Sleep-duration abnormality, depressive symptoms, and cardiometabolic indicators: evidence from NHANES 2013–2023 Authors: Lei Chen, Hongru Deng, Lingbing Meng Published: 2026 ABSTRACT AbstractBackgroundSleep-duration abnormality and depressive symptoms are both associated with cardiometabolic health, and they often co-exist. Current research has largely focused on the separate associations of abnormal sleep duration or depressive symptoms with cardiometabolic indicators, whereas fewer studies have examined their joint association and potential interaction across population subgroups. This study aimed to evaluate the co-occurrence of sleep-duration abnormality and depressive symptoms in relation to cardiometabolic indicators and to explore whether these associations differed by gender and age using NHANES 2013–2023 data.MethodsThis study utilized publicly available data from the National Health and Nutrition Examination Survey (NHANES) 2013–2023. Adult participants with complete data on sleep-duration variables, depression assessment, and cardiometabolic indicators were included. Sleep exposure was operationalized using NHANES Sleep Questionnaire variables for usual weekday/workday sleep duration (SLD012) and weekend/non-workday sleep duration (SLD013). Sleep-duration abnormality was defined as short sleep duration (< 7 h) or long sleep duration (> 9 h) on either weekday/workday or weekend/non-workday measures. Depressive symptoms were evaluated using the Patient Health Questionnaire-9 (PHQ-9). Cardiometabolic indicators included systolic blood pressure (SBP), diastolic blood pressure (DBP), body mass index (BMI), total cholesterol, and high-sensitivity C-reactive protein (hs-CRP). Survey-weighted multivariable logistic regression was used to examine the associations between sleep-duration abnormality, depressive symptoms, and cardiometabolic indicators. The adjusted models included demographic, socioeconomic, lifestyle, and clinical covariates, including age, gender, race/ethnicity, education level, household income, smoking status, alcohol consumption, diabetes history, and BMI where appropriate. Multiplicative interaction was assessed by including a product term between sleep-duration abnormality and depressive symptoms in the regression models. For hypertension, additive interaction was further evaluated using the relative excess risk due to interaction (RERI) and the attributable proportion due to interaction (AP). Gender- and age-stratified analyses were also conducted.ResultsParticipants with both sleep-duration abnormality and depressive symptoms had higher levels or prevalence of several adverse cardiometabolic indicators, including hypertension, obesity, and elevated hs-CRP, than those without either condition. For hypertension, the odds ratio associated with the co-occurrence of sleep-duration abnormality and depressive symptoms was 3.02 (95% CI: 2.56–3.56), compared with 1.48 (95% CI: 1.26–1.74) for sleep-duration abnormality alone and 1.62 (95% CI: 1.38–1.91) for depressive symptoms alone. Evidence of interaction for hypertension was observed on both the multiplicative scale (interaction P < 0.001) and additive scale (RERI = 1.21, 95% CI: 0.93–1.49; AP = 40.1%, 95% CI: 31.5–48.7%). Stratified analyses suggested that the additive interaction for hypertension was more apparent among women and older adults.ConclusionsThe co-occurrence of sleep-duration abnormality and depressive symptoms was associated with less favorable cardiometabolic indicators in this cross-sectional NHANES analysis. Evidence of additive interaction was observed for hypertension, particularly among women and older adults. These findings support the need to consider abnormal sleep duration and depressive symptoms jointly when evaluating cardiometabolic health, while causal and mechanistic interpretations require confirmation in longitudinal studies.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s13019-026-04469-z.Keywords: Sleep-duration abnormality, Depressive symptoms, Sleep-depression co-occurrence, Cardiometabolic indicators, Additive interaction, Multiplicative interaction, Gender differences, Age stratification, NHANES Introduction Cardiovascular disease (CVD) remains one of the leading contributors to global mortality and disease burden [1]. According to the World Health Organization, CVDs are the leading cause of death globally and accounted for an estimated 19.8 million deaths in 2022, representing approximately 32% of all global deaths [2]. Beyond conventional risk factors, increasing evidence suggests that psychological and behavioral factors are closely related to cardiometabolic health [3]. In particular, sleep disturbance and abnormal sleep duration have been associated with adverse cardiometabolic profiles and cardiovascular outcomes [4, 5], while depressive symptoms and depression have been linked to poorer cardiovascular health and adverse cardiovascular prognosis [6, 7]. Sleep-duration abnormality represents an important dimension of sleep disturbance [8, 9]. Previous systematic reviews and dose-response meta-analyses have shown that short sleep duration is associated with an increased risk of hypertension, while both short and long sleep duration have been linked to higher risks of coronary heart disease and other adverse cardiovascular outcomes [10, 11]. Depressive symptoms have also been independently associated with the development and prognosis of cardiovascular diseases, including coronary heart disease, myocardial infarction, and mortality among patients with established coronary heart disease [12, 13]. In addition, sleep problems and depressive symptoms frequently co-occur. Sleep disturbance, including insomnia and hypersomnia, is a core symptom and common clinical feature of depression [14], and bidirectional relationships between sleep disturbance and depression have been reported in previous reviews [15, 16]. Therefore, individuals with both sleep-duration abnormality and depressive symptoms may represent a subgroup with less favorable cardiometabolic profiles. Several biological