- Email: [email protected]
Outer Segment Junction in Patients with Diabetic Macular Edema
Association of Retinal Sensitivity to Integrity of Photoreceptor Inner/Outer Segment Junction in Patients with Diabetic Macular Edema Jithin Yohannan, BA,1,2 Millena Bittencourt, MD,1 Yasir Jamal Sepah, MBBS,1 Elham Hatef, MD, MPH,1 Raafay Sophie, MBBS,1 Ahmadreza Moradi, MD,1 Hongting Liu, MD,1 Mohamed Ibrahim, MD,1 Diana V. Do, MD,1,3 Elizabeth Coulantuoni, PhD,2 Quan Dong Nguyen, MD, MSc1,3 Purpose: To evaluate the relationship between retinal sensitivity and the photoreceptor inner segment/outer segment (IS/OS) layer status in patients with diabetic macular edema (DME). Design: Cross-sectional study. Participants: Twenty-five adult patients (37 eyes) diagnosed with DME and managed at the Wilmer Eye Institute, Johns Hopkins University (Baltimore, MD). Methods: We obtained simultaneous fundus microperimetry (MP) and optical coherence tomography (OCT) of patients with DME using a combined MP/OCT system. The device recorded retinal sensitivity and retinal thickness on a 3-dimensional tomography map, and we performed a point-by-point analysis of the IS/OS layer integrity at every MP point. We also reviewed OCT scans to determine the type of DME, cystoid macular edema, or diffuse macular edema (absence of any cysts). In addition, fixation stability and fixation location were analyzed. Main Outcome Measures: Retinal point sensitivity measured by MP. Results: Twenty-five patients (37 eyes: 29 male and 8 female; mean age, 64.16 years) with DME were enrolled. Fixation was centric in 30 eyes (81%), paracentric in 3 eyes (8%), and eccentric in 4 eyes (11%). Twenty-seven eyes had cystoid macular edema, and 10 eyes had diffuse macular edema. Mean central subfield thickness was 325 m. We analyzed a total of 1036 individual MP points. Mean point sensitivity was 10.51 dB. A total of 793 points (76.5%) had IS/OS layer present, and 243 points (23.5%) had IS/OS layer disrupted. A mixed linear model, constructed to adjust for potential confounders and account for dependence between retinal points, revealed that the absence of the IS/OS junction was significantly associated with a 3.28-dB decrease in retinal point sensitivity (P⬍0.001). Conclusions: This novel index study demonstrates that disruption of the IS/OS junction is correlated with a significant decrease in point sensitivity in eyes with DME. Further studies are indicated to confirm and validate this relationship. Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references. Ophthalmology 2013;120:1254 –1261 © 2013 by the American Academy of Ophthalmology.
Diabetic macular edema (DME) is a complication of diabetes mellitus that currently affects an estimated 8% to 10% of 23.6 million diabetic persons in the United States.1 Because the prevalence of diabetes is expected to grow by 50% worldwide to more than 300 million diabetic persons by 2025,2 the prevalence of DME is also expected to grow substantially. The development of DME is initiated by hyperpermeability of the retinal vasculature, which allows extracellular fluid accumulation within the retina. Ischemic dysfunction of the neuroglial cells due to macular edema results in visual disturbance.3 Therapeutic strategies, such as anti-vascular endothelial growth factor drugs,4 steroids,5 and laser photocoagulation,6 have been used to control DME. However, complete resolution of the disease is not always possible, which has led to a need for better understanding of the pathophysiology of DME.
1254
© 2013 by the American Academy of Ophthalmology Published by Elsevier Inc.
Recent advances in OCT technology, such as development of high-resolution spectral-domain OCT, have facilitated the visualization of outer retinal changes that may be important indicators of functional integrity of the visual pathway. The disruption of the junction of the inner segment/outer segment (IS/OS) layers of photoreceptors and the external limiting membrane has been shown to have an important role in retinal dysfunction.7–12 Initially, studies of retinal vein occlusion demonstrated that disruption of the IS/OS junction is highly correlated with decreased visual acuity (VA)7 and worse visual outcome.8 Subsequent studies demonstrated similar associations between IS/OS status and VA in age-related macular degeneration9 and DME.11,12 Such studies of DME have relied on VA, which is a clinically useful, global measure of macular function. However, VA may poorly reflect focal changes in retinal function and structures in more localized diseases. ISSN 0161-6420/13/$–see front matter http://dx.doi.org/10.1016/j.ophtha.2012.12.003
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME To our knowledge, no study has been conducted to evaluate the correlation of local retinal sensitivity to the corresponding local IS/OS changes in DME. The development of microperimetry (MP) has made it possible to evaluate retinal point sensitivity in areas of macular edema.13 Moreover, combined automatic fundus perimetry/tomography systems, such as the spectral-domain optical coherence tomography/scanning laser ophthalmoscope (Optos Spectral OCT/SLO system; Optos, Dunfermiline, Scotland, UK), allow one to investigate the findings on spectral-domain OCT associated with changes in point retinal sensitivity. Hatef et al14 demonstrated a strong relationship between macular sensitivity and point thickness in DME, whereas Sepah et al15 illustrated similar findings in eyes with uveitic macular edema. In another recent study,10 the integrity of the photoreceptor layer was found to be an important determinant of the effect of macular thickness on VA in patients with DME. Visual acuity was correlated with parafoveal thickness in patients who had an intact photoreceptor layer, whereas VA was not associated with parafoveal thickness in patients who had a disrupted photoreceptor layer. It is not known whether the presence or absence of the IS/OS layer affects the relationship between point sensitivity and point thickness.14 In this index study, we aim to assess the association between IS/OS integrity and retinal point sensitivity in patients with DME using the Spectral OCT/SLO system and to quantify the effect of presence versus absence of the IS/OS layer on the relationship between macular thickness and point sensitivity.
