Advertisement

Measurable Aspects of the Retinal Neurovascular Unit in Diabetes, Glaucoma, and Controls

      Purpose

      To study the structural and angiographic optical coherence tomography (OCT) data of the macula from controls, patients with diabetes, and patients with glaucoma to evaluate neurovascular and structural relationships.

      Methods

      This was a retrospective study of 89 eyes from 49 patients in a community-based retinal referral practice with diabetes, glaucoma, and normal controls. The patients were evaluated with OCT to include retinal nerve fiber layer (RNFL) thickness measurement and ganglion cell layer (GCL) volume determination. The vascular density of the radial peripapillary capillary network and the vascular plexuses in the macula were evaluated with OCT angiography. The main outcome measures were the data obtained per disease state and the interrelationships the data displayed.

      Results

      The mean GCL volumes were significantly lower than the control group in both the diabetic (P = .016) and glaucoma (P < .001) groups. The difference between the diabetic and glaucoma groups was not significant (P = .052). The mean global vascular density was greater in the control group than the diabetic group (P = .002) and the glaucoma group (P < .001). The mean RNFL thicknesses were lowest in the glaucoma group. Both the diabetic and glaucoma groups had significantly lower radial peripapillary network and deep vascular plexus density values compared to controls.

      Conclusions

      Although there are important differences in disease pathogenesis between diabetes and glaucoma, they share certain similarities in the structural and angiographic abnormalities eventually produced. This suggests that, in addition to canonical pathways of disease, a component of both could represent neurodegenerative disease, offering the possibility for the development of new treatments. NOTE: Publication of this article is sponsored by the American Ophthalmological Society.
      Diabetic retinopathy is a complication of diabetes in which the retinal vascular system is thought to be the primary site of injury. Classification of disease severity of diabetic retinopathy is based on vascular changes and their consequences.
      Early Treatment Diabetic Retinopathy Study Research Group
      Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10.
      Diabetic retinopathy is separated into non-proliferative and proliferative retinopathy, based on the presence of retinal neovascularization. The non-proliferative phase includes microaneurysms, dot-and-blot hemorrhages, and venous beading, with or without regional blood flow abnormalities, evident as cotton wool spots. With sufficient ischemic load, new vessels may grow, leading to proliferative diabetic retinopathy. The fibrovascular tissue that grows is susceptible to bleeding, which can lead to scar tissue formation. Contracture of the scar and fibrovascular tissue can produce tractional retinal detachments. Treatment of diabetic retinopathy starts with strict control of metabolic parameters of diabetes and blood pressure to help prevent vascular complications and the use of intravitreal medications and laser photocoagulation if the vascular problems occur. Surgery is used in some patients who have problems that are not amenable to medical therapy.
      Glaucoma is a family of optic neuropathies in which the primary site of injury is thought to be at the optic nerve head, a place where the retinal nerve fibers—the axons of the retinal ganglion cells—deflect at right angles and then snake their way through the pores of the lamina cribrosa. Glaucoma causes thinning posterior displacement, bowing, and development of focal dehiscences of the lamina cribrosa.
      • Downs J.C.
      • Girkin C.A.
      Lamina cribrosa in glaucoma.
      Axonal injury from mechanical factors leads to axonal degeneration and ultimately retinal ganglion cell death. Distal to the site of injury, Wallerian degeneration occurs
      • Kanamori A.
      • Catrinescu M.M.
      • Belisle J.M.
      • et al.
      Retrograde and Wallerian axonal degeneration occur synchronously after retinal ganglion cell axotomy.
      and can be followed by transsynaptic degeneration and atrophic changes in more central portions of the visual pathway. Treatment of glaucoma centers on reduction of intraocular pressure using topical and systemic medications, with laser, shunt, incisional surgical approaches reserved for those who show disease progression on medical therapy.
      Although diabetic retinopathy and glaucoma are canonically described in terms of retinal vascular or axonal injury, respectively, these descriptions occur against a broader, and dramatically more complicated, backdrop.
      • Antonetti D.A.
      • Barber A.J.
      • Bronson S.K.
      • et al.
      Diabetic retinopathy: seeing beyond glucose-induced microvascular disease.
      Prior to developing observable retinal vascular disease, diabetic eyes show decreased dark adaptation,
      • Henson D.B.
      • North R.V.
      Dark adaptation in diabetes mellitus.
      • Midena E.
      • Segato T.
      • Giuliano M.
      • Zucchetto M.
      Macular recovery function (nyctometry) in diabetics without and with early retinopathy.
      reduced light sensitivity,
      • Joltikov K.A.
      • de Castro V.M.
      • Davila J.R.
      • et al.
      Multidimensional functional and structural evaluation reveals neuroretinal impairment in early diabetic retinopathy.
      • Ostroy S.E.
      • Frede S.M.
      • Wagner E.F.
      • et al.
      Decreased rhodopsin regeneration in diabetic mouse eyes.
      hue discrimination,
      • Trick G.L.
      • Burde R.M.
      • Gordon M.O.
      • et al.
      The relationship between hue discrimination and contrast sensitivity deficits in patients with diabetes mellitus.
      contrast sensitivity,
      • McAnany J.J.
      • Park J.C.
      Reduced contrast sensitivity is associated with elevated equivalent intrinsic noise in type 2 diabetics who have mild or no retinopathy.
      • Sokol S.
      • Moskowitz A.
      • Skarf B.
      • et al.
      Contrast sensitivity in diabetics with and without background retinopathy.
      • Ghafour I.M.
      • Foulds W.S.
      • Allan D.
      • McClure E.
      Contrast sensitivity in diabetic subjects with and without retinopathy.
      • Della Sala S.
      • Bertoni G.
      • Somazzi L.
      • et al.
      Impaired contrast sensitivity in diabetic patients with and without retinopathy: a new technique for rapid assessment.
      perimetry
      • Gella L.
      • Raman R.
      • Kulothungan V.
      • et al.
      Retinal sensitivity in subjects with type 2 diabetes mellitus: Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetics Study (SN-DREAMS II, Report No. 4).
      and microperimetry defects,
      • Midena E.
      • Vujosevic S.
      Microperimetry in diabetic retinopathy.
      and changes in pattern, full field,
      • Yang Q.
      • Xu Y.
      • Xie P.
      • et al.
      Retinal neurodegeneration in db/db mice at the early period of diabetes.
      • Tzekov R.
      • Arden G.B.
      The electroretinogram in diabetic retinopathy.
      • Pescosolido N.
      • Barbato A.
      • Stefanucci A.
      • Buomprisco G.
      Role of electrophysiology in the early diagnosis and follow-up of diabetic retinopathy.
      • Zhou W.
      • Rangaswamy N.
      • Ktonas P.
      • Frishman L.J.
      Oscillatory potentials of the slow-sequence multifocal ERG in primates extracted using the Matching Pursuit method.
      and multifocal electroretinograms.
      • Di Leo M.A.
      • Caputo S.
      • Falsini B.
      • et al.
      Presence and further development of retinal dysfunction after 3-year follow up in IDDM patients without angiographically documented vasculopathy.
      • Bearse Jr., M.A.
      • Adams A.J.
      • Han Y.
      • et al.
      A multifocal electroretinogram model predicting the development of diabetic retinopathy.
      • Tanvir Z.
      • Nelson R.F.
      • DeCicco-Skinner K.
      • Connaughton V.P.
      One month of hyperglycemia alters spectral responses of the zebrafish photopic electroretinogram.
      • Ewing F.M.
      • Deary I.J.
      • Strachan M.W.
      • Frier B.M.
      Seeing beyond retinopathy in diabetes: electrophysiological and psychophysical abnormalities and alterations in vision.
      Glaucoma, in addition to decreases in retinal ganglion cell counts, retinal nerve fiber layer (RNFL) thickness, and visual field parameters, is also associated with decreased macular pigment
      • Siah W.F.
      • O'Brien C.
      • Loughman J.J.
      Macular pigment is associated with glare-affected visual function and central visual field loss in glaucoma.
      and abnormalities in hue discrimination,
      • Pacheco-Cutillas M.
      • Edgar D.F.
      • Sahraie A.
      Acquired colour vision defects in glaucoma-their detection and clinical significance.
      • Greenstein V.C.
      • Halevy D.
      • Zaidi Q.
      • et al.
      Chromatic and luminance systems deficits in glaucoma.
      • Alvarez S.L.
      • Pierce G.E.
      • Vingrys A.J.
      • et al.
      Comparison of red-green, blue-yellow and achromatic losses in glaucoma.
      • Kurtenbach A.
      • Kernstock C.
      • Zrenner E.
      • et al.
      Electrophysiology and colour: a comparison of methods to evaluate inner retinal function.
      • Zanen A.
      • Baeck A.
      Color vision and contrast sensitivity tests in early diagnosis of primary open-angle glaucoma.
      • Niwa Y.
      • Muraki S.
      • Naito F.
      • et al.
      Evaluation of acquired color vision deficiency in glaucoma using the Rabin cone contrast test.
      contrast sensitivity,
      • McKendrick A.M.
      • Sampson G.P.
      • Walland M.J.
      • Badcock D.R.
      Contrast sensitivity changes due to glaucoma and normal aging: low-spatial-frequency losses in both magnocellular and parvocellular pathways.
      • Atkin A.
      • Bodis-Wollner I.
      • Wolkstein M.
      • et al.
      Abnormalities of central contrast sensitivity in glaucoma.
      • Hawkins A.S.
      • Szlyk J.P.
      • Ardickas Z.
      • et al.
      Comparison of contrast sensitivity, visual acuity, and Humphrey visual field testing in patients with glaucoma.
      and full field, multifocal, and pattern electroretinogram results.
      • Karaśkiewicz J.
      • Drobek-Słowik M.
      • Lubiński W.
      Pattern electroretinogram (PERG) in the early diagnosis of normal-tension preperimetric glaucoma: a case report.
      • Wilsey L.J.
      • Fortune B.
      Electroretinography in glaucoma diagnosis.
      • Wilsey L.
      • Gowrisankaran S.
      • Cull G.
      • et al.
      Comparing three different modes of electroretinography in experimental glaucoma: diagnostic performance and correlation to structure.
      The panoply of vision abnormalities in both diabetes and glaucoma argue against a simplistic notion of vascular or RNFL damage occurring in isolation. The retina is a complex system of neural, glial, and vascular elements integrally tied together forming the neurovascular unit. Diseases of the retina are likely to involve all of these elements to varying degrees.
      Both diabetic retinopathy and glaucoma are associated with ganglion cell loss. The soma of the ganglion cell lies in a region quantifiable by optical coherence tomography (OCT), called the ganglion cell layer (GCL). The axons of the ganglion cells produce the RNFL, which is quantifiable by OCT. The RNFL is supplied by the radial peripapillary capillary network, and the GCL is supplied by the superficial vascular plexus (SVP). The intermediate and deep capillary plexuses perfuse the mid-portions of the retina. All these vascular layers can be evaluated and quantified by OCT angiography. Analysis of the quantifiable parameters in the retina by high-resolution imaging has the potential to elucidate interrelationships among subsystems that are involved in blinding diseases and offer new insights into treatment possibilities. The present study examined eyes from three groups—normal controls, patients with mild diabetic retinopathy, and patients with glaucoma—to evaluate these parameters.

