261

Soy and the exercise-induced inflammatory

response in postmenopausal women
Kristen M. Beavers, Monica C. Serra, Daniel P. Beavers, Matthew B. Cooke, and
Darryn S. Willoughby

Abstract: Aging is associated with increasing inflammation and oxidative stress in the body, both of which can have nega-
tive health effects. Successful attenuation of such processes with dietary countermeasures has major public health implica-
tions. Soy foods, as a source of high-quality protein and isoflavones, may improve such indices, although the effects in
healthy postmenopausal women are not well delineated. A single-blind, randomized controlled trial was conducted in 31
postmenopausal women who were assigned to consume 3 servings of soy (n = 16) or dairy (n = 15) milk per day for
4 weeks. Parameters of systemic inflammation (tumor necrosis factor-a (TNF-a), interleukin-1b (IL-1b), and interleukin-6
(IL-6)) and the oxidative defense system (superoxide dismutase (SOD), glutathione peroxidase, cyclooxygenase-2) were
measured post supplementation, before and after an eccentric exercise bout performed to elicit an inflammatory response.
A significant group-by-time effect for plasma TNF-a was observed (p = 0.02), with values in the dairy group increased
post supplementation and then decreasing into the postexercise period. Additionally, significant time effects were observed
for plasma SOD (p < 0.0001) and IL-6 (p < 0.0001) in the postexercise period. Overall results from our study do not sup-
port the notion that 4 weeks of daily soy milk ingestion can attenuate systemic elevations in markers of inflammation or
oxidative defense. However, data do suggest that the downhill-running protocol utilized in this study can be effective in al-
tering systemic markers of inflammation and oxidative defense enzyme activity, and that the ingestion of soy may help
prevent fluctuations in plasma TNF-a.
Key words: soy milk, inflammation, oxidative defense, downhill running, aging, isoflavone.
Résumé : Le vieillissement est associé à plus d’inflammation et de stress oxydatif dans l’organisme, ce qui entraine des effets
négatifs sur la santé. L’atténuation de ces effets au moyen de mesures alimentaires s’avère de la plus haute importance en
matière de santé publique. Les produits de soja, sources de protéines de grande qualité et d’isoflavones, recèlent un potentiel
d’amélioration de ces processus; cependant on ne connait pas bien les effets chez les femmes postménopausées en bonne
santé. Dans un essai clinique comparatif à simple insu et d’une durée de quatre semaines, on répartit au hasard 31 femmes
postménopausées; 16 acceptent de consommer tous les jours trois portions de produits de soja et 15, des produits laitiers. Les
paramètres de l’inflammation systémique (TNF-a, IL-1b, IL-6) et du système de défense contre l’oxydation (SOD, GPx,
COX-2) sont évalués après la supplémentation, et ce, avant et après une séance d’exercices pliométriques suscitant une
réponse inflammatoire. On observe un effet significatif du traitement en fonction du temps : les valeurs de TNF-a plasmatique
(p = 0,02) du groupe consommant des produits laitiers augmentent après la supplémentation puis diminuent après la séance
d’exercices. De plus, on observe un effet significatif en fonction du temps au sujet du SOD (p < 0,0001) et des IL-6 (p <
0,0001) plasmatiques après la séance d’exercices. D’après ces observations, la consommation quotidienne de boisson de soja
durant quatre semaines n’atténue pas les augmentations des marqueurs de l’inflammation et du système de défense antioxyda-
tive. En contrepartie, ces observations suggèrent que la course sur pente négative utilisée dans cette étude est efficace dans la
modification des marqueurs systémiques de l’inflammation et de l’activité enzymatique suscitée pour la défense antioxydative
et que la consommation de soja peut contribuer à la prévention des fluctuations plasmatiques de TNF-a.
Mots-clés : boisson de soja, inflammation, défense antioxydative, course en pente négative, vieillissement, isoflavone.
[Traduit par la Rédaction]

Received 21 September 2009. Accepted 18 February 2010. Published on the NRC Research Press Web site at apnm.nrc.ca on 13 May
2010.
K.M. Beavers,1,2 M.C. Serra,3 M.B. Cooke, and D.S. Willoughby. Department of Health, Human Performance, and Recreation, Baylor
University, Waco, TX. 76798-7313, USA.
D.P. Beavers.4 Department of Statistics, Baylor University, Waco, TX 76798-7313, USA.
1Corresponding author (e-mail: kbeavers@wfubmc.edu).
2Present address: Gerontology and Geriatric Medicine, Wake Forest University, School of Medicine, Medical Center Boulevard,
Winston-Salem, NC 27157, USA.
3Present address: Department of Gerontology, University of Maryland School of Medicine, 655 West Baltimore Street, Baltimore,

MD 21201-1559, USA.
4Present address: Department of Biostatistical Sciences, Wake Forest University, School of Medicine, Medical Center Boulevard,

Winston-Salem, NC 27157-1063, USA.

Appl. Physiol. Nutr. Metab. 35: 261–269 (2010) doi:10.1139/H10-015 Published by NRC Research Press
262 Appl. Physiol. Nutr. Metab. Vol. 35, 2010

