论文翻译

论文如下

trans Fatty Acids. 4. Effects on Fatty Acid Composition
of Colostrum and Milk
Jan Peltersen and Johannes Opstvedt*
Norwegian Herring Oil and Meal Industry Research Institute, Bergen, Norway

trans Isometric fatty acids of partially hydrogenated fish
oil (PHFO) consist of trans 20:1 and trans 22:1 in addition
to the trans isomers of 18:1, which are abundant in
hydrogenated vegetable oils, such as in partially
hydrogenated soybean oil (PHSBO). The effects of dietary
trans fatty acids in PHFO and PHSBO on the fatty acid
composition of milk were studied at 0 {colostrum} and 21
days postpartum in sows. The dietary fats were PHFO
(28% trans), or PHSBO (36% trans) and lard. Sunflower
seed oil (4%) was added to each diet. The fats were fed
from three weeks of age throughout the lactation period
of Experiment 1. In Experiment 2 PHFO or "fully"
hydrogenated fish oil (HFO) t19% trans), in comparison
with coconut oil (CF) (0% trans}, was fed with two levels
of dietary linoleic acid, 1 and 2.7%, from conception
throughout the lactation period. Feeding trans~contalning
fats led to secretion of trans fatty acids in the milk lipids.
Levels of trans 18:1 and trans 20:1 in milk lipids, as
percentages of total c/s + trans 18:1 and c/s + trans 20:1,
respectively, were about 60% of that of the dietary fats,
with no significant differences between PHFO and
PHSBO. The levels were similar for colostrum and milk.
Feeding HFO gave relatively less trans 18:1 and trans
20:1 fatty acids in milk lipids than did PHFO and
PHSBO. Only low levels of cis + trans 22:1 were found
in milk lipids. Feeding trans-containing fat had no consistent
effects on the level of polyenoic fatty acids but
reduced the level of saturated fatty acids and increased
the level of cis + trans monoenoic fatty acids. Increasing
the dietary level of linoleic acid had no effect on the
secretion of trans fatty acids but increased the level of
linoleic acid in milk. The overall conclusion was that the
effect of dietary fats containing trans fatty acids on the
fat content and the fatty acid composition of colostrum
and milk in sows were moderate to minor.
Lipids 26, 711-717 11991).

Increasing attention has been given to the nutritive value
of human milk. As milk is the sole or main source of
nutrients for an infant during the first six months of life,
it is important that the nutrients in milk meet the
substantial requirements of the rapidly developing infant.
The milk should contain sufficient quantities of essential
fatty acids to meet the demand for synthesis of membrane
tissues in the central nervous and vascular systems.
The content of long-chain polyenoic fatty acids le.g.,
22:6n-3) is particularly important for the development
*To whom correspondence should be addressed at: Norwegian Herring
Oil and Meal Industry Research Institute, N-5033 Fyllingsdalen,
Bergen, Norway.
Abbreviations: CF, coconut oil; EFA, essential fatty acids; GLC,
gas-liquid chromatography; HFO, hydrogenated fish oil; HPLC, highperformance
liquid chromatography; PHFO, partially hydrogenated
fish oil; PHSBO, partially hydrogenated soybean oil; PUFA, polyunsaturated
fatty acids; SFO, sunflower seed oil; VLDL, very low density
lipoproteins.

