Association between eating speed, body composition, and physical activity: a cross-sectional study in Gujarat, India

Abstract

This study aims to investigate the relationship between eating speed, body composition, and physical activity among adults in Gujarat, India. By analyzing anthropometric and metabolic parameters across different eating-speed groups, the study seeks to identify whether eating speed influences obesity-related body composition indicators and how it interacts with levels of physical activity. A cross-sectional study was conducted among 465 adults (240 males, 225 females) aged 18–65 years in Gujarat, India. Anthropometric and body-composition measures (BMI, body-fat %, visceral-fat %, resting metabolism) were recorded using a bioelectrical impedance analyzer. Eating speed was assessed via a self-structured questionnaire, and physical activity was measured using the International Physical Activity Questionnaire – Short Form (IPAQ-SF). Fast eaters demonstrated significantly higher BMI and visceral fat levels compared with moderate and slow eaters (BMI: p = 0.0179; visceral fat: p = 0.0166), indicating a positive association between rapid eating and greater adiposity. No significant associations were observed for body-fat percentage (p = 0.6815) or resting metabolism (p = 0.0657). There was no association between eating habits and physical activity. Eating speed was positively associated with BMI and visceral fat, indicating that faster eating contributes to greater adiposity. Therefore, modifying eating speed may act as a feasible behavioral intervention to help reduce obesity, especially when combined with regular physical activity.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-39798-5.

Introduction

Overweight or obesity can be defined as an abnormal or excessive accumulation of fat in the body¹. Metabolic risk is strongly linked to excessive body fat, suggesting that there must be both preventive and curative public-health strategies to mitigate obesity-related diseases². In recent times, obesity and weight gain pose a major health hazard in both developed and developing nations around the globe, leading to comorbidities such as diabetes and hypertension³.

Reports from the studies conducted recently suggest that more than 1.9 billion adults are underweight and 650 million are obese globally. Additionally, obesity and being overweight are responsible for nearly 2.8 million deaths⁴. Because of unhealthy eating habits, sedentary lifestyles, lack of health care services, and financial support, the developing countries are getting more prone to adverse consequences of obesity⁵. More than 135 million individuals in India suffer from obesity. Age, gender, geographical environment, socio-economic status, eating behavior, etc. are the suspected causes of the prevalence of obesity in our country⁶. The National Family Health Survey-5 (2019–2021) indicates that the prevalence of abdominal obesity is 40% in women and 12% in men. The results indicate that 50–60% of women aged 30–49 are abdominally obese⁷. According to the study of ICMR-INDIAB in 2015, the prevalence rate of central obesity varies from 16.9% to 36.3% and 11.8% to 31.3%, respectively⁸. Nowadays, obesity is one of the major medical and financial burdens for governments. This growing problem can be mitigated not only by adopting healthy food habits, lifestyle, and physical activity but also by promoting these behaviors at the community level⁹.

Eating behavior, particularly eating speed, can be considered a factor in the development of obesity^10–12. Other factors may include irregularities in the quantity of food eaten daily, the number of meals taken in a day, the amount of consumption of junk food, or skipping (any meal) breakfast¹³. The speed of eating a meal has been found to influence gut hormones, which are associated with satiety signals, which in slower eaters leads to more anorexigenic response than in fast eaters. Fast eating speed is positively associated with energy intake and shows an association with BMI, obesity, and metabolic disease. Whereas limited studies on eating rates show that slow eating speed is helpful in reducing energy intake¹⁴. The probable mechanism between the rate of eating and the weight gain is associated with a lack of satiety¹⁵, insulin resistance, and decreased insulin sensitivity¹⁶. The lack of satiety leads to excessive eating before the stomach senses fullness and thus leads to weight gain¹⁵. However, there are no references or any clinical guidelines that define an appropriate assessment of eating rate, which raises more demand for systemic assessment of eating speed on body composition¹⁷.

Another modifiable risk factor for obesity is lack of physical activity. Physical inactivity leads to cardiovascular disease and various other non-communicable diseases (NCDs) like diabetes mellitus, obesity, hypertension, or joint diseases. Physical fitness is defined as a physiological state of well-being that enables one to fulfil the demands of activity of daily living¹⁸. In recent times, physical activity at home and in the workplace has declined considerably, and more time is spent on sedentary activities, involving sitting for prolonged periods, which contributes to a decline in physical activity energy expenditure; therefore, obesity might be expected¹⁹. Studies have shown that an increase in physical fitness will subsequently lead to improvements in health status¹⁸.

