Triglyceride/High-Density Lipoprotein Cholesterol Ratio Is Associated with In-Hospital Mortality in Acute Type B Aortic Dissection.

Background Triglyceride/high-density lipoprotein cholesterol (TG/HDL-c) ratio varies with vascular and other metabolic diseases. However, its role in acute type B aortic dissection is not well understood. In the current study, we evaluated the relationship between TG/HDL-c ratio and in-hospital mortality in type B aortic dissection. Methods We performed a retrospective analysis of consecutive patients between January 2015 and December 2018, by targeting dependent (TG/HDL-c ratio) and independent (in-hospital mortality) variables. TG/HDL-c ratio was determined as a division of TG levels by HDL-c levels. Results Of 523 patients in the study, we found a mean age of 55.00 ± 11.74 years, 15.68% of them being female. A fully-adjusted model revealed a positive relationship between TG/HDL-c ratio and in-hospital mortality in acute type B aortic dissection after adjusting confounders (OR = 2.08, 95% CI 1.32 to 3.27). This relationship was also nonlinear, with a point of 2.05. OR values (and confidence intervals) for the right (>2.05) and left (≤2.05) sides of the inflection point were 1.0 (0.580-1.26, P = 0.983) and 3.17 (1.54-6.57, P = 0.001), respectively. Conclusions The TG/HDL-c ratio and in-hospital mortality in type B AAD have a nonlinear relationship among Chinese population. This ratio increased in-hospital mortality when it is less than 2.05.


Introduction
Acute aortic dissection (AAD) is a rare, life-threatening condition associated with high morbidity and mortality rates [1,2]. In-hospital mortality rates of patients with AAD is up to 27.4% [2,3], with type B AAD reported to account for one-third of all AAD cases [4]. Despite the recent improvements in management of type B AAD, mortality rates and postoperative complications of this disease remain high [5]. Studies have reported that medical treatment has been accepted as standard management for type B AAD, although its 5-year total survival rate stands at 60% [6]. It is, therefore, urgent to investigate the risk factors of type B AAD mortality and develop an effective intervention. Studies have reported several risk factors for type B AAD in-hospital mortality, including age, hypertension, ischemic complications, and genetic factors [7].
To date, dyslipidemia is a well-documented risk factor for stroke [8,9], metabolic syndrome [10], and cardiovascular disease [11,12]. Moreover, results from a meta-analysis suggested that lipid-modifying therapy is a protective factor on mortality after abdominal aortic aneurysm repair [13]. A cross-sectional study implicated dyslipidemia aortic stiffness [14], with international guidelines stating that lipid profiles should be managed [15]. Atherogenic dyslipidemia, a combination of low high-density lipoprotein cholesterol (HDL-c) and high triglyceride (TG) with elevated small dense lowdensity lipoprotein (LDL) particles and apolipoprotein B, is an important part of the metabolic syndrome and a powerful predictor of cardiovascular disease [16][17][18]. However, the costs of performing the LDL phenotyping are expensive and have not been standardized [19,20]. As a possible alternative, the TG/HDL-c ratio has been considered an easily available atherosclerotic marker [21]. Previous studies have reported that a higher TG/HDL-C ratio is related with LDL phenotype B, small insulin resistance, HDL particles, and cerebro/cardiovascular diseases [16]. In the current study, we aimed to evaluate the relationship between TG/HDL-c ratio and inhospital mortality in patients with type B AAD.

