Z Lebensm Unters Forsch (1996) 202:205-210 9 Springer-Verlag 1996

Gitte Sorensen 9 Soren Storgaard Jorgensen

A critical examination of some experimental variables in the

2.thiobarbituric acid (TBA) test for lipid oxidation in meat products

Received: 27 March 1995/Revised version: 9 June 1995

Abstract By comparison of various procedures for the semi-quantitative estimation of lipid oxidation in
determination of TBARS (2-thiobarbituric-acid-reactive foods. The method is based on the spectrophotometric
substances) in non-cured meat samples, it was found that measurement of a red chromophore formed by the
a direct extraction procedure with trichloroacetic acid reaction of TBA with secondary products from
gave significantly higher reaction yields than a rapid lipid oxidation of unsaturated fatty acids, with
Kjeltec steam distillation procedure (50 ml distillate in malondialdehyde (MDA) being used as a calibration
2.5 min). When optimizing the distillation procedure standard.
with regard to distillation principle (heating by thermal The TBA test has been described with various differ-
conductivity or steam injection) and rate, the reaction ences by many workers, the most common procedure
yield was more affected by changes in distillation prin- used being that described by Tarladgis et al. [1], which
ciple than by changes in distillation rate. Simple distilla- involves distillation of an acidified sample in order to
tion (conduction heating) yielded significantly higher separate the TBA-reactive substances (TBARS) from
TBA values than steam distillation at the same distilla- the food matrix. A more simple procedure is the
tion rate. The reproducibility of the distillation proced- so-called extraction method in which the acidified
ure was found to be about 5-10 times poorer than of the sample is filtered to give a clear aqueous solution [2].
extraction procedure and, because of the high reaction There has been great interest in optimizing the TBA
yields and robustness of the extraction procedure, the method, particularly the more generally applicable dis-
use of the extraction procedure is recommended with tillation procedure, since the extraction procedure may
food matrices such as non-cured meat samples with less be inadequate in cases of coloured samples and samples
than 14-18% fat and no artificial colorants. high in fat ( > 10%) [-3]. In the original work of Tar-
ladgis et al. [1], the distillation principle is based on
Key words 2-Thiobarbituric-acid-reactive simple distillation, in which case the sample mixture is
substances' Malondialdehyde 9Non-cured meat" heated by conductivity and not by steam. In more
TBA-test 9TBARS recent work almost all distillation procedures have
been based on steam distillation where it is possible to
obtain very high distillation rates with the automated
distillation apparatus (Kjeldahl stills) commonly found
Introduction in modern laboratories. Thus Hoyland and Taylor 1-4]
have described a rapid steam distillation procedure in
The 2-thiobarbituric acid (TBA) test is probably which 50 ml distillate is collected in about 4 min as
the most extensively used chemical method for the opposed to 10 min by simple distillation [1]. In our
laboratory, distillation rates of up to 20 ml/min (50 ml
collected in 2.5 min) are a common procedure using the
G. Sorensen Tecator Kjeltec distillation apparatus.
The Royal Veterinary and Agricultural University,
KVL Center for Food Research, Thorvaldsensvej 40, The results obtained by various TBA procedures are
DK-1871 Frederiksberg C, Denmark known to be both method and operator dependent
[5, 6]. In the extraction procedure, only TBARS that
S. Storgaard Jorgensen (12N)
The Royal Veterinary and Agricultural University, are instantaneously released from the food matrix into
Chemistry Department, Thorvaldsensvej 40, an acid aqueous phase at room temperature are deter-
DK-1871 Frederiksberg C, Denmark mined. In a distillation procedure, those TBARS that
206

