Table 3.

Risk of incident diabetes according to plasma fatty acid biomarkers of dairy fat consumption, evaluated continuously, among 3,333 men and women in the NHS (N=184 cases), HPFS (N=93 cases), and both cohorts combined.

	Multivariable HR (95%CI) standardized to difference between midpoints of highest vs. lowest quartiles (87.5th minus 12.5th percentiles)
	NHS	HPFS	Pooled	P–value
15:0	range=0.10*	range=0.10
Multivariable HR (95%CI)†	0.63 (0.42–0.92)	0.61 (0.36–1.03)	0.62 (0.46–0.85)	<0.01
+ BMI‡	0.60 (0.40–0.89)	0.73 (0.43–1.25)	0.64 (0.47–0.89)	0.01
17:0	range=0.13	range=0.14
Multivariable HR (95%CI)	0.71 (0.49–1.04)	0.63 (0.39–1.02)	0.68 (0.50–0.91)	0.01
+ BMI	0.76 (0.52–1.10)	0.69 (0.42–1.11)	0.73 (0.55–0.99)	0.04
t-16:1n-7	range=0.15	range=0.12
Multivariable HR (95%CI)	0.52 (0.36–0.76)	0.58 (0.36–0.93)	0.54 (0.40–0.73)	<0.001
+ BMI	0.51 (0.36–0.74)	0.61 (0.37–0.99)	0.54 (0.41–0.73)	<0.001
14:0	range=0.74	range=0.65
Multivariable HR (95%CI)	0.92 (0.64–1.32)	0.61 (0.35–1.09)	0.82 (0.60–1.11)	0.19
+ BMI	0.90 (0.63–1.31)	0.75 (0.42–1.32)	0.85 (0.63–1.16)	0.32

^*The difference in % of total fatty acids between midpoint of highest vs. lowest quartile.

^†Multivariable adjustments as in Table 2.

^‡Further adjusted for BMI (kg/m²) as a potential mediator or confounder of the association.