Abstract Caroli’s disease (CD) and Caroli’s syndrome (CS) are rare congenital fibropolycystic liver disorders. While CS is known to be associated with more severe hepatic complications than isolated CD, direct longitudinal comparisons between the two entities remain limited due to their rarity. This single-center, retrospective, longitudinal observational cohort study descriptively compared clinical features, complications, and long-term outcomes in 19 patients diagnosed with CD ( n = 7) or CS ( n = 12) between November 2015 and December 2022. Diagnoses were confirmed by magnetic resonance cholangiopancreatography. Patients underwent standardized clinical, laboratory, endoscopic, and imaging assessments during follow-up. Outcomes included recurrent cholangitis, portal hypertension–related complications, hepatic decompensation, cholangiocarcinoma, and all-cause mortality. Statistical analyses were descriptive and exploratory. Patients with CS exhibited more frequent portal hypertension, lower platelet counts, higher bilirubin and international normalized ratio levels, and lower serum albumin compared with patients with CD. All cases of esophageal varices, hepatic decompensation, cholangiocarcinoma ( n = 3), and death ( n = 3) occurred exclusively in the CS group. Exploratory analyses demonstrated monotonic associations between adverse outcomes and higher aspartate aminotransferase levels, increased frequency of cholangitis episodes, thrombocytopenia, and elevated CA 19 − 9 levels. Kaplan–Meier analysis showed shorter median survival in CS compared with CD (45.0 vs. 59.0 months; log-rank p = 0.020), which should be interpreted descriptively given the limited sample size. In this descriptive longitudinal cohort, Caroli’s syndrome was associated with a more severe clinical course and poorer outcomes than isolated Caroli’s disease, consistent with existing literature. These findings provide detailed comparative outcome data and highlight clinically relevant patterns that may inform surveillance strategies and guide future multicentre, hypothesis-driven studies. Graphical Abstract
Hanafy et al. (Thu,) studied this question.