BACKGROUND AND AIMS

There is increasing interest in expanding the use of lumen-apposing metal stents (LAMSs) in patients with pancreatic fluid collections (PFCs). The aim of this study was to determine whether there is a hospital volume threshold for which patient outcomes could be optimized.

METHODS

Data from a large multicenter series of patients with PFCs treated with LAMSs were retrieved. Rate of adverse events (AEs) was the primary outcome. Multivariable models with restricted cubic splines were used to identify a hospital volume threshold by plotting hospital volume against the log odds ratio (OR) of AE rate. Propensity score matching was applied to obtain 2 well-balanced groups according to hospital volume, and univariate/multivariate logistic regression analysis was performed to identify significant predictors of AEs.

RESULTS

Overall, 516 patients were included. Increasing hospital volume was associated with a reduced AE rate (P = .03), and the likelihood of experiencing an AE declined as hospital volume increased up to 15 cases. After propensity score matching, 175 patients in the high-volume (>15 cases) and 132 in the low-volume hospital group were compared. Overall, 41 AEs were observed (13.3%), of which 14 (8%) and 27 (20.4%) occurred at high-volume and low-volume centers, respectively (P = .001). Severe and fatal events were observed more frequently in low-volume centers (6% vs 1.7% and 2.2% vs 0%, respectively; P = .05). In multivariate analysis, main pancreatic duct injury (OR, 2.62; 95% confidence interval [CI], 1.26-4.67; P = .02), presence of abnormal vessels (OR, 2.93; 95% CI, 1.41-5.02; P = .006), and institutional experience (OR, 2.95; 95% CI, 1.48-5.90; P = .002) were significant predictors of AEs.

CONCLUSIONS

With 15 procedures representing the minimum number of cases associated with the lowest risk for postprocedural AEs, hospital volume is associated with improved outcomes. (Clinical trial registration number: NCT03903523.).