Individualized walking cadence thresholds of moderate to vigorous intensity for Taiwanese obese young adults

Abstract

Background

Given the altered gait patterns and metabolic demands in obese individuals, population-specific cadence thresholds (steps per minute) are essential for accurate intensity classification and effective exercise prescription. This study examined the relationship between body composition and Moderate-to-Vigorous Physical Activity (MVPA) cadence in obese individuals and establish cadence thresholds for moderate-intensity physical activity (MPA) (3 metabolic equivalents [METs]), MPA_young (moderate-intensity physical activity for young adults, 4.8 METs), and vigorous physical activity (VPA) (6 METs).

Methods

A total of 48 obese young adults participated in this study. The VO₂ was collected at rest in seated position for 10 min, followed by an incremental walking exercise at 3.2, 4.0, 4.8, 5.6, and 6.4 km/h with 5 min duration in each stage. Walking cadence and oxygen consumption were recorded and converted to steps per minute and METs.

Results

Body fat significantly related to 3 METs cadence in male (R² = 0.156, p < .05). The receiver operating characteristic (ROC) models demonstrated a good discrimination (area under the curve [AUC] = 0.85–0.88 in males, 0.82–0.89 in females). The optimal cadence thresholds for males were 114 steps/min for MPA (positive predictive value [PPV]: 96.97%, negative predictive value [NPV]: 47.30%), 119 steps/min for MPA_young (PPV: 50.94%, NPV: 95.40%), and 124 steps/min for VPA (PPV: 21.21%, NPV: 99.07%). In females, the optimal cadences were 115 steps/min for MPA (PPV: 98.08%, NPV: 36.96%), 125 steps/min for MPA_young (PPV: 68.75%, NPV: 93.94%), and 131 steps/min for VPA (PPV: 38.10%, NPV: 98.70%).

Conclusion

The effect of body fat percentage on gait adjustment mechanisms during MPA are different between genders in Taiwanese young adults. A practical cadence target range is 119–124 steps/min for males and 125–131 steps/min for females, based on the 4.8- to 6-METs thresholds; cadences within these ranges increase the likelihood of exceeding 3 METs.

Trial registration

ClinicalTrials.gov, number NCT06883253 (Retrospectively registered on March 19, 2025, for a study conducted between November 15, 2017, and November 8, 2018).

Background

Obesity has become one of the most significant global health challenges, acting as a major risk factor for various non-communicable diseases [1]. According to the World Health Organization (WHO), approximately 43% of adults aged 18 years and older were classified as overweight in 2022, with 16% categorized as obese [2]. The COVID-19 pandemic further underscored the urgency of addressing obesity, as it revealed strong associations between obesity and severe COVID-19 outcomes, along with comorbidities such as hypertension and type 2 diabetes [3, 4]. The pandemic significantly disrupted physical activity behaviors, with studies reporting a 33.5% decline in physical activity levels and a 28.6% increase in sedentary behavior during quarantine periods. These behavioral changes were predicted to raise the prevalence of type 2 diabetes from 7.2% to 9.6% and elevate all-cause mortality from 9.4% to 12.5% [5]. Recent evidence also indicates a sustained increase in obesity rates in the post-pandemic period, highlighting its persistent impact on public health [6, 7]. Given these developments, there is an urgent need for proactive and effective strategies to promote physical activity as a key intervention for managing overweight and obesity. Such initiatives are critical not only for reducing the growing burden of obesity-related diseases but also for enhancing overall health outcomes in the post-pandemic era.

However, the prevention and treatment of obesity emphasize on reducing sedentary behaviors [8] and promoting physical activity [9], which should be prioritized as a primary strategy. According to the American College of Sports Medicine (ACSM) guidelines for exercise testing and prescription, overweight and obese adults are recommended to engage in 250 to 300 min of MVPA per week to facilitate and sustain long-term weight loss [10]. Recent studies have reported a significant decline in time spent on MVPA during the COVID-19 pandemic [11]. Consequently, there is an urgent need in the post-pandemic phase to actively restore and promote physical activity among overweight and obese individuals [3, 5, 12–15]. Physical activity intensity can be effectively enhanced by monitoring cadence. Studies indicate that cadence is a reliable predictor of physical activity intensities (measured in METs), with strong validity (R² = 0.79–0.91) [16, 17]. Furthermore, research has identified 100 steps/min and 130 steps/min as reasonable cadence thresholds for MVPA in adults [17–21]. Additionally, evidence suggests that individual differences significantly influence cadence [22]. As a result, recent studies have increasingly focused on establishing personalized MVPA cadence thresholds for specific populations, such as children, older adults, pregnant women, and individuals with Parkinson’s disease [23–26].

Previous research has identified factors influencing MVPA cadence thresholds, including height, leg length, body mass index (BMI), and energy expenditure [18–20], with height and leg length considered the most critical determinants. Studies have confirmed that differences in physical morphology and physiological metabolism exist across regions, ethnicities, and populations [27]. Therefore, establishing MVPA cadence thresholds tailored to specific demographic characteristics is essential. While previous research has established MVPA cadence thresholds for normal-weight young adults [28], evidence addressing these thresholds for overweight and obese young populations remain limited. Furthermore, the relationship between body composition variables and MPA/VPA cadence remain relatively limited and requires further investigation.

