J. Agric. Food Chem.

1994, 42, 1931-1937 1931

Rapid, Sensitive, and Specific Thiobarbituric Acid Method for

Measuring Lipid Peroxidation in Animal Tissue, Food, and
Feedstuff Samples
Nickos A. Botsoglou,' Dimitrios J. Fletouris,*r$ Georgios E. Papageorgiou,§
Vassilios N. Vassilopoulos,t Antonios J. Mantis,$and Antonios G. Trakatelliss
Laboratory of Nutrition and Laboratory of Milk Hygiene and Technology, School of Veterinary Medicine,
and Laboratory of Biological Chemistry, School of Medicine, Aristotle University,
54006 Thessaloniki, Greece

A rapid aqueous acid extraction thiobarbituric acid method for measuring malondialdehyde as a
marker of lipid peroxidation in animal tissue, food, and feedstuff samples has been developed.
Sample is homogenized with aqueous trichloroacetic acid in the presence of hexane and butylated
hydroxytoluene, and the homogenate is centrifuged. Following reaction with thiobarbituric acid
reagent, malondialdehyde is directly quantified on the basis of the third-derivative absorption
spectrum of the pink complex formed. Further purification is not required because the derivative
transformation of the conventional analytical band at around 532 nm virtually eliminates spectral
interferences arising from other compounds. The effect of antioxidants and the optimum conditions
for the reaction have been established, and the analytical performance of the new method has been
evaluated. The applicability of the method on various animal tissue, food, and feedstuff samples
has been also tested. Owing to its simplicity and increased sensitivity and specificity, the method
may be preferred over other methods for estimating the extent of lipid peroxidation.

Keywords: Thiobarbituric acid method; derivative spectrophotometry; lipid peroxidation; malon-

dialdehyde

INTRODUCTION with TBA followed by separation of the pink complex

produced (Turner et al., 1953;Yu and Sinnhuber, 1957;
Monitoring and control of lipid peroxidation in food Sinnhuber and Yu, 19581, (b) by distillation of the
are of great importance due to increasing demand for sample followed by reaction of the distillate with TBA
good quality products. Lipid peroxidation, the oxidative (Tarladgis et al., 1960; Rhee, 1978; Yamauchi et al.,
deterioration of the polyunsaturated lipids of food, leads 1982; Ke et al., 19841, (c) by extraction of the lipid
through formation of hydroperoxides to short-chain portion of the sample with organic solvents and reaction
aldehydes, ketones, and other oxygenated compounds of the extract with TBA (Pikul et al., 1983, 1989), and
which are considered to be responsible for the develop- (d) by extraction of MDA using aqueous trichloroacetic
ment of rancidity in stored foods (Gray, 1978; Melton, (Witte et al., 1970; Sinnhuber and Yu, 1977; Siu and
1983; Wong, 1989) and related to experimental heart Draper, 1978; Newburg and Concon, 1980) or perchloric
disease, cancer, and aging in animals (Chio and Tappel, (Salih et al., 1987; Pikul et al., 1989) acid and reaction
1969; Kaunitz and Johnson, 1973; Cutler and Hayward, with TBA.
1974).
Although the distillation method is the mast fre-
Malondialdehyde (MDA),a major degradation product quently used procedure and may be regarded as the
of lipid hydroperoxides, has attracted much attention standard method for MDA analysis, it is more cumber-
as a marker for assessing the extent of lipid peroxidation some and requires more time than the aqueous acid
(Raharjo and Sofos, 1993). The compound is of particu- extraction method. Furthermore, heating during distil-
lar concern since it has been shown t o be mutagenic lation enhances the degradation of existing lipid hydro-
(Basu and Marnet, 1984) and carcinogenic (Shamberger peroxides and, thus, additional MDA and other TBA-
et al., 1974) and implicated in other pathological pro- reactive substances (TBARS) may be formed even in the
cesses such as the formation of fluorescent pigments presence of metal chelators or phenolic antioxidants
typical of cellular aging (Bidlack and Tappel, 1973;
(Gutteridge and Quinlam, 1983; Raharjo and Sofos,
Trombly and Tappel, 1975).
1993). The aqueous acid extraction method is also
The most common method for measuring MDA in food preferred by many workers because it is simple and
products and biological samples seems to be the thiobar- gives results that are highly correlated with those of
bituric acid (TBA) test, which is based on spectropho- distillation (Pikul et al., 1989) and sensory evaluation
tometric quantitation of the pink complex formed after (Salih et al., 1987) methods. In general, however, all
reaction of MDA with two molecules of TBA. The TBA versions of the TBA test have been criticized as being
test can be performed (a) by directly heating the sample insensitive for the detection of low levels of MDA and
nonspecific (Hackett et al., 1988). Other TBARS can
* Author to whom correspondence should be ad- interfere with the analysis, overestimating the results
dressed. (Marcuse and Johansson, 1973; Patton, 1973; Kosugi
Laboratory of Nutrition. et al., 1987).
$ Laboratory of Milk Hygiene and Technology. To eliminate interferences, Bird et al. (1983) used
5 Laboratory of Biological Chemistry. high-performance liquid chromatography (HPLC) for
0021-8561/94/1442-1931$04.50/0 0 1994 American Chemical Society
1932 J, Agric. Food Chem., Vol. 42,No. 9, 1994 Botsoglou et al.

analyzing MDA in food and feed samples. Other work-

ers applied an additional cleanup step prior to HPLC
(Draper and Hadley, 1990; Squires, 1990) or spectro-
photometric (Raharjo et al., 1992, 1993) determination
using solid-phase extraction (CIScartridge) to remove
interfering substances from the reaction mixture. These
procedures can provide increased specificity and sensi-
tivity over the conventional TBA methods, but they are
rather time-consuming and not practical for routine
analysis. Therefore, a simple, rapid, sensitive, and
specific method for measuring the extent of lipid per-
oxidation could be of value.
