The study subjects were selected from patients who were diagnosed with chronic kidney disease (CKD) by blood and/or urine examination, from June 2016 to June 2018, at the Faculty of Medicine, University of Yamanashi and the National Hospital Organization Kofu Hospital. For diagnosis, CKD was defined as persistent proteinuria or a decrease in the renal function.
In this study, the decreased muscle mass group comprised of patients with severe motor impairment and the inability of walking on their own. Nine patients with decreased muscle mass aged 6–64 years were included in this study. Of these patients, four were younger than 18 years of age. Patients with the following conditions were excluded: dehydration, fluid overload, estimated glomerular filtration rate (eGFR) calculated with serum cystatin C (eGFRCysC) > 180 mL/min/1.73 m2, 24H creatinine clearance (24HCCr) > 216 mL/min/1.73 m2, or eGFRCysC to 24HCCr ratio > 1.5 were excluded. For confirmation whether the e24HUCr formula from this study was applicable to normal muscle mass patients, 14 patients with normal muscle mass, aged 4–16 years were also investigated. The normal muscle mass group included patients who had a normal motor function.
Measurement of serum cystatin C and urine creatinine and calculations of eGFR and 24HCCr
Serum samples were obtained during 24H urine collection, and serum creatinine (sCr) and serum cystatin C (sCysC) were measured. The sCr level was determined by enzymatic methods, and sCysC was measured at the SRL Inc., Tokyo, Japan. Given that the use of eGFR with sCr was not suitable for patients with decreased muscle mass, eGFR values in this population were calculated using sCysC-based formulas shown in the upper row of Table 1 . In patients with normal muscle mass, eGFR was calculated using sCr-based formulas shown in the lower part of Table 1 [8, 9].
24H urine samples were collected to obtain total volume, urine creatinine (UCr) concentration, and UP concentration. The spot urine sample was obtained from the first voiding urine after 24H urine collection, and spot UCr (sUCr) and spot UP (sUP) were then measured. In patients who cannot urinate independently, 24H urine specimen was collected using urethral catheterization.
The relationship between eGFRCysC and 24HCCr was plotted by scatter gram, and correlation coefficient (R) was obtained. Because eGFRCysC and 24HCCr have the following relationship, eGFR = R × 24HCCr, we defined this formula as follows: eGFR = R × [(UCr × 24H urine volume × 1.73)] / [(sCr × 1440 × body surface (m2)]. UCr multiplied by 24H urine volume is equal to 24HUCr. Then, we redefined this formula as follows:
24HUCr (mg) = [eGFR × sCr × 1440 × body surface (m2)] / (1.73 × 100 × R)
Correlation of 24HUCr obtained from 24H urine collection and e24HUCr obtained from various formulas.
We evaluated 24HUCr obtained from 24H urine collection (actual 24HUCr) and estimated 24HUCr obtained from various formulas (e24HUCr) in both patients with decreased muscle mass and normal muscle mass. The formulas used for e24HUCr are shown below:
・Cockcroft-Gault formula: (24HUcr (mg)) = [28-(0.2 × age)] × weight (kg) × (0.85 if female) 
・Ghazali-Barratt formula: (24HUcr (mg)) = [0.46(age) + 15.4] × weight (kg) 
・New formula: (24HUCr(mg)) = [eGFR × sCr × 1440 × body surface (m2) ] / (1.73 × 100 × R)
Linear regression analysis was performed to evaluate relations between e24HUCr derived from the new formula and the actual 24HUCr. The Bland-Altman plot was used to determine the agreement between the actual 24HUCr and e24HUCr obtained from several estimation formulas.
Correlation of 24HUP and sUP/Cr ratio modified with e24HUCr.
The sUP/Cr ratio was corrected with e24HUCr derived from the Cockcroft-Gault formula, the Ghazali-Barratt formula, and the new formula developed in this study. We compared the sUP/Cr ratio modified with e24HUCr (modified 24HUP) and 24HUP with 24H urine collection (actual 24HUP). The Bland-Altman plot was used to determine the agreement between the modified 24HUP and the actual 24HUP.
All analysis was conducted using Microsoft Excel 2010 and the JMP8 tactical software (SAS Institute Inc., Cary, NC, USA). P < 0.05 was considered statically significant.