Diurnal Variation in Retinal Thickening Measurement by Optical Coherence Tomography in Center-Involved Diabetic Macular Edema FREE

Optical coherence tomography (OCT) is a noninvasive method for measuring the thickness of the central retina. It has become a standard tool in the management of patients with diabetic macular edema (DME). This imaging technique uses low-coherence interferometry to produce cross-sectional tomograms of the posterior segment eye structures. An 850-nm diode laser is used to measure the delay in light backscattering from different layers in the retina from which the retinal thickness is calculated. Retinal thickness is determined from multiple individual axial scans arrayed in a radial pattern along each of 6 linear segments all intersecting at the center. A computer algorithm is used to determine the inner and outer retinal boundaries for each scan.

Several small studies have reported that retinal thickening may vary during the day in some eyes with DME, generally being thickest in the morning and thinner later in the day.^1- 3 Because OCT is a very sensitive method used to measure retinal thickening and judge response to DME treatment in the clinic and in clinical trials, it is important to have reliable information on the diurnal variation in retinal thickening. These findings are essential to the appropriate interpretation of OCT results and the design of studies using this technique. We designed the current study to evaluate retinal thickening measured by OCT throughout the day in eyes with DME.

The study was conducted at 25 sites participating in the Diabetic Retinopathy Clinical Research Network (DRCR.net). The study protocol was approved by the institutional review boards, and each subject gave written informed consent for participation in the study.

To be eligible, a subject was required to have at least one eye with (1) definite retinal thickening due to DME based on clinical examination involving the center of the macula, (2) OCT central subfield of 225 μm or greater, (3) pupil dilation of 5 mm or larger, and (4) no treatment for DME within the prior 3 months. Subjects were excluded if there was a history of chronic renal failure requiring dialysis or kidney transplant, congestive heart failure currently under treatment, or blood pressure higher than 180/110 mm Hg.

Retinal thickness measurements were made with the OCT3 system (Stratus OCT3; Carl Zeiss Meditec, Dublin, Calif) at 8 AM, 9 AM, 10 AM, 12 noon, 2 PM, and 4 PM (within ± 30-minute time windows). In addition, at 8 AM, 12 noon, and 4 PM, the following were measured: visual acuity (following a refraction) using the electronic ETDRS (Early Treatment Diabetic Retinopathy Study) method,⁴ blood glucose using a home glucose meter, and blood pressure. On the day of testing, or within 1 month of the visit, fundus photographs were obtained. Hemoglobin A_1c was measured if not obtained during the prior 3 months. Additional health and past ocular history data were captured on electronic case report forms (Table 1).

Optical coherence tomography images were obtained on each eye following pupil dilation by a certified operator using the OCT3 system. The same operator obtained the OCT images once at each time point. A second measurement at each time point was made by the same operator or a different operator. Scans of inadequate quality (eg, standard deviation >10% or signal strength score <6) were to be repeated until they met these specified quality metrics. Scans were sent to the DRCR.net Reading Center at the University of Wisconsin–Madison for grading. For scans with standard deviations of the center point thickness that were 10.5% or greater, scans with inaccurately drawn automated boundary lines, or scans with decentration,⁵ center point thickness was measured manually with calipers from the retina thickness map scans (13% of images). For such scans, central subfield thickness was imputed from the manually determined center point thickness using a regression equation (because the correlation of the 2 measurements was 0.98 for 897 automated measurements).

The sample size was estimated to be 120 eyes such that the half-width of a 95% confidence interval on a 0.20-point estimate of the proportion of eyes experiencing meaningful diurnal variation would be 0.07 and on a 0.10-point estimate, 0.05.

Change in central subfield thickening was the primary outcome variable. Only eyes with both 8 AM and 4 PM OCT scans in which the 8 AM central subfield thickness was 225 μm or greater and the initial OCT scan was obtained within 45 minutes of 8 AM (an analysis window 30 minutes longer than the protocol-specified time window) were included in the analysis. Eleven subjects did not have either eye meeting these criteria and were excluded from analyses. Optical coherence tomography measurements obtained twice at the same time point were averaged for analysis.

Retinal thickening, ie, excess retinal thickness, was defined as the observed thickness minus the mean normal central subfield thickness of 202 μm (based on unpublished data provided by Carl Zeiss Meditec from a study of 260 nondiabetic eyes with normal maculae). Therefore, relative change in retinal thickening was calculated as (4 PM thickness minus 8 AM thickness)/(8 AM thickness minus mean normal thickness). The mean change in absolute and relative retinal thickening in the central subfield between 8 AM and 4 PM was computed and a 95% confidence interval constructed, accounting for the intereye correlation in subjects with 2 study eyes using repeated measures mixed models.

