Journal of Analytical Toxicology, Vol.

26, March 2002

Determination of Methemoglobin and Total

Hemoglobin in ToxicologicalStudiesby Derivative
Spectrophotometry
A. Cruz-Landeira, M.J. Bal, O. Quintela, and M. t6pez-Rivadulla
Forensic Toxicology Service, Legal Medicine Institute, University of Santiago of Compostela, San Francisco s/n, 15705.Santiago
of Compostela, Spain

Abstract ]
A method for the determination of methemoglobin in the
presence of other hemoglobin subforms (i.e., oxy-, deoxy-, and

carboxyhemoglobin) by use of derivative spectrophotometry is

proposed. The method, which uses the first-derivative of the
spectrum at 645 nm, is straightforward and expeditious, so it is of
special interest to forensic toxicology laboratories as it allows the
simultaneous determination of the methemoglobin saturation
percentage and the hemoglobin concentration. This facilitates
interpretation of the results and provides a better understanding of
the significance of methemoglobin saturation in specific cases.
Based on an analysis of interferences, the presence of other
hemoglobin subforms or of endogenous components of plasma
does not detract in any way from the performance of the method.

quently found in blood (i.e., oxyhemoglobin,deoxyhemoglobin,

and carboxyhemoglobin).The method uses the first derivative
of the absorption spectrum for the sample. Potential interferences from other frequently reported components including
bilirubin, lipids, and Methylene Blue (a therapeutic agent for
methemoglobin) were investigated. The method allows one to
determine the methemoglobin saturation percentage and the
total amount of hemoglobin, which facilitates interpretation of
the analytical results.

Experimental
Apparatus

Introduction
Methemoglobinemia is a pathological condition in which
iron in hemoglobin is in its tervalent oxidation state (Fe3
rather than its divalent one (Fe2 this results in the formation
of a hemoglobin subform that is unfit for transporting oxygen.
Methemoglobinin blood can originate from in vivo exposure to
oxidants (1-3), which gives rise to a variably serious picture of
tissue hypoxia, from heating of blood (e.g., in charred bodies)
(4) or even from specific storage conditions (e.g., refrigeration, freezing) (5-8).
In any case, determinations of methemoglobin are relatively
commonplace and useful to toxicological laboratories, so a
need exists for simple and reliable analytical methods for the
identification and quantitation of this hemoglobin subform in
blood. A large number of methods for this purpose have so far
been reported, the most common of which involve spectrophotometric multicomponent analysis (9-11).
This paper reports a method for the determination of methemoglobin in the presence of the hemoglobin subforms most fre-

A PerkinElmer Lambda 2 ultraviolet-visible (UV-VIS)spectrophotometer interfaced to a computer running the software

PECCS, which affords direct acquisition of analytical data and
storage of spectra, was used. Glass cuvettes of 1-cm light path
and distilled water as sample blank were employed. An Izasa
STKS autoanalyzer was used to determine hemoglobin.

Samples
The samples studied consisted of fresh blood collected in
Venojecttubes (Terumo)containing EDTAas anticoagulant. All
were obtained from the BiochemistryLaboratoryof the Galician
General Hospital.

Reagents
The reagents used included pure oxygen, carbon monoxide,
and nitrogen (all obtained from Air Liquide);a 13% hemoglobin
standard supplied by the Biochemistry Laboratory of the Galician General Hospital; a 5-mg/dL standard of bilirubin from
Bayer; potassium ferricyanide [K3Fe(CN)6],sodium dithionite
(Na2S204), crystalline sodium nitrite (NaNO3),Methylene Blue
for microscopy (16316), 37% hydrochloric acid, and sodium
hydroxide from Merck.

Reproduction(photocopying)of editorial content of this journal is prohibited without publisher'spermission.

67

Journal of AnalyticalToxicology,Vol. 26, March 2002

moglobin in the presence of other hemoglobin subforms. Finally, the quantitative study in the first-derivative mode (D1) ,
between 700 and 500 nm, was done.

