Lipid and Cholesterol Oxidation in Whole Milk Powder during
Processing and Storage
S. MC CLUSKEY, J.F. CONNOLLY, R. DEVERY, B. O’BRIEN, J. KELLY, D. HARRINGTON, and C. STANTON
ABSTRACT
The influence of animal feed quality on lipid and cholesterol oxidation
in whole milk powder was investigated. Powders from a herd receiving
a ‘supplemented’ diet showed reduced PV (p , 0.01) and TBARS (p ,
0.09) compared to a ‘restricted’ herd, after storage in both vacuum and
sachet-packs and less (p , 0.003) cholesterol oxidation products (COPs).
High pre-heating temperatures resulted in higher levels of PV, TBARS
and COPs in fresh whole milk powders than low pre-heat temperatures,
but after storage the reverse occurred. Superior animal feed quality and
proper control of processing and storage conditions enhanced oxidative
stability of whole milk powder. Lipid and cholesterol oxidation were
positively correlated (p , 0.001).
Key Words: milk powder, processing, cholesterol oxidation, TBARS,
peroxide value.
INTRODUCTION
CHOLESTEROL is an unsaturated lipid which is readily suscep-
tible to oxidation leading to chemically labile hydroperoxides,
which are decomposed to form secondary cholesterol oxidation
products (COPs). Various food processing and storage treat-
ments can lead to oxidation of cholesterol in the presence of
oxygen, heat, light or radiation to yield COPs. Although more
than 70 such oxidation products have been identified (Smith,
1981), only about eight have been detected in food (Sander et
al., 1989; Paniangvait et al., 1995). Major COPs in foodstuffs
are 25-hydroxycholesterol, cholestanetriol, hydroxycholesterol
derivatives, a- and b-epoxides and 7-ketocholesterol and their
presence in the human diet may have potentially negative health
implications (Emanuel et al., 1991). Several studies have impli-
cated COPs in a wide range of adverse biological effects both
‘in vivo’ and ‘in vitro’, including atherosclerosis, carcinogenesis,
mutagenicity and cytotoxicity (Bosinger et al., 1993).
Much research regarding cholesterol oxidation in foodstuffs
has focused on eggs and cheese (Tsai and Hudson, 1984; Mor-
gan and Armstrong, 1992; Nielsen et al., 1995). Some studies
have reported levels of COPs in whole milk powder (Nourooz-
Zadeh and Appelqvist, 1988; Sarantinos et al., 1993; Van de
Bovenkamp et al., 1988; Chan et al., 1993; Rose-Sallin et al.,
1995). These have shown that while no detectable levels of
COPs were evident in freshly prepared milk powders, aged
products contained varying amounts, the levels of which were
dependent on several factors, including drying technology and
storage conditions. Nourooz-Zadeh and Appelqvist (1988)
showed that while fresh milk powders prepared by spray-drying
in a low-medium heat process exhibited no detectable levels of
COPs, high-heat powders contained higher levels (2.8 ppm in
the lipid extract). Although systematic storage studies were not
undertaken, results indicated that storage duration and condi-
Authors Mc Cluskey, Connolly, Kelly, and Stanton are with Tea-
gasc, Moorepark, National Dairy Products Research Center, Fer-
moy, Co. Cork, Ireland. Author Devery is with the School of
Biological Sciences, Dublin City Univ., Glasnevin, Dublin 11, Ire-
land. Author O’Brien is with Teagasc, Dairy Husbandry Dept.,
Moorepark, Fermoy, Co. Cork, Ireland. Author Harrington is with
Teagasc HQ, 19 Sandymount Ave., Dublin 4, Ireland. Direct in-
quiries to Dr. Catherine Stanton, Teagasc, Moorepark, National
Dairy Products Research Center, Fermoy, Co. Cork, Ireland.
Volume 62, No. 2, 1997—JOURNAL OF FOOD SCIENCE—331
tions were important in COPs development. Subsequently, sev-
eral studies confirmed that reduced temperatures and exclusion
of oxygen reduced development of COPs in milk powders dur-
ing storage for prolonged periods (Sander et al., 1989; Chan et
al., 1993; Rose-Sallin et al., 1995). Previous findings indicated
that pre-heat treatment of milk at high heat (1207C) improved
oxidative stability compared to low-medium heat (80–957C)
(Tuohy, 1987; Van Mil and Jans, 1991). Heating leads to de-
naturation of whey proteins, especially b-lactoglobulin and thus
to exposure of sulfhydryl groups which can act as reducing
agents, resulting in a decrease of the autoxidation rate in bio-
logical and other systems (Taylor and Richardson, 1980; Walstra
FoodScience95 2126 Mp
and Jenness, 1984).
Based on several studies (Nourooz-Zadeh and Appelqvist,
1988; Van de Bovenkamp et al., 1988; Sarantinos et al., 1993;
Chan et al., 1993; Rose-Sallin et al., 1995) the most abundant
cholesterol oxides in aged whole milk powder are the 7-hydroxy
derivatives, the 7-keto-derivatives and the epoxides. There are
large discrepancies among reported values for COPs in dairy
products; such variations are attributable to the many methods
for COPs analysis (McCluskey and Devery, 1993). Variations
in temperature, light and oxygen availability during sample han-
331
dling all contribute to artifact formation; variations in mineral
composition among products also has an impact on ratio of 7-
Wednesday Apr 09 10:05 AM
keto to 7-hydroxycholesterol (Smith, 1981). In addition, stability
of cholesterol oxides is greatly influenced by pH (Kim and Na-
war, 1991). A standardized method for quantification of oxys-
terols has not been established. A variety of analytical methods
are used including HPLC with UV detection (Ansari and Smith,
1979; Sallin et al., 1993) and capillary GC (Park and Addis,
1985), which in conjunction with GC/MS is an effective ap-
proach for identification/confirmation (Rose-Sallin et al., 1995)
at low limits of detection with high specificity (Nielsen et al.,
1995).
In countries with a seasonal milk production pattern such as
Ireland, New Zealand and Australia, studies have shown a sea-
sonal variation in milk processability due primarily to seasonal
changes in animal diet and stage of lactation (Phelan et al. 1982;
Kefford et al., 1995). Variations in milk processability lead to
inconsistent quality of dairy products. The absolute criterion for
suitability of milk for processing is its stability during manufac-
ture and storage (often over extended periods). Antioxidants can
delay milkfat oxidation during storage (Touhy, 1987; Atwal et
al., 1991) and seasonal variations in animal feed quality have
been found to influence off-flavor development during storage
of whole milk powders (Steen, 1977). Our objective was to de-
termine the influence of animal feed quality on lipid and cho-
lesterol oxidation in stored whole milk powder. In particular,
the effects of pre-heating, packaging and storage conditions on
primary and secondary products of lipid and cholesterol oxida-
tion in whole milk powder produced from dietary restricted and
supplemented animals were investigated.
MATERIALS & METHODS
Reagents
Cholest-5-ene-3b-ol (cholesterol), 3b-hydroxycholest-5-en-7-one (7-
ketocholesterol), cholest-5-ene-3b, 25-diol (25-hydroxycholesterol), cho-
lesterol 5a, 6a-epoxide (a-epoxycholesterol), 5-cholestene-3b,19-diol
LIPID/CHOLESTEROL OXIDATION IN WHOLE MILK POWDER . . .

