Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry

ABSTRACT Early diagnosis of glaucoma requires evaluation of the retinal nerve fiber layer (RNFL) to pick-up subtle changes before visual field defects. Scanning laser polarimetry using the GDx VCC offers a unique tool for imaging of RNFL changes in glaucoma. It is useful for early diagnosis and also to detect progression of glaucoma. The following review covers the basic principles, interpretation, clinical utility of this technology and reviews the literature on its current applications Keywords: Scanning laser polarimetry, RNFL, GDx, Glaucoma. INTRODUCTION Even though field defects on full threshold central perimetry are considered the gold standard for the diagnosis of glaucoma, the analysis of RNFL may soon overtake the role of visual fields in the early diagnosis of glaucoma. The retinal nerve fiber layer (RNFL) assessment for glaucoma diagnosis and follow-up has several distinct advantages over current diagnostic approaches as RNFL defects occur prior to visual field loss.1-3 As many as half of all ganglion cells can be lost before a defect is detected by the visual field.4 It has also been documented that RNFL changes can occur prior to optic nerve head (ONH) changes.2,5 Also, RNFL evaluation has been found to be more sensitive for predicting future visual field loss compared to ONH evaluation, and is a better predictor of damage than C/D ratio.6-11 Red-free RNFL photography has been used to study the RNFL, but the subjective interpretation of the results and the practical problems of the method limit its usefulness. Principle of Scanning Laser Polarimetry The retinal nerve fiber layer (RNFL) is made of highly ordered parallel axon bundles which contain microtubules, cylindrical intracellular organelles with diameters smaller than the wavelength of light. The highly ordered (paralleled) structure of the microtubules is the source of RNFL birefringence which is the splitting of a light wave by a polar material into two components. These components travel at different velocities which creates a relative phase shift termed retardation. This retardation is proportional to the thickness of the RNFL.12,13 A scanning laser polarimeter is basically a confocal scanning laser ophthalmoscope with an integrated ellipsometer to measure retardation. Retinal scanning laser polarimetry (SLP) determines the RNFL thickness, point by point in the peripapillary region, by measuring the total retardation in the light reflected from the retina. Polarized light passes through the eye and is reflected off the retina.14-17 Because the RNFL is birefringent, the two components of the polarized light are phase shifted relative to each other (Fig. 1) and this is captured by a detector, and converted into thickness (in microns).12 Fig. 1: Two orthogonal components of polarized light pass through the RNFL (a birefringent medium) and one component is retarded proportionally to the RNFL thickness Anterior Segment Birefringence In addition to the RNFL, the anterior segment (the cornea and lens) is birefringent. The total retardation of a subject's eye is the sum of the cornea, lens and RNFL birefringence. Compensation of anterior segment birefringence is necessary to isolate RNFL birefringence. Early scanning laser polarimeters (e.g. the GDx NFA and the GDx access) compensated for anterior segment birefringence based on fixed values for the axis and magnitude of the anterior segment birefringence. This, however, varies for each individual. Journal of Current Glaucoma Practice Ajay Sharma et al Variable Corneal Compensation The GDx variable corneal compensation (VCC) measures and individually compensates for anterior segment birefringence for each eye (Fig. 2).16 For this, the specific axis and magnitude of the anterior segment birefringence is determined by first imaging the eye without compensation. The uncompensated image presents total retardation from the eye and includes retardation from the cornea, lens and RNFL. The macular region of this image is then analyzed to determine the axis and magnitude of the anterior segment birefringence. The macular region birefringence is uniform and symmetric due to the radial distribution of Henle's fiber layer. However, in uncompensated scans, a non-uniform retardation pattern is present in the macula due to the birefringence from the anterior segment. The axis and magnitude values from the anterior segment can be computed by analyzing the non-uniform retardation profile around the macula. The axis of the anterior segment birefringence is determined by the orientation of the 'bow-tie' birefringent pattern (Fig. 3) in the macula and the magnitude of the anterior segment birefringence is calculated by analyzing the circular profile of the birefringence in the macula according to standard equations.12 In cases of macular pathology, an alternative method is available that accurately compensates for the anterior segment birefringence.18 A Comparison of VCC Technology with FCC Technology If the anterior segment birefringence values for a given eye deviate from the assumed values of the fixed compensator, the FCC image will be less comparable. As the VCC individually measures and compensates for the anterior segment birefringence for each eye, discrepancies between scan modes (VCC vs FCC) are the result of incorrect FCC compensation. The VCC scan therefore results in a more accurate RNFL measurement, and is now universally accepted as the standard measurement strategy. Fig. 2: Scanning laser polarimeter (GDx VCC from Laser Diagnostics Inc.) Fig. 3: "Bow-tie" pattern (arrow) seen in the macula in an uncompensated scan RNFL Measurements The GDx VCC measurements are taken by scanning the beam of a near-infrared laser (780 nm) in a raster pattern13 which captures an image with a field 40° horizontally by 20° vertically, and including both the peripapillary and the macular region.19 Total scan time is 0.8 seconds. For each measurement, the GDx VCC generates two images: A reflectance image and a retardation image (Figs 4A and B). The reflectance image is generated from the light reflected directly back from the surface of the retina, and is displayed as the fundus image on the device screen and printouts. The retardation image is the map of retardation values and is converted into RNFL thickness based on a conversion factor of 0.67 nm/μm. Each image is made up of 256 (horizontal) ×128 (vertical) pixels, or 32,768 total pixels. For an emmetropic eye, 1 pixel is 0.0465 mm in size, and the total scan field is 11.9 mm (horizontal) ×5.9 mm (vertical).20 JAYPEE Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry Figs 4A and B: Images generated by the GDx VCC: (A) The reflectance image, which is displayed as a colored intensity map (greater reflectance corresponds to a lighter color). (B) The retardation map converted to RNFL thickness. The RNFL thickness is color-coded based on the color spectrum with thinner regions displayed in blue and green and thicker regions displayed in yellow and red Measurement Technique Measurement is performed with an undilated pupil of at least 2 mm diameter and takes only about a second to capture the image. Total time for the examination and output is less than 3 minutes for both eyes. The test is totally objective and the reproducibility of images is 5 to 8 micron per measured pixel. A warning is given if image fails to meet requisite criteria. The quality of image is affected by cataracts and poor media clarity. Looking at image allows one to see if the ellipse was placed properly and the ellipse can be manually aligned to conform the disk margin. The diameter of the ellipse is displayed in microns and gives an idea about the actual disk diameter. Clinical Interpretation of the GDx VCC Printout For each GDx VCC scan, an age-matched comparison is made to the normative database and any significant deviations from normal limits are flagged as abnormal with a p-value. Quantitative RNFL evaluation is provided through four key elements of the printout (Fig. 5): Thickness map Deviation map TSNIT graph Parameter table. The Thickness Map The thickness map shows the RNFL thickness using a color scale that follows the color spectrum going from blue to red. Thick RNFL values are colored yellow, orange and red while thin RNFL values are colored dark blue, light blue and green. The color scale follows the color spectrum (blue to red) up to 120 microns. Each quadrant is analyzed and actual deviation from normal, in microns, is displayed. Deviations from normal are highlighted in yellow if they are borderline (p < 0.10) or in red if they are outside normal limits (p < 0.05). The normal pattern is a symmetrical hourglass shape of bright colors superior and inferior and dark colors nasal and temporal. An Abnormal pattern may include any/all of the following: Diffuse loss of RNFL Focal defects are seen as concentrated dark areas (visible on fundus image as well) Asymmetry between superior and inferior quadrants Asymmetry between the two eyes Higher than normal nasal and temporal thickness. The Deviation Map The deviation map reveals the location and magnitude of RNFL defects over the entire thickness map. The deviation map analyzes a 128 ×128 pixel region (20° ×20°) centered on the optic disk. To reduce variability due to slight anatomical deviations between individuals, the 128 ×128 pixel thickness map is averaged into a 32 ×32 square grid, where each square is the average of a 4 ×4 pixel region (called super pixels). For each scan, the RNFL thickness at each super pixel is compared to the age-matched normative database, and the super pixels that fall below the normal range are flagged by colored squares based on the probability of normality. Dark blue squares represent areas where the RNFL thickness is below the 5th percentile of the normative database, i.e. there is only 5% probability that the RNFL thickness in this area is within the normal range. Light blue squares represent deviation below the 2% level, yellow represents deviation below 1%, and red represents deviation below 0.5%. The deviation map uses a grayscale fundus image of the eye as a background, and displays abnormal grid values as colored squares over this image (Fig. 6). The TSNIT Map The TSNIT stands for temporal-superior-nasal-inferiortemporal and displays the RNFL thickness values along the calculation circle starting temporally and moving superiorly, nasally, inferiorly and ending temporally. In a normal eye, the TSNIT plot follows the typical 'double hump' pattern with thick RNFL measures superiorly and inferiorly and thin RNFL values nasally and temporally. The TSNIT graph shows the curve (or function) of the actual values for that eye along with a shaded area which represents the 95% normal range for that age. In a healthy eye, the TSNIT curve will fall within the shaded area. When there is RNFL loss, the TSNIT curve will fall below this shaded area, especially in the superior and inferior regions. In the center of the printout at the bottom, the TSNIT graphs for both eyes are displayed together. In a healthy eye, there is good symmetry between the TSNIT graphs of the two eyes and the two curves will overlap. However, in glaucoma, one eye often has more advanced RNFL loss and, therefore, the two curves will have less overlap. A dip in the curve of one eye relative to another is indicative of RNFL loss (Fig. 7). Journal of Current Glaucoma Practice Ajay Sharma et al Fig. 5: Various parameters for quantitative RNFL evaluation JAYPEE Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry Fig. 6: Deviation map The Parameter Table The TSNIT parameters are summary measures based on RNFL thickness values within the calculation circle. The calculation circle is a fixed circle (a fixed size band) centered on the optic nerve head (ONH) which is 0.4 mm wide with outer and inner diameters of 3.2 and 2.4 mm respectively (see Fig. 7). These parameters are automatically compared to the normative database and are quantified in terms of probability of normality. Normal parameter values are displayed in white, abnormal values are color-coded based on their probability of normality. The probability levels used are the same as the deviation map: Dark blue represents 5% likelihood of being normal, light blue represents 2% level, yellow 1% and red 0.5%. The five TSNIT parameters are: TSNIT average, superior average, inferior average, TSNIT standard deviation (TSNIT SD) and intereye symmetry. TSNIT average: The average RNFL thickness around the entire calculation circle. Superior average: The average RNFL thickness in the superior 120° region of the calculation circle. Inferior average: