Intracranial pressure monitoring in acute liver failure: a review of 10 years experience

Introduction

Intracranial hypertension (ICH) complicates around 25% of grade III/IV encephalopathy in acute liver failure. Intracranial pressure (ICP) monitoring is controversial in this coagulopathic population.

Methods

We examined the ITU charts and prospectively maintained an ITU database of all patients admitted with acute liver failure who underwent ICP monitoring from December 1998 to December 2008. Data are presented as median and range.

Results

One hundred and fifteen patients were identified. Aetiology was acetaminophen (78), seronegative (14), drug induced (7), viral (14) and autoimmune (two). All patients had a Camino ICP bolt (subdural) inserted by the intensive care specialist. Coagulation support fresh frozen plasma (10 ml/kg) and platelets if the count was <100 were given. Recombinant factors were not given. Overall 66 survived and 49 died. Seventy patients underwent transplantation with 77% survival. Age was 31 (24 to 42), International Normalised Ratio (INR) 5.6 (3.9 to 9.2), creatinine 150 μmol/l (105 to 143), Na 139 (135 to 143). INR pre insertion of the bolt was 3.5 (2.2 to 7.3), platelets 109 (51 to 137). At bolt insertion, pulse was 100 (89 to 111), MAP 80 (72 to 89), ICP 19.5 (12 to 25), CO₂ 4.8 (4.3 to 5.5), lactate 3.5 (2.2 to 5.3), jugular venous saturation 76 (67 to 84), temperature 35.5 (34.8 to 36.4). Norepinephrine dose was 0.12 μg/kg/minute (0.02 to 0.4). ICP fell to 14 (9 to 19) at 6 hours and 13 (8 to 20) at 24 hours. Peak ICP was 30 (20 to 39), number of surges >20 per patient was 2 (0 to 7). Minimum cerebral perfusion pressure (CPP) was 45 (38 to 52). The pre-liver-transplant peak ICP was 27 (19 to 35), post was 23 (20 to 33). There were no correlates between clinical parameters, ICP or CPP at 0, 6, 12 or 24 hours. No difference in insertion ICP and CPP was found between groups. Sixteen patients had postmortem examination. Cerebral oedema was noted in seven, and in two there was evidence of an intracerebral bleed, fronto-parietal region (bolt site). Fifteen underwent CT scanning - of whom three had evidence of diffuse ischemia, oedema and tentorial herniation. One case had a small fronto-parietal bleed (survived). Mode of death was cerebral in 14 cases. In those who had a CPP <45 46% survived, whilst in those with a CPP >45 66% survived.

Conclusions

ICP monitoring is a relatively safe procedure in this high-risk cohort of patients. Three ICP bolt-related complications were observed, two died and one survived. In those who survived there were no clinically apparent neurological sequelae. Mortality is increased in those with CPP <45 but is not inevitable and should not preclude transplantation and/or aggressive medical management.