Abstract
Purpose
To compare evaluation by clinical examination with image grading at a reading center (RC) for the classification of diabetic retinopathy (DR) and diabetic macular edema (DME).
Methods
ACCORD and FIND had similar methods of clinical and fundus photograph evaluation. For analysis purposes the photographic grading scales were condensed to correspond to the clinical scales and agreement between clinicians and reading center classification were compared.
Results
6902 eyes of ACCORD participants and 3638 eyes of FIND participants were analyzed for agreement (percent, kappa) on DR on a 5 level scale. Exact agreement between clinicians and RC on DR severity category was 69% in ACCORD and 74% in FIND (Kappa 0.42 and 0.65). Sensitivity of the clinical grading to identify presence of mild nonproliferative retinopathy or worse was 0.53 in ACCORD and 0.84 in FIND. Specificities were 0.97 and 0.96, respectively. DME agreement in 6649 eyes of ACCORD participants and 3366 eyes of FIND participants was similar in both studies (Kappa 0.35 and 0.41). Sensitivities of the clinical grading to identify DME were 0.44 and 0.53 and specificities were 0.99 and 0.94, respectively.
Conclusion
Our results support the use of clinical information for defining broad severity categories, but not for documenting small to moderate changes in DR over time.
Keywords: ACCORD, FIND, Diabetic retinopathy, fundus photography
Most studies evaluating the sensitivity and specificity of color fundus photographs compared to clinical grading of diabetic retinopathy (DR) have been oriented toward diabetes screening and clinical care rather than classification for clinical trials. Clinical trials generally use more complex levels for the classification of disease for eligibility and disease progression, whereas screening or clinical care is often oriented toward detection of a threshold that would prompt an intervention. Consequently, photographic protocols for clinical studies are generally complex and technically difficult in order to maximize the detection and severity of individual lesions upon which the Early Treatment Diabetic Retinopathy Study (ETDRS) scale is based, versus simpler imaging methods that are less time consuming and more suitable for cost-efficient sampling of retinal lesions are used for photographic screening. The performance of clinician classification of DR for clinical trials should be examined carefully because of the expense and logistical complexity involving in obtaining and grading fundus photographs. In two large NIH-supported studies that enrolled patients with varying degrees of diabetic retinopathy, the Family Investigations of Nephropathy in Diabetes(FIND) Eye substudy1 and the Action to Control Cardiovascular Risk in Diabetes (ACCORD) Eye substudy,2,3 the availability of ophthalmoscopic examinations recorded on standard forms by clinicians (ophthalmologists and optometrists) and of 7-field fundus photographs collected concurrently and graded centrally provided an opportunity to examine degree of agreement between clinical and photographic assessments. The results of this analysis may guide policy makers and grant writers to determine when clinical evaluation would suffice for a given trial.
Methods
FIND and ACCORD studies
The design and methodology of FIND and ACCORD Eye studies have been published previously.21-23 Briefly, the goal of the cross-sectional FIND Study was to determine genetic risk factors for diabetic nephropathy. FIND enrolled families with probands who had advanced diabetic nephropathy, along with their diabetic siblings (>10 years history of diabetes). All FIND family study members were invited to participate in the FIND-Eye Study in an attempt to localize genes underlying susceptibility to DR. ACCORD was a randomized clinical trial to test the effects of lowering blood glucose, blood pressure and management of the dyslipidemia on the development and progression of diabetic complications in patients with Type 2 diabetes who had cardiovascular disease or risk factors for it. ACCORD participants who had a history of proliferative diabetic retinopathy (PDR) that had been treated with laser photocoagulation or vitrectomy were excluded from the ACCORD Eye Study. Only the baseline ACCORD data were included in the analyses for this manuscript. Both studies were compliant with the tenets of the Declaration of Helsinki, and the Health Insurance Portability and Accountability Act, and both were approved by the institutional review boards of the participating centers. Written informed consent was obtained from the subjects in both studies.
Clinical evaluation
All study participants received a comprehensive eye examination by a study retina specialist (ACCORD) or a retina specialist, ophthalmologist or optometrist (FIND) and underwent dilated fundus photography. The examiner was instructed to provide a clinical grade for (1) the severity of DR, and (2) the presence and location of diabetic macular edema (DME).
