Article Text

Download PDFPDF

Effects of exercise on markers of oxidative stress: an Ancillary analysis of the Alberta Physical Activity and Breast Cancer Prevention Trial
  1. Christine M Friedenreich1,2,
  2. Vincent Pialoux3,
  3. Qinggang Wang1,
  4. Eileen Shaw1,
  5. Darren R Brenner1,2,
  6. Xavier Waltz4,
  7. Shannon M Conroy5,
  8. Rhys Johnson1,
  9. Christy G Woolcott6,
  10. Marc J Poulin4,
  11. Kerry S Courneya7
  1. 1Department of Cancer Epidemiology and Prevention Research, CancerControl Alberta, Alberta Health Services, Calgary, Alberta, Canada
  2. 2Departments of Oncology and Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada
  3. 3Inter-University Laboratory of Human Movement Biology, Université Claude Bernard Lyon 1, University of Lyon, Villeurbanne, France
  4. 4Departments of Physiology and Pharmacology and Clinical Neurosciences, Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada
  5. 5Cancer Prevention Institute of California, Fremont, California, USA
  6. 6Departments of Pediatrics and Obstetrics and Gynecology, Faculty of Medicine, Dalhousie University, Halifax, Nova Scotia, Canada
  7. 7Faculty of Physical Education and Recreation, University of Alberta, Edmonton, Alberta, Canada
  1. Correspondence to Dr Christine Friedenreich; christine.friedenreich{at}


Background Oxidative stress may contribute to cancer aetiology through several mechanisms involving damage to DNA, proteins and lipids leading to genetic mutations and genomic instability. The objective of this study was to determine the effects of aerobic exercise on markers of oxidative damage and antioxidant enzymes in postmenopausal women.

Methods The Alberta Physical Activity and Breast Cancer Prevention Trial (ALPHA) was a two-centre, two-armed randomised trial of 320 inactive, healthy, postmenopausal women aged 50–74 years. Participants were randomly assigned to a year-long exercise intervention (225 min/week) or a control group while being asked to maintain a normal diet. Fasting blood samples were obtained and plasma concentrations of two oxidative damage markers (8-hydroxy-2′-deoxyguanosine (8-OHdG) and 8-isoprostaglandin F2α (8-Iso-PGF2α)) and two antioxidant enzymes (superoxide dismutase and catalase) were measured at baseline, 6 months and 12 months. Intention-to-treat (ITT) and per-protocol analyses were performed using linear mixed models adjusted for baseline biomarker concentrations. A further exercise adherence analysis, based on mean minutes of exercise per week, was also performed.

Results In the ITT and per-protocol analyses, the exercise intervention did not have any statistically significant effect on either oxidative damage biomarkers or antioxidant enzyme activity.

Conclusions A year-long aerobic exercise intervention did not have a significant impact on oxidative stress in healthy, postmenopausal women.

Trial registration number NCT00522262.

  • Cancer
  • Randomised controlled trial
  • Antioxidants
  • Exercise physiology

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Summary box

  • This is the first large-scale randomised controlled trial of 320 postmenopausal women to examine the effect of a year-long aerobic exercise intervention on markers of oxidative damage and antioxidant enzyme activity.

  • Overall, we did not observe any effect of physical activity on 8-hydroxy-2′-deoxyguanosine (8-OHdG), 8-isoprostaglandin F2α (8-Iso-PGF2α), superoxide dismutase or catalase in this study.

  • Oxidative stress may not be a predominant mechanism by which physical activity decreases breast cancer risk.


Breast cancer is the most common cancer among women worldwide.1 ,2 Since 2004, breast cancer incidence rates have stabilised in North America.3 However, the absolute number of cases being diagnosed is still increasing because of population growth and ageing.1 Consistent epidemiological evidence exists that physical activity reduces postmenopausal breast cancer risk by 20–25%.4 Several inter-related biological mechanisms are hypothesised to explain this association.4 Until now, five randomised controlled exercise trials have investigated the direct effect of exercise on hypothesised biomarkers of postmenopausal breast cancer.5–9 These studies have found that exercise decreases endogenous oestrogens, adiposity, leptin and markers of insulin resistance and inflammation.10–12

