Introduction

The prevalence of obesity among adults is rising globally1,2. By 2030, it is projected that 3 out of every 10 Brazilian adults will be affected by obesity1. If this trend continues, it could lead to approximately 808,600 deaths from chronic non-communicable diseases associated with excess weight3. Additionally, São Paulo, one of the largest cities in Latin America, has seen a concerning rise in obesity prevalence in recent years4,5,6,7. The prevalence of obesity (defined as a body mass index [BMI] of ≥ 30 kg/m2) increased from 11.1% in 2006 to 19.8% in 20196. Obesity can lead to an increased risk of developing other diseases, including cardiovascular disease8, type 2 diabetes mellitus8 and cancer9.

Non-pharmacological strategies to reduce obesity include engaging in moderate to vigorous physical activity (MVPA) and minimizing sedentary behavior10. According to the World Health Organization (WHO), adults should aim for at least 150 min/week of MVPA to achieve significant health benefits11. However, global estimates show that physical inactivity (< 150 min/week of MVPA) affects 30.3% of adults12. In São Paulo, the prevalence of physical inactivity is even higher, at 38.3%13 .

Several studies on physical activity and obesity have examined only physical activity or sedentary time to predict obesity outcomes, while adjusting for covariates such as age, gender, smoking, and others14,15. However, since a day consists of only 24 h, individuals allocate this time across various behaviors, including sleep, sedentary activities in different contexts (e.g., work, screen time, studying, driving)16, and physical activities across different domains (e.g., work, leisure, transportation)17,18 and intensities (e.g., light, moderate, vigorous)19,20. Due to the finite nature of the 24-hour day, increasing time spent on one behavior (e.g., screen time) often results in a decrease in time allocated to another behavior (e.g., light-intensity physical activity)21.

In this context, isotemporal substitution analysis examines the theoretical effect of “substituting” equal amounts of time spent on one behavior (e.g., 60 min of screen time) with another behavior (e.g., 60 min of light physical activity) on health outcomes22,23,24. Additionally, isotemporal substitution analysis assumes that time-use behaviors, such as sedentary time and physical activity, are interdependent. This approach suggests that reallocating time from one behavior inevitably results in a reduction of another22,23,24.

A growing body of evidence shows that replacing various durations of sedentary behavior (10 min/day or 60 min/day) with physical activity is associated with improvements in various health markers25,26,27,28,29,30,31,32. For instance, Cao and colleagues demonstrated that substituting 60 min/day of sedentary time with 60 min/day of light or vigorous-intensity physical activity was associated with a reduced risk of COPD, depression, and diabetes25. Additionally, another study examining the substitution of 15, 30, 45, and 60 min/day of sedentary time for light and MVPA found negative associations with fat mass (assessed by bioelectrical impedance) and BMI in European adults33. However, most of these studies have been conducted in high-income33,34,35,36.

In contrast, there is limited evidence on the use of isotemporal substitution for obesity markers in Brazil, with most studies focusing on children and adolescents37,38,39. Furthermore, a study that analyzed isotemporal substitution in obesity among Brazilian adults considered only 10 min/day for the analysis40. Therefore, investigating the association between substituting physical activity and sedentary time for different periods could provide further insights into the ideal balance between these behaviors for improving body composition. Thus, the aim of this study was to evaluate the associations between substituting 10, 30 and 60 min/day of physical activity and sitting time with BMI, waist circumference and body fat among workers in São Paulo.

Methods

Study design and sample

This cross-sectional study was conducted among employees of the Regional Health Department (RHD) aged 18 to 73, residing in various regions of the state of São Paulo. Data collection took place between July and September 2018 at each RHD for convenience.

This study is part of the “Agita São Paulo” program, a state initiative aimed at combating sedentary lifestyles and physical inactivity in São Paulo41. Details of the “Agita São Paulo” program have been published previously41,42,43. As an extension of the program’s efforts, surveys are periodically conducted to assess levels of physical activity, sedentary behavior, and health-related factors (e.g., obesity, hypertension, functionality, and self-perceived health) among the population of São Paulo44.

The state of São Paulo is divided into 17 RHD, and 15 15 of these departments from different regions participated in the study. A total of 1,056 workers met the eligibility criteria: (a) aged 18 years or older, (b) employed by an RHD, and (c) without clinical or functional limitations. Individuals with missing or inconsistent data on physical activity (n = 176), sitting time (n = 134), BMI (n = 74), waist circumference (n = 70), body fat (n = 76), age (n = 14), socioeconomic status (n = 51), marital status (n = 36), and educational level (n = 31) were excluded from the analysis. The final sample consisted of 394 adults (76.6% women).

