Antioxidant Supplementation Lowers Exercise-Induced Oxidative Stress in Young Overweight Adults


University of Virginia, Center for the Study of Complementary and Alternative Therapies, P.O. Box 800905, The Blake Center, Charlottesville, VA 22908-0905. E-mail:


Objective: To determine whether antioxidant (AOX) supplementation attenuates post-exercise oxidative stress and contributors to oxidative stress (inflammation, blood lipids) in overweight young adults.

Research Methods and Procedures: This was a randomized, double-blind, controlled study. Overweight (BMI, 33.2 ± 1.9 kg/m2) and comparative normal-weight (BMI, 21.9 ± 0.5 kg/m2) adults 18 to 30 years old (total N = 48) were enrolled. Participants received either daily antioxidant (AOX) treatment (800 IU of vitamin E, 500 mg of vitamin C, 10 mg of β-carotene) or placebo (PL) for 8 weeks for a total of four groups. All participants completed a standardized 30-minute cycle exercise bout at baseline and 8 weeks. Exercise-induced changes in lipid hydroperoxide (ΔPEROX), C-reactive protein (ΔCRP), interleukin-6 (ΔIL-6), cholesterol subfractions, triglycerides, total AOX status (ΔTAS), and adiponectin were assessed.

Results: Exercise-induced ΔPEROX was lower in the overweight-AOX group (0.09 nM/kg per min) compared with PL-treated overweight and normal-weight groups (0.98, 0.53 nM/kg per min) by 8 weeks (p < 0.05). Adiponectin was increased in both overweight and normal-weight AOX groups (22.1% vs. 3.1%; p < 0.05) but reduced in PL groups. ΔIL-6, Δtotal cholesterol, and Δlow-density lipoprotein-cholesterol concentrations during exercise were lower in the AOX-treated groups compared with PL groups (all p < 0.05). After controlling for BMI, the Δtotal cholesterol, Δlow-density lipoprotein-cholesterol, Δadiponectin, and ΔTAS explained 59.1% of the variance of the regression model of the ΔPEROX by 8 weeks (total model R2 = 0.600; p = 0.015).

Discussion: AOX lowers exercise-induced oxidative stress in overweight adults. Inflammatory and lipid markers may also be attenuated with AOX. Further studies are needed to determine whether AOX may be used in cardiovascular disease prevention in the overweight population.


Oxidative stress is an imbalance between the antioxidant (AOX)1 and pro-oxidant processes that occur in metabolism, an imbalance that causes excessive production of free radicals and taxes the systemic AOX defenses (1). The net result is accumulation of oxidative stress by-products and oxidative damage to tissues. Acute exercise increases oxidative stress, especially when the exercise intensity is high (2). In normal healthy physiology, free radicals can be combated with endogenous AOX defenses. In obesity, AOX defenses are compromised (3, 4, 5), and several sources of free radical production might contribute to exacerbation of oxidative damage during exercise.

Oxidative stress is exacerbated in both younger and older obese persons after acute exercise compared with their non-obese counterparts (6, 7). Several potential mechanisms for exercise-induced oxidative stress in the obese population are systemic inflammation (8), increased lipid substrate available for oxidation (9, 10), and compromised AOX status (11). Inflammatory cytokines such as interleukins (ILs) 1 and 6, tumor necrosis factor α, and C-reactive protein (CRP) are elevated during acute exercise and normalize hours to days after the exercise bout (12). Although some of the cytokine release (IL-6) is attributable to contracting skeletal muscle (13), excessive adipose tissue may exacerbate the inflammatory cytokine response to acute exercise because adipose tissue also produces these cytokines (14). Alternatively, suppressed production of anti-inflammatory adipocyte-derived adiponectin in obesity may permit inflammation during exercise to be enhanced (15). Plasma lipids are oxidized at a faster rate in obese persons compared with those of non-obese persons (10). In obesity, elevated lipid pools in the adipose tissue depots or in the blood are targets for free radical attack (16). During exercise when free radicals are generated at higher rates, plasma lipid pools in obesity may be susceptible to oxidative damage during exercise. Lastly, blood levels of specific AOXs such as carotenoids (β-carotene) and vitamins C and E are lower in obese than non-obese persons (17, 18). Total AOX status (TAS), an estimate of overall AOX capacity of the blood, is also low in obesity (11). During acute exercise, dietary AOX insufficiency in obesity likely compromises the protective margin of the overall AOX defense system and permits free radical damage.

AOX supplementation with a combination of AOXs may be one intervention to reduce exercise oxidative stress in the vulnerable obese population. Although supplementation with individual nutrients such as vitamin C (19) or vitamin E (20, 21) may protect against exercise oxidative stress, the major dietary AOXs, vitamins E and C and β-carotene, work together within intracellular AOX defense systems (22) and may be more effective when administered together. Evidence suggests that administration of one single AOX may even cause pro-oxidation processes (23). It is presently unknown whether the combination of vitamins E and C and β-carotene can reduce exercise-induced oxidative stress in the susceptible overweight/obese population. Therefore, the purpose of this study was to determine whether AOX supplementation (vitamins E and C and β-carotene) lowers exercise-induced oxidative stress in overweight young adults. A secondary purpose was to determine whether two major mechanisms underlying oxidative stress in obesity, inflammation (imbalance of inflammatory cytokines and adiponectin), blood lipids, and low TAS are attenuated by AOX supplementation in this population.

