Comparative efficacy of pharmacological agents on reducing the risk of major adverse cardiovascular events in the hypertriglyceridemia population: a network meta-analysis

Background Hypertriglyceridemia (HTG) is considered an independent risk factor for major adverse cardiovascular events (MACE). Methods This study analyzed the effects of various agents on MACE risk reduction in HTG (serum triglyceride ≥ 150 mg/dl) populations by performing a network meta-analysis. We performed a frequentist network meta-analysis to conduct direct and indirect comparisons of interventions. PubMed, EMBASE, and the Cochrane library were searched for trials until Jul 6, 2020. Randomized controlled trials that reported MACE associated with agents in entire HTG populations or in subgroups were included. The primary outcome was MACE. Results Of the 2005 articles screened, 21 trials including 56,471 patients were included in the analysis. The network meta-analysis results for MACE risk based on frequency data showed that eicosapentaenoic acid (EPA) (OR: 1.32; 95% CI 1.19–1.46), gemfibrozil (OR: 1.53; 95% CI 1.20–1.95), niacin plus clofibrate (OR: 2.00; 95% CI 1.23–3.25), pravastatin (OR: 1.32; 95% CI 1.15–1.52), simvastatin (OR: 2.38; 95% CI 1.55–3.66), and atorvastatin (OR: 0.55; 95% CI 0.37–0.82) significantly reduced the risk of MACE compared to the control conditions. In the subgroup analysis of HTG patients with triglycerides ≥ 200 mg/dL, bezafibrate (OR: 0.56; 95% CI 0.33–0.94), EPA (OR: 0.72; 95% CI 0.62–0.82), and pravastatin (OR: 1.33; 95% CI 1.01–1.75) significantly reduced the MACE risk. Conclusions Simvastatin had a clear advantage in reducing the risk of MACE in the entire HTG population analyzed in this meta-analysis. EPA, but not omega-3 fatty acid, was considered an effective HTG intervention. Among fibrates, gemfibrozil was most effective, though bezafibrate may significantly reduce the risk of MACE in populations with triglyceride levels of 200–300 mg/dL. Trial registration retrospectively registered in PROSPERO (CRD42020213705).

adverse cardiovascular events (MACE) [1,2]. According to Task Force recommendations, people with mild or moderate HTG (triglycerides between 150 mg/dl [1.69 mmol/L] and 999 mg/dl [11.3 mmol/L]) have an increased risk of cardiovascular disease, and people with severe HTG (triglycerides of > 1000 mg/dl [11.3 mmol/L]) have an increased risk of acute pancreatitis [3][4][5]. Realworld data based on the CANHEART cohort showed that HTG was very common in the atherosclerotic cardiovascular disease population, and elevated triglycerides (TGs) increased the risk of arteriosclerotic cardiovascular diseases (ASCVDs) [6].
Serum TGs are the main components of chylomicrons and very low-density lipoproteins (VLDLs) [7,8]. VLDLs and chylomicrons can permeate the arterial intima and selectively deposit, eventually causing the accumulation of cholesterol in the arterial intima and plaque formation [9]. In addition, elevated TGs have an important effect on coagulation and fibrinolysis, inhibiting fibrinolysis, increasing blood viscosity, and promoting thrombosis [10]. Therefore, HTG has a direct effect on atherosclerosis, increasing the risk of major cardiovascular events [11]. For mild to moderate HTG (177 mg/dl [2.0 mmol/L] to 885 mg/dl [10.0 mmol/L]), statins are considered the first-line drug. Although statins reduce low-density lipoprotein cholesterol (LDL-C) to a greater extent than they reduce TGs, a small reduction in TGs may further reduce the residual risk of cardiovascular disease, which indicates the cardiovascular risk among statin-treated individuals [12,13]. In addition to statins, lifestyle modifications, such as cessation of alcohol consumption, reduced intake of rapidly metabolized carbohydrates, weight loss, and blood sugar control are highly effective ways to lower TG levels [14]. For many HTG patients, a further reduction in TGs may reduce the risk of residual cardiovascular disease [15]. Therefore, guidelines recommend fibrates, niacin, and omega-3 fatty acids if HTG persists despite the application of high-intensity statin therapy [14] A previous meta-analysis showed that marine omega-3 fatty acids, especially high eicosapentaenoic acid (EPA)content agents, have obvious TG-lowering effects; moreover, TG reduction was associated with major vascular event risk reduction [16]. Omega-3 fatty acids are generally well tolerated [17]. In a study of a human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) population with HTG, the intake of EPA and docosahexaenoic acid (DHA) significantly reduced the risk of coronary heart disease (CHD) [18,19]. It is believed that fibrates can reduce the risk of cardiovascular disease in high-risk cardiovascular disease populations, such as patients with HTG or with atherogenic dyslipidemia (triglycerides > 150 mg/ dl [1.69 mmol/l] and high density lipoprotein cholesterol < 40 mg/dl [1.03 mmol/l] in men or < 50 mg/dl [1.29 mmol/l] in women) [20][21][22].
Currently, studies on the effects of various drugs on MACE outcomes in HTG populations are still insufficient; in many studies, HTG populations have been analyzed separately from major cardiovascular disease risk populations in only subgroup or post hoc analyses. This study will analyze the effects of various agents on reducing the MACE risk in HTG populations by a network meta-analysis to identify efficacious agents for clinical application.

