To compare the clinical course of Nocardia species keratitis with keratitis resulting from other bacterial organisms and to assess the effect of corticosteroids as adjunctive therapy using data collected from the Steroids for Corneal Ulcers Trial.
Subgroup analysis of a randomized controlled trial.
setting: Multicenter randomized controlled trial. study population: Five hundred patients with bacterial keratitis randomized 1:1 to topical corticosteroid or placebo who had received at least 48 hours of topical moxifloxacin. intervention/observation procedure: Topical prednisolone phosphate 1% or placebo and clinical course of Nocardia keratitis. main outcome measures: Best spectacle-corrected visual acuity and infiltrate or scar size at 3 months from enrollment.
Of 500 patients enrolled in the trial, 55 (11%) had a Nocardia corneal ulcer. Patients with Nocardia ulcers had better presentation visual acuity compared with non- Nocardia ulcers (median Snellen visual acuity, 20/45, compared with 20/145; P < .001) and comparable 3-month visual acuity (median, 20/25, vs 20/40; P = .25). Nocardia ulcers had approximately 2 lines less of improvement in visual acuity compared with non- Nocardia ulcers (0.21 logarithm of the minimal angle of resolution; 95% confidence interval, 0.09 to 0.33 logarithm of the minimal angle of resolution; P = .001). This difference may reflect the better starting visual acuity in patients with Nocardia ulcers. In Nocardia ulcers, corticosteroids were associated with an average 0.4-mm increase in 3-month infiltrate or scar size (95% confidence interval, 0.03 to 0.77 mm; P = .03).
Nocardia ulcers responded well to treatment. They showed less overall improvement in visual acuity than non- Nocardia ulcers, but had better presentation acuity. Corticosteroids may be associated with worse outcomes.
Nocardia species are a common cause of infectious keratitis in South Asia, where the species have been found to be ubiquitous in the soil. Several large case series have shown that patients with Nocardia keratitis typically have good visual outcomes with appropriate diagnosis and therapy. However, there is little evidence regarding outcomes of Nocardia keratitis compared with keratitis resulting from other bacterial organisms, and there is little information on the usefulness of topical corticosteroids as adjunctive therapy in these cases. Corticosteroids may reduce immune-mediated tissue damage in bacterial corneal ulcers and may improve clinical outcomes; however, case reports have suggested that topical corticosteroids may cause recurrent infection in Nocardia keratitis. Additionally, most prior reports focused primarily on the use of amikacin in the treatment of Nocardia keratitis, but there are few reports on the clinical response to fluoroquinolones. The Steroids for Corneal Ulcers Trial (SCUT) was a randomized controlled trial assessing clinical outcomes in patients with bacterial corneal ulcers who received adjunctive topical corticosteroids versus placebo. Herein, we report the clinical outcomes of Nocardia keratitis cases enrolled in the trial, the susceptibility and response of these cases to fluoroquinolones, and the effect of adjunctive topical corticosteroid therapy on clinical outcomes.
The SCUT was a National Institutes of Health/National Eye Institute-funded, multicenter, randomized, placebo-controlled, double-masked clinical trial comparing clinical outcomes in patients receiving topical corticosteroid or topical placebo as adjunctive therapy in the treatment of bacterial corneal ulcers. Detailed methods for the trial have been described previously. In brief, 500 patients with culture-proven bacterial corneal ulcers were randomized to receive topical prednisolone phosphate (Bausch & Lomb Pharmaceuticals Inc, Tampa, Florida, USA) or topical placebo (0.9% NaCl and preservative, prepared by Leiter’s Pharmacy, San Jose, California, USA) after they had received at least 48 hours of topical 0.5% moxifloxacin (Vigamox; Alcon, Fort Worth, Texas, USA). The corticosteroid and placebo regimens consisted of 1 drop applied topically 4 times daily for 1 week, and then twice daily for 1 week, and then once daily for 1 week. The moxifloxacin regimen for both arms consisted of 1 drop applied topically every hour while awake for the first 48 hours, then 1 drop applied every 2 hours until re-epithelialization, and then 4 times daily until 3 weeks from enrollment. Treating physicians were allowed to change or discontinue adjunctive medications and were allowed to discontinue the study medication if they believed that it was medically necessary. Exclusion criteria included: evidence of fungus on potassium hydroxide preparation, Giemsa stain, or culture; evidence of acanthamoeba by stain; evidence of herpetic keratitis by history or examination; use of a topical corticosteroid or systemic prednisolone during the course of the present ulcer; and impending perforation. Enrollment centers included the Aravind Eye Care System (Madurai, Coimbatore, Tirunelveli), Dartmouth-Hitchcock Medical Center, and the Francis I. Proctor Foundation at the University of California, San Francisco.