Research Methods and Procedures


Forty-eight apparently healthy young men and women (18 to 30 years old, 18 men, 30 women) volunteered as participants for this study. Participants were recruited by flyers, newspaper, institutional Internet sites, and clinic advertisements from the central Virginia region. All participants had to meet the following criteria before enrollment in the study: no participation in regular physical activity (vigorous exercise two times or more per week); no chronic health problems or current smoking; no history of cardiovascular, metabolic, or respiratory disease; no consumption of AOX supplements within the past 6 months; and no current usage of any form of hormonal forms of contraceptives. All participants read and signed a written informed consent statement consistent with university policy on protection of human subjects. The protocol of the study was approved by the Institutional Review Board for Studies Involving Human Subjects at the University of Virginia (UVA) and conforms to the guidelines involving the use of human participants as outlined by the American College of Sports Medicine.

Experimental Design

Participants and Study Groups

After the investigators had determined BMI values [normal weight (BMI < 25 kg/m2) and overweight and obese (BMI ≥ 25 kg/m2)], participants were stratified into normal-weight or overweight groups. Within each group, participants were randomized by the UVA investigational pharmacist to receive either AOX treatment or placebo (PL). Hereafter, the participant stratifications will be referred to as normal-weight, AOX-treated (N-AOX); normal-weight, PL (N-PL); overweight, AOX-treated (O-AOX); and overweight, PL (O-PL). The first character refers to the adiposity status (normalweight, overweight), and the second characters refer to the AOX treatment or PL (AOX, PL). A total of 65 participants were enrolled in the study; among all normal-weight and overweight participants, those who withdrew were 11 participants who had overcommitted themselves and requested to withdraw. Three were withdrawn by the investigators for failure to comply with the study visits. Two participants completed two visits and stopped responding to investigator contact. One had child care issues and could not complete the study.

AOX Intervention

Groups assigned to the AOX treatment were prescribed a mixture of vitamin E (800 IU/d), vitamin C (500 mg/d), and β-carotene (10 mg/d) by the physician investigator. The supplements or PL were administered by the UVA investigational pharmacist. Supplements were provided in opaque bottles. These dosages of AOXs were chosen based on previous studies that showed reductions in lipid peroxidation with this range of combined supplements (24). PLs were provided in identical opaque bottles with color-matched, odorless capsules similar to those of the AOX. Treatment was administered for 8 weeks, a time by which plasma concentrations of vitamin E stabilize with supplementation (25).

Testing Schedule

All participants visited the testing area five times (three pre-intervention, two post-intervention) and were acclimated to the UVA Exercise Physiology Laboratory at the General Clinical Research Center (GCRC). During Visit 1, body composition and vital signs were measured, and 3-day dietary record forms were provided to participants. During Visit 2, aerobic fitness levels [peak oxygen consumption (Vo2peak)] were estimated using a load-incremented cycle ergometer protocol. During Visit 3, a constant-load cycle ergometer test was administered to each participant with blood draws before and after the constant-load test. All visits were completed within a 14-day period. Visits 4 and 5 were repeats of Visits 2 and 3 that occurred after the supplementation period.

Anthropometric Measures

Height and weight were measured using a standard medical grade scale. For classification of obesity, waist and hip girths were measured using a soft cloth measuring tape at anatomical landmarks described by the American College of Sports Medicine (26). BMI values were determined by the following: BMI = weight (kilograms)/height (meters)2. Body volume was estimated using air displacement plethysmography in a BodPod device (Bod-Pod; Life Measurement Instruments, Concord, CA) corrected for thoracic gas volume; body density was calculated and used to predict body fat using the Siri equation (27).

Dietary Analysis

To determine whether nutritional intake was different between groups or over time, 3-day dietary record forms were provided to each participant with standard instructions on how to complete the record before Visit 3. Participants were instructed to estimate servings of foods using household measurements (volume) as described in national dietary guidance documents as previously described (28). Each participant received individual training sessions with the same investigator with regard to measuring technique and volume estimation. Picture books of portion sizes were also provided after the dietary estimation training session (28). Diet records were assessed by the same investigator using Nutritionist Pro Software (version 2.1.13; First DataBank, San Bruno, CA) and were analyzed for macronutrient, AOX, and caloric intake. To ensure the stability of the habitual diet of the subjects, all subjects completed a second 3-day dietary record 8 weeks later (28).

Vo2peak/Lactate Threshold (LT) Test

To reduce the potential effects of the weight-bearing effect on lipid hydroperoxide (PEROX) responses, a cycle ergometry mode of exercise was selected for this study. After a 12-hour overnight fast, participants arrived at the GCRC. A venous catheter was inserted into a forearm vein. Participants completed a Vo2peak/LT test on an electronically braked cycle ergometer (Ergo Metrics 800S; Sensor Medics, Yorba Linda, CA). The initial power output (PO) was set at 20 W, and the PO was increased 15 W every 3 minutes until volitional fatigue. Blood samples were taken at rest and during the last 15 seconds of each exercise stage for the measurement of blood lactate concentration (model 2700; YSI Instruments, Yellow Springs, OH). Heart rates (HRs), blood pressures, and ratings of perceived exertion were collected at every exercise stage (26).