Methods
This network meta-analysis of published randomized controlled trials (RCTs) was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses incorporating Network Meta-Analyses (PRISMA-NMA) statement.

Data sources, search strategy and selection criteria
Data from relevant studies were obtained by searching the PubMed, EMBASE, and Cochrane library databases without language restrictions. No initial date restriction was applied, and the search end date was Jul 6, 2020. Keywords included "hypertriglyceridemia", "high triglyceride", "high triacylglycerols", "high triacylglycerides", "survival", "survivors", "death", "die", "mortality", "cardiovascular", "MACE", "random*", "randomized", and "randomized". The search strategy details were list in Additional file 1: Table S1. Reference lists from identified trials and reviews were manually screened to identify additional trials.

Inclusion and exclusion criteria
The inclusion criteria included the following: 1, a RCT design; 2, the inclusion of an HTG population; 3 a separate analysis of the HTG population in subgroup or post hoc analyses; and 3, the reporting of MACE-related outcomes. The exclusion criteria included the following: 1, studies without the inclusion of an HTG population; 2, studies that did not report the HTG population separately; 3, dosage-related studies; 4, lifestyle behavior or nursing-related intervention studies; or 5, studies that did not report MACE-related outcomes. Additionally, studies reporting zero MACE events in both groups were excluded because of small sample sizes or short followup periods. Conference summaries, editorial comments, and comments were also excluded.

Data extraction and study quality assessment
Two authors independently extracted relevant information from the included studies. The extracted content included the name of the first author or collaborative organization, publication time, sample size, average age of the population, study abbreviations, intervention and control agents, outcome assessment, and follow-up period. MACE outcomes mainly included sudden cardiac death (defined as the unexpected death of an individual not attributable to an extracardiac cause, usually within one hour of symptom onset; it's a consequence of many cardiovascular conditions, and the leading cause of SCD is coronary heart disease (CHD), which accounts for over 70% of SCD cases) [23,24], fatal/nonfatal myocardial infarction (defined as a condition where there is interruption of blood supply to a part of the heart that led to a fatal or severe nonfatal event), and nonfatal stroke (defined as non-convulsive loss of neurological function due to brain ischemia or intracranial hemorrhages that led to a fatal or severe nonfatal event). Individual study definitions were used for the outcome analysis (see Additional file 2: Table S2 for details). The potential risk of bias of the included RCTs was evaluated using the risk assessment tool recommended by the Cochrane Collaboration guidelines (assessing random sequence generation; allocation concealment; blinding of participants and personnel; blinding of outcome assessment; incomplete outcome data; selective reporting; and other risk) [25].

Statistical analysis
Dichotomous data were combined to produce odds ratios (ORs), and 95% confidence intervals (CIs). Hazard ratios (HRs) and 95% CIs for Cox regression results were also calculated if raw event frequency data was not reported. A frequentist framework network meta-analysis was performed for mixed multiple treatment comparisons that combined direct and indirect evidence to obtain effect size [26]. Cochran's Q test was used to assess homogeneity in the whole network and the homogeneity/consistency between designs. The pairwise comparison results based on the network meta-analysis were obtained, and the effective ranking of various agents was performed based on the P-score [27]. A comparison-adjusted funnel plot was used to assess funnel plot asymmetry for the network meta-analysis. A p value less than 0.05 was considered to indicate a significant difference. The analyses were performed using the "netmeta" package in R language (version 4.0.2).

