This cross-sectional study investigates which features of a typical postfungal keratitis corneal scar are most associated with vision loss and evaluates the use of corneal imaging metrics as cornea-specific outcomes for clinical studies.
Key Points
Question
Which features of a postfungal keratitis corneal scar contribute most to vision loss?
Findings
In this ancillary cross-sectional study of a subset of 71 patients treated for fungal keratitis in the Mycotic Ulcer Treatment Trial I, irregular astigmatism and scar density were the features most strongly associated with vision loss. The thinnest point of the cornea was the metric that best discriminated between the natamycin- and voriconazole-treated ulcers.
Meaning
The findings of this study suggest that irregular astigmatism, scar density, and the thinnest point of the cornea may be meaningful cornea-specific metrics that could be used as outcomes in clinical research.
Abstract
Importance
Corneal opacity is a leading cause of visual impairment worldwide; however, the specific features of corneal scars, which decrease visual acuity, have not been well characterized.
Objective
To investigate which features of a postfungal keratitis corneal scar contribute to decreased visual acuity after an episode of infectious keratitis and evaluate whether any corneal features may be used as outcomes for clinical trials.
Design, Setting, and Participants
In this ancillary, prospective cross-sectional study, a subset of study participants treated for fungal keratitis (n = 71) as part of the Mycotic Ulcer Treatment Trial I (MUTT I) underwent best spectacle-corrected visual acuity (BSCVA) and best contact lens–corrected visual acuity examination, Scheimpflug imaging, and anterior segment optical coherence tomography at a referral hospital in India approximately 2 years after enrollment. Data were collected from December 3, 2012, to December 19, 2012, and analyses were performed from December 2, 2013, to October 2, 2019.
Main Outcomes and Measures
Linear regression models were used to evaluate the importance of various corneal features for BSCVA and to assess whether these features could be used to differentiate the 2 treatment arms of the MUTT I trial.
Results
Seventy-one patients (42 men [59.1%]; median age, 48 [range, 39-60] years) were examined at a median (IQR) time of 1.8 (1.4-2.2) years after enrollment. The mean (SD) logMAR BSCVA was 0.17 (0.19) (Snellen equivalent, 20/32). In multivariable linear regression models, BSCVA was most associated with irregular astigmatism (1.0 line of worse BSCVA per 1-line difference between BSCVA and contact lens visual acuity; 95% CI, 0.6-1.4) and corneal scar density (1.5 lines of worse vision per 10-unit increase in the mean central corneal density; 95% CI, 0.8-2.3). The thinnest point of the cornea was the metric that best discriminated between the natamycin- and voriconazole-treated ulcers in MUTT I, with 29.3 μm (95% CI, 7.1-51.6 μm) less thinning in natamycin-treated eyes.
Conclusions and Relevance
Both irregular astigmatism and corneal scar density may be important risk factors for BSCVA in a population with relatively mild, healed fungal corneal ulcers. The thinnest point of the corneal scar may be a cornea-specific outcome that could be used to evaluate treatments for corneal ulcers.
Introduction
Corneal scarring is one of the leading causes of vision impairment worldwide.1 Corneal scars are generally thought to reduce visual acuity by several mechanisms including blocking light, causing light scatter, and inducing irregular astigmatism. Although all of these mechanisms are recognized causes of vision loss in corneal disease, the relative importance of these 3 mechanisms after corneal infection has not been well characterized.2 The advent of new corneal imaging devices, including the Scheimpflug camera and anterior segment optical coherence tomography (AS-OCT), has made it possible to more rigorously evaluate the features of a corneal scar.3 The aims of the present study were to investigate which features of a typical postfungal keratitis corneal scar contribute to vision loss and to evaluate whether any corneal imaging metrics would be useful as cornea-specific outcomes for clinical studies.
Methods
This ancillary, prospective, cross-sectional study included a subset of the study participants enrolled in the Mycotic Ulcer Treatment Trial I (MUTT I), the results of which have been reported elsewhere.4 The present study was conducted at the Aravind Eye Hospital-Madurai (India) study site from MUTT I and was approved by the institutional review boards at the Aravind Eye Hospital and University of California, San Francisco. Written informed consent was obtained from the patients in their native language.
