The mean age of 14,561 participants was 44.6 years. About half of the participants were female, and 67% were non-Hispanic whites. The prevalence of inadequate, average, and optimum CVH was 6.1%, 59.7%, and 34.2%; 14.9% and 7.8% adults had mild and moderate or severe depression, respectively. Only 2.4% of participants met the ideal diet criteria. The prevalence of participants meeting the ideal level for the rest of CVH metrics were smoking, 78%; diabetes, 75%; total cholesterol, 50%; blood pressure, 45%; physical activity, 40%; BMI, 30%.

Table 2 shows the baseline characteristics of participants by the status of depressive symptoms. Women, non-Hispanic blacks, other race or ethnicity group, and participants with less than high school attainments were more likely to have self-reported depression. The prevalence of mild or moderate or severe depressive symptom was significantly lower among non-Hispanic whites and among participants with more than high school education. Ideal health behaviors, including normal weight, never smoking, healthy diet and active physical activity, and ideal diabetes status, were significantly associated with lower prevalence of mild or moderate/severe depressive symptom, whereas blood pressure and total cholesterol status were not.

Compared with those with available CVH metrics and depression scores, participants with missing CVH metrics and depression scores were younger, more likely to be women, non-Hispanic blacks or “other” race or ethnicity group, and have education less than 12 years (Supplemental Table 1).

Compared to participants with optimum CVH, the PRs for moderate or severe depression were 4.39 (95% CI, 3.32–5.80) and 2.64 (2.15–3.24), respectively, for those with inadequate and average CVH. The corresponding PRs for mild depression were 2.11 (1.77–2.52) and 1.36 (1.19–1.55; Table 3). Compared to participants meeting 3–4 ideal health behaviors, the PRs for moderate or severe depression were 4.97 (3.12–7.94) and 2.55 (1.54–4.22), respectively, for those meeting 0–1 or 2 ideal health behaviors. The corresponding PRs were 1.80 (1.47–2.21) and 1.30 (1.04–1.61) for mild depression. Ideal health factors were not associated with depressive symptom (Table 3).

Figure 1 showed adjusted PRs (95% CI) for mild and moderate or severe depressive symptoms by individual CVH component comparing poor or intermediate to ideal health. In general, individual health behaviors were associated with moderate or severe or mild depressive symptom, whereas individual health factors were not except for diabetes status.

Table 4 presented the PRs of depression for inadequate and average health compared with optimum health by subgroups. The associations were largely consistent by age, gender, race or ethnicity, and education subgroups.

In a sensitivity analysis, we included participants with CVD or cancer, and results remained largely consistent with PRs excluding participants with a history of CVD or cancer (Supplemental Tables 2 and 3).

In this nationally representative sample of adults in the United States, our analysis showed a strong and graded association between CVH metrics and both mild and moderate or severe depressive symptoms among U.S. adults. Having fewer ideal CVH metrics was associated with greater risk of having depressive symptoms. The associations were consistent across various subgroups.

Our results were consistent with several previous studies. In the Aerobics Center Longitudinal Study, Espana-Romero et al. followed 5110 participants for a mean period of 6.1 years and reported that ideal CVH, especially health behaviors, showed an inverse relationship with depressive symptoms [21]. However, most participants were male (80%) and Caucasian, relatively well-educated, and from middle to upper socioeconomic strata. A cross-sectional study of 6851 Chinese adults suggested better CVH was associated with a lower prevalence of depression, particularly among male and younger Chinese persons [22]. Participants included in this study were those with relatively high income and education. Other studies suggested that the association might be bidirectional. The Reasons for Geographic and Racial Differences in Stroke Study [27] and Mathews et al. [28] have reported that people with depressive symptoms were less likely to have optimal or adequate CVH.

The associations between depression and CVH metrics appeared to be stronger for the CVH behaviors (physical inactivity, smoking, obesity, and diet) than health factors (cholesterol, diabetes, and blood pressure). It could be that the relationship between depression and CVH behaviors is more immediate and more likely to be captured in a cross-sectional survey such as NHANES compared with the relationship between depression and health factors. For example, an episode of major depression might immediately impact a person’s likelihood of engaging in physical activity (thus leading to a stronger apparent association in a cross-sectional study), whereas a potential effect on blood glucose could take more time to develop (such as if a person experiencing depression skips a primary care appointment and thus does not have a renewed prescription for glucose-lowering medication when needed, months later). Also, it should be noted that bidirectional relationships between CVH metrics and depression are possible. Depression may lead to poorer CVH metrics, and also poor CVH metrics (such as inadequate physical activity or existing chronic conditions such as hypertension or diabetes) may lead to depression. For example, evidence suggests that having depression may increase the risk of developing diabetes, and having diabetes may increase the risk of developing depression, perhaps pertaining to increased stress and financial burden associated with having a chronic condition [29,30].

