Clinical Phenotypes of Diabetic Retinopathy

It is apparent, from the data available from a variety of large longitudinal studies, that the evolution and progression of diabetic retinopathy vary between the two types of diabetes involved and between different patients even when belonging to the same type of diabetes, and does not necessarily progress to clinically significant macular edema (CSME) or proliferative retinopathy leading to vision loss.

NATURAL HISTORY

Diabetic retinopathy shows initially minimal fundus abnormalities and progresses over time to more and more advanced microvascular changes. The main alterations occurring in the diabetic retina are: breakdown of the blood–retinal barrier (BRB), evidenced by abnormal vascular leakage and capillary closure leading to progressive tissue ischemia. These two main alterations lead, as they progress, to the two major complications of diabetic retinal disease which are associated with vision loss: CSME and proliferative retinopathy.

The retinal changes that result from diabetes before the development of the two main complications referred above are conventionally described under the name of nonproliferative diabetic retinopathy (NPDR).

The initial stages of NPDR are, therefore, characterized by the presence of microaneurysms (MA), hemorrhages, hard exudates or cotton-wool spots, indirect signs of vascular hyperpermeability, and capillary closure.

These are the alterations that dominate the initial stages of NPDR, and it is crucial to analyze their development and progression, in order to clarify their relative importance in the progression of diabetic retinopathy [2]. They are not present in every patient in the same way nor at the same rate.

It is fundamental to realize that the course and rates of progression of the retinopathy vary between patients. MA, for example, may come and go. Once you get an MA, you do not necessarily continue to have that MA. MA may disappear due to vessel closure, which is an indication of worsening of the retinopathy because of progressive vascular closure [3]. Hemorrhages will obviously come and go as the body heals them. Frequently, there is apparent clinical improvement with fewer lesions visible, but in reality, it masks worsening of the disease.

A prominent feature of diabetic retinopathy, focal edema, can spontaneously resolve itself. Indeed, it is resolved in approximately a third of patients over a period of 6 months, without any intervention [4].

The initial pathological changes occurring in the diabetic retina are characteristically located in the small retinal vessels of the posterior pole of the retina, i.e., in the macular area. The structural changes in the small vessels include endothelial cell and pericyte damage and thickening of basement membrane [2, 5].

Retinal vascular endothelium is a fundamental part of the BRB, which has many similarities with the blood–brain barrier. It functions as a selective barrier which has shown to be altered in experimental and human diabetes [6]. It is altered in the early stages of diabetic retinal disease.

Pericyte damage has also been reported as one of the earliest findings in diabetic retinal disease since the introduction of retinal digest studies [7]. However, pericyte damage may be more prominent just because it is more readily detectable than endothelial cell damage, because the pericytes are encased in basement membrane and thus less

accessible to the clearing effect of blood flow, whereas dying endothelial cells slough off into the capillary lumen and are rapidly cleared by the blood stream.

The simplest paradigm that explains the initial retinal microvascular changes in diabetes, capillary hyperpermeability, and capillary closure is damage to the vascular endothelium. In the retina, endothelial cells are the site of the BRB, a specific blood–tissue barrier, and, as in all vessels, provide a nonthrombogenic surface for blood flow. Both these properties are compromised by diabetes from the initial stages of diabetic retinal disease.

In addition, diabetes also affects the neural and glial cells of the retina. Consequently, we have an initial pathological picture characterized by endothelial and pericyte alterations associated with basement membrane thickening and MA formation, together with retinal tissue changes.

These alterations when seen as a whole are characteristic for NPDR, particularly the alteration of the BRB, the pericyte damage, and the MA formation, but they also occur in a variety of retinal diseases unrelated to diabetes. There is clear site specificity, not disease specificity [2].

Which are then the features of the retinal circulation which are specific to the retina and may be responsible for the site specificity of diabetic retinopathy? They are the BRB and the autoregulation of retinal blood flow. Both serve the needs of the neuronal and glial cells of the retina.

An abnormality of the BRB, demonstrated both by vitreous fluorometry and fluorescein angiography, has repeatedly been demonstrated to be an early finding both in human and in experimental diabetes [6, 8, 9]. Loss of retinal blood flow autoregulation contributes to capillary closure that ultimately leads to retinal ischemia and to one of the two major complications of diabetic retinal disease, proliferative retinopathy, which causes the most tragic outcomes: vitreous hemorrhage, rubeosis iridis, retinal detachment, etc.

It is becoming apparent that at least three processes can contribute to retinal capillary occlusion and obliteration in diabetes: proinflammatory changes, microthrombosis, and apoptosis [10].

MA FORMATION AND DISAPPEARANCE RATES

MA and hemorrhages are the initial changes seen on ophthalmoscopic examination and fundus photography (FP). MA counting has been suggested as an appropriate marker of retinopathy progression [11, 12].

It must be realized that MA formation and disappearance are dynamic processes. During a 2-year follow-up of 24 type 1 diabetics with mild background diabetic retinopathy using fluorescein angiography, Hellstedt and Immonen [13] observed 395 new MA and the disappearance of 258 previously identified.

Generally, the disappearance of an MA is not a reversible process and indicates vessel closure and progressive vascular damage. Therefore, to assess progression of retinopathy, MA counting should take into account every newly developed MA identified in a new location.

We have developed software for MA counting in fundus-digitized images where the location of each MA is taken into account and registered [14]. In a follow-up study with repeated fundus images obtained at regular intervals, all MAs in the fundus were counted and added as they became visible in new locations in the retina. The results of MA counting using this method, in a 2-year follow-up study of a series of eyes with

Fig. 1. Microaneurysm analysis.

mild nonproliferative retinopathy in subjects with type 2 diabetes, maintaining a stable metabolic control during the period of the study, suggested that MA counting may be a good marker of disease progression in the initial stages of NPDR [15].

