The growing prevalence of type 2 diabetes is one consequence of the obesity epidemic. Because the overwhelming majority of persons with type 2 diabetes are overweight or obese, the two disorders share common risk factors. Although the genetic determinates of both diseases cannot be discounted, the primary contributors to overweight and diabetes are adverse lifestyle factors. Specifically, the positive energy imbalance resulting from a hypercaloric diet unabated by increased levels of physical activity is directly associated with weight gain and associated comorbidities such as type 2 diabetes. Clinical trials have demonstrated that dietary modifications that lower energy intake while increasing energy expenditure through physical activity lower the incidence of diabetes in high- and normal-risk persons (9,29).
A number of nutritional factors have been evaluated in relation to diabetes development including overall energy (kilocalorie) intake, timing of meals, and specific foods. With the exception of strong evidence relating kilocalorie restriction to a reduction in diabetes incidence, conclusions from studies evaluating the association between specific foods and diabetes are inconsistent (10,19). At the opposite end of the spectrum, physical activity of varying frequency, duration, and intensity has consistently demonstrated an inverse association with diabetes incidence (2,15). Negative energy imbalance leading to weight loss is the most biologically plausible reason for this observation; however, many studies have demonstrated an association independent of weight loss.
The inverse association between physical activity and diabetes that remains after accounting for weight loss is described as the "residual effect" of physical activity. The presence of a residual effect of activity implicates mechanisms in addition to weight loss that protect against diabetes development. Some previously described mechanisms that could directly or indirectly lower diabetes risk include increased glucose uptake and transport, improved muscle insulin sensitivity and endothelial function, and reduced inflammation (Fig. 1) (15,26). Less attention has been paid to an equally biologically plausible and potentially important mechanism, autonomic nervous system function.
In this article, we will explore the hypothesis that improved autonomic nervous system function is one mechanism by which higher levels of physical activity are inversely associated with diabetes development. We selected this particular understudied mechanism because prior research from our team (4,5,7) and others (1,8,11,16,17,22,24) suggests that autonomic nervous system function may underlie obesity, insulin sensitivity, endothelial function, and inflammation - all mechanisms that are associated with the development of diabetes (11). We will advance our argument by summarizing findings from clinical and population studies on the associations among the following: 1)activity and diabetes development, 2) autonomic nervous system function and diabetes development, and, 3) the role of physical activity or fitness on autonomic nervous system function. We expect that this article will provide evidence for a plausible and modifiable mechanism involved in the development of diabetes.
PHYSICAL ACTIVITY AND DIABETES DEVELOPMENT
Exercise plays a crucial role in regulating glucose metabolism. Skeletal muscle is the largest insulin-sensitive tissue in the body and accounts for the greatest proportion of insulin-mediated glucose disposal (26). All movements that use the skeletal muscle, ranging from deliberate aerobic exercise training to weight training, and the movements that define activities of daily living stimulate the uptake of circulating blood glucose and improve muscle utilization of glucose to promote glucose homeostasis. Regular physical activity improves muscle insulin sensitivity, thus requiring lower levels of insulin secretion by the pancreas to control diabetes.
Higher levels of physical activity and cardiorespiratory fitness are consistently associated with a lower risk of diabetes development in observational studies and clinical trials (2,15). Leisure time, occupational, and active transportation activities are each inversely associated with the development of diabetes in a dose-dependent manner. Those few longitudinal studies that included an objective measure of cardiorespiratory fitness confirmed the inverse association of fitness with diabetes development and generally report findings of an even stronger magnitude because of the added precision of measuring fitness instead of self-reported activity. The residual effect of activity on diabetes development after multivariable statistical adjustment for body weight is even more apparent in studies using measures of fitness as compared with physical activity.
Additionally, studies that collected measures of either physical activity or fitness at multiple time points demonstrated that participants whose activity levels or fitness improved were less likely to develop diabetes than those participants whose fitness or activity remained the same (3). Randomized trials such as the Diabetes Prevention Program (DPP) demonstrated that high-risk adults (i.e., overweight with impaired glucose tolerance) who met the recommendations for 150 min·wk−1 of moderate physical activity and modified their diet had a lower incidence of type 2 diabetes than participants undergoing pharmacological therapy (9). While most participants in the intervention arm lost weight, the diabetes-lowering effect remained fairly strong. The substantial variation in weight loss in the DPP, and the range of weight change in observational studies, prompted exploration of additional mechanisms to explain the association between activity and diabetes (2,3,14,15).
