Percutaneous drainage (PCD) is the first-line treatment of intraabdominal abscesses. EUS-guided drainage (EUS-D) has also been employed in postoperative abdominal and liver abscesses. There are potential advantages of EUS-D over PCD, however the evidence to supportEUS-D is scant. To compare feasibility, safety, and clinical outcomes of EUS-D versus PCD of upper abdominal abscesses. Single center retrospective cohort study of all consecutive patients undergoing PCD or EUS-D of upper abdominal abscesses between January 2012-June 2017. Patients presenting at least one abdominal abscess proximal to the Treitz ligament were included. Abscess was defined by the presence of fever, pain and consistent imaging. Pancreatic fluid collections and liver transplant patients were excluded. Primary endpoints: technical success, clinical success and adverse event rates. Secondary endpoints: time to resolution, number of interventions and relapse rates. Technical success: ability to access and drain the abscess by placement of a drainage catheter/stent, with purulent effluent. Clinical success: complete resolution of clinical symptoms with disappearance or greater than 50% decrease in abscess size on imaging. To compare primary and secondary endpoints Pearson Chi-squared test or Fisher exact test were applied as warranted. 80 patients met inclusion criteria, 18 patients underwent EUS-D and 62 PCD. Baseline demographics and clinical characteristics are shown in Table 1. There were no statistically significant differences in age, sex, immunosuppressive therapy, number of lesions and etiology. Size was larger in the PCD group (80 vs 65.5 mm, p=0.04) and location around the gallbladder was more frequent in the PCD group (24.2% vs 11.1%, p=0.003). The transgastric route was chosen in 14 (77.8%) EUS-D. Covered metal stents were placed in 16 (89.9%) subjects (9 lumen apposing metal stents and 7 tubular self-expandable metal stents). Double pig-tail plastic stents were placed in 6 (33.3%) patients. Through-the-stent lavage/debridement was performed in 5 (27.8%). A naso-cavitary drainage catheter was placed in one patient for continuous irrigation. Procedure outcomes are shown in Table 2. There were no significant differences in primary or secondary endpoints. In the EUS-D group there were mild or moderate adverse events in four cases (22.2%): three patients presented gastrointestinal bleeding requiring conservative management and one presented a perforation which was closed during the procedure. In the PCD group, there were complications in 13 patients (21%), 7 of them severe. EUS-D presents a similar efficacy compared to PCD, with a non-significant reduction in the number of relapses. EUS-D affords large-diameter stents and through-the-stent intervention. These promising results warrant formal comparison in randomized trials.Tabled 1Table 1: Baseline characteristics of patientsEndoscopic drainage (n=18)Percutaneous drainage (n=62)PMale sex, n (%)12 (66.7%)45 (72.6%)0.63Age, mean (SD)66 (16.5)66.9 (15)0.83Immunosupressive therapy, n (%)1 (5.6%)1 (1.6%)0.40Size (mm), median (IQR)65.5 (51-80)80 (60-97.5)0.04Encapsulated, n (%)11 (61.1%)27 (43.6%)0.28Multiple lesions, n (%)2 (11.1%)17 (27.4%)0.21Location--0.003-Right hepatic lobe, n (%)5 (27.8%)22 (35.5%)-Left hepatic lobe, n (%)4 (22.2%)5 (8.1%)-Perivesicular, n (%)2 (11.1%)15 (24.2%)-Perigastric, n (%)6 (33.3%)6 (9.7%)-Subphrenic, n (%)014 (22.6%)-Left pararenal, n (%)1 (5.6%)0Etiology,--0.10-Biliary, n (%)9 (50%)12 (19.4%)-Postsurgical, n (%)6 (33.3%)31 (50%)-Other, n (%)1 (5.6%)9 (14.5%)-Unknown, n (%)2 (11.1%)10 (16.1%)Previous bile duct manipulation,--0,03-None, n (%)9 (50%)23 (37.1%)-ERCP, n (%)6 (33.3%)8 (12.9%)-Surgical, n (%)1 (5.6%)22 (35.5%)-Multiple, n (%)2 (11.1%)9 (14.5%) Open table in a new tab Tabled 1Table 2: ResultsEndoscopic drainage (n=18)Percutaneous drainage (n=62)PTechnical success, n (%)16 (88.9%)60 (96.8%)0.22Clinical success, n (%)16 (88.9%)51 (82.3%)0.50Time until clinical success (days), median (IQR)17.5 (5.5-25.5)17 (7-27)0.71Time until drainage removal (days), median (IQR)92 (31.5-124)10 (6-21)<0.001Number of procedures, median (IQR)1 (1-1)1 (1-1)0.23Relapse, n (%)010 (16.1%)0.11Time from procedure until relapse (days), median (IQR)37.5 (17-66)Adverse events,--0.91-Mild, n (%)4 (22.2%)6 (9.8%)-Severe, n (%)07 (11.2%) Open table in a new tab