Abstract
Background
Depression is a leading cause of global disability. Physical activity (PA) has been found to be associated with lower risk of depression, but the underlying biological mediating factors remain understudied. HDL cholesterol (HDL-C) is an important biomarker with neuroprotective properties, however, its potential mediating role in the relationship between PA and depression has not been explored.
Objectives
Our study aimed to investigate the role of HDL cholesterol in the association between PA and depressive symptoms.
Methods
A total of 16,535 adults from the NHANES (2009–2016) were analyzed. Depression was assessed with the Patient Health Questionnaire-9, and data on PA and HDL-C were collected for each participant. Multivariate logistic regressions, subgroup analysis, mediation analysis, and sensitivity analysis were performed.
Results
After adjusting for covariates, individuals with sufficient PA had 57% lower prevalence of depression [odds ratio (OR): 0.43; 95% confidence interval (CI): 0.35, 0.53] compared with those reporting insufficient PA. Of note, each unit increase in HDL-C associated with 23% lower odds of depression (OR: 0.77; 95% CI: 0.65, 0.90). Mediation analysis suggested a small but statistically significant indirect effect of HDL-C in the association between PA and depressive symptoms (indirect effect: −0.03; 95% CI: −0.04, −0.01), accounting for 3.3% of the total effect.
Conclusions
PA showed a protective association with depression through mediation of HDL-C, highlighting a metabolic pathway in the association between PA and mental health. Promoting PA and optimizing HDL-C concentration may help prevent depression at population levels.
Keywords: physical activity, HDL-C, depression, mediation analysis, NHANES
Introduction
Depression is a multifactorial mental disorder and ranks among the leading causes of disability worldwide [1]. According to the WHO, >280 million people experience depression each year [2]. Depression significantly reduces patients’ quality of life and imposes a huge socioeconomic burden [3,4]. The etiology of depression is related to multiple factors such as genetics, environment, and lifestyle [5]. Physical activity (PA), as a modifiable behavioral factor, has attracted much attention for its role in alleviating depressive symptoms and improving mental health [6]. A large body of studies have shown that PA can significantly reduce depressive symptoms, improve mood, and reduce the risk of relapse [7,8]. Many biomarkers have been shown to mediate the association between PA and depression. PA may alleviate depressive symptoms by reducing the inflammatory response [9]. PA could also reduce depression by enhancing neuroplasticity [10]. Brain-derived neurotrophic factor (BDNF), a key factor in the promotion of neuroplasticity, has also been suggested to potentially ameliorate depressive symptoms through regular PA [11].
HDL cholesterol (HDL-C) is an important biomarker in lipid metabolism, with both cardiovascular and neural protective effects [12]. Both cohort and cross-sectional studies have indicated that low HDL-C concentrations could elevate the risk of depression, anxiety, and stress-related disorders [13,14]. HDL-C may have a positive effect on brain health by lowering systemic inflammation and regulating oxidative stress [15]. Despite the progress in the abovementioned areas, no research has explicitly investigated the mediating effect of HDL-C in the association between PA and depression. Although HDL-C has attracted attention for its cardiovascular metabolic benefits and possible effects on brain health, its specific contribution to the protective effect of PA on mental health has not been fully verified. This problem is particularly prominent in large-scale nationwide population studies, which provide systematic and representative evidence.
Our study aims to fill this gap by exploring whether HDL-C plays a mediating role between PA and depressive symptoms in adults using a nationally representative sample. Findings of our work may help to enhance the comprehension of the mechanisms by which PA benefits mental health and provide novel insights for developing targeted strategies to promote both physical and psychological well-being.
Methods
Data source and participant selection
This research utilized data from the NHANES, an ongoing national program led by the National Center for Health Statistics (NCHS) to evaluate the health and nutritional status of the US population. NHANES implements an intricate, stratified, multiphase sampling strategy to guarantee data representativeness and uphold high standards of quality. Trained professionals carried out the survey and gathered comprehensive information, including demographics, socioeconomic status, dietary habits, and health-related factors. All participants gave informed consent before participation. The NCHS Research Ethics Committee evaluated and authorized the study design [16].
Our study used NHANES data from 2009 to 2016 and restricted the analytic sample to participants aged ≥20 y. We chose this age threshold to align with NHANES/Centers for Disease Control and Prevention (CDC) adult lipid reporting conventions and our prespecified age strata (20–39, 40–59, and ≥60 y) [[17], [18], [19]], rather than to redefine adulthood. The initial sample included 50,588 participants. Our study excluded several groups: participants without Patient Health Questionnaire-9 (PHQ-9) data (n = 24,196), participants missing PA data (n = 8944), and individuals lacking HDL-C test (n = 913). After applying these criteria, the final sample included 16,535 adults (Figure 1).
FIGURE 1.
Flowchart of the study design and participants selection. HDL-C, HDL cholesterol; PHQ-9, Patient Health Questionnaire-9.