AUTONOMIC NERVOUS SYSTEM FUNCTION
The autonomic nervous system is a division of the peripheral nervous system that controls automated body functions including heart rate, blood pressure, digestion, and metabolism. The autonomic nervous system is subdivided into the parasympathetic and sympathetic components that work antagonistically to provide a very fine degree of control over their target organs. In general, the parasympathetic nervous system predominates during rest by slowing heart rate, lowering blood pressure, and promoting digestion. The sympathetic nervous system is responsible for mounting responses to physical and psychological stimuli. In response to a challenge, the sympathetic nervous system boosts heart rate and blood pressure and directs blood flow away from digestion to maintain glucose in the blood stream so that it can be used for immediate energy for the well-known "fight or flight" response.
Autonomic function can be measured directly in animals and smaller studies of adults that permit more direct and invasive methodology such as the measurement of muscle sympathetic nerve activity (MSNA). In population studies, autonomic function is most often estimated noninvasively by measuring heart rate variability and the heart rate response to challenges. Heart rate variability refers to the beat-to-beat variability of heart rate as measured for lengths of time ranging from a 10-s standard 12-lead to 24-h Holter recordings. A healthy heart rate is not fixed but rather varies in milliseconds in response to moment-to-moment physiological changes, and low heart rate variability reflects generally poor autonomic tone (27). The autonomic response to a challenge can be determined using heart rate recovery (HRR) from a graded exercise treadmill test. HRR is the difference between maximum heart rate and heart rate 1- or 2-min after cessation of the treadmill test. The principle determinant of HRR is thought to be parasympathetic reactivation. Smaller (slower) values of HRR indicate a smaller parasympathetic response (21).
Autonomic dysfunction, impaired or improper autonomic responsiveness to challenge, is correlated with a number of adverse health behaviors and diseases. Diabetes and hypertension are most commonly associated with autonomic dysfunction. In persons with diabetes, prolonged hyperglycemia leads to degradation of the microvasculature, leading to a specific form of autonomic dysfunction termed "diabetic autonomic neuropathy". Although diabetic autonomic neuropathy is an established complication of diabetes (12), there is sufficient evidence from animal, clinical, and population studies implicating autonomic dysfunction in the development of diabetes.
Autonomic Nervous System Influence over Glucose Metabolism
Parasympathetic and sympathetic nerve fibers are present in each of the major organs and body systems that control metabolism (Fig. 2). Glucose and insulin are regulated by a feedback cycle that is controlled, in part, by autonomic inputs. As plasma glucose concentrations rise with feeding, a healthy autonomic response is characterized by parasympathetic neurons stimulating the pancreas to produce insulin. The liver simultaneously increases insulin sensitivity and shuts down glucose production in favor of glucose uptake. Conversely, sympathetic activation blunts the pancreatic secretion of insulin. Hepatic glucogenesis is stimulated, and glucose uptake slows down. These antagonistic actions provide a fine degree of control over energy regulation, making more energy available during times of high physical or psychological demands (sympathetic activation) and storing energy when needs are low (parasympathetic activation). If these systems do not work appropriately, glucose homeostasis can be disrupted, and weight gain can be stimulated.
Evidence for the association between autonomic dysfunction and weight gain was generated from a number of animal and clinical studies. Landsberg (16) hypothesized that feeding-induced thermogenesis and insulin secretion are sympathetically stimulated and are exacerbated by overeating. Van Vliet and colleagues (28) fed dogs a high-fat diet for 12 wk and found that the difference they observed in resting heart rate between normal weight and obese dogs was abolished after administration of an autonomic blocking drug. The authors concluded that the development of obesity in dogs was accompanied by impaired autonomic regulation. In humans, MSNA was elevated in obese volunteers as compared with nonobese volunteers, and MSNA was reduced after weight loss (13). Snitker and colleagues (25) conducted a comprehensive review of studies investigating the association between sympathetic nervous system activity and obesity and concludedthat the sympathetic system plays a critical role in the development of obesity.
