Lyme disease to address this problem (Table 1) (6). Because this definition is to be used for epidemiologic surveillance and not as a guide to clinical management, it is biased toward certainty in diagnosis and relies heavily on laboratory confirmation, particularly in cases occurring in nonendemic counties. A uniform definition is essential for case counting and for comparing the results of treatment in different trials; however, the clinical application of the CDC criteria has several difficulties. The definition recommends serologic confirmation of erythema chronicum migrans occurring in patients in nonendemic counties. Although such confirmation would increase the likelihood that patients with other skin lesions would not be miscounted as cases of Lyme disease, many patients with erythema chronicum migrans alone do not have a positive serologic response and, therefore, would not be considered as cases. Second, a skin lesion must be 5 cm in diameter to be counted. Residents of endemic areas are so sensitized to the risk for Lyme disease that they often seek medical attention with skin lesions of only 2 to 3 cm. Prompt antibiotic therapy in such situations no doubt results in the unnecessary treatment of some persons, but also has the greatest likelihood of curing Lyme disease. In the neurologic realm, encephalomyelitis is to be attributed to Lyme disease only if it is confirmed by the presence of locally produced antibodies against Borrelia burgdorferi in the spinal fluid. The necessary assay is not readily available in most areas, and its results are not always positive in cases of Lyme disease. If clinical manifestations strongly suggest the diagnosis of Lyme disease but laboratory confirmation is lacking, clinical judgment rather than a rigid adherence to a case definition must direct decisions about treatment. Underdiagnosis and unnecessary delay in the initiation of therapy with antibiotics thus may be avoided. Laboratory confirmation must be regarded as an adjunct in diagnosis. The clinical manifestations of Lyme disease have been reviewed comprehensively recently and therefore are not reviewed here (7-9). Application of Diagnostic Tests Antibody Response A specific immune response against B. burgdorferi is usually detectable by indirect immunofluorescence as American College of Physicians

2 say or enzyme-linked immunosorbent assay (ELISA) within weeks of the onset of the disease (10). Enzymelinked immunosorbent assays are preferred, because their sensitivity and reproducibility are better than those of immunofluorescence assays (11). Immunoglobulin M (IgM) antibody generally first develops within 2 to 4 weeks after the onset of erythema chronicum migrans, peaks after 6 to 8 weeks of illness, and declines to the normal range after 4 to 6 months of illness in most patients. In some patients, the IgM antibody level remains elevated for many months or IgM antibody reappears late in illness; these phenomena predict continued infection. The immunoglobulin G (IgG) antibody level is usually elevated within 6 to 8 weeks after the onset of the disease, peaks after 4 to 6 months of illness, and remains elevated indefinitely in patients with continued infection (11). At present, the methods for the serologic confirmation of Lyme disease are not standardized. The tests offered by various laboratories differ in sensitivity and specificity. (12, 13). Identical serum samples sent simultaneously to different laboratories may yield different results. This problem will likely be rectified within the foreseeable future; the CDC is currently collecting a set of reference sera for use in a standardization procedure. At present, interpreting laboratory results is difficult unless one is familiar with the reference laboratory used and its validation procedures (14). Most ELISAs use extracts of sonicated whole B. burgdorferi as antigen. Modifications of this technique to improve specificity or sensitivity (or both) have included the use of an IgM capture technique that improved the sensitivity of detection of IgM antibody early in the course of the disease (15); the adsorption of test serum with Escherichia coli to decrease nonspecific binding to cross-reactive bacterial epitopes (16); the preparation of purified flagellin protein for use as antigen, which may improve the specificity of a positive IgM antibody early in the course of the disease, because the 41-kD flagellar antigen is the major antigenic target of IgM (17); and the dissociation of putative immune complexes in serum before doing an ELISA to detect antibody previously sequestered in the immune complexes, a procedure recently reported to yield positive results in a very high percentage of a small group of patients who were seronegative by conventional serologic testing (18). These modifications all have merits, but some require confirmation and none is broadly available. Enzyme-linked immunosorbent assays that use whole organisms in antigen preparations are currently the standard, with the understanding that results may vary widely, unless the laboratory used has an aggressive quality-assurance program. True seronegativity is uncommon in patients with clinical manifestations of disseminated or chronic Lyme disease. Western Blotting Western blotting may aid in distinguishing true-positive from false-positive ELISA results. In Lyme disease, antibodies develop in a characteristic sequence against some immunogenic epitopes on B. burgdorferi (19). A 41-kD flagellar antigen, the usual primary target Table 1. Lyme Disease National Surveillance Case Definition* Lyme disease is a systemic, tick-borne disease with protean manifestations, including dermatologic, rheumatologic, neurologic, and cardiac abnormalities. The best clinical marker for the disease is the initial skin lesion, erythema migrans, that occurs in 60% to 80% of patients. Case definition for the national surveillance of Lyme disease: 1. A person with erythema migrans; or 2. A person with at least one late manifestation and laboratory confirmation of infection. General definitions: 1. Erythema migrans: For purposes of surveillance, erythema migrans is a skin lesion that typically begins as a red macule or papule and expands over a period of days or weeks to form a large round lesion, often with partial central clearing. To be considered to be erythema migrans, a solitary lesion must measure at least 5 cm. Secondary lesions may also occur. Annular erythematous lesions developing within several hours of a tick bite represent hypersensitivity reactions and do not qualify as erythema migrans. In most patients, the expanding erythema migrans lesion is accompanied by other acute symptoms, particularly fatigue, fever, headache, mildly stiff neck, arthralgias, and myalgias. These symptoms are typically intermittent. The diagnosis of erythema migrans must be made by a physician. Laboratory confirmation is recommended for patients with no known exposure. 