Human Nutrition and Metabolism

Gender May Affect the Action of Garlic Oil on Plasma Cholesterol

and Glucose Levels of Normal Subjects1,2,3
Xiao-Hua Zhang, Derek Lowe,* Paul Giles, Stephen Fell,
Martin J. Connock** and David J. Maslin4
Division of Biomedical Sciences and Human Biology, School of Health Sciences, University of
Wolverhampton, Wolverhampton WV1 1DJ, UK; *Astraglobe Limited, Mossley, Cheshire, UK;

Biochemistry Department, Walsall NHS Trust, The Manor Hospital, Walsall WS2 9PS, UK;
and **School of Applied Sciences, University of Wolverhampton, Wolverhampton WV1 1SB, UK

KEY WORDS:

plasma lipids

antioxidants

gender

Reduction in coronary heart disease (CHD)5 risk is an

objective of affluent societies and has become a major reason

garlic

normal men and women

Downloaded from jn.nutrition.org by on October 19, 2010

ABSTRACT Early trials of garlic preparations on blood lipids mainly supported a lipid-lowering effect, whereas later
well-designed garlic tablet trials were mainly entirely null. However, enteric simulation tests suggest that this discordance may result from ineffective delivery of bioactive agents from the brands of garlic powder (GP) and cyclodextrinbound garlic oil (GO) tablets tested in some recent negative trials. In contrast, enteric simulation tests show that the
preformed bioactive agents present in traditional gelatin capsules of GO are efficiently released, although such
capsules have rarely been investigated in lipid-lowering trials. It was hypothesized that gelatin capsules of GO given to
normal subjects would improve specified coronary heart disease risk factors. Effects of a GP preparation were also
investigated. Subjects (n 51; men and women, mean age 27 y) were randomly assigned to receive either 8.2 mg/d of
GO (allyl sulfides) or placebo for 11 wk. Another 27 subjects received garlic powder (GP) of similar biopotential (7.8 mg
allicin/d). Outcome measures were 95% confidence intervals (CI) between GO and placebo groups for differences
between baseline and subsequent sample times. Men and women combined showed no significant differences save for
an improved total antioxidant capacity at 6 wk (P 0.01). Hence, no benefit from GO after 11 wk is one plausible
conclusion. However, there were significant differences in effect of GO between men and women for HDL cholesterol
(HDL-C) (P 0.004) and total cholesterol (TC)/HDL-C (P 0.003). Women showed favorable effects in terms of CHD
risk factors (i.e., increases in HDL-C and reductions in TC/HDL-C), whereas men had small adverse effects. There was
a significant difference in the GO effect for glucose (P 0.006), with a reduction seen for men and an increase for
women. The gender effects were unexpected and such analyses were not planned in advance. Confirmation of these
findings with larger numbers of subjects would have importance for the use of garlic against CHD and for the design of
future garlic studies. J. Nutr. 131: 14711478, 2001.

for garlic consumption (1). The pharmacologic activities of

garlic, including its claimed lipid-lowering effects, depend
upon sulfur-rich compounds, most notably allicin (2,3). Allicin and its derivatives give rise to the characteristic odor and
taste of garlic (3). Allicin (allyl 2-propene thiosulfinate) is
formed instantaneously when garlic is crushed because cellular
disruption brings the enzyme alliinase into contact with its
substrate alliin (allylcysteine sulfoxide). Allicin itself has limited stability; thus, commercial products containing garlic
powder (GP), the preparations used in most human trials (2),
rely for their allicin delivery, not upon allicin content, but
upon the retention of their alliin and alliinase intact until the
intestine is reached. Heat, as applied during cooking or the
commercial steam distillation of garlic to produce steam-distilled garlic oil (GO), converts the allicin of crushed garlic to
allyl sulfides (2).
Numerous studies have demonstrated biological activity for
allicin and a variety of its derivatives including allyl sulfides
(2,3). However, in vivo, such garlic derivatives have no detectable metabolites in blood or urine (3) so that direct pharmacokinetic studies are lacking, and the active components of

1
Presented in part in abstract form at the XXIII Annual Conference, Association
of Clinical Biochemists, 2730 December 1996, Kottayam, India [Maslin, D. J. &
Zhang X.-H. (1996) The effects of garlic oil upon the plasma lipid profile of healthy
humans. Abs. 006, p. 5] and [Zhang X.-H., Maxwell, S.R.J., Thorpe, G.H.G., Thomason, H., Rea, C. A., Connock, M. J. & Maslin, D. J. (1997) The action of garlic oil
upon plasma total antioxidant capacity. Biochem. Soc. Trans. 25: 523s (abs.)].
2
Doses and funding from Seven Seas, Limited (UK) and from BioCare Limited
(UK).
3
See NAPS document No. 05595 for two pages of supplementary material.
This is not a multi-article document. Order from NAPS c/o Microfiche Publications, 248 Hempstead Turnpike, West Hempstead, New York 11552. Remit in
advance in U.S. funds only $15.00 for photocopies or $5.00 for microfiche. There
is a $25.00 invoicing charge on all orders filled before payment. Outside U.S. &
Canada add postage of $4.50 for the first 20 pages and $1.00 for each ten pages
thereafter, or $5.00 for the first microfiche and $1.00 for each fiche thereafter.
4
To whom correspondence should be addressed.
E-mail: D.Maslin@wlv.ac.uk.
5
Abbreviations used: AMPK, AMP-dependent kinase; CHD, coronary heart
disease; CI, confidence interval; GO, garlic oil; GP, garlic powder; HDL-C, HDL
cholesterol; HMGR, hydroxy-methyl-CoA-reductase; HRT, hormone-replacement
therapy; LDL-C, LDL cholesterol; TAOC, total antioxidant capacity; TC, total
cholesterol; TG, triglycerides.

0022-3166/01 $3.00 2001 American Society for Nutritional Sciences.

Manuscript received 20 September 2000. Initial review completed 26 October 2000. Revision accepted 5 February 2001.
1471

1472

ZHANG ET AL.

