In the global context, the prevalence of metabolic syndrome (MetS) had ranged from 20 to 25% and exhibited gradually increased over the past several years (1, 2). MetS had been associated with an elevated risk of developing various illnesses, including a 29% increase in the risk of major coronary events, a 30% increase in the risk of recurrent myocardial infarction, and a 12% increase in the risk of dementia (3, 4). Meanwhile, MetS had been linked to a rising mortality rate, with a 24% increase in all-cause mortality and a 44% increase in mortality due to heart disease (5).

The level of education, as an indicator of socioeconomic status, has the potential to impact both the prevalence of cardiovascular disease and the prevalence of MetS (6, 7). Individuals with a college education demonstrated a 57% lower risk of developing MetS than those with only a primary school education (8). Recent research findings had indicated that while several studies had explored the association between education and MetS, a definitive causal relationship between the two had not been found.

Mendelian randomization (MR) was initially introduced by Katan in 1986, utilizing genome-wide association study(GWAS) data to employ single nucleotide polymorphisms(SNPs) as instrumental variables(IVs) for investigating causal relationships between exposure and outcome (9). This method distinguishes itself from others by its independence from confounding variables and ability to avoid reverse causation (10).

This study utilized a 2-sample MR method to estimate the association between education and MetS.

Among the 418 education-associated variants, 3 SNPs were not available for MetS; one SNP was not available for WC, hypertension or HDL-C; and 76 SNPs were not available for FBG, which is a no missing variant in triglycerides. Palindromic SNPs were excluded from the analysis. A total of 404 SNPs were utilized as genetic instruments for MetS, 356 for WC, 406 for hypertension, 333 for FBG, 407 for TG and 356 for HDL-C in this study. The R² and F statistics were calculated to assess the efficacy of the genetic instruments (Table 1).

The results were shown in Table 2 and Figure 1. The scatter plots were presented in Supplementary Figure S3. The forest plots were presented in Supplementary Figure S4. The results of the IVW analysis [odds ratio (OR) = 0.55, 95% confidence interval (CI) = 0.48–0.63, p = 2.18E−51], MR–Egger regression (OR = 0.70, 95% CI = 0.41–0.99, p = 1.66E−2) and weighted median analysis (OR = 0.58, 95% CI = 0.49–0.67, p = 6.28E−34) indicated each one-s.d.-higher education level was negatively association with MetS (Table 2). The IVW analysis yielded Cochran’s Q test p values (p < 0.001), suggesting substantial heterogeneity in the results. However, the funnel plot did not display asymmetry (Supplementary Figure S1), indicating no evidence of publication bias. Furthermore, MR–Egger intercept values (intercept = −0.003; p = 0.098) indicate the absence of significant horizontal pleiotropy in this study. Furthermore, the LOO analysis demonstrated no significant changes upon the removal of any IVs (Supplementary Figure S2).

In terms of its constituent elements, each one-s.d.-higher education level was negatively correlated with WC, hypertension, FBG, and triglycerides but positively correlated with HDL-C. In the IVW model, the ORs with 95% CIs per log-odds increase in genetically predicted education level were 0.80 (95% CI: 0.76–0.83; p = 4.98E–33) for WC, 0.96 (95% CI: 0.95–0.97; p = 4.54E–10) for hypertension, 0.94 (95% CI: 0.91–0.97; p = 7.58E–6) for FBG, 0.83 (95% CI: 0.79–0.87; p = 7.87E–18) for triglycerides, and 1.22 (95% CI, 1.18–1.25; p = 1. 45E–31) for HDL-C (see Table 2). The findings were consistent across the various models. Potential horizontal pleiotropy was not found in the IVW model (all p values >0.05), as shown in Table 2. The results of the Cochran Q test revealed significant heterogeneity in WC, hypertension, TG, and HDL-C, but not FBG. Additionally, analysis of the funnel plot (Supplementary Figure S1) did not detect any visible asymmetry, suggesting a lack of evidence for heterogeneity. Furthermore, the stability of the results remained consistent even after the exclusion of any single SNP through LOO analysis. (Supplementary Figure S2).

