We describe a 26-year-old woman from Suffolk County, New York, who had pars planitis caused by Borrelia burgdorferi infection, as confirmed by polymerase chain reaction (PCR) on vitreous fluid. Results of serum Lyme enzyme-linked immunosorbent assay (ELISA), VDRL, and Mantoux tests were nonreactive. Serum IgG immunoblot for Lyme disease showed faint reactivity to the 34-, 41-, 60-, and 61-kDa proteins. Doxycycline (100 mg orally twice daily) was given as empirical treatment for Lyme disease. Visual acuity in the right eye improved from 20/400 to 20/200. After 2 months, however, the patient's vision in the left eye deteriorated to 20/80, with keratitis and inflammation.

Intravenous ceftriaxone (2 g/d) was initiated, and symptoms dramatically improved (visual acuity in the left eye, 20/50; visual acuity in the right eye, 20/25 after 4 days). On the tenth day, inflammation had regressed, but ceftriaxone therapy was discontinued because of severe thrombocytopenia. A 2-month course of treatment with oral macrolides was completed. Results of a repeat serum ELISA and immunoblot were positive (23-, 27-, 41-, 66-, 75-, and 83-kDa bands).

Because of visual deterioration during the next 6 months, vitrectomy was done on the patient's right eye. Culture of vitreous fluid was negative, but B. burgdorferi DNA was detected by PCR. The target was a 232 base-pair segment specific for B. burgdorferi [1]. Precautions used to avoid false-positive results were positive displacement pipettes and separate rooms for set up, reaction, and analysis of products. To avoid false-negative results, we used reaction mix without template and haemophilus DNA. We also ran two controls on patients with diabetic retinopathy. Products were visualized on an ethidium bromide agarose gel and were confirmed by Southern blot using an oligonucleotide probe.

Many reports have described ocular Lyme disease, some with negative vitreous cultures [2-5]. This condition may result from antibiotic therapy or a paucity of spirochetes. Specific demonstration of the organism is important, because seropositivity to B. burgdorferi may simply indicate previous or resolved infection. We propose that PCR be added to the diagnostic tests done on vitreous fluid in patients evaluated for ocular Lyme disease.