The proportions of study eyes with change in central subfield thickening meeting the composite threshold end point of at least 25% and at least 50 μm at consecutive time points or between 8 AM and 4 PM were calculated. The proportion of eyes with a change in central subfield thickening of at least 50 μm was also calculated. The threshold of 25% was arbitrarily chosen as a relative change that was likely to be clinically meaningful. The threshold of 50 μm was likewise considered to be clinically meaningful and exceeded the 95% half-width confidence interval for detection of change of 40 μm (based on reproducibility data from 1147 eyes). The composite end point was chosen to use the relative change measurement while assuring that the magnitude of change exceeded this 95% confidence interval and was clinically meaningful.

Results were evaluated in 3 subgroups based on the baseline central subfield retinal thickness: 225 to 300 μm, 301 to 450 μm, and greater than 450 μm. The effect of baseline thickness on the change in thickening from 8 AM to 4 PM was assessed using a repeated measures model adjusting for intereye correlation in subjects with 2 study eyes. The effects of additional clinical and ocular characteristics on the mean change in central subfield thickening were evaluated by analysis of covariance also adjusting for 8 AM central subfield thickness and accounting for intereye correlation. Correlations were used to define the relationship between change in central subfield thickening and the change in visual acuity or blood glucose. They were calculated in repeated measures models (to account for the correlation within subjects with 2 study eyes) based on the likelihood ratio as defined by Magee.⁶ To satisfy parametric assumptions, change in visual acuity was truncated at 3 standard deviations from the mean; truncation did not meaningfully change the results. Statistical analyses were performed using SAS software version 9.1 (SAS Inc, Cary, NC).

The study included 96 subjects, of whom 36 (38%) had 1 study eye and 60 (63%) 2 study eyes for a total of 156 study eyes. The baseline characteristics of the subjects and study eyes are shown in Table 1. Among the 156 study eyes, the mean ± SD central subfield retinal thickness at baseline (8 AM) was 368 ± 132 μm. There were 71 study eyes (46%) with baseline thickness of 225 to 300 μm, 53 (34%) with a thickness of 301 to 450 μm, and 32 (21%) with a thickness greater than 450 μm. Of the 96 subjects included in the analyses, all with 8 AM and 4 PM measurements, 1 subject missed both 9 AM measurements and 2 subjects missed the second 10 AM measurement. All 96 subjects completed both measurements at 12 PM and 2 PM.

The mean change in relative central subfield retinal thickening from 8 AM to 4 PM was a decrease of 6% (95% confidence interval, −9% to −3%) and the mean absolute change was a decrease of 13 μm (95% confidence interval, −17 to −8) (Figure 1 and Table 2). Five eyes (3%; 95% confidence interval, 1% to 8%) of 4 subjects reached the predefined composite outcome of 25% or greater decrease in central retinal thickening and 50 μm or greater decrease in central retinal thickening between 8 AM and 4 PM (Table 3). No additional eyes met these criteria at 2 consecutive earlier time points and not between 8 AM and 4 PM. Among the 5 eyes reaching the composite decrease in central subfield thickening, this decrease was first achieved at 2 PM in 2 eyes and 4 PM in 3 eyes. The maximum relative decrease was observed at 4 PM in all 5 eyes (Figure 2). Two eyes (1%; 95% confidence interval, 0% to 5%) of 2 subjects reached the predefined composite outcome for an increase in central subfield thickening, 25% or greater and 50 μm or greater. Tables 2 and 3 detail these changes subgrouped by 8 AM central subfield thickness.

On average, central subfield thickening decreased throughout the day by approximately 9.2 μm (95% confidence interval, 6.1 to 12.2) for every 100 μm of increased central subfield thickness at 8 AM, but there was considerable variation between individuals (Figure 3). The absolute change was significantly greater in retinas that were thicker at 8 AM (P<.001), but the relative change was not (P = .14) (Table 2). No additional factors were found to be significantly associated with the degree of diurnal change (P≥.01, Table 4).