Materials and Methods We performed a cross-sectional study of adult patients diagnosed with DME and managed at the Wilmer Eye Institute, Johns Hopkins University (Baltimore, MD). Our research followed the tenets of the Declaration of Helsinki and was approved by the Johns Hopkins Institutional Review Board/Ethics Committee. Informed consent was obtained from all participating subjects. Inclusion criteria included diagnosis of DME confirmed by a retina specialist using contact lens biomicroscopy and OCT findings indicating clinically significant macular edema, increased central point thickness, and the presence of fluid. Leakage on fluorescein angiography was also assessed to support the diagnosis of DME. The OCT scans were included in the study only if the signal strength was greater than 6/10. Exclusion criteria included focal or grid laser therapy that produced scarring in the area of the MP grid because we believe extensive laser treatment may artificially disrupt the IS/OS junction. Study data collected at the time of imaging visit included best-corrected VA measured using the Early Treatment Diabetic Retinopathy Study chart and any prior treatments for DME. We also collected demographic data, including age, gender, and race, from the patients’ medical records. We obtained superimposed fundus OCT and MP scans of each of patient using the Spectral OCT/SLO system. Optical coherence tomography was captured using the unit’s integrated OCT module. This module performs 28 000 A-scans per second over 2 seconds to generate a 3-dimensional retinal map of 128 longitudinal OCT cuts over a 5-mm macular area.16 Microperimetry was obtained using the POLAR III circular test pattern, which consists of 28 dots with 4 central dots within 4 degrees of the center of the fovea, 12 mid-periphery dots within 8
degrees, and 12 outer dots within 12 degrees. Additional parameters that were programmed into the device included Goldman III stimulus size, 200-ms stimulus duration, and 1000-ms interval between stimuli presentation. During OCT and MP scans, the Spectral OCT/SLO device selected retinal landmarks to automatically track the eye movements. The software provided with the device performed semiautomatic alignment of MP and OCT images and generated a 3dimensional superimposed image.16 The bundled software (version 1.87) allows the user to adjust horizontal and vertical scroll lines on a 2-dimensional retinal tomography map with superimposed MP values. Such establishment allows the user to view horizontal and vertical OCT cross-sections corresponding to a specific MP point. At each of the 28 MP points, the device measured retinal sensitivity on a scale of 0 to 20 dB, with 0 representing the brightest stimulus and 20 being the dimmest. A trained grader assessed the presence or absence of the IS/OS junction at the OCT cross-sections corresponding to each MP point. The IS/OS junction was defined as the hyper-reflective line between the inner and the outer segment of the photoreceptors, just above the hyperreflective line representing the retinal pigment epithelium and just below the hyper-reflective line representing the external limiting membrane. The IS/OS junction status, presence or absence, was recorded as a binary data point (1 or 0, respectively). If the initial grader was unable to accurately determine whether the IS/OS junction was disrupted at a particular point, a second expert grader was consulted and the presence or absence of the IS/OS junction was agreed on by consensus. If there was overlying vessel, fluid, blood, or exudate, then a shadow may fall on the IS/OS junction and mimic disruption. However, in such cases, nearby retinal layers also would demonstrate shadowing. Therefore, we graded points in which there was shadowing in the IS/OS junction and overlying or underlying retinal layers as having an intact IS/OS junction. The Spectral OCT/SLO software automatically calculated retinal thickness, defined as the distance between inner limiting membrane and retinal pigment epithelium, at each point in micrometers.16 Central subfield thickness for each eye was calculated using the mean of the 4 points in the central ring of the MP grid. Therefore, 3 data values were obtained for every MP point: retinal sensitivity (decibels), presence or absence of the IS/OS layer (1 or 0), and retinal thickness (micrometers). Figure 1 demonstrates images from a single retina being analyzed for point sensitivity and presence or absence of the IS/OS layer at 2 different test points. In addition, an expert grader reviewed all longitudinal cuts of a patient’s OCT and determined the type of DME present. We defined diffuse edema as an increase in retinal thickness with reduced intraretinal reflectivity in the absence of cysts; meanwhile, cystoid macular edema was defined as the presence of intraretinal cystoid spaces in the macular area.17
Table 1. Baseline Demographics of Study Subjects No. of patients Mean age at time of visit Male, n (%) Race Caucasian, n (%) African American, n (%) Asian, n (%) No. of eyes Right, n (%) Left, n (%)
25 64.16 yrs 19 (76) 19 (76) 5 (20) 1 (4) 37 23 (62) 14 (38)
1255
Ophthalmology Volume 120, Number 6, June 2013 Table 2. Characteristics of 37 Eyes and 1036 Retinal Points Examined* Mean VA in logMAR (SE) Type of macular edema Cystoid, n (%) Diffuse, n (%) Fixation stability Stable, n (%) Relatively unstable, n (%) Unstable, n (%) Fixation centricity Centric, n (%) Paracentric, n (%) Eccentric, n (%) Mean retinal point thickness in m (SE)* Mean central subfield thickness in m (SE)† Mean percentage of IS/OS layer disrupted (SE)* Mean retinal point sensitivity in dB (SE)*
0.40 (0.05) 27 (73) 10 (23) 25 (68) 12 (32) 0 (0) 30 (81) 3 (8) 4 (11) 325.2 (2.79) 324.1 (9.36) 23.46 (0.49) 10.51 (0.15)
IS/OS ⫽ inner segment/outer segment; logMAR ⫽ logarithm of the median angle of resolution; SE ⫽ standard error; VA ⫽ visual acuity. *These data were calculated using all microperimetry points. Total sample size is 37 ⫻ 28 ⫽ 1036 points. † Central subfield thickness was measured using the thickness values from the MP grid. Total sample size is 37 ⫻ 4 ⫽ 148 points.
We determined fixation location and fixation stability for each eye using fixation pattern information recorded automatically by the device. We classified fixation location into 1 of 3 categories:14 ● ● ●
Centric: ⬎50% of fixation points within 0.5 mm of the foveal center Pericentric: 25% to 50% of fixation points within 0.5 mm of foveal center Eccentric: ⬍25% of fixation points within 0.5 mm of foveal center.
Likewise, we classified fixation stability into 1 of 3 categories:14 ● ●
●
Stable: ⬎75% of fixation points located within a 2-degree circle, regardless of their position from the foveal center Relatively unstable: ⬍75% of fixation points were located within a 2-degree circle but ⬎75% were located in a 4-degree circle, regardless of their position from the foveal center Unstable: ⬍75% of points are located within a 4-degree circle, regardless of their position from the foveal center.
Statistical Analysis We constructed summary tables describing patient (age, sex, and race) and eye (fixation location, fixation stability, and type of DME) characteristics. We performed bivariable analysis using t tests that included robust variance estimates for clustering within eyes to compare (1) the main study variables in points with IS/OS present versus IS/OS absent and (2) the main study variables in eyes with cystoid macular edema versus eyes with diffuse edema.
We explored the association between point sensitivity and retinal thickness for points with IS/OS present and IS/OS absent using scatterplots with Lowess smooth average curves. Linear mixed models were used to (1) analyze association between point sensitivity and IS/OS status and (2) quantify the association between point sensitivity and retinal thickness stratified by the status of IS/OS junction (present and absent). On the basis of the visualization of the relationship between point sensitivity and retinal thickness, we added interaction terms to the models to allow for different linear slopes for thickness values less than and more than 300 m. The models adjusted for confounding factors, including the patient and eye characteristics described earlier. We assumed there would be correlation of retinal points at 3 different levels: eyes, patients, and ring of MP grid; therefore, we included a random intercept for each of these 3 levels. We performed all statistical analysis using R version 2.14.0 (The R Foundation for Statistical Computing; Vienna, Austria).