Constant-Load Cycle Ergometer Test

To standardize the duration and relative intensity of the exercise challenge among all subjects, a 30-minute constant-load exercise test was performed in a fasting state. The PO for the 30-minute constant-load PO (CLPO) aerobic exercise sessions (see below) was calculated as follows: CLPO = PO at LT + 0.50 (PO at Vo2peak − PO at LT) (1.5 LT). Each individual self-selected his/her seat height and the front console option during this exercise test. Each participant pedaled at ∼50% of the CLPO during Minute 1 and at ∼75% of the CLPO during Minute 2, and CLPO was attained by Minute 3. Subjects were required to pedal between 60 and 100 rpm. If pedaling cadence dropped to <60 rpm, the PO was reduced, but the flywheel resistance was increased by the ergometer to keep the work consistent at all times.

Metabolic, Hemodynamic, and Perceptual Responses

Metabolic data were collected during both the Vo2peak/LT protocol and the CLPO tests, using standard open-circuit spirometric techniques (Vmax 229; Sensor Medics). Ratings of perceived exertion were assessed at the end of each stage during the Vo2peak/LT protocol and every 10 minutes during the CLPO protocol using the Borg Scale. HR was determined electrocardiographically (Marquette Max-1 electrocardiograph, Marquette, WI). HR, blood pressures, and lactate were assessed every 10 minutes during the CLPO test.

Blood Sampling

Blood samples were collected from the participants’ venous catheter into heparinized Vacutainer tubes before and immediately post-CLPO test. Blood samples were analyzed for PEROX, TAS, cholesterol, and lipid subfractions. Other portions of whole blood were analyzed for glucose and hemoglobin A1c (HbA1c) levels, and inflammatory cytokines and adiponectin. A portion of the blood was immediately centrifuged at 1500g for 5 minutes to separate plasma from red blood cell pellets. Plasma samples were immediately frozen and stored at −70 °C until analysis.

Cholesterol, Glucose, and HbA1c

To document levels of lipid substrates in the blood available for oxidation, plasma cholesterol subfractions [total cholesterol (TC), high-density lipoprotein-cholesterol (HDL-C), triglycerides] were analyzed by the UVA Health System Clinical and Toxicology Laboratories using standard automated spectrophotometric laboratory procedures (Olympus AU640, Olympus calibrator catalog no. DR0040, Olympus, Tokyo, Japan; and Genzyme HDL-C calibrator catalog no. 80-4529-00, Genzyme, Cambridge, MA). Low-density lipoprotein-cholesterol (LDL-C) was estimated from the following equation: LDL-C = TC − HDL-C − (triglycerides/5). Blood glucose was assessed using an Olympus AU640 procedure, in which glucose was phosphorylated by hexokinase in the presence of ATP and magnesium. The resultant glucose-6-phosphonate dehydrogenase oxidized glucose-6-phosphonate to 6-phosphogluconate and reduced nicotinamide adenine dinucleotide to NADH. The change in absorbance at 340/380 was proportional to the amount of glucose in the sample (Olympus Glucose Reagent, Calibrator catalog no. DR0040). HbA1c was analyzed using automated high-performance liquid chromatography (Tosoh G7 Automated HPLC Analyzer, using TSKgel G and HSi elution columns). All samples were performed in duplicate.


As an estimate of the total AOX capacity of the plasma, a colorimetric commercial kit was used (Randox Laboratories Ltd., Antrim, UK; TAS catalog no. NX2332). In brief, 2,2-azion-di-[3-ethylbenzenthiazoline sulfonate] (ABTS) was incubated with a peroxidase and hydrogen peroxide to produce a radical cation ABTS. The suppression of the ABTS radical in vitro was proportional to the AOX level in the plasma samples. Samples were read in a single batch at 600 nm on a spectrophotometer. Samples were expressed in AOX capacity in millimolar plasma. All samples were performed in duplicate. The coefficient of variation for this assay was 4%.

Inflammatory Cytokines

Inflammatory cytokines IL-6 and CRP were measured as the major adipokines in this study. Cytokines were measured in the GCRC Core Laboratory using the following techniques: an enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN) for measurement of IL-6 and an Immunolite 2000 ELISA for high-sensitivity (hs)-CRP. All samples were performed in duplicate. Measurement of adiponectin was made using an ELISA technique (LINCO Research Inc., St. Charles, MO). Because adiponectin levels are not acutely changed with exercise (29), only resting adiponectin levels at baseline and 8 weeks were collected.

Lipid Peroxidation Measurements

PEROX was quantified using the colorimetric ferrous oxidation/xylenol orange spectrophotometric technique previously described, where cumene hydroperoxide was used as the standard for this assay, and samples were read at 580 nm (30). All samples were performed in triplicate in a single batch analysis. The coefficient of variation for this assay was 4%. The PEROX values were adjusted by Vo2 values as previously described (7) to account for possible differences in the oxygen exposure during exercise.


