Serum bilirubin and early mortality after transjugular intrahepatic portosystemic shunts: results of a multivariate analysis.

PURPOSE: To examine the prognostic utility of the serum bilirubin level before transjugular intrahepatic portosystemic shunt (TIPS) creation as an independent predictor of 30-day mortality in patients who underwent TIPS creation for treatment of variceal hemorrhage.
MATERIALS AND METHODS: Multiple covariates from a cohort of 220 consecutive patients undergoing TIPS creation were analyzed with use of univariate and multivariate logistic regression. These included pre-TIPS total bilirubin levels, modified Child-Pugh class, APACHE II score, intubation status, etiology of liver disease, and acute versus elective shunting.
RESULTS: The mean pre-TIPS serum total bilirubin level was 3.2 mg/dL (range, 0.4-40.3 mg/dL). The bilirubin level was <3 mg/dL in 102 patients, > or = 3.0 mg/dL in 58, > or = 4.0 mg/dL in 34, and > or = 5.0 mg/dL in 27. Each 1.0-mg/dL increase in total bilirubin was associated with 40% greater odds of 30-day mortality (odds ratio = 1.4; 95% CI = 1.2-1.7). Using each threshold as its own referent, bilirubin levels at or greater than 3.0, 4.0, and 5.0 mg/dL stratified patients into increased odds of early death by 5.7, 9.7, and 19.2 times, respectively (all P <.001). A pre-TIPS APACHE II score of >18 increased the odds of early death by a factor of 5.6 (95% CI = 2.4-8.7); modified Child-Pugh class C (vs classes A and B combined) alone increased the odds by a factor of 8.1 (95% CI = 3.6-18.1). Only one of 20 patients (5%) with a pre-TIPS bilirubin level >6.0 mg/dL survived more than 30 days after TIPS creation. In acutely bleeding patients (n = 122) undergoing TIPS creation, bilirubin levels > or = 3.0, > or = 4.0, and > or = 5.0 mg/dL stratified patients into odds ratios of 4.4, 7.1, and 9.8, respectively, compared with 7.1, 13.2, and 9.2 for patients undergoing elective TIPS creation. Combining endotracheal intubation (n = 72) and bilirubin strata yielded mortality odds of 8.3, 12.5, and 20.8 compared with odds of 2.3, 4.6, and 11.2 in nonintubated patients. Combining alcoholic cirrhosis (n = 129) with bilirubin levels yielded mortality odds of 8.0, 10.6, and 18.0 compared with other etiologies of liver disease (odds ratios = 2.9, 7.3, and 22.7).
CONCLUSION: An elevated pre-TIPS bilirubin level is a powerful independent predictor of 30-day mortality after TIPS creation with a 40% increased risk of death for each 1-mg/dL increase above 3.0 mg/dL. The predictive value of this criterion is increased in patients who undergo TIPS procedures electively. The magnitude of the effect on mortality is similar to that of APACHE II scores and modified Child-Pugh class but is simpler to ascertain.