Methods

This retrospective, noncomparative case series included 16 eyes of 14 patients treated between May 2006 and August 2011 at the Department of Ophthalmology, Taipei Veterans General Hospital. Microsporidial keratoconjunctivitis was diagnosed based on the presence of microsporidial spores by staining procedures. All patients underwent repeated corneal swabbing. Patients who underwent keratectomy with a blade to remove all the epithelium within the involved area were excluded.

For collecting specimens, a few whitish epithelial lesions were scraped from each patient after topical anesthesia using either a no. 26 needle or sterile McPherson Micro-Tying forceps (Albert Heiss, Tuttlingen, Germany) under a slit-lamp biomicroscope. This caused only small epithelium defects. The scraped epithelial specimens were smeared on slides and stained with Gram stain, modified Kinyoun acid-fast stain, or both. The remaining whitish epithelial lesions on the cornea were wiped gently with cotton swabs. This action was like erasing a chalkboard. The cotton swab was put in a 1.5 mL microcentrifuge tube (Molecular BioProducts, San Diego, California, USA) for polymerase chain reaction (PCR) analysis. DNA was extracted using the Gentra Puregene kit (Qiagen, Valencia, California, USA). Microsporidial PCR analysis using panmicrosporidian primers with targeting on the 16S rRNA was processed according to procedures described by Joseph and associates. The PCR products were cloned with a TOPO TA cloning kit (Invitrogen, Carlsbad, California, USA) and were sequenced at Mission Biotech (Taipei, Taiwan).

Topical antibiotics were prescribed 4 times daily to prevent secondary infection after corneal scraping. The topical antibiotics were tapered or discontinued according to epithelium healing condition. Each patient was followed up every day or every other day in the first week and then once or twice weekly until their epithelial lesions disappeared completely. Repeated swabbing was performed whenever the epithelial lesions were observed. If the intraocular pressure (IOP) was normal (≤ 21 mm Hg), low-potency topical corticosteroid (0.02% fluorometholone) 4 times daily was prescribed at the beginning to decrease ocular surface inflammation and to prevent subsequent development of nummular anterior stromal infiltration. The topical corticosteroid was discontinued after resolution of the keratoconjunctivitis.

Results

The Table shows the clinical characteristics, treatment details, and outcomes. Of these 14 patients, Patients 1, 2, 3, 5, 6, and 7 were reported previously. They are also included here with more detailed information of treatment. There were 8 (57%) men and 6 (43%) women with a mean age of 52.2 years, ranging from 23 to 78 years. Nine (64.3%) of these patients had a history of exposure to hot springs, 1 (7.1%) patient wore contact lenses, 2 (14.3%) patients had soil contact, and the other 2 (14.3%) patients had a history of swimming in public pools. None had a history of trauma or contact with animals. Eleven (78.6%) of these patients were misdiagnosed in another hospital: 8 (57.1%) patients as having acute conjunctivitis, 1 (7.1%) patient as having scleritis, and 2 (14.3%) patients as having herpes keratitis. Patient 6 had a long history of infection and she was treated for more than 1 month before her first visit to our department. Both Patients 6 and 12 had bilateral involvement. Patient 6 demonstrated infection in both eyes at presentation. Patient 12 initially showed infection in his left eye. Then, at follow-up 13 days later, the right eye was also involved. The follow-up periods in this study ranged from 5 days to 47 months, with median of 1 month.

Initial symptoms included red eye, pain, blurry vision, epiphora, and swelling of the eyelid. Tenderness of the preauricular lymph nodes was noted in 2 (14.3%) patients. Slit-lamp biomicroscopy disclosed mild to severe conjunctivitis in all patients. All patients manifested with typical multiple, whitish, raised epithelial lesions, varying in size and without stromal involvement. Four typical patterns of the epithelial lesions were noted: (1) smaller, faint white dot or round lesions, (2) denser white round-to-oval lesions, (3) bizarre shape (pleomorphic lesions, varying in size, and usually larger than the round-to-oval type), and (4) target-like lesions with a transparent hollow center ( Figure 1 ). After swabbing, the denser white lesions usually disappeared. Some left a punched-out epithelial defect ( Figure 2 ). The defects with a white rim resembled the target-like lesions, which always had disappeared at follow-up. The smaller faint lesions that could not be removed with swabbing became larger and more prominent. Some new faint lesions developed immediately after swabbing. Both lesions usually also had disappeared at follow-up. The noninvolved epithelium normally remained intact after a gentle swabbing. At follow-up, the lesions that reappeared were not always at the same location as previously observed ( Figure 2 ). In addition to the typical lesions, the right eye of Patient 12 also had tiny elevated epithelial lesions with a coarse punctate surface that stained poorly with fluorescein. These lesions had similar morphologic features as those of Thygeson superficial punctate keratitis.