Systemic oxidative stress has been implicated in many diseases and disorders, including the pathogenesis and progression of breast cancer.13–15 Oxidative stress refers to the excessive production of reactive oxygen species (ROS), particularly from oxygen radicals, during a period of increased exposure to environmental stress. ROS are a normal by-product of metabolism and are necessary components of both cell signalling and homoeostasis;16 however, their presence in excess amounts can have a negative physiological effect. ROS can induce damage to lipids, proteins and DNA, in turn leading to genetic mutations and genomic instability, thus contributing to carcinogenesis.17 In addition, ROS have been shown to be involved in the signalling pathways of neoangiogenesis, a mechanism that aids tumour growth and metastasis development.18

The balance of oxidative stress factors is mainly determined by endogenous enzymatic mechanisms, although exogenous lifestyle factors such as dietary intake, physical activity and medication use also play a major role.19 ,20 A single bout of intense exercise causes a transient increase in ROS.16 However, as part of a favourable biological adaptive response referred to as the ‘exercise-induced oxidative stress paradox’,21 exercise training enhances antioxidant and oxidative damage repair enzyme capacity,22 ,23 and subsequently reduces overall oxidative damage.24 ,25 Enzymatic antioxidants may decrease breast cancer development by neutralising the over-generated ROS.26 Hence, antioxidant enzymes such as superoxide dismutase (SOD) and catalase (CAT) are established markers for assessing the beneficial effect of physical training on antioxidant status.25 ,27 It is also hypothesised that oxidative damage markers can be affected by physical exercise training through the enhanced activity of DNA damage repair mechanisms.28 These include: 8-hydroxy-2′-deoxyguanosine (8-OHdG), an accurate marker of DNA oxidation that is associated with breast cancer risk,29 ,30 and 8-isoprostaglandin F2α (8-Iso-PGF2α), a lipid peroxidation product that is a reliable marker of in vivo oxidative stress and is also related to breast cancer risk.31

Until now, very few studies conducted in humans have examined the effect of long-term exercise on markers of oxidative stress. Of those studies, most did not focus on cancer-free, previously inactive, postmenopausal women who are of greatest relevance to postmenopausal breast cancer risk. One randomised controlled trial did examine the effects of a year-long exercise intervention on oxidative stress in postmenopausal women, but investigated only one biomarker via urine, F2-isoprostane.24 To the best of our knowledge, no other randomised controlled trials conducted in humans have studied the effects of long-term exercise on accurate and reliable markers of DNA oxidation, lipid peroxidation and antioxidant enzyme levels. The Alberta Physical Activity and Breast Cancer Prevention (ALPHA) Trial was a randomised controlled trial that examined how a 1 year aerobic exercise intervention influenced hypothesised biomarkers that can modulate risk of postmenopausal breast cancer compared to a sedentary lifestyle.6 This analysis uses data from the ALPHA Trial to describe the effects of a year-long, moderate-to-vigorous intensity, aerobic (225 min/week) exercise intervention on the oxidative stress markers SOD, CAT, 8-iso-PGF2α and 8-OHdG, which were secondary outcomes from the ALPHA Trial.


Study population

Methods for the ALPHA Trial have been previously reported.6 Briefly, women were recruited from May 2003 to June 2005 through targeted mailings to participants in the Alberta Breast Screening Programme, brochures distributed to family physicians and not currently undertaking or planning to undertake a weight loss programme, through media campaigns in Calgary and Edmonton, Alberta. Specific eligibility criteria included: inactive women (<90 min/week recreational activity over the past year or, if between 90 and 120 min/week of physical activity had a VO2max <34.5 mL/kg/min); aged 50–74 years; postmenopausal for at least 24 months; no previous cancer diagnosis besides non-melanoma skin cancer; no major comorbidities; acceptable heart and lung function as assessed by baseline fitness test; physician clearance for unrestricted physical activity; normal fasting lipids, glucose thyroid stimulating hormone and alanine aminotransferase; body mass index (BMI) 22–40 kg/m2; non-smoker; <14 drinks of alcohol/week; without diabetes; breast tissue density greater than a zero density level; no medications or exogenous hormones that might influence oestrogen metabolism; not currently or planning to undertake a weight loss programme and no planned, extended absences in the 18 months subsequent to enrolment.


Women were randomised to either a year-long, moderate-to-vigorous intensity aerobic exercise intervention of 225 min/week (n=160) or to a control group that was asked to maintain their current activity levels (n=160). The final exercise prescription was 45 min long sessions, five times per week at 70–80% heart rate reserve for 1 year. Of these sessions, at least three were facility-based with on-site exercise trainers while the remaining sessions were home based. The initial exercise prescription of three weekly sessions of 15–20 min duration at an intensity of 50–60% heart rate reserve was gradually increased over a period of 3 months to the final prescription. To monitor heart rate, participants wore Polar A3 heart rate monitors throughout the exercise period. Participants in both groups were also asked not to alter their diet for the duration of the study. Ethics approval was obtained from ethics review boards of the University of Calgary, the University of Alberta and the Alberta Cancer Board and each participant provided written informed consent.

