From the bottom of an old jar: A fluorometric method for the determination of creatinine in human serum.

In this paper we have investigated and optimized a non-enzymatic fluorometric creatinine assay. The method was originally described by Blass in the 90s, but we found that besides the reagents mentioned in the paper, an addition of hydrogen peroxide is required to obtain a fluorescent compound. The excitation and emission maxima of the fluorophore are 405 nm and 475 nm, respectively. The optimal conditions for creatinine quantification are as follows: 25 mmol L 3,5-dinitrobenzoic acid dissolved in 1,4-butanediol with 58 mmol L HO and aqueous solution of 2 mol L NaOH mixed in 1:1 ratio. A linear calibration curve (y = 18,7x + 446 for 300 s of incubation) was obtained in the range from 2.6 to 750 μmol L of creatinine with LOD and LOQ equal to 0.7 and 2.6 μmol L, respectively. The method was found to be selective towards the analyte in the presence of compounds such as urea, uric acid, bilirubin, albumin and glucose. The developed protocol was applied for creatinine determination in 13 real human serum samples. A correlation (y = (0.94 ± 0.03) x + (3.66 ± 3.22) with Pearson's r 0.996) and statistical agreement (two-tailed Student's t-test at 95% confidence interval with 12 degrees of freedom) was reached between the obtained results and the reference enzymatic method.