Physical activity and health: introduction to the dose-response symposium

How much physical activity is needed in adults to prevent morbidities, disabilities, and premature death? This is one of the most important questions for all those who have an interest in the topic of physical activity and health and who believe that the promotion of a physically active lifestyle should be the cornerstone of contemporary public health programs. The evolution from performance- and fitness-centered exercise prescriptions toward health-related physical activity recommendations over the last decades had a dramatic influence on this topic. This is best illustrated by the differences in the recommendations contained in the 1978 Position Statement of the ACSM on the issue (1) and in the 1996 Report of the Surgeon General of the U.S. (24). Recognizing the importance of the issue and the uncertainties concerning the amount of physical activity necessary to generate health benefits, a group of Canadians, among them those who organized the 1988 and 1992 International Conferences on exercise or physical activity and health outcomes (7,8), organized a Symposium entirely devoted to dose-response relationships. The Symposium was held from October 11 to 15, 2000, at Hockley Valley Resort, near Toronto. Participation was by invitation only and 24 experts from six countries were asked to review the evidence for a dose-response relationship between regular physical activity and health outcomes. A Consensus Committee composed of highly respected scientists from other fields was asked to evaluate the evidence and write a Consensus Statement. The Committee was chaired by Dr. Antero Kesaniemi, professor of internal medicine, University of Oulu, Finland. Ample time was set aside for discussion and for the members of the Consensus Committee to ask questions and request clarification on any relevant issues. ASSESSING THE QUALITY OF THE EVIDENCE Scientists and practitioners alike understand that the quality of the scientific evidence behind a commonly accepted view or a generalized clinical practice may vary considerably. The evidence is at times very strong but can be rather tenuous in other situations. The “gold standard,” i.e., the highest level of evidence, is thought to be when a solid body of data from several randomized controlled trials (RCTs) is available. Even though it would be desirable to make recommendations based only on such a high standard level of evidence, such data are not always available. Moreover, there are situations in which RCTs cannot even be contemplated. In such cases, one has to rely on other lines of evidence and on the degree of concordance or discordance among a variety of study designs. Controlled but not randomized studies, small experimental research focusing on mechanisms, prospective observational studies, cross-sectional observational research, animal model observations, case studies, surveys of expert views, etc., are typically used to substitute or complement the evidence from RCTs. For the Dose-Response Symposium, we have elected to use a system to qualify the level of evidence that was recently developed at the National Institute of Health (see below). The topic is further addressed in the paper of David Schriger published in this Supplement. At the initiative of the National Heart, Lung, and Blood Institute (NHLBI), a group of experts from many disciplines developed evidence-based guidelines for the prevention and treatment of obesity and its comorbidities that were published in 1998 (16). The experts invited to the Dose-Response Symposium were instructed to use the four evidence categories as defined in the NHLBI report to assess the level and quality of evidence for each particular issue they were addressing and in developing a series of summary statements. Table 1 highlights the evidence categories from the NHLBI report. Evidence Category A is attained when there is a rich body of data from RCTs. The evidence is from endpoints of well-designed RCTs that provide a consistent pattern of findings. Category A therefore requires a substantial number of studies involving a substantial number of participants. Table 1: The evidence categories with the sources of evidence.Evidence Category B is reached when there is a limited body of data from RCTs. It is applicable if few randomized trials exist, they are small in size, trial results are somewhat inconsistent, or trials were undertaken in populations that differ from the target population. Category B may also be attained based on the results of meta-analysis of RCTs. Evidence Category C is granted when the data supporting the conclusion are from uncontrolled or nonrandomized trials, or from cross-sectional or prospective observational studies. Finally, Evidence CategoryD can be given when the provision of some guidance is deemed valuable but there is no compelling scientific or clinical data to justify the use of categories A to C. Category D results from the expert judgment of participants and panel members. DOSE-RESPONSE: EVOLUTION OF CONCEPTS The quantification of the amount of physical activity has been the focus of many studies in the past century. Early on the research emphasis was on the measurement and quantification of the energy