For the present study, a random sample of 71 MUTT I patients returned for further testing. We excluded 1 patient with a phthisical eye and 1 patient with a white cataract. The logMAR best spectacle-corrected visual acuity (BSCVA) and best contact lens–corrected visual acuity were measured by trained refractionists using tumbling E charts, with the same methods used in MUTT I.4 Slitlamp examination included an assessment of cataract, retinal, and optic nerve pathologic characteristics. A single physician (S.A.M.) obtained Scheimpflug camera images from each eye with the Pentacam (Oculus), using the manufacturer’s software settings for 50 images in 2 seconds. Scheimpflug camera images were obtained with a black hood covering the patient, in a room without windows, and with all lights turned off except for the light from the Pentacam computer monitor. Scans were repeated until the quality specification was reported as “OK” or until 5 scans had been attempted. The same physician obtained AS-OCT images of the corneal scar using the Spectralis system (Heidelberg Engineering) with the automatic real time setting at 30 frames.
Several topographical variables reported by the Pentacam were used in the analysis, including the maximal keratometry of the anterior surface of the cornea and the horizontal curvature radius of the central 3 mm of the cornea. In addition, the mean corneal density of the central 2-mm-diameter circle and corneal density of the annulus from the 2- to 6-mm-diameter of the central cornea were reported by the Scheimpflug camera in standardized density gray scale units ranging from 0 to 100; these measurements were combined in a 1:8 ratio to form a single measurement of corneal density in the central 6-mm-diameter of the cornea. Both AS-OCT scans were reviewed by 3 trained graders, who measured (1) the area of the scar as calculated from the manufacturer’s software from a tracing of the outer extent of the scar; (2) the depth at the deepest part of the scar, assessed as the shortest distance from the endothelium to the posterior aspect of the scar; and (3) the proportion of the central cornea with scarring, with the central cornea defined as the area of a 4-mm-diameter circle centered around the corneal light reflex. The mean value from the 3 graders was used for analysis. All corneal measurements had acceptable repeatability, as reported in a previous study.5
To assess which corneal features were most associated with visual acuity, we constructed bivariate linear regression models with BSCVA as the response variable. We constructed similar multivariable models by including all variables with P < .10 in the bivariate models and performing a backward stepwise algorithm until all variables in the multivariable model had a P < .10. We reported the squared partial correlation (pR2) of each covariate in the final model to provide some indication of its relative importance; this number indicates the proportion of variance in the outcome not explained by the other covariates in the model that is explained by the factor of interest.6 Given the statistically significant difference in the 3-month visual acuity between the natamycin and voriconazole groups found in MUTT I, we were interested in whether any of the corneal measurements would be significantly associated with the treatment allocation.4 Therefore, we constructed linear regression models with each of the corneal measurements as the response variable and treatment allocation as the explanatory variable, adjusting for enrollment visual acuity and fungal organism (Fusarium species vs non-Fusarium species). All analyses were considered hypothesis-generating, with a 2-sided significance level set at P = .05.
Results
Seventy-one patients (men, 42 [59.1%]) with a prior fungal corneal ulcer who had enrolled in the MUTT I trial were included in this ancillary cross-sectional study. The median age of all patients was 48 (range, 39-60) years. The median (IQR) time since enrollment in the trial was 1.8 (1.4-2.2) years. Among these patients, Fusarium species had caused ulcers in 28 (39.4%), Aspergillus species had caused ulcers in 10 (14.1%), and other fungal species caused ulcers in the remaining patients. The median (IQR) enrollment BSCVA was 0.50 logMAR (0.32-0.84) (Snellen equivalent 20/63, IQR 20/40-20/160) and the median (IQR) enrollment infiltrate size on slitlamp biomicroscopy was 3.0 mm (2.37-3.38 mm) measured as the geometric mean of the longest perpendicular meridians of the scar.7 As part of MUTT I, 35 study participants (49.3%) had been randomized to voriconazole and 36 (50.7%) to natamycin.
Each participant had BSCVA and best contact lens–corrected visual acuity assessed and was examined with slitlamp biomicroscopy, AS-OCT, and the Scheimpflug imaging. The AS-OCT successfully captured images for all patients, whereas the Pentacam had inadequate quality specification readings for 5 study participants (7.0%). The mean (SD) logMAR BSCVA in the population was 0.17 (0.19) (Snellen equivalent, 20/32). Hard contact lens correction improved visual acuity by a mean (SD) of 0.7 (0.8) lines compared with BSCVA, with the maximum keratometric value providing the strongest indication of potential improvement (0.7 lines of improvement for each 10 diopters [D] of maximum keratometry; 95% CI, 0.3-1.0) (eTable 1 in the Supplement). Table 1 provides summary statistics for the results of the various tests performed at the 2-year follow-up visit.
