Instrumental variable analysis using offspring BMI in childhood as an indicator of parental BMI in relation to mortality (original) (raw)
Related papers
International journal of epidemiology, 2017
The observational association between mortality and body mass index (BMI) is U-shaped, leading to highly publicized suggestions that moderate overweight is beneficial to health. However, it is unclear whether elevated mortality is caused by low BMI or if the association is confounded, for example by concurrent ill health. Using HUNT, a Norwegian prospective study, 32 452 mother-offspring and 27 747 father-offspring pairs were followed up to 2009. Conventional hazard ratios for parental mortality per standard deviation of BMI were estimated using Cox regression adjusted for behavioural and socioeconomic factors. To estimate hazard ratios with reduced susceptibility to confounding, particularly from concurrent ill health, the BMI of parents' offspring was used as an instrumental variable for parents' own BMI. The shape of mortality-BMI associations was assessed using cubic splines. There were 18 365 parental deaths during follow-up. Conventional associations of mortality from ...
PLOS Medicine
Background High body mass index (BMI) is associated with mortality, but the pervasive problem of confounding and reverse causality in observational studies limits inference about the direction and magnitude of causal effects. We aimed to obtain estimates of the causal association of BMI with all-cause and cause-specific mortality. Methods and findings In a record-linked, intergenerational prospective study from the general population of Sweden, we used two-sample instrumental variable (IV) analysis with data from 996,898 fathers (282,407 deaths) and 1,013,083 mothers (153,043 deaths) and their sons followed up from January 1, 1961, until December 31, 2004. Sons' BMI was used as the instrument for parents' BMI to compute hazard ratios (HRs) for risk of mortality per standard deviation (SD) higher parents' BMI. Using offspring exposure as an instrument for parents' exposure is unlikely to be affected by reverse causality (an important source of bias in this context) and reduces confounding. IV analyses supported causal associations between higher BMI and greater risk of all-cause mortality (HR [95% confidence interval (CI)] per SD higher fathers'
Journal of Epidemiology & Community Health, 2012
Background Adult body mass index (BMI) has been consistently related to mortality, but little is known about the impact of earlier life BMI on adult mortality. The aim is to investigate the impact of childhood, adolescent and early adult BMI on premature adult all-cause mortality. Methods The British 1946 cohort study was used to assess the association of BMI in childhood, adolescence and adulthood with mortality 26e60 years (332 deaths). 4462 (83%) respondents were available for analysis at age 26 years. Splines were used in Cox regression to model the associations between BMI and mortality. Results In both genders, adult BMI from 20 years onwards showed a consistent U-shaped relationship with adult mortality (overall p value <0.05 for BMI at ages 20, 26 and 36 years). In women, a similar relationship was observed for adolescent BMI at 15 years (p¼0.02); the HR comparing women with low BMI (2 SDs below mean) versus mean BMI was 2.96 (95% CI 1.26 to 6.97). The corresponding HR for women with BMI 2 SDs above the mean was 1.97 (0.95 to 4.10). BMI in childhood was generally not associated with adult mortality except female BMI at 4 years where a U-shaped relationship was observed (p¼0.02); HR for BMI 2 SDs below mean versus mean was 2.13 (0.97 to 4.70) and the corresponding HR for 2 SDs above the mean was 1.67 (0.85 to 3.28). This association was not attenuated by subsequent BMI change or mediators. Conclusions High and low BMI from early adulthood were related to adult premature mortality suggesting that promoting a normal weight in early adulthood could prevent premature mortality.
American Journal of Epidemiology, 2003
A large number of studies have explored the relation between body mass index (BMI) and mortality in adults. The relation between BMI in adolescence and mortality has been investigated to a lesser extent. It has been suggested that all-cause mortality is elevated among those who were overweight during adolescence, but the limitation of previous studies has been study size. The present study explored this relation in a Norwegian cohort of 227,003 boys and girls, aged 14-19 years, whose height and weight were measured during tuberculosis screening in 1963-1975. These persons were followed for an average of 31.5 years (about 7.2 million personyears). A total of 7,516 deaths were registered. Multivariate Cox proportional hazards regression models were used in the analyses. An increasing risk of death by increasing BMI in adolescence was observed. Mortality among males whose baseline BMI was between the 85th and 95th percentiles and above the 95th percentile in the US reference population was 30% and 80% higher, respectively, than that among those whose baseline BMI was between the 25th and 75th percentiles. The corresponding rates among females were 30% and 100%. The excess mortality among adolescents whose BMI was high was not clearly manifested before they reached their thirties. Hence, BMI in adolescence is predictive of adult mortality.