Measurements
Depression
The PHQ-9 was used to assess depressive symptoms, which is a reliable tool for detecting and assessing the severity of depressive symptoms within the past 2 wk [20]. It is widely used due to its excellent validity, reliability, free access, and simplicity [21]. The PHQ-9 contains 9 items, and the symptoms of each item cover the criteria for diagnosing depression as defined in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders [22,23]. Each item adopts a 4-point score between 0 (not at all) and 3 (almost every day) according to the frequency of symptoms, and the overall score spans from 0 to 27 [24]. In our study, participants with a total PHQ-9 score of ≥10 were classified as having depressive symptoms [25], and the Cronbach’s α was 0.85, indicating good internal consistency [26].
HDL-C
Participants fasted for ≥8.5 h before blood sample collection. After qualified samples were selected, they were first subjected to a magnesium sulfate/dextran solution to create a water-soluble complex with non–HDL-C part to prevent it from participating in subsequent reactions. After adding relevant reagents, polyethylene glycol (PEG) cholesterol esterase converted HDL-C esters into HDL-C. PEG cholesterol oxidase then catalyzed the production of hydrogen peroxide, which reacted with 4-aminoantipyrine and HSDA to generate purple or blue pigments. These pigments were measured at 600 nm (secondary wavelength 700 nm), and the results were collected and recorded as millimoles per liter. Finally, the participants’ serum HDL-C values were obtained. More specific laboratory test information is available on the CDC website [27,28].
PA
PA was assessed with Global Physical Activity Questionnaire [27,29], which collected information on vigorous or moderate-intensity exercise, fitness routines, and recreational activities during a typical week. All of the above activities must last ≥10 min. The formula used to calculate total PA time per week is: Total PA time (minutes per week) = (2 × days of vigorous PA × average daily minutes of vigorous PA) + days of moderate PA × average daily minutes of moderate PA [30]. In our study, participants’ PA patterns were divided into insufficient (0 ≤ total PA time < 150 min) and sufficient (total PA time ≥150 min) based on United States PA recommendation guidelines [30].
Covariates
Based on previous studies [31,32], the covariates in our study included: sex (male, female), age (20–39, 40–59, ≥60 y), educational level (less than high school graduate, high school graduate, some college or associate degree, college graduate or higher), race (Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, other races), income level [measured by the family poverty income ratio (PIR): low income, PIR ≤1.3; middle income, 1.3 < PIR < 3.5; high income, PIR ≥3.5] [33], marital status (married, widowed, divorced, separated, never married, living with a partner), and BMI, which was grouped into normal and low weight (BMI <25.0 kg/m2), overweight (25.0 kg/m2 ≤ BMI <30.0 kg/m2), obesity class I (30.0 kg/m2 ≤ BMI <35.0 kg/m2), obesity class II and higher (BMI ≥35.0 kg/m2) [34]. Individuals who consumed a minimum of 12 glasses of any type of alcoholic beverage in any given year were considered drinkers. Smokers were included in the study if they had smoked ≥100 cigarettes over the course of their lifetime. Data on coronary artery disease, diabetes, and hypertension were obtained from self-reported medical histories of patients diagnosed by professional physicians. In addition, our study also included total cholesterol (TC, millimoles per liter) as a covariate because HDL-C is an important component of TC. Covariate selection was further informed by a literature-based directed acyclic graph (DAG), which was constructed to identify sufficient adjustment sets and avoid overadjustment by illustrating the hypothesized relationships among PA, HDL-C, depressive symptoms, and the selected covariates. The DAG and supporting references are provided in Supplemental Figure 1.
Statistical analysis
To ensure the national representativeness, our study processed data using the weighting method officially recommended by NHANES. First, participants were divided into 2 groups based on their PHQ-9 score. The Mann-Whitney U-test was used to compare continuous variables with skewed distribution, and the chi-square test was used to compare differences in categorical variables between the 2 groups. To explore the association between PA and HDL-C and depression, we applied 3 multivariate logistic regression models to estimate the adjusted odds ratio (OR) and its 95% confidence interval (CI). Model 1 was a crude model; model 2 was adjusted for age, sex, race, education level, marital status, and family PIR; and model 3 was further adjusted for smoking, drinking, hypertension, BMI, diabetes, coronary artery disease, and TC based on model 2. In these models, PA was considered as a binary variable (sufficient or insufficient), whereas HDL-C was modeled primarily as a continuous variable to preserve information and maintain statistical efficiency. Quartile-based analyses (Q1, Q2, Q3, and Q4) were performed as secondary analyses to examine the robustness and potential dose–response pattern of the association across increasing HDL-C concentrations; both approaches are reported to allow comparison and enhance interpretability [35]. To evaluate the independent effects of PA and HDL-C on depression in different subgroups, our study also conducted subgroup analyses, including age group (20–39, 40–59, ≥60 y), sex (male, female), education level (less than high school graduate, high school graduate, some college or associate degree, college graduate or higher), race (Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black), family PIR (PIR <1.3, 1.3 < PIR <3.5, PIR ≥3.5), marital status (married/living with partner, widowed/divorced/separated/never married), BMI (BMI <25.0 kg/m2, 25.0 kg/m2 ≤ BMI <30.0 kg/m2, 30.0 kg/m2 ≤ BMI<35.0 kg/m2, BMI ≥35.0 kg/m2), drinking, smoking, and disease conditions (hypertension, diabetes, and coronary artery disease). Subsequently, the R “mediation” package was applied to analyze the mediation effects [36], estimating the indirect, direct, and total effects of PA on depression in both the crude model and the fully adjusted model. The mediation effect ratio of HDL-C was calculated by dividing the indirect effect by the total effect, and the 95% CI of the mediation ratio was estimated. Furthermore, we conducted sensitivity analyses of multivariate logistic regression models and mediating effects after excluding participants with missing values in any covariates. R (version 4.4.0; R Core Team, R Foundation for Statistical Computing, Vienna, Austria) was used to perform complete statistical analyses. A statistically significant finding was indicated by a 2-sided P value < 0.05.