Blood collection and biomarker assays

Fasting blood was collected at baseline (60 mL), 6 and 12 months (40 mL) after a minimum 10 hours fast. Participants were instructed not to exercise or consume alcohol within 24 hours of the test. All blood samples were collected, processed and stored within 4 hours of collection at −86°C at the Holy Cross Centre in Calgary until the time of the assays. The laboratory assays described below were carried out at the University of Calgary with the oversight of VP (Lyon). All three samples for each participant were included in a single batch with equal numbers of samples from exercisers and controls and two pooled quality control samples. Samples were randomly allocated within each plate. Laboratory personnel were blinded to the samples. All samples were assayed in duplicate. For any sample, if the within-individual CV between repeats was above 20%, we used the value that was closest to the mean concentration for the plate. Values above four SDs from the mean of each plate were defined as outliers and removed. Values below the limit of detection (ie, below the optical density of the negative control) were removed from analyses for each assay. In order to adjust for potential plate effects, concentration values were adjusted for the overall within-plate mean concentrations.

Plasma 8-OHdG and 8-Iso-PGF2α were measured using ELISA kits (Cell Biolabs, San Diego, California, USA).23 ,32 Detection limits, intra-assay and interassay coefficients of variation (CV) were 0.1 ng/mL, 5% and 5%, respectively, for 8-OHdG and 0.1 ng/mL and 6% and 6%, respectively, for 8-iso-PGF2α. The quantity of 8-OHdG and 8-iso-PGF2α was determined by comparing its absorbance at 450 nm in a sample with those of a standard curve using an M2e SpectraMax microplate reader.

Plasma CAT and SOD activity were measured spectrophotometrically using standard quantification methods.33 Catalase activity was quantified by the method of Johansson and Borg using hydrogen peroxide as a substrate and formaldehyde as a standard.34 SOD activity was quantified using the Beauchamp and Fridovich method,35 slightly modified by Oberley and Spitz.36 The intra-assay and interassay CVs were 5% and 12%, respectively, for catalase and 6% and 7%, respectively, for SOD.

Statistical methods

All biomarker levels were subjected to log-transformation to obtain an approximately symmetrical distribution after observing data to be askew. The primary ITT analysis included all values that were available at each study interval. The per-protocol analysis included all control group participants and only those in the exercise group who completed ≥90% of the exercise prescription (≥204 min/week) for weeks 13–52 of the exercise intervention (excluding the 12-week ramp-up period). A further analysis was performed based on exercise adherence, using mean minutes of activity per week for all 52 weeks of the study. Exercise adherence was classified into the following categories based on previously defined public health guidelines: <150, 150–225 and >225 min/week.37

General linear mixed models were used for all biomarkers to observe the intervention effects using the measures at 6 and 12 months as repeated measurers. These models included the main effects of intervention and time, as well as their interaction term, and were adjusted for baseline values of each biomarker. Treatment effect ratios (TERs) were estimated as a geometric mean ratio of the exercisers compared to the controls from the linear mixed models. A TER > 1.0 indicates higher oxidative stress biomarker or antioxidant enzyme activity levels among exercisers compared to controls.

The secondary analysis examined if the effect of exercise on the biomarkers of interest was mediated by exercise adherence, changes in body composition (assessed by BMI, body fat percentage and intra-abdominal fat area), overall physical fitness throughout the study, (assessed by VO2max) and diet or vitamin intake. Effect modification (moderation) was evaluated based on statistical significance of the interaction term (Pheterogeneity) between the intervention group assignment and each proposed moderator at baseline in the ITT analysis models.38 Hypothesised moderators included baseline levels of: fitness (VO2max), age, BMI, body fat percentage, intra-abdominal fat area and oxidative stress biomarkers or antioxidant enzyme activity. All statistical analyses completed for this study were performed with the use of SAS software (V.9.2; SAS Institute, Cary, North Carolina, USA).


Of the 320 women initially randomised, 311 (97%) completed the trial. Baseline characteristics of the study population have previously been described.6 At baseline, there were limited differences between the exercise and control groups with respect to baseline concentrations of the biomarkers of interest (table 1).