The retinopathy level for each eye was determined by the presence of the most severe lesion as follows
| 1: No retinopathy |
| 2: Microaneurysms (Ma) only |
|
3: Mild to moderate non-proliferative retinopathy, defined as Ma plus retinal hemorrhages and/or exudates (lipid deposits and/or cotton wool spots) |
|
4: Moderately severe to severe non-proliferative retinopathy, defined as at least one of the following |
| a. Definite venous beading |
| b. Obvious intraretinal microvascular abnormalities |
|
c. Hemorrhages/Microaneurysms ≥ standard photograph 2A in at least two quadrants (Figure 1) |
| 5:Proliferative retinopathy or scars of panretinal (scatter) photocoagulation |
Macular edema was defined as retinal thickening within one disc diameter of the center of the macula, with or without lipid deposits. Involvement of the center of the macula was also evaluated. No specific minimum area of thickening was specified for this classification. DME was classified as absent, present (some thickening within one disc diameter of the center) but not center involved, and center-involved. Training of clinicians was not a part of these studies and clinicians performed the assessments according to their standard practices. No specific certification or reproducibility assessment of clinicians was performed.
Figure 1.
This grading system differs slightly from the International Classification of Diabetic Retinopathy, Appendix A4 Table 1), in that the more severe levels from Moderately Severe NPDR are binned into the Severe category and the level is named “Moderately Severe to Severe NPDR”. This provides a cut-point of similar clinical importance that is easier to describe and recognize clinically than is severe NPDR alone. The FIND/ACCORD scale also includes scars of scatter (panretinal) photocoagulation as part of the PDR category.
Table 1.
Corresponding categories on clinical scale and condensed ETDRS scale
| ACCORD/FIND Clinical Scale | Corresponding ETDRS Level |
|---|---|
| 1 –No Retinopathy | 10-15 |
| 2 – Microaneurysms Only | 20 |
| 3 – Microaneurysms plus retinal hemorrhages and/or exudates (lipid deposits and/or cotton wool spots) |
35-43, 47A |
| 4 – At least one of: venous beading, obvious IRMA, Hemorrhages/Microaneurysms >=Std. photograph 2A in at least 2 quadrants |
47B - 53 |
| 5 – Proliferative retinopathy or status post panretinal photocoagulation |
60 + |
IRMA – intraretinal microvascular abnormalities
Fundus Photography
All patients underwent pharmacologic pupil dilation followed by modified 7-field stereoscopic photography using 30 or 35 degree field. Images were obtained in film format and sent to the Reading Center according to standard procedures. The fundus photographs were evaluated centrally by the Reading Center using the same standardized protocol. Film sets were viewed upon a standard light box (6500° K color temperature), using a Donaldson stereo viewer (5X).
Grading of the Fundus
Photographs Stereoscopic fundus photographs in FIND and at baseline in ACCORD were graded according to the ETDRS protocol and analyzed according to the modified ETDRS severity scale5 categorized as follows: DR absent (levels 10-15), very mild DR (level 20), mild to moderate non-proliferative diabetic retinopathy (NPDR) (levels 35, 43 and 47A), moderately severe to severe NPDR (levels 47B- 53), and proliferative diabetic retinopathy (PDR) (levels 60-85). DME was graded from stereoscopic views of the macula, using ETDRS macular grids to demarcate subfields. DME features were graded according to the modified ETDRS procedure5 for area of retinal thickening within each subfield of the grid, clinically significant macular edema (CSME) and retinal thickening at the center of the macula.6 For this analysis, photographic grading steps were simplified to match the clinical grading scheme: macular edema was scored on the clinical evaluations using a three-step ordinal scale of absent, definitely present within 1 disc diameter (DD) from the center of the macula but without definite involvement of the center, and DME present with center involved.
The eye exam form used by the examining ophthalmologist provided scoring of diabetic retinopathy on a 5 step ordinal scale; which was mapped to the corresponding ETDRS levels for comparison as shown in Table 1.