An interesting alternative to conventional spectro-
photometric methods, in cases where spectral interfer-
ences obscure the analytical band, is derivative spec- -0.02
trophotometry (O’Haver and Green, 1976). This tech-
nique offers, in some instances, improved specificity and --- -, I 0.0
sensitivity without any need for lengthy sample clari- 450 500 550 600
fication (Botsoglou et al., 1993; Fletouris et al., 1993). Wavelength, nm
In this study, the performance of derivative spectropho-
tometry in developing a new, more reliable method for Figure 1. Normal (- - -,right Y-axis)and third-derivative (--,
left Y-axis)spectra of MDA-TBA reaction mixtures containing
the determination of MDA in animal tissue, food, and 31.9 (a) and 2.4 ng of MDMmL (b).
feedstuff samples has been investigated.
the reaction mixture was submitted to third-derivative spec-
MATERIALS AND METHODS trophotometry against blank reaction mixture.
Determination. Aliquots of standard solutions were pi-
Instrumentation. A Shimadzu Model W-160A double- petted into screw-capped tubes and diluted to 2.5-mL volume
beam spectrophotometer with 1-cm absorption cells was used with 5% TCA. A 1.5-mL volume of 0.8% TBA was added in
for all measurements. Third-derivative spectra were produced each tube, and the reaction was carried out as prescribed.
by electronic differentiation (d3A/dd3,where A is the absor- Calibration curves were constructed by plotting values of peak
bance and 1is the wavelength) of the normal spectra obtained height a t 521.5 nm, as they are printed on the instrumental
a t a scanning speed of 480 n d m i n , using a derivative chart in arbitrary units, versus known concentration of MDA
difference ( M )setting of 21 nm. An Ultra-Turrax blender in the final reaction mixtures. The concentration of MDA in
(Janke & Kunkel, GmbH, Germany), a Centra-MP4 centrifuge sample extracts was calculated on the basis of the slope and
(IEC), a vortex blender (Heidolph, Germany), and a thermo- intercept data of the computed least-squares fit of calibration
stated water bath, type 3044 (Kottermann, Germany), were curve. In case the absorbance value exceeded 1.0, sample
also used for sample treatment. extract was appropriately diluted with water before final
Reagents. Most chemicals and solvents used in this study measurement. MDA was determined in samples using the
were of ACS reagent grade. Butylated hydroxytoluene (BHT), formula
2-thiobarbituric acid (TBA), ethylenediaminetetraacetic acid
(EDTA), ascorbic acid (AA),bovine serum albumin (BSA),and MDA content, ppb = 16CV/W
the malondialdehyde precursor, 1,1,3,3-tetraethoxypropane
(TEP), were all obtained from Sigma Chemical Co. (St. Louis, where C is the MDA concentration (ng/mL) in the sample
MO), whereas trichloroacetic acid (TCA) was from Merck extract according to the calibration curve, V is the dilution
(Germany). factor of sample extract, if any, and W is the weight (g) of the
Preparation of MDA Standards. A quantity (73.2 mg) sample.
of TEP was accurately weighed in a screw-capped test tube,
diluted with 10 mL of 0.1 N HC1, immersed into a boiling water RESULTS AND DISCUSSION
bath for 5 min, and quickly cooled under tap water. Stock
solution of MDA (239 pg/mL) was prepared by transferring The performance of derivative versus normal spec-
the hydrolyzed TEP solution into a 100-mL volumetric flask trophotometry in the identification of the MDA-TBA
and diluting t o volume with water. Working MDA solution complex is shown in Figure 1. When an MDA-TBA
(2.39 pg/mL) was prepared by pipetting a 1-mL aliquot of the reaction mixture containing 31.9 ng/mL standard MDA
stock solution into another 100-mL volumetric flask and is submitted to third-derivative spectrophotometry, a
diluting to volume with water.
series of peaks and troughs appears that may be used
Animal Tissue, Food, and Feedstuff Samples. Cow,
pork, lamb, and chicken muscle, liver and kidney tissues, to locate and resolve the analytical band; the peak at
sardine, mackerel, and cod fish tissues, meat meal, fish 521.5 nm and the trough at 544 nm correspond t o the
(herring) meal, fish meal, soybean meal, infant milk formula, inflection points of the conventionally recorded analyti-
unsweetened condensed milk, blue cheese, feta cheese, kasseri cal band, while the zero-crossing observed a t 532 nm
cheese, processed cheese, coffee cream powder, and butter were represents the absorption maximum (Figure la). The
all purchased from a local market. Rat muscle and kidney appearance of such extremes improves the resolution
tissues were from the experimental animals of our laboratories. so that, while the normal spectrum of a further diluted
Analytical Procedure. A 2-g sample was transferred into (2.4 ng/mL) solution gives absolutely no information, its
a 25-mL centrifuge tube, and volumes of 5% aqueous TCA (8 derivative spectrum really does (Figure lb).
mL) and 0.8% BHT in hexane (5 mL) were successively added. Research based on normal spectrophotometry has,
The content of the tube was Ultra-Turraxed for 30 s at high already, pointed out that temperature, time of heating,
speed and centrifuged for 3 min a t 3000g, and the top hexane
layer was discarded. The bottom aqueous layer was made to and pH of the reaction mixture are all critical param-
10-mL volume with 5 8 TCA, and a 2.5-mL aliquot was eters to the completeness of the MDA-TBA reaction
pipetted into a screw-capped tube to which a volume (1.5 mL) (Raharjo and Sofos, 1993). However, the recommended
of 0.8% aqueous TBA was also added. Following incubation optimum conditions vary greatly among authors. Heat-
for 30 min at 70 “C, the tube was cooled under tap water, and ing the reaction mixture a t 70 or 80 “C for 30 (McCoy
Lipid Peroxidation in Tissue, Food, and Feed J. Agric. Food Cbem., Vol. 42,No. 9,1994 1933

Table 1. Effect of pH upon the Formation of the Witte et al. (1970). Using hexane, some purification of
MDA-TBA Complex the extract could be effected, and the appearance of
heighta of the heighta of the turbidity, the major shortcoming observed in all extrac-
third-derivative third-derivative tion methods for samples high in fat, was eliminated.