Methods

Preparationof standards
Oxyhemoglobin(OxyHb),deoxyhemoglobin(DeoxyHb),carboxyhemoglobin (COHb), and methemoglobin (MetHb) standards were prepared from a stock standard containing 13 g% of
hemoglobin. Followingdilution to 1% (v/v) in distilled water, a
100% saturated solution of oxyhemoglobin was obtained by
bubbling oxygen for 10 min, excess O2 being removed by bubbling N2for 5 rain. The resulting solution was used to make the
following four aliquots: aliquot A, which consisted of oxyhemoglobin; aliquot B (deoxyhemoglobin),which was obtained by
adding sodium nitrite; aliquot C (carboxyhemoglobin),which
was prepared by bubbling CO for 5 rain, followedby N2 to remove excess gas from the solution; and aliquot D (methemoglobin), which was obtained by oversaturation with
potassium ferricyanide.

Recordingof spectra
Absorbance and derivative spectra (D1, D2, and D3) for the
pure standards were recorded in order to determine the best
spectrophotometric conditions for the determination of methe-

Table I. Compositionof SolutionsContaining Increasing

Amounts (%) of MetHb
MetHb
%
0
2
4
6
8
10
20
40
60
80
100

[Hb] in the
solution
(mg/mL)

Vol Standard A
(100% MetHb)
(mL)

Vol StandardB
(100% OxyHb)
(mL)

1.3
1.3
1.3
1.3
1.3
1.3,
1.3
1.3
1.3
1.3
1.3

0.0
0.2
0.4
0.6
0.8
1
2
4
6
8
10

10
9.8
9.6
9.4
9.2
9
8
6
4
2
0

Table II. Compositionof SolutionsContaining Variable

Concentrations of Hemoglobin (MetHb)

MetHb
%
100
100
100
100
100
I00
100
100
100

68

[Hb]
in the
sample
(g/dL)
2.6
5.2
7.8
10.4
13.0
15.6
18.2
20.8
23.4

[Hb] in the
solution(after
Vol.
1% dilution) StandardC
(mg/mt)
(mL)
0.26
0.52
0.78
1.04
1.30
1.56
1.82
2.08
2.34

1
2
3
4
5
6
7
8
9

Vol.
Distilled
water
(mL)
9
8
7
6
5
4
3
2
1

Procedures
Determination of the methemoglobin saturation percentage.
Following dilution of the 13 g% hemoglobin (Hb) standard to
1% in distilledwater and 100% saturation with oxygen, the followingtwo working standards were prepared: standard A, which
contained 1.3 mg/mL MetHb (followingsaturation with potassium ferricyanide)and standard B, which contained 1.3 mg/mL
OxyHb.
Mixtures of the previous two standards in appropriate proportions gave the solutions listed in TableI, which contained increasing amounts of MetHb but the same concentration ofHb.
Determination of the total hemoglobin concentration. The 13
g% hemoglobinstandard was used to prepare another standard,
C, containing2.6 mg/mL methemoglobin(i.e., 100% MetHb).By
appropriate dilution, the solutions listed in TableII, Whichcontained variable concentrations of MetHb,were prepared.

Study of interferences
The interferences examined were those of the other
hemoglobin subforms, potassium ferricyanide, other plasma
components, and even a therapeutic agent used to treat methemoglobinemia (MethyleneBlue).