2. Late manifestations: These manifestations include any of the following when an alternate explanation is not found'. a. Musculoskeletal system: Recurrent, brief attacks (lasting weeks or months) of objective joint swelling in one or a few joints sometimes followed by chronic arthritis in one or a few joints. Manifestations that are not considered to be criteria for diagnosis include chronic progressive arthritis that is not preceded by brief attacks and chronic symmetric polyarthritis. Additionally, arthralgias, myalgias, or fibromyalgia syndromes alone are not accepted as criteria for musculoskeletal involvement. b. Nervous system: Lymphocytic meningitis, cranial neuritis, particularly facial palsy (may be bilateral), radiculoneuropathy, or, rarely, encephalomyelitis alone or in combination. Encephalomyelitis must be confirmed with evidence of antibody production against Borrelia burgdorferi in the cerebrospinal fluid, shown by a higher titer of antibody in the cerebrospinal fluid than in the serum. Headache, fatigue, paresthesias, or mildly stiff neck alone are not accepted as criteria for neurologic involvement. c. Cardiovascular system: Acute-onset, high-grade (second- or third-degree) atrioventricular conduction defects that resolve in days to weeks and are sometimes associated with myocarditis. Palpitations, bradycardia, bundle-branch block, or myocarditis alone are not accepted as criteria for cardiovascular involvement. 3. Exposure: Exposure is defined as having been in wooded, brushy, or grassy areas (potential tick habitats) in an endemic county no more than 30 days before the onset of erythema migrans. A history of tick bite is not required. 4. Endemic county: A county in which at least two definite cases have been previously acquired or in which a tick vector has been shown to be infected with B. burgdorferi. 5. Laboratory confirmation: Laboratory confirmation of infection with B. burgdorferi is established when a laboratory isolates the spirochete from tissue or body fluid, detects diagnostic levels of immunoglobulin M or immunoglobulin G antibodies to the spirochete in the serum or the cerebrospinal fluid or detects an important change in antibody levels in paired acute and convalescent serum samples. States may determine the criteria for laboratory confirmation and diagnostic levels of antibody. Syphilis and other known biologic causes of false-positive serologic test results should be excluded, when laboratory confirmation is based on serologic testing alone. * This epidemiologic case definition is intended for surveillance purposes only. 15 March 1991 Annals of Internal Medicine Volume 114 Number 6 473

3 Table 2. Evaluation of Laboratory Aids in Diagnosis Procedure Indication Culture Not readily available; low yield except in skin, where biopsy is rarely necessary for diagnosis. Antibody titers in Adjunct in diagnosis after the first 2 to 4 serum weeks of illness. Response may be ablated by antibiotic therapy administered early in the course of the disease. Tests are not yet standardized among different laboratories. Antibody titers in Adjunct in diagnosis of central nervous the cerebrospinal system involvement. A positive result fluid is strong evidence for central nervous system infection. Western blots Adjunct in diagnosis when antibody response is equivocal or when false-positive result is suspected. The result may be equivocal, especially in first few months of illness. T-cell blastogenic Not readily available; specificity unclear; response not recommended for diagnosis at present. Urine antigen No clinical validation; sensitivity and testing specificity unclear; not recommended for diagnosis at present. Polymerase chain reaction Under development for identification of spirochetal DNA in ticks and in patient tissues andfluids;no clinical validation. of IgM antibody, is seen first (20). Antibodies subsequently appear against outer surface proteins with molecular weights of approximately 31 kd (OspA) and 34 kd (OspB) as well as against a 60-kD common antigen believed to be a heat shock protein and several proteins with lower and higher molecular weights (19). The function of these bacterial surface proteins, which vary somewhat with different isolates, is unknown. An ELISA result can be validated by a Western blot result if the serum sample contains antibodies against epitopes that are unique to B. burgdorferi. Less dramatic results are more difficult to interpret. Using this criterion, a Western blot may not have a positive result until after many months of illness, because of the late appearance of antibodies against OspA or OspB (or both) (19). Cell-Mediated Immunity Testing Cell-mediated immunity testing has been used as a means of identifying exposure to B. burgdorferi; the earliest immune response after the onset of infection is a T-cell response (21). Doing these tests is technically difficult and labor-intensive, and their clinical usefulness has not been confirmed. The frequency with which patients with this disease have a positive T-cell response and a negative B-cell response is very low, so there is no indication for routine testing of the T-cell immune response (21). Such testing should not be used routinely in clinical diagnosis and management. Diagnostic Tools under Development Borrelia burgdorferi antigens may be excreted in the urine of experimentally infected animals and naturally infected humans (22). At the present time, urine antigen tests have not been clinically validated and should not be used in the diagnosis or management of patients with Lyme disease. Tests to identify B. burgdorferi DNA in ticks using the polymerase chain reaction are also being developed and may soon be available as an adjunct in assessing the likelihood of acquiring Lyme disease after ixodid tick bites (23). This technology may eventually prove useful for the identification of spirochetal DNA in human tissues or fluids, but it has not been validated for this purpose (24, 25). Recommendations The combination of an ELISA for detecting IgM and IgG anti-z?. burgdorferi antibodies and a Western blot to confirm questionable ELISA results (especially unexpected low-titer positive results) offers the greatest sensitivity and specificity for the laboratory diagnosis of Lyme disease at present (Table 2). Pitfalls of Antibody Testing The predictive accuracy of a positive or negative serologic result depends on the pretest likelihood of Lyme disease being present. If serologic tests are requested indiscriminately, a high rate of false positivity, especially at low titers, must be expected. The illnesses known to be associated with false-positive serologic results include infectious mononucleosis, rheumatoid arthritis, systemic lupus erythematosus, and other spirochetal diseases such as periodontal disease (11, 26, 27). Additionally, a true-positive serologic result merely indicates