Skip to main content
Download PDF

Causal relationship of sleep duration on risks for metabolic syndrome: a Mendelian randomization study

Diabetology & Metabolic Syndrome volume 17, Article number: 70 (2025) Cite this article

Abstract

Background

The cluster of cardiovascular risk factors, referred to as metabolic syndrome (MetS), represents a substantial risk factor for cardiovascular diseases and presents a significant public health challenge. However, previous epidemiological investigations exploring the link between sleep duration and MetS lack experimental evidence to establish a causal relationship. Hence, he objective of this study is to examine the association between sleep duration and MetS by employing the Mendelian randomization (MR) approach.

Methods

A cross-sectional study was conducted utilizing the Taiwan Biobank database, which comprised 33,270 predominantly Han Chinese individuals aged 30–70 years with no history of cancer and enrolled between 2008 and 2020. This study was conducted using Taiwan Biobank database. In MR analysis, we constructed weighted and unweighted genetic risk scores by calculating the SNP alleles significantly associated with sleep duration. Two-stage regression analysis was used to estimate odds ratio (OR) and 95% confidence interval (CI).

Results

In the observational epidemiologic study, after multivariate adjustment, the OR for sleep durations of < 5, 8–9 and > 9 h compared to those with a sleep duration of 7 h were 1.23 (95% CI: 1.07, 1.43), 1.15 (95% CI: 1.06, 1.24) and 1.84 (95% CI: 1.43, 2.36), respectively. In the MR analyses after multivariate adjustment, the ORs of MetS per 1 standard deviation increase in the estimated sleep duration and the probability of long and short sleep durations derived from weighted genetic risk scores were 0.64 (95% CI: 0.63, 0.66), 1.55 (95% CI: 1.51, 1.59), and 1.66 (95% CI: 1.62, 1.70), respectively.

Conclusions

Observational and MR analyses demonstrated that short and long sleep durations are potential causal risk factors for MetS. Therefore, long and short sleep durations should be considered as risk factors in MetS-prevention strategies.

Background

The global escalation in the prevalence and incidence of metabolic syndrome (MetS) over recent years is alarming, posing a substantial challenge to public health initiatives. MetS affects a significant portion of the population across various regions, with estimates ranging from 13.6 to 25.5% in Taiwan, 11.9–37.1% in Asia, 11.6–26.3% in Europe, and 20–25% globally [1]. MetS is characterized by a complex interplay among neurobiological abnormalities, inflammatory responses, and endocrine dysregulation. MetS and its components serve as a well-established precursor to several debilitating conditions, including kidney disorders, type 2 diabetes, atherosclerosis, cardiovascular disease, cancer, and premature mortality in general population [2, 3] and individuals with hypertension and obstructive sleep apnea [4, 5]. Given its widespread impact and multifaceted implications for public health, identifying factors associated with MetS has become an urgent priority in global health agendas.

The burden of diseases related to MetS is notably high in economically developed countries, but it is rapidly increasing in economically developing nations, such as China and India. This trend is primarily attributed to changes in lifestyle behaviors resulting from urbanization and an aging population [6]. MetS is widely acknowledged as a multifactorial disorder closely associated with lifestyle factors, such as poor sleep hygiene [7, 8]. Recent meta-analyses have underscored the significant role of unhealthy lifestyle behaviors in MetS development [9]. Numerous epidemiological studies, including cross-sectional studies [10,11,12,13,14,15], follow-up studies [16], and systematic reviews with meta-analyses [17,18,19,20], have examined the relationship between sleep duration and MetS and its components. However, prior epidemiologic studies have yielded conflicting findings [13, 21,22,23,24,25,26,27,28,29]; some studies have reported no significant association between short sleep duration and MetS [21, 25, 27, 28], while others have found a positive association [22,23,24]. Similarly, while some studies have found no significant association between long sleep duration and MetS [26, 28, 29], others have reported a significant association [13, 25]. Furthermore, prior epidemiologic studies lack experimental evidence to establish causal inference. Establishing evidence on causality is essential for the effective prevention and treatment of metabolic diseases. Genetic studies employing Mendelian randomization (MR) analysis, which is based on Mendel’s second law, can help to assess causality assessment. Therefore, the objective of this study is to evaluate the association between sleep duration and MetS through epidemiological and MR approaches.

Methods

Study subjects and data source

The data of the study participants were obtained from the Taiwan Biobank database, which represents Taiwan’s community population and mainly consists of individuals of Han Chinese ethnicity aged 30–70 years without a history of cancer. The Taiwan Biobank, launched in 2012, is a large-scale, population-based resource designed to collect a wide variety of genetic data, lifestyle behaviors, environmental risk factors, and family history of common, complex diseases from Taiwan individuals. TWB has specified the inclusion criteria for adult individuals aged 20–70 with the full capacity to provide informed consent. The exclusion criteria were individuals previously diagnosed with cancer, a major non-communicable disease whose incidence is expected to rise significantly after mid-life and is a leading cause of death in Taiwan. The present study further refined the inclusion criteria to adults aged 30–70, as this age group is at higher risk for MetS. Participants are recruited from over 30 sites across Taiwan, with recruitment sites strategically distributed based on population density of different counties and cities. The biobank aims to include a representative sample of the Taiwanese population, with efforts made to ensure the sample reflects various demographic characteristics such as age, gender, and regional distribution. Participants are recruited from across Taiwan, with the inclusion of individuals from urban and rural areas, which helps to enhance the generalizability of the data. The sample size of the Taiwan Biobank is substantial, involving over 200,000 participants. This allows for a more comprehensive analysis of the Taiwanese population, accounting for factors like socioeconomic status, health conditions, and lifestyle choices. To further ensure representativeness, the recruitment process is designed to minimize selection bias by targeting individuals across various backgrounds, while also considering genetic diversity within Taiwan. However, like any biobank, it may still have limitations in fully capturing all subgroups within the population. Because the variable for sleep duration was added to the questionnaire in later years, the present study included data from 45,295 individuals and 9,814,944 genetic variants identified from the genome-wide database TWB 2.0 (Fig. 1). After subjects whose genome-wide association study (GWAS) data did not meet quality control criteria were excluded, 37,705 study participants with 2,583,353 variants remained. Individuals exhibiting extreme heterozygosity rates (N = 318), duplicated or related individuals (N = 7272), and single-nucleotide polymorphisms (SNPs) with high missing genotype rates, low frequency (< 1%), or deviation from Hardy-Weinberg equilibrium (variants = 7,231,591) were also excluded. Subsequently, 299 sleep-related variants identified in the literature and 98 variants extracted from the dataset were analyzed. Finally, after 4435 individuals with 81 variants were excluded, the final cohort consisted of 33,270 individuals with 17 variants. Approval for the study was obtained from the human research committee of China Medical University Hospital (CMUH109-REC3-187), and the study was conducted in accordance with relevant regulations and guidelines.

