13. Cobb WE, Lamberton RP, Jackson IMD. Use of a rapid, 16. Klingler W, Ball P.

er W, Ball P. Die Bestimmung des Thyreotropin (TSH) in

sensitive immunoradiometric assay for thyrotropin to distinguish
normal from hyperthyroid subjects. Cm Chem 1984;30:1558-60.
14. Allen KR, Scott RDM, Hewitt JV, Watson D. Clinical value of a
sensitive immunoradiometric assay for TSH. Ann Clin Biochem
1985;22:506-8.
15. Caldwell G, Cow SM, Sweeting VM, et al. A new strategy for
thyroid function testing. Lancet 1985;i:1117-9.
Serum mit einer sensitiven immunoradiometrischen Methode.
Arztl Lab 1985;31:219-23.
17. Woodhead JS, Weeks I. Circulating thyrotropin as an index of
thyroid function. Ann Clin Biochem 1985;22:455-9.
18. O’Malley BP, Richardson A, Cook N, Swart S, Rosenthal FD.
Circadian rhythms of serum thyrotrophin and body temperature in
euthyroid individuals and their responses to warming. Clin Sci
1984;67:433-7.

Downloaded from https://academic.oup.com/clinchem/article-abstract/32/3/518/5652453 by guest on 13 February 2020

CLIN.CHEM.32/3, 518-521 (1986)

BilirubinInterferencewith Determinationof UricAcid,Cholesterol,and

Triglyceridesin CommercialPeroxidase-CoupledAssays,and the Effectof
Ferrocyanide
Mark A. Spain1 and Alan H. B. Wu2

We examined the effect of bilirubin on measurement of Peroxidase-coupled methods having become commonplace
serum uric acid, cholesterol, and triglyceride in the BMD- in the clinical laboratory, we investigated the effect of
8700, Hitachi 705, Multistat III, CentrifiChem, Cobas-Bio, bilirubin on results of several automated tests for uric acid,
Ektachem 400, aca II, and RA-1000 analyzers. In general, cholesterol, and triglyceride and compared these results
bilirubininterferesmore in peroxidase-coupledassays than with others obtained with different assay systems. In addi-
in either direct methods or those coupled to NAD(P). The tion, we evaluated ferrocyanide for its potential to minimize
the chemical interference from biirubin.
degree of interference, which can be spectral or chemical,
variesand dependsonthe chromogenand wavelengthused. Materialsand Methods
We present data to show the origin and extent of the
interferenceby bilirubinin these systems and demonstrate Samples. Bilirubin (Fluka AG, Buchs, Switzerland) was
how ferrocyanide can provide a satisfactory resolution in added to pooled adult-human serum: 0,50,80,160,240,320,
some cases. and 400 mg of biirubin per liter, by a modification of the
standard-addition technique (5) for bilirubin. These seven
AddftlonalKeyphrases:analytical error intermethod compati- specimens were stored in aliquots in capped test tubes at
son -20 “C until use.
Procedure. Uric acid, cholesterol, and triglyceride were
The interference by biirubin in peroxidase-coupled enzy- assayed with various automated analyzers according to the
matic assays involving the spectrophotometric measure- manufacturers’ instructions: BMD-8700 (Boehringer Mann-
ment of hydrogen peroxide is well recognized (1-3). It is heim Diagnostics, Indianapolis, IN 46250), the Hitachi 705
caused by two basic mechanisms. Spectral interference (Boehringer Mannheim), the Multistat ifi (Instrumentation
results when the absorbance spectra of bilirubin and of the Laboratory, Lexington, MA 02173), the CentrifiChem
chromophore produced in the reaction overlap; this type of (Baker Instruments, Allentown, PA 18001), the Cobas-Bio
interference can be eliminated by measuring the absorbance (Roche Diagnostic Systems, Nutley, NJ 07110), the Ekta-
at a longer wavelength such that it excludes any contribu- diem 400 (Eastman Kodak Co., Rochester, NY 14650), the
tion from biirubin. Chemical interference is thought to aca U (Du Pont Instruments, Wilmington, DE 19898), and
occur when a portion of the reaction intermediate is de- the RA-1000 (Technicon Instruments Corp., Tarrytown, NY
stroyed by high bilirubin concentrations, thereby decreasing 10591). Table 1 summarizes the methods involved. All
the amount of chromophore formed; this type of interference spectral absorbance curves were obtained with a Cary 118
can be eliminated by adding ferrocyanide to the reaction spectrophotometer (Varian Instruments, Palo Alto, CA
mixture, which stabilizes the reaction intermediate (4). 94303) with a tungsten light source and a 1-cm pathlength.
The Behring “X.L.R.” uric acid reagent contains a liquid
ferrocyanide supplement, which we added as directed by the
Department of Pathology and Laboratory Medicine, University of manufacturer to yield a ferrocyanide concentration of 18
Texas Health Science Center, Medical School, Houston, TX 77025. pniollL. We then used this working reagent in the Hitachi
‘Current address: Boehringer Mannheim Diagnostics, 9115 705 according to instructions provided by Behring Diagnos-
Hague Rd., Indianapolis, IN 46250. tics. We also prepared an aqueous 18 pmol/L stock solution
2Add correspondence to this author.
Presented at the 37th national meeting of the AACC, Atlanta, of potassium ferrocyanide (the trihydrate; ACS grade) and
GA.,July 1985. added it to Boehringer Mannheim reagents for uric acid,
Received July 8, 1985; accepted December 2, 1985. cholesterol, and triglyceride assays to give final concentra-