      Methods

      This retrospective study involved patients examined by the author at Vitreous Retina Macula Consultants of New York, a community-based retinal practice, over a 5-month period. The study design was approved by the Western Institutional Review Board (Puyallup, WA, USA) and was compliant with the Health Insurance Portability and Accountability and Act.

       Study Design

      The entry criteria were eyes of normal volunteers and patients with a history of glaucoma or diabetes. Exclusion criteria included those eyes with evidence of any neovascularization, any inflammatory disease involving the eye, history of any other retinal disease, macular edema, panretinal photocoagulation, myopia requiring -6 or more diopters of spherical equivalent, inability to accurate fixation, any media opacity that reduced the imaging quality or signal strength, or any movement artifacts. Accurate retinal segmentation can be problematical in eyes with dissociation of the retinal inner layers (DRILs), epiretinal membranes, or atrophy at the level of the retinal pigment epithelium. Therefore, these were considered exclusionary features. All diabetic patients were diagnosed and treated for diabetes by their internist or endocrinologist. The level of diabetic retinopathy was graded on the Diabetic Retinopathy Severity Scale.
      Early Treatment Diabetic Retinopathy Study Research Group
      Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12.
      The diagnosis and treatment of the glaucoma patients (all of whom had primary open angle glaucoma) was made by their attendant glaucoma specialists. All patients had 24-2 visual field testing with the HFA II-i Humphrey Field Analyzer using the standard Swedish Interactive Threshold Algorithm (Carl Zeiss Meditec, Inc., Dublin, CA, USA). The normal controls were free of any eye disease.

       Imaging

       Fundus photography

      Digital fundus photographs and fluorescein angiograms were taken with a Topcon TRC-50IX retinal camera (Topcon Medical Systems, Oakland, NJ, USA) using a 50° field of the posterior pole and recorded on a MegaPlus II ES 11000 (Redlake, Inc., Tuscon AZ, USA).

       Optical coherence tomography and layer assessment

      Images were obtained with both the Heidelberg Spectralis (Heidelberg Engineering, Franklin, MA, USA) and the Optovue RTVue XR Avanti (Optovue, Inc., Freemont, CA, USA). The Heidelberg Spectralis is a spectral domain instrument with a scanning speed of 85,000 A-scans per second, and it also has eye tracking. A 25 x 30-degree scan centered on the fovea was obtained with 31 sections. The layers of the retina were segmented with the Heidelberg HRA/Spectralis Viewing module 6.8.3, and the GCL was selected. Data recorded from this layer include the mean thickness in the center 1 mm, 4 quadrants within the inner and outer rings of the Early Treatment of Diabetic Retinopathy Study (ETDRS) grid overlay, as well as a global volume measurement (Figure 1). The RNFL was measured using the Heidelberg Spectralis, which performs a circle scan 3.45 mm in diameter centered on the nerve. The instrument provided RNFL thickness measurements of the global, temporal superior, nasal superior, temporal, nasal, temporal inferior, and nasal inferior sectors (Figure 1). The contained SD-OCT age-adjusted reference database was used to calculate deviations from normal means of the measurements. The superotemporal, temporal, and inferotemporal RNFLs were summed to produce a variable called macular nerve fiber layer. Every section of every scan was analyzed and manually corrected if there was a segmentation error.
      Figure thumbnail gr1
      Figure 1The selection and measurement of structural features. (A) The ganglion cell layer is segmented, and an en face map can be seen in this control eye. For analysis, the underlying data were used. (B) A B-scan through the nerve with the retinal nerve fiber layer highlighted in green. A circular area 3.45 mm in diameter was sampled, and the thicknesses were calculated. A graphical representation of the thickness can be visualized, but the underlying data were used in this study. The asterisks show the sectors for a right eye pooled together to produce the macular nerve fiber layer variable.

       Optical coherence tomography angiography

      The instrument used for OCT angiography images is based on the Optovue RTVue XR Avanti (Optovue, Inc.) to obtain amplitude-decorrelation angiography images. Exclusion criteria included a quality score of less than 6 as supplied by the Optovue instrument and gross eye motions not corrected by the motion control software in the Optovue system. The Optovue has an A-scan rate of 70,000 scans per second and was used in conjunction with its angiographic capability in a 6 x 6 mm area centered on the fovea. En face slab images through the retina were obtained to delineate the superficial vascular plexus. The intermediate capillary plexus and deep capillary plexus are found in close association, separation of which by imaging is dependent on precise segmentation. Given the potential difficulty with this segmentation in diseased eyes, the intermediate and deep plexuses were grouped together as per Snodderly et al
      • Snodderly D.M.
      • Weinhaus R.S.
      • Choi J.C.
      Neural-vascular relationships in central retina of macaque monkeys (Macaca fascicularis).
      and called the deep vascular plexus (DVP). The radial peripapillary capillary network was imaged by doing a 4.5 x 4.5-mm OCT angiographic scan centered on the optic nerve.
      The vascular images were thresholded using the Phansalkar algorithm for auto local threshold using a radius of 5 pixels in the FIJI program (Fiji Is Just ImageJ, National Institutes of Health, Rockville, MD, USA). The thresholded image was skeletonized, also using ImageJ. The scaled overlay of the ETDRS grid was overlaid on the vascular image as a mask, and for each sector the proportion of white compared to the black pixels was calculated and tabulated as a percentage (Figure 2).
      Figure thumbnail gr2
      Figure 2Imaging findings in glaucoma. (A) The radial peripapillary capillary network image obtain by optical coherence tomography angiography. (B) The angiographic image was thresholded and skeletonized. The grid corresponding to the nerve fiber layer was scaled to the size of the circle scanned and overlaid on the vascular image. The vascular density was calculated in each sector. (C) The 6 x 6 mm macular angiographic scan. (D) The macular angiographic scan was thresholded and skeletonized. The ETDRS grid was superimposed on the image, and the vascular density in each sector was measured.

       Statistical Methods

      Summary statistics were calculated for ocular and patient-specific variables. The correlation between variables was evaluated using generalized estimating equations (GEEs) except for simple exploratory analysis done with bivariate correlations and group mean comparisons done with one-way analysis of variance. To compare correlations, the Fisher r-Z transformation was used.
      • Warner R.M.
      Applied Statistics: From Bivariate Through Multivariate Techniques.
      • Preacher K.J.
      (2002, May). Calculation for the test of the difference between two independent correlation coefficients [Computer software].
      Linear discriminant analysis with prior probabilities computed from group sizes was performed with stepwise selection of variables with a F-value of entry and removal being 3.84 and 2.71, respectively. The Bonferroni approach was used to make alpha corrections for multiple comparisons. SPSS-IBM version 21 software (IBM Corp., Armonk, NY, USA) was used.

      Results

      There were 89 eyes of 49 subjects with a mean age of 59.1 years; 21 were males. The mean age of the 20 controls was 52.4 years; the 15 diabetic patients, 58.7 years; and the 14 glaucoma patients, 69.1 years. The difference in ages was significant by one-way analysis of variance (P = .006). The Games-Howell post hoc test, which does not assume either equal variances or sample sizes, revealed a significant difference between the ages of the control and glaucoma groups (P = .002). There was no significant difference between the diabetic and either the control (P = .462) or glaucoma groups (P = .076). In the diabetic group, the median Diabetic Retinopathy Severity Scale was 35, or mild non-proliferative diabetic retinopathy, found in 11 eyes (44%). Five patients (20%) had absent diabetic retinopathy (a score of 10), 5 had microaneurysms alone (a score of 20), 1 (4%) had moderate non-proliferative diabetic retinopathy (a score of 43), and 3 (12%) had moderately severe non-proliferative diabetic retinopathy (a score of 47). The average mean deviation of the Humphrey 24-2 was -3.36 (standard deviation [±] 2.14) dB, and the mean pattern standard deviation was 3.61 ± 3.21 dB.

       Optical Coherence Tomography Measurements

      The mean GCL volume was 0.98 ± 0.12 mm3. The values for the control, diabetic, and glaucoma groups were 1.06, 0.97, and 0.87 mm3, respectively (Table 1). Because of multiple comparisons, a P value of less than .0167 was required to meet statistical significance. The difference between the control and glaucoma groups was significant (P < .001, GEE), whereas the difference between the between the control and diabetic groups barely reached significance (P = .016, GEE) and the diabetic and glaucoma groups (P = .052, GEE) was not significant, after controlling for age. The mean global RNFL thickness measurements for the control, diabetic, and glaucoma groups were 98.8, 91.9, and 76.2 μm, respectively. Neither the difference in the mean global RNFL thickness measurements between the control and diabetic groups was not significant (P = .11, GEE) nor was the difference between the diabetic and glaucoma groups (P = .021, GEE), whereas the glaucoma group was significantly different from the control group (P < .001, GEE, after controlling for age). The macular nerve fiber layer thickness was considered to be the summation of the superotemporal, temporal, and inferotemporal RNFL thicknesses. The macular nerve fiber layer thickness showed no significant differences between the control and diabetic groups (P = .21, GEE), whereas there was a significant difference between the control and glaucoma groups (P < .001, GEE) and the diabetic and glaucoma groups (P < .001, GEE).
      Table 1Nerve Fiber Layer Thicknesses, Ganglion Cell Layer Volumes, and Associated Vascular Densities
      Control GroupDiabetic GroupGlaucoma GroupP Values
      Because of multiple comparisons, P must be less than .0167 to reach significance. Significant P values are highlighted in bold. All comparisons were made controlling for age.
      Group Comparison (Generalized Estimating Equations)
      Global Retinal NFL Thickness (μm)98.891.976.2.11Control-Diabetic
      < .001Control-Glaucoma
      .021Diabetic-Glaucoma
      Macular NFL Thickness.21Control-Diabetic
      < .001Control-Glaucoma
      < .001Diabetic-Glaucoma
      Peripapillary Capillary Density (% coverage)7.66.675.79.018Control - Diabetic
      < .001Control-Glaucoma
      .068Diabetic-Glaucoma
      GCL Volume (mm3)1.060.970.87.016Control - Diabetic
      < .001Control-Glaucoma
      .052Diabetic-Glaucoma
      SVP Density (% coverage)6.565.835.17.002Control - Diabetic
      < .001Control-Glaucoma
      .056Diabetic-Glaucoma
      DCP Density (% coverage)6.985.535.07.001Control - Diabetic
      < .001Control-Glaucoma
      .89Diabetic-Glaucoma
      DCP = deep capillary plexus; GCL = ganglion cell layer; NFL = nerve fiber layer; SVP = superficial vascular plexus.
      a Because of multiple comparisons, P must be less than .0167 to reach significance. Significant P values are highlighted in bold. All comparisons were made controlling for age.

