Lyme disease is the most common tick-borne disease in North America. From 1982 to 1994, more than 70 000 cases were reported in North America, most of them in the United States. It is important that clinicians diagnose Lyme disease correctly because efficacious therapy is available and delayed or inadequate treatment may lead to various morbid sequelae. On the other hand, inappropriate testing and therapy are costly and expose the patient to risk of the adverse effects of administered antibiotics.

Lyme disease is a complex, multisystem disease caused by the spirochete Borrelia burgdorferi. It affects persons of all ages and both sexes. Since the disease was recognized in Connecticut in 1975, areas in which it is endemic have been identified in several regions in the United States and Canada. In more restricted areas in some northeastern and upper midwestern states, Lyme disease has assumed the characteristics of an emerging epidemic. Its true incidence is almost certainly underestimated as a result of under-reporting.

In the classic presentation, patients develop a distinctive rash, erythema migrans, which is accompanied by such constitutional symptoms as fatigue, headache, mild stiff neck, musculoskeletal aches, and fever. Some weeks after initial exposure, untreated patients may develop symptoms and signs of disseminated disease, particularly neurologic, cardiac, or articular disease.

The Centers for Disease Control and Prevention (CDC) has developed a set of diagnostic criteria for Lyme disease for surveillance purposes [1.3]. These criteria are also applicable to the clinical diagnosis of Lyme disease.

Requests for laboratory testing for Lyme disease are rapidly increasing. In most tested patients, the clinical presentation consists of nonspecific symptoms, such as headache, fatigue, myalgia, and arthralgia; in such cases, even in highly endemic areas, the probability of a false-positive test result is higher than the probability of a true-positive test result. This lack of clinical specificity has been compounded by the lack of standardized serologic tests for Lyme disease; test comparisons of different laboratories have shown poor reliability and accuracy (that is, many false-positive and false-negative results).

General Recommendations

Clinical Diagnosis

According to the CDC surveillance criteria, the clinical diagnosis of early Lyme disease (that is, diagnosis by a physician in a patient who is from an area in which Lyme disease is endemic and who presents with erythema migrans) does not require laboratory confirmation. The confirmation of late disease requires objective evidence of a clinical manifestation of disseminated disease plus laboratory evidence of infection.

Cultures

Isolation of the organism is the best available evidence of causality. Although it is not part of the CDC surveillance criteria, skin culture may be useful in patients in whom primary erythema migrans lesions are suspected. Published evidence indicates that saline-lavage needle aspiration and 2-mm punch biopsies of the leading edge of suspected lesions successfully obtain organisms in 60% to 80% of cases. Although B. burgdorferi grows well in the laboratory, it is not easily recovered from clinical specimens other than biopsy samples of erythema migrans lesions [3.2].

Immunologic Tests

A central message of these recommendations is the importance of appreciating the limitations of laboratory tests for Lyme disease. This is especially true when the pretest probability of Lyme disease is low (<0.20), as judged from the incidence in the community and the absence of objective clinical signs [3.7]. Patients presenting with such symptoms as arthralgia, myalgia, headache, fatigue, or palpitations alone, without the objective signs of Lyme disease, have an extremely low probability of Lyme disease and should not be referred for laboratory testing. Even though the enzyme-linked immunosorbent assay (ELISA) alone has a sensitivity of 89% and a specificity of 72%, a positive test result in patients with a pretest probability of Lyme disease of less than 0.20 is more likely to be false positive than true positive. It is not appropriate to perform laboratory testing when the pretest probability is too low for the predictive value of a positive test result to be clinically useful.

Clinicians must be mindful of the resulting high rate of false-positive test results and inappropriate use of intensive antibiotic therapy, with its attendant risks and cost. However, because a negative test result in low-prevalence situations effectively rules out the presence of Lyme disease, such a result may be clinically useful if it helps to avoid inappropriate empirical antibiotic therapy. Tests for Lyme disease should definitely not be used to screen for disease in healthy persons because the proportion of false-positive results in these persons is even higher.

