Shiu Lun Au Yeung1, Chaoqiang Jiang2, Kar Keung Cheng3, Weisen Zhang2, Tai Hing Lam4, Gabriel Matthew Leung1, C Mary Schooling5. 1. School of Public Health, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China. 2. Guangzhou Number 12 Hospital, Guangzhou, China. 3. Department of Public Health and Epidemiology, University of Birmingham, UK. 4. School of Public Health, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China. Electronic address: hrmrlth@hku.hk. 5. School of Public Health, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China; City University of New York Graduate School of Public Health and Health Policy, New York, NY, USA.
Abstract
OBJECTIVE: The role of estrogen in cognitive function and depressive symptoms is controversial due to discrepancies between results from randomized controlled trials (RCT) and observational studies. Mendelian randomization analysis may provide further insights concerning the role of estrogen in these outcomes as it assesses the effect of lifelong endogenous exposure but is less vulnerable to confounding than observational studies. METHOD: We used separate sample instrumental variable analysis to estimate the association of log 17β estradiol with cognitive function (Delayed 10 word recall, and Mini Mental State Examination (MMSE)) and depressive symptoms (Geriatric Depression Scale (GDS)) in older Chinese women of the Guangzhou Biobank Cohort Study (GBCS, n=3086). The estimate was derived based on the Wald estimator, the ratio of the association of genetic determinants (rs1008805 and rs2175898) of log 17β-estradiol with cognitive function and depressive symptoms in GBCS and the association of log 17β-estradiol with genetic determinants in the sample of young women in Hong Kong (n=236). RESULTS: Genetically predicted 17β-estradiol was not associated with delayed 10-word recall (0.42 words per log increase in 17β-estradiol (pmol/L), 95% confidence interval (CI) -0.49 to 1.34) MMSE (0.39 per log increase in 17β-estradiol (pmol/L), 95% CI -0.87 to 1.65) or GDS (0.24 per log increase in 17β-estradiol (pmol/L), 95% CI -0.57 to 1.05). CONCLUSION: These results were largely consistent with evidence from RCTs and did not show any beneficial effect of estrogen on cognitive function and depressive symptoms. However, larger Mendelian randomization analyses are needed to identify any minor effects.
OBJECTIVE: The role of estrogen in cognitive function and depressive symptoms is controversial due to discrepancies between results from randomized controlled trials (RCT) and observational studies. Mendelian randomization analysis may provide further insights concerning the role of estrogen in these outcomes as it assesses the effect of lifelong endogenous exposure but is less vulnerable to confounding than observational studies. METHOD: We used separate sample instrumental variable analysis to estimate the association of log 17β estradiol with cognitive function (Delayed 10 word recall, and Mini Mental State Examination (MMSE)) and depressive symptoms (Geriatric Depression Scale (GDS)) in older Chinese women of the Guangzhou Biobank Cohort Study (GBCS, n=3086). The estimate was derived based on the Wald estimator, the ratio of the association of genetic determinants (rs1008805 and rs2175898) of log 17β-estradiol with cognitive function and depressive symptoms in GBCS and the association of log 17β-estradiol with genetic determinants in the sample of young women in Hong Kong (n=236). RESULTS: Genetically predicted 17β-estradiol was not associated with delayed 10-word recall (0.42 words per log increase in 17β-estradiol (pmol/L), 95% confidence interval (CI) -0.49 to 1.34) MMSE (0.39 per log increase in 17β-estradiol (pmol/L), 95% CI -0.87 to 1.65) or GDS (0.24 per log increase in 17β-estradiol (pmol/L), 95% CI -0.57 to 1.05). CONCLUSION: These results were largely consistent with evidence from RCTs and did not show any beneficial effect of estrogen on cognitive function and depressive symptoms. However, larger Mendelian randomization analyses are needed to identify any minor effects.