Regardless of participant characteristics, most studies have established MVPA cadence thresholds based on the lower limit of MPA at ≥ 3 METs [22]. According to the ACSM guidelines for exercise testing and prescription, MPA for young adults is defined as ≥ 4.8 METs [10]. Bae et al. found that the cadence corresponding to MPA during normal walking in healthy young males was approximately 128.9 steps/min [29], significantly higher than the widely proposed 100 steps/min threshold for 3 METs in earlier studies [17–21]. This discrepancy indicates that using 3 METs as a benchmark for establishing cadence thresholds may underestimate MPA intensity in real-world contexts. Bae et al. further argued that reliance on 3 METs as a standard for physical activity guidance has limitations due to variability in individual fitness levels [29]. Elevating the threshold to 4.8 METs could ensure that physical activity intensity is not underestimated in practical applications. Previous studies have confirmed that obesity affects gait efficiency and energy expenditure during walking [30, 31]. Therefore, it is important to establish cadence thresholds tailored to the characteristics of obese young adults. Based on these findings, this study aims to explore the relationship between body composition variables and MVPA cadence and to establish appropriate cadence thresholds for obese young adults in Taiwan. This approach seeks to provide more personalized physical activity recommendations to better address the specific needs of the obese population.

Methods

Participant characteristics

This study recruited 48 obese young adults who did not regularly engage in exercise. Participants were classified as obese based solely on the criteria of the Ministry of Health and Welfare in Taiwan, which defines obesity as a BMI ≥ 27 kg/m² [32]. After reviewing sample sizes reported in recent cadence-threshold studies [25, 26], we conducted an a-priori power analysis using G*Power 3.1.9.7, assuming a medium effect size (α = 0.05, 1-β = 0.80, f = 0.25, five repeated measures), which indicated a minimum of 21 participants. Therefore, 48 obese young adults were enrolled to ensure sufficient statistical power. Exclusion criteria included a history of smoking, hypertension, diabetes, cardiovascular or respiratory diseases, neuromuscular or musculoskeletal disorders, or any condition deemed unsuitable for exercise by a physician (e.g., severe joint disorders, recent orthopedic surgeries, or unstable medical conditions). The study protocol was approved by the Fu Jen University Institutional Review Board for Human Research (Approval Number: C105137). All participants were fully informed about the study procedures and provided written informed consent prior to participation.

Anthropometrics and body composition measurements

Participants arrived at the lab fasted for a minimum of 3 h, abstained from exercise for 12 h, avoided caffeine for 6 h, and refrained from alcohol consumption for 12 h. Data collection for each subject occurred during the morning hours, as close as possible to awakening from sleep (i.e., from 7:00 AM to 10:00 AM). Height was measured using a digital stadiometer accuracy nearest to ± 0.1 cm (Kongho-Super View, Kong Ho Instruments Co., Ltd., New Taipei, Taiwan). Body composition was assessed using bioelectrical impedance analysis (BIA) with the InBody 720 device (Biospace Co., Ltd., Seoul, South Korea). Skin resistance was measured through an eight-point tactile-electrode system at different frequencies (1 kHz, 5 kHz, 50 kHz, 250 kHz, 500 kHz, 1 MHz). Lean body mass and fat mass were determined from the impedance values using the manufacturer’s regression equations, based on the principle that different tissues have varying water content and, consequently, different electrical conductivity [33]. The components used in further analysis included weight, BMI, body fat mass (BFM), body fat percentage (FAT), visceral fat area (VFA), and skeletal muscle mass (SMM).

Incremental submaximal exercise testing protocol

Participants were instructed to refrain from vigorous exercise, smoking, and the consumption of caffeine-containing foods and beverages for 24 h prior to the experiment. During the test, each participant was equipped with a gas analyzer (MetaMax 3B, Cortex, Leipzig, Germany) and a heart rate monitor (Polar H10, Electro Oy, Kempele, Finland). The protocol began with a 10-minute seated rest period to measure resting oxygen consumption. Subsequently, participants completed an incremental submaximal treadmill walking test consisting of five stages, with speeds set at 3.2, 4.0, 4.8, 5.6, and 6.4 km/h. Each stage lasted 5 min, and participants were instructed to maintain a natural walking pattern throughout the test. Oxygen consumption was continuously recorded during the exercise phases using the gas analyzer. The total step count for each stage was recorded with a digital camera (Canon EOS 700D, Canon, Lake Success, NY, USA) positioned laterally at hip height and set to 1280 × 720 pixels at 50 fps to ensure accurate step identification.