Introduction flammation and muscular damage by increasing both muscle

proteases and ROS (Féasson et al. 2002; Peake et al. 2005b).
The process of senescence is associated with increasing As little as 30 min of exercise elevates biomarkers of oxida-
inflammation and oxidative stress in the body, both of which tive stress in the plasma, thereby eliciting an acute-phase re-
play a role in the progression of cardiovascular disease sponse (Bloomer et al. 2005). With regard to soy
(CVD). Systemic inflammation is reflected, in part, by ele- supplementation and acute exercise, the literature is scant
vations in proinflammatory cytokines. Elevations in these (Chen et al. 2004; Hill et al. 2004; Box et al. 2005),
cytokines, specifically tumor necrosis factor-a (TNF-a) and although generally favorable. However, data on dietary in-
interleukin-6 (IL-6), predict all-cause, as well as cardiovas- terventions using whole soy foods to achieve these outcomes
cular, mortality (Volpato et al. 2001; Bradley 2008). Accu- are lacking in the literature and, in light of increasing use, it
mulating evidence also indicates that oxidative stress plays is important to understand the health effects of soy milk.
a major role in the initiation and progression of cardiovascu- Therefore, the overall purpose of this study was to investi-
lar dysfunction (Dröge 2002; Taniyama and Griendling gate the effects of soy milk (compared with dairy milk) on
2003), as the generation of reactive oxygen species (ROS) plasma markers of inflammation and oxidative defense in re-
causes endothelial cell apoptosis, increased monocyte adhe- sponse to a single bout of eccentric exercise designed to
sion, and angiogenesis (Taniyama and Griendling 2003). elicit an acute-phase response in postmenopausal women.
Sobering statistics suggest that CVD will affect more than
1 in 3 American adults during their lifetimes (Rosamond et
Materials and methods
al. 2007). Although myocardial infarction risk is greater in
men, cerebrovascular accidents occur more commonly in Study design
women, and in women over the age of 40 years, CVD is Table 1 shows the general research design and time
the leading cause of death (Rosamond et al. 2007). This course of assessments administered in this study. The inde-
risk especially increases after menopause and is thought to pendent variables were the supplement groups (soy milk or
be caused, at least in part, by decreased estrogen production. dairy milk), the 4-week supplementation period, the eccen-
Given the prevalence of CVD and the increased longevity of tric exercise protocol, and the blood-sampling times during
women, CVD risk reduction in postmenopausal women rep- the course of this study. Dependent variables included the
resents a significant public health issue. information collected on the 24-h recall and 4-day food re-
Identification of nutritional countermeasures that can at- cords, delayed-onset muscle soreness (DOMS), and plasma
tenuate the progression of oxidative stress and inflammation inflammation and enzymatic oxidative defense variables
has major public health implications. Soy, as a unique (TNF-a, interleukin-1b (IL-1b), IL-6, superoxide dismutase
source of high-quality protein and isoflavones (phytoestro- (SOD), glutathione peroxidase (GPx), and cyclooxygenase-2
gens with antioxidant properties), is speculated to exert (COX-2)). Supplementation effects prior to eccentric exer-
such benefits, especially in postmenopausal women (Han et cise have already been reported (Beavers et al. 2009).
al. 2002). Compelling in vitro and animal data support this
notion (Elia et al. 2006; Oh et al. 2007; Martinez- Participants
Villaluenga et al. 2009), although data in humans are less An a priori power calculation showed that 14 participants
clear. Out of 6 studies assessing the effect of long-term (4 per group were necessary to detect a significant difference
to 24 weeks) soy consumption on expression of IL-6 and between groups in markers of oxidative defense (SOD effect
TNF-a (Jenkins et al. 2002; Hermansen et al. 2005; Hilpert size of 0.8 U
g–1; (Chen et al. 2004)) and inflammation
et al. 2005; Fanti et al. 2006; Ryan-Borchers et al. 2006; (IL-6 effect size of 0.24 pg
mL–1 (Jenkins et al. 2002)),
Azadbakht et al. 2007), only 2 found a statistically signifi- given a type I error rate of 0.05 and a power of 0.80. Partic-
cant effect of soy, with 1 study finding an increase in the ipants were recruited from the Central Texas area and were
expression of IL-6 (Jenkins et al. 2002) and the other a mod- considered eligible to participate in the study if they were
est decrease in TNF-a (Azadbakht et al. 2007). physically active but not trained (not engaged in an exercise
Given the inconsistencies in the clinical literature, program involving either resistance or endurance
coupled with the positive in vitro and animal data findings, training >3 times per week for 1 year); were postmenopausal
it has been suggested that basal concentrations of inflamma- (amenorrheic >1 year prior to start of study); and could ad-
tory cytokines and the activities of oxidative stress–defense here to the study protocol. Exclusionary criteria included ac-
biomarkers may be significant confounding factors in clini- tive use of hormone-replacement therapy, presence of
cal trial success. Dietary interventions may be efficacious coronary artery diseases and (or) other significant uncon-
only in populations with elevations of such biomarkers at trolled cardiovascular, renal, hepatic, gastrointestinal, men-
baseline. Indeed, results from our laboratory suggest that tal, and endocrine disorders, including diabetes mellitus, a
the mere consumption of soy milk for 4 weeks does not af- body mass index (BMI) of <19 kg
m–2 or >35 kg
m–2, active
fect the circulating concentrations of inflammatory cytokines smoking status (within the past 3 years), an average intake
or activities of oxidative defense enzymes in a healthy pop- of >2 alcoholic drinks per day, and (or) the consumption of
ulation (Beavers et al. 2009). Therefore, studying these phe- any dietary supplements that could affect antioxidant status
nomena in healthy individuals may require the employment (excluding multivitamins) 3 months prior to beginning the
of methodology known to upregulate such processes in a study.
timely and consistent manner.
Intense physical exercise, especially that requiring eccen- Familiarization session
tric muscle contraction or new training regimens, causes in- Prior to starting the study, all participants went through a