All data are expressed as mean ± standard error. Data were analyzed using the SPSS/PC statistical program (version 12.0 for Windows; SPSS, Inc., Chicago, IL). Descriptive variables were analyzed using a two-way ANOVA. If differences did not exist between groups at baseline, repeated measures ANOVAs were performed for the change scores (Δ values from pre- to post-CLPO exercise at baseline and 8 weeks) for the inflammatory cytokines, blood lipids, PEROX, and TAS. The between-group factors were adiposity status (non-obese, overweight) and treatment (AOXs, PL), and the within-group factor was time (pre- and post-exercise, baseline, and 8 weeks). When baseline differences existed for blood measures, two-way analyses of covariance were performed using the baseline value as the covariate. The between-group factors were adiposity status (non-obese, overweight) and treatment (AOXs, PL).

Lastly, a hierarchical regression analysis was performed on variables postulated to contribute to exercise induced oxidative stress in obesity to determine which variables contribute most to the ΔPEROX/ΔVo2 (the change in adjusted lipid peroxidation during exercise from baseline to 8 weeks). The level of significance was set at 0.05 for all statistical tests.


Subject Characteristics and Initial Blood Measurements

The subject characteristics are presented in Table 1 for all experimental groups. As expected, several indices of obesity were different between the normal-weight and overweight groups. In addition, resting blood pressures and aerobic fitness (Vo2peak) levels were different from the normal-weight groups (p > 0.05).

Table 1.  Subject characteristics at baseline. Values are means ± standard error
 AOX (n = 12)PL (n = 13)AOX (n = 12)PL (n = 11)
  • AOX, antioxidant; PL, placebo; WHR, waist-to-hip ratio; SBP, systolic blood pressure; DBP, diastolic blood pressure; Vo2peak, peak oxygen consumption.

  • *

    Different from normal-weight group at p < 0.05.

Age22.3 ± 0.722.3 ± 0.722.5 ± 0.926.5 ± 0.8
Height (cm)173.5 ± 2.7169.9 ± 2.1171.0 ± 3.0171.1 ± 2.4
Weight (kg)68.9 ± 2.861.9 ± 2.394.0 ± 5.5*99.5 ± 6.6*
Fat mass (kg)14.1 ± 1.213.9 ± 0.946.4 ± 4.9*47.7 ± 5.6*
Fat free mass (kg)54.8 ± 2.647.9 ± 2.257.0 ± 4.255.2 ± 4.0
BMI (kg/m2)22.7 ± 0.521.4 ± 0.532.7 ± 1.933.9 ± 2.0
Waist circumference (cm)75.5 ± 1.770.4 ± 1.695.7 ± 3.9*100.2 ± 4.1*
WHR0.76 ± 0.010.74 ± 0.10.83 ± 0.02*0.85 ± 0.02*
SBP (mm Hg)116.1 ± 4.1113.1 ± 3.1125.1 ± 4.2*133.1 ± 3.2*
DBP (mm Hg)69.3 ± 2.868.1 ± 2.472.0 ± 2.178.5 ± 2.9*
Vo2peak (mL/kg per min)33.4 ± 1.830.9 ± 1.720.5 ± 2.1*19.1 ± 1.4*

HDL and triglyceride concentrations and LDL-to-HDL ratios were higher in the overweight groups compared with the normal-weight groups, respectively (p < 0.05, Table 2). Also, basal concentrations of CRP and IL-6 were higher in the overweight groups than the normal-weight groups (p < 0.05). Basal adiponectin levels were lower in the overweight groups compared with the normal-weight groups (p < 0.05).

Table 2.  Fasting glucose, blood lipid measurements, and inflammatory cytokine measures in normal-weight and overweight groups by AOX treatment at baseline. Values are means ± standard error
 AOX (n = 12)PL (n = 13)AOX (n = 12)PL (n = 11)
  • AOX, antioxidant; PL, placebo; HbA1c, hemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; PEROX, lipid hydroperoxides; TAS, total antioxidant status; IL-6, interleukin 6; hs-CRP, high sensitivity C-reactive protein.

  • *

    Different from normal-weight groups at p < 0.05.

Glucose (mM/L)4.4 ± 0.14.4 ± 0.14.7 ± 0.35.0 ± 0.3
HbA1c (%)5.0 ± 0.15.0 ± 0.14.9 ± 0.15.0 ± 0.1
Total cholesterol (mM/L)4.3 ± 0.13.8 ± 0.24.0 ± 0.34.3 ± 0.3
HDL-C (mM/L)173.5 ± 2.7169.9 ± 2.1171.0 ± 3.0171.1 ± 2.4
LDL-C (mM/L)68.9 ± 2.861.9 ± 2.394.0 ± 5.5*99.5 ± 6.6*
Triglycerides (mM/L)14.1 ± 1.213.9 ± 0.946.4 ± 4.9*47.7 ± 5.6*
LDL/HDL ratio1.6 ± 0.11.5 ± 0.22.0 ± 0.2*2.3 ± 0.3*
PEROX (mM/mL)2.9 ± 0.32.8 ± 0.23.0 ± 0.32.8 ± 0.3
TAS (mM/mL)1.4 ± 0.11.3 ± 0.11.3 ± 0.11.3 ± 0.2
IL-6 (pg/mL)1.0 ± 0.31.4 ± 0.32.8 ± 0.5*2.2 ± 0.4*
hs-CRP (mg/dL)0.06 ± 0.020.18 ± 0.080.55 ± 0.17*0.67 ± 0.20*
Adiponectin (ng/mL)12,100.1 ± 2,05613,039 ± 3,8149,027 ± 1,472*8,597 ± 1,1,53*

