BACKGROUND: Urinary albumin excretion is a sensitive diagnostic and prognostic marker for renal disease. Therefore, measurement of urinary albumin must be accurate and precise. We have developed a method to quantify intact urinary albumin with a low limit of quantification (LOQ). METHODS: We constructed an external calibration curve using purified human serum albumin (HSA) added to a charcoal-stripped urine matrix. We then added an internal standard, 15N-labeled recombinant HSA (15NrHSA), to the calibrators, controls, and patient urine samples. The samples were reduced, alkylated, and digested with trypsin. The concentration of albumin in each sample was determined by liquid chromatography-tandem mass spectrometry (LC-MS/MS) and linear regression analysis, in which the relative abundance area ratio of the tryptic peptides 42LVNEVTEFAK51 and 526QTALVELVK534 from albumin and 15NrHSA were referenced to the calibration curve. RESULTS: The lower limit of quantification was 3.13 mg/L, and the linear dynamic range was 3.13-200 mg/L. Replicate digests from low, medium, and high controls (n = 5) gave intraassay imprecision CVs of 2.8%-11.0% for the peptide 42LVNEVTEFAK51, and 1.9%-12.3% for the 526QTALVELVK534 peptide. Interassay imprecision of the controls for a period of 10 consecutive days (n = 10) yielded CVs of 1.5%-14.8% for the 42LVNEVTEFAK51 peptide, and 6.4%-14.1% for the 526QTALVELVK534 peptide. For the 42LVNEVTEFAK51 peptide, a method comparison between LC-MS/MS and an immunoturbidometric method for 138 patient samples gave an R2 value of 0.97 and a regression line of γ = 0.99x + 23.16. CONCLUSIONS: Urinary albumin can be quantified by a protein cleavage LC-MS/MS method using a 15NrHSA internal standard. This method provides improved analytical performance in the clinically relevant range compared to a commercially available immunoturbidometric assay.