Table 3. Causal Estimates Derived From Instrumental Variable Analysis for an SD Increase in BMI and Risk of Cardiometabolic Traits Using Individual Participant Data.

Variable	No. (%)	Partially Adjusted Model, OR (95% CI)a	P Value	Fully Adjusted Model, OR (95% CI)b	P Value
Stroke	1958 (1.6)	1.15 (0.81 to 1.63)	.43	1.02 (0.71 to 1.46)	.93
CHD	5760 (4.8)	1.49 (1.20 to 1.83)c	2.4 × 10−04	1.35 (1.09 to 1.69)c	.007
Hypertension	32 874 (27.5)	1.65 (1.49 to 1.83)c	1.6 × 10−21	1.64 (1.48 to 1.83)c	1.1 × 10−19
Type 2 diabetes	6290 (5.3)	2.75 (2.24 to 3.37)c	3.9 × 10−22	2.53 (2.04 to 3.13)c	1.5 × 10−17
Blood pressured
Systolic	111 637 (93.1)	β = 1.53 (0.69 to 2.36)c	3.3 × 10−04	β = 1.65 (0.78 to 2.52)c	2.0 × 10−04
Diastolic	111 638 (93.1)	β = 1.23 (0.76 to 1.69)c	2.5 × 10−07	β = 1.37 (0.88 to 1.85)c	3.6 × 10−08
Pulse rated	111 638 (93.1)	β = −0.21 (−0.75 to 0.32)	.44	β = −0.46 (−1.02 to 0.10)	.12

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CHD, coronary heart disease; OR, odds ratios.

^aAdjusted for age and sex plus 10 genetic principal components, assessment center, and batch effects.

^bAdditionally adjusted for smoking history, alcohol intake, and Townsend scores.

^cSignificant at P < .01.

^dAdditionally corrected for use of antihypertensive medication; reported as unstandardized β coefficient.