Fig. 1
figure 1

The study flowchart for study subject and SNP selection

Full size image

Measurements

Independent variable: sleep duration

Sleep duration was operationalized as the average number of hours slept per day on workdays over the month preceding the self-reporting date provided by the respondents. This information was gathered through face-to-face interviews utilizing a standardized questionnaire, and a systematic and consistent process was used to reduce errors. Participants were asked the question, “How many hours do you typically sleep in a 24-hour period, including daytime naps, on workdays?” This method relies on the assumption that respondents can reliably estimate and summarize their sleep patterns into a single figure.

Dependent variable: metabolic syndrome

MetS comprises central obesity, elevated lipids, elevated blood pressure, and elevated glucose levels, which were assessed using a combination of laboratory measurements, physical examinations, and questionnaire responses. Body weight, height, waist circumference, and blood pressure were collected during physical examinations. Laboratory measurements included fasting plasma glucose (FPG), fasting triglycerides (TG), and high-density lipoprotein cholesterol (HDL-C) levels. Information on diagnosed diseases, prescribed medications, and medical history was obtained through standardized questionnaire responses.

The MetS was defined as having three or more of the following conditions: (1) obesity, which was indicated by a waist circumference greater than 90 cm in men or greater than 80 cm in women, or body mass index (BMI) greater than 24 kg/m2; (2) a low HDL-C level, which was indicated by a fasting HDL-C level of less than 40 mg/dL in men or less than 50 mg/dL in women; (3) hypertriglyceridemia, which was indicated by diagnosed hyperlipidemia, or fasting TG equal to or greater than 150 mg/dL; (4) hyperglycemia, which was indicated by a history of diabetes, diagnosed diabetes, or a FPG equal to or greater than 110 mg/dL; and (5) hypertension, indicated by a history of hypertension, diagnosed hypertension, or blood pressure reading of 130/85 mm Hg.

SNPs genotyping for genetic instruments in MR analysis

Genetic data were obtained from DNA samples genotyped using the TPM array and processed using the Axiom genome-wide array plate system (Affymetrix, Santa Clara, CA, USA). Each SNP was assessed for Hardy–Weinberg equilibrium (HWE) using PLINK (v2.0) [30] in the present study. Pairwise linkage disequilibrium among SNPs was quantified according to correlation coefficient r2 with Haploview (v4.2) [31]. Database imputation was conducted using IMPUTE2 [32] and a reference derived from the 1000 Genomes Project.

Genetic variants were selected by using candidate gene [33] and GWAS approaches [34,35,36,37,38,39,40,41] found in the literature. A list of selected genetic variants (i.e., SNPs) for sleep duration was compiled. We included all SNPs identified in the genomics dataset and excluded those with minor allele frequencies of < 5%. The remaining SNPs were pruned for linkage disequilibrium (Supplementary Fig. 1). Ultimately, 11 SNPs were associated with sleep duration, 2 with long sleep duration, and 4 with short sleep duration and 17 SNPs did not meet SNP-level MR assumption 1.

Statistical analysis

The Hardy-Weinberg equilibrium was assessed in participants by using the Chi-square test for goodness of fit. Individuals with and without MetS were compared in terms of sociodemographic factors, lifestyle behaviors, laboratory biomarkers, comorbidities, and medications by using a two-sample t-test and Chi-square test.

Moreover, we examined the association between sleep duration and MetS through unconditional logistic regression analysis and with a traditional epidemiological study. All analyses were adjusted for age, gender, and other potential confounders. Then, we investigated the relationship between sleep duration and its related SNPs in additive models or genetic risk scores (GRSs) to satisfy MR assumption 1 (relevance) using logistic regression models for long or short sleep (binary response variable) and linear regression models for sleep duration (continuous response variable). Only SNPs satisfy MR assumptions were selected to construct genetic risk scores. All SNPs that met the MR assumptions were tested for horizontal pleiotropy using MR-Egger regression. To validate the suitability of selected SNPs as instrumental variables for MR analysis, we assessed the associations between these SNPs and sleep variables (sleep duration, long sleep, and short sleep), coding the SNPs as 0, 1, or 2 according to the number of minor alleles. An unweighted GRS was generated by tallying the alleles of SNPs individually associated with sleep duration, and a weighted GRS was calculated by summing the number of minor alleles multiplied by their estimated coefficients from the logistic regression models and divided by the sum of weights. The weighted and unweighted GRSs were treated as continuous and categorical variables based on quartiles. Additionally, we utilized linear regression models for sleep duration and logistic regression models for long and short sleep to evaluate the associations of sleep variables with potential confounders, such as sociodemographic factors, lifestyle behaviors, laboratory biomarkers, comorbidities, and medication (MR assumption 2: independence). Lastly, logistic regression models adjusting for multiple variables were employed to estimate odds ratios (ORs) and explore the associations of individual SNPs and weighted and unweighted GRS with MetS (MR assumption 3: exclusion restrictions).