(19-hydroxycholesterol), cholestane-3b, 5a, 6b-triol (cholestanetriol),

(all of .99% purity) and the silylating agent N,O-bis (trimethylsilyl)
trifluoracetamide (BSTFA) and cysteine were purchased from Sigma (St.
Louis, MO) and b-epoxide (cholesterol 5b, 6b-epoxide) was obtained
from Steraloids Inc. (Wilton, NH). The cholesterol kit (Cat. No. 139
050) was supplied by Boehringer Mannheim (Mannheim, Germany). Ell-
man’s reagent (5, 5 dithiobis-(2-nitrobenzoic acid)) was purchased from
Aldrich-Chemie (Steinheim, Germany). All other reagents, including
thiobarbituric acid and ammonium thiocyanate were Analar grade and
were purchased from BDH (Dorset, UK). Chloroform and methanol were
HPLC grade, while all other solvents were Analar grade and were pur-
chased from Labscan Ltd. (Stillorgan, Co. Dublin, Ireland).

Herd rations
The cows (32) were assigned to two dietary treatments, denoted ‘re-
stricted’ (R) and ‘supplemented’ (S) for a 28-wk period (April–October).
During this time, the grass allocation was 4.0 and 3.57 cow/ha for ‘re-
stricted’ and ‘supplemented’ herds, respectively, with the ‘supplemented’
herd receiving, in addition, 3 kg/cow/day of concentrate supplement. The
concentrate supplement furnished 186 g/kg protein, 38 g/kg oil, 100 g/kg
crude fiber and 91 g/kg ash. The vitamin E concentration of the concen-
trate supplement was 7,500 IU/metric ton and supplementation was
therefore 22.5 IU vitamin E/cow/day. The herds were grazed separately
with a residence time of 3 days/plot.