Results
Characteristics of study participants
A total of 16,535 participants were included in our study, representing 115.7 million noninstitutionalized American adults. The mean age (SD) of the participants was 48.61 ± 18.45 y, with 49.02% being male. Among the participants, 1755 (10.6%) showed symptoms of depression. Individuals with depression showed a significantly lower level of PA compared with those without depression (P < 0.001). The serum concentration of HDL-C in participants with depression was lower than that in nondepressed individuals (P < 0.001). There were also significant differences in age group, sex, education level, race, PIR, marital status, BMI status, smoking status, coronary artery disease, and diabetes between participants with depression and those without depression (P < 0.001) (Table 1).
TABLE 1.
Weighted characteristics of the study participants. (n = 16,535)
| Characteristic | Total participants |
Depression |
Nondepression |
P value |
|---|---|---|---|---|
| (n = 16,535) | (n = 1755) | (n = 14,780) | ||
| Weighted sample size | 115,680,604 | 11,709,494 | 103,971,110 | |
| Age (y), mean (SD) | 48.61 (18.45) | 49.04 (16.61) | 48.56 (18.66) | 0.209 |
| Age group (y), n (%) | <0.001 | |||
| 20–40 | 5077 (32.2) | 479 (28.2) | 4598 (32.7) | |
| 40–60 | 5296 (33.6) | 695 (40.9) | 4601 (32.7) | |
| ≥60 | 5405 (34.3) | 524 (30.9) | 4881 (34.7) | |
| Sex, n (%) | <0.001 | |||
| Male | 8106 (49.0) | 614 (35.0) | 7492 (50.7) | |
| Female | 8429 (51.0) | 1141 (65.0) | 7288 (49.3) | |
| Educational level, n (%) | <0.001 | |||
| Less than high school graduate | 4483 (28.4) | 670 (39.5) | 3813 (27.1) | |
| High school graduate | 3647 (23.1) | 403 (23.7) | 3244 (23.1) | |
| Some college or associate degree | 4469 (28.3) | 470 (27.7) | 3999 (28.4) | |
| College graduate or higher | 3167 (20.1) | 155 (9.1) | 3012 (21.4) | |
| Race/ethnicity, n (%) | <0.001 | |||
| Mexican American | 2812 (17.0) | 293 (16.7) | 2519 (17.0) | |
| Other Hispanic | 1888 (11.4) | 264 (15.0) | 1624 (11.0) | |
| Non-Hispanic White | 6829 (41.3) | 731 (41.7) | 6098 (41.3) | |
| Non-Hispanic Black | 3428 (20.7) | 359 (20.5) | 3069 (20.8) | |
| Other races | 1578 (9.5) | 108 (6.2) | 1470 (9.9) | |
| Family poverty income ratio, n (%) | <0.001 | |||
| <1.3 | 5413 (35.9) | 891 (56.2) | 4522 (33.5) | |
| 1.3–3.5 | 5606 (37.1) | 512 (32.3) | 5094 (37.7) | |
| ≥3.5 | 4080 (27.0) | 183 (11.5) | 3897 (28.8) | |
| Marital status, n (%) | <0.001 | |||
| Married | 8032 (50.9) | 609 (35.9) | 7423 (52.7) | |
| Widowed | 1313 (8.3) | 180 (10.6) | 1133 (8.0) | |
| Divorced | 1751 (11.1) | 292 (17.2) | 1459 (10.4) | |
| Separated | 563 (3.6) | 120 (7.1) | 443 (3.1) | |
| Never married | 2817 (17.9) | 331 (19.5) | 2486 (17.7) | |
| Living with partner | 1296 (8.2) | 164 (9.7) | 1132 (8.0) | |
| BMI (kg/m2), mean (SD) | 29.32 (7.11) | 31.08 (8.54) | 29.11 (6.89) | <0.001 |
| BMI group (kg/m2), n (%) | <0.001 | |||
| <25 | 4686 (28.7) | 411 (23.8) | 4275 (29.3) | |
| 25–30 | 5294 (32.4) | 462 (26.7) | 4832 (33.1) | |
| 30–35 | 3468 (21.2) | 425 (24.6) | 3043 (20.8) | |
| ≥35 | 2896 (17.7) | 432 (25.0) | 2464 (16.9) | |
| Alcohol user, n (%) | 11,352 (70.0) | 1212 (70.2) | 10,140 (70.0) | 0.864 |
| Smoker, n (%) | 7278 (45.3) | 1034 (60.1) | 6244 (43.5) | <0.001 |
| Hypertension, n (%) | 6009 (36.4) | 859 (49.1) | 5150 (34.9) | <0.001 |
| Diabetes mellitus, n (%) | 2202 (13.3) | 365 (20.8) | 1837 (12.4) | <0.001 |
| Coronary artery disease, n (%) | 668 (4.3) | 119 (7.1) | 549 (3.9) | <0.001 |
| TC (mmol/L), mean (SD) | 4.96 (1.09) | 5.07 (1.17) | 4.95 (1.08) | <0.001 |
| PHQ-9 score, mean (SD) | 3.56 (4.58) | 14.30 (4.01) | 2.29 (2.50) | <0.001 |
| Physical activity, n (%) | <0.001 | |||
| Insufficient (<150 min/wk) | 13,002 (78.6) | 1617 (92.1) | 11,385 (77.0) | |
| Sufficient (≥150 min/wk) | 3533 (21.4) | 138 (7.9) | 3395 (23.0) | |
| HDL cholesterol (mmol/L), mean (SD) | 1.35 (0.42) | 1.31 (0.42) | 1.36 (0.42) | <0.001 |
Abbreviations: PHQ-9, Patient Health Questionnaire-9; TC, total cholesterol.