We conducted a formal MR analysis to assess the causal effect of sleep variables on MetS using instrumental variable analysis and a two-stage regression approach with multivariate adjustment. In the first stage, a linear regression model was applied to continuous exposure (sleep duration), while logistic regression models were used for binary exposure (long or short sleep). These models regressed the exposure on both weighted and unweighted genetic risk scores, which provided the genetic estimates of sleep duration or the likelihood of long or short sleep duration. In the second stage, logistic regression was utilized to evaluate the association between outcome (MetS) and the genetically predicted sleep duration or genetically predicted likelihood of long or short sleep durations, as estimated in the first stage. In the first stage, the covariates adjusted in the regression model included sociodemographic factors of gender, age, education attainment, marriage status, living alone, life-style behaviors of smoking experience, alcohol drinking, and leisure-time physical activity, albumin and eGFR while the second stage adjusted for the first two principle components of principle component analysis for GWAS data and the residuals from the first stage. The linearity assumption for linear regression analysis and linearity in logit assumption for logistic regression analysis were tested.

Both unweighted and weighted genetic scores have distinct advantages and disadvantage. The advantages of an unweighted genetic score include its simplicity, ease of calculation, and its straightforward application. However, its disadvantages include a lack of precision and an inability to account for the significance of individual variants. In contrast, a weighted genetic score incorporates the individual effect sizes of each SNP. The advantages of a weighted genetic score include higher accuracy, improved predictive power, and greater utility for complex traits. However, the disadvantages include increased complexity, reliance on data quality, and higher computational cost. In the present study, the sensitivity analysis is conducted using weighted genetic scores to assess the robustness of the study’s findings. All reported p-values were two-sided, and the level of significance was set at 0.05.

Results

A total of 33,270 participants with a mean age of 54.93 years (standard deviation: 10.18 years) were examined, and 36.03% being men. Table 1 presents the comparisons of sociodemographic factors, lifestyle behaviors, clinical and biochemical markers, and comorbidities according to MetS status. Supplementary Table 1 displays the ORs of MetS for sleep duration variables estimated in the observational epidemiologic study. The prevalence of MetS was statistically higher in individuals with sleep durations of < 5, 8–9 and > 9 h than those with a sleep duration of 7 h after multivariate adjustment (all p < 0.01). The ORs were 1.23 (95% confidence interval (CI): 1.07, 1.43), 1.15 (1.06, 1.24), and 1.84 (1.43, 2.36), respectively. Long sleep was significantly associated with MetS (OR: 1.73 [1.35, 2.22]), and short sleep was also associated with MetS (OR: 1.04 [0.89, 1.21]).

Table 1 Comparisons of sociodemographic factors, lifestyle behaviors, clinical and biochemical markers, and comorbidities according to metabolic syndrome
Full size table

Then we examined the SNP-level MR assumptions 1 and Supplementary Table 2 presents the regression coefficients, F values, and p values for sleep related SNPs that are significantly associated with sleep duration, long sleep, and short sleep. Supplementary Table 3 shows the function, minor allele, and prevalence of minor allele frequency in East Asian and the present study for these significant sleep-associated SNPs. To test the horizontal pleiotropy, MR-Egger regression was performed. The intercepts for sleep duration, long sleep, and short sleep were − 0.0070, 0.0067, and 0.0043, respectively. Two of the intercepts were not significantly different from zero (p = 0.8424 for sleep duration and p = 0.8250 for short sleep), suggesting no apparent horizontal pleiotropy. The significance test could not be performed for long sleep due to a zero standard error, although the intercept value is close to zero. Then the assumption 3 was assessed and none of the SNPs were associated with MetS (Supplementary Fig. 2). Only SNPs meet assumptions were selected to construct genetic risk scores. We then examined the GRS-level MR assumption 1, that is, the associations of weighted and unweighted GRSs with the sleep variables (Table 2). Whether the GRSs were considered as continuous or categorical variables, the unweighted and weighted GRSs were significantly associated with sleep duration, long sleep, and short sleep, thereby meeting MR assumption 1.

Table 2 Association of weighted and unweighted genetic risk scores with sleep duration, long sleep and short sleep (genetic risk score level of MR assumption 1)
Full size table

Then, we assessed MR GRS-level MR assumption 2, focusing on the associations between weighted and unweighted GRSs and covariates. Covariates included sociodemographic factors, lifestyle behaviors, and biomarkers. The weighted and unweighted GRSs derived from sleep duration (Supplementary Table 4), long sleep duration (Supplementary Table 5), and short sleep duration (Supplementary Table 6) were not associated with any covariate. Thus, all these nonsignificant covariates were deemed to satisfy assumption 2, thereby warranting consideration for adjustment in the first stage of the model for the derivation of the likelihood of sleep variables with the GRSs.

Table 3 presents the ORs of MetS in relation to unweighted and weighted GRSs for various sleep variables, including sleep duration, long and short sleep durations. Assumption 3 was evaluated at the GRS level (Supplementary Fig. 2). After multivariate adjustment, MetS did not show statistically significant association with either unweighted or weighted GRSs whether in continuous or categorical form for sleep duration and for long and short sleep durations. Thus, assumption 3 was satisfied.

