Vascular Remodeling of Visceral Arteries Following Interruption of the Splenic Artery During Liver Transplantation.

ABSTRACT

Splenic artery (SA) ligation can be performed during liver transplantation (LT) to avoid portal hyperperfusion, which is involved in the pathogenesis of both small-for-size and SA syndrome. The SA can also be used as an inflow for arterial reconstruction. Exceptionally, SA interruption or agenesis has been associated with positive remodeling of collateral arteries supplying the spleen via the left gastric artery (LGA), short gastric vessels, and the gastroepiploic arcade (GEA), with subsequent severe upper gastrointestinal (GI) bleeding. To determine incidence, magnitude, predictors, and clinical implications of vascular remodeling after SA interruption during LT, we identified 465 patients transplanted in the period 2007-2017 who had the SA ligated or interrupted at LT. Among them, 88 had a computed tomography angiography suitable for evaluation of vascular remodeling after LT. The presence of prominent gastric arterial collaterals and the increase in LGA and GEA diameter were evaluated on 2-dimensional axial images and multiplanar reconstructions. Of the 88 patients, 28 (31.8%), 32 (36.4%), and 22 (25.0%) developed gastric collateralization graded as mild, moderate, or severe. Of the patients for whom comparison with pre-LT imaging was possible (n = 54), 51 (94.4%) presented a median 37% and 55% increase in LGA and GEA diameter, respectively. Severe gastric collateralization was associated with lower body mass index (odds ratio, 0.84; 95% confidence interval [CI], 0.71-0.98; P = 0.03), whereas a GEA caliper measurement increase was positively correlated with Model for End-Stage Liver Disease score (r2 = 0.12; 95% CI, 0.65-4.15; P = 0.008). Out of 465 patients, 2 (0.43%) had severe episodes of arterial upper GI bleeding, possibly exacerbated by vascular remodeling. In conclusion, vascular remodeling after SA interruption during LT is frequent and can aggravate GI bleeding during follow-up.