Associations of PA and HDL-C with depression
After adjusting for all covariates (model 3), individuals with sufficient PA had 57% lower odds of depression (OR: 0.43; 95% CI: 0.35, 0.53; P < 0.001) (Table 2). As a continuous variable (model 3), the odds of depression decreased by 23.0% for each unit increase in HDL-C (OR: 0.77; 95% CI: 0.65, 0.90; P = 0.001). In addition, compared with the lowest HDL-C serum concentration (Q1), participants with the highest HDL-C serum concentration (Q4) had a 27% lower likelihood (OR: 0.73; 95% CI: 0.61, 0.87; P < 0.001) (Table 2).
TABLE 2.
Associations of physical activity and HDL cholesterol with depression in adults1
| Model 1 |
Model 2 |
Model 3 |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | OR | 95% CI | P value | |
| Physical activity | |||||||||
| Insufficient (<150 min/wk) | Ref. | Ref. | Ref. | ||||||
| Sufficient (≥150 min/wk) | 0.28 | (0.23, 0.34) | <0.001 | 0.39 | (0.31, 0.47) | <0.001 | 0.43 | (0.35, 0.53) | <0.001 |
| HDL cholesterol | |||||||||
| As continuous | 0.73 | (0.64, 0.83) | <0.001 | 0.65 | (0.56, 0.76) | <0.001 | 0.77 | (0.65, 0.90) | 0.001 |
| Q1 (<1.06 mmol/L) | Ref. | Ref. | Ref. | ||||||
| Q2 (1.06–1.29 mmol/L) | 0.80 | (0.70, 0.92) | 0.001 | 0.77 | (0.66, 0.90) | <0.001 | 0.85 | (0.73, 1.00) | 0.056 |
| Q3 (1.29–1.58 mmol/L) | 0.75 | (0.65, 0.86) | <0.001 | 0.69 | (0.59, 0.81) | <0.001 | 0.81 | (0.69, 0.96) | 0.016 |
| Q4 (>1.58 mmol/L) | 0.69 | (0.60, 0.79) | <0.001 | 0.60 | (0.51, 0.70) | <0.001 | 0.73 | (0.61, 0.87) | <0.001 |
| P-trend | <0.001 | <0.001 | <0.001 | ||||||
Abbreviations: CI, confidence interval; OR, odds ratio; Q, quartile; Ref, reference.
The associations of physical activity and HDL cholesterol with depression are presented as OR (95% CI). Model 1 did not adjust for any covariates. Model 2 adjusted for age, sex, race, education level, marital status, and the family poverty income ratio. Model 3 further adjusted for total cholesterol, smoking cigarettes, alcohol use, hypertension, BMI, hypertension, diabetes, and coronary artery disease based on model 2.
The results of PA and HDL-C were verified in sensitivity analysis after excluding participants with any missing values (Supplemental Table 1).
Subgroup analysis
The results showed that PA was associated with a lower risk of depression in all subgroups except those without coronary artery disease and less than high school education (Figure 2). Of note, the protective effect was significant for the age group 20 to 39 y, females, races other than non-Hispanic Black, low (PIR <1.3) or high family income (PIR >3.5), whether living with a partner or living alone, BMI <25 or 30 ≤ BMI <35 kg/m2, drinking, whether smoking or not, and people without hypertension, diabetes, or coronary artery disease (Figure 2). In addition, there was a significant interaction between age group and the relationship between PA and depression (P-interaction = 0.031).
FIGURE 2.
Subgroup analysis of the association between physical activity, HDL cholesterol, and depression. CI, confidence interval; OR, odds ratio.
Mediation analysis
The analysis results showed that in the crude model, PA had a significant indirect effect on depression through HDL-C (indirect effect: −0.03; 95% CI: −0.04, −0.02), and the indirect effect accounted for 2.0% of the total effect (P < 0.001). After adjustment for all covariates, HDL-C accounted for a small proportion of the association between PA and depressive symptoms in the adjusted mediation model (indirect effect: −0.03; 95% CI: −0.04, −0.01), accounting for 3.3% of the total effect (Figure 3). After excluding missing values, the indirect effect was still significant (Supplemental Figure 2). In addition, the mediating effect of HDL-C was significant in people aged 20 to 59 y, females, races other than non-Hispanic Blacks, low (PIR <1.3) or high family income (PIR >3.5), individuals who were married or living with a partner, BMI <35 kg/m2, and people without hypertension, diabetes, or coronary artery disease (Supplemental Table 2).