and behavioral pathways may link sleep-duration abnormality and depressive symptoms to cardiometabolic dysregulation [17, 18]. Sleep disturbance and depression have both been associated with hypothalamic-pituitary-adrenal axis activation, sympathetic nervous system dysregulation, autonomic imbalance, inflammatory activation, and circadian rhythm disruption [19]. These pathways may contribute to elevated blood pressure, metabolic abnormalities, endothelial dysfunction, and systemic inflammation [20]. Meanwhile, gender- and age-related differences may modify these associations [21]. Sex-related biological factors, differences in stress responses, and social determinants may influence the relationship between sleep, depression, and cardiometabolic health [22]. In older adults, accumulated chronic disease burden, reduced physiological reserve, and age-related inflammatory activation may further contribute to adverse cardiometabolic profiles [23]. Although sleep health and psychological health are increasingly recognized as important components or contextual determinants of cardiovascular and cardiometabolic health [3, 5, 24], many previous studies have examined sleep disturbance or depression separately, and comparatively fewer studies have evaluated their co-occurrence or interaction in relation to cardiometabolic indicators [25, 26]. Importantly, the interpretation of a combined association depends on the statistical scale used to evaluate interaction. Multiplicative interaction reflects departure from multiplicativity in regression models, whereas additive interaction, commonly assessed by the relative excess risk due to interaction (RERI) and the attributable proportion due to interaction (AP), evaluates whether the combined association exceeds the sum of the separate associations on an additive scale [27, 28]. Clarifying this distinction is necessary to avoid overinterpreting joint associations as biological synergy. The National Health and Nutrition Examination Survey (NHANES) provides standardized multidimensional health data by combining interviews, physical examinations, and laboratory tests, and therefore offers an opportunity to examine these associations in a large population-based sample[38]. Using NHANES 2013–2023 data, this study aimed to: (1) evaluate the co-occurrence of sleep-duration abnormality and depressive symptoms in relation to cardiometabolic indicators; (2) assess multiplicative and additive interaction for hypertension; and (3) explore whether these associations and interaction estimates differed by gender and age. Materials and methods Data source and study population This study used publicly available data from NHANES 2013–2023. NHANES employs a multi-stage, stratified sampling design to represent the health status of the non-institutionalized civilian population in the United States. All participants provided informed consent, and the study protocol was approved by the Ethics Review Board of the National Center for Health Statistics [15]. Adults aged ≥ 18 years were eligible for inclusion. Participants were included if they had complete data on sleep-duration variables from the NHANES Sleep Questionnaire, PHQ-9 depressive symptom assessment, cardiometabolic indicators, and key covariates used in the adjusted models. Required cardiometabolic variables included body mass index, blood pressure, blood lipid indicators, and hs-CRP. Participants were excluded if they were younger than 18 years, were pregnant at the time of examination, or had missing data on key exposure, outcome, or covariate variables required for the complete-case analysis. Pregnant women were excluded because pregnancy is associated with substantial physiological changes in blood pressure, lipid metabolism, inflammatory markers, body weight, and sleep patterns, which could affect the interpretation of cardiometabolic indicators. Severe cognitive impairment and acute psychiatric disorders were not retained as independent exclusion criteria because they could not be consistently operationalized using specific NHANES variables across all included cycles. Instead, participants with missing key exposure, outcome, or covariate data were excluded through complete-case analysis. Participants with self-reported established cardiovascular disease, including prior myocardial infarction, stroke, or heart failure, were not excluded from the primary analysis. These conditions were not treated as primary outcomes in the present study because the analysis focused on individual cardiometabolic indicators rather than incident cardiovascular events. Their potential influence on baseline cardiometabolic profiles was addressed as a source of residual confounding in the Discussion. Existing psychiatric diagnoses were not used as exclusion criteria. Depressive symptom status was operationalized using PHQ-9 scores rather than clinical psychiatric diagnoses. Prescription medication use, including antidepressants, antipsychotics, sedatives, anxiolytics, and antihypertensive medications, was not included in the primary adjustment models because medication data were incomplete and inconsistently available across NHANES cycles. Among chronic conditions, diabetes history was included as an adjustment variable. Other chronic conditions, such as chronic kidney disease or chronic inflammatory diseases, were not used as exclusion criteria in the primary analysis and were considered potential sources of residual confounding. Variable definitions and measurements Core exposure variables Sleep exposure was assessed using sleep-duration variables from the NHANES Sleep Questionnaire (SLQ). Specifically, usual weekday/workday sleep duration was assessed using SLD012, and usual weekend/non-workday sleep duration was assessed using SLD013. According to the NHANES documentation, SLD012 and SLD013 are derived variables calculated from reported usual sleep and wake times during the main sleeping period and represent hours of sleep on weekdays/workdays and weekends/non-workdays, respectively. Therefore, in the present study, the sleep exposure