Table 3. Cox Proportional Hazards Regression Analysis of the Association of Local and Systemic Recurrence With Patient and Surgical Factorsa.
| Characteristic | Hazard ratio (95% CI) | ||
|---|---|---|---|
| Local recurrence, univariate | Systemic recurrence | ||
| Univariate | Multivariate | ||
| Body mass index | |||
| Underweight or normal | 1 [Reference] | 1 [Reference] | NA |
| Overweight | 0.96 (0.38-2.45) | 0.99 (0.48-2.04) | NA |
| Obese | 0.41 (0.09-1.86) | 0.59 (0.33-1.05) | NA |
| Sex | |||
| Female | 1 [Reference] | 1 [Reference] | NA |
| Male | 1.62 (0.85-3.07) | 0.94 (0.59-1.47) | NA |
| Neoadjuvant therapy | |||
| No | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Yes | 0.75 (0.31-1.83) | 1.99 (1.27-3.13)b | 1.70 (0.90-3.21) |
| AJCC stage (clinical) | |||
| I | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| II | 0.60 (0.36-1.01) | 1.35 (0.73-2.49) | 0.72 (0.34-1.55) |
| III | 1.42 (0.83-2.44) | 2.30 (1.31-4.03)b | 1.17 (0.46-2.98) |
| IV | 1.26 (0.14-10.99) | NAc | NA |
| Tumor height | 0.92 (0.76-1.12) | 1.03 (0.93-1.14) | NA |
| Quality of mesorectum | |||
| Incomplete | 1 [Reference] | 1 [Reference] | NA |
| Complete | 0.63 (0.08-5.06) | 0.49 (0.13-1.76) | NA |
| CRM | |||
| Negative | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Positive | 4.19 (2.86-6.15)b | 4.00 (2.22-7.19)b | 2.95 (1.26-6.91)b |
| DRM | |||
| Negative | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Positive | NAd | 4.88 (1.55-15.33) | 3.37 (0.72-15.74) |
Abbreviations: AJCC, American Joint Committee on Cancer; CRM, circumferential radial margin; DRM, distal resection margin; NA, not applicable.
a
Both models were clustered by site, and the multivariable model was adjusted for neoadjuvant therapy, stage, CRM, and DRM.
b
Significant association.
c
Patients with disease stage IV were removed before data were analyzed.
d
A reliable estimate could not be derived owing to a lack of cases.