       Vascular Measurements

       Radial peripapillary capillary layer

      The mean vascular density in the radial peripapillary capillary layer of the groups was 6.8 (±.013)%. The values for the control, diabetic, diabetic, and glaucoma groups were 7.6%, 6.67%, and 5.79%, respectively. The difference between the control group and the glaucoma group was significant (P < .001, GEE), after controlling for age. The difference between the control and diabetic groups was close to significance (P = .018), and there was no difference between the diabetic and glaucoma groups (P = .068, GEE), after controlling for age.

       Superficial vascular plexus

      The mean global vascular density percentage of the SVP was 5.95 (±0.06)%. The values for the control, diabetic, and glaucoma groups were 6.56%, 5.83%, and 5.17%, respectively. The control group was greater than the diabetic group (P = .002, GEE) and the glaucoma group (P < .001, GEE), after controlling for age. The diabetic group was not different than the glaucoma group (P = .056), after controlling for age.

       Deep capillary plexus

      The mean global vascular density DVP percentage was 6.01 (±0.069)%. The values for the control, diabetic, and glaucoma groups were 6.98%, 5.53%, and 5.07%, respectively. The control group was greater than the diabetic group (P = .001, GEE) and the glaucoma group (P < .001, GEE), after controlling for age. There was no significant difference between the diabetic and glaucoma groups in the vascular density of the deep capillary plexus (DCP) (P = .89, GEE).

       Interrelationships

      There was a significant correlation between the ganglion cell volume and the mean global RNFL thickness for the group (r = .81, P < .001). Curiously, the r values of the correlation were lowest for the glaucoma group (r = .47, P = .016) compared to the control group (r = .63, P < .001) or the diabetic group (r = .89, P < .001). Using the Fisher r-to-Z transformation, the correlation of the ganglion cell volume with the global nerve fiber layer thickness in the diabetes group was significantly higher than those seen in the control (P = .012) or glaucoma (P = .002) groups.
      As would be suspected, the GCL volume was highly correlated with the macular nerve fiber layer thickness (r = .83, P < .001). There were highly significant correlations between the nerve fiber layer parameters and the vascular densities of the peripapillary capillary layer. There were significant correlations between the ganglion cell volume and the vascular density of the SVP and DVP layers (Table 2) and the vascular densities among the various vascular layers with each other.
      Table 2Correlations Among Nerve Fiber Layer, Ganglion Cell Layer, and Associated Vascular Densities
      Global NFL ThicknessMacular NFL ThicknessPeripapillary Capillary DensityGCL VolumeSVP DensityDCP Density
      Global NFL Thickness
       Pearson Correlation.87.81.81.56.44
      Macular NFL Thickness
       Pearson Correlation.87.71.83.62.45
      Peripapillary Capillary Density
       Pearson Correlation.81.71.72.67.54
      GCL Volume
       Pearson Correlation.81.83.72.63.47
      SVP Density
       Pearson Correlation.56.62.67.63.64
      DCP Density
       Pearson Correlation.44.45.54.47.64
      All correlations significant with P < .001.
      DCP = deep capillary plexus; GCL = ganglion cell layer; NFL = nerve fiber layer; SVP = superficial vascular plexus.

       Disease Classification

      The universe of variables measured in this study had significant correlation to the disease categories as well as with each other. Linear discriminant analysis uses linear combinations of predictor variables to produce discriminant functions that seek to maximize the intergroup variances compared to the intragroup variance according to each function. It seeks to classify cases by creation of latent variables that maximize the group differences on each respective function, which are orthogonal. The variables selected by the stepwise method with all possible variables included parameters related to sectors of the DVP. These measurements were obtained from Optovue projection artifact resolved images. DVP images are subject to segmentation errors in retinal diseases, and, if present, the method of projection artifact removal used by other instruments will likely differ. Therefore, to produce a more generalizable result, the DCP data were not used in the final model. Of the 47 remaining measured variables, three were selected by the program to create two functions, demonstrating the dimensionality reduction inherent in linear discriminant analysis. The three selected variables were the macular RNFL, the GCL thickness in the outer superior sector of the ETDRS grid, and the vascular density of the SVP in the inner superior ETDRS grid. The discriminant function coefficients used are shown in Table 3A . The weightings of these two functions are shown in Table 3B. Using a leave-one-out approach, in which each case is classified by the functions derived from all cases other than that case, 69.7% of cases were correctly classified. This was greater than the model using all potential variables (data not shown). The linear discriminant models had a difficult time classifying the diabetic cases correctly (Table 4).
      Table 3A. Standardized Canonical Discriminant Function Coefficients
      Function
      12
      Macular RNFL.927−.771
      GCL Outer Superior−.275.816
      SVP Density Inner Superior.471.617
      Tabled 1B. Functions at Group Centroids
      Diagnostic GroupFunction
      12
      Control1.229.259
      Diabetic.064−.632
      Glaucoma−1.858.229
      GCL = ganglion cell layer; RNFL = retinal nerve fiber layer; SVP = superficial vascular plexus. The sector locations refer to the ETDRS grid.
      Unstandardized canonical discriminant functions evaluated at group means.
      Table 4Cross-validated Linear Discriminant Analysis Results
      Diagnostic GroupPredicted Group Membership
      ControlDiabeticGlaucoma
      Control32 (84.2%)5 (13.2%)1 (2.6%)
      Diabetic9 (36%)11 (44%)5 (20%)
      Glaucoma0 (0%)7 (26.9%)19 (73.1%)
      Each case is classified by the functions derived from all cases other than that case; 69.7% of cross-validated cases were correctly classified.

       Clinical Examples

      Although the data can communicate the underlying concepts, clinical imaging makes the observations readily apparent. The study relied on the underlying data, not graphical representations of the data. However, seeing the imaging can help reinforce the conclusions based on the data and can show relevant patterns that may be recognized in the clinic. Figure 3 shows a patient with a sector RNFL defect related to glaucoma. The arcuate swath of damage extends back to the optic nerve where a defect is seen in the radial peripapillary capillary network. The GCL shows a defect arcing inferotemporally; by comparison, the superior macular GCL is intact. The SVP shows a hypovascular region corresponding to the area of ganglion cell loss. Note that neither the radial peripapillary capillary network nor the SVP abnormalities follow the distribution of any retinal blood vessel. The zone involved can be identified in the color photograph, particularly if its existence is suspected. Figure 4 shows a comparison between a normal control and a patient with glaucoma. Either one in isolation may appear normal, but the comparison shows the decreased vascularity in the glaucoma eye. Figure 5 shows mild diabetic retinopathy with microaneurysms and small areas of non-perfusion. The GCL abnormalities are evident in the en face color rendering, but the RNFL did not show as much of a defect as would be expected in an eye with glaucoma with similar GCL deficits.
      Figure thumbnail gr3
      Figure 3Clinical example of the angiographic and structural changes seen in this patient with glaucoma. (A) The patient has a defect in the vascularity of the radial peripapillary capillary network (red arrow). The inset shows the graphical representation of the nerve fiber layer with the inferotemporal quadrant being in the first percentile for thickness (inset, red sector). (B) The graphical representation of the ganglion cell layer. Note the sector of decreased ganglion cell thickness (black arrow). (C) The optical coherence tomography angiographic image shows a swath of decreased vascularity corresponding to the region of decreased ganglion cell layer thickness. (D) The color fundus photograph shows decreased sheen in the region of nerve fiber loss.
      Figure thumbnail gr4
      Figure 4Superficial vascular plexus in a normal control (A) versus a patient with glaucoma (B).
      Figure thumbnail gr5
      Figure 5Mild diabetic retinopathy. (A) There are several small areas of signal void consistent with flow deficits (yellow arrows) and several small microaneurysms (white arrows). (B) A color map showing the ganglion cell thickness. A normal eye would show a bright yellow donut, whereas this eye has decreased ganglion cell layer thickness. The inset shows the color map of the nerve fiber layer thickness, which uses yellow to show areas in the age-adjusted second through fifth percentile and green to show everything above the fifth percentile. For both the ganglion cell and nerve fiber layer, the underlying data were used in the analysis.