Results We analyzed the records and images of 25 patients (37 eyes). Our patients were predominantly elderly (mean age, 64.16 years) white men (Table 1). Table 2 summarizes the characteristics of the 37 eyes. Seventy-three percent (27) of the eyes had cystoid macular edema, with 68% fixation stable and 81% fixation centric. Among the 37 eyes, a total of 1036 retinal points with MP and OCT data were examined. The mean central subfield thickness of the points examined was 324.1 (standard error [SE], 2.79 m). On average, 23.46% (SE, 0.49) of IS/OS layers were disrupted. The mean retinal point sensitivity was 10.51 dB (SE, 0.15). Bivariate analyses comparing mean retinal thickness and mean point sensitivity in points with and without the IS/OS layer are summarized in Table 3. Compared with points with a normal IS/OS layer, points with IS/OS disruption had on average 4.23 dB lower sensitivity (P⬍0.001) and 62.48 m increased thickness (P⬍0.001). The mean central subfield thickness, mean percentage of IS/OS points disrupted, and mean retinal point sensitivity in eyes with cystoid macular edema versus eyes with diffuse macular edema are described in Table 3. Compared with diffuse macular edema, cystoid macular edema was associated with a 106-m increase in mean central subfield thickness (P⬍0.001), a 13% higher proportion of points with IS/OS layers disrupted (P ⫽ 0.11), and a 2.18-dB decrease in mean point sensitivity (P⬍0.001). Scatter plots of the relationship between retinal sensitivity and thickness in the presence versus the absence of the IS/OS are presented in Figure 2. We found that in points with an IS/OS layer present, there was a peak in retinal sensitivity at a retinal thickness of approximately 300 m. We observed little to no association between sensitivity and thickness when the IS/OS layer was absent. The multilevel linear models with random intercepts (Table 4) demonstrated that absence of the IS/OS layer results in a significant 3.28-dB (P⬍0.001) decrease in retinal sensitivity after controlling for retinal thickness, type of DME, age, sex, race, fixation location, and fixation stability. For points with an IS/OS layer present and a retinal thickness ⬍300 m, we found
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 Figure 1. A, Combined microperimetry (MP)/optical coherence tomography (OCT) scan of a patient with diabetic macular edema (DME). The green vertical and blue horizontal lines indicate the planes of the horizontal and vertical retinal cross-sections shown to the right and below the image, respectively. The cross-sections are centered on a point with an MP sensitivity of 14 decibels (dB). The inner segment/outer segment (IS/OS) layer is a thin line above the hyper-reflective retinal pigment epithelium, as indicated by the red arrows. It is intact at this specific retinal point. (1) A magnified view of the IS/OS layer in the horizontal cross-section. B, Scan of the same retina shown in A. The horizontal and vertical cross-sections are centered on a point with an MP sensitivity of 2 dB. The IS/OS layer indicated by the red arrows is disrupted at this point. (2) A magnified view of the IS/OS layer in the horizontal cross-section.
1256
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME
Figure 1.
1257
Ophthalmology Volume 120, Number 6, June 2013 Table 3. Retinal Characteristics on Optical Coherence Tomography Point Thickness and Sensitivity by IS/OS Layer Status Mean retinal thickness in m (SE) Mean retinal point sensitivity in dB (SE)
All
IS/OS Present
IS/OS Absent
P Value
325.2 (2.8) 10.51 (0.15)
310.6 (2.6) 11.50 (0.15)
373.1 (7.5) 7.27 (0.30)
⬍0.001 ⬍0.001
Retinal Characteristics by Type of Diabetic Macular Edema Present Cystoid Macular Edema
Diffuse Edema
352.7 (11.3) 27.0 (0.7) 9.9 (0.2)
247.1 (8.1) 13.9 (1.1) 12.1 (0.2)
Mean central subfield thickness in m (SE) Percentage of IS/OS disrupted (SE) Mean retinal point sensitivity in dB (SE)
⬍0.001 0.11 0.006
IS/OS ⫽ inner segment/outer segment; SE ⫽ standard error. The upper portion demonstrates retinal thickness and sensitivity for all points, points with IS/OS layer present, and points with IS/OS layer absent. The lower portion shows retinal thickness and sensitivity for microperimetry points with cystoid macular edema and diffuse edema.
that a 1-m increase in thickness resulted in a 0.04-dB (P⬍0.001) increase in retinal sensitivity. On the contrary, for points with an IS/OS layer present and a retinal thickness ⬎300 m, we found a 0.01-dB (P ⫽ 0.01) decrease in retinal sensitivity for every 1-m increase in thickness. When considering points with the IS/OS layer disrupted, we found that for points ⬍300 m, the thickness–sensitivity relationship was weakened to a 0.02-dB increase in sensitivity for every 1-m increase in thickness. For points ⬎300 m, the relationship remained similar to points with the IS/OS layer present. Finally, the presence of cystoid macular edema was associated with a significant 2.71-dB (P⬍0.001) decrease in retinal sensitivity.
Discussion In this index study, we have used a novel approach to analyze the effect of IS/OS junction status on retinal sensitivity and to investigate the thickness–sensitivity relationship in patients with DME. Disruption of the IS/OS junction is associated with a significant 3.28-dB decrease in retinal sensitivity after adjusting for confounding variables. In addition, we found that IS/OS disruption results in a weaker thickness–sensitivity association in retinal points that were thinner than 300 m. However, we also found that IS/OS disruption has no effect on the relationship between thickness and sensitivity in points thicker than 300 m. Our study suggests that the status of the IS/OS junction is a predictor of retinal sensitivity in patients with DME. Prior studies have shown that disruption of the IS/OS junction11,12 and external limiting membrane (a close correlate of the IS/OS junction)9 at the fovea results in a decreased VA measured by standard testing. To our knowledge, our study is the first to use MP testing to demonstrate that a similar relationship exists for points throughout the retina. Furthermore, we were able to obtain a large dataset because MP allows for 28 points of retinal data per eye, which allowed us to create a multilevel linear model to account for dependency between MP points and control for various factors that might confound the relationship between IS/OS status and MP sensitivity. The relationship between sensitivity and IS/OS status was maintained after correcting for
1258
confounding. The finding suggests that local photoreceptor integrity, as indicated by IS/OS status, is a strong predictor of local visual function throughout the retina. Our analysis of the thickness–sensitivity relationship suggests that there is a peak retinal sensitivity at approximately 300 m, which is similar to prior findings.14 When the IS/OS layer is disrupted and thickness is ⬍300 m, the thickness– sensitivity relationship was weaker than when the IS/OS layer is present. Such a relationship suggests that once the IS/OS layer is disrupted in retinas thinned by atrophic degeneration, thickness is no longer such a strong predictor of visual function. However, when the IS/OS layer is disrupted in points ⬎300 m, thickness remains an important predictor of retinal sensitivity. Therefore, in edematous retinas, increased thickness tends to have an effect independent of the effect of photoreceptor disruption. This finding may lend credence to the standard approach of today’s retina specialists to maximize resolution of edema in patients with DME. Recent studies have suggested that IS/OS status is a good prognostic indicator for DME therapy. One study18 found that patients treated with triamcinolone tended to do better when the IS/OS layer was intact. Another recent study found that IS/OS status was a good predictor of visual outcome in patients treated with pars plana vitrectomy.19 Although we agree that patients with IS/OS disruption may have reduced VA, our results suggest that retinal thickness remains an important predictor of visual function in patients with edema independent of IS/OS status. Therefore, in edematous patients, the clinician should strive to reduce macular edema regardless of a patient’s IS/OS status. We found that the presence of cysts in the edema (cystoid macular edema) results in decreased sensitivity, similar to findings in other studies.10,20 We found that cysts were significantly associated with retinal thickness and increased IS/OS disruption, albeit statistically insignificant. Our multilevel model revealed that the presence of cysts was associated with decreased sensitivity in DME independently of their effect of increasing retinal thickness and IS/OS disruption. Cysts, which usually occur in the inner nuclear layer or outer plexiform layer,21,22 have been hypothesized to have their effect by disrupting bipolar or ganglion cells.10
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME
Figure 2. Scatter plots of retinal thickness versus retinal sensitivity stratified by inner segment/outer segment (IS/OS) layer status. Lowess curve was fit in both graphs (thick black line).