Physical activity and sitting time

Participants reported their levels of physical activity and sitting time using the short version of the International Physical Activity Questionnaire (IPAQ), which has been validated in several countries, including Brazil45,46.

The data collection team instructed participants to report the frequency (days per week) and duration (≥ 10 min performed continuously) of walking, moderate physical activity (MPA), and vigorous physical activity (VPA) performed in the previous week. Average walking time, MPA, and VPA were calculated by multiplying frequency by duration and dividing by 7. IPAQ data were expressed in min/day25,47. Additionally, data where the total sum of walking time, MPA, and VPA exceeded 960 min/day were excluded from the analysis. Finally, data with at least 10 min/day of reported activity were considered valid for analysis48.

Participants also self-reported sitting time spent in various contexts (home, work, free time, studying, lying down watching television) during the last seven days. For the analysis, we used substitutions of 10, 30, and 60 min/day of walking, MPA, VPA, and sitting time. These periods were specifically chosen based on previous studies that have demonstrated significant associations with health outcomes, including type 2 diabetes, metabolic syndrome, glycated hemoglobin, and waist circumference49,50.

Obesity indicators

BMI, waist circumference, and body fat were used as indicators of obesity. Height (cm) was measured using a Seca 213 portable stadiometer (Seca 213®, Hamburg, Germany) with an accuracy of 0.5 cm. Participants were instructed to stand barefoot in an upright position, aligned with the Frankfurt plane.

Body weight (kg) was measured using a Tanita SC-240 bioelectrical impedance device (TANITA Corporation, Japan), which has a maximum capacity of 200 kg. BMI (kg/m²) was calculated by dividing body weight (kg) by height (m²).

Waist circumference (cm) was measured using an inelastic tape measure (Sanny, São Paulo, Brazil) with an accuracy of 0.1 cm. The tape was placed horizontally at the level of the umbilical scar while the participant stood in an upright position. Three measurements were taken, and the average was calculated.

Body fat (kg) was measured using bioelectrical impedance with the Tanita SC-240 portable body composition analyzer (TANITA Corporation, Japan). The analyzer uses four electrodes placed on the feet, with a frequency of 50 kHz and an electric current of 90 milliamps. Participants’ height, age, and gender were entered into the scale’s software before measurement. Once the information was input, participants were instructed to stand straight in the Frankfurt plane. Bioelectrical impedance analysis has demonstrated acceptable validity and reproducibility in adults when compared to dual-energy X-ray absorptiometry51.

Correlates

A standardized questionnaire was used across all departments during the interviews to collect sociodemographic data. Participants were categorized by age (< 65 and ≥ 65 years), sex (male and female), and marital status (single, married, divorced, and widowed). In addition, participants were classified by socioeconomic status (low, medium, and high) according to the Brazilian income classification index52. Educational level was categorized as low (up to elementary school), medium (completed elementary school), and high (university degree, master’s degree, PhD, or post-doctorate).

Statistical analyses

The Kolmogorov-Smirnov test was used to assess the normality of the data distribution. Frequency and percentage (%) were used to describe categorical variables. Given the non-normal distribution of data on physical activity, sitting time, and obesity indicators, continuous variables were reported as medians and interquartile ranges (25th and 75th percentiles).

Multivariate linear regression analysis (β coefficient) and 95% confidence intervals (95% CI) were used to assess the associations between different durations (10, 30, and 60 min/day) of sitting time, walking, MPA, and VPA with obesity indicators (BMI, waist circumference, and body fat). Due to the non-parametric distribution of sitting time, physical activity variables, and obesity indicators, we transformed the non-parametric data using a log10 transformation.

Linear models for single activities, partitioning, and isotemporal substitution were employed, following the method proposed by Merkay and colleagues22. The single-activity model assessed the associations of each behavior independently, adjusting for covariates, without including other behaviors in the statistical model. In contrast, the partition model considers all activities that contribute to the total time spent and evaluates the effect of each time-use behavior on the outcome variable independently. However, the partition model is not isotemporal, as it does not include the total time variable (the sum of sitting time, walking, MPA, and VPA) in the analysis.

The isotemporal model, however, assumes the effect of substituting one specific behavior for another within the same time period. For instance, when analyzing the effect of substituting 60 min/day of sitting time with 60 mi/day of walking or VPA on BMI, sitting time is excluded from the analysis. The coefficients generated for walking and VPA represent the effect of replacing sitting time with these activities, while keeping covariates and total time within the analysis.

