Evaluation of a model to predict poor survival in patients undergoing elective TIPS procedures

To validate a previously published model to predict the probability of patient death within 3 months after an elective transjugular intrahepatic portosystemic shunt (TIPS) procedure. The model is implemented with use of a nomogram or a formula.

Patients who underwent an elective TIPS procedure between May 1, 1999, and May 1, 2001, were selected. Patients who underwent emergency TIPS creation and patients with serum creatinine levels greater than 3.0 mg/dL were excluded. A total of 72 patients met the inclusion criteria. The patients were divided into two groups: group A (ethanol-induced cirrhosis; n = 23) and group B (non-ethanol-induced cirrhosis; n = 49). The model was applied and the predicted probability of death was compared to actual patient survival. A high risk score (R > or = 1.8) is associated with a high risk of death within 3 months after TIPS creation. Survival curves were estimated with use of Kaplan-Meier product limit estimates and were compared with use of the log-rank test. The model's accuracy was evaluated with use of the c-statistic. P values lower than.05 indicated statistical significance.

The technical success rate was 98.7%. The 3-month survival rate for the whole group was 79.7%. The predicted mortality rate was higher than the observed mortality rate. The c-statistic was 0.65 for the formula and 0.66 for the nomogram. Patients with a risk score of at least 1.8 had a 3-month survival rate of 54.6% and patients with a risk score lower than 1.8 had a 3-month survival rate of 84.9% (P =.037).

These results confirm that, after an elective TIPS procedure, patients with risk scores of at least 1.8 have a significantly lower 3-month survival rate than patients with risk scores lower than 1.8.