Images showing 4 patterns of epithelial lesions of microsporidial keratoconjunctivitis: (Left) smaller, faint-white dot or round lesions (solid arrow); denser, white round-to-oval lesions (dot-dash arrow); (Right) bizarre-shaped lesions; and (Left) target-like lesions with transparent hollow center (arrowhead). (Right) The bizarre-shaped lesions also may have transparent hollow defects as the target-like lesions.

Slit-lamp photographs of microsporidial keratoconjunctivitis in Case 1. (Left) Before first swabbing; (Middle) immediately after first swabbing; and (Right) 2 days after first swabbing. (Left) Some lesions (arrowheads) were removed by corneal swabbing and left a small epithelial defect surrounded by a ring of diseased whitish epithelium, which resembled (Middle) the target-like lesions (solid arrows). (Left) Some white lesions (a dark circle) were removed and (Middle) the underlying epithelium remained intact (a dark circle). (Middle) New lesions (dot-dash arrows) may develop immediately after swabbing, but (Right) they tended to disappear at follow-up.

Results of microsporidial PCR studies on corneal scrapes obtained from 11 eyes of 10 patients revealed 250- to 280-bp amplicon. The gene sequences in all cases disclosed similarity of 99% to 100% compared with the Vittaforma corneae ( V. corneae ) 16S rRNA ( www.ncbi.nLm.nih.gov/GenBank ). The sequences of clinical specimens from the left eye of Patient 12 and from Patients 5 and 10 were identical to the GenBank sequence of V. corneae (accession number HM854229.1). The other sequences have been deposited into GenBank (accession numbers JX049113, JX049114, JX049115, and JX049116).

The mean number of corneal swabbings was 3.0 in patients with 0.25% chloramphenicol and 3.5 in patients with 0.3% norfloxacin ( Table ). All except Patient 14 were prescribed 0.02% fluorometholone. Rapid improvement in clinical signs and symptoms was observed after repeated swabbing. The resolution time, defined as the time to reach complete disappearance of epithelial lesions, was approximately 1 week (mean, 6.6 days; range, 4 to 11 days). An identical mean of 6.6 days was obtained from patients treated with norfloxacin and patients treated with chloramphenicol. No systemic antibiotic was prescribed to any of these patients. No patient lost best-corrected visual acuity or had recurrence at the last follow-up.

Elevation of IOP was noted in Patients 13 and 14, who had no history of glaucoma. Both patients showed no inflammation in the anterior chamber. Patient 13 had an IOP of 23 mm Hg in both eyes at presentation. The IOP elevated to 39 mm Hg in the left diseased eye on the third day of treatment and the IOP in the other eye was 20 mm Hg. Brimonidine tartrate ophthalmic solution 0.15% then was prescribed and fluorometholone was discontinued. The active corneal lesions subsided after treatment for 4 days and the IOP decreased to 21 mm Hg. Without using any antiglaucoma agents, the IOP remained normal during the remaining 1 month of follow-up. Patient 14 had elevated IOP to 26 mm Hg in the diseased right eye and 14 mm Hg in the left eye at the first clinic visit. According to his old chart, IOP was normal in both eyes previously. The latest measurement was 14 mm Hg in the right eye and 13 mm Hg in the left 7 months previously. We did not prescribe him any antiglaucoma agent. The high IOP persisted for 1 week and then declined to 14 mm Hg after resolution of the keratitis. Fluorometholone was not prescribed until the IOP returned to normal.

Fine keratic precipitates were observed in 5 eyes in the late stages of the disease. Nummular anterior stromal infiltrates were noted in 2 eyes. Both findings disappeared gradually with treatment with 0.02% fluorometholone.