ORIGINAL STUDY
Comparison of Optic Nerve Head Topographic Parameters in Patients With Primary Open-Angle Glaucoma With and Without Diabetes Mellitus Serkan Akkaya, MD,* Ertug˘rul Can, MD,w and Faruk O¨ztu¨rk, MDz
Purpose: To evaluate the optic nerve head (ONH) parameters of primary open-angle glaucoma (POAG) patients with and without diabetes and to investigate the effect of the metabolic control of diabetes on ONH topography. Materials and Methods: A study group of 60 eyes of 60 POAG patients with type 2 diabetes mellitus and a control group of 41 eyes of 41 POAG patients without diabetes were recruited for the study. Complete ophthalmic examinations of all patients were performed and the quantitative optic disc parameters were evaluated with Heidelberg retina tomography (HRT) III. HbA1c measurements of diabetic patients were assessed on the same day when the ophthalmic assessments were performed. Results: Mean rim area and rim volume of the study group was statistically higher than the control group (P = 0.04 for rim area and P = 0.02 for rim volume). The difference in other parameters of the HRT between the groups were insignificant (P > 0.05). In the study group, the duration of the diabetes was not significantly correlated to rim area and rim volume (r = 0.03, P = 0.81 for rim area; r = 0.03, P = 0.79 for rim volume). Analysis of covariance showed insignificant effects of age, sex, MD, and PSD values over rim area and rim volume (P > 0.05). There were weak but statistically significant correlations between HbA1c levels and some HRT parameters including disc area, cup area, cup volume, and cup shape measure (r = 0.35 and P = 0.006 for disc area; r = 0.35 and P = 0.006 for cup area; r = 0.32 and P = 0.01 for cup volume; r = 0.32 and P = 0.01 for cup shape measure). Conclusions: The results of this study imply the protective effect of diabetes over glaucomatous optic nerve damage in POAG patients. However, larger and controlled studies are warranted to confirm those findings. Key Words: glaucoma, Heidelberg retinal tomography, diabetes
(J Glaucoma 2016;25:49–53)
P
rimary open-angle glaucoma (POAG) is a chronic progressive optic neuropathy that causes characteristic acquired optic atrophy due to the loss of retinal ganglion cells (RGCs) and their axons.1 Diabetes mellitus (DM) is a systemic disease characterized by hyperglycemia resulting from insulin deficiency or insulin resistance. Increased apoptosis of RGCs Received for publication July 21, 2013; accepted July 4, 2014. From the *Department of Ophthalmology, Dıs¸kapı Yıldırım Beyazıt Training and Research Hospital; zDepartment of Ophthalmology, Faculty of Medicine, Yıldırım Beyazıt University, Ankara; and wDepartment of Ophthalmology, Faculty of Medicine, Ondokuz Mayıs University, Samsun, Turkey. Disclosure: The authors declare no conflict of interest. Reprints: Serkan Akkaya, MD, Department of Ophthalmology, Dıs¸kapı Yıldırım Beyazıt Eg˘itim ve Aras¸tırma Hastanesi, Ankara 06610, Turkey (e-mail: [email protected]). Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/IJG.0000000000000096
J Glaucoma
Volume 25, Number 1, January 2016
in experimental diabetes has been reported.2 It has also been reported that several proapoptotic molecules are expressed by ganglion cells of diabetic retinas and RGC death occurs early in diabetic eyes.3 Experimental studies on animal models showed that the retrograde axonal transport in large-sized and medium-sized RGCs are damaged just before the optic nerve involvement.4,5 Several factors such as ischemia, increased activity of the sorbitol pathway, abnormal myoinositol metabolism, and decreased protein synthesis are discussed in the pathogenesis of diabetes and these factors may probably lead to damage of peripheral nerve fibers.6 Increased accumulation of advanced glycation end products in cribriform plates and in the vicinity of optic nerve vessels may play a role in the development of optic neuropathy in diabetes.7 The Heidelberg retina tomography (HRT) III is a confocal scanning laser ophthalmoscope that uses laser diode light. It takes data from tomographic images of the optic nerve head (ONH) and provides a retinal nerve fiber layer (RNFL) thickness based upon the retinal surface height, aligns the images, and computes a 3D topographic map of the surface of the retina. The aim of our study herein is to investigate the ONH topographic images in POAG patients with and without DM, and to evaluate the effects of disease duration and metabolic control of diabetes on ONH topography.