However, to our knowledge, no study has examined the relationship between eating speed and body composition in the Indian population. Active individuals tend to have better sensitivity to satiety signals, potentially helping them pace their meals better. Given the scarcity of literature, investigating how eating speed influences weight gain may contribute to primordial prevention of obesity by promoting slower eating habits and regular physical activity from an early age. The present study aims to examine the association between eating speed, body composition, and physical activity among adults in Gujarat, India.

Null hypothesis

There is no significant association between eating speed and body-composition indicators or physical-activity levels among adults.

Methods

Type of study

A cross-sectional study.

Study setting

Shree Krishna Hospital, Birla Vishwakarma Mahavidyalaya, and the vicinity of Vallabh Vidyanagar and Karamsad, Gujarat, India.

Study sample

Participants were recruited primarily from hospital and nearby community settings, which could introduce selection bias toward more health-conscious or health-aware individuals. As a result, the sample may not fully represent individuals from different socioeconomic or occupational backgrounds.

Ethical clearance

It was obtained from Institutional Ethics Committee – 2, HM Patel Centre for Medical Care and Education, Karamsad (IEC/HMPCMCE/93/Faculty/2/84/18) before the conduct of the study. All methods were performed in accordance with the relevant guidelines and regulations of the institutional and national research committees, in line with the ethical standards set forth in the Declaration of Helsinki.

Sample size

Assuming a small ANOVA effect (Cohen’s f = 0.15), α = 0.05, power = 0.80, and three groups, the required n ≈ 348; our achieved n = 465 aged 18–65 years, comprising 240 males (51.6%) and 225 females (48.4%), were included in the study.

Sampling method

The present study employed a convenience sampling method, which may limit the generalizability of the findings to the wider population.

Inclusion criteria

Healthy adult individuals aged 18–65 years.

Exclusion criteria

Individuals with any acute illness, individuals following any diet plan, individuals with any eating disorders, individuals consuming medicines that affect appetite, chronic alcoholics, and individuals with dental problems.

Instruments

Omron HBF-375 bioelectrical impedance analyzer (BIA) to measure body composition parameters. Bioelectrical Impedance Analysis (BIA) is a non-invasive method that uses a small, safe electrical current to measure how resistant (impedance) body tissues are to it. This lets you figure out how much fat and muscle you have. Because lean tissue conducts electricity better than fat tissue, changes in impedance can be used to figure out body fat percentage, visceral fat, and other parts. BIA is a useful and non-invasive way to estimate visceral fat and basal metabolic rate (BMR), but these values are based on algorithms and may not fully reflect true physiological measures. Omron BIA device shows moderate correlation (r = 0.65–0.73) with CT-derived visceral fat measurements^20–22. While useful for large-scale screening, BIA-derived visceral fat estimates have substantial individual-level measurement error and should be interpreted cautiously.

Procedure for study

Prior to data collection, each participant received a clear explanation of the study objectives, procedures, and confidentiality measures. Written informed consent was obtained from all participants before enrolment. Participation was entirely voluntary, and respondents were informed that they could withdraw from the study at any time without penalty or loss of benefits. The participant underwent a self-structured questionnaire via Google Forms targeting their demographic data, including age, gender, and occupation. Information on type of diet (vegetarian or non-vegetarian) and frequency of fried or junk food intake was also recorded to account for dietary variations.

Anthropometric measurements such as height and waist circumference were measured by standard measuring tape. A wall-mounted measuring tape was used for measuring height, for which subjects were asked to stand straight touching the wall, feet together, knees extended, and heels, buttocks, and shoulder blades touching the wall. Waist circumference of the subjects was measured by standard measuring tape at the end of normal expiration. It was measured at the level between the lower rib cage and iliac crest, keeping the measuring tape parallel to the floor. Weight, body mass index (BMI), body fat percentage, visceral fat level, and basal metabolic rate of the subjects were determined using Bioelectrical Impedance Analysis Omron HBF-375²³. Participants were asked to remove their shoes and socks, and then they were asked to stand on the bioelectrical impedance analyser with both feet al.igned properly, knees extended, both shoulders flexed 90°, elbows extended, and hands holding the extender of the machine. Dietary habits, especially their self-reported eating speed and consumption of fried food and junk food on a weekly basis.

