Sir—We read with great interest the recently published brief report entitled “Death from Inappropriate Therapy for Lyme Disease” by Dr. Robin Patel et al. . The purpose of our letter is to describe a similar, although less tragic, experience and to reinforce basic principles of test ordering and interpretation with regard to the clinical management of patients with suspected Lyme disease.
A 45-year-old man presented for evaluation for suspected Lyme arthritis. At the time of initial presentation (March 2000), the patient complained of chronic pain in the left knee, which had been diagnosed as mild osteoarthritis and which had recently worsened after the patient spent a weekend playing soccer with his teenage sons. The patient had no rashes, fever, or other systemic abnormalities, and he otherwise felt healthy. He described having a sensation of grinding and grating, but he experienced no buckling out of the knee. Physical examination showed mild-to-moderate bony hypertrophic changes; synovial fluid was aspirated from a very small left-knee effusion. No serologic test was performed. Results of PCR performed on the synovial fluid specimen (at Boston Biomedica, New Britain, Connecticut; table 1, laboratory 1) were positive for the outer-surface protein A (ospA) gene of Borrelia burgdorferi. The patient was told that he would need to receive long-term antibiotic treatment, but first he was referred for a second opinion.
At the time of our evaluation (May 2000), we confirmed the patient's history and noted the results of the physical examination. His past medical history was remarkable for erythema migrans that had been diagnosed by a physician 5 years earlier (in August 1995), confirmed serologically, and successfully treated with no clinical sequelae or residual complaints. Synovial fluid was again aspirated from the left knee, and the results of PCR performed on this specimen were negative for the ospA gene of B. burgdorferi (table 1). Furthermore, the patient was found to be seronegative for B. burgdorferi, according to the criteria of the Centers for Disease Control and Prevention (Atlanta) and the Association of State and Territorial Public Health Laboratory Directors (ASTPHLD), and an x-ray film of his knee demonstrated findings that were typical of mild osteoarthritis. Osteoarthritis was diagnosed, and the patient was advised that he did not have Lyme arthritis. He was not treated with antibiotics.
The patient was seen again 7 months later (December 2000) and was found to have the same mild knee abnormalities and no symptoms of Lyme disease. A synovial fluid specimen was again aspirated from the affected joint and was found to be negative for the ospA gene by PCR; results of repeated serologic studies were also negative. All serum specimens were frozen, archived, and tested concurrently. The negative PCR results for these synovial fluid specimens were also found to be negative at another laboratory facility (table 1).
The diagnostic issues in this case are similar to those related to the case presented by Patel et al. . In the evaluation of patients with suspected Lyme disease, priority should be given to careful clinical assessment and judgment, which can be supplemented by appropriate laboratory testing . This patient clearly did not have Lyme arthritis, and the first PCR result was false positive and clinically misleading. The first test of choice for evaluation of a patient with suspected Lyme arthritis should be serologic examination for B. burgdorferi. Since Lyme arthritis is nearly always a complication of infections that last for months to years, false-negative results of serologic examination are rarely encountered if the testing is performed in a competent laboratory. Seronegative status is extremely unusual among patients with late-stage Lyme disease; therefore, negative serologic test results should essentially rule out the diagnosis of Lyme arthritis, and the differential diagnosis should be revisited . We are unaware of any well-documented case of Lyme arthritis in a seronegative patient. If the first treating physician in this case had performed serologic examinations at the time of initial presentation, the patient would have received the correct diagnosis, and no further workup or treatment would have been suggested.
Virtually all studies of the use of PCR for analysis of synovial fluid or tissue specimens have involved patients who are strongly seroreactive and who undergo testing for the monitoring of disease activity . PCR testing of synovial fluid samples is of limited usefulness for the diagnosis of seronegative patients with suspected Lyme disease, although it is of potential value in the treatment of patients with established Lyme arthritis. Furthermore, it deserves to be emphasized that PCR is technically demanding and that any lapse in technique may result in false-positive test results . Of note, the laboratory that reported the false-positive PCR result in this case is the same one that reported the positive PCR result in the case reported by Dr. Patel et al. . Our patient, however, was more fortunate in that he was finally appropriately convinced that he did not have Lyme disease and did not receive unnecessary therapy.