Our patient with refractory HE and a large, occult SPSS successfully underwent ARTO with benefits persisting both radiologically and clinically at 6 weeks. This case is unique in several aspects including a) the size of the SPSS at 10x the threshold for risk of HE and death, b) difficulty in identification of the SPSS, and c) documented clinical and cognitive improvement following angiographic intervention.
At 809mm2, the SPSS identified in this case is far greater than the threshold identified by Praktiknjo et al, who additionally found that a total surface area of greater than 83mm2 was associated with higher model for end-stage liver disease score, history of overt HE, and lower 1-year survival compared to smaller SPSS(7). In addition to surface area, a diameter threshold of 8mm(3) compared to 23mm in our patient, as well as total number of SPSS, are both associated with presence of overt HE.
SPSS are present in 30–60% of those with cirrhosis(3). Common sites include gastrorenal, splenorenal, paraumbilical, oesophageal SPSS and additional complications include variceal bleeding, portal vein thrombosis and deterioration in liver disease(8). Most SPSS can be visualised on CT, as demonstrated in a study of 222 patients with cirrhosis who underwent CT abdomen with portal venous phase contrast, with an SPSS identified in 63.5% by two experienced radiologists(9). As demonstrated in our case, diagnosis can be challenging yet identification can have profound implications if angiographic intervention is feasible.
Balloon-occluded retrograde transvenous has long been used for treatment of gastric varices following haemorrhage and is being increasingly used for management of HE with large SPSS. Previous studies have used ammonia levels(5), presence/grade of HE, and hospitalisation(10) as markers of success. We used two forms of neurocognitive testing to contrast the patient’s ability pre- and post-ARTO. The Number Connection Test is a widely used assessment of visuospatial orientation and psychomotor speed, and is a component of the Psychometric Hepatic Encephalopathy Score(11). The patient was shown a sheet of paper with 15 numbered circles randomly spread across the paper and asked to draw a line connecting the circles in order from 1–15. A healthy subject should be able to complete the task within 30 seconds. Although there was an improvement in our patient from 62 seconds to 55 seconds, the persistent impairment may reflect minimal HE or another cognitive impairment. The baseline clockface drawing demonstrated constructional apraxia and conceptional deficits (incomplete numbers, inaccurate hand placement) associated with a positive predictive value of 0.96 for HE in patients with cirrhosis(12), with subsequent improvement in visuospatial construction. Additionally, the reduction in emergency hospital admissions in the three months following ARTO may reflect an improvement in functional status.