of insulin were observed in ethanol-fed rats given GO (24,25),

although when given in very large doses (100 mg/kg), a major
constituent of GO, diallyl disulfide, was reported to be hyperglycemic (26).
In vitro and whole-animal studies indicate an antioxidant
action of garlic (2,27). Reduced susceptibility of LDL to copper-catalyzed oxidation was reported for samples from healthy
human subjects administered Kwai GP tablets in an early study
(28). Negative results with Kwai that followed (16,29) are
questionable because of likely low allicin bioavailability (see
above, 15) which may also explain the inability of Harris to
replicate his earlier positive findings (Harris W. S., Lipid and
Diabetes Research Center, Saint Lukes Hospital, Kansas City,
MO, personal communication). Null effects of GO on plasma
and LDL total antioxidant capacity (TAOC) were recently
reported for male runners (30).
Aspects of dose strategy may be important factors influencing the response of human subjects to garlic, but few trials
have compared different dose levels (3133). Moreover, most
trials included in meta-analyses have provided garlic only as a
single daily dose of 600 900 mg Kwai GP (equivalent to
1.52.3 g fresh clove with allicin potential of 3.6 5.4 mg).
However, these garlic doses could be rather low because, on
the basis of traditional knowledge, daily doses of 4 g fresh
garlic cloves or 8 mg volatile oil [i.e., GO] has been recommended for qualification as a pharmaceutical garlic product
(34).
Garlic dose periods of 3 mo or less have been used in most
lipid-lowering studies (2), and the greatest cholesterol-lowering response has been reported for dose periods of this duration
(1). Longitudinal studies (35,36) have been infrequent, and
data describing dose washout responses are unavailable, thus
potentially limiting the validity of garlic crossover trials
(13,16,33,37,38). Few trials of GO (13, 39) and only one other
randomized controlled trial of GO in its traditional gelatin
capsule format (in which allicin-derived sulfides are diluted in
vegetable oil) (30) have been reported. This is remarkable
because GO is long-established as a garlic preparation (40) and
tests indicate that traditional GO capsules possess good bioavailability (41). Although garlic is widely consumed (41) by
those without overt CHD risk and trials of normal healthy
persons are especially relevant, only a few such trials have
been reported.(30,42 44)
In summary, although 60 trials of lipid-lowering have
been reported, our knowledge of the response to garlic is
limited with respect to effective doses, types of garlic preparation and human subject categories. Here we have sought to
monitor any beneficial influence of garlic preparations on a
range of CHD risk factors. Thus, decreased CHD risk is a
feasible benefit of reductions in plasma TC and LDL-C (45),
TG (46) and fasting glucose (47). Similarly, increases in
plasma HDL cholesterol (HDL-C) (48) and plasma antioxidant status (49) measured as plasma TAOC [a potential
marker of antioxidant protection of LDL against oxidative
modification (50)] would likely have benefit. Our study measured these variables and included several novel features, i.e.,
the use of traditional GO capsules in parallel with GP, the use
of traditional dose levels of 8 mg allicin equivalent (34)
with quantitation of the trial dose, inclusion of measurements
obtained on occasions during and after dosing and the use of
subjects lacking overt CHD risk. The aim here was to test the
hypothesis that garlic supplements would reduce specified risk
factors for CHD in male and female subjects at low risk of
CHD. Our main findings were that traditional GO capsules
had no significant effects but that gender differences may occur
in the influence of garlic on CHD risk factors.

Downloaded from jn.nutrition.org by on October 19, 2010

garlic derivable from allicin remain uncertain. Recent breath

analysis studies using gas chromatography indicated rapid metabolism of garlic and appear to offer a feasible means of
assessing the bioavailability of garlic materials (3,4).
Many animal studies (2,3,5) have suggested that garlic
favorably influences CHD risk factors. Of the 60 human
trials of the lipid-lowering effects of garlic, most provided
alliin/allicin standardized GP tablets, and most were positive
in outcome (3), as were two meta-analyses (1,6) showing
significant decreases in total cholesterol (TC) averaging
10%. However, many studies, including some incorporated
into meta-analyses, have been criticized for shortcomings of
design and description (7,8). The latest meta-analysis (9)
revealed a relatively small (5.8%) overall reduction in TC,
reflecting the inclusion of several recent well-designed trials
that were entirely negative (10 13). Ten of the 13 trials in
this meta-analysis, including all but one of the recent entirely
negative trials, tested the same alliin-standardized GP tablet
product (Kwai, Lichtwer Pharma GmbH, Berlin, Germany).
For 23 trials of Kwai reported between 1986 and 1994, almost
uniformly significant reductions (11%) in serum cholesterol
were reported (3). However, scrutiny of label descriptions
revealed that the formulation of Kwai was altered in 1992
1993, possibly affecting the bioavailability of allicin (14,15).
That this was the case is supported by the investigation of
batches of tablets from 1994 to 1997 using USP Method 724A
to simulate gastrointestinal conditions. This found very low
release, with only 14 18% of the potential allicin yield. In
contrast, batches dating from 1989 to 1992 released 36 61%
of their possible allicin yield. In view of their timing, it seems
likely that the wholly negative studies of Kwai conducted since
1995 (10 12,16,17) utilized the reformulated tablets that apparently had very low allicin availability. A negative study of
GO tablets (13), which was also included in the most recent
meta-analysis (9) was the first well-designed trial of the TClowering effects of GO. However, subsequent tests using USP
Method 724A as well as fecal examination indicated low
bioavailability of GO from the tablets used, in which GO is
bound to cyclodextrin (40). Thus, the smaller TC reductions
compared with earlier estimates that were identified in the
most recent meta-analysis (9) may merely reflect poor bioavailability of alliin-derived garlic components in several of
the included trials.
In the two most recent trials of lipid-lowering, results for
GP tablets were null (18) and positive (19), respectively.
Application of USP Method 724A revealed no detectable
allicin for the unnamed GP tablet brand of the null trial,
whereas in the positive trial, the significant reduction of TC
and LDL cholesterol (LDL-C) reported was associated with
intake of enteric-coated (acid-resistant) GP tablets for which
complete allicin release was indicated (15).
Summarizing, the current controversy associated with recent studies reporting null effects of garlic may have resulted
mainly from low bioavailability of alliin derivatives in the
garlic preparations consumed; thus, they should not overinfluence judgments concerning the TC-lowering ability of garlic.
Relative to TC, less attention has been paid to effects of
garlic on plasma triglycerides (TG), although a meta-analysis
of eight studies (1) indicated a significant 13% reduction, and
30 earlier (i.e., pre-1993) studies almost uniformly showed
reductions that were mostly significant (2).
In an early study (20), large reductions in the blood and
urine sugar levels of two diabetic subjects were observed after
ingestion of 10 15 g fresh garlic/d. More recent animal studies
(21,22) and a single recent human trial using GP tablets (23)
indicated that garlic can lower blood glucose. Increased levels

GARLIC OIL INTAKE AND CHD RISK FACTORS

SUBJECTS AND METHODS

statistical analysis. The distributions of age and triglyceride measurements were skewed; thus, the median and interquartile range were
used to describe these.
The main outcomes concerned the changes in plasma values from
the beginning to the end of the 11-wk treatment period. Mean
changes in plasma for subjects receiving GO were compared with
those for subjects receiving placebo by the use of 95% confidence
intervals (CI) and by two-sample t tests with probability values
considered significant at P 0.05.
Subgroup analyses by gender, age and body weight were not
specified in advance of analysis. For this reason and because many t
tests for interaction were done, statistical significance was set at P
0.01; accordingly, care should be taken in the interpretation of results
of borderline significance. The t tests of garlic effect by gender
interaction were done according to the method described by Pocock
(55). Gender differences in the effect of garlic were computed with
95% CI. Similar analyses were done for age and body weight groups
defined by the median cut-off points for all subjects in the randomized
trial.

RESULTS
Seventy-eight volunteers were recruited into the study.
Fifty-one of these were entered into the randomized trial to
compare GO and placebo, and 27 were recruited separately to
receive GP.
An overview of recruitment, loss to follow-up and key
stages of data collection and follow-up is given in Figure 1.
Sixty subjects (GO, 19/25; placebo, 21/26; GP, 20/27)
remained in the study to the end of the 11-wk treatment
period. In terms of measured baseline characteristics, those lost
to follow-up were not dissimilar from those who remained
(Table 1).

