Table 3 Findings of two-sample MR investigating effects of genetically predicted GrimAA on cognitive function based on primary method and pleiotropy-robust methods

Outcome	Exposure genetic instruments	Primary analysis				Sensitivity analysis
Outcome	Exposure genetic instruments	Egger intercept P	Method	Effect^a (95% CI)	P^b	Method	Effect^a (95% CI)	P
Trail making tests	GrimAA 4 SNPs	0.34	IVW	0.024 (0 to 0.049)	0.05	Weighted median	0.015 (−0.015 to 0.045)	0.32
Parts A (N = 99,479)	GrimAA 4 SNPs	0.34	IVW	0.024 (0 to 0.049)	0.05	Weighted mode	0.014 (−0.031 to 0.059)	0.59
Trail making tests	GrimAA 4 SNPs	0.54	IVW	0.022 (−0.015 to 0.059)	0.25	Weighted median	0.011 (−0.021 to 0.043)	0.49
Parts B (N = 99,477)	GrimAA 4 SNPs	0.54	IVW	0.022 (−0.015 to 0.059)	0.25	Weighted mode	0.007 (−0.043 to 0.058)	0.79
Numeric memory (N = 106,162)	GrimAA 4 SNPs	0.55	IVW	0.97 (0.94 to 1.01)	0.14	Weighted median	0.98 (0.94 to 1.02)	0.31
Numeric memory (N = 106,162)	GrimAA 4 SNPs	0.55	IVW	0.97 (0.94 to 1.01)	0.14	Weighted mode	0.98 (0.93 to 1.04)	0.61

CI confidence interval, IVW inverse variance-weighted, SNP single nucleotide polymorphism.
^aEffect is per 1-year increase of EAA. Effects for trail making tests are β coefficients based on the GWAS using inverse rank normal transformed duration to complete path. Effect for numeric memory is odds ratio based on the GWAS using maximum digits remembered correctly that is modeled as ordered categorical variable.
^bAll p-values reported are unadjusted for multiple comparisons.

Quick links

Search