Eight randomised controlled trials (7 published and 1 presented as an abstract), 1 quality-of-life paper and 1 practice guideline were eligible for inclusion in the systematic review of the evidence.
The pooled data suggested no significant difference between mucositis scores, whether or not amifostine was used (OR 0.11, 95% confidence interval, CI: 0.01, 1.26, P=0.08; chi-squared 13.31, d.f.=3, P=0.004). These data were based on the 4 studies which reported standard outcome measures. A pre-specified subgroup analysis found that amifostine was beneficial in patients undergoing radiochemotherapy (2 studies; OR 0.03, 95% CI: 0.00, 0.83, P=0.04; chi-squared 2.07, d.f.=1, P=0.15). This analysis was conducted for the first edition of the review. In the update, 2 additional studies were added, but one did not report any results for this outcome and the other had only four cases; as such the meta-analysis was not re-run. Two trials in patients having radiochemotherapy found a beneficial effect of amifostine (OR 0.03, 95% CI: 0.00, 0.83, P=0.04).
The pooled data suggested that amifostine was beneficial in acute xerostomia (OR 0.10, 95% CI: 0.02, 0.48, P=0.004), but significant heterogeneity was present (chi-squared 6.87, d.f.=2, P=0.032). These data were based on the 3 studies which reported standard outcome measures. Studies reporting non-standard outcome measures also reported benefits of the drug. The pooled data also suggested that amifostine was beneficial in late xerostomia (OR 0.19, 95% CI: 0.05, 0.64, P=0.008), but again, significant heterogeneity was present (chi-squared 5.32, d.f.=2, P=0.07). These data were also based on the 3 studies which reported standard outcome measures.
The results indicated that amifostine does not affect the anti-tumour effectiveness of radiotherapy with or without concurrent chemotherapy with carboplatin.
In terms of side-effects, nausea, vomiting, hypotension, asthenia and allergic reactions were the most commonly reported side-effects of amifostine, but they were rarely severe (grade 3 and above).
One study reported quality of life data. No differences were seen at baseline between patients with or without amifostine, but those treated with amifostine had a significantly better quality of life scores at 1, 7 and 11 months than those patients not treated with the drug.
In terms of the route of administration, similar results were found in one small study of patients treated with subcutaneous (19% incidence) and intravenous (23% incidence) amifostine (P-value or CIs were not reported).
The results of the analysis of publication bias were not presented. However, the authors reported that while the funnel plots appeared to be asymmetrical, Egger's and Begg's tests did not prove publication bias.