Table 3 Odds ratios of metabolic syndrome with unweighted and weighted GRSs for sleep duration, and long and short sleep duration (genetic risk score level of MR assumption 3)
Full size table

Table 4 displays the ORs of MetS associated with genetic-estimated sleep variables obtained from unweighted and weighted GRSs with and without adjustment. These genetic-estimated sleep variables, denoted as hat or phat, were derived from regressing sleep parameters on the unweighted and weighted GRSs and represented the predicted sleep duration or the likelihood of experiencing long and short sleep durations. Following multivariate adjustment, the OR of MetS per 1 standard deviation increase in estimated sleep duration derived from the weighted GRS was 0.64 (95% CI: 0.63, 0.66). Upon grouping the estimated sleep duration derived from the weighted GRS with adjustment into quartiles, the highest prevalence rate of MetS was observed in quartile 1 (36.84%), while the lowest was in quartile 4 (15.77%). Relative to quartile 1 as the reference group, the adjusted ORs of MetS for quartiles 2, 3, and 4 of estimated sleep duration derived from the weighted GRS were 0.60 (95% CI: 0.56, 0.65), 0.42 (95% CI: 0.39, 0.45), and 0.32 (95% CI: 0.29, 0.34), respectively. After multivariate adjustment, the OR of MetS per 1 standard deviation increase in the phat of long sleep duration derived from weighted GRS was 1.55 (95% CI: 1.51, 1.59). After the grouping of the phat of long sleep duration derived from the weighted GRS with adjustment based on quartiles, the adjusted ORs of MetS for quartiles 2, 3, and 4 of the phat of long sleep duration derived from weighted GRS were 1.41 (95% CI: 1.30, 1.53), 2.08 (95% CI:1.93, 2.25), and 3.32 (95% CI: 3.07, 3.58), respectively. For short sleep duration, the OR of MetS per 1 standard deviation increase in the phat of short sleep duration derived from weighted GRS was 1.66 (95% CI: 1.61, 1.70) after multivariate adjustment. Relative to quartile 1, the adjusted ORs of MetS for quartiles 2, 3, and 4 of the phat of short sleep duration derived from weighted GRS were 1.73 (95% CI: 1.59, 1.89), 2.60 (95% CI: 2.40, 2.83), and 4.36 (95% CI: 4.03, 4.73), respectively.

Table 4 Odds ratios of metabolic syndrome for predictive sleep duration, and predictive likelihood of long sleep and short sleep derived from unweighted and weighted GRS
Full size table

Discussion

The present study revealed a J-shaped relationship between sleep duration and MetS in observational epidemiological research involving 33,270 individuals, indicating that individuals with either short or long sleep duration face an elevated risk of MetS. After an MR approach was used, individuals in quartiles 2, 3 and 4 of estimated sleep duration derived from weighted GRS exhibited a 0.60-, 0.42-, and 0.32-fold increase in MetS risk, respectively, compared with quartile 1, suggesting that a negative linear relationship between genetically predisposition to sleep duration and MetS. Furthermore, individuals in the highest quartiles of the estimated likelihood of long sleep from unweighted and weighted GRSs showed a 3.37- and 3.32-fold increase in MetS risk, respectively. Moreover, those in the highest quartiles of the estimated likelihood of short sleep from unweighted and weighted GRSs showed a 4.38- and 4.36-fold increase in MetS risk, respectively, compared with those in the lowest quartile.

Previous studies that employed the MR approach presented genetically predicted long or short sleep durations in relation to adverse cardiovascular outcomes, including heart failure [42]; myocardial infarction [43]; total and ischemic stroke, and intracerebral hemorrhage [44]; congestive heart failure, hypertension, obesity, and type 2 diabetes [45]; type 2 diabetes and biomarkers of FPG and HbA1c [46]; cardiovascular risk factors and lipid levels [47]; and stroke [48]. Some of these studies considered MetS biomarkers, such as hypertension, obesity, and type 2 diabetes [45]; FPG and HbA1c biomarkers [46]; and cardiovascular risk factors and lipid levels [47]. None specifically focused on MetS.

Moreover, the MR studies did not find an association between sleep duration and certain cardiovascular outcomes, such as total stroke or stroke types [44]; diabetes, FPG, and HbA1c [46]; cardiovascular risk factors and lipid levels [47]; and cardio-embolic or large artery stroke [48]. Conversely, other MR studies found associations between sleep duration and cardiovascular outcomes, such as heart failure in genome-wide association studies comprising 47,309 cases and 930,014 controls of individuals of European descent [42], myocardial infarction in 461,347 UK Biobank participants free of relevant cardiovascular disease [43], and congestive heart failure (CHF) and hypertension in the participants of Partners Biobank, which was a hospital-based cohort study [45]. In these significant associations, some studies found that genetically predicted long sleep duration was associated with a low probability of heart failure [42] and a decreased risk of CHF and hypertension [45], and another study reported that short sleep duration was associated with an increased risk of myocardial infarction [43]. In contrast to these findings, our study discovered that genetic predisposition to long and short sleep durations was associated with increased likelihood of MetS and genetic estimated long and short sleep duration was associated with increased likelihood of MetS.

Observational epidemiological studies and MR analyses showed that the results aligned between the two approaches [43]. However, inconsistency was observed [44, 45]. Daghlas [43] demonstrated that observational and MR analyses supported the association of short sleep duration with an increased risk of myocardial infarction. Conversely, Titova [44] found a link between long sleep duration and increased risk of total and ischemic stroke and between short sleep and increased risk of intracerebral hemorrhage in observational epidemiological studies. However, MR analysis did not find a significant association between short or long sleep duration and the risk of total stroke or stroke subtypes. Similarly, Dashti [45] showed that short and long sleep durations were associated with increased likelihood of hypertension, obesity, and type 2 diabetes in observational epidemiological studies. In MR analyses, long sleep duration was found to be associated with a decreased risk of CHF and hypertension [45]. In our study, observational epidemiological analyses indicated that long and short sleep durations were linked with an increased likelihood of MetS.

