Abstract:

Objective To evaluate the potential causal relationship between tinnitus and the risk of Alzheimer’s disease （AD） onset using the two-sample Mendelian randomization （MR） method and to clarify its mechanism of action， so as to provide new ideas for early warning of AD. Methods Genome-wide association study （GWAS） database was used to search the keywords “tinnitus” and “Alzheimer” to obtain the related datasets of exposure factor tinnitus and outcome AD； the tinnitus datasets included ukb-d-4803_11， ukb-d-4803_12， ukb-d-4803_13， ukb-b-14254 and ukb-a-384； the AD datasets included ieu-b-5067， ieu-b-2， ieu-a-297， ebi-a-GCST90027158 and ebi-a-GCST002245. The single nucleotide polymorphisms （SNPs） closely and independently associated with tinnitus were screened as instrumental variables （IVs）， and the SNPs associated with AD were used as outcomes. Inverse variance weighted （IVW） method was used to conduct MR analysis to evaluate its odds ratio （OR） value， 95% confidence interval （CI） and P value； P<0.05 indicated significant causal relationship. Sensitivity detection used Cochran’s Q test to detect the heterogeneity of IVs to evaluate its Q value， df value and P value； when IVW method P>0.05， it indicated no significant heterogeneity； MR-Egger intercept was used to detect horizontal pleiotropy； when the intercept was 0 or close to 0 and P>0.05， it indicated no significant horizontal pleiotropy； meanwhile， leave-one-out method was used for sensitivity analysis. Finally， visualization results were performed using forest plot， scatter plot， funnel plot and leave-one-out plot. Results A total of 286 SNPs were screened as IVs. All instrumental variables satisfied F>10， suggesting no weak instrumental variable； after screening by PhenoScanner web tool， all SNPs were unrelated to confounding factors. When the tinnitus and AD datasets were ukb-d-4803 and ebi-a-GCST90027158 respectively， there was a significant positive correlation between tinnitus and the risk of AD onset （IVW： OR=1.842， 95%CI：1.065-3.188， P=0.029）； Cochran’s Q test suggested no significant heterogeneity of IVs （Q=9.788， df=10.000， P=0.459）； MR-Egger intercept indicated no horizontal pleiotropy （Egger intercept=-0.006， P=0.147）； leave-one-out method showed stable results， and no SNP with significant influence on the results was detected. Conclusion There is a positive causal relationship between tinnitus and the risk of AD onset. Neuroinflammation accompanied by persistent microglial activation to varying degrees may be the common pathogenesis of tinnitus and AD； in addition， depression may also act as an upstream factor to hyperactivate the hypothalamic-pituitary-adrenal （HPA） axis， leading to the progression of relationship between tinnitus and AD.

Key words: Tinnitus, Alzheimer disease, Mendelian randomization, Genome-wide association study, Inverse variance weighted method