TECHNICAL ARTICLE

Evaluation of New Rapid Office Test for Microalbuminuria and Its Comparison to Fully Quantitative Radioimmunoassay

A novel enzyme-linked immunosorbent assay (ELISA [Dialbumin]) for rapid office measurement of microalbuminuria was evaluated and its performance compared with that of a commercially available radioimmunoassay (double-antibody albumin). Urine samples containing between 0.75 and 1800 |xg/ml of albumin were obtained from 31 diabetic patients and assayed by both methods. A comparison of the paired values obtained from the two methods gave a correlation coefficient of >0.99. The Dialbumin assay, which used detachable eight-well strips (1 strip/sample), 10-min incubation, tap water wash, and a 2-min color development step, was read on both an ELISA reader and a hand-held analytical device (Acc-U-Dial) designed specifically for this test. The findings of this study indicate that the Dialbumin assay, used in conjunction with the Acc-U-Dial device, affords a rapid, convenient, and sensitive method for quantitative determination of a broad range (0.3-1280 fxg/ml) of urinary albumin levels in the office setting. Diabetes Care 13:1069-73, 1990

M

edical literature on the development of diabetic nephropathy and the recent introduction of several relevant commercial products have generated increased awareness of the importance of routine testing for microalbuminuria. The appearance of microalbuminuria (20-200 jji,g/min) has been established as an important marker of incipient From Exocell and the Department of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; and the Department of Endocrinology, University of Parma, Parma, Italy. Address correspondence and reprint requests to R. Glenn Neuman, MS, Exocell, Inc., 3508 Market Street, Room 420, Philadelphia, PA 19104. Received for publication 8 September 1989 and accepted in revised form 9 May 1990.

DIABETES CARE, VOL. 13, NO. 10, OCTOBER 1990

R. Glenn Neuman, MS Luigi V. Bonomini, MD Seth N. Braunstein, MD

nephropathy (1-4), and a recent article describing glomerular structural changes characteristic of diabetic nephropathy in a group of diabetic patients, with a mean albumin excretion rate of 68 (xg/min, underscores the clinical relevance of detecting and monitoring microalbuminuria (5). If progression to irreversible renal involvement is to be prevented, renal dysfunction at its earliest and potentially most reversible form must be detected, allowing identification of patients in whom aggressive therapy might halt the development of this complication. To determine which clinical treatment will best prevent or retard the development of diabetic nephropathy, careful and frequent monitoring of patients under various treatment regimens should be encouraged. If the data from such studies are to be comparable and meaningful, standardized methods for measuring urinary albumin should be adopted. Of the numerous methods available, none have been quantitative enough to detect small changes and simple enough to be performed by office staff. In this study, the performance of a novel quantitative enzyme-linked immunosorbent assay (ELISA [Dialbumin]) and a hand-held reading device, Acc-UDial, suitable for office-based measurement of microalbuminuria, was compared with that of a commercially available radioimmunoassay (RIA). The Dialbumin assay was also evaluated with respect to incubation time and interference by substances that may be present in the urine, and the accuracy of the Acc-U-Dial was compared with that of an ELISA reader.

RESEARCH DESIGN AND METHODS Dialbumin kits and the Acc-U-Dial hand-held reading device were obtained from Exocell (Philadelphia, PA).

