urines containing the 5.0 g/L concentrations of glucose
Glucosuria: Accuracy and precision of
produced readings 2.5 g/L. Neither method showed inter-
of
ference from 0.1 or 0.2 g of ascorbic acid per liter.
On the basis of these interference data, we predict that
the BM33071 would not present false-negative results un-
der the conditions studied. The effect of ascorbic acid should
be minimal, given the normal excreted concentration of
ascorbic acid of usually less than 0.25 gIL.
References
1. Peterson JI, Young DS. Evaluation of the hexokinaae/glucose-6-
phosphate dehydrogenine method of determination of glucosein
urine. Anal Bio.chem 23, 301-316 (1968).
CLIN. CHEM. 31/1, 92-94 (1985)
Indican Interference with Six Commercial Procedures for Measuring Total
Bilirubin
R. Poon and I. H. Hlnberg1
We have studied the effect of indican on six commercial
procedures for the measurement of total bilirubin in serum.
Total bilirubin measured by the Bilirubin A-Gent (Abbott)
2,4-dichlorophenyl diazonium procedure increased by 50
mg/L for each 1 mmol/L of added indican. Similarly, total
bilirubin measured by the Bilirubin C-System (Boehnnger
Mannheim) 2,5-dichiorophenyl diazonium procedure in-
creased by 33 mg/L per mmol/L of indican. Indican also
interfered with the Micro Bilirubin Reagent Sett (Harleco)
Malloy-Evelyn procedure, but to a much lesser extent. The
Jendrassik Bilirubin Reagent System#{176} (American Monitor)
and a modified Jendrassik-Grof procedure (Hoffmann-
LaRoche) adapted to the Cobas Blo analyzer were unaffect-
ed by the presence of indican. The amount of interference
with the 2,5-dichiorophenyl diazonium procedure increased
significantly with color development time and was twice the
initial amount after 30 mm. Concentrations of indican as high
as 0.38 mmol/L have been found in sera of patients with renal
failure, which would increase total bilirubin values measured
by the first two procedures above by 19 and 12 mg/L,
respectively. Users of these procedures should therefore be
suspicious of unexpectedly high bilirubin values obtained
with sera from patients with chronic renal disease.
Addftlonal Keyphrases: renal disease analytical error
Fifty-six years ago, Harrison and Bromfield (1) reported
that indican (indol-3-yl sulfate), a natural metabolite that
accumulates in the sera of patients with chronic renal
failure (2,3), interfered with the colorimetric determination
of total bilirubin. However, except for one report by Ertings-
Division of Research and Standards, Bureau of Medical Devices,
Environmental Health Centre, Tunneys Pasture, Ottawa, Ontario
K1A 0L2, Canada.
Address correspondenceto this author.
Received March 19, 1984; accepted August 30, 1984.
92 CLINICALCHEMISTRY, Vol. 31, No. 1, 1985
2. James GP, Bee DE.
laboratory diagnosis by dipstick analysis. Clin Chem 25, 996-1001
(1979).
3. Kutter D. Rapid Clinical Diagnostic Tests, Urban and Schwar-
zenberg, Muenchen-Wien-Baltimore, 1977,pp 16-17.
4. Dane LNW, Juell A. Ascorbicacid and test strip reactions for
haematuria. Scand J Clin Lab Invest 43, 267-269 (1983).
5. Bandi ZL, Myers JL, Bee DE, et al. Evaluation of determination
of glucosein urine with somecommerciallyavailable dipsticksand
tablets. Clin Chern 28, 2110-2115 (1982).
6. GoodmanLA, Kruskal WH. Measures of association for cross
classifications ifi: Approximatesamplingtheory. JAm Stat Assoc
58, 310-364 (1963).
hausen et al. (4), that the determination of total bilirubin by
a procedure involving 2,4-dichlorophenyl diazonium (2,4-
DCPD) gave falsely increased results in the presence of
indican, no attention has since been paid to this interferent.
We have studied the effect of indican on six commercial
colorimetric procedures for the measurement of total biliru-
bin in serum and report our findings.
Materials and Methods
Indican was obtained from Sigma Chemical Co., St. Louis,
MO 63178; reference grade bilirubin from Pfanstiehl Labo-
ratories Inc., Waukegan, IL 60085; bovine serum albumin
from Armour Pharmaceutical Co., Kankakee, IL 60901; and
2,4-dichloroaniline and 2,5-dichloroaniline from Aldrich
Chemical Co., Milwaukee, WI 53201. All other chemicals
used were reagent grade, obtained from Fisher Scientific
Co., Ottawa, Ontario, K2E 7L6.
We prepared samples with indican concentrations be-
tween 0 and 0.92 mmol/L by adding appropriate volumes of
a 47 mmol/L solution of indican to aliquots of pooled human
serum. The manual and automated colorimetric bilirubin
measurement procedures we studied are described in Table
1. The kit manufacturers directions for use were strictly
followed. Where calibrators were not provided, we used
bilirubin standards in bovine serum albumin, 40 g/L, pre-
pared as described by Perry et al. (5).
Stabilizer-free 2,4-DCPD reagent was prepared at room
temperature by adding sodium nitrite (final concentration,
10 MlnolJL) to a solution containing 2 minol of 2,4-dichloro-
aniline and 69 minol of sulfamic acid per liter of water!
methanol (1/1, by vol). We used the reagent 2 mm after
adding the sodium nitrite.
To measure absorbance, we used a Beckman DU-8B
spectrophotometer (Beckman Instruments Inc., Fullerton,
CA 92634); difference spectra were obtained with a Cary 219
spectrophotometer (Varian Associates Inc., Palo Alto, CA
94303).
 Table 1. Bllirubin Test Kits Studied
KIt
A. A-Gent BilirubinTest (AbbottLaboratories,Mis-
sissauga,OntarioL4W 2S7)
B. BilirubinC-System (BoehringerMannheimCan-
ada, Dorval,Quebec H9P 1A9)
C. MicroBilirubinReagent Set (Harleco, Gibbs-
town, NJ 08027)
D. JendrassikBilirubinReagent System (American
MonitorCorp., Indianapolis,IN 46268)
E. CobasTotal Bilirubin(Hoffmann-LaRoche
Ltd., Etobicoke,OntarioM9C 5J4)
F. JendrassikBilirubinReagent Systemadapted
to the Cobas Bio analyzer (by Roche Analyti-
cal InstrumentsInc., Nutley,NJ 07110)
ConditIonsfor szobHlrubln tennatlon
Reactionwith stabilized 2,4-DCPD reagent at acidicpH in the
presenceof methanol
Reaction with 2,5-DCPD reagentat acidicpH in the presenceof a
detergent
Reactionwith p-diazobenzenesulfonicreagent at acidicpH in the
presenceof methanol(Malloy-Evelynmethod)
Reactionwith p-diazobenzenesulfonicreagent at acidicpH in the
presenceof caffeinereagentfollowedby an alkaline reagent ad-
dition(Jendrassik-Grofmethod)
Reactionwith a stabilizedp-diazobenzenesulfonicreagentat pH
1.6 in the presenceof dimethylsulfoxide(modifiedMalloy-Eve-
lyn method).
Reactionwith p-diazobenzenesulfonicreagent at acidicpH in the
presenceof caffeinereagent,but withoutadditionof alkalinerea-
gent (modified Jendrassik-Grofmethod)

