Prognostic factors and predictive models in hot gallbladder surgery: A prospective observational study in a high-volume center

Abstract

Backgrounds/aims: The standard treatment for acute cholecystitis, biliary pancreatitis and intractable biliary colics ("hot gallbladder") is emergency laparoscopic cholecystectomy (LC). This paper aims to identify the prognostic factors and create statistical models to predict the outcomes of emergency LC for "hot gallbladder."

Methods: A prospective observational cohort study was conducted on 466 patients having an emergency LC in 17 months. Primary endpoint was "suboptimal treatment," defined as the use of escape strategies due to the impossibility to complete the LC. Secondary endpoints were postoperative morbidity and length of postoperative stay.

Results: About 10% of patients had a "suboptimal treatment" predicted by age and low albumin. Postop morbidity was 17.2%, predicted by age, admission day, and male sex. Postoperative length of stay was correlated to age, low albumin, and delayed surgery.

Conclusions: Several predictive prognostic factors were found to be related to poor emergency LC outcomes. These can be useful in the decision-making process and to inform patients of risks and benefits of an emergency vs. delayed LC for hot gallbladder.

Keywords: Acute cholecystitis; Laparoscopic cholecystectomy; Prognostic factors.

Fig. 1

ROC curve (model “suboptimal treatment”). AUC = 0.757. ROC, receiver operating characteristic; AUC, area under the curve.

Fig. 2

ROC curve (model “postoperative morbidity”). AUC = 0.621. ROC, receiver operating characteristic; AUC, area under the curve.

Fig. 3

ROC curve (model “readmission”). AUC = 0.622. ROC, receiver operating characteristic; AUC, area under the curve.

Fig. 4

ROC curve (model “LOS”). AUC = 0.700. ROC, receiver operating characteristic; AUC, area under the curve; LOS, lenght of stay.