Double-blind randomized phase II study of hyperbaric oxygen in patients with radiation-induced brachial plexopathy

Background: Radiation-induced brachial plexopathy (RIBP) is an untreatable complication of curative radiotherapy for early breast cancer, characterize d by chronic neuropathic pain and limb paralysis. Hyperbaric oxygen (HBO2) therapy is known to promote healing of tissue rendered ischaemic by radioth erapy, but is untested in RIBP. Methods: Thirty four eligible research volunteers suffering from RIBP were randomized to HBO2 or control group. The HBO2 group breathed 100% oxygen fo r 100 min in a multiplace hyperbaric chamber on 30 occasions over a period of 6 weeks. The control group accompanied the HBO2 group and breathed a gas mixture equivalent to breathing 100% oxygen at surface pressure. All volun teers and investigators, except the operators of the hyperbaric chamber and the trial statistician, were blind to treatment assignments. The warm sens ory threshold, which measures the function of small sensory fibres, was sel ected as the primary endpoint. Findings: Pre-treatment neurophysiological tests were grossly abnormal in t he affected hand compared to the unaffected hand in both HBO2 and control g roups, as expected, but no statistically significant differences were noted in either group at any time up to 12 months posttreatment. However, normal ization of the warm sensory threshold in two of the HBO2 group was reliably recorded. Two cases with marked chronic arm lymphoedema reported major and persistent improvements in arm volume for at least 12 months after treatme nt with HBO2. Interpretation: There is no reliable evidence to support the hypothesis tha t HBO2 therapy slows or reverses RIBP in a substantial proportion of affect ed individuals, although improvements in warm sensory threshold offer some suggestion of therapeutic effect. Improvement in long-standing arm lymphoed ema was not anticipated, and justifies further investigation. (C) 2001 Else vier Science Ireland Ltd. All rights reserved.