Association between macro- and microvascular damage and the triglyceride glucose index in community-dwelling elderly individuals: the Northern Shanghai Study

The Northern Shanghai Study (NSS) (ClinicalTrials.gov Identifier: NCT02368938) is a prospective, ongoing, and multistage population study. The study protocol has been published previously [20]. Briefly, participants were enrolled if they (1) were 65 years or older, (2) resided in the urban communities in Northern Shanghai, and (3) were willing to sign the informed consent and participate in long-term follow-up. The exclusion criteria were (1) New York Heart Association functional class IV or end-stage renal disease, (2) malignant tumor or life expectancy < 5 years, and (3) stroke history in the previous 3 months. From June 2014 to August 2018, 3092 subjects were invited, and 2830 (91.5%) participants were enrolled.

A standard questionnaire was used to collect information, including age, sex, smoking habit, history of hypertension, previous cardiovascular diseases, history of diabetes mellitus, cerebrovascular disease, renal disease, and use of antihypertensive or antidiabetic drugs or insulin. Alcohol consumption habits were categorized into three groups: no alcohol consumption, moderate alcohol consumption (daily alcohol intake 0.10–19.99 g for women and 0.10–39.99 g for men), and excessive alcohol consumption (daily alcohol intake at least 20 g for women and at least 40 g for men) [21]. Body weight, body height and waist circumference (WC) were measured by professionals, and body mass index (BMI) was computed. Blood and urine samples were obtained when the participants were in a fasting state in the morning of the exam day. Biological markers were tested in the center of Shanghai Tenth People’s Hospital. The TyG index was calculated as ln[fasting triglycerides (mg/dL) × fasting glucose (mg/dL)/2]. eGFR was determined using the Chinese modified Chronic Kidney Disease Epidemiology Collaboration [22].

The participants’ blood pressure was measured while they were in a supine position by a semiautomatic oscillometric device (Omron, Japan) after 5 min of rest. Blood pressure was calculated as the average value after repeating the measurement three times at intervals of 2 min. Hypertension was defined as systolic and/or diastolic blood pressure greater than 140/90 mmHg or the current use of an antihypertensive medication. Arterial stiffness was evaluated by carotid-femoral pulse wave velocity (cf-PWV) and brachial-ankle pulse wave velocity (ba-PWV). The carotid-femoral PWV (cf-PWV) was measured by applanation tonometry (SphygmoCor, AtCor Medical, Sydney, Australia). The brachial-ankle PWV (ba-PWV) and ankle–brachial index (ABI) were recorded automatically and calculated by a VP1000 system (Omron, Kyoto, Japan). Both cf-PWV and ba-PWV were tested after the participants rested for at least 5 min. Carotid intima–media thickness (CMT) and carotid plaque were determined by the MyLab 30 Gold cardiovascular system (ESAOTE SpA, Genoa, Italy) with a 7.5-MHz probe.

In the present study, macrovascular damage included arterial stiffness (cf-PWV > 10 m/s or ba-PWV > 1800 cm/s), lower extremity atherosclerosis (ABI < 0.9), carotid hypertrophy (CMT > 0.9 mm) and carotid plaque. Microvascular damage included CKD (eGFR ≤ 60 mL/min per 1.73 m²) and MAU (urinary albumin to creatinine ratio (UACR) > 30 mg/g).

Quantitative parameters are shown as the mean ± standard deviation, and qualitative parameters are presented as numbers with the percentages in parentheses. The quantitative parameters of males and females were compared by Student’s t test, and qualitative parameters were compared by the Chi squared test. We used Pearson’s correlation to assess the correlation between the TyG index and cardiometabolic risk factors. The TyG index was grouped by quartiles when analyzing the relationship between the TyG index and macro- and microvascular damage risk through a logistic regression model. We performed two models: (a) crude; and (b) adjusted for age, sex, BMI, WC, smoking habit, hypertension, family history of premature CVD, diabetes, HDL-C, LDL-C, insulin therapy and statin therapy. We performed C-statistic values for the TyG index, fasting glucose, triglycerides, LDL-C, and non-HDL-C in a multivariable logistic regression model after adjusting for age, sex, BMI, WC, smoking habit, hypertension, family history of premature CVD, diabetes, HDL-C, insulin therapy and statin therapy. Statistical analyses were performed using SAS software version 9.4 (SAS Institute, Inc., Cary, North Carolina, USA). The significance level was set at P < 0.05.

