Content uploaded by Robert Hovenier
Author content
All content in this area was uploaded by Robert Hovenier on Feb 04, 2015
Content may be subject to copyright.
Content uploaded by Peter C H Hollman
Author content
All content in this area was uploaded by Peter C H Hollman on Dec 03, 2014
Content may be subject to copyright.
ABSTRACT Intake of dietary flavonols and flavones was
inversely associated with risk for cardiovascular disease in sev-
eral epidemiologic studies. This may have been due to effects on
hemostasis because flavonoids have been reported to inhibit
platelet aggregation in vitro. We indeed found that 2500 mmol/L
of the flavonol quercetin and the flavone apigenin significantly
inhibited collagen- and ADP-induced aggregation in platelet-rich
plasma and washed platelets by <80–97%. However, lower con-
centrations, such as might occur in vivo, had no effect. To test
this in vivo we fed 18 healthy volunteers 220 g onions/d provid-
ing 114 mg quercetin/d, 5 g dried parsley/d providing 84 mg api-
genin/d, or a placebo for 7 d each in a randomized crossover
experiment with each treatment period lasting 2 wk. Onion con-
sumption raised mean plasma quercetin concentrations to 1.5
mmol/L; plasma apigenin could not be measured. No significant
effects of onions or parsley were found on platelet aggregation,
thromboxane B
2
production, factor VII, or other hemostatic vari-
ables. We conclude that the antiaggregatory effects of flavonoids
seen in vitro are due to concentrations that cannot be attained in
vivo. Effects of dietary flavonols and flavones on cardiovascular
risk are possibly not mediated by hemostatic variables. Am J
Clin Nutr 1998;67:255–62.
KEY WORDS Apigenin, diet, factor VII, fibrinogen,
flavonoid, hemostasis, humans, plasminogen activator inhibitor
1, PAI-1, plasminogen, platelet aggregation, quercetin, throm-
boxane, in vitro study, onions, parsley
INTRODUCTION
Flavonoids are polyphenolic compounds that occur ubiquitous-
ly in plant foods. Flavonols and flavones are subclasses of
flavonoids (Figure 1) (1–4). In 1992 the average daily intake of
the flavonols quercetin, kaempferol, and myricetin from the Dutch
diet was 16, 4, and 1 mg, respectively; the average intake of the
flavones apigenin and luteolin was 1 mg each (5). The intake of
these five dietary flavonoids was associated with a reduced risk for
ischemic heart disease and stroke in several (6–9), although not all
(10, 11), epidemiologic studies. Two types of mechanisms have
been proposed to explain this protective effect: inhibition of LDL
oxidation and inhibition of platelet aggregation (12, 13). Results
obtained by incubation of human platelets or animal cells with iso-
lated flavonoids suggest that flavonoids inhibit platelet aggrega-
tion probably by inhibition of cyclooxygenase activity. Conflicting
observations on the effects of flavonoids on in vitro platelet aggre-
gation have been made (14–21). It has been stated that flavones are
potent and flavonols moderate inhibitors of cyclooxygenase, and
that flavonoid glycosides are less effective inhibitors than their
aglycones—the sugar-free part of the flavonoid molecule (14, 21).
This is relevant because flavonoids in foods are present mainly as
glycosides (1); we showed earlier that absorption of flavonoid glu-
cosides is more effective than absorption of aglycones (22). Holl-
man et al (23) found peak plasma quercetin concentrations of 0.6
mmol/L (196 ng/mL) in humans after consumption of 64.2 mg
quercetin from onions. Flavonoid concentrations used in in vitro
studies ranged between 10 and 1000 mmol/L (14–21), which is
thus 10–1000 times higher than plasma concentrations reached
after oral intake. There are no data on plasma concentrations of
apigenin in humans. Thus, there are several reasons why the effect
of flavonoids on platelets in vivo might differ from those observed
in vitro.
Flavonoids may also affect the activity or the concentration of plas-
ma coagulation or fibrinolysis factors such as fibrinogen, factor VII,
and plasminogen (24–26). The formation of a thrombus in athero-
sclerotic coronary arteries gives rise to acute ischemic heart disease,
and coagulation and fibrinolysis factors play a key role in the control
of thrombus formation (27, 28). Several studies indeed showed that
plasma fibrinogen concentration is an independent risk factor for
ischemic heart disease (29–31). Factor VII and plasminogen activity
were also associated with ischemic heart disease risk (30) and plasma
plasminogen activator inhibitor 1 (PAI-1) activity was associated with
increased risks of myocardial (re)infarction (31, 32).
Am J Clin Nutr 1998;67:255–62. Printed in USA. © 1998 American Society for Clinical Nutrition
Effects of the flavonoids quercetin and apigenin on hemostasis
in healthy volunteers: results from an in vitro and a dietary
supplement study
1–3
PLTM Karin Janssen, Ronald P Mensink, Frank JJ Cox, Jan L Harryvan, Robert Hovenier, Peter CH Hollman,
and Martijn B Katan
255
1
From the Department of Human Nutrition, Agricultural University,
Wageningen, Netherlands; the Department of Human Biology, Maastricht
University, Maastricht, Netherlands; and the State Institute for Quality Con-
trol of Agricultural Products (RIKILT-DLO), Wageningen, Netherlands.
2
Supported by grants from the Netherlands Heart Foundation (93.084)
and the Netherlands Foundation for Nutrition and Health Research, and a
PhD fellowship from Wageningen Agricultural University (to PLTMKJ).
3
Address reprint request to MB Katan, Department of Human Nutrition
and Epidemiology, Agricultural University, Bomenweg 2, 6703 HD,
Wageningen, Netherlands.
Received March 10, 1997.
Accepted for publication September 2, 1997.
by guest on July 14, 2011www.ajcn.orgDownloaded from
We carried out both an in vitro study and a dietary supplement
study in healthy volunteers. In the in vitro study we investigated
whether a test tube addition of flavonoids in the estimated phys-
iologic range (0–2.5 mmol/L) inhibited in vitro platelet aggrega-
tion, and we included unphysiologically high concentrations of
flavonoids (>2.5 mmol/L) to enable comparisons with published
studies. In the dietary study we examined whether administration
of foods rich in flavonoids affected platelet aggregation and
coagulation and fibrinolysis factors in healthy volunteers.