Many prior studies have examined associations between individual CVH behaviors and depression. Studies have consistently reported that physical activity has an inverse association with depressive symptom [31-33]. Physical activity likely produces endorphins (the chemicals in the brain that act as natural painkillers) and also improves the ability to sleep, which, in turn, reduces stress and depressive symptoms [34,35]. Many studies have found an association between cigarette smoking and depression, with results replicated in adolescents, adults, and the elderly [36,37]. Nicotine might damage certain pathways in the brain that regulate mood and might trigger mood swings. Nicotine’s influence on neuro-transmission pathways implicated in affective disorders provides a potential mechanism for such a relationship [38-40]. Studies have also suggested that obesity may have a direct relationship with depressive symptoms [41]. Putative mechanisms involve behavioral, physiologic, and genetic pathways, as well as iatrogenic effects of medications [42]. Nutrition factors have been found to be associated with depression symptoms; for example, Mediterranean diets might confer protection against the development of depression [43]. Consistent with the findings of previous studies, our results suggested that individual health behaviors appeared to have a stronger association with depressive symptoms in this nationally representative sample.

Few studies have examined the relationship between biological health factors and depressive symptoms. A meta-analysis showed a negative correlation between total cholesterol and depression [44]. Our study demonstrated this association for mild, but not for moderate or severe depression. Our study found the relationship between poor diabetes health and depression, which is consistent with the findings by Marano et al. [45]. However, other studies did not find associations between total cholesterol or fasting glucose and depression [21,22]. For the association between blood pressure and depression, several studies showed low or ideal blood pressure was associated with psychological or depressive symptoms [22,46,47]. A proposed mechanism for this relationship is that neurons that control blood pressure by releasing neuropeptide Y might both lower blood pressure and induce anxiety [48].

To our knowledge, this is the first study to assess the association between CVH metrics and depressive symptoms in a large nationally representative sample of U.S. adults. In addition, we were able to conduct a detailed assessment of depression and seven CVH metrics and control for several important confounding variables using NHANES data.

Our study has several limitations. First, HEI-2010 was from the first-day 24-hour dietary recall, which might not reflect individuals’ usual dietary intakes. A previous validation study using 24-hour dietary recalls suggested that energy intake may be under-estimated by as much as 11% [49]. Second, severely depressed persons may have disproportionately chosen not to participate in the survey or health examination, which included administration of the PHQ-9; therefore, the prevalence estimates in our study may underestimate the actual prevalence of depression. In addition, participants being successfully treated for depression would not be identified as having depression by the PHQ-9. Third, PHQ-9–based depressive symptoms, rather than clinical depression, can be highly variable. There could be differences in interpreting and answering the questions, especially for participants with different backgrounds and beliefs. However, PHQ-9 scores of 10 or higher have a sensitivity of 88% and a specificity of 88% for major depression [26]. Fourth, we cannot fully exclude potential effects of unmeasured confounding factors, such as anxiety disorder, general- and work-related stress, and so on, and the association between CVH and depression could be overestimated in our analysis. Fifth, smoking, physical activity, diet, and alcohol use were self-reported and subject to misclassification and recall biases. The association between CVH and depression might be over- or under-estimated due to the misclassification and recall biases of the self-reported variables. Sixth, we used HbA1c as a proxy of fasting glucose, which might misclassify the status of diabetes for some participants. Also, we used HEI-2010 as a proxy for the AHA’s healthy diet score. It is not clear how accurately the HEI-2010 healthy diet classification compared with the AHA five-component healthy diet score. However, a prior study suggested that the HEI-2010 can reliably detect the meaningful differences in diet quality in a population [50]. Furthermore, participants with missing CVH metrics and depression scores were younger, more likely to be women, non-Hispanic blacks or “other” race or ethnicity group, and have education less than 12 years. The association between CVH and depression could be biased due to the missing CVH metrics and depression scores. Finally, our study was cross-sectional; thus, the associations between CVH and depression cannot be interpreted as directly causal.

In conclusion, depression is common among U.S. adults and might be an independent risk factor for CVD. Our results suggested that there was a graded association between CVH metrics, particularly for the health behaviors, and both mild and moderate or severe depressive symptoms among U.S. adults. Our analyses extend the findings from previous studies by comprehensively assessing this relationship in a large, diverse, and nationally representative sample. Given the relationship between CVH and depression, public health efforts to achieve the AHA 2020 goals of improving CVH by 20% by 2020 might be supported by efforts to reduce the prevalence of depression.