The Circadian Syndrome Is a Significant and Stronger Predictor for Cardiovascular Disease than the Metabolic Syndrome—The NHANES Survey during 2005–2016

In the current analysis, we used data from six survey cycles (2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, and 2015–2016) with 12,156 participants (Figure 1). Participants below the age of 20 years were excluded from the study. NHANES has collected sociodemographic, lifestyle (including sleep), disease history, and health-related data that include clinical measures of blood pressure, fasting blood glucose, serum lipids, including triglycerides and high-density lipoprotein (HDL) cholesterol, and self-reported drug use for health conditions. Participants with missing values of components used to define MetS or CircS components were excluded from the study. The final analytical sample for the cross-sectional analysis included 12,156 participants (Figure 1), but seven individuals had missing data on mortality.

Depression was measured using the Patient Health Questionnaire (PHQ-9). It is a nine-item screening instrument that asks questions about the frequency of depression symptoms over the past two weeks. Participants with a PHQ-9 score of ≥5 were defined as having depression symptoms.

MetS was defined using the harmonized criteria proposed in the joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention, National Heart, Lung, and Blood Institute, American Heart Association, World Heart Federation, International Atherosclerosis Society, and International Association for the Study of Obesity [2] (Table 1). An individual was categorized as having MetS if he/she had ≥3 of the following metabolic disorder components: (1) elevated waist circumference (≥102 cm in men, ≥88 cm in women); (2) high blood pressure (Systolic ≥130 mm Hg and/or diastolic ≥85 mm Hg) or drug treatment for hypertension; (3) low HDL-cholesterol (<40 mg/dL in men and <50 mg/dL in women) or drug treatment for dyslipidemia; (4) high triglycerides (TG) (≥150 mg/dL) or drug treatment for dyslipidemia; or (5) elevated fasting glucose (≥100 mg/dL, drug treatment of elevated glucose is an alternate indicator).

CircS was assessed based on seven components including self-reported short sleep (<6 h/day), depression symptoms and the five components used to define MetS. A cut-off for CircS was set as ≥4 components (Table 1). Based on MetS and CircS, a third variable was constructed for each individual with a possible value of: normal, MetS alone, CircS alone, and both MetS and CircS.

Participants were asked whether they had ever been diagnosed with congestive heart failure, coronary heart disease, angina, heart attack, or stroke by a doctor or other health professional. A positive answer to any of the conditions was defined as having prevalent CVD.

Mortality data were obtained via probabilistic matching to the death certificates from the National Death Index recorded up to 31 December 2019. In the analyses, the cause of death was coded according to the International Classification of Diseases, Tenth Revision (ICD-10). CVD mortality was defined based on ICD codes I00–I99. This method has been validated and widely used [15,16].

The following self-reported variables were treated as covariates: age (years), sex, race (whites, blacks, Mexican Americans, other race), education level (recoded as lower than high school, graduated from high school or equivalent institution, any college, and college graduate or above), physical activity (based on the Global Physical Activity Questionnaire (GPAQ),the metabolic equivalent of task (MET) minutes was calculated and recoded into three levels: <600, 600–1200 and >1200 MET min/week), smoking status (non-smoker, ex-smoker and current smoker), and alcohol drinking. Poverty Income Ratio (PIR) was calculated by dividing the income of the family by the poverty threshold of the family and was categorized on three levels: <1.3 (low), 1.3–3.5 (moderate), and >3.5 (high) [17].

Sex-specific analyses were conducted. Multivariable logistic regression or Cox regression analyses were used to examine the associations between MetS or CircS and prevalent CVD and CVD mortality, respectively. Survey weight was used in the multivariable model to account for the complex multistage probability sampling. Two multivariable logistic models were used: model 1, adjusted for age; and model 2, adjusted for age, education, race, income-poverty ratio, physical activity, smoking, and alcohol drinking. We calculated the positive predictive value (PPV) and negative predictive value (NPV) of MetS and CircS for prevalent CVD and CVD mortality. All analyses were conducted using STATA 17.0 (Stata Corporation, College Station, TX, USA). Statistical significance was considered when p < 0.05 (two sided). Stata codes for the analyses are available online (Supplementary Materials Code S1).

In this study, the MetS was associated with a tripled risk of prevalent CVD. The finding is in line with previous studies. In a meta-analysis of 87 studies, Mottillo et al. found that MetS had a relative risk of 2.35 (95%CI 2.02–2.73) for CVD [4]. This is the third study that has assessed the prevalence of CircS and the association between CircS and CVD. Two additional components, short sleep and depression, were added into the components of MetS to construct the CircS, which is also defined by more stringent criteria requiring ≥4 instead of ≥3 components. Indeed, using this new definition of CircS, 1131 (9.3%) participants (MetS alone) who were defined as having MetS in this study were found to be unqualified as regards having CircS (Table 2). Although it requires a higher stringency to define CircS, 259 (2.1%) additional participants (CircS alone) were identified, who were otherwise found to have no MetS. This reflects the additional value of the components of short sleep and depression added to define CircS. Existing evidence suggests that both short sleep and depression are associated with increased risk of CVD [18,19]. Using data from NHANES 2005–2018, Wang et al. found that the CircS was associated with stroke [20]. Consistent with our previous study in the Chinese population [12], the CircS was associated with CVD in the current study in the population of USA. Furthermore, the CircS had a better predictive value for prevalent CVD than the MetS as indicated by PPV.