Results and Discussion

Distilled water was used as the blood diluting medium because it was found to providestable solutions while resulting in
osmotic lysis of red cells. In addition, aqueous solutions are
easier to handle than those used by other authors (e.g., phosphate buffer, Sterox-E). A 1% dilution was found to be the
most effectivewith a viewto examiningthe spectrophotometric
absorption band on which the Studywas based. Higher dilutions
resulted in decreased absorbance, and lower ones distorted the
band of interest. Other authors (12,13) use dilutions as high as
1:1000; this entails operating at lower wavelengths (405-425
nm), where the hemoglobin molecule exhibits a much higher
absorbance.
Figures 1 and 2 show the absorbance spectra for OxyHb,
COHb, DeoxyHb,and MetHb, and their first derivativesoverthe
range 500-700 nm, respectively. Notwithstanding the strong
overlap of bands in the absorbance spectra (Figure 1), only
MetHb absorbs at 645 nm (D1),with a positive peak that is not
affected by the other components (Figure 2). Some authors
(12) use the band at approximately 400 nm (Soret region) to
quantitate MetHb because its absorbance is 50 times higher
than in the 630 nm region. We discardedthis region becauseof
the strong overlap between the spectra for the different
hemoglobin subforms. In addition, some of the substances
studied (e.g., bilirubin, which exhibits an absorbance peak at
450 nm) might also have interfered in this region.
As can be seen from the spectra of Figure 3 and the data of
Table III, 1D64snm (the quantity used to determine methe-

Journal of Analytical Toxicology, Vol. 26, March 2002

moglobin) was proportional to the MetHb content at a given

concentration of hemoglobin. The corresponding linear regression equation was
Y-- 0.025X + 0.060 (r = 0.9998)
where Yis the ]D64snmvalue andX the MetHb percent content.
Table III. Quantitation of MetHb in Solutions Containing
Increasing Concentrations of this Hemoglobin Subform
MetHb
%
0
2

4
6
8
10
20
40
60
80
100

1D64snm

SD
(n = 5)

CV(%)
(n = 5)

0.0500
0.1229
0.1551
0.1999
0.2641
0.3252
0.5603
1.0753
1.5448
2.0574
2.5710

0.003
0.004
0.010
0.006
0.013
0.0t0
0.013
0.005
0.022
0.012
0.001

5.15
3.44
6.41
2.75
4.86
3.10
2.35
0.43
1.42
0.60

The precision of the method was studied at three different

MetHb levels (i.e., 20, 55, and 100%). The coefficient of variation never exceeded 2.92 (n = 25).
The limit of detection could not be determined as all samples
contained some methemoglobin. Instead, the basal MetHb percentage was determined in 79 samples from blood donors and
found to be 0.5% (SD = 0.0234).
At a given MetHb saturation percentage, peak D 1 was proportional to the hemoglobin concentration. This is clearly apparent from Figure 4 and Table IV, which show the 1D645nm
values for solutions of increasing concentrations of Hb 100%
saturated with MetHb (the solutions in Table II).
The corresponding linear regression equation for the
hemoglobin concentration (g/dL) versus ]D64snmplot is
Y2= 0.1999;(2 + 0.08 (r = 0.9999)
where )'2 is the ]D64snmvalue and)(2 the real hemoglobin concentration (g/dL) in the sample concerned.
The results were compared with those provided by a Coulter
STKS autoanalyzer for 15 fresh blood samples from donors.
Table V lists the hemoglobin concentrations provided by the
proposed method (rightmost column) and those obtained with

0.05

4,MO
Z,4000

l,.f~lm

8.nnan

o,gM
c~

8.5510

"4,8000

6,44N

-Z,4000
O.~ii

9,4,NOt t

~,O

O,MM
Qi.I

m.O

~,0

M,O

nm

rim

Figure 1, Absorbancespectrafor OxyHb (A), COHb (g), DeoxyHb (C), and

MetHb (D).

Figure 3. D1 spectra for solutions containing a constant concentration of

Hb (1.3 mg/mL) and increasing amounts of MetHb (0, 20, 40, 60, 80, and
100%). See the increasing value for the 645 nm positive peak.

'"
5 . i

"1

II.U

t 6 z'm
"ll.U
0.9609

-2.4El

4,~0

-Z.IIN I

SN.II

~.11

.
r.li.O

.
r.,,~,11

780.0

nm

Figure 2. Dn spectra for OxyHb (A),COHb (B), DeoxyHb(C), and MetHb

(D),

Figure 4. D 1 spectra for solutions containing increasing concentrations of

hemoglobin as MetHb (100% saturation).