was operationalized as sleep-duration abnormality rather than a clinical diagnosis of insomnia or a global measure of sleep quality. Sleep duration was categorized as short sleep duration (< 7 h), normal sleep duration (7–9 h), or long sleep duration (> 9 h). Sleep-duration abnormality was defined as short or long sleep duration on either weekday/workday sleep duration (SLD012) or weekend/non-workday sleep duration (SLD013). Participants with sleep duration within 7–9 h on both available sleep-duration measures were classified as having no sleep-duration abnormality. Depressive symptoms were assessed using the Patient Health Questionnaire-9 (PHQ-9), which evaluates depressive symptoms over the past two weeks with a total score ranging from 0 to 27. Depression severity was categorized as no depression (0–4 points), mild depressive symptoms (5–9 points), and moderate-to-severe depressive symptoms (≥ 10 points). Sleep-depression co-occurrence was defined as the simultaneous presence of sleep-duration abnormality and moderate-to-severe depressive symptoms. Cardiometabolic indicators The outcomes in this study were individual cardiometabolic indicators and biomarkers rather than incident cardiovascular events or validated cardiovascular risk scores. These indicators included blood pressure, obesity-related measures, lipid indicators, and systemic inflammatory status. Blood pressure indicators included systolic blood pressure (SBP) and diastolic blood pressure (DBP). Hypertension was defined as SBP ≥ 140 mmHg, DBP ≥ 90 mmHg, or self-reported history of hypertension[5–7]. This threshold was selected to maintain consistency with commonly used epidemiological definitions in population-based studies and to facilitate comparison with previous NHANES-based analyses. Because some contemporary clinical guidelines use lower thresholds, such as 130/80 mmHg, the present definition should be interpreted as an epidemiological operational definition rather than a comprehensive clinical diagnosis. Obesity-related indicators included body mass index (BMI) and waist circumference. Obesity was defined as BMI ≥ 30 kg/m². Metabolic indicators included total cholesterol, triglycerides, and low-density lipoprotein cholesterol (LDL-C). Systemic inflammation was assessed using high-sensitivity C-reactive protein (hs-CRP), with elevated hs-CRP defined as ≥ 3 mg/L. Stratification variables and covariates Stratification variables included gender and age group. Gender was categorized as male or female. Age was categorized as young adults (18–44 years), middle-aged adults (45–64 years), and older adults (≥ 65 years). Covariates included demographic variables, socioeconomic indicators, lifestyle factors, and clinical variables. Specifically, the main adjustment variables included age, gender, race/ethnicity, education level, household income, smoking status, alcohol consumption, and diabetes history. BMI was additionally adjusted for in models where BMI or obesity was not the outcome variable. In stratified analyses, the corresponding stratification variable was not included as an adjustment variable within that subgroup model. Established cardiovascular disease history, including prior myocardial infarction, stroke, and heart failure, was not used as an exclusion criterion or primary adjustment variable. Existing psychiatric diagnoses were not used to define the depression exposure, which was based on PHQ-9 depressive symptom scores. Prescription medication use, including antidepressants, antipsychotics, sedatives, anxiolytics, and antihypertensive medications, was not included as a covariate in the primary analyses due to substantial missing data. Other chronic conditions, such as chronic kidney disease or chronic inflammatory diseases, were not excluded in the primary analysis and were considered potential sources of residual confounding. Statistical analysis methods Statistical analyses were performed using R software (version 4.2.0), accounting for the complex sampling design of NHANES by applying examination weights (WTMEC2YR), primary sampling units (SDMVPSU), and stratification variables (SDMVSTRA)[14–17]. Descriptive statistics: Continuous variables were presented as mean ± standard deviation or median (interquartile range), and categorical variables as frequency (percentage). Group comparisons: t-tests, analysis of variance (ANOVA), or Kruskal-Wallis tests were used to compare differences between groups. Missing data handling: Participants with missing data on key exposure variables, PHQ-9 depressive symptom assessment, cardiometabolic indicators, or covariates required for the adjusted models were excluded using a complete-case approach. Because complete-case analysis may introduce selection bias if missingness is not completely random, this issue was acknowledged as a limitation. Multivariate logistic regression: Associations between sleep-duration abnormality, depressive symptoms, and cardiometabolic indicators were examined using survey-weighted multivariable logistic regression, with odds ratios (ORs) and 95% confidence intervals (CIs) calculated. The primary adjusted models included age, gender, race/ethnicity, education level, household income, smoking status, alcohol consumption, diabetes history, and BMI where appropriate. Established cardiovascular disease history, psychiatric diagnoses, prescription medication use, chronic kidney disease, and chronic inflammatory diseases were not included in the primary adjustment set because of variable availability, missingness, or inconsistent measurement across NHANES cycles. For subgroup analyses stratified by gender or age, the corresponding stratification variable was omitted from the adjustment set within that subgroup model. Interaction assessment: Interaction between sleep-duration abnormality and depressive symptoms was evaluated on both the multiplicative and additive scales. Multiplicative interaction was assessed by including a product term between sleep-duration abnormality and depressive symptoms in the survey-weighted logistic regression models. Additive