Hormone replacement therapy and cognition: Systematic review and meta-analysis

Context: Some observational data suggest that hormone replacement therapy (HRT) may reduce the risk of cognitive decline and dementia but results have been conflicting. Objective: To review and evaluate studies of HRT for preventing cognitive decline and dementia in healthy postmenopausal women. Data Sources: Studies with English-language abstracts identified in MEDLINE (1966-August 2000), HealthSTAR (1975-August 2000, PsychINFO (1984-August 2000); Cochrane Library databases; and articles listed in reference lists of key articles. Study Selection: Randomized controlled trials and cohort studies were reviewed for the effects of HRT on cognitive decline; cohort and case-control studies were reviewed for dementia risk. No randomized controlled trials regarding dementia risk were identified. Data Extraction: Twenty-nine studies met inclusion criteria and were rated. Two reviewers rated study quality independently and 100% agreement was reached on Jadad scores and 80% agreement was reached on US Preventive Services Task Force quality scores. A final score was reached through consensus if reviewers disagreed. Data Synthesis: Studies of cognition were not combined quantitatively because of heterogeneous study design. Women symptomatic from menopause had improvements in verbal memory, vigilance, reasoning, and motor speed, but no enhancement of other cognitive functions. Generally, no benefits were observed in asymptomatic women. A meta-analysis of observational studies suggested that HRT was associated with a decreased risk of dementia (summary odds ratio, 0.66; 95% confidence interval, 0.53-0.82). However, possible biases and lack of control for potential confounders limit interpretation of these studies. Studies did not contain enough information to assess adequately the effects of progestin use, various estrogen preparations or doses, or duration of therapy. Conclusions: In women with menopausal symptoms, HRT may have specific cognitive effects, and future studies should target these effects. The meta-analysis found a decreased risk of dementia in HRT users but most studies had important methodological limitations.