Table 4.
Results of multiple logistic regression model for the top discriminating metabolites SDMA and glutamine in addition to breed, sex, age and fasting status.
| Analysis of maximum likelihood estimates | |||
|---|---|---|---|
| Variable | B | SE | p |
| Intercept | −3.326 | 2.311 | 0.150 |
| Breed (GD) | −3.183 | 1.954 | 0.103 |
| Sex (female) | −1.085 | 1.356 | 0.424 |
| Age | −0.036 | 0.032 | 0.255 |
| Fasting (yes) | 1.989 | 1.443 | 0.168 |
| SDMA | −0.229 | 0.307 | 0.456 |
| Glutamine | 1.146 | 0.475 | 0.016 |
| SDMA*breed (GD) | 1.030 | 0.499 | 0.039 |
| Parameter | OR | Lower 95%CI | Higher 95%CI |
| SDMA (breed = GD) | 2.23 | 1.181 | 4.206 |
| SDMA (breed = GS) | 0.80 | 0.436 | 1.451 |
| Glutamine | 3.15 | 1.240 | 7.986 |
B = coefficient, SE = standard error, GD = Great Dane, SDMA = symmetric dimethylarginine. DF = 1. Odds ratio estimates of the significant variables in the model. Odds ratio of SDMA is given in the subgroups of Great Danes and German Shepherd because of the interaction between breed and SDMA.
OR = odds ratio, CI = confidence interval, SDMA = symmetric dimethylarginine, GD = Great Dane, GS = German Shepherd.
Pearson goodness-of-fit statistics: value/DF = 0.781, p = 0.806
Model fit statistics: AIC = 39.651
ROC association statistics: AUC = 0.930, lower 95%CI = 0.850, higher 95%CI = 1.000.