Study questionnaire

Eating speed was assessed using a self-administered questionnaire (attached in the supplementary file) adapted from Hamada et al. (2017), which classified participants into slow (> 20 chews/bite), moderate (10–20 chews/bite), and fast (< 10 chews/bite) groups²⁴. Although self-reported measures may involve subjective estimation and recall bias, prior studies have shown reasonable validity when compared to observed eating rates^12,17. The selected cut-offs were consistent with existing literature on eating-rate categorization. These questions were included in a self-structured questionnaire. The physical activity was measured by the International Physical Activity Questionnaire (IPAQ)—Short Form. The IPAQ-SF questionnaire was helpful in finding out the level of physical activity that people do as part of their daily routine lives, and it was divided into mild activity, moderate activity, and vigorous activity according to Metabolic Equivalent of Task (MET)-minutes/week based on activity they had done in the previous 7 days²⁵. As the study focused on behavioral classification rather than development or validation of the instrument, reliability testing of the questionnaire was not performed.

All anthropometric and questionnaire data were collected by a single trained investigator to minimize inter-rater variation. Each anthropometric measurement was taken twice, and the mean value was used for analysis to ensure measurement reliability. Anthropometric measurements followed standardized procedures based on the ISAK (International Society for the Advancement of Kinanthropometry) guidelines for consistency and reliability.

Statistical analyses

It were performed using IBM SPSS Statistics for Windows, Version 31.0 (IBM Corp., Armonk, NY, USA). Descriptive statistics were expressed as mean ± SD for continuous variables and frequency (percentage) for categorical variables. One-way ANOVA with Tukey’s HSD post-hoc test was used for group comparisons, and p-values were adjusted using the Bonferroni correction to control for multiple testing. Effect sizes were calculated for each inferential test (η² for ANOVA, d for pairwise comparisons, OR with 95% CI for logistic models). Model fit was evaluated using Hosmer–Lemeshow, Cox & Snell R², and Nagelkerke R². Chi-square test and Fisher’s exact test to find associations between eating speed categories, BMI, and physical activity. Binary logistic regression analyses were conducted to identify predictors of high visceral fat (≥ 10%) and overweight/obesity (BMI ≥ 25 kg/m²). Predictors were entered hierarchically in three blocks¹:demographic variables (age, sex)², lifestyle factors (eating speed, physical activity, sleep duration), and³physiological measures (body fat %, BMR). Model diagnostics included the Box–Tidwell test for linearity of continuous predictors with the logit, Hosmer–Lemeshow goodness-of-fit, and variance inflation factor (VIF) values to assess multicollinearity (VIF < 2.0, tolerance > 0.5 considered acceptable). Significance was set at p < 0.05 after Bonferroni correction. Incomplete questionnaires (< 5%) were excluded via listwise deletion; no variable exceeded 2% missingness, and assumption testing confirmed MCAR (Little’s χ² = 12.4, p = 0.19). To examine whether the association between eating speed and body composition differed by physical activity level, an exploratory two-way ANOVA was conducted using eating-speed category (slow, moderate, fast) and physical-activity level (mild, moderate, vigorous; based on IPAQ MET-classification) as fixed factors. Dependent variables included BMI, body-fat percentage, and visceral-fat percentage. Interaction terms (eating speed × physical activity) were assessed to identify potential moderation effects. As these analyses were exploratory in nature and intended to supplement the primary objective of assessing the direct association between eating speed and body composition, the full statistical outputs of the two-way ANOVA—including interaction effects and effect sizes—are presented in the Supplementary File (Table S1).

Results

Demographic, Lifestyle, and Eating-Behavior characteristics of participants

The study included 465 adults, of whom 240 (51.6%) were male and 225 (48.4%) were female. Nearly half of the participants were below 35 years of age (49%), followed by those aged 36–50 years (29%) and 51–65 years (22%). The largest occupational group comprised service employees (37.85%), followed by students (28.60%), other occupations such as peon, tailor, carpenter, housewife, and vendor (28.39%), and professors (5.16%). Regarding eating habits, 25.16% of participants completed a meal in less than 10 min, while 44.73% took 10–15 min, 21.08% took 15–20 min, and 9.03% required more than 20 min. Based on chewing frequency, 10.97% were categorized as slow eaters, 44.73% as moderate eaters, and 44.30% as fast eaters. Fried-food consumption was highest among those consuming it 3–4 times per week (30.32%), followed by occasional consumers (29.03%), those eating it 1–2 times per week (28.39%), and daily consumers (12.26%). Junk food intake showed that 26.82% consumed it weekly, 16.99% occasionally, and 12.91% often. Physical-activity levels assessed using the IPAQ indicated that most participants had moderate activity (55.70%), followed by mild (35.91%) and vigorous activity (8.39%). (Table 1)

Table 1.