The average RNFL thickness in the inferior 120° region of the calculation circle. TSNIT SD: This measure captures the modulation (peak to trough difference) of the double-hump pattern. A normal eye will have high modulation in the double-hump RNFL pattern, while a glaucoma eye will typically have low modulation in the double-hump pattern (Fig. 8). Intereye symmetry: Measures the degree of symmetry between the right and left eyes by correlating the TSNIT functions from the two eyes. Values range from -1 to 1, where values near one represent good symmetry. Normal eyes have good symmetry with values around 0.9. The nerve fiber indicator (NFI): The NFI is a global measure based on the entire RNFL thickness map and is calculated using an advanced form of neural network, called a support vector machine (SVM). It utilizes information from the entire RNFL thickness map to optimize the discrimination between healthy and glaucomatous eyes. The output of the NFI is a single value that ranges from 1 to 100 indicating the overall integrity of theRNFL with classification based on the ranges: 1 to 30 as normal, 31 to 50 as borderline and 51+ as abnormal. Fig. 7: Calculation circle, characteristic double hump pattern Fig. 8: Deviation map Clinical research has shown that the NFI is the best parameter for discriminating normal from glaucoma22 with sensitivity and specificity of the NFI reported to be as high as 89% and 98% respectively. Abnormal Scan Although there is no consensus on definition of an abnormal scan, the following guidelines can be used (see Figs 5 to 7). TSNIT average, superior average, inferior average, TSNIT standard deviation, intereye symmetry or NFI are abnormal at p < 1% level. They are considered borderline at p < 5% level (in general if NFI is > 47 at the p < 1% level or >30 at p < 5% level, the scan is abnormal). The normal values of the GDx VCC parameters in the Indian population (40-70 years) according to our database of 200 subjects (40-60 years) is as follows : TSNIT average = 54.8 ± 4.1 (45.6-66.8) microns Superior average = 66.8 ± 6.7 (55.1-85) microns Inferior average = 62.1 ± 6.6 (38.9-74.3) microns NFI = 17.2 ± 6.9 (4-35) Additional Diagnostic Parameters For an extended analysis, the following parameters are also available on this machine: Symmetry-superior quadrant thickness/inferior quadrant thickness Superior ratio-superior quadrant thickness/temporal quadrant thick Inferior ratio-inferior quadrant thickness/temporal quadrant thickness Maximum modulation-thickest quadrant/thinnest quadrant within image Ellipse modulation-thickest quadrant/thinnest quadrant within ellipse. In eyes with advanced chorioretinal degeneration or peripapillary atrophy the GDx VCC image may show very high retardation values (supranormal) with a pink color depicting a thickness > 140 microns. This occurs due to additional birefringence from the sclera and such abnormal scans should not be used for interpretation of the RNFL. In eyes with peripapillary atrophy, the default scan diameter can be manually increased to fall outside the atrophic area around the disk. However, the normative database comparisons are affected, if the calculation circle is reset. Detecting Progression of RNFL Loss: Serial Analysis The serial analysis printout has five key elements that should be considered when assessing RNFL change over time (Fig. 8): Thickness maps, deviation maps, deviation from reference maps, parameters tables and TSNIT graph. A change probability map has also been added in the new software. The serial analysis can compare up to four exams. The first exam is the baseline or reference exam, and all follow-up exams are compared to this. A colored rectangle to the left of the thickness map contains the date and quality score of each exam. The same color is used in the TSNIT graph to indicate which TSNIT curve corresponds to which exam (see Fig. 8). The deviation from reference map displays the RNFL difference of the follow-up exam compared to the baseline. If the difference exceeds 20 microns at any pixel, it is color coded. The areas of RNFL change shown on the deviation from reference map frequently, but not always correspond to the areas of loss detected by the deviation map, because the deviation map shows loss compared to the normative database while the deviation from reference maps shows RNFL change over time in the same eye. The TSNIT Graph shows the TSNIT curves for all exams, its color corresponding to the color of the vertical rectangle next to each exam. The TSNIT curves are overlaid on the shaded area representing the normal range for that age. RNFL loss results in a lower TSNIT curve on the follow-up exam compared to baseline (Figs 8 and 9). JAYPEE Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry Fig. 9: GDx VCC scan in advanced glaucoma with diffuse RNFL loss Journal of Current Glaucoma Practice Ajay Sharma et al Thus, progression of the RNFL over a period provides key data regarding: Identification of RNFL defect Rate of progression of RNFL Assessment of treatment effectiveness. The new software which helps in investigating the progression is GPATM (progression analysis for GDx) (Figs 9 and 10). Guided progression analysis (GPA) compares measurements over time and determines, if the differences are statistically significant. GDx GPA reports "possible progression" when significant change is detected and "likely progression" when significant change is confirmed. Possible progression requires a minimum of three visits, and likely progression requires a minimum of four. The GDx GPA algorithms are designed to have 95% specificity for likely progression. This means theoretically that GPA will correctly identify 95% of stable eyes as not changing. At this time, there is no quantitative clinical data on the sensitivity of GPA. Progression analysis has two modes: Fast and extended. Fast mode is for analyzing data sets that include single measurements. It compares change to the predetermined average measurement variability derived from a sample population. In contrast, extended mode requires means of three measurements, and GPA calculates the individual measurement variability of each eye for a selected patient. It measures and detects the progression based on three different parts of the analysis: Image change map TSNIT change graph Summary of parameter charts. Image Change Map Image change map recognizes the change in the reflectance image. The minimal cluster size considered is 150 pixels which is 2% of image area. Any significant change in the image is depicted on the progression map. "Possible progression" areas are shown in yellow, "likely progression" areas in red, and "possible increase" areas in purple. It can detect narrower and deeper defects. This design has specificity of 95%. TSNIT Progression Graph The ring around the optic nerve is divided into 64 equal segments and compared on follow-up. If three adjacent segments show significant change on follow-up, the progression is indicated. Areas between the current baseline set and the current exam that report significant change are displayed with likely progression shown in red, possible progression shown in yellow, and possible increase shown in purple. Fig. 10: Normal and glaucomatous GDx JAYPEE Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry Fig. 11: Serial analysis to detect progression Parameter Progression Chart TSNIT average, superior average and inferior average are compared. On the chart regression line is drawn to show likely progression and p < 5% (Fig. 11). This design also has 95% specificity. This can detect diffuse changes in the RNFL better. This parameter can also compare the rate of progression before and after treatment, thus helpful in guiding the treatment line. GDx GPA uses two different algorithms to determine significant change, based on GPA mode. CFB (Change from baseline): Based on changes from two baseline exams compared to measurement variability. It is most sensitive when there is little variability between baselines. Mean readings are treated as single data points. SIM (statistical image mapping): Based on trend analysis. All visits contribute to change detection, as opposed to CFB in which the data from the first two and last two visits are used to determine, if change occurred. Therefore, SIM is able to detect progression between the first two visits better than CFB. Journal of Current Glaucoma Practice Ajay Sharma et al Advantages of GDx VCC Easy to operate Does not require pupillary dilatation Good reproducibility Does not require a reference plane Can detect glaucoma on the first exam Early detection before standard visual field Comparison with age-matched normative database It is independent of the optical resolution of the human eye. Limitations Does not measure actual RNFL thickness (inferred value) Measures RNFL at different locations for each patient Does not differentiate true biological change from variability Limited use in moderate/advanced glaucoma Requires a wider database from the Indian population Fourth machine prototype (cannot update earlier versions) Affected by anterior and posterior segment pathology like: Ocular surface disorders, macular pathology, cataract and refractive surgery, refractive errors (false positive in myopes), peripapillary atrophy (scleral birefringence interferes with RNFL measurement), etc. PRACTICAL TIPS Verify image quality. In case image is of poor quality, a flag is displayed at the top of the page. Discard images with poor registration, Q < 7, or TSS < 40 whenever possible or interpret with caution. Review the summary box. A possible progression flag indicates additional follow-up visits are recommended to confirm change. A likely progression flag indicates statistically significant change is detected in GDx measurements. A possible increase flag could indicate high measurement variability, especially when increase and progression are flagged simultaneously. Correlate GDx results with other clinical tests to detect glaucomatous progression. Rate of progression, locations of the detected progression, age of the patient, stage of the disease and other clinical factors should be considered before taking a clinical decision. Instrument or calibration change is indicated in the summary parameter charts by a blue asterisk at the top of the charts where a GDx instrument has changed. Typical scan score (TSS) provides a measure of the "typicality" of the RNFL image. In an atypical scan, the retardance profile does not match the known anatomical RNFL distribution and can be characterized by a variable retardance pattern. Atypical scans are more common in pale fundi, high myopes and elderly eyes. TSS ranges from 0 (very atypical) to 100 (very typical). Exams with TSS < 40 should be interpreted with caution. CLINICAL STUDIES USING GDx VCC TECHNOLOGY The diagnostic accuracy of the GDx VCC for identification of eyes with glaucoma has also been shown to be quite good in various studies.15-17 Weinreb et al15 found that significantly higher sensitivity and specificity with GDx VCC compared to the GDx FCC. RNFL thickness measures from the GDx VCC also have an improved correlation with visual fields.18-20 In an animal model with the lens and cornea removed, Weinreb showed that the retardation is linearly related to the thickness of the RNFL with excellent correlation (r = 0.83) between retardation and the histopathologic measurement of RNFL thickness. The resolution of measurements in vitro was estimated to be 13 m.12 Reus et al determined the diagnostic accuracy of the GDx VCC in the diagnosis of glaucoma in a prospective case series.21 and found that NFI was the best discriminating parameter with a sensitivity and specificity of 89.0 and 95.9% respectively. At the cut-off level of > 40, the sensitivities of the NFI for correctly identifying glaucoma patients with mild, moderate and severe damage were 83.8, 92.9 and 90.1% respectively. Reus et al also compared scanning laser polarimetry (SLP) measurements of retinal nerve fiber layer (RNFL) thickness in perimetrically unaffected eyes of glaucoma patients with those in their fellow eyes with field loss and eyes of healthy subjects.22 They found that GDx VCC measurements showed more RNFL thinning in the perimetrically unaffected eyes of glaucoma patients than in the healthy control eyes. The RNFL in the perimetrically unaffected eyes of glaucoma patients was thicker than that in their fellow eyes with field loss. Henderson et al studied the relationship between central corneal thickness and retinal nerve fiber layer thickness in ocular hypertensive patients (OHP)23 and found that ocular hypertension patients with thinner corneas had