Table 2.
Multivariable Cox regression model for disease free survival and overall survival at five years follow-up.
Outcome of interest | Complete Cases Model | Missing Data Multiple Imputation | ||||
---|---|---|---|---|---|---|
Disease Free Survival Model | N = 135 Events = 30 | N = 174 Events = 40 | ||||
Hazard Ratio | Lower-Upper 95% CI | p-value | Hazard Ratio | Lower-Upper 95% CI | p-value | |
MOR expression | 0.791 | 0.603–1.039 | 0.092 | 1.062 | 0.930–1.212 | 0.376 |
First postoperative 96 h transfusion (yes) | 0.991 | 0.392–2.503 | 0.985 | 1.060 | 0.701–1.603 | 0.784 |
ASA (Reference category = 1) | ||||||
2 | 0.707 | 0.155–3.223 | 0.654 | 0.854 | 0.427–1.710 | 0.657 |
3 | 0.936 | 0.195–4.481 | 0.934 | 0.994 | 0.475–2.080 | 0.986 |
4 | 1.322 | 0.159–11.007 | 0.796 | 0.517 | 0.129–2.069 | 0.351 |
Preoperative Hemoglobin (g∙dL−1) | 1.043 | 0.846–1.287 | 0.693 | 1.012 | 0.919–1.117 | 0.807 |
Number of affected lymph nodes | 1.283 | 0.921–1.788 | 0.141 | 1.028 | 0.780–1.322 | 0.828 |
CEA at diagnosis (U∙mL−1) | 1.811 | 1.245–2.635 | 0.002 | 1.058 | 0.877–1.28 | 0.557 |
Age (years) | 1.010 | 0.970–1.052 | 0.638 | 1.005 | 0.987–1.022 | 0.591 |
Overall survival model | N = 135 Events = 29 | N = 174 Events = 40 | ||||
MOR-1 expression | 1.023 | 0.784–1.335 | 0.869 | 1.031 | 0.906–1.173 | 0.645 |
First postoperative 96 h transfusion (yes) | 1.556 | 0.658–3.682 | 0.314 | 1.004 | 0.670–1.503 | 0.986 |
ASA score (Reference category = 1) | ||||||
2 | 0.954 | 0.119–7.629 | 0.965 | 0.898 | 0.479–1.685 | 0.737 |
3 | 1.948 | 0.247–15.357 | 0.527 | 1.072 | 0.538–2.138 | 0.843 |
4 | 2.375 | 0.208–27.07 | 0.486 | 0.832 | 0.183–3.786 | 0.812 |
Preoperative Hemoglobin (g∙dL−1) | 0.911 | 0.729–1.139 | 0.415 | 1.016 | 0.925–1.115 | 0.743 |
Number of affected lymph nodes | 1.482 | 1.110–1.978 | 0.008 | 0.971 | 0.774–1.218 | 0.800 |
CEA at diagnosis (U∙mL−1) | 1.485 | 1.017–2.170 | 0.041 | 1.031 | 0.859–1.24 | 0.746 |
Age (years) | 1.031 | 0.989–1.074 | 0.147 | 1.003 | 0.986–1.020 | 0.746 |
MOR-1 expression is introduced in both models as a 0 to 6 ordinal variable. The effect estimate is thus to be interpreted as the difference in hazard in the monitored time period when MOR-1 expression increases one level.