Fueled by findings from smaller studies, a number of investigators tested the cross-sectional association among autonomic function, obesity, and glucose and insulin metabolism in larger epidemiological cohort studies. Central fat, and in particular, visceral fat, demonstrates the strongest inverse association with sympathetic nervous system functioning (1). Both Liao et al. (17) and Dekker et al. (8) described poorer overall autonomic balance and decreased parasympathetic control in persons with elevated fasting insulin and glucose in the nondiabetic range. In at least two reports (5,17), the association between parasympathetic dysfunction and hyperinsulinemia was stronger and more consistent than the association of parasympathetic function with hyperglycemia. Participants who were free from diabetes in the Lipid Research Clinics Prevalence study but who fell in the highest quartile of triglyceride-high-density lipoprotein cholesterol ratio (a marker of insulin resistance) had the highest overall prevalence of abnormally slow HRR from a graded exercise treadmill test, indicative of impaired parasympathetic reactivation (24). A similar association was observed with the relation between fasting glucose and abnormal HRR in the same study (22). Findings from clinical and population studies describe an association of autonomic dysfunction with obesity and insulin resistance that is present in the absence of diabetes. This observation suggests that autonomic dysfunction is present before the development of diabetes.
Autonomic Dysfunction and Diabetes Development
In three different population-based longitudinal studies each using different measures of autonomic function, we tested the hypothesis that autonomic function could be identified before the onset of diabetes. In the Atherosclerosis Risk in Communities study of 8185 healthy middle-aged adults who were free from diabetes at baseline, we observed that participants with higher heart rates (>73 bpm) were 1.6 times more likely to develop diabetes over approximately 8 yr of follow-up than participants with heart rates of 60 bpm or lower. Participants with low 2-min low-frequency heart rate variability, a marker of the joint contribution of sympathetic and parasympathetic function, were also at increased risk of developing diabetes (4).
We replicated these findings in 3295 men and women aged 18 to 30 yr from the Coronary Artery Risk Development in Young Adults (CARDIA) study cohort using HRR taken 2 min after cessation of a symptom-limited exercise treadmill test (5). Participants free from diabetes at baseline with HRR of less than 42 bpm were more likely to develop diabetes over 15 yr than were participants with HRR of greater than 42 bpm. However, we found that this association was strongest among participants who jointly had low cardiorespiratory fitness. In those participants with both low cardiorespiratory fitness and slow HRR, the odds of developing diabetes was 3.4 times higher than among participants with faster HRR. Low fitness was undoubtedly more strongly associated with diabetes development than poorer parasympathetic reactivation, but slow HRR demonstrated a "modifying" effect in the presence of low fitness that identified participants at an even greater likelihood for developing diabetes.
The DPP was a randomized controlled trial initiated to test whether the administration of metformin or lifestyle modifications to reduce sedentary behavior and decrease overall energy intake were effective in preventing the development of diabetes in 2980 high-risk persons (i.e., overweight and with impaired glucose tolerance) who were followed for 3 yr. One of the objectives of our post hoc analysis in the DPP was to test whether autonomic function parameters at baseline or the changes in parameters over follow-up were associated with the development of diabetes (7). Baseline heart rate was associated with incident diabetes after adjustment for age, race, sex, and weight change in the lifestyle and metformin arms (hazard ratios, 1.19 and 1.17 per 11 bpm, respectively; both P < 0.05). We additionally observed that decreases in heart rate and increases in heart rate variability over 3 yr were both associated with a lower risk of developing diabetes over time (18).
AUTONOMIC FUNCTION AS A MECHANISM
Physical Activity and Autonomic Function
Autonomic function is influenced by current disease and lifestyle factors including physical activity. Small studies conducted in selected samples (e.g., endurance runners and volunteers) suggested that physical activity is inversely associated with measures of autonomic function. To address the paucity of research addressing this question in diverse population-based study samples and using longitudinal study designs, we tested whether higher levels of physical activity were associated with faster HRR from an exercise treadmill test in participants from the CARDIA study. Using multiple measurements collected over time, we additionally tested whether increasing or maintaining physical activity was associated with more favorable changes in HRR over time.