Statistical analysis
Calculated measures of agreement included the percent agreement (both exact agreement and agreement within 1 step), and simple and weighted kappa (κ) statistics. The kappa statistics were on the eye level. The weights used were Cicchetti-Allison weights with 1.0 for complete agreement and a 1/(L-1) decrease for each step disagreement where L was the number of categories. Landis and Koch’s guidelines 7 were used to evaluate simple and weighted kappa statistics: 0.00-0.20, slight agreement; 0.21-0.40, fair; 0.41-0.60, moderate; 0.61-0.80, substantial; and 0.81-1.00, almost perfect agreement. Only eyes that were graded both at the clinics and the reading center were included in calculating measures of agreement. We calculate sensitivity and specificity (and 95% confidence intervals) of the clinical scales using techniques to account for correlation between eyes8-11. We dichotomized both the clinical scale and ETDRS scales. Our analysis compared absence versus presence of any DR and presence of lower levels of retinopathy versus moderately severe NPDR or higher, but we also examined other cut points. For DME, we compared absence versus presence of any DME and the absence versus presence of DME at the center. We calculated the area under the receiver operating characteristic (ROC) curve (AUC) for the most severe eye using the 5-level ACCORD clinical scale (treated as an ordinal variable).
Results
ACCORD
Included in analysis for DR severity agreement were 6902 eyes (3474 participants) that had been assessed at the clinics as well as at the reading center (Table 2). DR was graded as absent by both assessments in 3715 eyes (53.8%). Exact agreement on corresponding DR category was seen in 4762 eyes (69%) and agreement within one step in 6301 eyes (91%). Agreement by kappa statistic was moderate – simple kappa 0.42 (95% CI 0.40-0.44) and weighted kappa 0.54 (95% CI 0.52-0.55). Reading center grades tended to be higher than clinical grades (P<0.01). The largest disagreements were in the 1070 eyes with microaneurysms only by photo grading, 817 of which (76%) were judged to have no retinopathy clinically, and in the 122 eyes with PDR by photo grading, only 26 of which (21%) had PDR by clinical assessment.
Table 2.
Diabetic Retinopathy Severity comparison of the collapsed ETDRS photographic scale and the ACCORD clinical scale: N (%)
| ACCORD Clinical Scale |
ETDRS Severity Level |
Total | ||||
|---|---|---|---|---|---|---|
| 10-15 | 20 | 35-43,47A | 47B-53 | 60+ | ||
| 1 | 3715 (53.8) |
817 (11.8) |
404 (5.9) |
3 (0.04) |
41 (0.6) |
4980 (72.2) |
| 2 | 128 (1.9) |
209 (3.0) |
412 (6.0) |
8 (0.1) |
14 (0.2) |
771 (11.2) |
| 3 | 101 (1.5) |
44 (0.6) |
785 (11.4) |
22 (0.3) |
25 (0.4) |
977 (14.2) |
| 4 | 2 (0.03) |
0 (0) |
97 (1.4) |
27 (0.4) |
16 (0.2) |
142 (2.1) |
| 5 | 0 (0) |
0 (0) |
3 (0.04) |
3 (0.04) |
26 (0.4) |
32 (0.5) |
|
| ||||||
| Total | 3946 (57.2) |
1070 (15.5) |
1701 (24.7) |
63 (0.9) |
122 (1.8) |
6902 |
Presence and center involvement of DME was analyzed in 6649 eyes (3367 participants, Table 3). DME was absent by both assessments in 6352 eyes (95.5%) but elsewhere agreement was poor and kappa was 0.35 (95% CI 0.30-0.40).
Table 3.
Macular Edema Severity comparison of the photographic assessment and ACCORD clinical scale: N (%)
| ACCORD Clinical Scale |
Photographic DME Severity Level |
Total | ||
|---|---|---|---|---|
| Not Present | Present without center involvement |
Present with center involvement |
||
| 1 | 6352 (95.5) |
14 (0.2) |
110 (1.7) |
6476 (97.4) |
| 2 | 52 (0.8) |
5 (0.1) |
46 (0.7) |
103 (1.6) |
| 3 | 25 (0.4) |
4 (0.1) |
41 (0.6) |
70 (1.1) |
|
| ||||
| Total | 6429 (96.7) |
23 (0.2) |
197 (3.0) |
6649 |
Table 4 presents the sensitivities of the clinical grading to identify presence of selected levels of retinopathy, which ranged from 0.21 (95% CI 0.13 – 0.30) for any retinopathy to 0.57 (95%CI 0.55-0.59) for PDR. The sensitivity of the clinical scale to identify presence of any DME was 0.44 (95% CI 0.36 – 0.51), specificity was 0.99 (95% CI 0.99-0.99), and AUC was 0.72 (95% CI-0.68-0.76). Sensitivity for center-involved DME was 0.21 (95% CI 0.14-0.27), specificity was 0.99 (95 % CI 0.99-0.99), and AUC was 0.73 (95% CI 0.69 – 0.77).
