This cross‑sectional analysis used enrollment data from the Haiti Cardiovascular Disease Cohort to assess whether depressive symptoms (DS) are associated with prevalent cardiovascular disease (CVD) in urban Haiti. The analysis included adults aged ≥18 years enrolled between March 2019 and August 2021. The authors sought to quantify prevalence differences, adjust for potential confounders, and examine sex differences in the DS–CVD relationship.
The analytic sample comprised 2,995 participants with a mean age of 41.9 years; 58.0% were female. The study was conducted in Port‑au‑Prince, Haiti, as part of an epidemiologic cohort designed to evaluate cardiovascular disease burden in a low‑resource urban setting.
Depressive symptoms were measured using the Patient Health Questionnaire–9 (PHQ‑9). Scores were categorized into two groups: none–mild (PHQ‑9 <10) and moderate–severe (PHQ‑9 ≥10). Prevalent CVD was adjudicated using epidemiologic definitions consistent with international cohort practices and included angina, myocardial infarction, transient ischemic attack or stroke, and heart failure.
Prevalence ratios (PRs) for the association between DS category and prevalent CVD were estimated using generalized estimating equation Poisson models with a log link. Models adjusted for a set of sociodemographic, behavioral, and clinical covariates: age, sex, education, income, food insecurity, smoking, alcohol use, physical activity, stress, and body mass index (BMI). The authors assessed effect modification by sex on both multiplicative and additive scales.
Overall, 16.2% of participants (95% CI: 14.8–17.5; n=484) had moderate–severe DS. Prevalence of moderate–severe DS differed by sex: 22.0% (95% CI: 19.8–23.6) in females versus 8.5% (95% CI: 7.0–10.1) in males. The prevalence of CVD was higher among participants with moderate–severe DS compared with those with none–mild DS in both sexes. Reported prevalence comparisons were: males with moderate–severe DS 21.5% vs none–mild 10.6%; females with moderate–severe DS 23.3% vs none–mild 15.3%.
After adjustment for the listed covariates, moderate–severe DS remained associated with higher prevalence of CVD. The adjusted prevalence ratio (aPR) comparing moderate–severe to none–mild DS was 1.36 (95% CI: 1.08–1.71), indicating a statistically significant association in this cross‑sectional sample.
In sex‑stratified adjusted models the aPR was 1.38 (95% CI: 1.06–1.78) for females and 1.25 (95% CI: 0.75–1.99) for males. Although point estimates suggested a stronger association among women, formal evidence for interaction by sex on the additive scale was limited; the reported relative excess risk due to interaction (RERI) was 0.07 (95% CI: −0.74 to 0.88). The authors therefore describe results as consistently stronger among women but note limited statistical evidence of effect modification by sex.
The study reports an independent association between moderate–severe depressive symptoms and higher prevalence of cardiovascular disease in an urban Haitian cohort, with higher DS prevalence among women. The authors suggest that integrating depression screening and management into CVD prevention efforts could help address both conditions in resource‑limited settings. They emphasize the need for prospective studies to clarify temporal relationships and underlying mechanisms linking depression and CVD in this population.
As a cross‑sectional analysis, the study cannot establish temporality or causation between depressive symptoms and prevalent CVD. The article is presented as a preprint and has not been peer reviewed; readers and clinicians should interpret findings accordingly. The authors call for longitudinal and mechanistic studies to better understand causal pathways.
The research received approval from the Institutional Review Board of Weill Cornell Medicine and the Ethics Committee of GHESKIO. All participants provided written informed consent and the study received community stakeholder input. The authors declared no competing interests. Funding sources included the National Institutes of Health (R01HL143788). This manuscript is a medRxiv preprint and has not undergone peer review.