Hyperhomocysteinemia (HHcy) is supposed to be one of the modifiable risk factors that, if treated, may delay the onset of Alzheimer's disease (AD). The relation between serum homocysteine (Hcy) and vitamin levels during AD and its preclinical phase was systematically reviewed. Searches through large literature and trial databases were conducted. Data were extracted from studies and, after quality assessment, analyzed using a meta-analysis software package. Nine qualitatively good case-control studies were identified. The pooled standardized mean difference (PSMD) of Hcy levels (631 patients, 703 controls) was 1.04 (0.44-1.63), indicating higher Hcy levels in AD patients. Levels of folate (PSMD=0.65) (0.34-0.95) (387 patients, 312 controls) and vitamin B(12) (PSMD=0.50) (-0.05-1.06) (387 patients, 312 controls) were lower in AD patients. Vitamin B(6) levels were evaluated in 1 case-control study and were not significantly lower in AD patients. Analysis of prospective cohort studies (2569 subjects) revealed a pooled relative risk for AD in HHcy of 2.5 (1.38-4.56, p<0.01). No specific randomized controlled trials (RCTs) concerning Hcy-lowering therapy and AD-risk were identified. Prospective studies on the relation between folate, vitamins B(6) and B(12) levels and the risk of developing AD are warranted, preferably in the form of RCTs.