Endosonography-Guided Versus Percutaneous Gallbladder Drainage Versus Cholecystectomy in Fragile Patients with Acute Cholecystitis-A High-Volume Center Study

Abstract

Background and Objectives: Acute cholecystitis is a frequent cause of admission to the emergency department, especially in old and frail patients. Percutaneous drainage (PT-GBD) and endosonographic guided drainage (EUS-GBD) could be an alternative option for relieving symptoms or act as a definitive treatment instead of a laparoscopic or open cholecystectomy (LC, OC). The aim of the present study was to compare different treatment groups. Materials and Methods: This is a five-year monocentric retrospective study including patients ≥65 years old who underwent an urgent operative procedure. A descriptive analysis was conducted comparing all treatment groups. A propensity score was estimated based on the ACS score, incorporated into a predictive model, and tested by recursive partitioning analysis. Results: 163 patients were included: 106 underwent a cholecystectomy (81 laparoscopic (LC) and 25 Open (OC)), 33 a PT-GBD and 21 EUS-GBD. The sample was categorized into three prognostic groups according to the adverse event occurrence rate. All patients treated with EUS-GBD or LC resulted in the low risk group, and the adverse event rate (AE) was 10/96 (10.4%). The AE was 4/28 (14.2%) and 21/36 (58.3%) in the middle- and high-risk groups respectively (p < 0.001). These groups included all the patients who underwent an OC or a PT-GBD. The PT-GBD group had a lower clinical success rate (55.5%) and higher RR (16,6%) when compared with other groups. Conclusions: Surgery still represents the gold standard for AC treatment. Nevertheless, EUS-GBD is a good alternative to PT-GBD in terms of clinical success, RR and AEs in all kinds of patients.

Keywords: ACS score; cholecystectomy; cholecystitis; endosonography-guided gallbladder drainage; frailty; percutaneous gallbladder drainage.

Figure 1

Variable importance estimated by permutation-based mean decreased accuracy importance considering adverse event rate as outcome.

Figure 2

Recursive partitioning classification tree for adverse event occurrence. The terminal nodes categorized the study sample into three prognostic groups according to the adverse event occurrence rate. Adverse event rate was 10/96 (10.4%) in the low-risk group, 4/28 (14.2%) and 21/36 (58.3%) in the middle and high-risk groups, respectively (p < 0.001). Abbreviations: VLS: Videolaparoscopy, PT-GBD: Percutaneous Gallbladder drainage; EUS-GBD: Endoscopic ultrasound guided gallbladder drainage; ACS: American College of Surgeons risk score; AE: Adverse Event; SAE: Severe adverse event; Succ: Clinical success (CS); Rec: Recurrence (RR).

Figure 3

Receiver operating characteristic (ROC) curve and the corresponding area of the predictive model. Red line corresponded to the analysis before internal validation and green line corresponded to the analysis after bootstrapping-based internal validation.

Figure 4

Calibration plot. Smoothed (loess) calibration plots reporting increasing predicted probability of adverse events by the assessed model. The diagonal line indicates the ideal line of perfect correspondence of predicted to observed adverse event rate. Mean absolute error rate was 3.3%.