Skip to main content
. 2021 Sep 29;22(19):10541. doi: 10.3390/ijms221910541

Table 4.

Multinomial logistic regression of VAT and SAT gene expression associated with NAFLD.

Variable B S.E. (B) Exp B (OR) (95% C.I) Sig. R2 (%) Correct Prediction (%) M. Sig.
MODEL 1
SAT-TGFB1 1.397 0.642 4.04 (1.15–14.23) 0.030 45.2 80 0.000
VAT-P53 −1.400 0.634 0.25 (0.07–0.85) 0.027
Age −0.016 0.040 0.98 (0.91–1.06) 0.683
MODEL 2
SAT-TGFB1 110.07 8455.23 6.33 (0.00–>50.00) 0.790 100 100 0.000
VAT-P53 −56.76 5653.40 <0.001 (–) 0.792
VAT-CIDEA −390.28 13264.12 <0.001 (–) 0.777
SAT-UCP2 264.87 9103.50 10.73 (0.00–>50.00) 0.777
SAT-ATG5 −54.44 2596.91 <0.001 (–) 0.783
SAT-MRC1 −97.42 7094.69 <0.001 (–) 0.789
VAT-ANGPT2 −149.39 5819.60 <0.001 (–) 0.780
Age 8.92 312.56 74.90 (0.00–>100.00) 0.777

B, Beta coefficient; S.E. (B), Standard error of Beta; Exp B (OR) (95% C.I), Exponent beta with 95% confidence intervals; Sig, p value tested for each independent variable enter in the model; R2, Nagelkerke R Square of the model; Correct prediction (%), Percentage of patients correctly classified for NAFLD according to model; M. Sig, Model significance based on Omnibus test of the model coefficients.