Dietary Analyses

Correlation values (r) between specific macro- and micronutrient intake between the first and second dietary records ranged from 0.70 to 0.93, indicating a stable intra-subject dietary pattern. The only significant differences with regard to dietary intakes were found with saturated fat and copper (Table 3). The O-PL had higher saturated fat intake than the N-PL group (p < 0.05). The O-AOX had lower copper intake compared with the N-AOX group (p < 0.05). No other significant differences were found between groups for other nutrient intakes.

Table 3.  Average dietary intakes of normal-weight and overweight groups by AOX treatment. Values are means ± standard error
  • AOX, antioxidant; PL, placebo; N-PL, normal-weight, PL; N-AOX, normal-weight, AOX-treated.

  • *

    Different from N-PL at p < 0.05.

  • **

    Different from N-AOX at p < 0.05.

Energy intake (kcal)2357 ± 1711940 ± 2102104 ± 2132422 ± 223
Fat (% of kcal)35.4 ± 2.428.3 ± 2.330.0 ± 2.336.3 ± 3.3
Fat (g/d)95.6 ± 2.964.5 ± 12.274.1 ± 10.2100.8 ± 12.9
Saturated fat (g/d)30.8 ± 4.619.7 ± 2.923.9 ± 3.835.9 ± 4.5*
Monounsaturated fat (g/d)31.1 ± 4.819.5 ± 6.021.6 ± 3.628.3 ± 3.9
Polyunsaturated fat (g/d)17.5 ± 2.610.1 ± 3.39.3 ± 1.415.0 ± 2.1
Protein (% of kcal)13.4 ± 0.815.1 ± 1.115.5 ± 1.116.1 ± 1.5
Protein (g/d)82.6 ± 10.671.6 ± 7.176.7 ± 9.298.0 ± 12.9
Carbohydrate (% kcal)48.8 ± 2.953.1 ± 2.755.0 ± 2.646.6 ± 4.1
Carbohydrate (g/d)286.3 ± 16.8261.6 ± 26.0287.8 ± 28.3281.8 ± 32.2
Fiber (g/d)19.1 ± 2.116.3 ± 2.514.8 ± 1.717.2 ± 2.3
Vitamin C (mg)123.9 ± 33.9113.4 ± 28.9102.5 ± 28.098.4 ± 26.1
Vitamin E (mg)10.9 ± 2.35.3 ± 1.36.5 ± 1.96.1 ± 0.8
α-tocopherol (mg)4.4 ± 1.21.9 ± 0.52.6 ± 0.72.8 ± 0.4
β-carotene (mg)2847.3 ± 1193743.2 ± 311.8806.9 ± 201.31267 ± 386.8
Zinc (mg)10.1 ± 1.46.9 ± 1.18.8 ± 1.311.4 ± 1.8
Copper (mg)1.1 ± 0.10.8 ± 0.10.7 ± 0.1**1.0 ± 0.1
Manganese (mg)2.9 ± 0.62.4 ± 0.51.5 ± 0.31.8 ± 0.4
Selenium (μg)85.7 ± 15.157.0 ± 8.566.9 ± 12.994.8 ± 17.8

Exercise-induced Changes in PEROX and TAS

Resting levels of PEROX were not different between normal-weight and overweight groups at baseline or at 8 weeks (Table 2, p > 0.05). Figures 1and 2 illustrate the changes in PEROX and TAS, respectively, during each CLPO exercise bout at baseline and Week 8. All overweight participants had significantly higher ΔPEROX values (increase in PEROX from pre- to post-CLPO exercise) at baseline than normal-weight groups (p < 0.05, Figure 1). Within the AOX-treated group, ΔPEROX was significantly reduced during exercise compared with the PL group (p < 0.05). Within the AOX group, there was no statistical difference in the ΔPEROX at 8 weeks for normal weight and overweight.

Figure 1.

The change in PEROX (Δ scores from pre- to post-exercise, adjusted for the exercise change in Vo2) during constant-load cycle exercise in normal-weight and overweight young adults. Values are means ± standard error of Δ scores at baseline and after 8 weeks of AOX or PL intervention. * Different from comparative normal-weight group. ^ Different from baseline value at p < 0.05. PEROX, lipid hydroperoxide; N-AOX, normal-weight, antioxidant-treated; N-PL, normal-weight, placebo; O-AOX, overweight, antioxidant-treated; O-PL, overweight, placebo.

Figure 2.

