| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
St. Bartholomew's and the Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, UK;
2
Bayer plc, Diagnostics Division, Bayer House, Strawberry Hill, Newbury, Berks RG14 1JA, UK;
3
SW Thames Institute for Renal Research, St. Helier NHS Trust, Wrythe Lane, Carshalton, Surrey SM5 1AA, UK;
a address
correspondence to this author at: Department of Clinical Biochemistry, St. Bartholomew's and the Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, UK
Microalbuminuria is a risk factor for the development of overt nephropathy in type 1 and type 2 diabetic patients (1)(2). Importantly, improvement of glycemic control and early intervention with antihypertensive drugs can retard the development of microalbuminuria and possibly its progression towards overt nephropathy (3)(4). Microalbuminuria is defined as an increased excretion of albumin above the reference range for healthy nondiabetic subjects, but which is undetectable by the Albustix dipstick test (5). It has also been defined as a urinary albumin excretion rate between 20 and 200 µg/min in an overnight or 24-h sample on at least two of three occasions within a period of 6 months (6). This is equivalent to 30–300 mg/24 h or 3–30 g/mol creatinine.
Although there are no clinical data to define the analytical requirements for the detection and monitoring of microalbuminuria, it has been suggested that a method should have an interassay precision of <12% over the concentration range 5–200 mg/L and be sensitive enough to reliably detect changes of 10 mg/L in the concentration range 5–35 mg/L (5). The method must also be reproducible over long periods because patients may be screened only two or three times throughout the year.
Albumin may be measured in urine by radial immunodiffusion methods, which include a rapid, gold sol labeled variant of the technique capable of semiquantitative measurement (7), and by direct immunoturbidimetric and nephelometric methods that enable automation and rapid turnaround of results. The latter methods are prone to a "hook effect" at high antigen concentrations (8). Latex-enhanced methods, in addition to providing improved sensitivity, also allow the use of an inhibition format, thereby avoiding a hook effect (9).
The DCA 2000TM desktop microalbumin system (Bayer plc, Newbury, UK) is a disposable cartridge device encapsulating an immunoturbidimetric assay for albumin and a colorimetric assay for creatinine, together with a programmable photometer with an incubation chamber. The cassette includes separate reservoirs that contain buffer and antibody reagent and reagents for a creatinine assay. Each batch of reagents is calibrated, and the algorithm is encoded in a bar code strip on the cartridge. The analyzer prints out the albumin:creatinine ratio as well as the concentrations of albumin and creatinine. The analyzer can also be used for the quantitation of HbAIC (10). We have evaluated the performance of this system and compared it with a quantitative particle enhanced immunoinhibition method (9) (Dade Behring aca IV® analyzer, Dade Behring Inc., Wilmington, DE).
In the DCA method, albumin-specific goat anti-human polyclonal antibody binds to albumin in the presence of polyethylene glycol. Formation of albumin-antibody complexes increases turbidity, which is measured by the absorbance at 531 nm. The preprogrammed calibration curve uses calibrator values assigned by the use of dilutions of the Reference Preparation for Proteins in Human Serum (CRM470) (11).
In the kinetic creatinine assay, creatinine complexes with 3,5-dinitrobenzoic acid at an alkaline pH to form a red chromogen, which is measured at 531 nm (12). The calibration curve is programmed into the analyzer by a bar code.
To perform a test, the cartridge is removed from the foil and the bar code is scanned in the track located at the side of the analyzer. Urine (40 µL) is collected with a capillary action sampler provided as part of the assay kit and introduced into the cartridge. The cartridge is then inserted into the analyzer; the fluid release tab is pulled, releasing alkaline nitrobenzaldehyde; and the lid closed to start the analysis. The reactions are performed automatically, and the results displayed in ~7 min.
Comparative albumin analyses were performed on a Dade Behring aca IV discrete clinical analyzer (Dade Behring Inc.) (9), which uses albumin covalently coupled to latex particles and a monoclonal antibody against human albumin. The aggregation, which decreases with increasing urine albumin concentration, was monitored at 340 nm. The creatinine assay used a kinetic Jaffe procedure.
We analyzed 96 urine samples (from 24-h collections), from patients with diabetes and/or renal failure selected from the routine workload of the laboratory to represent a wide range of analyte concentrations for the comparison study. The samples were stored for a maximum of 1 week at 4 °C; the samples were warmed to room temperature and centrifuged at 1000g for 5 min before analysis.
The within-run imprecision (CV) was assessed by the assay of two
control materials and two urine samples in duplicate 15 times in a
randomized batch. The between-run imprecision was assessed with two
control materials over 20 days. The within- and between-run CVs for
albumin, creatinine, and the calculated ratio are shown in Table 1
.
|
The precision profile derived from duplicate analysis on the DCA and aca methods was used to assess the working range of the assays at a CV of <10%; the ranges were 5–300 mg/L for albumin and 1–60 mg/mmol for the albumin:creatinine ratio for the DCA method; the ranges were 5–300 mg/L and 0.5–60 mg/mmol, respectively, for the aca methods.
For the assessment of method linearity, five urine samples with albumin concentrations within the microalbuminuric range (60.7–263 mg/L) were diluted 1/10 to 9/10 in a solution containing 9 g/L NaCl and 0.01 g/L Brij 35 and analyzed using the DCA 2000. Linear regression of the observed vs expected values for each of the dilutions yielded slope values within the range 0.968–0.994, with r2 values between 0.996 and 0.998.
Ten of the urine samples collected were assayed undiluted, using the DCA 2000 method. They were then mixed with either a 50 or 100 mg/L solution of albumin (code ORHA 20/21; Dade Behring) in 9 g/L sodium chloride, using one volume of albumin solution to nine volumes of urine. The mixtures were reassayed, and the analytical recovery for each sample was calculated; the mean recovery was 107.3% (SD, 5.3%) for the 50 mg/L albumin solution and 108.5% (SD, 6.3%) for the 100 mg/L albumin solution.
