Physical activity has emerged as one of the most potent therapeutic lifestyle agents in the promotion of improved glucose regulation and the prevention of type 2 diabetes, with observational, mechanistic, and clinical interventions all supporting its efficacy (1,9,10,23). Furthermore, observational research has consistently demonstrated that higher levels of physical activity are associated with a lower risk of cardiovascular disease and mortality (1,13,20). Physical activity promotion is, therefore, a cornerstone of lifestyle programs aimed at the prevention or treatment of type 2 diabetes and other chronic diseases (15). However, over the last two decades, there has been increasing evidence that not all participants respond to moderate or vigorous intensity physical activity with increased cardiorespiratory fitness or improved metabolic health (5,12,14,17,18). More worryingly, a recent analysis of pooled exercise training trials suggested that exercise may actually result in an adverse metabolic response in a minority of participants, with between 5% and 15% experiencing substantial worsening in a given marker of metabolic health (4). This effect is thought to be quantitatively different to the rare acute risks of physical exertion such as cardiac death (4). This suggests that rather than existing on a spectrum from null to substantial metabolic benefit, increased physical activity may actually cause chronic metabolic harm to a significant minority of individuals. If true, this has serious public health implications by suggesting that universal physical activity promotion may be inappropriate and that developing methods of screening for individuals who are at high risk of an adverse metabolic response after increased physical activity is a research priority (6).
However, the area of adverse metabolic responses after increased physical activity is controversial and lacks biological plausibility, and evidence to date has been flawed by a failure to consider or compare against responses that occur naturally under control conditions. Markers of metabolic health are affected by multiple factors beyond physical activity; therefore, it is crucial that studies investigating the risk of an adverse metabolic response following a physical activity intervention quantify this in relation to a control group. Failure to do this could result in erroneous conclusions being disseminated, leading to a negative effect on public health programs and perceptions around the health benefits of engaging in a healthy lifestyle. In addition, it is important to investigate this issue in clinically relevant populations undergoing an intervention that is reflective of clinical practice. Using secondary analysis of the PREPARE randomized controlled trial, we investigate the occurrence of adverse responses in an overweight and obese population with impaired glucose tolerance (IGT) after usual care or two physical activity promotion programs, one of which was based on pedometer use and successfully increased levels of physical activity by around 2000 steps per day over 12 months compared with control conditions (23). Adverse responses are investigated for glucose regulation, the primary focus of the intervention, along with blood lipids through HDL-cholesterol and triglycerides.
This study uses secondary analysis of the PREPARE trial, which was designed to evaluate the effectiveness of a brief theory-driven group-based physical activity intervention at improving glucose tolerance in overweight and obese individuals with IGT. The primary outcome was 2-h postchallenge glucose (2-h glucose). The full methods, participant flow, and 3, 6, and 12 month results have been described in detail elsewhere (22,23). In brief, participants were recruited from a population-based type 2 diabetes screening program between September 2006 and March 2007 at Leicestershire, United Kingdom. Participants were included if they met WHO 2006 criteria for IGT (21) and were overweight or obese. Participants were randomized to one of three groups. The control group received usual care and an advice leaflet. Intervention 1 received a standard 3-h group-based structured education program, delivered by two trained educators to groups of up to 10 individuals. The program was aimed at promoting increased physical activity, particularly walking activity, to levels that were consistent with minimum recommendations for health. The program was designed to address key perceptions and knowledge of IGT and to target the perceived effectiveness of increased physical activity as a treatment for IGT, walking self-efficacy beliefs, barriers to walking, and self-regulatory strategies such as action planning and goal setting. Intervention 2 received the same 3-h structured education program, but with an enhanced self-regulation section that incorporated personalized steps per day targets and pedometer use to aid goal setting and self-monitoring. Both intervention conditions also received brief one-to-one counseling at 3 and 6 months. The study was completed in April 2008.
All applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during this research, including informed consent. This study was approved by the Leicestershire, Northamptonshire, and Rutland NHS Research Ethics Committee in June 2006.
Assessment of biochemical and demographic data.