This is the first article to elucidate the causal relationship between education and metabolic syndrome, using Mendelian randomization method. In this 2-sample MR study, each addition 4.2 years of education decreased 45% risk of MetS, had a negative association with WC, hypertension, FBG, and TG, a positive association with HDL-C.

Higher education levels are linked to improved health habits and lower rates of MetS, as well as a decreased risk of cardiovascular and cerebrovascular diseases such as heart disease and stroke (18–20). A cross-sectional analysis revealed a negative association between education and MetS (21), which is consistent with the findings of our own study. A study conducted in Korea demonstrated that education level has both direct and indirect impacts on MetS, with higher levels of education indirectly influencing food choices to decrease the prevalence of MetS. Specifically, individuals with higher education levels were more likely to adopt healthier food habits, such as consuming fruits, vegetables, and milk, with this effect being more pronounced in females (21). Previous research has revealed sex differences in the onset and progression of MetS. Interestingly, these differences also extend to the association between education level and MetS, with a more pronounced effect observed in females (8). A study focusing on a cohort of middle-aged women revealed that higher education levels were associated with a decreased prevalence of MetS. Specifically, individuals who had completed college or higher were 2.7 times more likely to have MetS (22). Prior research has generally posited that education is among the factors influencing metabolic syndrome. However, our study indicates that higher levels of education are associated with a reduced incidence of metabolic syndrome.

As the most important component of MetS, WC can also be influenced by education level. In the European cohort, higher levels of education were found to have a significant effect on reducing WC (23). A cross-sectional study conducted on elderly adults in Taipei further supports this finding, showing a positive correlation between obesity and lower education levels, particularly among females, with no significant association with income level (24). These results are consistent with our own study, which revealed a negative relationship between years of education and WC. In a prior study conducted by Jose P. Lopez-Lopez et al., the prevalence of hypertension was significantly higher among populations of African origin than among other populations. Additionally, the study identified an association between the risk of hypertension and lower levels of education (25). Another MR study also identified a lower level of education as a risk factor for hypertension (26). This finding is consistent with the results of our study. A study conducted by Braverman-Bronstein et al. demonstrated that the association between education level and diabetes is modulated by sex and region (27). Specifically, a negative correlation was detected in females, whereas males presented varying correlations across different regions (27). The study revealed a negative correlation in countries such as Argentina, Brazil, Colombia, Chile, and Mexico, whereas a positive correlation was observed in Peru, Panama, and El Salvador (27). A propensity score-based analysis conducted in Bangladesh revealed no significant influence of education level on the prevalence of diabetes and hypertension (28). This finding conflicts with our study. In our study, we observed that a longer education duration was associated with a lower risk of diabetes and hypertension. The discrepancies in findings across studies may be attributed to variations in study populations. A study conducted by Christopher R. Stephens demonstrated that an increase in educational duration was associated with a reduction in WC, blood pressure, FBG, and TG, alongside an increase in HDL-C. However, subgroup analysis revealed that WC increased with education duration exclusively in females (20). Additionally, a cross-sectional analysis of an Australian cohort by L. A. Simons indicated that TG and HDL-C levels were negatively correlated with educational attainment in females but not in males (29). In our study, we observed that extended educational duration was associated with increased WC, FBG, TG, and hypertension risk, as well as decreased HDL-C levels.

First, MR was used in this study, reducing the influence of confounders. Second, GWAS data include comprehensive and summary-level data for exposure (education) and outcome (MetS, WC, hypertension, FBG, TG, HDL-C) and therefore have high investigative power. Second, multiple statistical methods were used in this study, increasing the credibility of the results.

First, the influence of IVs on outcomes remains uncertain. Second, owing to the unavailability of data (For example, GWAS data for education is available only for overall population, not separated by sex), subgroup analyses, such as examinations of sex differences, could not be conducted. Finally, significant heterogeneity was identified in this study on the basis of the Cochran’s Q value, but LOO analysis revealed that the result was credible.

This study employed MR methodology to investigate the association between levels of education and MetS. A low-level education population should be focused on reducing the incidence of MetS, obesity, hypertension, FBG, TG and HDL-C.