There were two main findings in our study. First, new-onset diabetes is associated with higher RCF compared to the RCF recorded in the patients with a longer duration of the disease. Second, structural arterial remodeling assessed as an increased WLR and WT is evident in the patients with a diabetes duration of 10 years or longer. These results could not be explained age, HBA1c or presence of hypertension. Neither RCF nor the structure of microcirculation was associated with any of the following: patient age, HbA1c concentration, and SBP. The only variable differentiating RCF and WLR was diabetes duration. These findings indicate that structural retinal changes observed in long-term type 1 diabetes are preceded by functional changes during early phase of the disease
Hyperglycemia has a profound impact on the cardiovascular system(15). A meta-analysis of 102 prospective studies prepared by the Emerging Risk Factors Collaboration showed only modestly a higher risk of coronary heart disease among people with no history of diabetes and a fasting glucose level below 7 mmol/L but a substantially higher risk in those with a fasting glucose of at least 7 mmol/L (hazard rate 1.17 vs 1.78; 1.61 vs 2.36 in people with diabetes and the same fasting glucose values)(2).
Our knowledge regarding cardiovascular risk in T1DM is mainly based on the results of the Diabetes Control and Complications Trial (DCCT). The DCCT was a clinical trial conducted between 1983 and 1993 to check the hypothesis that the complications of T1DM could be delayed or prevented with the intensification of glucose control through intensive treatment. In total, 1441 patients were randomized either to the intensive treatment (INT) group or the conventional therapy (CON) group. The mean HbA1C at the end of the study was 7.2% (0.9%) for the INT group and 9.1% (1.3%) for the CON group. After the termination of the DCCT, the Epidemiology of Diabetes Interventions and Complications (EDIC) study was performed as a follow-up(16). The intima-media complex was measured at years 1 and 6 of the EDIC. Even though there were no significant differences between the diabetic and nondiabetic patients at year 1 of follow-up, after 6 years, the intima-media thickness of both the common and internal carotid arteries was significantly higher in the diabetic cohort, and the progression was more intensive in the CON group(17). The patients who were assigned to the INT group also had a much lower rate of cardiovascular events (i.e., nonfatal myocardial infarction, stroke, or death from cardiovascular disease) compared to the CON group (0.38 vs 0.80, respectively; p = 0.007) even though the risk was higher compared to the nondiabetic patients. No patients had hypertension or hypercholesterolemia at the beginning of the DCCT. The patients who experienced cardiovascular events during the EDIC were older, had a longer duration of diabetes, had higher HbA1C and total and LDL cholesterol levels, a history of smoking, and a family history of myocardial infarction(18). Despite the cardiovascular risk, the presence of microangiopathic complications was also lower in the INT group(19). The follow-up period was similar in length to the duration of diabetes in group C in our study.
In our study, retinal flow was highest among the patients with new-onset diabetes and lower in the groups that had a longer duration of diabetes, among patients with > 1 diabetes duration it was even lower then comparing to nondiabetic participants. These results could not be explained by the age of the patients as the patients enrolled in the study were within a narrow age interval.
Additionally, the patients who had recently been diagnosed with diabetes had higher glucose levels and significantly higher HbA1C concentrations compared to glucose levels. This phenomenon may be similar to glomerular hyperfiltration, which is observed among patients with diabetes and even prediabetes(20). The mechanism is not completely understood but seems to be related to tubular factors and glomerular hemodynamics. This condition is present in about 50% of adolescents and young adults with diabetes. Hyperfiltration is thought to be a very important factor in diabetic kidney disease, likely as a reflection of the increased intraglomerular pressure resulting from structural changes(21). A significant decline in GFR is observed following hyperfiltration, which appears to present as normalized kidney function. Nonetheless, it is merely an interim state that further exacerbates the kidney damage, and this kidney damage is even greater compared to patients who do not present with hyperfiltration(5). This observation could be similar to our results describing retinal circulation. Chronic enzymatic glycation decreases endothelium-derived NO and increases not only the production of vasoconstrictor prostanoids and endothelin, but also the expression of cyklooxygenase-2. Furthermore, it impairs autonomic nervous function and alters the vascular smooth muscle synthesis of collagen. All the above mechanisms may contribute to vascular dysfunction(3). Our study provides novel insights into the relationship between T1DM and microcirculation impairment in patients without other cardiovascular risk factors such as hypertension, obesity, and aging.
Our study showed that the WLR was significantly higher among the patients with at least a 10-year course of diabetes compared to those with a shorter history of the disease and the healthy controls. Numerous studies have described the relationship between changes in the small arteries and increased cardiovascular risk leading to organ damage(22–24). Studies on hypertension have shown a significant correlation between BP and the WLR of retinal arterioles. Ritt et al. examined patients with never-treated essential hypertension and normotensive patients using SLDF(25). The arteriolar WLR in the retina among the patients with hypertension was 0.36, and among the normotensive volunteers, it was 0.28 (the standard deviation was ± 0.1 in both groups, p = 0.028). There were also no significant differences in RCF (334 ± 84 AU and 340 ± 57 AU, respectively; p = 0.739). A possible cause of the increased WLR in hypertensive patients may be hypertrophy of the smooth cell layer and/or remodeling. Additionally, apoptosis, inflammation, and fibrotic processes could contribute to arterial structure, with an abnormal balance between growth and apoptosis in hypertension(26). All these changes seem to be an adaptive response necessary to maintain an optimum level of wall tension. Despite the initial physiological adaptive response in chronic exposure to elevated BP levels, this leads to a maladaptive response and vascular complications of hypertension(25). Similar results were observed by Salvetti et al., who also examined normotensive patients in comparison to patients with treated or untreated hypertension(27). The WLR measured in the retinal arterioles in the patients without hypertension was 0.23 ± 0.13, while in the hypertensive individuals who were or were not receiving antihypertensive treatment, the WLR was 0.29 ± 0.18 and 0.28 ± 0.18, respectively. There was no significant difference between the hypertensive patients, but a significant difference was present when comparing the hypertensive patients and the normotensive individuals. An increased WLR of the retinal arterioles has also been found among patients diagnosed with primary aldosteronism or pheochromocytoma or after cerebrovascular events(9, 28, 29). Among the patients enrolled in our study, BP was in the normotensive range. Nevertheless, the patients with a longer history of diabetes had significantly higher SBP, DBP, and MAP compared to the patients with new-onset diabetes, and SBP, DBP, and MAP in these patients tended to be higher than in the group of non-diabetic patients.
Strengths and limitations
We presented well-selected, relatively large groups of patients who did not differ in age or BMI. There was only a slight difference in smoking prevalence and antihypertensive treatment between the groups, with only a few participants either smoking or receiving antihypertensive medications (i.e., a low dose of ACE inhibitors to prevent albuminuria). SLDF is one of the most accurate methods for evaluating retinal microcirculation. Nevertheless, it is a somewhat subjective method operator-dependent. Therefore, we made every effort to reduce potential bias by carefully preparing and implementing a strict protocol, which has previously been used in a number of SLDF-based studies. We took at least 3 images of the retinal arterioles of each participant, and the data were averaged to reduce the possibility of measurement errors. All the images were blindly cross-checked by an investigator with long-standing expertise in the SLDF technique to identify images of unsatisfactory quality, which were subsequently excluded from further analysis. Potential limitations include lack of ambulatory blood pressure monitoring. Furthermore, our study was cross-sectional, and prospective observations are needed to gain more insights into the mechanisms underlying retinal microvascular abnormalities in type 1 diabetes.