MATERIALS AND METHODS Subjects were prospectively selected among patients from the glaucoma clinic of the Ulucanlar Training and Research Hospital, Eye Clinic, Ankara, Turkey. The Ethics Committee of University of Ankara, Turkey approved the study and written informed consent was obtained from each participant after study had been fully explained. The subjects were divided into a study group of 60 eyes of 60 POAG patients with type 2 DM and a control group of 41 eyes of 41 POAG patients without diabetes. Glaucoma diagnosis was made in the presence of increased intraocular pressure, visual field defects, and increased cup-to-disc ratio. Demographic data were collected and the patients were asked about the duration of DM and existence of any systemic disease or family history of glaucoma. The same examiner performed all the assessments. All patients had a complete ophthalmic examination, including refraction, visual acuity, slit-lamp examination, and fundus examination. Intraocular pressure measurements were performed with Goldmann applanation tonometer. Iridocorneal angle examinations were performed using a Goldmann 3-mirror lens. Ultrasonic pachymetry (Nidek UP-1000 Ultrasonic Pachymeter; Nidek Co. Ltd., Japan) was used for central corneal thickness measurements. Visual field tests were performed using 24-2 SITA standard test strategy on the Humphrey field analyzer 750 (Zeiss/Humphrey Systems, www.glaucomajournal.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
49
Akkaya et al
J Glaucoma
Volume 25, Number 1, January 2016
TABLE 1. The Differences in Spherical Equivalent, Keratometry, Central Corneal Thickness, and Visual Field Parameters Between the Groups
Study Group Parameters Spherical equivalent (D) Axial length (mm) Keratometry (mm) Central corneal thickness (mm) Mean deviation of Humprey VF PSD of Humprey VF Intraocular pressure (mm Hg)
Mean ± SD
Median
0.15 ± 1.43 22.85 ± 0.82 7.69 ± 0.26 544.72 ± 31.45 3.35 ± 1.63 2.96 ± 1.21 15.03 ± 3.07
0.06 22.91 7.64 544.50 3.06 2.68 14
Control Group Range 4.62 21.06 7.29 465 5.38 1.34 8
to to to to to to to
4.00 24.72 8.48 624 0.73 6.49 25
Mean ± SD
Median
0.25 ± 1.81 23.10 ± 0.95 7.72 ± 0.28 538.59 ± 36.65 3.42 ± 1.83 3.15 ± 1.34 13.98 ± 3.08
0.12 23.15 7.67 534 3.38 3.10 14
Range
P
to to to to to to to
0.26 0.15 0.61 0.37 0.83 0.45 0.18
5.00 20.88 7.13 464 5.18 1.28 7
2.87 25.00 8.50 607 0.41 7.93 20
VF indicates visual field test.
Dublin, CA) and ONH were analyzed with the Heidelberg Retina Tomograph-III (HRT-III, software version 3.0; Heidelberg Engineering GMBH, Heidelberg, Germany). HbA1c measurements of diabetic patients were made on the same day when the ophthalmic assessments were performed. All type 2 DM patients were receiving oral antidiabetic medication or insulin therapy. Inclusion criteria were defined as follows: absence of any corneal pathology or any previous eye surgery, spherical refraction within ± 5.0 D and cylinder correction within ± 3.0 D, best corrected visual acuity of 20/40 or better, absence of severe media opacities such as corneal opacity, mature cataract, vitreous opacity, absence of optic nerve or retina pathologies other than glaucomatous optic neuropathy, and absence of any inflammatory eye disease or systemic collagen disease. Patients with POAG were graded according to the Hodapp-Anderson-Parrish grading scale with their visual field assessments. Only the patients with early stage of POAG were recruited into the study.8
STATISTICAL METHODS Analyses were performed using SPSS for Windows (version 15.0, SPSS Inc.). Normality of the groups was evaluated using Kolmogorov-Smirnov test. The student t test was used for parametric data and the Mann-Whitney U test was used for nonparametric data to compare the groups. Pearson correlation coefficient was used when both variables were measured and Spearman correlation coefficient was used when variables were nonparametric. Analysis of covariance
(ANCOVA) test was used to account the effects of age, sex, MD, and PSD values over significant differences between the groups. Age, sex, MD, and PSD values were set as covariates in Analysis of covariance (ANCOVA) test. Logarithmic transformation was used to convert nonparametric data to parametric data where needed. Results were reported as mean ± SD with 95% confidence intervals and P = 0.05 was considered as statistically significant.