Preparation of whole milk powders

Bulk milk from the two herds was collected at one morning and one
evening milking on 14 occasions and whole milk powder was manufac-
tured (Fig. 1). The milk from each herd was standardized (to yield whole
milk powder containing 26% fat) and pre-heated at low (757C 3 10 sec)
and high (1207C 3 2 min) heat. Milk standardization was achieved by
removal of a milkfat fraction by centrifugation to yield a percent fat
content (w/v) which was equivalent to 26% fat in dry matter. Milk con-
centrates underwent two-stage homogenization (80 bar, first stage; 20
bar, second stage) prior to spray-drying in a pilot-scale Anhydro Lab 3
spray-drier, with pneumatic nozzle atomization. The powders were sub-
sequently sachet-packed in foil-lined sample bags or vacuum-packed in
foil bags using a Webomatic vacuum packer and stored under ambient
(157C) or ‘hot room’ (307C) conditions in the dark up to 12 mo.
Determination of undenatured whey protein nitrogen
The undenatured whey protein nitrogen index (WPNI) was used to
categorize the powders into low-heat and high-heat classes. Low-heat
powder was classified as containing ≥ 6.0 mg undenatured whey pro-
tein/g powder and high-heat powder as containing ≤ 1.5 mg undenatured
whey protein/g powder. WPNI was determined using the method out-
lined by American Dry Milk Institute (ADMI, 1971). Undenatured whey
protein was expressed as mg undenatured whey protein/g powder.
Measurement of lipid oxidation
The peroxide value (PV) was determined using the ferric thiocyanate
method (IDF, 1991) on milkfat samples, isolated from reconstituted milk
powder using the procedure of Murphy (1978). Results were expressed
as mEq O2/kg fat. Secondary lipid oxidation products were measured by
the 2-thiobarbituric (TBA) method of Tarladgis et al. (1964). TBA re-
active substances (TBARS) were expressed as absorbance units at 530
nm.
Measurement of sulfhydryl groups and Vitamin E
Total sulfhydryl groups were measured according to the method out-
lined by Beveridge et al. (1974) using a 20% solution of whole milk
powder and reactive sulfhydryl groups according to the method of Kalab
(1970). Total and reactive sulfhydryl groups were expressed as µM/g
powder. Vitamin E concentration in the milk powder was determined
using the method outlined by Buttriss and Diplock (1984). Results were
expressed as mg vitamin E/100g milk powder.
Extraction and derivatization of COPs
The Boehringer Mannheim cholesterol assay kit (Cat No. 139 050)
was used for total cholesterol estimation. 19-Hydroxycholesterol was the
internal standard of choice, which was used to monitor recoveries and
Fig. 1—Schematic of whole milk powder manufacture. Milk was
standardized (to yield whole milk powder containing 26% fat) and
pre-heated at low (75&C) and high (120&C) heat. Milk concentrates
were homogenized prior to spray-drying in a pilot-scale Anhydro
Lab 3 spray-drier, with pneumatic nozzle atomization.
correct for losses during sample preparation for COPs analysis. The
method of Folch et al. (1957), as modified by Sander et al. (1989), was
used in the extraction of lipids for COPs analysis. Powder, (1g) contain-
ing 0.05 mg 19-hydroxycholesterol, was homogenized in 10 mL chlo-
roform/methanol (2/1 v/v) for 1 min; and 15 mL chloroform/methanol
(2/1 v/v) was used to rinse the homogenizer and re-extract the residue.
The homogenate was then filtered through a Whatman No. 1 filter,
washed with 5 mL distilled water and subsequently centrifuged at 1000
3 g for 20 min. The total lipid extract (bottom organic layer) was dried
in a rotary evaporator at 307C and subsequently redissolved in 5 mL
hexane. This was then applied to a Sep-Pak silica cartridge (Waters
Associates, Milford, MA, USA) which had been previously equilibrated
with 5 mL hexane. This was washed successively with 10 mL hex-
ane/diethyl ether (95:5 v/v), 25 mL hexane/diethyl ether (90:10 v/v) and
15 mL hexane/diethyl ether (80:20 v/v). The COPs fraction was then
eluted with 10 mL acetone and dried under nitrogen. The COPs fraction
was redissolved in 250 µL BSTFA and derivitization completed by heat-
ing at 1207C for 30 min.
Quantification of COPs
Separation of COPS was performed using a Varian 3500 gas liquid
chromatograph (Varian, Harbor City, CA, USA) fitted with a flame ion-
ization detector (FID). Separation was performed on a DB5 capillary
column (J & W Scientific, Folson, CA, USA) (30 m 3 0.32 mm i.d.,
1.0 mm film thickness). Samples were injected onto the column using a
SPI injector, using He as the carrier gas and N2 as make-up gas and a
flow rate of 5 mL/min. Injector temperature was 1007C increasing to a
final temperature of 3007C at 307C/min. Initial column temperature was
1007C with a hold time of 40 mins, then increased to 3007C at 107C/min.
The detector was at 3007C. Results were expressed as ppm in the lipid
extract.
GC/MS analysis of COPs
The structures of 19-hydroxycholesterol and 25-hydroxycholesterol
were confirmed by GC/MS, using a Varian Saturn 4D GC/MS (Varian,

