Skip to main content

Table 6 Estimated log relative risk per unit change of true long-term homocysteine concentrations, treated as a latent variable in a single compartment linear-kinetics model; data simulated assuming homocysteine is the causal variable.

From: Use of pathway information in molecular epidemiology

Sampling

scheme

Subsample

size

 

Biomarker(s) measured

 
  

Homocysteine

TS enzyme

Both

Random

80

1.92 (0.21)

-

2.68 (0.46)

  

-

2.71 (0.43)

-0.04 (0.20)

 

200

1.82 (0.15)

-

2.28 (0.21)

  

-

3.26 (0.54)

-0.04 (0.14)

Stratified by G, E, and Y

8 × 10 = 80

1.77 (0.25)

-

2.62 (0.94)

  

-

2.47 (0.64)

-0.05 (0.28)

 

8 × 25 = 200

1.82 (0.16)

-

2.03 (0.20)

  

-

2.93 (0.39)

 
    

-0.14 (0.14)

  1. The simulated coefficients are 2.0 for homocysteine and 0 for TS