are released during heat treatment at 100 ~ and distil- Table 1 Factors and levels in the optimization design experiment
led with water vapour are determined. Earlier research
[2, 7-10] has concluded that the distillation procedure Factor Description Level 1 Level 2
yields greater TBA values than does the extraction A TBA procedure Extraction Distillation"
procedure. This indicates that more TBARS are re- B Pretreatment of TBA
leased at 100 ~ than at room temperature, but it is also reaction 5 ~ for 1 h b None
generally accepted that when using the distillation C Reaction temperature 70 ~ 100 ~
procedure the total amount of TBARS in the sample D Sensitivity of TBA
reaction to light Light No light
mixture is not transferred to the collected distillate.
Calculated TBA values are therefore corrected by a Steam distillation on Kjeltec System 1002 Distilling Unit at a rate
means of an experimentally determined percentage of 20 ml/min
recovery, in order to get a more reliable estimate of the b According to [11]
total amount of TBARS and to be able to use TBA
values on a comparative basis. Obviously this requires collected in each distillation) were compared to simple distillation
a stable and reproducible distillation procedure. In (the mixture was heated by thermal conductivity) with distillation
contrast, the extraction procedure will normally yield rates of 3.3, 5, 10 and 20 ml/min.
about 100% recovery. In our laboratory, low and All treatment parameters were tested on both meat and standard
samples with two levels of TBARS.
poorly reproducible recoveries of added malondi-
aldehyde (MDA) were experienced when using the very
rapid Kjeltec distillation procedure. Extraction procedure
The purpose of this work was to compare various
procedures for the determination of TBARS in meat The extraction procedure was modified according to previous publi-
samples with respect to reaction yields, reproducibility cations [2, 12]. Of the sample 10 g was homogenized with 30 ml of
and ease of operation. The following variables were a 7.5% trichloroacetic acid (TCA) solution containing
examined: (1) extraction versus distillation; (2) reaction 0.1% propylgallate (PG) and 0.1% ethylenediaminetetraacetic
acid, disodium salt (EDTA) for 30 s in an Ultra Turrax blender
temperature; (3) the influence on reaction of light and (9500rpm) (Janke and Kunkel, Germany) and filtered through
of a cold pretreatment [11]. a Whatman filter no. 42.
An attempt has been made to optimize the
distillation procedure with regard to distillation
principle (conduction and steam heating) and rate. Distillation procedure

The distillation procedure was modified according to previous

publications [1, 13]. Of the sample, 10 g was homogenized with
Materials and methods 50 ml distilled water containing 0.1% P G and 0.1% EDTA for 1 rain
in an Ultra Turrax blender (9500 rpm.). The mixture was transferred
Sample preparation quantitatively into a distillation tube by washing with an additional
47.5 ml distilled water containing 0.1% PG and 0.1% EDTA. The
Meat. Commercially sold fresh ground beef with 14-18% fat was mixture was acidified with 2.5 ml of a HC1 solution (concentrated
used. To produce meat samples with two different levels of lipid HCI: distilled water, 1 : 2) and five drops of Antifoam A were added.
oxidation, one half of the sample was immediately vacuum-packed Steam distillation was carried out using an automated distillation
in plastic bags and stored at - 18 ~ The other half was exposed to apparatus (Kjeltec System 1002 Distilling Unit, Tecator, Sweden)
air and light for 2 days to accelerate lipid oxidation and then with a distillation rate of 20 ml/min, or using a laboratory designed
vacuum-packed in plastic bags and stored at - 18 ~ On the day of glass apparatus with distillation rates of 3.3, 5 and 10 ml/min. Simple
analysis the meat samples were thawed in lukewarm water and distillation was carried out with common glass equipment with
homogenized before analysis. distillation rates of 2.1 and 4.2 ml/min. Volumes of 50 ml were
collected in each distillation.
Standards. A stock solution of 0.002 M TEP (1,1,3,3-tetraethoxy-
propane, Merck 805797) in distilled water was prepared. Aliquots
of the TEP shock solution were used to make standard samples Quantification of TBARS
with concentrations corresponding to 2 x 10 - 6 M and 10 x 1 0 - 6 M
TEP. Equal 5 ml volumes of filtrate or distillate and 0.02 M TBA solution
(Merck 8180, in distilled water) were mixed in glass stoppered tubes
and incubated in a water bath at 70 ~ or 100 ~ for 40 rain before
cooling to room temperature under running cold tap water. All
Experimental design samples were analysed in duplicate. In accordance with the optim-
ization design experiment (Table 1), some samples were subjected to
Optimization. In order to optimize the TBA method, a fractional a cold pretreatment at 5 ~ for 1 h before incubation in the water
factorial experiment was designed with 4 factors at 2 levels each. bath. Sensitivity of the reaction to light was tested either with
Treatment factors and levels are shown in Table 1, a metal lid placed over the water bath (no light) or with two 60 W
light bulbs placed directly (5 cm) above the samples in the water bath
Comparison of distillation principles and distillation rates. Steam and covered with a reflecting lid to give an even distribution of light.
distillation (the mixture was heated by steam injection) with A calibration curve for each of the extraction and the distillation
distillation rates of 2.1 and 4.2 ml/min (volumes of 50 ml were procedures was constructed from a dilution series of the 0.002 M
 207