of the brain and the retina (1}. Since in vitro studies have
indicated that trans fatty acids may influence essential
fatty acid metabolism, the effect on the fatty acid composition
of human milk from ingestion of trans-containing
partially hydrogenated fats has been of concern (2}.
Although the transfer of trans fatty acids from diet and
the adipose tissue into milk has been demonstrated in
humans {3-8} and rats {9-11}, no significant trans fatty
acid effects on the content of essential fatty acids have
been found. The studies used elaidic acid (transoctadecenoic
acid, trans 18:1n-9} or partially hydrogenated
vegetable fats containing trans fatty acids with chain
length of 18 carbon atoms or less. However, partially
hydrogenated fish oils contain trans fatty acids with a
chain length of 20 and 22 carbon atoms {12}, and the
biological effects of these acids are only poorly
understood. We have previously reported the effects of
dietary trans fatty acids in PHFO in comparison with
PHSBO on brain weight, peroneal histology and nerve
conduction velocity (13), and on the fatty acid profile of
brain lipids and other organ lipids in sows (14) and offsprings
(15). The present work was conducted to study
the extent to which dietary trans fatty acids, and especially
long-chain {i.e., with 20 and 22 carbon atoms} trans monoenoic
fatty acids, are incorporated into milk lipids under
chronic dietary regimens (i.e., when the contents of trans
fatty acids have reached a steady state condition). It was
further intended to find out if dietary trans fatty acids
affect the overall fatty acid composition of milk lipids and
essential fatty acid (EFA) content. Since an interrelationship
between the metabolism of trans fatty acids and EFA
has been indicated (16}, the trans fatty acids were fed at
two levels of dietary linoleic acid. Because of the many
physiological similarities between pig and humans (17,18},
we used pigs as models in these studies.

MATERIALS AND METHODS
Generalprocedures. A composite sample of about 10 mL
of milk was obtained from all lactating teats of each
primiparous sow on day 0 (colostrum) and 21 days
postpartum. Experiment 1 used four sows that had been
fed 14 wt% of lard (control, diet 1}, PHFO (partially
hydrogenated capelin oil, mp 30-32~ diet 2} or PHSBO
{partially hydrogenated soybean oil, mp 40~ diet 3) plus
4 wt% of SFO {sunflower seed oil, 65% linoleic acid} as
additions to diets otherwise composed of solvent extracted
soybean meal and cereals. The diets were fed from
three weeks of age throughout the lactation period.
Experiment 2 used four sows that had been fed according
to a 2 X 3 factorial design with 14 wt% of CF
{coconut fat, diets 1.1 and 2.1}, PHFO (diets 1.2 and 2.2}
or HFO Ifully hydrogenated fish oil, mp 50-52~ diets
1.3 and 2.3}, either without (diets 1.1, 1.2 and 1.3) or with
{diets 2.1, 2.2 and 2.3) an extra 3 wt% addition of SFO
to diets otherwise composed of solvent extracted soybean
meal and cereals. The diets were fed from gestation at
about six months of age throughout the lactation period.

J. PETTERSEN AND J. OPSTVEDT
In both experiments the sows were held in individual pens
and offered the feed twice daily in amounts determined
by their weights, stage of pregnancy and number of offspring
{feeding standard}. The design of the experiments
and the fatty acid composition of the diets are shown in
Table 1. Details of diets, feeding and management have
been described earlier {13).
Chemical methods. Lipids from milk and colostrum
were prepared by diethyl-ether extraction {19}, and the
lipid extracts were stored at -25~ pending further
analysis. Total fat content was determined gravimetrically.
Fatty acid composition was analyzed by capillary gasliquid
chromatography {GLC} using a 50-m fused silica
capillary column (CP Sfl 88, Chrompack, Middleburg, The
Netherlands) as described earlier (14). The content of trans
16:1 was only analyzed in the dietary fats of Experiment
2 (Table 1).
Statistical methods. The statistical model used was a
linear mixed model with diets as fixed, and sows as random
effects. The various subgroups contained varying
numbers of experimental units, making the statistical
estimation and testing procedures numerically rather complex.
The computer program selected among the few
available at the time was General Mixed Model Analysis
of Variance (BMDP P3V) (20). Estimation is Maximum
Likelihood, and testing of effects uses chisquares in
Likelihood Ratio Tests. Other methods used were simple
one-way analysis of variance models with varying group
sizes. The pooled standard deviation, rather than the individual
group standard deviations, was used in assessing
the variability. Since testing of effects uses a one~way
analysis of variance model, the underlying assumption
that groups may vary according to location, but not in
terms of spread, implies that the estimate of spread, found
by MS error, comprises the total information on variability.
Accordingly, in assessing differences in the group
means the standard error of the different means were computed,
making use of this common measure of spread for
individual experimental units.