Table 1

Baseline characteristics of study participants, the Alberta physical activity and breast cancer prevention trial, 2003–2007, n=320

In the ITT analysis of the data (table 2), we did not observe any statistically significant differences in levels for the oxidative stress biomarkers or antioxidant enzyme activities between the two groups. Similarly, in the per-protocol analysis (table 3), no biomarkers showed a statistically significant difference between groups following the intervention.

Table 2

Intention-to-treat analysis of circulating oxidative stress biomarker levels for exercisers and controls in the ALPHA Trial

Table 3

Per-protocol* analysis of circulating oxidative stress biomarker levels for exercisers and controls in the ALPHA Trial

In the exercise adherence analysis (table 4), no differences were observed in these biomarker levels between any levels of exercise adherence compared to the control group. Finally, we did not observe any statistically significant mediation or effect modification by fitness (VO2max), age, BMI, body fat percentage, intra-abdominal fat area and oxidative stress biomarkers or antioxidant enzyme activity.

Table 4

Adherence level* analyses of oxidative stress biomarker levels at baseline and 12 months in exercisers and controls in the ALPHA Trial


Our year-long moderate-to-vigorous intensity exercise intervention among previously inactive, postmenopausal, cancer-free women did not cause any statistically significant effects compared to the controls for biomarkers of oxidative stress or antioxidant enzyme activity.

Oxidative stress is one of the more recently hypothesised mechanisms whereby physical activity can influence the risk of breast cancer and other chronic diseases. While there have been a limited number of studies in this area of research, the findings of this trial differ from previous research. Exercise has generally been found to increase antioxidant enzyme capacity while decreasing lipid peroxidation and DNA damage markers in healthy populations of men and women.39–51 However, not all studies showed consistency across biomarkers,40 ,42 ,45 ,46 ,52 and some studies have shown no effect of exercise on oxidative stress or antioxidant capacity.53 ,54 In these findings, study designs and populations are heterogeneous. Very few randomised controlled studies have examined these markers of oxidative stress in inactive, postmenopausal women and none considered all of the above markers simultaneously. Of these studies, most have shown a beneficial impact of physical activity in reducing level of oxidative stress or increasing antioxidant enzyme activity,24 ,52 ,55–59 but not all are consistent.56 Only one trial was comparable to the ALPHA Trial in terms of size, duration and scope.10 While this study showed a non-statistically significant decrease in urinary F2-isoprostane, no antioxidant enzymes were assayed.

The role of physical activity, specifically exercise training, has been attributed to increased antioxidant capacity and decreased oxidative stress biomarkers in what has been described as the ‘oxidative stress paradox’.21 When exercise is performed, ROS concentrations increase; however, over time, this increase produces a compensation mechanism, by which adaptation occurs to higher levels of oxidative stress by upregulating the amount and efficiency of antioxidant enzymes. Increased ROS species have been attributed to a multitude of different cancers and other chronic diseases.13–15 Additionally, oxidative stress increases with increasing age,60 and there is evidence that postmenopausal women have higher levels of lipid peroxidation compared to premenopausal women.61 ,62 Thus, their reduction due to physical activity could serve as a plausible mechanism of cancer prevention, although this effect was not observed in our trial.

The study population of the ALPHA Trial consisted of healthy, postmenopausal women. Healthier individuals may have lower levels of oxidative stress; hence, an exercise intervention may not have had a significant impact in reducing ROS or increasing antioxidant enzyme activity. This hypothesis is supported by the most comparable study of a year-long exercise intervention in postmenopausal women that showed only a modest improvement in urinary F2-isoprostane with exercise training.24 Nevertheless, we did not find any statistically significant interaction with baseline values of the oxidative stress markers.

The medium in which antioxidant enzyme activities were measured may have played a role in the lack of effect seen in this study. Most studies have shown an increase in antioxidant activity in erythrocytes39 ,42 ,46 ,50 ,52 ,55 ,57 ,59 or whole blood.44 ,46 ,49 ,50 Studies using plasma measures of activity have shown inconsistent results,45 ,46 ,56 with one study demonstrating a differential effect depending on whether enzyme activities were measured in plasma or in erythrocytes.46 Furthermore, plasma levels of antioxidant activity may not correlate with tissue levels. Similarly, in terms of oxidative stress biomarkers, urinary measures of F2-isprostanes have been found to show daily variability.63 In addition, concentrations of F2-isprostanes in urine are 20–50 times lower than in plasma and dependent on creatinine concentration, both suggesting that urine may be more sensitive to smaller changes.64 It is possible that antioxidant enzyme capacities and DNA damage biomarkers in plasma may not be as responsive to aerobic exercise training as those in erythrocytes, whole blood or urine.