Results 0.2-

Our results (Figure 1) show that total bilirubin measured 0

U
with the Bilirubin C-System, which involves use of 2,5- C
dichlorophenyl diazonium (2,5-DCPD), increased by 33 mgI 0 /
L for each millimole of indican added per liter. Similarly, .0 0_i.
S
0 -.-- --
S
total bilirubin values obtained with the A-Gent bilirubin U) S
kit, a 2,4-DCPD procedure, increased by 50 mgfL for each .0 S.

millimole of indican per liter. Indican also interfered, but to

a much lesser extent, with the Micro Bilirubin Reagent Set, 0-
which involves a Malloy-Evelyn procedure. In contrast, the I I I I I

Jendrassik Bilirubin Reagent System, a Jendrassik-Grof 425 475 525 575 625 675
method, was not affected by the presence of indican. Wavelength(nm)
Roche Analytical Instruments has adapted this last proce-
dure to the Cobas Bio Centrifugal Analyzer by omitting the 0.4-
final alkaline reagent addition step; this modified procedure
was similarly unaffected by indican. The Cobas Total Biliru-
bin Reagent procedure, a modified Malloy-Evelyn method I;
in which the accelerating agent is dimethyl sulfoxide in- 0.3-
stead of methanol, was also unaffected by indican.
We also confirmed the findings of Ertingshausen et al. (4)
U
that the 2,4-DCPD reagent reacts with bilirubin to form a C
color complex with a maximum absorbance at 540 nm. 0.2-
.0
Reaction of the 2,4-DCPD reagent with indican, on the other 0
hand, produced a color complex with an absorbance band .0
centered around 480 nm (Figure 2, top). The significant
0.1-
absorbance of this band at 540 nm was responsible for the
erroneously high total bilirubin values obtained with 2,4-

0.
I
350 400
I I
450
I
500 550
I -- 600
I
650
Wav#{149}Iength (nm)
Fig. 2. Absorbancespectra of color complexes formed by reaction of
2,4-DCPD (top) or 2,5-DCPD (bottom) with bilirubin or indican
Tpe, reactionof 2,4-DCPDwith 40 mg of bilirubin (- - -) or with 0.46 mmcl of
indican(-) per liter of bovine serum albumin(40 g/L); referencecell: diazonium
reagentplusalbumin(40 g/L). Bottom, reactionof2,5-DCPDwith samesolutions
of bilirubin (-) and jndican (- - -, 5 mm after mixing; -., 30 mm after mixing).
Referencecell:diazoniumreagentplusalbumin(40 g/L)

