This review is not appropriate for update since radiotherapy is now routinely combined with cisplatin and there have been no further studies using hydroxyurea.

Seven studies were found to be suitable for inclusion from 33 identified as relevant. None of the trials provided adequate evidence to support the use of hydroxyurea owing to small sample size, large numbers of post-randomisation exclusions and questionable rules for censoring, particularly a failure to include treatment-related deaths in the survival analysis. Details of statistical analysis were limited and often confusing, and we felt meta-analysis would lead to unreliable and invalid conclusions. Most studies appeared to be double blind placebo-controlled studies but none give details of power calculations or reasons for stopping recruitment. Only two studies had more than 50 patients. Patients were excluded from analysis in most trials for treatment-related reasons; in one, less than half those recruited were used in the analysis, the remainder having been excluded because of tumour progression or treatment-related conditions e.g. septicaemia, worsening renal/hepatic function. In another trial five out of 20 in the hydroxyurea group died of treatment-related complications but the five-year survival group was presented as 94%.