Glucose and insulin may play an important role in the pathophysiology and symptomatology of Alzheimers disease (AD), and prior studies suggest interactions among glucose, insulin, gender and apolipoprotein E genotype. We analyzed the relationship between steroid-induced glucose elevation and gender, presence of the apolipoprotein E epsilon 4 (APOE-4) allele and cognition using data from a multicenter trial of prednisone therapy in AD. The low-dose prednisone regimen (initial dose: 20 mg/day, maintenance dose: 10 mg/day) caused a moderate increase in random blood glucose (mean post-baseline glucose 115 mg/dl). There was a significant interaction between rise in glucose, gender and presence of the APOE-4 allele. There was no important relationship between glucose and cognitive function at baseline or with prednisone treatment. Meta-analysis including data from three other AD trials showed a small influence of random blood glucose on cognitive scores. These results support a relationship between gender, apolipoprotein E genotype and glucose metabolism, but do not indicate that mild changes in glucose have an important impact on cognitive function.