significantly thinner RNFL values than OHT patients with thicker corneas and healthy control subjects. RNFL defects as assessed by the GDx VCC may therefore represent early glaucomatous damage in OHT eyes. Reus et al24 found a statistically significant correlation in most sectors between standard automated perimetry and GDx VCC measurements in patients with glaucoma. Based on the observed relationships between function and structure, the authors concluded that patients with mild to moderate visual field loss in glaucoma may be better monitored with the GDx VCC and patients who have severe loss with perimetry. Medeiros et al25 compared the ability of scanning laser polarimetry with variable corneal compensation (GDx VCC), confocal scanning laser ophthalmoscopy [HRT II (Heidelberg Retina Tomograph)] and optical coherence tomography (Stratus OCT) to discriminate between healthy eyes and eyes with glaucomatous visual field loss. No statistically significant difference was found between the areas under the receiver operating characteristic curves (AUCs) for the best parameters from the GDx VCC (nerve fiber indicator, AUC = 0.91), Stratus OCT (retinal nerve fiber layer inferior thickness, AUC = 0.92), and HRT II (linear discriminant function, AUC = 0.86). Abnormal results for each of the instruments, after comparison with their normative databases, were associated with strong positive likelihood ratios. -JAYPEE Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry The AUCs and the sensitivities at high specificities were similar among the best parameters from each instrument. Abnormal results (as compared with each instrument's normative database) were associated with high likelihood ratios and large effects on post-test probabilities of having glaucomatous visual field loss. The authors concluded that calculation of likelihood ratios may provide additional information to assist the clinician in diagnosing glaucoma with these instruments. Aung et al evaluated the changes in retinal nerve fiber layer (RNFL) thickness in the first 16 weeks after acute primary angle closure (APAC)26 and found that after an episode of APAC, superior and inferior average RNFL thickness decreases significantly from week 2 to 16. We evaluated the retinal nerve fiber layer (RNFL) thickness parameters with optical coherence tomography (OCT) using four scan diameters to study the effect of radius of measurement on the RNFL values and then correlated with scanning laser polarimetry (GDx VCC) in 74 eyes.27 The three measured parameters (superior RNFL, inferior RNFL and average RNFL thickness) showed a significant positive correlation when measurements of OCT3 and GDx VCC were compared. Highest degree of correlation for all 3 parameters was observed with a circular scan radius of 1.73 on OCT3 and Gdx VCC (superior r = 0.74, p < 0.001; inferior r = 0.54, p < 0.001; average r = 0.53, p < 0.001). The following regression equation was obtained: OCT fast RNFL thickness = 56.17 + 0.613 GDx TSNIT thickness All the RNFL thickness parameters on OCT showed a decrease in magnitude with an increase in the size of scan radius. Since scan diameters are fixed for both instruments irrespective of the disk diameter, measurements closer to the disk margin in large sized disks will give higher thickness values as compared to measurements at the same radius (but further away from disk margin) in small sized disks. This is one important drawback of the current imaging technologies and normative databases for RNFL thickness on OCT and GDx VCC should take the disk size into account. Parravano et al reported that average peripapillary RNFL thickness was reduced in patients of diabetes mellitus type 1. Matrix MD, HFA MD, PSD, average peripapillary and superior retinal nerve fiber layer (RNFL) were significantly reduced in patients with HbA1c > 7% compared to controls.28 They concluded that functional and structural retinal testing by Humphrey-Matrix and GDx VCC could be useful for the identification of early retinal impairment in DM1 patients with no sign of retinal vasculopathy. Martinez et al compared scanning laser polarimetry measurements of RNFL thickness in eyes of migraine patients with those in eyes of age-matched, healthy subject, and reported that the mean RNFL average thickness parameter was found to be thinner in migraine patients.29 In addition, there was a strong correlation between migraine severity and RNFL average thickness parameters. Grabska-Liberek evaluated the applicability of selected methods in glaucoma diagnosis in a patient with optic disk drusen.30 The scanning laser polarymetry showed extensive losses in nerve fiber layer of retina and the retinal thickness analysis showed a reduction of the retina thickness in the posterior pole. Jankowska-Lech et al reported that evaluation with scanning polarymetry laser might be precious method in discovering retinal nerves fiber layer damage in the course of multiple sclerosis.31 Presence of defects in retinal nerves fiber layer in patients suffering from multiple sclerosis with no history of retrobulbar neuritis may suggest subclinical damage of optic nerve. Zaveri et al also concluded that scanning laser polarimetry with variable corneal compensation measurements of RNFL thickness corroborates OCT evidence of visual pathway axonal loss in MS and provides new insight into structural aspects of axonal loss that relate to RNFL birefringence (microtubule integrity).32 These results support validity for RNFL thickness as a marker for axonal degeneration and support use of these techniques in clinical trials that examine neuroprotective and other disease-modifying therapies. Garcia-Medina concluded that serial analyses with GDx VCC may be used as objective and quantitative tests to assess the progression of chorioretinal dystrophies like choroideremia.33 The groningen longitudinal glaucoma study II. A prospective comparison of frequency doubling perimetry, the GDx nerve fiber analyzer and standard automated perimetry in glaucoma suspect patients concluded that the most frequent finding after a 4-year follow-up of a cohort of glaucoma suspects was conversion on GDx.34 Hlavakova et al found a statistically significant decrease of RNFL thickness after LASIK in every single quadrant (Fig. 12). Clinically, the differences in RNFL thickness before and after LASIK were minimal.35 They proposed that the measurements by means of GDX are influenced by changes in the polarization features of the cornea caused by LASIK procedure. Iester et al reported that the VCC algorithm is able to compensate for most of the changes in corneal birefringence induced by corneal refractive surgery if the polarization has been recalculated.36 Because mild changes in GDx parameters could affect the interpretation of the results in some patients, a new postoperative baseline macular image should be acquired. Arraes et al reported that moderate degrees of PCO and/or acceptable images in pseudophakic patients do not alter the analysis of nerve fiber layer by GDx.37 Only intense degrees of PCO that hinder analyzable images make the examination impracticable. Chen et al found that there was no significant difference between the HT-POAG and PACG eyes as far as the various parameters were concerned.38 GDx VCC shows fair discriminating ability in distinguishing normal from POAG and PACG eyes in Taiwan Chinese population. Journal of Current Glaucoma Practice Ajay Sharma et al Fig. 12: Stable and progressive RNFL loss GDx -Enhanced Corneal Compensation (GDx-ECC) Scanning laser polarimetry measures the strength of the retinal birefringence measurement relative to optical and digital noise. Its sensitivity can be enhanced using a software algorithm (ECC) which measures the birefringence of the cornea and retina concurrently, as opposed to canceling out the corneal measurement with variable corneal compensation (VCC). This alternate method results in high-quality scans of all subjects. A baseline image, which consists of the mean of three scans, is analyzed. The computerized export of the temporal-superiornasal-inferior-temporal (TSNIT) plots on the GDx-ECC printout includes the mean RNFL thickness from 64 polar sectors (5.625°/arc). The mean for each of these sectors is computed along a 3.2 mm diameter measurement circle surrounding the optic nerve head. The mean RNFL thickness for the superior (0-180°) and inferior (181-360°)retinal region is computed separately by averaging the corresponding mean sectors. Retinal nerve fiber layer images obtained using enhanced corneal compensation show a stronger structure-function relationship with standard automated perimetry, thereby demonstrating a highers visual field sensitivity compared with variable corneal compensation. Madieros et al concluded that GDx-ECC performed significantly better than GDx VCC in glaucoma detection in patients with more severe atypical retardation patterns. For lower values of TSS and lower AGIS scores, GDx-ECC performed significantly better than GDx VCC and at earlier stages of disease.39 Mai et al reported that RNFL measurements by SLP ECC had, in general, a good measurement repeatability, although some parameters seemed to be less stable in glaucomatous eyes than in healthy eyes and eyes with OHT. SLP ECC may therefore probably be employed for the detection of glaucomatous progression.40 They also reported that structure-function relationship between RNFL retardation and SAP VF sensitivity was stronger in images obtained with the GDx-ECC than with the GDx VCC. ABPs, which appeared more markedly with VCC than with ECC, weakened the structure-function relationship. When eyes with marked ABP images were removed from the analysis, the structure-function relationship with VCC improved, and no statistically significantly differences were found in the relationships between VCC and ECC.41 Morishita et al compared the results of scanning laser polarimetry (GDx) with variable corneal compensation (VCC) and enhanced corneal compensation (ECC) when applied to myopic glaucomatous eyes. They reported that mean typical scan score is significantly lower (p < 0.0001) and the prevalence of atypical retardance pattern is significantly higher (p < 0.0001) by VCC scans than by ECC scans. TSNIT average and temporal average thickness show significantly higher values (p < 0.001) by VCC than by ECC. A statistically significant association was observed between TSNIT average and mean deviation of SAP by ECC scan. They therefore concluded that ECC scans showed a better retardation pattern and structure-function relationship than did VCC, and ECC appears to be more suitable for RNFL assessment in glaucomatous eyes that are moderately to highly myopic.42 JAYPEE Evaluation of Retinal Nerve Fiber Layer using Scanning Laser Polarimetry Toth et al found that the intervisit standard deviation, ISD of GDx-ECC NFI but not GDx VCC NFI, was significantly higher in progression than in the stable glaucoma group.43 Also, several other ISD values tended to increase in the progressing group. Inferior average, and average thickness along the measuring ellipse (OR = 2.00, p = 0.042), as determined with GDx-ECC (but not with GDx VCC), were associated with visual field progression, independently of patient age. They concluded that with GDx-ECC, increase of ISD is an early sign of glaucoma progression, precedes the development of detectable parameter changes and is associated with visual field progression. CONCLUSIONS The use of GDx VCC for RNFL assessment in glaucoma enables the clinician to pick-up preperimetric glaucoma and provides objective and quantitative information of the RNFL that is highly reproducible. It can discriminate normal from glaucoma with a high degree of accuracy. The procedure is easy to perform does not need pupillary dilatation, and clinical interpretation of the results is simple and direct. The quantitative RNFL assessment aids the clinician in the diagnosis and management of glaucoma, and should be used in conjunction with other diagnostic information when making clinical decisions. Treatment should not be started based on GDx VCC parameters alone and the results of other anatomical and functional investigations must be taken into account. An abnormality on the GDx VCC implies that the patients require a closer follow-up to detect progression and confirmation of glaucomatous damage. Nonglaucomatous causes for optic neuropathy must be ruled out by a thorough clinical examination and appropriate investigation. Further long-term studies are required before the GDx VCC technology becomes accepted as the gold standard for making a diagnosis of glaucomatous optic neuropathy and detecting progression.