Our findings, that cysts have an effect independent of their effect on retinal thickness and photoreceptor disruption, support such hypothesis.
Study Limitations Our study has provided valuable insights into the determinants of visual disruption in DME. However, the study also
has several limitations. We had a relatively small sample size of patients. In addition, there is always the possibility of residual confounding in our analysis. We did not have the data to control or model for duration or severity of diabetes, prior treatments for DME, and any concurrent ophthalmologic condition. Finally, because our study was a crosssectional study, it is difficult to establish temporality of the exposure (IS/OS status) and outcome (retinal sensitivity).
1259
Ophthalmology Volume 120, Number 6, June 2013 Table 4. Factors Affecting Retinal Point Sensitivity in Patients with Diabetic Macular Edema Factor Disrupted IS/OS layer* Thickness–sensitivity relationship† Points with IS/OS layer intact Increase in 1 m of thickness ⬍300 m Increase in 1 m of thickness ⬎300 m Points with IS/OS layer disrupted Increase in 1 m of thickness ⬍300 m Increase in 1 m of thickness ⬎300 m Cystoid macular edema†
Mean Change in Sensitivity (dB) ⫺3.28 0.04 ⫺0.01
0.02
95% CI
P Value
⫺3.89 to ⫺2.68
⬍0.001
0.02 to 0.06
⬍0.001
⫺0.02 to ⫺0.001
0.006 to 0.03
0.001
0.006
⫺0.01
⫺0.02 to ⫺0.003
⬍0.001
⫺2.71
⫺4.22 to ⫺1.21
⬍0.001
CI ⫽ confidence interval; IS/OS ⫽ inner segment/outer segment. Results from multilevel linear models with random intercepts built to assess the influence of various factors on retinal point sensitivity in patients with DME. In addition to the factors indicated, adjustments were also made for age, sex, race, and fixation location and stability. *Results are from a multilevel model without an interaction term for thickness and IS/OS status. † Results are from a multilevel model with an interaction term for thickness and IS/OS status.
In conclusion, we have demonstrated through our study that the absence of the IS/OS layer has a significant impact on retinal sensitivity in patients with DME. We have also shown that retinal thickness is an important predictor of retinal function in edematous patients even when the IS/OS layer is disrupted. Therefore, although active monitoring of the IS/OS layer by retina physicians may provide further insight into the potential for visual function of a patient, IS/OS disruption should not be a contraindication for treatments to reduce retinal edema. As our understanding of the pathophysiologic effect of DME on the retina broadens, on the basis of what we learn from the roles of retinal structures such as the IS/OS, we may be able to use our current therapies more appropriately, as well as design more targeted and effective future treatments for DME.
References 1. Chen E, Looman M, Laouri M, et al. Burden of illness of diabetic macular edema: literature review. Curr Med Res Opin 2010;26:1587–97. 2. Sjostrand J, Popovic Z, Conradi N, Marshall J. Morphometric study of the displacement of retinal ganglion cells subserving cones within the human fovea. Graefes Arch Clin Exp Ophthalmol 1999;237:1014 –23. 3. Antonetti DA, Barber AJ, Bronson SK, et al, JDRF Diabetic Retinopathy Center Group. Diabetic retinopathy: seeing beyond glucose-induced microvascular disease. Diabetes 2006; 55:2401–11. 4. Diabetic Retinopathy Clinical Research Network, Elman MJ, Aiello LP, Beck RW, et al. Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema. Ophthalmology 2010;117:1064 –77. 5. Martidis A, Duker JS, Greenberg PB, et al. Intravitreal triamcinolone for refractory diabetic macular edema. Ophthalmology 2002;109:920 –7.
1260
6. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema: Early Treatment Diabetic Retinopathy Study report no. 4. Int Ophthalmol Clin 1987;27:265–72. 7. Ota M, Tsujikawa A, Murakami T, et al. Association between integrity of foveal photoreceptor layer and visual acuity in branch retinal vein occlusion. Br J Ophthalmol 2007;91: 1644 –9. 8. Shin HJ, Chung H, Kim HC. Association between integrity of foveal photoreceptor layer and visual outcome in retinal vein occlusion [report online]. Acta Ophthalmol 2011;89:e35– 40. Available at: http://onlinelibrary.wiley.com/doi/10.1111/j. 1755-3768.2010.02063.x/full. Accessed October 7, 2012. 9. Oishi A, Hata M, Shimozono M, et al. The significance of external limiting membrane status for visual acuity in agerelated macular degeneration. Am J Ophthalmol 2010;150: 27–32. 10. Murakami T, Nishijima K, Sakamoto A, et al. Association of pathomorphology, photoreceptor status, and retinal thickness with visual acuity in diabetic retinopathy. Am J Ophthalmol 2011;151:310 –7. 11. Maheshwary AS, Oster SF, Yuson RM, et al. The association between percent disruption of the photoreceptor inner segment-outer segment junction and visual acuity in diabetic macular edema. Am J Ophthalmol 2010;150:63–7. 12. Otani T, Yamaguchi Y, Kishi S. Correlation between visual acuity and foveal microstructural changes in diabetic macular edema. Retina 2010;30:774 – 80. 13. Vujosevic S, Midena E, Pilotto E, et al. Diabetic macular edema: correlation between microperimetry and optical coherence tomography findings. Invest Ophthalmol Vis Sci 2006; 47:3044 –51. 14. Hatef E, Colantuoni E, Wang J, et al. The relationship between macular sensitivity and retinal thickness in eyes with diabetic macular edema. Am J Ophthalmol 2011;152:400 –5. 15. Sepah YJ, Hatef E, Colantuoni E, et al. Macular sensitivity and fixation patterns in normal eyes and eyes with uveitis with and without macular edema. J Ophthalmic Inflamm Infect 2012;2:65–73.
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME 16. Landa G, Rosen RB, Garcia PM, Seiple WH. Combined three-dimensional spectral OCT/SLO topography and microperimetry: steps toward achieving functional spectral OCT/SLO. Ophthalmic Res 2010;43:92– 8. 17. Panozzo G, Gusson E, Parolini B, Mercanti A. Role of OCT in the diagnosis and follow up of diabetic macular edema. Semin Ophthalmol 2003;18:74 – 81. 18. Shin HJ, Lee SH, Chung H, Kim HC. Association between photoreceptor integrity and visual outcome in diabetic macular edema. Graefes Arch Clin Exp Ophthalmol 2012;250: 61–70.
19. Sakamoto A, Nishijima K, Kita M, et al. Association between foveal photoreceptor status and visual acuity after resolution of diabetic macular edema by pars plana vitrectomy. Graefes Arch Clin Exp Ophthalmol 2009;247:1325–30. 20. Soliman W, Hasler P, Sander B, Larsen M. Local retinal sensitivity in relation to specific retinopathy lesions in diabetic macular oedema. Acta Ophthalmol 2012;90:248 –53. 21. Tso MO. Pathology of cystoid macular edema. Ophthalmology 1982;89:902–15. 22. Yanoff M, Fine BS, Brucker AJ, Eagle RC Jr. Pathology of human cystoid macular edema. Surv Ophthalmol 1984;28(Suppl):505–11.
