BACKGROUND: Asymmetric dimethylarginine (ADMA) is increasingly being investigated as a renal and cardiovascular risk factor in chronic kidney disease (CKD). We assessed the degree of agreement of an enzyme-linked immunosorbent assay (ELISA) of ADMA and the gold standard liquid chromatography-electrospray tandem mass spectrometry (LC-MS/MS). METHODS: ADMA was measured in an incident cohort of 126 stable CKD patients. Correlations between methods were studied by estimating the interclass Pearson coefficient (IPC), Lin's concordance correlation coefficient (CCC) and the maximum likelihood intraclass correlation coefficient (ICC). Limits of agreement (LOA) were estimated using the Bland-Altman method. RESULTS: ADMA values were normally distributed with means of 0.78 ± 0.16 (ELISA) and 0.59 ± 0.09 mol/L (LC-MS/MS). IPC was 0.69 (95% confidence interval [95% CI], 0.59-0.78). Overall CCC was 0.29 (95% CI, 0.23-0.37), with a difference in means of 0.19 mol/L (95% LOA, -0.043 to 0.43), delta slope of 0.577 and delta intercept of 0.14 (vs. perfect agreement line). Data were similar across categories of clinical characteristics. Mixed models provided an ICC of 0.58 (95% CI, 0.46-0.69). Bias was larger among patients with glomerular filtration rate (GFR) <30 ml/min. When values obtained from ELISA were corrected using the slope and intercept estimates from concordance analyses, the adjusted ICC improved to 0.67 (95% CI, 0.57-0.76), and bias modification by GFR levels canceled out. CONCLUSIONS: In CKD, ELISA overestimates ADMA concentration as compared with LC-MS/MS. Appropriate calibration is needed when ADMA is measured by ELISA in CKD patients.
BACKGROUND: Asymmetric dimethylarginine (ADMA) is increasingly being investigated as a renal and cardiovascular risk factor in chronic kidney disease (CKD). We assessed the degree of agreement of an enzyme-linked immunosorbent assay (ELISA) of ADMA and the gold standard liquid chromatography-electrospray tandem mass spectrometry (LC-MS/MS). METHODS:ADMA was measured in an incident cohort of 126 stable CKDpatients. Correlations between methods were studied by estimating the interclass Pearson coefficient (IPC), Lin's concordance correlation coefficient (CCC) and the maximum likelihood intraclass correlation coefficient (ICC). Limits of agreement (LOA) were estimated using the Bland-Altman method. RESULTS:ADMA values were normally distributed with means of 0.78 ± 0.16 (ELISA) and 0.59 ± 0.09 mol/L (LC-MS/MS). IPC was 0.69 (95% confidence interval [95% CI], 0.59-0.78). Overall CCC was 0.29 (95% CI, 0.23-0.37), with a difference in means of 0.19 mol/L (95% LOA, -0.043 to 0.43), delta slope of 0.577 and delta intercept of 0.14 (vs. perfect agreement line). Data were similar across categories of clinical characteristics. Mixed models provided an ICC of 0.58 (95% CI, 0.46-0.69). Bias was larger among patients with glomerular filtration rate (GFR) <30 ml/min. When values obtained from ELISA were corrected using the slope and intercept estimates from concordance analyses, the adjusted ICC improved to 0.67 (95% CI, 0.57-0.76), and bias modification by GFR levels canceled out. CONCLUSIONS: In CKD, ELISA overestimates ADMA concentration as compared with LC-MS/MS. Appropriate calibration is needed when ADMA is measured by ELISA in CKDpatients.