Table 5 Comparative model fitting for BMI as a continuous moderator of SBP and DBP

From: Genetic and environmental influences on blood pressure and body mass index in Han Chinese: a twin study

 

Model fitting

 

Comparative model fitting

 

Model

−2LL

d.f.

Δ-2LL

Δd.f.

P-value

AIC

SBP

 1.Full

ACETUVB

6067.531

2054

    

 2.T=U=V=0

ACEB

6092.519

2057

24.988

3

<0.001

18.988

 3.V=0

ACETUB

6074.529

2055

6.999

1

0.008

4.999

 4.U=0

ACETVB

6068.573

2055

1.042

1

0.307

−0.958

 5.T=0

ACEUVB

6067.540

2055

0.010

1

0.922

−1.990

 6.U=V=0

ACETB

6074.561

2056

7.031

2

0.030

3.031

 7.T=V=0

ACEUB

6077.025

2056

9.494

2

0.009

5.494

 8.T=U=0

ACEVB

6074.966

2056

7.436

2

0.024

3.436

DBP

 1.Full

ACETUVB

6355.688

2054

    

 2.T=U=V=0

ACEB

6363.211

2057

7.524

3

0.057

1.524

  1. Abbreviations: −2LL, −2-log likelihood; A, additive genetic variance; AIC, Akaike's Information Criterion; B, linear effects of BMI on means of the outcome variables; BMI, body mass index; C, common environmental variance; DBP, diastolic blood pressure; d.f., degrees of freedom; E, unique environmental variance; SBP, systolic blood pressure; T, moderation of additive genetic variance by BMI; U, moderation of common environmental variance by BMI; V, moderation of unique environmental variance by BMI.
  2. All models were compared with the full model (model 1).
  3. SBP and DBP were log-transformed and adjusted for age, sex and study site.
  4. Best fitting models are in bold.
Back to article page