Data collection and processing

Oxygen consumption and step count measurements data were extracted from the stabilized period (3rd to 4th minute) of each stage in the submaximal exercise test. This period was selected to ensure that data were collected during a steady-state phase, after initial adaptation. The mean oxygen consumption of each stage was converted into METs as the physical activity intensity. Oxygen consumption were divided by 3.5 mL/kg/min for males and 3.81 mL/kg/min for females. According to the ACSM’s Guidelines for Exercise Testing and Prescription, Physical activity intensity thresholds were defined as follows: MPA for all age groups was ≥ 3 METs, MPA for young adults aged 20–39 (MPA_young) was ≥ 4.8 METs, and VPA was ≥ 6 METs [10]. Two trained researchers independently reviewed all videos and manually counted steps during the stabilized period (minutes 3–4 of each stage); a step was defined as one complete left–right footfall cycle. When the counts differed by more than one step, a third researcher adjudicated the final value. The verified step counts directly represented the cadence for each stage, expressed as steps per minute. Inter-rater agreement across all videos was excellent, with an ICC (2, k) = 0.993 (95% CI 0.991–0.994, p <.001).

Analytic sample

All 48 participants provided data across five treadmill walking stages. Two female participants did not complete the final stage (6.4 km/h) due to physical discomfort, and those incomplete observations were excluded. As a result, the final analytic sample consisted of 140 stage-level observations for males (28 participants × 5 stages) and 98 for females (20 participants × 5 stages, minus 2 observations), totaling 238 observations.

Statistical analyses

All data were analyzed using the Statistical Package for the Social Sciences (SPSS, Version 29; SPSS Inc., Chicago, IL, USA). Inter-rater reliability of the manual step counts was assessed using an intraclass correlation coefficient, ICC (2, k), based on a two-way random-effects model with absolute agreement. An independent samples t-test was conducted to compare gender differences across all variables. The Pearson product-moment correlation was employed to assess the associations among body composition variables, cadence, and METs. Based on previous studies that have established linear relationships between cadence and METs [16, 17], linear regression was used to develop individualized METs–cadence equations for each participant, which were then applied to estimate cadence values corresponding to 3, 4.8, and 6 METs. Additionally, multiple stepwise regression analyses were performed to examine whether, and which, body composition variables significantly influenced the cadence corresponding to 3, 4.8, and 6 METs. Furthermore, Receiver Operating Characteristic (ROC) curve analyses were conducted to identify optimal cadence thresholds for classifying MPA, MPA_young, and VPA. Cadence (steps/min) served as the predictor variable in all ROC models. For each intensity classification, METs values were dichotomized to create binary classification outcomes: < 3 METs = 0 and ≥ 3 METs = 1 for MPA; < 4.8 METs = 0 and ≥ 4.8 METs = 1 for MPA_young; < 6 METs = 0 and ≥ 6 METs = 1 for VPA. Each treadmill stage was treated as an independent observation. The optimal cadence value was defined as the point that maximized the sum of sensitivity (Se) and specificity (sp), equivalent to maximizing Youden’s index [34]. To further assess the accuracy of each model, To further assess the accuracy of each model, performance indicators including AUC, Se (correct identification of individuals who reached the intensity), Sp (correct exclusion of those who did not), PPV, NPV, overall accuracy (ACC), and AUC. Based on previously standards, AUC values were interpreted as excellent (≥ 0.90), good (0.80–0.89), fair (0.70–0.79), and poor (< 0.70) [35]. Statistical significance was set at α = 0.05.

Results

The anthropometric characteristics of overweight and obese young adults in this study are presented in Table 1. Body composition analysis revealed that males had significantly greater height (cm), weight (kg), and SMM (kg) compared to females (p <.001), whereas females exhibited significantly higher FAT (%) and BFM (kg) than males (p <.05). Further analysis of gender differences in VO₂, METs, and cadence across various speeds revealed no significant differences in VO₂ or METs between males and females. (p >.05). However, females demonstrated significantly higher cadence than males at speeds of 4.0, 4.8, 5.6, and 6.4 km/h (p <.05). Detailed statistical results are provided in Table 2.

Table 1.

Anthropometry and body composition differences between men and women

Variables	All (N = 48)	Male (N = 28)	Female (N = 20)	p
Age (year)	20 ± 1.4	20.1 ± 1.6	20 ± 1	-
Height (cm)	166.4 ± 8.4	171.4 ± 6.6	159.4 ± 4.8	< 0.001
Weight (kg)	86 ± 12.5	91.1 ± 11.4	78.9 ± 10.5	< 0.001
BMI (kg/m2)	31 ± 3	31 ± 2.7	31 ± 3.4	0.971
BFM (kg)	31.2 ± 7.3	28.9 ± 5.9	34.3 ± 8.0	0.010
FAT (%)	36.4 ± 7.6	31.7 ± 4.8	43.1 ± 5.4	< 0.001
VFA (cm2)	104.4 ± 30.9	102 ± 33.9	107.7 ± 26.5	0.539
SMM (kg)	30.7 ± 6.8	35.2 ± 5	24.4 ± 2.6	< 0.001
VO2 rest (ml−1∙kg−1∙min−1)	3.91 ± 0.59	4.0 ± 0.5	3.8 ± 0.7	0.337

p <.05: significantly different from women and men

Table 2.