SUBJECTS AND METHODS
Both protocols were approved by the Medical Ethics Com-
mittee of the Department of Human Nutrition of Wageningen
Agricultural University and were fully explained to the partici-
pants, who gave their written, informed consent. Participants
were all nonsmokers. The in vitro study was carried out in the
summer of 1995 and the dietary supplement study from October
until December 1995.
In vitro study
Subjects
Four men from the Department of Human Biology, Maastricht
University, aged 24, 29, 35, and 47 y served as blood donors. All
were healthy according to a medical questionnaire and ate an ordi-
nary Western diet. None of the blood donors took any medication
for 2 wk preceding the study until the end of the measurements.
Methods
We investigated the effects on platelet aggregation of six con-
centrations (0, 0.25, 2.5, 25, 250, and 2500 mmol/L) of pure
quercetin-3-glucoside (Apin Chemicals LTD, Abingdon Oxon,
United Kingdom); quercetin aglycone, ie, quercetin without a
sugar moiety (quercetin dihydrate, no. 75670; Fluka Chemika,
Buchs, Germany); apigenin aglycone (no. 10798; Fluka Chemika,
Meppel, Netherlands); and catechin [(+)-catechin hydrate; Fluka
Chemika, Meppel, Netherlands]. It was not possible to study in
vitro effects of apigenin glycosides because these glycosides are
not available commercially. Effects were studied in both platelet-
rich plasma and washed platelets to exclude possible effects of
plasma factors on aggregation. Effects were not studied in whole
blood because whole blood is only stable for <30 min.
Indomethacin (I-7278; Sigma Chemical Co, St Louis), a spe-
cific inhibitor of the enzyme cyclooxygenase, was used as a pos-
itive control. Effects of final indomethacin concentrations of 0,
0.1, 1, 10, 100, and 1000 mmol/L on in vitro platelet aggregation
in platelet-rich plasma and washed platelets were tested by using
citrated blood from one of the donors. Aggregations in platelet-
rich plasma were stimulated with final concentrations of 2 mg
collagen/L (Collagen Horm, München, Germany) or 2.5 mmol
ADP/L (A6521; Sigma Chemical Co). Aggregations in washed
platelets were induced by 2.5 mg collagen/L or 18 mmol ADP/L.
Because polyphenols are known to bind proteins, catechin—a
flavonoid with a high protein binding capacity (33) but hardly any
effect on platelet aggregation or prostaglandin synthesis (14, 20,
34)—served as a control for nonspecific effects. Each donor gave
blood four times within a period of 1–6 wk. Effects of apigenin,
quercetin, quercetin-3-glucoside, and catechin were tested in random
order for all four donors. Effects of the six different concentrations
of each compound were tested on the same day in random order.
After the participants had fasted overnight, free-flowing
venous blood was sampled without stasis (Strauss Kanüle, 1.2-
mm syringe; Luer, Wächterbach, Germany) with the subject in a
supine position. The first 3 mL blood was discarded. Blood was
collected into tubes prefilled with a sodium citrate solution (final
concentration: 10.9 mmol/L, pH 7.3; Merck BV, Amsterdam).
Platelet-rich plasma was centrifuged at 150–160 3 g for 15 min
at room temperature and diluted with autologous platelet-poor
plasma to a final concentration of 185 3 10
9
platelets/L. Washed
platelets were prepared by mixing 5.8 mL of 80 mmol trisodium
citrate/L, 52 mmol citric acid/L, and 183 mmol glucose [acid cit-
rate dextrose (ACD)]/L with 29.2 mL blood. This solution was
centrifuged for 15 min at 160 3 g at room temperature and
platelet-rich plasma was removed. Subsequently, 25 mL platelet-
rich plasma was mixed with 1 mL ACD, centrifuged for 15 min
at 610 3 g at room temperature, and the platelet pellet was resus-
pended in 2 mL HEPES buffer (pH 6.6; 136 mmol NaCl/L, 2.7
mmol KCl/L, 10 mmol HEPES/L, 2 mmol MgCl
2
·6H
2
O/L, 1 g
glucose/L, and 1 g bovine serum albumin/L); HEPES buffer was
added to a total volume of 30 mL, and 0.07 L ACD/L buffer was
added. The suspension was centrifuged at 610 3 g for 15 min at
room temperature, and the platelet pellet was resuspended in
HEPES buffer (pH 7.45). Glucose and bovine serum albumin
were added to the buffer just before use.
After the first blood donation, optimal doses of the stimuli
collagen and ADP were determined for each person by using 5
mL blank solvent dimethylsulfoxide (DMSO; Fluka Chemika,
Meppel, Netherlands) instead of a flavonoid solution, under con-
ditions as described below. The optimal stimulus doses, defined
as the final concentrations leading to a maximal aggregation of
65%, were as follows for the aggregations in platelet-rich plas-
ma of the four donors: 2.0, 4.0, 2.0, and 4.0 mg collagen/L and
9.0, 5.0, 2.5, and 2.5 mmol ADP/L. The optimal doses for aggre-
gation in washed platelets of the four donors were 7.5, 7.5, 2.5,
and 7.5 mg collagen/L and 18.0, 18.0, 18.0, and 6.0 mmol
ADP/L, respectively. These doses were used throughout.
For each aggregation measurement, 400 mL diluted platelet-
rich plasma or washed platelets were incubated in an aggre-
gometer at 37 °C at 1000 rpm (Chronolog Corporation, Haver-
town, PA). Flavonoids were dissolved in DMSO and 5 mL was
256 JANSSEN ET AL
FIGURE 1. Structure of flavonols and flavones. Flavonols: X = OH
(quercetin: R
1
= OH, R
2
= H; kaempferol: R
1
= H, R
2
= H; and myricetin:
R
1
= OH, R
2
= OH); flavones: X = H (apigenin: R
1
= H, R
2
= H; and lute-
olin: R
1
= OH, R
2
= H).
by guest on July 14, 2011www.ajcn.orgDownloaded from
added to the platelets to produce final concentrations of 0, 0.25,
2.5, 25, 250, and 2500 mmol/L. Ten microliters of collagen sus-
pension or ADP was added exactly 10 min after addition of the
flavonoid. We added fibrinogen (final concentration: 0.5 g/L,
fraction 1, type 4 bovine plasma; Sigma Chemie, Brunschwig,
Amsterdam) during the ADP-induced aggregation of washed
platelets. The change in percentage of transmitted light was
monitored continuously for 7 min. Light transmitted was set at
100% for platelet-poor plasma and HEPES buffer and at 0% for
platelet-rich plasma and the washed platelet suspension.