DCPD procedures in the presence of indican. Both color

Concentrition
Fig. 1. Effect of added indicanon total bilirubinmeasured by six
commercial procedures
A-F as entified in Table 1
complexes were stable for at least 30 mm.
Although the 2,5-DCPD reagent is structurally similar to
the 2,4-DCPD reagent, it forms different color complexes
with bilirubin and indican (Figure 2, bottom). As shown
previously by Wahlefeld et al. (7), the 2,5-DCPD reagent
Os
ol Indican (mmol/L) reacted with bilirubin to produce a color complex with
maximum absorbance at 520 rim that was stable for at least
30 mm. In contrast, the color complex initially formed by
reaction of the 2,5-DCPD reagent with indican had absor-

CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985 93

bance bands with maximum absorbances at 375, 460, and
530 nm; the absorbance of the 460 nm band decreased with
time, while that of the 530 nm band increased to almost
twice its initial value after 30 mm (Figure 2, bottom).
Because it is this band at 520 nm that is responsible for the
interference of indican with the measurement of total biliru-
bin by 2,5-DCPD procedure, the amount of interference will
increase significantly as color development is prolonged.
The absorbance spectra in Figure 2 were obtained with
commercial 2,4-DCPD and 2,5-DCPD reagents (Table 1).
Because the commercial 2,4.-DCPD reagent used contained a
stabilizer, probably 1,5-naphthalenedisulfonate (4), we also
conducted studies with stabilizer-free 2,4-DCPD reagent.
This reagent reacted with bilirubin and indican to produce
color complexes with absorbance spectra identical to those
in Figure 2 (top).
In the Malloy-Evelyn procedure, bilirubin reacts with p-
diazobenzenesulfonic reagent to form a color complex with a
maximum absorbance at 560 nun. Indican also reacts with
the reagent, to give peak absorbances at 380 and 490 rim.
However, the 490-nun peak had near-baseline absorbance at
560 nm and therefore only minimally influenced biirubin
measurement.
In the Jendrassik-Grof procedure, bilirubin reacts to form
a broad absorbance band from 400 to 700 nun with peak
absorbance at 595 nm. Indican does not react to produce a
color complex in this wavelength range.
Discussion
Only the 2,4-DCPD and 2,5-DCPD procedures for deter-
mining bilirubin were markedly affected by indican, which
reacted with the reagents to form color complexes that
absorbed significantly at the wavelengths used for the
measurement of bilirubin, thereby producing falsely high
results.
Concentrations of indicart as high as 0.38 mmol/L have
been found in sera of patients with chronic renal failure (4).
Our results show that this concentration would increase
total bilirubin values measured by the 2,4-DCPD and 2,5-
DCPD diazonium procedures by more than 19 and 12 mg/L,
respectively. Because a total bilirubin value of 15 mg/L is
generally considered abnormally high (8), the interference
94 CLINICALCHEMISTRY, Vol. 31, No. 1, 1985
of indican with these procedures may lead to false presump-
tive diagnoses of liver abnormalities in these patients,
necessitating time-consuming, expensive, and hazardous
additional diagnostic procedures.
Because of their relative stability and the speed with
which they react with bilirubin, both the 2,4-DCPD and 2,5-
DCPD procedures are widely used in automated and man-
ual procedures for total bilirubin (4, 9, 10). Users of these
reagents should be suspicious of unexpectedly high values
for bilirubin in sera from uremic patients.
References
1. Harrison GA, Bromfield RJ. VII. The causeof Andrewess diazo-
test for renal inefficiency. Biochem J 22, 43-45 (1928).
2. Ludwig GD, Senesky D, Bluemle LW Jr, Elkington JR. Indoles
in uremia: Identification by countercurrent distribution and paper
chromatography. Am J Clin Nut,- 21, 436-450 (1968).
3. Swan JS, Kragten EY, Veemng H. Liquid-chromatographic
studyof fluorescent materials in uremic fluids. Clin Chem 29, 1082-
1084 (1983).
4. Ertingshausen G, Fabiny Byrd DL, Tiffany TO, Casey SJ.
Single-reagent method for rapid determination of total bilirubin
with the CentrifiChem analyzer. Clin Chem 19, 1366-1369
(1973).
5. Perry BW, Doumas BT, Bayse DD, et al. A candidate reference
method for determination of bilirubin in serum. Test for transfer-
ability. Clin Chem 29, 297-301 (1983).
6. Cross RE, Heintges MG, Savory J, Wentz PW. Adaptation of
blank-corrected methods for measurement of total and conjugated
bilirubin and uric acid to the centrifugal analyzer. Clin Chem 22,
429-433 (1976).
7. Wahlefeld AW, Herz G, Bernt E. Modification of the Malloy-
Evelyn method for a simple, reliable determinationof total biiru-
bin in serum. Scand J Clin Lab Invest 29, Suppl 126 (1972).
Abstract.
8. Henry RJ, Cannon DC, Winkelman JW, Eds. Clinical Chemis-
try. Principles and Technics, 2nd ed., Harper and Row, New York,
NY, 1974, p 1060.
9. Kelley A, McKenna JP, McLelland A, et al. A bichromatic
methodfor total bilirubin with a CentriflChem 400. Clin Chem 25,
1482-1484 (1979).
10. Kineiko RW, Floering DA, Morrissey M. Laboratory evaluation
of the Boehringer Mannheim Hitachi 705 automatic analyzer.
Clin Chem 29, 688-691 (1983).