518 CLINICALCHEMISTRY,Vol.32, No.3, 1986

 Table 1. Summary of Instrument Methods’ and TypIcal Preclslonb
Uric acid Trlglyc.rldes

InstrumeM Method A, nm#{176} Ref. CV, % A, nm’ CV, % Method A, nm#{176} Ref. CV,%
BMD-8700 3 540/675 6 1.2 600/505 1.4 3 600/505 10 0.8
Hftachi 705 3 660/570’ 6 1.2 600/505 0.7 3 700/505 10 1.0
600/505f
MultistatIII 2 7 3402.0 690/500 5.3 2 340 11 1.4
CentrlfiChem 1 8 2922.4 520 1.3 4 520 12 1.9
Cobas-Blo 2 340 7 0.9 500 0.5 2 340 11 2.0
Ektachem 400 3 670 6 0.9 540 1.9 3 540 10 0.7
acall 1 293 8 1.4 600/540 0.4 3 550 10 1.1
RA-1000 2 340 7 1.2 500 0.9 3 550 10 0.9
‘Methods: 1 = directspectrophotometrlc, 2 = NAD/NADP-coupIed, 3 = peroxldase-coupled,4 = tetrazolium dye-coupled. bAverage CV of triplicate

Downloaded from https://academic.oup.com/clinchem/article-abstract/32/3/518/5652453 by guest on 13 February 2020

determinationsat each bilirubin concentration. “For blchromatlc analysis, the lower numberis the pdmaiy wavelength. dFor all Instruments,the peroxidase-
coupled method BMDreagent. Behring X.L.R. reagent
(.