FIGURE 3.
The mediation analysis of HDL-C in the association between PA and depression. ∗∗∗P < 0.001. CI, confidence interval; HDL-C, HDL cholesterol; PA, physical activity.
Discussion
Our study explored the mediating role of HDL-C in the relationship between PA and depressive symptoms using a nationwide sample of adults. We found that PA and HDL-C showed significant protective effects on depression, and HDL-C showed a modest indirect effect in the association between PA and depressive symptoms. Given the cross-sectional design, this finding should be interpreted cautiously and requires confirmation in longitudinal studies. In addition, subgroup analysis revealed differences in protective and mediating effects in specific populations (ages 20–59 y, females, those with BMI <35 kg/m2, and those without coronary artery disease, diabetes, and hypertension), providing a new perspective for understanding the complex interactions between PA, HDL-C, and depression.
Higher PA was associated with lower odds of depressive symptoms in our study, which is consistent with previous prospective evidence showing that physically active individuals are less likely to develop depression. A meta-analysis of prospective cohort studies by Schuch et al. [37] showed that higher levels of PA can reduce the risk of depression by 17%–41%. Previous prospective evidence suggests an inverse curvilinear dose–response association between PA and incident depression, with the greatest marginal benefits observed when moving from no activity to low-to-moderate activity levels [38]. A systematic review [39] found that adequate PA can reduce the risk of depression by 57% and further pointed out that PA may alleviate depressive symptoms through multiple mechanisms such as regulating neurotransmitters (such as serotonin and dopamine), enhancing the expression of neurotrophic factors (such as BDNF), and improving the function of the hypothalamus-pituitary-adrenal axis.
Moreover, higher concentrations of HDL-C were associated with a significantly lower risk of depression. The results of a cohort study [13] showed that for each unit increase in HDL-C, the risk of depression decreased by 12%, which was consistent with our study. The neuroprotective effect of HDL-C may be related to the following pathways: first, where HDL-C binds to paraoxonase-1 to clear oxidized low-density lipoprotein, thereby reducing oxidative damage to the central nervous system [40]. Additionally, HDL-C inhibits macrophage and microglial activation in hypercholesterolemia, lowering proinflammatory factor release and neuroinflammation [41].
Our study further suggested that HDL-C accounted for a small but statistically significant indirect effect in the association between PA and depressive symptoms, suggesting that PA may indirectly affect depressive symptoms by regulating lipid metabolism. Previous studies have demonstrated that HDL-C concentrations in healthy populations were significantly elevated after aerobic exercise [42,43]. The findings of this study extend to older demographics [44] and individuals afflicted with cardiovascular disease [45], where PA also led to a notable enhancement in functions of HDL-C. On the other hand, chronic inflammatory states have been shown to be closely related to the occurrence of depression [46]. PA may exert an antidepressant effect by dynamically regulating the inflammatory-immune balance. For example, regular exercise can reduce the concentrations of proinflammatory factors (such as IL-6 and TNF-α) and upregulate anti-inflammatory factors (such as IL-10) [47]. As mentioned above, HDL-C also reduces the risk of depression through an anti-inflammatory pathway. Given the cross-sectional design and the modest magnitude of the indirect effect, the mediating role of HDL-C should be interpreted with caution. Although the indirect effect is relatively small, recent methodological guidance also suggests that mediation results should not be evaluated solely by the proportion mediated, as this quantity may be unstable and should be interpreted alongside the size and potential clinical relevance of the indirect effect [48]. Thus, HDL-C may be regarded as one possible, but limited, pathway linking PA with depressive symptoms rather than a dominant mechanism.
Our findings also highlight the differential physiological and metabolic effects of HDL-C in different subpopulations. First, the mediating effect of HDL-C in the 20–59 y old population may be related to more active lipid metabolism and less age-related decline in HDL-C concentrations [49]. Young people generally have higher HDL-C concentrations and more efficient cholesterol transport, which may enhance the protective effect of HDL-C against depression. The significant mediation effect observed in females may be related to hormonal differences, especially the role of estrogen. Studies have shown that estrogen can positively regulate HDL-C concentrations, enhance its efflux and function, and enhance its anti-inflammatory and antioxidant functions [50], which may contribute to the stronger association of HDL-C with depression prevention in females. In addition, women generally have a higher incidence of depression [51], and estrogen can also promote the occurrence of depression [52], making the protective effect of HDL-C more significant in this subgroup.
Our study has several strengths, including a nationally representative sample; HDL-C is a laboratory indicator, and the data obtained are objective; multiple confounding factors such as age, sex, BMI, coronary artery disease, diabetes, and hypertension were fully controlled in the analysis, thereby enhancing the reliability of the conclusions. However, our study also has certain limitations. First, due to the cross-sectional data, it is difficult to determine the causal path relationship of HDL-C as a mediating variable, and the small indirect effect of HDL-C should therefore be interpreted cautiously. Second, PA and depression symptoms rely on self-reported data, which may be subject to memory bias or social desirability effects. Third, given the multifactorial nature of depressive symptoms, HDL-C is likely to represent only one of several possible pathways linking PA with depressive symptoms. Future longitudinal studies are needed to further clarify the temporal sequence and to verify the potential mediating role of HDL-C. Future studies should verify the mediating role of HDL-C through longitudinal design, and more objective assessments of depression, such as structured clinical interviews, are also warranted.
