Nonlinear causal effects of estimated glomerular filtration rate on myocardial infarction risks: Mendelian randomization study

Table 2 Meta-regression results of the causal estimates from nonlinear MR analysis by fractional polynomial method

Genetically predicted exposure	Adjusted covariates	Quadratic P value	β	Fractional polynomial model power	Estimated beta	Estimated standard error	Estimated P value
Creatinine-based eGFR	Age, sex, and 10 PCs	< 0.001	β1	1	− 5.36E−2	1.61E−3	< 0.001
	Age, sex, and 10 PCs	< 0.001	β2	3	2.31E−6	6.53E−7	< 0.001
	Age, sex, 10 PCs, clinical covariates (e.g., BMI, hypertension, diabetes, dyslipidemia, and albuminuria)	0.02	β1	0.5	− 8.87	3.57	0.013
		0.02	β2	log 0.5	1.38	0.55	0.013
Cystatin C-based eGFR	Age, sex, and 10 PCs	0.01	β1	2	− 1.48E−3	5.51E−4	0.007
	Age, sex, and 10 PCs	0.01	β2	log 2	2.96E−4	1.11E−4	0.008
	Age, sex, 10 PCs, clinical covariates (e.g., BMI, hypertension, diabetes, dyslipidemia, and albuminuria)	0.02	β1	0	− 1.44	0.67	0.03
		0.02	β2	3	8.85E−7	3.60E−7	0.01

Clinical covariates included in the adjusted model were body mass index, systolic blood pressure, hypertension medication history, diabetes mellitus diagnosis, hemoglobin A1c, medication history for dyslipidemia, triglycerides, high-density lipoprotein and low-density lipoprotein cholesterols, and urine microalbumin levels
eGFR estimated glomerular filtration rate, PC principal components, BMI body mass index

ISSN: 1741-7015