peakb at 521.5 nm peakb at 521.5 nm The method of Witte et al. (1970) did not include
DH (arbitrary units) pH (arbitrary units) antioxidants during sample blending, and neither do
~~~

3.1 0.220 1.6 0.885 many aqueous extraction methods in current use (Sham-
2.9 0.298 1.2 1.042 berger et al., 1977; Newburg and Concon, 1980; Bird et
2.7 0.356 1.1 1.047 al., 1983; Csallany et al., 1984; Lee and Csallany, 1987;
2.4 0.509 0.5 1.044
2.1 0.689 Schmedes and Holmer, 1989). Some investigators,
however, have stated that lipid peroxidation may be a
a Average of triplicate analyses. The reaction mixture contains
serious concern unless antioxidants such as BHT, AA,
147.6 ng of MDNmL.
butylated hydroxyanisole, propyl gallate, sodium sulfite,
and tocopherol or chelating agents such as EDTA are
et al., 1988) or 90 min (Siu and Draper, 19781, respec- added before sample processing (Siu and Draper, 1978;
tively, or boiling for 10 (Buttkus and Bose, 1972), 15 Rhee, 1978; Pikul et al., 1983; Ke et al., 1984; Salih et
(Shamberger et al., 19771, 20 (Kosugi et al., 19891, 25 al., 1987, 1989; Squires, 1990). To determine whether
(Braddock and Petrus, 1971), 30 (Salih et al., 1987), 35 autoxidation occurred during sample treatment, various
(Sidwell et al., 1955),40 (Yu and Sinnhuber, 19571, 45 concentrations of selected antioxidants and EDTA were
(Ke et al., 1984),55(Tarladgis et al., 19601, and 60 min tested using 2-days-refrigerated and 2-months-frozen
(Pikul et al., 1989) are all suggested to be optimum stored mackerel and cod samples, respectively. The
heating conditions. On the other hand, some of the hydrophobic BHT and the hydrophilic AA were used as
proposed optimum pH conditions include acidification test antioxidants because the former is considered to
of the reaction mixture at pH values of 3.5 (Ohkawa et be more effective than other commonly used antioxi-
al., 19791, 3.0 (Draper et al., 19841, and 1.5 (Tarladgis dants (Khayat and Schwall, 1983) while the latter,
et al., 1960). Due to this inconsistency, reexamination although antioxidant, exhibits also prooxidant activity
of the optimum reaction conditions through the use of (Benedict et al., 1975).
derivative spectrophotometry was carried out.
Table 4 shows that significant lipid peroxidation
The data presented in Tables 1 and 2 do not fully occurred when samples were blended with TCA. Hex-
support the above view. Lowering the pH of the ane added before blending was effective for reducing
reaction mixture causes the derivative response to further lipid peroxidation, but it had no effect when
arrive at a maximum (Table l), whereas when the added after blending, a finding suggesting that blending
temperature is decreased from 100 to 70 "C, the yield was the major cause of sample autoxidation. On the
of the reaction becomes higher but the reaction time is other hand, addition of EDTA before blending in the
also markedly increased (Table 2). Heating times longer presence of hexane had a more pronounced effect on the
than 30 min at 100 "C or 50 min at 80 "C were found to suppression of lipid peroxidation. This is consistent
result in turbidity development and, consequently, in with previous findings suggesting that the prooxidant
high erroneous values when measurements were made action of some metal ions in foods can be greatly
by normal spectrophotometry. This might account for retarded by the addition of EDTA due t o formation of
the results presented by Tarladgis et al. (1962) and Yu complexes with low reduction potentials favoring an
and Sinnhuber (1964), who also observed spectral antioxidant effect (Richardson and Korycka-Dahl, 1983).
interferences to the TBA reaction during heat treatment Table 4 also shows that AA exhibited a definite prooxi-
in the presence of acids and attributed them to decom- dant action. The univalent reduction of metal ions by
position of the TBA reagent. Derivative processing, AA followed by the reduction of hydroperoxides to yield
however, eliminated this turbid background, and quan- hydroxyl radical may provide one basis for the prooxi-
titation of the MDA-TBA complex could be made dant function of AA (Kanner et al, 1977; Kanner, 1994).
possible even in such turbid solutions. However, combinations of AA and metal ions can be
Summarizing the data of Tables 1 and 2, it can be prooxidant or antioxidant depending upon their relative
assumed that a pH value lower than 1.2 and a heating concentrations (Kanner et al., 1977). A number of
time of 30 min at 70 "C might be the optimum conditions possible mechanisms have been suggested for this
for MDA-TBA reaction. Regression analysis of the data apparent paradox, including reductive activation of
obtained by running a series of working MDA solutions metal ions, increased levels of the supposed prooxidant
showed the response to be linear in the range examined ascorbyl radical, and formation of unspecified ascor-
(Table 3). Under the mentioned conditions, as low as bate-metal complexes that may differentially affect
2.6 ng of MDNmL of reaction mixture could be precisely propagation and termination reactions of lipid oxidation
measured on the basis of the height of the third- (Richardson and Korycka-Dahl, 1983; Kanner, 1994).
derivative peak at 521.5 nm (Table 3). Considering Numerous studies in this area have not succeeded in
sample size and total extract dilution during the ana- defining the complex behavior of AA in the oxidative
lytical process, this corresponds to a limit of determina- stability of lipids.
tion (ca. 20 pglkg of sample), which is much lower than On the other hand, BHT showed an outstanding
that that has been estimated for the conventional antioxidant activity. At the lower BHT level tested,
spectrophotometric methods (Raharjo et al., 1993) and some decomposition of lipid peroxides still occurred,
comparable to that for solid-phase extraction (Raharjo whereas at higher levels decomposition was fully sup-
et al., 1993) and HPLC (Bird et al., 1983; Squires, 1990) pressed even when no EDTA was concurrently added.
methods. These findings lend support to previous results sug-
The extraction of MDA from samples was carried out gesting that full protection against autoxidation can be
with 5% TCA in the presence of BHT and hexane, a assured when BHT is added to the sample before the
modification of the aqueous acid extraction method of homogenization process (Pikul et al., 1983, 1989).