The primary outcome for the trial was best spectacle-corrected visual acuity (BSCVA) at 3 months from enrollment. Secondary outcomes included infiltrate or scar size at 3 months from enrollment, time to re-epithelialization, and proportion of adverse events, including corneal perforation, in each arm. Patients were evaluated every 3 ± 1 days until re-epithelialization. Infiltrate or scar size and epithelial defect size were measured by slit-lamp examination as the longest diameter by the longest perpendicular to that diameter. BSCVA was measured at enrollment and at 3 months using an Early Treatment Diabetic Retinopathy Study Tumbling E chart. All measurements were performed by masked examiners certified for the study.
Corneal scraping was performed at presentation. Two scrapings were smeared for Gram stain and KOH wet mount. Three scrapings were inoculated onto sheep blood agar, chocolate agar, and potato dextrose agar or Sabouraud agar. The criterion for positive bacterial culture results was growth of the organism on 1 solid medium at the site of inoculation, with the exception of diphtheroids and coagulase-negative Staphylococcus , where more stringent criteria were required. Antibiotic susceptibility testing for moxifloxacin was performed using the Etest method (AB BIODISK, Solna, Sweden). Kirby-Bauer disk diffusion (Hi-media, Mumbai, India) was used to test antibiotic susceptibility to amikacin, ciprofloxacin, ofloxacin, gatifloxacin, and moxifloxacin, according to Clinical and Laboratory Standards Institute standards. Patients with mixed infections were excluded from this ancillary study.
The Nocardia strains were speciated and subtyped by polymerase chain reaction and bidirectional sequencing using methods previously described. The sequences obtained then were used to perform nucleotide–nucleotide searches using the BLASTn database at the National Center for Biotechnology Information web site ( www.ncbi.nlm.nih.gov/BLAST/ ). BLAST outputs were sorted based on maximum identity, and identifications were made when BLAST searches yielded at least 85% and for closely related species with more than 90% of query coverage.
Demographic and clinical characteristics across Nocardia species were analyzed with a Fisher exact test for categorical variables or analysis of variance for continuous variables. Comparisons of 3-month BSCVA and infiltrate or scar size in Nocardia species keratitis versus all other causative organisms were analyzed using multivariate linear regression with a dichotomous term for Nocardia species versus other causative organisms, controlling for corticosteroid treatment and baseline characteristics (BSCVA or infiltrate scar or size, respectively). The effect of corticosteroid in Nocardia species on BSCVA and infiltrate or scar size at 3 months was analyzed using a multivariate linear regression analysis controlling for whether the patient had amikacin added to their treatment regimen and baseline BSCVA and infiltrate or scar size, respectively. A sensitivity analysis was performed on the entire data set, using an interaction term for Nocardia species by treatment arm (corticosteroid or placebo). A Cox proportional hazards model controlling for baseline epithelial defect size and addition of amikacin to the treatment regimen was used to assess time to re-epithelialization. The geometric mean of the 2 measurements obtained for infiltrate or scar size and epithelial defect size was used in the analyses. The proportion of cases perforating was assessed using a Fisher exact test. All analyses were performed in Stata software version 10.0 (StataCorp, College Station, Texas, USA).
Of the 500 corneal ulcers enrolled in the trial, 56 (11%) were identified as being the result of Nocardia species. All Nocardia cases were isolated in India. One patient had a mixed infection with Nocardia and was excluded. Nocardia species were the third most commonly isolated organism in the trial overall. Of the 55 enrolled Nocardia cases, 32 (58%) reported a history of trauma in the affected eye. Of these, 11 (34%) were the result of vegetative matter injury. Nearly half (n = 26; 47%) of the patients were agricultural workers. Median age at presentation was 48 years (interquartile range [IQR], 40 to 60 years). Thirty (55%) of the patients were male. The median duration of symptoms before presentation was 10 days (IQR, 4 to 15 days).
