Fourteen RCTs were included in the review (n=8,122 patients). The trial sample size ranged from 96 to 1,388 patients. There were 17 comparisons. Follow-up ranged from 16 to 88 months.
Disease-free survival: Interferon-alpha treatment was statistically significantly superior to control treatments in terms of disease recurrence (HR 0.82, 95% CI 0.77 to 0.87; I2=24%; 14 RCTs). Sensitivity analysis, leaving trials out, resulted in hazard ratio estimates that ranged from 0.81 to 0.83. Subgroup/meta-regression analyses did not show any significant differences according to interferon-alpha regimen, disease stage, study design, year of publication, length of follow-up, planned treatment duration, percentage of patients enrolled with lymph node disease, or discontinuation
Overall survival: Interferon-alpha was statistically significantly superior to control in terms of risk of death (HR 0.89, 95% CI 0.83 to 0.96; I2=18%; 12 RCTs). Sensitivity analysis, leaving trials out, resulted in hazard ratio estimates that ranged from 0.87 to 0.91. Subgroup/meta-regression analyses did not show any significant differences according to interferon-alpha regimen, disease stage, study design, year of publication, length of follow-up, planned treatment duration, percentage of patients enrolled with lymph node disease, or discontinuation.
Publication bias was found to be statistically significant (Begg’s test p=0.04). The trim-and-fill method estimated that two further trials would be needed to obtain funnel plot symmetry, but their inclusion in the meta-analysis would not change the overall significance of results.