Visual Field Characteristics of Type I Boston Keratoprosthesis Patients Without Glaucoma

Purpose: To determine visual field findings in Boston type 1-KPro (BI-KPro) patients without glaucoma. Characterize normal threshold values and global indices using standard automated perimetry and characterize visual field amplitude using Goldmann’s manual perimetry. Methods: This cross-sectional prospective noninterventional study included patients (n=6 patients, 6 eyes) with BI-KPro who had normal optical coherence tomography and fundoscopic evaluation of the optic disc and retina. None had a previous history of glaucoma. Visual acuity, reliable and reproducible standard automated perimetry (24-2 and 30-2), and manual perimetry examinations were obtained from all patients. Each patient answered the National Eye Institute Visual Function Questionnaire, and the results were correlated with visual field indices. Results: The mean visual acuity was 0.35±0.31 logMAR (0.84 to 0.10). All visual fields had good reliability indices. The standard automated perimetry mean deviation values were −7.25±3.63 decibels (dB) and −7.75±3.23 (24-2 and 30-2 values, respectively), whereas pattern SD values were 2.72±0.82 and 3.30±1.13 (24-2 and 30-2, respectively). The manual visual field mean values of the 4 quadrants (superior, temporal, inferior, and nasal), were 39.7±4.5, 61.8±6.2, 54.0±4.3, and 48.2±7.6 degrees, respectively. The authors found a significant correlation between the VFQ-25 indexes of general sight and close-range activities with the values of total deviation at 10 degrees. VFQ-25 peripheral vision indexes also correlated significantly with values of total deviation at 30 degrees (outermost locations in the 30-degree area). Conclusions: Patients with BI-KPro presented reliable and reproductive visual field measurements. The authors found a consistent reduction in visual field extension and a global sensitivity reduction in these patients. Despite visual field changes, our patients had a good quality of life scores. Overall, these results could be useful to improve early glaucoma diagnosis and to follow-up BI-KPro patients.

K eratoprosthesis (KPro) is an artificial device used to replace an opaque cornea, thereby allowing light to reach the retina. It is considered an alternative when conventional penetrating keratoplasty has a poor prognosis. 1,2 Boston type I keratoprosthesis (BI-KPro) is currently the most implanted device, with appreciable anatomic and functional results, and with relatively low complication rates. [3][4][5] Between 2002 and 2014,~11,000 devices were implanted. 6 Despite positive anatomic and functional results, complications such as ocular infection (25% to 60%), 6 membrane formation (5.4% to 60%), 4,7 and glaucoma 8,9 must be continuously monitored and prevented. Crnej et al 10 reported a preoperative glaucoma prevalence of 66%, and Talajic and colleagues reported a postoperative prevalence of 26%, with a permanent visual acuity decrease in a large percentage of the patients. 11 Glaucoma is currently considered a real and continuous threat to irreversible visual loss in these patients, and its detection and adequate follow-up are usually challenging in these cases. Intraocular pressure measurements are largely imprecise, fundoscopy can be very difficult, and visual field (VF) findings can be influenced by the KPro optic properties. 12 The KPro fixed optic diameter can even limit the VF and mimic a VF defect. 13 Definitive identification and follow-up of glaucoma in KPro patients is key to sustained and adequate visual outcomes. A lack of studies characterizing the VF in these patients makes the task even more challenging. The objective of this study was therefore to determine the VF findings in BI-KPro patients without glaucoma, including the normal threshold values and global indexes using standard automated perimetry. Also, to determine VF extension using Goldmann manual perimetry (GMP). In addition, quality of life scores was correlated with VF data.

METHODS
This was a cross-sectional prospective noninterventional study conducted at the Ophthalmology and Visual Sciences Department of the Federal University of São Paulo. The study protocol was approved by the local ethics committee, adhered to the tenets of the Declaration of Helsinki, and written informed consent was signed by all patients before data collection and analyses.
We included 8 patients with reliable and reproducible VFs, visual acuity better than 20/200, and no previous history of glaucoma. Two glaucoma specialists evaluated the optic discs to exclude patients with glaucomatous neuropathy (a cup/disc ratio > 0.7; localized thinning of the neural rim, localized defects of retinal fiber layer, or papillary hemorrhage). In addition, all patients had a normal retinal fiber layer thickness evaluated by spectral domainoptical coherence tomography (Software version 5.8.3; Heidelberg Engineering, Dosseinheim, Germany).
After the initial evaluation, 2 patients were excluded because they had spectral domain-optical coherence tomography findings that suggested glaucoma or another retinopathy. The 6 remaining patients were tested with an Early Treatment Diabetic Retinopathy Study chart (Precision Vision, Woodstock, IL) to determine the best-corrected visual acuity. VF examinations were performed using strategies 24-2 and 30-2 SITA Standard using Humphrey automated perimetry (HFA) model 750i (Carl Zeiss Meditec, Dublin, CA) and GMP with stimuli V4e (Haag-Streit, Bern, Switzerland).
Only reliable VFs were included (fixation losses under 20%, false positive and negative responses under 33% for HFA). 14,15 In each quadrant, normative data were used to compare with obtained GMP values using V4e stimuli for normal and KPro participants. 16 The rationale for using the V4e stimulus for goldman visual field analyzes was to optimize the acquisition of the maximum peripheral VF extension. To avoid examiner interference, all examinations were performed by the same technician using pre-established techniques such as standardization of stimulus speed presentation and test-retest strategy. All left eye VF tests involved a transposition of values to correspond to the VF of the right eye. A validated and translated version of the National Eye Institute Visual Function Questionnaire (NEI-VFQ-25) 17