costs and physiological demands of occupational tasks, personal chores, sports, and leisure-time physical activities. Subsequently, the focus shifted on the conditions under which regular exercise leads to improvement in physical fitness or optimization of physical performance. More recently, we have seen a growing interest for the assessment of physical activity level and measurement of energy expenditure in free-living individuals under a variety of circumstances. At present, a key question is how much physical activity is needed to experience health benefits and avoid premature death. The Hockley Valley Resort 2000 Symposium was designed to review and qualify the evidence bearing on the latter question. Scandinavian and German physiologists and physicians were the first to investigate the topic of the amount of training necessary to improve fitness and performance. One of the dominant research themes in the 1960s was the differential effects of intermittent and continuous exercise or exercise training (4,9,15,21,22). Another major issue was the difference in cardiovascular and metabolic adaptation to physical work performed at various intensities and durations with the goal of defining the threshold above which performance time was considerably diminished, the so-called “Ausdauerleistungsgrenze”(14). Over the years, various definitions of the concept of “intensity threshold” were proposed but were generally applied more to training and performance issues than to health-related fitness and health outcomes. During the same period, the issue of the minimal amount of regular exercise needed to generate significant health benefits with an emphasis on cardiovascular fitness began to be investigated (5). How much physical activity with comprehensive experimental manipulations of intensity, frequency, and duration of sessions was first asked in the context of training-induced changes in V̇O2max by Shephard in 1968 (23) (predicted V̇O2max) and Davies and Knibbs in 1971 (10) (measured V̇O2max). Later, Nördesjo (18) in 1974 investigated the issue in terms of short (6 min) and long (90 min) maximal performance tests. Subsequently, the attention shifted to the conditions under which submaximal working capacity (PWC at a heart rate of 170) could be improved (6). These studies were complemented by a large number of experiments in which one, or at times two, of the three dimensions of intensity, frequency, and duration of sessions were experimentally altered to assess their effects on an outcome that was generally V̇O2max. These studies up to 1972 were reviewed by Pollock (20). An important landmark in the history of the evolution of concepts on the dose-response relationships was the publication in 1978 by ACSM of the Position Statement on the Recommended Quantity and Quality of Exercise for Developing and Maintaining Fitness in Healthy Adults (1). The recommendations can be summarized as follows: intensity of 60 to 90% of maximum heart rate reserve, frequency of 3–5 d·wk-1, and duration of 15–60 min per session. Locomotor activities and other activities requiring the involvement of large muscle masses were recommended. In 1990, the Position Statement was revised (2). Although it remained essentially unchanged for the intensity, frequency, and duration recommendations, the document recognized that these exercise recommendations were designed to improve physical fitness and cardiorespiratory endurance rather than health-related fitness. It also emphasized that health benefits could be obtained from engaging in regular exercise performed under conditions that differed from those described in the ACSM Position Statement. This was followed by a series of very influential reports and pronouncements on the nature of the relationship between regular physical activity and health outcomes and public health messages. A group of about 20 experts met in 1994 at the invitation of the Centers for Disease Control and Prevention and the ACSM to review the evidence and develop a concise public health message. The results of these deliberations were summarized in JAMA in 1995 (19). The recommendation was that “every U.S. adult should accumulate 30 min or more of moderate-intensity physical activity on most, preferably all, days of the week.” A year later, the very influential report of the Surgeon General of the United States on Physical Activity and Health was published (24). The same recommendation was made with the addition that it applied to people of all ages. The same public health message was adopted by the 13-member panel of the NIH Consensus Development Conference on Physical Activity and Cardiovascular Health (17). An underlying assumption made at all these gatherings of experts and behind the resulting public health message is that the relation between physical activity and health outcomes, particularly mortality rates, is not linear. The common view is that most of the benefits of a regularly active lifestyle can be obtained at low to moderate volumes of physical activity and at less than vigorous intensity (12). In 1998, ACSM published a third Position Statement on the Recommended Quantity and Quality of Exercise for Developing and Maintaining Fitness in Healthy Adults (3). In general, the recommendations