69

Journal of Analytical Toxicology, Vol. 26, March 2002

the method used by the Clinical Analysis Laboratory (central

column). As can be seen from Figure 5, the results of both
methods are linearly correlated.
The proposed method allows one not only to determine the
proportion of methemoglobin but also to calculate the real
concentration of hemoglobin in each sample and hence predict
the functional implications of such a proportion in specific
cases. This can be used to determine total hemoglobin in toxicological laboratories, where an autoanalyzer is rarely available.
In addition, the proposed method surpasses the standard
methods based on the addition of KCN typically used by clinical
laboratories; in fact, it avoids the use of this reagent and hence
its toxic effects.
In practice, processing an unknown sample by using the proposed method involves the step sequence depicted in Scheme I.
The proportion of MetHb, X, can be calculated from a simple
rule of three:

The hemoglobin concentration in the working solution can

be calculated by substituting the value of B into Eq. 2 (Y2 =
0.1999)(2 + 0.008):
B = 0.19992(2 + 0.008
)(2 (g/dL) = (B - 0.008) / 0.1999

For example, if the "A value" for an unknown is 0.7890, and

after the saturation of the sample with potassium ferricyanide
the "B value" (the 100% of metHb for the sample) is 2.650,
then:
%MetHb = (0.7890 x 100)/2.650 = 78.90/2.650 = 29.7% of
MetHb in the sample.
[Hb] g/dL = (2.650 - 0.008)/0.1999 = 2.642/0.1999 --- 13.2
g/dL of Total Hb in the sample.

~'
Table IV. Quantitation of MetHb in Solutions Containing
Increasing Concentrations of Hemoglobin
(Methemoglobin)
[Hb] in the [Hb] in the
MetHb sample(real) solution
%
g/dL
mg/mL
100
100
100
100
100

100
100
100

1O0

2.6
5.2
7.8
10.4
13.0
15.6
18.2
20.8
23.4

0.26
0.52
0.78
1.04
1.30

1.56
1.82
2.08

2.34

1D64snm
0.5145
1.0552
1.5324
2.0698
2.5704
3.1191
3.6265
4.1623
4.6744

SD
(n = 5)

CV (%)
(n = 5)

0.0096
0.0121
0.0160
0.0217
0.0199

1.87
1.15
1.04
1.05
0.53

0.0232

0.74

0.0292
0.0218

0.80
0.52

0.0207

0.44

Table V. Comparisonof the ResultsProvidedby the

ProposedD1-Methodand the Coulter Method
Sample

[Hb] D1.Method

[Hb] Coulter

1D645nm

(g/dL)

(g/dL)

1
2
3
4
5

1.3472
1.5786
1.9967
1.9580
2.0454

6.7
7.9
9.9
9.8
10.2

6.4
7.6
9.8
10.3
10.5

6
7
8
9
10

1.9533
2.2936
2.4561
2.9429
2.7879

9.7
11.4
12.2
14.7

11.1
12.2
13.6
14.4

11
12
13
14
15

2.8969
2.8549
2.9127
2.9871
3.1909

13.9
14.5
14.2
14.5
14.9
15.9

14.5
14.8
14.8
15.2
15.5
16.2

NO

70

(Eq

18

% MetHb (X) = 100 xA/B

y = 0.9634X
r = 0.9841

16

.:3:. 8
~10

=/

0/0

6
6.0

8.0

10.0

12.0

14.0

16.0

18.0

[Hb] Coulter (g/dL)

Figure5. Correlation of the results obtained in the determination of the
hemoglobin concentration usingthe Coulter STKSmethod and the proposed
(DI) method.