Downloaded from jn.nutrition.org by on October 19, 2010

Subjects. Subjects were recruited by personal invitation; most

were students or academic staff of the University of Wolverhampton.
The study was conducted from October 1994 to March 1995. Subjects
gave informed consent following the procedures laid down in the
Helsinki agreement with revisions. Eligibility criteria included the
age range 18 65 y and (self-identified) good health. Exclusions were
medical diagnosis of any of the following: diabetes mellitus; cardiovascular disease; hyperlipoproteinemia; bleeding disorders; infectious
blood diseases; and heavy consumption of aspirin, onion or garlic.
Ethical permission for the study was given by the Research Ethics
Committee of the University of Wolverhampton. Body weights were
recorded at the beginning of the study.
Diet control and assessment. Diet control consisted of asking
subjects at recruitment to keep to their usual diet throughout the
study and especially not to alter their garlic or onion intake. Diet
assessment took the form of a qualitative assessment of dietary change
by questionnaire at the end of the dose period.
Design. The sample size was comparable to that of studies
reviewed in two meta-analyses (1,6); it was greater than that of earlier
trials of normal persons (2) and was chosen to be similar to one such
trial (43). Volunteers were allocated randomly, in a double-blind
manner, to GO or placebo by a member of the study team uninvolved
in recruitment. Blinding in placebo-controlled garlic trials is complicated by the odorous nature of garlic, and few previous studies have
attempted to do this. The GO gelatin capsules used in this study
(Cardiomax, Seven Seas, Hull, UK) were of the same size, ovoid
shape and yellow color as placebo capsules, and both contained
peppermint oil. The GO dose (2 capsules/d) was determined by
HPLC (51) to deliver 8.2 mg [4.09 mg/capsule, CV 0.9%, n 3] of
allyl sulfide/d (52).
Volunteers for the GP study were recruited separately. The GP
gelatin capsules (Garlicin, Biocare, Birmingham, UK) were of similar size to GO and placebo capsules, but were of distinctive appearance, and had little taste or odor of garlic. A daily dose of 1 g of GP
(2 capsules/d) was given and GP samples from the same supplier were
determined by HPLC (51) to release 7.8 mg allicin/g (7.77 mg/g, CV
1.4%, n 3) in water. Thus GO and GP treatments provided similar
allicin-equivalent doses. Subjects were asked to compensate for any
missed daily dose by taking an extra half-dose on the next 2 d.
Reviews of subjects were at 0, 4 and 6 wk, at the end of the
intervention phase at 11 wk and at 14 15 and 1718 wk after the
start of the intervention phase during the dose washout phase. During
the initial visit, subjects were asked to return their dose boxes,
together with any remaining capsules at the end of the treatment
period. Subjects were questioned about compliance during reviews
and those admitting to having not compensated for any missed doses
more than occasionally were treated as lost from the study; those who
had not fasted for 12 h were asked to return in a fasting state the next
day. Venous blood samples were normally collected in the morning at
0800 1000 h, transferred to lithium heparin tubes and stored on ice;
plasma was separated within 2 h. At the end of the dose period (end
of wk 11), subjects were asked to complete a questionnaire to determine their use of medicines, to further check for dose compliance and
to assess major dietary changes.
Laboratory analyses. Plasma analyses were made on a Beckman
Synchron CX7 (Beckman Instruments, High Wycombe, UK) in the
Department of Chemical Pathology, the Manor Hospital, Walsall.
TC/HDL-C ratios were calculated from the measurement of TC by
cholesterol esterase/cholesterol oxidase/peroxidase (CV 1.8% at
4.5 mmol/L) and HDL-C was measured using dextran sulfate/manganese precipitation (CV 6% at 1.4 mmol/L). LDL-C was calculated
by means of the Friedewald formula (53) from TC, HDL-C and TG
values. TG was measured using lipase/glycerol kinase/glycerol phosphate oxidase/peroxidase (CV 2% at 1.4 mmol/L). Glucose was
measured using a glucose oxidase/conductivity electrode method (CV
1.5% at 6.3 mmol/L). Plasma TAOC measurements were performed using the enhanced chemiluminescence method (54) (CV
5.2% at 408 mmol Trolox equivalent/L) in the Wolfson Laboratory, Birmingham University, UK.
Statistical analyses. The Statistical Package for the Social Sciences (SPSS, Version 7.5, SPSS, Woking, UK) was used for the

1473

FIGURE 1 Study profile showing recruitment and losses of subjects, treatment duration and blood sampling times throughout the trial.

ZHANG ET AL.

1474

TABLE 1
Baseline characteristics of subjects completing treatment
and those lost to follow-up at the end of treatment1,2
Lost to follow-up in GO
and placebo groups
No

Yes

Lost to follow-up in GP
group
No

Yes

n
40
11
20
7
Men, %
50
55
35
14
TC, mmol/L
4.9 0.8
5.1 0.7
4.3 0.9
5.4 1.8
HDL-C, mmol/L 1.30 0.26 1.21 0.22 1.18 0.22 1.38 0.24
TC/HDL-C
3.9 1.0
4.4 1.2
3.8 0.9
4.0 1.2
LDL-C, mmol/L
3.1 0.8
3.3 0.7
2.8 0.8
3.5 1.5
TG, mmol/L
1.07 0.50 1.19 0.39 0.86 0.36 1.21 0.45
Glucose, mmol/L 5.0 0.5
5.1 0.3
5.1 0.4
5.1 0.3
1 Data are means SD.
2 Abbreviations used: GO, garlic oil; GP, garlic powder; HDL-C, HDL

cholesterol; LDL-C, LDL cholesterol; TC, total cholesterol; TG, triglycerides.