Individuals underwent an oral glucose tolerance test after standardized procedures at 3, 6, and 12 months. Plasma glucose was measured using a glucose oxidase method on the Beckman AutoAnalyzer (Beckman, High Wycombe, UK). Serum cholesterol was analyzed using a cholesterol enzymatic assay (Abbott Clinical Chemistry, Chicago, IL). HDL-cholesterol was analyzed using the ultra-HDL assay (Abbott Clinical Chemistry). Serum triglyceride was analyzed using the triglyceride glycerol phosphate oxidase assay (Abbott Clinical Chemistry). Biochemical measurements were performed in the same laboratory located within Leicester Royal Infirmary using stable methodology standardized to external quality assurance reference values. Body weight (Tanita TBE 611; Tanita, West Drayton, UK) and height were measured to the nearest 0.1 kg and 0.5 cm, respectively, and BMI was calculated. Ethnicity was determined by self-report. Social deprivation was determined by assigning an Index of Multiple Deprivation (IMD) score to each participant’s postcode (19). IMD is publicly available continuous measures of compound social and material deprivation and is calculated using a variety of data including current income, employment, health, education, and housing.
Sealed piezoelectric pedometers with a 7-d memory (NL-800, New-lifestyles, USA) were used to measure ambulatory activity (23); a minimum of 3 d were required for valid pedometer data. Measurement pedometers were different from the motivational instruments used in the primary intervention condition and have been shown to be more accurate than traditional spring-levered pedometers in overweight and obese individuals (7,8).
Definition of adverse response.
Consistent with a recent study (4), an adverse response for included variables was defined as two times the technical error. The technical error represents within-subject SD derived from repeated measures (or assays) over a short period. On the basis of previously reported data (4,11,16), we estimated that an adverse response was an increase of ≥0.8 mmol·L−1 for fasting glucose, ≥1.3 mmol·L−1 for 2-h glucose, ≥0.42 mmol·L−1 for triglycerides, and −0.12 mmol·L−1 or less for HDL-cholesterol.
This secondary analysis was conducted on the full study cohort; each group included 29 participants with data at each time point (23). Differences between groups in change in physical activity at 3, 6, and 12 months were conducted using ANCOVA procedures with baseline data entered as a covariate (23).
Differences in the rates of adverse responses between groups were assessed using logistic regression models and reported as odds ratios. To assess the effect of change in physical activity on the chance of an adverse event occurring, the study cohort was combined and stratified by whether participants recorded increased or decreased physical activity (compared with baseline values) at each follow-up time point (3, 6, and 12 months). As randomization was broken for this analysis, adjustment was undertaken for basic demographic data; age, sex, ethnic origin, and social deprivation. Differences in baseline characteristics between those observed to have an adverse response and those without an adverse response were tested using independent t-tests, Mann–Whitney tests for nonparametric variables, and chi-squared tests for categorical data. Differences in change in physical activity between those observed to have an adverse response and those without an adverse response were tested using ANOVA. Finally, the relation between change in physical activity and the number of adverse events reported for each participant was tested using Spearman correlation. All analyses were two tailed and carried out on PASW Statistics version 18 for Windows (SPSS, Chicago, IL) in 2013.
Analyses were repeated with adverse response defined as three times the technical error, changing the definition to the following: ≥1.2 mmol·L−1 for fasting glucose, ≥2.0 mmol·L−1 for 2-h glucose, ≥0.63 mmol·L−1 for triglycerides, and −0.18 mmol·L−1 or less for HDL-cholesterol.
Cohort characteristics and results for change in physical activity and biochemical variables at 3, 6, and 12 months have been reported previously and are not replicated here (23). Briefly, intervention 2 resulted in increased ambulatory activity across all time points; values for change were 2395 (95% CI, 1285–3505), 2093 (95% CI, 944–3242), and 1039 (95% CI, 135–1943) steps per day at 3, 6, and 12 months, respectively. The corresponding results for intervention 1 were as follows: 1362 (95% CI, 1502–2221), 870 (95% CI, −54 to 1793), and 549 (95% CI, −290 to 1390) steps per day. Results for the control group were as follows: 552 (95% CI, −290 to 1393), −152 (95% CI, −778 to 573), and −940 (95% CI, −1574 to −307) steps per day. Results for intervention 2 were significantly different to the control group across all time points (P = 0.005, P = 0.001, and P ≤ 0.001, respectively). In contrast, results for intervention 1 were only different to the control group at 12 months (P = 0.006). Significant decreases in both 2-h glucose and fasting glucose were seen in intervention 2 compared with the control group across all time points. No differences in glucose parameters were observed across any time point in intervention 1. No other biochemical differences were observed in either group.
