Table 3 Associations between zBMI at age 13 and adult mortality: hazard ratios (95% confidence intervals) per zBMI estimated from conventional analyses of own zBMI and from analyses using offspring zBMI at age 13 as instrumental variable (IV).

From: Instrumental variable analysis using offspring BMI in childhood as an indicator of parental BMI in relation to mortality

Cause of death

Model1

P-values

Conventional (C)

Offspring zBMI at age 13 as IV (IV)

PIV vs C2

Pnon-linearity for own zBMI3

Pnon-linearity in IV 4

Women and men

All-cause

1.07 (1.04–1.09)

1.22 (1.14–1.31)

 < 0.001

0.003

0.670

Cardiovascular disease

1.16 (1.11–1.20)

1.36 (1.22–1.53)

0.002

0.045

0.681

Cancer

1.03 (0.99–1.07)

1.08 (0.97–1.20)

0.346

0.989

0.796

Women

All-cause

1.11 (1.08–1.15)

1.30 (1.19–1.42)

 < 0.001

0.085

0.785

Cardiovascular disease

1.18 (1.11–1.24)

1.49 (1.28–1.74)

0.001

0.114

0.613

Cancer

1.06 (1.01–1.11)

1.13 (0.99–1.29)

0.304

0.724

0.723

Men

All-cause

1.03 (0.99–1.06)

1.15 (1.04–1.29)

0.027

0.047

0.516

Cardiovascular disease

1.15 (1.09–1.20)

1.30 (1.10–1.54)

0.122

0.133

0.664

Cancer

0.99 (0.94–1.05)

1.02 (0.85–1.23)

0.747

0.874

0.397

  1. 1Adjusted for offspring and parental birth cohort, parental and offspring zheight and parental sex.
  2. 2P-values from Durbin-Wu-Hausman test.
  3. 3P-values from a Wald test of non-linearity in the conventional association.
  4. 4Linear trend test from a meta-regression of stratum specific IV estimates on the stratum-specific mean zBMI.
  5. BMI body mass index, IV instrumental variable.
Back to article page