Authors' response to: A Mendelian randomization study of Alzheimer's disease and daytime napping

Dear Editor, We thank the interest from Ran et al. and their thoughtful letter in response to our research article.1 Ran et al. conducted a Mendelian randomization (MR) analysis to examine the bidirectional relationship between daytime napping and Alzheimer's disease (AD). The Authors claim that based on their bidirectional MR findings, daytime napping was causally associated with AD, partially consistent with our observations. Given the absence of important details in their MR analysis, extreme caution should be exercised when interpreting the provided results. For example, Ran et al. listed three publicly available genome-wide association studies (GWASs) retrieved from the IEU OpenGWAS project for daytime napping and excessive daytime sleepiness. It is unclear how a genetic instrument for daytime napping was constructed by aggregating the three GWASs assessing two different napping phenotypes. The source of the 33 single-nucleotide polymorphisms (SNPs) and the corresponding effect estimates included are unclear, and the appropriateness of including genetic variants related to daytime sleepiness/narcolepsy in the genetic instrument is questionable. Furthermore, it lacks information on how much variance of the daytime napping phenotype is explained by these SNPs or any metrics reflecting the robustness of the genetic instrument (e.g., F-statistic). It is unclear why the Authors did not leverage the largest GWAS study for daytime napping which identified 123 genetic variants for the trait.2 The same concern also holds for the GWAS of AD. The reverse MR in the study of Ran et al. showed “genetic predisposition to AD could lower the risk of excessive daytime napping,” which is surprising, implying that AD leads to fewer daytime naps. Thus, their conclusion of “a bidirectional causal relationship between excessive daytime napping and Alzheimer's dementia [disease]” seems inappropriate and misleading. In addition, the Authors did not provide results from their mentioned sensitivity analysis (based on MR Egger and weighted media estimator), and therefore, it is not possible to evaluate whether the results are robust or prone to pleiotropy. Results from a 2021 study suggested potential protective effects of daytime napping on AD using MR and leveraging robust datasets.3 Daytime napping was also shown to be causally associated with larger total brain volume in a more recent MR study.4 While these results seemingly conflict with our observational findings, the genetic instrument used in these two studies only focused on daytime napping frequency. In the letter by Ran et al., the Authors referred to the phenotype as “excessive daytime napping,” but the GWASs also analyzed the frequency of daytime napping. Therefore, direct inference with regard to “excessive daytime napping” is not appropriate, and the interpretation needs to be cautious. Ran et al. also highlighted the conflicting findings from another recent observational study by Wong et al.,5 in which daytime napping was shown not to be associated with the risk for dementia. It is worth noting that the study used the Million Women Study cohort, in which only women were enrolled. In our study,1 no sex differences were observed, but women dominated our cohort (ie, the proportion of women was > 75%). Therefore, besides sex differences, other factors should contribute to the discrepancies. For instance, a crucial factor can be age.6 Daily behaviors including napping can change over time, as we demonstrated in our study.1 The cohort of the Million Women Study has a mean age of 60 years, which is significantly younger than participants in our study, with a mean age of over 80 years. The other important factor is the approach to assessing daytime napping. Subjective self-reported daytime napping behavior was used in Wong et al. study, whereas, in our study, daytime napping was objectively assessed based on continuously monitored actigraphy data. Recall bias, especially in older adults, may limit self-report accuracy. Additional instruments that capture other dimensions of daytime napping, such as the timing and regularity of scheduled versus impromptu naps,7 are needed. In a previous study using cluster analysis, three subclusters of genetic variants were identified, which suggests that daytime napping may have different effects on various subgroups of people.2 Future analyses should consider subgroups/clusters of daytime napping variants and their associations with AD. Further work is also needed to understand the clinical implications of different characteristics of daytime napping on cognition and AD risk. Cultural differences in daytime napping across different ancestry backgrounds and geographic locations may play a significant role in the relationship between daytime napping and AD. Finally, we encourage mechanistic studies to enhance the understanding of the causal links between daytime napping and AD. The viewpoint in this response was conceived by all authors. P.L. drafted the manuscript. All authors edited the manuscript. This research did not receive any specific grant from federal funding agencies or commercial or not-for-profit sectors. P.L. is funded by the BrightFocus Foundation under grant no. A2020886S and the Fund to Sustain Research Excellence from the Brigham Research Institute. L.G. is funded by the Alzheimer's Association under grant no. AACSF-23-1148490. H.S.D is funded by the National Institutes of Health (NIH) grant number R00HL153795. K.H. is funded by NIH RF1AG064312. Y.L. is funded by NIH R00AG056598. The funders played no role in the writing of this article. The authors declare no conflicts of interest. Author disclosures are available in the supporting information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.