Microalbuminuria is considered a marker of diabetic nephropathy. Early detection of microalbuminuria allows for early intervention with the goal of delaying the onset of overt diabetic nephropathy.
The aim of this study was to compare the analysis of urinary albumin and creatinine by both Microalbustix and Clinitek microalbumin urinary dipsticks compared with high-performance liquid chromatography (HPLC) analysis.
Diabetic patients attending the Endocrine Clinic at Austin & Repatriation Medical Center, Heidelberg, Australia, were studied (56% normoalbuminuric, 34% microalbuminuric, and 10% macroalbuminuric by immunoturbidimetry). Albumin and creatinine concentrations were determined in 24-h urine samples by Microalbustix and Clinitek microalbumin dipsticks using a Clinitek 50 Autoanalyzer (Bayer, Elkhart, IN) and by HPLC, as previously described (1). Urinary creatinine concentration was also determined using a Hitachi 971 Autoanalyzer.
Using an upper limit of normal of <30 mg/l for albumin concentration, 56 of 115 and 48 of 98 urines were normal by Microalbustix and Clinitek microalbumin, respectively, and also by HPLC analysis. The Microalbustix and Clinitek microalbumin gave false-negative results (<30 mg/l) for 20 of 115 (17.4%) and 16 of 98 (16.3%) urines, respectively. There were 21.0% false-negative results for diabetic urines measured by immunoturbidimetry compared with HPLC analysis. In urine tested by HPLC from nondiabetic volunteers, 73 of 106, 32 of 106, and 1 of 106 had albumin concentrations of <10, 10–30, and 30–80 mg/l, respectively. There were 3.7% false negatives for albumin concentration in nondiabetic urines measured by immunoturbidimetry compared with HPLC analysis. These results demonstrate that HPLC analysis does not overestimate albumin concentrations since results similar to conventional immunoturbidimetry were obtained for nondiabetic urine. The differences obtained for diabetic urine are due to changes in albumin processing in the diabetic state that are undetectable by immunochemical assays (2).
Using an upper limit of normal for the albumin-to-creatinine ratio (ACR) of <30 mg/g, 34 of 115 and 29 of 98 urines were normal by Microalbustix and Clinitek microalbumin, respectively, and by HPLC analysis. The Microalbustix and Clinitek microalbumin gave false-negative results (<30 mg/g) for 42 (36.5%) and 42 (42.9%) urines, respectively, compared with HPLC analysis (>30 mg/g). There were 36.3% false-negative results for diabetic urines measured by immunoturbidimetry compared with HPLC analysis. For the 106 nondiabetic volunteers, 100 of 106 and 6 of 106 were found to have ACRs <30 mg/g and 30–300 mg/g when albumin concentration was measured by HPLC. There were 0% false negatives for ACR in nondiabetic urine measured by immunoturbidimetry compared with HPLC analysis.
Development of a quantitative test for urinary albumin is difficult. Intact albumin filtered by the kidney is biochemically modified, resulting in the excretion of a complex mixture of <1% intact albumin and >99% albumin-derived fragments <10kDa (2). Conventional immunoassays and dye binding methods cannot detect albumin-derived fragments (2,3), and a proportion of the intact albumin excreted has also been shown to be immunounreactive (1,4). Immunoassays measure active epitope in urine; however, this epitope may not be exclusively associated with intact albumin (1–4). HPLC analysis of urinary albumin is able to detect both intact immunoreactive and intact immunounreactive albumin (1) and is priced <$10 per urine sample. All other available methods for measuring urinary albumin cannot detect albumin fragments.
Microalbustix and Clinitek microalbumin dipsticks exhibit poor sensitivity in detecting early changes in the kidney that result in microalbuminuria. Both dipsticks were similar, which is not surprising. They are based on the same sulfonephthalein dye binding method that does not detect albumin-derived fragments (3) because an inhibitor is present in the dipstick formulation that inhibits urinary protein fragments binding to it (5). The American Diabetes Association recommends that semiquantitative or qualitative screening tests for microalbuminuria have a detection rate for abnormal samples of >95% for patients with microalbuminuria to be useful for screening (6). However, in 2002 the American Diabetes Association could find no published study that fulfilled these criteria for qualitative (or semiquantitative) dipstick tests (6).
A satisfactory screening test for microalbuminuria needs to be at least as sensitive as a laboratory method. Subsequent laboratory testing can eliminate false positives, but false negatives could result in the delay of beneficial early treatment that may stop or reverse the progression to overt diabetic nephropathy.