Footnotes and Financial Disclosures Originally received: May 14, 2012. Final revision: November 25, 2012. Accepted: December 4, 2012. Available online: March 15, 2013.
Manuscript no. 2012-696.
Financial Disclosure(s): The author(s) have made the following disclosure(s): The Johns Hopkins University has received research support from OPKO, Inc, and Optos, Inc.
Retinal Imaging Research and Reading Center, Wilmer Eye Institute, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Presented at the Association for Research in Vision and Ophthalmology Annual Meeting, May 6 –10, 2012, Fort Lauderdale, Florida.
1
2
Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland.
3
Stanley M. Truhlsen Eye Institute, University of Nebraska Medical Center, Omaha, Nebraska.
Correspondence: Quan Dong Nguyen, MD, MSc, Stanley M. Truhlsen Eye Institute, University of Nebraska Medical Center, 3902 Leavenworth Street, Omaha, Nebraska 68105. E-mail: [email protected].
1261
Diabetic macular edema (DME) is a complication of diabetes mellitus that currently affects an estimated 8% to 10% of 23.6 million diabetic persons in the United States.1 Because the prevalence of diabetes is expected to grow by 50% worldwide to more than 300 million diabetic persons by 2025,2 the prevalence of DME is also expected to grow substantially. The development of DME is initiated by hyperpermeability of the retinal vasculature, which allows extracellular fluid accumulation within the retina. Ischemic dysfunction of the neuroglial cells due to macular edema results in visual disturbance.3 Therapeutic strategies, such as anti-vascular endothelial growth factor drugs,4 steroids,5 and laser photocoagulation,6 have been used to control DME. However, complete resolution of the disease is not always possible, which has led to a need for better understanding of the pathophysiology of DME.
1254
© 2013 by the American Academy of Ophthalmology Published by Elsevier Inc.
Recent advances in OCT technology, such as development of high-resolution spectral-domain OCT, have facilitated the visualization of outer retinal changes that may be important indicators of functional integrity of the visual pathway. The disruption of the junction of the inner segment/outer segment (IS/OS) layers of photoreceptors and the external limiting membrane has been shown to have an important role in retinal dysfunction.7–12 Initially, studies of retinal vein occlusion demonstrated that disruption of the IS/OS junction is highly correlated with decreased visual acuity (VA)7 and worse visual outcome.8 Subsequent studies demonstrated similar associations between IS/OS status and VA in age-related macular degeneration9 and DME.11,12 Such studies of DME have relied on VA, which is a clinically useful, global measure of macular function. However, VA may poorly reflect focal changes in retinal function and structures in more localized diseases. ISSN 0161-6420/13/$–see front matter http://dx.doi.org/10.1016/j.ophtha.2012.12.003
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME To our knowledge, no study has been conducted to evaluate the correlation of local retinal sensitivity to the corresponding local IS/OS changes in DME. The development of microperimetry (MP) has made it possible to evaluate retinal point sensitivity in areas of macular edema.13 Moreover, combined automatic fundus perimetry/tomography systems, such as the spectral-domain optical coherence tomography/scanning laser ophthalmoscope (Optos Spectral OCT/SLO system; Optos, Dunfermiline, Scotland, UK), allow one to investigate the findings on spectral-domain OCT associated with changes in point retinal sensitivity. Hatef et al14 demonstrated a strong relationship between macular sensitivity and point thickness in DME, whereas Sepah et al15 illustrated similar findings in eyes with uveitic macular edema. In another recent study,10 the integrity of the photoreceptor layer was found to be an important determinant of the effect of macular thickness on VA in patients with DME. Visual acuity was correlated with parafoveal thickness in patients who had an intact photoreceptor layer, whereas VA was not associated with parafoveal thickness in patients who had a disrupted photoreceptor layer. It is not known whether the presence or absence of the IS/OS layer affects the relationship between point sensitivity and point thickness.14 In this index study, we aim to assess the association between IS/OS integrity and retinal point sensitivity in patients with DME using the Spectral OCT/SLO system and to quantify the effect of presence versus absence of the IS/OS layer on the relationship between macular thickness and point sensitivity.
Materials and Methods We performed a cross-sectional study of adult patients diagnosed with DME and managed at the Wilmer Eye Institute, Johns Hopkins University (Baltimore, MD). Our research followed the tenets of the Declaration of Helsinki and was approved by the Johns Hopkins Institutional Review Board/Ethics Committee. Informed consent was obtained from all participating subjects. Inclusion criteria included diagnosis of DME confirmed by a retina specialist using contact lens biomicroscopy and OCT findings indicating clinically significant macular edema, increased central point thickness, and the presence of fluid. Leakage on fluorescein angiography was also assessed to support the diagnosis of DME. The OCT scans were included in the study only if the signal strength was greater than 6/10. Exclusion criteria included focal or grid laser therapy that produced scarring in the area of the MP grid because we believe extensive laser treatment may artificially disrupt the IS/OS junction. Study data collected at the time of imaging visit included best-corrected VA measured using the Early Treatment Diabetic Retinopathy Study chart and any prior treatments for DME. We also collected demographic data, including age, gender, and race, from the patients’ medical records. We obtained superimposed fundus OCT and MP scans of each of patient using the Spectral OCT/SLO system. Optical coherence tomography was captured using the unit’s integrated OCT module. This module performs 28 000 A-scans per second over 2 seconds to generate a 3-dimensional retinal map of 128 longitudinal OCT cuts over a 5-mm macular area.16 Microperimetry was obtained using the POLAR III circular test pattern, which consists of 28 dots with 4 central dots within 4 degrees of the center of the fovea, 12 mid-periphery dots within 8
degrees, and 12 outer dots within 12 degrees. Additional parameters that were programmed into the device included Goldman III stimulus size, 200-ms stimulus duration, and 1000-ms interval between stimuli presentation. During OCT and MP scans, the Spectral OCT/SLO device selected retinal landmarks to automatically track the eye movements. The software provided with the device performed semiautomatic alignment of MP and OCT images and generated a 3dimensional superimposed image.16 The bundled software (version 1.87) allows the user to adjust horizontal and vertical scroll lines on a 2-dimensional retinal tomography map with superimposed MP values. Such establishment allows the user to view horizontal and vertical OCT cross-sections corresponding to a specific MP point. At each of the 28 MP points, the device measured retinal sensitivity on a scale of 0 to 20 dB, with 0 representing the brightest stimulus and 20 being the dimmest. A trained grader assessed the presence or absence of the IS/OS junction at the OCT cross-sections corresponding to each MP point. The IS/OS junction was defined as the hyper-reflective line between the inner and the outer segment of the photoreceptors, just above the hyperreflective line representing the retinal pigment epithelium and just below the hyper-reflective line representing the external limiting membrane. The IS/OS junction status, presence or absence, was recorded as a binary data point (1 or 0, respectively). If the initial grader was unable to accurately determine whether the IS/OS junction was disrupted at a particular point, a second expert grader was consulted and the presence or absence of the IS/OS junction was agreed on by consensus. If there was overlying vessel, fluid, blood, or