332—JOURNAL OF FOOD SCIENCE—Volume 62, No. 2, 1997

Harbor City, CA). The conditions used were as previously outlined
(Rose-Sallin et al., 1995). Separation was performed on a DB5 fused
silica capillary column (J & W Scientific, Folson, CA) (30 m 3 0.32
mm i.d., 0.25 mm film thickness), following on-column injection, using
He as carrier gas at 0.7 bar. The initial oven temperature was 607C with
a hold time of 1 min, increasing at 307C/min to 2007C which was then
increased to 3007C at 57C/min and held for 5 min. The mass spectrom-
eter was operated under electron ionization conditions with an electron
energy of 70 ev.
Statistical analysis
Whole milk powder was manufactured on 14 occasions which were
grouped into three periods. The duration of each period was selected so
that powders made during a specific period were manufactured from milk
of similar composition, and thus could be used as replicates for statistical
analysis. Milk powder was manufactured on five occasions during period
1 (April 20–May 26), on five occasions during period two (June 16–July
21) and on four occasions during period three (August 3–October 21).
For each period COPs, WPNI, PVs and TBARS were analyzed as a
randomized block design with animal feed quality (‘restricted’ vs. ‘sup-
plemented’) and heat treatment as a factorial combination. COPs at 12
mo, PV and TBARS at 2, 4, 6 and 12 mo were analyzed as a split-plot
design with animal feed quality and heat treatment combined factorially
in the main plots and temperature and package type as the sub-plot
factorial combination. Sulfhydryl groups were analyzed as a randomized
block design with diet and heat treatment as the factors. This was done
for a mid-lactation period with milk powder made on four separate oc-
casions as replicates and for a late lactation period using three separate
occasions as replicates.
RESULTS
Undenatured whey protein nitrogen index
The effects of animal feed quality and pre-heating tempera-
tures on WPNI values of whole milk powders were compared
(Fig. 2). Animal feed quality had an effect on the levels of
undenatured whey protein during period 1 (p , 0.022) but had
no effect on powders produced in periods two and three. The
low-heat powders of the ‘restricted’ and ‘supplemented’ herds
had mean WPNI values of 6.9 5 0.9 and 7.8 5 1.1 mg un-
denatured whey protein/g powder, respectively, compared to
values of 1.2 5 0.3 and 1.1 5 0.2 mg undenatured whey pro-
tein/g powder, respectively, for corresponding high-heat pow-
ders. These results are similar to previous findings, where values
of 8.38 and 1.27–1.6 mg undenatured whey protein/g powder
were reported for low-heat and high-heat powders, respectively
(Sanderson, 1970; Van Mil and Jans, 1991). Levels of whey
proteins in raw milk vary significantly (Harland et al., 1955)
leading to slight variations in WPNI values, thereby limiting this
test as a means of evaluating the heat treatment of milk. All
low-heat and high-heat powders manufactured throughout the
study fell within the limits of each class according to the WPNI.
As expected, the high pre-heating temperature (1207C) caused
significantly more denatured protein (Fig. 2) than the low pre-
heating temperature (757C).
Lipid oxidation
PV and TBARS values for high-heat and low-heat powders
from the different experimental samples were compared (Fig. 3,
a–d). In general, milk powders from the ‘supplemented’ herd
exhibited lower lipid oxidation during storage, as indicated by
lower PV and TBARS values than powders of the ‘restricted’
herd, suggesting that animal feed quality had a protective effect
in PV and TBARS stability. The higher quality of the ‘supple-
mented’ diet, which included elevated levels of vitamin E com-
pared to the ‘restricted’ diet contributed to improved lipid
stability. The vitamin E concentration of the concentrate sup-
plement fed to the ‘supplemented’ herd was 7,500 IU/metric ton
or 22.5 IU vitamin E/cow/day. Vitamin E levels in the milk
powders from ‘restricted’ and ‘supplemented’ herds were 1.1
and 1.4 mg/100g, respectively. Animals supplemented with vi-
Volume 62, No. 2, 1997—JOURNAL OF FOOD SCIENCE—333
Fig. 2—Undenatured whey protein nitrogen index (WPNI) levels
in low-heat (a) and high-heat (b) whole milk powders of ‘re-
stricted’ and ‘supplemented’ herds. Powders from the ‘supple-
mented’ herd are represented by the continuous lines
( ) and powders from the ‘restricted’ herd by the broken
( ------ ) lines.
tamin E have shown increased milk vitamin E contents com-
pared to milk of grazing animals (Atwal et al., 1991). Several
studies have shown that vitamin E improved the stability of veal
fat, frozen poultry, milk and pork (Buckley et al., 1989; Buckley
and Connolly, 1980). Other dietary antioxidants may include
natural phenolic compounds found in the majority of dietary
plant materials and conjugated dienoic isomers of linoleic acid
found in milkfat. These are all known to show antioxidant ac-
tivity (Bonorden and Pariza, 1994) and increased lipid stability
of milk powder but they were not considered here.
Our data indicate that high pre-heating conditions (1207C for
2 min) led to improved oxidative stability in whole milk pow-
ders during subsequent storage than did low pre-heat treatment
(757C for 10 sec). The PV values for low-heat powders were
higher (p , 0.002) than those of high-heat powders after 6 mo
at the temperatures examined (157C and 307C; Fig. 3a–d). While
PVs increased seven-fold in low-heat powders, which were sa-
chet-packed and stored at 157C for 6 mo, an increase of only
3.5-fold was observed in high-heat powders under similar
packaging and storage (Fig. 3a). Similarly, TBARS levels were
reduced (p , 0.01) in high-heat powders compared to low-heat
powders during the 12-mo storage (Fig. 3). It has been shown
that high pre-heating temperatures resulted in less oxidation in
whole milk powder during subsequent storage than did low pre-
heat treatment (Boon, 1976; Tuohy, 1987).
The levels of ‘free’ sulfhydryl groups in milk powder were
also compared (Table 1). High pre-heat treatment of milk re-
sulted in increased levels of ‘free’ sulfhydryl groups (p , 0.002)
compared to low pre-heat treatment. The ‘free’ sulfhydryl con-
tent of high-heat powder of the ‘restricted’ herd was on average,
0.28 µM/g, while corresponding low-heat powders resulted in
only 0.07 µM/g powder. The inverse of this was found for ‘total’
sulfhydryl levels, which decreased as pre-heat temperature in-
creased. The ‘total’ sulfhydryl groups for low-heat powder of
the ‘restricted’ herd was 15.9 µM/g compared to 0.14 µM/g for
high-heat powder. The observation that heat treatment increased
‘free’ sulfhydryl groups, but decreased ‘total’ sulfhydryl groups
has been reported (Taylor and Richardson, 1980). Increased ox-
idative stability in whole milk powder has been observed as a
result of high pre-heat treatment, presumably due to protein de-
naturation and consequently increased levels of ‘free’ sulfhydryl
groups (Boon, 1976; Van Mil and Jans, 1991). In addition, high
pre-heat treatment may produce more Maillard browning reac-
tion products, which have been reported to contribute to in-
creased milkfat stability (Wyott and Day, 1965; Ericksson,
1982). Animal feed quality also had an effect (p , 0.05) on
levels of sulfhydryl groups in milk powder. Powder of the ‘sup-
LIPID/CHOLESTEROL OXIDATION IN WHOLE MILK POWDER . . .