RESULTS Sixty eyes of 60 POAG patients with diabetes (study group) and 41 eyes of 41 POAG patients without diabetes (control group) were included. Mean age of the study and the control groups were 59.1 ± 7.4 and 58.6 ± 7.7 years, respectively (P = 0.696). Female/male ratios were 38/22 and 25/16 for study and control groups, respectively (P = 0.837). In the study group, mean HbA1c level during HRT measurements was 8.00 ± 1.2% and mean disease (DM) duration was 9.2 ± 5.8 years. The differences in spherical equivalent, keratometry, central corneal thickness, visual field MD, and PSD differences between the groups were statistically insignificant (Table 1). HRT-based measurements (disc area, cup area, rim area, cup volume, rim volume, cup/disc area ratio, linear cup/disc ratio, mean cup depth, cup shape measure, mean RNFL thickness, and RNFL cross-section area) between the groups were investigated. Rim area and rim volume were significantly higher in POAG patients with diabetes. The remaining HRT differences were insignificant (Table 2). Statistical power of the
TABLE 2. The Differences in HRT Parameters Between the Groups
Study Group Parameters 2
Disc area (mm ) Cup area (mm2) Rim area (mm2) Cup volume (mm3) Rim volume (mm3) CD area ratio Linear CD ratio Mean cup depth (mm) Maximum cup depth (mm) Cup shape measure Mean RNFL thickness (mm) RNFL cross-sectional area (mm2) Height variation contour (mm)
Mean ± SD
Median
2.39 ± 0.51 0.75 ± 0.44 1.63 ± 0.42 0.21 ± 0.17 0.46 ± 0.20 0.30 ± 0.15 0.53 ± 0.15 0.27 ± 0.10 0.70 ± 0.21 0.15 ± 0.08 0.25 ± 0.09 1.38 ± 0.57 0.46 ± 0.18
2.39 0.73 1.63 0.18 0.43 0.32 0.56 0.29 0.71 0.13 0.26 1.42 0.43
Control Group Range
1.37 0.00 0.67 0.00 0.07 0.00 0.03 0.03 0.12 0.39 0.01 0.04 0.20
to to to to to to to to to to to to to
3.70 1.97 0.60 0.79 1.04 0.64 0.80 0.57 0.27 0.02 0.56 0.20 1.20
Mean ± SD
Median
2.31 ± 0.42 0.83 ± 0.40 1.47 ± 0.28 0.25 ± 0.18 0.36 ± 0.10 0.34 ± 0.13 0.57 ± 0.11 0.29 ± 0.11 0.75 ± 0.22 0.15 ± 0.07 0.23 ± 0.07 1.24 ± 0.35 0.41 ± 0.12
2.26 0.81 1.42 0.22 0.35 0.35 0.59 0.29 0.83 0.16 0.23 1.21 0.39
Range 1.62 0.17 0.84 0.02 0.20 0.09 0.30 0.11 0.37 0.30 0.12 0.57 0.20
to to to to to to to to to to to to to
3.80 2.00 2.06 0.91 0.67 0.64 0.80 0.56 1.29 0.03 0.45 2.41 0.80
P 0.36 0.38 0.04 0.18 0.02 0.16 0.19 0.33 0.26 0.44 0.23 0.12 0.18
CD indicates cup disc; HRT, Heidelberg retina tomography; RNFL, retinal nerve fiber layer.
50 | www.glaucomajournal.com Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma
Volume 25, Number 1, January 2016
differences in main outcome measures (disc area, cup area, rim area, cup volume, rim volume, cup-disc area ratio, mean cup depth, maximum cup depth, cup shape measure, mean RNFL, RNFL cross-sectional area, and height variation contour) ranged from 21.7% (disc area) to 100% (RNFL cross-sectional area). The statistical powers for 2 significantly different results were 0.74 and 0.95 for rim area and rim volume, respectively. The effects of possible confounding factors (age, sex, MD, and PSD) on the significant differences between the groups (rim area and rim volume) were investigated using ANCOVA test after logarithmic transformation of nonparametric data to parametric data. None of the possible confounding factors had significant effect on initial statistical comparisons. The statistical significances of age, sex, MD, and PSD after ANCOVA test were 0.957, 0.109, 0.932, and 0.595 for rim area and 0.359, 0. 153, 0.478, and 0.828 for rim volume, respectively. The correlation analyses in diabetic group showed that duration of diabetes and HbA1c levels were not significantly correlated to rim area and rim volume (Fig. 1).
Optic Nerve Head in Diabetic Glaucoma Patients
HbA1c levels were significantly correlated to disc area, cup area, cup volume, and cup shape measure (Fig. 2). Diabetic patients were divided into 2 groups according to disease duration. There were 27 patients with disease duration of r8 years and 33 patients with disease duration of Z10 years. Rim area and rim volume differences between those groups were also insignificant (P = 0.191, P = 0.203, respectively). Similarly, the DM patients were divided into 2 groups according to HbA1c levels. Mean ± SD for HbA1c levels were 6.97 ± 0.59 (n = 30) and 9.01 ± 0.82 (n = 30) for the groups. Rim area and rim volume did not differ significantly between the groups (P = 0.912, P = 0.894, respectively).
