The Management of Alcoholic Hepatitis
The Management of Alcoholic Hepatitis
Unselected emergency admissions to Glasgow Royal Infirmary with a clinical diagnosis of alcoholic hepatitis were studied. These patients were admitted between February 2005 and April 2009. The diagnosis of alcoholic hepatitis was based on clinical history of excessive alcohol consumption (>80 g/day until at least 3 weeks before admission), supportive biochemical and clinical picture (new onset jaundice: serum bilirubin >80 μmol/L, AST and ALT <500 IU) and the absence of acute drug toxicity, autoimmune liver disease, acute viral hepatitis, biliary obstruction or hepatocellular carcinoma. All patients were screened for sepsis with analysis and culture of urine, blood and ascitic fluid as appropriate, and with chest radiography. Patients presenting with acute upper gastro-intestinal haemorrhage or with chronic viral hepatitis were excluded.
The unit protocol was that only patients with a GAHS ≥9 were considered for treatment with either corticosteroids (prednisolone 40 mg/day) or pentoxifylline (400 mg t.d.s) at the discretion of the responsible consultant. No patient was started on combination treatment as first-line therapy. The GAHS was used as a trigger for treatment such that if the GAHS exceeded eight at any stage during admission, the patient was regarded as having entered a poor prognostic group. Thus, once this threshold was breached, treatment could be initiated even if subsequently the value fell below nine. Corticosteroid response was assessed 7–9 days after initiation of prednisolone treatment with a 25% fall in bilirubin from the immediately pre-treatment value taken as indicating response. The decision to discontinue or continue corticosteroids, or to add in or change to pentoxifylline was again at the discretion of the responsible consultant. The Lille Score and ECBL were also calculated, but did not determine clinical management of the patients.
Disease severity was determined using the GAHS, DF, MELD and ABIC scores. In some cases, it was not possible to calculate the MELD and ABIC scores on account of the Jaffe reaction which, for many analysers, leads to inaccurate measurement of creatinine in the context of hyperbilirubinaemia. The prognostic thresholds for each score was as previously published or recommended by guidelines: GAHS ≥9; DF ≥32; MELD ≥18; ABIC <6.71, 6.71–8.99, ≥9.0.
Data were gathered on all patients from the time of admission and during the first 7–9 days thereafter. Survival was determined by review of hospital and community records. For the purposes of DF, MELD, ABIC and GAHS assessment, survival from the date of admission was recorded. For assessment of corticosteroid response, survival from the date of initiating treatment was recorded.
Statistical analysis was performed using MedCalc® v10.4.0.
Approval was obtained to gather clinical and laboratory information as part of BIAS (Bloods in Alcoholic Hepatitis Study) from the West Glasgow Ethics Committee 1 in 2006. Data were held on a secure database and patient anonymity was preserved.
Methods
Unselected emergency admissions to Glasgow Royal Infirmary with a clinical diagnosis of alcoholic hepatitis were studied. These patients were admitted between February 2005 and April 2009. The diagnosis of alcoholic hepatitis was based on clinical history of excessive alcohol consumption (>80 g/day until at least 3 weeks before admission), supportive biochemical and clinical picture (new onset jaundice: serum bilirubin >80 μmol/L, AST and ALT <500 IU) and the absence of acute drug toxicity, autoimmune liver disease, acute viral hepatitis, biliary obstruction or hepatocellular carcinoma. All patients were screened for sepsis with analysis and culture of urine, blood and ascitic fluid as appropriate, and with chest radiography. Patients presenting with acute upper gastro-intestinal haemorrhage or with chronic viral hepatitis were excluded.
The unit protocol was that only patients with a GAHS ≥9 were considered for treatment with either corticosteroids (prednisolone 40 mg/day) or pentoxifylline (400 mg t.d.s) at the discretion of the responsible consultant. No patient was started on combination treatment as first-line therapy. The GAHS was used as a trigger for treatment such that if the GAHS exceeded eight at any stage during admission, the patient was regarded as having entered a poor prognostic group. Thus, once this threshold was breached, treatment could be initiated even if subsequently the value fell below nine. Corticosteroid response was assessed 7–9 days after initiation of prednisolone treatment with a 25% fall in bilirubin from the immediately pre-treatment value taken as indicating response. The decision to discontinue or continue corticosteroids, or to add in or change to pentoxifylline was again at the discretion of the responsible consultant. The Lille Score and ECBL were also calculated, but did not determine clinical management of the patients.
Disease severity was determined using the GAHS, DF, MELD and ABIC scores. In some cases, it was not possible to calculate the MELD and ABIC scores on account of the Jaffe reaction which, for many analysers, leads to inaccurate measurement of creatinine in the context of hyperbilirubinaemia. The prognostic thresholds for each score was as previously published or recommended by guidelines: GAHS ≥9; DF ≥32; MELD ≥18; ABIC <6.71, 6.71–8.99, ≥9.0.
Data were gathered on all patients from the time of admission and during the first 7–9 days thereafter. Survival was determined by review of hospital and community records. For the purposes of DF, MELD, ABIC and GAHS assessment, survival from the date of admission was recorded. For assessment of corticosteroid response, survival from the date of initiating treatment was recorded.
Statistical analysis was performed using MedCalc® v10.4.0.
Approval was obtained to gather clinical and laboratory information as part of BIAS (Bloods in Alcoholic Hepatitis Study) from the West Glasgow Ethics Committee 1 in 2006. Data were held on a secure database and patient anonymity was preserved.
Source...