Aggregation measurements were completed within 2 h after
blood sampling. We found in earlier studies (unpublished obser-
vations, 1995) that platelets were stable during this period.
Platelets were always handled in plastic material and at room
temperature. Maximal aggregation was calculated for all mea-
surements. For ADP-induced aggregations, maximal aggregation
of the first wave was used as the outcome variable. Maximal
aggregation on 5 mL solvent DMSO using optimal individual
stimulus concentrations was chosen as 100%. Means and SDs of
the maximal aggregation values were calculated for each
flavonoid concentration as the means and SDs of the four donors.
Statistics
Within-person effects of the addition of flavonoids to the
platelet-rich plasma compared with no flavonoid addition were
calculated for each flavonoid concentration tested. We then cal-
culated the mean effects (n = 4) and the 95% CIs of these effects.
For platelet-rich plasma and washed platelets the power to detect
a difference between two aggregation measurements, by using a
within person CV of 11%, was 90% to detect a difference of 18%
(a = 0.05).
Dietary supplement study
Subjects
Ten men and 12 women were recruited through announce-
ments in the University newspaper and posters in student dormi-
tories. One man was excluded because of elevated alanine
aminotransferase and g-glutamyltransferase concentrations and
one woman withdrew on her own accord. From the remaining 20
subjects, 9 men and 9 women were selected by drawing lots. One
man dropped out during the first week of the study for personal
reasons unrelated to the study. A woman who had participated in
the screening replaced him; she started at day 7 of the study.
There were no significant differences in the outcomes of the
dependent variables when the data of this woman were excluded
(data not shown). All 18 volunteers successfully completed the
study.
Participants were healthy on the basis of a medical question-
naire and had normal values for urinary protein and glucose,
hematocrit, hemoglobin, white blood cell and platelet counts,
mean red cell volume, plasma alanine aminotransferase, plasma
g-glutamyltransferase, serum creatinine, prothrombin, and acti-
vated partial thromboplastin time. The mean (±SD) age of the
subjects was 25 ± 8 y and body mass index (kg/m
2
) was 22 ± 1.
Design
We investigated the effects of daily consumption of flavonoid-
rich dietary supplements on indexes for hemostasis in a random-
ized, placebo-controlled, multiple crossover study involving
three treatments, each treatment period lasting 2 wk. All subjects
participated simultaneously. Participants consumed supplements
daily during weeks 2, 4, and 6. Week 1 served as the run-in peri-
od and weeks 3 and 5 as washout periods. Supplements were
given in random order.
Supplements
Supplements consisted of 400 g bouillon, to which was added
either 220 g cooked yellow onions (Favorit MSP, class II 60–80
mm; Luctor, Dronten, Netherlands), 4.9 g dried Western European
parsley (Verstegen Specerijen, Rotterdam, Netherlands), or nothing
as a placebo. Onions, parsley, and bouillon powder (Maggi Bouil-
lonkorrels; Nestlé Foodservice Catering, Amsterdam) were bought
in one batch just before the study and stored in the dark at 4°C.
The onions were thinly peeled 1 wk before the study started,
cut into pieces of 8 3 8 3 8 mm in a blender (Robot-Coupe SA,
Montceau-en-Bourgogne, France), heated in portions of 600 g in
a microwave oven for 5 min at 800 W or 7 min at 500 W, mixed,
and heated for another 4 min at 800 W or 5 min at 500 W. Por-
tions of 220 g were weighed on a digital scale to a precision of
0.01 g (model 1203 MP; Sartorius, Gottingen, Germany) and
stored at 220°C. Onions were thawed in the dark at room tem-
perature overnight before consumption. Before the study started
the parsley was mixed, weighed out in portions of 4.9 g, and
stored in the dark at room temperature until consumption. Bouil-
lon was prepared each day before consumption by using 16.7 g
bouillon powder/kg boiling water; 400-g portions were weighed
out and stored in the dark at 4°C. Duplicate portions of the sup-
plements were prepared daily and stored at 220 °C until ana-
lyzed (35). The onion supplement contained 114 ± 3 mg
quercetin (377 ± 10 mmol; n = 15) and the placebo 0.015 ± 0.004
mg quercetin (0.05 ± 0.01 mmol; n = 6). The parsley supplement
contained 84 ± 6 mg apigenin (n = 15).
During weeks 2, 4, and 6, fasted subjects came to the depart-
ment daily between 0730 and 0900 on working days and between
0800 and 1000 on weekend days to consume their supplements.
Parsley and onions were mixed with the bouillon just before con-
sumption. The onion soup was heated for 6 min at 800 W or for
7 min at 500 W; the parsley soup and placebo bouillon were
heated for 4.5 min at 500 W or 3.5 min at 800 W. Participants
were not allowed to eat or drink anything except mineral water
until 2 h after consumption of their supplements.
Food consumption, physical activity, and medication
Participants were urged not to consume any fruit or vegetables
containing >15 mg quercetin or apigenin/kg (eg, apples, endive,
beans, broccoli, celery, cherries, cloves, grapes, leeks, onions,
parsley, and tomatoes), any beverages containing > 4 mg
quercetin or apigenin/L (eg, tea and wine) (35–38), or any fatty
fish. Subjects were asked to maintain their normal eating and
drinking habits and physical activity levels during the study as
much as possible, given the restrictions mentioned above.
Subjects were instructed to avoid taking traditional and homeo-
pathic medicines and vitamin and mineral supplements from 1 mo
preceding the study until the end of the study. Use of oral contra-
ceptives was permitted. Participants were supplied with aceta-
minophen (paracetamol; Samenwerkende Apothekers, Utrecht,
Netherlands), which could be used for pain relief. Subjects were
urged to record health complaints, medications taken, and any
deviations from their normal physical activity and dietary habits in
FLAVONOIDS AND HEMOSTASIS 257
by guest on July 14, 2011www.ajcn.orgDownloaded from
a diary. We determined the body weights of the participants and
checked the diaries weekly at days 1, 8, 15, 22, 29, 36, and 42.