Downloaded from jn.nutrition.org by on October 19, 2010

For the GO trial, the baseline TAOC mean for the GO

group was 0.5 SD below that for the placebo group (Table 2).
The GO group also contained a slightly higher proportion of
men and was younger on average by 5 y. Otherwise, baseline
characteristics did not differ in the two trial groups. The GP
subjects had similar characteristics to the groups in the GO
trial except that they weighed less and there were more
women.
Over 95% of questionnaires were returned in each group at
the end of treatment. In matters of taste, the placebo capsules
were rated as nice by 3 subjects and OK by 17 respondents.
No one rated them as unpleasant or really bad. GO capsules were rated as OK by 12 subjects, unpleasant by 7 and
really bad by 1. GP was also OK for 12 and unpleasant
for 7. Blinding was only partially effective because 28, 67 and
71% of placebo, GO and GP subjects, respectively, felt that
they were consuming garlic. Too few dose pots were returned
to yield useful information about compliance, but the regular
(unrecorded) questioning of subjects at each review suggested
good compliance with both fasting and dose taking.
There was a low intake of prescribed drugs [contraceptive
pill, hormone replacement therapy (HRT)] and vitamin supplements, which did not differ for the GO, placebo and GP
groups. The contraceptive pill had been taken by 4 (placebo),
2 (GO) and 3 (GP) subjects during treatment, whereas 2
(placebo) and 1 (GO) subjects were receiving female HRT,
and 1 (placebo), 3 (GO) and 6 (GP) had taken occasional or
regular vitamin supplements. No subject took prescribed medicines.
During the fall months, changes in diet were reported as
slight, but diets tended to become richer over the Christmas/
New Year period as shown by increases in intakes of fatty foods
and alcohol and by body weight gain in the GO and placebo
groups. Body weight increase was recorded by 11 subjects (7
placebo, 2 GO and 2 GP) during the dose period. Detailed diet
composition was not documented, although questionnaire responses did indicate that a few changes had occurred.
Changes in measured values for CHD risk factors during the
treatment period are summarized in Table 3. There was a
mean reduction between baseline (treatment wk 0) and the
end of treatment (treatment wk 11) of 0.24 mmol/L in
plasma TC in subjects taking GO, and an average reduction in
plasma TC of 0.04 mmol/L in subjects taking placebo. The

effect of GO on TC was estimated by the difference between

these two means, 0.24 (0.04) 0.20 mmol/L, i.e., the
effect of GO was to reduce plasma TC levels by 0.20 mmol/L
on average. This difference was not significant and the 95% CI
(0.59 0.19 mmol/L) included zero difference, as did the CI for
GO on other outcome variables and for GP on all outcome
variables. One plausible explanation of these results is that
there are no beneficial effects of garlic on the plasma CHD risk
factors measured.
At the end of the intervention period, measured effects of
garlic over placebo for the cholesterol-related variables and
glucose were in opposite directions for men and women (Table
4). Significant differences in effect of GO between men and
women (P 0.01) were observed for HDL-C (P 0.004) and
TC/HDL-C (P 0.003). Women showed favorable effects in
terms of CHD risk (i.e., increase in HDL-C and reduction in
TC/HDL-C), whereas men had small adverse effects. There
was a significant difference in GO effect for glucose (P
0.006), with a reduction seen for men and an increase for
women. For TAOC, there was no evidence of a difference in
GO effect between men and women (P 0.95). Effects due to
GP were not significant (P 0.07 0.95).
Age and body weight values were each dichotomized using
the median cut-off values (22 y, 68.0 kg) for those subjects in
the randomized trial (Tables A and B deposited with NAPS).
In contrast to gender, there were no significant interactions
with age or body weight.
The effect of garlic at each time point (Table C deposited
with NAPS) was estimated as for treatment wk 11 (Table 3).
No significant effect of GO or GP was discernible in wk 4 and
6 (Table C deposited with NAPS) except for TAOC at 6 wk
for which an early downward trend in the placebo group was
not observed with GO (P 0.01, Table C deposited with
NAPS). The results depicting the influence of gender interaction throughout the trial include those for the post-treatment washout period (Table 5).

TABLE 2

Baseline characteristics including plasma lipids, glucose and

total antioxidants of subjects completing the 11-wk treatment
involving garlic oil, placebo or garlic powder1,2

n
Men, %
Age, y
Body weight, kg
TC, mmol/L
HDL-C, mmol/L
TC/HDL-C
LDL-C, mmol/L
TG, mmol/L
Glucose, mmol/L
TAOC, mmol Trolox
equiv/L

GO group

Placebo
group

GP group

19
58
24 7
22 (2023)3
71 16
4.8 0.9
1.28 0.25
3.9 1.1
3.0 0.9
1.19 0.62
1.0 (0.81.4)3
5.1 0.3

21
43
29 9
27 (2133)
70 14
4.9 0.7
1.32 0.28
3.9 0.9
3.1 0.6
0.96 0.34
1.0 (0.61.2)
5.0 0.6

20
35
26 8
23 (2129)
63 10
4.3 0.9
1.18 0.22
3.8 0.9
2.8 0.8
0.86 0.36
0.8 (0.61.1)
5.1 0.4

372 75

413 78

366 92

1 Data are means SD unless indicated.

2 Abbreviations used: equiv, equivalent; GO, garlic oil; GP, garlic

powder; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; TAOC, total

antioxidant capacity; TC, total cholesterol; TG, triglycerides.
3 Median (interquartile range) for skewed data.

GARLIC OIL INTAKE AND CHD RISK FACTORS

1475

TABLE 3
Effects of garlic oil (GO) or garlic powder (GP) intake for 11 wk by men and women relative to placebo upon selected
CHD risk factors including plasma lipids, glucose and total antioxidants1
Plasma variable

Change2 in GO group
(n 1819)

TC, mmol/L

0.24 (0.52, 0.05)

0.04 (0.32, 0.25)

HDL-C, mmol/L

0.05 (0.12, 0.02)

0.08 (0.15, 0.01)

TC/HDL-C

0.06 (0.29, 0.42)

0.22 (0.23, 0.68)

LDL-C, mmol/L

0.09 (0.37, 0.20)

0.06 (0.26, 0.38)

TG, mmol/L

0.22 (0.42, 0.02)

0.04 (0.21, 0.12)

Glucose, mmol/L

0.16 (0.30, 0.01)

0.00 (0.20, 0.21)

TAOC, mmol Trolox equiv/L

Change2 in placebo
group (n 1921)

26 (3, 55)

Garlic oil effect3

32 (1, 64)

0.20 (0.59, 0.19)

P 0.31
0.03 (0.07, 0.12)
P 0.56
0.16 (0.73, 0.41)
P 0.57
0.15 (0.57, 0.27)
P 0.48
0.18 (0.42, 0.07)
P 0.16
0.16 (0.41, 0.08)
P 0.19
6 (48, 36)
P 0.77

Change2 in GP
group (n 1920)

Garlic powder
effect3

0.09 (0.33, 0.15)

0.05 (0.42, 0.31)

P 0.78
0.05 (0.05, 0.14)
P 0.31
0.18 (0.68, 0.31)
P 0.45
0.10 (0.46, 0.27)
P 0.59
0.00 (0.20, 0.20)
P 0.98
0.04 (0.29, 0.20)
P 0.73
10 (38, 57)
P 0.68

0.03 (0.10, 0.03)

0.04 (0.17, 0.25)
0.04 (0.23, 0.15)
0.05 (0.17, 0.08)
0.04 (0.18, 0.11)
42 (4, 79)

DISCUSSION
The main hypothesis was not supported by statistical tests.
One plausible explanation is that garlic lacks lipid-lowering
efficacy, as was concluded for other recent negative garlic
studies (10,11,13,16,17) and the latest meta-analysis (9). This
conclusion differs from that of earlier meta-analyses (1,6) and
some recent metabolic studies, which indicate a lipid-lowering
action of garlic. For example, inhibition of cholesterol biosynthesis by garlic was observed in rat hepatocytes (56), and garlic
intake by rats produced elevated catecholamine levels associ-