Strain typing results were available for 52 (95%) of the 55 samples. The most common species were Nocardia cyriacigeorgica (n = 18; 35%), Nocardia pneumoniae (n = 11; 21%), and Nocardia asteroides (n = 10; 19%; Table 1 ). Ninety-eight percent of Nocardia species were susceptible to amikacin. The fluoroquinolones tested by Etest and disk diffusion had variable activity against the Nocardia species, ranging from 45% of N. pneumoniae isolates susceptible to moxifloxacin and ciprofloxacin to 100% susceptibility against Nocardia farcinica. There was no difference across Nocardia species in terms of history of trauma ( P = .88), agricultural work ( P = .43), age ( P = .38), gender ( P = .24), or duration of symptoms ( P = .50 ). N. cyriacigeorgica ulcers were associated with a vegetative matter injury ( P = .03). There were no contact lens wearers.
|Nocardia Species Strain||Total No. (%)||Moxifloxacin MIC 50 (μg/mL)||Susceptible Isolates by Disc Diffusion, No. (%)|
|Nocardia cyriacigeorgica||18 (35)||2.5||14 (78)||18 (100)||11 (65) a||13 (72)||11 (65) a|
|Nocardia pneumoniae||11 (21)||12||5 (45)||10 (91)||5 (45)||6 (55)||5 (45)|
|Nocardia asteroids||10 (19)||1||8 (80)||10 (100)||6 (60)||9 (90)||6 (60)|
|Nocardia farcinica||5 (10)||0.125||5 (100)||5 (100)||5 (100)||5 (100)||5 (100)|
|Nocardia ignorata||3 (6)||1.5||2 (66)||3 (100)||2 (67)||2 (67)||2 (67)|
|Nocardia araoensis||2 (4)||5.7||1 (50)||2 (100)||1 (50)||2 (100)||1 (50)|
|Nocardia asiatica||1 (2)||32||0 (0)||1 (100)||0 (0)||0 (0)||0 (0)|
|Nocardia blacklockiae||1 (2)||0.5||1 (100)||1 (100)||1 (100)||1 (100)||1 (100)|
|Nocardia wallacei||1 (2)||32||0 (0)||1 (100)||0 (0)||0 (0)||0 (0)|
|Total||52||2||36 (69)||51 (98)||31 (61)||38 (73)||31 (61)|
Nocardia species cases had a median duration of symptoms of 10 days (IQR 4 to 15 days) versus 4 days (IQR, 3 to 7 days) in non- Nocardia cases. Median enrollment BSCVA of Nocardia species cases was 0.34 logarithm of the minimal angle of resolution (logMAR; approximate Snellen equivalent, 20/45; IQR, 0.1 to 1.2 logMAR; Snellen IQR, 20/25 to 20/300), compared with 0.86 logMAR (approximate Snellen equivalent, 20/145; IQR, 0.42 to 1.7 logMAR; Snellen IQR, 20/50 to counting fingers; P < .001) in non- Nocardia cases. Median infiltrate or scar size at enrollment in Nocardia species cases was 2.7 mm (IQR, 1.9 to 4.1 mm) and 2.7 mm (IQR, 1.9 to 4.0 mm; P = .75) in non- Nocardia cases. Of the 55 Nocardia cases, 51 (93%) returned for their 3-month follow-up visit. At 3 months, median BSCVA in Nocardia keratitis cases was 0.12 logMAR (approximate Snellen equivalent, 20/25; IQR, 0.00 to 0.74 logMAR; Snellen IQR, 20/20 to 20/100), compared with 0.30 logMAR (approximate Snellen equivalent, 20/40; IQR, 0.06 to 0.68 logMAR; Snellen IQR, 20/22 to 20/95; P = .25) in all other organisms. Median infiltrate or scar size at 3 months was 2.7 mm (IQR, 1.6 to 3.7 mm) in Nocardia cases and 2.7 mm (IQR, 1.8 to 3.9 mm; P = .9) in non- Nocardia cases. Controlling for baseline characteristics, Nocardia species ulcers were associated with 2 lines less of improvement in 3-month BSCVA (0.21 logMAR; 95% confidence interval [CI], 0.09 to 0.33 logMAR; P = .001; Table 2 ) and 0.2 mm less improvement in infiltrate or scar size (95% CI, 0.02 to 0.4 mm; P = .03; Table 3 ) compared with other bacterial ulcers.
|Covariate||Coefficient (logMAR)||95% Confidence Interval||P Value|
|Nocardia species (vs all other organisms)||0.21||0.09 to 0.33||.001|
|Corticosteroid (vs placebo)||−0.02||−0.09 to 0.06||.64|
|Enrollment BSCVA (logMAR)||0.64||0.58 to 0.70||<.001|
|Covariate||Coefficient (mm)||95% Confidence Interval||P Value|
|Nocardia species (vs all other organisms)||0.20||0.02 to 0.39||.03|
|Corticosteroid (vs placebo)||0.06||−0.05 to 0.18||.29|
|Enrollment infiltrate/scar size (geometric mean in mm)||0.91||0.88 to 0.95||<.001|