From 8 AM to 4 PM, best corrected electronic ETDRS visual acuity letter score increased on average by 1.0 letters (95% confidence interval, −0.05 to +2.1 letters) (Figure 4), improving by 10 or more letters in 16 eyes (10%) and worsening by 10 or more letters in 9 eyes (6%). Of the 5 eyes with the composite outcome for a reduction in retinal thickening, visual acuity improved 5 or more letters in 1 eye, worsened 5 or more letters in 2 eyes, and changed less than 5 letters in 2 eyes. Change in retinal thickening from 8 AM to 4 PM correlated poorly with change in visual acuity (r = −0.05) (Figure 5), reflecting the large proportion of eyes with little change in thickening and the variability inherent in the measurements of both acuity and retinal thickening.

Blood glucose concentration increased from 8 AM to 4 PM by 50 mg/dL (2.78 mmol/L) or more in 21 of 95 subjects (8 AM blood glucose for 1 subject was missing) and decreased by 50 mg/dL (2.78 mmol/L) or more in 23 subjects. Change in retinal thickening from 8 AM to 4 PM correlated weakly (r = −0.13) with change in blood glucose. Results for other OCT measures such as center point thickness, average of central and inner subfield thickness, and macular volume were similar to the results for the central subfield (data not shown).

This study demonstrates that although on average there are slight decreases in retinal thickening during the day, most eyes with DME have little meaningful change in OCT central subfield thickening between 8 AM and 4 PM. Only 3% of eyes (5 of 156) met the composite threshold of both 25% or greater and 50 μm or greater decrease in retinal thickening at 2 consecutive time points or between 8 AM and 4 PM. However, 2 eyes (1%) increased by these amounts over this period as well. Although a change in retinal thickening of 50 μm may be a meaningful clinical change in an individual patient, with current OCT technology it is not possible to differentiate true change in retinal thickening from the variability of measurement when the difference is less than 40 μm, the measurement of variability.

A study by Frank et al¹ suggested that OCT-measured retinal thickening in DME varied according to time of day. In that study of 10 subjects, retinal thickness was measured at 8:00 AM, 11:00 AM, 2:00 PM, and 5:00 PM. Four of the 10 subjects had decreased thickening in 1 or more macular subfields in one or both eyes over the course of the day. The eyes with greater initial retinal thickening tended to show a greater absolute decline over the day than did the eyes with less initial thickening. The magnitude of the decrease in retinal thickening was relatively small (estimated at approximately a mean of 15 μm from the published figures), although statistically significant by paired t test comparing 8 AM with later times. In the study by Frank et al,¹ almost all of the change in retinal thickening occurred by 11 AM. In contrast, the 5 eyes in our study with a composite decrease in retinal thickening changed throughout the data collection period through the final 4 PM evaluation (Figure 2). An earlier study by Sternberg et al⁷ that used psychovisual assessments of 3 patients with DME noted that visual function in patients with macular edema may vary with time of day, being generally worse in the morning. Such a finding supports the anatomic observations in the study by Frank et al.¹ Both authors suggested that the diurnal change may be related to nocturnal recumbency playing a role in retinal vascular homeostasis. In both studies, patients were admitted to the hospital so that the measurements could be obtained immediately on arising, a methodological difference from the current study.

Larsen et al² evaluated change in retinal thickening from evening to morning in 12 eyes of 12 patients with DME. There was a statistically significant mean increase in thickening overnight of 21 μm in the DME group, which was statistically significant compared with the results in 14 eyes of 7 nondiabetic control subjects. Reproducibility of measurements was not addressed in that study using the OCT1 machine. Patients with DME also demonstrated a mean loss of 5 ETDRS letters of acuity overnight. Recently, Polito et al³ published data from 15 eyes of diabetic macular edema patients and 10 eyes of nondiabetic controls who had OCT3 fast macular thickness map scans at 4 time points between 9 AM and 6 PM on a single day. They did not detect any suggestion of diurnal variation in the control eyes or the macular edema eyes with center subfield thickness less than 300 μm. However, they found that among the 9 eyes with center subfield thickness of 300 μm or greater, 4 demonstrated a consistent decrease in thickness over the course of the day, and this subgroup had a mean decrease of central subfield thickness of 9.4%. Two of these 4 eyes had a 5-letter increase in visual acuity.

Our data suggest that decreasing retinal thickening in eyes with DME is a real phenomenon that may occur over the course of a single day in some patients. This change is generally of small magnitude and only occasionally achieves levels that may be considered clinically relevant. Indeed, the mean change of approximately −15 μm measured in a small cohort of DME patients by Frank et al¹ and the 9.4% reduction measured among 9 subjects by Polito et al³ during daytime hours are similar to the mean change in the 301- to 450-μm group in our study.