previous immunologic exposure; the symptoms under evaluation need not be related to active Lyme disease. In two epidemiologic studies conducted in the United States, a substantial number of asymptomatic seropositive patients in groups at high risk for exposure to B. burgdorferi have been identified (28, 29). When such patients seek medical attention for any reason, a serologic test for Lyme disease would uncover their seropositivity. Careful clinical evaluation is the only means of determining whether the complaints under evaluation are likely to be related to Lyme disease. Serologic testing should be ordered only when clinical evidence suggests the diagnosis of Lyme disease, in which case a positive test result should be considered confirmatory of the diagnosis. Treatment Recommendations The treatment recommendations presented here are based on a review of reported data on the treatment of patients with Lyme disease in various stages in the United States and on our extensive experience in treating this illness at Yale University. The review was largely limited to North American treatment results because of the somewhat different clinical course of Lyme disease in Europe. This variation may reflect biologic differences in infecting organisms that could affect treatment results. Common clinical practice frequently deviates, often with reason, from the regimens used in clinical trials. The treatment recommendations presented here and summarized in Table 3 are only guidelines; as March 1991 Annals of Internal Medicine Volume 114 Number 6

4 more data about treatment emerge, these guidelines will likely require modification. Recommendations are presented according to the clinical manifestations when therapy is initiated. Drug dosage should be adjusted for children. Early Lyme Disease Literature Review The first randomized, prospective study of antibiotic therapy for early Lyme disease showed that phenoxymethyl penicillin, tetracycline, and erythromycin (each administered as 250 mg orally four times daily for 10 days) all shortened the duration of erythema chronicum migrans and reduced the incidence of progression to major late sequelae of infection (nervous system manifestations, carditis, and recurrent arthritis) (30). However, minor late sequelae, including headache, fatigue, arthralgias, transient facial palsy, and one or a few brief attacks of arthritis, occurred with all three antibiotics. Tetracycline did best, as measured by complete protection against major sequelae, and, by the same measure, failures were more common with erythromycin. Extending the course of antibiotic therapy with tetracycline to 20 days did not alter the occurrence of minor late sequelae. Patients with severe (disseminated) early disease before treatment were twice as likely to have minor late sequelae, as were patients who presented with a single skin lesion alone. In a second study, 72 patients were randomly assigned to treatment with either the combination of amoxicillin and probenecid (500 mg each three times daily) or doxycycline (100 mg twice daily) for 21 days. (In practice, amoxicillin has largely supplanted penicillin because of its better in-vitro activity against B. burgdorferi [31]. Similarly, doxycycline has replaced tetracycline because of its twice-daily dose schedule and because its gastrointestinal absorption, tolerability, and penetration into the central nervous system are better than those of tetracycline.) The response rate was 100%, as measured by resolution of erythema chronicum migrans and protection against major late sequelae, but 15% of patients had post-treatment fatigue or arthralgias (or both) lasting at least 3 months. Nine of thirty-one patients with several erythema chronicum migrans lesions or prominent systemic symptoms had persistent post-treatment symptoms compared with 2 of 41 patients who presented with erythema chronicum migrans alone (32). Whether these post-treatment symptoms result from continuing infection or from a parainfectious mechanism is unclear, but they rarely, if ever, respond to repeated courses of antibiotics, and the majority of patients with these symptoms return to normal health over a period of months. Some patients develop a typical fibromyalgia syndrome that responds to the usual treatment for this disorder (33, 34). Some physicians have adopted the practice of using probenecid with penicillin or amoxicillin to extend the serum half-life of the antibiotic; however, no direct comparative trial has been conducted to evaluate whether this practice increases the rate of cure of early Lyme disease (32, 35). Rashes occur more commonly with the use of a combination of probenecid and penicillin than with the use of penicillin alone. The use of tetracyclines is generally contraindicated in children younger than 9 years and in pregnant or lactating women. Additionally, sun sensitivity and photosensitive rashes occur with the use of tetracyclines. Jarisch- Herxheimer-like reactions may occur after the administration of the first or second dose of antibiotics. Clinical trials are in progress to evaluate oral cefuroxime, a second-generation cephalosporin, and azithromycin, a macrolide that is not available in the United States at present (36). Both look promising. Treatment Recommendations for Early Lyme Disease Our current recommendations for the treatment of early Lyme disease are shown in Table 3. Erythromycin is a less desirable alternative to amoxicillin or doxycycline. The optimal dose and the optimal duration of therapy have not been established for any agent. The duration of antibiotic administration should be individ- Table 3. Recommendations for Antibiotic Treatment* Early Lyme diseaset Doxycycline, 100 mg twice daily for 10 to 21 days Amoxicillin, 500 mg three times daily for 10 to 21 days Erythromycin, 250 mg four times daily for 10 to 21 days (less effective than doxycycline or amoxicillin) Lyme carditis Ceftriaxone, 2 g daily intravenously for 14 days Penicillin G, 20 million units intravenously for 14 days Doxycycline, 100 mg orally twice daily for 14 to 21 days, may sufficed Amoxicillin, 500 mg orally three times daily for 14 to 21 days, may sufficed Neurologic manifestations Facial nerve paralysis For an isolatedfinding,oral regimens for early disease, used for at least 21 days, may suffice. For afindingassociated with other neurologic manifestations, intravenous therapy {see below) Lyme meningitis Ceftriaxone, 2 g daily by single dose for 14 to 21 days Penicillin G, 20 million units daily in divided doses for 10 to 21 days Possible alternatives for Lyme meningitis Doxycycline, 100 mg orally or intravenously for 14 to 21 days Chloramphenicol, 1 g intravenously every 6 hours for 10 to 21 days Lyme arthritis Doxycycline, 100 mg orally twice daily for 30 days Amoxicillin and probenecid, 500 mg each orally four times daily for 30 days Penicillin G, 20 million units intravenously in divided doses daily for 14 to 21 days Ceftriaxone, 2 g intravenously daily for 14 to 21 days In pregnant women For localized early Lyme disease, amoxicillin, 500 mg three times daily for 21 days For disseminated early Lyme disease or any manifestation of late disease, penicillin G, 20 million units daily for 14 to 21 days For asymptomatic seropositivity, no treatment necessary * These guidelines are to be modified by new findings and should always be applied with close attention to the clinical course of individual patients. t Shorter courses are reserved for disease that is limited to a single skin lesion only. X Oral regimens are reserved for mild cardiac involvement {see text). Regimens for radiculononeuropathy, peripheral neuropathy, and encephalitis are the same as those for meningitis. 15 March 1991 Annals of Internal Medicine Volume 114 Number 6 475