1934 J. Agric. Food Chem., Vol. 42, No. 9,1994 Botsoglou et al.

Table 2. Effect of Temperature and Reaction Time upon the Formation of the MDA-TBA Complex

temp heighta of the third-derivative peakb at 521.5 nm (arbitrary units) at reaction time of
("C) 5min 10min 15min 20min 25min 30min 40min 50min 60min 80min 100min 120min
100 0.349 0.414 0.417 0.420 0.416 0.415 0.407 0.398 0.395
80 0.305 0.401 0.422 0.433 0.436 0.434 0.432 0.431 0.423 0.423 0.423 0.409
70 0.196 0.339 0.390 0.423 0.448 0.449 0.450 0.448 0.449 0.452 0.449 0.448
Average of triplicate analyses. The reaction mixture contains 63.8 ng of MDA/mL.
Table 3. Raw Data and Regression Equation of Table 5. Precision and Accuracy Data for the
Calibration Curve for MDA Quantitation by Determination of Malondialdehyde in Cow Muscle,
Third-DerivativeSpectrophotometry Mackerel Tissue, and Milk Samples by Derivative
Spectrophotometry
concn of standard
(ng of MDA/mL of peak height," concn of MDA concn found" RSD mean
reaction mixture) mean f SD (n = 5) RSD (%) sample added (ppb) (ppb f SD) (%) recovery (a)
2.6 0.017 f 0.001 5.9 cow muscle 0 88 f 4.0 4.5
5.3 0.038 f 0.002 5.3 647 613 f 7.3 1.2 81.1 (98.3)b
10.6 0.069 f 0.001 0.7 998 890 f 23.1 2.6 80.4 (99.5)
31.9 0.219 f 0.004 1.8 1285 1161 f 19.9 1.7 83.5 (96.7)
42.6 0.301 f 0.002 0.7
53.2 0.378 f 0.003 0.8 mackerel 0 2093 f 39.3 1.9
79.8 0.565 f 0.002 0.4 tissue 364 2373 f 49.8 2.1 76.9 (95.4)
106.4 0.758 f 0.001 0.1 715 2620 f 21.0 0.8 73.5 (97.5)
133.0 0.936 f 0.002 0.2 1079 2894 f 33.5 1.1 74.2 (97.9)
a Peak height is expressed in arbitrary units. Regression 0 0
equation: Y = 4 2 . 7 x + (7.1 x 10-3)X,where Y represents 349
699
330 f 8.6
654 f 12.4
2.6
1.9
94.6 (99.2)
93.6 (95.6)
peak height and X the concentration of MDA in reaction mixture
(ng/mL). Correlation coefficient: 0.9999, 1048 969 f 17.4 1.8 92.5 (97.4)
a Average of five replicates. Values in parentheses represent
Table 4. Effect of Adding AA, BHT, or EDTA before the recoveries found when unhydrolyzed standard solution (TEP)
Blending on Malondialdehyde Values of Mackerel and was used for sample spiking.
Cod Samules
fresh mackerel frozen cod used for recovery determination are performed in the
concn of additive MDA concn" RSD MDA conma RSD same way. In most aqueous extraction methods, how-
(mglg of sample) (ppb f SD) (%) (ppb f SD) (%) ever, recovery experiments have been performed by
0 2289 f 322.7 14.1 2066 f 264.4 12.8 using TEP, the tetraethylacetal of MDA, instead of pure
Ob 2722 f 862.9 31.7 2250 f 751.5 33.4 MDA for sample spiking (Witte et al., 1970; Siu and
AA Draper, 1978; Csallany et al., 1984; Pikul et al., 1989;
1 2843 f 272.9 9.6 2616 f 217.1 8.3 Salih et al., 19891, whereas Squires (1990) used pure
3 2995 f 212.6 7.1 2544 f 198.4 7.8
BHT MDA, as we also did, but added it to an extract of the
5 1207 f 42.2 3.5 1478 f 29.6 2.0 sample, not to the sample itself (Raharjo et al., 1992,
10 1142 f 18.3 1.6 1333 f 29.3 2.2 1993). Considering the high reactivity of pure MDA
20 1114 f 18.8 1.7 1304 f 5.2 0.4 toward proteins and amino acids (Buttkus, 1967; Craw-
30 1126 f 22.5 2.0 1313 f 14.4 1.1 ford et al., 1967; Chi0 and Tappel, 19691, it seems
50 1117 f 12.3 1.1 1315 f 11.7 0.9 reasonable that these recovery procedures may not
EDTA
5 1571 f 110.0 7.0 1688 f 96.2 5.7
adequately represent the actual status of endogenous
10 1603 f 113.8 7.1 1616 f 132.5 8.2 MDA in real samples.
