Table 2.
Associations of the two candidate metabolites with eGFR in hyperglycemic individuals. Regression coefficients with 95% CI and p-values of weighted multivariable linear regression after inverse probability weighting are shown. The basic model was adjusted for age, sex, BMI, systolic blood pressure, triglyceride, total cholesterol, HDL cholesterol, and HbA1C. The full model was additionally adjusted for smoking status, use of lipid-lowering drugs, and antihypertensive medication, and urinary albumin-to-creatinine ratio. p-values shown in bold represent statistical significance at the 0.05 level. Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; SM, sphingomyelin; PC aa, phosphatidylcholine diacyl.
| Models | SM C18:1 | PC aa C38:0 | ||
|---|---|---|---|---|
| Effect Estimate (95% CI) | p-Value | Effect Estimate (95% CI) | p-Value | |
| Adjusted imbalanced covariates | −1.51 (−2.92 to −0.1) a | 3.624 × 10−2 | −1.82 (−3.04 to −0.59) b | 3.757 × 10−3 |
| Basic model | −1.83 (−2.98 to −0.68) | 1.879 × 10−3 | −1.91 (−3.11 to −0.72) | 1.784 × 10−3 |
| Full model | −1.76 (−2.9 to −0.62) | 2.499 × 10−3 | −1.81 (−2.99 to −0.63) | 2.607 × 10 −3 |
a with adjustments for sex, systolic blood pressure, total cholesterol, smoking status, and use of antihypertensive medication; b with adjustments for age, HDL cholesterol, and smoking status.