Creatine (methyl-d3) dilution in urine for estimation of total body skeletal muscle mass: accuracy and variability vs. MRI and DXA

A noninvasive method to estimate muscle mass based on creatine (methyl-d(3)) (D-3-creatine) dilution using fasting morning urine was evaluated for accuracy and variability over a 3-to 4-mo period. Healthy older (67-to 80-yr-old) subjects (n = 14) with muscle wasting secondary to aging and four patients with chronic disease (58-76 yr old) fasted overnight and then received an oral 30-mg dose of D-3-creatine at 8 AM (day 1). Urine was collected during 4 h of continued fasting and then at consecutive 4-to 8-h intervals through day 5. Assessment was repeated 3-4 mo later in 13 healthy subjects and 1 patient with congestive heart failure. Deuterated and unlabeled creatine and creatinine were measured using liquid chromatographytandem mass spectrometry. Total body creatine pool size and muscle mass were calculated from D-3-creatinine enrichment in urine. Muscle mass was also measured by whole body MRI and 24-h urine creatinine, and lean body mass (LBM) was measured by dual-energy X-ray absorptiometry (DXA). D-3-creatinine urinary enrichment from day 5 provided muscle mass estimates that correlated with MRI for all subjects (r = 0.88, P < 0.0001), with less bias [difference from MRI = -3.00 +/- 2.75 (SD) kg] than total LBM assessment by DXA, which overestimated muscle mass vs. MRI (-22.5 +/- 3.7 kg). However, intraindividual variability was high with the D-3-creatine dilution method, with intrasubject SD for estimated muscle mass of 2.5 kg vs. MRI (0.5 kg) and DXA (0.8 kg). This study supports further clinical validation of the D-3-creatine method for estimating muscle mass. NEW & NOTEWORTHY Measurement of creatine (methyl-d(3)) (D-3-creatine) and D-3-creatinine excretion in fasted morning urine samples may be a simple, less costly alternative to MRI or dualenergy X-ray absorptiometry (DXA) to calculate total body muscle mass. The D-3-creatine enrichment method provides estimates of muscle mass that correlate well with MRI, and with less bias than DXA. However, intraindividual variability is high with the D-3-creatine method. Studies to refine the spot urine sample method for estimation of muscle mass may be warranted.