Dilute 0.1 mL of the unknown blood sample

until 10 mL with distilled water

Record D1 spectrum
between 700-500 nm
MeasureID64snm ~

Saturatethe
dilutionto 100% ~
with 02

Saturateto 100%
with Potassium
Ferricyanide

Avalue ]
Record DI spectrum
between 700-500 nm
Measure 1D64snm

SchemeL Procedure used to calculate the MetHb and Hb contents in a

real sample.

Journal of Analytical Toxicology,Vol. 26, March 2002

Interferences

Other hemoglobin subforms. Any spectral interference from

carboxyhemoglobin,oxyhemoglobin,or deoxyhemoglobinwas
discarded earlier in establishing the most suitable operating
conditions. As can be seen from Figure 2, which shows the D1spectra for these hemoglobin subforms and that for methemoglobin, the spectra for COI-Ib,OxyHb,and DeoxyHbdo not
interfere with the determination of MetHb based on the first
derivative of the spectra in the region of interest (645 nm).
This was why parameter 1DB45nmwas chosen to quantitate
MetHb.
Potassium ferricyanide. Becausepotassium ferricyanidegives
colored aqueous solutions, it is essential to obtain their spectra
in order to discard potential interferences with the measurement of MetHb as it is used to saturate samples in the proposed
method. A potassium ferricyanide solution at the usual concentration (10 mg/mL) did not interfere with the determination
of MetHb as its Dl-spectrum was fiat in the measuring region
(645 nm).
Bilirubin. In order to determine the potential interference of
bilirubin with the determination of MetHb from D], a solution
containing a bilirubin concentration equivalentto that in blood
from a patient with severe jaundice (22 mg/dL) was prepared
and its spectrum recorded. Although the D1 spectrum exhibits
a relativelystrong peak in the band at 500 nm (specifically,at
approximately480 nm), its is virtually fiat between 500 and 700
nm. We quantitated the potential interference by adding increasing concentrations of bilirubin from 0 to 50 mg/dL to a
blood samplewith a basal MetHb concentration of 0.5%. The interference was found to be quite weak at bilirubin concentrations up to 40 mg/dL, which increased the MetHb saturation
percentage to 0.8%. Abovesuch a concentration, the interference was substantial (e.g., it raised the MetHb saturation percentage to 1.2% at 50 mg/dL). According to Zwart et al. (10),
the strongest interferences are observed in blood with a high
proportion of Hb F, which often contains a high enough
bilirubin concentration to simulate an appreciable proportion
of MetHb. The difficultyof obtaining specimensof this type prevented us from confirming this assertion except in fetal blood,
which exhibited no signs of interference.
Methylene Blue. It was important to establish the potential
influence of Methylene Blue on the determination of the
hemoglobin subforms studied as this dye is the treatment of
choice for methemoglobinemia,so the need frequentlyexists to
measure one substance in the presence of the other. To some
authors (13), the presence of MethyleneBlue can result in significant errors in the conventional spectrophotometric determination of MetHb. The D1 spectrum for an aqueous solution
of MethyleneBlue at over the wavelength range 500-700 nm
was foundto exhibit a maximum at 679 nm and a minimum at
645 nm that bore a more or less constant relationship to each
other. This suggests that Methylene Blue interferes severely
with the determination of methemoglobin. However, taking
into account that therapeutic dose of the dye is 1-2 mg/kg, its
concentration will normally lie below28 mg/mL.No significant
interference was observed over the concentration range tested
(0-50 mg/mL) as the increase in the MetHb saturation percentage never exceeded 0.2%.

Effect of turbidity. In order to determine the potential influence of turbidity on the determination of methemoglobin,
serum with a high triglycerideconcentration was supplied to diluted blood at increasing concentrations from 0 to 3075 mg/dL.
Interferences were found to be significant above a triglyceride
concentration of 800 mg/dL---equivalent of a high sample
opacity--which caused the percent MetHb saturation to rise to
1%; in fact, a 3000-mg/dL concentration simulated a MetHb
saturation of 5.5%. We should note, however, that normal
triglyceride levels in blood are below 160 mg/dL and that concentrations above 300 mg/dL are definitely pathological.