The single-activity, partition, and isotemporal models were applied to the periods of 10, 30, and 60 min per day of sitting time and physical activity to assess obesity outcomes. All analyses were adjusted for age, sex, marital status, socioeconomic status, and educational level. A significance level of p < 0.05 was considered, and analyses were conducted using SPSS, version 24 (SPSS Inc., IBM Corp., Armonk, Nova York, NY, EUA).

Ethics approval and consent to participate

All aspects of the study were in accordance with the Declaration of Helsinki. The study was approved by the Ethics Committee of Universidade Municipal de São Caetano do Sul under protocol number 2.531.002. All participants provided written informed consent or assent before participating in the study. To ensure confidentiality, participants’ data was anonymized using numeric identification codes instead of names.

Results

The characteristics of the participants are summarized in Table 1. Overall, the sample was predominantly female (76.6%), with a medium socioeconomic level (59.1%), married (52.5%), and highly educated (63.2%). Participants self-reported a median of 150 min/day of walking (IQR: 60.0 − 360.0), 180 min/day of MPA (IQR: 78.7 − 450.0), and 60 min/day of VPA (IQR: 0.0 − 240.0). In contrast, the median sitting time was 445.7 min/day (IQR: 317.1 − 565.7).

Table 1 Descriptive characteristics of the participants (median [P25-P75] or n [%]).
Full size table

Table 2 presents the results of the single and partial activity analyses examining the associations between different patterns of physical activity and sitting time with obesity indicators. In the single activity models, both 10 and 60 min/day of walking were negatively associated with waist circumference (β: -0.125; 95% CI: -0.025; -0.004; β: -0.135; 95% CI: -0.026; -0.005, respectively) and body fat (β: -0.101; 95% CI: -0.067; -0.001; β: -0.111; 95% CI: -0.068; -0.003, respectively). No significant associations were found between 30 min/day of physical activity or sitting time and obesity indicators. In the partitioned model, only 10 min/day of walking was negatively associated with waist circumference (β: -0.129; 95% CI: -0.028; -0.001).

Table 2 Single and partition analysis of the association between patterns (10, 30, 60 min/day) of physical activity and sitting time with obesity indicators in workers.
Full size table

No statistically significant association was found between substituting 10 min/day of physical activity or sitting time with BMI (Table 3). On the other hand, substituting 10 min/day of MPA and 10 min/day of VPA for 10 min/day of walking (β: -0.176; 95% CI: -0.037; -0.002; β: -0.183; 95% CI: -0.037; -0.004) was associated with waist circumference. Additionally, substituting 10 min/day of sitting time for 10 min/day of walking (β: -0.216; 95% CI: -0.041; -0.007) was negatively associated with waist circumference. Similarly, substituting 10 min/day of VPA with 10 min/day of walking (β: -0.172; 95% CI: -0.106; -0.005) was associated with a decrease in body fat. Furthermore, substituting 10 min/day of sitting time for 10 min/day of walking (β: -0.203; 95% CI: -0.119; -0.012) or MPA (β: -0.208; 95% CI: -0.127; -0.008) was negatively associated with body fat (Table 3).

Table 3 Isotemporal substitution for 10 min/day activities and obesity indicators in workers.
Full size table

No associations were found between substituting 30 min/day of physical activity or sitting time with BMI, waist circumference, or body fat (Table 4).

Table 4 Isotemporal substitution for 30 min/day activities and obesity indicators in workers.
Full size table

Table 5 shows the associations between the substitution of 60 min/day of physical activity and sitting time with indicators of obesity. Substituting 60 min/day of VPA for 60 min/day of walking (β: -0.151; 95% CI: -0.033; -0.001) was negatively associated with waist circumference. In addition, substituting 60 min/day of sitting time for MPA (β: -0.206; 95% CI: -0.123; -0.008) was negatively associated with body fat.

Table 5 Isotemporal substitution for 60 min/day activities and obesity indicators in workers.
Full size table

Discussion

The present study examined the associations between substituting 10, 30 and 60 min/day of physical activity and sitting time with BMI, waist circumference and body fat among workers in São Paulo. Overall, the results of this study suggest that substituting different patterns of sitting time with various forms of physical activity may be associated with a reduction in obesity indicators in São Paulo workers.

