Literature was searched using Medline, Embase, PsycInfo, Cochrane Library, Allied and Complementary Medicine Database (AMED), and Cumulative Index to Nursing and Allied Health Literature (CINAHL) up to September 2011. Treatment effects were summarized across cognitive subdomains, and effect sizes were estimated using Hedge's g and random effects modeling.
Ten RCTs were combined quantitatively. There was no effect of n-3 FAs on composite memory (g = 0.04 [95%CI: 鈭?.06-0.14], N = 934/812, p = 0.452). When examined by domain, no overall benefit for immediate recall (0.04 [鈭?.05-0.13], N = 934/812, p = 0.358) was detected; however, an effect in CIND subjects (0.16 [0.01-0.31], N = 349/327, p = 0.034) was found. A benefit for attention and processing speed was also detected in CIND (0.30 [0.02-0.57], N = 107/86, p = 0.035), but not healthy subjects. Benefits for delayed recall, recognition memory, or working memory and executive function were not observed. Treatment did not benefit AD patients as measured by the Mini-Mental State Examination (MMSE) or Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-cog). No differences in adverse events (AE), dropout, or dropout due to AE between groups were observed.
These results suggest an effect of n-3 FAs within specific cognitive domains in CIND, but not in healthy or AD subjects.