An interesting finding of our study was that CircS alone in women was associated with a high likelihood of prevalent CVD and risk of CVD mortality, especially in those aged 40–59 years at baseline. It may be speculated that women with menopause, with a short sleep duration and depression are placed at higher risk of CVD. This emphasizes the importance to incorporate depression and short sleep into the MetS cluster that will construct the CircS. The difference between the current study in the US and the previous study in China is that the prevalence of CircS alone was lower in the US population (2.1%) than in China (4.6%). It means that in the US population, the majority of individuals with CircS had MetS, suggesting a more severe consequence of circadian disruption.

The circadian system has been highlighted as a major regulator of almost every aspect of human health and metabolism [21]. The challenge of using this knowledge in clinical medicine is the difficulty of detecting and diagnosing circadian disruption. The concept of CircS was initially proposed in 2019 based on the finding that circadian disruption was linked to many modern chronic conditions including the MetS [11]. In 2017, the Nobel Prize for Physiology or Medicine was awarded to discoveries of molecular mechanisms controlling the circadian rhythm [22]. The link between disruption of circadian rhythms and MetS has been well-established [23]. It has been reported that CircS is a better substitute for the MetS in predicting CVD [12] as well as lower urinary tract symptoms [13] in the Chinese population. The circadian rhythm disruption could underline the common etiology of MetS and play a key pathological role in the newly defined CircS as well as in other modern diseases. In modern life, the extensive use of artificial light, exposure to low light levels indoors during the day, controlled ambient temperature and constant food availability, shift work, and jet travel over many different time zones, and even daylight saving time, can disrupt circadian rhythms [11,24,25,26]. All these modern lifestyle impacts have been shown to be associated with the components used to define CircS in the current study, as well as with CVD. For example, shift work and sleep disorders have been found to cause circadian misalignment. This in turn increases the risk of cardiovascular diseases [27]. Artificial light at night has adverse effects on psychological, cardiovascular and metabolic functions [28], and sleep deprivation increases the risk of type 2 diabetes and obesity [29]. Furthermore, in the Multi-Ethnic Study of Atherosclerosis, increased variability in sleep duration and timing (based on 7-day actigraphy measurements) was positively associated with metabolic abnormalities [30]. Both meal timing and quality affect sleep quality [31,32]. A causal link between circadian misalignment and metabolic homeostasis has been found in clinical laboratory research [33], and even a mild disturbance such as daylight saving time resulted in an increase in the occurrence of acute myocardial infarction, especially in the first week after the spring shift [26].

The limitation of this study is the lack of repeated measure of CircS and MetS and some of their potential covariates. Sleep duration in the present study was self-reported rather than objectively measured. Future studies should focus on measuring sleep patterns objectively using polysomnography [34,35] and measuring the circadian pattern of cortisol or of hormones involved in metabolic physiopathology (such as thyroid hormones [36]), which can objectively ascertain whether the chrono organization of the patients is actually altered or not. The cut-off used for depression symptoms based on PHQ-9 was 5 instead of 10. This is used for mild depressive symptoms, not for clinical depression. The advantage of the use of PHQ-9 is that it is relatively easy to use and takes less time. Non-alcoholic fatty liver disease (NAFLD) was part of the original proposed CircS [11]. In NHANES 1999–2012, the prevalence of NAFLD was 30% among adults aged 18 years and above [37]. In order to compare NHANES results with our previous study in China, we did not include NAFLD in the definition of CircS. The use of self-reported CVD is another limitation. However, self-reported CVD has been shown to be reliable to ascertain symptomatic non-fatal events [38,39]. Furthermore, CVD mortality was obtained via data linkage to the national mortality registry. Based on the components of MetS, we only added short sleep and depressive syndrome to define CircS. This could explain the lack of increased mortality of CircS as compared with MetS. The strength of the present study is the use of NHANES database, as it is representative of the US general population and has a large sample size. Our findings can be generalized to the US population but studies are needed to validate whether these can be generalized to other populations. The findings of this study confirm our previous study in China on the association between CircS and CVD risk. This is the first study to examine the association between CircS and CVD mortality. It supports the use of CircS in the context of the CVD prevention in the clinical settings, as the assessment of sleep duration and depression symptom is relatively easy and inexpensive based on the measure of MetS. In practice, the use of CircS will enable the identification of those with CircS alone as a high risk group for CVD prevention. However, more studies are needed to validate the findings. To date, this is the third study comparing the use of CircS and MetS in predicting health outcomes.

The data used in the study are publicly available from the NHANES website.