Published by NRC Research Press

Beavers et al. 263

familiarization session to learn about the study protocol, ob- of perceived exertion (RPE) ‡ 19 on the RPE scale, and
tain medical clearance from their personal physician, and (iii) maximum heart rate within 10 beats of age-predicted
sign university institutional review board-approved informed maximum. The highest level of VO _ 2 was defined as
consent documents. Participants were then assigned to begin _ 2 max. Heart rate was calculated from a continuously
VO
a 2-week washout period, in which they were instructed to monitored heart rate monitor (Polar Electro, Lake Success,
consume their usual diet but limit the consumption of all N.Y.) and blood pressure was determined with a mercurial
soy-based or dairy-based products. Specifically, during the sphygmomanometer every 3 min during the exercise ses-
study, each participant was instructed not to consume any sion. Information from this test was used in participant
additional soy- or isoflavone-containing foods, to limit dairy matching as well as in determining the speed to use during
intake to 2 servings per day, and to avoid dairy milk. In ad- the eccentric exercise bout.
dition, a 24-h recall was performed by a registered dietitian
utilizing a multiple pass method to determine ‘‘typical’’ cal- Supplementation protocol
orie and protein intakes (Rumpler et al. 2008). In a single-blind format, participants were matched for
_ 2 max) and baseline protein intake
physical fitness level (VO
Baseline assessment of maximum oxygen uptake and diet –1 –1
(g
kg
day ), and randomly placed into 1 of 2 dietary
Prior to beginning the supplementation protocol, each par- groups. Group A was instructed to consume a commercial
ticipant was scheduled to come to the Exercise and Sport vanilla soy milk product and group B was instructed to con-
Nutrition Laboratory at Baylor University (Waco, Tex.) for sume a commercial reduced-fat dairy milk product. The sup-
baseline testing. During this session, participants had their plement groups consumed 3 servings of their respective
body mass, height, and BMI determined. Each subject was supplements per day for a period of 4 weeks. Servings of
then assessed for aerobic fitness by performing an incremental- both supplements were matched as closely as possible for
load exercise test of their maximal oxygen uptake total caloric (soy milk, 130 kcal (1 kcal = 4.186 kJ) vs. dairy
_ 2 max) on a treadmill ergometer (Quinton, Cardiac Sci-
(VO milk, 120 kcal) and protein content (soy milk, 6 g vs. dairy
ence, Bothell, Wash.). After baseline measurements at rest milk, 8 g), with the soy milk providing an additional
were completed, the treadmill test was started at a velocity *30 mg isoflavones per serving. Participants were asked to
of 8 km
h–1 and a 0% incline. The running speed was in- record daily supplement consumption on an adherence log
creased by 2 km
h–1 every 3 min until exhaustion. Oxygen provided to them at the baseline testing session. Finally, a
uptake (VO_ 2) was measured every 30 s via an open-circuit reported side-effects questionnaire was administered at 2, 4,
sampling system (Parvo Medics, Sandy, Utah). The test and 6 weeks to determine if any significant negative side ef-
was accepted as a VO _ 2 max if 2 of the following criteria fects occurred because of the dietary supplementation during
were met: (i) respiratory exchange ratio ‡ 1.15, (ii) rating the course of the study.

Published by NRC Research Press

264 Appl. Physiol. Nutr. Metab. Vol. 35, 2010

Dietary records Table 2. Baseline demographic statistics by treatment

group.
In an attempt to determine compliance with the dietary
prescriptions, and also to assess the average daily macronu- Soy group Dairy group
trient consumption of fat, carbohydrate, and protein, each Variable (n = 16) (n = 15)
participant was asked to keep 4-day dietary records during Age (y) 53.9±3.7 55.0±3.1
weeks 2 and 6 of the study. Each participant was instructed BMI (kg
m–2) 25.36±4.05 26.31±3.96
by a registered dietitian on how to fill out the record prior to V̇O2 max (mL
kg–1min–1) 25.60±4.86 27.22±5.07
use. The dietary records were analyzed with the Food Pro- Protein intake (g
kg–1) 1.16±0.48 0.98±0.28
cessor Dietary Assessment Software program (ESHA Re-
search Inc., Salem, Ore.). Note: Data are presented as means ± SD. No differences
were noted between treatment groups at baseline (all p > 0.05).
BMI, body mass index; V̇O2 max, maximal oxygen uptake.
Eccentrically biased downhill run
After the 4-week supplementation period, each participant
liers (Stevens 2002) and those determined to be faulty by in-
reported back to the laboratory for a downhill run–walk test.
strumentation error were excluded from analysis.
Participants were instructed to refrain from aerobic exercise
for 48 h prior to the exercise session. A 5-min warm-up
(–10% grade and 3.5 km
h–1 speed) was performed on the Statistical analyses
treadmill prior to the 45-min downhill running–walking ex- Baseline demographic, health, and dietary characteristics
ercise. The treadmill was then accelerated and participants were summarized as sample means and SDs. Means were
were instructed to run or walk at 60% of their VO_ 2 max (de- compared between groups using independent t tests to assess
termined from the baseline cardiopulmonary test) for the balance achieved via randomization. Study compliance,
45 min. During the exercise protocol, 60% of VO _ 2 max was defined as the percentage of prescribed beverage consumed,
maintained by measuring VO _ 2 every 3 min and adjusting was assessed across groups using a 2-sample Wilcoxon
the treadmill speed accordingly. Between 15 and 18 min rank-sum test because these data were demonstrably non-
and 27 and 30 min of exercise, heart rate and blood pressure normal. Statistical analysis for the biomarkers of interest
were recorded. After the exercise session, subjects donated were performed by utilizing repeated-measures (group-by-
blood, received a small breakfast (granola bar (90 kcals and time) analyses of variance (ANOVAs) to control for intrain-
1 g protein)), and remained rested until a 4-h blood sample dividual variability. Because traditional post-hoc testing in a
was taken. repeated-measures framework does not fully account for in-
traindividual variability, significant differences in mean val-
Blood-collection procedure ues for main effects or interactions were explored further
Venous blood samples were drawn from the antecubital using profile plots and independent t tests. All statistical
vein into a 10-mL collection tube using a standard vacu- procedures were performed using SAS, version 9.1.3 soft-
tainer apparatus. Blood samples were allowed to stand at ware (Cary, N.C.) and a probability level of <0.05 was
room temperature for 10 min and were then centrifuged. adopted throughout.
For each sample, the plasma was removed and frozen
at –80 8C for later analysis. Blood samples were collected Results
and analyzed after the 4-week supplementation period, im-
mediately post exercise, and at 4, 24, and 48 h after the ex- Retention and baseline characteristics
ercise session. Except for the 4-h postexercise sample, all Thirty-three postmenopausal women were enrolled in this
blood samples were obtained after a 12-h fast and standar- study; one was unable to finish because of an unrelated
dized to the same time of day. musculoskeletal injury and one was dropped because of diet-
ary noncompliance. Data from only those participants who
Assessment of plasma markers of inflammation and were at least 80% compliant with the dietary protocol (n =
oxidative stress 31) were retained for later analysis. All baseline demo-
All biochemical analytical techniques were carried out ac- graphic data for these participants are presented in Table 2.
cording to standard procedures (Orenstein et al. 2009; Row- According to a Wilcoxon rank-sum test, there was no statis-
sey et al. 2009). Briefly, the activities of the plasma tically significant difference in dietary compliance between
oxidative defense enzymatic markers (SOD, GPx, and the treatment groups (p = 0.54). Additionally, at baseline
COX-2) were assessed using enzyme-linked immunoabsorb- there were no significant differences between the soy milk
ent assay (ELISA) technology with a microplate reader and dairy milk group with regard to age (p = 0.37), BMI
(Wallac Victor-1420, Perkin-Elmer Life Sciences, Boston, (p = 0.52), cardiopulmonary fitness level (p = 0.37), or base-
Mass.). TNF-a, IL-1b, and IL-6 concentrations were as- line protein intake (p = 0.22). All study participants were of
sessed using the Bio-Plex bead-based multiplex assays Caucasian descent and no significant body weight changes
(Luminex xMAP technology, Bio-Rad Laboratories, Inc., were observed over the course of the intervention (p =
Hercules, Calif.). All ELISA kit and Bio-Plex data were an- 0.22). All women participated in all blood draws (except for
alyzed in duplicate to reduce within-subject variability. The 1 participant who missed the immediately postexercise
intra-assay coefficients of variation for each biomarker were blood draw) during the course of the study.
as follows: SOD, 3.7%; COX-2, 8.7%; GPx, 7.2%; and Reported side-effects questionnaires administered at base-
IL-1b, IL-6, TNF-a were all <20%. Data points falling line and 2 and 4 weeks after consumption of soy or dairy
3 SDs above or below the group mean were labeled as out- milk revealed minimal and similar side effects in the