In conclusion, our study showed that HDL-C has a significant mediating role between PA and depressive symptoms. These findings provide a new perspective for exploring the metabolic mechanisms of depression and emphasize the importance of promoting PA and improving HDL-C concentrations in the prevention and treatment of depression.
Author contributions
The authors’ responsibilities were as follows – JZ: wrote, reviewed, and edited the original draft, curated data, responsible for formal analysis, visualization, methodology, software, and conceptualization; BG: wrote, reviewed, and edited the manuscript, responsible for project administration, supervision, and conceptualization; and both authors read and approved the final manuscript.
Data availability
Data in our study can be accessed directly from the NHANES website: https://wwwn.cdc.gov/nchs/nhanes/Default.aspx.
Funding
The authors reported no funding received for this study.
Conflict of interest
The authors report no conflicts of interest.
Acknowledgments
We sincerely appreciate the efforts of the Centers for Disease Control and Prevention’s National Center for Health Statistics in compiling and sharing the NHANES dataset for public research purposes.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.cdnut.2026.107691.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
References
- 1.GBD 2019 Mental Disorders Collaborators Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry. 2022;9(2):137–150. doi: 10.1016/s2215-0366(21)00395-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.World Health Organization . 2023. Depressive disorder (depression)https://www.who.int/news-room/fact-sheets/detail/depression [cited 22 December, 2024]. Available from: [Google Scholar]
- 3.Richards D. Prevalence and clinical course of depression: a review. Clin. Psychol. Rev. 2011;31(7):1117–1125. doi: 10.1016/j.cpr.2011.07.004. [DOI] [PubMed] [Google Scholar]
- 4.GBD 2021 Diseases and Injuries Collaborators Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2024;403(10440):2133–2161. doi: 10.1016/s0140-6736(24)00757-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Saveanu R.V., Nemeroff C.B. Etiology of depression: genetic and environmental factors. Psychiatr. Clin. North Am. 2012;35(1):51–71. doi: 10.1016/j.psc.2011.12.001. [DOI] [PubMed] [Google Scholar]
- 6.Rahmati M., Lee S., Yon D.K., Lee S.W., Udeh R., McEvoy M., et al. Physical activity and prevention of mental health complications: an umbrella review. Neurosci. Biobehav. Rev. 2024;160 doi: 10.1016/j.neubiorev.2024.105641. [DOI] [PubMed] [Google Scholar]
- 7.Harris E. Meta-analysis: exercise as effective as therapy for treating depression. JAMA. 2024;331(11):908. doi: 10.1001/jama.2024.1121. [DOI] [PubMed] [Google Scholar]
- 8.Heissel A., Heinen D., Brokmeier L.L., Skarabis N., Kangas M., Vancampfort D., et al. Exercise as medicine for depressive symptoms? A systematic review and meta-analysis with meta-regression. Br. J. Sports Med. 2023;57(16):1049–1057. doi: 10.1136/bjsports-2022-106282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cotman C.W., Berchtold N.C., Christie L.A. Exercise builds brain health: key roles of growth factor cascades and inflammation. Trends Neurosci. 2007;30(9):464–472. doi: 10.1016/j.tins.2007.06.011. [DOI] [PubMed] [Google Scholar]
- 10.Smith P.J., Merwin R.M. The role of exercise in management of mental health disorders: an integrative review. Annu. Rev. Med. 2021;72:45–62. doi: 10.1146/annurev-med-060619-022943. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zarza-Rebollo J.A., López-Isac E., Rivera M., Gómez-Hernández L., Pérez-Gutiérrez A.M., Molina E. The relationship between BDNF and physical activity on depression. Prog. Neuropsychopharmacol. Biol. Psychiatry. 2024;134 doi: 10.1016/j.pnpbp.2024.111033. [DOI] [PubMed] [Google Scholar]
- 12.Tran-Dinh A., Levoye A., Couret D., Galle-Treger L., Moreau M., Delbosc S., et al. High-density lipoprotein therapy in stroke: evaluation of endothelial SR-BI-dependent neuroprotective effects. Int. J. Mol. Sci. 2020;22(1):106. doi: 10.3390/ijms22010106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chourpiliadis C., Zeng Y., Lovik A., Wei D., Valdimarsdóttir U., Song H., et al. Metabolic profile and long-term risk of depression, anxiety, and stress-related disorders. JAMA Netw. Open. 2024;7(4) doi: 10.1001/jamanetworkopen.2024.4525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Morgan R.E., Palinkas L.A., Barrett-Connor E.L., Wingard D.L. Plasma cholesterol and depressive symptoms in older men. Lancet. 1993;341(8837):75–79. doi: 10.1016/0140-6736(93)92556-9. [DOI] [PubMed] [Google Scholar]