TEP stock solution to a range from 2 x 10 6 M to 10 x 10 . 6 M Table 2 Estimated significance levels of treatment parameters from
TEP. The dry matter content of the beef samples was measured the analysis of variance of the results from the optimization design
on an automatic moisture analyser (Sartorius MA30, Sartorius, experiment
G6ttingen, Germany). All absorbances were converted to concentra-
tions of MDA by reference to the calibration curves and the corres- Factor Description Meat Standard
ponding TBA values for standards were expressed as micromoles
MDA and for meat as micromoles MDA equivalents per kilogram A TBA procedure *** ***
sample (TBARS) according to: (extraction/distillation)
B Cold pretreatment n.s. n.s.
TBARS (Extraction) C Reaction temperature * *
D Sensitivity to light n.s. n.s.
i ( 1 0 0 % + % drymatter
CMDA" VTCA "Jr- 100% AxC n.s. **
A x Concentration *** n.s.
Ms
Significance levels: *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not
CMDA " Vdis significant
TBARS (Distillation) (2)
Ms
where CMDAis the concentration of MDA (gM) as read from the
calibration curve; Ms is the mass of the sample (g); VTCA is the ~" 1 0 - - -
volume of TCA (ml); and Vdisis the volume of the distillate (ml).

Spectrophotometric measurements
6
A HP 8452 A diodearray photometer (Hewlett Packard, Germany)
and 1-cm quartz cuvettes were used for all measurements. Absorb-
4
ances were recorded in the wavelength range 300 nm to 650 nm
against a reagent blank which had been subjected to the same
analytical procedure as described above for the samples. TBA values 2 2
were calculated from the absorbance measured at 532 nm, but the
entire spectrum was checked for the possible formation of other
components. 0--
Extraction Distillation
70oC 10OOC 70~C 100~