exudate, then a shadow may fall on the IS/OS junction and mimic disruption. However, in such cases, nearby retinal layers also would demonstrate shadowing. Therefore, we graded points in which there was shadowing in the IS/OS junction and overlying or underlying retinal layers as having an intact IS/OS junction. The Spectral OCT/SLO software automatically calculated retinal thickness, defined as the distance between inner limiting membrane and retinal pigment epithelium, at each point in micrometers.16 Central subfield thickness for each eye was calculated using the mean of the 4 points in the central ring of the MP grid. Therefore, 3 data values were obtained for every MP point: retinal sensitivity (decibels), presence or absence of the IS/OS layer (1 or 0), and retinal thickness (micrometers). Figure 1 demonstrates images from a single retina being analyzed for point sensitivity and presence or absence of the IS/OS layer at 2 different test points. In addition, an expert grader reviewed all longitudinal cuts of a patient’s OCT and determined the type of DME present. We defined diffuse edema as an increase in retinal thickness with reduced intraretinal reflectivity in the absence of cysts; meanwhile, cystoid macular edema was defined as the presence of intraretinal cystoid spaces in the macular area.17
Table 1. Baseline Demographics of Study Subjects No. of patients Mean age at time of visit Male, n (%) Race Caucasian, n (%) African American, n (%) Asian, n (%) No. of eyes Right, n (%) Left, n (%)
25 64.16 yrs 19 (76) 19 (76) 5 (20) 1 (4) 37 23 (62) 14 (38)
1255
Ophthalmology Volume 120, Number 6, June 2013 Table 2. Characteristics of 37 Eyes and 1036 Retinal Points Examined* Mean VA in logMAR (SE) Type of macular edema Cystoid, n (%) Diffuse, n (%) Fixation stability Stable, n (%) Relatively unstable, n (%) Unstable, n (%) Fixation centricity Centric, n (%) Paracentric, n (%) Eccentric, n (%) Mean retinal point thickness in m (SE)* Mean central subfield thickness in m (SE)† Mean percentage of IS/OS layer disrupted (SE)* Mean retinal point sensitivity in dB (SE)*
0.40 (0.05) 27 (73) 10 (23) 25 (68) 12 (32) 0 (0) 30 (81) 3 (8) 4 (11) 325.2 (2.79) 324.1 (9.36) 23.46 (0.49) 10.51 (0.15)
IS/OS ⫽ inner segment/outer segment; logMAR ⫽ logarithm of the median angle of resolution; SE ⫽ standard error; VA ⫽ visual acuity. *These data were calculated using all microperimetry points. Total sample size is 37 ⫻ 28 ⫽ 1036 points. † Central subfield thickness was measured using the thickness values from the MP grid. Total sample size is 37 ⫻ 4 ⫽ 148 points.
We determined fixation location and fixation stability for each eye using fixation pattern information recorded automatically by the device. We classified fixation location into 1 of 3 categories:14 ● ● ●
Centric: ⬎50% of fixation points within 0.5 mm of the foveal center Pericentric: 25% to 50% of fixation points within 0.5 mm of foveal center Eccentric: ⬍25% of fixation points within 0.5 mm of foveal center.
Likewise, we classified fixation stability into 1 of 3 categories:14 ● ●
●
Stable: ⬎75% of fixation points located within a 2-degree circle, regardless of their position from the foveal center Relatively unstable: ⬍75% of fixation points were located within a 2-degree circle but ⬎75% were located in a 4-degree circle, regardless of their position from the foveal center Unstable: ⬍75% of points are located within a 4-degree circle, regardless of their position from the foveal center.
Statistical Analysis We constructed summary tables describing patient (age, sex, and race) and eye (fixation location, fixation stability, and type of DME) characteristics. We performed bivariable analysis using t tests that included robust variance estimates for clustering within eyes to compare (1) the main study variables in points with IS/OS present versus IS/OS absent and (2) the main study variables in eyes with cystoid macular edema versus eyes with diffuse edema.
We explored the association between point sensitivity and retinal thickness for points with IS/OS present and IS/OS absent using scatterplots with Lowess smooth average curves. Linear mixed models were used to (1) analyze association between point sensitivity and IS/OS status and (2) quantify the association between point sensitivity and retinal thickness stratified by the status of IS/OS junction (present and absent). On the basis of the visualization of the relationship between point sensitivity and retinal thickness, we added interaction terms to the models to allow for different linear slopes for thickness values less than and more than 300 m. The models adjusted for confounding factors, including the patient and eye characteristics described earlier. We assumed there would be correlation of retinal points at 3 different levels: eyes, patients, and ring of MP grid; therefore, we included a random intercept for each of these 3 levels. We performed all statistical analysis using R version 2.14.0 (The R Foundation for Statistical Computing; Vienna, Austria).
Results We analyzed the records and images of 25 patients (37 eyes). Our patients were predominantly elderly (mean age, 64.16 years) white men (Table 1). Table 2 summarizes the characteristics of the 37 eyes. Seventy-three percent (27) of the eyes had cystoid macular edema, with 68% fixation stable and 81% fixation centric. Among the 37 eyes, a total of 1036 retinal points with MP and OCT data were examined. The mean central subfield thickness of the points examined was 324.1 (standard error [SE], 2.79 m). On average, 23.46% (SE, 0.49) of IS/OS layers were disrupted. The mean retinal point sensitivity was 10.51 dB (SE, 0.15). Bivariate analyses comparing mean retinal thickness and mean point sensitivity in points with and without the IS/OS layer are summarized in Table 3. Compared with points with a normal IS/OS layer, points with IS/OS disruption had on average 4.23 dB lower sensitivity (P⬍0.001) and 62.48 m increased thickness (P⬍0.001). The mean central subfield thickness, mean percentage of IS/OS points disrupted, and mean retinal point sensitivity in eyes with cystoid macular edema versus eyes with diffuse macular edema are described in Table 3. Compared with diffuse macular edema, cystoid macular edema was associated with a 106-m increase in mean central subfield thickness (P⬍0.001), a 13% higher proportion of points with IS/OS layers disrupted (P ⫽ 0.11), and a 2.18-dB decrease in mean point sensitivity (P⬍0.001). Scatter plots of the relationship between retinal sensitivity and thickness in the presence versus the absence of the IS/OS are presented in Figure 2. We found that in points with an IS/OS layer present, there was a peak in retinal sensitivity at a retinal thickness of approximately 300 m. We observed little to no association between sensitivity and thickness when the IS/OS layer was absent. The multilevel linear models with random intercepts (Table 4) demonstrated that absence of the IS/OS layer results in a significant 3.28-dB (P⬍0.001) decrease in retinal sensitivity after controlling for retinal thickness, type of DME, age, sex, race, fixation location, and fixation stability. For points with an IS/OS layer present and a retinal thickness ⬍300 m, we found
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 Figure 1. A, Combined microperimetry (MP)/optical coherence tomography (OCT) scan of a patient with diabetic macular edema (DME). The green vertical and blue horizontal lines indicate the planes of the horizontal and vertical retinal cross-sections shown to the right and below the image, respectively. The cross-sections are centered on a point with an MP sensitivity of 14 decibels (dB). The inner segment/outer segment (IS/OS) layer is a thin line above the hyper-reflective retinal pigment epithelium, as indicated by the red arrows. It is intact at this specific retinal point. (1) A magnified view of the IS/OS layer in the horizontal cross-section. B, Scan of the same retina shown in A. The horizontal and vertical cross-sections are centered on a point with an MP sensitivity of 2 dB. The IS/OS layer indicated by the red arrows is disrupted at this point. (2) A magnified view of the IS/OS layer in the horizontal cross-section.
