BACKGROUND: Amyotrophic lateral sclerosis, also known as motor neuron disease, is a progressive neuromuscular disease that causes disability and eventual death. Various amino acid preparations, the three branched-chain amino acids (L-leucine, L-valine and L-isoleucine) or, alternatively, L-threonine have been used as experimental therapy. OBJECTIVES: To examine the efficacy of amino acid therapies in prolonging survival and/or slowing the progression of amyotrophic lateral sclerosis/motor neuron disease. SEARCH STRATEGY: We searched the Cochrane Neuromuscular Disease Group trials register (searched February 2003), MEDLINE (from January 1966 to December 2002) and EMBASE (from January 1980 to December 2002) databases and reports of specialist conferences. Authors of known studies were contacted. SELECTION CRITERIA: We included randomised or quasi-randomised trials of participants with a clinical diagnosis of amyotrophic lateral sclerosis/motor neuron disease treated with all combinations of amino acids. Our primary outcome measure was survival determined by a pooled hazard ratio of all studies. Our secondary outcome measures were (in order of priority): survival at six and 12 months, muscle strength, any validated rating scale of physical function, quality of life, proportion of patients completing therapy and proportion of patients reporting adverse events attributable to treatment. DATA COLLECTION AND ANALYSIS: We identified six eligible trials and rejected a further seven because of incomplete data or inadequate duration. Eligible studies were rated for methodological quality and missing data sought from the authors. After this examination two studies were excluded from analysis. Our pooled survival analysis was performed by the Parmar method, other statistical calculations were done using the Review Manager 4.2 software package. MAIN RESULTS: No benefit could be demonstrated for either branched-chain amino acids or L-threonine in improving survival in amyotrophic lateral sclerosis/motor neuron disease. Neither could we find evidence of an effect of either treatment on muscle strength or disability as measured by functional rating scales. No study assessed quality of life. Both branched-chain amino acids and L-threonine appeared well tolerated and caused a degree of adverse events comparable to that of the control medication. REVIEWER'S CONCLUSIONS: There is no evidence to support a beneficial effect of either branched-chain amino acids or L-threonine in amyotrophic lateral sclerosis/motor neuron disease.