The characteristics of the participants are presented in Table 1. Of the participants, 1571 (55.51%) were men, 1862 (65.84%) had hypertension, and 650 (22.97%) had diabetes mellitus. A total of 432 men consumed alcohol (289 consumed a moderate amount of alcohol and 143 consumed an excessive amount of alcohol). Compared with men, women were significantly less likely to be smokers (2.16% vs. 54.73%, P < 0.001) and have a family history of premature CVD (24.22% vs. 19.06%, P = 0.001). Additionally, women had lower diastolic blood pressure (DBP) (77.4 ± 9.4 vs. 79.6 ± 9.6 mmHg, P < 0.001) and WC (85.40 ± 10.06 vs. 88.21 ± 9.43 cm, P < 0.001) and higher total cholesterol (195.8 ± 36.7 vs. 176.6 ± 36.9 mg/dL, P < 0.001), HDL-C (54.54 ± 13.41 vs. 47.58 ± 12.48 mg/dL, P < 0.001), LDL-C (119.8 ± 32.5 vs. 108.8 ± 31.4 mg/dL, P < 0.001), triglycerides (148.6 ± 95.2 vs. 137.7 ± 91.2 mg/dL, P = 0.002), and TyG index (8.78 ± 0.57 vs. 8.71 ± 0.57, P < 0.001). In terms of vascular damage, women showed lower CMT (0.62 ± 0.15 vs. 0.66 ± 0.17 mm, P < 0.001), lower incidence of carotid plaque (60.09% vs. 66.88%, P < 0.001) and lower extremity atherosclerosis (12.86% vs. 16.76%, P = 0.004) but higher ba-PWV (1907.4 ± 453.2 vs. 1827.5 ± 387.7 cm/s, P < 0.001), and eGFR (73.9 ± 12.4 vs. 67.8 ± 10.8 mL/min per 1.73 m², P < 0.001).

In the age-, sex- and education-adjusted correlation analysis, the TyG index was correlated with SBP (r = 0.13), DBP (r = 0.07), WC (r = 0.29), BMI (r = 0.25), HDL-C (r = − 0.49), LDL-C (r = 0.15) and total cholesterol (r = 0.124) (P < 0.001 for all) (Table 2).

As shown in Table 3 and Fig. 1, in the univariate logistic regression, with Q1 (first quartile) set as the reference, the TyG index levels in Q4 (fourth quartile) were associated with an increased OR (odds ratio) for cf-PWV > 10 m/s (OR = 2.48, 95% confidence interval [95% CI] 1.97–3.13, P_{for trend} < 0.001), ba-PWV > 1800 cm/s (OR = 1.76, [95% CI] 1.43–2.18, P_{for trend} < 0.001), lower extremity atherosclerosis (OR = 1.77, [95% CI] 1.31–2.40, P_{for trend} < 0.001), MAU (OR = 1.70, [95% CI] 1.27–2.12, P_{for trend} < 0.001) and CKD (OR = 1.41, [95% CI] 1.07–1.84, P_{for trend} < 0.01) but not for carotid hypertrophy (OR = 1.27, [95% CI] 0.81–1.99, P_{for trend} = 0.26) or carotid plaque (OR = 1.12, [95% CI] 0.90–1.39, P_{for trend} = 0.17).

As shown in Table 4 and Fig. 2, after adjustment for age, sex, BMI, WC, smoking habit, hypertension, family history of premature CVD, diabetes, HDL-C, LDL-C, insulin therapy and statin therapy, we found that the TyG index levels in Q4 were associated with an increased OR for cf-PWV > 10 m/s (OR = 1.86, 95%, [95% CI] 1.37–2.53, P_{for trend} < 0.001), ba-PWV > 1800 cm/s (OR = 1.39, [95% CI] 1.05–1.84, P_{for trend} = 0.02), MAU (OR = 1.61, [95% CI] 1.22–2.13, P_{for trend} < 0.001) and CKD (OR = 1.67, [95% CI] 1.10–1.50, P_{for trend} = 0.02) compared with the TyG index levels in Q1, but the levels were not associated with lower extremity atherosclerosis (OR = 1.11, [95% CI] 0.75–1.63, P_{for trend} = 0.66), carotid hypertrophy (OR = 0.91, [95% CI] 0.52–1.60, P_{for trend} = 0.83) or carotid plaque (OR = 0.93, [95% CI] 0.71–1.23, P_{for trend} = 0.81).

