Table 3.
Multiple Logistic Regression Results for Elevated Serum GM-CSF Ab Concentration (≥ 5mcg/mL)
| Factors | P-values | ||
|---|---|---|---|
| Age at diagnosis (>16yrs) | 0.6758 | ||
| Gender | 0.8317 | ||
| NOD2 risk allele carriage | 0.7367 | ||
| Cohort | 0.7530 | ||
| Smoking | 0.0686 | ||
| Ileal Location | 0.0086 | ||
| Perianal | 0.3149 | ||
| Disease duration (>3yrs) | 0.5496 | ||
| Cohort*smoking | 0.0322 | ||
| smoking*Ileal Location | 0.0015 | ||
| Cohort*smoking*Ileal Location | 0.0038 | ||
| Factors | By Ileal location | By Cohort | OR (95%CI) |
| Smoking Yes vs. No | Yes | Brisbane | 0.70 (0.35–1.41) |
| NIDDK | 0.32 (0.17–0.61) | ||
| No | Brisbane | 0.63 (0.21–1.89) | |
| NIDDK | 4.35 (0.83–20.4) | ||
A multiple logistic regression model was fitted with elevated GM-CSF-Ab as the response variable. Main effects for the variables age at diagnosis, gender, NOD2 allele, Cohort, Smoking, Ileal location, Perianal location and disease duration were included in the model automatically, while interaction terms between these main effects were included only if significant after backward elimination. Table 3 displays the p-values for all fitted terms in the final model, as well as the estimated odds ratios (OR) for significant association of Elevated GM-CSF-Ab with smoking. The odds ratio estimates were broken down by Ileal location and cohort as the significant association was defined by interactions of smoking with Ileal location and cohort factors.