Table 4.
Sensitivity and specificity of clinical scale in ACCORD, as compared to ETDRS
| Sensitivity | Specificity | AUC | ||||
|---|---|---|---|---|---|---|
|
|
||||||
| Outcome | estimate | 95% CI | estimate | 95% CI | Estimate | 95% CI |
| Diabetic Retinopathy | ||||||
| Any Retinopathy (level 1 vs. 2-5) |
0.213 | 0.128 – 0.298 | 0.999 | 0.998 – 1.000 | 0.75 | 0.737 – 0.763 |
| Moderate NPDR or higher (level 1-2 vs. 3-5) |
0.532 | 0.505 – 0.560 | 0.971 | 0.966 – 0.976 | 0.838 | 0.825 – 0.852 |
| Moderately Severe NPDR or higher (level 1-3 vs. 4-5) |
0.389 | 0.300 – 0.478 | 0.985 | 0.981 – 0.988 | 0.775 | 0.732 – 0.817 |
| PDR (level 1-4 vs. 5) |
0.572 | .0.550 – 0.594 | 0.941 | 0.933 – 0.950 | 0.722 | 0.667 – 0.776 |
| Macular Edema | ||||||
| Present | 0.44 | 0.36 – 0.51 | 0.99 | 0.985 – 0.991 | 0.72 | 0.68 – 0.76 |
| Center Involved | 0.21 | 0.14 – 0.27 | 0.996 | 0.994 – 0.997 | 0.73 | 0.69 – 0.77 |
NPDR – non proliferative diabetic retinopathy
PDR – proliferative diabetic retinopathy
AUC – area under curve
FIND
Included in analysis for DR severity agreement were 3638 eyes (1893 participants) that had been assessed at the clinics and graded at the reading center (Table 5). DR was absent by both assessments in 1104 eyes (30.4%). There was exact agreement in 2906 eyes (74%) and agreement within one step in 3361 eyes (92.4%); kappa 0.65 (95% CI 0.63-0.67), weighted kappa 0.80 (95% CI 0.79-0.81), indicating substantial agreement. In comparison with ACCORD by photographic grading the proportion of eyes with microaneurysms only was smaller (8% vs. 16%) and the proportion of eyes with PDR much larger (29% vs. 2%). In addition, in both of these categories clinical-photographic agreement appeared to be somewhat greater in FIND than in ACCORD.
Table 5.
Diabetic Retinopathy comparison of ETDRS condensed scale and FIND clinical scale: N (%)
| FIND Clinical Scale |
ETDRS Severity Level |
Total | ||||
|---|---|---|---|---|---|---|
| 10-15 | 20 | 35-43,47A | 47B-53 | 60+ | ||
| 1 | 1104 (30.4) |
160 (4.4) |
94 (2.6) |
4 (0.1) |
18 (0.5) |
1380 (37.9) |
| 2 | 63 (1.7) |
111 (3.1) |
194 (5.3) |
7 (0.2) |
22 (0.6) |
397 (10.9) |
| 3 | 27 (0.7) |
18 (0.5) |
539 (14.8) |
47 (1.3) |
72 (2.0) |
703 (19.3) |
| 4 | 0 (0) |
0 (0) |
98 (2.7) |
58 (1.6) |
72 (2.0) |
228 (6.3) |
| 5 | 7 (0.2) |
1 (0.0) |
25 (0.7) |
13 (0.4) |
884 (24.3) |
930 (25.6) |
|
| ||||||
| Total | 1201 (33.0) |
290 (8.0) |
950 (26.1) |
129 (3.6) |
1068 (29.4) |
3638 |
Presence and center involvement of DME was analyzed in 3366 eyes (1790 participants), shown in Table 6. Agreement between the two methods was moderate with simple kappa of 0.41 (95% CI 0.36-0.45) and weighted kappa of 0.43 (95% CI 0.38-0.48).
