When Work Wears You Down: Why Exercise at Your Job May Not Protect Your Heart

For decades, public health guidelines have treated physical activity as a single, undifferentiated good: move more, in any context, and your heart and metabolism will thank you. A new study from Japan challenges that assumption with unusual precision. Researchers led by Rina So of the National Institute of Occupational Safety and Health, Japan, together with colleagues at the National Research Centre for the Working Environment in Copenhagen, have found that among Japanese workers, those whose main physical activity comes from their jobs show cardiometabolic risk profiles that are just as poor as those of sedentary colleagues, and clearly worse than those who are active in their leisure time. The findings, published in BMC Public Health, provide some of the strongest device-based evidence yet that the domain in which we move matters as much as how much we move.

The concept at the heart of the study is known as the physical activity paradox. Epidemiological studies across Europe and North America have repeatedly shown that occupational physical activity, the movement performed during paid work, is associated with elevated rather than reduced risks of cardiovascular disease and premature mortality, while leisure-time physical activity of similar or even lower volume confers protection. Explanations typically point to the different character of the two activity types: work-related movement tends to be prolonged, repetitive, low-intensity, and performed without the recovery periods, autonomy, and psychological enjoyment that accompany voluntary exercise. Yet the paradox has been difficult to test rigorously in Japan, a country where long working hours are culturally entrenched and where the structure of daily life differs markedly from Western populations.

To address that gap, the research team conducted a cross-sectional analysis of 1,233 Japanese workers, using hip- or waist-worn accelerometers to objectively measure physical activity and sedentary behavior for more than seven consecutive days. This device-based approach is a critical methodological strength. Self-reported activity questionnaires are notoriously biased, particularly when respondents conflate work and leisure movement, and they systematically underestimate sedentary time. Accelerometers, by contrast, capture the raw temporal architecture of the day: how long a person accumulates movement, at what intensity, and in which blocks of the daily schedule. By segmenting each participant’s 24-hour record into work, leisure, and sleep, the researchers could assign every minute of activity to its proper domain.

The time-use data alone paint a striking portrait of contemporary Japanese working life. On average, participants spent 9.7 hours per day working, 6.5 hours per day in leisure, and 7.8 hours per day sleeping. That means work occupied a larger share of waking life than leisure did, leaving comparatively little discretionary time in which health-promoting behaviors such as deliberate exercise, meal preparation, and adequate rest could take place. In most high-income countries, leisure comfortably exceeds work time; in this Japanese cohort, the relationship was reversed, which is precisely why the researchers argue that Japan is an especially important setting for testing whether occupational activity can substitute for recreational exercise.

Rather than simply comparing hours of activity, the team used the accelerometer records to classify participants into three behavioral clusters based on their daily activity patterns. The first cluster, labeled Movers, was characterized by high leisure-time physical activity. The second, labeled Grinders, showed high occupational physical activity, meaning these workers accumulated substantial movement on the job but little during their free time. The third cluster, the Chillers, was defined by high sedentary behavior across the day. This clustering strategy preserves the natural structure of real-world behavior, in which activity choices are bundled across the whole day, rather than forcing each behavior into a separate statistical model where domain-specific effects can be obscured.

Cardiometabolic health was assessed using five standard indicators drawn from annual health examinations: waist circumference, systolic and diastolic blood pressure, triglycerides, high-density lipoprotein cholesterol, and fasting plasma glucose. Together these markers capture the core physiological disturbances of metabolic syndrome, including central adiposity, hypertension, dyslipidemia, and impaired glucose regulation. The researchers combined them into a composite cardiometabolic risk score, a technique that increases statistical power and reflects the fact that these risk factors tend to cluster within individuals. Associations with the activity clusters were then evaluated using multiple linear regression, adjusted for a range of potential confounders including demographic and lifestyle variables.

The results were unambiguous. Compared with the Movers, both the Chillers and the Grinders exhibited significantly higher cardiometabolic risk scores. The sedentary Chillers showed a beta coefficient of 0.19, with a 95 percent confidence interval of 0.11 to 0.27, while the physically hard-working Grinders showed a nearly identical coefficient of 0.20, with a confidence interval of 0.10 to 0.31. In practical terms, a worker whose activity comes almost entirely from the job carries roughly the same elevated cardiometabolic risk as a worker who sits most of the day, once other factors are held constant. The similarity of the two coefficients is the study’s most consequential finding, because it demonstrates that high occupational activity does not merely provide a smaller benefit than leisure activity; in this dataset it provided no detectable protective benefit at all.

The physiological literature offers several plausible mechanisms for this asymmetry. Leisure-time exercise typically involves bouts of moderate-to-vigorous intensity that elevate heart rate and cardiac output, improve endothelial function, enhance insulin sensitivity in skeletal muscle, and stimulate parasympathetic recovery afterward. Occupational activity, by contrast, is often dominated by light-intensity, static or highly repetitive loading, prolonged standing, heavy lifting without adequate rest, and sustained elevated blood pressure during tasks such as carrying or pushing. Workers in physically demanding jobs also tend to have less recovery time, less control over their work pace, and higher levels of psychological strain, all of which are independently associated with cardiovascular risk. The paradox, in other words, may reflect not just the type of movement but the total physiological context in which it occurs: intensity, recovery, autonomy, and stress.

The study’s cross-sectional design means that causality cannot be established, and the authors themselves are careful on this point. It remains possible that workers with poorer underlying health select into, or remain in, more physically demanding occupations, a phenomenon sometimes called the healthy worker survivor effect in reverse. Reverse causation of this kind could inflate the apparent risk associated with occupational activity. The authors therefore call for prospective investigations of the physical activity paradox in Japanese working populations, ideally longitudinal designs that follow workers over years and can disentangle the direction of the relationships. Still, the consistency of the Japanese findings with the broader international literature, and the use of objective measurement rather than self-report, substantially strengthens the case that the paradox is real in this population.

The public health implications are considerable. If occupational activity does not protect cardiometabolic health, then current guidelines that count any movement toward weekly activity targets may inadvertently reassure workers in physically demanding jobs who are, in metabolic terms, no better off than their desk-bound counterparts. The authors argue that because work accounts for such a large proportion of daily life in Japan, understanding and addressing the domain-specific effects of activity is especially urgent there. Practical responses could include restructuring work to include recovery periods, reducing prolonged static loading, and designing interventions that add genuine moderate-to-vigorous leisure activity for workers in high-occupational-activity jobs. For the millions of people worldwide whose daily movement comes from labor rather than recreation, the message of this study is uncomfortable but clear: a hard day’s physical work is not a workout, and heart health may depend on finding time to move for its own sake.

Subject of Research: Associations between domain-specific physical activity at work and during leisure and cardiometabolic risk in Japanese workers

Article Title: The physical activity paradox in Japan: associations with cardiometabolic risk

Article References: So, R., Matsuo, T., Murai, F., Holtermann, A., & Gupta, N. (2026). The physical activity paradox in Japan: associations with cardiometabolic risk. BMC Public Health. https://doi.org/10.1186/s12889-026-29606-0

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29606-0

Keywords: physical activity paradox, occupational physical activity, leisure-time physical activity, cardiometabolic risk, accelerometer, sedentary behavior, metabolic syndrome, Japan, occupational health, workers, BMC Public Health, cardiovascular disease