tions of 18, 18, and 36 pmol/L, respectively. These modified uricase (CentrifiChem, aca H). Bilirubin has no effect on the
reagents were used in the BMD-8700 and Hitachi 705 just uricase reaction, and any endogenous absorption at this
as their unmodified counterparts were. wavelength is corrected with an initial sample blank read-
ing. In three other uric acid procedures evaluated-.Multi-
Results stat ifi, Cobas-Bio, and RA-1000-the uricase reaction is
Figure 1 shows how assay results for uric acid, cholester- coupled to the reduction of NADP, which is monitored at
ol, and triglyceride are affected by the presence of bilirubin 340 nm. Blank corrections are again integral to each
in a serum sample. In general, the peroxidase-coupled procedure; even so, only assay with the Cobas-Bio reagent
assays are the most seriously compromised. Notable excep- was unaffected by bilirubin, the other two showing obvious
tions are the Ektachem procedures, which involve a unique positive spectral bias (Table 2). The values for uric acid
chromogen that produces a reaction intermediate that evi- obtained with the peroxidase-coupled BMD-8700 and Hita-
dently is not susceptible to interference by bilirubin. More- dii 705 tests were decreased by small increases of bilirubin.
over, because the molar absorptivity of the chromophore in In contrast, use of the Behring X.L.R. reagent (which
the Ektachem methods is at least twice that of those used in includes ferrocyanide) on the Hitachi 705 resulted in in-
any of the other analyses, spectrophotometric measure- creasing values for apparent uric acid with increasing
ments can be made at wavelengths where there is no concentrations of biirubin.
biirubin absorption. All of the cholesterol tests studied involve peroxidase-
The uric acid assay system least affected by bilirubin coupled assays; most are affected by chemical interference of
interferences was the direct spectrophotometric measure- bilirubin, except for the aca, Ektachem, and Multistat. In
ment at 292 (or 293) nm after enzymatic oxidation by the aca an aniline rather than a phenol derivative is
oxidatively coupled with aminoantipyrine-an approach
CHOLESTEROL TRIGLYCERIDE URIC ACID a”'”
that, as in the Ektachem, may not afford an opportunity for
.40 bilirubin involvement. We have no explanation for the lack
U of bilirubin interference in the Multistat; however, the
-30 precision of that method in our hands was relatively poor,
z and any interference may have been masked.
2 -20 The triglyceride assays we examined demonstrate trends
4
similar to those observed for uric acid and cholesterol.
>.
Ui S - Ultraviolet-absorbances procedures appear not to be affected
0 #{149} o
Ui C #{149} 0 by bilirubin, whereas peroxidase-coupled methods, except in
0 *
4 0 0#{149} the Ektachem, show chemical interference. A tetrazoliuni
5 0. 0e 0
dye-coupled method (CentrifiChem) was also found to exhib-
2 : o: ____
______ _______
Ui
-20
U U .
Table 2. Billrubin Concentrations (mg/L) at WhIch
& #{149} Interference Becomes Significant’
-30 Instrument Uric acid Cholesterol Triglycerides
BMD-8700 <25 (-) <25 (-) <25 (-)
-40 BMD-8700 + ferrocyanide b 50 (+) 240 (+)
Hitachi705 50(-) 80(-) 80(-)
AB DEFO I J ABC DEFGHI J BCDEFGI K HitachI 705 + ferrocyanide 50 (-) 240 (+) 400 (0)
b
Hitachi 705 Behnngreagent 50 (+) I)
Fig. 1. Percent deviation (average of hiplicate determinations) from MuttistatIII 160 (+) 400 (0) 400 (0)
analyte values obtained for a serum specimen with bilirubin<5 mgL CentrifiChem 400 (0) 50 (-) <25 (-)
(cholesterol 1800 mg/L; trIglycerides 1600 mg/L; uric acid 50 mg/L) to Cobas-Bio 400 (0) <25 (-) 240 (-)
whIch 50(0), 80(0), 160 (a), 240 (#{149}),
320 (#{149}),
or 400 (A) mg of Ektachem400 400 (0) 400 (0) 400 (0)
billrubinwasaddedperliter acall 400(0) 400(0) 400(0)
A.CentrlffChem;B, MultistatIII; C, Cobas-Bio;0, Ektad’iem 400; E. ace II; F. RA- RA-1000 320(+) 240(+) <25(-)
1000; G, BMD8700; H, BMD-8700 + ferrocyanide; I, HitachI705; J, HItachi705 ‘Direction of interferenceIndicated In parentheses. b tested.
+ ferrocyanide;K Hitachi705 wIthBehrlngX.LR.reagent