Perfluoron and Silicone Oil for Complex Retinal Detachments

The development of perfluorocarbon liquids for use in vitreoretinal surgery by Stanley Chang, MD, revolutionized the treatment of complex retinal detachments.¹ Perfluoron (perfluoro-n-octane; Alcon), transparent and heavier than water, facilitates intraoperative retinal attachment in a supine patient. Perfluoron's low viscosity makes it easy to inject into the eye and to manipulate during surgery.² Its high boiling point allows the use of endolaser therapy through Perfluoron during surgery.³ Since laser photocoagulation does not form an immediate retinochoroidal adhesion, and since Perfluoron cannot be safely left in the eye postoperatively, gas or silicone oil must be exchanged for the Perfluoron once the retina is attached.^4,5
Silicone oil or gas can safely remain in the eye to support the retina while a laser adhesion matures. Silicone oil is a useful postoperative tamponade in eyes of patients who cannot position postoperatively (such as children), eyes of patients who need to fly soon after surgery, and commonly in eyes with proliferative vitreoretinopathy.⁶ Five thousand-centistoke silicone oil may resist emulsification better than less-viscous 1,000-centistoke silicone oil and may therefore be preferable in eyes requiring long-term tamponade.^7,8

TWO APPROACHES TO REATTACHMENT

There are two avenues for successfully progressing from a situation where the retina is attached under Perfluoron to a situation where the retina is attached under silicone oil. First, Perfluoron can be exchanged for air, and then the airfilled eye can be filled with silicone oil. Second, Perfluoron can be exchanged for silicone oil directly.⁹ In either case, the eye is initially filled not with Perfluoron alone, but rather with a combination of Perfluoron and balanced salt solution (BSS). If either the Perfluoron or the BSS migrates under the retina during the exchange, the retina will redetach.
Before these two techniques of Perfluoron-silicone oil exchange are discussed, the use of Perfluoron and the use of silicone oil will be briefly reviewed.

