Researchers are beginning to assess how genetic variation influences sex differences in Alzheimer’s disease. True, women who have an ApoE4 allele are more likely to develop Alzheimer’s than are male carriers, but what about the many other variants that have been implicated in this genetically complex disease? In the June 27 Brain, researchers led by Anders Dale at the University of California, San Diego, report that polygenic hazard scores for AD are sex-specific, i.e., different when calculated for men and women separately. Sex-matched scores better predicted disease onset, progression, and neuropathology than did scores calculated for all study participants together, as is usual in genetics studies in AD. “The finding suggests that sex differences in genetic risk go beyond ApoE4,” Chun Chieh Fan, co-first author, told Alzforum.

  • Sex differences in genetic risk for AD are uncertain.
  • Researchers calculated sex-specific polygenic hazard scores.
  • These better predict clinical diagnosis, pathology, and cognitive decline.

Fan, co-first author Sarah Banks, and colleagues capitalized on a polygenic hazard score developed by Dale and the late Rahul Desikan (Mar 2017 newsAug 2019 news). To identify sex differences in genetic risk, the researchers correlated genetic variants with AD in men and women separately by way of genome-wide association of 7,158 men and 10,697 women who had volunteered for the Alzheimer’s Disease Genetic Consortium. Using this sex-based GWAS data, the researchers calculated polygenic hazard scores (PHS) and polygenic risk score (PRS), again separately, for 2,628 men and 3,448 women in the National Alzheimer’s Coordinate Cohort (NACC). Then, in a crossover design, they calculated how polygenic scores for women predicted outcomes in women and in men, and ditto for the polygenic scores for men.

The crossover analysis was based on genetic, diagnostic, and pathology data from NACC and from ROSMAP, a combination of the Religious Orders Study and Memory and Aging Project both being conducted at Rush University, Chicago.

It turned out that the male and female polygenic scores had stronger predictive power when applied to their matched gender. One standard-deviation uptick in the sex-matched PHS increased the odds of a clinical AD diagnosis by 26 percent, whereas for male scores applied to female cohorts, or vice versa, the hazard ratio increase was only 14 percent.

The sex-matched PHS also better predicted neuropathologically confirmed AD and annual progression of dementia, as judged by change in CDR sum-of-boxes scores. In the ROSMAP data, the matched PHS associated much more strongly with plaque burden, neurofibrillary tangles, Braak staging, and CERAD neuropathology scores than did the mismatched PHS. 

All told, the study indicates that the genetic risk associated with AD is tied to a person’s sex. This seems to affect only progression, not prevalence, because male and female PRS scores equally foretold AD in men and women. While PRS predict whether a person will get AD, PHS predicts at what age he or she is likely to become symptomatic.

What variants are involved in this sex difference? Figuring this out will require more study and many more samples. Still, some hints emerged from the male/female GWAS. Effect sizes for BIN1, MS4A6A, DNAJA2, and FERMT2 were all higher in women, while FAM193B, C2ord47, and TYW5 variants appeared stronger in men. To get a handle on this, Fan believes it will be important to correlate sex-dependent cognitive scores and genetic data with biomarkers and gene-expression analysis. “Only then can we be confident about the underlying biology we predict from genetics,” he said.—Tom Fagan


  1. For anyone interested, along the lines of this excellent new study, we also recently published a paper using family-based association to identify novel AD genetic risk factors that are sex-specific, in this case actually conferring opposite risk effects in men and women.

    The top hit was ZBTB7C, which encodes the Zinc Finger and BTB Domain Containing 7C, a transcriptional repressor of membrane metalloproteases, implicated in AD neuropathology.


    . Identification of Novel Alzheimer's Disease Loci Using Sex-Specific Family-Based Association Analysis of Whole-Genome Sequence Data. Sci Rep. 2020 Mar 19;10(1):5029. PubMed.

  2. C.C. Fan et al. developed and examined a sex-stratified polygenic score for prediction of AD age of onset and other AD phenotypes, including AD pathological hallmarks. Their study expands recently published GWAS papers, exploring genes that were associated with AD phenotypes either in men or in women, not in both.

    Fan’s paper has two main findings. First, polygenic hazard score, developed through conduction of Cox hazard function, predicted AD phenotypes better than polygenic risk score, developed by conduction of logistic regression. Second, sex-matched polygenic scores, which are developed and applied in the same sex, outperformed sex-mismatched polygenic scores, which are developed in one sex and applied in another sex, in prediction of AD.

    Currently, polygenic scores are gaining widespread attention and use in the field of medicine as most common human diseases have genetic susceptibility originating from cumulative effects of thousands of SNPs, each having a very small effect on the disease of interest. In the field of AD, several groups have developed an AD polygenic score. Fan’s paper suggests that AD polygenic risk scores should be developed separately in men and women, the same path that other investigators followed previously, including those who developed the Framingham Cardiovascular risk scores for prediction of vascular events.

  3. The findings of this paper are interesting and add some further food for thought in the search for an explanation as to why we observe differences between men and women in the onset and progression of AD. It is interesting to note that there was no difference in the prediction accuracy of the polygenic risk score (PRS, a measure of general risk for AD diagnosis at any time within a person's lifetime, or at death) but the more “temporal” polygenic hazard score (PHS) showed sex-specific prediction accuracy. These findings, in combination with the AD endophenotype studies, demonstrate why it is so important that researchers move toward the biologically based ATN research framework proposed by Cliff Jack and colleagues (Jack et al., 2018). Further research is necessary to determine why there is this growing evidence for sex differences in AD risk and progression. In terms of purely strict genetic inheritance, the sex chromosomes are the only difference between biological males and females, which means that what we are seeing in these sex-specific genetic associations are reflections of biological, environmental, social, and/or even cohort differences between men and women.

    As a side note, I like that they use the E2 and E4 dosages for their APOE genotype correction instead of the simple E4-carrier status that many groups use, as this takes into account the nonlinear relative risk of the APOE genotype.


    . NIA-AA Research Framework: Toward a biological definition of Alzheimer's disease. Alzheimers Dement. 2018 Apr;14(4):535-562. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Genetic Risk Score Combines AD GWAS Hits, Predicts Onset
  2. Colleagues Mourn Loss of Rahul Desikan, Age 41, to ALS

Further Reading


  1. . Sex-specific genetic predictors of Alzheimer's disease biomarkers. Acta Neuropathol. 2018 Dec;136(6):857-872. Epub 2018 Jul 2 PubMed.

Primary Papers

  1. . Sex-dependent autosomal effects on clinical progression of Alzheimer's disease. Brain. 2020 Jul 1;143(7):2272-2280. PubMed.