CLINICALCHEMISTRY, Vol.32,No.3, 1986 519

it some chemical interference from bilirubin.
In the last segment of our study, we examined the use of
ferrocyanide to remove the chemical interference of biiru-
bin from the uric acid, cholesterol, and triglyceride tests of
the BMD-8700 and Hitachi 705, both of which are currently
in operation in our laboratory. Adding ferrocyanide to the
uric acid reagents manufactured for these instruments had
no effect (data not shown). Use of the ferrocyanide-supple-
mented Behring X.L.R. uric acid reagent on the Hitachi 705
also failed to remove the interference; the chemical interfer-
ence was apparently diminished, but underlying spectral
interference then became a factor, as discussed below. A
similar phenomenon was observed for the BMD 8700 choles-
terol test. However, we have had some success with adding
ferrocyanide to cholesterol reagent for the Hitachi 705 and
to the triglyceride reagents for the Hitachi 705 and BMD-
8700 analyzers; in each case, the apparent concentration of
the analyte remained the same despite increasing biirubin
concentrations as compared with unsupplemented reagents.
Discussion
Our studies indicate that there indeed are differences in
response to increases in bilirubin concentrations, not only
between peroxidase-coupled and other methods for mecha-
nized determination of serum uric acid, cholesterol, and
triglycerides, but also among various peroxidase-coupled
methods. The choice of the chromogen determines whether
or not chemical interference from bilirubin will be a factor
and, if so, to what degree supplements such as ferrocyanide
will be useful in its elimination. The triarylimidazole com-
pound in the three Ektachem tests and the aniline deriva-
tive in the aca cholesterol test either do not chemically react
with bilirubin or else other factors in the reaction sequence
are involved to prevent the interference by bilirubin-
factors such as the ifitering layers on the Ektachem slide or
the specific buffers, their pH, and (or) ionic strength used for
the aca. However, all of the other peroxidase-coupled tests
evaluated do involve chroinogens having some reactivity
with bilirubin.
The wavelength chosen for measurement determines the
extent of the contribution of bilirubin to the absorbance
measurement. As shown in Figure 2A, measurements made
0.
Wavelength (iwn)
Fig. 2.Absorbance spectrum after incubatIon of 16 pl. of pooled serum
for 10 mm at 37#{176}C
with800 1iL 01(A) isotonic saline, added bilirubin
100 mg/L; ( Behnng Diagnostics X.LR. uric acid reagent, serum
bilirubin<5 mgfL;(C reagent plus added bliirubin 200 mg/L; (L
reagent plus added bllirubin 200 mg/L and ferrocyanide 18 ixnoUL
Dashedilne Is at 515 rim
520 CLINICALCHEMISTRY,Vol. 32, No. 3, 1986
above about 540 rim will exclude most bilirubin absorption,
but most of the chromophores used in peroxidase-coupled
procedures have absorbance marima below 540 rim; increas-
ing the wavelength to improve specificity will thus yield a
concomitant loss in sensitivity.
Additionally, wavelength selection directly influences the
effect of agents such as ferrocyanide on the problem of
chemical interference from biirubin. Figure 2B shows the
absorbance maximum of the Behring uric acid chromophore
(515 rim) in the presence of a normal concentration of
bilirubin (<5 mg/L). Addition of bilirubin to 200 mg/L
(Figure 2C) produces a negative chemical interference and a
decrease in absorbance. When 18 pmol of ferrocyanide is
Downloaded from https://academic.oup.com/clinchem/article-abstract/32/3/518/5652453 by guest on 13 February 2020
then added per liter (Figure 2D), the chemical interference
is removed and the absorbance is restored to near its
original value. However, not all wavelengths are available
on most multichannel analyzers; e.g., measurements in the
