Fig. 3.

Insulin tolerance tests (ITT) were conducted after fasting a) at week 0, b) after 12 weeks of LFD and HFD, and c) after 24 weeks of LFD and HFD. The corresponding AUC (right) was calculated (n = 43) (a), n = 10–12 for LFD and n = 31–32 for HFD (b), n = 10–17 per genotype (WT and SmPparγ^−/−) and diet (c). Values were normalized to baseline, fasting glucose levels, which were arbitrarily set as 100%. For better visualization of glucose curves, significances (p < 0.05) between LFD and HFD of the same genotypes are not depicted. A linear two factor repeated measurement model with interaction was performed. The intervention (ITT at 12 weeks) as within-subject factor was significant (p < 0.001). The interaction of ITT at 12 weeks with diet was significant with p = 0.043, whereas the interaction with genotype achieved a p = 0.072/p = 0.94. The three-way interaction of the ITT challenge with diet and genotype was borderline with p = 0.057/p = 0.061. For further testing the with-in subject effects at 12 weeks testing for sphericity was performed by a Mauchly test (p = 0.039). The Huynh Feldt correction was used for the within-subject effects. The intervention (ITT at 24 weeks) as within-subject factor was significant (p < 0.001). The interaction of ITT at 24 weeks with diet was significant with p = 0.025, whereas the interaction with genotype achieved a p = 0.067. The three-way interaction of ITT with diet and genotype was borderline in this test (p = 0.079). Homogeneity of variance between within-subject contrasts was asserted by a Levene test in ITT at 24 weeks. Significance of contrasts was assessed by an F-test. In ITT at 24 weeks the change of contrast was significant at 15, 30, 60 and 90 min (p = 0.003, p < 0.001, p < 0.001, p = 0.045). Diet has a significant impact on measurements at 30, 60 and 90 min (p = 0.03, p = 0.008). There was a significant interaction of diet with genotype in ITT at 15, 30, 60, 90 and 120 min (p = 0.042, p = 0.028, p = 0.035, p = 0.015, p = 0.014)