Statistical analysis Fig. 1 Results from the optimization design experiment for both
standards and meat samples expressed as mean 2-thiobarbituric acid
(TBA) values sorted according to TBA procedure and reaction
Statistical analysis was performed with SAS version 6.0. (SAS Insti- temperature. Standards: 9 2x 10 -6 M MDA; [] 10x 10 -6 M
tute, N.C., USA). Treatment effects were tested with standard analy-
MDA. Meat samples: 9 low oxidative level; [] high oxidative level
sis of variance (General Linear Models), and Tukey's (HSD) Studen-
tized test was used to test for significant differences between mean
TBA values at the 5% level.
s a m p l e size a n d f i l t r a t e / d i s t i l l a t e v o l u m e . T h e T B A
v a l u e s a r e n o t c o r r e c t e d for p e r c e n t a g e r e c o v e r y o f t h e
Recommended procedure (extraction, non-cured meat samples)
T B A m e t h o d , as is u s u a l l y d o n e , b e c a u s e of t h e p o o r
Homogenize 10 g of sample with 30 ml of a 7.5% TCA solution reproducibility of the steam distillation procedure. The
containing 0.1% PG and 0.1% EDTA for 30 s in an Ultra Turrax reproducibility of the distillation procedure, expressed
blender (9500 rpm) and filter through a Whatman filter no. 42. Mix as t h e r e l a t i v e s t a n d a r d d e v i a t i o n , is a b o u t 5 - 1 0 t i m e s
equal 5 ml volumes of filtrate and 0.02 M TBA solution in glass p o o r e r t h a n t h a t o f t h e e x t r a c t i o n p r o c e d u r e a n d fur-
stoppered tubes and incubate in a water bath at 100 ~ for 40 rain
ther results indicate (data not shown) that when con-
before cooling to room temperature under running cold tap water.
Measure the absorbance at 532 nm against a reagent blank which structing a calibration curve with samples of TEP that
has been subjected to the same analytical procedure as described h a v e b e e n distilled, t h e w i t h i n - s a m p l e r e p r o d u c i b i l i t y is
above for the sample. Prepare a calibration curve from a dilution e q u a l l y p o o r a n d t h e r e g r e s s i o n coefficient o f t h e c a l i b -
series of 0.002M TEP to cover a range from l x 10-6M to r a t i o n c u r v e is v e r y low, at a b o u t 0.88. W h e n c o n s t r u c t -
10x 10 6 M TEP, where the TEP aliquots are diluted with a
7.5% TCA solution containing 0.1% PG and 0.1% EDTA. The i n g a l i n e a r c a l i b r a t i o n c u r v e in t h e s a m e f a s h i o n for t h e
calibration curve should be linear in this range. Calculate TBARS e x t r a c t i o n p r o c e d u r e , t h e r e g r e s s i o n coefficient is 0.99.
from the equation given above. I n T a b l e 2 a r e l i s t e d s i g n i f i c a n c e levels of t h e t r e a t -
ment parameters from the analysis of variance which
were performed on the estimated TBA values from the
Results and discussion optimization design experiment. It was found that the
p a r a m e t e r s p r o c e d u r e a n d t e m p e r a t u r e h a v e a signifi-
I n o r d e r t o q u a n t i f y t h e T B A R S in t e r m s o f T B A c a n t effect o n t h e m e a s u r e d T B A v a l u e s for b o t h s t a n -
values, t h e m e a s u r e d a b s o r b a n c e s h a v e b e e n c o n v e r t e d dards and meat samples. Figure 1 shows the results
to e q u i v a l e n t a m o u n t s o f M D A a n d t h e n c o r r e c t e d for f r o m t h e o p t i m i z a t i o n o f T B A m e t h o d s , a n d it is seen
208

Table 3 Reproducibility of percentage recovery of malondialdehyde 1.6

t
(MDA) from standard samples for 3 different TBA procedures 2.4

TBA procedure Percentage recovery of MDA (%)

J 2.2 e
Concentration of M D A in standard 0.8:
sample
2 ~tM 10 pM 0.4
1.8
Extraction 102.7 _+ 5.6 (5.5) 98.9 __ 1.0 (0.9)
Kjeltec Steam distillation 22.8 __+8.5 (37.3) 18.2 _+ 2.5 (13.7) S 0 \ 1.6 .~
(20 ml/min) %
Simple distillation 76.1 • 8.5 (11.2) 72.5 i 0.8 (1.1) -0.4 1.4
(4.2 ml/min) Extraction Distillation
T B A procedure
Results are expressed as the mean percentage recovery _+ SD of
4 determinations. In parentheses are shown the coefficient of vari- Fig. 2 Interaction effect between TBA procedure and reaction
ation (%) temperature shown for both standards and meat samples. Stan-
d a r d s : - D - 70 ~ - I 1 - 100 ~ Meat samples: - - * - - 70 ~ - - A - -
100 ~

