From June 2008 to February 2012, of the 301 patients admitted to this tertiary referral center with a diagnosis of STEMI for primary PCI (pPCI), 247 had a lipid profile checked within 24 hours of admission so were enrolled and categorized into either the lower (≦150 mg/dl, N = 163, or 66%) or the higher (>150 mg/dl, N = 84, or 34%) triglyceride group according to the baseline level of serum triglycerides. Baseline characteristics, angiographic findings, and clinical outcomes were obtained from electronic or written medical records as well as the cardiac catheterization laboratory database and were compared between groups. The study protocol was approved by the human research committee of Taichung Veterans General Hospital (approval No. CE12289), and requirement of informed consent was waived by the institutional review board.

The diagnosis of STEMI was made based on conventionally standardized criteria [2]. Coronary angiography was performed from either a radial or femoral approach at the discretion of the interventional cardiologists. Grade of blood flow was determined by TIMI classification system [14]. Angiographic stenosis was defined as a luminal diameter reduction of ≥50% by quantitative coronary angiography, and critical stenosis was defined as ≥70% narrowing of the coronary artery luminal diameter. Complete coronary occlusion was defined as absence of antegrade flow of contrast media beyond a specific vascular segment. The infarct-related lesion was determined by the relevant ECG leads showing ST-segment elevation (or ST-segment depression along with a tall R wave in lead V1 for posterior infarction) and by the typical vascular morphological features of thrombus-laden or hazy filling defects along with compromise of distal flow. Once angiographic occlusion or critical stenosis of any of the coronary arteries was demonstrated, PCI with or without stenting was performed for the infarct-related artery but not for any other bystander vessel unless hemodynamic instability persisted despite opening of the culprit lesion. The success of pPCI was defined as achievement of the TIMI flow of the infarct related artery to grade II or III. Patients were then taken to our intensive coronary care unit for post-procedural care. Pharmacological treatment, including dual-antiplatelet medication, was in accordance with current guidelines [2]. In-hospital complications, including cardiogenic shock requiring intra-aortic balloon pumping support, new-onset atrial fibrillation, ventricular arrhythmia, and death served as in-hospital outcomes.

Discharged patients were followed up by medical records of clinic visits or telephone contacts for MACE, including data regarding non-fatal myocardial infarction, target vessel revascularization (TVR), need of coronary artery bypass grafting, cardiac death and all-cause mortality. Use of lipid-lowering agents (chiefly statins and fibrates) and non–lipid-lowering drugs (aspirin, clopidogrel, cilostazol, β-blockers, angiotensin-converting enzyme inhibitors, angiotensin receptor antagonists, aldactone receptor blockers) was compared between groups.

Continuous variables were expressed as mean ± standard deviation (SD). Normally distributed continuous data were compared between groups by unpaired student’s T-test, while non-parametric continuous data were compared by Mann–Whitney U test. Categorical variables were compared by χ² analysis with Fisher’s exact correction. Univariate followed by multivariate logistic regression analyses were used to identify independent correlates of in-hospital mortality in all patients. Rationale variables selected for regression analysis included the following: demographic characteristics, co-morbid illnesses, echocardiographic indicators, and angiographic features. Variables were tested in a backward conditioned multivariate logistic regression model if their univariate P values were <0.20. Statistical significance was defined as a multivariate P <0.05. The odds ratios and their 95% confidence intervals (CIs) from the multivariate logistic regression analysis were used as estimates of relative risk. Kaplan-Meier survival curves for components of MACE and overall MACE rate were constructed and compared between groups with the Log-Rank test. Multivariate Cox proportional hazard analysis was used to determine the independent predictors of TVR and overall MACE after adjustment for baseline and angiographic variables with unequal distribution. A p-value <0.05 was considered significant for all analyses. Statistical analysis was done using SPSS 11.5 or SAS 9.3.

The demographic data of both groups of patients are listed in Table 1. Patients in the lower-TG group were older (67 vs. 56 years, p < 0.01) and had lower body mass index, worse estimated glomerular filtration rate, higher high-density lipoprotein (HDL) as well as lower total cholesterol levels than those in the higher-TG group.

The territory of myocardial infarction determined by ECG was mostly located in the anterior wall in both groups (Table 2). The severity of overall coronary artery disease, the culprit lesion vessel, the ECG-to-balloon time and the therapeutic modalities of pPCI in terms of balloon angioplasty, thrombectomy, and endovascular stenting were similar between the two groups (p = NS in all). Success of PCI (post-procedural TIMI-blood flow to ≧grade 2) was accomplished in most of the patients in both groups (p = NS). Though post-procedural left ventricular ejection fraction estimated with either ventriculography or echocardiography was more depressed in the low-TG group (44.8% vs. 46.9%, p = 0.031), myocardial infarct size estimated by peak creatinine kinase (CK) levels was comparable between the two groups. Occurrence of new cardiogenic shock, respiratory failure, atrial fibrillation and ventricular arrhythmia, as well as requirement of emergency coronary bypass surgery was also statistically equivalent in the two groups (p = NS in all). Though in-hospital mortality happened in 6 patients in the lower TG (5 from ventricular arrhythmia and 1 from refractory pumping failure) but 0 in the higher TG group, the difference was not statistically significant (p = 0.098). Further univariate followed by multivariate regression analysis identified peak CK and CAD number >1 as positive, whereas serum TG level as negative predictors of in-hospital death for all of these patients (Table 3).

