Demographic features
There were 10 male and 6 female CS patients, with a male to female ratio of 5:3. The median age of onset of symptoms was 7 years of age (range: 1 month to 38 years old, IQR: 15.5 years old), the median age of first clinic visit was 10 years old (range: 1 month to 38 year old, IQR: 17.6 years old), and the median age at diagnosis was 15 years old (range: 8 months to 39 years old, IQR: 24.5 years old). The median diagnostic delay, which was defined as the time between initial clinical presentation and final diagnosis, was 1.75 years (range: 1 month to 29 years, IQR: 6.2 years). Sex, age of onset of symptoms, age at first clinic visit, and age at diagnosis were not significantly associated with diagnostic delay (Table 1).
Initial clinical presentation
Initial symptoms of the 16 CS patients included fever (n=6), abdominal pain (n=3), abdominal distention (n=8), gastrointestinal bleeding (n=1), and fatigue (n=1). Most initial symptoms led directly to a first clinic visit, with the exception of three patients who had abdominal distention as the initial symptom but only presented later to hospital for other reasons (Table 2). None of these initial symptoms were associated with diagnostic delay (Table 1). Three patients had a positive family history, but similarly this was not associated with diagnostic delay (Table 1).
Laboratory findings
Laboratory findings of the 16 CS patients at diagnosis included decreased peripheral white blood cell count (n=11), anemia (n=11), and decreased platelet count (n=10). Among them, ten patients had pancytopenia. Some patients had abnormal liver function, with elevated serum alanine aminotransferase (ALT) (n=2), elevated serum bilirubin (n=2), decreased serum albumin (n=2), and prolonged prothrombin time (PT) (n=5). Of note, the two patients with elevated ALT had cholangitis when admitted, and their ALT levels returned to normal after anti-microbial treatment. Two patients had increased serum creatinine levels, both of whom had multiple renal cysts.
Positive autoantibodies were present in four CS patients: antinuclear antibodies (ANA) were tested in ten patients and two were positive; anti-mitochondrial antibodies (M2 subtype; AMA-M2) were tested in seven patients and two were positive; and anti-smooth muscle antibodies (SMA) were tested in seven patients and one was positive. Two patients were initially misdiagnosed as autoimmune disease. The first misdiagnosed patient was a 21-year-old man with positive ANA and slightly elevated aspartate transaminase (AST) and immunoglobulin G (IgG), who was misdiagnosed with autoimmune hepatitis (AIH) and was treated with corticosteroids and azathioprine for one month. There was no clinical improvement, and histopathological examination of his liver biopsy secured a diagnosis of CS. The other misdiagnosed patient was a 33-year-old woman who presented with low-grade fever and abdominal distention who had a positive ANA and AMA-M2 and slightly elevated gamma-glutamyltransferase (GGT) and alkaline phosphatase (ALP). She was initially misdiagnosed with primary biliary cirrhosis and treated with corticosteroids plus ursodeoxycholic acid for three months. Again, there was no clinical improvement and she developed recurrent cholangitis. MRCP was performed, and the diagnosis of CS was made according to typical imaging results. Her symptoms were soon controlled with antibiotics and her GGT and ALP returned to normal.
Lower white blood cell (WBC) counts (5 years vs. 4 months; p=0.01), lower platelet (PLT) counts (4.5 years vs. 6.5 months; p=0.01), and pancytopenia (4.5 years vs. 6.5 months; p=0.01) at diagnosis were associated with longer diagnostic delay. There were no other significant associations between laboratory findings and diagnostic delay (Table 1).
Imaging manifestations
All 16 CS patients were examined by US before diagnosis, 13 were examined by CT with or without contrast, and 12 by magnetic resonance cholangiopancreatography (MRCP). Some patients had several imaging studies at different hospitals before the final diagnosis was made. On average, US was performed 1.9 times and CT 1.3 times per patient before diagnosis. The accuracy of US, CT, and MRCP modalities was 25%, 69.2%, and 83.3%, respectively (Table 3). Fourteen patients were diagnosed by imaging, while two patients were finally diagnosed from the histopathological appearances on liver biopsy.
All 16 patients had US performed during the first visit to a hospital, while eight patients also had CT scans due to suspicions raised by physicians and radiologists. CT performed at the first hospital visited was associated with statistically significant shorter diagnostic delay (p=0.021; Table 1). The median diagnostic delay for patients with CT performed or not performed at the first hospital visit was 7.4 months and 6 years, respectively.
We examined the US signs in patients at their first hospital visit. Diffusive hepatic lesions, hepatic cysts, splenomegaly, and renal cysts were reported in 10 (62.5%), 10 (62.5%), 15 (93.8%), and 8 (50%) patients, respectively. The combination of diffuse hepatic lesions plus splenomegaly, hepatic cysts plus splenomegaly, and renal cysts plus splenomegaly were found in 10 (62.5%), 9 (56.3%), and 8 (50%) patients, respectively (Table 4).
Diagnostic timelines
The diagnostic timelines of all 16 patients were drawn to visualize important diagnostic time points, disease phase at diagnosis, and imaging modalities used to make the diagnosis (Figure 1). We divided the CS course into three phases: phase 1 (four patients, 25%), no proof of portal hypertension (i.e., hypersplenism); phase 2 (ten patients, 62.5%), discovered complications of portal hypertension without gastrointestinal varicosity bleeding; and phase 3. (2 patients, 12.5%), at least one recorded variceal bleed.