DOI: 10.1002/alz.074418 ISSN: 1552-5260

Causal effects of Alzheimer’s Disease on snoring may be driven by Body Mass Index: A bidirectional Mendelian Randomization Study

Yaqing Gao, Shea J Andrews, Kristine Yaffe, Yue Leng
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology

Abstract

Background

Loud snoring is a symptom commonly associated with obesity and obstructive sleep apnea (OSA), which have both been suggested as risk factors for Alzheimer’s disease (AD). However, it remains unclear whether snoring might be causally linked to AD and if body mass index (BMI) may play a role in this relationship. Here we use bidirectional Mendelian randomization (MR) analysis to investigate the causal relationship between snoring and AD.

Method

We performed two sample MR using summary statistics for genome‐wide association studies of AD (n = 94,437; cases = 35,274) and snoring (n∼408,000; snorers∼152,000). Genetic instruments were obtained by selecting independent genome‐wide significant SNPs (p<5e‐8, r2 = 0.001, window = 10Mb) for each exposure and harmonizing their effects with each outcome. We used fixed‐effects inverse‐variance weighted (IVW) meta‐analysis as the primary method, and MR‐Egger, Weighted mode, Weighted median, radial‐MR, estimators as sensitivity analyses. LHC‐MR was used to investigate bidirectional effects accounting for potential heritable confounders. Heterogeneity was assessed using Cohcrans Q test, and pleiotropy using the MR‐Egger intercept. We used multivariable MR to estimate direct effects of AD on snoring independent from BMI.

Result

No causal effect of snoring (OR [95%CI] = 0.842 [0.391, 1.798], p = 0.653) on AD was observed. In the reverse direction, we did not observe significant effects of AD on snoring using IVW (OR [95%CI] = 0.991 [0.979, 1.002], p = 0.114) or LHC‐MR (OR [95%CI] = 1.022 [0.805, 1.298], p = 0.850). However, suggestive causal effects were observed in the MR‐Egger (OR [95%CI] = 0.974 [0.958, 0.990], p = 0.004), Weighted mode (OR [95%CI] = 0.981 [0.965, 0.997], p = 0.031), and Weighted Median (OR [95%CI] = 0.979 [0.963, 0.996], p = 0.013) analyses. There was no evidence of heterogeneity, but some evidence of pleiotropic effects (p = 0.009). Multivariable MR indicated that the effect of AD on snoring was likely driven by BMI.

Conclusion

Genetic liability to AD reduced snoring, possibly through lower BMI due to AD. Further analyses are needed to determine if OSA is causally related to AD and the role of BMI in this relationship.

More from our Archive