If the exposure history and clinical presentation suggest that the pretest probability of late Lyme disease is greater than 0.20 (a value that requires clear evidence of the characteristic symptoms and signs of Lyme disease [this is almost never present for patients with nonspecific, diffuse symptoms of myalgia]), then the following laboratory testing is recommended.

Testing with ELISA is the cornerstone of laboratory diagnosis for Lyme disease; when ELISA results are indeterminate, the Western blot is used. It is essential that the tests be carefully standardized and that correct cut-offs be used for clinical diagnosis. Studies have shown that when the prior probability of true disease is 0.20 or greater, the following classification produces the most useful results in terms of both accuracy and cost-effectiveness [3.12]. When appropriately applied, an ELISA with a "positive" result suggests that the probability of Lyme disease is high enough to indicate treatment. An indeterminate result should be followed by a Western blot [3.17]. A positive result on Western blotting "confirms" infection and is an indication for therapy. A negative result on Western blotting greatly decreases the probability of true Lyme disease; thus, therapy is not indicated. In expert hands, the immunofluorescent assay may produce levels of accuracy similar to those of ELISA. However, the immunofluorescence assay is less easily automated, is more difficult to interpret, and is associated with an unacceptably low level of reproducibility [3.29]. Therapy may be warranted if patients with the characteristic signs and symptoms of late Lyme disease and a negative ELISA result have a positive result on a T-cell proliferative assay or polymerase chain reaction of serum or cerebrospinal fluid. Such tests need further validation [3.29].

In cases of suspected neuroborreliosis, testing of cerebrospinal fluid is not necessary [3.25]. The CDC surveillance criteria require either "diagnostic levels of IgM or IgG antibodies to the spirochete in serum or [cerebrospinal fluid]." Because almost all patients with neuroborreliosis who are described in the literature have positive results on serum serologic testing, testing of cerebrospinal fluid is not required to diagnose Lyme disease. Determination of intrathecal antibody production may be useful in identifying meningeal involvement, although estimates of sensitivity and specificity must await prospective studies with adequate sample sizes. These and other recommendations for the diagnosis of Lyme disease must be reassessed periodically as new information becomes available.

Recommendations for the Use of Serologic Testing in Suspected Lyme Disease

1. The initial step should be to determine the pretest probability of Lyme disease in the diagnostic evaluation of a patient for Lyme disease. This determination should be based on the findings of a thorough clinical examination and a knowledge of the incidence of Lyme disease in the population represented by the patient.

2. The decision to use serologic testing in the diagnostic evaluation should be based on an understanding of the patient's pretest probability of Lyme disease and the performance characteristics of the test to be used.

3. In appropriate patients with objective clinical signs (in whom the pretest probability is 0.20 to 0.80), serologic testing for Lyme disease should follow a two-test protocol, under which either ELISA or an immunofluorescence assay is followed by Western blotting of all specimens found to be indeterminate by the first test. Positive results provide clinically useful increases in the likelihood of disease that help diagnose the disease; negative results are similarly useful for ruling out disease.

4. For a patient who presents with symptoms of myalgia (in whom the pretest probability of Lyme disease is low), we recommend therapy for the symptoms but do not recommend testing or antibiotic therapy for Lyme disease. If testing is performed in this situation, a positive result should not be viewed as an indication for antibiotic therapy; a negative result, however, can be a useful aid in the decision not to proceed with empirical antibiotic treatment.

5. For a patient who presents with a rash resembling erythema migrans or with arthritis, a history of a rash resembling erythema migrans, and a previous tick bite (in whom the pretest probability of Lyme disease is high), we recommend empirical antibiotic therapy.

6. For a patient who presents with other objective clinical signs (in whom the pretest probability is intermediate), we recommend a two-step testing strategy.

7. It is important that laboratories conducting Lyme disease testing conform to a quality-control standard and an improved proficiency program.

8. Further serologic testing should be performed with suitable controls to improve the estimates of both sensitivity and specificity in different patient care settings.

9. Recommendations for the laboratory diagnosis of Lyme disease should be reassessed and revised accordingly as new information on testing becomes available.