Influences of place of residence on risk factors for atherosclerotic cardiovascular diseases in South India

There has been much debate about the role of place of residence (rural vs. urban) on non-communicable disease outcomes but the potential pathways are relatively poorly quantified. This PhD explores the links between place of residence and cardiovascular diseases (CVDs) in rapidly urbanising India in three connected research papers. The first paper systematically reviewed the available published evidence on the associations of engaging in agriculture with CVD incidence and prevalence and CVD risk factors in low- and middle-income countries (LMICs). The review included 15 observational studies, and high-quality evidence was lacking. Thirteen studies from five LMICs suggested that agricultural workers living in rural areas had a lower prevalence of some important CVD risk factors (hypertension and high body mass index [BMI]) but higher prevalence of others (smoking and underweight) than non-agricultural workers mainly living in urban areas. Building on these initial findings, the second paper estimated the association of urbanisation level with a range of CVD risk factors using data from the third wave of the Andhra Pradesh Children and Parents Study (APCAPS) (n=6236). Remote sensing night-time light intensity (NTLI) data (unitless digital numbers) provided a continuous proxy measure of levels of urbanisation for 27 APCAPS villages. Mixed-effects linear regression models with log-transformed NTLI were used in analysis. Increasing NTLI was associated with rises in mean BMI and systolic blood pressure (SBP), but not low-density lipoprotein (LDL) or fasting plasma glucose (FPG), after adjusting for confounders and these increases were greater among participants aged above 40 years. Using this dataset in the third research paper, the total effect of increasing urbanisation level on mean SBP was decomposed into direct and indirect effects via hypothesised pathways. Mediation analysis was performed using mixed-effects linear regression models for SBP, log-transformed NTLI (that ranged from 4.1 to 7.0) and three composite mediators summarising (i) socio-demographic (e.g. occupation and education), (ii) lifestyle, mental health (e.g. diet and depression), and (iii) metabolic factors (e.g. BMI and LDL). All models were gender-stratified and adjusted for age and other confounders. Mean SBP was 122.7mmHg (±15.7) among men and 115.8mmHg (±14.2) among women. A one unit (integer) increase in log-NTLI was associated with a rise in SBP by 2.0mmHg (95% CI 0.4, 3.5) among men and 1.3mmHg (95% CI 0.006, 2.6) among women. A considerable indirect effect via the metabolic pathway elevated SBP among men by 4.6mmHg (95% CI 2.0, 7.3) and a smaller SBP rise among women by 0.7mmHg (95% CI 0.1, 1.3) per one log-NTLI increase. Among men, but not women, NTLI 4 acted indirectly via the lifestyle and mental health pathway to elevate SBP by 0.7mmHg (95% CI 0.1, 1.3) per one log-NTLI increase. The total effect among both genders and the indirect effect via metabolic factors among men approximately doubled among participants aged above 40 years. This PhD identified night-time light intensity as a potentially important continuous proxy indicator of urbanisation levels in India and formally tested potential causal pathways linking urbanisation level with CVD risk factors. Increasing level of urbanisation was associated with greater mean SBP and BMI at early stages of urbanisation in South India. The findings offer new insights into possible pathways through which urbanisation may act on CVD risk factors. Pathways via metabolic factors independent of socio-demographic, lifestyle and mental health factors emerged as particularly interesting. These findings identified a need to understand better the indirect effects of urbanisation-related upstream determinants on CVD risk factors in India, independent of socio-demographic, lifestyle and mental health factors. Mediation analysis may be a useful approach to inform strategies to mitigate the expected large public health and economic impacts from continued, rapid urbanisation in India. I recommend implementing and scaling available, evidence-based, population-wide, primordial prevention strategies in India that target upstream determinants to modify behavioural risk factors for NCDs. In line with WHO’s best buys for NCD prevention and recommendations of the WHO and the Climate and Clean Air Coalition, strategies could include (but are not limited to), advertisement restriction, health promotion campaigns; taxation of alcohol, petrol and food content; improved enforcement of the Cigarettes and Other Tobacco Products Act and shifts to low-emission transport options. Evaluation of the ongoing Smart City Mission could identify innovative and effective interventions for cleaner, greener, and more environmentally sustainable urban development throughout India. In order to tackle expected rises in CVDs with urbanisation in India (and elsewhere in LMICs), I further call for a new interdisciplinary urbanisation science and collaborations to scale-up national surveillance systems and periodical surveys on environment, demographics, phyco-social factors and NCDs.