Table 1. Results of Testing in 71 Patients Treated for Fungal Keratitis in the Mycotic Ulcer Treatment Trial I.
| Test Result | Median (IQR) |
|---|---|
| Vision | |
| logMAR | |
| BSCVA | 0.16 (0 to 0.30)a |
| Best hard contact lens–corrected visual acuity | 0.10 (0 to 0.20)b |
| Difference between best contact lens–corrected visual acuity and BSCVA, lines | 0.8 (0 to 1) |
| Refraction | |
| Spherical equivalent, D | –0.50 (–1.75 to 0.50) |
| Astigmatism, D | 1.00 (0.00-2.00) |
| >2, No. (%) | 12 (16.9) |
| >3, No. (%) | 3 (4.2) |
| Scheimpflug imaging | |
| Horizontal curvature, D | 44.5 (43.1 to 46.4) |
| Maximum keratometry, D | 50.3 (47.8 to 54.5) |
| Mean density, central 6 mm | 18.8 (17.3 to 25.0) |
| AS-OCT | |
| Total area of corneal scar, mm2 | 9.3 (5.4 to 13.9) |
| Proportion of central cornea with scarring, % | 21.1 (8.2 to 50.1) |
| Thinnest cornea within scar, μm | 416 (386 to 447) |
| Deepest part of scar, μm from endothelium | 192 (137 to 256) |
| Slitlamp examination, No. (%) | |
| Cataract | |
| Grade nuclear 2 or greaterc | 10 (14.1) |
| White or brunescent | 0 |
| Retinal or optic nerve pathology | 0 |
Abbreviations: AS-OCT, anterior segment optical coherence tomography; BSCVA, best spectacle-corrected visual acuity; D, diopter; IQR, interquartile range.
Snellen equivalent: 20/32 (20/20-20/40).
Snellen equivalent: 20/25 (20/20-20/32).
Based on Lens Opacity Classification System II.8
The associations of the measured parameters with BSCVA in bivariate models are given in Table 2. A multivariable regression model was constructed that included all corneal tests with a bivariate of P < .10 as explanatory variables. After a backward stepwise selection algorithm, this multivariable model found associations of BSCVA with irregular astigmatism (1.0 lines of worse BSCVA per 1-line difference between the BSCVA and best contact lens–corrected visual acuity; 95% CI, 0.6-1.4), horizontal curvature radius (1.6 lines of worse vision per 10 D; 95% CI, 0.5-2.6), scar density (1.5 lines of worse vision per 10-unit increase in the mean central corneal density; 95% CI, 0.8-2.3), and corneal thinning (0.7 lines of worse vision per 100 μm of corneal thinning; 95% CI, 0.1-1.4) (Table 3). This model performed well in terms of describing differences in BSCVA, with an adjusted R2 of 0.59. The squared partial correlation values were highest for irregular astigmatism (pR2 = 0.30) and scar density (pR2 = 0.23), suggesting that these were the most important variables of the multivariable model. We subsequently constructed a second multivariable model that included not only the corneal tests but also a variable for cataract. Cataract was associated with a 2.3-line decrement in BSCVA (95% CI, 1.7-3.0) after the backward stepwise algorithm (Table 3). The importance of cataract for visual acuity was evidenced by several observations. First, including a term for cataract improved the ability of the model to distinguish changes in BSCVA (ie, R2 = 0.75 vs 0.59). Second, the cataract term had the highest squared partial correlation (pR2 = 0.47). Third, including cataract in the model attenuated the magnitude of association for each of the corneal tests compared with the other regression model (ie, the lines of worse vision per unit of each test were lower in the model that included a term for cataract). Sensitivity analyses restricted to participants without cataract were consistent, with similar regression coefficients (eTable 2, eTable 3, and eTable 4 in the Supplement).
