Vascular risk factors for depression and apathy | Part 1 50 Potential confounders (covariates) Age, sex, general health-related variables (functional limitations and cognitive impairments), and established stroke risk factors (smoking, obesity, diabetes mellitus, and hypertension) were considered potential confounders and as such were included in the analyses 18. Functional limitations were scored as none, 1, or >=2, using a 3-item questionnaire 27. Cognition was measured with the Mini-Mental State Examination (MMSE) 28. The variable smoking included current smoking. Obesity was defined as a body mass index of 30 kg/ m2 or greater 29. Diabetes mellitus (yes/no) was based on self-report data, the use of antidiabetic agents, or a GP diagnosis 17. Blood pressure (mm Hg) was measured with an oscillometric blood pressure monitor (HEM-706; Omron Corporation, Tokyo, Japan) after 5 minutes of rest. Out of the 3 measurements, a mean systolic blood pressure of 140-159 mm Hg or a mean diastolic blood pressure of 90-99 mm Hg was categorized as stage 1 hypertension. A mean systolic blood pressure of >=160 mm Hg or a mean diastolic blood pressure of >=100 mm Hg was categorized as stage 2 hypertension 30. Antidepressant use was established by visually checking all of the participants’ medications during interview at their homes. Statistical methods Differences between groups were explored by calculating descriptive statistics (e.g., means, SDs, and frequencies) and performing t tests for continuous measures with normal distributions, Mann-Whitney U tests for continuous measures with skewed distributions, and [chi]2 tests for categorical variables. We checked the primary variables for normality, collinearity, and proportionality of hazards. Neuroticism was not normally distributed; therefore we classified respondents as low or high on neuroticism based on the median split (=5) in order to prevent influential outliers from affecting results. We also performed sensitivity analyses by repeating all analyses on the log-transformed continuous neuroticism score. The predictive effect of depression on incidence of stroke was tested with multiple Cox regression analyses with time to a fatal or nonfatal stroke as the dependent variable and corrected for age, sex, global cognitive functioning (MMSE score), one or more functional limitations, smoking, hypertension (stage1or 2), diabetesmellitus, andobesity. Depression was examined both as a continuous measure based on the CES-D total sum score as well as dichotomized (>=16), indicative of clinically relevant depressive symptoms. We first checked for an interaction between depression and the presence of cardiac disease using Cox proportional hazards regression models with stroke as the dependent variable. In the fully adjustedmodels, the hazard ratio (HR) for clinically relevant depressive symptoms by cardiac disease status was 4.03 (95% CI 1.22-13.28) (p = 0.022) and HR for
RkJQdWJsaXNoZXIy MjY0ODMw