      Discussion

      This study confirmed long-held concepts that glaucoma is related to lower RNFL and GCL parameters and that diabetic retinopathy is associated with retinal vascular abnormalities compared to controls. However, the glaucoma group had significantly decreased radial peripapillary capillary, SVP, and DCP vascular densities. The diabetic group had a decreased ganglion cell thickness compared to the control group. Interestingly, the RNFL was not significantly less in the diabetic group compared to the control group. Thus, the correlation between the ganglion cell volume and the nerve fiber layer was different in the diabetic group than in the glaucoma group. The measured variables were highly collinear. Linear discriminant analysis using a stepwise selection strategy found three variables out of the potential 47 that were then reduced to two functions. Although the overall success rate of disease classification was nearly 70%, this model had difficulty predicting the classification of patients from the diabetic group and placed many in the control and glaucoma groups. Expansion of these points and their greater clinical significance will occur in the paragraphs to follow.
      The measurement of any given layer includes more components than its name would imply. The nerve fiber layer has nerve fibers, but a significant proportion of the layer is Müller cells, astrocytes, microglia, and the radial peripapillary capillary network.
      • Büssow H.
      The astrocytes in the retina and optic nerve head of mammals: a special glia for the ganglion cell axons.
      The glial content of the RNFL varies from 18% to 42%, with the Müller cell fraction accounting for 15.2–21.3%.
      • Ogden T.E.
      Nerve fiber layer of the primate retina: thickness and glial content.
      In a similar fashion, the GCL has the soma of the ganglion cells, Müller cells, astrocytes, microglia, and displaced amacrine cells.
      • Frenkel S.
      • Goshen G.
      • Leach L.
      • et al.
      Peripapillary distribution of Muller cells within the retinal nerve fiber layer in human eyes.
      • Curcio C.
      • Allen K.
      Topography of ganglion cells in human retina.
      Jeon et al
      • Jeon C.J.
      • Strettoi E.
      • Masland R.H.
      The major cell populations of the mouse retina.
      used electron microscopy and determined in the C57 mouse that 59% of the total cells in the GCL were actually displaced amacrine cells. Snodderly et al
      • Snodderly D.M.
      • Weinhaus R.S.
      • Choi J.C.
      Neural-vascular relationships in central retina of macaque monkeys (Macaca fascicularis).
      reported that the volume occupied by the capillaries was 1.6–1.7% of the GCL; the aggregate volume of the larger vessels was not reported.
      Loss of layer volume is a gross measurement that has several implications. Increased intraocular pressure or increased glucose levels can lead to reduction of the size of the ganglion cell soma dendritic arborization and synaptic connections, effectively reducing the volume of the GCL before loss of actual ganglion cell numbers.
      • Sohn E.H.
      • van Dijk H.W.
      • Jiao C.
      • et al.
      Retinal neurodegeneration may precede microvascular changes characteristic of diabetic retinopathy in diabetes mellitus.
      • Weber A.J.
      • Kaufman P.L.
      • Hubbard W.C.
      Morphology of single ganglion cells in the glaucomatous primate retina.
      Loss of one cell type in the retina is often met with changes in neighboring cells of that layer. In a murine model of glaucoma, there was loss of axons of the ganglion cells but an increase in the glial content, with hypertrophic astrocytes, reactive astrocytes, and activated microglia forming gliotic and scar domains.
      • Bosco A.
      • Breen K.T.
      • Anderson S.R.
      • et al.
      Glial coverage in the optic nerve expands in proportion to optic axon loss in chronic mouse glaucoma.
      Cull et al
      • Cull G.A.
      • Reynaud J.
      • Wang L.
      • et al.
      Relationship between orbital optic nerve axon counts and retinal nerve fiber layer thickness measured by spectral domain optical coherence tomography.
      investigated a rhesus macaque model of laser-induced glaucoma and found that the axon count was equal to 12,336 × RNFL thickness in μm - 257,050. This implies that the RNFL could measure nearly 21 μm thick and contain no axons. Loss of any component of a layer, adjusting for potential induced proliferation of other cell types, could lead to layer thinning. Attribution of the loss would implicitly be to the name of the cell layer, but it is possible that the proportions of cell loss may vary from one disease to another. Thus, the proportions of ganglion cell loss in the GCL compared to the other constituents of that layer may be different in glaucoma compared to diabetes. This is suggested by the difference in the RNFL-GCL ratio in the present study. Greatly complicating the issue is that each of the cell types present have been classified into many different subtypes. There are dozens of different types of ganglion cells,
      • Sanes J.R.
      • Masland R.H.
      The types of retinal ganglion cells: current status and implications for neuronal classification.
      for example. Knowledge of the total numbers of ganglion cells lost does not necessarily mean we know the proportionate loss of each ganglion cell type in any given disease.
      The retinal cells are located in strata and can contain specific layers of blood vessels seemingly adapted to local needs. The RNFL has a high metabolic demand and a specialized layer of capillaries—the radial peripapillary capillary network—to supply its needs. Reduction of the nerve fibers in the layer would be expected to reduce the metabolic load and thus the necessity of maintaining a dense capillary network. Numerous investigators have established a decrease in the radial peripapillary capillary vascular density of nerve fiber loss.
      • Yarmohammadi A.
      • Zangwill L.M.
      • Diniz-Filho A.
      • et al.
      Relationship between optical coherence tomography angiography vessel density and severity of visual field loss in glaucoma.
      • Yarmohammadi A.
      • Zangwill L.M.
      • Diniz-Filho A.
      • et al.
      Optical coherence tomography angiography vessel density in healthy, glaucoma suspect, and glaucoma eyes.
      • Chung J.K.
      • Hwang Y.H.
      • Wi J.M.
      • et al.
      Glaucoma diagnostic ability of the optical coherence tomography angiography vessel density parameters.
      • Rao H.L.
      • Pradhan Z.S.
      • Weinreb R.N.
      • et al.
      Regional comparisons of optical coherence tomography angiography vessel density in primary open-angle glaucoma.
      • Liu L.
      • Jia Y.
      • Takusagawa H.L.
      • et al.
      Optical coherence tomography angiography of the peripapillary retina in glaucoma.
      • Takusagawa H.L.
      • Liu L.
      • Ma K.N.
      • et al.
      Projection-resolved optical coherence tomography angiography of macular retinal circulation in glaucoma.
      • Chen H.S.
      • Liu C.H.
      • Wu W.C.
      • et al.
      Optical coherence tomography angiography of the superficial microvasculature in the macular and peripapillary areas in glaucomatous and healthy eyes.
      • Triolo G.
      • Rabiolo A.
      • Shemonski N.D.
      • et al.
      Optical coherence tomography angiography macular and peripapillary vessel perfusion density in healthy subjects, glaucoma suspects, and glaucoma patients.
      • Chen C.L.
      • Zhang A.
      • Bojikian K.D.
      • et al.
      Peripapillary retinal nerve fiber layer vascular microcirculation in glaucoma using optical coherence tomography-based microangiography.
      • Sakaguchi K.
      • Higashide T.
      • Udagawa S.
      • et al.
      Comparison of sectoral structure-function relationships in glaucoma: vessel density versus thickness in the peripapillary retinal nerve fiber layer.
      • Scripsema N.K.
      • Garcia P.M.
      • Bavier R.D.
      • et al.
      Optical coherence tomography angiography analysis of perfused peripapillary capillaries in primary open-angle glaucoma and normal-tension glaucoma.
      • Shin J.W.
      • Sung K.R.
      • Lee J.Y.
      • et al.
      Optical coherence tomography angiography vessel density mapping at various retinal layers in healthy and normal tension glaucoma eyes.
      • Chihara E.
      • Dimitrova G.
      • Amano H.
      • Chihara T.
      Discriminatory power of superficial vessel density and prelaminar vascular flow index in eyes with glaucoma and ocular hypertension and normal eyes.
      • Kumar R.S.
      • Anegondi N.
      • Chandapura R.S.
      • et al.
      Discriminant function of optical coherence tomography angiography to determine disease severity in glaucoma.
      Histopathologic support for this imaging finding of radial peripapillary capillary loss was published a half century ago.
      • Kornzweig A.L.
      • Eliasoph I.
      • Feldstein M.
      Selective atrophy of the radial peripapillary capillaries in chronic glaucoma.
      Loss of ganglion cells in diabetes has been seen in experimental animal models and in humans.
      • van Dijk H.W.
      • Verbraak F.D.
      • Kok P.H.
      • et al.
      Decreased retinal ganglion cell layer thickness in patients with type 1 diabetes.
      • Srinivasan S.
      • Pritchard N.
      • Sampson G.P.
      • et al.
      Focal loss volume of ganglion cell complex in diabetic neuropathy.
      • Hegazy A.I.
      • Zedan R.H.
      • Macky T.A.
      • Esmat S.M.
      Retinal ganglion cell complex changes using spectral domain optical coherence tomography in diabetic patients without retinopathy.
      • Carpineto P.
      • Toto L.
      • Aloia R.
      • et al.
      Neuroretinal alterations in the early stages of diabetic retinopathy in patients with type 2 diabetes mellitus.
      Ganglion cell complex thinning was more evident in diabetic patients who had diabetic polyneuropathy compared to those who did not.
      • Salvi L.
      • Plateroti P.
      • Balducci S.
      • et al.
      Abnormalities of retinal ganglion cell complex at optical coherence tomography in patients with type 2 diabetes: a sign of diabetic polyneuropathy, not retinopathy.
      Experimental and human diabetes increases TUNEL-positive cells compared to non-diabetics.
      • Barber A.J.
      • Lieth E.
      • Khin S.A.
      • et al.
      Neural apoptosis in the retina during experimental and human diabetes. Early onset and effect of insulin.
      Human diabetic autopsy eyes show increased expression of caspase-3, caspase-9, and Bax, which are all pro-apoptotic markers.
      • Oshitari T.
      • Yamamoto S.
      • Hata N.
      • Roy S.
      Mitochondria- and caspase-dependent cell death pathway involved in neuronal degeneration in diabetic retinopathy.
      Valverde et al
      • Valverde A.M.
      • Miranda S.
      • García-Ramírez M.
      • et al.
      Proapoptotic and survival signaling in the neuroretina at early stages of diabetic retinopathy.
      found an imbalance of pro- to anti-apoptotic markers in retinas from human donors compared to controls. Diabetes also causes microvascular abnormalities, which in turn could affect the viability of the cells in the retina. However, neuronal loss occurs early after the onset of diabetes, even with no detectable vessel loss.
      • Barber A.J.
      • Lieth E.
      • Khin S.A.
      • et al.
      Neural apoptosis in the retina during experimental and human diabetes. Early onset and effect of insulin.
      • Kern T.S.
      • Barber A.J.
      Retinal ganglion cells in diabetes.
      As early as 1961, Wolter found loss of both ganglion cells and photoreceptors in diabetic eyes obtained at autopsy.
      • Wolter J.R.
      Diabetic retinopathy.
      Hyperglycemia in animal models leads to apoptosis of photoreceptors and ganglion cells, even prior to observable retinopathy. The RNFL has been found to be decreased in patients with diabetes compared to non-diabetics in some studies.
      • Joltikov K.A.
      • de Castro V.M.
      • Davila J.R.
      • et al.
      Multidimensional functional and structural evaluation reveals neuroretinal impairment in early diabetic retinopathy.
      • Park S.H.
      • Park J.W.
      • Park S.J.
      • et al.
      Apoptotic death of photoreceptors in the streptozotocin-induced diabetic rat retina.
      • Chen X.
      • Nie C.
      • Gong Y.
      • et al.
      Peripapillary retinal nerve fiber layer changes in preclinical diabetic retinopathy: a meta-analysis.
      • Dumitrescu A.G.
      • Istrate S.L.
      • Iancu R.C.
      • et al.
      Retinal changes in diabetic patients without diabetic retinopathy.
      • Jeon S.J.
      • Kwon J.W.
      • La T.Y.
      • et al.
      Characteristics of retinal nerve fiber layer defect in nonglaucomatous eyes with type II diabetes.
      In the present study, the difference did not reach statistical significance. These findings suggest that a component of diabetic retinopathy is a neurodegenerative condition.