1069

RAPID MEASUREMENT OF DIABETIC NEPHROPATHY

Dialbumin is a competitive immunoassay, wherein the intensity of colored product decreases in relation to increasing amounts of albumin between 0.1 and 10 |xm/ ml. By serially diluting a single urine sample into eight consecutive microtiter wells, as described in the kit insert, the range can be extended to 2560 |xg/ml. The Acc-U-Dial reader is a hand-held device in which, after completion of the assay, an eight-well strip of serially diluted sample is inserted alongside an eight-well strip of simultaneously run serially diluted standard. According to the manufacturer's instructions, the device's dial is turned, causing an alignment of colored wells in the sample with those in the standard. When the colors in the sample and standard wells are matched well for well, the appropriate concentration (in (xg/ml) appears in the window of the device. Both the Acc-U-Dial and an ELISA reader (MR700, Dynatech, Alexandria, VA) were used to determine urinary albumin concentrations for samples run with the Dialbumin assay. Double-antibody albumin RIA kits were purchased from Diagnostic Products (Los Angeles, CA). The assay was performed as described in the kit insert. Gamma emissions from the RIA were counted on a Mini Instruments Well Probe type 5-43 (Essex, UK). Urine samples, where indicated, were centrifuged on an HBA microcentrifuge for 2 min at 13,000 rpm. Human serum albumin, gentamicin, creatinine, and urea were purchased from Sigma (St. Louis, MO). Urine samples from 31 diabetic men and women, ranging from 18 to 80 yr of age, were randomly collected during routine scheduled visits to the diabetic clinic at the Hospital of the University of Pennsylvania. The samples were stored at -20°C for up to 2 wk and then assayed by RIA and Dialbumin. Before being assayed, the samples were thawed at room temperature and mixed, and a portion was withdrawn and centrifuged. Only centrifuged samples were run in the RIA, as recommended in the kit insert. Samples were run in the ELISA both before and after centrifugation, because the Dialbumin instructions did not stipulate a need for centrifugation. An additional 30 samples were used to determine intra- and interpersonnel variation for the Acc-U-Dial reader. This group of samples included one nondiabetic urine that was divided into seven different samples consisting of dilutions of 1:16 and 1:4 in water, an undiluted sample, and four additional samples to which 20, 80, 320, and 1280 |xg/ml of urine were added. Twentythree randomly collected diabetic urines, which were stored at 4°C and assayed within 3 days, were included in this latter group. A final urine sample from a nondiabetic man, used to determine albumin recovery from urine, was divided into three portions to which human serum albumin was added in the amounts of 0, 67, and 167 |xg/ml and stored at -80°C. Intraobserver variation in Acc-U-Dial readings was studied by having a single individual read 22 different sample strips, which were stored at 4°C in a humidified container, on 3 different days. Interobserver variation

1070

was assessed by assaying duplicates of 15 samples that were chosen to cover the entire range of the Acc-U-Dial readout and presenting them in random order to 10 different observers.

RESULTS Urine samples from the 31 diabetic patients, assayed in the double-antibody RIA and Dialbumin with an ELISA reader and the Acc-U-Dial, were found to contain between 0.75 and 1800 (xg/ml of albumin (Table 1). For the total sample population, a correlation of >0.99 was found when Acc-U-Dial values were paired with either RIA values or ELISA reader values (Fig. 1). An intraclass correlation between Acc-U-Dial and RIA was performed for the normal values (200 jig/ml, n = 5), resulting in correlation coefficients of 0.82, 0.89, and 0.98, respectively. The effect that sample centrifugation has on assay results was determined for all 31 diabetic urine samples by assaying the samples within the Dialbumin kit both before and after centrifugation. Assay results were then read both on the ELISA reader and the Acc-U-Dial. There were no significant differences in results obtained with centrifuged versus uncentrifuged samples (data not TABLE 1 Comparative values of urinary albumin from diabetic patients, as determined with Dialbumin kit, Acc-U-Dial and enzyme-linked immunosorbent assay (ELISA) readers, and double-antibody radioimmunoassay kit Dialbumin Acc-U-Dial (|xg/ml)

Dialbumin/ELISA reader

Patient

Double-antibody radioimmunoassay ((xg/ml)

1 2 3 5 7 8 9 10 11 12 13 18 19 20 22 23 25 26 27 28 29 30 31

7.4 276.0 6.6 6.9 77.0 1.0 13.4 384.0 1760.0 120.0 2.7 34.6 603.0 181.0 1.0 9.7 707.0 7.5 87.0 46.0 7.6 3.8 40.0

5.0 220.0 5.0 5.0 60.0 1.3 10.0 320.0 1280.0 160.0 2.5 40.0 640.0 160.0 0.3 10.0 640.0 10.0 80.0 20.0 10.0 2.5 40.0