EDTA and BHT To determine whether the loss of the pure MDA added
10,20 1127 f 21.4 1.9 1318 f 22.4 1.7 to samples was due to partial binding with sample
" Mean of three replicates. Processing of these samples differs constituents or to incomplete extraction, all samples
in that hexane addition was carried out after initial blending with were further fortified with TEP and resubmitted to
TCA solution. analysis. The recovery values found were all higher
than 95.4% (Table 5, values in parentheses). Similar
To evaluate the precision and accuracy of the method, results were obtained when standard MDA solutions
cow and mackerel muscle tissues and milk samples, were submitted to analysis through the whole proce-
fortified with standard MDA solution at different levels, dure, a finding suggesting that incomplete extraction
were analyzed according to the procedure. The results does not occur.
based on five independent determinations at each Since Kwon et al. (1965) have specified that heating
fortification level are summarized in Table 5. The is needed to release MDA from its bound forms, ad-
precision was found t o be comparable to that reported ditional research was also directed toward examining
by Pikul et al. (1989), and the accuracy was quite whether heating could eliminate the binding of MDA
acceptable. Previously reported aqueous extraction to sample proteins. In this experiment, BSA samples
methods state recovery values of 62% (Newburg and fortified with pure MDA were either vortexed with 5%
Concon, 19801, 69% and 74.7% (Raharjo et al., 19931, TCA and then directly heated with TBA at 100 "C for
76% and 78% (Raharjo et al., 1992), 83% (Siu and 15 min or analyzed according to the proposed procedure.
Draper, 19781, 93% (Salih et al., 19871, 94% (Witte et The recoveries found were in the ranges 47.2-55.3% for
al., 1970; Pikul et al., 1989),94.8%(Squires, 1990), and the former samples and 86.4-89.1% for the latter ones.
98-100% (Csallany et al., 1984). Different recovery These results indicate that heating does not promote
values may be the result of different types of samples the release of MDA from its bound forms but may cause
or analytical methods, provided that the procedures further binding, a process which may account for the
Lipid Peroxidation in Tissue, Food, and Feed J. Agric. Food Chem., Vol. 42, No. 9,1994 1935

0.00 "
\
' -- 1.0
4

- 0.4 1
I 1 I U, I 10.0
450 500 550 600 450 500 550 600
Wavelengt h, nm Wavelength, nm
Figure 2. Representativenormal (- - -,right Y-axis)and third- Figure 3. Normal (- - -,right Y-axis)and third-derivative(-,
derivative (-, left Y-axis)spectra of sample extracts: (a)milk left Y-axis) spectra of reaction mixtures containing 15 pg of
not containingMDA, (b) fish meal containing 623 ppb of MDA. bilirubin (a) and 11 mg of sucrose (b).

lower recoveries reported for the direct heating (Rahaqjo (Marcuseand Johansson, 1973;Pryor, 1980; Esterbauer,
et al., 1993) and distillation methods (Siu and Draper, 1982; Yu et al., 1986; Kosugi et al., 1987, 1989).
1978) in comparison to our method. However, the identity of these red pigments has not
Figure 2 illustrates normal and third-derivative spec- been verified. The extent to which these aldehydes
tra of milk and fish meal samples. A secondary absorp- contribute to the red color formation in the TBA test is
tion band at around 451 nm appears in the normal also not known, but the red color yielded from these
spectra of both samples, with different intensities. The aldehydes is generally much lower than that from MDA
compound producing this absorbance was at such a high (Kosugi et al., 1987).
concentration that the initially colorless milk extracts The applicability of the method in various tissue, food,
turned deep yellow aRer the TBA reaction. Appearance and feedstuff samples is shown in Table 6. Values of
of such an absorption band has been previously reported MDA as determined by the derivative method were far
for many products (Tarladgis et al., 1964; Crackel, below the values found when samples were reanalyzed
1986). Several compounds may cause this secondary according to the method of Witte et al. (1970). To assess
peak and include carbohydrates, furfural, (hydroxy- this inconsistency, all final reaction mixtures taken by
methyl)furfural, alkenals, alkadienals and other alde- the method of Witte et al. (1970)were further submitted
hydes and ketones (Marcuse and Johansson, 1973; to derivative processing and evaluation. The difference
Sinnhuber and Yu, 1977; Pryor, 1980). The appearance between these results and those of our method does
of this interfering absorption band substantiates the point out that the protective action of BHT is against
thesis, advanced by Igene et al. (1985) and supported lipid peroxidation for the various types of products.
by Salih et al. (19871, that all spectrophotometric acid Thus, for animal tissues the antioxidant effect of BHT
extraction TBA methods should only be used if the was rather low, while for fish tissues it markedly
compounds producing this secondary band are absent increased due, obviously, t o the high levels of polyun-
or present only in small quantities that do not overlap saturated fatty acids in these samples. On the other
the main MDA-TBA absorption band at 532 nm. hand, the high difference between the derivative and
Figure 2 shows, however, that such a limitation does absorbance measurements of all samples processed
not exist when normal spectra are submitted to deriva- according to the method of Witte et al. (1970) cannot
tive processing. Any substance that produces a constant be considered unexpected when one examines both the
background absorption or a gradual background varia- coloration and the normal spectra of sample extracts
tion can be normalized by the third-derivative function, before and after the TBA reaction. Serious interferences
which takes zero value. due to the presence of colored endogenous compounds
Since there may have been additional compounds were noted in the case of fish tissue and feedstuff
interfering with the TBA reaction, an interference test samples, whereas interferences due to yellow pigment
was also carried out. Various compounds with known formation after the TBA reaction also appeared in dairy
interfering action on the spectrophotometric assay such products.