An experienced dietitian trained one of the human nutrition stu-
dents involved in the study to determine food intake. The food
intake of each participant was measured on 1 d in week 2, 1 d in
week 4, and 1 d in week 6 by using 24-h dietary recalls (39); 29%
of all recalls were from weekend days whereas 71% were from
weekdays. The dietitian and student were responsible for collecting
food intake data. Recalls of a particular subject were handled by the
same person during the whole study. We regularly checked the
between-interviewer variation, which turned out to be negligible.
The habitual energy intake of the subjects was 9.3 ± 2.7 MJ/d (2214
± 642 kcal/d), of which 31 ± 9% was provided by fat, 14 ± 3% by
protein, 53 ± 10% by carbohydrate, and 1 ± 3% by alcohol, with no
changes during the study. Mean body weight decreased by 0.3 ± 1.4
kg (NS) during the study. Subjects did not consume any fatty fish.
There was no evidence from the dietary recall data of changes in
physical activity patterns or any deviations that might have affected
the results. One subject took iron tablets (Ferrofumarat, three 200-
mg tablets/d; FNA, The Hague) from day 33 until the end of the
study because of low hemoglobin concentrations. Subjects took no
medications during the study, except for the acetaminophen sup-
plied by us. All participants denied having taken acetylsalicylic acid
from 1 mo preceding the study until the end of the study.
Blood sampling
Blood samples were taken on days 14, 28, and 42 as described
above. About 25 mL blood was drawn 90 min after subjects had
consumed the supplement. Subjects were in a supine position for
20 min before until the end of venipuncture. Blood was drawn
into tubes containing a final concentration of 10.9 mmol sodium
citrate/L. Samples for coagulation and fibrinolysis measure-
ments were put on ice immediately.
Analyses
Immediately after venipuncture 1 mL citrated blood was incu-
bated in a prewarmed aggregometer (37°C) and stirred at 1000
rpm. Exactly 5 min later 10 mL collagen suspension (final con-
centration: 2 mg/L) was added and whole-blood aggregation was
recorded for 10 min on a computerized system with an imped-
ance method. Maximal aggregation was measured in whole
blood to study effects of the supplements on platelet aggregation
under most physiologic conditions.
Platelet-rich plasma containing 250 3 10
9
platelets/L was pre-
pared as described previously (40). Some diluted platelet-rich
plasma was frozen in fluid nitrogen and stored at 280 °C for
determination of flavonoid concentrations. HPLC separation was
used combined with fluorescence detection for determination of
quercetin (41) and with ultraviolet detection for apigenin con-
centrations (35). The interassay CV for quercetin in plasma was
<10% and the intraassay CV was 4%.
For each aggregation, 400 mL platelet-rich plasma was incu-
bated in a prewarmed (37 °C) aggregometer and stirred at 1000
rpm. Exactly 5 min later, 10 mL collagen (final concentration:
2.0 mg/L) or ADP (final concentration: 1.5 and 3.0 mmol/L) was
added to induce aggregation. The change in percentage of trans-
mitted light was monitored continuously for 10 min and maximal
aggregation was calculated as described for the in vitro study.
Aggregations in platelet-rich plasma were carried out to compare
our results with the results of others and with the results of our
in vitro data. Maximally stimulated thromboxane B
2
production
in platelet-rich plasma was measured as described earlier (40) as
a specific measure of active cyclooxygenase present. All samples
of one subject were analyzed within one run.
Immediately after blood sampling, plasma for measurements of
coagulation and fibrinolysis indexes was separated by centrifuga-
tion at 1500 3 g for 20 min at 4°C. It was divided into aliquots,
snap-frozen in liquid nitrogen, and stored at 280 °C. Factor VII
and plasminogen activity were determined with a laser-nephelo-
metric centrifugal ACL-200 analyzer (Instrument Laboratory,
Milano, Italy). Factor VII clotting time was determined in a stan-
dard one-stage assay: plasma samples and factor VII–deficient
plasma were mixed, the clotting process was initiated, and clotting
time was measured by using PT-fibrinogen and thromboplastin,
both from the Instrumentation Laboratory (IJsselstein, Nether-
lands), and factor VII from Organon Technica (Oss, Netherlands).
Plasminogen activity was determined according to the test manu-
facturer’s instructions by using COATEST antiplasmine and strep-
tokinase from Chromogenix (Amsterdam). Blood for a normal
plasma pool was donated by 40 healthy volunteers. Factor VII and
plasminogen results were expressed in percentages relative to val-
ues for this pool. Standards for factor VII and plasminogen mea-
surements were derived from BIOPOOL (hemostasis reference
plasma; Kordia, Leiden, Netherlands). Plasma fibrinogen concen-
trations were determined with an STA II coagulation analyzer
(STA-fibrinogen; Diagnostica Stago, Boehringer Mannheim,
Mannheim, Germany): a fixed surplus of thrombin was added to
diluted platelet-poor plasma samples and the clotting time of a
series of dilutions of a human plasma pool with a known fibrino-
gen concentration was measured. Standards for fibrinogen mea-
surements were derived from Boehringer Mannheim (STA-Preci-
clot I and II). PAI-1 activity was measured with a chromogenic
assay (Spectrolyse/pL PAI; Biopool, Umea, Sweden). Plasmino-
gen and factor VII measurements were done in duplicate; fibrino-
gen and PAI-1 measurements were done once.
Statistics
We checked the data for normality using residual analysis
(42). Effects of onions and parsley were analyzed by using the
general linear model (analysis of variance, fixed effect) of the
Statistical Analysis System (a = 0.05) (43), with subject and
treatment as class variables. A period term was introduced into
the model to check for time effects, and a treatment-by-period
interaction term to check for carryover effects. Subsequently,
95% CIs were calculated for the effects of onions and parsley.