5 ualized according to the severity of illness and the rapidity of clinical response. No reported study has examined courses of oral antibiotics exceeding 21 days for the treatment of early Lyme disease. Minor sequelae have occurred with all antibiotics studied to date and have not been shown to result from continuing infection or to respond to longer or repeated courses of therapy. The decision of whom to treat is at least as difficult as the decision of how to treat. Early Lyme disease is diagnosed by clinical presentation alone and cannot be confirmed serologically in many circumstances (10). Patients may be seronegative at presentation, and curative antibiotic therapy given at this stage of disease may truncate the development of a mature immune response and prevent serconversion after treatment. The threshold for treatment of a patient with a skin lesion compatible with erythema chronicum migrans and no other manifestations of Lyme disease must be individualized and, in general, should be lower in endemic than in nonendemic areas. However, the decision to initiate therapy with antibiotics on minimal evidence should not commit the physician to lengthy or repeated courses of treatment when evidence in support of the diagnosis remains scant. Carditis Literature Review Lyme carditis is uncommon, occurring in 8% of patients in an epidemiologic study conducted in Connecticut (37). In patients with antecedent erythema chronicum migrans, carditis becomes apparent clinically a median of 3 weeks later, with a broad range (38). By far the most common abnormalities are varying degrees of atrioventricular block causing palpitations, syncope, exertional dyspnea or light-headedness, or fatigue. Atrioventricular conduction may fluctuate between first-degree atrioventricular block, Mobitz 1 second-degree block, and complete heart block. Most patients require hospitalization for cardiac monitoring, and temporary pacing may be necessary for those with complete heart block complicated by long pauses or slow escape rhythms. Even in the pre-antibiotic era of Lyme disease, carditis was generally self-limited, with complete recovery the rule. In the original report of Lyme carditis, 18 of 20 patients had varying degrees of heart block; the other 2 had transient ST-T abnormalities and reversible impairment of ventricular performance (38). Nine patients with heart block were treated with prednisone and completely recovered within 1 to 2 weeks. Nine patients received only aspirin and recovered within 1 to 6 weeks. Both patients with ST-T changes also recovered, 1 receiving aspirin alone and 1 receiving no therapy. Although other manifestations of late Lyme disease occurred commonly in this group of patients, cardiac abnormalities did not recur despite management without antibiotics. In a second report, four patients with Lyme carditis with high-grade atrioventricular block were treated with various antibiotics, including oral tetracycline, 250 mg orally four times daily for 2 weeks; penicillin G, 12 to 24 million units daily intravenously for 2 weeks; and ceftriaxone, 1 g intravenously twice daily for 2 weeks (39). Three patients had prompt, complete resolution of conduction abnormalities; the fourth had persistent Mobitz 1 second-degree block 16 months after the completion of therapy, but was otherwise well. Less commonly, patients have had more diffuse cardiac involvement, with reversible ventricular dysfunction and electrocardiographic changes suggestive of myopericarditis. A case report from the European literature documented B. burgdorferi in the myocardium (by transvenous endomyocardial biopsy) in a patient who had dilated cardiomyopathy for more than 2 years; no improvement occurred after antibiotic therapy with ceftriaxone, 1 g daily for 14 days (40). In a separate report from The Netherlands, three patients who had electrophysiologic testing and transvenous endomyocardial biopsy had histologic evidence of endomyocarditis, and two of them had conduction abnormalities distal to the atrioventricular node (41). The only reported death associated with Lyme disease occurred suddenly in a patient with concurrent babesiosis who was subsequently shown to have a pancarditis containing spirochetes (42). Diffuse myocarditis may be a more frequent occurrence than clinical evidence alone would suggest. Treatment Recommendations Despite the apparent benignity of Lyme carditis in most patients and the absence of data supporting the need for aggressive antibiotic therapy, administering intravenous antibiotic therapy to patients with all but the mildest forms of cardiac involvement (first-degree atrioventricular block of less than 0.40 s) has become common clinical practice. This practice is warranted in view of the data substantiating that B. burgdorferi can invade the myocardium directly and that other manifestations of disseminated infection occur commonly in these patients. Our current recommendations are offered in Table 3. Neurologic Manifestations Therapy for the neurologic manifestations of Lyme disease is structured according to the particular lesion present. Minor neurologic symptoms, including headache, stiff neck, and irritability, occur commonly during the early course of the disease and are not considered here (30). However, patients with prominent neurologic symptoms accompanying early Lyme disease may be at an increased risk for failing oral therapy and require more study (36). Considered separately are the wellcharacterized neurologic syndromes of Lyme disease: Bell palsy, lymphocytic meningitis, radiculoneuritis and peripheral neuropathy, and encephalopathy and encephalomyelitis. Bell Palsy Facial nerve paralysis typically occurs early in the course of Lyme disease, often appearing while erythema chronicum migrans is still present; the mean time between the onset of erythema chronicum migrans and the appearance of Bell palsy was 20 days in a retrospective series of 124 palsies in 101 patients (43). Of the 122 paralyses for which the outcome was known, March 1991 Annals of Internal Medicine Volume 114 Number 6