2.1.
Participants. This was a retrospective study, in which we nonselectively and consecutively evaluated 1526 patients between January 2015 and December 2018 at the Second Xiangya Hospital of Central South University. The Diagnostic process was mainly based on guidelines proposed by ESC 2014 on the treatment and diagnosis of aortic diseases [22]. Among the patients, 140 of them who showed presence of symptoms for more than 14 days were excluded. Similarly, those whose covariate data was missing (uncompleted serum triglyceride or high-density lipoprotein cholesterol tests) were also excluded. According to the Stanford classification, which is based on anatomical categorization [23], a total of 564 type A AAD patients were excluded. Finally, 523 type B AAD patients were enrolled in the study (Figure 1). The study was approved by the Institutional Review Board at the Second Xiangya Hospital of Central South University.
2.3. Statistical Analysis. Continuous variables were expressed as mean ± standard deviations of the mean (normal distribution) or median with interquartile range (IQR) (Skewed distribution), while categorical variables were expressed as frequencies or percentages. Kruskal-Wallis H (skewed distribution) and One-Way ANOVA tests (normal distribution) or χ 2 (categorical variables) were performed, applied, and compared with Q1, Q2, Q3, and Q4 (TG/HDL-c ratio categorized as quartiles). To investigate the relationship between TG/HDL-c ratios and in-hospital mortality, we performed statistical analyses based on 3 points. First, we use univariate and multivariate linear regression model analyses. A total of three models were established: model I, in which only sociodemographic data were adjusted; model II, the model I added other covariates shown in Table 1. Secondly, in order to solve the nonlinearity of TG/HDL-c ratio and in-hospital mortality, a smooth curve fitting (penalized spline method) was conducted. If  2 BioMed Research International nonlinearity was detected, we first calculated the inflection point using the recursive algorithm and then constructed a two-piecewise linear regression on both sides of the inflection point. This was followed by determination of the best fit model on the basis of P values for log-likelihood ratio test. The third step involved subgroup analysis based on stratified linear regression models. Continuous variables were initially transformed into categorical variables using the clinical quartile of cut point, followed by interaction tests. Tests for effect modification across subgroup indicators were followed by analysis of likelihood ratios. Sensitivity analyses were then performed, in which the TG/HDL-c ratio was converted to categorical variables for calculation of the P for trend to ensure the robustness of the analysis. The aim of this step was to validate the results of TG/HDL-c ratio as continuous variables and to detect non-linearity. These analyses were carried out in Empower Stats (http://www.empowerstats.com, X&Y Solutions Inc., Boston, MA) [24] and R (http://www.r-projec.org, The R Foundation). Statistical significance was determined at P < 0:05.

Baseline Characteristics and Univariate Analyses.
A total of 523 type B AAD patients were enrolled in the study, and this resulted a median age of 55:00 ± 11:74 years. 84.32% of these were male. A flow chart outlining their selection is presented in Figure 1, while baseline features of the enrolled patients are outlined in Table 1. In-hospital mortality was found in 33 (6.31%) patients. Univariate analyses indicated that the outcome variable was associated with CRI,   (Table 2).

TG/HDL-c Ratio Increased the In-Hospital Mortality.
Results from the Kaplan-Meier analysis showed a significantly higher cumulative in-hospital survival rate in the Q1 relative to the other groups (log-rank χ 2 = 12:96, P = 0:005) (Figure 2).

The Relationship between TG/HDL-C Ratio and
In-Hospital Mortality. We established three models in order to examine the independent effects of TG/HDL-c ratio on inhospital mortality after adjusting for confounding factors (Table 3). Model III resulted in a TG/HDL-c ratio (OR = 2:08, 95% CI = 1:32 to 3.27, P = 0:002) remained an important predictor of in-hospital mortality after all adjusted covariates are showed in Table 1. We also converted TG/HDL-c ratio from continuous to categorical variable (quartiles), the P for the trend of categorized TG/HDL-c ratio in the fully adjusted model matched with the result when TG/HDL-c ratio is a continuous variable. However, when the TG/HDL-c ratio enters the fully-adjusted model as a categorical variable, the trend of the effective value in the different TG/HDL-c ratio group had nonequidistant changes. There may be a nonlinear relationship between TG/HDL-c ratio and in-hospital mortality according to this kind of nonequidistant changes in effect size.

The Nonlinearity of TG/HDL-c Ratio and In-Hospital
Mortality. We performed further analysis on the nonlinear relationship of TG/HDL-c ratio with in-hospital mortality (Table 4, Figure 3). Results of the smooth curve indicated nonlinear relationship (adjusted for other covariates listed in Table 1) between TG/HDL-c ratio and in-hospital mortality. We used the two-piecewise linear regression and the linear regression models to fit the association between TG/HDL-c ratio and in-hospital mortality, respectively. The P value for the log-likelihood ratio test was 0.017, indicating that the two-side linear regression was more appropriate for fitting the association between TG/HDL-c ratio and inhospital mortality because it can accurately represent the relationship between them. The inflection point, determined  Though they are all consistent with an increased risk of in-hospital mortality in the right side of the inflection point, there is a lack of a statistically significant association with the TG/HDL-c ratio.