There was also little correlation between change in retinal thickening and change in visual acuity. This is in agreement with the study of Larsen et al² and not surprising given the small absolute changes in retinal thickening noted in this study. A separate article authored by the DRCR.net details the relationship between central retinal thickening and visual acuity both before and after laser photocoagulation.⁸

In summary, although some patients with clinically significant macular edema appear to have progressive decreases in retinal thickening during the course of the day, the extent and frequency of these changes do not appear to be of a sufficient magnitude to substantially impact clinical trials (especially if there is a randomly assigned control group and no systematic difference between treatment groups in the time of day of measurement). Our study did not address whether the meaningful diurnal changes exhibited by the small number of patients would be consistently present on a daily basis. However, the clinical impact of these changes is likely to be small.

Correspondence: Ronald P. Danis, MD, Jaeb Center for Health Research, 15310 Amberly Dr, Suite 350, Tampa, FL 33647 (drcrnet@jaeb.org).

Financial Disclosure: None reported.

Funding/Support: This study was supported through a cooperative agreement from the National Eye Institute and the National Institute of Diabetes and Digestive and Kidney Diseases (EY14231, EY14269, and EY14229).Article

Clinical Sites That Participated in This Protocol. Sites are listed in order by number of subjects enrolled into the study. The number of subjects enrolled is noted in parentheses preceded by the site name and location. Personnel are listed as (I) for investigator, (C) for coordinator, (V) for visual acuity tester, and (P) for photographer.

Joslin Diabetes Center, Boston, Mass (13): George S. Sharuk (I); Paul G. Arrigg (I); Sabera T. Shah (I); Timothy J. Murtha (I); Deborah K. Schlossman (I); Ann Kopple (C); Margaret E. Stockman (C); Leila Bestourous (V); Jerry D. Cavallerano (V); Robert W. Cavicchi (P); James Strong (P); International Eye Center, Tampa, Fla (12): Don John Perez Ortiz (I); Madelyn Alvarez (C); Rita L. Johnson (C); Ross Jarrett (P); Lena M. Yagecic (P); Retina Consultants, Providence, RI (11): Magdalena G. Krzystolik (I); Harold A. Woodcome (I); Paul B. Greenberg (I); Emiliya German (C); Sandra Henriques (V); Claudia Salinas (V); Erika Banalewicz (V); Mark Hamel (P); Texas Retina Associates, Dallas (10): Gary E. Fish (I); Rajiv Anand (I); Robert C. Wang (I); Jean Arnwine (C); Brenda Sanchez (V); Sally Arceneaux (V); Penny Ellenich (P); Diana Jaramillo (P); Hank Aguado (P); University of Washington Medical Center, Seattle, Wash (7): James L. Kinyoun (I); Susan A. Rath (C); Chuck Stephens (P); Retina Consultants, PLLC, Slingerlands, NY (6): Paul M. Beer (I); Eugenia Olmeda (C); Robert Davis (P); Joe Fischer (P); Texas Retina Associates, Lubbock (5): Michel Shami (I); Phyllis Pusser (C); Linda Squires (V); Thom F. Wentlandt (P); Charlotte Eye, Ear, Nose and Throat Associates, PA, Charlotte, NC (4): David J. Browning (I); Andrew N. Antoszyk (I); Angela K. Price (C); Heather L. Murphy (V); Michael D. McOwen (P); Linda M. Davis (P); Loraine M. Clark (P); Palmetto Retina Center, Columbia, SC (4): W. Lloyd Clark (I); John A. Wells (I); Mallie M. Taylor (C); Robbin Spivey (V); Amy B. Hickman (P); Paducah Retinal Center, Paducah, KY (4): Carl W. Baker (I); Tracey M. Caldwell (C); Lynnette F. Lambert (V); Dawn D. Smith (P); Retina Research Center, Austin, Tex (3): Brian B. Berger (I); Margaret Rodriguez (C); Steve A. Jeffers (C); Diana S. Dorn (V); Bobbi Gallia (V); Ben Ostrander (P); Ophthalmic Consultants of Boston, Boston, Mass (3): Trexler M. Topping (I); Lori A. Tyler (C); Kathleen Marino (V); Margie Graham (P); Medical College of Wisconsin, Milwaukee (3): Judy E. Kim (I); William J. Wirostko (I); Christine Y. Lange (C); Troy S. Drescher (C); Kathy J. Selchert (P); Joseph R. Beringer (P); Sarasota Retina Institute, Sarasota, Fla (3): Keye L. Wong (I); John H. Niffenegger (I); Christine Holland (C); Mark Sneath (P); Retina-Vitreous Surgeons of Central New York, PC, Syracuse (3): G. Robert Hampton (I); Paul F. Torrisi (I); Bryan K. Rutledge (I); Cindy J. Grinnell (C); Lynn M. Kwasniewski (V); Peter B. Hay (P); Lynn A. Capone (P); Northwestern Medical Faculty Foundation, Chicago, Ill (2): Alice T. Lyon (I); Jeevan R. Mathura (I); Annmarie C. Munana (C); Lori A. Kaminski (C); Jonathan Shankle (P); Carolina Retina Center, Columbia, SC (2): Jeffrey G. Gross (I); Barron C. Fishburne (I); Peggy W. Cummings (C); Regina A. Gabriel (V); Randall L. Price (P); Denver Health Medical Center, Denver, Colo (2): Antonio P. Ciardella (I); Dorothy L. Thomas (C); Colleen J. Smith (V); Debbie M. Brown (P); Central Florida Retina Institute, Lakeland, Fla (2): Scott M. Friedman (I); Steve Carlton (C); Damanda A. Fagan (V); Valley Retina Institute, McAllen, Tex (2): Victor H. Gonzalez (I); Jessica Herrera (C); Alma Herrera (V); Cassandra Garza (P); Rosie M. Corona (P); Daniel Cuellar (P); Vanderbilt University Medical Center, Nashville, Tenn (2): Franco M. Recchia (I); Genise G. Mofield (V); Tony Adkins (P); Cynthia C. Recchia (P); Retinal Consultants, Inc, Dublin, Ohio (1): Frederick H. Davidorf (I); Robert B. Chambers (I); Jill D. Milliron (C); Jerilyn G. Perry (V); Scott J. Savage (P); Vitreoretinal Consultants, Houston, Tex (1): David M. Brown (I); Rebecca Delagarza (C); Karin A. Mutz (V); Eric N. Kegley (V); Central Florida Retina, Lake Mary, Fla (1): Preston P. Richmond (I); Douglas F. Lieb (I); Laverne Denise Davila (C); Ross Jarrett (P); Retina Northwest, PC, Portland, Ore (1): Mark A. Peters (I); Stephen Hobbs (C); Katie J. Reichenberger (C); Marcia Kopfer (V); and Howard Daniel (P).