1256
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME
Figure 1.
1257
Ophthalmology Volume 120, Number 6, June 2013 Table 3. Retinal Characteristics on Optical Coherence Tomography Point Thickness and Sensitivity by IS/OS Layer Status Mean retinal thickness in m (SE) Mean retinal point sensitivity in dB (SE)
All
IS/OS Present
IS/OS Absent
P Value
325.2 (2.8) 10.51 (0.15)
310.6 (2.6) 11.50 (0.15)
373.1 (7.5) 7.27 (0.30)
⬍0.001 ⬍0.001
Retinal Characteristics by Type of Diabetic Macular Edema Present Cystoid Macular Edema
Diffuse Edema
352.7 (11.3) 27.0 (0.7) 9.9 (0.2)
247.1 (8.1) 13.9 (1.1) 12.1 (0.2)
Mean central subfield thickness in m (SE) Percentage of IS/OS disrupted (SE) Mean retinal point sensitivity in dB (SE)
⬍0.001 0.11 0.006
IS/OS ⫽ inner segment/outer segment; SE ⫽ standard error. The upper portion demonstrates retinal thickness and sensitivity for all points, points with IS/OS layer present, and points with IS/OS layer absent. The lower portion shows retinal thickness and sensitivity for microperimetry points with cystoid macular edema and diffuse edema.
that a 1-m increase in thickness resulted in a 0.04-dB (P⬍0.001) increase in retinal sensitivity. On the contrary, for points with an IS/OS layer present and a retinal thickness ⬎300 m, we found a 0.01-dB (P ⫽ 0.01) decrease in retinal sensitivity for every 1-m increase in thickness. When considering points with the IS/OS layer disrupted, we found that for points ⬍300 m, the thickness–sensitivity relationship was weakened to a 0.02-dB increase in sensitivity for every 1-m increase in thickness. For points ⬎300 m, the relationship remained similar to points with the IS/OS layer present. Finally, the presence of cystoid macular edema was associated with a significant 2.71-dB (P⬍0.001) decrease in retinal sensitivity.
Discussion In this index study, we have used a novel approach to analyze the effect of IS/OS junction status on retinal sensitivity and to investigate the thickness–sensitivity relationship in patients with DME. Disruption of the IS/OS junction is associated with a significant 3.28-dB decrease in retinal sensitivity after adjusting for confounding variables. In addition, we found that IS/OS disruption results in a weaker thickness–sensitivity association in retinal points that were thinner than 300 m. However, we also found that IS/OS disruption has no effect on the relationship between thickness and sensitivity in points thicker than 300 m. Our study suggests that the status of the IS/OS junction is a predictor of retinal sensitivity in patients with DME. Prior studies have shown that disruption of the IS/OS junction11,12 and external limiting membrane (a close correlate of the IS/OS junction)9 at the fovea results in a decreased VA measured by standard testing. To our knowledge, our study is the first to use MP testing to demonstrate that a similar relationship exists for points throughout the retina. Furthermore, we were able to obtain a large dataset because MP allows for 28 points of retinal data per eye, which allowed us to create a multilevel linear model to account for dependency between MP points and control for various factors that might confound the relationship between IS/OS status and MP sensitivity. The relationship between sensitivity and IS/OS status was maintained after correcting for
1258
confounding. The finding suggests that local photoreceptor integrity, as indicated by IS/OS status, is a strong predictor of local visual function throughout the retina. Our analysis of the thickness–sensitivity relationship suggests that there is a peak retinal sensitivity at approximately 300 m, which is similar to prior findings.14 When the IS/OS layer is disrupted and thickness is ⬍300 m, the thickness– sensitivity relationship was weaker than when the IS/OS layer is present. Such a relationship suggests that once the IS/OS layer is disrupted in retinas thinned by atrophic degeneration, thickness is no longer such a strong predictor of visual function. However, when the IS/OS layer is disrupted in points ⬎300 m, thickness remains an important predictor of retinal sensitivity. Therefore, in edematous retinas, increased thickness tends to have an effect independent of the effect of photoreceptor disruption. This finding may lend credence to the standard approach of today’s retina specialists to maximize resolution of edema in patients with DME. Recent studies have suggested that IS/OS status is a good prognostic indicator for DME therapy. One study18 found that patients treated with triamcinolone tended to do better when the IS/OS layer was intact. Another recent study found that IS/OS status was a good predictor of visual outcome in patients treated with pars plana vitrectomy.19 Although we agree that patients with IS/OS disruption may have reduced VA, our results suggest that retinal thickness remains an important predictor of visual function in patients with edema independent of IS/OS status. Therefore, in edematous patients, the clinician should strive to reduce macular edema regardless of a patient’s IS/OS status. We found that the presence of cysts in the edema (cystoid macular edema) results in decreased sensitivity, similar to findings in other studies.10,20 We found that cysts were significantly associated with retinal thickness and increased IS/OS disruption, albeit statistically insignificant. Our multilevel model revealed that the presence of cysts was associated with decreased sensitivity in DME independently of their effect of increasing retinal thickness and IS/OS disruption. Cysts, which usually occur in the inner nuclear layer or outer plexiform layer,21,22 have been hypothesized to have their effect by disrupting bipolar or ganglion cells.10
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME
Figure 2. Scatter plots of retinal thickness versus retinal sensitivity stratified by inner segment/outer segment (IS/OS) layer status. Lowess curve was fit in both graphs (thick black line).
Our findings, that cysts have an effect independent of their effect on retinal thickness and photoreceptor disruption, support such hypothesis.
Study Limitations Our study has provided valuable insights into the determinants of visual disruption in DME. However, the study also
has several limitations. We had a relatively small sample size of patients. In addition, there is always the possibility of residual confounding in our analysis. We did not have the data to control or model for duration or severity of diabetes, prior treatments for DME, and any concurrent ophthalmologic condition. Finally, because our study was a crosssectional study, it is difficult to establish temporality of the exposure (IS/OS status) and outcome (retinal sensitivity).