The addition of bilirubin (final concentration, 44 and 88 mg/L), glucose (final concentration 2500 and 5000 mg/L), and hemoglobin [final concentration, 300 mg/L (30 mg/dL) and 600 mg/L (60 mg/dL)] had no significant effect on measured albumin and creatinine values.
Passing-Bablock regression analysis (13) for the DCA 2000 (y-axis) and aca IV albumin assays yielded a slope of 0.84 [95% confidence interval (CI), 0.82–0.87] and intercept of 0.57 (95% CI, -0.17 to 1.34).
The slope of 0.84 could be related to calibration because the DCA 2000 was calibrated against the international reference material for serum proteins (CRM470) and the Dade aca was calibrated against an artificial calibrator containing purified albumin. To examine this possibility, the aca calibrators were assayed on the DCA 2000 and the results were plotted against the assigned values for the aca calibrators; this yielded a slope of 0.79 and an intercept of 3.66 (r2 = 1.00). The reciprocal of the slope of this graph was used as a correction factor to reassign DCA 2000 results to the aca calibrator values for the purposes of the comparison study.
Passing-Bablock regression with the corrected albumin values yielded a
slope of 1.02 (95% CI, 1.00–1.05) and an intercept of 1.40 (95% CI,
0.64–2.00; Fig. 1
A). The results for the DCA 2000 creatinine method showed close
agreement with the aca IV method, with a slope of 1.06 (95% CI,
1.02–1.09) and an intercept of 0.13 (95% CI, -0.11 to 0.28). The
albumin:creatinine ratios calculated by the DCA 2000 (with the
corrected albumin value for the purposes of comparison) showed close
agreement with the ratios calculated by the aca method, with a slope of
0.97 (95% CI, 0.94–1.00) and an intercept of 0.13 (95% CI, -0.08 to
0.39; Fig. 1B
).
|
When we considered the DCA 2000 result as a two-class test with a cutoff of 20 mg/L for albumin, the sensitivity was 91.9%, the specificity was 100%, and the positive predictive value was 100% when compared with the laboratory procedure. There is less agreement on a lower cutoff value for the albumin:creatinine ratio, although a value of 2.65 mg/mmol creatinine (23.5 mg/g) has been proposed by the American Diabetes Association (14). Analysis of data with this cutoff for the DCA 2000 system gave a sensitivity of 100%, a specificity of 92%, and a positive predictive value of 97%.
Some authors have proposed that the imprecision of an albumin assay to be used in the detection of microalbuminuria should be <12%, with the capability of detecting changes of 10 mg/L in the concentration range 5–35 mg/L (5). The DCA method meets these proposed criteria; there have been no specific proposals made with regard to the albumin:creatinine ratio. Although the DCA method showed a good correlation of results with a laboratory procedure, there appears to be a systematic error that was possibly attributable to differences in the calibration procedures. These observations support the case for an international reference preparation for urine albumin (15)(16). If a patient had been monitored using both of the systems studied here, perhaps alternately in the primary care and outpatient clinic setting, there would have been an inherent change in the result of 15% on average, equivalent to ± 3mg/L at the cutoff concentration, irrespective of any pathological change.
The data show that there are more points above the upper limit of the microalbuminuria range when it is expressed as 200 mg albumin/L compared with the albumin:creatinine ratio of 30 mg/mmol (two vs nine, respectively). The difference probably reflects a higher variability in the constituents in the presence of renal pathology. The final choice among proposed cutoff values will need be based on large outcome studies. To ensure wide applicability of these studies, it is important to achieve comparability of results between analytical systems; this again suggests the need for an international reference preparation for urine albumin.
The advent of a therapeutic intervention that delays the progression of diabetic nephropathy provides a stronger argument for the use of a fully quantitative technique once excretion above a threshold value has been established; this can be achieved with the DCA 2000 system, which also has the capability for measuring HbA1c.
Acknowledgments
The costs of this work were defrayed by a grant from Bayer plc; D.J.N. was supported by a grant from Dade Behring Inc.
Footnotes
fax 44 171 377 1544, e-mail c.p.price{at}mds.qmw.ac.uk
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  K. M. Aakre, G. Thue, S. Subramaniam-Haavik, T. Bukve, H. Morris, M. Muller, M. V. Lovrencic, I. Plum, K. Kallion, A. Aab, et al. Postanalytical External Quality Assessment of Urine Albumin in Primary Health Care: An International Survey Clin. Chem., October 1, 2008; 54(10): 1630 - 1636. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Pugia, G. Blankenstein, R.-P. Peters, J. A. Profitt, K. Kadel, T. Willms, R. Sommer, H. H. Kuo, and L. S. Schulman Microfluidic Tool Box as Technology Platform for Hand-Held Diagnostics Clin. Chem., October 1, 2005; 51(10): 1923 - 1932. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J.G. Hanley, S. B. Harris, M. Mamakeesick, K. Goodwin, E. Fiddler, R. A. Hegele, J. D. Spence, A. A. House, E. Brown, B. Schoales, et al. Complications of Type 2 Diabetes Among Aboriginal Canadians: Prevalence and associated risk factors Diabetes Care, August 1, 2005; 28(8): 2054 - 2057. [Full Text] [PDF]
|
|
|
|
|
|
Q.-P. Qin, O. Peltola, and K. Pettersson Time-resolved Fluorescence Resonance Energy Transfer Assay for Point-of-Care Testing of Urinary Albumin Clin. Chem., July 1, 2003; 49(7): 1105 - 1113. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
