Table 5.
Multiple regression analysis to evaluate changes in serum cytokine concentrations as determinants of the changes in insulin sensitivity resulting from weight loss.
Variable | Entry into model | Coefficient | SE | P-value | R2 |
---|---|---|---|---|---|
Full model | ---- | ---- | ---- | <0.0001 | 0.63 |
Intercept | ---- | −0.373 | 0.39 | 0.34 | ---- |
Pentraxin-3 change, ng/mL | 1st | 2.776 | 1.067 | 0.01 | 0.28 |
Leptin change, pg/mL | 2nd | −0.023 | 0.006 | 0.001 | 0.17 |
GIP AUC, pmol/L·min | 3rd | −0.001 | 0.0003 | 0.0002 | 0.14 |
HMW adiponectin change, μ/mL | 4th | −2.404 | 1.132 | 0.039 | 0.04 |
Regression analysis was performed using a forward selection approach, only including cytokines that were significant predictors of changes in insulin sensitivity, based on bivariate correlations as shown in Table 4. Analyses were performed on transformed data (log or power transformations, as appropriate for each variable) to normalize data distributions.