Published by NRC Research Press

Beavers et al. 265

Table 3. Plasma oxidative stress and inflammatory marker data by treatment group.

Marker T1 T2 T3 T4 T5 p value
SOD
Soy 1.25±0.46 0.92±0.36 0.9±0.59 1.22±0.66 1.3±0.41 Group: 0.88
Dairy 1.09±0.38 1.05±0.44 0.9±0.57 1.33±0.47 1.29±0.48 Time: <0.0001*
GxT: 0.34
COX-2
Soy 1.74±1.23 1.76±1.65 1.74±1.57 1.83±1.60 2.07±1.82 Group: 0.18
Dairy 1.08±0.96 1.50±1.38 1.29±0.87 1.45±1.05 1.15±1.42 Time: 0.84
GxT: 0.62
GPx
Soy 140.24±32.31 141.57±25.24 143.4±28.80 148.44±35.30 151.71±19.22 Group: 0.24
Dairy 137.46±33.54 125.70±36.81 140.07±33.68 139.33±32.97 132.99±27.21 Time: 0.28
GxT: 0.42
TNF-a
Soy 2.71±1.01 2.77±1.16 2.86±1.02 3.24±1.52 2.75±0.72 Group: 0.31
Dairy 3.53±0.85 3.09±1.15 2.67±0.77 3.15±0.83 3.04±1.02 Time: 0.17
GxT: 0.04*
IL-1b
Soy 0.73±0.29 0.77±0.34 0.74±0.27 0.76±0.35 0.81±0.37 Group: 0.82
Dairy 0.77±0.22 0.75±0.22 0.78±0.22 0.7±0.23 0.73±0.30 Time: 0.81
GxT: 0.11
IL-6
Soy 2.10±0.78 2.95±1.43 4.68±2.78 2.69±1.50 2.39±1.26 Group: 0.71
Dairy 2.48±1.63 2.94±3.73 4.43±2.96 2.39±1.23 2.30±1.02 Time: <0.0001*
GxT: 0.43
Note: Data are presented as means ± SD. GxT denotes the group-by-time interaction, Group denotes the main effect for group, and
Time denotes the main effect for time. T1, pre-exercise; T2, immediately post exercise; T3, 4-h post exercise; T4, 24-h post exercise;
T5, 48-h post exercise; SOD; superoxide dismutase (U
mL–1); COX-2, cyclooxygenase-2 (pg
mL–1); GPx, glutathione peroxidase
(nmol
min–1
mL–1); TNF-a, tumor necrosis factor-a (pg
mL–1); IL-1b, interleukin-1b (pg
mL–1); IL-6, interleukin-6 (pg
mL–1).
*Significant (p < 0.05).

2 groups. All participants completed the required exercise Discussion

protocol and reported significant increases in DOMS (pre-
Overall, our study does not support the notion that 4 weeks
exercise, 0.69 ± 1.16 AU vs. 24 h post exercise, 4.17 ±
of daily soy milk ingestion can attenuate systemic elevations
2.37 AU; p < 0.01). Lastly, no significant differences were
in markers of inflammation or oxidative defense in response
observed between groups for total daily caloric or macronu-
to eccentric exercise. However, data do suggest that the
trient intake at baseline or post supplementation.
downhill-running protocol utilized in this study can be effec-
tive in altering systemic activity and concentrations of select
Effects of the intervention on select biomarkers biomarkers (specifically, SOD, TNF-a, and IL-6), and that
A 2-way (treatment groups (2)  time point (5)) repeated- the ingestion of soy may help prevent fluctuations in plasma
measures ANOVA to control for within-individual variation TNF-a concentration after exposure to a stress-inducing
was conducted to evaluate the effects of the different sup- stimulus. Potential reasons for these findings are discussed
plements (soy milk or dairy milk) on the measured blood below and future research directions are presented
markers. All plasma marker data are shown in Table 3, and
significant results are displayed pictorially in Fig. 1. Briefly, Plasma markers of oxidative stress
no significant group effects were noted; however, a signifi- Harman (1956) proposed that ROS formed during normal
cant group-by-time interaction was observed for plasma oxygen metabolism induce macromolecular damage and that
TNF-a (p = 0.02) with values in the dairy group increased accumulation of such products accounts for the progressive
post supplementation and then dropping through the post ex- deleterious health changes we experience during aging. In
ercise period. TNF-a values for the soy group stayed more the body, the activity of certain antioxidant enzymes (e.g.,
consistent and slightly lower throughout most of interven- SOD, GPx, and COX-2) can be assessed as indicators of the
tion period. Significant time effects (p < 0.0001) were ob- oxidative stress imposed on the tissue.
served for plasma SOD and IL-6, with SOD activity In the present study, there was no significant effect of soy
decreasing from pre-exercise (T1) to the 4-h (T3) time pe- supplementation on the activity of any of the plasma oxida-
riod, and concentration of IL-6 increasing from pre-exercise tive defense enzymes assessed. However, we did note a time
(T1) to the 4-h (T3) time period. No significant effects in effect for SOD, in which SOD activity was shown to de-
response to diet or exercise were seen for plasma GPx, crease 4-h post exercise and return to baseline 2 days later.
COX-2, or IL-1b. This is contrary to what may be expected, because SOD ac-