Table 2. Bivariate Analyses for Factors Associated With Decreased logMAR Best Spectacle-Corrected Visual Acuity.
| Variable | Lines of Worse BSCVA per Unit of Variable (95% CI) | P Value |
|---|---|---|
| Vision, per logMAR line | ||
| Difference between best hard contact lens–corrected visual acuity and BSCVA | 1.1 (0.5 to 1.6) | <.001 |
| Scheimpflug imaginga | ||
| Horizontal curvature radius, per 10 D | 2.1 (0.5 to 3.7) | .01 |
| Maximum keratometry, per 10 D | 2.2 (1.4 to 3.0) | <.001 |
| Mean density in central 6 mm, per 10 units | 2.4 (1.6 to 3.2) | <.001 |
| AS-OCT | ||
| Total area of corneal scar, per mm2 | 0.1 (0.01 to 0.2) | .02 |
| Proportion of central cornea with scarring, per % | 1.1 (–0.6 to 2.8) | .21 |
| Thinnest cornea within scar, per 100 μm of thinning | 2.0 (1.2 to 2.7) | <.001 |
| Deepest aspect of scar, per 100 μm closer to endothelium | 1.1 (0.7 to 1.6) | <.001 |
| Slitlamp examination | ||
| Presence of grade nuclear 2 or greater cataractb | 3.0 (1.9 to 4.0) | <.001 |
Abbreviations: AS-OCT, anterior segment optical coherence tomography; BSCVA, best spectacle-corrected visual acuity; D, diopter.
Restricted to the 66 tests captured without error.
Based on Lens Opacity Classification System II.8
Table 3. Multivariable Analyses for Factors Associated With Decreased logMAR Best Spectacle-Corrected Visual Acuitya.
| Covariate | Lines of Worse Vision per Unit of Covariate (95% CI) | P Value | pR2b |
|---|---|---|---|
| Cornea risk factors only (R2 = 0.59) | |||
| Difference between best hard contact lens–corrected visual acuity and BSCVA, per logMAR line | 1.0 (0.6-1.4) | <.001 | 0.30 |
| Horizontal curvature, per 10 D | 1.6 (0.5-2.6) | .006 | 0.12 |
| Mean Scheimpflug density in central 6 mm, per 10 units | 1.5 (0.8-2.3) | <.001 | 0.23 |
| Thinnest cornea within scar, per 100 μm of thinning on AS-OCT | 0.7 (0.1-1.4) | .03 | 0.08 |
| Cornea and lens risk factors (R2 = 0.75) | |||
| Difference between best hard contact lens–corrected visual acuity and BSCVA, per logMAR line | 0.9 (0.6-1.2) | <.001 | 0.37 |
| Mean Scheimpflug density in central 6 mm, per 10 units | 1.4 (0.8-2.0) | <.001 | 0.29 |
| Thinnest cornea within scar, per 100 μm of thinning on AS-OCT | 0.5 (–0.03 to 1.0) | .06 | 0.06 |
| Presence of grade nuclear 2 or greater cataractc | 2.3 (1.7-3.0) | <.001 | 0.47 |
Abbreviations: AS-OCT, anterior segment optical coherence tomography; BSCVA, best spectacle-corrected visual acuity; D, diopter; pR2, squared partial correlation.
Multivariable linear regression models included the 66 eyes with reliable Scheimpflug camera readings.
Squared partial correlation, which can be interpreted as the amount of variance in BSCVA not explained by any of the other covariates in the model that is explained by the variable of interest.
Based on Lens Opacity Classification System II.8
Given the association between cataract and BSCVA in this population of study participants with known corneal scars, we hypothesized that visual acuity measurement may not be the optimal test to assess differences in randomized clinical trials of corneal ulcer treatments. Therefore, we constructed linear regression models with each of the corneal tests as the response variable and the treatment assignment from the underlying clinical trial (natamycin vs voriconazole) as the explanatory variable, adjusted for BSCVA at the time of enrollment and causative organism (Table 4). We found that the thinnest area of the corneal scar was the metric that best discriminated between eyes treated with natamycin and those treated with voriconazole (crude mean thickness, 424 μm; 95% CI, 409-439 μm in natamycin-treated eyes; crude mean thickness, 397 μm; 95% CI, 378-417 μm in voriconazole-treated eyes). After adjusting for ulcers caused by Fusarium species (which accounted for 17 of 35 [48.6%] voriconazole-treated ulcers and 11 of 36 [30.6%] natamycin-treated ulcers) and visual acuity at enrollment, natamycin-treated eyes were a mean of 29.3 μm (95% CI, 7.1-51.6 μm) less thin than voriconazole-treated eyes. In contrast, BSCVA was not statistically significantly different between the natamycin- and voriconazole-treated ulcers in this subpopulation of MUTT I (Table 4). These findings did not change in a sensitivity analysis restricted to the population without cataract (eTable 5 in the Supplement). To demonstrate this difference between the thinnest corneal thickness variable and BSCVA, the distributions of each variable stratified by treatment assignment were plotted (Figure).