1259
Ophthalmology Volume 120, Number 6, June 2013 Table 4. Factors Affecting Retinal Point Sensitivity in Patients with Diabetic Macular Edema Factor Disrupted IS/OS layer* Thickness–sensitivity relationship† Points with IS/OS layer intact Increase in 1 m of thickness ⬍300 m Increase in 1 m of thickness ⬎300 m Points with IS/OS layer disrupted Increase in 1 m of thickness ⬍300 m Increase in 1 m of thickness ⬎300 m Cystoid macular edema†
Mean Change in Sensitivity (dB) ⫺3.28 0.04 ⫺0.01
0.02
95% CI
P Value
⫺3.89 to ⫺2.68
⬍0.001
0.02 to 0.06
⬍0.001
⫺0.02 to ⫺0.001
0.006 to 0.03
0.001
0.006
⫺0.01
⫺0.02 to ⫺0.003
⬍0.001
⫺2.71
⫺4.22 to ⫺1.21
⬍0.001
CI ⫽ confidence interval; IS/OS ⫽ inner segment/outer segment. Results from multilevel linear models with random intercepts built to assess the influence of various factors on retinal point sensitivity in patients with DME. In addition to the factors indicated, adjustments were also made for age, sex, race, and fixation location and stability. *Results are from a multilevel model without an interaction term for thickness and IS/OS status. † Results are from a multilevel model with an interaction term for thickness and IS/OS status.
In conclusion, we have demonstrated through our study that the absence of the IS/OS layer has a significant impact on retinal sensitivity in patients with DME. We have also shown that retinal thickness is an important predictor of retinal function in edematous patients even when the IS/OS layer is disrupted. Therefore, although active monitoring of the IS/OS layer by retina physicians may provide further insight into the potential for visual function of a patient, IS/OS disruption should not be a contraindication for treatments to reduce retinal edema. As our understanding of the pathophysiologic effect of DME on the retina broadens, on the basis of what we learn from the roles of retinal structures such as the IS/OS, we may be able to use our current therapies more appropriately, as well as design more targeted and effective future treatments for DME.
References 1. Chen E, Looman M, Laouri M, et al. Burden of illness of diabetic macular edema: literature review. Curr Med Res Opin 2010;26:1587–97. 2. Sjostrand J, Popovic Z, Conradi N, Marshall J. Morphometric study of the displacement of retinal ganglion cells subserving cones within the human fovea. Graefes Arch Clin Exp Ophthalmol 1999;237:1014 –23. 3. Antonetti DA, Barber AJ, Bronson SK, et al, JDRF Diabetic Retinopathy Center Group. Diabetic retinopathy: seeing beyond glucose-induced microvascular disease. Diabetes 2006; 55:2401–11. 4. Diabetic Retinopathy Clinical Research Network, Elman MJ, Aiello LP, Beck RW, et al. Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema. Ophthalmology 2010;117:1064 –77. 5. Martidis A, Duker JS, Greenberg PB, et al. Intravitreal triamcinolone for refractory diabetic macular edema. Ophthalmology 2002;109:920 –7.
1260
6. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema: Early Treatment Diabetic Retinopathy Study report no. 4. Int Ophthalmol Clin 1987;27:265–72. 7. Ota M, Tsujikawa A, Murakami T, et al. Association between integrity of foveal photoreceptor layer and visual acuity in branch retinal vein occlusion. Br J Ophthalmol 2007;91: 1644 –9. 8. Shin HJ, Chung H, Kim HC. Association between integrity of foveal photoreceptor layer and visual outcome in retinal vein occlusion [report online]. Acta Ophthalmol 2011;89:e35– 40. Available at: http://onlinelibrary.wiley.com/doi/10.1111/j. 1755-3768.2010.02063.x/full. Accessed October 7, 2012. 9. Oishi A, Hata M, Shimozono M, et al. The significance of external limiting membrane status for visual acuity in agerelated macular degeneration. Am J Ophthalmol 2010;150: 27–32. 10. Murakami T, Nishijima K, Sakamoto A, et al. Association of pathomorphology, photoreceptor status, and retinal thickness with visual acuity in diabetic retinopathy. Am J Ophthalmol 2011;151:310 –7. 11. Maheshwary AS, Oster SF, Yuson RM, et al. The association between percent disruption of the photoreceptor inner segment-outer segment junction and visual acuity in diabetic macular edema. Am J Ophthalmol 2010;150:63–7. 12. Otani T, Yamaguchi Y, Kishi S. Correlation between visual acuity and foveal microstructural changes in diabetic macular edema. Retina 2010;30:774 – 80. 13. Vujosevic S, Midena E, Pilotto E, et al. Diabetic macular edema: correlation between microperimetry and optical coherence tomography findings. Invest Ophthalmol Vis Sci 2006; 47:3044 –51. 14. Hatef E, Colantuoni E, Wang J, et al. The relationship between macular sensitivity and retinal thickness in eyes with diabetic macular edema. Am J Ophthalmol 2011;152:400 –5. 15. Sepah YJ, Hatef E, Colantuoni E, et al. Macular sensitivity and fixation patterns in normal eyes and eyes with uveitis with and without macular edema. J Ophthalmic Inflamm Infect 2012;2:65–73.
Yohannan et al 䡠 IS/OS Status and Retinal Sensitivity in DME 16. Landa G, Rosen RB, Garcia PM, Seiple WH. Combined three-dimensional spectral OCT/SLO topography and microperimetry: steps toward achieving functional spectral OCT/SLO. Ophthalmic Res 2010;43:92– 8. 17. Panozzo G, Gusson E, Parolini B, Mercanti A. Role of OCT in the diagnosis and follow up of diabetic macular edema. Semin Ophthalmol 2003;18:74 – 81. 18. Shin HJ, Lee SH, Chung H, Kim HC. Association between photoreceptor integrity and visual outcome in diabetic macular edema. Graefes Arch Clin Exp Ophthalmol 2012;250: 61–70.
19. Sakamoto A, Nishijima K, Kita M, et al. Association between foveal photoreceptor status and visual acuity after resolution of diabetic macular edema by pars plana vitrectomy. Graefes Arch Clin Exp Ophthalmol 2009;247:1325–30. 20. Soliman W, Hasler P, Sander B, Larsen M. Local retinal sensitivity in relation to specific retinopathy lesions in diabetic macular oedema. Acta Ophthalmol 2012;90:248 –53. 21. Tso MO. Pathology of cystoid macular edema. Ophthalmology 1982;89:902–15. 22. Yanoff M, Fine BS, Brucker AJ, Eagle RC Jr. Pathology of human cystoid macular edema. Surv Ophthalmol 1984;28(Suppl):505–11.
Footnotes and Financial Disclosures Originally received: May 14, 2012. Final revision: November 25, 2012. Accepted: December 4, 2012. Available online: March 15, 2013.
Manuscript no. 2012-696.
Financial Disclosure(s): The author(s) have made the following disclosure(s): The Johns Hopkins University has received research support from OPKO, Inc, and Optos, Inc.
Retinal Imaging Research and Reading Center, Wilmer Eye Institute, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Presented at the Association for Research in Vision and Ophthalmology Annual Meeting, May 6 –10, 2012, Fort Lauderdale, Florida.
1
2
Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland.
3
Stanley M. Truhlsen Eye Institute, University of Nebraska Medical Center, Omaha, Nebraska.
Correspondence: Quan Dong Nguyen, MD, MSc, Stanley M. Truhlsen Eye Institute, University of Nebraska Medical Center, 3902 Leavenworth Street, Omaha, Nebraska 68105. E-mail: [email protected].
1261