Fifty-two studies, enrolling 3,749 patients: 2,927 patients received interferon and 822 had either not been treated or had received placebo.
Twenty-nine studies evaluated interferon alpha-2b, 12 studies evaluated interferon alpha-2a, 11 studies evaluated interferon alpha-n, 1 study evaluated interferon alpha-2c, and in 1 the type of interferon is not given; 2 studies evaluated 2 types of interferon.
Interferon-alpha therapy caused a normalisation of serum transaminases in 51.2% of patients (1,499 out of 2,927).
Only 21.7% of patients (482 out of 2,218) still had normal serum transaminases for at least 3 months after discontinuation of therapy. Spontaneous long-term normalisation was 2.7% (22 out of 822) in untreated or placebo patients. Long-term benefit of interferon therapy is more than 8-fold higher than in the control group (chi-squared=156.1, p<0.00001).
Short-term responses were approximately doubled when weekly doses of 9 MU or more were compared to doses of 6 MU or less. Doses of 12 MU or more did not show an increase compared to 9 MU. Response rates were 25.8% for 6 MU or less, 53.7% for 9 MU, and 57.1% for 12 MU or more (Pearson's chi-squared=87.2, p<0.00001).
The long-term response gradually increased with increasing doses of interferon, with response rates of 9.2% for 6 MU or less, 16.7% for 9 MU and 31.3% for 12 MU or more (Pearson's chi-squared=59.9, p<0.00001).
Treatment for 9 months or more doubled the long-term response rate (29.4%), compared to treatment for 6 months or less (14.6%; chi-squared=64.0, p<0.00001).
Similarly, the long-term response was approximately 3-times better for patients receiving a total dose of more than 240 MU (30.6%), compared to those receiving less than 240 MU (10.8%), (chi-squared=103.3, p<0.00001).
There was no significant association between the long-term response rate and the type of interferon given.
There was no significant difference in the results from studies published in full and abstracts, or between escalating or constant dose.
Identifying the effects of cirrhosis as a pre-treatment selection criterion, both short- (28.9%) and long-term (12.2%) response rates were significantly lower in patients with cirrhosis present, compared to those without cirrhosis (59.8 and 31.5% respectively; chi-squared=59.1 and 12.1, p<0.00001 and p=0.00013, respectively).