Introduction

Investigators over the last two decades have produced an increasing body of knowledge on racial/ethnic hypertension disparities [1–4] and hypothesized that depression plays a mediating role in elevated blood pressure among adult women. [5–10] Depression and depressive symptoms are risk factors for hypertension (HTN) and cardiovascular disease (CVD). However the pathophysiologic mechanisms linking depressive symptoms to cardiovascular outcome has yet to be understood among Hispanic women. [11–14] Depression and anxiety are more common in women than men, increase with age, and menopause doubles the risk of hypertension even after adjusting for factors such as age and body mass index. [5,15] Several hypotheses have been proposed to explain this relationship: (1) Hypertension may be a mechanism through which depressive symptoms influence CVD pathogenesis; [7] (2) depression may contribute to hypertension, or conversely (3) women who are hypertensive may exhibit higher rates of depression. [15] Hispanic women (specifically of Puerto Rican and Mexican ancestry) across the life course have lower-income, less education, lower perceptions of social support that contribute to higher rates of depression and depressive symptoms compared to non-Hispanic White and other Hispanic subgroups. [16–19] A paucity of research exists that investigates the association between incident prehypertension and hypertension with depression among postmenopausal Hispanic women, and few include multiple psychosocial stressors, or use a longitudinal study design. [6]

Women’s response to life stressors play a role in the higher rates of depression. High rates of depression for Hispanic women include daily stressors such as living in low-resourced neighborhoods, and low social support with high levels of caregiving. [16,20,21] Hispanic women are more likely to experience stressful life events, less likely to receive mental health services, least likely to receive psychiatric medication during treatment, [16,22, 23] and are less likely to be aware of their HTN; receive treatment and or have their HTN controlled compared to other racial/ethnic women. [24] Comprehensive reviews of the role of psychosocial factors, namely low levels of social support, [25] stressful life events, low optimism, depression, and increased rates of caregiving responsibilities, [8,26–31] provide convincing evidence of their link with higher levels of systolic blood pressure. [6,14] Although social support has been shown to reduce stress, improve adherence to medication, and buffer disease severity and depression, [32,33] increased care giving responsibilities of family members and a pessimistic attitude play a significant role in predicting hypertensive risk. [34,35] These psychosocial predictors of depression have not been included in most prior epidemiologic studies. The contribution of this study is its ability to assess known psychosocial correlates of depression that have a significant association with depression and hypertension. We seek to assess if an association exists between baseline depression and incident hypertension at 3 years follow-up among postmenopausal Hispanic women.

Methods

The Women’s Health Initiative (WHI) is a large, multiethnic, 40-center study funded by the National Heart, Lung, and Blood Institute (NHLBI)that focuses on strategies for preventing heart disease, breast and colorectal cancer, and osteoporotic fractures in postmenopausal women. The total sample includes 161,808 women aged 50–79 years. Self-reported race/ethnicity includes 82.5% non-Hispanic White; 9.0% Black or African American; 4.0% Hispanic, predominantly Mexican American; 2.6% Asian or Pacific Islander; and 1.5% other. For the analyses, we include 4,680 Hispanic women who participated in the WHI observational and clinical trial studies at baseline (1994–1998) and at third-year follow-up and for whom blood pressure was measured at baseline and year 3. Data were collected during a baseline screening visit and during a third year visit; the data included physical measurements and questionnaires related to medical history, family history, and behavioral factors. A full description of the WHI study is presented elsewhere. [36,37] The dependent variables were prehypertension and hypertension analyzed in separate models. Blood pressure was measured at the first screening clinic visit by certified staff with the use of standardized procedures and instruments; it was measured in the right arm with a mercury sphygmomanometer after the participant was seated and had rested for 5 minutes. [36] The cuff, was inflated to 30 mm Hg above palpated systolic blood pressure (SBP). SBP was defined as the pressure level at which the first of two Korotkoff sounds occurred in appropriate rhythm and diastolic blood pressure (DBP) was the phase V Korotkoff value. The average of 2 blood pressure readings obtained at least 30 seconds apart, was used for analysis. Hypertension was defined in women who reported that they were told by a doctor they had high blood pressure, and/or those who were prescribed medications for hypertension, and/or those whose systolic blood pressure was ≥140 mm Hg, and/or those whose diastolic blood pressure was ≥90 mm Hg. [1] Participants with prehypertension were those women whose systolic/diastolic blood pressure was 120-139/80-89 mm Hg with no self-report of medication prescribed for hypertension as defined by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC7). [38] Participants with normal blood pressure had systolic/diastolic blood pressure < 120/80 mm Hg and no self-report of medication for hypertension.