For aggregation in platelet-rich plasma, the power to detect a
difference between two aggregation measurements was 90% to
detect a difference of 9% (a = 0.05). In whole-blood aggregation
a difference of 7% could be detected (a = 0.05; within-person
variation: 8%). The difference in platelet thromboxane B
2
pro-
duction that could be detected in this study was 8% (power: 90%,
a = 0.05; within person variation: 10%).
RESULTS
In vitro study
Indomethacin at a concentration of 1 mmol/L inhibited maxi-
mal collagen-induced aggregation in human platelet-rich plasma
by 67% and in washed platelets by 74%; 1000 mmol/L inhibited
258 JANSSEN ET AL
by guest on July 14, 2011www.ajcn.orgDownloaded from
aggregation in platelet-rich plasma by 98% and in washed
platelets by 82% (data not shown). ADP-induced aggregation
was hardly affected.
Catechin inhibited collagen- and ADP-induced aggregation in
platelet-rich plasma significantly only at a concentration of 2500
mmol/L (Figures 2 and 3): average maximum collagen-induced
aggregation was inhibited by 59 ± 25% (95% CI: 19%, 99%),
and ADP-induced aggregation by 26 ± 17% (95% CI: 1%, 53%).
Results in washed platelets were similar to results found in
platelet-rich plasma (data not shown).
Apigenin at a concentration of 2.5 mmol/L inhibited collagen-
induced aggregation in platelet-rich plasma by 24 ± 34% (NS)
and ADP-induced aggregation by 22 ± 33% (NS). Concentra-
tions of 2500 mmol/L significantly inhibited collagen-induced
aggregation by 91 ± 4% (95% CI: 85%, 97%) and ADP-induced
aggregation by 80 ± 13% (95% CI: 59%, 101%) (Figures 2 and
3).
In platelet-rich plasma, 2500 mmol/L of the flavonol quercetin
significantly inhibited collagen-induced aggregation by 95 ± 4%
(95% CI: 89%, 101%) and ADP-induced aggregation by 97 ± 4%
(95% CI: 91%, 103%). Lower concentrations of quercetin were
ineffective (Figures 2 and 3).
Concentrations of 0.25-2500 mmol/L of the flavonol gluco-
side quercetin-3-glucoside did not significantly affect platelet
aggregation (Figures 2 and 3). Effects on aggregation in washed
platelets were comparable with those in platelet-rich plasma
(Figures 2 and 3).
Dietary supplement study
No adverse reactions to the supplements were reported,
although some subjects had difficulties in consuming the amount
of onions all at once. One subject consumed only part of the
onion supplement on one occasion; the leftovers contained 29
mg quercetin. Subjects consumed negligible amounts of
flavonol- or flavone-rich products during the study.
Concentrations of quercetin in platelet-rich plasma were 1.48
± 0.39 mmol/L (447 ± 117 ng/mL) 90 min after consumption of
onions and 0.02 ± 0.01 mmol/L (5 ± 4 ng/mL) after placebo. Api-
genin concentrations in platelet-rich plasma were all below the
limit of detection of 1.1 mmol/L, or 330 ng/mL.
Daily consumption of 220 g cooked onions or 4.9 g dried pars-
ley for 7 d did not significantly affect collagen-induced platelet
aggregation in whole blood or platelet-rich plasma; ADP-induced
platelet aggregation in platelet-rich plasma; thromboxane B
2
pro-
duction in platelet-rich plasma; platelet number; factor VII, plas-
minogen, and PAI-1 activity; or fibrinogen concentrations (Table
1). There were no treatment-sequence or time effects.
DISCUSSION
The in vitro study showed that a test tube addition of 0.25–250
mmol/L of apigenin, catechin, quercetin, or quercetin-3-gluco-
side did not inhibit collagen- or ADP-induced aggregation in
human platelets. Quercetin, apigenin, and catechin concentra-
tions of 2500 mmol/L inhibited platelet aggregation significantly
in vitro, whereas quercetin-3-glucoside still had no significant
effect. The dietary supplement study showed that administration
of large amounts of foods rich in apigenin or quercetin gluco-
sides did not affect platelet aggregation or other hemostatic vari-
ables in healthy volunteers.
In vitro study
We carried out highly standardized aggregation measurements
in both studies. Published in vitro studies were not always strict-
ly standardized as to platelet number and optimal stimulus con-
centration for each donor, and the experimental conditions var-
ied from one study to another, which probably explains the large
variation in outcome (14–21).
To check the validity of the in vitro aggregation measurements,
indomethacin was used as a positive and catechin as a negative
control. As expected, indomethacin—a specific inhibitor of
cyclooxygenase (44)—inhibited in vitro platelet aggregation
induced by collagen but not by ADP. We therefore conclude that
our assay could detect specific inhibiting effects on cyclooxyge-
nase activity in both platelet-rich plasma and washed platelets.
The in vitro study showed that concentrations of 0.25 and 2.5
FLAVONOIDS AND HEMOSTASIS 259
FIGURE 2. Relation between flavonoid concentration added to
platelet-rich plasma in vitro and maximal platelet aggregation induced
with collagen (n = 4). d, catechin; s, quercetin; ,, quercetin-3-gluco-
side; j, apigenin.
FIGURE 3. Relation between flavonoid concentration added to
platelet-rich plasma in vitro and maximal platelet aggregation induced
with ADP (n = 4). d, catechin; s, quercetin; ,, quercetin-3-glucoside;
j, apigenin.
by guest on July 14, 2011www.ajcn.orgDownloaded from
mmol catechin/L, which has a high affinity for proteins, did not
affect collagen- or ADP-induced platelet aggregation, although
unphysiologically high concentrations (2500 mmol/L) did (Fig-
ures 2 and 3). The latter may be a nonspecific inhibiting effect
because flavonoids readily bind to proteins (33). Thus, it is not
likely that nonspecific protein binding would explain the possible
effects of flavonoids at physiologic concentrations.
Dietary supplement study
Values of the hemostatic variables measured in the dietary
supplement study after treatment with placebo were in the nor-
mal range for healthy volunteers (Table 1). The design of the
dietary supplement study was not optimal: a controlled feeding
study would have reduced potential confounding from foods
eaten during the study. However, data from the diaries, body
weights, and 24-h recalls did not reveal any confounding effects.