Fig. 3—PV (---) and TBARS ( ) were monitored in high-heat and low-heat whole milk powders, manufactured from milk of ‘re-
stricted’ (R) and ‘supplemented’ (S) herds. a. Sachet-packed powders stored at 15&C; b. Sachet-packed powders stored at 30&C. c.
Vacuum-packed powders stored at 15&C. d. Vacuum-packed powders stored at 30&C.

plemented’ herd contained greater levels than the ‘restricted’ Table 1—Free and total sulfhydryl group concentrations in whole milk pow-
der
herd at both low and high-preheat temperatures (Table 1). Su-
lfhydryl groups are important in the antioxidant activity of milk Free sulfhydryl Total sulfhydryl
Herda Heatb µM/g powder µM/g powder
but other components, including the casein fraction of milk,
R low 0.07 15.91
which contains only a small amount of sulfhydryl groups, pro- S low 0.11 16.88
vide much of the antioxidant activity. Milk protein contents R high 0.28 0.14
were higher (p , 0.001) from the ‘supplemented’ herd and val- S high 0.46 0.16
ues ranged from 2.96–3.66% for ‘restricted’ herd milk and from n57
3.34–4.02% for ‘supplemented’ herd milk. In addition, casein a Herd R denotes a restricted grass supply; Herd S denotes a standard grass supply

levels were higher (p , 0.001) in milk from the ‘supplemented’
herd compared to the ‘restricted’ herd. This may also have been
a contributing factor to the increased oxidative stability in pow-
ders from this milk.
Storage temperature was important in lipid oxidation of milk
powder in our study, with higher storage temperatures leading
to increased PV levels (p , 0.001) and TBARS (p , 0.001)
(Fig. 3a, b). PV values increased at an accelerated rate compared
to TBARS, reaching a maximum at 6 mo at 157C and 307C for
all powders investigated, and declined thereafter (Fig. 3a–d). An
inverse relationship was observed between PV and TBARS after
6 mo at both 157C and 307C, indicating progression of oxidation
from a primary to a secondary state. An accelerated rate of PV
development was seen in powders stored at 307C compared to
plus a concentrate supplement.
b Low-heat: 75&C for 10 sec; High-heat: 120&C for 2 min.
storage at 157C in both sachet-packed and vacuum-packed prod-
ucts, although the rate was higher (p , 0.001) for sachet-packed
powders. Increasing the storage temperature from 157C to 307C
resulted in substantially higher PV in aged milk powders. A
seven-fold increase in PV values was observed due to storage
of low-heat powders at 157C for 6 mo but storage at 307C re-
sulted in a 10-fold increase (Fig. 3b). These results confirmed
previous studies which indicated that temperature was important
in the rate of autoxidation. A 107C increase in temperature re-
sulted in a 2-fold increase in reaction rate (Walstra and Jenness,
1984) and up to 2-fold increases in PV values occurred when