Table 6.
Macular Edema comparison of ETDRS and FIND clinical scale: N (%)
| FIND Clinical Scale |
Photographic DME Severity Level |
Total | ||
|---|---|---|---|---|
| Not Present | Present without center involvement |
Present without center involvement |
||
| 1 | 2840 (84.4) |
24 (0.7) |
144 (4.3) |
3008 (89.4) |
| 2 | 23 (0.7) |
9 (0.3) |
37 (1.1) |
69 (2.1) |
| 3 | 143 (4.3) |
17 (0.5) |
129 (3.8) |
289 (8.6) |
|
| ||||
| Total | 3006 (89.3) |
50 (1.5) |
310 (9.2) |
3366 |
Table 7 shows that the sensitivity of the clinical scale to identify selected levels of DR ranged from 0.89 (95% CI 0.87 – 0.90) for any retinopathy to 0.83 (95%CI 0.80-86) for PDR, with specificity ranging from 0.92 (95% CI 0.90-0.94) to 0.98 (95 % CI 0.98-0.99). Center-involved DME was found in 310 eyes according to the reading center and 289 eyes by clinicians – much higher than that observed in ACCORD. Sensitivity of the clinical scale to identify presence of any DME was 0.53 (95% CI 0.47 – 0.59), with specificity of 0.94 (95% CI 0.93-0.95), and AUC was 0.75 (95% CI-0.72-0.78). Sensitivity to identify center-involved DME was 0.41 (95 % CI 0.35-0.48) with specificity of 0.95 (95 % CI 0.94-0.96), and AUC was 0.74 (95% CI 0.71 – 0.78).
Table 7.
Sensitivity and specificity of clinical scale in FIND, as compared to ETDRS
| Outcome | Sensitivity | Specificity | AUC | |||
|---|---|---|---|---|---|---|
|
| ||||||
| estimate | 95% CI | estimate | 95% CI | Estimate | 95% CI | |
| Diabetic Retinopathy | ||||||
| Any Retinopathy (level 1 vs. 2-5) |
0.887 | 0.871 – 0.902 |
0.919 | 0.901 – 0.937 |
0.924 | 0.913 – 0.935 |
| Moderate NPDR or higher (level 1-2 vs. 3-5) |
0.842 | 0.823 – 0.861 |
0.964 | 0.954 – 0.975 |
0.952 | 0.943 – 0.961 |
| Moderately Severe NPDR or higher (level 1-3 vs. 4-5) |
0.858 | 0.833 – 0.883 |
0.946 | 0.935 – 0.957 |
0.953 | 0.943 – 0.962 |
| PDR (level 1-4 vs. 5) |
0.828 | 0.800 – 0.856 |
0.982 | 0.976 – 0.989 |
0.952 | 0.941 – 0.962 |
| Macular Edema | ||||||
| Present | 0.53 | 0.47 – 0.59 |
0.94 | 0.93 – 0.95 |
0.75 | 0.72 – 0.78 |
| Center Involved | 0.41 | 0.35 – 0.48 |
0.95 | 0.94 – 0.96 |
0.74 | 0.71 – 0.78 |
NPDR – non proliferative diabetic retinopathy
PDR – proliferative diabetic retinopathy
AUC – area under curve
Discussion
This report compares levels of agreement on DR severity between clinical ophthalmoscopic examinations and reading center grading of 7-field color stereoscopic fundus photographs in two large clinical studies of diabetic retinopathy that differed substantially in retinopathy severity. By photographic grading proportions of eyes with no retinopathy, microaneurysms only and PDR were, respectively, 57%, 20%, and 2% in ACCORD and 33%, 8%, 29% in FIND. Kappa statistics for agreement between clinical and photographic assessments were 0.42 (95% CI 0.40, 0.44) and 0.65 (95% CI 0.63, 0.67), respectively, and sensitivities for detecting mild or worse retinopathy were 0.532 and 0.842 respectively. Differences in clinical-photographic agreement on DR severity appeared to be related largely to differing distributions of severity level. The greater ease of detection of occasional microaneurysms in good quality photographs than by conventional ophthalmoscopy (as opposed to slit lamp biomicroscopy, which was not required in our studies) has been well documented.12-13 The equal ability of photos and ophthalmoscopy to demonstrate presence of the easily identifiable scars of scatter (panretinal) photocoagulation (provided they are not limited to areas of retina not included in the photographs) is also well recognized from clinical experience. In FIND 83.1% of the 1068 eyes classified by photos as PDR had such scars compared to 10.7% of the 122 eyes so classified in ACCORD. As noted for other lesions of DR, subtle early neovascularization is often difficult to identify by examination, especially in the retinal periphery examined by indirect ophthalmoscopy, which minifies the image compared to slit lamp or direct ophthalmoscope examination of the posterior pole.
