Additional key phrases: creatinine clearance; low molecular weight plasma proteins; proteinuria; tubular proteinuriaIn renal failure, the concentrations of low molecular weight (LMW) plasma proteins (MWt < 50 OOO) may be increased in the serum due to reduced filtration, and increased in the urine because of impaired tubular uptake., Because of the sensitivity of the changes in their concentrations in both serum and urine, LMW plasma proteins, particularly µglobulin (&M), have been used for the assessment of glomerular and tubular function. Following doubts about the reliability and stability of &-M measurements in the urine and the increase in serum concentration of &-M in several other clinical situations independent of renal function, retinol binding protein (RBP; MWt 21 OOO)z and a-1 microglobulin (alM; MWt 33OOO),' which are relatively more stable in body fluids, have been suggested as the preferred alternative indicators. We measured RBP and a,M levels in serum and urine in patients with varying degrees of renal failure in an attempt to assess their relative sensitivity in monitoring renal function.
PATIENTS AND METHODSThirty healthy adults (16 men, 14 women), with normal renal function and urinary protein < 150 mg/day, and 72 adult patients (40 men and 32 women) with impaired renal function of different aetiologies (20 were haemodialysis dependent, 52 suffered from chronic renal failure) were studied. The haemodialysis patients were dialysed on cellosic dialysers which were impermeable to both RBP and a,M.Correspondence: Dr J T C Kwan, Department of Nephrology, The Royal London Hospital, Whitechapel, London El IBB, UK.Twenty-four hour urine samples together with matching serum samples were collected, except in the haemodialysis patients from whom only pre-dialysis serum samples were collected. Creatinine was determined by the standard Jaffk Rate reaction on a Cobas Bio Analyser (Roche Diagnostics, Welwyn Garden City, UK) and total urinary protein by a dye-binding m e t h~d .~ RBP in serum and urine was measured by an enzymelinked immunosorbent (within-run and betweenrun CVs of 6.2% and 9 . 5 % , respe~tively).~ Serum and urinary a l M was estimated by the method of Mancini (within-run and betweenrun CVs of 3.2% and 4.6%. respectively).6 Rabbit anti-human RBP (code A M , lot 082) and rabbit anti-human a,M (code A256, lot 013B) were obtained from Dako Ltd (High Wycombe, UK).Data were compared using Student's t-test and correlated by linear regression analysis. Protein clearance was calculated in the same manner as that of creatinine clearance (Ccr). Fractional urinary clearance of protein was defined as the percentage ratio of protein clearance to creatinine clearance.
RESULTSThe results, stratified into six bands according to creatinine clearances of our subjects, are shown in Table 1. There were significant (n = 102, P < 0.001) correlations between serum RBP and plasma creatinine (r=0-69), serum a , M and plasma creatinine (r = 0 -85) and between serum RBP and serum a l M (r=0.75). Excluding the 20 di...