Alcohol and estrogen replacement therapy in postmenopausal women. Direct and mediated effects on cognitive component processes.

The literature remains contentious regarding the separate and combined effects of moderate drinking and estrogen replacement therapy (ERT) on cognition. In the current study, the authors sought to disentangle the predictive utility of alcohol use, ERT and their interaction on the episodic and semantic memory stores of postmenopausal women. It was predicted that relationships between moderate drinking, ERT and cognition would be attenuated by demographic and health-related factors. Postmenopausal women (n = 298) completed a battery of cognitive tests designed to assess speed and accuracy of episodic and knowledge-based cognitive processing. Potentially confounding variables were categorized and tested as mediators in hierarchical regression analyses. Moderate drinking was a weak predictor of episodic availability prior to removal of potential mediators. ERT use was a significant predictor of episodic and knowledge-based availability; no mediators were identified. Alcohol moderated ERT, as a combined alcohol/ERT variable was shown to be related to cognition. Neither moderate drinking nor ERT use was associated with cognitive speed. These findings suggest that positive relationships between alcohol and cognition are likely mediated by other variables, and should not be regarded as a benefit of drinking. Further, results support ERT as a predictor of knowledge-based and episodic availability, independent of mood stabilization or socioeconomic influences. Finally, alcohol and ERT appear to interact to impact both episodic and knowledge-based performance. Copyright 2008 S. Karger AG, Basel.