as sucrose, fructose, ribose, cystine, formaldehyde, acet- In conclusion, the results of the present study suggest
aldehyde, glyceraldehyde, and N-acetylneuraminic acid that the derivative method can be a rapid and efficient
(Shin et al., 1972; Patton, 1973; Cordis et al., 1993)were alternative to existing methods for estimating the extent
tested and found t o have no interfering action on the of lipid peroxidation in various products. Increased
derivative assay. Bilirubin could react with TBA to sensitivity and specificity can be readily attained through
produce a red pigment (Figure 31, but this interference the use of derivative processing. Specially trained staff
was practically eliminated by the hexane partitioning are not required, and the equipment needed is easily
and derivative processing. Red pigments with absorp- accessible, as most modern spectrophotometers allow
tion spectrum and HPLC retention time similar to those instant generation of derivative spectra. These advan-
of the MDA-TBA complex may also be produced by the tages make the new method particularly useful for
reaction of TBA with alkanals, alkenals, and alkadienals routine control. To maximize the safety of using hexane
1936 J. Agric. Food Chem., Vol. 42,No. 9,1994 Botsoglou et al.
Table 6. Malondialdehyde Content of Animal Tissue, Chio, K.; Tappel, A. Synthesis and Characterization of the
Food, and Feedstuff Samples As Determined by Different Fluorescent Products Derived from Malondialdehyde and
Procedures Amino Acids. Biochemistry 1969, 8, 2821-2826.
found malondialdehvde concn (ppb) Cordis, G. A,; Maulik, N.; Bagchi, D.; Engelman, R. M.; Das,
D. K. Estimation of the Extent of Lipid Peroxidation in the
extraction method Ischemic and Reperfused Heart by Monitoring Lipid Meta-
(Witte et al., 1970) bolic Products with the Aid of High-Performance Liquid
absorbance derivative derivative Chromatography. J . Chromatogr. 1993, 632, 97-103.
sample measurement measurement method Crackel, R. L. Stability of Lipids in Restructured Beef Steaks.
M.S. Thesis, Michigan State University, East Lansing, MI,
rat
muscle 529 64 37 1986.
kidney 1416 141 108 Crawford, D.; Yu, T.; Sinnhuber, R. Reaction of Malonaldehyde
cow with Protein. J . Food Sci. 1967, 32, 332-335.
muscle 749 190 138 Csallany, A. S.; Guan, M. D.; Manwaring, J. D.; Addis, P. B.
liver 2081 494 480 Free Malonaldehyde Determination in Tissues by High-
pork Performance Liquid Chromatography. Anal. Biochem. 1984,
muscle 621 209 20 1 14,277-283.
liver 971 173 138 Cutler, M. G.; Hayward, M. A. Effect of Lipid Peroxides on
kidney 870 138 106 Fat Absorption and Folic Acid Status in the Rat. Nutr.
chicken Metabol. 1974, 16, 87-93.
muscle 726 291 254 Draper, H. H.; Hadley, M. Malondialdehyde Determination as
liver 687 133 85 Index of Lipid Peroxidation. Methods Enzymol. 1990,186,
lamb 421-431.
kidney 638 68 166 Draper, H. H.; Polensek, L.; Hadley, M.; McGirr, L. G. Urinary
fish Malondialdehyde as an Indicator of Lipid Peroxidation in
frozen cod 14303 3624 2424 the Diet and in the Tissues. Lipids 1984,19, 836-843.
mackerel 19247 6357 2289
sardine 13325 4329 1565 Esterbauer, H. Aldehydic Products of Lipid Peroxidation. In
dairy products Free Radicals, Lipid Peroxidation and Cancer;McBrien, D.,
unsweetened 782 53 0 Slater, T., Eds.; Academic Press: London, 1982; pp 101-
condensed milk 128.
infant milk formula 2633 101 0 Fletouris, D. J.;Botsoglou, N. A.; Papageorgiou, G. E.; Mantis,
blue cheese -a 68 27 A. J. Rapid Determination of Tryptophan in Intact Proteins
kasseri cheese -a 40 40 by Derivative Spectrophotometry. J . AOAC Znt. 1993, 76,
feta cheese 875 146 110 1168-1173.
processed cheese 2191 1408 884 Gray, J. I. Measurement of Lipid Oxidation: A Review. J .
coffee cream powder 4578 56 30 Am. Oil Chem. SOC.1978,55, 539-546.
butter -b -b 1173 Gutteridge, J. M. C.; Quinlam, G. J. Malonaldehyde Formation
feedstuffs from Lipid Peroxides in the Thiobarbituric Acid Test: The
fish (herring) meal 33293 2019 1209 Role of Lipid Radicals, Iron Salts, and Metal Chelators. J .
fish meal 29475 3268 1012 Appl. Biochem. 1983,5, 293-299.
meat meal 10251 4899 4182 Hackett, C.; Linley-Adams, M.; Lloyd, B.; Walker, V. Plasma
soybean meal 11356 1842 1087 Malondialdehyde: A Poor Measure of in Vivo Lipid Peroxi-
Measurement could not be made because of high turbidity. dation. Clin. Chem. 1988,34, 208.
Butter cannot be analyzed by this method. Igene, J. 0.;Yamauchi, K.; Pearson, A. M.; Gray, J. I.; Aust,
S. D. Evaluation of a 2-Thiobarbituric Acid Reactive Sub-
in sample treatment, use of an explosion-proof blender stances (TBRS) in Relation to Warmed-over Flavor (WOF)
Development in Cooked Chicken. J . Agric. Food Chem.
and screw-capped centrifuge tubes is highly recom- 1985,33,364-367.
mended. Kanner, J. Oxidative Processes in Meat and Meat Products:
Quality Implications. Meat Sci. 1994, 36, 169-189.
LITERATURE CITED Kanner, J.; Mendel, H.; Budowski, P. Prooxidant and Anti-
oxidant Effects of Ascorbic Acid and Metal Salts in A
Basu, A.; Marnett, L. Unequivocal Demonstration that Mal- ,Warotene-Linoleate Model System. J . Food Sci. 1977,42,
onaldehyde is a Mutagen. Carcinogenesis 1984, 4, 331- 60-64.