Differences in measured variables at each speed

Speeds	Variables	ALL	Male	Female	p
(km/h)		M ± SD	M ± SD	M ± SD
3.2	VO2 (ml−1∙kg−1∙min−1)	11.11 ± 1	11.26 ± 1.05	10.90 ± 0.92	0.211
	METs	2.84 ± 0.26	2.87 ± 0.37	2.91 ± 0.38	0.684
	Cadence (step/min)	100.57 ± 7.86	99.59 ± 8.68	101.95 ± 6.52	0.310
4.0	VO2 (ml−1∙kg−1∙min−1)	12.24 ± 1.06	12.20 ± 1.15	12.29 ± 0.95	0.784
	METs	3.13 ± 0.27	3.11 ± 0.41	3.28 ± 0.46	0.164
	Cadence (step/min)	107.74 ± 6.77	105.88 ± 6.23	110.35 ± 6.77*	0.022
4.8	VO2 (ml−1∙kg−1∙min−1)	14.25 ± 1.28	14.21 ± 1.36	14.31 ± 1.21	0.788
	METs	3.64 ± 0.33	3.61 ± 0.48	3.83 ± 0.57	0.165
	Cadence (step/min)	114.69 ± 5.75	112.89 ± 5.97	117.2 ± 4.45*	0.009
5.6	VO2 (ml−1∙kg−1∙min−1)	17.40 ± 1.52	17.23 ± 1.71	17.65 ± 1.19	0.348
	METs	4.45 ± 0.39	4.40 ± 0.71	4.37 ± 0.69	0.112
	Cadence (step/min)	123.64 ± 8.58	120.91 ± 7.24	127.45 ± 9.02*	0.008
6.4	VO2 (ml−1∙kg−1∙min−1)	22.09 ± 2.28	21.6 ± 2.54	22.86 ± 1.54	0.067
	METs	5.65 ± 0.58	5.51 ± 0.91	6.03 ± 0.89	0.067
	Cadence (step/min)	133.67 ± 9.68	129.98 ± 7.49	139.42 ± 10.09*	< 0.001

*p <.05: significantly different from male and female

Correlation analysis revealed a strong relationship between METs and cadence for both males and females (r =.726, 0.737, p <.001). Additionally, a significant positive correlation was observed between males’ FAT and 3 METs cadence (r =.395, p <.005), while no significant correlations were identified among females. Detailed statistical results are provided in Table 3 (males) and Table 4 (females). Stepwise regression analysis indicated that FAT accounted for 16% of the variance in 3 METs cadence (R² = 0.156, β = 0.395, p =.037).

Table 3.

The correlation between body composition variables and Cadence during MPA, MPA_young, and VPA in males

Variables	Height	Weight	BMI	BFM	FAT	VFA	SMM	3 METs Cadence	4.8 METs Cadence	6 METs Cadence
Height	--
Weight	0.683**	--
BMI	0.069	0.773**	--
BFM	0.231	0.701**	0.753**	--
FAT	− 0.273	0.078	0.335	0.761**	--
VFA	0.332	0.675**	0.620**	0.771**	0.466*	--
SMM	0.742**	0.858**	0.534**	0.238	− 0.441*	0.365	--
3 METs Cadence	− 0.355	− 0.144	0.110	0.184	0.395*	0.185	− 0.320	--
4.8 METs Cadence	− 0.243	− 0.008	0.209	0.148	0.207	0.199	− 0.099	0.818**	--
6 METs Cadence	− 0.190	0.036	0.224	0.126	0.133	0.189	− 0.022	0.699**	0.983**	--

BMI Body mass index, BFM Body fat mass, FAT Body fat percentage, VFA Visceral fat area, SMM Skeletal muscle mass

3 METs Cadence: cadence threshold corresponding to moderate-intensity physical activity (MPA)

4.8 METs Cadence: cadence threshold for moderate-intensity physical activity in young adults (MPA_young)

6 METs Cadence: cadence threshold for vigorous-intensity physical activity(VPA)

* p <.05

** p <.01

Table 4.

The correlation between body composition variables and Cadence during MPA, MPA_young, and VPA in females

Variable	Height	Weight	BMI	BFM	FAT	VFA	SMM	3 METs Cadence	4.8 METs Cadence	6 METs Cadence
Height	--
Weight	0.594**	--
BMI	0.163	0.890**	--
BFM	0.308	0.915**	0.949**	--
FAT	− 0.053	0.616**	0.789**	0.876**	--
VFA	0.174	0.721**	0.802**	0.780**	0.683**	--
SMM	0.812**	0.650**	0.338	0.291	− 0.187	0.263	--
3 METs Cadence	− 0.063	0.027	0.071	− 0.048	− 0.154	− 0.021	0.171	--
4.8 METs Cadence	− 0.191	− 0.027	0.081	− 0.036	− 0.057	− 0.003	0.023	0.931**	--
6 METs Cadence	− 0.228	− 0.045	0.082	− 0.031	− 0.023	0.003	− 0.028	0.879**	0.992**	--