To give an idea of the public health impact of short sleep and long sleep duration on MetS in the population, this study also provides an estimate of population attributable risk (PAR), a measure commonly used in epidemiology to assess the proportion of a disease or health outcome in a population that can be attributed to a specific risk factor or exposure. For example, in our study, the prevalence of sleep duration of < 5 h, > 9 h and 8–9 h per day were 3.3%, 1.0% and 20.9%, respectively, and the relative risks of MetS associated with these durations were 1.23, 1.84 and 1.15, respectively (estimated using odds ratios). The corresponding PARs for sleep duration of < 5 h, > 9 h and 8–9 h per day would be 0.75%, 0.83% and 3.04%. This indicates the proportion of MetS in the population that could be prevented if long sleep duration were eliminated.

Several pathophysiological mechanisms have been proposed to elucidate the association between sleep duration and MetS. One such mechanism involves the impact of short sleep duration on metabolic and endocrine function. This effect is characterized by reduced leptin, elevated ghrelin level, increased BMI or adiposity, and impaired glycemic control [49,50,51,52,53]. The heightened hunger and appetite observed with increased BMI or adiposity may contribute to this effect [54]. Additionally, inflammation has been identified as another potential pathophysiological mechanism. Short sleep duration has been correlated with elevated levels of IL-6 and TNF-alpha [55, 56], which are associated with insulin resistance, a key characteristic of MetS. Various possible mechanisms linking long sleep and MetS have been discussed [57], including photoperiodic abnormalities, underlying disease processes, such as sleep apnea and heart disease, fatigue, sleep fragmentation, immune function, depression, and lack of challenge. These mechanisms bolster the biological plausibility of associations of long and short sleep durations with MetS.

The present study has several strengths. First, we are the first to assess the causal effects of sleep duration on the likelihood of MetS using a MR approach. The MR design minimizes potential biases because of confounding and reverse causality. Second, the sample size of the present study was large enough to provide enough power to detect weak-to-moderate associations. Third, a sensitivity analysis of deriving weighted and unweighted GRSs was performed to provide evidence of the robustness of the results. Fourth, multiple SNPs from various genes were used as instruments for sleep duration and short sleep, which can help detect potential pleiotropy and lead to strong genetic instruments. Lastly, the study restricted the population to Chinese Han individuals, thereby minimizing potential bias due to population stratification.

However, some limitations should be considered. First, sleep duration was measured through self-report by the participants, which may introduce recall bias and be less accurate than objective methods, such as actigraphy or polysomnography. Studies assessing the concordance between these two measurement approaches have found acceptable correlations between self-reported sleep duration and actigraphy [58]. Moreover, the low accuracy of self-reporting can be partially mitigated by standardized measurement. To minimize these potential biases, one strategy was implemented, including clear instructions by providing with clear, standardized instructions on how to report their sleep duration. This included specific definitions and guidelines on what to include (e.g., time spent in bed versus actual sleep time). Additionally, certain sleep quality characteristics were not considered in the present study, such as daytime nap duration, changes in sleep rhythm, sleep apnea, snoring, and dozing. A prior study reported that poor sleep quality is associated with new-onset hypertension in a diverse young and middle-aged population [16]. Second, our study focused on participants from the Chinese Han population to enhance ethnic homogeneity and limit the potential for population stratification. Caution should be taken when generalizing the findings of our study to other populations. Further investigation involving other ethnicities is warranted. Third, the age range of the study participants is specified as 30 to 70 years. Therefore, the generalizability of our results to other age groups necessitates further investigation. Nevertheless, the findings of our study can be generalized to populations within a similar age group. Fourth, the quality of MR studies depends on the selection of SNPs as instruments. Although we thoroughly searched the literature for SNPs associated with sleep duration, we may have missed some important SNPs from our GWAS data. SNPs identified in GWAS typically have small effect sizes, which can lead to errors because of weak instrumental effects. However, we derived GRS to address this issue. Fifth, SNPs with an F-statistic below 10 may not be suitable for MR analysis due to the risk of weak instrument bias. However, none of the SNPs have F-statistic values greater than 10. While there is a potential risk of weak instrument bias, its impact would be to attenuate the associations between SNP-predicted sleep-related variables and metabolic syndrome detected in this study. Despite this, the study was still able to identify an association between SNP-predicted sleep-related variables and metabolic syndrome, suggesting that the actual relationship is likely stronger. The biased results in the effect is toward the null, which is a lesser threat to validity. Lastly, while we adjusted for as many confounders as possible, some confounding factors may have been insufficiently adjusted.

Conclusions

The observational and MR analyses demonstrated that short and long sleep durations are causal risk factors for MetS. Therefore, long and short sleep durations should be considered as risk factors in MetS prevention strategies. Maintaining a healthy sleep duration may help mitigate the risk of MI in individuals at high genetic risk.

Data availability

Data described in the manuscript, code book, and analytic code will be made available from the Taiwan Biobank upon request (https://www.twbiobank.org.tw/), pending application and approval. The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Abbreviations

MetS:

Metabolic syndrome

MR:

Mendelian randomization

GWAS:

Genome-wide association study

SNPs:

Single-nucleotide polymorphisms

FPG:

Fasting plasma glucose

TG:

Triglycerides

HDL-C:

High-density lipoprotein cholesterol

BMI:

Body mass index

HWE:

Hardy–Weinberg equilibrium

GRSs:

Genetic risk scores

ORs:

Odds ratios

CI:

Confidence interval

CHF:

Congestive heart failure

References

  1. Ranasinghe P, Mathangasinghe Y, Jayawardena R, Hills AP, Misra A. Prevalence and trends of metabolic syndrome among adults in the asia-pacific region: a systematic review. BMC Public Health. 2017;17(1):101.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. O’Neill S, O’Driscoll L. Metabolic syndrome: a closer look at the growing epidemic and its associated pathologies. Obes Rev. 2015;16(1):1–12.