After adjustment for age, race, sex, body mass index, smoking status, and diastolic blood pressure, participants in the highest tertile of physical activity had significantly faster HRR 2 min after the exercise treadmill test (45.1 bpm vs 41.8 bpm in the lowest tertile of activity; P < 0.01). We also found that participants whose physical activity levels remained the same or increased over the 7 yr before treadmill fitness testing had significantly smaller declines in HRR (−1.8 bpm and −1.3 bpm, respectively) than those participants whose activity levels declined (−3.6 bpm; P < 0.01 vs increase or stable). We further evaluated whether the association between physical activity and HRR was consistent by race, sex, weight strata, smoking status, blood pressure categories, and glycemic status. Physical activity was directly associated with HRR across demographic and disease strata (6). Although it was notable that physical activity is associated with better autonomic function in the presence of indicators of poor health such as diabetes or hypertension, it is even more notable that physical activity and autonomic function are associated in the absence of preexisting diseases known to influence autonomic function.
Our findings from the CARDIA study prompted us to hypothesize that participants in the intervention arm of the DPP trial would have more favorable changes in autonomic function during the study (7). Participants in the intensive lifestyle modification arm were counseled to engage in 150 min of moderate-to-vigorous exercise each week and simultaneously reduce dietary fat intake with the goal of losing and maintaining the loss of 7% of their present body weight. As we hypothesized, heart rate declined, and heart rate variability increased to the greatest extent among participants in the intensive lifestyle modification arm, which reflected improved parasympathetic function. Participants in the medication (metformin) arm of the trial had a lower likelihood of developing diabetes but experienced relatively smaller nonsignificant changes in autonomic function. It is possible that improved autonomic function attributable to physical activity further lowered the risk of diabetes in the lifestyle modification arm.
Despite the consistency of prior findings, it is not possible to determine which components of the lifestyle intervention in the DPP influenced changes in autonomic function. Prior research clearly identified a role for increased physical activity, and 74% of participants in the lifestyle arm reported meeting their activity goals 24 wk postrandomization (9). However, those same participants may have changed their dietary intake, and a higher dietary intake of omega-3 fatty acids is associated with improvements in autonomic function (20). Alternatively, weight loss experienced by participants in the lifestyle modification arm may have contributed to the observed improvements in autonomic function. Future studies controlling for dietary intake, weight change, and activity are needed to isolate and quantify the relative contributions of each to changes in autonomic function. Table summarizes population studies relating autonomic function with diabetes incidence and physical activity with autonomic nervous system function.
STRENGTHS AND LIMITATIONS
The association between autonomic function and incident diabetes is relatively smaller than the association between weight loss and incident diabetes, but the potential role of autonomic function is important. Autonomic dysfunction influences other mechanisms that modify diabetes risk, namely insulin sensitivity, glucose regulation, and inflammation. Inflammation and endothelial dysfunction co-occur with autonomic dysfunction; however, this could be attributable to a common antecedent such as obesity. Autonomic function is relatively easy to measure and standardize in large observational studies. Resting heart rate variability measurement requires little participant burden and time and can provide valuable information about the relative contributions of the sympathetic and parasympathetic divisions through mathematical manipulations of the rhythm strip. Although HRR requires greater resources and participant burden because dynamic exercise tests are administered, treadmill tests are commonly done in large populations to collect other prognostic information. Heart rate is a standard simple procedure that can provide crude information about the combination of fitness and autonomic function that has reliably been associated with diabetes incidence and mortality. Perhaps most importantly, autonomic function can be modified through lifestyle changes that include more physical activity, or it can be modified pharmacologically.
