Impact of Metabolic Syndrome and It's Components on Prognosis in Patients With Cardiovascular Diseases: A Meta-Analysis

Background: Patients with metabolic syndrome (MetS) have a higher risk of developing cardiovascular diseases (CVD). However, controversy exists about the impact of MetS on the prognosis of patients with CVD. Methods: Pubmed, Cochrane library, and EMBASE databases were searched. Cohort Studies and randomized controlled trials post hoc analyses that evaluated the impact of MetS on prognosis in patients (≥18 years) with CVD were included. Relative risk (RR), hazard rate (HR) and 95% confidence intervals (CIs) were calculated for each individual study by random-effect model. Subgroup analysis and meta-regression analysis was performed to explore the heterogeneity. Results: 55 studies with 16,2450 patients were included. Compared to patients without MetS, the MetS was associated with higher all-cause death [RR, 1.220, 95% CI (1.103 to 1.349), P, 0.000], CV death [RR, 1.360, 95% CI (1.152 to 1.606), P, 0.000], Myocardial Infarction [RR, 1.460, 95% CI (1.242 to 1.716), P, 0.000], stroke [RR, 1.435, 95% CI (1.131 to 1.820), P, 0.000]. Lower high-density lipoproteins (40/50) significantly increased the risk of all-cause death and CV death. Elevated fasting plasma glucose (FPG) (>100 mg/dl) was associated with an increased risk of all-cause death, while a higher body mass index (BMI>25 kg/m2) was related to a reduced risk of all-cause death. Conclusions: MetS increased the risk of cardiovascular-related adverse events among patients with CVD. For MetS components, there was an increased risk in people with low HDL-C and FPG>100 mg/dl. Positive measures should be implemented timely for patients with CVD after the diagnosis of MetS, strengthen the prevention and treatment of hyperglycemia and hyperlipidemia.


Impact of Metabolic Syndrome and It's Components on Prognosis in Patients With Cardiovascular
Diseases: A Meta-Analysis INTRODUCTION Cardiovascular disease (CVD) has attracted worldwide attention and accounts for 46.2% of deaths from non-communicable diseases (1). CVD is one of the main causes of premature death and disability. Metabolic syndrome (MetS), including dysglycemia, obesity (especially central obesity), high blood pressure, low high-density lipoprotein cholesterol (HDL-C), and elevated triglyceride levels, is a complex of risk factors for type 2 diabetes and CVD (2). Patients with MetS have a higher risk of developing CVD compared with those without MetS in the next 5-10 years, and the long-term risk is even higher (3). The National Cholesterol Education Program (NCEP) Adult Treatment Panel (ATP) III criteria also considered MetS as the second major target for CVD prevention (4).
The prevalence of MetS is higher in patients with CVD than in patients without MetS. The prevalence of MetS in hospitalized patients with acute myocardial infarction (AMI) is 46%, similar to that of the acute coronary syndrome (43.4%) (5); This finding indicates that MetS is associated with CVD. Boulon et al. (6) suggested that despite active management, patients with MetS have a higher long-term risk of cardiovascular events (6). However, Selcuk et al. (7) suggested that the main determinant of long-term prognosis of AMI is heart failure rather than metabolic disorder (7). But some researchers suggested that MetS does not increase the mortality among patients with CVD (8). Therefore, controversy exists about the impact of MetS on patients with CVD.
MetS is a disease associated with multiple factors, and the main diagnostic indicators (components) include blood pressure, overweight and obesity, HDL-C, and fasting blood glucose (9,10). Most studies have focused on the overall effect of MetS on the prognosis of CVD. However, whether a correlation exists between each component and prognosis and which factor is more important have not been elucidated. Considering these inconsistencies, we performed a meta-analysis of cohort studies and RCT post-hoc analysis from CVD patients to evaluate associations between different definitions of MetS and the risk of all cause death, CV death and cardiovascular events.

METHODS
The study was registered with PROSPERO (CRD42021147609), and reported in accordance with the PRISMA statement (11).

Eligible Studies
(1) Influencing factors and study types: Studies that evaluated the influence of MetS and its components on patients with CVD were included. We included cohort and randomized controlled trials post hoc analyses and excluded single-group observational studies. (2) Types of patients: Patients with CVD were aged ≥18. (3) Outcomes: Primary outcomes were all-cause death, cardiovascular (CV) death, incidence of MI and stroke. Secondary outcomes were TVR, heart failure, cardiac arrest, angina pectoris, cardiogenic shock. All-cause death of high TG, low HDL-C, high BP, FPG>100 mg/dl, BMI>25kg/m 2 , high WC. CV death of high TG, low HDL-C, high BP, FPG>100 mg/dl, BMI>25kg/m 2 .
The definition of cardiovascular disease in this metaanalysis was history (comorbidity) of cardiovascular or cardiac disease. Hypertension/Cardiovascular Infections/Cardiovascular Abnormalities/Pregnancy Cardiovascular Complications/cardiomyopathy in specific terms was excluded because these diseases often overlap and potentially result in overestimation of cases.