The change in TAS (Δ scores from pre- to post-exercise) during constant-load cycle exercise in normal-weight and overweight young adults. Values are means ± standard error of Δ scores at baseline and after 8 weeks of AOX or PL intervention. TAS, total antioxidant status; N-AOX, normal-weight, antioxidant-treated; N-PL, normal-weight, placebo; O-AOX, overweight, antioxidant-treated; O-PL, overweight, placebo.

Resting TAS levels were not different between normal and overweight groups at baseline but tended to increase with AOX supplementation by Week 8 (p = 0.08). The following were the TAS values for the N-AOX (1.38 ± 0.13, 1.43 ± 0.15 nM/mL), N-PL (1.24 ± 0.13, 1.27 ± 0.15 nM/mL), O-AOX (1.32 ± 0.11, 1.41 ± 0.13 nM/mL), or O-PL (1.30 ± 0.12, 1.19 ± 0.13 nM/mL). During exercise, the O-AOX group demonstrated an improvement in the ΔTAS during exercise at Week 8, although this improvement did not reach significance (p > 0.05, Figure 2).

Exercise-induced Changes in Inflammatory Cytokines and Lipids

After controlling for baseline values, adiponectin levels after the intervention period were not statistically different based on adiposity or AOX treatment status. However, clinically important changes in adiponectin occurred. With AOX, adiponectin increased by 2974 ± 1168 and 57 ± 527 ng/mL in the normal-weight and overweight groups, respectively. Within the PL group, adiponectin levels were reduced by 1203 ± 2696 and 570 ± 508 ng/mL in the normal-weight and overweight groups, respectively. These changes did not reach significance, however (p = 0.124).

Within the AOX group, resting IL-6 values were not significantly different at baseline or 8 weeks in the normal-weight (1.0 ± 0.4, 1.6 ± 0.3 pg/mL) or overweight (2.7 ± 0.4, 2.7 ± 0.3 pg/mL) groups. Within the PL group, resting IL-6 was also not different in the normal-weight and overweight groups (1.4 ± 0.3, 1.6 ± 0.3 and 2.2 ± 0.4, 2.2 ± 0.3 pg/mL, respectively). Within the AOX group, resting hs-CRPs were not different at baseline or 8 weeks in the normal-weight or overweight groups (0.06 ± 0.1, 0.9 ± 0.1 and 0.55 ± 0.1, 59 ± 0.1 mg/dL, respectively). Within the PL group, hs-CRP was not different at baseline or 8 weeks (0.17 ± 0.1, 0.11 ± 0.1 and 0.67 ± 0.1, 0.58 ± 0.1 mg/dL).

The cytokine values (Δ scores) presented in Table 4 represent the difference in cytokine levels from pre- to post-CLPO exercise at baseline and Week 8. After controlling for baseline values, the AOX-treated group had significantly lower ΔIL-6 scores at Week 8 compared with the PL normal-weight and overweight groups (p < 0.05). Within the AOX, the Δhs-CRP scores were lowered at Week 8 in both normal-weight and overweight groups, although this reduction did not achieve statistical significance (p = 0.110).

Table 4.  Cytokine concentration changes during the constant load exercise in normal-weight and overweight groups by AOX treatment at baseline and week 8. Values are the Δ scores (difference in cytokine levels from pre- to post-exercise)
  • AOX, antioxidant; PL, placebo; hs-CRP, high sensitivity C-reactive protein; IL-6, interleukin 6.

  • *

    Significant reduction in cytokine level compared with normal-weight group at p < 0.05.

 Baseline0.008 ± 0.0040.008 ± 0.0030.070 ± 0.0370.040 ± 0.013
 8 weeks0.004 ± 0.0020.044 ± 0.0300.033 ± 0.0280.078 ± 0.040
 Baseline0.336 ± 0.1280.218 ± 0.1830.080 ± 0.1460.105 ± 0.145
 8 weeks−0.035 ± 0.147*0.032 ± 0.108−0.206 ± 0.173*1.206 ± 1.073

The changes in cholesterol and lipid fractions are presented in Table 5. After controlling for baseline values, the exercise-induced changes of TC and triglycerides at Week 8 in the AOX-treated group were lower than those of the PL-treated groups (p < 0.05). There were no significant exercise-induced changes in HDL or LDL at 8 weeks between the groups.

Table 5.  Cholesterol and triglyceride concentration changes during the constant load exercise in normal-weight and overweight groups by AOX treatment at baseline and week 8. All values are expressed in mM/L. Values are the Δ scores (difference in lipid levels from pre- to post-exercise)
Time pointBaseline8 weeksBaseline8 weeksBaseline8 weeksBaseline8 weeks
  • AOX, antioxidant; PL, placebo; TC, total cholesterol; HDL-C, high-density lipoprotein; LDL-C, low-density lipoprotein.

  • *

    Different from PL group at p < 0.05.