The number of adverse responses reported for fasting glucose, 2-h glucose, triglycerides, and HDL-cholesterol at each measurement time point is presented in Table 1. Table 2 shows the summary data for the number of participants who had an adverse response at least once during the trial in the assessed variables. In total, 22 (76%) individuals in the control group experienced an adverse response in at least one variable during the trial compared with 23 (79%) in intervention 1 and 12 (41%) in intervention 2. For intervention 2 compared with the control group, the odds of an adverse response occurring at least once during the trial was significantly reduced (0.22; 95% CI, 0.07–0.69).
Table 3 shows the numbers of adverse responses in measured variables stratified by those that increased and decreased their levels of objectively assessed ambulatory activity physical activity at each time point. Over the course of the trial, those that increased their ambulatory activity from baseline at all time points (3, 6, and 12 months) had a significantly reduced odds of having an adverse response in any variable compared with those that did not (0.30; 95% CI, 0.10–0.93), after adjustment for age, sex, social deprivation, and ethnicity.
Baseline characteristics stratified by whether or not an adverse event was detected in any variable are displayed in Table 4; those with an adverse event were more likely to be from a minority ethnic group and were from more socially deprived areas; 2-h glucose levels were also lower in this group.
Table 5 shows change in objectively assessed physical activity from baseline in those who did and did not have an adverse response at 3, 6, and 12 months. There was no significant difference in change in ambulatory activity between these groups at any follow-up time point.
The majority of individuals with an adverse response during the study had multiple events resulting from adverse responses in the same variable at different follow-up time points and/or adverse responses in more than one variable; Figure 1 shows the distribution in the total number of adverse events observed in each participants during the trial, stratified by study group. The total number of adverse events observed in each participant was not significantly correlated with change in physical activity at any time point (ρ = −0.14, P = 0.21) or at 3 (ρ = 0.06, P = 0.59), 6 (ρ = −0.12, P = 0.28), and 12 (ρ = −0.05, P = 0.64) months.
The pattern of results across intervention groups was largely unaffected when an adverse response was defined as three rather than two times the technical error (see Table, Supplemental Digital Content 1 (Supplemental Digital Content 1, http://links.lww.com/MSS/A352)—number of adverse responses occurring in fasting glucose, 2-h glucose, HDL-cholesterol and triglycerides following a physical activity intervention or control condition at 3, 6 and 12 months; adverse responses defined as three times the technical error).
This secondary analysis of a three-armed intervention trial found that although 41% of participants randomized to a successful pedometer-based physical activity intervention had an adverse response in fasting glucose, 2-h glucose, triglycerides, or HDL-cholesterol during the study, rates were significantly higher in the control group where 76% of participants had an adverse response. Similarly, 79% of participants experienced an adverse response in the other intervention group, which did not demonstrate sustained increases in physical activity. The odds of the successful pedometer-based intervention group experiencing an adverse response were reduced by 78% in comparison with the control group. Results were similar in the combined cohort stratified by change in activity levels; those who increased their physical activity at each time point had a 70% reduced odds of an adverse response compared with those who did not.
The proportion of participants with an adverse event in the pedometer-based intervention was broadly consistent with pooled data from exercise training studies that reported an 8%, 13%, and 10% adverse response rate in fasting insulin, HDL-cholesterol, and triglycerides, respectively, at a single time point following a supervised exercise training program with 31% experiencing an adverse response in at least one of four reported metabolic variables (4). The authors of the study concluded that adverse responses to exercise occur and that developing methods of identifying those who are likely to develop unwarranted responses is a matter of priority (4). However, the study only included data for those undertaking exercise training and failed to report or consider whether adverse responses were also seen under control conditions.