Downloaded from jn.nutrition.org by on October 19, 2010

1 Abbreviations used: CHD, coronary heart disease; equiv, equivalent; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; TAOC, total antioxidant
capacity; TC, total cholesterol; TG, triglycerides.
2 Changes from baseline values at the end of the treatment period (wk 11) were calculated by subtracting the measured values of baseline from
those at 11 wk; therefore a positive figure implies a decrease and a negative figure implies an increase in values, on average, during the treatment
period; data are presented as means (95% confidence intervals).
3 Effects of garlic intake were calculated by subtracting changes from baseline in the placebo group from those in the garlic group; therefore any
favorable effects of garlic were indicated by negative figures for TC, TC/HDL-C, LDL-C and glucose and by positive figures for HDL-C and TAOC;
data are presented as means (95% confidence intervals).

ated with increased fat catabolism (57). Moreover, exhalation

of acetone by human subjects has been shown to be stoichiometrically related to the intake of allyl groups present in
crushed raw garlic, allicin (3) and the major diallyl components of garlic oil (diallyl disulfide and diallyl trisulfide) (Lawson, L. D., Plant Bioactives Research Institute, Orem, UT,
personal communication), which suggests that garlic acts to
increase hepatic fatty acid oxidation.
Noticeable in the present study were opposing influences of
GO on men and women. By canceling each other out, these

TABLE 4
Influence of gender interaction upon effects of garlic oil (GO) or garlic powder (GP) intake for 11 wk by men and women relative
to placebo on selected CHD risk factors including plasma lipids, glucose and total antioxidants1
Garlic oil
Mean effects2

Plasma variable
TC, mmol/L
HDL-C, mmol/L
TC/HDL-C
LDL-C, mmol/L
TG, mmol/L
Glucose, mmol/L
TAOC, mol Trolox equiv/L

All
Women
Men
(19/21)3 (8/12)3 (11/9)3
0.20
0.03
0.16
0.15
0.18
0.16
6

0.50
0.16
0.94
0.63
0.06
0.16
7

0.11
0.08
0.56
0.29
0.23
0.49
4

Garlic powder
Mean effects2

Gender interaction
Mean (95% CI)4

P5

0.60 (1.38, 0.18)

0.24 (0.08, 0.40)
1.51 (2.46, 0.55)
0.92 (1.70, 0.14)
0.17 (0.31, 0.66)
0.66 (0.20, 1.11)
3 (85, 80)

0.13
0.004
0.003
0.02
0.47
0.006
0.95

All
Women
Men
(20/21)3 (7/9)3 (13/12)3
0.05
0.05
0.18
0.10
0.00
0.04
10

0.22
0.05
0.40
0.28
0.00
0.15
5

0.19
0.05
0.08
0.16
0.03
0.31
35

Gender interaction
Mean (95% CI)4

P5

0.41 (1.12, 0.29)

0.00 (0.17, 0.18)
0.47 (1.38, 0.44)
0.43 (1.12, 0.26)
0.03 (0.34, 0.41)
0.46 (0.04, 0.96)
40 (154, 74)

0.24
0.95
0.30
0.21
0.86
0.07
0.48

1 Abbreviations used: CHD, coronary heart disease; 95% CI, 95% confidence intervals; equiv, equivalent; HDL-C, HDL cholesterol; LDL-C, LDL
cholesterol; TAOC, total antioxidant capacity; TC, total cholesterol; TG, triglycerides.
2 Effects of garlic intake were calculated as for Table 3.
3 Number in garlic group/number in placebo group.
4 Gender interactions on garlic effects were calculated as the mean difference (95% CI) between women and men (women men).
5 P probability value of t test for gender interaction; underlined values indicate significant interaction (P 0.01). Null hypothesis: effect of garlic
on response variable is independent of gender.

ZHANG ET AL.

1476

TABLE 5
Influence of gender interaction on the effects of garlic oil (GO) or garlic powder (GP) intake relative to placebo throughout the trial
upon selected CHD risk factors including plasma lipids, glucose and total antioxidants1,2
Gender interaction upon effects of garlic intake

Plasma variable
TC, mmol/L
HDL-C, mmol/L
TC/HDL-C
LDL-C, mmol/L
TG, mmol/L
Glucose, mmol/L
TAOC, mmol Trolox equiv/L

Type of
garlic intake

4 wk

GO
GP
GO
GP
GO
GP
GO
GP
GO
GP
GO
GP
GO
GP

0.26
0.29
0.01
0.06
0.27
0.00
0.33
0.21
0.17
0.06
0.27
0.03
28
43

6 wk
0.23
0.52
0.09
0.14
0.32
0.25
0.46
0.41
0.33
0.07
0.48
0.80
20
105

11 wk
(end of treatment)

14/15 wk
(postdose 1)

17/18 wk
(postdose 2)

0.60
0.41
0.24
0.00
1.51
0.47
0.92
0.43
0.17
0.03
0.66
0.46
3
40

0.58
0.39
0.07
0.07
0.93
0.72
0.83
0.44
0.39
0.10
0.50
0.34
0
42

0.22
0.43
0.30
0.11
1.18
0.59
0.73
0.59
0.45
0.10
0.41
0.20
26
12

effects generated a broadly neutral set of results overall. The

influence of gender on the response to garlic observed in our
study was comparable to that observed with the potent inhibitors of cholesterol biosynthesis, fluvastatin (58) and simvastatin (59). This effect of garlic was surprising because although
some earlier studies described garlic affecting reproductive
activities of both male and female fish (60) and mimicry of sex
hormones in mice (61), no reports described gender differences in the response of CHD risk factors to garlic intake.
Gender differences in the incidence of hyperlipidemia (62)
and CHD (63) are well known. In men, insulin-induced peripheral uptake of glucose and associated reduction in fasting
plasma glucose levels (64, 65) as well as catecholamine stimulation of hormone-sensitive lipase activity are each greater
than in women (66 68). Such differences in metabolic regulation processes could underlie differences in response to lipidlowering agents between men and women. One endogenous
lipid-regulating process influenced by garlic that appears to
provide a feasible basis for gender effects is protein phosphorylation by AMP-dependent kinase (AMPK) (69). The actions
of this enzyme include inhibition of hydroxy-methyl-CoAreductase (HMGR) and acetyl-CoA-carboxylase (70), which
are rate-controlling enzymes for hepatic synthesis of cholesterol and fatty acid, respectively. The activity of AMPK is
conditional upon cell energy charge; in rat hepatocytes, it
increases together with cellular levels of fructose and fatty
acyl-CoA synthesis, both of which deplete ATP and raise
AMP. Because inhibition of HMGR by garlic is amplified in
the presence of fructose (69) and palmitate (71), it has been
proposed that garlic promotes the phosphorylating action of
AMPK (69). This metabolic basis for any lipid-lowering
effects of garlic would likely be significantly influenced by
cellular fatty acid availability. The liver clears long-chain
fatty acids from plasma more rapidly in women than men
(72), and hepatocytes derived from female rats exhibit a
twofold faster uptake of long-chain fatty acids than those
from male rats (73). Greater activation of AMPK in female
liver cells is therefore feasible and could be expected to