4.0 275.2 7.2 4.0 84.4 0.9 10.8 312.1 1461.1 121.9 2.4 40.1 513.1 255.6 0.5 8.3 482.2 6.7 84.5 21.4 8.0 3.8 33.6

(jjLg/ml)

DIABETES CARE, VOL. 13, NO. 10, OCTOBER 1990

R.G. NLUMAN, L.V. BONOMINI, AND S.N. BRAUNSTEIN

1000

TABLE 2 Day-to-day interday intraobserver variation in readings for 22 unknown sample strips that were preserved and read by single observer

100

1

10

100

1000

DIALBUMIN with ACC-U-DIAL,

FIG. 1. Linear regression of paired urinary albumin values obtained from Dialbumin assay and double-antibody radioimmunoassay (DPC RIA) kit. Acc-U-Dial reader was compared with both DPC RIA (A) and enzyme-linked immunosorbent assay (ELISA; V) reader. Results from 30 diabetic patients are depicted.

shown). The average ratio of centrifuged to uncentrifuged values, calculated for all 31 diabetic samples, was 0.99 with Acc-U-Dial and 1.05 with the ELISA reader. A normal urine sample containing 5.4 |xg/ml of excreted albumin and 0, 67, and 167 |xg/ml of added albumin was assayed in Dialbumin. When read on the ELISA reader, these samples were found to contain 5.4, 70.9, and 162.2 (xg/ml, respectively, giving recoveries of 100, 97.9, and 94.1%, respectively. When the same samples were read on the Acc-U-Dial, values of 5, 80, and 160 |xg/ml were obtained, giving respective recoveries of 92.6, 110.5, and 92.8%. The effect of incubation time in the Dialbumin assay was examined by incubating standards for 2, 5, 10, 30, and 60 min and reading the wells on the ELISA reader. Virtually no difference in the standard curve could be seen for incubation times between 5 and 60 min, but 100 75 55 O i—i

50

CQ t—i

g 25

-25

0.10

1.00 HSA, /zG/mL

10.00

FIG. 2. Effect of incubation time on relative percent inhibition produced by increasing amounts of human albumin standard (HSA) in Dialbumin assay. Incubation with antibody was allowed to proceed for 2 (o), 5 (A), 10 (•), 15 (V), 30 (•), and 60 (O) min at room temperature.

DIABETES CARE, VOL. 13, NO. 10, OCTOBER 1990

Sample

Day 1

Day 2

Day 3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Mean

20.0 5.0 10.0 2.0 5.0 0.3 10.0 2.5 1.3 80.0 60.0 320.0 40.0 1.3 40.0 0.3 2.5 10.0 2.5 2.5 160.0 40.0

10.0 5.0 10.0 2.5 5.0 0.3 10.0 2.5 1.3 80.0 60.0 320.0 40.0 1.3 30.0 0.3 2.5 10.0 1.3 3.5 120.0 60.0

10.0 5.0 10.0 1.3 5.0 0.3 10.0 2.0 1.3 80.0 60.0 320.0 40.0 1.3 30.0 0.3 2.0 10.0 1.3 2.5 160.0 60.0

Mean ± SE 13.3 5.0 10.0 1.9 5.0 0.3 10.0 2.3 1.3 80.0 60.0 320.0 40.0 1.3 33.3 0.3 2.3 10.0 1.7 2.8 146.7 53.3

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

2.7 0.0 0.0 0.3 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 2.7 0.0 0.1 0.0 0.3 0.3 10.9 5.4