      One hallmark of glaucoma is the loss of ganglion cells. The susceptibility to this damage seems to run in families and seems to share mechanisms of cell injury and death with central neurodegenerative diseases such as Alzheimer's and Parkinson's diseases.
      • Eraslan M.
      • Çerman E.
      • Çekiç O.
      • et al.
      Neurodegeneration in ocular and central nervous systems: optical coherence tomography study in normal-tension glaucoma and Alzheimer disease.
      • Kaur M.
      • Saxena R.
      • Singh D.
      • et al.
      Correlation between structural and functional retinal changes in Parkinson disease.
      • Jones-Odeh E.
      • Hammond C.J.
      How strong is the relationship between glaucoma, the retinal nerve fibre layer, and neurodegenerative diseases such as Alzheimer's disease and multiple sclerosis?.
      • Marziani E.
      • Pomati S.
      • Ramolfo P.
      • et al.
      Evaluation of retinal nerve fiber layer and ganglion cell layer thickness in Alzheimer's disease using spectral-domain optical coherence tomography.
      Of interest, there are abnormalities of GCL thickness in neurodegenerative brain changes associated with bipolar disorder,
      • Kalenderoglu A.
      • Sevgi-Karadag A.
      • Celik M.
      • et al.
      Can the retinal ganglion cell layer (GCL) volume be a new marker to detect neurodegeneration in bipolar disorder?.
      schizophrenia,
      • Celik M.
      • Kalenderoglu A.
      • Sevgi Karadag A.
      • et al.
      Decreases in ganglion cell layer and inner plexiform layer volumes correlate better with disease severity in schizophrenia patients than retinal nerve fiber layer thickness: findings from spectral optic coherence tomography.
      multiple sclerosis,
      • Tugcu B.
      • Soysal A.
      • Kılıc M.
      • et al.
      Assessment of structural and functıonal vısual outcomes ın relapsıng remıttıng multıple sclerosıs wıth vısual evoked potentıals and optıcal coherence tomography.
      and obstructive sleep apnea.
      • Huseyinoglu N.
      • Ekinci M.
      • Ozben S.
      • et al.
      Optic disc and retinal nerve fiber layer parameters as indicators of neurodegenerative brain changes in patients with obstructive sleep apnea syndrome.
      The Rotterdam Study found that thinner retinal nerve fiber as well as ganglion cell and inner plexiform layers were associated with smaller gray-matter and white-matter volume.
      • Mutlu U.
      • Bonnemaijer P.W.M.
      • Ikram M.A.
      • et al.
      Retinal neurodegeneration and brain MRI markers: the Rotterdam Study.
      Central nervous system white-matter microstructural abnormalities as determined by fractional anisotropy and mean diffusivity of diffusion tensor magnetic resonance imaging was associated with decreased retinal nerve fiber and GCL thicknesses.
      • Nork T.M.
      • Ver Hoeve J.N.
      • Poulsen G.L.
      • et al.
      Swelling and loss of photoreceptors in chronic human and experimental glaucomas.
      Animal and human glaucoma can have signs of photoreceptor abnormalities.
      • Nork T.M.
      • Ver Hoeve J.N.
      • Poulsen G.L.
      • et al.
      Swelling and loss of photoreceptors in chronic human and experimental glaucomas.
      These findings suggest that glaucoma includes a component of a neurodegenerative disease that can affect the retina, optic nerve, and potentially more central portions of the nervous system.
      The present study found that the vascular layers were individually decreased in the radial peripapillary, SVP, and DCP layers in both the diabetic and glaucoma groups. The loss of layers of retinal vascularity is a common feature in the retinas affected by diabetes. The vascular involvement in glaucoma is more controversial. Alnawaiseh et al
      • Alnawaiseh M.
      • Lahme L.
      • Müller V.
      • et al.
      Correlation of flow density, as measured using optical coherence tomography angiography, with structural and functional parameters in glaucoma patients.
      and Lommatzsch et al
      • Lommatzsch C.
      • Rothaus K.
      • Koch J.M.
      • et al.
      OCTA vessel density changes in the macular zone in glaucomatous eyes.
      found similar results to the present study; Takusagawa et al
      • Takusagawa H.L.
      • Liu L.
      • Ma K.N.
      • et al.
      Projection-resolved optical coherence tomography angiography of macular retinal circulation in glaucoma.
      found abnormalities in the SVP but did not find abnormalities in either the intermediate or deep vascular plexus; Chung et al
      • Chung J.K.
      • Hwang Y.H.
      • Wi J.M.
      • et al.
      Glaucoma diagnostic ability of the optical coherence tomography angiography vessel density parameters.
      found decreased total macular vascular density; Chen et al
      • Chen H.S.
      • Liu C.H.
      • Wu W.C.
      • et al.
      Optical coherence tomography angiography of the superficial microvasculature in the macular and peripapillary areas in glaucomatous and healthy eyes.
      found that the SVP vascular density was decreased but did not mention the deeper vascular layers; and Triolo et al
      • Triolo G.
      • Rabiolo A.
      • Shemonski N.D.
      • et al.
      Optical coherence tomography angiography macular and peripapillary vessel perfusion density in healthy subjects, glaucoma suspects, and glaucoma patients.
      did not find macula vascular density defects in glaucoma patients. Given that the vascular abnormalities were found to be correlated to the RNFL and GCL volume in the present study, there is a possibility that the correlations or lack thereof in various studies may be a function of the variation in disease severity.
      Vascular abnormalities have been considered to be the most salient aspect of diabetic retinopathy; intuitively, one could assume that the retinal injury in diabetes is secondary to the vascular problems. As already reviewed in this article, numerous abnormalities occur in diabetic patients prior to observed vascular defects. This would suggest that some of the vascular abnormalities could be secondary rather than primary. In a similar fashion, this study found that there is potential for widespread vascular abnormalities in the retina in glaucoma. A reductivist approach would seek to isolate if the pressure or the vascular compartment was the primary insult. However, this approach would ignore the complex interrelationships that exist in the retina. Any one subsystem is integrally intertwined with all the remaining subsystems.
      The argument that glaucoma was due to either pressure (von Graefe
      • Graefe A.
      On iridectomy in glaucoma and the glaucomatous process.
      ) or ischemia (von Jaeger
      • von Jaeger E.
      Ueber Glaucom und seine Heilung durch Iridectomie.
      ) was framed in the nineteenth century. This argument, in various forms, persists because it is a paradox created by the oversimplification inherent in reductionism. It is much like arguing if snow occurs because the temperature is cold or if there are clouds. That does not mean that weather may not be modifiable in the future by targeting one aspect, seeding clouds for example, but this would occur in the total context of the weather system attributes. Of the many potential targets in treating glaucoma, control of intraocular pressure has been shown to be only moderately effective. In the Early Manifest Glaucoma Trial,
      • Heijl A.
      • Leske M.C.
      • Bengtsson B.
      • et al.
      Reduction of intraocular pressure and glaucoma progression: results from the Early Manifest Glaucoma Trial.
      patients in the treatment group showed an average reduction of intraocular pressure of 25% over 6 years of follow-up. Progression was seen in 45% of the controls compared to 62% of the observation group (P = .007). This converts to a number needed to treat of 5.9. Treatment delayed, but did not prevent, progression in many patients. It is possible that a greater proportion of patients would not have had progression with greater pressure reduction, but the results suggest that intraocular pressure reduction does not address all of the factors involved in perimetric progression in glaucoma.
      Much of the recent efforts in treating diabetic macular edema have focused on the vascular abnormalities, first starting with proliferative diabetic retinopathy being treated with panretinal photocoagulation,
      The Diabetic Retinopathy Study Research Group
      Photocoagulation treatment of proliferative diabetic retinopathy. Clinical application of Diabetic Retinopathy Study (DRS) findings, DRS Report Number 8.
      then diabetic macular edema treated with scatter laser photocoagulation,
      Early Treatment Diabetic Retinopathy Study Research Group
      Early photocoagulation for diabetic retinopathy. ETDRS report number 9.
      and, more recently, both conditions being treated with anti-vascular endothelial growth factor (VEGF) agents.
      • Nguyen Q.D.
      • Brown D.M.
      • Marcus D.M.
      • et al.
      Ranibizumab for diabetic macular edema: results from 2 phase III randomized trials: RISE and RIDE.
      Even with the most modern approaches, patients with diabetic retinopathy do not have normal visual acuity.
      • Gross J.G.
      • Glassman A.R.
      • Liu D.
      • et al.
      Five-year outcomes of panretinal photocoagulation vs intravitreous ranibizumab for proliferative diabetic retinopathy: a randomized clinical trial.
      Deviations from normal may be attributed to ischemia in the absence of edema, but given the large number of abnormalities that can occur in the context of diabetes, this assumption is not likely to be completely correct. Even if there are vascular abnormalities associated with decreased visual performance, it is not established that residual vascular defects have a purely primary or a secondary relationship to the retinal degeneration.
      The present study found considerable overlap in the structural and angiographic features measured by OCT or OCT angiography in glaucoma and diabetes. The features used by linear discriminant analysis were macular RNFL, the GCL thickness in the outer superior sector of the ETDRS grid, and the vascular density of the SVP in the inner superior ETDRS grid. Inspecting the variables selected can suggest mechanistic reasons for their selection and may help in hypothesis generation for future studies. Selection of one variable may occur not because that variable is affected in a particular disease but because it is not affected in a competing group. However, the exact reason for the selection of variables may not be readily apparent as they are used in ratiometric ways in factor creation in the context of creating discriminant functions. The macular RNFL is decreased predominantly in glaucoma but to a lesser extent in diabetes.
      Glaucoma can affect the inferotemporal GCL preferentially.
      • Hood D.C.
      • De Moraes C.G.
      Challenges to the common clinical paradigm for diagnosis of glaucomatous damage with OCT and visual fields.
      Thus, the superior GCL could be a way to differentiate diabetes from glaucoma. Although it is possible to think of ways that the vascular density of the superior outer ring of the ETDRS grid could differentiate the groups, the exact mechanism by which this is used is not known. The control and glaucoma groups were classified with a higher accuracy than the diabetic group. Diabetic patients may have no observable retinopathy and therefore would be difficult to differentiate from normals. Diabetics with more advanced disease have more prominent structural and angiographic findings, some of which overlap those found in eyes with glaucoma.
      Diabetes and glaucoma are two of the most prevalent blinding diseases. Each has some modifiable factors that can be used to reduce the risk of vision loss, but each of these treatment modalities addresses one subset of the pathology that may be involved. A larger universe of neurodegenerative and neuroinflammatory processes
      • Williams P.A.
      • Marsh-Armstrong N.
      • Howell G.R.
      • et al.
      Neuroinflammation in glaucoma: a new opportunity.
      • de Hoz R.
      • Rojas B.
      • Ramírez A.I.
      • et al.
      Retinal macroglial responses in health and disease.
      may be functioning to lead to tissue damage.
      • Barber A.J.
      A new view of diabetic retinopathy: a neurodegenerative disease of the eye.
      The results of the linear discriminant analysis show the difficulty in separating eyes based on only the measured structural and vascular density data because of these similarities. Although diabetes and glaucoma are viewed as discrete diseases, potentially driven in part by mechanical or vascular pathology, they may share similar pathways in retinal neurodegeneration. Addressing the neurodegeneration, then, would seem to offer an opportunity to advance treatment beyond the current level of success.