    Article  PubMed  Google Scholar 

  3. Marazziti D, Rutigliano G, Baroni S, Landi P, Dell’Osso L. Metabolic syndrome and major depression. CNS Spectr. 2014;19(4):293–304.

    Article  PubMed  Google Scholar 

  4. Cai X, Hu J, Wen W, Wang J, Wang M, Liu S, Zhu Q, Hong J, Dang Y, Yao X et al. Associations of the Cardiometabolic Index with the Risk of Cardiovascular Disease in Patients with Hypertension and Obstructive Sleep Apnea: Results of a Longitudinal Cohort Study. Oxid Med Cell Longev 2022, 2022:4914791.

  5. Hu J, Cai X, Li N, Zhu Q, Wen W, Hong J, Zhang D, Yao X, Luo Q, Sun L. Association between triglyceride glucose Index-Waist circumference and risk of first myocardial infarction in Chinese hypertensive patients with obstructive sleep apnoea: an observational cohort study. Nat Sci Sleep. 2022;14:969–80.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases: part I: general considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation. 2001;104(22):2746–53.

    Article  CAS  PubMed  Google Scholar 

  7. Penninx BW. Depression and cardiovascular disease: epidemiological evidence on their linking mechanisms. Neurosci Biobehav Rev. 2017;74(Pt B):277–86.

    Article  PubMed  Google Scholar 

  8. Slanovic-Kuzmanović Z, Kos I, Domijan AM. Endocrine, lifestyle, and genetic factors in the development of metabolic syndrome. Arh Hig Rada Toksikol. 2013;64(4):581–91.

    Article  PubMed  Google Scholar 

  9. Yamaoka K, Tango T. Effects of lifestyle modification on metabolic syndrome: a systematic review and meta-analysis. BMC Med. 2012;10:138.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Arora A, Pell D, van Sluijs EMF, Winpenny EM. How do associations between sleep duration and metabolic health differ with age in the UK general population? PLoS ONE. 2020;15(11):e0242852.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Feng X, Wu W, Zhao F, Li X, Han D, Li C, Xu F, Lyu J. Relationship between self-reported sleep duration during week-/work-days and metabolic syndrome from NHANES 2013 to 2016. Sleep Breath. 2022;26(4):1593–601.

    Article  PubMed  Google Scholar 

  12. Katsuura-Kamano S, Arisawa K, Uemura H, Van Nguyen T, Takezaki T, Ibusuki R, Suzuki S, Otani T, Okada R, Kubo Y, et al. Association of skipping breakfast and short sleep duration with the prevalence of metabolic syndrome in the general Japanese population: baseline data from the Japan Multi-Institutional collaborative cohort study. Prev Med Rep. 2021;24:101613.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Kim CE, Shin S, Lee HW, Lim J, Lee JK, Shin A, Kang D. Association between sleep duration and metabolic syndrome: a cross-sectional study. BMC Public Health. 2018;18(1):720.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Rafati S, Isheh M, Azarbad A, Ghadiri Soufi F, Rahimi A, Kheirandish M. The association of sleep duration and metabolic syndrome in the Bandare-Kong cohort study, a cross-sectional survey (finding from PERSIAN cohort study). Diabetol Metab Syndr. 2021;13(1):114.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Smiley A, King D, Bidulescu A. The association between sleep duration and metabolic syndrome: the NHANES 2013/2014. Nutrients. 2019;11(11):2582.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Yuan Y, Heizhati M, Wang L, Li M, Lin M, Gan L, Cai X, Yang W, Yao L, Wang Z, et al. Poor sleep quality is associated with new-onset hypertension in a diverse young and middle-aged population. Sleep Med. 2021;88:189–96.

    Article  PubMed  Google Scholar 

  17. Xie J, Li Y, Zhang Y, Vgontzas AN, Basta M, Chen B, Xu C, Tang X. Sleep duration and metabolic syndrome: an updated systematic review and meta-analysis. Sleep Med Rev. 2021;59:101451.

    Article  CAS  PubMed  Google Scholar 

  18. Che T, Yan C, Tian D, Zhang X, Liu X, Wu Z. The association between sleep and metabolic syndrome: A systematic review and Meta-Analysis. Front Endocrinol (Lausanne). 2021;12:773646.

    Article  PubMed  Google Scholar 

  19. Hua J, Jiang H, Wang H, Fang Q. Sleep duration and the risk of metabolic syndrome in adults: A systematic review and Meta-Analysis. Front Neurol. 2021;12:635564.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Iftikhar IH, Donley MA, Mindel J, Pleister A, Soriano S, Magalang UJ. Sleep duration and metabolic syndrome. An updated Dose-Risk metaanalysis. Ann Am Thorac Soc. 2015;12(9):1364–72.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Choi KM, Lee JS, Park HS, Baik SH, Choi DS, Kim SM. Relationship between sleep duration and the metabolic syndrome: Korean National health and nutrition survey 2001. Int J Obes (Lond). 2008;32(7):1091–7.

    Article  CAS  PubMed  Google Scholar 

  22. Hall MH, Muldoon MF, Jennings JR, Buysse DJ, Flory JD, Manuck SB. Self-reported sleep duration is associated with the metabolic syndrome in midlife adults. Sleep. 2008;31(5):635–43.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Kobayashi D, Takahashi O, Deshpande GA, Shimbo T, Fukui T. Relation between metabolic syndrome and sleep duration in Japan: a large scale cross-sectional study. Intern Med. 2011;50(2):103–7.

    Article  PubMed  Google Scholar 

  24. Najafian J, Toghianifar N, Mohammadifard N, Nouri F. Association between sleep duration and metabolic syndrome in a population-based study: Isfahan healthy heart program. J Res Med Sci. 2011;16(6):801–6.