interaction was evaluated for hypertension using the relative excess risk due to interaction (RERI) and the attributable proportion due to interaction (AP). In these analyses, participants without sleep-duration abnormality or depressive symptoms served as the reference group, while participants with sleep-duration abnormality alone, depressive symptoms alone, and both conditions were compared with this reference group. RERI was calculated as OR11 − OR10 − OR01 + 1, where OR11 represents the odds ratio for participants with both sleep-duration abnormality and depressive symptoms, OR10 for sleep-duration abnormality alone, and OR01 for depressive symptoms alone. AP was calculated as RERI/OR11. RERI > 0 and AP > 0 were interpreted as evidence of positive additive interaction. Because additive interaction is scale-dependent, these results were interpreted cautiously and were not used to infer causality. Stratified analysis: Stratified analyses were conducted by gender and age group to explore whether the associations and interaction estimates differed across population subgroups. Visualization: Box plots, line plots, and heat maps were used only to descriptively present group differences and age- or gender-related patterns in cardiometabolic indicators. Some descriptive figures used graphical age categories displayed within the figure to illustrate age-related trends. These graphical categories were not used for the formal age-stratified regression analyses, which were conducted using the prespecified age groups of 18–44, 45–64, and ≥ 65 years. Formal interaction results were primarily reported using model-based estimates, including interaction P values, RERI, and AP. Exploratory ANOVA-based visualizations were provided as supplementary materials and were not interpreted as formal interaction analyses. Results Baseline characteristics of the study population Distribution of sleep-duration abnormality and depressive symptom co-occurrence Participants were categorized into four mutually exclusive groups according to sleep-duration abnormality and depressive symptom status: participants without either condition, participants with sleep-duration abnormality only, participants with depressive symptoms only, and participants with both sleep-duration abnormality and depressive symptoms. The prevalence of co-occurring sleep-duration abnormality and depressive symptoms was 15.3%. The prevalence was higher among women than among men (18.9% vs. 11.8%, P < 0.001). Across age groups, the prevalence was 8.2% among adults aged 18–44 years, 16.5% among those aged 45–64 years, and 21.3% among those aged ≥ 65 years (trend test P < 0.001). Demographic and clinical characteristics Baseline characteristics differed across the four exposure groups. Participants with both sleep-duration abnormality and depressive symptoms were older than participants without either condition, with a median age of 58.3 years (IQR: 45.6–70.1) compared with 47.6 years (IQR: 36.2–58.9) (P < 0.001). The proportion of women was also higher among participants with both conditions. Educational attainment and lifestyle factors differed across groups: 42.3% of participants with both conditions had less than a high school education, and the current smoking rate was 35.6%, both higher than those observed among participants without either condition. Cardiometabolic indicators also differed across groups. The prevalence of hypertension was 51.2% among participants with both sleep-duration abnormality and depressive symptoms, compared with 38.4% among those with sleep-duration abnormality only, 41.6% among those with depressive symptoms only, and 25.8% among participants without either condition (between-group comparison P < 0.001). Obesity was also more frequent among participants with both conditions than among the other three groups. Median hs-CRP was 3.8 mg/L (IQR: 2.4–5.6) among participants with both conditions, compared with 2.1 mg/L (IQR: 1.3–3.4) among those with sleep-duration abnormality only, 2.6 mg/L (IQR: 1.7–3.9) among those with depressive symptoms only, and 1.3 mg/L (IQR: 0.8–2.2) among participants without either condition. Detailed baseline characteristics are shown in Tables 1 and 2. Table 1Baseline Characteristics of Participants According to Sleep-Duration Abnormality and Depressive Symptom StatusVariable CategorySpecific IndicatorNeither Condition (n = 3215)Sleep-Duration Abnormality Only (n = 1562)Depressive Symptoms Only (n = 842)Both Conditions (n = 935)P-valueDemographic CharacteristicsAge, median(IQR)47.6 (36.2–58.9)52.3 (41.5–63.2)50.8 (39.7–61.9)58.3 (45.6–70.1)< 0.001Male, n(%)1571 (48.9%)698 (44.7%)352 (41.8%)320 (34.2%)< 0.001Education< High School, n(%)923 (28.7%)586 (37.5%)346 (41.1%)395 (42.3%)< 0.001Physical MeasurementsBMI, median(IQR)26.8 (23.5–30.2)28.4 (25.1–32.6)28.1 (24.8–31.9)30.5 (27.2–34.8)< 0.001ComorbiditiesHypertension, n(%)830 (25.8%)600 (38.4%)350 (41.6%)478 (51.2%)< 0.001Diabetes, n(%)386 (12.0%)281 (18.0%)168 (20.0%)280 (30.0%)< 0.001Dyslipidemia, n(%)675 (21.0%)452 (28.9%)231 (27.4%)374 (40.0%)< 0.001LifestyleCurrent Smoker, n(%)620 (19.3%)437 (28.0%)252 (29.9%)333 (35.6%)< 0.001Routine Lab TestsFasting Glucose, median (IQR)5.3 (4.9–5.8)5.7 (5.2–6.3)5.6 (5.1–6.2)6.1 (5.5–6.9)< 0.001LDL-C, median (IQR)2.6 (2.1–3.1)2.9 (2.4–3.5)2.8 (2.3–3.4)3.2 (2.6–3.9)< 0.001Inflammatory Markershs-CRP, median (IQR)1.3 (0.8–2.2)2.1 (1.3–3.4)2.6 (1.7–3.9)3.8 (2.4–5.6)< 0.001White Blood Cell Count, median (IQR)6.2 (5.1–7.3)6.7 (5.6-8.0)6.8 (5.7–8.1)7.4 (6.2–8.8)< 0.001Neutrophil Percentage, median (IQR)58.5 (53.0–64.0)60.8 (55.5–66.5)61.2 (56.0–67.0)63.8 (58.5–69.5)< 0.001BMI = body mass index; LDL-C = low-density lipoprotein cholesterol; hs-CRP = high-sensitivity C-reactive protein. Continuous variables not normally distributed are expressed as median (interquartile range); categorical variables are expressed as number (percentage) Baseline Characteristics of Participants According to Sleep-Duration Abnormality and Depressive Symptom Status BMI = body mass index; LDL-C = low-density lipoprotein cholesterol; hs-CRP = high-sensitivity C-reactive protein. Continuous variables not normally distributed are expressed as median (interquartile range); categorical variables are expressed as number (percentage) Table 2Cardiometabolic Indicators According to Sleep-Duration Abnormality and Depressive Symptom StatusVariable CategorySpecific IndicatorNeither Condition (n = 3215)Sleep-Duration Abnormality Only (n = 1562)Depressive Symptoms Only (n = 842)Both Conditions (n = 935)P-value (Between-group)Blood Pressure IndicatorsSystolic BP, mean ± SD (mmHg)121.8 ± 13.5126.5 ± 15.2128.3 ± 15.7134.2 ± 16.8< 0.001Diastolic BP, mean ± SD(mmHg)75.3 ± 9.178.2 ± 9.879.6 ± 10.184.6 ± 10.3< 0.001Hypertension, n (%)830 (25.8%)600 (38.4%)350 (41.6%)478 (51.2%)< 0.001Metabolic & Obesity IndicatorsBMI, mean ± SD (kg/m²)26.8 ± 4.728.4 ± 5.328.1 ± 5.130.5 ± 5.8< 0.001Obesity (BMI ≥ 30), n (%)869 (27.0%)625 (40.0%)370 (44.0%)521 (55.7%)< 0.001Total Cholesterol, median (IQR)(mmol/L)4.8 (4.2–5.4)5.1 (4.5–5.8)5.0 (4.4–5.7)5.5 (4.8–6.2)< 0.001LDL-C, median (IQR) (mmol/L)2.6 (2.1–3.1)2.9 (2.4–3.5)2.8 (2.3–3.4)3.2 (2.6–3.9)< 0.001Inflammatory Markerhs-CRP, median (IQR) (mg/L)1.3 (0.8–2.2)2.1 (1.3–3.4)2.6 (1.7–3.9)3.8 (2.4–5.6)< 0.001Elevated hs-CRP (≥ 3 mg/L), n (%)643 (20.0%)562 (36.0%)378 (44.9%)598 (64.0%)< 0.001BMI = body mass index; LDL-C = low-density lipoprotein cholesterol; hs-CRP = high-sensitivity C-reactive protein. Continuous variables with approximately normal distributions are presented as mean ± standard deviation; otherwise, they are expressed as median (interquartile range). Categorical variables are expressed as number (percentage) Cardiometabolic Indicators According to Sleep-Duration Abnormality and Depressive Symptom Status BMI = body mass index; LDL-C = low-density lipoprotein cholesterol; hs-CRP = high-sensitivity C-reactive protein. Continuous variables with approximately normal distributions are presented as mean ± standard deviation; otherwise, they are expressed as median (interquartile range). Categorical variables are expressed as number (percentage) Association between sleep-depression co-occurrence and cardiometabolic indicators Blood pressure trends Blood pressure levels differed across the four exposure groups. Mean SBP was 134.2 ± 16.8 mmHg among participants with both sleep-duration abnormality and depressive symptoms, compared with 126.5 ± 15.2 mmHg among those with sleep-duration abnormality only, 128.3 ± 15.7 mmHg among those with depressive symptoms only, and 121.8 ± 13.5 mmHg among participants without either condition (P < 0.001). DBP showed a similar pattern, with the highest mean value observed among participants with both conditions. Figures 1 and 2 provide descriptive visualizations of SBP and DBP patterns by age, gender, and sleep-depression status using graphical age categories displayed within the figures. These visualizations were used only to illustrate descriptive trends and were not interpreted as formal age-stratified or interaction analyses. Descriptive distribution of systolic blood pressure according to sleep-duration abnormality and depressive symptom status by gender. Age categories shown in the figure were used only for graphical visualization and were not used for formal age-stratified regression analyses Descriptive trends in systolic and diastolic blood pressure according to age, gender, and sleep-depression status. Age categories shown in the figure were used only for graphical visualization and were not used for formal age-stratified regression analyses Hypertension prevalence and interaction analysis Survey-weighted multivariable logistic regression was used to evaluate the association between sleep-depression co-occurrence and hypertension. After adjustment for demographic, socioeconomic, lifestyle, and clinical covariates, participants with both sleep-duration abnormality and depressive symptoms had higher odds of hypertension (OR = 3.02, 95% CI: 2.56–3.56). Sleep-duration abnormality only (OR = 1.48, 95% CI: 1.26–1.74) and depressive symptoms only (OR = 1.62, 95% CI: 1.38–1.91) were also associated with higher odds of hypertension, but with smaller point estimates. The detailed estimates are shown in Table 3. Table 3Gender- and age-stratified logistic regression and interaction analysis for hypertension according to sleep-duration abnormality and depressive symptom statusStratification VariableSubgroupSleep-Duration Abnormality Only OR (95% CI)Depressive Symptoms Only OR (95% CI)Both Conditions OR (95% CI)Multiplicative Interaction P-valueRERI (95% CI)AP (95% CI)Overall1.48 (1.26–1.74)1.62 (1.38–1.91)3.02 (2.56–3.56)< 0.0011.21 (0.93–1.49)40.1% (31.5–48.7%)GenderMale1.42 (1.16–1.74)1.52 (1.24–1.86)2.51 (2.10-3.00)0.020.97 (0.65–1.29)32.1% (23.5–40.7%)Female1.55 (1.28–1.88)1.75 (1.44–2.12)3.45 (2.89–4.12)< 0.0011.43 (1.12–1.74)45.3% (36.8–53.8%)Age GroupYoung adults (18–44 years)1.32 (1.02–1.71)1.41 (1.08–1.85)1.89 (1.32–2.71)0.120.58 (0.15–1.01)20.5% (8.5–32.5%)Middle-aged adults (45–64 years)1.51 (1.25–1.82)1.68 (1.38–2.04)2.98 (2.35–3.78)0.0021.05 (0.72–1.38)35.2% (26.8–43.6%)Older adults (≥ 65 years)1.61 (1.31–1.98)1.85 (1.50–2.28)3.56 (2.87–4.42)< 0.0011.52 (1.18–1.86)42.7% (34.5–50.9%)Models were adjusted for age, gender, race/ethnicity, education level, household income, smoking status, alcohol consumption, diabetes history, and BMI where appropriate. In gender- and age-stratified analyses, the corresponding stratification variable was omitted from the adjustment set within that subgroup model. Prescription medication use was not included in the primary adjustment set due to substantial missing data. RERI = relative excess risk due to interaction; AP = attributable proportion due to interaction. RERI and AP were used to evaluate additive interaction for hypertension Gender- and age-stratified logistic regression and interaction analysis for hypertension according to sleep-duration abnormality and depressive symptom status Models were adjusted for age, gender, race/ethnicity, education level, household income, smoking status, alcohol consumption, diabetes history, and