Exclusion Criteria
(1) Studies that had incomplete or unavailable original data. (2) The diagnostic criteria for MetS were not specified. (3) Repeated published data. (4) Studies that evaluated the relationship between MetS and congenital heart disease.

Study Selection
Endnote X9 was used to manage and screen the literature. Title, abstract, and full texts were selected based on inclusion/exclusion criteria. We designed a standardized form to extract data including study characteristics, diagnostic criteria, characteristics of the study population, risk of bias, and outcome measures.

Risk of Bias Analysis
We used the Newcastle-Ottawa Scale (NOS) to assess the quality of the cohort studies (12). To be specific, studies with scores >7 were treated as high quality, 4-6 as medium quality, and below 4 as low quality (13). Cochrane Collaboration's tool for assessing the risk of bias was applied to determine the quality of the included RCT post-hoc studies (12,14). Two researchers (X Li, YJ Zhai) independently screened and extracted the data, and a third researcher (J Lyu) resolved any disagreements. Quality evaluation results are reported in Supplementary Table 2.
The diagnostic criteria for MetS vary among different regions and institutions, but the majority of them included central obesity, hypertension, low HDL-C, and high TG and fasting blood glucose (FBG) levels. Other diagnostic criteria also included dyslipidemia, chronic mild inflammation, endothelial dysfunction, insulin resistance, increased oxidative stress. The diagnostic criteria used in the included studies were NCEP2001 criteria (9), NCEP2005 criteria (4), and The International Diabetes Federation (IDF) criteria (10) (details reported in Supplementary Table 3). For specific diagnostic criteria, we compared the above criteria and divide into subgroups based on the comparison results.

Statistical Analysis
Statistical analysis was performed using STATA 13 and R software. For dichotomous outcomes (all-cause death, CV-death, the incidence of MI, stroke, TVR, heart failure, cardiac arrest, angina pectoris, and cardiogenic shock), relative risk (RR) and 95% confidence intervals (CIs) were calculated for each individual study. For the impact of MetS components on patients with CVD (all-cause death and CV death), hazard rate (HR) and 95% confidence intervals (CIs) were determined for each study. The heterogeneity across studies was examined using the Chi-square test and I-square statistics. The results were pooled by the D-L random-effect model due to the large statistical heterogeneity among the studies.
To explore the sources of clinical heterogeneity and methodological heterogeneity, we performed subgroup analysis based on the following: (1) diagnostic criteria, studies were divided into four subgroups (NCEP2001, NCEP2005, IDF and "others") and (2) study type, studies were divided into three subgroups (prospective cohort study, retrospective cohort study, and RCT post-hoc study). Meta-regression analysis of three covariates (follow-up time, male proportion, and patient age) was performed to explore the size and source of heterogeneity.
Effect measures [risk ratio (RR) vs. odds ratio (OR) vs. risk difference (RD)] and statistical models (D-L random-effects model vs. M-H fix-effects model) were used to examine the robustness of the results. We evaluated publication bias by Begg's tests and drew contour-enhanced funnel plots to assess whether the asymmetry of the funnel plots was caused by publication bias or other biases.

Overview of the Characteristics of the Studies
A total of 5,028 unique records were identified from the literature search. After excluding 226 duplicate articles, 125 studies were initially included by reading the title and abstract. Fifty-five studies were finally included after further reading the full text, including six RCT post-hoc studies (15)(16)(17)(18)(19)(20) and 49 cohort studies (3, 5-8, 21-64) (Figure 1).