333. Kaunitz, H.; Johnson, R. E. Exacerbation of Heart and Liver
Benedict, R. C.; Strange, E. D.; Swift, C. E. Effect of Lipid Lesions in Rats by Feeding of Various Mildly Oxidized Fats.
Antioxidants on the Stability of Meat During Storage. J . Lipids 1973,8, 329-336.
Agric. Food Chem. 1975,23, 167-170. Ke, P. J.; Cervantes, E.; Robles-Martinez, C. Determination
Bidlack, W.; Tappel, A. Fluorescent Products of Phospholipids of Thiobarbituric Acid Reactive Substances (TBARS)in Fish
During Lipid Peroxidation. Lipids 1973, 8, 203-207. Tissue by an Improved Distillation Spectrophotometric
Bird, R. P.; Hung, S. S. 0.; Hadley, M.; Draper, H. H. Method. J . Sci. Food Agric. 1984, 35, 1248-1254.
Determination of Malonaldehyde in Biological Materials by Khayat, A.; Schwall, D. Lipid Oxidation in Seafood. Food
High-pressure Liquid Chromatography. Anal. Biochem. Technol. 1983,37, 130-140.
1983,128, 240-244. Kosugi, H.; Kato, T.; Kikugawa, K. Formation of Yellow,
Botsoglou, N. A.; Fletouris, D. J.;Papageorgiou, G. E.; Mantis, Orange, and Red Pigments in the Reaction of Alk-2-enals
A. J. Derivative Spectrophotometric Method for the Analysis with 2-Thiobarbituric Acid. Anal. Bwchem. 1987,165,456-
of Tyrosine in Unhydrolyzed Protein, Food, and Feedstuff 464.
Samples. J . Agric. Food Chem. 1993,41, 1635-1639. Kosugi, H.; Kojima, T.; Kikugawa, K. Thiobarbituric Acid-
Braddock, R. J.; Petrus, D. R. Malonaldehyde in Aqueous Reactive Substances from Peroxidized Lipids. Lipids 1989,
Orange Juice Essences. J . Food Sci. 1971,36,1095-1097. 24, 873-881.
Buttkus, H. The Reaction of Myosine with Malonaldehyde. J . Kwon, T. W.; Menzel, D. B.; Olcott, H. S. Reactivity of
Food Sci. 1967,32, 432-434. Malonaldehyde with Food Constituents. J . Food Sci. 1965,
Buttkus, H.; Bose, R. J. Amine-Malonaldehyde Condensation 30, 808.
Products and Their Relative Color Contribution in the Lee, H.-S.; Csallany, A. S. Measurement of Free and Bound
Thiobarbituric Acid Test. J . Am. Oil Chem. SOC.1972,49, Malondialdehyde in Vitamin E -Deficient and -Supplement-
440 -443. ed Rat Liver Tissues. Lipids 1987,22, 104-107.
Lipid Peroxidation in Tissue, Food, and Feed J. Agric. Food Chem., Vol. 42,No. 9, 1994 1937

Marcuse, R.; Johansson, L. Studies on the TBA Test for Sidwell, C. G.; Salwin, H.; Mitchell, J. H. Measurement of
Rancidity Grading. 11. TBA Reactivity of Different Alde- Oxidation in Dried Milk Products with Thiobarbituric Acid.
hyde Classes. J . Am. Oil Chem. Soc. 1973, 50, 387-391. J . Am. Oil Chem. Soc. 1955,32, 13-16.
McCoy, R. N.; Hill, K. E.; Ayon, M. A,; Stein, J. H.; Burk, R. Sinnhuber, R. 0.;Yu, T. C. 2-Thiobarbituric Acid Method for
F. Oxidant Stress Following Renal Ischemia: Changes in the Measurement of Rancidity in Fishery Products. 211.
the Glutathione Redox Ratio. Kidney Int. 1988, 33, 812- The Quantitative Determination of Malonaldehyde. Food
817. Technol. 1958,12,9-12.
Melton, S. L. Methodology for Following Lipid Oxidation in Sinnhuber, R. 0.;Yu, T. C. The 2-Thiobarbituric Acid Reaction,
Muscle Foods. Food Technol. 1983,37, 105-116. an Objective Measure of the Oxidative Deterioration Oc-
Newburg, D. S.; Concon, J. M. Malonaldehyde Concentrations curring in Fats and Oils. J . Jpn. Soc. Fish. 1977,26,259-
in Food are Affected by Cooking Conditions. J . Food Sci. 267.
1980,45, 1681-1687. Siu, G. M.; Draper, H. H. A Survey of the Malonaldehyde
O’Haver, T. C.; Green, G. L. Numerical Error Analysis of Content of Retail Meats and Fish. J . Food Sci. 1978, 43,
Derivative Spectrometry for the Quantitative Analysis of 1147- 1149.
Mixtures. Anal. Chem. 1976, 48, 312-318. Squires, E. G. High Performance Liquid Chromatographic
Ohkawa, H.; Ohishi, N.; Yagi, K. Assay for Lipid Peroxides in Analysis of the Malondialdehyde Content of Chicken Liver.
Animal Tissues by Thiobarbituric Acid Reaction. Anal. Poult. Sci. 1990, 69, 1371-1376.
Biochem. 1979,95, 351-358. Tarladgis, B. G.; Watts, B. M.; Younathan, M. T.; Dugan, L.
Patton, S. Malonaldehyde, Lipid Oxidation, and Thiobarbituric R. A Distillation Method for the Quantitative Determination
Acid Test. J . Am. Oil Chem. Soc. 1973, 51, 114. of Malonaldehyde in Rancid Foods. J . Am. Oil Chem. Soc.