Table 4. Association Between Measures of Corneal Health and Treatment Assignment From a Sample of 71 Study Participants From the Mycotic Ulcer Treatment Trial I.
| Response Variable | Treatment Assignment, Coefficient (95% CI)a | P Value |
|---|---|---|
| Vision, logMAR lines | ||
| BSCVA | 0.12 (–0.71 to 0.95) | .78 |
| CLVA | –0.02 (–0.78 to 0.73) | .96 |
| Difference between best hard contact lens–corrected visual acuity and BSCVA | 0.14 (–0.22 to 0.50) | .44 |
| Scheimpflug camera, Db | ||
| Horizontal curvature radius, D | 0.57 (–0.78 to 1.91) | .40 |
| Maximum keratometry front, D | –1.02 (–3.28 to 1.23) | .37 |
| Mean density in central 6 mm, units | –0.07 (–2.07 to 1.92) | .94 |
| AS-OCT | ||
| Total area of corneal scar, mm2 | –2.08 (–4.71 to 0.54) | .12 |
| Proportion of central cornea with scarring, % | –0.03 (–0.15 to 0.08) | .63 |
| Thinnest cornea within scar, μm | 29.3 (7.1 to 51.6) | .01 |
| Deepest aspect of scar, μm from endothelium | 20.2 (–17.4 to 57.8) | .29 |
Abbreviations: AS-OCT, anterior segment optical coherence tomography; BSCVA, best spectacle-corrected visual acuity; D, diopter.
Numbers represent the difference in the response variable for natamycin-treated fungal ulcers relative to voriconazole-treated ulcers in a linear regression model adjusted for enrollment BSCVA and genus of fungus (Fusarium species vs non-Fusarium species).
Restricted to the 66 tests captured without error.
Figure. Distribution of Visual Acuity and Corneal Thinness Measurements.
The kernel density plots represent the distribution of best spectacle-corrected visual acuity (BSCVA) testing (A) and assessment of the thinnest cornea within the corneal scar (B), stratified by treatment group in the Mycotic Ulcer Treatment Trial I. The density function represents the probability per logMAR line or per µm thinning,.
Discussion
This study found that irregular astigmatism, scar density, and corneal thinning were the key features of a corneal scar contributing to reduced visual acuity that could be measured using the Scheimpflug camera, AS-OCT, and contact lens overrefraction. The maximum keratometric value was the metric most strongly associated with visual acuity improvement owing to a hard contact lens, indicating that this value may be used to identify patients who would benefit from using a contact lens. The thinnest area of the corneal scar was the metric that best differentiated between the 2 MUTT I treatment arms, suggesting that this metric may be a sensitive outcome for observational studies and randomized trials.
Corneal measurements taken with a variety of tools were able to detect a change in BSCVA with high accuracy. Only a few corneal measurements were necessary to explain almost 60% of the variance in BSCVA. The most useful corneal measurement according to the squared partial correlation coefficients was the difference between best contact lens–corrected visual acuity and BSCVA. This measure should capture most of the corneal astigmatism induced by a corneal scar. The horizontal curvature radius and thinnest corneal measurement were also likely indicators of irregular astigmatism and remained relevant factors associated with decreased BSCVA even in the multivariable models. Finally, corneal scar density as measured by the Scheimpflug camera had a high partial correlation with BSCVA in the multivariable models. Increased corneal scar density may be associated with reduced vision through light scatter, although denser scars could block entry of light rays altogether.
In the present study, several factors thought to influence vision were not statistically significantly associated with BSCVA in the multivariable models, including scar depth and the proportion of central cornea with scarring. These factors may still be important if their role in BSCVA is mediated by one of the other metrics in the study (eg, irregular astigmatism). Nonetheless, the present study suggests that as metrics for predicting visual acuity, scar depth and the proportion of central cornea with scarring are less important. Of note, these results may also reflect the mild nature of residual scarring 2 years after healed fungal keratitis in this population. The inclusion criteria for MUTT I included patients whose presenting visual acuity was 20/40 to 20/400, and the patients in the present study had relatively good visual acuity (mean Snellen equivalent, 20/32). It is possible that corneal metrics not found to have a strong association with decreased visual acuity in this population may be found to have a strong association in eyes with more severe corneal scars.