At baseline, participants completed a psychosocial questionnaire using 6-items from the Center for Epidemiological Studies Depression Scale (CES-D) scale to assess depressive symptoms in the past week and two items from the Diagnostic Interview Schedule (DIS) regarding a history of depressive symptoms in the past year. [19,39,40] Women were asked in the CES-D to report how often, in the past week, they had felt depressed; had restless sleep; enjoyed life; had crying spells; felt sad; and felt that people disliked them. Items were scored as: rarely or none of the time (<1 day); some or a little of the time (1–2 days); occasionally or a moderate amount of time (3–4 days); and most or all of the time (5–7 days). The 2 items derived from the DIS asked: “‘In the past year, have you had 2 weeks or more during which you felt sad, blue, or depressed, or lost pleasure in things that you usually cared about or enjoyed?”; ‘‘Have you had 2 years or more in your life when you felt depressed or sad most days, even if you felt okay sometimes?”; and ‘‘If yes, have you felt depressed or sad much of the time in the past year?”

Based on prior research, [9,19,41] we constructed three measures of depression. The CES-D items measured current depressive symptoms at baseline and were scored from 0 to 3, with a score of 0 indicating the participant never or rarely experienced the depressive symptoms and a score of 3 indicating the symptom was experienced most or all of the time. A summary measure of current depressive symptoms at baseline summed the responses to the CES-D and categorized a score of 5 or higher (of a possible total 18) as indicative of symptoms of depression. In a second measure, we combined the DIS items, so that a woman who responded ‘yes’ to both questions was considered to have a prior history of depression. We constructed a third measure that included items from the CES-D and DIS and a logistic regression algorithm [39] was used to compute the Burnam score indicating probability of depression. The Burnam score ranges from 0 to 0.99 with a score higher than 0.06 defining high depressive symptomatology. [19] This threshold score is not itself a measure of clinically diagnosed depression but is well correlated with clinical depression. [41] We coded antidepressant use at baseline (yes/no) from a detailed medication inventory at that visit.

Our assessment of psychosocial factors included four validated scales that measured stressful life events, social support, optimism, and caregiving responsibilities. An index of stressful life events was created with 11 items, derived from the Alameda County Epidemiologic Study, [42] and modified for the Beta Blocker Heart Attack Trial. [43] A summary score (range: 0–33) was computed as the count of the number of events weighted by the participant’s appraisal of stress attributed to each event with higher scores indicating a greater number of stressful events. The stress-weighted number of life events was divided into four categories of events (0, 1–2, 3–4, 5–33). Social support was assessed using 9 items from the Medical Outcomes Study questionnaire. [44] Participants ranked on a 5-point scale how often emotional, affection, and tangible and positive interaction support domains were available. The total summary score ranged from 9 to 45, where a higher score indicated more social support. For analysis, we recoded the scores as lower (9 to 33), medium (34 to 40), and higher levels of support (41 to 45). Optimism was measured using the Life Orientation Test-Revised (LOT-R) six item scale. [45] Responses were coded on a 5-point scale, and categorized as low (6–23), medium (24, 25), and high (26–30). Caregiving responsibility was based on a two-part item derived from the Cardiovascular Health Survey (none/less than once a week, 1–2 times per week, ≥3 times per week). [46]

All participants provided written informed consent. Institutional review board approval was obtained from each of the participating study centers and from the Fred Hutchinson Cancer Research Center, which currently serves as the IRB of record for the WHI.

