DM affects all structural layers of the cornea, including the corneal epithelium, corneal nerves, tear film, and corneal endothelium.17 Clinically observed corneal diabetic complications include recurrent epithelial erosion, delayed wound repair, neurotrophic ulcers, and decreased corneal sensitivity. Morphological and functional changes also occur in the corneal endothelium, and diabetes carries an increased risk of endothelial complications after intraocular surgery. Although the pathophysiological mechanisms underlying diabetic damage of the corneal endothelium are not yet clear, osmotic damage due to excessive sorbitol accumulation, oxidative damage due to glycation end-product accumulation, and the reduction of corneal endothelial cell adhesion due to direct adhesive protein modifications of Descemet's membrane by AGEs has been proposed.18–20
Several studies have evaluated the endothelium in patients with type 2 diabetes and normal subjects. Most of them agree that the corneal endothelium of diabetic patients changes morphologically.10,18,21−23 Su et al.18 reported that diabetics had thicker corneas and decreased ECD than persons without diabetes. Roszkowska et al.23 reported a significant decrease in cell density in type 2 diabetic patients. Other studies also showed a significant decrease in ECD and hexagonality and an increase in CV in type 2 DM patients.12,14 These results are consistent with the findings of this study. However, several studies found no difference between the ECDs of diabetics and normal subjects.24,25 The effect of DM on the corneal endothelium remains controversial.
In the present study, it was detected that ECD, CV, hexagonality, and CCT showed significant differences between long durations of diabetes (≥ 10 years) and control in all age (Fig. 1) However, only CV and CCT showed significant differences between short durations of diabetes (< 10 years) and control. In addition, high HbA1c (≥ 7%) patients showed differences in ECD, CV, and CCT, and low HbA1c patients (< 7%) showed only differences in CV and CCT. Therefore, it can be assumed that the duration of DM or poor blood glucose control had a greater effect on corneal endothelial change. Previous studies also reported that changes in ECD, CV, and CCT are associated with the duration of DM, and patients who had had DM for more than 10 years had more significant changes.14,26 Corneal endothelial cell change in diabetes is a chronic microvascular complication, and complications occur more frequently when DM is prolonged and patients are chronically exposed to hyperglycemic conditions.27 On the other hand, HbA1c reflects short-term DM control, and it is reported that CCT may change within a short period depending on the HbA1c level.14 In our study, long-standing diabetic patients showed more significant differences in overall corneal parameters between DM and control groups than in high Hb1Ac patients. This suggests that the duration of diabetes is more strongly associated with chronic corneal endothelial damage than HbA1c. Therefore, the duration of diabetes is recommended over HbA1c for predicting corneal endothelial cell damage.
According to the results of the correlation analysis, age was the most important factor, and it correlated with all corneal parameters, and DM duration correlated with ECD. On the other hand, HbA1c did not show significant correlations with the corneal parameters; This is because HbA1c reflects short-term diabetic control, and it does not reflect chronic corneal endothelial cell damage.
When the groups were stratified by age, ECD, hexagonality, and CCT decreased whereas CV tended to increase with aging, which was consistent with the findings of previous studies.28 A few papers have compared diabetic patients by dividing them into age groups. In 2012, Rachapalle et al.10 analyzed the data of 1191 DM patients and 121 non-diabetic patients stratified by age. They reported that CV and hexagonality were greater in diabetic patients than in controls aged between 50 and 69 years, and CCT was thicker in diabetic patients than in controls aged between 60 and 69 years. In this study, most corneal parameters showed significant differences in the DM and control groups after the age of 60 years. The reason for the pronounced difference between the DM and control groups as the age increased was not revealed. This is presumed to be because the DM duration is more likely to become longer as the elderly age, and the older person is vulnerable to corneal endothelial cell damage.
When the DM duration and HbA1c of the DM and control groups stratified by age were analyzed, CCT showed a significant difference after 50 years in patients with long-standing diabetes (≥ 10 years) and ECD, CV, and hexagonality showed significant differences after 60 years. Similar trends were observed in the high HbA1c group (≥ 7%); CCT showed a difference at the age of 40 years, whereas ECD, CV, and hexagonality showed significant differences at the age of 60. Although the overall corneal parameters were more affected by DM duration than HbA1c, but HbA1c can affect CCT for a short period of time. Therefore, it is assumed that there was a difference between the two groups even at age 40.
The limitation of this study is that the results were cross-sectional, and prospective changes were not confirmed. In future studies, it is necessary to observe consecutive changes of corneal parameters over time. Despite these limitations, the relatively large sample of more than 1000 patients subdivided into age groups in this study is an advantage.
In conclusion, the corneas of patients with type 2 DM showed an increase in thickness and CV and reductions in ECD and hexagonality compared with the healthy controls. These changes were more in older type 2 DM patients (≥ 60 years) with a long duration of disease (≥ 10 years) and high HbA1c (≥ 7%). In the correlation analysis, age and the duration of DM affected ECD. These results are from a large population-based sample, and which support the theory that type 2 DM can affect corneal endothelial cells. Therefore, it is necessary to evaluate not only the retina but also the corneal endothelial cells regularly during follow-up for type 2 DM patients.