Furthermore, all 18 subjects ate the supplements under our
supervision and plasma quercetin concentrations after treatment
with placebo indicated negligible dietary quercetin contents in
the background diets. These findings indicate that there was no
evidence for confounding in our study.
Plasma quercetin concentrations after treatment with the
onion supplement indicated that quercetin was absorbed. We
bought all onions, parsley, and bouillon before the study to
exclude differences between batches. We stored all onion sup-
plements at –20°C and parsley in the dark at room temperature
to diminish changes in composition. We also prepared all sup-
plements in a standardized way (see Methods) to prevent differ-
ences in composition of the supplements during the study.
Comparison with results of earlier studies
Gryglewski et al (17) reported stimulation of cyclooxygenase
activity after adding quercetin or rutin to ram seminal vesicle
microsomes. In contrast, results of other structure-function stud-
ies suggest that flavones such as apigenin are strong inhibitors
and flavonols such as quercetin are moderate inhibitors of
cylooxygenase. Those studies also suggested that glycosylated
compounds are less potent inhibitors of cyclooxygenase than
their aglycones (14, 21). A major drawback of in vitro studies
done earlier is that only the effects of unphysiologically high
flavonoid contents (> 10 mmol/L) were studied (14–21). It was
shown recently that plasma peak quercetin concentrations in
humans were 0.6 mmol/L after consumption of 64 mg quercetin
from onions (23). Although no data on apigenin absorption are
available, our data showed that plasma peak apigenin concentra-
tions in humans were < 1.1 mmol/L. We found that the flavone
apigenin and the flavonol quercetin significantly inhibited colla-
gen-induced aggregation in vitro only at concentrations of 2500
mmol/L, whereas quercetin-3-glucoside, the form in which
quercetin appears in foods and in which it is probably absorbed
(22), did not affect collagen- or ADP-induced aggregation. The
quercetin concentration in platelet-rich plasma after onion con-
sumption was comparable with data reported earlier (23). The
apigenin concentration in platelet-rich plasma could not be quan-
tified because of the high limit of detection (1.1 mmol/L, or 330
ng/mL). Thus, it is unclear whether apigenin was absorbed. It is
unlikely that apigenin absorption was inhibited by dietary pro-
teins because the subjects were not allowed to eat or drink any-
thing until 2 h after consumption of the supplements, and the
supplements contained hardly any protein. The plasma concen-
tration of quercetin suggests that these measures were adequate.
260 JANSSEN ET AL
TABLE 1
Hemostatic variables in healthy volunteers after consumption of placebo or 220 g onions or 4.9 g parsley daily for 7 d
1
Effect (treatment2placebo)
Variables Placebo
2
Onions
3
Parsley
4
Factor VII activity (n = 18) (%) 92 ± 25 24 ± 13 (210, 2) 21 ± 2 (22, 0)
Fibrinogen (n = 18) (g/L) 2.4 ± 0.6 0.1 ± 0.7 (20.2, 0.4) 0.2 ± 0.6 (20.1, 0.5)
Plasminogen activity (n = 17) (%) 87 ± 19 1 ± 8 (23, 5) 0 ± 6 (23, 3)
PAI-1 activity (n = 17) (kU/L) 4.35 ± 3.54 20.90 ± 3.17
5
(22.66, 0.86) 0.31 ± 3.97
6
(21.73, 2.35)
Maximally stimulated thromboxane
B
2
production in platelet-rich plasma (n = 18) (nmol/10
11
platelets) 2502 ± 593 2122 ± 438 (2340, 96) 2131 ± 503 (2381, 119)
Maximal collagen-induced aggregation in whole blood
(final 2 mg/L; n = 18) (V) 14 ± 4 20.6 ± 5.2 (23.2, 2.0) 20.1 ± 5.7 (22.9, 2.7)
Maximal collagen-induced aggregation in platelet-rich plasma
(final 2 mg/L; n = 13) (%) 84 ± 13 26.4 ± 14.5 (215.2, 2.4) 210.5 ± 27.4 (227.1, 6.1)
Maximal ADP-induced aggregation in platelet-rich plasma
(final 1.5 mmol/L; n = 18) (%)
7
23 ± 8 7 ± 20
8
(24, 18) 6 ± 19
6
(24, 16)
Maximal ADP-induced aggregation (%) in platelet-rich plasma
(final 3 mmol/L; n = 16) (%)
7
57 ± 28 28 ± 22
5
(220, 4) 26 ± 26
5
(220, 8)
Platelet number (n = 18) (310
9
/L) 235 ± 47 0 ± 28 (214, 14) 9 ± 25 (23, 21)
1
x
–
± SD; 95% CIs in parentheses.
2
Quercetin concentration in platelet-rich plasma: 0.02 ± 0.01 mmol/L.
3
Quercetin concentration in platelet-rich plasma: 1.48 ± 0.39 mmol/L.
4
Apigenin concentrations in platelet-rich plasma were lower than the limit of detection (1.1 mmol/L).
5
n = 15.
6
n = 17.
7
First wave.
8
n = 16.
by guest on July 14, 2011www.ajcn.orgDownloaded from
It is possible that plasma peak values and elimination curves of
apigenin are different from quercetin (23) and that blood sam-
pling at a different time point would reveal apigenin concentra-
tions higher than the detection limit. In any case, if apigenin is
absorbed, the prolonged supplementation would ensure interac-
tion with the blood components. Epidemiologic studies showed
that the inverse association between dietary flavonoid intake and
ischemic heart disease risk was strongest with quercetin (9).
Also, average daily dietary intake of quercetin was high (16 mg),
compared with intake of apigenin (1 mg) in the Netherlands (5).
In accordance with our in vitro data, no effect of consumption
of apigenin- or quercetin-rich foods was found on collagen-
induced aggregation in whole blood and platelet-rich plasma, on
ADP-induced aggregation in platelet-rich plasma, on thrombox-
ane production, or on other hemostatic variables (Table 1). These
findings indicate that daily consumption of large amounts of
quercetin- or apigenin-rich foods may not be effective in inhibit-
ing cyclooxygenase activity or platelet aggregation in human
volunteers.
