Risk factors and outcomes associated with blood pressure metrics in Nepal

Abstract

Background: Hypertension is a common and established risk factor for cardiovascular disease and a potentially modifiable risk factor for dementia. There is growing evidence that blood pressure (BP) variability, independent of BP, is also linked to poor cardiovascular and cognitive outcomes. This is particularly important for populations with a high prevalence of uncontrolled hypertension such as those in Nepal and other low- and middle-income countries. The objectives of this dissertation were to (1) validate a picture-based cognitive test, (2) evaluate the associations of cardiovascular health (CVH) with hypertension and BP variability, and (3) assess the associations of hypertension and BP variability with cognitive tests among Nepalese adults. Methods: We used data from the second wave (2022-2023) of data collection for the Dhulikhel Heart Study (DHS), a community-based prospective study of Nepalese adults aged 18 years or older living in the town of Dhulikhel in central Nepal. The validity of the Picture Symbol Substitution Test (PSST), developed to increase participation of people with low education in cognitive evaluation, was examined by measuring the Pearson correlation coefficient between the PSST and the Digit Symbol Substitution Test (DSST). The test-retest reliability of the PSST was assessed by calculating the intraclass correlation coefficient (ICC) between repeated measurements of the PSST. Agreement between the PSST and DSST and between repeated PSST measurements was examined by a Bland-Altman plot and method (95% limits of agreement). We conducted cross-sectional studies of the associations of CVH with hypertension and BP variability and the associations of hypertension and BP variability with cognitive tests (Mini-Mental State Exam, PSST, Digit Span Forward, and Digit Span Backward). Generalized estimating equations (GEEs) for multivariate log-linear regression with robust standard errors were used to estimate prevalence ratios (PRs) and 95% confidence intervals (CIs) of the hypertension prevalence. GEEs for multivariate linear regression were used to estimate beta coefficients and 95% CIs for BP variability (coefficient of variation) and cognitive test scores. The unadjusted and adjusted associations of ideal CVH and hypertension were assessed using three models: unadjusted (model 1), age, sex, ethnicity, education; additionally, number of household poverty measures (model 2); and additionally, number of standard drinks per week, smoking, and physical activity (model 3). The adjusted associations of hypertension and BP variability with cognitive function were assessed using three models: age, sex, ethnicity, and education (model 1); additionally, number of household poverty measures (model 2); and additionally, BMI, physical activity, number of standard drinks per week, diabetes, and smoking (model 3). Results: The Pearson correlation coefficient for the association between the PSST and DSST was 0.90 (95% CI: 0.87, 0.92). The ICC of the PSST repeated measurements was 0.95 (95% CI: 0.92, 0.97). Results were similar across age, sex, and education subgroups. A higher number of ideal CVH metrics was associated with a lower prevalence of hypertension. The adjusted PR (95% CI) for hypertension (ref. no hypertension) was 0.87 (0.80, 0.95; p-value=0.0028). We did not find statistical evidence of an association between ideal CVH and BP variability. We found that hypertension was associated with lower PSST scores (adjusted beta coefficient [95% CI], -2.85 [-4.97, -0.72]; p-value=0.0087), cognitive global z-scores of three tests (adjusted beta coefficient [95% CI], -0.069 [-0.14, -0.0015]; p-value=0.045), and cognitive global z-scores of four tests (adjusted beta coefficient [95% CI], -0.10 [-0.20, -0.0043]; p-value=0.041). A 1-unit increase in the coefficient of variation was associated with lower cognitive test scores, although associations were not statistically significant. Conclusion: The PSST performed sufficiently well among those with low and high levels of education/literacy. Ideal CVH was associated with a lower prevalence of hypertension, and hypertension was associated with lower PSST scores and cognitive global z-scores. While we did not find statistically significant associations between CVH and BP variability or between BP variability and cognitive tests, the latter findings suggested an inverse relationship, consistent with previous studies. Findings from this dissertation highlight the need for better and greater screening of hypertension which is not only related to CVH but also cognitive health. Future research is needed to identify how optimal CVH may influence the risk of incident hypertension or BP variability and how hypertension and BP variability may contribute to a change in performance on cognitive tests.