6 105 patients had complete recovery, 16 patients had near-complete recovery (6 of the 16 had a spinal fluid pleocytosis and other neurologic abnormalities at the time of presentation), and 1 patient had severe residual weakness. Treatment had no clear effect on the outcome of the facial palsy per se. The median time to recovery (24 days) was identical, regardless of whether patients were treated with antibiotics alone, corticosteroids alone, or antibiotics plus corticosteroids. Untreated patients had full recovery within a median of 26 days. The only poor outcome was in a patient treated with corticosteroids. Although the study sample was small, there was no difference in outcome among patients, regardless of whether they received oral or intravenous penicillin, oral tetracycline, or oral erythromycin. Because the outcome of Bell palsy in Lyme disease is favorable in most instances, with or without antibiotic therapy, the primary rationale for treatment is the prevention of other manifestations of disseminated disease. Corticosteroids have no identified role. Some authorities now recommend lumbar puncture for all patients with Lyme disease-related Bell palsy and treatment with intravenous antibiotics if a pleocytosis is present. Most patients with Bell palsy without other clinically apparent manifestations of neurologic dissemination, however, remain well after oral antibiotic therapy. The implication of an asymptomatic pleocytosis therefore is unknown. We recommend oral antibiotic therapy with amoxicillin or doxycycline for 21 to 30 days unless symptoms suggestive of meningitis or radiculoneuritis are present, in which case we recommend evaluating the patient appropriately and treating the patient accordingly (see below). Because facial nerve palsy may occur within the first month of illness, not all patients with Lyme disease are seropositive at the time of presentation. The decision to treat a patient with facial palsy alone with antibiotics for Lyme disease must be based on epidemiologic factors. Serologic testing can be repeated 6 weeks later in questionable cases. Meningitis Lymphocytic meningitis, with or without cranial or peripheral neuropathies and mild encephalopathy, occurred in 7% of patients seen at the Yale Lyme Disease Clinic from 1976 to 1978 (44). Meningitis resolved in these patients without antibiotic therapy, although often only after a fluctuating course for many months. Corticosteroid therapy alleviated headache but did not clearly shorten the duration of meningitis, and many patients eventually also developed frank arthritis (45). Lyme meningitis was subsequently shown to result from the spread of spirochetal infection to the meninges (5), and a series of 12 patients at Yale were treated with a 10-day course of intravenous penicillin G (20 million units per day) (45). The mean duration of neurologic abnormalities before the initiation of treatment was 6 weeks, and 3 patients had previously received prednisone. Ten patients had superimposed cranial or peripheral neuropathy or both. Meningitis resolved promptly in all patients, but sensory and motor deficits persisted afterward for 1 to 24 weeks, and 3 patients had persistent musculoskeletal pain and fatigue after the completion of treatment. Most clinicians favor using longer courses of therapy at present, but whether such a practice reduces the incidence of persistent symptoms is unclear. Ceftriaxone (2 g intravenously daily for 14 days) has been curative for Lyme meningitis, but few instances of its use in the United States have been reported (46). In Europe, intravenous penicillin, ceftriaxone, and cefotaxime have been shown to be effective for the treatment of radiculoneuropathy and meningitis (47, 48). In a single report in the American literature, chloramphenicol (1 g intravenously every 6 hours for 10 days) was noted to be effective (49). Doxycycline (100 mg twice daily orally or intravenously for 14 days) was used successfully in Sweden in a series of nine patients with neurologic manifestations of Lyme disease, five of whom had a pleocytosis in the cerebrospinal fluid when therapy was initiated (50); however, the use of this agent for treating Lyme meningitis in the United States has not been reported. Antibiotic therapy shortens the duration of Lyme meningitis. All suspected cases should be confirmed by doing a diagnostic lumbar puncture and showing a lymphocytic pleocytosis in the cerebrospinal fluid. When possible, anti-b. burgdorferi antibody levels in the serum and the cerebrospinal fluid should be measured simultaneously; hyperconcentration of antibody in the cerebrospinal fluid is strong evidence in favor of the diagnosis of Lyme meningitis, although this finding is not always present. Intravenous ceftriaxone or penicillin is the current agent of first choice; ceftriaxone's once-daily dose schedule (which makes outpatient treatment feasible) and better in-vitro performance make it the preferred agent (31). The duration of antibiotic therapy should, to some extent, be individualized, but documenting the complete clearance of the pleocytosis in the cerebrospinal fluid before the discontinuation of therapy with antibiotics is not necessary. Radiculoneuritis and Peripheral Neuropathy These abnormalities have a greater tendency for chronicity than do cranial nerve palsies or meningitis. The symptoms of these abnormalities include paresthesias, numbness, and, less often, motor weakness in the territory supplied by a nerve root or peripheral nerve. These abnormalities can be confirmed by electrophysiologic testing (44). Nerve conduction studies and electromyography have shown subclinical deficits in some patients (51). Although the pathogenesis of radiculoneuropathy is unclear, with some evidence favoring autoimmunity (52), there is mounting evidence of antibiotic responsiveness. In a series of 14 symptomatic patients treated with various regimens, including penicillin G (24 million units daily intravenously for 10 days), amoxicillin plus probenecid (500 mg each orally four times daily for 30 days), ceftriaxone (2 g intravenously twice daily for 14 days), penicillin VK and probenecid (1000 mg and 500 mg orally four times daily for 30 days), and chloramphenicol (1 g intravenously four times daily for 10 days), the conditions of 13 patients responded, with significant improvement in both clinical and electrophysiologic evidence of peripheral neuropathy. Three 15 March 1991 Annals of Internal Medicine Volume 114 Number 6 477