Published by NRC Research Press

266 Appl. Physiol. Nutr. Metab. Vol. 35, 2010

Fig. 1. Column charts of significant changes in biomarker activity and concentration. Data are presented as means ± SD. From top, absolute
plasma superoxide dismutase (SOD) activity (U
mL–1), tumor necrosis factor-a (TNF-a) (pg
mL–1), and interleukin-6 (IL-6) (pg
mL–1) va-
lues are represented by time. T1, pre-exercise; T2, immediately post exercise; T3, 4-h post exercise; T4, 24-h post exercise; T5, 48-h post
exercise. The following sample sizes were observed for the soy group for T1 to T5, respectively: 16, 15, 15, 16, and 16. The following
sample sizes were observed for the dairy group for T1 to T5, respectively: 14, 15, 15, 15, and 15. *, Significant time effect (p < 0.0001);
{, significant group-by-time interaction (p < 0.05).

tivity is thought to increase during times of physical stress, recent animal study by Itoh et al. (2004), dietary calcium re-
like downhill running (Hollander et al. 2001). However, striction was shown to significantly upregulate the activities
based on the increases in plasma IL-6 and the responses to of manganese-SOD and copper zinc-SOD. Furthermore,
the DOMS questionnaires, it is believed that an inflamma- acute exercise, in addition to calcium restriction, decreased
tory state was induced, and that the decrease in the activity both SOD isoenzymes in the diaphragms of calcium-
of this marker may be due to a collective protective effect of restricted rats. Thus, it is speculated that in our study, be-
the supplements, or to inherent variability in measuring SOD cause of the calcium loading associated with consumption
activity. of both beverages, SOD activity may have been suppressed,
We could be observing a protective trend, with both the which would explain the observed results. Unfortunately,
soy and dairy milk products aiding in the conversion of without a placebo group, we cannot infer the exact reason
ROS to more inert byproducts, thereby decreasing the need for the time effect, although it is interesting to consider that
(and subsequent activity) of systemic SOD. From a dietary an unaccounted-for dietary factor may have played a role.
standpoint, similarities do exist between the soy and dairy Furthermore, because the main decline in SOD activity
milk products, including their carbohydrate, calcium, and vi- occurred 4-h post exercise, it may be that the carbohydrate-
tamin D composition, all of which may confer benefit. In a rich granola bar given to both groups immediately after ex-

Published by NRC Research Press

Beavers et al. 267

ercise caused a blunting of SOD activity. There is limited induced inflammatory response in postmenopausal women,
evidence in the animal literature to support this conjecture there are limitations worthy of mention. Previous literature
(Lygren and Hemre 2002). does exist to support our chosen supplementation amount
Additionally, the large variability in the activity of the en- and duration of study (Mitchell and Collins 1999; Djuric et
zymatic markers assessed in our study may have contributed al. 2001), but success is clearly contingent on participant ad-
to this counterintuitive finding. Conflicting findings have herence to the dietary protocol. Although self-report is an
been reported with regard to SOD, with studies showing in- accepted method for process evaluation (Jasti et al. 2005),
creases (Buczynski et al. 1991; Chen et al. 1994), decreases and adherence logs indicated excellent compliance among
(Tozzi-Ciancarelli et al. 2002), and no change (Elosua et al. both groups, because of funding constraints we do not have
2003) post exercise. Similar findings have also been reported any information regarding the isoflavone levels in biological
for GPx (Buczynski et al. 1991; Laaksonen et al. 1999; fluid (i.e., urine and blood). Moreover, the considerable
Akova et al. 2001; Miyazaki et al. 2001; Vider et al. 2001). interindividual variability in the bioavailability of soy isofla-
In a comparably designed study, Chen et al. (2004) found vones (Hendrich 2002) may have contributed to the reported
that isoflavone supplementation significantly increased pre- lack of effect. Future studies should consider using more ob-
exercise erythrocyte SOD activity and prevented the jective measures of adherence in their methodology, and
exercise-induced decrease in activities of GPx. The authors should account for individual differences in isoflavone bioa-
concluded that the results suggested that isoflavones can re- vailability when possible. Additionally, because SOD, GPx,
store the redox homeostasis of antioxidant enzymes altered and COX-2 are enzyme components of the oxidative defense
because of exercise. Although the results from the present system, and not direct markers of oxidative stress, it is
study do not support the findings of Chen et al., differences possible that we may not have captured the true effect of
in the intervention product (isolated soy isoflavone vs. soy supplementation on oxidative stress. Because of the var-
whole soy food), isoflavone dosage (150 mg
d–1 vs. iability in assessing these enzymes, future studies may con-
*90 mg
d–1), and exercise modality (cycling vs. downhill sider assessing more direct measures (i.e., F2-isoprostanes or
running) may explain the discrepancy. 8-oxo-2’-deoxyguanosine) of oxidative damage. Lastly,
although we selected sampling time points known to corre-
Plasma markers of inflammation
spond to peak concentrations–activities of biomarkers of in-
The cytokines IL-1b, IL-6, and TNF-a are pleiotropic terest (Peake et al. 2005a; Peake et al. 2005b), it is possible
molecules that play major roles in the inflammatory process that we failed to capture the peak concentrations–activities
(Trikha et al. 2003; Bradley 2008) and are linked to cardio- of all the biomarkers measured. Inferences, therefore, can
vascular morbidity (Mendall et al. 1997; Ridker et al. only be drawn from our sampling points.
2000a; 2000b) and mortality (Volpato et al. 2001). In our
study, we observed a significant group-by-time interaction
for the proinflammatory cytokine TNF-a, with postsupple- Conclusions
mentation values of TNF-a elevated in the dairy group com- Overall, results from our study do not support the notion
pared with the soy group, followed by a decrease in group that 4 weeks of daily soy milk ingestion can attenuate sys-
mean values to T3 and a return to postsupplementation val- temic elevations in markers of inflammation or oxidative de-
ues for the dairy group by 24-h post exercise. The soy fense activity. However, data do suggest that the downhill-
group, however, did not fluctuate in the postsupplementation
running protocol utilized in this study can be effective in al-
or postexercise periods with regard to TNF-a concentration.
tering systemic markers of inflammation and oxidative de-
Thus it appears that with regard to this proinflammatory cy-
fense. The findings from the plasma cytokine analysis
tokine, of which elevations are thought to promote chronic
portion of our study add to a mixed body of literature that
disease, the soy group fared better overall than the dairy
group. hints at the potential of soy foods to influence proinflamma-
tory cytokines, although results are far from conclusive. Fu-
Although a significant group-by-time interaction was not
observed for plasma IL-6 values, a marked increase in post- ture studies examining the ability of soy ingestion to reduce
exercise IL-6 concentrations (specifically T3) was observed systemic inflammation should address potential confounding
in both groups, suggesting that the exercise bout success- dietary factors by utilizing an inert control group to deter-
fully induced an acute inflammatory response. No time or mine if a different nutritional compound found in both soy
group trends were observed for plasma IL-1b; however, it and dairy milk could be a confounding, protective factor.
has been shown clearly that IL-6 rises in accordance with Despite the subdued findings in this study, the displacement
both IL-1b and TNF-a, because each stimulates the synthe- and lipid-lowering effects of soy consumption are well
sis and release of IL-6 (Hildebrand et al. 2005). The synthe- known. Coupled with the potential for improvement in sys-
sis of IL-6 then dampens the synthesis of IL-1b and TNF-a temic cytokine profiles, the results from this study still war-
(Barton 1997). Therefore, it is speculative that there may rant inclusion of soy foods in a heart-healthy diet.
have been a time effect for IL-1b and TNF-a, although the
time points from which we sampled may not have corre- Acknowledgements
sponded with the peak concentrations of these biomarkers. We thank Dr. Barbara J. Nicklas for thoughtfully review-
ing and significantly contributing to the completion of this
Limitations manuscript. Funding for this study was provided by White-
Although the present study contributes novel findings re- Wave Foods Inc., as well as by a dissertation grant provided
garding the effects of soy consumption on the exercise- by Baylor University.