PERFLUORON OVERVIEW

Perfluoron displaces subretinal fluid anteriorly. Therefore, during Perfluoron infusion, the eye should be rolled so that any retinal defect, such as a retinotomy or retinal tear, is as anterior as possible. Then, as the Perfluoron is injected into the vitreous cavity, the retinal defect remains open, allowing egress of displaced subretinal fluid (Figure 1). If Perfluoron occludes the retinal defect before all the posterior subretinal fluid drains, a pocket of subretinal fluid will be trapped posteriorly, preventing complete intraoperative retinal attachment. That trapped subretinal fluid makes retinal slippage more likely when Perfluoron is removed from the eye.
Figure 1. Intraoperative photograph of Perfluoron bubble on macula with retinal detachment and retinotomy on the left of photo.
Subretinal migration of Perfluoron and emulsification of Perfluoron should be avoided during surgery. Subretinal migration of Perfluoron occurs when it is injected beyond the edge of a retinal defect (retinal tear, retinal hole, retinotomy, or retinectomy) that is under traction and cannot lie flat against the underlying retinal pigment epithelium. Just like when you overfill a glass of milk, the milk spills over the edge of the glass, Perfluoron that is filled beyond a tractionally elevated retinal defect will spill over the edge of the defect. The Perfluoron will then settle posteriorly underneath the retina. Similarly, subretinal Perfluoron migration can occur if the globe is rolled during surgery to visualize peripheral retina, causing the Perfluoron to "slosh around" in the eye and spill through a tractionally elevated retinal defect.
Emulsification of Perfluoron occurs when the BSS rapidly flows through the eye during the time Perfluoron is in use. Powerful infusion jets can form in the eye when a sclerotomy or trochar is not occluded. Flow through a sclerotomy occurs when it is not occluded by an instrument or plug or when it has enlarged to a size at which it remains open, even while occupied with an instrument. An eye filled with Perfluoron up to the level of the infusion cannula is also subject to violation of the Perfluoron bubble and emulsification. High-flow infusion aimed at the Perfluoron bubble can break off tens or hundreds of tiny Perfluoron bubbles inside the eye.
Removing Perfluoron completely once it has emulsified is difficult or impossible. Small residual droplets that remain in the eye postoperatively can migrate under the retina or cause chronic inflammation. Some patients are also bothered postoperatively because they see the Perfluoron droplets moving around in their superior visual field.

SILICONE OIL OVERVIEW

If surgeons are fortunate to work in an operating room sufficiently equipped with small-gauge instruments and need to repair a complex retinal detachment using the 23-gauge or 25-gauge vitrectomy systems, then they can use silicone oil as a postoperative tamponade. Many surgeons prefer using 20-gauge systems when working with silicone oil because they infuse more easily through a larger-gauge cannula. The resistance to flow through 23- and 25-gauge infusion tubing is much greater than the resistance to flow through 20-gauge infusion tubing. This is because the resistance to viscous flow through a tube is inversely proportional to the fourth power of the radius of the tube. In other words, a reduction of the inner diameter of a tube by half increases resistance to flow through the tube by a factor of 16. Despite this drawback, several publications have proven that Perfluoron and silicone oil can be successfully used with 25-gauge and 23-gauge vitrectomy systems.
Silicone oil overfill, underfill and anterior chamber fill need to be avoided during surgery. While infusing silicone oil, it is important to periodically palpate the eye to make sure the intraocular pressure is reasonable. During surgery, the silicone oil infusion cannula or needle must be angled away from the anterior chamber to avoid filling the chamber with oil during surgery.
In eyes that are pseudophakic with questionable capsular or zonular integrity or eyes that are aphakic, an inferior iridectomy large enough to stay open postoperatively usually prevents migration of silicone oil into the anterior chamber postoperatively (Figure 2).¹⁰ If oil migrates into the anterior chamber during surgery, the oil will sometimes exit the anterior chamber if the surgeon injects viscoelastic into the anterior chamber, removes oil from the posterior chamber, or completely removes the oil from the eye. Ocucoat (Bausch & Lomb) can be left in place; removal causes oil to come forward.
Figure 2. Image showing generous inferior peripheral iridectomy. This patient recovered 20/30 visual acuity after several retinal reattachment surgeries for retinal detachment and proliferative vitreoretinopathy. His silicone oil has been removed. Subtle droplets of silicone oil can be seen on the intraocular lens.
When operating on an eye where silicone oil injection is planned, the surgeon should avoid violating the anterior chamber of the eye. Any anteriorchamber incision that is not completely sealed at the end of surgery (including needle tracks) can allow aqueous to drain from the anterior chamber, facilitating the migration of silicone oil from the posterior chamber into the anterior chamber.

PERFLUORON–AIR–SILICONE OIL EXCHANGE

Most retinal detachments have an anterior component that retains subretinal fluid after Perfluoron has been infused to displace the posterior component of the retinal detachment. As air is infused into the eye, a soft-tipped cannula can be used to remove as much BSS anterior to the Perfluoron as possible. Subsequently, while minimally rolling the eye to maintain the Perfluoron bubble posteriorly, the soft-tip cannula can be used to drain subretinal fluid from the edges of any retinal defects. Often, subretinal fluid that was pushed anteriorly by the Perfluoron or that was there from the initial retinal detachment will be forced posteriorly as the air pressure reattaches the anterior retina. That fluid should be removed through any opening in the retina prior to removing the Perfluoron (Figure 3).⁶

FIGURE 3 IS REPRINTED FROM RETINA, 4TH ED., RYAN SJ, HINTON DR, SCHACHAT AP, WILKINSON CP, EDS., COPYRIGHT 2005, WITH PERMISSION FROM ELSEVIER.Figure 3. Perfluorocarbon liquid–silicone oil exchange: "sandwich" technique for subretinal fluid removal. Subretinal fluid loculated anterior to the retinal break is displaced posteriorly by silicone oil infusion (or air infusion) and removed through the retinal break.
After the anterior retina has reattached and the edges of any retinal defects have been dehydrated, the Perfluoron can be removed by placing the soft-tip cannula over the optic nerve. Since Perfluoron's refractive index is 1.27 and the refractive index of BSS is 1.3345, the edge of the Perfluoron bubble becomes visible during its removal. Sometimes, the residual preretinal Perfluoron can be washed off of the retinal surface by first allowing small amounts of BSS through the infusion cannula to wet the retina and subsequently removing the fluid from the eye and any residual Perfluoron it contains. Finally, to obtain a complete fluid-air exchange, a five- to 15-minute waiting period can be observed, and then the residual preretinal fluid can be removed from over the optic nerve with a soft-tipped extrusion cannula.
During the Perfluoron-air exchange, there is a greater chance that the edge of a giant retinal tear, retinotomy, or retinectomy may slip posteriorly than during a Perfluoron�silicone oil exchange (Figure 4). The force acting on the edge of the retinal defect as the air-BSS interface passes over it is generated primarily by the surface tension of BSS in air, which is approximately 60 dyne/cm (600 μN/cm).¹¹ This is much greater than the force acting on the edge of the retinal defect as silicone oil-BSS interface passes over it. This force is generated primarily by the interfacial tension of silicone oil in BSS, which is approximately 33 dyne/cm (330 μN/cm).¹¹
Figure 4. A retinal tear edge can be seen in white, and it has been pushed posteriorly by air.
If a retinectomy edge slips posteriorly after Perfluoronair exchange, the surgeon can attempt to reposition the slipped edge. The retina can be gently unfolded anteriorly by stroking it with a silicone-tipped extrusion cannula under gentle suction.¹² Alternatively, the eye can be slightly underfilled with silicone oil (this is only advisable for giant tears or retinotomies that are not located in the inferior half of the retina), and the patient can be instructed postoperatively to perform rolling maneuvers that can smooth the slipped retina back into place. If all else fails, the eye can be refilled with BSS and then reattached with Perfluoron, and a direct Perfluoron-silicone oil exchange can subsequently be performed.
After a complete Perfluoron-air exchange, silicone oil is injected into the air-filled eye. The nasal sclerotomy can be sutured or left open for air egress. Then the globe is rotated slightly nasally so the infusion cannula is as anterior as reasonably as possible. Some surgeons preplace sclerotomy closing sutures in the open sclerotomy and sometimes also around the infusion sclerotomy. Then the air infusion pressure is elevated briefly so the needle on the silicone oil syringe can be inserted into the superotemporal sclerotomy. Subsequently, the air infusion pressure is decreased and the oil is infused. Infusion of especially high-viscosity silicone oil is faster if a trimmed 18- or 20-gauge angiocatheter is used instead of the blunt needle provided with the silicone oil kit (Figure 5). The intraocular pressure can be checked periodically with palpation.
Figure 5. The cut edge of a 20-gauge angiocatheter has a larger inner diameter than the 20-gauge silicone oil needle.
When the eye is nearly filled with oil, the oil will move across the pupil or the back of the lens, and then the infusion cannula will start to fill with oil. At that moment, the infusion is stopped, the oil needle is removed, and the sclerotomy is closed. If there is a trapped air bubble in the front of the vitreous cavity, a 30-gauge needle can be inserted into the air to remove it from the eye, which allows oil that has refluxed into the low-pressure, air-filled infusion cannula to fill the eye. Then the infusion cannula is removed and the sclerotomy quickly closed.