Results

The mean age at baseline of participants by hypertension status was: normotensive: 58.3 years (SD 6.17), prehypertensive: 60.2 years (SD 6.7) and hypertensive: 62.3 years (SD 6.8). Prevalence of prehypertension and hypertension at baseline was lower in participants with some college education or higher compared to those with a lower education attainment (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline Characteristics of Hispanic Women in the WHI Cohorts, by Baseline Hypertension Status.

https://doi.org/10.1371/journal.pone.0152765.t001

Women enrolled in WHI at an older age, those with insurance, a BMI ≥ 30, taking high cholesterol and diabetes medications, a family and/or personal history of cardiovascular disease were more likely to have prehypertension or hypertension at baseline. Most participants were never or past smokers and were slightly more likely to have prehypertension or hypertension. However, women who had <7 drinks per week were less likely to have prehypertension or hypertension. Participants with ≥ 2 episodes of moderate-strenuous physical activity ≥ 20 minutes or ≥ 11.75 MET/hr/week were less likely to have prehypertension or hypertension. Of 4,680 Hispanic women, 26% met current depression criteria and 19% had a history of depression at baseline. About 12% reported both current depression and history of depression. There were no differences in baseline history of depression, current depressive status, or use of antidepressants at baseline across hypertension categories. Finally, there were no differences across measures of stressful life events, social support, optimism and caregiving by hypertension status.

After adjustment for the full set of demographic and clinical variables, we observed a modest association between a history of depression and prehypertension at baseline (OR 1.27, 95% CI: 1.01, 1.61). Age-adjusted multinomial models found an association between baseline depression (measured by shortened CES-D/DIS (score > = 0.06) and baseline hypertension (OR 1.25, 95% CI: 1.04, 1.51) (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Associations of Depression with Baseline Prehypertension and Hypertension.

https://doi.org/10.1371/journal.pone.0152765.t002

A similar association between history of depression and baseline hypertension was observed (OR 1.23, 95% CI: 1.02, 1.49). These findings were attenuated to the null in fully adjusted models. No associations were found between baseline hypertension status and measures of stressful life events, social support, optimism and caregiving (not shown), nor with baseline depression and incident prehypertension at year 3. Incident hypertension at year 3 was associated with baseline depression after adjustment for demographic characteristics (age, education), health care insurance, and behavioral factors (OR 1.74, 95% CI: 1.10, 2.74), but the association was attenuated after further adjustment for clinical factors such as body mass index (BMI) (OR = 1.53; 95% CI 0.95–2.46) (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Association of Baseline Depression ^a and Incident Prehypertension and Hypertension at Year 3 in Women who were Normotensive at Baseline.

https://doi.org/10.1371/journal.pone.0152765.t003

While the psychosocial variables of social support, optimism and stressful life events were not found to be associated with baseline hypertension, they were associated with depression and therefore could be effect modifiers on the relationship of depression with prehypertension and hypertension (p < .001). We therefore stratified our results by social support, optimism, and stress life events and evaluated the association of depression on incident prehypertension and hypertension. Stratified analyses by social support, optimism, and stress life events and the association of depression on incident prehypertension and hypertension demonstrated no effect modification by these factors.

Discussion

Our analyses reveal that Hispanic women have high levels of baseline depression relative to other Hispanic cohorts, [48] and education moderates these depression rates. In this study, baseline depression and history of depression were not associated with incident prehypertension or hypertension. However, psychosocial variables (life events, social support and optimism) were associated with baseline depression in unadjusted analyses. This is the first study to include additional psychosocial variables in analyses of hypertension risk and these psychosocial variables have been observed to be associated with depression among urban, racial/ethnic and poor women, [49] and with cardiovascular events. [7,50] Among Hispanics, high rates of depression compared to other racial/ethnic groups nationally are associated with history of CVD, and increasing number of CVD risk factors that strongly correlate with psychosocial stressors and socioeconomic status. [10,17]

Hispanic women across the life course have consistently demonstrated high rates of depression (ranges from 22.3% to 43.1%) that are associated with increased hypertension risk.[8,10,17,51] We observed only modest associations of depression at baseline and incident hypertension in 3-year follow-up. Our findings are consistent with prior longitudinal studies [52] which have found mixed associations between baseline depressive symptoms and incident hypertension. The Whitehall II cohort [53] and CARDIA demonstrated a positive association [54] while the Multi-Ethnic Study of Atherosclerosis (MESA) cohort found small associations between baseline depressive symptoms and increases in blood pressure but no statistically significant association with incident hypertension. [51] Although our study is unique by focusing only on Hispanics, it is unknown if the differences in our findings could be attributable to differences in age distributions in other population cohorts, follow-up time, depression measure used, or possibly socio-ethnic differences in the experience and reporting of depression.