Published by NRC Research Press

268 Appl. Physiol. Nutr. Metab. Vol. 35, 2010

References Transplant. 21(8): 2239–2246. doi:10.1093/ndt/gfl169. PMID:

16766544.
Akova, B., Sürmen-Gür, E., Gür, H., Dirican, M., Sarandöl, E., and
Féasson, L., Stockholm, D., Freyssenet, D., Richard, I., Duguez, S.,
Küçükoglu, S. 2001. Exercise-induced oxidative stress and mus-
Beckmann, J.S., and Denis, C. 2002. Molecular adaptations of
cle performance in healthy women: role of vitamin E supple-
neuromuscular disease-associated proteins in response to ec-
mentation and endogenous oestradiol. Eur. J. Appl. Physiol.
centric exercise in human skeletal muscle. J. Physiol.
84(1–2): 141–147. doi:10.1007/s004210000331. PMID:
543(Pt. 1): 297–306. doi:10.1113/jphysiol.2002.018689. PMID:
11394244.
12181300.
Azadbakht, L., Kimiagar, M., Mehrabi, Y., Esmaillzadeh, A., Hu,
Han, K.K., Soares, J.M., Jr., Haidar, M.A., de Lima, G.R., and
F.B., and Willett, W.C. 2007. Soy consumption, markers of in-
Baracat, E.C. 2002. Benefits of soy isoflavone therapeutic regi-
flammation, and endothelial function: a cross-over study in post-
men on menopausal symptoms. Obstet. Gynecol. 99(3): 389–
menopausal women with the metabolic syndrome. Diabetes
394. doi:10.1016/S0029-7844(01)01744-6. PMID:11864664.
Care, 30(4): 967–973. doi:10.2337/dc06-2126. PMID:17392557.
Harman, D. 1956. Aging: a theory based on free radical and radia-
Barton, B.E. 1997. IL-6: insights into novel biological activities.
Clin. Immunol. Immunopathol. 85(1): 16–20. doi:10.1006/clin. tion chemistry. J. Gerontol. 11(3): 298–300. PMID:13332224.
1997.4420. PMID:9325064. Hendrich, S. 2002. Bioavailability of isoflavones. J. Chromatogr. B
Beavers, K.M., Serra, M.C., Beavers, D.P., Cooke, M.B., and Analyt. Technol. Biomed. Life Sci. 777(1–2): 203–210. doi:10.
Willoughby, D.S. 2009. Soymilk supplementation does not alter 1016/S1570-0232(02)00347-1. PMID:12270213.
plasma markers of inflammation and oxidative stress in Hermansen, K., Hansen, B., Jacobsen, R., Clausen, P., Dalgaard,
postmenopausal women. Nutr. Res. 29(9): 616–622. doi:10. M., Dinesen, B., et al. 2005. Effects of soy supplementation on
1016/j.nutres.2009.09.002. PMID:19854376. blood lipids and arterial function in hypercholesterolaemic sub-
Bloomer, R.J., Goldfarb, A.H., Wideman, L., McKenzie, M.J., and jects. Eur. J. Clin. Nutr. 59(7): 843–850. doi:10.1038/sj.ejcn.
Consitt, L.A. 2005. Effects of acute aerobic and anaerobic exer- 1602151. PMID:15900307.
cise on blood markers of oxidative stress. J. Strength Cond. Res. Hildebrand, F., Pape, H.C., and Krettek, C. 2005. The importance
19(2): 276–285. doi:10.1519/14823.1. PMID:15903362. of cytokines in the posttraumatic inflammatory reaction. Unfall-
Box, W., Hill, S., and DiSilvestro, R.A. 2005. Soy intake plus chirurg, 108(10): 793–794, 796–803. [In German.] doi:10.1007/
moderate weight resistance exercise: effects on serum concentra- s00113-005-1005-1. PMID:16175346.
tions of lipid peroxides in young adult women. J. Sports Med. Hill, S., Box, W., and DiSilvestro, R.A. 2004. Moderate intensity
Phys. Fitness, 45(4): 524–528. PMID:16446685. resistance exercise, plus or minus soy intake: effects on serum
Bradley, J.R. 2008. TNF-mediated inflammatory disease. J. Pathol. lipid peroxides in young adult males. Int. J. Sport Nutr. Exerc.
214(2): 149–160. doi:10.1002/path.2287. PMID:18161752. Metab. 14(2): 125–132. PMID:15118187.
Buczynski, A., Kedziora, J., Tkaczewski, W., and Wachowicz, B. Hilpert, K.F., Kris-Etherton, P.M., and West, S.G. 2005. Lipid re-
1991. Effect of submaximal physical exercise on antioxidative sponse to a low-fat diet with or without soy is modified by C-