Statistics
SPSS statistical software for Windows, version 22.0 (IBM, Armonk, NY) was used for all statistical analyses, with a significance level of P < 0.05. Sample normality was not demonstrated because of the small sample size. The paired nonparametric Wilcoxon test was therefore used for perimetry comparisons. Pearson's correlation test was used to analyze VF data and quality of life indices. Four VFQ-25 outputs (general visual acuity, close-range activities, farrange activities, and peripheral vision) were selected. VF indices used were the mean reduction index (mean reduction of thresholds for each quadrant of the VF) and TD 10 degrees and TD 30 degrees.
The mean visual acuity was 0.35 ± 0.31 logMAR (0.84 to 0.10 logMAR). All VFs fields had good reliability indices and were reproducible. Figure 1 shows the mean thresholds in decibels.
Foveal thresholds and mean and pattern SD values using strategies 24-2 and 30-2 are shown in Table 1. No statistically significant differences were found between the strategies when comparing the VF indices. Concerning the visual acuity, the only significant correlation was found for the foveal threshold (r = −0.06; P = 0.04).
Regarding the manual VF, the mean values in degrees representing the 4 quadrant amplitudes (superior, temporal, inferior, and nasal) were 39.7 ± 4.5, 61.8 ± 6.2, 54.0 ± 4.3, and 48.2 ± 7.6 degrees, respectively. The measured VFs and theoretical VF extensions for the normal population are shown in Figure 2, and the mean values of the constriction of the VF are shown in degrees. This value was obtained by the difference between the mean value of the KPro's VF and the theoretical VF extension in the 4 quadrants (at 0, 90, 180, and 270 degrees).
In the VFQ-25, the mean value of the total score was 82.28 ± 7.72. Table 2 shows the mean scores for the 4 outputs related to vision symptoms and their correlations with the mean reduction index and TD 10 degrees and TD 30 degrees. We found a significant correlation between the VFQ-25 indexes of general sight and close-range activities with the values of TD 10 degrees and between the peripheral vision index and TD 30 degrees.

DISCUSSION
The prevalence of glaucoma post-KPro is a significant problem and an important cause of irreversible visual acuity  loss after the procedure. 18,19 Elevated intraocular pressure has a prevalence as high as 64%, 20 and glaucoma surgery before or concurrent with the KPro procedure is as high as 78% among these patients. 12 Monitoring and detection of glaucoma in these patients is a challenge, mainly because of the difficulty in early diagnosis because of the barrier presented by the prosthesis during standard screening. 10,20 In our patients, the VF from KPro patients had a reduction in amplitude and lower sensitivity compared with the normal population. The standard automated perimetry results showed an MD below the normal range (−7.25 dB) and a relatively lower pattern standard deviation, but still above the normal range (2.72 dB). These findings suggested a diffuse loss of sensitivity without significant localized defects. Elevated pattern standard deviation values can be indicative of glaucoma, but in patients with BI-KPro, other factors could be influencing these results.
Using GMP, we observed a mean VF extension of 91.34 degrees in our patients. In 2010, Sayegh et al, 13 studying KPro patients, reported a mean VF extension of 95 degrees, which was consistent with our results. We found a significant reduction in the VF area when comparing it with the normal population (30.78%), specifically in the temporal region (40.51%). We could not find any similar results reported in previous studies for comparison, but these findings could be related to the fixed aperture and the circular shape of the BI-KPro optical piece.
Despite the VF restriction and the reduction of sensitivity, the quality of life scores was high (mean, 82.28). This finding was consistent with a recent study from Cortina and Hallak. 21 We speculate that the positive impact of the visual acuity improvement after the BI-KPro implant could compensate the VF restriction and loss of sensitivity, influencing the VFQ-25 results. A positive correlation was found between the overview index and between TD 10 degrees, which indicates a greater influence of the macular sensitivity than the other parameters analyzed in the perception of the quality of life of these patients.
Concerning the indexes correlated to the VF, we observed that the indexes of close-range activities and peripheral vision were higher in subjects who had less reduction in the region of the central 10 and 30 degrees, respectively, in the general sight.
The present study had a few limitations. The small sample size could interfere and affect the statistical analyses, exemplified by the need to use a nonparametric test to compare the perimetry results. This population had a high rate of ocular comorbidities, usually underwent multiple surgeries, and had a high prevalence of glaucoma. Therefore, even a large BI-KPro series may not allow VF characterization in a reasonable normal (nonglaucomatous KPro patients) population assuming an elevated prevalence of glaucoma in these patients. Also, the rarity of other studies with a similar population (BI-KPro patients without glaucoma) made it difficult to compare our results with other studies. Multicenter studies might therefore be necessary to better characterize VF metrics in BI-KPro patients without glaucoma and other comorbidities.
In conclusion, our study demonstrated reliable and reproductive VF estimations in BI-KPro patients without glaucoma. We found a consistent reduction in VF and a relative global sensitivity reduction in our patients. Despite the VF alterations, our patients had a good quality of life scores using the National Eye Institute Visual Function Questionnaire. We believe this study could be useful to improve the early diagnosis of glaucoma and to follow-up glaucoma in this population.