To the editor:
The incidence of cerebral oedema as a complication of fulminant liver failure (FHF) has been reported to be as high as 80% (1). Recently, there has been a significant improvement in the supportive treatment of FHF and an observed fall in the incidence of cerebral oedema (61% in 1987 to 45% in 1993) (2). In this retrospective analysis of trends in mortality from FHF, we aimed to determine the extent of cerebral oedema as the primary cause of death in FHF during 6 years of follow-up in a single centre.
The Queen Elizabeth Hospital Birmingham serves a large area of England and Wales. Transfer of FHF patients to the Liver Unit is based on local guidelines. Between 1/1/1996 and 31/12/2001, 503 FHF patients were admitted, 118 died (62 males; median age: 44 years, range: 19–73; liver transplant recipients: 24) and were included in this study. All 118 had grade III/IV hepatic encephalopathy, were electively ventilated and admitted to ITU. Treatment for prevention of raised intracranial pressure (ICP) included the routine insertion of ICP bolts in all patients at risk of cerebral oedema (29 patients (25%)).
Data were collected from medical notes, the Liver Unit database and autopsy reports (33 patients). A broad definition of death associated with cerebral oedema was used, this included patients who had: (i) subdural bolts recording intracranial hypertension (ICH), (ii) subdural bolts and intermittent ICH without another cause of death, (iii) autopsy evidence of cerebral oedema/ICH, (iv) clinical signs of ICH, (v) a hyperacute presentation without sepsis, pneumonia, adult respiratory distress syndrome (ARDS), disseminated intravascular coagulation (DIC) or gastrointestinal haemorrhage. Multiorgan failure (MOF) was considered the cause of death if three of the following were present: (i) renal failure requiring continuous veno-venous haemofiltration (CVVH), (ii) sepsis with positive blood culture(s), (iii) respiratory failure secondary to pneumonia/ARDS, (iv) persistent DIC despite treatment in the context of systemic sepsis.
The most common cause of FHF was paracetomol overdose (POD) (n=57; 48.3%) followed by non-A–non-B hepatitis (n=37; 31.3%). MOF was the cause of death in 80 (67.7%) and cerebral oedema in 29 (24.5%). Of these, 14 had an ICP bolt showing ICH, and 13 had autopsy findings consistent with ICH. Figure 1 shows the causes of death for each aetiology of FHF. POD patients were significantly more likely to die of MOF (n=33; 57.9%) than cerebral oedema (n=20; 35.1%; P<0.05) as were non-A–non-B hepatitis patients (n=28; 75.6% vs. n=4; 10.8%).
This study demonstrates that MOF is the most common cause of death in FHF and that this is independent of the aetiology of FHF. The overall incidence of cerebral oedema continues to fall and there are several possible explanations. Firstly, a change in guidelines has led to earlier referral and earlier treatment with N-acetyl-cysteine (NAC). Indeed, even late administration of NAC in paracetomol-induced FHF, improves outcome (3–5) and decreases the incidence of cerebral oedema (4). Secondly, the supportive medical management of FHF has also improved, guided by local protocols. Finally, current transplant criteria encourage early listing of patients with FHF, before they are at risk of developing cerebral oedema.
In conclusion, this study both reiterates the prominence of MOF as a cause of death in FHF and the falling incidence of death from cerebral oedema. Our findings emphasise the importance of continually modifying and reviewing protocols for the management of MOF focusing on the prevention of infection.