Relationship between triglyceride glucose index, retinopathy and nephropathy in Type 2 diabetes

Abstract Aims To explore the relationship between TyG index, diabetic retinopathy (DR) and nephropathy. Methods This was a cross‐sectional observational study that examined 1413 subjects with type 2 diabetes (both known and newly diagnosed). Subjects underwent a detailed standard evaluation to detect diabetic retinopathy (fundus photography) and nephropathy (defined as urinary albumin excretion ≥ 30 mg/24 h). The TyG index was calculated as ln (fasting triglycerides (mg/dL) × fasting glucose (mg/dL)/2) and stratified into 4 quartiles (TyG‐Q). The baseline characteristics of the study population in the four TyG‐Q (Q1 (≤7.3) n = 349, Q2 (>7.3 to ≤ 7.5) n = 358, Q3 (>7.5 to ≤ 8.0) n = 354, and Q4 (>8.0) n = 352) were analysed. Variables associated with the presence of DR and nephropathy were assessed using a stepwise binary logistic regression analysis. Results The presence of DR was associated with higher TyG index (OR = 1.453, P =.001) and longer duration of diabetes (OR = 1.085, P < .001). The presence of nephropathy was associated with a higher TyG index (OR = 1.703, P < .001), greater age (OR = 1.031, P < .001), use of insulin (OR = 1.842, P = .033), higher systolic BP (OR = 1.015, P < .001), and the presence of DR (OR = 3.052, P < .001). Higher TyG‐Q correlated with the severity of DR (P = .024), presence of nephropathy (P = .001), age (P < .001) and diastolic blood pressure (P = .006). Conclusions A higher TyG index is associated with the presence of retinopathy and nephropathy in individuals with diabetes and could be used for monitoring metabolic status in clinical settings.


| INTRODUC TI ON
Vascular complications are the leading cause of mortality and morbidity in type 2 diabetes, affecting smaller and larger blood vessels. 1,2 Microvascular complications may include but not limited to diabetic retinopathy (DR), diabetic neuropathy and diabetic nephropathy, while macrovascular complications include diseases of the coronary, peripheral and cerebral arteries. An optimal control of serum glucose has been the mainstay in the prevention of microvascular and macrovascular complications of diabetes. Nevertheless, abnormal plasma triglycerides have been associated with metabolic disorders and cardiovascular disease due to their interaction with raised glucose levels in fat, muscle and beta cells of the pancreas. 3 We previously reported that the poor control of serum triglycerides is associated with progression to proliferative DR (PDR). 4 Likewise, studies have shown that higher triglycerides predict future risk of albuminuria progression in patients with diabetes. 5 Wiggin et al 6 found that in patients with mild to moderate diabetic neuropathy, elevated triglyceride levels correlated with progressive myelinated fibre density loss which was independent of disease duration, age and glycaemic control. In effect, a product of fasting triglyceride and glucose (called TyG index) has been proposed and utilized to identify metabolically unhealthy individuals. 7 The triglyceride glucose (TyG) index has been demonstrated as a novel marker for its association with insulin resistance and the risk of cardiovascular disease. 1,8,9 However, there is no study demonstrating the relationship between TyG index and microangiopathy in diabetes. Therefore, we aimed to study the relationship between TyG index, diabetic retinopathy (DR), diabetic neuropathy and diabetic nephropathy.

| ME THODS
Study participants were recruited from the Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetic Study (SN-DREAMS-1). The study design and research methodology are described in detail elsewhere. 10 In summary, the study area was the Chennai metropolis with a population of 4.3 million distributed in 155 divisions of 10 zones. As a sample, a total of 5999 subjects selected from the general population aged 40 years or above were enumerated; multistage stratified random sampling was performed on the basis of economic criteria. The data were compared between responders (1563 who visited the base hospital) and nonresponders (253 who did not visit the base hospital)

| Definitions of biochemical variables
After 8 hours of overnight fasting, a blood sample was taken for estimating the plasma glucose and serum lipids. For those with provisional diabetes, confirmation of diabetes was done by re-estimation of the fasting blood glucose by enzymatic assay; glucose was oxidized by glucose oxidase and produced gluconate and hydrogen peroxide, which was then analysed photometrically. Patients were considered to have 'known diabetes' if they were using hypoglycaemic drugs, either oral or insulin, or both.