BMI where appropriate. In gender- and age-stratified analyses, the corresponding stratification variable was omitted from the adjustment set within that subgroup model. Prescription medication use was not included in the primary adjustment set due to substantial missing data. RERI = relative excess risk due to interaction; AP = attributable proportion due to interaction. RERI and AP were used to evaluate additive interaction for hypertension Interaction analysis was performed for hypertension. In the multiplicative interaction model, the product term between sleep-duration abnormality and depressive symptoms was statistically significant (interaction P < 0.001). Additive interaction analysis showed a positive RERI of 1.21 (95% CI: 0.93–1.49) and an AP of 40.1% (95% CI: 31.5–48.7%), consistent with positive additive interaction for hypertension. In gender-stratified analyses, the RERI and AP were numerically higher among women (RERI = 1.43, 95% CI: 1.12–1.74; AP = 45.3%, 95% CI: 36.8–53.8%) than among men (RERI = 0.97, 95% CI: 0.65–1.29; AP = 32.1%, 95% CI: 23.5–40.7%). Metabolic and inflammatory indicators Metabolic and inflammatory indicators differed across exposure groups. Mean BMI was highest among participants with both sleep-duration abnormality and depressive symptoms (30.5 ± 5.8 kg/m²), compared with the other three groups (P < 0.001). Lipid indicators also differed across groups, with higher total cholesterol and LDL-C levels observed among participants with both conditions. Median hs-CRP was 3.8 mg/L (IQR: 2.4–5.6) among participants with both conditions, compared with 2.1 mg/L (IQR: 1.3–3.4) among those with sleep-duration abnormality only, 2.6 mg/L (IQR: 1.7–3.9) among those with depressive symptoms only, and 1.3 mg/L (IQR: 0.8–2.2) among participants without either condition. Figure 3 shows descriptive age-related patterns of cardiometabolic indicators using graphical age categories displayed within the figure. This figure was used for descriptive visualization only and was not used as the basis for formal age-stratified inference. Descriptive age-related patterns of cardiometabolic indicators according to sleep-duration abnormality and depressive symptom status. Age categories shown in the figure were used only for graphical visualization and were not used for formal age-stratified regression analyses Gender- and age-stratified interaction analysis for hypertension Gender differences In gender-stratified analyses, the association between co-occurring sleep-duration abnormality and depressive symptoms and hypertension was observed in both men and women. The OR for hypertension associated with both conditions was 3.45 (95% CI: 2.89–4.12) among women and 2.51 (95% CI: 2.10–3.00) among men. Additive interaction estimates were also numerically higher among women, as reflected by the RERI and AP values shown in Table 3. Age differences Age-stratified analyses showed that the association between co-occurring sleep-duration abnormality and depressive symptoms and hypertension differed across age groups. Among adults aged 18–44 years, the OR for hypertension was 1.89 (95% CI: 1.32–2.71), and the multiplicative interaction term was not statistically significant (interaction P = 0.12). Among adults aged 45–64 years, the corresponding OR was 2.98 (95% CI: 2.35–3.78), with evidence of interaction (interaction P = 0.002). Among adults aged ≥ 65 years, the OR was 3.56 (95% CI: 2.87–4.42), with a statistically significant interaction term (interaction P < 0.001). Additive interaction estimates showed a similar increasing pattern across age strata, as reflected by the RERI and AP values in Table 3. Exploratory ANOVA-based visualization of SBP differences is provided in Supplementary Figure S1 and was interpreted only as a descriptive supplement. Descriptive comparison of inflammatory markers hs-CRP levels differed significantly among the four exposure groups (Kruskal-Wallis test P < 0.001). Participants with both sleep-duration abnormality and depressive symptoms had the highest hs-CRP levels, followed by those with depressive symptoms only and sleep-duration abnormality only. The median hs-CRP level was 3.8 mg/L (IQR: 2.4–5.6) among participants with both conditions, compared with 2.6 mg/L (IQR: 1.7–3.9) among those with depressive symptoms only, 2.1 mg/L (IQR: 1.3–3.4) among those with sleep-duration abnormality only, and 1.3 mg/L (IQR: 0.8–2.2) among participants without either condition. These results were descriptive and were not interpreted as evidence of mediation or a biological mechanism. Discussion Summary of key findings This study examined the joint association of sleep-duration abnormality and depressive symptoms with cardiometabolic indicators using NHANES 2013–2023 data. The main finding was that participants with both sleep-duration abnormality and depressive symptoms had less favorable cardiometabolic profiles, including higher prevalence of hypertension, obesity, and elevated hs-CRP. For hypertension, interaction analyses suggested evidence of interaction on both the multiplicative and additive scales, as reflected by the product-term P value, RERI, and AP. These findings suggest that the co-occurrence of sleep-duration abnormality and depressive symptoms may be associated with a greater hypertension burden than either condition alone, with more apparent interaction estimates among women and older adults in stratified analyses. The observed gender and age differences should be interpreted cautiously. Previous studies have suggested that sex-related biological factors, psychosocial stress responses, and age-related physiological changes may influence the relationships among sleep disturbance, depressive symptoms, inflammation, and cardiometabolic health [29–31]. Our findings are generally consistent with this literature, but the cross-sectional design prevents causal or mechanistic interpretation. Therefore, the present results should be understood as hypothesis-generating evidence supporting further longitudinal research. Potential biological and behavioral pathways Several