Fatty acid groups and ratios j
^{3}

External validation with biomarkers (M/N)

Internal repeat measurements (Q_{2}/Q_{3})



Attenuation factor ${\widehat{\lambda}}_{j}$

Standard error (SE ${\widehat{\lambda}}_{j}$)

RCAL odds ratio (95% CI)^{2}

Attenuation factor ${\widehat{\lambda}}_{j}$

Standard error (SE ${\widehat{\lambda}}_{j}$)

RCAL odds ratio (95% CI)^{2}


6p

0.230

(0.020)

20.3 (11.8, 35.0)

0.605

(0.031)

3.1 (2.7, 3.7)

3p

0.090

(0.022)

>100

0.456

(0.033)

4.6 (3.5, 5.9)

6H

0.046

(0.028)

>100

0.653

(0.027)

2.9 (2.6, 3.2)

3H

0.408

(0.020)

5.5 (4.5, 6.6)

0.717

(0.025)

2.6 (2.4, 2.9)

6

0.269

(0.020)

13.1 (8.8, 19.7)

0.603

(0.031)

3.2 (2.7, 3.7)

3

0.353

(0.021)

7.1 (5.5, 9.2)

0.632

(0.027)

3.0 (2.6, 3.4)

3p/6p

0.251

(0.021)

15.8 (9.8, 25.6)

0.588

(0.031)

3.2 (2.8, 3.8)

3H/6H

0.426

(0.020)

5.1 (4.3, 6,1)

0.655

(0.028)

2.9 (2.5, 3.2)

3/6

0.401

(0.021)

5.6 (4.6, 6.9)

0.655

(0.024)

2.9 (2.6, 3.2)


^{1} Q_{1}: Food frequency questionnaire measurement at visit 1 of fatty acid group intake as % of energy intake or ratio of n3 to n6 groups. M: biomarker of fatty acid intake in cholesteryl ester fraction of plasma; N: biomarker of fatty acid intake in phospholipids fraction of plasma; Q_{2}: Repeat of Q_{1} measured at visit 2 among a subset of the cohort; Q_{3}: Repeat of Q_{1} measured at visit 3 among the surviving baseline cohort. RCAL: regression calibrated estimate.

^{2} Naïve odds ratio assumed to be equal to 2.00 (1.64, 2.43), with Log_{e}(odds ratio_{naive}) = β
_{naive} = 0.693 with a SE(β
_{naive}) = 0.100, in the following bivariate model: $\text{E(}Y\text{}{Q}_{j}\text{)}={\widehat{\beta}}_{\text{0}}+{\widehat{\beta}}_{\text{(na\xefve,j)}}{Q}_{j}$. The odds ratio is for any binary outcome (0,1) vs. fatty acid exposure expressed as a zscore (increase in odds of outcome for each 1 SD in fatty acid exposure). Refer to Eq. 4.1–4.4.

^{3} See table 2.