Although the studies that provide evidence in support of our argument meet many causal criteria including biological plausibility, temporality, and consistency, the findings are hindered by some limitations. First, few scientists outside of our investigative team have tested this hypothesis using a longitudinal study design. Evidence for the consistency of associations would be even stronger if this hypothesis was tested in different study populations. Second, the measures of autonomic function administered in population studies are all based on alterations in heart rate. No studies to date have evaluated changes in catecholamine levels, nor have investigators assessed sympathetic nerve activity directly. Future observational studies and trials should assess autonomic function using measures that capture both the parasympathetic and sympathetic inputs of function. Finally, in our studies to date, we have not had glucose tolerance testing data available to rule out diabetes that was not detected using fasting glucose measures. Thus, it is possible that diabetes status is misclassified in some participants at baseline, which could bias estimates away from the null and overestimate the effect sizes. Despite these limitations, the biological plausibility of our hypothesis, along with the unexplained residual effect of physical activity in existing studies of physical activity and diabetes, warrants further investigation of this topic.
The autonomic nervous system is an attractive mechanism to evaluate in the association between physical activity and diabetes incidence. One of the most compelling reasons is the responsiveness of the autonomic nervous system to lifestyle changes. Increasing physical activity levels has numerous primary and secondary benefits that are realized independent of the challenging goal of weight loss. The positive impact of physical activity on autonomic nervous system functioning may prove to be a key benefit in diabetes prevention.
1. Alvarez, G.E., S.D. Beske, T.P. Ballard, and K.P. Davy. Sympathetic neural activation in visceral obesity. Circulation
. 106:2533-2536, 2002.
2. Bassuk, S.S., and J.E. Manson. Epidemiological evidence for the role of physical activity
in reducing risk of type 2 diabetes
and cardiovascular disease. J. Appl. Physiol
. 99:1193-1204, 2005.
3. Carnethon, M.R., S.S. Gidding, R. Nehgme, S. Sidney, D.R. Jacobs Jr, and K. Liu. Cardiorespiratory fitness
in young adulthood and the development of cardiovascular disease risk factors. JAMA
. 290:3092-3100, 2003.
4. Carnethon, M.R., S.H. Golden, A.R. Folsom, W. Haskell, and D. Liao. Prospective investigation of autonomic nervous system function
and the development of type 2 diabetes
: the Atherosclerosis Risk In Communities study, 1987-1998. Circulation
. 107:2190-2195, 2003.
5. Carnethon, M.R., D.R. Jacobs, Jr., S. Sidney, and K. Liu. Influence of autonomic nervous system dysfunction on the development of type 2 diabetes
: the CARDIA study. Diabetes Care
. 26:3035-3041, 2003.
6. Carnethon, M.R., D.R. Jacobs, Jr., S. Sidney, B. Sternfeld, S.S. Gidding, C. Shoushtari, and K. Liu. A longitudinal study of physical activity
and heart rate recovery: CARDIA, 1987-1993. Med. Sci. Sports Exerc
. 37:606-612, 2005.
7. Carnethon, M.R., R.J. Prineas, M. Temprosa, Z.M. Zhang, G. Uwaifo, and M.E. Molitch. The association among autonomic nervous system function
, incident diabetes, and intervention arm in the diabetes prevention program. Diabetes Care
. 29:914-919, 2006.
8. Dekker, J.M., E.J. Feskens, E.G. Schouten, P. Klootwijk, J. Pool, and D. Kromhout. QTc duration is associated with levels of insulin and glucose intolerance. The Zutphen Elderly Study. Diabetes
. 45:376-380, 1996.
9. Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes
with lifestyle intervention or metformin. N. Engl. J. Med
. 346:393-403, 2002.
10. Dunstan, D.W., R.M. Daly, N. Owen, D. Jolley, M. de Courten, J. Shaw, and P. Zimmet. High-intensity resistance training improves glycemic control in older patients with type 2 diabetes
. Diabetes Care
. 25:1729-1736, 2002.
11. Egan, B.M. Insulin resistance and the sympathetic nervous system. Curr. Hypertens. Rep
. 5:247-254, 2003.
12. Ewing, D., I.W. Campbell, and B.F. Clarke. The natural history of diabetic autonomic neuropathy. Q. J. Med
. 49:95-108, 1980.
13. Grassi, G., G. Seravalle, M. Colombo, G. Bolla, B.M. Cattaneo, F. Cavagnini, and G. Mancia. Body weight reduction, sympathetic nerve traffic, and arterial baroreflex in obese normotensive humans. Circulation
. 97:2037-2042, 1998.