protection of human blood platelets. Int. J. Sports Med. 12(1): reactive protein status in moderately hypercholesterolemic
52–54. doi:10.1055/s-2007-1024655. PMID:2030060. adults. J. Nutr. 135(5): 1075–1079. PMID:15867284.
Chen, C.Y., Bakhiet, R.M., Hart, V., and Holtzman, G. 2004. Iso- Hollander, J., Fiebig, R., Gore, M., Ookawara, T., Ohno, H., and Ji,
flavones restore altered redox homesostasis of antioxidant en- L.L. 2001. Superoxide dismutase gene expression is activated by
zymes in health young men undergoing 80% peak oxygen a single bout of exercise in rat skeletal muscle. Pflugers Arch.
consumption (VO2pk) exercise. Nutr. Res. 24(5): 347–359. 442(3): 426–434. doi:10.1007/s004240100539. PMID:11484775.
doi:10.1016/j.nutres.2004.02.001. Itoh, M., Oh-Ishi, S., Hatao, H., Leeuwenburgh, C., Selman, C.,
Chen, M.F., Hsu, H.C., and Lee, Y.T. 1994. Effects of acute exer- Ohno, H., et al. 2004. Effects of dietary calcium restriction and
cise on the changes of lipid profiles and peroxides, prostanoids, acute exercise on the antioxidant enzyme system and oxidative
and platelet activation in hypercholesterolemic patients before stress in rat diaphragm. Am. J. Physiol. Regul. Integr. Comp.
and after treatment. Prostaglandins, 48(3): 157–174. doi:10. Physiol. 287(1): R33–R38. PMID:14764436.
1016/0090-6980(94)90016-7. PMID:7809382. Jasti, S., Siega-Riz, A.M., Cogswell, M.E., Hartzema, A.G., and
Djuric, Z., Chen, G., Doerge, D.R., Heilbrun, L.K., and Kucuk, O. Bentley, M.E. 2005. Pill count adherence to prenatal multivita-
2001. Effect of soy isoflavone supplementation on markers of min/mineral supplement use among low-income women. J.
oxidative stress in men and women. Cancer Lett. 172(1): 1–6. Nutr. 135(5): 1093–1101. PMID:15867287.
doi:10.1016/S0304-3835(01)00627-9. PMID:11595123. Jenkins, D.J., Kendall, C.W., Connelly, P.W., Jackson, C.J., Parker,
Dröge, W. 2002. Free radicals in the physiological control of cell T., Faulkner, D., and Vidgen, E. 2002. Effects of high- and low-
function. Physiol. Rev. 82(1): 47–95. PMID:11773609. isoflavone (phytoestrogen) soy foods on inflammatory biomar-
Elia, D., Stadler, K., Horváth, V., and Jakus, J. 2006. Effect of soy- kers and proinflammatory cytokines in middle-aged men and
and whey protein-isolate supplemented diet on the redox para- women. Metabolism, 51(7): 919–924. doi:10.1053/meta.2002.
meters of trained mice. Eur. J. Nutr. 45(5): 259–266. doi:10. 33352. PMID:12077742.
1007/s00394-006-0593-z. PMID:16575496. Laaksonen, D.E., Atalay, M., Niskanen, L., Uusitupa, M.,
Elosua, R., Molina, L., Fito, M., Arquer, A., Sanchez-Quesada, Hänninen, O., and Sen, C.K. 1999. Blood glutathione homeosta-
J.L., Covas, M.I., et al. 2003. Response of oxidative stress bio- sis as a determinant of resting and exercise-induced oxidative
markers to a 16-week aerobic physical activity program, and to stress in young men. Redox Rep. 4(1–2): 53–59. doi:10.1179/
acute physical activity, in healthy young men and women. 135100099101534648. PMID:10714277.
Atherosclerosis, 167(2): 327–334. doi:10.1016/S0021-9150(03) Lygren, B., and Hemre, G. 2002. Influence of dietary carbohydrate
00018-2. PMID:12818416. on antioxidant enzyme activities in liver of atlantic salmon.
Fanti, P., Asmis, R., Stephenson, T.J., Sawaya, B.P., and Franke, Aquacult. Int. 9(5): 421–427. doi:10.1023/A:1020530432508.
A.A. 2006. Positive effect of dietary soy in ESRD patients with Martinez-Villaluenga, C., Dia, V.P., Berhow, M., Bringe, N.A., and
systemic inflammation–correlation between blood levels of the Gonzalez de Mejia, E. 2009. Protein hydrolysates from beta-
soy isoflavones and the acute-phase reactants. Nephrol. Dial. conglycinin enriched soybean genotypes inhibit lipid accumula-