334—JOURNAL OF FOOD SCIENCE—Volume 62, No. 2, 1997


Fig. 4—Gas chromatogram of (a) standard mix of cholesterol ox-
ides, including the internal standard, 19-hydroxycholesterol, (b)
chromatogram of COPs in fresh low-heat whole milk powder and
(c) chromatogram of COPs in 12-month-stored low-heat whole
milk powder.
storage was increased from 227C to 307C (Boon, 1976). In ad-
dition, TBARS increased '3-fold during storage of whole milk
powder for 6 mo at 407C compared to 207C (Chan, 1992).
Vacuum-packaging of whole milk powder resulted in lower
PV and TBARS values compared to sachet-packing (p , 0.001)
during subsequent storage at 157C and 307C (Fig. 3c-d). Vac-
uum-packing reduces the oxygen content by evacuation of air
and compression of powder (Touhy, 1984). In our study,
TBARS values of sachet-packed powders of the ‘restricted’ herd
increased 5-fold during storage for 12 mo at 157C (Fig. 3a) and
15-fold during storage at 307C (Fig. 3b), while only three and
four-fold increases were observed when powders were vacuum-
packed and stored at 157C and 307C (Fig. 3c–d), respectively.
Chan et al. (1993) reported that oxidative stability of lipids and
cholesterol in whole milk powder could be maintained for ex-
tended periods by exclusion of oxygen during storage. However,
residual oxygen may still cause lipid oxidation of milk powder
in enclosed vacuum-packed systems. Tamsa et al. (1961) noted
that extremely low oxygen concentrations were required to
Volume 62, No. 2, 1997—JOURNAL OF FOOD SCIENCE—335
avoid oxidative flavors in reconstituted foam-spray-dried pow-
der.
The data indicated some evidence of a seasonal effect on lipid
oxidation in the milk powders. Animal feed quality had an effect
(p , 0.02) on PV values of fresh whole milk powders manu-
factured in periods one and two but no significant effect was
found for powders manufactured in period three. Grass quality
can deteriorate in late autumn/early winter, which can lead to a
late-lactation effect in the milk at that season. After 6 mo storage
of the powders, animal feed quality had an effect on PV levels
for powders manufactured in period 1 (p , 0.001), but no effect
on those produced in periods two and three. An effect (p ,
0.09) was observed on TBARS levels of fresh powders produced
in periods one and two but none was observed on powders pro-
duced in period three. In contrast to PV values, TBARS for
powders from the ‘supplemented’ herd were lower (p , 0.09)
for those produced in all three periods than those of the ‘re-
stricted’ herd after 12 mo storage (Fig. 3). This indicated that
improved animal feed quality was protective against lipid oxi-
dation during storage of whole milk powder.
Cholesterol oxidation
‘Total COPs’ refers to the quantification of 5 of the most
commonly found COPs in whole milk powder (a- and b-epox-
ides, 7-ketocholesterol, 25-hydroxycholesterol, and cholestane-
triol). These were reviewed by Paniangvait et al. (1995). A
typical GC chromatogram of the standard COPs mixture (Fig.
4a) was based on 19-hydroxycholesterol as the internal standard.
The limit of detection of this method for COPs quantification is
0.1 ppm in the lipid extract. A chromatogram of a fresh low-
heat whole milk powder, manufactured from milk of the ‘sup-
plemented’ herd during period one (Fig. 4b) exhibited no
detectable levels of COPs. A chromatogram of a similar powder,
stored for 12 mo (Fig. 4c) indicated the appearance of quanti-
fiable levels of several cholesterol oxides. The structures of
some of these COPs was confirmed by GC/MS including 19-
hydroxycholesterol and 25-hydroxycholesterol (Fig. 5a, b).
The total COPS levels were compared (Fig. 6) in freshly man-
ufactured and 12-mo-stored low-heat and high-heat powders
from milk of ‘restricted’ and ‘supplemented’ herds. Animal feed
quality had no significant effect on levels of COPs in fresh
whole milk powders but upon storage a significant effect was
found. Low-heat and high-heat powders from milk of the ‘re-
stricted’ herd had higher (p , 0.003) COPs levels than corre-
sponding powders from the ‘supplemented’ herd after 12 mo
storage. This was found for both sachet-packed and vacuum-
packed powders which were stored at 157C and 307C. Thus, the
‘supplemented’ diet had a protective effect against cholesterol
oxidation during whole milk powder processing.
High pre-heat treatment resulted in higher levels (p , 0.001)
of COPs than low pre-heat treatment in freshly prepared milk
powders. Values for total COPs of fresh high-heat powders of
the ‘restricted’ herd were 1.6 5 0.55 ppm in the lipid extract,
compared to 0.66 5 0.24 ppm in the lipid extract for low-heat
powders (Fig. 6). COPs levels of the corresponding high-heat
and low-heat powders of the ‘supplemented’ herd were 1.07 5
0.60 and 0.43 5 0.24 ppm in the lipids, respectively. Nourooz-
Zadeh and Appleqvist (1988) reported higher COPs levels in
freshly prepared high-heat powders (2.8 ppm in the lipids) com-
pared to low-heat powders (none detectable).
Pre-heating temperatures and storage conditions had an effect
on subsequent COPs development during storage of milk pow-
ders for 12 mo. The fresh high-heat powders showed higher
levels of COPs than corresponding low-heat powders. However,
upon storage COPs development was greater (p , 0.001) in the
low-heat powders than in the high-heat powders (Fig. 6). Sa-
chet-packed low-heat powder of the ‘restricted’ herd developed
the highest levels of cholesterol oxides after 12 mo storage at
307C, increasing from 0.66 to 37.20 ppm in lipids upon storage
for 12 mo. Values of only 11.52 ppm were reached in high-heat
LIPID/CHOLESTEROL OXIDATION IN WHOLE MILK POWDER . . .

Fig. 6—Total COPs in fresh and stored (12 months) low-heat and
high-heat whole milk powder manufactured from milk of ‘re-
stricted’ and ‘supplemented’ herds.

Fig. 5—Mass spectra of (a) 25-hydroxycholesterol and (b) 19-hy-

droxycholesterol as trimethyl silyl (TMS)-ethers identified in 12-
month-stored low-heat whole milk powder.

powders under similar packaging and storage conditions (Fig.

6). As seen for lipid oxidation, cholesterol oxidation in milk
powders also increased (p , 0.001) upon storage at 307C com-
pared to 157C (Fig. 6). Similar to the data for lipid oxidation,
cholesterol oxidation was enhanced by storage in sachet-packs
compared to storage under vacuum (Fig. 6). The increased in-
cidence of COPs in sachet-packed compared to vacuum-packed
powders may be attributed to the greater volume of headspace
oxygen in sachet-packs.
A linear relationship was observed between lipid oxidation, Fig. 7—Correlation between COPS and TBARS in whole milk
quantified by the TBA method and cholesterol oxidation with powder stored for 12 months, (a) sachet-packed and (b) vacuum-
correlation coefficients of 0.73 (p , 0.001) and 0.55 (p , 0.001) packed.
for sachet-packed and vacuum-packed powders, respectively
(Fig. 7a and b). This indicated that the oxidative rate of choles-
terol during storage of whole milk powder coincided with (Nourooz-Zadeh and Appelqvist, 1988; Morgan and Armstrong
TBARS and was more highly correlated for sachet-packed pow- 1992; Sander et al., 1989; Chan et al., 1993; Rose-Sallin et al.,
ders. Previously, secondary oxidation products (TBARS) have 1995) which further contributes to variations reported in COPs
been positively correlated with COPs production during storage values.
in cooked pork (Monahan et al., 1992) and whole milk powder
(Chan et al., 1993). COPs production in dairy spreads has been
CONCLUSIONS
linked to the early stages of lipid oxidation (Nielsen et al.,
1996). IN MILK POWDERS stored up to 12 mo, primary stage oxidation
The different methods reported for COPs quantification may was dominant during the first 6 mo and secondary stage oxi-
make unreliable any direct comparisons between the data from dation dominated during the subsequent 6 mo storage. Milk
different studies. Our study employed the method of Sander et powders from the ‘supplemented’ herd resulted in less protein
al. (1989) for COPs quantification by GC and yielded similar denaturation in low-heat powders and higher levels of reactive
values to those reported by Rose-Sallin et al. (1995) for milk sulfhydryl groups than those from the ‘restricted’ herd. This may
powder stored at ambient temperature for 12-mo, although there have contributed to the lower levels of lipid and cholesterol
were considerable variations between methods. Variation also oxidation found in fresh and stored powders from the ‘supple-
exist in the standard mix used for total COPs quantification mented’ herd. The elevated antioxidant levels in whole milk