Both 10 min/day and 60 min/day of walking were associated with lower waist circumference and body fat in the single and partition models, but no associations were found with higher-intensity activities. These findings are consistent with studies in high- and middle-income countries33,40,53,54, which have shown that light-intensity physical activity (10 and 60 min/day), as measured by accelerometry and self-report, is negatively associated with BMI, body fat, and waist circumference. Walking, in particular, can be a more accessible form of physical activity, as it requires no prior experience and involves no financial cost55. Furthermore, the potential benefits of light physical activity on obesity and cardiovascular risk, as well as its significance for public health, have already been well established19,56.

In the isotemporal substitution model, beneficial associations were found when substituting at least 10 min/day of sitting time with 10 min/day of walking and 10 min/day of MPA, which were linked to lower waist circumference and body fat, respectively. A study by Gonze and colleagues40, conducted with 780 Brazilian adults (493 women), found that substituting 10 min/day of sedentary time with 10 min/day of light physical activity or MVPA was associated with reductions in BMI (β = -0.104) and fat percentage (β = -0.383). Sitting time, a form of sedentary behavior, is strongly associated with an increase in waist circumference, visceral adipose tissue, and body fat, which in turn are linked to the development of certain cancers (including digestive system and colorectal cancers) and all-cause mortality14,57,58. Our findings suggest that reducing sitting time and increasing physical activity levels can positively impact obesity indicators and may help reduce the risk of adverse health outcomes.

Curiously, our results showed that substituting 10 min/day of MPA and VPA with 10 min/day of walking was associated with a reduction in obesity indicators (waist circumference and body fat). These findings are not consistent with the WHO’s recommendation of at least 60 min/day of MVPA11. This discrepancy may be due to the discomfort higher-intensity activities may cause compared to light physical activity59,60. Additionally, walking is a natural activity that helps break sedentary behavior, improves lipid metabolism, and can be sustained for long periods61,62,63. However, it is important to highlight the potential benefits of higher-intensity activities, such as greater energy expenditure and improved health markers, which have been widely demonstrated in the literature34,39,64,65.

We found no association between substituting 30 min/day of physical activity and sitting time with obesity indicators. These results contrast with existing literature49,50,66,67. Additionally, a systematic review study49 examining the effect of isotemporal substitution on various obesity indicators found that some studies included in the review analyzed the effects of 30 min/day and reported associations with different obesity indicators, such as BMI, waist circumference, and body fat percentage. We hypothesize that methodological differences between our study and the cited studies may explain the discrepancies. For example, the sample size in our study was smaller than in other studies50,66. Additionally, objective methods, such as accelerometry, were used to measure physical activity in these studies50,67, which was not feasible in our study. While questionnaires are commonly used in large-scale studies due to their practicality, accelerometry helps reduce memory bias by directly measuring physical activity and sedentary time without relying on self-reports68,69. Finally, the studies mentioned above50 estimated body composition objectively, which was not possible in our study. These differences in measurement methods could lead to variations in body composition estimates and impact the analysis70.

Finally, a previous population-based study found that substituting 60 min/day of MVPA for sedentary time (assessed using accelerometry) was negatively associated with fat mass and body fat, as measured by bioelectrical impedance33. Similarly, Wang and colleagues observed beneficial effects on overweight and obesity when sedentary time was replaced with MVPA, as assessed by the IPAQ34. These findings support our results and emphasize the need for effective strategies to encourage reallocating sedentary time to more active behaviors.

Educating the public about the distribution of time spent on various behaviors can have positive implications for public health71. Given the challenges middle-aged and older adults face in performing VPA for extended periods72,73, replacing sedentary time with at least 10 min/day of light physical activity could be a useful starting point for reducing overweight and obesity.

This study has several limitations. Due to its cross-sectional design, we cannot establish causal relationships. Additionally, our results may not be generalizable to other populations, as the sample consisted mainly of Brazilian workers with a high socioeconomic status and education level. We also could not assess the amount of sleep as part of the full 24-hour movement cycle. While the IPAQ’s validity has been established, memory bias may have led participants to over report physical activity levels and underreport sedentary time74. Body fat was measured indirectly using bioelectrical impedance, so the estimates are prone to error75,76. Finally, we were unable to account for other potential confounders, such as race/ethnicity, household income, energy consumption, and the built environment.

Conclusion

The results of this study suggest that substituting 10 min/day and 60 min/day of sitting time for 60 min/day of walking or MPA is beneficial for reducing waist circumference and body fat. These findings align with existing literature on the positive effects of walking on body composition and support lifestyle interventions that promote replacing sitting time with physical activity to reduce and prevent obesity. Furthermore, our results underscore the need for longitudinal studies to evaluate the long-term effects of these time-use behavior substitutions on obesity indicators.