Our results for DR severity are similar to those from two recent reports in which assessments by large groups of clinicians who used their usual ophthalmoscopic methods and by photograph grading were compared. By photographic grading proportions of eyes with no retinopathy, microaneurysms only and PDR were, respectively, 32%,18% and 5% in the Veterans Affairs Diabetes Trial14 and 0%, 0% and 32%, respectively, in a Phase 2 study of intravitreal bevacizumab for DME conducted by the Diabetic Retinopathy Clinical Research Network.15 In the former study exact agreement between clinical and photographic assessments was 43% and unweighted kappa was 0.21 (95% CI 0.15, 0.27) while in the latter study these values were 75% and 0.55 (95% CI 0.41, 0.68). The sensitivity for ophthalmoscopy compared to grading of fundus photographs for the detection of any retinopathy in the Veterans Affairs Diabetes Trial14 was 51% and specificity was 91%. For proliferative diabetic retinopathy (PDR), sensitivity was 61% and specificity 98%. As in the studies reported herein, differences in distributions of DR severity appeared to explain, at least in part, these differences in degree of agreement.
Macular edema was infrequent in both of our studies and, when present, usually involved the center of the macula. By photographic grading macular edema without and with involvement of the center of the macula was present in 0.2% and 3.0% of eyes in ACCORD vs. 1.5% and 9.2%, respectively, in FIND. Kappa statistics were 0.35 (95% CI 0.30, 0.40) and 0.41 (90% CI 0.36, 0.45), respectively. Prior studies comparing photographic and ophthalmoscopic assessments of presence and severity of DME separately from assessments of DR, rather than as components used in defining a level of severity warranting referral for consideration of treatment, have generally used slitlamp biomicroscopy as the ophthalmoscopic method and, as might be expected because both high magnification and stereopsis similar to that obtained in stereoscopic photographs are provided by this method, have reported somewhat better agreement than that observed in our studies (kappa values 0.44-0.61).16-18 Although both slit lamp biomicroscopy and photography continue to be useful in assessing various clinical features of DME, optical coherence tomography, which provides quantitative measures of retinal thickness, is now the principal method for assessing severity of DME and its changes over time.
The implication of these findings for clinical care of patients is speculative. The increased sensitivity of lesion detection, particularly at the lower and highest ends of the scale, by 7 field fundus photography must be balanced against the availability of equipment, cost, and available technical expertise to acquire the images. While photography may allow earlier detection of threshold PDR, for instance, more frequent examinations as is customary for eyes with severe NPDR are likely to result in detection of neovascularization at an early stage as well. It is important to note that photographic diabetic retinopathy screening programs generally are directed toward detection of thresholds for referral for care, with the indications for referral varying with the needs of the population, and the full ETDRS scale is not employed. Additionally, screening programs usually employ a much simpler photographic protocol for costs and efficiency which covers a smaller area of the retinal periphery, likely decreasing the detection of DR lesions, Therefore, results of this study cannot be used to compare clinical evaluation to telemedicine screening protocols.
In conclusion, our findings in two large studies, including a wide spectrum of diabetic retinopathy severity, with many participating clinicians support prior conclusions that the moderate degree of agreement on severity of DR between clinical examinations and grading of 7-field stereoscopic photographs supports the use of this clinical information for defining broad severity categories. Clinician grading tends to be less sensitive for early DR and for early PDR. However, for documenting small to moderate changes in DR severity over time as is required in clinical research comparing either different modalities of treatment or natural history, detailed masked evaluation of fundus photographs is likely to yield better quality of data.