| Diabetic retinopathy assessment
All patients had their retina photographed using the 45° fourfield stereoscopic digital photography (Carl Zeiss Fundus Camera,

What is already known?
• The triglyceride glucose (TyG) index has been demonstrated as a novel marker for insulin resistance and cardiovascular disease. Both, glycaemic control and serum triglycerides are related to microangiopathy.
• The correlation with TyG index with diabetic retinopathy or nephropathy has not been explored.

What this study has found?
• We aimed to study the relationship between TyG index and diabetic retinopathy (DR), diabetic neuropathy and diabetic nephropathy.
• We observed that TyG index is associated with diabetic retinopathy as well as nephropathy.

What are the clinical implications of the study?
• TyG index could be used for monitoring metabolic status in clinical settings.
The presence and the severity of DR were noted based on the modified Klein classification (Modified Early Treatment Diabetic Retinopathy Study scales). 12 This grading was done by two independent observers in a masked fashion; the grading agreement was high (k = 0.83).

| Diabetic nephropathy assessment
Albuminuria estimation was done by a semi-quantitative procedure (Bayer Clinitek 50 Urine Chemistry Analyzer) with the first morning urine sample. The patient was considered to have normoalbuminuria, if urinary albumin excretion (UAE) was < 30 mg/24 hour; microalbuminuria, if UAE was 30-300 mg/24 hours; and macroalbuminuria, if UAE was > 300 mg/24 hours. 10

| Diabetic neuropathy assessment
Diabetic neuropathy assessment was done by measuring vibration perception threshold (VPT) using a sensitometer. The VPT was

| Statistical analysis
Continuous variables were assessed for normality of distribution. Inter-group comparisons were assessed using ANOVA tests and were corrected for multiple testing using Tukey's method.
Variables associated with the presence of DR and albuminuria were assessed using a stepwise binary logistic regression analysis. Statistical analysis was performed using the SPSS statistical package, version 25.0 (SPSS Inc, Chicago, IL, SPSS). P values <0.05 were considered statistically significant. Variables that were significant (P < .05) on univariate analysis were entered into the stepwise logistic regression.
The mean (SD) for nephropathy present was 7.6 (0.6) and 7.4 (0.7) for nephropathy absent, P < .001. The mean (SD) for DR present was 7.6 (0.7) and 7.4 (0.7) for DR absent, P = .005. Furthermore, Table 2 shows diabetic microangiopathy distribution in the four TyGI-Q. The TyGI-Q were significantly different with the severity of DR (P = .024), albuminuria (P =.001), but not in the presence of diabetic neuropathy (P = .35). Among those with DR (n = 255), 81 (31.7%) had diabetic macular oedema (DME). The mean (SD) of TyG index was 7.7 (0.6) in those with DME and 7.5 (0.7) in those with no DME, P = .196. Proportion of subjects with DME in the four quartiles is shown in Table 2.