TC0.22 ± 0.040.08 ± 0.04*0.10 ± 0.040.17 ± 0.040.21 ± 0.040.13 ± 0.04*0.16 ± 0.040.15 ± 0.04
Triglycerides0.12 ± 0.050.03 ± 0.05*0.08 ± 0.050.17 ± 0.060.13 ± 0.050.11 ± 0.06*0.11 ± 0.060.19 ± 0.06
HDL-C0.05 ± 0.010.03 ± 0.010.04 ± 0.010.05 ± 0.010.04 ± 0.010.03 ± 0.010.05 ± 0.010.03 ± 0.01
LDL-C0.15 ± 0.040.05 ± 0.030.05 ± 0.030.09 ± 0.030.15 ± 0.040.08 ± 0.030.09 ± 0.040.09 ± 0.03

Hierarchal Regression Analysis

The results of the hierarchical regression for the change in the dependent variable, the lipid peroxidation response (ΔPEROX/ΔVo2 from baseline to 8 weeks), are found in Table 6. The model included BMI, exercise-induced changes in cholesterol, LDL-C, TAS, and adiponectin. After controlling for BMI, the remaining four variables explained 59.1% of the variance in the ΔPEROX/ΔVo2. In exploratory analyses with other study variables not shown in Table 6, the overall regression model predicting the ΔPEROX/ΔVo2 was not significantly improved. These data suggest that the adjusted exercise-induced change in lipid peroxidation is likely due to the combined influence of several factors including lipid substrate, shifts in cytokine profile, and AOX availability.

Table 6.  Hierarchal regression analysis for the change in the lipid peroxidation response from baseline to 8 weeks in non-obese and overweight subjects. Each step includes the listed variable and the addition of the previous factor(s) and represents a separate regression equation.
StepVariablerR2p value
  • LDL-C, low-density lipoprotein; TAS, total antioxidant status.

  • *

    Denotes a significant contributor to the model.

  • Denotes a trend at p = 0.07.

2ΔCholesterol during exercise0.6800.4620.005*
3ΔLDL-C during exercise0.6920.4790.013
5ΔTAS during exercise0.7750.6000.015


Main Findings

This study examined whether short-term AOX supplementation reduced exercise-induced oxidative stress in overweight young adults. This study also tested whether AOX supplementation attenuated potential contributors to exercise-induced oxidative stress: inflammatory cytokines and blood lipids. To our knowledge, these data are the first to show that short-term AOX intervention (combined vitamins E and C and β-carotene) lowers the ΔPEROX during exercise in both normal-weight and overweight healthy young adults. In addition, hs-CRP and IL-6 concentrations were reduced during exercise after the treatment. Finally, TC and triglyceride elevations were attenuated in AOX-treated groups compared with controls.

AOX Supplementation and Oxidative Stress

Resting PEROX values were not found to be different between AOX- and PL-treated groups. Comparative data regarding the effects of this AOX combination in obesity or on oxidative stress are scarce. However, our finding is in contrast to existing previous data in athletes or older adults. Schroder et al. (24) showed that vitamins E and C and β-carotene reduced resting lipid peroxidation by 23% after 35 days of treatment in aerobic athletes. Nagyova et al. (31) found that these same AOXs and selenium together reduced malondialdehyde and conjugated dienes in older healthy men and those who were post-myocardial infarction, with the greater effect occurring in the post-myocardial infarction group. In another study, 6 weeks of α-tocopherol (592 Eq), vitamin C (1000 mg), and β-carotene (30 mg) reduced resting malondialdehyde and pentane production in supplemented healthy men but did not prevent the exercise-induced increase in lipid peroxidation (32). Discrepancies among these studies may be due to the differences among population characteristics (healthy, athletes, post-myocardial infarction, overweight), measurements of oxidative stress (type of lipid peroxidation biomarker, AOX capacity), time-points for measurement (rest, post-exercise), and AOX dosages. For example, the AOX dose used in an earlier study (24) was lower in vitamin E (400 IU) but substantially higher in vitamin C (1000 mg) and β-carotene (24 mg), which may, in part, explain the changes in PEROX at rest compared with the lack of differences at rest with our study.

It is possible that early adaptations to chronic oxidative stress in these young overweight adults may have counteracted free radical damage and elevation of lipid peroxidation at rest. Up-regulation of specific components of tissue AOX enzymes such as superoxide dismutase, catalase, glutathione peroxidase, and glutathione reductase might have occurred in these young adults as has been shown in animal models of obesity development (33).

To our knowledge, these are the first data to demonstrate that a combination of vitamins E and C and β-carotene have the collective potential to reduce post-exercise oxidative stress in overweight adults. During the CLPO exercise, a change in PEROX after the CLPO was detected in the overweight groups who were vulnerable to oxidative stress. The CLPO bout might have been sufficient to create an AOX–pro-oxidant imbalance that favored elevation of PEROX in our overweight groups (34). The ability of the AOX treatment to lower ΔPEROX with exercise may be due either to direct scavenging of free radicals (2) or to up-regulation of AOX enzymes by the AOX supplements (22, 35). The present data suggest that the utility of AOX supplementation in young overweight adults likely occurs in situations where free radicals are generated at higher rates or the AOX defense is excessively taxed, such as vigorous exercise.

