Body-mass index, blood pressure, and cause-specific mortality in India: a prospective cohort study of 500 810 adults.

BACKGROUND: The association between cause-specific mortality and body-mass index (BMI) has been studied mainly in high-income countries. We investigated the relations between BMI, systolic blood pressure, and mortality in India. METHODS: Men and women aged 35 years or older were recruited into a prospective study from the general population in Chennai, India between Jan 1, 1998, and Dec 31, 2001. Participants were interviewed (data collected included age, sex, education, socioeconomic status, medical history, tobacco smoking, and alcohol intake) and measured (height, weight, and blood pressure). Deaths were identified by linkage to Chennai city mortality records and through active surveillance by household visits from trained graduate non-medical fieldworkers. After the baseline survey, households were visited once in 2002-05, then biennially until 2015. During these repeat visits, structured narratives of any deaths that took place before March 31, 2015, were recorded for physician coding. During 2013-14, a random sample of participants was also resurveyed as per baseline to assess long-term variability in systolic blood pressure and BMI. Cox regression (standardised for tobacco, alcohol, and social factors) was used to relate mortality rate ratios (RRs) at ages 35-69 years to systolic blood pressure, BMI, or BMI adjusted for usual systolic blood pressure. FINDINGS: 500 810 participants were recruited. After exclusion of those with chronic disease or incomplete data, 414 746 participants aged 35-69 years (mean 46 [SD 9]; 45% women) remained. At recruitment, mean systolic blood pressure was 127 mm Hg (SD 15), and mean BMI was 23·2 kg/m2 (SD 3·8). Correlations of resurvey and baseline measurements were 0·50 for systolic blood pressure and 0·88 for BMI. Low BMI was strongly associated with poverty, tobacco, and alcohol. Of the 29 519 deaths at ages 35-69 years, the cause was vascular for 14 935 deaths (12 504 cardiac, 1881 stroke, and 550 other). Vascular mortality was strongly associated with systolic blood pressure: RRs per 20 mm Hg increase in usual systolic blood pressure were 2·45 (95% CI 2·16-2·78) for stroke mortality, 1·74 (1·64-1·84) for cardiac mortality, and 1·84 (1·75-1·94) for all vascular mortality. Although BMI strongly affected systolic blood pressure (an increase of about 1 mm Hg per kg/m2) and diabetes prevalence, BMI was little related to cardiac or stroke mortality, with only small excesses even for grade 1 obesity (ie, BMIs of 30·0-35·0 kg/m2). After additional adjustment for usual systolic blood pressure, BMI was inversely related to cardiac and stroke mortality throughout the range 15·0-30·0 kg/m2: when underweight participants (ie, BMI 15·0-18·5 kg/m2) were compared with overweight participants (ie, BMI 25·0-30·0 kg/m2), the blood-pressure-adjusted RR was 1·28 (95% CI 1·20-1·38) for cardiac mortality and 1·46 (1·22-1·73) for stroke mortality. INTERPRETATION: In this South Asian population, BMI was little associated with vascular mortality, even though increased BMI is associated with increased systolic blood pressure, which in turn is associated with increased vascular mortality. Hence, some close correlates of below-average BMI must have important adverse effects, which could be of relevance in all populations. FUNDING: UK Medical Research Council, British Heart Foundation, Cancer Research UK.

Tobacco chewing and adult mortality: a case-control analysis of 22,000 cases and 429,000 controls, never smoking tobacco and never drinking alcohol, in South India.

BACKGROUND: Tobacco is consumed in both smoking and smokeless forms in India. About 35-40% of tobacco consumption in India is in the latter. The study objective was to describe the association between chewing tobacco and adult mortality. MATERIALS AND METHODS: A case-control study was conducted in urban (Chennai city) and rural (Villupuram district) areas in Tamil Nadu state in South India. Interviewed in 1998-2000 about 80,000 families (48,000 urban and 32,000 rural) with members who had died during 1995-1998. These were the cases and their probable underlying cause of death was arrived at by verbal autopsy. Controls were 600,000 (500,000 urban, 100,000 rural) individuals from a survey conducted during 1998-2001 in the same two study areas from where cases were included. RESULTS: Mortality analyses were restricted to non-smoking non-drinkers aged 35-69. The age, sex, education and study area adjusted mortality odds ratio was 30% higher (RR:1.3, 95%CI:1.2-1.4) in ever tobacco chewers compared to never chewers and was significant for deaths from respiratory diseases combined (RR:1.5, 95%CI:1.4-1.7), respiratory tuberculosis (RR:1.7, 95%CI:1.5-1.9), cancers all sites combined (RR:1.5, 95%CI:1.4-1.7) and stroke (RR:1.4, 95%CI:1.2-1.6). Of the cancers, the adjusted mortality odds ratio was significant for upper aero-digestive, stomach and cervical cancers. Chewing tobacco caused 7.1% of deaths from all medical causes. CONCLUSIONS: The present study is the first large study in India analysing non-smoking non-drinkers. Statistically significant excess risks were found among ever tobacco chewers for respiratory diseases combined, respiratory tuberculosis, stroke and cancer (all sites combined) compared to never tobacco chewers.

Physical activity levels among urban and rural women in south India and the risk of breast cancer: a case-control study.

Breast cancer (BC) incidence in India is approximately 100% higher among urban women than rural women. The role of household activities (HA) among urban and rural women in the development of BC was investigated. The study was conducted between 2002 and 2005, at the Regional Cancer Center, Trivandrum and three cancer hospitals in Chennai. Cases (735 urban and 1131 rural) were women with histologically confirmed incident BC. Controls (631 urban and 1242 rural) were age-matched women, who accompanied cancer patients to the hospital. Using in-person interview, information on time spent on HA and potential confounding variables was collected. Odds ratio (OR) and 95% confidence intervals (CI) of BC were estimated according to the time spent on HA through logistic regression models. Time spent on HA was longer among rural women (urban vs. rural: 48 vs. 59% in Trivandrum and 31 vs. 41% in Chennai for 5 or more hours/day on HA). The risk of BC declined with increasing time spent on HA. Compared with less than 3 h/day, the ORs for 5-6 h/day were 0.48 (95% CI: 0.32-0.72) in premenopausal and 0.49 (95% CI: 0.34-0.72) in postmenopausal women. Corresponding ORs for 6 or more hours/day were 0.70 (95% CI: 0.48-1.02) and 0.51 (95% CI: 0.35-0.73). The study supports the hypothesis that a high level of physical activity (PA) contributes to the difference in BC risk between urban and rural women in India. The proportion of BC avoided because of moderate or high PA was estimated to be 19% in urban women and 38% in rural women.

Breastfeeding and breast cancer risk in India: a multicenter case-control study.

Breast cancer incidence is low in India compared with high-income countries, but it has increased in recent decades, particularly among urban women. The reasons for this pattern are not known although they are likely related to reproductive and lifestyle factors. Here, we report the results of a large case-control study on the association between breastfeeding and breast cancer risk. The study was conducted in 2 areas in South India during 2002-2005 and included 1,866 cases and 1,873 controls. Detailed information regarding menstruation, reproduction, breastfeeding and physical activity was collected through in-person interview. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by unconditional logistic regression models. Breastfeeding for long duration was common in the study population. Lifetime duration of breastfeeding was inversely associated with breast cancer risk among premenopausal women (p-value of linear trend, 0.02). No such protective effect was observed in postmenopausal women, among whom a protective effect of parity was suggested. A reduction of breast cancer risk with prolonged breastfeeding was shown among premenopausal women. Health campaign focusing on breastfeeding behavior by appropriately educating women would contribute to reduce breast cancer burden.