Since the ALPHA Trial was a year-long intervention of 225 min/week of aerobic exercise, this time frame may be too long to observe an effect of the exercise intervention. Most trials that have shown increases in antioxidant enzyme capacity with exercise were much shorter in duration (8–24 weeks)45 ,47 ,49 ,51 ,56 ,59 and we hypothesise that without an increase in aerobic exercise volume or intensity, the exercise effect on oxidative stress biomarkers and antioxidant enzyme activity could have an adaptive effect over time. It is possible that an increase in antioxidant activity as a result of the intervention would have occurred within the first few weeks of the aerobic exercise training. Since oxidative stress is a balance between ROS and the antioxidant system,65 we hypothesise that this early increase in antioxidant enzyme activity could be matched by a decrease in ROS generation in response to the exercise training intervention such that, over time, the initial increase in antioxidant enzyme activity would begin to diminish and return to basal levels because of the decreased stimulus. Indeed, studies of prolonged aerobic exercise training do not always show an improvement in total antioxidant capacity.66–68 The lack of effect on antioxidant enzyme capacity due to decreased ROS is most likely complemented by the decreased adiposity and C reactive protein (CRP) levels seen in the exercise group of the ALPHA Trial.11 ,69 Adipose tissue has been shown to be a potent source of ROS,70 ,71 while increased CRP has been correlated with increased ROS.72

Overall, this study showed no effect of a year-long exercise intervention on these four markers of oxidative stress in healthy, postmenopausal women. While we did not find any statistically significant effects of aerobic exercise training on oxidative stress biomarkers in this study, based on the current literature and biological models, oxidative stress remains a plausible mechanism by which physical activity could influence breast cancer risk and further research is warranted.


The study set-up of the ALPHA trial was performed by Kim van der Hoek and Marla Orenstein. The Study Coordinators were Rosemary Crosby, Ame-Lia Tamburrini and Sarah MacLaughlin. Fitness Centre Managers were Ben Wilson, Lisa Workman and Diane Cook. Exercise Trainers were Shannon Hutchins, Kathy Traptow, Shannon Brown, Susan Daniel, Parissa Gillani, Stephanie Sanden, Karen Mackay and Sandra Olsen. Data preparation was conducted by Sandra Blitz and Sony Brar. Blood sample storage, maintenance and retrievals were carried out by the Alberta Cancer Research Biobank staff. Dr Scott Forbes conducted part of the laboratory assays for this study that were performed in the laboratory of Dr Katherine Wynne-Edwards at the University of Calgary.


View Abstract


  • Contributors CMF, KSC, SMC, CGW, VP and MJP took part in study concept, design and funding. CMF, KSC and CGW participated in data collection. VP, XW and MJP took part in laboratory analyses. QW, ES, DRB, CMF and KSC carried out data analysis. CMF, RJ, ES, DRB and KSC reviewed the manuscript. All authors reviewed and approved the final manuscript.

  • Funding The Alberta Physical Activity and Breast Cancer Prevention (ALPHA) Trial was funded by research grants from the Canadian Breast Cancer Research Alliance (number 13576 and number 017468), the Alberta Cancer Foundation (number 22170) and the ALPHA Ancillary Study was funded by the Canadian Institutes for Health Research (CIHR) (number MOP-130238). CMF holds a Health Senior Scholar Award from Alberta Innovates-Health Solutions (AI-HS) and is the Alberta Cancer Foundation Weekend to End Women's Cancers Breast Cancer Chair. VP was supported by the Institut Universitaire de France. SMC was supported by postdoctoral fellowships (AI-HS) and the CIHR. DRB was supported by a postdoctoral fellowship from AI-HS and a career award from the Canadian Cancer Society Research Institute. XW was supported by postdoctoral fellowships from the Hotchkiss Brain Institute, AI-HS and the CIHR. MJP holds the Brenda Strafford Foundation Chair in Alzheimer Research. KSC was supported by the Canada Research Chairs Programme.

  • Competing interests None declared.

  • Ethics approval Alberta Cancer Research Ethics Board (Alberta Cancer Board), the Conjoint Health Research Ethics Board (University of Calgary) and the Health Research Ethics Board (University of Alberta).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data are available.