Literature screening
Through public database searching, 695 studies were obtained from PubMed, 1661 studies were obtained from EMBASE, and 654 studies were obtained from the Cochrane library. After removing duplications, 2005 studies were obtained. A total of 1838 trials were excluded by title and abstract screening. A total of 167 full-text articles were reviewed. Additional studies were excluded on the basis of the following: the studies did not report the HTG population in their entirety or separately (55); the studies did not report MACE-related outcomes or reported no events in either arm (31); the studies were duplicated (18); the studies were reviews (17); the studies were conference summaries (13); the studies had non-RCT designs (5); the studies evaluated lifestyle behavior or nursing interventions (4); the studies were protocols (2); and the studies were dosage-related studies (1). For included post hoc studies, the latest reported results were included in the analysis. Finally, a total of 21 articles including 56,471 HTG patients were included in the analysis ( Fig. 1) .
Included studies were published from 1988 to 2019. In terms of sample size, with the exception of one study that included only 59 patients [42], the studies included more than 300 patients. The average age of the population was approximately 60 years old, except in one article in which the average age was approximately 80 years old [29]. The shortest follow-up period among the included studies was 1 year [42] (Table 1). In most studies, the most common MACE were cardiovascular death, nonfatal myocardial infarction (MI), nonfatal stroke, acute coronary syndrome, and coronary revascularization. However, there were studies that reported only death events [31,34,35,42,47] (Additional file 2: Table S2). Regarding the research quality assessment, only two articles did not state their blinding method [40,47], and two articles performed only blinding of the assessors [29,37]. The rest of the study designs were of high quality. Thus, the overall quality of included studies was relatively satisfactory (Fig. 2).
The studies with MACE frequency data were analyzed first. The interventions included atorvastatin, bezafibrate, EPA, ezetimibe, omega-3 fatty acid, fenofibrate, gemfibrozil, niacin, niacin plus clofibrate, niacin plus laropiprant, pravastatin, simvastatin, and control (Fig. 3, A). Due to the included trials comparing interventions and control directly head to head, as well as the Q test results (p = 0.744), this study used consistency fixed-effect models. In pairwise comparisons, compared to control, EPA  Table 2). In omega-3 fatty acid research, due to the smaller number of patients and zero-events in the intervention arm [42], the accuracy of the omega-3 fatty acid results was low with a wide 95% CI. P-score ranking results showed that simvastatin (0.899), niacin plus clofibrate (0.812), and atorvastatin (0.767) have relative advantages. A comparison-adjusted funnel plot showed no publication bias (Fig. 4a). This study also conducted an analysis based on the HR results of the Cox regression analysis as a supplement to the results of the event frequency analysis because several studies did not report event frequency results but did report HR results [35,36,38]. The interventions included atorvastatin, EPA, ezetimibe, omega-3 fatty acids, fenofibrate, niacin, simvastatin and controls (Fig. 3b) (Table 3). P-score ranking showed that simvastatin (0.979), ezetimibe (0.739), and EPA (0.690) had relative advantages. A comparison-adjusted funnel plot showed no publication bias (Fig. 4b).
We further classified the agents for analysis. The classifications included omega-3 fatty acids, fibrates, niacin, niacin plus fibrates, niacin plus laropiprant, selective cholesterol absorption inhibitors (SCAIs), and statins (Fig. 3c). Of these, statins are inhibitors of the hydroxymethylglutaryl-CoA (HMG-CoA) reductase enzyme to interfere with the endogenous synthesis of cholesterol. Ezetimibe inhibits the Niemann-Pick C1-Like 1 (NPC1L1) receptor to block intestinal cholesterol absorption. Omega-3 fatty acids are polyunsaturated fatty acids that regulate lipid metabolism by inhibiting lipogenesis and promoting lipolysis and are also involved in inflammatory regulation. Fibrates are a class of   Tables 4 and 5).