Study Characteristics
A total of 162,450 patients from 25 countries and regions were included, the sample size for each individual study varies from 57 to 44 548. Forty-one studies (145,390 patients) evaluated the risk of all-cause death among patients with CVD and MetS. Twentyone studies with 95,049 patients reported CV death, 23 studies with 77,618 patients reported the incidence of MI, and 11 studies with 59,770 patients reported the incidence of stroke.
Twenty-six studies adopted NCEP-ATPIII (2005) criteria, 21 studies mainly adopted NCEP-ATPIII (2001) criteria, and 7    Table 2). The cohort studies comprised 16 medium-quality studies, and 33 high-quality studies. For RCT post-hoc studies, the risk of bias was deemed low in 2 studies and moderate in 4 studies.  Table 3). Among different study types, the cohort study subgroup was in the same direction with the overall results. No statistically significant difference was found in the RCT post-hoc studies. Diagnostic criteria and study type were the factors that affected heterogeneity. Meta-regression showed that the follow-up time and male proportion were not the sources of heterogeneity (P > 0.05), and age only explained 1.6% of the heterogeneity (P = 0.022). The Begg's test result showed bias (P = 0.012), and the contour-enhanced funnel plots showed that the bias may be due to other reasons rather than publication bias. Twenty-one studies with 94,542 patients reported CV-related death. The MetS group had higher CV death than the non-MetS group [RR = 1.360, 95% CI (1.152, 1.606), P = 0.000] according to the heterogeneity test I 2 =87.0% (Table 2, Figure 3). Subgroup analysis showed that among different diagnostic criteria of MetS, NCEP-ATPIII (2001) and NCEP-ATPIII (2005) subgroups were consistent with the overall result ( Table 3). Among different study types, the subgroups were consistent with the overall results. Diagnostic criteria affected the heterogeneity. Meta-regression showed that follow-up time, age, and male proportion were not the source of heterogeneity (P > 0.05). The Begg's test and the contour-enhanced funnel plots showed that bias may be caused by publication bias and other reasons.  Figure 4). Subgroup analysis showed that among the diagnostic criteria of MetS, the results of NCEP-ATPIII (2001) and NCEP-ATPIII (2005) were consistent with the overall results ( Table 3). Other subgroups had no statistically significant difference. Among the study types, the subgroup results were in the same direction as the overall results. Meta-regression showed that follow-up time, age, and male proportion were not the source of heterogeneity (P > 0.05). The Begg's test and the contour-enhanced funnel plots reported no publication bias (P = 0.125).

Risk of MI and Stroke
Eleven studies with 60,297 patients reported the risk of stroke. Patients with CVD and MetS had a higher risk of stroke [RR = 1.435, 95% CI (1.131, 1.820), P = 0.000] according to the heterogeneity test I 2 = 75% (Table 3, Figure 5). The Begg's test and the contour-enhanced funnel plots showed that the bias may be caused by other reasons rather than publication bias.

Other Adverse Cardiovascular Indicators
The results of the TVR (13 studies) reported that patients with CVD and MetS had a higher risk to develop TVR [RR = 1.241, 95% CI (1.063, 1.448), P = 0.000]. Subgroup analysis showed that diagnostic criteria and study type explained the partial heterogeneity. The risk of heart failure was evaluated in eight studies. Patients with CVD and MetS were more likely to have heart failure [RR = 1.497, 95% CI (1.116, 2.007), P = 0.000]. Subgroup analysis showed that diagnostic criteria partly explained the heterogeneity ( Table 2).
Other indicators include risk of cardiac arrest (4 studies), angina pectoris (3 studies), and cardiogenic shock (3 studies). We found no statistically significant difference in the risk of cardiac arrest [RR = 1.

Impact of MetS Component
Among MetS components, low HDL (40/50) was significantly associated with increased risks of all-cause death and CV death. Elevated FPG (>100 mg/dl) was significantly associated with an increased risk of all-cause death, whereas body mass index (BMI) > 25 kg/m 2 was related to a reduced risk of all-cause death ( Table 4).

Sensitivity Analysis
We examined the robustness of our results. The sensitivity analysis of the effect measures showed that the OR increased the effect size and did not change the direction of the results, except for angina pectoris. The RD did not change the direction of the results. The sensitivity analysis of the statistical models did not change the direction of the results. Hence, the results of this meta-analysis were robust.

Summary of Main Results
Fifty-five studies with 162,450 patients from 25 countries or regions were included. Most studies defined MetS using NCEP2001, NCEP2005, and IDF criteria, and other works   adopted specific diagnostic criteria. Our results suggested that patients with CVD and MetS had an increased risk of all-cause death, CV-related death, MI, stroke, TVR, and heart failure. In the analysis of MS components, BMI>25 kg/m 2 was negatively correlated with the prognosis of patients with CVD. Dyslipidemia and abnormal glucose metabolism were the main risk factors for the prognosis of CVD. Different spectrum within patients with cardiovascular diseases may be the sources of heterogeneity.