       Limitations and Future Challenges

      This study sought to evaluate multiple parameters evident in structural and angiographic OCT to determine similarities and differences between the eyes of patients with diabetes and glaucoma compared to controls. There are obvious limitations inherent in this approach in evaluating the neurovascular unit. Neurovascular coupling is the linkage in magnitude and spatial location of blood flow with neural activity. Neurovascular coupling is a function of the neurovascular unit and is abnormal in the retinas in both diabetes and glaucoma.
      • Prada D.
      • Harris A.
      • Guidoboni G.
      • et al.
      Autoregulation and neurovascular coupling in the optic nerve head.
      Unfortunately, this information is not currently available in current structural or angiographic OCT testing. The neurovascular unit also contains microglia and inflammatory cells.
      • Silverman S.M.
      • Wong W.T.
      Microglia in the retina: roles in development, maturity, and disease.
      These are important in the pathogenesis of both diabetes and glaucoma
      • Gauthier A.C.
      • Liu J.
      Neurodegeneration and neuroprotection in glaucoma.
      • Zeng H.L.
      • Shi J.M.
      The role of microglia in the progression of glaucomatous neurodegeneration- a review.
      but also cannot be quantified with current imaging technology. One focus of the present study is on the RNFL and GCL; each of these has numerous cells other than its namesake. The numbers of the constituent cells cannot be determined in humans yet with current imaging. Each of these cells has many subclasses that represent further future challenges in quantifying the cells present.
      • Vlasits A.L.
      • Euler T.
      • Franke K.
      Function first: classifying cell types and circuits of the retina.
      The functioning of the retina depends on the synaptic microcircuitry of the retina, and, with changes in cell numbers, subclasses, or interconnectivity, the functional capability will likely change as well.
      The present study had a limited number of patients evaluated. Expansion of the present study can help gain a clearer picture of the interrelationships among the tested variables. The present study looked at generic vascular density findings. Using vascular features such as microaneurysms would likely improve the classification capabilities. However, this would not mask the overlap between the structural findings in the retina in the diabetes versus the glaucoma groups. The diabetic group in this study could not have any edema, disorganization of the retinal inner layers, epiretinal proliferation, or atrophy of the retinal pigment epithelium (such as from laser). Therefore, the findings expected in a group of diabetics encountered in a clinical setting may include those with more severe disease and more pronounced findings than those in the present study.
      This study examined the eyes of diabetic and glaucoma patients and compared them to controls. It found overlapping clinical findings in those with diabetes and glaucoma. Although the mechanism inciting injury may be different in these two conditions, some of the induced pathologic processes may result in similar modes of neuroretinal inflammation and degeneration.
      • Rübsam A.
      • Parikh S.
      • Fort P.E.
      Role of inflammation in diabetic retinopathy.
      • Vohra R.
      • Tsai J.C.
      • Kolko M.
      The role of inflammation in the pathogenesis of glaucoma.
      Attacking each of these diseases can be done by targeting certain salient modes of disease expression, but it is likely that a broader conceptual approach is needed to address all the generated problems. A stronger anti-VEGF agent or a more potent prostaglandin analogue is not likely to dramatically repair all of the problems in patients with diabetes or glaucoma. Anti-VEGF agents or prostaglandin analogues are important current treatments, but to get significant improvements in treatment efficacy, other aspects of disease expression, particularly neurodegeneration,
      • Gupta N.
      • Yücel Y.H.
      Glaucoma as a neurodegenerative disease.
      • Danesh-Meyer H.V.
      • Levin L.A.
      Glaucoma as a neurodegenerative disease.
      • Barber A.J.
      Diabetic retinopathy: recent advances towards understanding neurodegeneration and vision loss.
      • Simó R.
      • Hernández C.
      European Consortium for the Early Treatment of Diabetic Retinopathy
      Neurodegeneration in the diabetic eye: new insights and therapeutic perspectives.
      need to be managed. Potential avenues include neuroprotection, decreasing neuroinflammatory activity,
      • Schröder S.
      • Palinski W.
      • Schmid-Schönbein G.W.
      Activated monocytes and granulocytes, capillary nonperfusion, and neovascularization in diabetic retinopathy.
      redirecting glial attack against retina cells,
      • Kadłubowska J.
      • Malaguarnera L.
      • Wąż P.
      • et al.
      Neurodegeneration and neuroinflammation in diabetic retinopathy: potential approaches to delay neuronal loss.
      • Inman D.M.
      • Lupien C.B.
      • Horner P.J.
      Manipulating glia to protect retinal ganglion cells in glaucoma.
      and altering the biochemical changes that promote cell degeneration and death.
      • Shoeb Ahmad S.
      • Abdul Ghani S.
      • Hemalata Rajagopal T.
      Current concepts in the biochemical mechanisms of glaucomatous neurodegeneration.
      • Ola M.S.
      • Nawaz M.I.
      • Siddiquei M.M.
      • Al-Amro S.
      • Abu El-Asrar A.M.
      Recent advances in understanding the biochemical and molecular mechanism of diabetic retinopathy.
      • Wang J.T.
      • Medress Z.A.
      • Barres B.A.
      Axon degeneration: molecular mechanisms of a self-destruction pathway.
      • Lorenzi M.
      The polyol pathway as a mechanism for diabetic retinopathy: attractive, elusive, and resilient.
      • Kuehn M.H.
      • Fingert J.H.
      • Kwon Y.H.
      Retinal ganglion cell death in glaucoma: mechanisms and neuroprotective strategies.
      • Brownlee M.
      Biochemistry and molecular cell biology of diabetic complications.
      • Almasieh M.
      • Wilson A.M.
      • Morquette B.
      • et al.
      The molecular basis of retinal ganglion cell death in glaucoma.
      • Nickells R.W.
      The cell and molecular biology of glaucoma: mechanisms of retinal ganglion cell death.
      It may be possible to halt disease progression at a stage where anti-VEGF agents are not the mainstay therapy for diabetic retinopathy, or pressure-reducing efforts are not the main treatment in glaucoma.
      Funding/Support: Supported in part by the Macula Foundation , New York, NY.