    PubMed  PubMed Central  Google Scholar 

  25. Ostadrahimi A, Nikniaz Z, Faramarzi E, Mohammadpoorasl A, Ansarin K, Somi MH. Does long sleep duration increase risk of metabolic syndrome in Azar cohort study population? Health Promot Perspect. 2018;8(4):290–5.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Potter GDM, Cade JE, Hardie LJ. Longer sleep is associated with lower BMI and favorable metabolic profiles in UK adults: findings from the National diet and nutrition survey. PLoS ONE. 2017;12(7):e0182195.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Titova OE, Lindberg E, Elmståhl S, Lind L, Schiöth HB, Benedict C. Associations between the prevalence of metabolic syndrome and sleep parameters vary by age. Front Endocrinol (Lausanne). 2018;9:234.

    Article  PubMed  Google Scholar 

  28. van der Pal KC, Koopman ADM, Lakerveld J, van der Heijden AA, Elders PJ, Beulens JW, Rutters F. The association between multiple sleep-related characteristics and the metabolic syndrome in the general population: the new Hoorn study. Sleep Med. 2018;52:51–7.

    Article  PubMed  Google Scholar 

  29. Zohal M, Ghorbani A, Esmailzadehha N, Ziaee A, Mohammadi Z. Association of sleep quality components and wake time with metabolic syndrome: the Qazvin metabolic diseases study (QMDS), Iran. Diabetes Metab Syndr. 2017;11(Suppl 1):S377–80.

    Article  PubMed  Google Scholar 

  30. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81(3):559–75.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21(2):263–5.

    Article  CAS  PubMed  Google Scholar 

  32. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 2009;5(6):e1000529.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Archer SN, Carpen JD, Gibson M, Lim GH, Johnston JD, Skene DJ, von Schantz M. Polymorphism in the PER3 promoter associates with diurnal preference and delayed sleep phase disorder. Sleep. 2010;33(5):695–701.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Dashti HS, Jones SE, Wood AR, Lane JM, van Hees VT, Wang H, Rhodes JA, Song Y, Patel K, Anderson SG, et al. Genome-wide association study identifies genetic loci for self-reported habitual sleep duration supported by accelerometer-derived estimates. Nat Commun. 2019;10(1):1100.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Gottlieb DJ, Hek K, Chen TH, Watson NF, Eiriksdottir G, Byrne EM, Cornelis M, Warby SC, Bandinelli S, Cherkas L, et al. Novel loci associated with usual sleep duration: the CHARGE consortium Genome-Wide association study. Mol Psychiatry. 2015;20(10):1232–9.

    Article  CAS  PubMed  Google Scholar 

  36. Allebrandt KV, Amin N, Müller-Myhsok B, Esko T, Teder-Laving M, Azevedo RV, Hayward C, van Mill J, Vogelzangs N, Green EW, et al. A K(ATP) channel gene effect on sleep duration: from genome-wide association studies to function in Drosophila. Mol Psychiatry. 2013;18(1):122–32.

    Article  CAS  PubMed  Google Scholar 

  37. Byrne EM, Gehrman PR, Medland SE, Nyholt DR, Heath AC, Madden PA, Hickie IB, Van Duijn CM, Henders AK, Montgomery GW, et al. A genome-wide association study of sleep habits and insomnia. Am J Med Genet B Neuropsychiatr Genet. 2013;162b(5):439–51.

    Article  PubMed  Google Scholar 

  38. Doherty A, Smith-Byrne K, Ferreira T, Holmes MV, Holmes C, Pulit SL, Lindgren CM. GWAS identifies 14 loci for device-measured physical activity and sleep duration. Nat Commun. 2018;9(1):5257.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Jones SE, Tyrrell J, Wood AR, Beaumont RN, Ruth KS, Tuke MA, Yaghootkar H, Hu Y, Teder-Laving M, Hayward C, et al. Genome-Wide association analyses in 128,266 individuals identifies new morningness and sleep duration loci. PLoS Genet. 2016;12(8):e1006125.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Spada J, Scholz M, Kirsten H, Hensch T, Horn K, Jawinski P, Ulke C, Burkhardt R, Wirkner K, Loeffler M, et al. Genome-wide association analysis of actigraphic sleep phenotypes in the LIFE adult study. J Sleep Res. 2016;25(6):690–701.

    Article  PubMed  Google Scholar 

  41. Jones SE, van Hees VT, Mazzotti DR, Marques-Vidal P, Sabia S, van der Spek A, Dashti HS, Engmann J, Kocevska D, Tyrrell J, et al. Genetic studies of accelerometer-based sleep measures yield new insights into human sleep behaviour. Nat Commun. 2019;10(1):1585.

    Article  PubMed  PubMed Central  Google Scholar 

  42. van Oort S, Beulens JWJ, van Ballegooijen AJ, Handoko ML, Larsson SC. Modifiable lifestyle factors and heart failure: A Mendelian randomization study. Am Heart J. 2020;227:64–73.

    Article  PubMed  Google Scholar 

  43. Daghlas I, Dashti HS, Lane J, Aragam KG, Rutter MK, Saxena R, Vetter C. Sleep duration and myocardial infarction. J Am Coll Cardiol. 2019;74(10):1304–14.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Titova OE, Michaëlsson K, Larsson SC. Sleep duration and stroke: prospective cohort study and Mendelian randomization analysis. Stroke. 2020;51(11):3279–85.

    Article  PubMed  PubMed Central  Google Scholar 

  45. Dashti HS, Redline S, Saxena R. Polygenic risk score identifies associations between sleep duration and diseases determined from an electronic medical record biobank. Sleep. 2019;42(3):zsy247.

    Article  PubMed  Google Scholar 

  46. Wang J, Kwok MK, Au Yeung SL, Li AM, Lam HS, Leung JYY, Hui LL, Leung GM, Schooling CM. Sleep duration and risk of diabetes: observational and Mendelian randomization studies. Prev Med. 2019;119:24–30.