The psychosocial variables considered shed some light on the complex yet indirect moderating role that these factors may exert on hypertension risk and confirm the findings of other studies. One study of 2,564 Mexican Americans aged 65 or older showed an association between high positive emotion and lower blood pressure among older Mexican Americans. [55] In contrast, pessimism heightens risk for self-reported, treated and incident hypertension, [56] and significant interactions are reported between pessimism, socioeconomic status (SES) and hypertension. [30,57] In our study, the majority of respondents who reported low rates of social support (61–63%) also had the highest rates of current depression and history of depression at baseline suggesting that social support may not be able to “buffer” stress from adverse life events. [58]

A recent study of 15,864 Hispanics confirms the association of depression with education level and Hispanic subgroup with higher rates observed for those with less than a high school education among those aged 45–64 similar to our study. [17] Another study of Hispanic women (<high school) found that 11% reported at least one major depressive disorder in the past 12 months while 12% rated their mental health status as fair or poor. [22] In our study about one-third of women with less than a high school education reported a history of (24.5%) and current depression (35.3%) compared to respondents who are college graduates or above (19.8% and 14.5% respectively). Unadjusted analyses confirm that education was inversely associated with depression with a significant decrease in rates of depression as education increases. These high rates of depression may be due to differences in awareness and reporting of depressive symptoms that are influenced by higher educational attainment and income, and higher rates of health care insurance coverage that may increase use of mental health services and psychotropic medications. [59] Alternatively Hispanic women’s higher prevalence of depression may reflect more exposure to stressful workplace events, discrimination [60,61] and living in unsafe neighborhood contexts. [62] In our study, higher levels of current depression at baseline were associated with lower levels of antidepressant use (7%) and lack of health insurance (15%). Moreover, Hispanic patients may have their medications less intensified than other racial/ethnic groups.[63]

Our study has several limitations. The Hispanic sample is predominantly of Mexican-origin and these data may not be representative of other Hispanic subgroups. All women were postmenopausal and our findings cannot be generalized to younger reproductive age women. This cohort were healthier at baseline than women of this age in the U.S. population, as indicated by lower baseline prevalence of diabetes, hypertension, CVD, high cholesterol levels that required medication, and cigarette use. [64] Education levels were higher among WHI participants and fewer women reported no leisure-time physical activity compared to other cohorts with Hispanics. These demographic and health findings may be attributed to the voluntary recruitment strategy that was used in the WHI studies. Although we accounted for traditional CVD risk factors in multivariable regression models, we were unable to include all possible confounders thus preventing any causal conclusions about incident hypertension. Additionally, we did not include use of medication in management of prehypertension or initial drug therapy with compelling indications as suggested by the JNC7 report. Notwithstanding these limitations, our longitudinal study is the largest study to date of older Hispanic women which included physiologic, behavioral and psychosocial moderators of depression.

Future studies need to be mindful of evaluating psychosocial factors such as social support to assess its moderating effect by SES and to parse out SES by nativity and ethnic background/ancestry to disentangle effects of depression on hypertension, [4,65] as Hispanic women report less control over their lives and experience more negative life events than men and majority groups. [66] Yet, questions remain regarding what are the direct and indirect effects of psychosocial variables on hypertension development, management and control for Hispanic populations. [6,14] Primary prevention is an important goal in the detection and management of HTN among Hispanic populations and differences by nativity, geographic region and Hispanic subgroup are significant moderators of hypertension. [3,17] We found low rates of antidepressant medication usage that confirms other studies suggesting an additional clinical intervention to improve the higher rates of depressive symptoms among Hispanic women. [17] Further analyses that included psychosocial variables correlated with depression did not alter findings suggesting the complex role these factors may exert on cardiovascular health. Elucidating the complex pathways of depression’s effect on hypertension may lead to appropriately tailored interventions among Hispanic women.

Acknowledgments

We gratefully acknowledge the initial contributions of Dr. Darrell Gaskin for concept and design, Dr. Guangyu Zhang for analysis and those of Laura A. Logie, PhD who assisted with review of literature. We thank the reviewers for all their thoughtful comments and suggestions. We acknowledge Women’s Health Initiative Centers and their principal investigators for their participation in this research, and the dedicated and committed participants of the WHI who made this study possible.