336—JOURNAL OF FOOD SCIENCE—Volume 62, No. 2, 1997

powder from cows fed a ‘supplemented’ diet undoubtedly also
contributed to reduced lipid and cholesterol oxidation. High pre-
heating temperatures resulted in higher levels of COPs in fresh
whole milk powders, but were more stable during subsequent
storage. Both lipid and cholesterol oxidation increased under all
conditions investigated, but these undesirable ageing defects
could be further minimized by exclusion of oxygen and reduced
storage temperatures.
REFERENCES
ADMI. 1971. Standards for Grades of Dry Milk Including Methods of Anal-
ysis, Bulletin 916. American Dry Milk Institute, Inc., Chicago.
Ansari, G.A.S. and Smith, L.L. 1979. High-performance liquid chromatog-
raphy of cholesterol autoxidation products. J. Chromatog. 175: 307–315.
Atwal, A.S., Hidiroglou, M., and Kramer, J.K.G. 1991. Effect of feeding Pro-
tect and a-tocopherol on fatty acid composition and oxidative stability of
cow’s milk. J. Dairy Sci. 74: 140–145.
Beveridge, T., Toma, S.T., and Nakai, S. 1974. Determination of SH- and
SS- groups in some food proteins using Ellman’s reagent. J. Food Sci. 39:
49–51.
Bonorden, W.R. and Pariza, M.W. 1994. Antioxidant nutrients and protec-
tion from free radicals. Nutritional Toxicology. F.N. Kotsonis, M. Mackey,
and J. Hjelle (Ed.), p. 19–48. Raven Press Ltd., New York.
Boon, P.M. 1976. The effect of pre-heat treatment on the storage stability of
whole milk powder. N. Z. J. Dairy Sci. Tech. 2: 278–280.
Bosinger, S., Luf, W., and Brandl, E. 1993. ‘Oxysterols’: Their occurrence
and biological effects. Int. Dairy Journal 3: 1–33.
Buckley, J. and Connolly, J.F. 1980. Influence of alpha-tocopherol (vitamin
E) on storage stability of raw pork and bacon. J. Food Protect. 43: 265–
267.
Buckley, D.J., Gray, J.J., Asgar, A., Price, J.F., Crackel, R.L., Boren, A.M.,
Pearson, A.M., and Miller, E.R. 1989. Effects of dietary antioxidants and
oxidized oil on membranal lipid stability and pork lipid quality. J. Food
Sci. 54: 1193–1197.
Buttriss, J.L. and Diplock, A.T. 1984. High-performance liquid chromatog-
raphy methods for vitamin E in tissues. Methods Enzymol. 105: 131–138.
Chan, S.H. 1992. Cholesterol oxidation in whole milk powder as influenced
by processing and packaging. MSc. thesis, Michigan State Univ., East Lan-
sing, MI.
Chan, S.H., Gray, J.I., Gomma, E.A., Harte, B.R., Kelly, P.M., and Buckley,
D.J. 1993. Cholesterol oxidation in whole milk powders as influenced by
processing and packaging. Food Chem. 47: 321–328.
Emanuel, H.A., Hassel, C.A., Addis, P.B., Bergmann, S.D., and Zavoral, J.H.
1991. Plasma cholesterol oxidation products (oxysterols) in human subjects
fed a meal rich in oxysterols. J. Food Sci. 56: 843–849.
Ericksson, C.E. 1982. Lipid oxidation catalysts and inhibitors in raw mate-
rials and processed foods. Food Chem. 9: 3–13.
Folch, J., Lees, M., and Sloane-Stanley, G.H. 1957. A simple method for the
isolation and purification of total lipids from animal tissues. J. Biol. Chem.
226: 497–509.
Harland, H.A., Coulter, S.T., and Jenness, R. 1955. Natural variation of milk
serum proteins as a limitation of their use in evaluating the heat treat-
ment of milk. J. Dairy Sci. 38: 858–869.
IDF. 1991. Standard 74A. International Dairy Federation, Brussels, Bel-
gium.
Kalab, M. 1970. Factors affecting the Ellman determination of sulfhydryl
groups in skimmilk powder and gels. J. Dairy Sci. 53: 711–717.
Kefford, B., Christian, M.P., Sutherland, B.J., Mayes, J.J., and Grainger, C.
1995. Seasonal influences on Cheddar cheese manufacture: influence of
diet quality and stage of lactation. J. Dairy Res. 62: 529–537.
Kim, S.K. and Nawar, W.W. 1991. Oxidative interactions of cholesterol with
triacylglycerols. J.O.A.C.S. 68: 931–934.
Mc Cluskey, S. and Devery, R. 1993. Validation of chromatographic analysis
of cholesterol oxides in dried foods. Trends Food Sci. Technol. 4: 175–178.
Monahan, F.J., Gray, J.I., Boren, A.M., Miller, E.R., Buckley, D.J., Morris-
sey, P.A., and Gomma, E.A. 1992. Influence of dietary treatment on lipid
and cholesterol oxidation in pork. J. Agric. Food Chem. 40: 1310–1315.
Volume 62, No. 2, 1997—JOURNAL OF FOOD SCIENCE—337
Morgan, J.N. and Armstrong, D.J. 1992. Quantification of cholesterol oxi-