Metabolic markers of health status are influenced by a wide variety of factors beyond physical activity, including assay variability, dietary intake, medication status, illness, and life events that cause stress or alterations in habitual behavior. Therefore, it is not possible, and academically disingenuous, to attribute casualty to adverse responses observed in a minority of free living individuals undertaking an exercise training program, unless carefully compared with control conditions. By demonstrating more adverse responses under control conditions, the present study highlights the importance of this concept and suggests that physical activity actually reduces the risk of an adverse metabolic response. This consistent pattern of results would be unexpected with a generally effective intervention, which nonetheless actively caused adverse metabolic responses in a significant minority of participants.
The area of responders and nonresponders in physical activity research, and in particular, the hypothesis that some individuals may actually suffer from metabolically deleterious effects, is a complex and controversial area (3). Although it is biologically plausible and widely reported that response to physical activity is heterogeneous and exists on a broad continuum ranging from negligible to substantial benefit for any given factor, it is less plausible that physical activity may cause chronic deleterious adaptations to metabolic health. High levels of physical activity were essential for survival for the vast majority of human evolution, and physical activity has consistently demonstrated to have pleiotropic benefits across a wide spectrum of chronic diseases and conditions (2). Against this backdrop, only controlled trials, or other appropriate experimental methodologies, should be used to confirm or refute the hypothesis that adverse metabolic responses are seen with increased physical activity. To date, this level of evidence has not been presented in the literature.
This study has strengths and limitations. The strengths include the randomized nature of the trial, the rigorous and standardized trial follow-up visits, and the high-risk nature of the cohort, which was broadly representative of individuals who are likely to be referred in a lifestyle intervention aimed at preventing chronic disease. The limitations include the small sample and the fact that the trial was not designed to investigate adverse responses. Nonetheless, although this trial cannot be used to reject the hypothesis that some individuals may experience an adverse response to increased physical activity, it is sufficient to highlight the methodological importance of only considering such analyses in relation to control conditions. In addition, we acknowledge that the definition of an adverse response adopted here, and used elsewhere (4), utilizes a cut-point on a continuous response continuum defined by assay variability rather than clinical significance per se.
In conclusion, although some individuals experienced adverse metabolic responses after receiving a physical activity intervention program, rates were greater in the control group. This study helps highlight the importance of considering control conditions when investigating whether increased physical activity leads to adverse metabolic responses in some individuals. Until this has been undertaken and considering the wide ranging evidence underpinning the therapeutic efficacy of physical activity, it should continue to form the cornerstone of lifestyle interventions aimed at the prevention and treatment of chronic disease.
The authors would like to acknowledge the participants who volunteered for this study.
This trial was funded by a grant from Diabetes UK and the current project is supported by the Leicester Clinical Trials Unit, University of Leicester, and the NIHR Leicester–Loughborough Diet, Lifestyle and Physical Activity, Biomedical Research Unit, which is a partnership between University Hospitals of Leicester NHS Trust, Loughborough University, and the University of Leicester.
There are no perceived conflicts of interest associated with this research.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Bassuk SS, Manson JE. Epidemiological evidence for the role of physical activity
in reducing risk of type 2 diabetes and cardiovascular disease. J Appl Physiol
. 2005; 99: 1193–204.
2. Booth FW, Chakravarthy MV, Gordon SE, Spangenburg EE. Waging war on physical inactivity: using modern molecular ammunition against an ancient enemy. J Appl Physiol
. 2002; 93: 3–30.
3. Booth FW, Laye MJ. The future: genes, physical activity
and health. Acta Physiol
. 2010; 199: 549–56
4. Bouchard C, Blair SN, Church TS, et al. Adverse metabolic response
to regular exercise: is it a rare or common occurrence? PLoS One
. 2012; 7: e37887.
5. Boule NG, Weisnagel SJ, Lakka TA, et al. Effects of exercise training on glucose
homeostasis: the HERITAGE Family Study. Diabetes Care
. 2005; 28: 108–14.