Conclusions
The proposed spectrophotometric method, based on the first
derivative of the spectrum at 645 nm, allows the determination
of the methemoglobin saturation percentage and the total
hemoglobin concentration in the sample. This affordsbetter interpretation of the analytical results because it allows the absolute amount of functional hemoglobin in an individual to be
determined. Basedon the results, neither the presence of other
hemoglobin subforms nor that of endogenous plasma components (bilirubin and triglycerides) or even therapeutic agents
such as Methylene Blue interferes with the proposed method.

References
1. S.M. Bradbery, R.M. Whittington, D.A. Parry, and J.A. Vale. Fatal
methemoglobinemia due to inhalation of isobutyl nitrite. Clin.
Toxicol. 32(2): 179-184 (1994).
2. P.M. Wax and R.S. Hoffman. Methemoglobinemia: an occupational hazard of phenyl-propanolamine production. Clin. Toxicol. 32(3): 299-303 (1994).
3. C.L. French, S.S. Yaun, L.A. Baldwin, D.A. Leonard, X.Q. Zhao,
and E.J. Calabrese. Potency ranking of methemoglobin-forming
agents. J. Appl. Toxicol. 15(3): 167-174 (1995)
4. G.P. Fechner and D.J. Gee. Study on the effects of heat on blood
and on the post-mortem estimation of carboxyhaemoglobin and
methaemoglobin. Forensic Sci. Int. 40:63-67 (1989).
5. K. Sato, K. Tamaki, H. Okajima, and Y. Katsumata. Long-term
storage of blood samples as whole blood at extremely low temperatures for methemoglobin determination. Forensic Sci. Int. 37:
99-104 (1988).
6. K. Sato, K.Tamaki, H.Tsutsumi, H.Okajima, and Y. Katsumata.
Storage of blood for methemoglobin determination: comparison of
storage with a cryoprotectant at-30~ and without any additions
at -80~ or -196~ Forensic Sci. Int. 45:129-134 (1990).
7. I. Uchida, C. Tashido, Y.H. Koo, T. Mashimo, and I. Yoshiya. Carboxyhemoglobin and methemoglobin levels in banked blood.
J. Clin. Anesth. 2:86-90 (1990).
8. G.L. Moore, A. Zegna, M.E. Ledford, J.P. Huling, and R.M.
Fishman. Evaluation of methemoglobin formation during the
storage of various hemoglobin solutions. Artif. Organs 16(5):
513-518 (1992).
9. J.J. Mahoney, H.J. Vreman, D.K. Stevenson, and A.L. Van Kessel.
Measurement of carboxyhemoglobin and total hemoglobin by
five specialized spectrophotometers (CO-oximeters) in comparison
with reference methods. Clin. Chem. 39(8): 1693-1700 (1993).
10. A. Zwart, E.J.Van Kampen, and W.G. Zijlstra. Results of routines

71

Journal of Analytical Toxicology,Vol. 26, March 2002

determination of clinically significant hemoglobin derivatives by
multicomponent analysis. Clin. Chem. 32(6): 972-978 (1986).
11. M.L. Shih, W.D. Korte, and C.R. Clark. Multicomponent spectroscopic assay for hemoglobin and ferrihemoglobin species in
methemoglobin treatment of cyanide poisoning. J. Anal. Toxicol.
21:543-547 (1997).
12. A. Taulier, R Levillain, and A. Lemonnier. Determining methemoglobin in blood by zero-crossing-point first derivative spec-

72

trophotometry. Clin. Chem. 33(10): 1767-I 770 (1987).

13. J. Dotsch, S. Demirakca, D. Hahn, N. Katz, P.G. Kuhl, and W.
Rascher. Accuracy of methemoglobin measurements: comparison
of six different commercial devices and one manual method. Crit.
Care Med. 27(6): 1191-I 194 (I 999).
Manuscript received August 14, 2000;
revision received January 22, 2001.