DRCR.net Coordinating Center, Tampa, Fla: Roy W. Beck (Executive Director), Kimberly E. McLeod (DRCR.net Associate Director), Kelly A. Blackmer, Brian B. Dale, Adam R. Glassman, Nicola B. Hill, Paula A. Johnson, Craig Kollman, Alisha N. Lawson, Brenda L. Loggins, Ana C. Perez, Apryl C. Quillen, Cynthia R. Stockdale, and Samara Strauber.

DRCR.net Chairman's Office, Boston, Mass: Neil M. Bressler, Baltimore, Md (Network Chair); and Lloyd P. Aiello, Boston, Mass (Network Chair, 2002-2005).

Fundus Photograph Reading Center, Madison, Wis: Matthew D. Davis (Director Emeritus), Ronald P. Danis (Director), Larry Hubbard (Associate Director), James Reimers (Lead Color Photography Evaluator), Pamela Vargo (Lead Photographer), Ericka Lambert (Digital Imaging Specialist), and Dawn Myers (Lead OCT Evaluator).

National Eye Institute, Bethesda, Md: Päivi H. Miskala and Donald F. Everett (2002-2004).

DRCR.net Executive Committee: Lloyd P. Aiello (Chair, 2002-2005), Roy W. Beck, Neil M. Bressler (Chair), David M. Brown, David J. Browning, Ronald P. Danis, Matthew D. Davis, Michael J. Elman, Frederick L. Ferris, Adam R. Glassman, Kimberly E. McLeod, and Päivi H. Miskala.

OCT Diurnal Variation Study Steering Committee: Lloyd P. Aiello, Roy W. Beck, Neil M. Bressler, Alexander J. Brucker, Steve Carlton, Emily Y. Chew, Ronald P. Danis (protocol chair), Frederick L. Ferris, Don S. Fong, Adam R. Glassman, Jeffrey G. Gross, Julia A. Haller, Helen K. Li, Kimberly McLeod, and Päivi H. Miskala.