were for a larger volume of activity performed at higher intensities than in the public health messages. Finally, Canada’s Physical Activity Guide was published in 1998 (13). The Guide recommended that adults should accumulate 60 min of physical activity every day. In this case, the assumption was that most people interpret the public health message in terms of predominantly light intensity activities, thus the necessity to recommend a larger daily volume. KEY ISSUES CONSIDERED IN THE CONSENSUS PROCESS The emphasis of the Symposium was on level of physical activity and health outcomes. Figure 1 depicts the basic paradigm underlying the consensus effort. Two paths link physical activity to health outcomes. The first is a direct path in which variation in physical activity level is thought to have an impact on health. The second path is one in which variation in physical activity level translates into changes in health-related fitness, which in turn influence health outcomes. Obviously, references were often made to fitness in the discussion but no systematic attempt was undertaken to arrive at a consensus on the relationships between physical or health-related fitness levels and health outcomes. FIGURE 1: The basic paradigm defining the paths from physical activity levels to health outcomes.The Consensus Committee was faced with the challenge of defining the nature of the relationships between regular physical activity and a whole series of health outcomes ranging from premature mortality to quality-of-life indicators. Even though these relationships can be described in detail by a complex family of curves (11), the problem can also be approached in terms of the three curves shown in Figure 2. The first pattern (curve B) is a linear relationship. This curve seems the most appropriate for the relationship between physical activity level and mortality rates as will be evident from the research summarized in this Supplement. The other two curves (A and C) also provide good fits with specific health outcomes. Curve A best describes the dose-response pattern upon which the current physical activity recommendations are based. It specifies that most of the benefits are attained at low to moderate levels of physical activity. One of the aims of the Consensus Symposium was to examine critically the evidence commonly cited in support of this dose-response pattern. In contrast, curve C specifies that the greatest benefits are obtained only when the level of physical activity is rather high. Some of the health outcomes conform to this curve. FIGURE 2: Schematic illustration depicting the relationships between physical activity level defined in minutes of participation per week or energy expended. See text for explanation.A large number of questions are addressed in the papers prepared by the invited experts and were discussed with the Consensus Committee during the Symposium. The definitions of physical activity, exercise, and fitness were considered, as were frequency and duration of sessions. The topic of fractionation into various physical activity periods was addressed. A considerable amount of time was devoted to the topic of absolute and relative intensity, the issue of thresholds, and the monitoring of intensity. Total amount or volume of physical activity, its quantification, and monitoring also received considerable attention. The benefits resulting from an acute exposure to physical activity versus those expected with regular participation were discussed. The levels of risks versus the anticipated benefits were taken into consideration for a variety of outcomes. Whether the apparent benefits associated with a physically active lifestyle could be imputed to physical activity or energy expenditure per se or to ensuing loss of adiposity was a topic that generated much interest. The participants also discussed whether there were any differences in the dose-response relationships between men and women, young versus older people or ethnic groups. In summary, the goal of the Consensus Symposium was to critically examine the evidence for the dose-response relationships between physical activity levels and health benefits and to identify key issues for future research. The key questions were: Is there a dose-response relation? Does it vary by outcome? What is the exact nature of the relationship? Is there evidence for a threshold? These central questions provided the impetus for an evidence-based Symposium. CONTENT OF THE SUPPLEMENT The various texts published in this Supplement of Medicine and Science in Sports and Exercise deal with these major questions. After this introduction, the Consensus Statement is presented. This is followed by the papers that focused on definitions and general issues. After a series of related topics (the equivalent of half a day of meeting time), a summary of the evidence as defined by the colleague who chaired the session is presented. Thus, the Supplement includes the Preface, Introduction, Consensus Statement, 24 papers, and 6 summaries from chairpersons. Eleven of these papers deal with specific health outcomes. Address for correspondence: Claude Bouchard, Ph.D., Pennington Biomedical Research Center, 6400 Perkins Road, Baton Rouge, LA 70808; E-mail: email protected