that the extraction procedure yields much higher TBA

values than the rapid Kjeltec steam distillation proced- a similar conclusion to that drawn by others [4] who
ure. For standards, steam distillation yields about 20% stress the importance of using a uniform temperature in
of extraction values while for meat the yield is about the water bath when using the distillation procedure.
42%. Although these findings are in contrast with those The optimization experiment showed no effect of
in several publications [-2,7 10], the conflicting results letting the filtrate or distillate react with TBA at 5 ~
are not surprising considering the missing correction for 1 h before incubation in water at 70 ~ or 100 ~
for percentage recovery. It was found that the extrac- just as no difference could be detected, neither at
tion procedure yields about 100% recovery while the 532 nm nor in any other part of the spectrum (300 nm
recovery is about 20% for the rapid steam distillation to 650 nm) between samples reacted in the dark or
procedure (Table 3), but with a large variation between under intense illumination.
determinations and some inconsistency for low and Figure 3a, b presents the results from the comparison
high concentrations of TEP. (Percentage recovery of between simple distillation and steam distillation with
M D A is calculated as the concentration of M D A in different distillation rates for either procedure. For
standard samples which have been extracted or distil- both standard and meat samples the TBA values from
led compared to the concentration of appropriate cal- simple distillation are significantly higher than those
ibration samples.) Reports relating to the recovery of obtained by steam distillation where the automated
a steam distillation method [-4] support these results, as Kjeltec procedure yields the significantly lowest TBA
the recoveries have a tendency to vary with the concen- values. With the distillation equipment available there
tration of TEP. seems to be no effect of lowering the distillation rate for
The reaction temperature was found to be significant simple distillation from 4.2 to 2.1 ml/min, just as there
at the 5% level, but the analysis of variance indicates is no significant difference between TBA values ob-
a statistical interaction between procedure and temper- tained with distillation rates of 3.3, 5 and 10 ml/min.
ature (Table 2). This is illustrated in Fig. 2 for both This could lead to the conclusion that it is the distilla-
meat and standard samples. The interaction effect is tion principle, rather than the distillation rate, which
found only to be significant for standard samples, but has the largest influence on the estimated TBA values.
Fig. 2 suggests that this could hold for meat samples as Our findings coincide with those of Tarladgis et al. [-1]
well since the interaction plot for meat samples is very who recommended the collection of the distillate from
similar to that of standard samples. The statistical simple distillation with the highest distillation rate pos-
interaction between procedure and temperature indi- sible. From relevant literature [-1,4, 14, 15] it is noticed
cates that while the TBA values found when using the that a higher distillation rate is usually obtained by
distillation procedure at a reaction temperature of using steam distillation instead of simple distillation,
70 ~ are lower than those found at a temperature of but, at the same time, this will result in lower reaction
100 ~ there seems to be no difference between TBA yields.
values found with the two temperatures when using the In Fig. 4 mean TBA values found with the two differ-
extraction procedure. This could be caused by a lower ent distillation principles with distillation rates of
pH in the filtrates after extraction due to the addition of 20 ml/min and 4.2 ml/min and the extraction proced-
TCA. This finding suggests that the distillation proced- ure are compared. The automated steam distillation
ure is more sensitive to temperature changes during procedure yields, as expected, the significantly lowest
analysis than the extraction procedure and supports TBA values for both meat and standard samples. For
 209