We think that the treatment periods of the dietary supplement
study were long enough to detect possible effects on blood
platelet function. Blood platelets have a mean lifetime of 7–10 d;
therefore, treatment for 7 d with high doses of flavonoids should
be sufficient to determine effects on platelet aggregation and
thromboxane B
2
production. We showed earlier in healthy vol-
unteers that doses as small as 3 mg acetylsalicylic acid/d inhib-
ited maximally stimulated platelet thromboxane production by
39 ± 8% (40). Therefore, we think that our study group was large
enough to detect relevant effects on thromboxane production and
platelet aggregation (Table 1). However, we might have missed
small but biologically relevant effects (29–32) of the dietary
flavonoids on the coagulation and fibrinolytic indexes measured
(Table 1); this possibility should be examined in a larger study
population.
The results of the present study agree with results from the
dietary supplement study of Srivastava (45). They found no
effects of consumption of 70 g raw onions/d for 7 d on platelet
thromboxane production in five healthy volunteers: the mean
effect of onion consumption on thromboxane production was 95
± 756 nmol/L serum (n = 5). However, these authors could have
missed biologically significant effects because of the small study
sample, the low dose, and the large variation in the outcome vari-
able. We did not find any effects in a larger group consuming 220
g onions/d. Hertog et al (5) showed that the average daily intake
with the Dutch diet was 16 mg quercetin/d and 1 mg apigenin/d
(5). In our dietary supplement study the participants consumed
high doses of flavonoids: 114 mg quercetin/d and 84 mg api-
genin/d. We think that consumption of higher doses is impracti-
cal in a normal mixed diet.
Conclusions
Suggestions for an antiaggregatory effect of flavonoids are based
on the in vitro use of concentrations that cannot be attained in vivo
by dietary consumption. Reported effects of dietary flavonoids on
ischemic heart disease risk are possibly not mediated through col-
lagen- or ADP-induced platelet aggregation or cyclooxygenase
activity. We cannot exclude small biologically relevant effects of
dietary flavonoids on known risk indicators for ischemic heart dis-
ease from the coagulation cascade or fibrinolytic system, which
should be examined in a larger population.
We thank the participants for their cooperation; M Buijsman, A Hansel-
man, M Mommersteeg, M Straatman, E Voorter, J Wierts, and I Zijp for tech-
nical assistance; N Duif for checking the medical questionnaires; S Meyboom
for developing the supplements and for coordinating the food-consumption
analyses; and EJM Feskens, JGAJ Hautvast, D Kromhout, WA van Staveren,
and JJ Zwaginga for helpful suggestions. We thank the Department of Hema-
tology, Academic Hospital, Utrecht, Netherlands, for making their equipment
available to us.
REFERENCES
1. Kühnau J. The flavonoids: a class of semi-essential food compo-
nents: their role in human nutrition. World Rev Nutr Diet
1976;24:117–91.
2. Markham KR. Flavones, flavonols and their glycosides. Methods
Plant Biochem 1989;1:197–235.
3. Herrmann K. On the occurrence of flavonol and flavone glycosides
in vegetables. Z Lebensm Unters Forsch 1988;186:1–5.
4. Herrmann K. Flavonols and flavones in food plants: a review. J Food
Technol 1976;11:433–48.
5. Hertog MGL, Hollman PCH, Katan MB, Kromhout D. Intake of
potentially anticarcinogenic flavonoids and their determinants in
adults in The Netherlands. Nutr Cancer 1993;20:21–9.
6. Keli SO, Hertog MGL, Feskens EJM, Kromhout D. Dietary
flavonoids, antioxidant vitamins, and incidence of stroke. Arch
Intern Med 1996;156:637–42.
7. Knekt P, Reunanen RJ, Maatela J. Flavonoid intake and coronary
mortality in Finland: a cohort study. Br Med J 1996;312:478–81.
8. Hertog MGL, Kromhout D, Aravanis C, et al. Flavonoid intake and
long-term risk of coronary heart disease and cancer in the seven
countries. Arch Intern Med 1995;155:381–6.
9. Hertog MGL, Feskens EJ, Hollman PCH, Katan MB, Kromhout D.
Dietary antioxidant flavonoids and risk of coronary heart disease:
the Zutphen Elderly Study. Lancet 1993;342:1007–11.
10. Rimm ER, Katan MB, Ascherio A, Stampfer MJ, Willett WC. Rela-
tion between intake of flavonoids and risk for coronary heart disease
in male health professionals. Arch Intern Med 1996;125:384–9.
11. Hertog MGL, Sweetnam PM, Fehily AM, Elwood PC, Kromhout D.
Antioxidant flavonols and ischemic heart disease in a Welsh popu-
lation of men: the Caerphilly Study. Am J Clin Nutr 1997;65:1489−
94.
12. Hollman PCH, Hertog MGL, Katan MB. Role of dietary flavonoids
in protection against cancer and coronary heart disease. Biochem
Soc Trans 1996;24:785–9.
13. Hertog MGL, Hollman PCH. Potential health effects of the dietary
flavonol quercetin. Eur J Clin Nutr 1996;50:63–71.
14. Landolfi R, Mower RL, Steiner M. Modification of platelet function
and arachidonic acid metabolism by bioflavonoids: structure-activ-
ity relations. Biochem Pharmacol 1984;33:1525–30.
15. Laughton MJ, Evans PJ, Moroney MA, Hoult JR, Halliwell B. Inhi-
bition of mammalian 5-lipoxygenase and cyclo-oxygenase by
flavonoids and phenolic dietary additives: relationship to antioxi-
dant activity and to iron ion-reducing ability. Biochem Pharmacol
1991;42:1673–81.
16. Mower RL, Landolfi R, Steiner M. Inhibition of in vitro platelet
aggregation and arachidonic acid metabolism by flavones. Biochem
Pharmacol 1984;33:357–63.
17. Gryglewski RJ, Korbut R, Robak J, Swies J. On the mechanism of
antithrombotic action of flavonoids. Biochem Pharmacol
1987;36:317–22.
18. Pace-Asciak CR, Hahn S, Diamandis EP, Soleas G, Goldberg DM.
The red wine phenolics trans-resveratrol and quercetin block human
platelet aggregation and eicosanoid synthesis: implications for pro-
tection against coronary heart disease. Clin Chim Acta
1995;235:207–19.
