High level of depressive symptoms as a barrier to reach an ideal cardiovascular health. The Paris Prospective Study III

We hypothesized that depression might represent a barrier to reach an ideal cardiovascular health (CVH) as estimated by the 7-item tool proposed by the American Heart Association. Between 2008 and 2012, 9,417 subjects 50–75 years of age were examined in a large health center and enrolled in the Paris Prospective Study III (PPS3). Participants with 0–2, 3–4 and 5–7 health metrics at the ideal level were categorized as having poor, intermediate and ideal CVH, respectively. Participants with a score ≥7 on the 13-item Questionnaire of Depression 2nd version, Abridged or who were on antidepressants were referred as having high level of depressive symptoms (HLDS). The mean age of the 9417 study participants was 59.57 (SD 6.28) years and 61.16% were males. A total of 9.55% had HLDS. Poor, intermediate and ideal CVH was present in 40.38%, 49.52% and 10.10% of the participants. In multivariate polytomous logistic regression analysis, HLDS was inversely associated with ideal CVH (odds ratio = 0.70; 95% CI: 0.55;0.90). This was driven by an association with the behavioural component of the CVH. Participants with HLDS had a substantial reduced chance of reaching an ideal CVH.


Food Frequency questionnaire
Sugar sweetened-beverages were estimated through questions related to the daily amount of sweetened drinks (never, less than 0.5 liters, 0.5 to 1 liter per day). Daily sodium intake was estimated through questions related to the habitual composition of breakfast (none or only a drink, continental and English breakfast), the daily consumption of butter (never, once, two to four times and more than four times per day), dairy products (never, once, twice and more than twice a day), bread (never, less than 50g, 50 to 200g and more than 200g per day), meat (never, less than 100g, between 100 and 200g and more than 200g per day), cheese (never, one portion, two to three portions, and more than three portions per day), and weekly consumption (never, less than twice, two to four and more than four times a week) of eggs, cooked meats and starchy food, using the References manual of Food (ANSES : French Agency for Food, Environmental and Occupational Health & Safety, www. anses.fr).
Information on the weekly consumption (never, less than twice, two to four and more than four times a week) of fish (all types), vegetables and fresh fruits were converted in average daily consumption.

Sensitivity analyses Linear regression analysis
In exploratory descriptive statistics, a linear trend was observed between depressive symptoms score and CVH score in participants with a depressive symptoms score ≥7 only.
We therefore run linear regression analysis in this subgroup and found that each additional point of the depressive symptoms score (from 7 to 13) was associated with a significantly lower global ideal cardiovascular health score (β =-0.05, p=0.004), behavioural score (β =-0.04, p=0.002) but not biological score (β =-0.009, p=0.32) after adjustment for age, sex, 1 0 global (β =-0.36, p<0.001) and the behavioural cardiovascular health score (β =-0.36, p<0.001) but not with the biological score (β <0.001, p=0.99) in linear regression analysis adjusted for age, sex, living alone and education.

Imputations of missing items of cardiovascular health
Missing data including depressive symptoms score, confounding variables (age, sex, education, living alone) and each health metric of the CVH were simultaneously imputed (15 times each) by multiple imputations with chained equations assuming that the data were missing at random. Predictive mean matching (continuous variable), logistic regression

Stratified analysis by country of birth of the participants and of their parents
These analyses were conducted only for the behavioural component of the CVH. The prevalence of HLDS and ideal CVH, together with the association between HLDS and ideal CVH by country of birth of the participants (proxy for the migrant status) are reported on supplementary Table 2. Despite some heterogeneity in the distribution of both HLDS and ideal CVH across countries of birth, the association between HLDS and ideal CVH was fairly consistent across strata. The prevalence of HLDS and ideal CVH, together with the association between HLDS and ideal CVH by country of birth of the parents of the participants (proxy for ethnicity) are reported on supplementary Table 3. Despite some heterogeneity in the distribution of both HLDS and ideal CVH across the ethnic groups, the association between HLDS and ideal CVH was fairly consistent across strata.

Cardiovascular health as a contributor to high level of depressive symptoms ?
When exploring the association between HLDS and CVH the other way around, we found   The country of birth of the participants was missing for n= 288 individuals and 9 subjects only were born in Oceania, so that analysis was carried out in n=9120 participants * High level of depressive symptoms defined as Questionnaire of Depression 2nd version, Abridged depressive symptoms score ≥ 7 or being under antidepressant. † Odds ratio and 95% Confident Interval were obtained by polytomous logistic regression and were adjusted for age and gender education and living alone status.