DISCUSSION The mechanism of pathologic changes in retinal neurons due to diabetes is extremely complex. Diabetes has several effects on different retinal cells such as neurons, vascular cells, glial cells, and microglial cells.9 It has been shown that some abnormalities occur in glial elements
FIGURE 1. The correlations between HbA1c and duration of diabetes to rim area and rim volume.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
www.glaucomajournal.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
51
Akkaya et al
J Glaucoma
Volume 25, Number 1, January 2016
FIGURE 2. The correlations between HbA1c and disc area, cup area, cup volume, and cup shape measure.
before the clinical manifestations of diabetic retinal vasculopathy.10–12 Barber et al10 reported the increase of apoptotic (TUNEL-positive) cells in human retinal neurons. They also proposed that significant decrease in the number of RGCs and diminution of the thickness of the inner plexiform and inner nuclear layers increase the likelihood that apoptotic cells in retina comprises ganglion cells and other neurons. In an experimental model of diabetes, Hammes et al2 showed that diabetes induces apoptosis in RGCs and Mu¨ller cells.2 In another study by Abu-El-Asrar et al,3 it has been reported that the number of promoting factors inducing apoptosis in RGCs is increased and the death of RGCs occur early in diabetic eyes. It has also been reported in experimental animal studies that diabetes induces impairment of retrograde axonal transport in largesized and medium-sized RGCs and this impairment precedes optic nerve involvement.4,5 There are limited published data which investigates the ONH parameters in diabetic POAG patients. Budde and Jonas compared the stereo disc photographs of 52 diabetic POAG patients with 312 nondiabetic POAG patients. There were no statistically significant differences regarding mean optic disc size, neural rim, alpha zone, and beta zone of parapapillary atrophy. Patients with proliferative diabetic retinopathy were not recruited in their study. The authors concluded that diabetes does not have a marked effect on ONH damage in POAG patients.13 Yu¨cel et al,14
reported that rim area, RNFL cross-sectional area, rim volume, and RNFL thickness correlated significantly with optic nerve fiber number in glaucomatous eyes. In our study, greater measurements of rim area and rim volume in diabetic POAG patients increases the likelihood that the number of optic nerve fibers in diabetic POAG patients is greater than that in nondiabetic POAG patients. It may be argued that optic nerve swelling may have caused increased rim area and rim volume measurements. However, in our study, rim area and rim volume measurements were not significantly correlated to HbA1c levels. In addition, rim area and rim volume differences between patients with disease duration of r8 and Z10 years, and between patients with lower and higher HbA1c levels were insignificant. This findings lessen (but does not omit) the possibility of the significant effect of subclinical optic nerve swelling on rim area and rim volume. Because, patients with higher HbA1c levels and longer disease duration are more likely to have optic nerve swelling. In this study, we did not find significant differences in cup area and cup volume between diabetic and nondiabetic POAG patients. In diabetic POAG patients, greater measurements of rim area and rim volume together with the absence of any increase in the optic cup depth lead us to consider that diabetes may cause protective effect on optic nerve and retinal nerve fibers in POAG. The persistence of the significant differences in rim area and rim volume
52 | www.glaucomajournal.com Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma
Volume 25, Number 1, January 2016
between the diabetic and nondiabetic patients after controlling the effects of age, sex, MD, and PSD values with ANCOVA test further stresses the validity of our findings. In our study, there was not any statistically significant difference between the duration of diabetes and topographic parameters of ONH. This result also lead us to consider that the duration of diabetes have no significant effect on ONH topography. There were weak but statistically significant correlations between HbA1c levels and HRT parameters including cup area, cup volume, and cup shape measure. This correlation implies that poor metabolic control of diabetes may have negative effect on optic cup depth. These findings show the necessity of longitudinal and prospective studies about the effect of the metabolic control of diabetes on ONH topography in POAG patients with diabetes. The limitations of the HRT technology include the need to outline the anterior scleral canal opening and a reference plane. Stereometric parameters are calculated upon this delineation. Coefficient of variation of 3 HRT measurements performed by the same technician were reported to be between 2.9% and 5.2%.15 However, interindividual variations are expected to be higher than those percentiles. In the most comprehensive model, it may be argued that the delineation of the anterior scleral canal in diabetic patients might have been assessed deeper than the nondiabetic patients (perhaps because they have some slightly higher degree of temporal RNFL loss) and this might have been resulted in higher rim area and rim volume. However, insignificant rim area and rim volume differences between patients with lower and higher HbA1c levels and between patients with shorter and longer disease duration makes this presumption almost improbable. We think that if diabetes would have caused temporal RNFL loss more than nondiabetic patients in our study, patients with higher HbA1c levels and longer disease duration would possibly have significantly lower rim area and rim volume. One of the limitations in this study is the low levels of statistical powers in some comparisons. Although statistical powers for rim area and rim volume comparisons were 0.74 and 0.95, most of the statistical powers for the remaining insignificant differences were
