Introduction
Cardiovascular diseases (CVD) are becoming increasingly frequent and associated with a high incidence of disability, and death. CVD are the first cause of mortality in the world, representing 31.5% of all global deaths. In 2015, 17.7 million people died by CVD according to the World Health Organization (WHO) [
1]. Prevention of CVD through the control of risk factors is a priority in most developed countries [
2].
The prevalence of T2DM is rising at an alarming rate globally. T2DM as a common and severe condition has posed a significant burden on patients, their families, and health care systems. T2DM is a significant risk factor for CVD, such as coronary heart disease (CHD), stroke, and peripheral arterial disease. Patients with T2DM have a 2–4 times higher risk of CVD incidence than those without diabetes [
3]. CVD is the leading cause of death among individuals with T2DM, accounting for approximately 70% of death [
4,
5].
It is well documented that elevated blood pressure, hyperglycemia and low-density lipoprotein cholesterol (LDL-C) abnormalities are vital contributors to the risk of CVD in individuals with diabetes [
6‐
8]. Several meta-analyses and reviews have highlighted the relationship between TG level, and risk of CVD among the general population, an increase in the TG concentration is exposed to the higher risk of CVD events [
9‐
11]. However, the association of elevated TG level and the risk of CVD in type 2 diabetic population is still not conclusive. Some studies found that there is no association between TG and the risk of CVD among T2DM [
12,
13], whereas some studies found the higher rate of CVD incidence in diabetic individuals [
14,
15]. In randomized trials, patients used medications that reduce triglyceride levels also had different results in CVD risk. Clinical trials of agents that lower TG, specifically fenofibrate and niacin, have failed to demonstrate a reduction in CVD outcomes when administered in addition to appropriate medical therapy [
16,
17]. Recent studies of n − 3 fatty acid products have not shown a benefit in patients receiving statin therapy [
18‐
20]. However, the Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention Trial (REDUCE-IT) study indicated that the risk of major ischemic events, including cardiovascular death, was significantly lower with icosapent ethyl compared to placebo in patients with elevated triglyceride levels [
21].
All these studies have yielded conflicting results. To our knowledge, there has been no systematic analysis of these evidence to date. Thus, we performed a systematic review and meta-analysis on the available studies to evaluate the relationship between circulating TG and CVD in patients with T2DM.
Discussion
In this systematic review and meta-analysis, we pooled the data for TG levels of 132, 044 T2DM subjects from 31 studies. To the best of our knowledge, this is the first meta-analysis to comprehensively investigated the predictive role of TG levels and assessed the factors influencing the predictive ability of TG. Our study findings provide useful implications for the development of disease prevention strategies. Firstly, the increased TG levels are associated with an increase in the risk of incident CVD. However, this association remained statistically insignificant when adjusted for other blood lipid parameters; it affirms the findings of previous studies reported that the TG is not an independent risk factor for incident CVD in the T2DM population [
30,
31]. In contrast, the data from the general population confirmed that fasting serum TG are independently associated with the risk of CVD, even after adjustment for other lipid parameters [
10,
32,
33]. Furthermore, our result indicated that a higher TG level tends to increase the risk of CVD mainly due to an increased risk of CHD but not a stroke. Secondly, the subgroup analyses on the impact of TG on CVD showed prominence among T2DM individuals with a mean age < 65 years. It is consistent with a recent cross-sectional study which found higher TG level was associated with an increased risk of CVD among the patients with short duration of diabetes, whereas, a lower TG level was associated with an increased risk of CVD in patients with longer duration of diabetes, and this association was mainly driven by CHD [
34]. Older age tends to have a longer duration of diabetes in general. It had been reported that triglyceride-rich lipoproteins are the primary source of energetic heart lipids [
35]. Moreover, it is well understood that elderly T2DM individuals have more severe insulin resistance and deteriorated Beta-cell function, resulting in poor glycemic control in patients with T2DM, which is a well‐established causal factor for CVD [
36], and it is important to have a sufficient supply of TG to the heart because it is harder for the heart to use glucose as fuel. Therefore, a higher blood TG level might not be a risk factor for elderly T2DM patients but an essential fuel for their heart. Thirdly, a subgroup analysis indicated a weaker effects of elevated TG levels on CVD risk in male patients with T2DM compared to the female counterparts, and this is in line with Framingham Heart Study that reported higher TG to correlate more strongly with CVD risk in women than in men [
37]. However, the reason behind this remains unknown, and future studies are invited to explore the pathogenic mechanism behind such inference. Moreover, in agreement with Liu’s meta-analysis in general population [
38], a stronger association was also found in Asians compared with Western populations, and participants with the prior macrovascular disease compared with no history of macrovascular disease participants. Lastly, our study found that T2DM patient with eGFR ≤ 60 min/ml/1.73 m
2 showed the blood TG as a protective factor against CVD, this finding coincides with the results of some observational studies based on hemodialysis patients [
39,
40], according to their findings a higher cholesterol level was shown to predict a better survival. However, the exact reasons are not known and require further elucidation.