Demographic, Lifestyle, and Eating-Behaviour characteristics of participants (N = 465).

Characteristic	Category	n (%)
Sex	Male	240 (51.6%)
	Female	225 (48.4%)
Age Group (years)	< 35 years	228 (49.0%)
	36–50 years	136 (29.0%)
	51–65 years	101 (22.0%)
Occupation	Service	176 (37.85%)
	Students	133 (28.60%)
	Other occupations	132 (28.39%)
	Professor	24 (5.16%)
Time Taken to Complete a Meal	< 10 min	117 (25.16%)
	10–15 min	208 (44.73%)
	15–20 min	98 (21.08%)
	> 20 min	42 (9.03%)
Eating-Speed Category	Slow (> 20 chews/bite)	51 (10.97%)
	Moderate (10–20 chews/bite)	208 (44.73%)
	Fast (< 10 chews/bite)	206 (44.30%)
Fried Food Consumption	Daily	57 (12.26%)
	3–4 times/week	141 (30.32%)
	1–2 times/week	132 (28.39%)
	Occasionally	135 (29.03%)
Junk Food Consumption	Weekly	125 (26.82%)
	Occasionally	79 (16.99%)
	Often	74 (12.91%)
Physical Activity (IPAQ)	Mild	167 (35.91%)
	Moderate	259 (55.70%)
	Vigorous	39 (8.39%)

Descriptive data

The average BMI, resting metabolism, body fat percentage, visceral fat percentage, and waist circumference of participants were 25.01 (± 4.89) kg/m², 1427.16 (± 265.63) kcal, 30.63 (± 7.56) %, 9.43 (± 6.44) %, and 90.42 (± 14.10) cm, respectively (Table 2).

Table 2.

Descriptive statistics for key variables.

Sr No.	Variable	Mean	SD
1	Age (years)	36.83	13.63
2	BMI (kg/m2)	25.01	4.89
3	Body fat (%)	30.63	7.56
4	Visceral Fat (%)	9.43	6.44
5	Resting Metabolism (Kcal)	1427.16	265.63
6	Waist circumference (cm)	90.42	14.10

Eating speed and BMI category

A Chi-square test of independence revealed a statistically significant association between eating speed and BMI category (p = 0.006). These findings indicate that faster eating behavior is significantly associated with higher BMI categories. (Table 3)

Table 3.

Association between Eating-Speed categories and BMI classification (Chi-square Test).

No. of chews/bite (Eating Speed)	BMI ≤ 18.5	BMI 18.6–24.9	BMI 25–29.9	BMI ≥ 30
Slow (> 20)	15.69%	43.14%	19.61%	27.57%
Moderate (10–20)	9.62%	47.12%	33.65%	9.62%
Fast (< 10)	9.22%	33.98%	36.89%	19.90%
Total	10.11%	40.86%	33.55%	15.48%

Note: Data represents the percentage of participants within each BMI category across eating-speed groups. Association tested using Pearson’s Chi-square test (χ²). Statistical significance was set at p < 0.05 (p = 0.006).

A one-way ANOVA revealed a statistically significant difference in mean BMI across eating-speed categories (F (2,462) = 4.09, p = 0.0179, η² = 0.017, 95% CI [0.002–0.044]). Although the effect-size was small, the findings suggest that faster eating behavior is associated with slightly higher BMI values compared to moderate-speed eaters. (Table 4). Post-hoc Tukey HSD tests (Table 5) were conducted to identify pairwise differences between eating-speed groups. A significant mean difference in BMI was observed between fast and moderate eaters (mean difference = 1.36%, 95% CI [0.29, 2.43], p = 0.013, d = 0.29 [0.10, 0.49]). No significant differences were found between moderate and slow eaters (p = 0.634) or between fast and slow eaters (p = 0.653). These results indicate that fast eating is associated with a modestly higher BMI than moderate eating, although the effect size was small.

