Deregulation of brain-derived neurotrophic factor (BDNF) is a possible contributor to the pathology and symptoms of Alzheimer's disease (AD). Most studies support an association between the dysfunction of BDNF and the pathogenesis of AD. This study aimed to evaluate the diagnostic value of peripheral BDNF levels in patients with AD and mild cognitive impairment (MCI) using meta-analytic techniques. A systematic search of the MEDLINE, EMBASE, ISI Web of Science, and the Cochrane Central database was performed and 34 eligible articles were identified for inclusion in the meta-analysis. Random-effects meta-analysis showed that AD patients had significantly decreased levels of peripheral BDNF compared with healthy control (HC) subjects (Hedges' g = - 0.725, 95% CI = -1.06 to - 0.39, p < 0.01). MCI patients showed a same trend with decreased BDNF levels compared with HC subjects (Hedges' g = - 0.296, 95% CI = - 0.57 to - 0.02, p < 0.01). Significant differences were found between AD and MCI subjects in peripheral BDNF levels (Hedges' g = - 0.462, 95% CI = - 0.95 to 0.03, p < 0.01). However, the ROC curve analysis revealed that the peripheral BDNF levels may not be an optimal biomarker potentially for AD and MCI diagnosis with a lower AUC (AD: 0.707; MCI: 0.573), less sensitivity (AD: 66.67%; MCI: 50.00%) and poor specificity (AD: 93.33%; MCI: 83.33%). These results suggested that AD or MCI is accompanied by reduction of peripheral BDNF, but the levels of circulating BDNF may not be suitable as a diagnostic marker for AD and MCI.
Keywords: Alzheimer’s disease; ROC analysis; brain-derived neurotrophic factor; meta-analysis; mild cognitive impairment.