6. Buford TW, Roberts MD, Church TS. Toward exercise as personalized medicine. Sports Med
. 2013; 43: 157–65.
7. Craig CL, Marshall AL, Sjöström M, et al. International physical activity
questionnaire: 12-country reliability and validity. Med Sci Sports Exerc
. 2003; 35 (8): 1381–95.
8. Crouter SE, Schneider PL, Bassett DR Jr. Spring-levered versus piezo-electric pedometer
accuracy in overweight and obese adults. Med Sci Sports Exerc
. 2005; 37 (10): 1673–9.
9. Gillies CL, Abrams KR, Lambert PC, et al. Pharmacological and lifestyle interventions to prevent or delay type 2 diabetes in people with impaired glucose
tolerance: systematic review and meta-analysis. BMJ
. 2007; 334: 299–302.
10. Hawley JA. Exercise as a therapeutic intervention
for the prevention and treatment of insulin resistance. Diabetes Metab Res Rev
. 2004; 20: 383–93.
11. Lacher DA, Hughes JP, Carroll MD. Estimate of biological variation of laboratory analytes based on the Third National Health and Nutrition Examination Survey. Clin Chem
. 2005; 51: 450–2.
12. Leon A, Gaskills S, Rice T, et al. Variability in the response of HDL cholesterol
to exercise training in the HERITAGE Family Study. Int J Sports Med
. 2002; 23: 1–9.
13. Moore SC, Patel AV, Matthews CE, et al. Leisure time physical activity
of moderate to vigorous intensity and mortality: a large pooled cohort analysis. PLoS Med
. 2012; 9 (11): e1001335.
14. Ruchat SM, Rankinen T, Weisnagel SJ, et al. Improvements in glucose
homeostasis in response to regular exercise are influenced by the PPARG Pro12Ala variant: results from the HERITAGE Family Study. Diabetologia
. 2010; 53: 679–89.
15. Schwarz P, Greaves C, Lindström J, Yates T, Davies MJ. Nonpharmacological interventions for the prevention of type 2 diabetes mellitus. Nat Rev Endocrinol
. 2012; 8: 363–73.
16. Selvin E, Crainiceanu CM, Brancati FL, Coresh J. Short-term variability in measures of glycemia and implications for the classification of diabetes. Arch Intern Med
. 2007; 167: 1545–51.
17. Sisson SB, Katzmarzyk T, Earnest CP, Bouchard C, Blair SN, Church TS. Volume of exercise and fitness non-response in sedentary, post-menopausal women. Med Sci Sports Exerc
. 2009; 41 (3): 539–45.
18. Skinner JS, Jaskolski A, Jaskolska A, et al. Age, sex, race, initial fitness, and response to training: the HERITAGE Family Study. J Appl Physiol
. 2001; 90: 1770–6.
20. Wen CP, Wai JP, Tsai MK, et al. Minimum amount of physical activity
for reduced mortality and extended life expectancy: a prospective cohort study. Lancet
. 2011; 378: 1244–53.
21. World Health Organization Diabetes Guideline Development Committee. Definition and Diagnoses of Diabetes Mellitus and Intermediate Hyperglycemia
. Geneva: WHO; 2006. p. 4.
22. Yates T, Davies M, Gorely T, Bull F, Khuni K. Rationale, design and baseline data from the PREPARE (Pre-diabetes Risk Education and Physical Activity
Recommendation and Encouragement) programme study: a randomized controlled trial. Patient Educ Couns
. 2008; 73: 264–71.
23. Yates T, Davies M, Gorely T, Bull F, Khuni K. Effectiveness of a pragmatic education programme aimed at promoting walking activity in individuals with impaired glucose
tolerance: a randomized controlled trial. Diabetes Care
. 2009; 32: 1404–10.
ADVERSE METABOLIC RESPONSE; CHOLESTEROL; GLUCOSE; INTERVENTION; PEDOMETER; PHYSICAL ACTIVITY; STEPS/DAY; TRIGLYCERIDES
Supplemental Digital Content
© 2014 American College of Sports Medicine