biological and behavioral pathways may help explain the observed associations between sleep-depression co-occurrence and less favorable cardiometabolic indicators. Both sleep disturbance and depressive symptoms have been linked to hypothalamic-pituitary-adrenal axis activation, sympathetic nervous system dysregulation, autonomic imbalance, and low-grade systemic inflammation. These pathways may be associated with elevated blood pressure, metabolic dysregulation, endothelial dysfunction, and inflammatory activation. In the present study, participants with both sleep-duration abnormality and depressive symptoms had higher hs-CRP levels than those without either condition or with only one condition. This finding is consistent with the possibility that inflammatory activation may accompany sleep-depression co-occurrence. However, because the present analysis was cross-sectional and did not perform mediation or mechanistic modeling, hs-CRP should be interpreted as an associated inflammatory marker rather than evidence of a causal biological pathway. Behavioral pathways may also contribute to the observed associations. Individuals with both sleep disturbance and depressive symptoms may have lower physical activity, poorer diet quality, reduced treatment adherence, or other unfavorable health behaviors due to fatigue, anhedonia, and impaired motivation. These factors could partly explain the less favorable cardiometabolic profile observed in this subgroup [32–34]. Future longitudinal and mechanistic studies are needed to clarify the temporal sequence and underlying pathways. Gender differences In gender-stratified analyses, the association between sleep-depression co-occurrence and hypertension was observed in both men and women, with a larger point estimate and higher additive interaction estimates among women. These findings suggest that women with co-occurring sleep-duration abnormality and depressive symptoms may represent a subgroup with a particularly unfavorable hypertension profile. Several factors may contribute to this pattern. Biological factors, including sex hormone changes, stress-related neuroendocrine responses, and inflammatory regulation, may influence the relationship between psychological distress, sleep disturbance, and cardiometabolic health. Psychosocial factors may also play a role, including differences in coping styles, caregiving burden, socioeconomic stressors, symptom reporting, and healthcare-seeking behavior. Previous studies have suggested that women may be more likely to report somatic symptoms such as insomnia and pain in the context of depression, whereas men may be more likely to underreport psychological distress [35]. Nevertheless, these explanations remain speculative in the present study. Because NHANES data do not allow direct testing of these pathways, the observed gender differences should be interpreted as subgroup associations rather than evidence of specific biological mechanisms. Age differences and public health implications Age-stratified analyses showed that the association between sleep-depression co-occurrence and hypertension was more apparent among older adults. This pattern may reflect the combined influence of age-related physiological changes, accumulated chronic disease burden, and reduced cardiometabolic reserve. Aging is often accompanied by vascular stiffness, autonomic changes, metabolic alterations, and low-grade inflammatory activation, which may increase vulnerability to adverse cardiometabolic profiles. However, these findings should be interpreted cautiously. Older adults are more likely to have comorbid chronic conditions and to use multiple medications, and medication use was not included in the main adjustment set due to substantial missing data. Therefore, residual confounding may partly explain the stronger associations observed in older participants. From a public health perspective, these results suggest that screening for sleep-duration abnormality and depressive symptoms may be particularly relevant in older adults, but longitudinal evidence is needed before making causal or intervention-specific conclusions [36]. Clinical and public health implications Our findings have several implications for clinical practice and public health. First, they suggest that sleep-duration abnormality and depressive symptoms may need to be considered together when evaluating cardiometabolic health, particularly among women and older adults. This interpretation is consistent with recent evidence emphasizing the close association between depressive symptoms and sleep disorders in high-risk populations [37]. Second, the observed subgroup differences suggest that stratified screening and prevention strategies may be useful. For individuals with co-occurring sleep-duration abnormality and depressive symptoms, clinical assessment may consider blood pressure, obesity-related indicators, and systemic inflammatory status in addition to mental and sleep health. Third, the association with elevated hs-CRP suggests that inflammatory markers may help characterize the cardiometabolic profile of this subgroup, although the present study cannot determine whether inflammation mediates the observed associations. Overall, these findings support a more integrated approach to cardiometabolic health evaluation that incorporates sleep, psychological symptoms, and conventional risk indicators. Study innovations, strengths and limitations This study has several strengths. First, it examined the joint association of sleep-duration abnormality and depressive symptoms with cardiometabolic indicators in a large population-based NHANES sample. Second, the study distinguished between multiplicative and additive interaction and used RERI and AP to evaluate additive interaction for hypertension. Third, the analysis incorporated multiple cardiometabolic and inflammatory indicators, including blood pressure, BMI, lipid measures, and hs-CRP. Fourth, gender- and age-stratified analyses provided additional information on potential population heterogeneity. Major strengths include the