Discussion
The current study resulted in three main findings. First, we found a positive relationship between TG/HDL-c ratio and in-hospital mortality, after adjusting other covariates. This indicated that an increase in TG/HDL-c ratio produces 2.08-fold increase in in-hospital mortality. Secondly, our analysis revealed nonlinearity between in-hospital mortality and TG/HDL-c ratio. Given that this correlation displayed a ratio-response pattern, our findings may provide clues for the potential pathophysiologic mechanisms. Thirdly, our results confirmed an association between in-hospital mortality and TG/HDL-C ratio in type B AAD in Chinese patients which can be used to screen high-risk individuals. To better understand the trend in this association in differ-ent populations, we performed subgroup analysis and found stable results.
Several studies have suggested high levels of TG/HDL-c ratio in cardiovascular diseases across Europe and Asian Pacific populations [25][26][27]. TG/HDL-c ratio indicates harmful small/dense LDL particles and has been found to have a positive association with atherosclerosis [28]. Moreover, a previous study [14] analyzing lipid parameters of 603 participants to evaluate aortic stiffness found a significant association between aortic stiffness with total cholesterol/HDL, total cholesterol, and non-HDL. Therefore, the cholesterol deposition in the aortic tissues may explain the development of type B AAD or abdominal aortic aneurysm. In additional, there is a large number of evidence that has suggested the role of TG/HDL-c ratio in prediction of poor outcomes in cardiovascular, especially myocardial infarction and ischemic stroke [28,29]. Furthermore, previous studies have also deduced that the TG/HDL-c ratio has a positive  Figure 3: Association between TG/HDL-c ratio and in-hospital mortality. A nonlinear association between TG/HDL-c ratio and in-hospital mortality was found. The solid red line represents the smooth curve fit between variables. Blue bands represent the 95% confidence interval from the fit. All adjusted for gender, age, BMI, hypertension, smoking, atherosclerosis, WBC, PLT, Hb, ALT, AST, Cr, UA, and TnT. association with the occurrence of ischemic [30], which is potentially the source of high TG/HDL-c ratio in type B AAD patients [31].
Although operation is still the most effective way to avoid in-hospital deaths in patients with AAD, other risk factors for the prediction of in-hospital mortality need to be considered. Our results suggested that TG/HDL-c ratio is one of the risk factors for in-hospital mortality. Clinical therapies should therefore arouse high attention to the ratio with these patients. Increased TG/HDL-c ratios are not only related to aortic stiffness, but also an indicator of inflammatory factors and oxidative stress, which show a risk of type B AAD and a poor prognosis. Recent evidence suggests that inflammatory factors and oxidative stress play a crucial role in the pathogenesis and progression of AAD [32,33]. For instance, Daiber et al. [34] reported the quality of high-density lipoprotein changes under oxidative stress conditions. Similarly, Kaye et al. [35] found that if the levels of HDL are insufficient, cholesterol and lipids in arterial walls trigger an exacerbated inflammatory response. Meta-analysis reported that low serum lipids may reduce patient mortality after abdominal aortic aneurysm repair [13]. Overall, increased TG/HDLc ratio in type B AAD is associated with aortic stiffness, inflammatory factors, and oxidative stress.
The advantages of our research are strict statistical adjustment, subgroup analysis, and large research population. However, there are some weaknesses in the study. Firstly, the participants herein were mainly Chinese AAD patients, recruited in central south China. As a result of this sampling, our study may not be generalizable to AAD populations in other regions as factors influencing cholesterol levels may vary. Secondly, we excluded chronic AD and type A AAD patients from analysis; therefore, the study conclusion does not apply to these people. Consequently, because it is a retrospective study, some variables such as physical activity, diet, and daily habits were not achieved. We also were not able to know if the patients had fibrates.   Nevertheless, our study results indicate that the TG/HDL-C ratio is a risk factor. If a patient takes lipid-lowering drugs, high TG/HDL-C ratio will have a strong positive association with in-hospital mortality.

Conclusions
The TG/HDL-c ratio and in-hospital mortality in type B AAD have a nonlinear relationship among Chinese population. This ratio increased in-hospital mortality when it is less than 2.05.

Data Availability
The datasets used and/or analyzed during the present study were availed by the corresponding author on reasonable request.