- 15.Morris G., Puri B.K., Bortolasci C.C., Carvalho A., Berk M., Walder K., et al. The role of high-density lipoprotein cholesterol, apolipoprotein A and paraoxonase-1 in the pathophysiology of neuroprogressive disorders, Neurosci. Biobehav. Rev. 2021;125:244–263. doi: 10.1016/j.neubiorev.2021.02.037. [DOI] [PubMed] [Google Scholar]
- 16.National Center for Health Statistics . 2024. NHANES. Ethics Review Board Approval.https://www.cdc.gov/nchs/nhanes/about/erb.html [cited 20 December, 2024]. Available from: [Google Scholar]
- 17.Carroll M.D., Kit B.K., Lacher D.A., National Center for Health Statistics . NCHS Data Brief No. 92; April 2012. Total and High-density Lipoprotein Cholesterol in Adults: National Health and Nutrition Examination Survey, 2009–2010.https://www.cdc.gov/nchs/products/databriefs/db92.htm Available from: [PubMed] [Google Scholar]
- 18.Carroll M.D., Fryar C.D., Kit B.K., National Center for Health Statistics . NCHS Data Brief No. 226; December 2015. Total and High-density Lipoprotein Cholesterol in Adults: United States, 2011–2014.https://www.cdc.gov/nchs/products/databriefs/db226.htm Available from: [PubMed] [Google Scholar]
- 19.Carroll M.D., Fryar C.D., National Center for Health Statistics Total and High-density Lipoprotein Cholesterol in Adults: United States, 2015–2018. https://www.cdc.gov/nchs/products/databriefs/db363.htm [Internet]. NCHS Data Brief No. 363, April 2020. Available from:
- 20.Negeri Z.F., Levis B., Sun Y., He C., Krishnan A., Wu Y., et al. Accuracy of the Patient Health Questionnaire-9 for screening to detect major depression: updated systematic review and individual participant data meta-analysis. BMJ. 2021;375 doi: 10.1136/bmj.n2183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ferenchick E.K., Ramanuj P., Pincus H.A. Depression in primary care: part 1-screening and diagnosis. BMJ. 2019;365 doi: 10.1136/bmj.l794. [DOI] [PubMed] [Google Scholar]
- 22.American Psychiatric Association . 5th ed. American Psychiatric Association; Washington, DC: 2013. Diagnostic and Statistical Manual of Mental Disorders. [Google Scholar]
- 23.Kroenke K., Spitzer R.L., Williams J.B. The PHQ-9: validity of a brief depression severity measure. J. Gen. Intern. Med. 2001;16(9):606–613. doi: 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Maurer D.M., Raymond T.J., Davis B.N. Depression: screening and diagnosis. Am. Fam. Physician. 2018;98(8):508–515. [PubMed] [Google Scholar]
- 25.Levis B., Benedetti A., Thombs B.D. DEPRESsion Screening Data (DEPRESSD) Collaboration, Accuracy of Patient Health Questionnaire-9 (PHQ-9) for screening to detect major depression: individual participant data meta-analysis. BMJ. 2019;365 doi: 10.1136/bmj.l1476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Johnson E. In: Encyclopedia of Autism Spectrum Disorders. Volkmar F.R., editor. Springer International Publishing; Cham, Switzerland: 2021. Cronbach’s alpha; pp. 1248–1249. [Google Scholar]
- 27.Centers for Disease Control and Prevention . 2024. National Health and Nutrition Examination Survey.https://www.cdc.gov/nchs/nhanes/index.html [cited 12 December, 2024]. Available from: [Google Scholar]
- 28.National Center for Health Statistics . 2017. Cholesterol - High-Density Lipoprotein (HDL)https://wwwn.cdc.gov/Nchs/Data/Nhanes/Public/2015/DataFiles/HDL_I.htm#LBDHDDSI [cited 12 December, 2024]. Available from: [Google Scholar]
- 29.National Center for Health Statistics . 2017. Physical Activity.https://wwwn.cdc.gov/Nchs/Data/Nhanes/Public/2015/DataFiles/PAQ_I.htm#PAQ620 [cited 15 December, 2024]. Available from: [Google Scholar]
- 30.Piercy K.L., Troiano R.P., Ballard R.M., Carlson S.A., Fulton J.E., Galuska D.A., et al. The Physical Activity Guidelines for Americans. JAMA. 2018;320(19):2020–2028. doi: 10.1001/jama.2018.14854. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Chen R., Wang K., Chen Q., Zhang M., Yang H., Zhang M., et al. Weekend warrior physical activity pattern is associated with lower depression risk: findings from NHANES 2007-2018, Gen. Hosp. Psychiatry. 2023;84:165–171. doi: 10.1016/j.genhosppsych.2023.07.006. [DOI] [PubMed] [Google Scholar]
- 32.Shi Y.Y., Zheng R., Cai J.J., Qian S.Z. The association between triglyceride glucose index and depression: data from NHANES 2005-2018. BMC Psychiatry. 2021;21(1):267. doi: 10.1186/s12888-021-03275-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Johnson C.L., Paulose-Ram R., Ogden C.L., Carroll M.D., Kruszon-Moran D., Dohrmann S.M., et al. National health and nutrition examination survey. Analytic guidelines, 1999-2010. Vital Health Stat. 2013;2(161):1–24. [PubMed] [Google Scholar]