dation products in egg yolk powder spray-dried with direct heating. J.
Food Sci. 57: 43–45.
Murphy, J.J. 1978. Biochemical studies on lipolysis in bovine milk and dairy
products. Ph. D. thesis, Univ. College, Cork, National University of Ire-
land.
Nielsen, J.H., Olsen, C.O., Duedahl, C., and Skibsted, L.H. 1995. Isolation
and quantification of cholesterol oxides in dairy products by selected mon-
itoring mass spectrometry. J. Dairy Res. 62: 101–113.
Nielsen, J.H., Olsen, C.E., Jensen, C., and Skibsted, L.H. 1996. Cholesterol
oxidation in butter and dairy spread during storage. J. Dairy Res. 63: 159–
167.
Nourooz-Zadeh, J. and Appelqvist, L.A. 1988. Cholesterol oxides in Swedish
foods and food ingredients: milk powder products. J. Food Sci. 53: 74–87.
Paniangvait, P., King, A.J., Jones, A.D., and German, B.G. 1995. Cholesterol
oxides in foods of animal origin. J. Food Sci. 60: 1159–1174.
Park, S.W. and Addis, P.B. 1985. HPLC determination of C-7 oxidized cho-
lesterol derivatives in foods. J. Food Sci. 50: 1437–1441.
Phelan, J.A., O’Keefe, M.K., Keogh, M.K., and Kelly, P.M. 1982. Seasonal
changes in the composition of Irish milk. Ir. J. Fd. Sci. Technol. 6: 1–11.
Rose-Sallin, C., Huggett, A.C., Bosset, J.O., Tabacchi, R., and Fay, L.B. 1995.
Quantification of cholesterol oxidation products in milk powders using
[2H7] cholesterol to monitor cholesterol autoxidation artifacts. J. Agric.
Food Chem. 43: 935–941.
Sallin, C., Baumann, E., Butikofer, U., Sieber, R., and Bosset, J.O. 1993.
Possibilities and limitations of reversed-phase HPLC techniques. Trav.
Chim. Alimant. Hyg. 84: 141–157.
Sander, B.D., Smith, D.E., Addis, P.B., and Park, S.W. 1989. Effects of pro-
longed and adverse storage conditions on levels of cholesterol oxidation
products in dairy products. J. Food Sci. 54: 874–877.
Sanderson, W.B. 1970. Seasonal variations affecting the determination of
the whey protein nitrogen index of skim milk powder. N. Z. J. Dairy Sci.
and Tech. 5: 48–52.
Sarantinos, J., O’Dea, K., and Sinclair, A.J. 1993. Cholesterol oxides in Aus-
tralian foods: Identification and quantification. Food Australia 45: 485–
490.
Smith, L.L. 1981. Cholesterol Oxidation. Plenum Press, New York.
Steen, K. 1977. Seasonal variations in the composition of dry whole milk
and its keeping qualities under storing. Maelkeritidende 90: 126–140,
152–168.
Tamsa, A., Pallansch, N.J., Mucha, T.J., and Patterson, W.I. 1961. Factors
relating to the flavor stability of foam-dried whole milk. J. Dairy Sci. 44:
1644–1649.
Tarladgis, B., Pearson, A.M., and Duggan, L.R. (Jr). 1964. The chemistry of
the 2-thiobarbituric acid test for the determination of oxidative rancidity
in foods. J.A.O.C.S. 39: 34–39.
Taylor, M.J. and Richardson, T. 1980. Antioxidant activity of skim milk:
Effect of heat and resultant sulfhydryl groups. J. Dairy Sci. 63: 1783–1795.
Tsai, L.S. and Hudson, C.A. 1984. Cholesterol oxides in commercial dry egg
products. Isolation and identification. J. Food Sci. 49: 1245–1248.
Tuohy, J.J. 1984. Developments in the packaging of whole milk powder.
Farm and Food Res. 15: 7–9.
Tuohy, J.J. 1987. Milkfat autoxidation in whole milk powder during storage.
Ph. D. thesis, Univ. College Cork, National University of Ireland.
Van de Bovenkamp, P., Kosmeijer-Schuil, T.G., and Katan, M.B. 1988.
Quantification of oxysterols in Dutch foods. Egg products and mixed diets.
Lipids 23: 1079–1085.
Van Mil, P.J.J.M. and Jans, J.A. 1991. Storage stability of whole milk pow-
der: Effects of process and storage conditions on product properties. Neth.
Milk Dairy J. 45: 145–167.
Walstra, P. and Jenness, R. 1984. Dairy Chemistry and Physics. Wiley &
Sons, New York.
Wyott, C.J. and Day, E.A. 1965. Evaluation of antioxidants in deodorized
and non-deodorized butteroil stored at 307C. J. Dairy Sci. 48: 682–686.
Ms received 5/10/96; revised 9/1/96; accepted 10/1/96.
We thank Drs. Brendan O’Kennedy and John Murphy of Teagasc, Moorepark Research
Center for helpful discussions. The technical assistance of Seamus Aherne and E. O’Neill is
gratefully acknowledged as well as the assistance of Ester O’Mahony of the Nutrition Dept.,
University College, Cork for the vitamin E analyses. This work was supported by a grant
from the research levy contributed by Irish dairy farmers.