14. Kriska, A.M., A. Saremi, R.L. Hanson, P.H. Bennett, S. Kobes, D.E. Williams, and W.C. Knowler. Physical activity
, obesity, and the incidence of type 2 diabetes
in a high-risk population. Am. J. Epidemiol
. 158:669-675, 2003.
15. LaMonte, M.J., S.N. Blair, and T.S. Church. Physical activity
and diabetes prevention. J. Appl. Physiol
. 99:1205-1213, 2005.
16. Landsberg, L. Diet, obesity and hypertension: an hypothesis involving insulin, the sympathetic nervous system, and adaptive thermogenesis. Q. J. Med
. 61:1081-1090, 1986.
17. Liao, D., J. Cai, F.L. Brancati, A. Folsom, R.W. Barnes, H.A. Tyroler, and G. Heiss. Association of vagal tone with serum insulin, glucose, and diabetes mellitus-The ARIC Study. Diabetes Res. Clin. Prac
. 30:211-221, 1995.
18. Lichtenstein, A.H., L.J. Appel, M. Brands, M. Carnethon, S. Daniels, H.A. Franch, B. Franklin, P. Kris-Etherton, W.S. Harris, B. Howard, N. Karanja, M. Lefevre, L. Rudel, F. Sacks, L. Van Horn, M. Winston, and J. Wylie-Rosett. Diet and lifestyle recommendations revision 2006: a scientific statement from the American Heart Association Nutrition Committee. Circulation
. 114:82-96, 2006.
19. Meyer, K.A., L.H. Kushi, D.R. Jacobs, Jr., J. Slavin, T.A. Sellers, and A.R. Folsom. Carbohydrates, dietary fiber, and incident type 2 diabetes
in older women. Am. J. Clin. Nutr
. 71:921-930, 2000.
20. Mozaffarian, D., R.J. Prineas, P.K. Stein, and D.S. Siscovick. Dietary fish and n-3 fatty acid intake and cardiac electrocardiographic parameters in humans. J. Am. Coll. Cardiol
. 48:478-484, 2006.
21. Nishime, E.O., C.R. Cole, E.H. Blackstone, F.J. Pashkow, and M.S. Lauer. Heart rate recovery and treadmill exercise score as predictors of mortality in patients referred for exercise ECG. JAMA
. 284:1392-1398, 2000.
22. Panzer, C., M.S. Lauer, A. Brieke, E. Blackstone, and B. Hoogwerf. Association of fasting plasma glucose with heart rate recovery in healthy adults: a population-based study. Diabetes
. 51:803-807, 2002.
23. Rennie, K.L., H. Hemingway, M. Kumari, E. Brunner, M. Malik, and M. Marmot. Effects of moderate and vigorous physical activity
on heart rate variability in a British study of civil servants. Am. J. Epidemiol
. 158:135-143, 2003.
24. Shishehbor, M.H., B.J. Hoogwerf, and M.S. Lauer. Association of triglyceride-to-HDL cholesterol ratio with heart rate recovery. Diabetes Care
. 27:936-941, 2004.
25. Snitker, S., I. Macdonald, E. Ravussin, and A. Astrup. The sympathetic nervous system and obesity: role in aetiology and treatment. Obes. Rev
. 1:5-15, 2000.
26. Stump, C.S., E.J. Henriksen, Y. Wei, and J.R. Sowers. The metabolic syndrome: role of skeletal muscle metabolism. Ann. Med
. 38:389-402, 2006.
27. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation
. 93:1043-1065, 1996.
28. Van Vliet, B.N., J.E. Hall, H.L. Mizelle, J.P. Montani, and M.J. Smith, Jr. Reduced parasympathetic control of heart rate in obese dogs. Am. J. Physiol
. 269:H629-H637, 1995.
29. Yates, T., K. Khunti, F. Bull, T. Gorely, and M.J. Davies. The role of physical activity
in the management of impaired glucose tolerance: a systematic review. Diabetologia
. 50:1116-1126, 2007.