FLAVONOIDS AND HEMOSTASIS 261
by guest on July 14, 2011www.ajcn.orgDownloaded from
19. Beretz A, Briancon Scheid F, Stierle A, Corre G,Anton R, Cazenave
JP. Inhibition of human platelet cyclic AMP phosphodiesterase and
of platelet aggregation by a hemisynthetic flavonoid, amentoflavone
hexaacetate. Biochem Pharmacol 1986;35:257–62.
20. Corvazier E, Maclouf J. Interference of some flavonoids and non-
steroidal anti-inflammatory drugs with oxidative metabolism of
arachidonic acid by human platelets and neutrophils. Biochim Bio-
phys Acta 1985;835:315–21.
21. Moroney MA, Alcaraz MJ, Forder RA, Carey F, Hoult JR. Selectiv-
ity of neutrophil 5-lipoxygenase and cyclo-oxygenase inhibition by
an anti-inflammatory flavonoid glycoside and related aglycone
flavonoids. J Pharm Pharmacol 1988;40:787–92.
22. Hollman PCH, de Vries JHM, van Leeuwen SD, Mengelers MJB,
Katan MB. Absorption of dietary quercetin glycosides and quercetin
in healthy ileostomy volunteers. Am J Clin Nutr 1996;62:1276–82.
23. Hollman PCH, van der Gaag MS, Mengelers MJB, de Vries JHM,
Katan MB. Absorption and disposition kinetics of the dietary
antioxidant quercetin in man. Free Radic Biol Med 1996;21:703–7.
24. Beretz A, Cazenave JP. The effect of flavonoids on blood-vessel
wall interactions. In: Cody V, Middleton E, Harborne JB, Beretz A,
eds. Plant flavonoids in biology and medicine. II: Biochemical, cel-
lular, and medicinal properties. New York: Alan R Liss,
1988:187–200.
25. Middleton EJ, Kandaswami C. Effects of flavonoids on immune and
inflammatory cell functions. Biochem Pharmacol 1992;43:1167–79.
26. Beretz A, Cazenave JP. Old and new natural products as the source
of modern antithrombotic drugs. Planta Med 1991;57:S68–72.
27. Badimon JJ, Fuster V, Chesebro JH, Badimon L. Coronary atheroscle-
rosis: a multifactorial disease. Circulation 1993;87(suppl):II-3–16.
28. Ross R. The pathogenesis of atherosclerosis: a perspective for the
1990’s. Nature 1993;362:801–9.
29. Meade TW, Cooper JA, Chakrabarti R, Miller GJ, Stirling Y,
Howarth DJ. Fibrinolytic activity and clotting factors in ischemic
heart disease in women. Br Med J 1996;312:1581.
30. Meade TW, Ruddock V, Stirling Y, Chakrabarti R, Miller GJ. Fibri-
nolytic activity, clotting factors, and long-term incidence of
ischemic heart disease in the Northwick Park Heart Study. Lancet
1993;342:1076–9.
31. Hamsten A, de Faire U, Walldius G, et al. Plasminogen activator
inhibitor in plasma: risk factor for recurrent myocardial infarction.
Lancet 1987;2:3–9.
32. Munkvad S, Gram J, Jespersen J. A depression of active tissue plas-
minogen activator in plasma characterizes patients with unstable
angina pectoris who develop myocardial infarction. Eur Heart J
1990;11:525–8.
33. Haslam E. Plant polyphenols: vegetable tannins revisited. Cam-
bridge, United Kingdom: Cambridge University Press, 1989.
34. Baumann J, von Bruchhausen F, Wurm G. Flavonoids and related
compounds as inhibition of arachidonic acid peroxidation.
Prostaglandins 1980;20:627–39.
35. Hertog MGL, Hollman PCH, Venema DP. Optimization of a quanti-
tative HPLC determination of potentially anticarcinogenic
flavonoids in vegetables and fruits. J Agric Food Chem
1992;40:1591–8.
36. Hertog MGL, Hollman PCH, van de Putte B. Flavonol and flavone
content of beverages. J Agric Food Chem 1993;41:1242–6.
37. Hertog MGL, Hollman PCH, Katan MB. Content of potentially
anticarcinogenic flavonoids of 28 vegetables and 9 fruits commonly
consumed in the Netherlands. J Agric Food Chem
1992;40:2379–83.
38. Blacquiere C, van de Pol L. Kruiden en specerijen: het voorkomen
van carcinogenen en anticarcinogenen, de consumptie, het
flavonoidgehalte. (Herbs and spices: the presence of carcinogens
and anticarcinogens, consumption, flavonoid content.) Wageningen,
Netherlands: Department of Human Nutrition, 1993.
39. Cameron ME, van Staveren WA. Manual on methodology for food
consumption studies. Oxford, United Kingdom: Oxford University
Press, 1988:83–8.
40. Janssen PLTMK, Akkerman JWN, Hollman PCH, van Staveren WA,
Zwaginga JJ, Katan MB. Pilot trial of the effects of low-dose acetyl-
salicylic acid on platelet thromboxane B
2
production. Eur J Clin
Nutr 1995;49:365–70.
41. Hollman PCH, Van Trijp JMP, Buysman MNCP. Fluorescence
detection of flavonoids in HPLC by postcolumn chelation with alu-
minum. Anal Chem 1996;68:3511−5.
42. Kleinbaum DG, Kupper LL, Muller KE. Applied regression analy-
sis and other multivariable methods. 2nd ed. Belmont, CA: Duxbury
Press, 1988.
43. SAS Institute Inc. SAS/STAT user’s guide 6. Cary, NC: SAS Insti-
tute Inc, 1989.
44. Maderna P, Caruso D, Tremoli E, Galli G. Differential effects of
oral administrations to human volunteers of acetylsalicylic acid,
sodium salicylate and indomethacin on 12-hydroxyeicosatetraenoic
acid formation by stimulated platelets. Thromb Res 1988;
52:197–206.
45. Srivastava KC. Effect of onion and ginger consumption on platelet
thromboxane production in humans. Prostaglandins Leukot Essent
Fatty Acids 1989;35:183–5.
262 JANSSEN ET AL
by guest on July 14, 2011www.ajcn.orgDownloaded from