Downloaded from jn.nutrition.org by on October 19, 2010

1 Gender interaction upon effects of garlic intake were calculated as for Table 4. Values underlined indicate significant interaction (P 0.01). Null
hypothesis: effect of garlic on response variable is independent of gender.
2 Abbreviations used: CHD, coronary heart disease; equiv, equivalent; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; TAOC, total antioxidant
capacity; TC, total cholesterol; TG, triglycerides.

amplify the inhibitory effects of garlic on lipid biosynthesis

considerably more in women than in men. Such marked
effects on lipid biosynthesis would likely influence the measures of blood lipids and therefore have the potential to
explain the gender differences in cholesterol parameters
observed in the present study.
The observation of initial TAOC declines in the placebo
group may have resulted from seasonal factors influencing
plasma vitamin C levels, which have been observed to decline
during fall and winter (74). The concurrent presence of a
significantly higher TAOC level in the GO group suggests an
improvement of plasma antioxidant status with garlic. Subsequent recovery of placebo values (at wk 11, mid-late January)
are a feasible consequence of the (recorded) dietary replenishments of the festive season.
Effects of GO intake on cholesterol-related variables at 4
and 6 wk in the mixed gender group (Table C deposited with
NAPS), although statistically null (range of P-values 0.31
0.57), were consistent in direction. The adverse trends observed are comparable to those reported from some previous
longitudinal garlic studies in which lipid-lowering was eventually observed (31,36,42,75,76). Such effects on plasma lipids
were construed by Bordia (75) to result from short-term lipid
mobilization by garlic. Persistence of gender influences into
the washout period (Table 5) would imply a carryover effect of
garlic. If supported experimentally, this could have implications for the interpretation of negative crossover studies
(13,16,33,37, 38). As with recent negative studies using reformulated Kwai GP tablets (14,15 and see Introduction), low
bioavailability of allicin from GP has to be considered a
possible factor in the present study. Thus, subsequent to the
commencement of our trial in 1994, similar gelatin-encapsulated GP preparations were shown to disintegrate rapidly in
simulated gastric fluid, resulting in alliinase inactivation (40).
In conclusion, the results obtained for the mixed gender
group with GO were statistically null, so that despite positive
overall trends for all lipid measurements and glucose, support
for the value of garlic supplementation in the reduction of

GARLIC OIL INTAKE AND CHD RISK FACTORS

CHD risk in normal subjects was lacking. Opposite responses

by men and women were a substantial confounder contributing to this null outcome. The results suggest that for healthy
subjects, any benefits of garlic to plasma lipid and glucose
levels may differ between men and women. The possible
gender effects observed indicate that earlier studies of combined male and female groups may have given a misleading
impression of the effects of garlic on CHD risk factors and that
the design of future studies into both the metabolic actions
and efficacy of garlic should take account of subject gender.
The results of this study and one other, which investigated
young male athletes (30), suggest that for normal men, GO
doses may have to be substantially greater than traditionally
recommended levels (34) to stand a chance of producing
significant improvements in plasma lipid risk factors.
ACKNOWLEDGMENTS

LITERATURE CITED
1. Silagy, C. & Neil, A. (1994) Garlic as a lipid lowering agenta metaanalysis. J. R. Coll. Physicians Lond. 28: 39 45.
2. Reuter, H. D., Koch, H. P. & Lawson, L. D. (1996) Therapeutic effects
and applications of garlic and its preparations. In: Garlic: The Science and
Therapeutic Application of Allium sativum L. and Related Species (Koch, P. H. &
Lawson, L. D., eds.), 2nd ed., pp. 135212. Williams & Wilkins, Baltimore, MD.
3. Lawson LD. (1998) Garlic: a review of its medicinal effects and indicated active compounds. In: Phytomedicines of Europe: Their Chemistry and
Biological Activity (Lawson, L. D. & Bauer, R., eds.), pp.176 209. ACS Symposium Series 691, American Chemical Society, Washington, DC.
4. Taucher, J., Hansel, A., Jordan, A. & Lindinger, W. (1996) Analysis of
compounds in human breath after ingestion of garlic using proton-transferreaction mass spectrometry. J. Agric. Food Chem. 44: 3778 3782.
5. Augusti, K. T. (1996) Therapeutic values of onion (Allium cepa L.) and
garlic (Allium sativum L.). Indian J. Exp. Biol. 34: 634 640.
6. Warshafsky, S., Kamer, R. S. & Sivak, S. L. (1993) Effect of garlic on
total serum cholesterol. A meta-analysis. Ann. Intern. Med. 119: 599 605.
7. Kleijnen, J., Knipschild, P. & Ter Riet, G. (1989) Garlic, onions and
cardiovascular risk factors. A review of the evidence from human experiments
with emphasis on commercially available preparations. Br. J. Clin. Pharmacol. 28:
535544.
8. Beaglehole, R. (1996) Garlic for flavour, not cardioprotection. The
Lancet 348: 186 187.
9. Stevinson, C., Pittler, M. H. & Ernst, E. (2000) Garlic for treating
hypercholesterolemia. A meta-analysis of randomised clinical trials. Ann. Intern.
Med. 133: 420 429.
10. Neil, H.A.W., Silagy, C. A., Lancaster, T., Hodgeman, J., Vos, K., Moore,
J. W., Jones, L., Cahill, J. & Fowler, G. (1996) Garlic powder in the treatment
of moderate hyperlipidaemia: a controlled trial and a meta-analysis. J. R. Coll.
Physicians Lond. 30: 329 334.
11. Isaacsohn, J. L., Moser, M., Stein, E. A., Dudley, K., Davey, J. A., Liskov,
E. & Black, H. R. (1998) Garlic powder and plasma lipids and lipoproteins: a
multicenter, randomized, placebo-controlled trial. Arch. Intern. Med. 158: 1189
1194.
12. McCrindle, B. W., Helden, E. & Conner, W. T. (1998) Garlic extract
therapy in children with hypercholesterolemia. Arch. Pediatr. Adolesc. Med. 152:
1089 1094.
13. Berthold, H. K., Sudhop, T. & von Bergmann, K. (1998) Effect of garlic
oil preparation on serum lipoproteins and cholesterol metabolism. J. Am. Med.
Assoc. 279: 1900 1902.
14. Lawson, L. D. (1998) Garlic powder for hyperlipidemiaanalysis of
recent negative results. Q. Rev. Nat. Med., Fall: 187189.
15. Lawson, L. D., Wang, Z. J. & Papadimitriou, D. (2001) Allicin release
under simulated gastrointestinal conditions from garlic powder tablets employed
in clinical trials on serum cholesterol. Planta Med. 67: 1318.