Randomized clinical trials of treatments for corneal ulcers have typically used re-epithelialization, healing, or BSCVA as the primary outcome measure.4,9,10,11,12,13,14 However, each of these factors has limitations. Reepithelialization is a prerequisite for healing, but it is unclear how longer reepithelialization times are associated with the final visual acuity outcome. Healing is a subjective measure and does not capture information about the severity of the ulcer or the resultant corneal scar. Best spectacle-corrected visual acuity is an important outcome measure for any eye disease, but is not a cornea-specific outcome. Cataract had a greater association with decreased BSCVA than did corneal metrics, even in this population of patients with corneal scars. A nonspecific outcome may increase statistical noise, which will reduce the power of the study. A cornea-specific outcome, in contrast, should be a more accurate and precise measurement for a corneal disorder. Of all the corneal tests evaluated in the present study, corneal thinning was the metric that best discriminated between the natamycin- and voriconazole-treated eyes. Corneal thinning is a plausible cornea-specific outcome because better treatments for corneal infections should result in faster clearance of the infection, less stromal inflammation, and therefore less corneal thinning. Moreover, all study participants were able to undergo AS-OCT, suggesting it would be feasible to obtain measurements even in patients with severe corneal disease. In contrast, the Scheimpflug camera was unable to obtain high-quality images in approximately 8% of eyes, perhaps because of difficulties fixating in an eye with a corneal opacity. The inability to image all eyes could limit the Scheimpflug camera’s use in a clinical trial of corneal ulcers, a setting in which many patients would be expected to have poor vision. The utility of corneal thinning as a clinical trial outcome must be confirmed in studies with different populations, especially because the present study evaluated patients with mild corneal ulcers from a single center.
Limitations
This study has several limitations. Only a subset of study participants from MUTT I were invited to participate, which limited our ability to assess which of the corneal metrics could best discriminate the 2 treatment groups. Moreover, only select corneal scar features were assessed. Other factors not evaluated in the present study could also be associated with visual acuity, such as the location of the scar on the cornea. In addition, lens opacity was measured subjectively using the slitlamp, which could introduce measurement error. Pentacam has software for a more objective assessment of lens opacity, but this software was not installed on the machine used for the present study. In addition, study participants generally had good vision (mean Snellen equivalent, 20/32), so it is unclear whether these findings are generalizable to eyes with more severe corneal scars. It would be instructive to repeat the present study in a group of patients with more severe corneal scars.
Conclusions
This study found that corneal metrics of irregular astigmatism and corneal scar density appeared to be the key features most strongly associated with decreased BSCVA in eyes with corneal scars caused by fungal keratitis. The presence of cataract was also associated with BSCVA in this population, suggesting that BSCVA is a poor outcome measure for specifically measuring corneal disease. Corneal thickness at the thinnest area of the scar may be a cornea-specific outcome, but the results from this study must be confirmed in other populations with corneal scars.
