Skip to main content
. 2019 Oct 29;20:397. doi: 10.1186/s12882-019-1584-7

Table 2.

Regression models for aGFR, 1/SCR and aGFR-eGFR

Variable Β SE P-value
A Log aGFR (log ml/min/1.73m2)
 Intercept 5.8502 0.0430
 Sex (female vs male) −0.2151 0.0128 < 0.0001
 Age (yrs) −0.0100 0.0005 < 0.0001
 Serum creatinine (mg/dl) −0.8639 0.0334 < 0.0001
 Fasting plasma glucose (10 mg/dl) 0.0061 0.0008 < 0.0001
 HbA1c (%) 0.0061 0.0030 0.0402
B 1/SCR (dl/mg)
 Intercept 1.4931 0.0566
 Sex (female vs male) 0.2821 0.0218 < 0.0001
 Age (yrs) −0.0082 0.0009 < 0.0001
 HbA1c (%) 0.0287 0.0033 < 0.0001
 Fasting plasma glucose (10 mg/dl) −0.0043 0.0007 < 0.0001
C aGFR-eGFR (ml/min/1.73m2)
 Intercept 39.7124 5.3606
 Sex (female vs male) −10.5330 1.5884 < 0.0001
 Age (yrs) −0.5928 0.0708 < 0.0001
 Serum creatinine (mg/dl) −11.2485 4.1942 0.0074
 Fasting plasma glucose (10 mg/dl) 0.7927 0.0976 < 0.0001
 HbA1c (%) 1.0216 0.3789 0.0071
 Duration of diabetes 0.1916 0.08687 0.0275

Beta is the regression coefficient and SE is its standard error. Regression models were fit using a mixed model procedure to account for multiple examinations within individuals. Results are shown for best fitting model (lowest AIC)