| D ISCUSS I ON
The TyG index has been examined in relation to metabolic disorders including diabetes and cardiovascular diseases in previous studies. [13][14][15][16][17][18][19] The current study examined the TyG index in relation to diabetic retinopathy, neuropathy and nephropathy, an association which has not been reported in literature.
It was observed that TyG index is independently associated with the presence of DR when adjusted for age, smoking and blood pressure. Similarly, a higher TyG index is associated with about twice the odds of having micro-and macroalbuminuria. However, TyG index is not related to diabetic neuropathy in our study.
We found that the proportion of those with and without DR differed significantly among the four quartiles. The formula for assessing TyG index incorporates both fasting glucose and triglyceride levels into consideration. We observed that the fasting blood glucose is higher with higher quartiles of TyG indices. The association between fasting blood glucose and DR is well established. Previous study reported increased prevalence of DR with fasting blood glucose greater than 7.03 mmol/L and for a HbA 1c more than 6.4% (46 mmol/mol). 20 This further confirms the association between higher fasting blood glucose, triglyceride and DR. It has been proposed that the triglycerides and triglyceride-rich lipoproteins are causal factors related to cardiovascular diseases 21 and microvascular complications of diabetes. 22 Although the exact mechanism is not known, triglyceride levels are proposed to be linked to microvascular complications by means of lipid peroxidation 23 and endothelial dysfunction, associated with retinal and renal complications in diabetes. 24 In addition, inflammatory markers including tumour necrosis factor-α, interleukins, leucocytes and fibrinogen may cause atherosclerotic plaque been proposed to play a crucial role in metabolic syndrome and related disorders, 25 previous report from the same population reported a prevalence of metabolic syndrome to be 73.3%. 33 The prevalence of DR in people without and with MS (21.3% and 16.9%, P =.057), and the prevalence of nephropathy (20.5% and 18.0%, respectively) (P =.296) did not differ. However, in the current study, we found a difference in DR and in diabetic nephropathy with respect to TyG index. In the current population, TyG index appears to be a better predictor for diabetic complications comparing to the aforementioned factors that also play a role in MS.
We found no correlation between TyG index and diabetic neuropathy. Kwai et al 34 also found that there was no association between changes in axonal function and triglyceride levels in a cohort of type 2 diabetic patients. It is likely that other mediating factors apart from hypertriglyceridaemia and hyperglycaemia may be involved in mediating diabetic neuropathy. Potential candidates may include increased circulating inflammatory markers including NF-κB, which is involved in altered thermonociception.
Our study was conducted in Indian population in Tamil  The mean (SD) of TyG index was 7.4 (0.7) and 7.7 (0.6) for the known diabetes and newly detected diabetes groups, respectively (P < .001). This only means that those with known diabetes must be on medication or some form of control and, therefore, may reflect a slightly lower TyG index values than the newly detected diabetes patients.
There were no significant differences in TyG index between those with and without diabetic macular oedema. One explanation could be due to small number of participants with diabetic macular oedema. Another likely explanation could be inconsistent association or differential association between the type of serum lipids and diabetic macular oedema. We previously reported that high serum low-density lipoprotein, non-high-density lipoprotein, and high cholesterol ratio were related to non-clinically significant macular oedema, and while high serum total cholesterol was related to clinically significant macular oedema. 40 In addition, we also reported that total cholesterol was associated with the incidence and all lipid types were associated with progression to sight-threatening retinopathy. 4 The strength of our study is that it is a large population-based study. The photographic standard way of grading diabetic retinopathy is another strength of this study. The number of patients with DR was not high. Nevertheless, we observed a significant association between the presence of DR and TyG index. There were several limitations in the present study. Firstly, the measurements of TG and fasting glucose had unavoidable intra-individual biological variation.
We did not study insulin resistance. Moreover, other confounding factors such as exercise habit, nutritional status and cardiorespiratory status were not included in the model. Although an association was observed between TyG-Q and the severity of DR (Table 2), there were a small number of participants with proliferative DR in all four quartiles. Therefore, the results must be interpreted with this observation in mind.
In conclusion, we found that a higher triglyceride glucose index is independently associated with retinopathy and nephropathy in individuals with diabetes. TyG may offer some clue about the presence and severity of DR, but whether it can be used as replacement to the existing risk factors is still questionable. But in the Indian population, where anaemia is a problem, anaemia can alter HbA 1c levels. 41 In this instance, TyG may be helpful in the assessment of DR. Further studies may be required to explore whether TyG index has a role in the incidence and progression of microvascular complications of diabetes.

CO N FLI C T O F I NTE R E S T
Nothing to declare. Subjects provided written informed consent, and the study was conducted according to the tenets of the Declaration of Helsinki.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.