use of standardized NHANES data, a relatively large sample size, and adjustment for the complex sampling design [38]. This study also has several limitations. First, its cross-sectional design precludes inferences about causality or temporal relationships between sleep-depression co-occurrence and cardiometabolic indicators. Therefore, all interpretations regarding mechanisms and effects should be understood as describing associations rather than establishing causation. Second, participants with established cardiovascular disease, including prior myocardial infarction, stroke, or heart failure, were not excluded from the primary analysis. Because these conditions may alter baseline cardiometabolic profiles, their inclusion may have introduced residual confounding. Third, depressive symptoms were defined using PHQ-9 scores rather than clinical psychiatric diagnoses, and existing psychiatric diagnoses were not separately modeled. Fourth, prescription medication use, including antidepressants, antipsychotics, sedatives, anxiolytics, and antihypertensive medications, was not included in the main adjustment set due to substantial missing data and inconsistent availability across NHANES cycles. This is an important limitation because medication use may influence sleep patterns, depressive symptoms, body weight, blood pressure, and inflammatory markers. Fifth, although diabetes history was included as an adjustment variable, other chronic conditions such as chronic kidney disease or chronic inflammatory diseases were not excluded or separately adjusted for in the primary analysis, which may also contribute to residual confounding. Sixth, participants with missing key exposure, outcome, or covariate data were excluded using complete-case analysis. If the missingness pattern was not completely random, selection bias may have been introduced. Seventh, sleep-duration abnormality and depressive symptoms were assessed using questionnaire-based measures, which may introduce information bias or exposure misclassification. Eighth, additive interaction metrics such as RERI and AP are scale-dependent and should be interpreted cautiously, particularly in a cross-sectional observational study. Finally, the outcomes analyzed in this study were individual cardiometabolic indicators and biomarkers rather than incident cardiovascular events, the 10-year ASCVD risk score, the Framingham risk score, or a validated composite cardiovascular risk metric. In addition, hypertension was defined using the ≥ 140/90 mmHg threshold or self-reported history of hypertension, which may differ from lower thresholds used in some contemporary clinical guidelines. Therefore, the findings should not be interpreted as direct evidence of increased clinical cardiovascular event risk. Future longitudinal studies with detailed disease history, medication data, repeated measurements, and clinical cardiovascular outcomes are needed to confirm these associations and clarify the underlying pathways. Conclusion In this cross-sectional NHANES analysis, the co-occurrence of sleep-duration abnormality and depressive symptoms was associated with less favorable cardiometabolic indicators, including higher prevalence of hypertension, obesity, and elevated hs-CRP. For hypertension, interaction was observed on both the multiplicative and additive scales, as reflected by the product-term P value, RERI, and AP. Stratified analyses suggested that the additive interaction for hypertension was more apparent among women and older adults. These findings support the importance of considering sleep-duration abnormality and depressive symptoms jointly when evaluating cardiometabolic health. However, causal and mechanistic interpretations require confirmation in longitudinal studies with repeated measurements, detailed medication information, and clinical cardiovascular outcomes Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Abbreviations American Heart Association Body mass index Cardiovascular diseases Centers for Disease Control and Prevention Disability-adjusted life years Global Burden of Disease High-sensitivity C-reactive protein Hypothalamic-pituitary-adrenal axis Low-density lipoprotein cholesterol Major depressive disorder National Center for Health Statistics National Health and Nutrition Examination Survey Patient Health Questionnaire-9 Poverty-to-income ratio Pittsburgh Sleep Quality Index Receiver operating characteristic Restricted cubic spline Sleep Questionnaire Socioeconomic status World Health Organization (WHO) Author contributions LC: Conceptualization, Data curation, Formal analysis, Methodology, Writing – original draft. LM: Investigation, Software. HD: Resources, Supervision, Writing – review & editing. All authors read and approved the final manuscript. Funding The research was funded by CAMS Innovation Fund for Medical Sciences (CIFMS) (2025-12 M-XHXX-034), National Natural Science Foundation of China (Grant no. 82400543). Data availability The datasets generated and/or analysed during the current study are available in the NHANES repository, https://www.cdc.gov/nchs/nhanes/index.htm. Additional data supporting the findings of this study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This study utilized de-identified, publicly available data from the National Health and Nutrition Examination Survey (NHANES). The NHANES protocol was approved by the Ethics Review Board of the National Center for Health Statistics (NCHS). All participants provided written informed consent. The analysis of this de-identified data was exempt from additional ethical review by our institutional review board. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Hongru Deng, Email: denghongru2023@163.com. Lingbing Meng, Email: menglingbing@fuwai.com. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets generated and/or analysed during the current study are available in the NHANES repository, https://www.cdc.gov/nchs/nhanes/index.htm. Additional data supporting the findings of this study are available from the corresponding author on reasonable request.