Acknowledgments
Financial support: National Eye Institute Intramural Research Program Sharon Adler: GCRC grant M01-RR00425. FIND grant R01-DK0069844 ACCORD grant N01-HC-95178
Appendix A
TABLE 1. International Clinical Diabetic Retinopathy Disease Severity Scale.
| Proposed Disease Severity Level | Findings Observable upon Dilated Ophthalmoscopy |
|---|---|
| No apparent retinopathy | No abnormalities |
| Mild NPDR | Microaneurysms only |
| Moderate NPDR | More than just microaneurysms but less than severe NPDR |
| Severe NPDR | Any of the following and no signs of proliferative retinopathy:
|
| PDR | One or both of the following:
|
IRMA = intraretinal microvascular abnormalities; NPDR = nonproliferative diabetic retinopathy; PDR = proliferative diabetic retinopathy
Reproduced with permission from Wilkinson CP, Ferris FL III, Klein RE, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology 2003;110:1679.
TABLE 2. International Clinical Diabetic Macular Edema Disease Severity Scale.
| Proposed Disease Severity Level | Findings Observable upon Dilated Ophthalmoscopy |
|---|---|
| Diabetic macular edema apparently absent | No apparent retinal thickening or hard exudates in posterior pole |
| Diabetic macular edema apparently present | Some apparent retinal thickening or hard exudates in posterior pole |
| If diabetic macular edema is present, it can be categorized as follows: | |
| Proposed Disease Severity Level | Findings Observable upon Dilated Ophthalmoscopy* |
| Diabetic macular edema present |
|
Hard exudates are a sign of current or previous macular edema. Diabetic macular edema is defined as retinal thickening; this requires a three-dimensional assessment that is best performed by dilated examination using slit-lamp biomicroscopy and/or stereoscopic fundus photography.
Reproduced with permission from Wilkinson CP, Ferris FL III, Klein RE, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology 2003;110:1680.
Footnotes
Conflict of interest: None of the authors have any conflict of interest pertinent to this manuscript.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Knowler WC, Coresh J, Elston RC, Freedman BI, Iyengar SK, Kimmel PL, Olson JM, Plaetke R, Sedor JR, Seldin MF, Family Investigation of Nephropathy and Diabetes Research Group The Family Investigation of Nephropathy and Diabetes (FIND): design and methods. J Diabetes Complications. 2005 Jan-Feb;19(1):1–9. doi: 10.1016/j.jdiacomp.2003.12.007. [DOI] [PubMed] [Google Scholar]
- 2.Chew EY, Ambrosius WT, Howard LT, Greven CM, Johnson S, Danis RP, Davis MD, Genuth S, Domanski M, ACCORD Study Group Rationale, design, and methods of the Action to Control Cardiovascular Risk in Diabetes Eye Study (ACCORD-EYE) Am J Cardiol. 2007 Jun 18;99(12A):103i–111i. doi: 10.1016/j.amjcard.2007.03.028. [DOI] [PubMed] [Google Scholar]
- 3.ACCORD Study Group; ACCORD Eye Study Group. Chew EY, Ambrosius WT, Davis MD, Danis RP, Gangaputra S, Greven CM, Hubbard L, Esser BA, Lovato JF, Perdue LH, Goff DC, Jr, Cushman WC, Ginsberg HN, Elam MB, Genuth S, Gerstein HC, Schubart U, Fine LJ. Effects of medical therapies on retinopathy progression in type 2 diabetes. N Engl J Med. 2010 Jul 15;363(3):233–44. doi: 10.1056/NEJMoa1001288. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wilkinson CP, Ferris FL, 3rd, Klein RE, Lee PP, Agardh CD, Davis M, Dills D, Kampik A, Pararajasegaram R, Verdaguer JT, Global Diabetic Retinopathy Project Group Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology. 2003 Sep;110(9):1677–82. doi: 10.1016/S0161-6420(03)00475-5. [DOI] [PubMed] [Google Scholar]
