Dose-Response Relation Between Physical Activity and Fitness: Title and subTitle BreakEven a Little Is Good; More Is Better

The concept of dose is important in clinical medicine. In the pharmacological treatment of many conditions, physicians typically start with a dose of a drug believed to be the minimum effective dose. If the patient does not respond, this initial dose may then be titrated upward to a maximum dose, beyond which the adverse effects of the drug are unacceptable for treatment. Thus, all marketed drugs require data on their efficacy and safety.

Physical activity, while not a drug, can behave like one—it causes many physiological changes in the body (often beneficial for health),¹ helps prevent the development of many chronic diseases,² and is a useful adjunct to drug treatment for many diseases, including cardiovascular disease, cancer, and diabetes.^{3 - 4} Health care professionals, including physicians, are encouraged to prescribe physical activity for health.^{3 - 4} It is plausible that there is a minimum dose of physical activity for health benefits, that these benefits increase with increasing dose, and that beyond a certain dose, adverse effects (eg, musculoskeletal injuries,⁵ sudden death⁶ ) outweigh benefits.

Unlike prescription drugs, however, the minimum dose, dose response, and maximum safe dose of physical activity are not well understood. An expert panel, charged with reviewing the evidence on the dose-response relation between physical activity and health outcomes, concluded in 2000 that it was unable to describe the relation because of the paucity of data.² Since then, several investigators^{7 - 9} have attempted to address dose response questions for physical activity in individual studies; however, it is difficult to synthesize the information across studies because investigators have measured physical activity in different ways and classified physical activity according to different dose schemes that often are difficult to compare directly. A meta-analysis examining the dose response between levels of physical activity or fitness and cardiovascular disease risk¹⁰ did not attempt to find congruence among the many different definitions and classifications of physical activity dose, choosing, instead, to categorize physical activity according to the distribution of study participants (eg, 20th percentile of the study population, instead of, say, 210 kcal/week of energy expended).

Are dose questions related to physical activity just fodder for academic rumination? By no means—these questions have practical implications for both patient and clinicians. Over the years, various expert groups, based on the best evidence available, have formulated different physical activity recommendations and guidelines, which require different doses of physical activity. Recommendations from the 1970s and 1980s prescribed vigorous exercise (eg, running) for 20 minutes continuously, 3 days a week¹¹ ; in the 1990s, suggested at least 30 minutes a day of accumulated moderate-intensity activity (eg, brisk walking) most days of the week^{12 - 14} ; in the 2000s, advocated at least 60 minutes a day of moderate activity.^{15 - 16} Predictably, many patients and clinicians are confused about what dose of physical activity is needed.¹⁷

In this issue of JAMA, the randomized clinical trial reported by Church and colleagues¹⁸ provides some clarification to dose-response questions for physical activity. In this trial, 464 sedentary postmenopausal women, with mean body mass index of 31.8 and mean systolic blood pressure of 139.8 mm Hg, were randomly assigned to a control group or to 3 groups with different exercise doses: 4, 8, or 12 kcal/kg per week. Adherence (exercise sessions were directly observed) and follow-up over the 6-month intervention were excellent. On average, the exercise groups carried out 72, 136, and 192 minutes per week, respectively, of moderate-intensity physical activity (cycling and walking). These correspond to, approximately, 50%, 100%, and 150% of the surgeon general's recommendation (30 minutes a day, most days of the week, generally accepted as 5 days a week, for a total of 150 minutes a week,¹³ similar to other expert recommendations^{12 ,14} ). Although the total minutes of exercise corresponded to about 50%, 100%, or 150% of the recommendation, the number of sessions per week were lower than recommended; 2.6 to 3.1, instead of 5. The primary outcome, physical fitness, showed a linear, dose response over the 3 exercise groups, with significant increases of 4.2% in the 4-kcal/kg, 6.0% in the 8-kcal/kg, and 8.2% in the 12-kcal/kg per week groups in peak absolute oxygen consumption. In contrast, other cardiovascular risk factors—blood pressure, lipid profile, and weight—showed no significant improvements with any exercise dose.

An interesting finding from this study is that even exercising at 50% of recommendations appeared sufficient to provide some improvement in fitness. Could this improvement be due not only to the 72 minutes a week (a very modest dose) of observed cycling and walking, but to additional physical activity outside of the intervention in which, perhaps, these newly empowered women engaged? This is unlikely because the pedometers worn by participants indicated that daily steps hovered around 5000 a day over the study. The observation of improved fitness with as little as 72 minutes a week of physical activity dovetails well with data from the Women's Health Study investigating a clinical end point (vs fitness, a risk factor); women who walked as little as 1 to 1.5 hours a week, also approximately equivalent to 50% of the surgeon general's recommendation, had half the risk of developing coronary heart disease than did sedentary women.⁸ These data should be encouraging to sedentary individuals because they indicate that an achievable dose of physical activity may be sufficient to begin reaping health benefits.

The trial by Church et al also found no significant interaction by race or weight for fitness improvements caused by exercise. These findings correspond well with data from other studies investigating clinical end points; the Women's Health Study and the Women's Health Initiative reported similar risk reductions for coronary heart disease or cardiovascular disease with physical activity, among women of different adiposities and races.^{8 - 9}

The investigators also report that other cardiovascular risk factors—blood pressure, lipid profile, and weight (fasting glucose showed some improvement, of borderline significance)—did not improve with exercise. Thus, dose-response relations between physical activity and different health outcomes are probably different. In particular, with obesity a major health problem worldwide,¹⁹ sedentary, overweight or obese individuals should not be lulled into believing that 72 minutes a week of physical activity will ameliorate their weight concerns. Even at 192 minutes a week of activity, the highest dose, trial participants, who were not asked to change their diet, did not lose weight (although waist circumference did decrease). Therefore, given the typical US diet, it is likely that for many individuals, guidelines suggesting 60 minutes or more of daily physical activity,^{15 - 16} which were developed with a focus on weight control, are more appropriate if weight control were the primary goal.

Although the trial by Church et al shows a linear dose-response relation between physical activity and improvements in physical fitness, with benefit observed beginning at 72 minutes a week of moderate activity, it is limited in its ability to provide direct answers for other patterns of physical activity. Understandably, because of cost and feasibility reasons, the trial mimicked only 3 physical activity patterns that can occur; in real life, there are infinitely more. For example, this trial does not provide information about whether vigorous activities, such as running, can improve the cardiovascular risk factors that the moderate cycling and walking, carried out by participants, did not. And, as for typical patterns of activity in everyday life, eg, accumulated bouts of cycling or walking in commuting, or walking the dog, will fitness improve to the extent seen in the trial if the physical activity was not conducted in a single session but was broken up over the day? Or, if rather than the average 2.6 to 3.1 sessions a week of physical activity undertaken by study participants, the activity was conducted in 1 to 2 sessions over the weekend (weekend warrior) by busy, working adults? These questions ask about dose(s) of physical activity, with dose(s) referring to additional components of activity apart from the component studied, total energy expended in moderate activity. Returning to the drug analogy, an important unanswered question is, what is the dose-response for adverse health effects—at what activity level do harmful effects negate beneficial health effects?

Although current knowledge regarding the dose-response relation between physical activity and health remains incomplete, the study by Church et al does provide important information on the dose of physical activity to improve physical fitness, a strong predictor of chronic disease and premature mortality.²⁰ This may be succinctly summarized for patients and clinicians as “Even a little is good; more may be better!”