eTable 1. Bivariate Predictors of Improvement in Visual Acuity With Hard Contact Lens Fitting
eTable 2. Results of Testing Done on 71 Patients Approximately 2 Years After Being Treated for Fungal Keratitis, Stratified by Cataract Status
eTable 3. Predictors of logMAR Best Spectacle Corrected Visual Acuity (BSCVA): Bivariate Analyses Among the 61 Participants Without Cataract
eTable 4. Predictors of logMAR Best Spectacle Corrected Visual Acuity (BSCVA): Multivariable Analyses in Participants Without Cataract
eTable 5. Association Between Various Measures of Corneal Health and Treatment Assignment From a Sample of 61 Study Participants Without Cataract From the Mycotic Ulcer Treatment Trial I
References
- 1.Foster A, Resnikoff S. The impact of Vision 2020 on global blindness. Eye (Lond). 2005;19(10):1133-1135. doi: 10.1038/sj.eye.6701973 [DOI] [PubMed] [Google Scholar]
- 2.Duke-Elder S, Leigh AJ. Diseases of the outer eye, part 2 In: System of Ophthalmology. Vol VIII St Louis: C.V. Mosby Co; 1965:632-645. [Google Scholar]
- 3.Otri AM, Fares U, Al-Aqaba MA, Dua HS. Corneal densitometry as an indicator of corneal health. Ophthalmology. 2012;119(3):501-508. doi: 10.1016/j.ophtha.2011.08.024 [DOI] [PubMed] [Google Scholar]
- 4.Prajna NV, Krishnan T, Mascarenhas J, et al. ; Mycotic Ulcer Treatment Trial Group . The mycotic ulcer treatment trial: a randomized trial comparing natamycin vs voriconazole. JAMA Ophthalmol. 2013;131(4):422-429. doi: 10.1001/jamaophthalmol.2013.1497 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Das M, Menda SA, Panigrahi AK, et al. Repeatability and reproducibility of slit lamp, optical coherence tomography, and Scheimpflug measurements of corneal scars. Ophthalmic Epidemiol. 2019;26(4):251-256. doi: 10.1080/09286586.2019.1600199 [DOI] [PubMed] [Google Scholar]
- 6.Cohen J, Cohen P. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. 2nd ed Hillsdale, NJ: L. Erlbaum Associates; 1983. [Google Scholar]
- 7.Diabetic Retinopathy Clinical Research Network Visual acuity score conversion chart. http://publicfiles.jaeb.org/drcrnet/Misc/VAScoreConversionChart.pdf. Accessed January 13, 2014.
- 8.Chylack LT Jr, Leske MC, McCarthy D, Khu P, Kashiwagi T, Sperduto R. Lens opacities classification system II (LOCS II). Arch Ophthalmol. 1989;107(7):991-997. doi: 10.1001/archopht.1989.01070020053028 [DOI] [PubMed] [Google Scholar]
- 9.O’Brien TP, Maguire MG, Fink NE, Alfonso E, McDonnell P. Efficacy of ofloxacin vs cefazolin and tobramycin in the therapy for bacterial keratitis: report from the Bacterial Keratitis Study Research Group. Arch Ophthalmol. 1995;113(10):1257-1265. doi: 10.1001/archopht.1995.01100100045026 [DOI] [PubMed] [Google Scholar]
- 10.Rahman MR, Johnson GJ, Husain R, Howlader SA, Minassian DC. Randomised trial of 0.2% chlorhexidine gluconate and 2.5% natamycin for fungal keratitis in Bangladesh. Br J Ophthalmol. 1998;82(8):919-925. doi: 10.1136/bjo.82.8.919 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Katz J, Khatry SK, Thapa MD, et al. A randomised trial of povidone-iodine to reduce visual impairment from corneal ulcers in rural Nepal. Br J Ophthalmol. 2004;88(12):1487-1492. doi: 10.1136/bjo.2004.044412 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Constantinou M, Daniell M, Snibson GR, Vu HT, Taylor HR. Clinical efficacy of moxifloxacin in the treatment of bacterial keratitis: a randomized clinical trial. Ophthalmology. 2007;114(9):1622-1629. doi: 10.1016/j.ophtha.2006.12.011 [DOI] [PubMed] [Google Scholar]
- 13.Prajna NV, Mascarenhas J, Krishnan T, et al. Comparison of natamycin and voriconazole for the treatment of fungal keratitis. Arch Ophthalmol. 2010;128(6):672-678. doi: 10.1001/archophthalmol.2010.102 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Srinivasan M, Mascarenhas J, Rajaraman R, et al. ; Steroids for Corneal Ulcers Trial Group . Corticosteroids for bacterial keratitis: the Steroids for Corneal Ulcers Trial (SCUT). Arch Ophthalmol. 2012;130(2):143-150. doi: 10.1001/archophthalmol.2011.315 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Bivariate Predictors of Improvement in Visual Acuity With Hard Contact Lens Fitting
eTable 2. Results of Testing Done on 71 Patients Approximately 2 Years After Being Treated for Fungal Keratitis, Stratified by Cataract Status
eTable 3. Predictors of logMAR Best Spectacle Corrected Visual Acuity (BSCVA): Bivariate Analyses Among the 61 Participants Without Cataract
eTable 4. Predictors of logMAR Best Spectacle Corrected Visual Acuity (BSCVA): Multivariable Analyses in Participants Without Cataract
eTable 5. Association Between Various Measures of Corneal Health and Treatment Assignment From a Sample of 61 Study Participants Without Cataract From the Mycotic Ulcer Treatment Trial I