- 5.Early Treatment Diabetic Retinopathy Study Research Group Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10. Ophthalmology. 1991 May;98(5 Suppl):786–806. [PubMed] [Google Scholar]
- 6.Gangnon RE, Davis MD, Hubbard LD, Aiello LM, Chew EY, Ferris FL, 3rd, Fisher MR, Early Treatment Diabetic Retinopathy Study Research Group A severity scale for diabetic macular edema developed from ETDRS data. Invest Ophthalmol Vis Sci. 2008 Nov;49(11):5041–7. doi: 10.1167/iovs.08-2231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977 Mar;33(1):159–74. [PubMed] [Google Scholar]
- 8.Zhou XH, Obuchowski NA, McClish DK. Statistical Methods in Diagnostic Medicine. Wiley; New York: 2002. [Google Scholar]
- 9.Obuchowski NA. Nonparametric Analysis of Clustered ROC Curve Data. Biometrics. 1997;53(2):567–578. [PubMed] [Google Scholar]
- 10.Efron B, Tibshirani RJ. An Introduction to the Bootstrap. Chapman & Hall; New York: 1993. [Google Scholar]
- 11.Mossman D. Resampling techniques in the analysis of non-binormal ROC Data. Medical Decision Making. 1995;15:358–366. doi: 10.1177/0272989X9501500406. [DOI] [PubMed] [Google Scholar]
- 12.Kinyoun JL, Martin DC, Fujimoto WY, Leonetti DL. Ophthalmoscopy versus fundus photographs for detecting and grading diabetic retinopathy. Invest Ophthalmol Vis Sci. 1992 May;33(6):1888–93. [PubMed] [Google Scholar]
- 13.Moss SE, Klein R, Kessler SD, Richie KA. Comparison between ophthalmoscopy and fundus photography in determining severity of diabetic retinopathy. Ophthalmology. 1985;92:62–67. doi: 10.1016/s0161-6420(85)34082-4. [DOI] [PubMed] [Google Scholar]
- 14.Emanuele N, Klein R, Moritz T, Davis MD, Glander K, Anderson R, Reda D, Duckworth W, Abraira C, VADT Study Group Comparison of dilated fundus examinations with seven-field stereo fundus photographs in the Veterans Affairs Diabetes Trial. J Diabetes Complications. 2009 Sep-Oct;23(5):323–9. doi: 10.1016/j.jdiacomp.2008.02.010. [DOI] [PubMed] [Google Scholar]
- 15.Scott IU, Bressler NM, Bressler SB, Browning DJ, Chan CK, Danis RP, Davis MD, Kollman C, Qin H, Diabetic Retinopathy Clinical Research Network Study Group Agreement between clinician and reading center gradings of diabetic retinopathy severity level at baseline in a phase 2 study of intravitreal bevacizumab for diabetic macular edema. Retina. 2008 Jan;28(1):36–40. doi: 10.1097/IAE.0b013e31815e9385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Broadbent DM, Scott JA, Vora JP, Harding SP. Prevalence of diabetic eye disease in an inner city population: the Liverpool Diabetic Eye Study. Eye (Lond) 1999 Apr;13(Pt 2):160–5. doi: 10.1038/eye.1999.43. [DOI] [PubMed] [Google Scholar]
- 17.Davis MD, Bressler SB, Aiello LP, Bressler NM, Browning DJ, Flaxel CJ, Fong DS, Foster WJ, Glassman AR, Hartnett ME, Kollman C, Li HK, Qin H, Scott IU, Diabetic Retinopathy Clinical Research Network Study Group Comparison of time-domain OCT and fundus photographic assessments of retinal thickening in eyes with diabetic macular edema. Invest Ophthalmol Vis Sci. 2008 May;49(5):1745–52. doi: 10.1167/iovs.07-1257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kinyoun J, Barton F, Fisher M, Hubbard L, Aiello L, Ferris F, 3rd, The ETDRS Research Group Detection of diabetic macular edema. Ophthalmoscopy versus photography--Early Treatment Diabetic Retinopathy Study Report Number 5. Ophthalmology. 1989 Jun;96(6):746–50. doi: 10.1016/s0161-6420(89)32814-4. discussion 750-1. [DOI] [PubMed] [Google Scholar]