7 patients had concurrent spinal fluid abnormalities that were not detailed in the report (51). In the series of 12 patients with meningitis discussed in the preceding section, several had superimposed cranial neuritis (4 patients), radiculoneuropathy (2 patients), or both (4 patients). Although the peripheral nervous system manifestations of all patients eventually completely resolved after intravenous penicillin therapy, the neuropathic process resolved much more slowly than the associated meningitis, requiring up to 24 months (45). At present, the evidence favors treating peripheral neuropathy or radiculoneuropathy in the same manner that Lyme meningitis is treated, especially because the two often occur together (Table 3). In some instances, however, particularly when there is no evidence of central nervous system involvement, oral therapy may suffice (50). Radiculoneuropathy and peripheral neuropathy may resolve slowly (perhaps reflecting the slow rate of recovery from axonal injury), and there is no evidence that prolonged courses of antibiotics (longer than 1 month) influence the rate of recovery. Encephalopathy and Encephalomyelitis Borrelia burgdorferi infection has been shown to cause a mild encephalopathy with deficits in memory and cognitive function (53, 54). Much less commonly, it has been associated with focal brain or spinal cord parenchymal inflammation mimicking multiple sclerosis (55-57). These abnormalities are associated with positive Lyme serologies and often with intrathecal synthesis of specific antibody which, if present, may be the best indicator of direct central nervous system invasion and antibiotic responsiveness (54). Lyme encephalitis may be accompanied by meningitis. Both intravenous penicillin G (57) and ceftriaxone (51, 54) have been reported to reverse cerebral dysfunction, but data on optimal treatment are lacking and the frequency of the occurrence of this syndrome is unclear. On one occasion, spirochetes morphologically compatible with B. burgdorferi were seen on silver staining of a temporal lobe biopsy specimen (57), but no culture-proved case of parenchymal brain or spinal cord infection has been reported. The antibiotic regimens used for the treatment of Lyme disease-associated encephalitis or encephalomyelitis have been selected empirically and are identical to the intravenous regimens used for meningitis (Table 3). The outcome is usually favorable, but cases of incomplete resolution have been reported (53, 57). Whether deficits may persist because of continued infection or a para-infectious mechanism is unclear. Suspected cerebral dysfunction associated with Lyme disease should be carefully evaluated and documented before the institution of antibiotic therapy, because the effect of therapy can be assessed adequately only if an anatomic or functional lesion has been identified before the institution of treatment. A complete evaluation should include, in addition to a neurologic examination, a spinal fluid examination if symptoms referable to the central nervous system are present, an electrophysiologic evaluation of any concomitant radiculoneuropathic symptoms, and formal mental status testing of cognitive and memory function. If any deficits are found with this evaluation, consideration should be given to brain imaging. Antibody titers against B. burgdorferi are generally elevated in the serum at this state in the course of the illness but rarely are antibodies found in the cerebrospinal fluid only (54). Chronic fatigue alone has not been shown to be a clinical manifestation of active B. burgdorferi infection. Although debilitating fatigue, musculoskeletal pain, and headache may persist for many months after otherwise successful treatment of Lyme disease, these symptoms do not appear to result from active infection or to respond to antibiotic therapy (33). If no objective abnormality is identified, antibiotic therapy for Lyme disease should, in general, be withheld. Patients with vague symptoms of cognitive dysfunction, no focal deficits or evidence of meningitis, and no evidence of seroreactivity or intrathecal antibody synthesis against B. burgdorferi almost certainly do not have Lyme disease of the central nervous system and should not be treated as though they do (54). Lyme Arthritis The rheumatic manifestations of Lyme disease include migratory musculoskeletal pain and intermittent brief attacks of mono- or oligo-articular arthritis of large joints (particularly the knee), which, less commonly, evolve into chronic synovitis. Chronic synovitis is usually preceded by brief self-limited attacks of arthritis (3). Although arthritis may occur as early as a few weeks after the onset of the disease, the average interval between the onset of the disease and the occurrence of arthritis is 6 months. Patients may present with arthritis alone, however, without antecedent manifestations of early Lyme disease. When arthritis is present, the serum almost always contains anti-#. burgdorferi antibodies. Lyme arthritis has been treated successfully with both oral and parenteral antibiotics. The earliest study of antibiotic therapy for Lyme arthritis was designed as a placebo-controlled trial of penicillin given intramuscularly (2.4 million units of benzathine penicillin weekly for 3 weeks). This therapy cured 7 of 20 patients (58). All patients in the placebo group continued to have recurrent bouts of arthritis or chronic synovitis during the follow-up period. In a second open-label phase, penicillin was given intravenously (20 million units daily for 10 days) to 20 patients, and the conditions of 11 patients improved. The duration of arthritis before the initiation of antibiotic therapy did not predict response, and the conditions of some patients with chronic synovitis improved. A trend suggested that previous intraarticular steroids increased the risk for antibiotic failure. Complete recovery usually occurred after the completion of the antibiotic course. Subsequently, in a separate randomized trial, intravenous ceftriaxone (4 g or 2 g daily for 14 days [the maximum currently recommended dose is 2 g daily]) was compared with penicillin (24 million units daily for 10 days) for the treatment of intermittent arthritis (46). In this study, only 2 of 7 penicillin-treated patients were cured 3 months after treatment, whereas the conditions March 1991 Annals of Internal Medicine Volume 114 Number 6