Comparison of Optic Nerve Head Topographic Parameters in Patients With Primary Open-Angle Glaucoma With and Without Diabetes Mellitus Serkan Akkaya, MD,* Ertug˘rul Can, MD,w and Faruk O¨ztu¨rk, MDz
Purpose: To evaluate the optic nerve head (ONH) parameters of primary open-angle glaucoma (POAG) patients with and without diabetes and to investigate the effect of the metabolic control of diabetes on ONH topography. Materials and Methods: A study group of 60 eyes of 60 POAG patients with type 2 diabetes mellitus and a control group of 41 eyes of 41 POAG patients without diabetes were recruited for the study. Complete ophthalmic examinations of all patients were performed and the quantitative optic disc parameters were evaluated with Heidelberg retina tomography (HRT) III. HbA1c measurements of diabetic patients were assessed on the same day when the ophthalmic assessments were performed. Results: Mean rim area and rim volume of the study group was statistically higher than the control group (P = 0.04 for rim area and P = 0.02 for rim volume). The difference in other parameters of the HRT between the groups were insignificant (P > 0.05). In the study group, the duration of the diabetes was not significantly correlated to rim area and rim volume (r = 0.03, P = 0.81 for rim area; r = 0.03, P = 0.79 for rim volume). Analysis of covariance showed insignificant effects of age, sex, MD, and PSD values over rim area and rim volume (P > 0.05). There were weak but statistically significant correlations between HbA1c levels and some HRT parameters including disc area, cup area, cup volume, and cup shape measure (r = 0.35 and P = 0.006 for disc area; r = 0.35 and P = 0.006 for cup area; r = 0.32 and P = 0.01 for cup volume; r = 0.32 and P = 0.01 for cup shape measure). Conclusions: The results of this study imply the protective effect of diabetes over glaucomatous optic nerve damage in POAG patients. However, larger and controlled studies are warranted to confirm those findings. Key Words: glaucoma, Heidelberg retinal tomography, diabetes
(J Glaucoma 2016;25:49–53)
P
rimary open-angle glaucoma (POAG) is a chronic progressive optic neuropathy that causes characteristic acquired optic atrophy due to the loss of retinal ganglion cells (RGCs) and their axons.1 Diabetes mellitus (DM) is a systemic disease characterized by hyperglycemia resulting from insulin deficiency or insulin resistance. Increased apoptosis of RGCs Received for publication July 21, 2013; accepted July 4, 2014. From the *Department of Ophthalmology, Dıs¸kapı Yıldırım Beyazıt Training and Research Hospital; zDepartment of Ophthalmology, Faculty of Medicine, Yıldırım Beyazıt University, Ankara; and wDepartment of Ophthalmology, Faculty of Medicine, Ondokuz Mayıs University, Samsun, Turkey. Disclosure: The authors declare no conflict of interest. Reprints: Serkan Akkaya, MD, Department of Ophthalmology, Dıs¸kapı Yıldırım Beyazıt Eg˘itim ve Aras¸tırma Hastanesi, Ankara 06610, Turkey (e-mail: [email protected]). Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/IJG.0000000000000096
J Glaucoma
Volume 25, Number 1, January 2016
in experimental diabetes has been reported.2 It has also been reported that several proapoptotic molecules are expressed by ganglion cells of diabetic retinas and RGC death occurs early in diabetic eyes.3 Experimental studies on animal models showed that the retrograde axonal transport in large-sized and medium-sized RGCs are damaged just before the optic nerve involvement.4,5 Several factors such as ischemia, increased activity of the sorbitol pathway, abnormal myoinositol metabolism, and decreased protein synthesis are discussed in the pathogenesis of diabetes and these factors may probably lead to damage of peripheral nerve fibers.6 Increased accumulation of advanced glycation end products in cribriform plates and in the vicinity of optic nerve vessels may play a role in the development of optic neuropathy in diabetes.7 The Heidelberg retina tomography (HRT) III is a confocal scanning laser ophthalmoscope that uses laser diode light. It takes data from tomographic images of the optic nerve head (ONH) and provides a retinal nerve fiber layer (RNFL) thickness based upon the retinal surface height, aligns the images, and computes a 3D topographic map of the surface of the retina. The aim of our study herein is to investigate the ONH topographic images in POAG patients with and without DM, and to evaluate the effects of disease duration and metabolic control of diabetes on ONH topography.