a) Standard further lipid oxidation during distillation despite the

a
use of ant 9 [10, 161.
This work corroborates the general experience that
the TBA test is highly influenced by procedural para-
.I meters, especially whether the TBARS are separated
from the sample matrix by distillation or by extraction.
The reaction yields from extraction are found to be
significantly higher than reaction yields found with
either distillation principle and it is seen that the reac-
5 10 15 20 tion yield from distillation is more affected by changes
Distillation rate in 9
in distillation principle than by changes in distillation
60 b) M e a t rate. It is notable that the Kjeltec steam distillation
procedure yields by far the lowest TBA values. This
could indicate that a u t o m a t e d Kjeldahl stills are not
4o
optimal for distillation of TBARS.
3?
<
The TBA test is often used as a measure of lipid
ca oxidation on a comparative basis, where a correction
20
u for percentage recovery of M D A is thought to justify
-------L the comparison of TBA values found with different
distillation procedures. Experimentally we found it dif-
5 10 15 20
Distillation rate in ml/min
ficult to establish a distillation procedure which was
stable and provided reliable and reproducible yields of
TBARS. This is particularly i m p o r t a n t for the compari-
Fig. 3a,b Comparison of different distillation principles and distilla-
tion rates. Mean TBA values labelled with the s a m e letter are not son of TBA values when the distillation procedure
significantly different (5% level), a Standards, simple distillation: * yields low amounts of TBARS and consequently a low
2 x 10.6 M MDA; 9 10 x 10-6 M MDA. Standards, steam distilla- percentage recovery as is typically seen at high distilla-
tion: [] 2 x 10.6 M MDA; 9 10 x 10-6 M MDA. b Meat samples, tion rates. As noted by Karastogiannidou and Ryley
simple distillation: * low oxidative levels; 9 high oxidative level.
Meat samples, steam distillation: [] low oxidative level; 9 high [15] there seems to be some inherent advantage in the
oxidative level reliability of a high reaction yield which, at the same
distillation rate, is shown to be obtained more easily
with simple distillation than with steam distillation.
C o n t r a r y to this, the extraction procedure has
10-
proven to be simple and very easy to operate. It was
30 possible to obtain clear filtrates from meat samples
with as much as 14-18% fat. Recoveries of M D A were
z" found to be a b o u t 100% and highly reproducible.
9
6- As mentioned earlier, the calculated TBA values in
z0
I this work have not been corrected for recovery of M D A
0~ and it is argued that due to the influence of procedural
4~ C~
-<
variables, the quantitative results from the TBA test are
~0
best evaluated without correction for recovery. As in all
empirical methods, it is essential to emphasize the im-
portance of a detailed description of all procedural
o-
Steam distill. Simple distill. Extraction details in order to evaluate the found TBA results
Kieltec 20 ml/min 4,2 ml/rnin
correctly. O u r results show that this should include the
Fig. 4 Comparison of two distillation procedures (rapid Kjeltec distillation principle as well.
steam distillation and simple distillation) and the extraction proced- T o our knowledge no single version of the TBA test
ure for both standards and meat samples. Mean TBA values labelled is suitable for application to all sample matrices, but
with the s a m e letter are not significantly different (5% level). Stan- considering the robustness of the extraction procedure
dards; 9 2x 10-6M MDA; [] 10x 10 6M MDA. Meat samples:
9 low oxidative level; [] high oxidative level c o m p a r e d to the distillation procedure, it is recom-
mended to use the extraction procedure with food
matrices such as non-cured meat products with less
than 1 4 - 1 8 % fat and no artificial colorants.
standard samples simple distillation gives significantly
lower TBA values than the extraction procedure, while,
for meat samples, no significant differences exist be- Acknowledgements We wish to thank Ib Skovgaard for helpful
advice with the experimental design and statistical treatment. This
tween the two procedures. In a food matrix this could work was financially supported by the Danish Ministry of Educa-
be due to insufficient extraction of the TBARS or tion and Research as a part of the FOTEK programme.
210

References 8. Siu GM, Draper HH (1978) J Food Sci 43:1147-1149

9. Salih AM, Smith DM, Price JF, Dawson LE (1987) Poult Sci 66:
1483-1488
1. Tarladgis BG, Watts BM, Younathan MT, Dugan LR (1960) 10. Shahidi F, Hong C (1991) J Food Biochem 15:97-105
J Am Oil Chem Soc 37:44-48 11. Kosngi H, Kojima T, Kikugawa K (1989) Lipids 24:873-881
2. Vyncke W (1975) Fette Seifen Anstrichm 77:239-240 12. Vyncke W (1970) Fette Seifen Anstrichm 72:1084 1087
3. Hoyland DV, Taylor AJ (1991) Food Chem 40:271 291 13. Tarladgis BG, Pearson AM, Dugan LR (1964) J Sci Food Agric
4. Hoyland DV, Taylor AJ, (1989) Int J Food Sci Technol 24: 35:1248-1254
153-161 14. Ke PJ, Cervantes E, Robles-Martinez C (1984) J Sci Food Agric
5. Gray JI, Monahan FJ (1992) Trends Food Sci Technol 3: 35:1248-1254
315-319 15. Karastogiannidou C, Ryley J (1994) Int J Food Sci Technol 29:
6. Draper HH, Squires EJ, Mahmoodi H, Wu J, Agarwal S, Had- 19-22
ley M (1993) Free Radical Biol Med 15:353-363 16. Rhee KS (1978) J Food Sci 43:1776-1781
7. Witte VC, Krause GF, Bailey ME (1970) J Food Sci 35:582-585