BMI Body mass index, BFM Body fat mass, FAT Body fat percentage, VFA Visceral fat area, SMM Skeletal muscle mass

3 METs Cadence: Cadence threshold corresponding to moderate-intensity physical activity (MPA);

4.8 METs Cadence: Cadence threshold for moderate-intensity physical activity in young adults (MPA_young); 

6 METs Cadence: Cadence threshold for vigorous-intensity physical activity (VPA).

** p <.01

The ROC analysis demonstrated the effectiveness of cadence models for 3, 4.8, and 6 METs. For males, the area under the curve (AUC) values were 0.82, 0.89, and 0.89 (p <.001), while for females, the AUC values were 0.85, 0.88, and 0.88 (p <.001). According to established interpretation criteria, AUC values between 0.80 and 0.89 indicate good discrimination ability, suggesting that the cadence thresholds derived from the ROC models are highly reliable in distinguishing individuals who meet versus do not meet the target METs intensities. Using Youden’s index, the optimal cadence thresholds for males were 113.75 steps/min for MPA (Se = 62.14%, Sp = 94.59%), 118.50 steps/min for MPA_young (Se = 87.10%, Sp = 76.15%), and 123.75 steps/min for VPA (Se = 87.50%, Sp = 80.30%). For females, the optimal cadence thresholds were 114.75 steps/min for MPA (Se = 63.75%, Sp = 94.44%), 124.50 steps/min for MPA_young (Se = 84.62%, Sp = 86.11%), and 130.50 steps/min for VPA (Se = 88.89%, Sp = 85.39%). Detailed results of the ROC analysis are provided in Table 5. The distribution of true and false positives and negatives is illustrated in Fig. 1.

Table 5.

Cadence thresholds for MPA, MPA_young, and VPA derived from ROC analyses

Intensities	Sex	Threshold (steps/min)	Se	Sp	PPV	NPV	ACC	AUC (95% CI)
MPA	Male	113.75	62.14%	94.59%	96.97%	47.30%	70.71%	0.85 (0.781–0.909)
	Female	114.75	63.75%	94.44%	98.08%	36.96%	69.39%	0.82 (0.727–0.920)
MPAyoung	Male	118.50	87.10%	76.15%	50.94%	95.40%	78.57%	0.88 (0.825–0.940)
	Female	124.50	84.62%	86.11%	68.75%	93.94%	85.71%	0.89 (0.831–0.957)
VPA	Male	123.75	87.50%	80.30%	21.21%	99.07%	80.71%	0.88 (0.802–0.959)
	Female	130.50	88.89%	85.39%	38.10%	98.70%	85.71%	0.89 (0.823–0.961)

MPAModerate-intensity Physical Activity (3 METs),MPA_young Moderate-intensity Physical Activity for young adults aged 20–39,(4.8 METs) VPA Vigorous-intensity Physical Activity (6 METs), Se Sensitivity, Sp Specificity, PPVPositive predictive value,NPVNegative predictive value, ACC Accuracy, AUC Area under the ROC curve

Fig. 1.

Classification accuracy of cadence thresholds for ROC curve analysis. Classification accuracy of cadence thresholds for ROC analysis in males (A1–A3) and females (B1–B3), at three intensity levels: 3 METs, 4.8 METs, and 6 METs. Each panel presents the optimal cadence cut-point (vertical dashed line) and corresponding MET threshold (horizontal dashed line). Thresholds were identified as follows: 114, 119, and 124 steps/min in males, and 115, 125, and 131 steps/min in females, corresponding to ≥3, ≥4.8, and ≥6 METs, respectively. Each panel includes counts and percentages of true positives, false positives, true negatives, and false negatives based on ROC-derived thresholds. Classification metrics: Males – MPA: AUC = 0.85, Se = 62.14%, Sp = 94.59%; MPA_young: AUC = 0.88, Se = 87.10%, Sp = 76.15%; VPA: AUC = 0.88, Se = 87.50%, Sp = 80.30%. Females – MPA: AUC = 0.82, Se = 63.75%, Sp = 94.44%; MPA_young: AUC = 0.89, Se = 84.62%, Sp = 86.11%; VPA: AUC = 0.89, Se = 88.89%, Sp = 85.39%

Discussion

This study determined cadence thresholds for MPA (3 METs), MPA for young adults (MPA_young, 4.8 METs), and vigorous-intensity physical activity (VPA, 6 METs) in overweight and obese young adults in Taiwan. The identified thresholds were 114, 119, and 124 steps/min for males, and 115, 125, and 131 steps/min for females. The findings also revealed significant gender differences in cadence adjustment mechanisms for MPA, with these adjustments influenced by body composition factors. However, at higher intensity levels, cadence appeared to remain unaffected by such factors.