Additionally, LDL-C and non-HDL-C were associated with carotid hypertrophy, carotid plaque and ABI < 0.9, while the TyG index, fasting glucose, and triglycerides were not. The TyG Index had a higher C-statistic for cf-PWV > 10 m/s and ba-PWV > 1800 cm/s but not for MAU or CKD after adjustment for age, sex, BMI, WC, smoking habit, hypertension, family history of premature CVD, diabetes, HDL-C, insulin therapy and statin therapy (Additional file 1).

In sensitivity analyses, the TyG index was significantly associated with cf-PWV > 10 m/s, ba-PWV > 1800 cm/s, and CKD in subjects without diabetes (n = 2180) and with cf-PWV > 10 m/s and CKD in subjects without hypertension (n = 968). However, the TyG index was less significantly associated with micro- and macrovascular damage in subjects without hypertension and diabetes (n = 803). This finding may be because of a much smaller sample size in these subgroups (Additional file 2).

IR plays a role in metabolic abnormalities. The TyG index not only serves as a simple and reliable surrogate marker of IR but is also associated with metabolic abnormalities. Simental et al. found that the TyG index was correlated with IR assessed by HOMA-IR in healthy subjects [23]. Fernando et al. reported that the TyG index was an accessible and inexpensive way to identify insulin resistance in the general population [9]. Similar results were also found in obese adolescents [11, 12]. A higher TyG index was significantly associated with a higher risk of diabetes [16, 24, 25], hypertension [17, 26] and metabolic syndrome [27]. In the present study, we also found that the TyG index was significantly correlated with cardiometabolic risk factors.

IR is an important mechanism of type 2 diabetes (T2D). Macrovascular disease (CVD, stroke and peripheral arterial disease) is one of the complications of T2D. Macrovascular disease is the main cause of morbidity and mortality in patients with T2D [28]. Recently, it was reported that a history of peripheral arterial disease was associated with the risk of lower‑extremity amputation in patients with T2D [29]. It is very important to evaluate the structure and function of macrovasculature in the prevention and treatment of T2D. The TyG index may be associated with macrovascular damage. Irace et al. showed that the TyG index was significantly associated with CMT [13]. However, we did not find that the TyG index was associated with CMT or carotid plaque. It was reported that the TyG index was independently associated with arterial stiffness, as assessed by ba-PWV in a Korean population [14, 15]. In the present study, we found that the TyG index was associated with abnormalities of cf-PWV and ba-PWV. Cf-PWV is used as the gold standard of arterial stiffness. Our paper is the first to report the association between the TyG index and cf-PWV. It was reported that the prevalence of coronary artery calcification increased with an increasing TyG index [30, 31]. Moreover, subjects with a higher TyG index had a higher risk of developing cardiovascular events after adjusting for conventional CVD risk factors [18, 32].

The UACR and eGFR are commonly used to assess renal function in patients with diabetes. MAU and CKD were associated with a higher risk of cardiovascular events [33]. Moreover, MAU is strongly associated with cardiac microvascular dysfunction [34]. However, the association between the TyG index and nephric microvascular damage remains unclear. Sikandar et al. reported that the TyG index demonstrated a higher positive linear correlation with UACR [35], but they did not perform further analysis. In the present study, we found that a higher TyG index was associated with a higher risk of CKD and MAU. The TyG index could be a predictor of incident CKD. The TyG index might play a role in nephric microvascular damage. In the future, more studies are needed to explore the relationship between the TyG index and microvascular damage, such as CKD, MAU, retinopathy, cardiac microvascular dysfunction and markers of endothelial injury.

The results of the present study need to be interpreted within the context of their limitations. First, this is a cross-sectional study, which limits what we can infer about the causality of the results. Second, the study included only older community-dwelling Chinese participants, so we should be cautious in extrapolating the present findings to other subjects. Third, we did not assess HOMA-IR in the present study, which is the gold standard method for measuring IR. However, the TyG index is certainly more convenient to measure in routine clinical practice. Finally, some metabolic conditions that might impact the results were not included in our investigation, such as primary triglyceride abnormalities.