16. Simons, L. A., Balasubramaniam, S., von Konigsmark, M., Parfitt, A.,
Simons, J. & Peters, W. (1995) On the effect of garlic on plasma lipids and
lipoproteins in mild hypercholesterolaemia. Atherosclerosis 113: 219 225.
17. Superko, H. R. & Krauss, R. M. (2000) Garlic powder, effect on plasma
lipids, post-prandial lipemia, low-density lipoprotein subclass distribution and
lipoprotein(a). J. Am. Coll. Cardiol. 35: 321326.
18. Gardner, C. D., Chatterjee, L. M. & Carlson, J. J. (2001) The effect of
a garlic preparation on plasma lipid levels in moderately hypercholesterolemic
adults. Atherosclerosis 154: 213220.
19. Kannar, D. (2001) Clinical evaluation of garlic powder tablets in moderate hyperlipidemia. J. Am. Coll. Nutr. (in press).
20. Mahler, P. & Pasterny, K. (1924) Klinische Beobachtung uber Insulinwirkung beim Diabetes mellitus. Med. Klin. 11: 335338.
21. Chang, M.L.W. & Johnson, M. A. (1980) Effect of garlic on carbohydrate and lipid synthesis in rats. J. Nutr. 110: 931936.
22. Sheela, C. G. & Augusti, K. T. (1992) Antidiabetic effects of S-allyl
cysteine sulphoxide isolated from garlic Allium sativum L. Indian J. Exp. Biol. 30:
523526.
23. Kiesewetter, H., Jung, F., Pindur, G, Jung, E. M., Mrowietz, C. & Wenzel,
E. (1991) Effect of garlic on thrombocyte aggregation, microcirculation, and
other risk factors. Int. J. Clin. Pharmacol. Ther. Toxicol. 29: 151155.
24. Devaki, T., Venmadhi, S. & Govindaraju, P. (1992) Alterations in protein metabolism in ethanol-ingested rats treated with garlic. Med. Sci. Res. 20:
725727.
25. Venmadhi, S. & Devaki, T. (1992) Studies on some liver enzymes in
rats ingesting ethanol and treated with garlic oil. Med. Sci. Res. 20: 729 731.
26. Pushpendran, C. K., Devasagayam, T. P. & Eapen, J. (1982) Agerelated hyperglycaemic effect of diallyl disulphide in rats. Indian J. Exp. Biol. 20:
428 429.
27. Rabinkov, A., Miron, T., Konstantinovski, L., Wilchek, M., Mirelman, D. &
Weiner, L. (1998) The mode of action of allicin: trapping of radicals and
interaction with thiol containing proteins. Biochim. Biophys. Acta 1379: 233244.
28. Phelps, S. & Harris, W. S. (1993) Garlic supplementation reduces the
susceptibility to oxidation of apoprotein B-containing lipoproteins. Lipids 28:
475 477.
29. Byrne, D. J., Neil, H. A., Vallance, D. T. & Winder, A. F. (1999) A pilot
study of garlic consumption shows no significant effect on markers of oxidation
or sub-fraction composition of low-density lipoprotein including lipoprotein(a)
after allowance for non-compliance and the placebo effect. Clin. Chim. Acta 285:
2133.
30. Zhang, X.-H., Lowe, D., Giles, P., Fell, S., Board, A. R., Baughan, J. A.,
Connock, M. J. & Maslin, D. J. (2001) A randomized trial of the effects of garlic
oil upon coronary heart disease risk factors in trained male runners. Blood
Coagul. Fibrinolysis 12: 1 8.
31. Bimmermann, A., Weingart, K. & Schwartzkopff, W. (1988) Allium
sativum: Studie zur Wirksamkeit bei Hyperlipoproteinamie. Therapiewoche 38:
38853890.
32. Brewitt, B. & Lehmann, B. (1991) Lipidregulierung durch standardisierte Naturarzneimittel. Multizentrische Langzeitstudie an 1209 Patienten. Kassenarzt 5: 4755.
33. Luley, C., Lehmann-Leo, W., Moller, B., Martin, T. & Schwartzkopff, W.
(1986) Lack of efficacy of dried garlic in patients with hyperlipoproteinemia.
Arzneim.-Forsch. 36: 766 768.
34. Bundesgesundheitsamt (German Health Department) (1988) Kommission E monograph. BGA, Berlin, Germany.
35. Adler, A. J. & Holub, B. J. (1997) Effect of garlic and fish-oil supplementation on serum lipid and lipoprotein concentrations in hypercholesterolemic
men. Am. J. Clin. Nutr. 65: 445 450.
36. Lash, J. P, Cardoso, L. R., Mesler, P. M., Walczak, D. A. & Pollak, R.
(1998) The effect of garlic on hypercholesterolemia in renal transplant patients.
Transplant. Proc. 30: 189 191.
37. Plengvidhya, C., Sitprija, S., Chinayon, S., Pasatrat, S. & Tankeyoon, M.
(1988) Effects of spray dried garlic preparation on primary hyperlipoproteinemia.
J. Med. Assoc. Thail. 71: 248 252.
38. Steiner, M., Khan, A. H., Holbert, D. & Lin, R. I. (1996) A double-blind
crossover study in moderately hypercholesterolemic men that compared the
effect of aged garlic extract and placebo administration on blood lipids. Am. J.
Clin. Nutr. 64: 866 870.
39. Schiewe, F. P. & Hein, T. (1995) Garlic in hyperlipidemia. Influence of
a garlic preparation on the lipid serum levels of patients with primary hyperlipidaemia. Z. Phytother. 16: 343348.
40. Lawson, L. D. (1996) The composition and chemistry of garlic cloves
and processed garlic. In: Garlic: The Science and Therapeutic Application of
Allium sativum L. and Related Species (Koch, P. H. & Lawson, L. D., eds.), 2nd
ed., pp. 37107. Williams & Wilkins, Baltimore, MD.
41. Lawson, L. D. (1998) Garlic oil for hypercholesterolemianegative
results. Q. Rev. Nat. Med., Fall: 185186.
42. Lau, B.H.S., Lam, F. & Wang-Chang, R. (1987) Effect of an odormodified garlic preparation on blood lipids. Nutr. Res. 7: 139 149.
43. Barrie, S. A., Wright, J. V. & Pizzorno, J. E. (1987) Effects of garlic oil
on platelet aggregation, serum lipids and blood pressure in humans. J. Orthomol.
Med. 2: 1521.
44. Saradeth, T., Seidl, S., Resch, K. L. & Ernst, E. (1994) Does garlic alter
the lipid pattern in normal volunteers? Phytomedicine 1: 183185.
45. Rosengren, A. (1998) Cholesterol: how low is low enough? Br. Med. J.
317: 425 426.

Downloaded from jn.nutrition.org by on October 19, 2010

Final Year Honors degree students Anthony Allen, Arvinder Gill,

Emma Kirkman, Jane Lewis, Samantha-Jo McKinlay, Rimi Obra and
Sylvia Perwaiz, who assisted in subject recruitment; L. D. Lawson,
Research Director of Plant Bioactives Research Institute, for analysis
of the GO capsules; L. D. Lawson, P. Murray and S. Dunmore for
critical reading of manuscripts; Seven Seas Limited and BioCare
Limited for provision of garlic materials and financial contributions;
G. Thorpe, S. Maxwell and Helen Thomason of the Wolfson Laboratory, University of Birmingham for assistance with TAOC analysis;
David Roser, MCIM, Director of the Garlic Research Bureau, Bury
St. Edmunds, UK; and Willem vas Dias of Seven Seas Limited for
encouraging our research into the biological effects of garlic oil.

