Table 5.
Predictors of extended times from CT to surgery
| Univariable analysis | Multivariable analysis | |||
|---|---|---|---|---|
| Odds ratio | P | Odds ratio | P | |
| Age at surgery (years) | 0·019 | 0·088 | ||
| < 65 | 1·00 (reference) | 1·00 (reference) | ||
| 65–74 | 1·00 (reference) | 0·574 | 2·47 (0·94, 6·50) | 0·066 |
| ≥ 75 | 1·00 (reference) | 0·007 | 3·26 (1·02, 10·36) | 0·046 |
| Female sex | 0·84 (0·51, 1·39) | 0·491 | n.s. | |
| BMI (kg/m2) | 0·536 | n.s. | ||
| < 25 | 1·00 (reference) | |||
| 25–29 | 0·80 (0·46, 1·39) | 0·424 | ||
| ≥ 30 | 1·17 (0·58, 2·39) | 0·656 | ||
| Smoking status | 0·992 | n.s. | ||
| Non‐smoker | 1·00 (reference) | |||
| Smoker | 0·98 (0·41, 2·37) | 0·963 | ||
| Ex‐smoker | 1·04 (0·49, 2·23) | 0·910 | ||
| Charlson Co‐morbidity Index > 2 | 1·37 (0·82, 2·31) | 0·231 | n.s. | |
| Diagnosis | < 0·001 | n.s. | ||
| Ductal adenocarcinoma | 1·00 (reference) | |||
| Cholangiocarcinoma | 4·00 (1·75, 9·12) | < 0·001 | ||
| Ampullary cancer | 2·79 (1·44, 5·43) | 0·002 | ||
| Jaundice | 1·22 (0·54, 2·76) | 0·631 | n.s. | |
| T3–4 category | 0·32 (0·18, 0·56) | < 0·001 | n.s. | |
| N1–2 category | 0·34 (0·20, 0·59) | < 0·001 | n.s. | |
| Lymph node ratio | < 0·001 | 0·033 | ||
| 0 | 1·00 (reference) | 1·00 (reference) | ||
| 0·01–0·20 | 0·31 (0·16, 0·60) | < 0·001 | 0·50 (0·19, 1·30) | 0·154 |
| > 0·20 | 0·31 (0·16, 0·58) | < 0·001 | 0·23 (0·08, 0·70) | 0·010 |
| No. of EUS scans | < 0·001 | < 0·001 | ||
| 0 | 1·00 (reference) | 1·00 (reference) | ||
| 1 | 9·6 (5·0, 18·2) | < 0·001 | 11·9 (4·7, 30·7) | < 0·001 |
| > 1 | 46·4 (10·5, 205·1) | < 0·001 | 67·2 (7·1, 635·8) | < 0·001 |
| MRI | 1·96 (1·07, 3·58) | 0·029 | n.s. | |
| PET | 6·6 (0·8, 54·8) | 0·079 | 9·4 (0·6, 142·6) | 0·108 |
| No. of PBD procedures | < 0·001 | < 0·001 | ||
| 0 | 1·00 (reference) | 1·00 (reference) | ||
| 1 | 11·5 (5·6, 23·8) | < 0·001 | 10·8 (4·2, 27·8) | < 0·001 |
| > 1 | 43·3 (15·1, 123·8) | < 0·001 | 29·9 (7·6, 118·6) | < 0·001 |
Values in parentheses are 95 per cent confidence intervals. Results are from binary logistic regression analysis, with Q4 versus Q1 as the dependent variable; hence odds ratios greater than 1 represent a greater chance of extended time to surgery. The multivariable model used a backwards stepwise approach to variable selection, with all factors analysed in the univariable analysis considered for inclusion. The final model was based on 221 patients, after exclusions owing to missing data. n.s., Not selected for inclusion by the stepwise procedure; EUS, endoscopic ultrasonography; PBD, preoperative biliary drainage.