Evidence showed a TG involvement in atherogenesis indirectly [
41,
42]. The progression of coronary atherosclerosis is powerfully stimulated by interactions between diabetes-associated factors and other factors such as abnormal lipid metabolism [
43], and hypertriglyceridemia involved in these two important factors. Hypertriglyceridemia deteriorates diabetes by impairing the function of β-cell and causing peripheral insulin resistance (IR) [
44]. TG overload in islets interferes with glucose metabolism, and the accumulation of metabolites derived from fatty-acid esterification impairs β-cell function [
45‐
47]. The impairment of β-cell function decreases the glucose-induced insulin secretion, resulting in an increased glycemic level in T2DM patients, which significantly increases the risk of cardiovascular disease. Hypertriglyceridemia has a unidirectional relationship with peripheral IR. The metabolites of TG such as free fatty acids, diacylglycerol and others can regulate insulin-signaling pathways through activating several serine/threonine kinases, which suppress insulin receptor and tyrosine phosphorylation of insulin receptor substrates, inducing peripheral IR [
48‐
50]. Many studies have indicated that IR leads to inflammation, altered coagulation and atherosclerosis [
51‐
54]. Furthermore, an independent association between IR and CVD has been reported [
55,
56]. TG involves in atherogenesis also by altering LDL-particle size. Some studies indicated that LDL-particle size showed a significantly negative correlation with the serum TG levels [
57,
58], this means that when serum TG is elevated, the LDL-particle size became smaller. The Québec Cardiovascular Study demonstrated that patients had LDL-particle size of 25.5 nm or smaller, the CHD incidence increased significantly as the serum LDL-C level increased, while in patients having large LDL-particle sizes of 26.0 nm or greater, no significant difference in CAD events was observed according to the absolute serum LDL-C level [
59]. Smaller LDL-particle size showed a powerful atherogenic effect [
41]. Moreover, TG involved in atherogenesis by other mechanisms. TG metabolites, i.e., chylomicrons, very low-density lipoprotein, and remnant-like particle cholesterol, which are TG-rich lipoproteins, and, apolipoprotein (apo) C-II, and apo C-III which are involved in the metabolic process, etc., have been demonstrated to be involved in the progression of atherosclerosis [
41]. In addition, mounting evidence suggested that TG may stimulate atherogenesis through the production of proinflammatory cytokines, fibrinogen and coagulation factors and impairment of fibrinolysis [
60,
61].
However, a series of experimental studies based on alloxan-diabetic rabbits demonstrated that the lipoproteins in the hypertriglyceridemic diabetic rabbits are much larger than in normotriglyceridemic rabbits, and the plasma lipoproteins that contain most of the plasma cholesterol are so large in size (the diameter larger than 75 nm) that they are not able to enter the arterial wall. A reduced aortic cholesterol influx means that the subendothelial macrophages and smooth muscle cells are exposed to relatively small amounts of cholesterol, this probably explains the slow development of atherosclerotic plaques in severely hypertriglyceridemic diabetic rabbits [
62‐
64]. However, these findings did not mean that the very large triglyceride-rich lipoproteins in diabetes are harmless, it means these are less atherogenic than smaller lipoproteins and may partially explain that TG is not an independent risk factor for CVD in diabetes patients.
Even in diabetes patients receiving appropriate treatment with statins, a substantial residual cardiovascular risk remains [
65‐
67]. Also, the American Diabetes Association (ADA) recommending TG-lowering as an important secondary target in patients with diabetes [
68]. There are different triglyceride-lowering drugs available. Clinically, the drugs that predominantly lower the triglyceride levels are fibrates, fish oils and nicotinic acid. The statins, ezetimibe, proprotein convertase subtilisin/kexin type 9 inhibitors are LDL-cholesterol lowering agents which also reduces the triglycerides level. Besides, there are newer triglyceride-lowering agents currently under evaluation. These novel specific agents for high TG include newer fibrates such as pemafibrate; newer omega-3 FAs such as eicosapentaenoic acid; combined peroxisome proliferator-activated receptor alpha/gamma agonists such as aleglitazar; and new drug classes of apolipoprotein CIII antisense therapies, microsomal triglyceride transfer protein inhibition such as lomitapide, and diacylglycerol acyltransferase inhibitors such as pradigastat; and anti-angiopoietin-like protein 3 or anti-angiopoietin-like protein 4. Furthermore, probiotics and gut microbiome is a new therapeutic modulation for hypertriglyceridemia. Many new TG reducing therapies are in clinical trials. In addition to the REDUCE-IT study, the Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High CV Risk Patients With Hypertriglyceridemia (STRENGTH) and the Pemafibrate to Reduce Cardiovascular Outcomes by Reducing Triglycerides In Patients With Diabetes (PROMINENT) studies are ongoing large CV outcomes trials involving high-risk CVD patients, including a large percentage of patients with diabetes, undergoing statin therapy [
17,
69‐
72].
We acknowledge that our study had several limitations. First, most of the included studies were observational studies, but several were randomized control trials, which might have led to the heterogeneities. Furthermore, because our study is a literature-based meta-analysis, the lack of access to individual patient data might lead to the heterogeneities in TG contrast level and adjustments. Meanwhile, the differences in the definition of CVD, effect size types (OR, RR, or HR) have to be mentioned as well. However, a random-effects model and subgroup analyses were conducted to minimize the impact of heterogeneity. Second, a single baseline TG estimation may have led to the misclassification of study participants in each category and multiple determination of TG during follow up would increase the precision. Third, although we used the most fully adjusted RR from each study to calculate the pooled result, there are so many confounders that may have an effect on the outcome, such as lifestyle and socio‐economic status, e.g., cigarettes smoking and alcohol drinking, antihypertension and lipid‐lowering drugs and antidiabetic medicine. Further well designed clinical trials are needed to overcome all these confounders. Finally, evidence suggests that postprandial (non-fasting) TG levels may have a stronger association with CVD than fasting levels [
73] in the general population, but in our study, only one study had reported on the relationship between postprandial TG and CVD, so we could not evaluate this issue very robustly.
Authors’ contributions
XFY searched databases, selected studies, extracted data, analyzed data and wrote the manuscript. WK searched databases, selected studies, extracted and analyzed data, contributed to the discussion, and reviewed and edited the manuscript. MIZ selected studies, extracted data and reviewed and edited the manuscript. LLC contributed to the design and discussion and reviewed and edited the manuscript. All authors read and approved the final manuscript.