Hitachi 705 for uric acid must be made at 505 nm. As shown
in Figure 2D, the chemical interference of bilirubin as
removed by ferrocyanide is replaced with significant spec-
tral interference at this wavelength.
The peroxidase-coupled methods for triglycerides and
cholesterol are slightly different from those for uric acid.
More chromophore is formed because the concentration of
cholesterol and triglycerides is higher than that of uric acid.
In addition, BMD makes use of a different chromophore, one
with a niirimum absorbance at 500 rim (Figure 3A). The
addition of ferrocyarnde (e.g., 36 Mmol/L for the triglyceride
assay) will partly correct for the chemical interference of
bilirubin (Figure 3C) without contributing to the spectrum.
The optimum wavelength for this chromophore would be
495 nm where the chemical and spectral interference exact-
ly offset each other.
In conclusion, bilirubin interference in peroxidase-cou-
pled assays is a serious problem, one that has been incom-
pletely resolved in commercial assays. The Ektachem meth-
ods involve a chromogen that reacts with hydrogen peroxide
but not with biirubin to produce a highly absorptive species
that can be measured at a wavelength at which the absor-
bance of bilirubin is small. With other chromogens, ferrocy-
snide can eliminate the chemical interference from biliru-
bin, but only if spectrophotometry is done at optimal wave-
lengths-a possibility precluded by the fixed configurations
of many automated analyzers. Other approaches proposed
for eliminating these problems include pretreating samples
with bilirubin oxidase, an enzyme that catalyzes the oxida-
Wavelength (rim)
Fig.3.Absorbance
spectrum of101L ofpooledserumIncubated for10
mmat 37#{176}C
wIth1000 $L of BMDtriglyceridereagent
(A) serum billrubin<5 mg/I., reagentwithoutferrocyanide; ( added bilIrubIn,
200 mg/I.. reagent without ferrocyanide; (C) added bilIntrIn, 200 mg&, with added
ferrocyanlde36 imot& Dashed line is at 505 nm
tive destruction of bilirubin (14), or eliminating peroxidase
itself by reacting directly with a titanium(1V) complex (15).
These alternatives must be investigated further, to provide
the clinical laboratory with analytical procedures that are
free from the chemical and spectral interferences of biliru-
bin.
We thank Gregory Buffone, Ching Ou, Thomas Spiliman, and
Loretta Chow for their assistance in this work. We also thank
Boehrunger Mannheim Diagnostics, Eastman Kodak, Roche Diag.
nostics, and Behring Diagnostics for generously supplying the test
reagents used in this study.
References
1. Witte DL, Brown LF, Feld RD. Effects of bilirubin on detection of
hydrogen peroxide by use of peroxidase. Clin Chem 197824:1778-
82.
2. Foesati P, Prencipe L La reazione di Emerson-Trinder studio
del van cromogeni e analisi delle principali interferenze. Quad
Sclavo Diagn 1978;14:164-77.
3. Glick MR. Ryder KW, Jackson SA. Comparisons of clinical
chemistry systems: response to interfering substances as shown by
analyte-specific “interferograms” [Abstract]. Clin Chem
1985;31:1015-6.
4. Foeaati P, Prencipe L, Berti G. Use of 3,5-dichloro-2-hydroxy-
benzenesulfonic acid/4-axninophenazone chromogenic system in di.
CLIN.CHEM.32/3, 521-522 (1986)
Rapid Determinationof Ironin Urine,in the Presenceof Deferoxamine,by
InductivelyCoupledPlasma EmissionSpectrometry
Paul Leflon and Roger Plaquet
Usinga spectrometerwith an argon plasma source coupled
to a high-frequency
magneticfield,we developed a direct
method for determining iron in urineof patientsbeingtreated
with deferoxamine. Thedetectionlimitfor ironwas 75 nmol/L;
added ironwas satisfactorilyrecovered; and we observedno
interferencefrom deferoxamine at its most commonlyused
concentrations.Values for between-run andwithin-runpreci-
sion (CV) was <5%. Correlation of results with those ob-