- 34.World Health Organization Obesity: preventing and managing the global epidemic [Internet] World Health Organization technical report series. 2000;894 https://iris.who.int/handle/10665/42330 Available from: [PubMed] [Google Scholar]
- 35.Turner E.L., Dobson J.E., Pocock S.J. Categorisation of continuous risk factors in epidemiological publications: a survey of current practice. Epidemiol. Perspect. Innov. 2010;7:9. doi: 10.1186/1742-5573-7-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Tofighi D., MacKinnon D.P. RMediation: an R package for mediation analysis confidence intervals. Behav. Res. Methods. 2011;43(3):692–700. doi: 10.3758/s13428-011-0076-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Schuch F.B., Vancampfort D., Firth J., Rosenbaum S., Ward P.B., Silva E.S., et al. Physical activity and incident depression: a meta-analysis of prospective cohort studies. Am. J. Psychiatry. 2018;175(7):631–648. doi: 10.1176/appi.ajp.2018.17111194. [DOI] [PubMed] [Google Scholar]
- 38.Pearce M., Garcia L., Abbas A., Strain T., Schuch F.B., Golubic R., et al. Association between physical activity and risk of depression: a systematic review and meta-analysis. JAMA Psychiatry. 2022;79(6):550–559. doi: 10.1001/jamapsychiatry.2022.0609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Mammen G., Faulkner G. Physical activity and the prevention of depression: a systematic review of prospective studies. Am. J Prev. Med. 2013;45(5):649–657. doi: 10.1016/j.amepre.2013.08.001. [DOI] [PubMed] [Google Scholar]
- 40.Morris G., Gevezova M., Sarafian V., Maes M. Redox regulation of the immune response. Cell. Mol. Immunol. 2022;19(10):1079–1101. doi: 10.1038/s41423-022-00902-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Tall A.R., Yvan-Charvet L. Cholesterol, inflammation and innate immunity. Nat. Rev. Immunol. 2015;15(2):104–116. doi: 10.1038/nri3793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Berg A., Frey I., Baumstark M.W., Halle M., Keul J. Physical activity and lipoprotein lipid disorders. Sports Med. 1994;17(1):6–21. doi: 10.2165/00007256-199417010-00002. [DOI] [PubMed] [Google Scholar]
- 43.Kokkinos P.F., Fernhall B. Physical activity and high density lipoprotein cholesterol levels: what is the relationship? Sports Med. 1999;28(5):307–314. doi: 10.2165/00007256-199928050-00002. [DOI] [PubMed] [Google Scholar]
- 44.Braga P.G.S., Freitas F.R., Bachi A.L.L., Amirato G.R., Baroni R.V., Alves M.J.N.N., et al. Regular practice of physical activity improves cholesterol transfers to high-density lipoprotein (HDL) and other HDL metabolic parameters in older adults. Nutrients. 2023;15(23):4871. doi: 10.3390/nu15234871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Hernáez Á., Soria-Florido M.T., Castañer O., Pintó X., Estruch R., Salas-Salvadó J., et al. Leisure time physical activity is associated with improved HDL functionality in high cardiovascular risk individuals: a cohort study. Eur. J. Prev. Cardiol. 2021;28(12):1392–1401. doi: 10.1177/2047487320925625. [DOI] [PubMed] [Google Scholar]
- 46.Guo X., Le Y. The triangular relationship of physical activity, depression, and inflammatory markers: a large cross-sectional analysis with NHANES data. J. Affect. Disord. 2024;367:589–597. doi: 10.1016/j.jad.2024.09.008. [DOI] [PubMed] [Google Scholar]
- 47.Silverman M.N., Deuster P.A. Biological mechanisms underlying the role of physical fitness in health and resilience. Interface Focus. 2014;4(5) doi: 10.1098/rsfs.2014.0040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Twisk J.W.R. Indirect effects in mediation analyses should not be tested for statistical significance. J. Clin. Epidemiol. 2024;171 doi: 10.1016/j.jclinepi.2024.111393. [DOI] [PubMed] [Google Scholar]
- 49.Wilson P.W., Anderson K.M., Harris T., Kannel W.B., Castelli W.P. Determinants of change in total cholesterol and HDL-C with age: the Framingham Study. J. Gerontol. 1994;49(6):M252–M257. doi: 10.1093/geronj/49.6.m252. [DOI] [PubMed] [Google Scholar]
- 50.Palmisano B.T., Zhu L., Eckel R.H., Stafford J.M. Sex differences in lipid and lipoprotein metabolism. Mol. Metab. 2018;15:45–55. doi: 10.1016/j.molmet.2018.05.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Kuehner C. Why is depression more common among women than among men? Lancet Psychiatry. 2017;4(2):146–158. doi: 10.1016/S2215-0366(16)30263-2. [DOI] [PubMed] [Google Scholar]
- 52.Barth C., Crestol A., de Lange A.G., Galea L.A.M. Sex steroids and the female brain across the lifespan: insights into risk of depression and Alzheimer’s disease. Lancet Diabetes Endocrinol. 2023;11(12):926–941. doi: 10.1016/s2213-8587(23)00224-3. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data in our study can be accessed directly from the NHANES website: https://wwwn.cdc.gov/nchs/nhanes/Default.aspx.