    Article  PubMed  Google Scholar 

  47. Zhuang Z, Gao M, Yang R, Li N, Liu Z, Cao W, Huang T. Association of physical activity, sedentary behaviours and sleep duration with cardiovascular diseases and lipid profiles: a Mendelian randomization analysis. Lipids Health Dis. 2020;19(1):86.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Lu H, Wu PF, Li RZ, Zhang W, Huang GX. Sleep duration and stroke: A Mendelian randomization study. Front Neurol. 2020;11:976.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Chaput JP, Després JP, Bouchard C, Tremblay A. Short sleep duration is associated with reduced leptin levels and increased adiposity: results from the Quebec family study. Obes (Silver Spring). 2007;15(1):253–61.

    Article  CAS  Google Scholar 

  50. Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet. 1999;354(9188):1435–9.

    Article  CAS  PubMed  Google Scholar 

  51. Taheri S, Lin L, Austin D, Young T, Mignot E. Short sleep duration is associated with reduced leptin, elevated Ghrelin, and increased body mass index. PLoS Med. 2004;1(3):e62.

    Article  PubMed  PubMed Central  Google Scholar 

  52. Van Cauter E, Spiegel K, Tasali E, Leproult R. Metabolic consequences of sleep and sleep loss. Sleep Med. 2008;9(Suppl 1):S23–28.

    Article  PubMed  PubMed Central  Google Scholar 

  53. Kamkar MZ, Mahyar M, Maddah SA, Khoddam H, Modanloo M. The effect of melatonin on quality of sleep in patients with sleep disturbance admitted to post coronary care units: A randomized controlled trial. Biomed (Taipei). 2021;11(1):34–40.

    Article  Google Scholar 

  54. Spiegel K, Tasali E, Penev P, Van Cauter E. Brief communication: sleep curtailment in healthy young men is associated with decreased leptin levels, elevated Ghrelin levels, and increased hunger and appetite. Ann Intern Med. 2004;141(11):846–50.

    Article  PubMed  Google Scholar 

  55. Irwin MR, Wang M, Campomayor CO, Collado-Hidalgo A, Cole S. Sleep deprivation and activation of morning levels of cellular and genomic markers of inflammation. Arch Intern Med. 2006;166(16):1756–62.

    Article  CAS  PubMed  Google Scholar 

  56. Vgontzas AN, Zoumakis E, Bixler EO, Lin HM, Follett H, Kales A, Chrousos GP. Adverse effects of modest sleep restriction on sleepiness, performance, and inflammatory cytokines. J Clin Endocrinol Metab. 2004;89(5):2119–26.

    Article  CAS  PubMed  Google Scholar 

  57. Grandner MA, Drummond SP. Who are the long sleepers? Towards an Understanding of the mortality relationship. Sleep Med Rev. 2007;11(5):341–60.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Cespedes EM, Hu FB, Redline S, Rosner B, Alcantara C, Cai J, Hall MH, Loredo JS, Mossavar-Rahmani Y, Ramos AR, et al. Comparison of Self-Reported sleep duration with actigraphy: results from the Hispanic community health study/study of Latinos Sueño ancillary study. Am J Epidemiol. 2016;183(6):561–73.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

This study was supported primarily by the Ministry of Science and Technology of Taiwan (MOST 108-2314-B-039-035-MY3 & MOST 110-2314-B-039-021), National Science and Technology Council (NSTC 112-2314-B-039-042- & NSTC 113-2314-B-039-042-) and China Medical University (CMU112-MF-80).

Author information

Author notes

  1. Cheng-Chieh Lin and Chuan-Wei Yang contributed equally to this work.

Authors and Affiliations

  1. School of Medicine, College of Medicine, China Medical University, Taichung, Taiwan

    Cheng-Chieh Lin, Chia-Ing Li, Chiu-Shong Liu & Chih-Hsueh Lin

  2. Department of Medical Research, China Medical University Hospital, Taichung, Taiwan

    Cheng-Chieh Lin, Chuan-Wei Yang, Chia-Ing Li & Chiu-Shong Liu

  3. Department of Family Medicine, China Medical University Hospital, Taichung, Taiwan

    Cheng-Chieh Lin, Chiu-Shong Liu & Chih-Hsueh Lin

  4. Department of Public Health, College of Public Health, China Medical University, No. 100, Sec. 1, Jingmao Rd., Beitun Dist, Taichung City, 406040, Taiwan

    Shing-Yu Yang & Tsai-Chung Li

  5. Department of Audiology and Speech-Language Pathology, Asia University, Taichung, Taiwan

    Tsai-Chung Li

Authors
  1. Cheng-Chieh Lin
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Chuan-Wei Yang
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Chia-Ing Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Chiu-Shong Liu
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Chih-Hsueh Lin
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Shing-Yu Yang
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Tsai-Chung Li
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

T.C.L., C.C.L. and C.W.Y. were responsible for drafting the article, the conception and design of the study. T.C.L., C.I.L. and S.Y.Y. acquired data and analysed data. C.S.L. and C.H.L. interpreted data. All authors revised the manuscript and approved the final version. T.C.L. and C.C.L. are responsible for the integrity of the work as a whole.

Corresponding author

Correspondence to Tsai-Chung Li.

Ethics declarations

Ethics approval and consent to participate

This present study was approved by the Ethics and Governance Council of Taiwan Biobank (approval number: TWBR10811-06) and the Ethical Review Board of China Medical University Hospital (CMUH109-REC3-187). All participants provided written informed consent. All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lin, CC., Yang, CW., Li, CI. et al. Causal relationship of sleep duration on risks for metabolic syndrome: a Mendelian randomization study. Diabetol Metab Syndr 17, 70 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-025-01643-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13098-025-01643-6

Keywords

Diabetology & Metabolic Syndrome

ISSN: 1758-5996

Contact us