The results of this study revealed that females exhibited significantly higher cadence than males at walking speeds of 4.0, 4.8, 5.6, and 6.4 km/h. Previous research has identified height and leg length as key determinants of MVPA cadence, with taller individuals typically having longer stride lengths, which reduces the cadence required to maintain the same walking speed [18, 19]. In this study, males were significantly taller than females (171 cm vs. 159 cm). We speculated that females may increase their cadence to compensate for shorter stride lengths and match the walking speed of males. These findings are consistent with previous research and underscore the importance of considering sex and anthropometric differences when establishing universal cadence thresholds. Additionally, it is possible that gait adaptations related to excess body fat, such as shorter stride length or compensatory movement patterns, may further contribute to the higher cadence observed in females. Further research is needed to clarify this potential explanation.

The correlation analysis revealed a significant and strong positive relationship between cadence and METs for both males and females (r =.726 and 0.737). However, these correlations were slightly lower than those reported in previous studies (r =.79–0.90) [16, 17]. This finding indicated that metabolic response enhanced as cadence increased. Previous research has highlighted that obesity can alter movement patterns [36, 37] and is closely linked to body size, weight, gait mechanics, and the metabolic energy expenditure of walking [37, 38]. Furthermore, studies have confirmed that overweight or obese adults exhibit significantly lower physical fitness levels compared to their normal-weight counterparts [39]. It can be speculated that individuals with higher body mass, as a result of reduced gait efficiency, expend more energy per step. As a result, they may achieve higher METs at lower cadences, which could weaken the correlation between METs and cadence.

This study identified a significant positive correlation between FAT and 3 METs cadence in males. This suggests that individuals with higher body fat may require a higher cadence to achieve the same metabolic intensity, possibly due to reduced gait efficiency and increased energy expenditure during walking. This finding is consistent with previous studies indicating altered gait mechanics in individuals with greater adiposity [37, 40]. Further analysis indicated that FAT accounted for 16% of the variance in 3 METs cadence. This indicated that body fat is a statistically significant predictor of MPA cadence adjustments in males. Indicating that individuals with higher body fat percentages require higher cadence strategies to maintain MPA. Importantly, this significant relationship (R² = 0.156) also identifies body-fat percentage as a confounding factor in the cadence–intensity nexus, underscoring the need to account for body composition when prescribing cadence-based guidelines. Previous research has shown that individuals with obesity often exhibit gait adaptations due to reduced stability, including shorter stride lengths, decreased single-leg support time, and increased cadence and step width [37, 40–42]. These findings are similar to our finding. Additionally, body fat and lower limb strength are critical factors in gait propulsion mechanisms [43]. Hills et al. suggested that insufficient lower limb strength may hinder the ability to effectively support body weight during walking [44]. Although this study found that males had higher body weight and muscle mass than females, these advantages may not fully offset the impact of body weight on MPA gait mechanics. Females had higher FAT and lower muscle mass compared to males, may rely on different gait adjustments. These adaptations enable meeting the demands of MPA cadence. However, the specific mechanisms underlying these adjustments in females warrant further investigation. These findings imply there may be a potential gender differences of the relationship between FAT and gait adjustment mechanisms.

According to the ROC curve analysis, the models for MPA, MPA_young, and VPA demonstrated a “good” discrimination for both males and females, with AUC ranging from 0.82 to 0.89. Optimal MPA cadence thresholds were 113.75 steps/min in male and 114.75 steps/min in female, with model sensitivities of 62.14% and 63.75% and specificities of 94.59% and 94.44%, respectively. This means the model was better at correctly excluding individuals who did not reach MPA than at detecting those who did. Wang et al. [28] reported an MPA cadence threshold of 105 steps/min for normal-weight young adults in China, with sensitivity and specificity values of 85% and 74%, respectively. In comparison, the models in this study showed relatively higher false-negative rates, with approximately 36.25–37.86% of individuals achieving MPA but incorrectly classified. The cadence thresholds in this study were determined objectively using Youden’s index, which balances sensitivity and specificity. Nevertheless, this relatively high false-negative rate may be attributed to the use of a single cadence threshold to represent moderate intensity, which may not fully capture inter-individual variations in gait efficiency, body composition, or energy cost. Future research may consider incorporating additional physiological indicators to improve classification accuracy. Nevertheless, it accurately identified individuals who did not reach MPA (sp: 94.59% and 94.44%), thereby contributing to a lower false-positive rate. The PPVs for identifying individuals achieving 3 METs were 96.97% for males and 98.08% for females, while the NPVs for identifying individuals not reaching 3 METs were only 47.30% and 36.96%, respectively. In comparison, Tudor-Locke et al. reported an MPA cadence threshold of 95.5 steps/min for healthy adults aged 21–40 years, with a sensitivity, specificity, PPV, and NPV values of 91.3%, 86.2%, 89.5%, and 88.5%, respectively [21]. The models exhibited higher specificity and PPV in current results. These results indicate that the models are highly reliable for accurately identifying individuals who meet MPA thresholds but less effective in detecting all individuals who achieve MPA. This discrepancy may be attributed to higher body fat percentages in obese populations, which affect the relationship between cadence and METs, reducing the accuracy in identifying individuals achieving MPA. Additionally, obese individuals may adopt gait adjustments to accommodate body weight, requiring higher cadences to achieve the same level of physical activity as normal-weight individuals. These findings underscore the importance of developing individualized cadence thresholds tailored to the unique characteristics of obese populations.

