Large-scale observational studies in diverse populations

Experience in Western countries has shown that large population-based prospective studies are a valuable way of investigating many slow-acting causes of common chronic diseases. The PHRU is collaborating with researchers in low and middle income countries to generate reliable evidence about the importance of common risk factors for premature death (such as smoking, alcohol, obesity and hypertension).
Our previous research has demonstrated the hazards of prolonged smoking and the benefits of stopping smoking in high-income countries; and ongoing studies, involving 1.4 million participants, are currently assessing the risks of smoking in Mexico, Cuba and India. In 2009 and 2014 we reported that excessive alcohol consumption (chiefly vodka) was responsible for most premature deaths in Russia and this immediately prompted new public health strategies to limit heavy alcohol consumption. Our study in Mexico (where one-third of men and one-half of women are obese) has shown that uncontrolled diabetes causes at least one third of all deaths between 35 and 74 years of age.

Over a period of decades, large population-based observational studies have helped identify several major 'classic causes' of premature death from non-communicable disease (NCD), including smoking, hypertension, obesity (adiposity), harmful alcohol use, dyslipidaemia, diabetes, poor diet and physical inactivity. However, the effects of such factors vary greatly from one population to another. There is still substantial uncertainty as to how important these factors are in different settings, and how their importance is changing with time. Direct evidence from low-and middle-income countries (LMICs) is important as it may in the long run improve local and regional NCD control strategies and actions. However, such evidence is lacking. Indeed, at present, global estimates of the importance of classic causes depend mainly on mathematical extrapolation from findings in high-income countries. LMIC findings may change global estimates significantly, thus strengthening our understanding of the NCD burden and the local relevance of causes.
During the 1990s/2000s, therefore, PHRU in collaboration with local investigators set up large prospective studies in Chennai (500,000 adults), Kerala (400,000), Russia (200,000 adults), Cuba (150,000) and Mexico City (150,000 with blood). Recruitment finished in Cuba in 2000, Chennai in 2001, Kerala in 2002, Mexico in 2004 and Russia in 2008. Follow-up involves fieldwork together, where possible, with electronic linkage to national mortality records. Resurveys among subsets of survivors have been conducted or are planned in each cohort. Some 53,000 deaths have already been recorded in Chennai, 33,000 in Kerala, 12,000 in Cuba and 20,000 in Mexico. In Mexico, further linkage to non-fatal health records is being conducted and validated, while the availability of blood samples collected at baseline in nearly all participants will allow (funding permitting) for the measurement of a large range of biochemical and genetic markers. These cohorts have already demonstrated the value of studying classic causes of premature death in different settings, for example by demonstrating the risks from hazardous drinking in Russia (which, together with smoking, may well account for 3/4 of all male deaths before age 75) and the risks from diabetes in Mexico (where diabetes was found to be a cause of at least one third of all deaths between 35 and 74 years of age, twice previous indirect estimates based on evidence from other populations).
The prospective studies PHRU has set up will help monitor the evolution of the worldwide epidemics of premature death from classic causes, while the availability of blood samples in the Mexico study affords the further opportunity for discovery of novel genetic and metabolic pathways and their relevance for premature death. The studies have been set up, follow-up has accumulated, collaboration is assured and PHRU will, during the next 5 years, continue to collaborate closely with the national PIs to generate reliable analyses and publications. These reports will reach the highest levels of the governments concerned as well as reaching the wider population.