PERFLUORON�SILICONE OIL EXCHANGE

Once the retina is attached under Perfluoron, the silicone oil syringe is connected to the high-pressure infusion cannula in place of the BSS. Care is taken to make sure the Luer-Lock is tight so the oil does not spray all over the operating room and not into the eye. The eye is briefly soft during this maneuver, so it must be done quickly. Then, as soon as the oil syringe is snugly attached to the infusion, the foot-pedal–controlled, high-pressure infusion pump can be used to infuse the oil into the vitreous cavity. The infusion needle must remain directed posteriorly or the anterior chamber will fill with oil. Using a wideangle viewing system, the BSS can be passively or actively removed from the vitreous space, just behind the infusing oil (Figure 6). I prefer the Charles 20-gauge flute needle attached to a backflush handpiece for these exchanges (Figure 7).
Figure 6. Oil is being infused from the right of the picture. The edge of the oil can be seen and the extrusion needle is being positioned with the tip immediately next to the infusing oil to remove the underlying BSS.
Figure 7. Charles fluted 20-gauge needle on backflush passive aspiration handpiece.
If the surgery is not done with trochars, the sclerotomies must be tight on the instruments to avoid excess spillage of silicone oil around the instruments through the sclerotomies and to assure adequate pressure inside the eye to drive the BSS and Perfluoron up the flute needle and out of the eye. Sometimes a suture needs to be placed to tighten a sclerotomy before commencing the silicone oil- Perfluoron exchange.
The tip of the aspiration needle is initially placed just behind the silicone oil. Silicone oil has a refractive index of 1.40. (BSS, as noted, has a refractive index of 1.3345.) This makes visualization of the interface between the oil and BSS possible until the oil fills the front of the eye and the posterior surface of the oil bubble-BSS interface flattens out.
Sometimes moving the needle from the back to the front and then to the back of the eye helps to find the interface. Once the oil has infused back to the edge of any retinal tears or retinotomies, time can be taken to remove as much BSS from those areas as possible to assure complete retinal reattachment (Figure 8).⁹ The eye is rotated as little as possible toward the breaks during this maneuver, so the benefit of the Perfluoron pressure on the posterior retina is maximized. After the meniscus of the silicone oil advances beyond the edge of the retinotomy or tear, the remaining Perfluoron can be removed from the posterior pole near the optic nerve.
Figure 8. Time is taken to drain as much BSS from the eye at the edge of the retinotomy, as the Perfluoron posteriorly and the silicone oil anteriorly push the subretinal fluid out from under the retina through the break.
The eye should be rotated, if possible, to avoid suction at or near the fovea. Intraocular pressure should be checked with palpation to make sure it is under control. Following complete removal of Perfluoron, the sclerotomies are closed while the oil infusion maintains a reasonable intraocular pressure. While closing the first sclerotomy, the second sclerotomy can be plugged.
Alternatively, if trochars were used, they are removed at this point and the integrity of the self-sealing sclerotomies is ascertained. Sutures can be placed in the trochar holes if necessary. Finally, when all other sclerotomies are sealed, the silicone oil infusion is removed and the infusion sclerotomy is closed as quickly as possible. If the intraocular pressure is high, oil can be passively leaked through this final sclerotomy prior to its closure. Anterior chamber-taps are not advisable since they can induce anterior-chamber migration of silicone oil.

FINISHING UP

At the conclusion of any case in which silicone oil is used, the surface of the eye must be thoroughly washed with salt solution to remove any residual oil. Postoperatively, the patient should be positioned face down to encourage resumption of aqueous humor flow into the anterior chamber and posterior settling of the silicone oil bubble. Using the techniques described in this article, it is possible to repair complex retinal detachments with proliferative vitreoretinopathy, sometimes achieving remarkable visual results, such as the retina pictured in Figure 9 of a patient with 20/30 visual acuity after four retinal surgeries. RP
Figure 9. Montage fundus photograph of a retina following retinal detachment repair, with an inferior retinectomy and silicone oil tamponade following silicone oil removal. The eye has 20/30 visual acuity even though there are small oil droplets on the posterior-chamber intraocular lens (seen in Figur

AWS-#2

	Print Friendly Page

Development of gene therapy for treatment of age-related macular degeneration.

Abstract

Intraocular neovascular diseases are the leading cause of blindness in the Western world in individuals over the age of 50. Age-related macular degeneration (AMD) is one of these diseases. Exudative AMD, the late-stage form, is characterized by abnormal neovessel development, sprouting from the choroid into the avascular subretinal space, where it can suddenly cause irreversible damage to the vulnerable photoreceptor (PR) cells essential for our high-resolution, central vision. The molecular basis of AMD is not well understood, but several growth factors have been implicated including vascular endothelial growth factor (VEGF), and the advent of anti-VEGF therapy has markedly changed the outcome of treatment. However, common to all current therapies for exudative AMD are the complications of repeated monthly intravitreal injections, which must be continued throughout one's lifetime to maintain visual benefits. Additionally, some patients do not benefit from established treatments. Strategies providing long-term suppression of inappropriate ocular angiogenesis are therefore needed, and gene therapy offers a potential powerful technique. This study aimed to develop a strategy based on RNA interference (RNAi) for the sustained attenuation of VEGF. We designed a panel of anti-VEGF short hairpin RNAs (shRNA), and based on the most potent shRNAs, microRNA (miRNA)-mimicked hairpins were developed. We demonstrated an additive VEGF silencing effect when we combined the miRNAs in a tricistronic miRNA cluster. To meet the requirements for development of medical treatments for AMD with long-term effects, the shRNA/miRNA is expressed from vectors based on adeno-associated virus (AAV) or lentivirus (LV). Both vector systems have been found superior in terms of transduction efficiency and persistence in gene expression in retinal cells. The capacity of AAV-encoded RNAi effector molecules to silence endogenous VEGF gene expression was evaluated in mouse models, including the model of laser-induced choroidal neovascularization (CNV), and we found that subretinal administration of self-complementary (sc)-AAV2/8 encoding anti-VEGF shRNAs can impair vessel formation. In parallel, a significant reduction of endogenous VEGF was demonstrated following injection of scAAV2/8 vectors expressing multiple anti-VEGF miRNAs into murine hind limb muscles. Furthermore, in an ongoing project we have designed versatile, multigenic LV vectors with combined expression of multiple miRNAs and proteins, including pigment epithelium-derived factor (PEDF), a multifunctional, secreted protein that has anti-angiogenic and neurotrophic functions. Co-expression of miRNAs and proteins from a single viral vector increases safety by minimizing the viral load necessary to obtain a therapeutic effect and thereby reduces the risk of insertional mutagenesis as well as the immune response against viral proteins. Our results show co-expression of functional anti-VEGF-miRNAs and PEDF in cell studies, and in vivo studies reveal an efficient retinal pigment epithelium (RPE)-specific gene expression following the incorporation of the vitelliform macular dystrophy 2 (VMD2) promoter, demonstrating the potential applicability of our multigenic LV vectors in ocular anti-VEGF gene therapy, including combination therapy for treatment of exudative AMD. In conclusion, these highly promising data clearly demonstrate that viral-encoded RNAi effector molecules can be used for the inhibition of neovascularization and will, in combination with the growing interest of applying DNA- or RNA-based technologies in the clinic, undoubtedly contribute to the development of efficacious long-term gene therapy treatment of intraocular neovascular diseases.