1477

1478

ZHANG ET AL.
62. Austin, M. A., King, M.-C., Vranisan, K. M. & Krauss, R. M. (1990)
Atherogenic lipoprotein phenotype. A proposed marker for coronary heart disease risk. Circulation 82: 495506.
63. Scandinavian Simvastatin Survival Study Group (1994) Randomised
trial of cholesterol lowering in 4444 patients with coronary heart disease: the
Scandinavian Simvastatin Survival Study (4S). The Lancet 344: 13831389.
64. Stark, B. & Keller, U. (1987) Alpha 1-adrenergic stimulation of ketogenesis and fatty acid oxidation is associated with inhibition of lipogenesis in rat
hepatocytes. Experientia 43: 1104 1106.
65. Falkner, B., Hulman, S. & Kushner, H. (1994) Gender differences in
insulin-stimulated glucose utilization among African-Americans. Am. J. Hypertens. 7: 948 952.
66. Brindle, N. P. & Ontko, J. A. (1988) Alpha-adrenergic suppression of
very-low-density-lipoprotein triacylglycerol secretion by isolated rat hepatocytes.
Biochem. J. 250: 363368.
67. Lonnqvist, F., Thorne, A., Large, V. & Arner, P. (1997) Sex differences
in visceral fat lipolysis and metabolic complications of obesity. Arterioscler.
Thromb. Vasc. Biol. 17: 14721480.
68. Wahrenberg, H., Bolinder, J. & Arner, P. (1991) Adrenergic regulation
of lipolysis in human fat cells during exercise. Eur. J. Clin. Investig. 21: 534 541.
69. Gebhardt, R. (2000) In vitro inhibition of cholesterol biosynthesis by
allicin and related garlic compounds. AHPA International Garlic Symposium 2000,
San Jose, CA.
70. Kemp, B. E., Mitchelhill, K. I., Stapleton, D., Michell, B. J., Chen, Z. P. &
Witters, L. A. (1999) Dealing with energy demand: the AMP-activated protein
kinase. Trends Biochem. Sci. 24: 2225.
71. Gebhardt, R. (1995) Amplification of palmitate-induced inhibition of
cholesterol biosynthesis in cultured rat hepatocytes by garlic-derived allylthiosulfinates. Phytomedicine 2: 29 34.
72. Hagve, T. A., Christenson, E., Gronn, M. & Christophersen, B. O. (1988)
Regulation of the metabolism of polyunsaturated fatty acids. Scand. J. Clin. Lab.
Investig. Suppl. 191: 33 46.
73. Luxon, B. A., Holly, D. C., Milliano, M. T. & Weisiger, R. A. (1998) Sex
differences in multiple steps in hepatic transport of palmitate support a balanced
uptake mechanism. Am. J. Physiol. 274: G52G61.
74. Johnston, C. S., Solomon, R. E. & Corte, C. (1998) Vitamin C status of
a campus population: college students get a C minus. J. Am. Coll. Health 46:
209 213.
75. Bordia, A. (1981) Effect of garlic on blood lipids in patients with
coronary heart disease. Am. J. Clin. Nutr. 34: 2100 2103.
76. Nitiyanant, W., Ploybutr, S., Wanuwat, S. & Tandhanand, S. (1987)
Effect of the dried powder extract, water soluble of garlic (Allium sativum) on
cholesterol, triglyceride and high density lipoprotein in the blood. J. Med. Assoc.
Thail. 70: 646 648.

Downloaded from jn.nutrition.org by on October 19, 2010

46. Hokanson, J. E. & Austin, M. A. (1996) Plasma triglyceride level is a

risk factor for cardiovascular disease independent of high-density lipoprotein
cholesterol level: a meta-analysis of population-based prospective studies.
J. Cardiovasc. Risk 3: 213219.
47. Bjornholt, J., Aaser, E., Sandvik, L., Nitter-Hauge, S., Erikkssen, G.,
Erikkssen, J. & Thaulow, E. (1999) Fasting blood glucose level is an independent predictor of cardiovascular mortality in healthy Norwegian men: results from
22 years of follow-up. Diabetes Care 22: 45 49.
48. Castelli, W. P. (1996) Lipids, risk factors and ischaemic heart disease.
Atherosclerosis 124 (suppl.): S1S9.
49. Steinberg, D. (1995) Clinical trials of antioxidants in atherosclerosis:
are we doing the right thing? The Lancet 346: 36 38.
50. Woodford, F. P. & Whitehead, T. P. (1998) Is measuring serum antioxidant capacity clinically useful? Ann. Clin. Biochem. 35: 48 56.
51. Lawson, L. D., Wang, Z. J. & Hughes, B. G. (1991) Identification and
HPLC quantitation of the sulfides and dialk(en)yl thiosulphinates in commercial
garlic products. Planta Med. 57: 363370.
52. Pentz, R. & Siegers, C.-P. (1996) Garlic preparations: methods for
qualitative and quantitative assessment of their ingredients. In: Garlic: The Science and Therapeutic Application of Allium sativum L. and Related Species (Koch,
P. H. & Lawson, L. D., eds.), 2nd ed., pp. 109 134. Williams & Wilkins, Baltimore,
MD.
53. Mackness, M. I. & Durrington, P. N. (1992) Lipoprotein separation and
analysis for clinical studies. In: Lipoprotein Analysis: A Practical Approach (Converse, C. A. & Skinner, E. R., eds.), pp.1 42. RIL Press, Oxford, UK.
54. Whitehead, T. P, Thorpe, G. H. & Maxwell, S.R.J. (1992) Enhanced
chemiluminescent assay for antioxidant capacity in biological fluids. Anal. Chim.
Acta 266: 265277.
55. Pocock, S. J. (1983) Clinical Trials, A Practical Approach. John Wiley
and Sons, New York, NY.
56. Gebhardt, R. (1993) Multiple inhibitory effects of garlic extracts on
cholesterol biosynthesis in hepatocytes. Lipids 28: 613 619.
57. Oi, Y., Kawada, T., Kitamura, K., Oyama, F., Nitta, M. & Kominato, Y.
(1995) Garlic supplementation enhances norepinephrine secretion, growth of
brown adipose tissue, and triglyceride catabolism in rats. Nutr. Biochem. 6:
250 255.
58. Leitersdorf, E. (1994) Gender-related response to fluvastatin in patients with heterozygous familial hypercholesterolaemia. Drugs 47 (suppl. 2):
54 58.
59. Clifton, P. M., Noakes, M. & Nestel, P. J. (1994) Gender and diet
interactions with simvastatin treatment. Atherosclerosis 110: 2533.
60. Glaser, E. & Drobnik, R. (1939) Beitrage zur Kenntnis der Wirkstoffe
des Knoblauchs. Arch. Exp. Pathol. Pharmakol. 193: 19.
61. Al-Bekairi, A. M., Shah, A. H. & Qureshi, S. (1990) Effect of Allium
sativum on epididymal spermatozoa, estradiol-treated mice, and general toxicity.
J. Ethnopharmacol. 29: 117125.