Correlation coefficient (%) 20.4 0 0 14.7 0 0 0 5.8 0 0 0 0 0 0.0 8.2 0.0 5.8 0.0 19.2 9.6 7.4 10.2 4.6

the 2-min incubation produced an erratic curve that showed significantly less inhibition (Fig. 2). The presence of urea, creatinine, and gentamicin, when added to the Dialbumin test kit in concentrations as high as 10 |xg/ml, had no effect on assay results (data not shown). Day-to-day intraobserver variation in Acc-U-Dial readings was studied with the use of 22 diabetic urine samples containing unknown levels of albumin. The samples were assayed with Dialbumin, after which the developed strips were preserved at 4°C in a humidified chamber, randomized, and read by the same individual on 3 separate days. From the 22 samples, 13 gave identical readings over the 3 days. Nine of 22 samples, most of which were in the 2.5-|xg/ml range, showed slight inconsistencies when read repeatedly (Table 2). Intraobserver variation in Acc-U-Dial sample reading was assessed by having 10 different observers from diverse backgrounds read the same group of samples, which contained 0-5000 |xg/ml albumin. The samples consisted of one normal sample that was divided into seven diluted or spiked samples containing between 0.3 and 1280 fxg/ml albumin, and six diabetic urine samples selected for their range of albumin content. Also included were a negative (0 fxg/ml) and positive (5000 |xg/ml) control. Each of the 15 samples was assayed twice in the Dialbumin kit, and each observer read the resulting 30 strips on the Acc-U-Dial. The resulting interobserver frequency data are shown in Fig. 3. Read-

1071

RAPID MEASUREMENT OF DIABETIC NEPHROPATHY

> 1,280 II

20 |

1,280 || 640 || 320 || 160

c,

S a

18

||

6

18

i| 40 jj

14

2

8O

9 6

20 ll 101| sj|

*

ii

=•

25

14

13

6

|

4

2

2 13 18

7

19 1 15 9 10

1.3 ||

1

0.6 ||

2 12

0.3 ||

3

2

16

60 |xg/ml required repeating the assay at several dilutions. Analysis of the data in Tables 1 and 2 and Fig. 3 suggests that variation in Dialbumin/Acc-U-Dial readings is greatest for samples containing albumin in the range of 20 (xg/ml. Figure 3 shows that 10% of the readers interpreted a 20-|xg/ml

T

1072

sample to be 40 fxg/ml. When a 30-|xg/ml sample was assayed, 65% of the readers reported a 40-|xg/ml reading and 35% reported a 20-|xg/ml reading. Consequently, patients having urinary albumin in the 20- to 40-|jLg/ml range should be retested at a later date to confirm the presence of persistent microalbuminuria. For those samples with values that appear to fall in between readings on the Acc-U-Dial, the higher value may more accurately reflect the true urinary albumin level. The studies reported herein were conducted with random urine samples collected at physicians' offices during scheduled visits. Provided that polyuria or oliguria is not present, microgram per mi I Ml iter concentration values (from single-void specimens) will closely approximate microgram per minute excretion rates (from timed collections), and in patients with urinary output of 60 ml/h or 1.44 L/day, the two values will be interchangeable (6). Exercise may augment albumin excretion, and measurement of postexercise urinary albumin can be used to unmask latent microalbuminuria diabetic subjects (7-9). Correction for creatinine excretion (e.g., creatinine-albumin ratios) adds little to information evident in concentration (|xg/ml) results and, in fact, may inappropriately distort results, because in athletically active individuals, postexercise increases in urinary creatinine lag 6-12 h behind increases in urinary albumin (7). Finally, as noted in the manufacturer's instructions, contamination of samples with blood or semen must be avoided. Urine albumin concentrations in single samples with blood or semen must be avoided. Urine albumin concentrations in single-void samples should be interpreted, in conjunction with the patient interview, to determine his/her daily routine, and a baseline range should be established from several samples taken within 1-2 wk. The results of this study show that the Dialbumin/ Acc-U-Dial office test has the attributes necessary to make it an effective tool for the detection and monitoring of microalbuminuria. One kit was able to process a standard and up to 23 patient samples in

Evaluation of new rapid office test for microalbuminuria and its comparison to fully quantitative radioimmunoassay.

A novel enzyme-linked immunosorbent assay (ELISA [Dialbumin]) for rapid office measurement of microalbuminuria was evaluated and its performance compa...
531KB Sizes 0 Downloads 0 Views