Transjugular intrahepatic portosystemic shunt induced short- and long-term cerebral blood flow variations in cirrhotic patients: an arterial spin labeling MRI study

Short- and long-term effects of transjugular intrahepatic portosystemic shunt (TIPS) on cerebral blood flow (CBF) in patients with cirrhosis are still unclear. The purpose of this longitudinal study was to explore CBF alteration patterns in cirrhotic patients after TIPS. Thirteen cirrhotic patients (7 male, 6 female, mean age 50.0 ± 9.3 years) underwent arterial-spin labeling (ASL) MRI 1–9 days (median 1 days) before TIPS. Follow-up MR examinations were performed about 1 week (median 6 days), 3 months (n = 6), 6–9 months (n = 5) and 12–18 months (n = 5) after TIPS. CBF, ammonia level, Child-Pugh score, number connection test type A (NCT-A) and digit symbol test (DST) scores were converted into relative values by dividing by his/her pre-TIPS values, and then, compared via one-way analysis of variance (ANOVA). Correlations between the pre- and post-TIPS changes of relative CBF (rCBF) and the changes of relative ammonia (rAmmonia), Child-Pugh (rChild-Pugh), and NCT-A/DST (rNCT-A/rDST) scores were calculated by crossing subjects. Compared with the pre-TIPS level, the global rCBF slightly increased by 10.9 % about 1 week later, then rapidly decreased by 14.2 % 3 months later, and flatly decreased by 17.2 % in 6–9 months and 18.0 % in 12–18 months following TIPS. The changes of 3-month rDST score were slightly correlated with 3-month rCBF rather than 1-week rCBF, (P < 0.1, FDR-corrected) No difference was found between the pre- and post-TIPS rAmmonia levels, rChild-Pugh and rNCT-A/rDST scores (Post-hoc P > 0.05). CBF measured at different time points after TIPS insertion showed different patterns, indicating varying longitudinal effects of TIPS on CBF. A sharp decline of rCBF was found in the 1 week to 3 months period after insertion, indicating that high event rate of hepatic encephalopathy might relate with the unadaptable CBF in patients undergoing TIPS insertion.