8 of 17 of 18 ceftriaxone-treated patients improved, including the conditions of 3 of 5 patients with chronic arthritis. Ceftriaxone was similarly superior, although less strikingly so, for the treatment of arthralgias alone. In this trial, previous intra-articular corticosteroid use was associated with ceftriaxone failure. In a third trial, reported to date only in abstract form, two oral regimens were compared as primary therapy for Lyme arthritis. Thirty-six patients were randomly assigned to receive either doxycycline (100 mg orally twice daily) or amoxicillin and probenecid (500 mg each orally four times daily) for 30 days. The response rates were 72% (13 of 18 patients) and 61% (11 of 18 patients), respectively, although the amoxicillin and probenecid regimen was associated with a higher incidence of side effects (59). In a separate phase of the study, 29 patients for whom oral antibiotic therapy had failed were randomly assigned to receive intravenous therapy with either penicillin G (20 million units) or ceftriaxone (2 or 4 g daily) for 14 days. The response rates were comparable and disappointingly low, with only 5 of 14 penicillin-treated patients (36%) and 6 of 15 ceftriaxonetreated patients (40%) being cured (57). Failures with either oral or intravenous therapy resulted from either continuing arthritis or the subsequent appearance of neurologic symptoms. The difference between the low ceftriaxone response rate found in this study and the rate of 94% reported in the earlier ceftriaxone trial is unexplained. These data indicate that the optimal therapy for Lyme arthritis has not yet been established. Several oral and parenteral alternatives exist, but failures occur with any regimen chosen. It is not clear that all patients with persistent arthritis after antibiotic therapy have continuing joint infection; rather, in some patients, synovitis may persist because of immunogenetic factors and autoimmunity (60) or a reactive inflammatory mechanism (61), although these suspicions need confirmation. Joint rest and aspiration of reaccumulated joint fluid are general measures that have been classically neglected in treating Lyme arthritis, because pain is often not severe; they merit more attention. We prefer initial therapy with an oral antibiotic regimen unless neurologic manifestations co-exist, in which case either intravenous ceftriaxone or penicillin should be chosen as firstline therapy. The response to antibiotics may occur 3 months or longer after the completion of therapy. Intra-articular corticosteroid injections should be avoided, at least during periods of treatment, because they may increase the risk for antibiotic failure (46, 58). The role of antibiotics for the treatment of arthralgias and myalgias alone in patients with a history of Lyme disease or with seroreactivity for B. burgdorferi is less well established than for the treatment of overt arthritis; repeated courses of antibiotic therapy for such patients should be discouraged. Thefibromyalgiasyndrome has been described as a sequela of Lyme disease that does not respond to antibiotic therapy but that responds to treatment with tricyclic antidepressants, adjunctive anti-inflammatory agents, exercise, and rest (33). Finally, patients with persistent Lyme arthritis of the knee after antibiotic therapy may be treated successfully with arthroscopic synovectomy (62). Pregnancy Both the maternal-fetal transmission of B. burgdorferi and the risk for congenital malformations in children born to women with Lyme disease during pregnancy have raised concern. Maternal-fetal transmission of B. burgdorferi has been reported in isolated cases in which symptomatic early Lyme disease developed during pregnancy and was either untreated or inadequately treated. Case reports have described the premature delivery of an infant who died shortly after birth from congenital heart disease (63), the still birth of a full-term infant (64), and a neonatal death from brain damage, with spirochetes seen in the brain at autopsy (65). In two follow-up studies, maternal Lyme disease was not directly implicated as a cause of fetal malformations (66), and no association was found between the presence of IgG antibody to B. burgdorferi in the cord blood at the time of delivery when congenital malformations were found (67). Pregnant women with active Lyme disease should be treated aggressively. Screening for serologic evidence of previous infection is unnecessary, because neither asymptomatic seropositivity nor a previous history of maternal Lyme disease is associated with any known risks to the fetus. Additionally, although the case reports of maternal-fetal transmission of B. burgdorferi are worrisome, there has been no report of fetal infection when pregnant women with Lyme disease have been treated with currently recommended antibiotics. These reassurances notwithstanding, the threshold for instituting antibiotic therapy for suspected Lyme disease in pregnant women should be lower than that in nonpregnant women. At present, we recommend intravenous antibiotic therapy for all cases of Lyme disease in pregnant women except in those who present with a single erythema chronicum migrans lesion and no associated symptoms of disseminated disease. Tick Bites No data support the common practice of administering antibiotics prophylactically after ixodid tick bites, even in areas known to be endemic for Lyme disease. The single placebo-controlled study of tick bites that has been completed to date was conducted in an area of Connecticut with a high rate of tick infection. Although the study sample was small, the risk for infection was low, occurring in only 1 of 20 untreated patients, and asymptomatic seroconversion did not occur. The use of prophylactic penicillin resulted in an allergic skin rash in 1 of 20 patients, but no case of early Lyme disease or seroconversion occurred in the treated group (68). The low risk for infection in the placebo group may be explained by the fact that nymphal Ixodes dammini (the primary vector in Connecticut) must, at least in mice, remain imbedded for more than 24 hours before they are generally capable of transmitting spirochetes (69). Persons who find a tick on themselves probably remove it before 24 hours of attachment, thus preventing the 15 March 1991 Annals of Internal Medicine Volume 114 Number 6 479

9 transmission of infection. In this study, the investigators concluded that although penicillin may be effective for prophylaxis after a tick bite, waiting for the appearance of symptoms was safe and effective. A study enrolling more patients for greater statistical power is currently in progress in Connecticut, and, to date, the findings support those of the previous study (70). The complications and the cost of antibiotic therapy may outweigh the benefit. Moreover, most patients who develop Lyme disease have been unaware of a tick bite and will not be helped by conclusions about whether to treat tick bites. Caution is in order in generalizing the data from prophylactic treatment trials. Infection rates in ticks vary widely, from 50% or more in /. dammini (71) to \% to 2% in /. pacificus (72), so data collected in one area with one species may not be applicable to other locales with different tick ecology. Antibiotics should not be administered routinely after ixodid tick bites. Patients should be counseled about the symptoms of early Lyme disease. The administration of antibiotics on an individualized basis (for example, to pregnant women) while more data are being collected from clinical studies may be appropriate. Grant Support: By grants AR and AR from the National Institutes of Health. Requests for Reprints: Daniel W. Rahn, MD, Yale University School of Medicine, Section of Rheumatology, P.O. Box 3333, 333 Cedar Street, New Haven, CT Current Author Addresses: Drs. Rahn and Malawista: Yale University School of Medicine, Section of Rheumatology, P.O. Box 3333, 333 Cedar Street, New Haven, CT References 1. Steere AC, Bartenhagen NH, Craft JE, et al. The early clinical manifestations of Lyme disease. Ann Intern Med. 1983;99: Steere AC, Malawista SE, Hardin JA, Ruddy S, Askenase PW, Andiman WA. Erythema chronicum migrans and Lyme arthritis: the enlarging clinical spectrum. Ann Intern Med. 1977;86: Steere AC, Schoen RT, Taylor E. The clinical evolution of Lyme arthritis. Ann Intern Med. 1987;107: Tsai TF, Bailey RE, Miller GL, Craven RB, Letson GW. National surveillance of Lyme disease, [Abstract]. IV International Conference on Lyme Borreliosis. Stockholm: 1990; Steere AC, Grodzicki RL, Kornblatt AN, et al. The spirochetal etilogy of Lyme disease. N Engl J Med. 1983;308: Connecticut Epidemiologist. 