. Author manuscript; available in PMC: 2008 Mar 1.

Published in final edited form as: J Clin Endocrinol Metab. 2006 Dec 19;92(3):948–954. doi: 10.1210/jc.2006-1390

TABLE 2.

Mixed regression models for predicting increase in natural log of BMI and increase in natural log of body fat mass

Effect	Estimate	SE	P value
Model predicting increase in natural log of BMI
Intercept	−0.1672	0.1419	0.2401
Follow-up age to the third order	−6.67e⁻⁵	1.43e⁻⁵	<0.0001
Sex (0 girls, 1 boy)	−0.05628	0.04751	0.2376
Baseline age	−0.05776	0.00443	<0.0001
Baseline BMI (log)	0.9806	0.04878	<0.0001
Baseline leptin (log)	0.06673	0.02711	0.0147
Follow-up age × sex (girls)	0.0921	0.007663	<0.0001
Follow-up age × sex (boys)	0.08541	0.007277	<0.0001
Baseline leptin (log) × follow-up age	−0.00599	0.001948	0.0022
Model predicting increase in natural log of body fat mass
Intercept	−0.913	1.1232	0.4176
Follow-up age to the third order	−0.00174	4.83e⁻⁴	0.0004
Follow-up age to the fourth order	5.3e⁻⁵	1.8e⁻⁵	0.0039
Sex (0 girls, 1 boy)	−0.3271	0.1492	0.03
Race (0 Caucasian, 1 other ethnicity)	0.4818	0.1522	0.0019
Baseline age	−0.1561	0.01557	<0.0001
Baseline total fat (log)	0.8561	0.1253	<0.0001
Baseline leptin (log)	0.1423	0.04099	0.0007
Baseline fat mass (log) × follow-up age	−0.0257	0.009007	0.0046
Follow-up age × sex (girls) × race (Caucasian)	0.7164	0.1092	<0.0001
Follow-up age × sex (girls) × race (other ethnicity)	0.7598	0.1087	<0.0001
Follow-up age × sex (boys) × race (Caucasian)	0.6928	0.107	<0.0001
Follow-up age × sex (boys) × race (other ethnicity)	0.7264	0.1061	<0.0001