To the best of our knowledge, this study is the first to establish MPA cadence thresholds for obese young adults at 4.8 METs. The findings showed that as intensity increased from MPA to MPA_young, male and female cadence thresholds elevated to 118.50 and 124.50 steps/min, respectively. While prior research reported a 5 METs cadence threshold of 116.5 steps/min in generally healthy adults, only half of the samples were normal weight [21]. This study, focused exclusively on obese young adults, found slightly higher thresholds at 4.8 METs. The difference may be due to obesity-related gait and metabolic factors, such as shorter stride length or higher energy cost per step.

Sensitivity improved from 62.14% to 63.75% to 87.10% and 84.62%, indicating that the model was more effective in identifying individuals who achieved higher intensities, despite a slight decrease in specificity. In practical terms, this means fewer active individuals were misclassified. Similarly, PPV decreased from 96.97% to 98.08% to 50.94% and 68.75%, while NPV increased to 95.40% and 93.94%. Although no previous studies have examined cadence thresholds at 4.8 METs, cadence thresholds for 5 METs has been established across different age groups (21–85 years). These studies consistently reported a decline in PPV and an increase in NPV as cadence thresholds increased from 3 METs to 5 METs, similar to our findings in this study [21, 24, 45]. The observed decrease in PPV and increase in NPV may be attributed to greater variability in gait, influenced by factors such as cadence, stride length, muscular strength, and physical fitness. This variability likely contributes to an increased false-positive rate. Meanwhile, the improved sensitivity reducing the false-negative rate, enhancing the model’s ability to accurately identify individuals who did not achieve 4.8 METs, thereby increasing the NPV.

This study identified the optimal VPA cadence thresholds for obese males and females as 123.75 and 130.50 steps/min, respectively, with sensitivities of 87.50% and 88.89% and specificities of 80.30% and 85.39% at 6 METs. Compared with previous studies, which reported sensitivities of 98.4% and specificities of 85.8% for adults aged 21–40, sensitivity of 97.4% and specificity of 82.9% for middle-aged adults (41–60 years), and sensitivity of 100.0% and specificity of 97.1% for older adults (61–85 years) [21, 24, 42]. The sensitivity of the current models for obese populations was slightly lower but still demonstrated good discrimination in identifying individuals achieving 6 METs. The specificity was comparable to those reported across various age groups, effectively excluding individuals who did not meet the threshold and minimizing false positives. However, the PPV analysis revealed relatively low percentage of 21.21% for males and 38.10% for females. Previous studies reported PPVs of 44.7% for adults aged 21–40 years, 27.6% for middle-aged adults (41–60 years), and 27.3% for older adults (61–85 years), which are similar to the current findings. The relatively low PPV likely reflects a higher rate of false positives when identifying individuals achieving 6 METs. In contrast, the high NPV observed in this study with 99.07% for males and 98.70% for females is consistent with previous findings (99.8%–100%) [21, 24, 45]. This indicates that the model effectively identifies individuals who did not achieve 6 METs, demonstrating high reliability in excluding non-qualified individuals.

We observed a decreasing tendency in PPV as physical activity intensity increased from 3 METs to 6 METs, which is consistent with previous research. Previous research across different age groups revealed that 72% of young adults aged 21–40 years achieved 6 METs [21], while this proportion dropped to 42% among middle-aged adults [42] and 6% among older adults [24]. In the present study, the percentage of young obese adults achieving ≥ 6 METs were 28.57% for males and 50.00% for females, both notably lower than those reported for young adults in general populations. These findings infer that the decline in PPV is closely associated with the reduced proportion of individuals meeting the threshold for higher-intensity activities, with a more pronounced decrease at 6 METs. Potential factors contributing to this result include the high body fat percentage, low muscle mass, and reduced exercise economy commonly observed in obese populations. These factors likely diminish their capacity to achieve VPA levels, leading to a reduced proportion of individuals reaching 6 METs and a corresponding decline in PPV.

In summary, the analytical model developed in this study demonstrated good discrimination and indicated that cadence thresholds increased when physical activity intensities are higher. The model exhibited a tendency for decreased PPV and increased NPV as activity intensity elevated. Specifically, the MPA cadence threshold showed high accuracy in identifying individuals achieving 3 METs, while the VPA cadence threshold was particularly effective in excluding individuals who did not reach 6 METs. This study is the first to establish an MPA_young cadence threshold model for obese young adults in Taiwan, providing individualized physical activity guidelines for walking activity. The MPA_young (4.8 METs) cadence threshold established in this study mitigates the risk of underestimating MPA (3 METs) intensity, thereby supporting appropriate physical activity among obese individuals and informing personalized exercise prescriptions.