tained with a colorimetricmethodinvoMng bathophenanthro-
line was good (r = 0.96).
Iron is usually determined in the urine of hemochromato-
sis patients who are being treated with deferoxarnine, to
monitor the elimination of iron from their bodies. Van
Stekelenburg et al. (1) established that colorimetric meth-
ods based on 2,4,5-tripyridyl-s-triazine (TPTZ) or bathe-
phenanthroline are not suitable for use with such patients
because of interference with the reagents by the deferoxa-
mine medication. For this reason they replaced it by a
ferrozine method, but it requires a 2-h pre-incubation at
37#{176}C
in the presence of citric acid in combination with
ascorbic acid and thiourea.
Laboratoire de Biochirnie, Centre Hospitalier Universitaire, 1,
Place Victor Pauchet, 80030 Amiens Cedex, France.
Received July 16, 1985; accepted December 9, 1985.
rect enzymic assay of uric acid in serum and urine. Ibid.
1980;26:228-.31.
5. Laessig RH, Schwartz TI!, Paskey TA. Determination of SMA.
Mark X set points for bilirubin by standard addition technique.
Ibid. 1972;18:48-51.
6. Thvedi R, Rebar L, Berta E, Stong L. New enzymatic method for
serum uric acid at 500 nra. Ibid. 1978;24:1908-11.
7. Haeckel R. The use of aldehyde dehydrogenase to determine
H202-producing reactions: 1. The determination of the uric acid
concentration. J Clin Chem Clin Biochem 1976;14:101-7.
8. Praetorius E, Poulsen H. Enzymatic determination of uric acid
with detailed directions. Scand J Clin Invest 1953;5:273-80.
9. Allain CC, Poon LS, Chan CSG, Richmond W, Fu PC. Enzymatic
determination of total serum cholesterol. Cliii Chem 197420:470-5.
Downloaded from https://academic.oup.com/clinchem/article-abstract/32/3/518/5652453 by guest on 13 February 2020
10. Spayd RW, Bruschi B, Burdik BA, et al. Multilayer ifim
elements for clinical analysis: applications to representative chemi-
cal determinations. Ibid. 197824:1343-50.
11. Bucolo G, David H. Quantitative determination of serum
triglycerides by the use of enzymes. Ibid. 19’73;19:476-82.
12. Megraw RE, Dunn DE, Biggs HG. Manual and continuous-flow
colorunetry of triacylglycerols by a fully enzymatic method. Ibid.
1979,25:273-8.
13. Grossman SI!, Mollo E, Ertingshausen G. Simplified, totally
enzymatic method for determination of serum triglycerides with a
centrifugal analyzer. Ibid. 1976;22:1310-3.
14. Artiss JD, McEnroe RJ, Zak B. Bilirubin interference in a
peroxidase-coupled procedure for creatinune eliminated by bilirubin
oxidase. Ibid. 1984;30:1389-92.
Alternatively, we describe here the first proposal of a
direct emission-spectrornetric method in argon plasma, in-
volving an instrument routinely in use in our laboratory for
determining aluminum in biological fluids from hemodia-
lyzed patients.
We optimized the conditions of iron measurement and
determined the lack of interference by added deferoxamine,
as well as the repeatability and reproducibility of the
method and its correlation with a colorirnetric technique
when no deferoxamine is present.
Materials and Methods
Apparatus. We used a Model J.Y. 38 P spectroanalyzer
with a Czerny Turner 1-rn (focal length) monochromator,
which includes a holographic grating ruled at 2400 grooves
per millimeter, and a “Plasmatherrn” source inductively
coupled to a high-frequency (27.12 MHz) magnetic field (1.5
kW maximum power). The sample solution was introduced
into the argon plasma by use of a pneumatic nebulizer. (The
whole equipment was supplied by Jobin & Yvon, Instru-
ment S.A., 91163 Longjumeau Cedex, France.)
Reagents: Deferoxamine (“Desferal”; Ciba, Base!, Switzer-
land); argon (argon 93”; Cornpagnie Francaise des produits
Oxyg#{233}nes,
Rouen, France); standard iron solution (Fitri-
aol”; Merck, Darmstadt, F.R.G.).
Prepare the iron stock solution by diluting a volume of
CLINICALCHEMISTRY,Vol. 32, No. 3, 1986 521