MATERIALS AND METHODS Subjects were prospectively selected among patients from the glaucoma clinic of the Ulucanlar Training and Research Hospital, Eye Clinic, Ankara, Turkey. The Ethics Committee of University of Ankara, Turkey approved the study and written informed consent was obtained from each participant after study had been fully explained. The subjects were divided into a study group of 60 eyes of 60 POAG patients with type 2 DM and a control group of 41 eyes of 41 POAG patients without diabetes. Glaucoma diagnosis was made in the presence of increased intraocular pressure, visual field defects, and increased cup-to-disc ratio. Demographic data were collected and the patients were asked about the duration of DM and existence of any systemic disease or family history of glaucoma. The same examiner performed all the assessments. All patients had a complete ophthalmic examination, including refraction, visual acuity, slit-lamp examination, and fundus examination. Intraocular pressure measurements were performed with Goldmann applanation tonometer. Iridocorneal angle examinations were performed using a Goldmann 3-mirror lens. Ultrasonic pachymetry (Nidek UP-1000 Ultrasonic Pachymeter; Nidek Co. Ltd., Japan) was used for central corneal thickness measurements. Visual field tests were performed using 24-2 SITA standard test strategy on the Humphrey field analyzer 750 (Zeiss/Humphrey Systems, www.glaucomajournal.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
49
Akkaya et al
J Glaucoma
Volume 25, Number 1, January 2016
TABLE 1. The Differences in Spherical Equivalent, Keratometry, Central Corneal Thickness, and Visual Field Parameters Between the Groups
Study Group Parameters Spherical equivalent (D) Axial length (mm) Keratometry (mm) Central corneal thickness (mm) Mean deviation of Humprey VF PSD of Humprey VF Intraocular pressure (mm Hg)
Mean ± SD
Median
0.15 ± 1.43 22.85 ± 0.82 7.69 ± 0.26 544.72 ± 31.45 3.35 ± 1.63 2.96 ± 1.21 15.03 ± 3.07
0.06 22.91 7.64 544.50 3.06 2.68 14
Control Group Range 4.62 21.06 7.29 465 5.38 1.34 8
to to to to to to to
4.00 24.72 8.48 624 0.73 6.49 25
Mean ± SD
Median
0.25 ± 1.81 23.10 ± 0.95 7.72 ± 0.28 538.59 ± 36.65 3.42 ± 1.83 3.15 ± 1.34 13.98 ± 3.08
0.12 23.15 7.67 534 3.38 3.10 14
Range
P
to to to to to to to
0.26 0.15 0.61 0.37 0.83 0.45 0.18
5.00 20.88 7.13 464 5.18 1.28 7
2.87 25.00 8.50 607 0.41 7.93 20
VF indicates visual field test.
Dublin, CA) and ONH were analyzed with the Heidelberg Retina Tomograph-III (HRT-III, software version 3.0; Heidelberg Engineering GMBH, Heidelberg, Germany). HbA1c measurements of diabetic patients were made on the same day when the ophthalmic assessments were performed. All type 2 DM patients were receiving oral antidiabetic medication or insulin therapy. Inclusion criteria were defined as follows: absence of any corneal pathology or any previous eye surgery, spherical refraction within ± 5.0 D and cylinder correction within ± 3.0 D, best corrected visual acuity of 20/40 or better, absence of severe media opacities such as corneal opacity, mature cataract, vitreous opacity, absence of optic nerve or retina pathologies other than glaucomatous optic neuropathy, and absence of any inflammatory eye disease or systemic collagen disease. Patients with POAG were graded according to the Hodapp-Anderson-Parrish grading scale with their visual field assessments. Only the patients with early stage of POAG were recruited into the study.8
STATISTICAL METHODS Analyses were performed using SPSS for Windows (version 15.0, SPSS Inc.). Normality of the groups was evaluated using Kolmogorov-Smirnov test. The student t test was used for parametric data and the Mann-Whitney U test was used for nonparametric data to compare the groups. Pearson correlation coefficient was used when both variables were measured and Spearman correlation coefficient was used when variables were nonparametric. Analysis of covariance
(ANCOVA) test was used to account the effects of age, sex, MD, and PSD values over significant differences between the groups. Age, sex, MD, and PSD values were set as covariates in Analysis of covariance (ANCOVA) test. Logarithmic transformation was used to convert nonparametric data to parametric data where needed. Results were reported as mean ± SD with 95% confidence intervals and P = 0.05 was considered as statistically significant.