      References

        • Early Treatment Diabetic Retinopathy Study Research Group
        Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10.
        Ophthalmology. 1991; 98: 786-806
        • Downs J.C.
        • Girkin C.A.
        Lamina cribrosa in glaucoma.
        Curr Opin Ophthalmol. 2017; 28: 113-119
        • Kanamori A.
        • Catrinescu M.M.
        • Belisle J.M.
        • et al.
        Retrograde and Wallerian axonal degeneration occur synchronously after retinal ganglion cell axotomy.
        Am J Pathol. 2012; 181: 62-73
        • Antonetti D.A.
        • Barber A.J.
        • Bronson S.K.
        • et al.
        Diabetic retinopathy: seeing beyond glucose-induced microvascular disease.
        Diabetes. 2006; 55: 2401-2411
        • Henson D.B.
        • North R.V.
        Dark adaptation in diabetes mellitus.
        Br J Ophthalmol. 1979; 63: 539-541
        • Midena E.
        • Segato T.
        • Giuliano M.
        • Zucchetto M.
        Macular recovery function (nyctometry) in diabetics without and with early retinopathy.
        Br J Ophthalmol. 1990; 74: 106-108
        • Joltikov K.A.
        • de Castro V.M.
        • Davila J.R.
        • et al.
        Multidimensional functional and structural evaluation reveals neuroretinal impairment in early diabetic retinopathy.
        Invest Ophthalmol Vis Sci. 2017; 58: BIO277-BIO290
        • Ostroy S.E.
        • Frede S.M.
        • Wagner E.F.
        • et al.
        Decreased rhodopsin regeneration in diabetic mouse eyes.
        Invest Ophthalmol Vis Sci. 1994; 35: 3905-3909
        • Trick G.L.
        • Burde R.M.
        • Gordon M.O.
        • et al.
        The relationship between hue discrimination and contrast sensitivity deficits in patients with diabetes mellitus.
        Ophthalmology. 1988; 95: 693-698
        • McAnany J.J.
        • Park J.C.
        Reduced contrast sensitivity is associated with elevated equivalent intrinsic noise in type 2 diabetics who have mild or no retinopathy.
        Invest Ophthalmol Vis Sci. 2018; 59: 2652-2658
        • Sokol S.
        • Moskowitz A.
        • Skarf B.
        • et al.
        Contrast sensitivity in diabetics with and without background retinopathy.
        Arch Ophthalmol. 1985; 103: 51-54
        • Ghafour I.M.
        • Foulds W.S.
        • Allan D.
        • McClure E.
        Contrast sensitivity in diabetic subjects with and without retinopathy.
        Br J Ophthalmol. 1982; 66: 492-495
        • Della Sala S.
        • Bertoni G.
        • Somazzi L.
        • et al.
        Impaired contrast sensitivity in diabetic patients with and without retinopathy: a new technique for rapid assessment.
        Br J Ophthalmol. 1985; 69: 136-142
        • Gella L.
        • Raman R.
        • Kulothungan V.
        • et al.
        Retinal sensitivity in subjects with type 2 diabetes mellitus: Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetics Study (SN-DREAMS II, Report No. 4).
        Br J Ophthalmol. 2016; 100: 808-813
        • Midena E.
        • Vujosevic S.
        Microperimetry in diabetic retinopathy.
        Saudi J Ophthalmol. 2011; 25: 131-135
        • Yang Q.
        • Xu Y.
        • Xie P.
        • et al.
        Retinal neurodegeneration in db/db mice at the early period of diabetes.
        J Ophthalmol. 2015; 2015: 757412
        • Tzekov R.
        • Arden G.B.
        The electroretinogram in diabetic retinopathy.
        Surv Ophthalmol. 1999; 44: 53-60
        • Pescosolido N.
        • Barbato A.
        • Stefanucci A.
        • Buomprisco G.
        Role of electrophysiology in the early diagnosis and follow-up of diabetic retinopathy.
        J Diabetes Res. 2015; 2015: 319692
        • Zhou W.
        • Rangaswamy N.
        • Ktonas P.
        • Frishman L.J.
        Oscillatory potentials of the slow-sequence multifocal ERG in primates extracted using the Matching Pursuit method.
        Vision Res. 2007; 47: 2021-2036
        • Di Leo M.A.
        • Caputo S.
        • Falsini B.
        • et al.
        Presence and further development of retinal dysfunction after 3-year follow up in IDDM patients without angiographically documented vasculopathy.
        Diabetologia. 1994; 37: 911-916
        • Bearse Jr., M.A.
        • Adams A.J.
        • Han Y.
        • et al.
        A multifocal electroretinogram model predicting the development of diabetic retinopathy.
        Prog Retin Eye Res. 2006; 25: 425-448
        • Tanvir Z.
        • Nelson R.F.
        • DeCicco-Skinner K.
        • Connaughton V.P.
        One month of hyperglycemia alters spectral responses of the zebrafish photopic electroretinogram.
        Dis Model Mech. 2018; 11
        • Ewing F.M.
        • Deary I.J.
        • Strachan M.W.
        • Frier B.M.
        Seeing beyond retinopathy in diabetes: electrophysiological and psychophysical abnormalities and alterations in vision.
        Endocr Rev. 1998; 19: 462-476
        • Siah W.F.
        • O'Brien C.
        • Loughman J.J.
        Macular pigment is associated with glare-affected visual function and central visual field loss in glaucoma.
        Br J Ophthalmol. 2018; 102: 929-935
        • Pacheco-Cutillas M.
        • Edgar D.F.
        • Sahraie A.
        Acquired colour vision defects in glaucoma-their detection and clinical significance.
        Br J Ophthalmol. 1999; 83: 1396-1402
        • Greenstein V.C.
        • Halevy D.
        • Zaidi Q.
        • et al.
        Chromatic and luminance systems deficits in glaucoma.
        Vision Res. 1996; 36: 621-629
        • Alvarez S.L.
        • Pierce G.E.
        • Vingrys A.J.
        • et al.
        Comparison of red-green, blue-yellow and achromatic losses in glaucoma.
        Vision Res. 1997; 37: 2295-2301
        • Kurtenbach A.
        • Kernstock C.
        • Zrenner E.
        • et al.
        Electrophysiology and colour: a comparison of methods to evaluate inner retinal function.
        Doc Ophthalmol. 2015; 131: 159-167
        • Zanen A.
        • Baeck A.
        Color vision and contrast sensitivity tests in early diagnosis of primary open-angle glaucoma.
        Bull Soc Belge Ophtalmol. 1992; 244: 73-83
        • Niwa Y.
        • Muraki S.
        • Naito F.
        • et al.
        Evaluation of acquired color vision deficiency in glaucoma using the Rabin cone contrast test.
        Invest Ophthalmol Vis Sci. 2014; 55: 6686-6690
        • McKendrick A.M.
        • Sampson G.P.
        • Walland M.J.
        • Badcock D.R.
        Contrast sensitivity changes due to glaucoma and normal aging: low-spatial-frequency losses in both magnocellular and parvocellular pathways.
        Invest Ophthalmol Vis Sci. 2007; 48: 2115-2122
        • Atkin A.
        • Bodis-Wollner I.
        • Wolkstein M.
        • et al.
        Abnormalities of central contrast sensitivity in glaucoma.
        Am J Ophthalmol. 1979; 88: 205-211
        • Hawkins A.S.
        • Szlyk J.P.
        • Ardickas Z.
        • et al.
        Comparison of contrast sensitivity, visual acuity, and Humphrey visual field testing in patients with glaucoma.
        J Glaucoma. 2003; 12: 134-138
        • Karaśkiewicz J.
        • Drobek-Słowik M.
        • Lubiński W.
        Pattern electroretinogram (PERG) in the early diagnosis of normal-tension preperimetric glaucoma: a case report.
        Doc Ophthalmol. 2014; 128: 53-58
        • Wilsey L.J.
        • Fortune B.
        Electroretinography in glaucoma diagnosis.
        Curr Opin Ophthalmol. 2016; 27: 118-124
        • Wilsey L.
        • Gowrisankaran S.
        • Cull G.
        • et al.
        Comparing three different modes of electroretinography in experimental glaucoma: diagnostic performance and correlation to structure.
        Doc Ophthalmol. 2017; 134: 111-128
        • Early Treatment Diabetic Retinopathy Study Research Group
        Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12.
        Ophthalmology. 1991; 98: 823-833
        • Snodderly D.M.
        • Weinhaus R.S.
        • Choi J.C.
        Neural-vascular relationships in central retina of macaque monkeys (Macaca fascicularis).
        J Neurosci. 1992; 12: 1169-1193
        • Warner R.M.
        Applied Statistics: From Bivariate Through Multivariate Techniques.
        Second Edition. Sage Publications, Thousand Oaks, CA2013
        • Preacher K.J.
        (2002, May). Calculation for the test of the difference between two independent correlation coefficients [Computer software].
        (Available at:) (Accessed July 5, 2019)
        • Büssow H.
        The astrocytes in the retina and optic nerve head of mammals: a special glia for the ganglion cell axons.
        Cell Tissue Res. 1980; 206: 367-378
        • Ogden T.E.
        Nerve fiber layer of the primate retina: thickness and glial content.
        Vision Res. 1983; 23: 581-587
        • Frenkel S.
        • Goshen G.
        • Leach L.
        • et al.
        Peripapillary distribution of Muller cells within the retinal nerve fiber layer in human eyes.
        Exp Eye Res. 2018; 166: 91-95
        • Curcio C.
        • Allen K.
        Topography of ganglion cells in human retina.
        J Comp Neurol. 1990; 300: 5-25
        • Jeon C.J.
        • Strettoi E.
        • Masland R.H.
        The major cell populations of the mouse retina.
        J Neurosci. 1998; 18: 8936-8946
        • Sohn E.H.
        • van Dijk H.W.
        • Jiao C.
        • et al.
        Retinal neurodegeneration may precede microvascular changes characteristic of diabetic retinopathy in diabetes mellitus.
        Proc Natl Acad Sci U S A. 2016; 113: E2655-E2664
        • Weber A.J.
        • Kaufman P.L.
        • Hubbard W.C.
        Morphology of single ganglion cells in the glaucomatous primate retina.
        Invest Ophthalmol Vis Sci. 1998; 39: 2304-2320
        • Bosco A.
        • Breen K.T.
        • Anderson S.R.
        • et al.
        Glial coverage in the optic nerve expands in proportion to optic axon loss in chronic mouse glaucoma.
        Exp Eye Res. 2016; 150: 34-43
        • Cull G.A.
        • Reynaud J.
        • Wang L.
        • et al.
        Relationship between orbital optic nerve axon counts and retinal nerve fiber layer thickness measured by spectral domain optical coherence tomography.
        Invest Ophthalmol Vis Sci. 2012; 53: 7766-7773
        • Sanes J.R.
        • Masland R.H.
        The types of retinal ganglion cells: current status and implications for neuronal classification.
        Annu Rev Neurosci. 2015; 38: 221-246
        • Yarmohammadi A.
        • Zangwill L.M.
        • Diniz-Filho A.
        • et al.
        Relationship between optical coherence tomography angiography vessel density and severity of visual field loss in glaucoma.
        Ophthalmology. 2016; 123: 2498-2508
        • Yarmohammadi A.
        • Zangwill L.M.
        • Diniz-Filho A.
        • et al.
        Optical coherence tomography angiography vessel density in healthy, glaucoma suspect, and glaucoma eyes.
        Invest Ophthalmol Vis Sci. 2016; 57: OCT451-OCT459
        • Chung J.K.
        • Hwang Y.H.
        • Wi J.M.
        • et al.
        Glaucoma diagnostic ability of the optical coherence tomography angiography vessel density parameters.
        Curr Eye Res. 2017; 42: 1458-1467
        • Rao H.L.
        • Pradhan Z.S.
        • Weinreb R.N.
        • et al.
        Regional comparisons of optical coherence tomography angiography vessel density in primary open-angle glaucoma.
        Am J Ophthalmol. 2016; 171: 75-83
        • Liu L.
        • Jia Y.
        • Takusagawa H.L.
        • et al.
        Optical coherence tomography angiography of the peripapillary retina in glaucoma.
        JAMA Ophthalmol. 2015; 133: 1045-1052
        • Takusagawa H.L.
        • Liu L.
        • Ma K.N.
        • et al.
        Projection-resolved optical coherence tomography angiography of macular retinal circulation in glaucoma.
        Ophthalmology. 2017; 124: 1589-1599
        • Chen H.S.
        • Liu C.H.
        • Wu W.C.
        • et al.
        Optical coherence tomography angiography of the superficial microvasculature in the macular and peripapillary areas in glaucomatous and healthy eyes.
        Invest Ophthalmol Vis Sci. 2017; 58: 3637-3645
        • Triolo G.
        • Rabiolo A.
        • Shemonski N.D.
        • et al.
        Optical coherence tomography angiography macular and peripapillary vessel perfusion density in healthy subjects, glaucoma suspects, and glaucoma patients.
        Invest Ophthalmol Vis Sci. 2017; 58: 5713-5722
        • Chen C.L.
        • Zhang A.
        • Bojikian K.D.
        • et al.
        Peripapillary retinal nerve fiber layer vascular microcirculation in glaucoma using optical coherence tomography-based microangiography.
        Invest Ophthalmol Vis Sci. 2016; 57: OCT475-OCT485
        • Sakaguchi K.
        • Higashide T.
        • Udagawa S.
        • et al.
        Comparison of sectoral structure-function relationships in glaucoma: vessel density versus thickness in the peripapillary retinal nerve fiber layer.
        Invest Ophthalmol Vis Sci. 2017; 58: 5251-5262
        • Scripsema N.K.
        • Garcia P.M.
        • Bavier R.D.
        • et al.
        Optical coherence tomography angiography analysis of perfused peripapillary capillaries in primary open-angle glaucoma and normal-tension glaucoma.
        Invest Ophthalmol Vis Sci. 2016; 57: OCT611-OCT620
        • Shin J.W.
        • Sung K.R.
        • Lee J.Y.
        • et al.
