Table 2.

Analyses of BMI-SDS changes in German obese children and adolescents who participated in the obesity intervention program 'Obeldicks'

Parameter	N1	Genotype	N (%)	Mean ± SD	Additive genetic model2
					Estimate	95% CI	p-value3
Δ BMI-SDS4,5	207	TT	53 (26)	0.33 ± 0.40	-0.032	-0.091...0.027	0.287
		AT	86 (41)	0.25 ± 0.27
		AA	68 (33)	0.27 ± 0.27

¹ total number of individuals who participated in the obesity intervention program; ² linear regression analyses for BMI-SDS http://www.mybmi.de including covariates age and sex- including baseline BMI-SDS did not substantially alter the result; the p-value for the sex main effect was 0.740; ³ two-sided p-value; ⁴ median time interval between repeated measurements 12 months; positive values for descriptive statistics indicate weight reduction in units of BMI-SDS; asymptotic two-sided p-value for Kruskal-Wallis Test 0.44; ⁵ A trend towards a deviation from Hardy-Weinberg equilibrium was observable for genotype frequencies for Δ BMI-SDS (exact p = 0.02). This finding, however, is not surprising and expected in case of a true genetic association and indeed the number of homozygotes for the at-risk A-allele was increased in these patients compared to controls.