RESULTS Sixty eyes of 60 POAG patients with diabetes (study group) and 41 eyes of 41 POAG patients without diabetes (control group) were included. Mean age of the study and the control groups were 59.1 ± 7.4 and 58.6 ± 7.7 years, respectively (P = 0.696). Female/male ratios were 38/22 and 25/16 for study and control groups, respectively (P = 0.837). In the study group, mean HbA1c level during HRT measurements was 8.00 ± 1.2% and mean disease (DM) duration was 9.2 ± 5.8 years. The differences in spherical equivalent, keratometry, central corneal thickness, visual field MD, and PSD differences between the groups were statistically insignificant (Table 1). HRT-based measurements (disc area, cup area, rim area, cup volume, rim volume, cup/disc area ratio, linear cup/disc ratio, mean cup depth, cup shape measure, mean RNFL thickness, and RNFL cross-section area) between the groups were investigated. Rim area and rim volume were significantly higher in POAG patients with diabetes. The remaining HRT differences were insignificant (Table 2). Statistical power of the
TABLE 2. The Differences in HRT Parameters Between the Groups
Study Group Parameters 2
Disc area (mm ) Cup area (mm2) Rim area (mm2) Cup volume (mm3) Rim volume (mm3) CD area ratio Linear CD ratio Mean cup depth (mm) Maximum cup depth (mm) Cup shape measure Mean RNFL thickness (mm) RNFL cross-sectional area (mm2) Height variation contour (mm)
Mean ± SD
Median
2.39 ± 0.51 0.75 ± 0.44 1.63 ± 0.42 0.21 ± 0.17 0.46 ± 0.20 0.30 ± 0.15 0.53 ± 0.15 0.27 ± 0.10 0.70 ± 0.21 0.15 ± 0.08 0.25 ± 0.09 1.38 ± 0.57 0.46 ± 0.18
2.39 0.73 1.63 0.18 0.43 0.32 0.56 0.29 0.71 0.13 0.26 1.42 0.43
Control Group Range
1.37 0.00 0.67 0.00 0.07 0.00 0.03 0.03 0.12 0.39 0.01 0.04 0.20
to to to to to to to to to to to to to
3.70 1.97 0.60 0.79 1.04 0.64 0.80 0.57 0.27 0.02 0.56 0.20 1.20
Mean ± SD
Median
2.31 ± 0.42 0.83 ± 0.40 1.47 ± 0.28 0.25 ± 0.18 0.36 ± 0.10 0.34 ± 0.13 0.57 ± 0.11 0.29 ± 0.11 0.75 ± 0.22 0.15 ± 0.07 0.23 ± 0.07 1.24 ± 0.35 0.41 ± 0.12
2.26 0.81 1.42 0.22 0.35 0.35 0.59 0.29 0.83 0.16 0.23 1.21 0.39
Range 1.62 0.17 0.84 0.02 0.20 0.09 0.30 0.11 0.37 0.30 0.12 0.57 0.20
to to to to to to to to to to to to to
3.80 2.00 2.06 0.91 0.67 0.64 0.80 0.56 1.29 0.03 0.45 2.41 0.80
P 0.36 0.38 0.04 0.18 0.02 0.16 0.19 0.33 0.26 0.44 0.23 0.12 0.18
CD indicates cup disc; HRT, Heidelberg retina tomography; RNFL, retinal nerve fiber layer.
50 | www.glaucomajournal.com Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma
Volume 25, Number 1, January 2016
differences in main outcome measures (disc area, cup area, rim area, cup volume, rim volume, cup-disc area ratio, mean cup depth, maximum cup depth, cup shape measure, mean RNFL, RNFL cross-sectional area, and height variation contour) ranged from 21.7% (disc area) to 100% (RNFL cross-sectional area). The statistical powers for 2 significantly different results were 0.74 and 0.95 for rim area and rim volume, respectively. The effects of possible confounding factors (age, sex, MD, and PSD) on the significant differences between the groups (rim area and rim volume) were investigated using ANCOVA test after logarithmic transformation of nonparametric data to parametric data. None of the possible confounding factors had significant effect on initial statistical comparisons. The statistical significances of age, sex, MD, and PSD after ANCOVA test were 0.957, 0.109, 0.932, and 0.595 for rim area and 0.359, 0. 153, 0.478, and 0.828 for rim volume, respectively. The correlation analyses in diabetic group showed that duration of diabetes and HbA1c levels were not significantly correlated to rim area and rim volume (Fig. 1).
Optic Nerve Head in Diabetic Glaucoma Patients
HbA1c levels were significantly correlated to disc area, cup area, cup volume, and cup shape measure (Fig. 2). Diabetic patients were divided into 2 groups according to disease duration. There were 27 patients with disease duration of r8 years and 33 patients with disease duration of Z10 years. Rim area and rim volume differences between those groups were also insignificant (P = 0.191, P = 0.203, respectively). Similarly, the DM patients were divided into 2 groups according to HbA1c levels. Mean ± SD for HbA1c levels were 6.97 ± 0.59 (n = 30) and 9.01 ± 0.82 (n = 30) for the groups. Rim area and rim volume did not differ significantly between the groups (P = 0.912, P = 0.894, respectively).
DISCUSSION The mechanism of pathologic changes in retinal neurons due to diabetes is extremely complex. Diabetes has several effects on different retinal cells such as neurons, vascular cells, glial cells, and microglial cells.9 It has been shown that some abnormalities occur in glial elements
FIGURE 1. The correlations between HbA1c and duration of diabetes to rim area and rim volume.