Prevention

• People with AMD may experience delay in progression of the disease with antioxidant vitamin and mineral supplementation.^[17] 
• The Age-Related Eye Disease Study (AREDS) classification of macular degeneration into early, intermediate and advanced forms revealed a beneficial effect of very high doses of antioxidants in reducing patient’s relative risk of progression to advanced AMD by 25%. These supplements may be indicated in patients with advanced AMD in the other eye.^[1] For example Viteyes 2 contains vitamin C 500 mg, vitamin E 400 IU, lutein 10 mg, zeaxanthin 2 mg, zinc minimum 25 mg and copper which is in line with the mix of antioxidants recommended by the AREDS.
• Studies have shown that increased intake of the macular carotenoids lutein and zeaxanthin and foods rich in these nutrients (eg, spinach and collard greens) is associated with a decreased risk of neovascular AMD. Dietary analysis of the observational component of AREDS also showed that lutein and zeaxanthin reduced AMD degeneration risk.^[2] 
• The strongest risk factor, age, is not preventable so currently the most important advice remains to focus on modifiable risk factors, such as control of hypertension, maintaining or achieving an ideal weight and smoking cessation. 

• Chemical (Alkali and Acid) Injury of the Conjunctiva and Cornea

Disease Entity

International Classification of Diseases

ICD-9-CM  940.2 alkaline chemical burn to cornea and conjunctiva, 940.3 acid chemical burn to the cornea and conjunctiva, 372.06 chemical conjunctivitis
ICD-10-CM T26.60XA Corrosion of cornea and conjunctival sac, unspecified eye, initial encounter. 

Epidemiology

Chemical injuries to the eye represent between 11.5%-22.1% of ocular traumas.[1]About two thirds of these injuries occur in young men. The vast majority occur in the workplace as a result of industrial accidents. A minority of injuries occur in the home or secondary to assault. Alkali materials are found more commonly in building materials and cleaning agents and occur more frequently than acid injuries.[2]

Etiology

Chemical injuries occur as a result of acid, alkali, or neurtral agents. Common causes of alkali and acid injuries are listed below.[2][3]       

  

Pathophysiology

Alkali

Alkali agents are lipophilic and therefore penetrate tissues more rapidly than acids. They saponify the fatty acids of cell cell membranes, penetrate the corneal stroma and destroyproteoglycan ground substance and collagen bundles. The damaged tissues then secrete proteolytic enzymes, which lead to further damage.[4][5]

Acids

Acids are generally less harmful than alkali substances. They cause damage by denaturing and precipitating proteins in the tissues they contact. The coagulated proteins act as a barrier to prevent further penetration (unlike alkali injuries).[5] The one exception to this is hydrofluoric acid, where the fluoride ion rapidly penetrates the thickness of the cornea and causes significant anterior segment destruction.[6]

Primary prevention

Since the majority of injuries occur at work, protective eye shields are mandatory when handling potentially corrosive substances (OSHA regulation, 1910.133). However, even protective goggles are no match for chemicals under high pressure.

Diagnosis

History

The severity of ocular injury depends on four factors: the toxicity of the chemical, how long the chemical is in contact with the eye, the depth of penetration, and the area of involvement. It is therefore critical to take a careful history to document these factors. The patient should be asked when the injury occurred, whether they rinsed their eyes afterwards and for how long, the mechanism of injury (was the chemical under high pressure?), the type of chemical that splashed in the eye, and whether or not they were wearing eye protection. If available, it is helpful to obtain the packaging of the chemical. There is often product information on this packaging including chemical composition. If this information is not immediately available, chemical information can be found by contacting the local poison control center at aapcc or 1 800-222-1222.

Physical examination

Prior to a full ophthalmic exam, the pH of both eyes should be checked. If the pH is not in physiologic range, then the eye must be irrigated to bring the pH to an appropriate range (between 7 and 7.2). It is recommended to wait at least five minutes after irrigation before checking the pH to ensure that the pH does not rise or fall secondary to retained particulate matter.
The physical exam should be used to assess the extent and depth of injury (see classification schemes below). Specifically, the degree of corneal, conjunctival and limbalinvolvement should be documented, as it can be used to predict ultimate visual outcome.[7]
The palpebral fissures should be checked and the fornices should be swept during the initial exam. Both the palpebral and bulbar conjunctiva should be examined with fluoresceinunder a cobalt blue light. As above, retained particulate matter can cause persistent damage, despite irrigation. The intraocular pressure should also be documented, as alkali injuries have been found to both acutely and chronically cause an elevation of IOP.[8]
Two major classification schemes for corneal burns are the Roper-Hall (modified Hughes) classification[9][10] and the Dua classification.[11] The Roper-Hall classification is based on the degree of corneal involvement and limbal ischemia. The Dua classification is based on an estimate of limbal involvement (in clock hours) and the percentage of conjunctival involvement. In a randomized controlled trial of acute burns, the Dua classification was found to be superior to the Roper-Hall in predicting outcome in severe burns.[7] However, both classification schemes are commonly employed in daily practice. 

                

Symptoms

The most common symptoms are severe pain, epiphora, blepharospasm, and reduced visual acuity.

Management

Irrigation

Early irrigation is critical in limiting the duration of chemical exposure. The goal of irrigation is to remove the offending substance and restore the physiologic pH. It may be necessary to irrigate as much as 20 liters to achieve this. To optimize patient comfort and ensure effective delivery of the irrigating solution, a topical anesthetic is generally administered. An eyelid speculum or Morgan Lens® (MorTan, Missoula MT) can be used to keep the eye open, while the irrigating solution is delivered through IV tubing. There has been some debate on the most effective irrigating solutions. A study by Herr et al. compared Normal Saline (NS), Normal Saline with Bicarbonate (NS + Bicarb),Lactated Ringer’s solution (LR), and Balanced Saline Solution Plus (BSS Plus, Alcon Laboratories, Fort Worth, TX) irrigating solutions to investigate which solution optimized patient comfort. They found that patients tolerated and preferred BSS irrigation compared to NS, NS + Bicarb, and LR.[12] In experiments in rabbit eyes following sodium hydroxide injury, a borate buffer solution called Cedderroth eye wash(Cedderroth Industrial Products, Upplands Vaasby Sweden) and a Diphthorine and Previn solution (Prevor, Cologne Germany) more efficiently normalized the pH compared to saline and phosphate buffer solutions.[13] Of course, early irrigation is paramount to limiting the duration of chemical exposure. If clean water is available at the site of injury and a standard irrigating solution is not, then the eyes should immediately be washed out with water.[14][15]

Medical therapy

Patients with mild to moderate injury (Grade I and II) have a good prognosis and can often be treated successfully with medical treatment alone. The aims of medical treatment are to enhance recovery of the corneal epithelium and augment collagen synthesis, while also minimizing collagen breakdown and controlling inflammation.[3]

Standard Treatment

Antibiotics- A topical antibiotic ointment like erythromycin ointment four times daily can be used to provide ocular lubrication and prevent superinfection. Stronger antibiotics (e.g. a topical fluoroquinolone) are employed for more severe injuries (e.g. Grade II and above).
Cycloplegic agents such as atropine or cyclopentolate can help with comfort.
Artificial tears- and other lubricating eye drops, preferably preservative free, should be used generously for comfort.
Steroid drops- In the first week following injury, topical steroids can help calm inflammation and prevent further corneal breakdown.[14] In mild injuries, topical prednisolone(Predforte) can be employed four times daily. In more severe injuries, prednisolone can be used every hour. After about one week of intensive steroid use, the steroids should be tapered because the balance of collagen synthesis vs. collagen breakdown may tip unfavorably toward collagen breakdown.[16]

Other Treatments:

