I had the privilege of participating in the study by Luft and colleagues [1]. The data on the rate of complete resolution of erythema migrans in amoxicillin recipients may have been overstated. My patient, a 75-year-old man, was enrolled in the study in July 1991 when he presented with an engorged, embedded deer tick surrounded by an erythema migrans rash in the right popliteal fossa. This patient received and was fully compliant with treatment. His rash cleared within a few weeks, and he was considered to be a complete responder. About 1 month later, he developed internuclear ophthalmoplegia associated with several areas of increased signal intensity on magnetic resonance imaging of the brain. This was thought to be due to microvascular disease. In October 1991, the patient developed complete heart block and required insertion of a pacemaker. The attending cardiologist did not consider Lyme disease, and I was not notified of this development until December 1991. The patient was next seen in May 1992, when he presented with synovitis and effusion involving his knees and ankles. By July 1992, I was concerned that this patient had not been cured of Lyme disease and communicated my concerns to Pfizer Central Research. It is not clear from the article by Luft and colleagues [1] whether this particular patient was still considered a complete responder, was considered a treatment failure, or was excluded from evaluation. Clarification of this issue is needed before the data on response rates can be considered reliable.

The above concerns notwithstanding. Luft and colleagues provide crucial confirmation of the existence of relapsing Lyme disease in seronegative patients, of the lack of reliability of serologic testing in patients treated with antibiotics, and of the failure of short-course antibiotic therapy to be universally effective even early in the course of disease.