Excess adiposity and physical inactivity are established risk factors for premature mortality and non-communicable diseases like CVD and cancer. Previous research, primarily relying on single baseline assessments, has yielded inconsistent findings on the joint effects of physical activity and adiposity on mortality. These studies are limited because both factors change over time, leading to potential biases like non-differential misclassification and regression dilution. A recent study in the UK Biobank showed positive changes in one factor attenuating negative changes in the other. However, there's a lack of longitudinal evidence, particularly from Asian populations, on the joint associations of physical activity and adiposity changes with cause-specific mortality (CVD and cancer). This study aims to address this gap by investigating these associations in a large cohort of Taiwanese adults, a population with a higher predisposition to cardiometabolic disorders and a high cancer mortality rate.

Existing literature examining the relationship between physical activity, adiposity, and mortality has yielded mixed results. Some studies suggest a synergistic relationship, where physical activity attenuates the increased risk associated with excess adiposity, while others report independent effects. Most prior studies relied on single baseline measurements of physical activity and adiposity, failing to account for the dynamic nature of these risk factors over time. This limitation has led to inconsistencies and challenges in determining the nuanced influence of each risk factor on mortality. The 2020 WHO Physical Activity Guidelines and the Physical Activity Guidelines for Americans acknowledge the need for more longitudinal research, especially in non-Western populations, to better understand the temporal relationship between these factors and mortality, and whether these associations vary across ethnicities. This is particularly important in Asian countries, where there is a greater predisposition to cardiometabolic disorders for a given adiposity level and where cancer has become a leading cause of death.

This study utilized longitudinal data from Taiwan's MJ Cohort, comprising 116,228 adults (46.2% female) recruited between 1998 and 2013. Participants underwent repeated health screenings approximately 4.6 years apart, with a follow-up period for mortality of 11.9 years. Physical activity was assessed using questionnaires, categorizing activities by intensity (METs) and calculating weekly physical activity volume (MET-h). Participants were classified as inactive, insufficiently active, or sufficiently active according to WHO guidelines. Changes in physical activity were categorized as decreased, stable, or increased. Adiposity was measured using BMI, waist circumference (WC), and body fat percentage (BF%), classified into categories based on public health and clinical guidelines. Changes in adiposity were also categorized as decreased, stable, or increased. Participants were further classified into nine mutually exclusive groups based on changes in physical activity and each adiposity indicator. Mortality data was obtained from the National Death file, categorizing deaths into all-cause mortality (ACM), CVD mortality, and cancer mortality. Cox proportional hazards regression models were used for ACM, while Fine-Gray subdistribution hazard models were used for CVD and cancer mortality (to account for competing risks). Models were adjusted for age, sex, smoking, alcohol consumption, sleep duration, diet, education, and baseline physical activity and adiposity. Sensitivity analyses were conducted using multiple imputation for missing data, restricting the analysis to participants aged 40 or older at baseline, and adjusting for potential mediators (hypertension and diabetes). E-values were calculated to assess the potential impact of unmeasured confounding, and negative control outcomes were used to assess potential bias.

Compared to stable-insufficient physical activity, increasing physical activity was consistently associated with lower risk of ACM, CVD, and cancer mortality. This reduction in risk was similar in magnitude to maintaining sufficient physical activity at both time points. Decreasing physical activity from any baseline level increased mortality risk. Increasing adiposity from a baseline of overweight/moderate was associated with increased mortality risk for all three outcomes, with the strongest effect on CVD mortality. Decreasing adiposity from overweight/obese levels attenuated but did not fully offset the increased mortality risk. Only maintaining healthy adiposity levels significantly reduced mortality risk. In joint change analyses, lower mortality risk was primarily driven by increases in physical activity, regardless of adiposity changes. Decreasing adiposity attenuated the negative associations of decreased physical activity. The combination of increased physical activity and decreased adiposity was associated with the lowest mortality risk across all three outcomes. The interaction between physical activity and adiposity changes was significant for all three outcomes (p<0.05). E-values indicated that substantial unmeasured confounding would be needed to explain away the observed associations. Negative control analyses suggested minimal confounding effects.

This study provides strong evidence that increasing physical activity is strongly associated with lower mortality risk, regardless of adiposity status. The benefits of increased physical activity appear to be more immediate and pronounced than those of adiposity reduction. Decreasing adiposity attenuated mortality risk but did not eliminate it, particularly when combined with decreased physical activity. These findings highlight the importance of promoting physical activity and maintaining healthy adiposity levels to reduce mortality risk, particularly from CVD and cancer. The results contrast with some previous studies which assumed constant or consistently ordered risk factor levels over time. The study supports public health guidelines promoting physical activity and emphasizes its critical role in preventing premature mortality. The lack of complete elimination of risk with decreasing adiposity, particularly in an Asian population, warrants further investigation into appropriate adiposity cut-offs for this population group.

This large-scale longitudinal study of an Asian population demonstrates the substantial benefits of increased physical activity in reducing all-cause, CVD, and cancer mortality. While decreasing adiposity offers some attenuation of risk, the immediate and significant impact of increasing physical activity underscores its crucial role in improving health outcomes. Future research should focus on refining adiposity cut-offs for Asian populations and investigating the interplay between physical activity, adiposity, and other risk factors in a wider range of ethnic groups.

As an observational study, this research cannot definitively establish causality. Although efforts were made to minimize confounding and reverse causation, residual confounding remains a possibility. The use of self-reported physical activity data may lead to some underestimation of the true association. The study also relied on a specific operationalization of change based on guidelines, which may not capture all relevant patterns of change. Future research incorporating objective measures of physical activity and more detailed assessment of adiposity changes is warranted.