Pikul, J.; Leszczynski, D. E.; Kummerow, F. A. Elimination 1960,37,44-48.
of Sample Autoxidation by Butylated Hydroxytoluene Ad- Tarladgis, B. G.; Pearson, A. M.; Dugan, L., Jr. The Chemistry
ditions before Thiobarbituric Acid Assay for Malonaldehyde of the 2-Thiobarbituric Acid Test for the Determination of
in Fat from Chicken Meat. J . Agric. Food Chem. 1983,31, Oxidative Rancidity in Foods. I. Some Important Side
1338-1342. Reactions. J . Am. Oil Chem. Soc. 1962, 39, 34-39.
Pikul, J.; Leszczynski, D. E.; Kummerow, F. A. Evaluation of Tarladgis, B. G.; Pearson, A. M.; Dugan, L. R. Chemistry of
Three Modified TBA Methods for Measuring Lipid Oxidation the 2-Thiobarbituric Acid Test for Determination of Oxida-
in Chicken Meat. J . Agric. Food Chem. 1989, 37, 1309- tive Rancidity in Foods. 11. Formation of the TBA-Mal-
1313. onaldehyde Complex Without Acid-Heat Treatment. J . Sci.
Pryor, W. A. In Molecular Basis of Environmental Toxicity; Food Agric. 1964,15, 602-607.
Bhatnagar, R. S., Ed.; Ann Arbor Science: Ann Arbor, MI, Trombly, R.; Tappel, A. Fractionation and Analysis of Fluo-
1980; pp 3-36. rescent Products of Lipid Peroxidation. Lipids 1975, 10,
Raharjo, S.; Sofos, J. N. Methodology for Measuring Malonal- 441-447.
dehyde as a Product of Lipid Peroxidation in Muscle Turner, E. W.; Paynter, W. D.; Montie, E. J.; Bessert, M. W.;
Tissues: A Review. Meat Sci. 1993,35,145-169. Struck, G. M.; Olson, F. C. Use of the 2-Thiobarbituric Acid
Raharjo, S.; Sofos, J. N.; Schmidt, G. R. Improved Speed, Reagent to Measure Rancidity in Frozen Pork. Food Tech-
Specificity, and Limit of Determination of an Aqueous Acid nol. 1964, 8, 326-330.
Extraction Thiobarbituric Acid418 Method for Measuring Witte, V. C.; Krause, G. T.; Bailey, M. E. A New Extraction
Lipid Peroxidation in Beef. J . Agric. Food Chem. 1992,40, Method for Determining 2-Thiobarbituric Acid Values of
2182-2185. Pork and Beef During Storage. J . Food Sci. 1970,35,582-
Raharjo, S.; Sofos, J. N.; Schmidt, G. R. Solid-Phase Acid 585.
Extraction Improves Thiobarbituric Acid Methods to De- Wong, D. W. S. Lipids. In Mechanism and Theory in Food
termine Lipid Oxidation. J . Food Sci. 1993, 58, 921-932. Chemistry; Van Nostrand Reinhold: New York, 1989; pp
Rhee, K. S. Minimization of Further Lipid Peroxidation in the 1-47.
Distillation 2-Thiobarbituric Acid Test of Fish and Meat. Yamauchi, K.; Nagai, Y.; Ohashi, T. Quantitative Relationship
J . Food Sci. 1978,43, 1776-1779. Between Alpha-Tocopheroland Polyunsaturated Fatty Acids
Richardson, T.; Korycka-Dahl, M. Lipid Oxidation. In Deuel- and its Connection to Development of Oxidative Rancidity
opments in Dairy Chemistry-2; Fox, P. F., Ed.; Applied in Chicken Skeletal Muscle. Agric. Biol. Chem. 1982, 46,
Science Publishers: London, 1983; pp 241-364. 2719-2724.
Salih, A. M.; Smith, D. M.; Price, J. F.; Dawson, L. E. Modified Yu, L. W.; Latriano, L.; Duncan, S.; Hartwick, R. A.; Witz, G.
Extraction 2-Thiobarbituric Acid Method for Measuring High-Performance Liquid Chromatography Analysis of the
Lipid Oxidation in Poultry. Poult. Sci. 1987,66,1483-1488. Thiobarbituric Acid Adducts of Malonaldehyde and Trans,
Salih, A. M.; Price, J. F.; Smith, D. M.; Dawson, L. E. Lipid Trans-Muconaldehyde. Anal. Bwchem. 1986,156,326-333.
Degradation in Turkey Breast Meat during Cooking and Yu, T. C.; Sinnhuber, R. 0. 2-Thiobarbituric Acid Method for
Storage. Poult. Sci. 1989, 68, 754. the Measurement of Rancidity in Fishery Products. Food
Schmedes, A.; Holmer, G. A New Thiobarbituric Acid (TBA) Technol. 1957, 11, 104-108.
Method for Determining Free Malondialdehyde (MDA) and Yu, T. C.; Sinnhuber, R. 0. Further Obervations on the
Hydroperoxides Selectively as a Measure of Lipid Peroxi- 2-Thiobarbituric Acid Method for the Measurement of
dation. J.Am. Oil Chem. SOC. 1989, 66, 813-817. Oxidative Rancidity. J . Am. Oil Chem. SOC.1964,41,540-
Shamberger, R.; Andreone, T.; Willis, C. Antioxidants and 543.
Cancer. IV. Initiating Activity of Malonaldehyde as a
Carcinogen. J . Natl. Cancer Inst. 1974,53, 1771-1773.
Shamberger, R. J.; Shamberger, B. A.; Willis, C. E. Malonal- Received for review February 18, 1994. Accepted June 9,
dehyde Content of Food. J . Nutr. 1977, 107, 1404-1409. 1994.@
Shin, B. C.; Huggins, J. W.; Carraway, K. L. Effect of pH,
Concentration, and Aging on the Malonaldehyde Reaction @Abstract published in Advance A C S Abstracts, Au-
with Proteins. Lipids 1972, 7,229-233. gust 1, 1994.