Published by NRC Research Press

Beavers et al. 269

tion and inflammation in vitro. Mol. Nutr. Food Res. 53(8): Greenlund, K., et al.; American Heart Association Statistics
1007–1018. doi:10.1002/mnfr.200800473. PMID:19603404. Committee and Stroke Statistics Subcommittee. 2007. Heart dis-
Mendall, M.A., Patel, P., Asante, M., Ballam, L., Morris, J., ease and stroke statistics–2007 update: a report from the Ameri-
Strachan, D.P., et al. 1997. Relation of serum cytokine concen- can Heart Association Statistics Committee and Stroke Statistics
trations to cardiovascular risk factors and coronary heart disease. Subcommittee. Circulation, 115(5): e69–e171. doi:10.1161/
Heart, 78(3): 273–277. PMID:9391290. CIRCULATIONAHA.106.179918. PMID:17194875.
Mitchell, J.H., and Collins, A.R. 1999. Effects of a soy milk sup- Rowsey, P.J., Metzger, B.L., Carlson, J., and Gordon, C.J. 2009.
plement on plasma cholesterol levels and oxidative DNA da- Long-term exercise training selectively alters serum cytokines
mage in men–a pilot study. Eur. J. Nutr. 38(3): 143–148. involved in fever. Biol. Res. Nurs. 10(4): 374–380. doi:10.1177/
doi:10.1007/s003940050055. PMID:10443336. 1099800408329409. PMID:19190031.
Miyazaki, H., Oh-ishi, S., Ookawara, T., Kizaki, T., Toshinai, K., Rumpler, W.V., Kramer, M., Rhodes, D.G., Moshfegh, A.J., and
Ha, S., et al. 2001. Strenuous endurance training in humans re- Paul, D.R. 2008. Identifying sources of reporting error using
duces oxidative stress following exhausting exercise. Eur. J. measured food intake. Eur. J. Clin. Nutr. 62(4): 544–552.
Appl. Physiol. 84(1-2): 1–6. doi:10.1007/s004210000342. doi:10.1038/sj.ejcn.1602742. PMID:17426745.
PMID:11394236. Ryan-Borchers, T.A., Park, J.S., Chew, B.P., McGuire, M.K.,
Oh, H.Y., Lim, S., Lee, J.M., Kim, D.Y., Ann, E.S., and Yoon, S. Fournier, L.R., and Beerman, K.A. 2006. Soy isoflavones modu-
2007. A combination of soy isoflavone supplementation and late immune function in healthy postmenopausal women. Am. J.
exercise improves lipid profiles and protects antioxidant de-
Clin. Nutr. 83(5): 1118–1125. PMID:16685055.
fense-systems against exercise-induced oxidative stress in ovar-
Stevens, J. 2002. Applied multivariate statistics for the social
iectomized rats. Biofactors, 29(4): 175–185. doi:10.1002/biof.
sciences. Lawrence Erlbaum Associates, Mahwah, N.J.
5520290402. PMID:18057549.
Taniyama, Y., and Griendling, K.K. 2003. Reactive oxygen species
Orenstein, S.B., Qiao, Y., Kaur, M., Klueh, U., Kreutzer, D.L., and
in the vasculature: molecular and cellular mechanisms. Hyper-
Novitsky, Y.W. 2009. Human monocyte activation by biologic
tension, 42(6): 1075–1081. doi:10.1161/01.HYP.0000100443.
and biodegradable meshes in vitro. [Epub ahead of print.] Surg.
09293.4F. PMID:14581295.
Endosc. 24(4): 805–811. doi:10.1007/s00464-009-0664-3.
Tozzi-Ciancarelli, M.G., Penco, M., and Di Massimo, C. 2002. In-
Peake, J., Nosaka, K., and Suzuki, K. 2005a. Characterization of
inflammatory responses to eccentric exercise in humans. Exerc. fluence of acute exercise on human platelet responsiveness: pos-
Immunol. Rev. 11: 64–85. PMID:16385845. sible involvement of exercise-induced oxidative stress. Eur. J.
Peake, J.M., Suzuki, K., Wilson, G., Hordern, M., Nosaka, K., Appl. Physiol. 86(3): 266–272. doi:10.1007/s00421-001-0542-8.
Mackinnon, L., and Coombes, J.S. 2005b. Exercise-induced PMID:11990737.
muscle damage, plasma cytokines, and markers of neutrophil ac- Trikha, M., Corringham, R., Klein, B., and Rossi, J.F. 2003. Tar-
tivation. Med. Sci. Sports Exerc. 37(5): 737–745. doi:10.1249/ geted anti-interleukin-6 monoclonal antibody therapy for cancer:
01.MSS.0000161804.05399.3B. PMID:15870626. a review of the rationale and clinical evidence. Clin. Cancer
Ridker, P.M., Hennekens, C.H., Buring, J.E., and Rifai, N. 2000a. Res. 9(13): 4653–4665. PMID:14581334.
C-reactive protein and other markers of inflammation in the pre- Vider, J., Lehtmaa, J., Kullisaar, T., Vihalemm, T., Zilmer, K.,
diction of cardiovascular disease in women. N. Engl. J. Med. Kairane, C., et al. 2001. Acute immune response in respect to
342(12): 836–843. doi:10.1056/NEJM200003233421202. PMID: exercise-induced oxidative stress. Pathophysiology, 7(4): 263–
10733371. 270. doi:10.1016/S0928-4680(00)00057-2. PMID:11228396.
Ridker, P.M., Rifai, N., Stampfer, M.J., and Hennekens, C.H. Volpato, S., Guralnik, J.M., Ferrucci, L., Balfour, J., Chaves, P.,
2000b. Plasma concentration of interleukin-6 and the risk of fu- Fried, L.P., and Harris, T.B. 2001. Cardiovascular disease, inter-
ture myocardial infarction among apparently healthy men. Cir- leukin-6, and risk of mortality in older women: the women’s
culation, 101(15): 1767–1772. PMID:10769275. health and aging study. Circulation, 103(7): 947–953. PMID:
Rosamond, W., Flegal, K., Friday, G., Furie, K., Go, A., 11181468.

Published by NRC Research Press