Ascorbic acid- is a cofactor in collagen synthesis and may be depleted following chemical injury. Ascorbic acid can be used as a topical drop (10% every hour) or orally (two grams, four times daily in adults). In one study, severe alkali burns in rabbit eyes were associated with reduced ascorbic acid levels in the aqueous humor. This reduction correlated with corneal stromal ulceration and perforation. Systemic administration of Vitamin C helped promote collagen synthesis and reduce the level of ulceration.[17]Care must be taken in patients with compromised renal function because high levels of Vitamin C are potentially toxic to the kidneys.[18]
Doxycycline- acts independently of its antimicrobial properties to reduce the effects of matrix metalloproteinases (MMPs), which can degrade type I collagen. The tetracyclineclass inhibits MMPs by restriction of the gene expression of neutrophil collagenase and epithelial gelatinase, suppression of alpha 1 antitrypsin degradation and scavenging reactive oxygen species, thereby reducing ocular surface inflammation.[19][20]
Citrate drops- histological sections of cornea from alkali burns reveal an intense polymorphonuclear infiltrate (PMN).[21] PMNs provide a major source of proteolytic enzymes, which can dissolve the corneal stromal collagen. Deficiency in calcium inhibits the PMNs from granulating and releasing proteolytic enzymes. Citrate is a potent chelator and can therefore decrease proteolytic activity. Citrate also appears to inhibit collagenases.[22][23]
1% Medroxyprogesterone- is a progestational steroid and has less anti-inflammatory potency than corticosteroids, but has a minimum effect on stromal repair.Medroxyprogesterone can therefore be substituted for cortical steroids after 10-14 days of steroid treatment.[2][24]
Platelet rich plasma eye drops- have been found to be rich in growth factors and platelet rich plasma eye drops can lead to faster epithelialization for certain classes of burns.[25]

Surgical Treatments

Debridement of necrotic epithelium- should be performed as early as possible because necrotic tissue serves as a source of inflammation and can inhibit epithelialization.[3]

Conjunctival/Tenon’s transposition (Tenonplasty)- in Grade IV burns, anterior segment necrosis can result from loss of limbal vascular blood supply. In severe limbal ischemia, a sterile corneal ulceration can ensue. After removal of necrotic tissue, a tenonplasty (advancement of the conjunctiva and Tenon’s to the limbus) can be employed to reestablish limbal vascularity and facilitate re-epithelialization.[26]

Amniotic membrane transplantation (AMT)- the purpose of AMT is to rapidly restore the conjunctival surface and to reduce limbal and stromal inflammation. The benefits are thought to be two fold: physical and biological. Physically, AMT has been shown to improve patient comfort by reduction of eyelid friction. Numerous studies have found a reduction in pain following AMT for moderate to severe burns.[27][28] Through its physical actions, AMT may also prevent symblepharon formation. Amniotic membrane is also felt to have biologic effects.[29] It expresses TGFB1 and epidermal growth factor, which have roles in wound healing.[30][31] It has also been found to have anti-inflammatory properties.[32][33][34] Taken together, these biological effects may dampen inflammation, promote epithelial growth, prevent scarring and prevent neovascularization. New delivery devices like ProKera® (Bio-Tissue, Miami, Florida), which consists of a piece of cryopreserved amniotic membrane clipped into a dual ring system, like a symblepharon ring, allows rapid and sutureless placement of amniotic membrane.[35] A recent Cochrane review found only one randomized controlled trial of amniotic membrane for treatment of chemical ocular burn in the first seven days following injury.[1]Patients with moderate burns were found to have a significantly better visual acuity following AMT compared to medical therapy alone.[36] However, this was an unmasked trial and there were uneven baseline characteristics of the control and treatment eyes.[1]While case series and reviews show great promise of AMT in the treatment of chemical burns, conclusive evidence is still lacking. 

Limbal stem cell transplant- Much of the damage following chemical injuries results from limbal ischemia and the subsequent loss of stem cells capable of repopulating the corneal epithelium. Limbal stem cell transplants have been employed to replace this critical group of cells. Limbal stem cells are located at the base of the limbal epithelium and are responsible for repopulation of cells in the corneal epithelium and inhibition of conjunctival growth over the cornea.[37] Limbal autografts can be used from the healthy contralateral eye if only one eye is injured in a chemical burn.[38] When both eyes are injured, transplants have been attempted from living related donors. In a recent study from China, a portion of the limbus of HLA matched living related donors (allograft) was transplanted following chemical injury. Patients experienced a reduction in vascularity, improved corneal opacity and corneal epithelialization without the need for systemic immunosuppression.[37] Another option is to use cadaveric donors. This requires systemic immunosuppression.[39] When possible, limbal stem cell transplantation should be delayed until ocular surface inflammation has quieted.[40][41]

Cultivated oral mucosal epithelial transplantation (COMET)- can also be used to promote re-epithelialization and reduce inflammation in corneal burns. The cells are harvested from the patient’s own buccal mucosa so that systemic immunosuppression is not necessary.[42][43]

Boston Keratoprosthesis- Severe chemical injury leads to chronic inflammation and scarring, making visual recovery challenging. In cases with severe inflammation, limbal stem cell transplants and corneal transplants do not survive. In these most difficult cases, the Boston Keratoprosthesis can be used. Because it is independent of stem cell function, it does not require systemic immunosuppression.[44]

Recommended Treatment

While there is variability in treatment strategies of chemical burns, most authors recommended a graded approach depending on the severity of injury. Mild burns (Roper-Hall grade I) respond well to medical treatments and lubrication, while more severe burns necessitate more intensive medical therapies and surgery. Below is a paradigm for the initial treatment of chemical injury based on the Roper-Hall grade of injury.[3][45]

Grade I

• Topical antibiotic ointment (erythromycin ointment or similar) four times a day

• Prednisolone acetate 1% four times a day
• Preservative free artificial tears as needed
• If there is pain, consider a short acting cycloplegic like cyclopentolate three times a day

Grade II

• Topical antibiotic drop like fluoroquinolone four times daily
• Prednisolone acetate 1% hourly while awake for the first 7-10 days. Consider tapering the steroid if the epithelium has not healed by day 10-14. If an epithelial defect persists after day 10, consider progestational steroids (1% medroxyprogesterone four times daily)
• Long acting cycloplegic like atropine
• Oral Vitamin C, 2 grams four times a day
• Doxycycline, 100 mg twice a day (avoid in children)
• Sodium ascorbate drops (10%) hourly while awake
• Preservative free artificial tears as needed
• Debridement of necrotic epithelium and application of tissue adhesive as needed

Grade III

• As for Grade II
• Consider amniotic membrane transplant/Prokera placement. This should ideally be performed in the first week of injury

Grade IV

• As for Grade II/III
• Early surgery is usually necessary. For significant necrosis, a Tenonplasty can help reestablish limbal vascularity. An amniotic membrane transplant is often necessary due to the severity of the ocular surface damage.

Stages of Ocular Recovery

Stages of ocular recovery following chemical injury- [3][6]

 

Figure E                                                             Figure F

*Images courtesy of Dr. Kathryn Colby (Massachusetts Eye and Ear Infirmary)

Follow up

With severe chemical burns, patients should initially be followed daily. If there is concern for compliance with medication or if the patient is a child, one should consider inpatient admission. Once the health of the ocular surface has been restored, follow up can be spread apart. However, even in the healthiest appearing eyes, patients need long term monitoring for glaucoma and dry eye as below.

Other long term complications

Glaucoma

Glaucoma is quite common following ocular injury, ranging in frequency from 15%-55% in patients with severe burns.[8] The mechanism of glaucoma is multifactorial and includes contraction of the anterior structures of the globe secondary to chemical and inflammatory damage, inflammatory debris in the trabecular meshwork, and damage to the trabecular meshwork itself.[46] More severe burns (Roper-Hall Grade III or IV) have been found to have significantly higher intraocular pressure at presentation and were more likely to require long term glaucoma medication and undergo glaucoma surgery than grade I or II injuries.[8] Glaucoma medications should be prescribed as necessary to maintain normal intraocular pressure 

Dry eye

Chemical injury can destroy conjunctival goblet cells, leading to a reduction or even absence of mucus in the tear film, and compromising the proper dispersion of the precorneal tear film. This mucus deficiency results in keratoconjunctivitis sicca (dry eye).[47] Even in well-healed eyes, chronic dry eye can cause significant morbidity because of discomfort, visual disturbance, and potential for damage of the ocular surface.

Damage to the eyelids or palpebral conjunctiva

Direct chemical damage to the conjunctiva can lead to scarring, forniceal shortening, symblepharon formation and ciccatricial entropion or ectropion. These entities are encountered weeks to months after injury and can be treated by suppressing inflammation and with early amniotic membrane transplantation or oral mucosal graft.[3][48][49]