Potential Biases in the Review Process
MetS and its components are a complex of risk factors for CVD and diabetes (21). Ford (65) reported that the population attributable fractions for CVD, diabetes, and all-cause death among patients with MS were 12-17%, 30-52%, and 6-7%, respectively (65). However, for patients with CVD, whether MetS and its components is associated with the risk of CV events remains controversial. Obesity is an independent risk factor for hypertension, CVD, and diabetes (66). Given the known association between obesity and CVD, the adverse consequences of obesity may persist after the onset of CVD. However, previous studies suggest a contradictory U-shaped relationship between obesity and CVDrelated death; hence, overweight and mild obesity are related to lower short-term and long-term mortality (67-69) based on the concept of "Obesity paradox" or "reverse epidemiology" (66). Although the setting of obesity indicators was involved in different MetS diagnostic criteria, the core of the diagnosis was consistent. In   Hence, we need to reconsider which diagnostic criteria can predict the prognosis of MetS among patients with CVD more accurately. The heterogeneity in this study may be associated with the proportion of obese patients included. AHA/NHLBI 2009 diagnostic criteria were not adopted in all of the included studies, which may be related to the fact that the indicators and numerical intervals of abdominal obesity were not clearly given in the criteria. BMI was used in most of the studies as a proxy for waist circumference, but the cutoffs for the inclusion criteria in each study were different. This phenomenon may be related to two factors: (1) BMI is easier to obtain than waist circumference, and (2) BMI can be effectively docked with the WHO's definition of obesity. However, existing evidence suggests that MetS might be caused by excessive central obesity (70). Therefore, in future research on MetS, we suggest that BMI and waist circumference data should be collected at the same time for strict implementation of MetS diagnostic criteria.

Impact of Follow-Up Time on Results
The span of follow-up time included was very large, ranging from 0.33 years to 12.6 years. A 32-year prospective cohort study of male residents without MI or stroke in the community showed that the CV-related mortality curves among patients with MetS varied at 10-15 years of follow-up (70). The findings of Kasai et al. (26) and Nigam et al. (46) show that MetS and its components had a significantly positive association with allcause death of patients with CVD during 4-5 years of follow-up (20,26). However, the impact of MetS on patients with CVD might be underestimated in these studies.

MS Components of Study
In this study, the potential influences of the five components of MetS [TG, HDL, BP, FPG, BMI/Waist circumference (WC)] on CVD prognosis was analyzed. We found that abnormal blood glucose and lipid metabolism are important factors that could lead to poor prognosis of CVD. As such, these factors should be considered as intervention targets for predicting the prognosis of patients with CVD. BMI was negatively correlated, which was manifested as the obesity paradox. Waist circumference was included in only two studies with relatively small sample sizes and conducted among Chinese patients only. Further studies are needed to explore the rationality, applicability, and the risk prognosis of BMI and waist circumference.
Prediabetes is an intermediate metabolic state between normoglycemia and diabetes, includes impaired glucose tolerance and impaired fasting glucose (71). Compared with NCEP-ATP III (2001) criteria, the NCEP-ATP III (2005) reduced the fasting plasma glucose from 6.1 mmol/L to 5.6 mmol/L. Our results showed that the two diagnostic criteria had the same contribution in predicting the prognosis of patients with CVD. The results of Huang 2016 also found that prediabetes with impaired fasting glucose or impaired glucose tolerance is associated with an increased risk of composite cardiovascular events, coronary heart disease, stroke, and all-cause mortality (71). Our findings indirectly supported the modification of the American ADA guidelines to reduce the standard of pre-diabetes from 6.1 mmol/L (72) to 5.6 mmol/L (73). In response to this result, lifestyle intervention is the fundamental management approach for prediabetes (73,74).

Limitations
The span of follow-up time included was very large, ranging from 0.33 to 12.6 years, most studies were followed up for <5 years, the impact of MetS on patients with CVD in this study might be underestimated. As one of the diagnostic indicators of MetS, WC was only included in two studies, reflected the problems in the implementation of MetS diagnostic criteria and possibly underestimated the impact of central obesity on patients with CVD.

CONCLUSIONS
This meta-analysis was conducted using cohort studies and RCT post-hoc studies. MetS was found to be associated with an increased risk of CV-related adverse events among patients with CVD. For MetS components, there was an increased risk in people with low HDL-C and FPG>100 mg/dl. Positive measures should be implemented timely for patients with CVD after the diagnosis of MetS to reduce risk factors and strengthen the prevention and treatment of hyperglycemia and hyperlipidemia. Further studies need to clarify the selection of MetS diagnostic indicators (particularly the BMI or waist circumference).

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
JL, AX, XL, and YZ conceived the study. YZ and HH designed the search strategy and XL performed the literature search. JZ and YZ screened studies for eligibility. HH, YLi, YLiu, and AF performed data extraction. XL, YZ, LL, and TH assessed the risk of bias. XL, YZ, HH, and JZ performed data analysis. JL interpreted the data analysis and assessed the certainty of evidence. XL and YZ wrote the first draft of the manuscript and all other authors revised the manuscript. All authors contributed to the article and approved the submitted version.