Discussion
TG level was an independent risk factor for cardiovascular events, despite adjustment for the LDL-C level [49]. Currently, there is a lack of meta-analyses of the effects of various agents on the MACE risk reduction in the HTG population. This study analyzed this topic by network meta-analysis. Statins, niacin, fibrates, omega-3 fatty acids, and ezetimibe were included in the analysis. It was found that simvastatin is superior to other agents for MACE risk reduction, and niacin plus clofibrate, EPA, and gemfibrozil also have relative advantages. The subgroup analysis of the population   13:15 with TGs ≥ 200 mg/dL showed that bezafibrate and omega-3 fatty acids have relative advantages.
Niacin plus clofibrate has relative advantages in reducing the MACE risk in HTG patients based on a study published in 1988 [47]. Clofibrate is a fibric acid derivative used as a hypertriglyceridemia therapy [50]. However, clofibrate was withdrawn from the market in 2002 because of concern about its side effects, including elevated transaminase and rare acute liver injury [51]. In the analysis based on HR results, ezetimibe was found to    13:15 reduce the risk of MACE, but the results were not significant. In the original study, ezetimibe reduced the risk of cardiovascular events in people with elevated LDL-C and those ≥ 75 years old. For those with LDL-C but without HTG, the effect of ezetimibe was obvious (HR: 0.65; 95% CI 0.47-0.91; p = 0.011) [29]. A systematic review also concluded that the combination of a statin and ezetimibe is more effective for lowering LDL-C levels than for lowering TG levels [52].
Two studies, the REDUCE-IT [28] and JELIS [37] studies, included EPA interventions. According to the results of these studies, EPA significantly reduced the risk of MACE compared to the control. However, omega-3 fatty acids showed no benefit in reducing the risk of MACE [35]. The results based on event frequency were also inaccurate due to zero events in the intervention arm [42]. However, it is generally believed that the effect of omega-3 fatty acids is inferior to that of EPA along, and DHA may cause an increase in LDL-C levels. This might be the reason why the Epanova (omega-3-carboxylic acids) study (NCT02104817) was discontinued; it had only limited benefits in mixed dyslipidemia patients [53]. In general, EPA is still considered an effective HTG treatment.
Fibrates including gemfibrozil, fenofibrate, and bezafibrate mainly act to clinically reduce TGs. Our results showed that gemfibrozil had a significant effect on reducing the risk of MACE compared with the control. However, its clinical application rate has gradually decreased, mainly due to muscle-related and blood-related side effects [54]. Therefore, gemfibrozil was mostly used for very high TG level intervention [55]. In the BIP study, bezafibrate was believed to be effective in those with TGs ≥ 200 mg/dL but not in those with TGs ≥ 150 mg/ dL in the multivariate analysis after adjusting for age, previous MI, use of non-study-related lipid-lowering medication, and diabetes mellitus [31]. It is possible that in moderate HTG patients, bezafibrate has good TG-lowering effects that further reduce the MACE  .084) 1.32 (1.19-1.46) 1.21 (1.05-1.40) 1.04 (0.76-1.41) 2.00 (1.23-3.25) 1.08 (0.94-1.24) 1.56 (0.95-2.56) 1.44 (1.27-1.63  risk. A post hoc study also showed that among patients with TGs ≥ 200 mg/dL, those whose TGs decreased by > 0.5 mmol/L after intervention attained a greater effect in reducing their MACE risk [56]. However, compared with those in other studies, the patients in the BIP study had a lower TG upper limit (TG ≤ 300 mg/dL), so the benefit of bezafibrate in those with TGs ≥ 300 mg/ dL is still unclear. Furthermore, the value of fenofibrate might have been under-estimated by population selection. In dyslipidemic diabetic patients, fenofibrate may be effective in reducing MACE risk [30,38]. Pemafibrate is also a novel, highly selective peroxisome proliferatoractivated receptor (PPAR)-α modulator (SPPARM) that can modulate lipid metabolism to decrease plasma triglyceride levels and increase high-density lipoprotein cholesterol levels [57,58]. It has been approved in Japan for hyperlipidemia treatment [59]. However, MACE results for RCTs investigating pemafibrate were not completed at the initiation of this study, so pemafibrate was not analyzed in this work [60].
It should be noted that nonstatin intervention studies did not restrict the administration of statins in either arm. In some studies, all patients in both arms were treated with statins [32,34]. Therefore, even if agents such as fibrates or EPA show benefits, the use of statins as first-line drugs will not change. Especially in comparisons with control groups, simvastatin showed obvious advantages.
In addition, familial hypertriglyceridemia (FHTG) is a familial autosomal dominant disease characterized by excessive triglyceride and VLDLs and is often found with polygenic genetic characteristics [61]. Due to insufficient investigation of the familial aggregation of HTG and the similar treatment strategy for reducing serum TG, there is still a lack of RCTs about FHTG. Thus, this work did not distinguish whether HTG in patients was familial or nonfamilial.
There were some limitations in this study. This study did not analyze the influence of patient characteristics and accompanying treatment on the results. Therefore, our study analyzed the effects of different agents in HTG populations. For patients with specific concomitant diseases, to account for accompanying treatment or type of dyslipidemia, more targeted studies are needed. In addition, this meta-analysis did not include studies analyzing lifestyle behavior interventions, although it is widely believed that good dietary and exercise habits are beneficial to TG reduction. Although the MACE criteria within individual studies were consistent, there were differences among the included studies. This study analyzed only the effects of various agents on reducing MACE risk in the HTG population, but side effects should be considered in clinical application when choosing personalized treatment regimens.

Conclusions
Overall, among the statins, simvastatin maintained a clear advantage in reducing the risk of MACE in the HTG population. EPA, but not omega-3 fatty acid, was considered an effective HTG intervention. Among the fibrates, gemfibrozil was most effective, and bezafibrate may significantly reduce the risk of MACE in populations with TG levels of 200-300 mg/dL.