In order to improve the identification and counting of MA on color fundus images, the software included algorithms for eye movement compensation, color correction, and identification of each MA by its coordinates.

Using the software’s ability to identify each MA as a single entity, in a specific location with identifiable coordinates, the following parameters were assessed: cumulative number of MA, MA formation rate, and MA disappearance rate.

In a study involving 50 eyes/patients over a period of 2 years, with examinations performed every 6 months, using the traditional procedure, the total amount of MA detected at every visit remained stable. However, using the software to identify MA location, the cumulative number of MA rose from 115 at the first visit to 505 at the last visit, showing a marked increase in new MA. It is now obvious that there were many more new MA in the fundus in this 2-year time period than expected using data for each examination separately.

One of the advantages of the method used is the ability to count the number of real new MA appearing at every visit (MA formation rate) (Fig. 1). The rate of formation (MA/year) ranged from 0 to 22. The results showed that eyes in the same retinopathy stage from different patients show very different MA formation rates. Values for MA formation rate higher than 3 MAs/year correlated well with increased fluorescein leakage measured by vitreous fluorometry and capillary closure identified by a damaged foveal avascular zone (FAZ), demonstrating a direct correlation with faster retinopathy progression [16].

The MA disappearance rate ranged from 0 to 16 MAs/year. MA disappearance rates also varied quite markedly in eyes from different patients and showed similar correlations.

MA formation represents particularly well diabetic retinopathy because MAs are associated with localized proliferation of endothelial cells, loss of pericytes, and alterations of the capillary basement membrane, alterations that occur in the initial stages of diabetic retinal disease and have been considered to be directly involved in its pathophysiology [2, 17, 18].

MA closure and their disappearance are most probably due to thrombotic phenomena leading to subsequent rerouting of capillary blood flow and progressive remodeling of the retinal vasculature in diabetes [19]. These thrombotic changes are probably enhanced by changes in the red and white cells occurring as a result of diabetes.

MA counting on fundus photographs and MA counting on fluorescein angiography have been proposed as predictive indicators of progression of diabetic retinopathy [20, 21]. The software developed by our research group allows the identification of the exact location of each MA in successive fundus photographs performed in each eye. The identification of the exact location of an individual MA is considered particularly important because a new MA is considered to develop only once in a specific location, its disappearance being generally associated with capillary closure, leaving in its place mainly remnants of basement membrane [2, 18].

Our studies demonstrated a steady turnover of MAs in the diabetic retina, even in the initial stages of retinopathy. In fact, most MAs show a lifetime of less than 1 year, with new ones being formed and disappearing at rates which vary between different patients, confirming previous reports [22].

Most interestingly, however, is the observation that some patients show much higher rates of MA formation and disappearance, suggesting that they may represent specific phenotypes of diabetic retinopathy. These eyes showed also faster progression in other retinal lesions, with increased fluorescein leakage, i.e., alterations of BRB, and progression in capillary closure.

Using this new methodology, we have recently analyzed data from a group of 113 type 2 diabetic patients with mild-to-moderate NPDR, followed up for 2 years as controls in diabetic retinopathy clinical trials, and thereafter, by usual care at the same institution for a period of 10 years [23].

MA turnover from the initial 2 years was correlated with the occurrence of CSME during the following 8 years.

Patients were maintained under acceptable metabolic control during this period, and underwent ophthalmological examinations (including color fundus photography) every month.

At baseline, all patients showed mild-to-moderate retinopathy and were classified as levels 20 (MA only) or 35 (MA/hemorrhages and/or hard exudates) according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) grading scale.

At the end of the 10-year follow-up period, 17 out of the 113 patients developed CSME needing photocoagulation.

When counting the total number of MA over the first 2 years of the follow-up, a significant increase in the number of MA was found for the CSME eyes ( p = 0.002), while for the non-CSME eyes, the number of MA remained relatively constant ( p = 0.647).

Fig. 2. Boxplot for the microaneurysm formation rate for clinically significant macular edema (CSME) and non-CSME eyes, and number of eyes for the different values of the microaneurysm formation rate.

When computing the MA turnover for the same period of time, a higher MA turnover was found in the group of patients/eyes that developed CSME (higher MA formation and disappearance rates). Formation and disappearance rates of 9.2 ± 18.2 and 7.5 ± 16.6 MAs/year, respectively, were found for the eyes that developed CSME, while rates of 0.5 ± 1.2 and 0.5 ± 1.2 MAs/year were found for the non-CSME eyes ( p < 0.001).

A MA turnover of at least 2 MAs/year was found in 12 of the 17 eyes that developed CSME (70.6%), whereas this was only found in 8 of the 96 eyes that did not develop CSME during the 10-year follow-up period (8.3%) (Fig. 2).

This study shows that in the initial stages of diabetic retinopathy, higher MA counts and MA turnover obtained from color fundus photography are good indicators of retinopathy progression and development of CSME needing photocoagulation.

We also found that MA turnover is more reliable than simple MA counts and that there was much better agreement between graders when determining MA turnover than MA counts.

Recently, Sharp et al. [24] found that the MA turnover varied widely between eyes of the same retinopathy level. This is also consistent with our findings. MA turnover has been shown in this study to vary between patients that were classified with the same retinopathy level. Particularly relevant is the finding that the patients who have higher MA turnover values are the ones that go on to develop CSME within a period of 10 years and show a more rapid retinopathy progression, particularly in association with poor metabolic control demonstrated by higher HbA1C values.