1990; 10:. 7. Steere AC. Lyme disease. N Engl J Med 1989;321: Sigal LH. Lyme disease, 1988: immunologic manifestations and possible immunopathogenetic mechanisms. Semin Arthritis Rheum. 1989;18: Rahn DW. Lyme disease: a review of clinical manifestations, diagnosis and treatment. Semin Arthritis Rheum [In press]. 10. Shrestha M, Grodzicki RL, Steere AC. Diagnosing early Lyme disease. Am J Med. 1985;78: Craft JE, Grodzicki RL, Steere AC. Antibody response in Lyme disease: evaluation of diagnostic tests. J Infect Dis. 1984;149: Schwartz BS, Goldstein MD, Ribeiro JM, Schultz TL, Shahied SI. Antibody testing in Lyme disease. JAMA. 1989;262: Luger SW, Krauss E. Serologic tests for Lyme disease: interlaboratory variability. Arch Intern Med. 1990;150: Magnarelli LA. Quality of Lyme disease tests [Editorial]. JAMA. 1989;262: Berardi VP, Weeks KE, Steere AC. Serodiagnosis of early Lyme disease: analysis of IgM and IgG antibody responses by using an antibody-capture enzyme immunoassay. J Infect Dis. 1988; 158: Fawcett PT, O'Brien AE, Doughty RA. An adsorption procedure to increase the specificity of enzyme-linked immunosorbent assays for Lyme disease without decreasing sensitivity. Arthritis Rheum. 1989; 32: Hansen K, Asbrink E. Serodiagnosis of erythema migrans and acrodermatitis chronica atrophicans by the Borrelia burgdorferi flagellum enzyme-linked immunosorbent assay. J Clin Microbiol. 1989; 27: Schutzer SE, Coyle PK, Belman AL, Golightly MG, Drulle J. Sequestration of antibody to Borrelia burgdorferi in immune complexes in seronegative Lyme disease. Lancet. 1990;336: Craft JE, Fischer DK, Shimamoto GT, Steere AC. Antigens of Borrelia burgdorferi recognized during Lyme disease. Appearance of a new immunoglobulin M response and expansion of the immunoglobulin G response late in the illness. J Clin Invest. 1986;78: Grodzicke RL, Steere AC. Comparison of immunoblotting and indirect enzyme-linked immunosorbent assay using different antigen preparations for diagnosing early Lyme disease. J Infect Dis. 1988; 157: Dattwyler RJ, Volkman DJ, Luft BJ, Halperin JJ, Thomas J, Golightly MG. Dissociation of specific T- and B-lymphocyte responses to Borrelia burgdorferi. N Engl J Med. 1988;319: Hyde FW, Johnson RC, White TJ, Shelburne CE. Detection of antigens in urine of mice and humans infected with Borrelia burgdorferi, etiologic agent of Lyme disease. J Clin Microbiol. 1989;27: Persing DH, Telford SR III, Spielman A, Barthold SW. Detection of Borrelia burgdorferi infection in Ixodes dammini ticks with the polymerase chain reaction. J Clin Microbiol. 1990;28: Persing DH, Rys PN, Van Blaricom G, et al. Multi-target detection of B. burgdorferi-assoc'iated DNA sequences in synovial fluids of patients with arthritis [Abstract]. Arthritis Rheum. 1990;33(Suppl): S Nishio MJ, Liebling MR, Rodriguez A, Louie JS. A sensitive method of identifying Borrelia burgdorferi using the polymerase chain reaction [Abstract]. Arthritis Rheum. 1990;33(Suppl):R Weiss NL, Philips MR, Sadock VA, Sigal LH, Merryman PF, Abramson SB. False positive seroreactivity to Borrelia burgdorferi in rheumatic disease: the value of immunoblot analysis [Abstract]. Arthritis Rheum. 1990;33(Suppl);S Magnarelli LA, Miller JN, Anderson JF, Rivere GR. Cross-reactivity of nonspecific treponemal antibody in serologic tests for Lyme disease. J Clin Microbiol. 1990;28: Hanrahan JP, Benach JL, Colemen JL, et al. Incidence and cumulative frequency of endemic Lyme disease in a community. J Infect Dis. 1984;150: Steere AC, Taylor E, Wilson ML, Levine JF, Spielman A. Longitudinal assessment of the clinical and epidemiological features of Lyme disease in a defined population. J Infect Dis. 1986; 154: Steere AC, Hutchinson GJ, Rahn DW, et al. Treatment of the early manifestations of Lyme disease. Ann Intern Med. 1983;99: Johnson RC, Kodner C, Russell M. In vitro and in vivo susceptibility of the Lyme disease spirochete, Borrellia burgdorferi, to four antimicrobial agents. Antimicrob Agents Chemother. 1987;31: Dattwyler RJ, Volkman D, Conaty S, Gorevic P, Luft B. Amoxicillin plus probenecid compared to doxycycline for the treatment of erythema migrans (EM) [Abstract]. Arthritis Rheum. 1990;33(Suppl): S Sigal LH. Summary of the first 100 patients seen at a Lyme disease referral center. Am J Med. 1990;88: Dinerman H, Steere AC. Fibromyalgia following Lyme disease: association with neurologic involvement and lack of response to antibiotic therapy [Abstract]. Arthritis Rheum. 1990;33(Suppl):S Berger BW. Treatment of erythema chronicum migrans of Lyme disease. Ann N Y Acad Sci. 1988;539: Massarotti E, Luger SW, Rahn DW, et al. A multicenter randomized trial of doxycycline, amoxicillin/probenecid, and azithromycin for the treatment of early Lyme disease [Abstract]. Arthritis Rheum. 1990;33(Suppl):S Steere AC, Broderick TF, Malawista SE. Erythema chronicum migrans and Lyme arthritis: epidemiologic evidence for a tick vector. Am J Epidemiol. 1978;108: Steere AC, Batsford WP, Weinberg M, et al. Lyme carditis: cardiac abnormalities of Lyme disease. Ann Intern Med. 1980;93: McAlister HF, Klementowicz PT, Andrews C, Fisher JD, Feld M, Furman S. Lyme carditis: an important cause of reversible heart block. Ann Intern Med. 1989;110: Stanek G, Klein J, Bittner R, Glogar D. Isolation of Borrelia burgdorferi from the myocardium of a patient with longstanding cardiomyopathy. N Engl J Med. 1990;322: van der Linde MR, Crijns HJ, de Koning J, et al. Range of atrioventricular conduction disturbances in Lyme borreliosis; a report of four cases and review of other published reports. Br Heart J. 1990;63: Marcus LC, Steere AC, Duray PH, Anderson AE, Mahoney EB. Fatal pancarditis in a patient with coexistent Lyme disease and babesiosis. Demonstration of spirochetes in the myocardium. Ann Intern Med. 1985;103: Clark JR, Carlson RD, Sasaki CT, Pachner AR, Steere AC. Facial paralysis in Lyme disease. Laryngoscope. 1985;95: Reik L, Steere AC, Bartenhagen NH, Shope RE, Malawista SE March 1991 Annals of Internal Medicine Volume 114 Number 6