In this retrospective study, 78 patients with GV bleeding who underwent TIPS creation with or without embolotherapy with metallic coils and/or plugs from 1999 to 2014 were identified. Individuals who had a bare-metal TIPS and/or lacked post-TIPS imaging or endoscopic follow-up were excluded. The final cohort included 26 patients (16 men; median age, 54 y; median Model for End-stage Liver Disease score, 16). Variceal types, supplying vessels, and postprocedure GV patency on cross-sectional imaging or endoscopy were assessed. The primary study outcome measure was GV patency rate as a surrogate for efficacy of TIPS creation with or without embolization.
GVs included gastroesophageal varix types 1 (n = 10) and 2 (n = 2), isolated GV types 1 (n = 4) and 2 (n = 2), and unspecified (n = 8). TIPS creation resulted in a median final portosystemic pressure gradient of 7 mm Hg. Multiple GV-supplying vessels (left/posterior/short gastric veins) were present in 65% of patients (n = 17). Embolization was performed in 69% (n = 18). Thirteen, four, and nine patients had imaging, endoscopic, or both imaging/endoscopic follow-up. GV patency rate was 65% (n = 17; 61%/75% with/without embolization) at a median of 128.5 days (range, 1–1,295 d) after TIPS creation. Incidence of recurrent bleeding was 27% (n = 7), and the 90-day mortality rate was 15% (n = 4).
In this study, most GVs showed persistent patency despite TIPS decompression and variceal embolization, and the incidence of recurrent bleeding was high. The findings suggest suboptimal efficacy for GVs, and indicate a need for study of alternative or adjunctive approaches to GV treatment, such as chemical obliteration.