PLOS Medicine
Background Pediatric obesity is associated with increased risk of premature death from middle age onward, but whether the risk is already increased in young adulthood is unclear. The aim was to investigate whether individuals who had obesity in childhood have an increased mortality risk in young adulthood, compared with a population-based comparison group. Methods and findings In this prospective cohort study, we linked nationwide registers and collected data on 41,359 individuals. Individuals enrolled at age 3-17.9 years in the Swedish Childhood Obesity Treatment Register (BORIS) and living in Sweden on their 18th birthday (start of followup) were included. A comparison group was matched by year of birth, sex, and area of residence. We analyzed all-cause mortality and cause-specific mortality using Cox proportional hazards models, adjusted according to group, sex, Nordic origin, and parental socioeconomic status (SES). Over 190,752 person-years of follow-up (median follow-up time 3.6 years), 104 deaths were recorded. Median (IQR) age at death was 22.0 (20.0-24.5) years. In the childhood obesity cohort, 0.55% (n = 39) died during the follow-up period, compared to 0.19% (n = 65) in the comparison group (p < 0.001). More than a quarter of the deaths among individuals in the childhood obesity cohort had obesity recorded as a primary or contributing cause of death. Male sex and low parental SES were associated with premature all-cause mortality. Suicide and self-harm with undetermined intent were the main cause of death in both groups. The largest difference between the groups lay within endogenous causes of death, where children who had undergone obesity treatment had an adjusted mortality rate ratio of 4.04 (95% CI 2.00-8.17, p < 0.001) compared with the comparison group. The main study limitation was the lack of anthropometric data in the comparison group.
American Journal of Epidemiology, 2008
The prevalence of obesity in childhood and adolescence has increased worldwide. Long-term effects of adolescent obesity on cause-specific mortality are not well specified. The authors studied 227,000 adolescents (aged 14-19 years) measured (height and weight) in Norwegian health surveys in [1963][1964][1965][1966][1967][1968][1969][1970][1971][1972][1973][1974][1975]. During follow-up (8 million person-years), 9,650 deaths were observed. Cox proportional hazards regression was used to compare causespecific mortality among individuals whose baseline body mass index (BMI) was below the 25th percentile, between the 75th and 84th percentiles, and above the 85th percentile in a US reference population with that of individuals whose BMI was between the 25th and 75th percentiles. Risk of death from endocrine, nutritional, and metabolic diseases and from circulatory system diseases was increased in the two highest BMI categories for both sexes. Relative risks of ischemic heart disease death were 2.9 (95% confidence interval (CI): 2.3, 3.6) for males and 3.7 (95% CI: 2.3, 5.7) for females in the highest BMI category compared with the reference. There was also an increased risk of death from colon cancer (males: 2.1, 95% CI: 1.1, 4.1; females: 2.0, 95% CI: 1.2, 3.5), respiratory system diseases (males: 2.7, 95% CI: 1.4, 5.2; females: 2.5, 95% CI: 1.4, 4.8), and sudden death (males: 2.2, 95% CI: 1.2, 4.3; females: 2.7, 95% CI: 1.1, 6.6). Adolescent obesity was related to increased mortality in middle age from several important causes. adolescent; body mass index; cause of death; follow-up studies; mortality; Norway; obesity Abbreviations: BMI, body mass index; CI, confidence interval; RR, relative risk.
The association between BMI value and long-term mortality
Int. J. Obes., 2009
Objectives: To study total mortality in different categories of BMI values, with adjustments for important covariates in a population-based 26-year mortality follow-up. Special interest will be given to gender differences and low BMI values. Methods: From a stratified sample in 1969 of 32 185 individuals aged 18-64 years from Stockholm County, 2422 underwent a health examination, with complete data obtained for 1020 subjects. BMI was classified as underweight (o20), normal (20-24.9), overweight (25-29.9) or obesity (X30). Participants were followed up in the National Cause of Death Register until the end of 1996. Multivariate analysis was performed by Cox regression for men and women separately, with different models, with step-wise adjustment for age, care need category, heart rate, hypertension, blood glucose, alcohol intake and smoking, with hazard ratios (HR) and 95% confidence interval (CI) and with normal weight as reference. Results: Among men, the age-adjusted HR was 1.68 (95% CI 1.10-2.57) for underweight and 1.62 (95% CI 1.08-2.43) for obesity, and among women it was 0.93 (95% CI 0.58-1.51) for underweight and 1.88 (95% CI 1.26-2.82) for obesity. In men, the significantly increased mortality remained when also adjusting for care need category, but not when adjusting for other factors, whereas the opposite was found regarding obesity. For women, underweight was significantly associated with decreased mortality when adjusting for smoking and for all factors together, whereas obesity was associated with increased mortality when adjusting for the different factors except for all factors together. Conclusions: Underweight was associated with higher mortality among men, but not when adjusting for covariates, whereas underweight was associated with lower mortality among women when adjusting for smoking.