Copyright
r
2014 Wolters Kluwer Health, Inc. All rights reserved.
www.glaucomajournal.com |
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
51
Akkaya et al
J Glaucoma
Volume 25, Number 1, January 2016
FIGURE 2. The correlations between HbA1c and disc area, cup area, cup volume, and cup shape measure.
before the clinical manifestations of diabetic retinal vasculopathy.10–12 Barber et al10 reported the increase of apoptotic (TUNEL-positive) cells in human retinal neurons. They also proposed that significant decrease in the number of RGCs and diminution of the thickness of the inner plexiform and inner nuclear layers increase the likelihood that apoptotic cells in retina comprises ganglion cells and other neurons. In an experimental model of diabetes, Hammes et al2 showed that diabetes induces apoptosis in RGCs and Mu¨ller cells.2 In another study by Abu-El-Asrar et al,3 it has been reported that the number of promoting factors inducing apoptosis in RGCs is increased and the death of RGCs occur early in diabetic eyes. It has also been reported in experimental animal studies that diabetes induces impairment of retrograde axonal transport in largesized and medium-sized RGCs and this impairment precedes optic nerve involvement.4,5 There are limited published data which investigates the ONH parameters in diabetic POAG patients. Budde and Jonas compared the stereo disc photographs of 52 diabetic POAG patients with 312 nondiabetic POAG patients. There were no statistically significant differences regarding mean optic disc size, neural rim, alpha zone, and beta zone of parapapillary atrophy. Patients with proliferative diabetic retinopathy were not recruited in their study. The authors concluded that diabetes does not have a marked effect on ONH damage in POAG patients.13 Yu¨cel et al,14
reported that rim area, RNFL cross-sectional area, rim volume, and RNFL thickness correlated significantly with optic nerve fiber number in glaucomatous eyes. In our study, greater measurements of rim area and rim volume in diabetic POAG patients increases the likelihood that the number of optic nerve fibers in diabetic POAG patients is greater than that in nondiabetic POAG patients. It may be argued that optic nerve swelling may have caused increased rim area and rim volume measurements. However, in our study, rim area and rim volume measurements were not significantly correlated to HbA1c levels. In addition, rim area and rim volume differences between patients with disease duration of r8 and Z10 years, and between patients with lower and higher HbA1c levels were insignificant. This findings lessen (but does not omit) the possibility of the significant effect of subclinical optic nerve swelling on rim area and rim volume. Because, patients with higher HbA1c levels and longer disease duration are more likely to have optic nerve swelling. In this study, we did not find significant differences in cup area and cup volume between diabetic and nondiabetic POAG patients. In diabetic POAG patients, greater measurements of rim area and rim volume together with the absence of any increase in the optic cup depth lead us to consider that diabetes may cause protective effect on optic nerve and retinal nerve fibers in POAG. The persistence of the significant differences in rim area and rim volume
52 | www.glaucomajournal.com Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Glaucoma
Volume 25, Number 1, January 2016
between the diabetic and nondiabetic patients after controlling the effects of age, sex, MD, and PSD values with ANCOVA test further stresses the validity of our findings. In our study, there was not any statistically significant difference between the duration of diabetes and topographic parameters of ONH. This result also lead us to consider that the duration of diabetes have no significant effect on ONH topography. There were weak but statistically significant correlations between HbA1c levels and HRT parameters including cup area, cup volume, and cup shape measure. This correlation implies that poor metabolic control of diabetes may have negative effect on optic cup depth. These findings show the necessity of longitudinal and prospective studies about the effect of the metabolic control of diabetes on ONH topography in POAG patients with diabetes. The limitations of the HRT technology include the need to outline the anterior scleral canal opening and a reference plane. Stereometric parameters are calculated upon this delineation. Coefficient of variation of 3 HRT measurements performed by the same technician were reported to be between 2.9% and 5.2%.15 However, interindividual variations are expected to be higher than those percentiles. In the most comprehensive model, it may be argued that the delineation of the anterior scleral canal in diabetic patients might have been assessed deeper than the nondiabetic patients (perhaps because they have some slightly higher degree of temporal RNFL loss) and this might have been resulted in higher rim area and rim volume. However, insignificant rim area and rim volume differences between patients with lower and higher HbA1c levels and between patients with shorter and longer disease duration makes this presumption almost improbable. We think that if diabetes would have caused temporal RNFL loss more than nondiabetic patients in our study, patients with higher HbA1c levels and longer disease duration would possibly have significantly lower rim area and rim volume. One of the limitations in this study is the low levels of statistical powers in some comparisons. Although statistical powers for rim area and rim volume comparisons were 0.74 and 0.95, most of the statistical powers for the remaining insignificant differences were
