Article Text


Population Based Studies: Intergenerational
OP63 Associations of All-Cause and Cause-Specific Mortality with body Mass Index in a Large Norwegian Cohort: Use of Offspring Body Mass Index as an Instrumental Variable
  1. DJ Carslake1,
  2. G Davey Smith1,
  3. D Gunnell2,
  4. TIL Nilsen3,
  5. P Romundstad4
  1. 1MRC Centre for Causal Analyses in Translational Epidemiology, School of Social and Community Medicine, University of Bristol, Bristol, UK
  2. 2School of Social and Community Medicine, University of Bristol, Bristol, UK
  3. 3Department of Human Movement Science, Norwegian University of Science and Technology, Trondheim, Norway
  4. 4Department of Public Health and General Practice, Norwegian University of Science and Technology, Trondheim, Norway


Background People with high body mass index (BMI) suffer elevated rates of mortality from a variety of causes, particularly from cardiovascular disease (CVD). Some studies have also found elevated mortality at low BMI, but it is unclear whether this represents a causal association. Potential confounding by existing ill-health is of particular concern. To avoid this problem, we used the BMI of an offspring as an instrument for the parent’s BMI, as well as conducting conventional analyses of parental mortality against their own BMI.

Methods We extracted 33,011 mother-offspring and 28,142 father-offspring pairs from HUNT, a large prospective cohort study conducted in Nord-Trøndelag county, Norway. Participants’ BMI was adjusted for age, sex, secular trends and smoking status. Associations of parental mortality rates with their own BMI were estimated as hazard ratios using Cox regression. Age was the time axis, and we adjusted for smoking, alcohol, exercise, education and employment. Additionally, adjusted causal hazard ratios between parental mortality and BMI were estimated using offspring BMI as an instrument for parental BMI and cubic splines were fitted to all associations to assess their linearity.

Results There were 9,271 maternal and 9,889 paternal deaths within the follow-up period of 1984–2009. The associations of all-cause, CVD and cancer mortality with parents’ own BMI were substantially nonlinear, with elevated mortality at both extremes and minima at 21–25 kg m–2. Hazard ratios for all-cause mortality per standard deviation (4.1 kg m–2) of own BMI were 1.04 (95% CI: 1.02, 1.06) and 1.05 (1.02, 1.08) in mothers and fathers, respectively. In contrast, associations of mortality with offspring BMI were approximately linear and positive. Causal hazard ratios per standard deviation of BMI estimated from the instrumental variable analyses were 1.20 (1.11, 1.30) and 1.11 (1.02, 1.22) for all-cause mortality in mothers and fathers, respectively. For CVD they were 1.26 (1.13, 1.41) and 1.15 (1.01, 1.30), and for cancer they were 1.23 (1.05, 1.45) and 1.07 (0.90, 1.27).

Conclusion These results confirm the elevated mortality from all causes, CVD and cancer in subjects with high or low BMI. The use of an offspring’s BMI as an instrument suggests that the elevated mortality at low BMI in this and other studies may be the result of confounding by existing ill-health, and that the causal relationship between BMI and mortality is positive and approximately linear.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.