It is important to note that this study excluded individuals with cardiometabolic, respiratory, and musculoskeletal disorders to ensure participant safety and internal validity. Cadence thresholds were derived on a level treadmill; overground and free-living conditions may alter cadence–MET relationships, so generalization should be made with caution. However, this approach may limit the generalizability of our findings to obese individuals with comorbidities. Moreover, the homogeneous nature of the sample—obese young adults without comorbidities—may limit the external validity of the findings. The cadence thresholds established in this study may not generalize to individuals with different health statuses, age groups, or fitness levels. Future studies should validate these thresholds in broader obese populations, including those with diverse clinical profiles. Additionally, although body fat percentage was associated with 3 METs cadence in males, body composition was not considered in the ROC analysis. This may limit the applicability of fixed cadence thresholds across individuals with varying body composition. Future studies may consider stratified or covariate-adjusted analyses to further examine this effect. Finally, leg length—a recognized determinant of cadence—was not measured in this study; therefore, its potential moderating role could not be quantified.

From a practical standpoint, a cadence of ≥ 114 steps/min can serve as an initial guideline for achieving MPA (3 METs) because it encompasses the minimal sex difference; however, the confounding effect of body-fat percentage on the 3 METs cadence in males means this threshold should be adjusted or stratified according to individual adiposity to avoid misclassifying exercise intensity. However, this study recommends sex-specific cadence thresholds for higher intensities: for MPA_young (4.8 METs: ≥119 steps/min for males and ≥ 125 steps/min for females), and for VPA (6 METs: ≥124 steps/min for males and ≥ 131 steps/min for females). Obese young adults are encouraged to adopt the 4.8 METs thresholds (≥ 119 and ≥ 125 steps/min) to increase the likelihood of exceeding 3 METs, preventing underestimation of physical activity intensity in practical settings. A simple method for self-monitoring is counting steps for 10 s and multiplying by 6 to estimate cadence. Alternatively, practical step-count goals can be set during walking sessions: for obese young males, 1,190 steps in 10 min (3,570 steps in 30 min) for 4.8 METs and 1,240 steps in 10 min (3,720 steps in 30 min) for 6 METs; for obese young females, 1,250 steps in 10 min (3,750 steps in 30 min) for 4.8 METs and 1,310 steps in 10 min (3,930 steps in 30 min) for 6 METs. These step-count targets provide feasible and accessible markers for moderate-to-vigorous intensity walking in health promotion and exercise interventions.

Conclusion

This study revealed the differences in body composition among obese individuals significantly affect gait adjustment mechanisms during moderate intensity walking activity (3 METs) along with notable gender differences. As physical activity intensity increased, the impact of body composition on obese males diminished. Factors such as cardiorespiratory fitness and muscular strength may play a greater role in determining cadence and energy expenditure at higher intensity. This study established cadence thresholds for Taiwanese obese young adults: 114 steps/min for male and 115 steps/min for female in MPA and 125 steps/min for males and 131 steps/min for females in VPA. Notably, it is the first study to establish MPA_young (4.8 METs) cadence thresholds for obese young adults. Thresholds were identified as 119 steps/min for male and 125 steps/min for female. These findings provide walking activity guidelines for obese young adults and address gaps in cadence research as well as the importance of physical activity for improving body composition and health.

Abbreviations

ACSM

American College of Sports Medicine

BFM

Body Fat Mass

BIA

Bioelectrical Impedance Analysis

BMI

Body Mass Index

FAT

Body Fat Percentage

METs

Metabolic Equivalents

MPA

Moderate-Intensity Physical Activity

MPA_young

Moderate-Intensity Physical Activity for Young Adults (Aged 20–39)

MVPA

Moderate-To-Vigorous Physical Activity

NPV

Negative Predictive Value

PPV

Positive Predictive Value

ROC

Receiver Operating Characteristic

SMM

Skeletal Muscle Mass

VFA

Visceral Fat Area

VPA

Vigorous Physical Activity

Authors’ contributions

T.-L. Chiang conceptualized the study, supervised data collection and experimental implementation, analyzed and interpreted the data, drafted the initial manuscript, and critically revised and finalized the manuscript. Y.-C. Lin designed the study methodology, supervised data analysis procedures, analyzed and interpreted the data, drafted sections of the manuscript, and critically revised and finalized the manuscript. C. Chen designed the study, analyzed and interpreted the data, and critically revised the manuscript. S.-H. Chan collected, analyzed, interpreted the data, and performed data visualization, and critically revised the manuscript. Y.-Y. Ye, C.-H. Hsu and P.-W. Su collected, analyzed, interpreted the data, and performed data visualization. H.-J. Wu designed the study, interpreted the data, critically revised the manuscript, provided expertise in physical activity and exercise research, and acquired funding. All authors have read and approved the final version of the manuscript and agree on the order of authorship.T.-L. Chiang and Y.-C. Lin contributed equally as co-first authors.

Funding

This study was supported by the Chinese Culture University and Ministry of Science and Technology, Taiwan (No. 105-2815-C-034-028-H).

Data availability

The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was developed in accordance with the guidelines proposed in the Declaration of Helsinki and was approved by Human Research Ethics Committee of the Institutional Review Board of Fu Jen Catholic University (No. C105137, approved November 9, 2017). All participants gave informed consent before their inclusion in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Conflict of interest

The authors declare there are no conflicts of interest in this study.