Table 2.
Comparative models fits for BMI as a continuous moderator of insulin sensitivity and lipid profile
| Variables | Comparative model fit | ||||
|---|---|---|---|---|---|
| Model | Δ-2LL | df | p value | Test | |
| QUICKI | |||||
| 1. Full | ACETUVB | ||||
| 2. U = 0 | ACETVB | 1.52 | 1 | 0.22 | 2 vs 1 |
| 3. T = 0 U = 0 | ACEVB | 0.44 | 1 | 0.51 | 3 vs 2 |
| 4. V = 0 T = U = 0 | ACEB | 25.2 | 1 | <0.001 | 4 vs 3 |
| 5. C = 0 T = U = 0a | AEVB | 0.42 | 1 | 0.52 | 5 vs 3 |
| Triacylglycerol | |||||
| 1. Full | ACETUVB | ||||
| 2. U = 0 | ACETVB | 0.01 | 1 | 0.96 | 2 vs 1 |
| 3. V = 0 U = 0 | ACETB | 0.04 | 1 | 0.82 | 3 vs 2 |
| 4. T = 0 U = V = 0 | ACEB | 14.7 | 1 | <0.001 | 4 vs 3 |
| 5. C = 0 U = V = 0a | AETB | 0.00 | 1 | 1.00 | 5 vs 3 |
| Total cholesterol | |||||
| 1. Full | ACETUVB | ||||
| 2. U = 0 | ACETVB | 2.27 | 1 | 0.13 | 2 vs 1 |
| 3. V = 0 U = 0 | ACETB | 0.01 | 1 | 0.96 | 3 vs 2 |
| 4. T = 0 U = V = 0a | ACEB | 3.59 | 1 | 0.06 | 4 vs 3 |
| 5. C = 0 T = U = V = 0 | AEB | 4.40 | 1 | 0.04 | 5 vs 4 |
| HDL-cholesterol | |||||
| 1. Full | ACETUVB | ||||
| 2. U = 0 | ACETVB | 0.12 | 1 | 0.73 | 2 vs 1 |
| 3. V = 0 U = 0 | ACETB | 0.18 | 1 | 0.67 | 3 vs 2 |
| 4. T = 0 U = V = 0a | ACEB | 3.09 | 1 | 0.08 | 4 vs 3 |
| 5. C = 0 T = U = V = 0 | AEB | 9.60 | 1 | 0.002 | 5 vs 4 |
aThe best fitting model
Δ-2LL, log likelihood; A, additive genetic variance; B, linear effects of BMI on means of the outcome variables; C, shared environmental variance; df, degrees of freedom; E, unique environmental variance; T, moderation of additive genetic variance by BMI; U, moderation of shared environmental variance by BMI; V, moderation of unique environmental variance by BMI