AOXs and Inflammation, TAS, and Lipids

The present data indicate that AOX may favorably shift the inflammatory cytokine profile of IL-6, hs-CRP, and adiponectin. First, contracting skeletal muscle can directly produce IL-6 and release this cytokine into the bloodstream (13). Fischer et al. (25) demonstrated that supplementation with vitamins C and E suppresses IL-6 gene expression and cytokine release of IL-6 into the bloodstream during acute exercise; lipid peroxidation levels (as measured by F2 isoprostanes) were also reduced during the same exercise bout. Second, several studies support that co-administration of AOXs reduces CRP production and release in vivo (25, 36, 37). CRP production is directly regulated by both IL-1β and IL-6; vitamin E reduces 5-lipoxygenase activity, which inhibits IL-1β (38). Down-regulation of both IL-1β and IL-6 by AOX likely reduced CRP in this study. Third, adiponectin expression may be affected by oxidative stress. Adiponectin exerts multiple anti-atherogenic activities at the endothelium and is associated with lower risk for cardiovascular disease; increasing adiponectin is protective against disease (39). Preliminary evidence showed that cultured murine adipocytes show decreased expression of adiponectin levels when exposed to free radicals in vitro (40). In our study, adiponectin levels tended to increase with AOX, whereas these levels decreased with PL by 8 weeks. AOX supplementation may have squelched free radicals and lowered oxidative stress in systemic tissues, thereby increasing adiponectin expression and release into the bloodstream. Therefore, the observed reduction in lipid peroxidation with AOX might have been, in part, through an overall restoration of a more favorable cytokine profile with AOX treatment.

In this study, small but non-significant improvements in TAS were detected at rest and after exercise in the two AOX-treated groups compared with controls. Other studies have shown that increased dietary intake of AOXs (vitamins E and C, β-carotene) over the short term increases plasma indices of overall total AOX capacity (31, 41), and/or blood levels of each AOX (42). The fact that we did not find a statistically significant increase in TAS during exercise in the AOX-treated groups suggests that although some components of TAS may have been improved with AOX supplementation, other components may have been down-regulated. For example, Brennan et al. (43) demonstrated that 42 days of combined vitamin C and E supplementation significantly reduced erythrocyte superoxide dismutase and glutathione peroxidase in healthy adults by 35.7% and 22.3%, respectively. Alternatively, AOX mobilization in response to exercise is dynamic, and different TAS components like vitamin E might increase (44), whereas other components, such as glutathione, decrease (45), thereby minimizing the potential improvement in an overall total AOX capacity measurement. Further research is required to determine what specific enzymatic, dietary, and non-enzymatic AOXs are acutely affected in overweight persons during exercise with and without supplementation.

Although it is well established that acute exercise increases HDL-C and decreases triglycerides and TC (46), it is unclear whether AOX supplementation influences this response. This study showed that AOX reduced TC and triglycerides during CLPO exercise, with no significant changes in the PL group. Atherosclerosis-prone murine and porcine animal models show that treatment with vitamins C and E is associated with smaller reductions in triglycerides and TC than those treated with PL (47, 48). Other studies showed no changes in these same blood lipids with AOX supplementation in overweight adults (49). It is unclear whether AOX influences enzymes that mobilize lipids during exercise. Additional data are needed to address these points.


Although these data show promise that AOX are protective against exercise oxidative stress in overweight young adults, this study was performed in a small sample and should be replicated with a larger sample size. Patterns of change in this study regarding adiponectin and exercise changes in hs-CRP may have been statistically significant with a larger study sample. Measures of specific plasma AOXs in relation to overall AOX capacity would have enhanced the understanding of the dynamics of AOX mobilization in overweight individuals. Finally, an additional group of very obese persons would have provided additional data regarding the utility of AOX supplementation in persons in the extreme end of the BMI spectrum to oxidative stress during exercise.


Eight weeks of AOX reduces the exercise-induced ΔPEROX in overweight young adults. Possible collective mechanisms to explain this finding include a shift in the cytokine profile from a pro-inflammatory to a less inflammatory profile (lowered IL-6, increased adiponectin), an attenuation of cholesterol and triglyceride levels during exercise, and a small increase in TAS.


The project described was supported, in part, by the National Center for Complementary and Alternative Medicine (Grants T32-AT00052 and K30-AT-00060) and by the UVA General Clinical Center (Grant 5 M01 RR000847). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Center for Complementary and Alternative Medicine or the NIH.


  • 1

    Nonstandard abbreviations: AOX, antioxidant; IL, interleukin; CRP, C-reactive protein; TAS, total AOX status; UVA, University of Virginia; PL, or placebo; N-AOX, normal weight, AOX treated; N-PL, normal weight, PL; O-AOX, overweight, AOX treated; O-PL, overweight, PL; GCRC, General Clinical Research Center; Vo2peak, peak oxygen consumption; LT, lactate threshold; PEROX, lipid hydroperoxide; PO, power output; HR, heart rate; CLPO, constant-load PO; HbA1c, hemoglobin A1c; TC, total cholesterol; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-density lipoprotein-cholesterol; ABTS, 2,2-azion-di-[3-ethylbenzenthiazoline sulfonate]; ELISA, enzyme-linked immunosorbent assay; hs, high sensitivity.

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.