        Optical coherence tomography angiography vessel density mapping at various retinal layers in healthy and normal tension glaucoma eyes.
        Graefes Arch Clin Exp Ophthalmol. 2017; 255: 1193-1202
        • Chihara E.
        • Dimitrova G.
        • Amano H.
        • Chihara T.
        Discriminatory power of superficial vessel density and prelaminar vascular flow index in eyes with glaucoma and ocular hypertension and normal eyes.
        Invest Ophthalmol Vis Sci. 2017; 58: 690-697
        • Kumar R.S.
        • Anegondi N.
        • Chandapura R.S.
        • et al.
        Discriminant function of optical coherence tomography angiography to determine disease severity in glaucoma.
        Invest Ophthalmol Vis Sci. 2016; 57: 6079-6088
        • Kornzweig A.L.
        • Eliasoph I.
        • Feldstein M.
        Selective atrophy of the radial peripapillary capillaries in chronic glaucoma.
        Arch Ophthalmol. 1968; 80: 696-702
        • van Dijk H.W.
        • Verbraak F.D.
        • Kok P.H.
        • et al.
        Decreased retinal ganglion cell layer thickness in patients with type 1 diabetes.
        Invest Ophthalmol Vis Sci. 2010; 51: 3660-3665
        • Srinivasan S.
        • Pritchard N.
        • Sampson G.P.
        • et al.
        Focal loss volume of ganglion cell complex in diabetic neuropathy.
        Clin Exp Optom. 2016; 99: 526-534
        • Hegazy A.I.
        • Zedan R.H.
        • Macky T.A.
        • Esmat S.M.
        Retinal ganglion cell complex changes using spectral domain optical coherence tomography in diabetic patients without retinopathy.
        Int J Ophthalmol. 2017; 10: 427-433
        • Carpineto P.
        • Toto L.
        • Aloia R.
        • et al.
        Neuroretinal alterations in the early stages of diabetic retinopathy in patients with type 2 diabetes mellitus.
        Eye (Lond). 2016; 30: 673-679
        • Salvi L.
        • Plateroti P.
        • Balducci S.
        • et al.
        Abnormalities of retinal ganglion cell complex at optical coherence tomography in patients with type 2 diabetes: a sign of diabetic polyneuropathy, not retinopathy.
        J Diabetes Complications. 2016; 30: 469-476
        • Barber A.J.
        • Lieth E.
        • Khin S.A.
        • et al.
        Neural apoptosis in the retina during experimental and human diabetes. Early onset and effect of insulin.
        J Clin Invest. 1998; 102: 783-791
        • Oshitari T.
        • Yamamoto S.
        • Hata N.
        • Roy S.
        Mitochondria- and caspase-dependent cell death pathway involved in neuronal degeneration in diabetic retinopathy.
        Br J Ophthalmol. 2008; 92: 552-556
        • Valverde A.M.
        • Miranda S.
        • García-Ramírez M.
        • et al.
        Proapoptotic and survival signaling in the neuroretina at early stages of diabetic retinopathy.
        Mol Vis. 2013; 19: 47-53
        • Kern T.S.
        • Barber A.J.
        Retinal ganglion cells in diabetes.
        J Physiol. 2008; 586: 4401-4408
        • Wolter J.R.
        Diabetic retinopathy.
        Am J Ophthalmol. 1961; 51: 1123-1141
        • Park S.H.
        • Park J.W.
        • Park S.J.
        • et al.
        Apoptotic death of photoreceptors in the streptozotocin-induced diabetic rat retina.
        Diabetologia. 2003; 46: 1260-1268
        • Chen X.
        • Nie C.
        • Gong Y.
        • et al.
        Peripapillary retinal nerve fiber layer changes in preclinical diabetic retinopathy: a meta-analysis.
        PLoS One. 2015; 10: e0125919
        • Dumitrescu A.G.
        • Istrate S.L.
        • Iancu R.C.
        • et al.
        Retinal changes in diabetic patients without diabetic retinopathy.
        Rom J Ophthalmol. 2017; 61: 249-255
        • Jeon S.J.
        • Kwon J.W.
        • La T.Y.
        • et al.
        Characteristics of retinal nerve fiber layer defect in nonglaucomatous eyes with type II diabetes.
        Invest Ophthalmol Vis Sci. 2016; 57: 4008-4015
        • Eraslan M.
        • Çerman E.
        • Çekiç O.
        • et al.
        Neurodegeneration in ocular and central nervous systems: optical coherence tomography study in normal-tension glaucoma and Alzheimer disease.
        Turk J Med Sci. 2015; 45: 1106-1114
        • Kaur M.
        • Saxena R.
        • Singh D.
        • et al.
        Correlation between structural and functional retinal changes in Parkinson disease.
        J Neuroophthalmol. 2015; 35: 254-258
        • Jones-Odeh E.
        • Hammond C.J.
        How strong is the relationship between glaucoma, the retinal nerve fibre layer, and neurodegenerative diseases such as Alzheimer's disease and multiple sclerosis?.
        Eye (Lond). 2015; 29: 1270-1284
        • Marziani E.
        • Pomati S.
        • Ramolfo P.
        • et al.
        Evaluation of retinal nerve fiber layer and ganglion cell layer thickness in Alzheimer's disease using spectral-domain optical coherence tomography.
        Invest Ophthalmol Vis Sci. 2013; 54: 5953-5958
        • Kalenderoglu A.
        • Sevgi-Karadag A.
        • Celik M.
        • et al.
        Can the retinal ganglion cell layer (GCL) volume be a new marker to detect neurodegeneration in bipolar disorder?.
        Compr Psychiatry. 2016; 67: 66-72
        • Celik M.
        • Kalenderoglu A.
        • Sevgi Karadag A.
        • et al.
        Decreases in ganglion cell layer and inner plexiform layer volumes correlate better with disease severity in schizophrenia patients than retinal nerve fiber layer thickness: findings from spectral optic coherence tomography.
        Eur Psychiatry. 2016; 32: 9-15
        • Tugcu B.
        • Soysal A.
        • Kılıc M.
        • et al.
        Assessment of structural and functıonal vısual outcomes ın relapsıng remıttıng multıple sclerosıs wıth vısual evoked potentıals and optıcal coherence tomography.
        J Neurol Sci. 2013; 335: 182-185
        • Huseyinoglu N.
        • Ekinci M.
        • Ozben S.
        • et al.
        Optic disc and retinal nerve fiber layer parameters as indicators of neurodegenerative brain changes in patients with obstructive sleep apnea syndrome.
        Sleep Breath. 2014; 18: 95-102
        • Mutlu U.
        • Bonnemaijer P.W.M.
        • Ikram M.A.
        • et al.
        Retinal neurodegeneration and brain MRI markers: the Rotterdam Study.
        Neurobiol Aging. 2017; 60: 183-191
        • Nork T.M.
        • Ver Hoeve J.N.
        • Poulsen G.L.
        • et al.
        Swelling and loss of photoreceptors in chronic human and experimental glaucomas.
        Arch Ophthalmol. 2000; 118: 235-245
        • Alnawaiseh M.
        • Lahme L.
        • Müller V.
        • et al.
        Correlation of flow density, as measured using optical coherence tomography angiography, with structural and functional parameters in glaucoma patients.
        Graefes Arch Clin Exp Ophthalmol. 2018; 256: 589-597
        • Lommatzsch C.
        • Rothaus K.
        • Koch J.M.
        • et al.
        OCTA vessel density changes in the macular zone in glaucomatous eyes.
        Graefes Arch Clin Exp Ophthalmol. 2018; 256: 1499-1508
        • Graefe A.
        On iridectomy in glaucoma and the glaucomatous process.
        in: Graefe A. Three Memoirs on Iridectomy: In Certain Forms of Iritis, Choroiditis, and Glaucoma. New Sydenhan Society, London1859 (translated by Windsor T)
        • von Jaeger E.
        Ueber Glaucom und seine Heilung durch Iridectomie.
        Z Ges Aerzte Wein. 1858; 14: 465-484
        • Heijl A.
        • Leske M.C.
        • Bengtsson B.
        • et al.
        Reduction of intraocular pressure and glaucoma progression: results from the Early Manifest Glaucoma Trial.
        Arch Ophthalmol. 2002; 120: 1268-1279
        • The Diabetic Retinopathy Study Research Group
        Photocoagulation treatment of proliferative diabetic retinopathy. Clinical application of Diabetic Retinopathy Study (DRS) findings, DRS Report Number 8.
        Ophthalmology. 1981; 88: 583-600
        • Early Treatment Diabetic Retinopathy Study Research Group
        Early photocoagulation for diabetic retinopathy. ETDRS report number 9.
        Ophthalmology. 1991; 98: 766-785
        • Nguyen Q.D.
        • Brown D.M.
        • Marcus D.M.
        • et al.
        Ranibizumab for diabetic macular edema: results from 2 phase III randomized trials: RISE and RIDE.
        Ophthalmology. 2012; 119: 789-801
        • Gross J.G.
        • Glassman A.R.
        • Liu D.
        • et al.
        Five-year outcomes of panretinal photocoagulation vs intravitreous ranibizumab for proliferative diabetic retinopathy: a randomized clinical trial.
        JAMA Ophthalmol. 2018; 136: 1138-1148
        • Hood D.C.
        • De Moraes C.G.
        Challenges to the common clinical paradigm for diagnosis of glaucomatous damage with OCT and visual fields.
        Invest Ophthalmol Vis Sci. 2018; 59: 788-791
        • Williams P.A.
        • Marsh-Armstrong N.
        • Howell G.R.
        • et al.
        Neuroinflammation in glaucoma: a new opportunity.
        Exp Eye Res. 2017; 157: 20-27
        • de Hoz R.
        • Rojas B.
        • Ramírez A.I.
        • et al.
        Retinal macroglial responses in health and disease.
        Biomed Res Int. 2016; 2016: 2954721
        • Barber A.J.
        A new view of diabetic retinopathy: a neurodegenerative disease of the eye.
        Prog Neuropsychopharmacol Biol Psychiatry. 2003; 27: 283-290
        • Prada D.
        • Harris A.
        • Guidoboni G.
        • et al.
        Autoregulation and neurovascular coupling in the optic nerve head.
        Surv Ophthalmol. 2016; 61: 164-186
        • Silverman S.M.
        • Wong W.T.
        Microglia in the retina: roles in development, maturity, and disease.
        Annu Rev Vis Sci. 2018; 4: 45-77
        • Gauthier A.C.
        • Liu J.
        Neurodegeneration and neuroprotection in glaucoma.
        Yale J Biol Med. 2016; 89: 73-79
        • Zeng H.L.
        • Shi J.M.
        The role of microglia in the progression of glaucomatous neurodegeneration- a review.
        Int J Ophthalmol. 2018; 11: 143-149
        • Vlasits A.L.
        • Euler T.
        • Franke K.
        Function first: classifying cell types and circuits of the retina.
        Curr Opin Neurobiol. 2018; 56: 8-15
        • Rübsam A.
        • Parikh S.
        • Fort P.E.
        Role of inflammation in diabetic retinopathy.
        Int J Mol Sci. 2018; 19 (pii: E942)
        • Vohra R.
        • Tsai J.C.
        • Kolko M.
        The role of inflammation in the pathogenesis of glaucoma.
        Surv Ophthalmol. 2013; 58: 311-320
        • Gupta N.
        • Yücel Y.H.
        Glaucoma as a neurodegenerative disease.
        Curr Opin Ophthalmol. 2007; 18: 110-114
        • Danesh-Meyer H.V.
        • Levin L.A.
        Glaucoma as a neurodegenerative disease.
        J Neuroophthalmol. 2015; 35: S22-S28
        • Barber A.J.
        Diabetic retinopathy: recent advances towards understanding neurodegeneration and vision loss.
        Sci China Life Sci. 2015; 58: 541-549
        • Simó R.
        • Hernández C.
        • European Consortium for the Early Treatment of Diabetic Retinopathy
        Neurodegeneration in the diabetic eye: new insights and therapeutic perspectives.
        Trends Endocrinol Metab. 2014; 25: 23-33
        • Schröder S.
        • Palinski W.
        • Schmid-Schönbein G.W.
        Activated monocytes and granulocytes, capillary nonperfusion, and neovascularization in diabetic retinopathy.
        Am J Pathol. 1991; 139: 81-100
        • Kadłubowska J.
        • Malaguarnera L.
        • Wąż P.
        • et al.
        Neurodegeneration and neuroinflammation in diabetic retinopathy: potential approaches to delay neuronal loss.
        Curr Neuropharmacol. 2016; 14: 831-839
        • Inman D.M.
        • Lupien C.B.
        • Horner P.J.
        Manipulating glia to protect retinal ganglion cells in glaucoma.
        in: Gunvant P. Glaucoma-Current Clinical and Research Aspects. InTech, Rijeka, Croatia2011: 26-50
        • Shoeb Ahmad S.
        • Abdul Ghani S.
        • Hemalata Rajagopal T.
        Current concepts in the biochemical mechanisms of glaucomatous neurodegeneration.
        J Curr Glaucoma Pract. 2013; 7: 49-53
        • Ola M.S.
        • Nawaz M.I.
        • Siddiquei M.M.
        • Al-Amro S.
        • Abu El-Asrar A.M.
        Recent advances in understanding the biochemical and molecular mechanism of diabetic retinopathy.
        J Diabetes Complications. 2012; 26: 56-64
        • Wang J.T.
        • Medress Z.A.
        • Barres B.A.
        Axon degeneration: molecular mechanisms of a self-destruction pathway.
        J Cell Biol. 2012; 196: 7-18
        • Lorenzi M.
        The polyol pathway as a mechanism for diabetic retinopathy: attractive, elusive, and resilient.
        Exp Diabetes Res. 2007; 2007: 61038
        • Kuehn M.H.
        • Fingert J.H.
        • Kwon Y.H.
        Retinal ganglion cell death in glaucoma: mechanisms and neuroprotective strategies.
        Ophthalmol Clin North Am. 2005; 18: 383-395.vi
        • Brownlee M.
        Biochemistry and molecular cell biology of diabetic complications.
        Nature. 2001; 414: 813-820
        • Almasieh M.
        • Wilson A.M.
        • Morquette B.
        • et al.
        The molecular basis of retinal ganglion cell death in glaucoma.
        Prog Retin Eye Res. 2012; 31: 152-181
        • Nickells R.W.
        The cell and molecular biology of glaucoma: mechanisms of retinal ganglion cell death.
        Invest Ophthalmol Vis Sci. 2012; 53: 2476-2481