The medications prescribed at discharge were similar in the two groups of patients surviving the STEMI episodes, except that fibrates were given more often in higher-TG patients with the intention to lower serum TG level (Table 4). Kaplan-Meier survival test showed that during a mean follow-up period of 1.23 years for lower-TG and 1.4 years for higher-TG patients (p = 0.126), lower-TG patients had significantly more incidences of TVR (21.7% vs. 9.5%, Log-Rank p = 0.0111) and in turn overall MACE (26.1% vs. 11.9%, Log-Rank p = 0.0137) compared to higher-TG patients, yet the rates of de novo lesions, non-fatal MI, cardiac deaths and all-cause mortality were comparable between groups (Table 4). Multivariate Cox proportional hazard model confirmed that, besides the number of diseased coronary arteries as a positive predictor, serum TG level inversely correlated with TVR (hazard ratio 0.993, 95% CI 0.988-0.998, p = 0.007) and overall MACE (hazard ratio 0.994, 95% CI 0.990-0.999, p = 0.016) in all of our study patients (Table 5).

The role of serum TG in the pathogenesis of CAD is not as clear as that of serum LDL. A number of clinical studies reported a positive association between TG level and severity of CAD [7, 15‐17], whereas others demonstrated conflicting results [18‐20]. Investigations involving histological research found that triglycerides played a constitutive role in atherosclerotic plaques yet were rarely identified in active coronary artery lesions as a culprit component [21]. These findings suggest that triglycerides might be merely an innocent bystander rather than a direct atherogenic mediator [22‐24], and this concept may offer some explanation for our findings that patients with higher TG levels are not at risk of poorer in-hospital outcome after STEMI. In fact, our results demonstrate that there exists a paradoxical trend toward more in-hospital deaths in patients with lower TG levels, though the between-group difference did not reach statistical significance possibly due to limited patient number and confounding by coexistent variables. The role that serum TG plays as an independent negative predictor for in-hospital deaths, finally shown by multivariate regression analysis in this study, has similarly been reported in patients with acute stroke [11‐13] and acute coronary syndrome [25], but is not easy to understand compared to the roles played by CK level and CAD number as positive predictors, since these two variables represent surrogates of infarct size and ischemic myocardial territory thus are rationally related with in-hospital mortality. One possible explanation for the inverse relation between TG level and in-hospital mortality comes from the findings that incidences of new atrial arrhythmia, ventricular arrhythmia, respiratory failure and new cardiogenic shock all tended to be higher in the lower TG group, along with the exclusive occurrence of lethal ventricular arrhythmia and refractory heart failure in 6 (3.7%) of lower TG patients but not in any of higher TG group. These results implicate the potential role of serum TG in stabilizing the infarcted, stunned or reperfused myocardium after an acute STEMI episode. Another possibility may be based on the concept that serum TG level is an indicator of nutritional status. Lower TG levels denote poorer nutritional condition which could halt myocardial and whole body recovery from STEMI, in turn subjecting patients to in-hospital deaths. This hypothesis gains support from some epidemiologic studies describing low body mass index and serum cholesterol level as markers of poor health status in older subjects [26‐28], and from clinical investigations reporting worse outcome in patients with advanced heart failure or various cardiovascular diseases and low serum cholesterol [29, 30] or TG concentrations [11‐13]. Further studies are needed to determine the exact pathophysiological relations between serum TG level and in-hospital outcome in patients with STEMI who receive pPCI treatment.

Previous studies have demonstrated that even after primary PCI a considerable proportion of patients with STEMI still succumb to subsequent MACE due to left ventricular remodeling and coronary restenosis [31]. The distinct pathogenesis of restenosis caused by smooth muscle cell proliferation rather than lipid-related neo-atherosclerosis [32] explains the poor preventive efficacy of lipid-lowering medications, particularly statins, on inhibition of neointimal formation in this entity of lesions [33‐35]. Nonetheless, whether higher TG-rich lipoprotein is [36, 37] or is not [34] related with the restenotic phenomenon remains debatable. The current study demonstrated that risk of de novo atherosclerosis was similar between groups with higher or lower TG levels but overall MACE rates, contributed mostly from incidence of TVR, were significantly lower in high-TG patients. These results suggest a potential action of serum TG against neointimal proliferation at the STEMI-related lesion sites ever treated by angioplasty or stenting. This phenomenon has ever been reported in an in vitro study describing the distinct effects of various TG particles on either stimulation or suppression of the proliferative activity of vascular smooth muscle cells [38]. Our findings further indicate that though hypertriglyceridemia is regarded as an independent risk factor for atherosclerosis [39] when coexistent with low HDL- and elevated LDL-cholesterol and needs to be treated after the LDL goal is reached [25], application of such a therapeutic strategy to patients with STEMI post-pPCI is not warranted. On the contrary, the paradoxically inverse relation between hypertriglyceridemia and better late outcome in patients with STEMI treated with pPCI may justify exclusion of specific TG-lowering therapy from the standard therapy in long-term management of such patients. Whether there exists a “J-curve relationship” between serum TG level and the outcome of patients with STEMI treated with pPCI awaits further clarification.