Introduction

Hypertension is the leading cause of cardiovascular events and all-cause mortality worldwide, which has become an emerging challenge for global public health [1]. In China, with an aging population and changing lifestyles, the prevalence of hypertension is also increasing year by year, with approximately one-third of the adult population, or more than 300 million people, suffering from hypertension between 2014 and 2015 [2, 3]. Therefore, early identifying the high-risk individuals and develo** effective primary prevention strategies are very urgent to reverse the rapidly rising trend of hypertension.

Disorders of lipoprotein metabolism, in particular elevated plasma triglycerides (TG), and elevated fasting blood glucose (FBG), are all established risk factors for cardiovascular disease, especially in hypertension [4, 5]. This could be explained by insulin resistance (IR) via at least three mechanisms: inflammatory endothelial dysfunction [6], ectopic synthesis of angiotensinogen [7], and hyperinsulinaemia overstimulating the renin-angiotensin–aldosterone system [8]. Recently, the triglyceride-glucose (TyG) index, which was calculated by using TG and FBG [9], has been proposed as a surrogate of IR, and correlated with various indices of IR [10, 11]. Previous research, however, has primarily focused on the association between TyG index and the incidence of prediabetes [12], diabetes [13], cardiovascular events [14, 15], and all-cause mortality [16, 17]. Very limited studies were conducted to explore the association of TyG index with new-onset hypertension [18,19,20,21,22,23,24,25,26,20,21,22,23], in children or adolescents [42], or in the elderly population [24]. Moreover, a prospective study in Spanish [18] reported a positive association between TyG index and hypertension in the general population during a long-term follow-up. Zheng et al. [19] also conducted another single-center, longitudinal study with 4,686 subjects followed up for 9 years, and demonstrated that TyG index could predict incident hypertension among the Chinese population. Of note, another study [28] reached a similar conclusion using a national cohort, but it was limited to middle-aged and older adults over 45 years. In addition, in a recent meta-analysis of 8 studies involving 200,044 general adult participants [43], the relative risk of hypertension was higher for the highest category of TyG index compared with the lowest. These studies were consistent with our study, which showed that high levels of TyG index were significantly associated with increased risk of new-onset hypertension and that a linear association was observed. Contrary to previous studies and our results, two cross-sectional studies [25, 26] did not find a significant association between TyG index and hypertension in obese or normal‐weight individuals. This may be attributed to the heterogeneity of the selected population and the sample size of the study. Therefore, further research on this topic is needed.

In stratified analysis, we found that most of the variables did not significantly modify the association between TyG index and new-onset hypertension, which indicates that the results of this study are applicable to the majority of the general population. However, the association between TyG index and new-onset hypertension was stronger in patients with low carbohydrate intake (< 288 g/day). Previous studies have reported that higher carbohydrate intake was related to a higher risk of hypertension [44, 45]. Therefore, individuals with higher carbohydrate intake will offset some of the risk from TyG index.

The potential mechanisms for the association between TyG index and new-onset hypertension may be explained by IR. IR has been confirmed by many studies to be significantly associated with hypertension by many mechanisms [6,7,8]. In recent years, many studies have concluded that TyG index is a surrogate of IR because it was correlated with various indices of IR, such as the M rates in the hyperinsulinaemic–euglycaemic clamp test [10] or the homeostasis of minimal assessment of insulin resistance (HOMA-IR) [46], and with the degree of carotid atherosclerosis [11]. More mechanistic studies are needed to further validate the relationship between TyG index, IR and hypertension.

The major strengths of this study were the 6-year national, longitudinal population-based study and the large number of subjects used to explore the relationship between TyG index and new-onset hypertension. However, certain limitations also existed in this study. First, although we adjusted for potential confounders such as demographics, dietary intake, and blood biochemical markers as much as possible, the E-values for the hazard ratio and lower confidence bound for the primary outcome were also small, which implies that little unmeasured confounding would be needed to reduce the observed association or its 95% CI to the null. However, residual confounding could not be completely eliminated. Second, limited by observational studies, we could not determine the causal relationship between TyG index and new-onset hypertension. Third, since the definition of hypertension was based on physician on-site blood pressure measurement data and questionnaires, there may be recall bias, but the findings remained consistent when we excluded questionnaire-based information on hypertension obtained in the sensitivity analysis. Forth, we could not explore the relationship between TyG index and different hypertension subtypes since information related to 24-h dynamic changes in blood pressure was not available. Fifth, considering that death was an inevitable competitive risk, we may underestimate the relationship between TyG index and new-onset hypertension. However, in our study population, only 74 (1.6%) participants died during the follow-up, which is unlikely to change the trend of results. Last, this study was limited to the Chinese general population, and more studies were needed to confirm the consistency of findings in other ethnic and national studies in the future.

Conclusions

In conclusion, we confirmed that high levels of TyG index were associated with higher risk of new-onset hypertension and showed a linear relationship through a longitudinal national cohort. Our findings suggest that maintaining a relatively low level of TyG index will help with the primary prevention of hypertension. In clinical practice, the TyG index is easily available, and clinicians could use this indicator to risk-stratify the general population in order to provide more personalized prevention or treatment.