We agree fully with Block et al. that it is simply not possible to make a conclusive statement with the data now available regarding the efficacy or safety of high-dose antioxidant supplementation concurrent with chemotherapy and/or radiation therapy. Indeed, we ( 1 ) arrived at our recommendation, primum non nocere, after considering the limited number of clinical trials ( 2 , 3 ) that showed diminished local tumor control and decreased overall survival following radiation therapy plus antioxidant supplementation as well as the absence of other well-designed or adequately powered trials addressing this topic. Furthermore, although the chemo- and radioprotective effects of antioxidants suggest they may confer a potential benefit for normal tissues, there are also plausible antagonistic interactions between these nutrients and tumor DNA damage and apoptosis that are produced by reactive oxygen species generated by radiation therapy and many chemotherapy agents.
The reanalysis of the trials reported by Bairati, Meyer, and colleagues indicated that patients who smoked cigarettes and received antioxidant treatment during radiation therapy had the poorest disease control and survival outcomes ( 4 ). Although this observation is provocative, such subgroup analyses are often subject to statistical limitations, which can lead to inflated false-positive rates ( 5 ). It is also important to note that cigarette smoking alone has been associated with decreased tumor control and decreased survival among patients who receive radiation therapy ( 6 , 7 ). Thus, these outcomes may not be due solely to interactions with the antioxidant treatment. These are some of the reasons we proposed that a robustly designed phase 3 clinical trial is necessary to draw firmer conclusions regarding these variables. Although we discourage concurrent high-dose antioxidant supplementation as a routine complementary treatment during cancer therapy, we also encourage relevant research in this area because there may be potential for clinically significant benefit. In the meantime, we feel that the benefit-to-risk ratio cannot be calculated from the available literature, and so it is best to err on the side of doing no harm.