Table 4.

Comparison of mean BMI across Eating-Speed categories using One-way ANOVA.

No. of chews/bite (Eating Speed)	n	BMI (Mean ± SD)	95% CI of mean	Test → p	Effect Size
Slow (> 20)	51	25.03 ± 6.32	23.27 to 26.79	-	-
Moderate (10–20)	208	24.34 ± 4.37	23.75 to 24.93	-	-
Fast (< 10)	206	25.70 ± 4.91	25.05 to 26.34	-	-
Total	465	25.01 ± 4.89	24.56 to 25.46	ANOVA: F (2,462) = 4.09, p = 0.0179	η²=0.017 (95% CI 0.002–0.044)

Note: Data expressed as Mean ± SD with 95% Confidence Intervals (CI). One-way ANOVA used to test differences across eating-speed categories. Effect size expressed as eta-squared (η²). Statistical significance set at p < 0.05.

Table 5.

Post-hoc tests within eating-speed groups (BMI as outcome).

Comparison	Statistical Test	Mean Difference (%)	95% CI for Mean Difference	Adjusted p (Tukey)	Effect Size (Cohen’s d)	95% CI for d
Moderate vs. Slow	Tukey HSD (ANOVA Post-hoc)	−0.69	[−2.54, 1.16]	0.634	0.14	[–0.16, 0.45]
Fast vs. Slow	Tukey HSD (ANOVA Post-hoc)	0.67	[−0.36, 1.69]	0.653	0.13	[–0.18, 0.43]
Fast vs. Moderate	Tukey HSD (ANOVA Post-hoc)	1.36	[0.29, 2.43]	0.013 *	0.29	[0.10, 0.49]

Note: *Significant at p < 0.05. Post-hoc comparisons derived from one-way ANOVA (F (2,462) = 4.09, p = 0.0179, η² = 0.018 [95% CI 0.002–0.045]). Effect sizes reported as Hedges-corrected Cohen’s d with 95% CIs.

Comparison between eating speed and body fat

A one-way ANOVA showed no statistically significant difference in mean body fat percentage among the three eating-speed categories (F(2, 462) = 0.38, p = 0.6815, η² = 0.002, 95% CI [0.000–0.015]). The effect size was negligible, indicating that eating speed had no meaningful association with body fat percentage in this sample. (Table 6)

Table 6.

Comparison of mean body fat across Eating-Speed categories using One-way ANOVA.

No. of chews/bite (Eating Speed)	n	Body fat (Mean ± SD)	95% CI of mean	Test → p	Effect Size
Slow (> 20)	51	31.22 ± 9.58	28.55 to 33.89	-	-
Moderate (10–20)	208	30.32 ± 7.05	29.35 to 31.29	-	-
Fast (< 10)	206	30.80 ± 7.52	29.77 to 31.83	-	-
Total	465	30.63 ± 7.56	29.95 to 31.31	One-way ANOVA F (2, 462) = 0.38, p = 0.6815	η² = 0.002 [95% CI 0.000–0.015]

Note: Data expressed as Mean ± SD with 95% Confidence Intervals (CI). One-way ANOVA was used to test differences across eating-speed categories. Effect size is reported as eta-squared (η²) with 95% CI. Statistical significance was set at p < 0.05.

Comparison between eating speed and visceral fat

A one-way ANOVA revealed a statistically significant difference in mean visceral fat percentage among the three eating-speed categories (F (2,462) = 4.12, p = 0.0166, η² = 0.017, 95% CI [0.002–0.044]). The effect size (η² = 0.017) indicated a small but meaningful effect (Table 7). These results suggest that faster eating behavior is associated with higher visceral fat accumulation. Post-hoc Tukey HSD comparisons (Table 8) revealed a statistically significant difference in visceral fat percentage between fast and moderate eaters (mean difference = 1.78%, 95% CI [0.39, 3.18], p = 0.013, d = 0.28 [0.09, 0.48]). No significant differences were found between slow and moderate eaters (p = 0.895) or between fast and slow eaters (p = 0.376). The small effect size suggests that, while statistically significant, the practical difference in visceral fat between fast and moderate eaters is modest.

Table 7.

Comparison of mean visceral fat across Eating-Speed categories using One-way ANOVA.

No. of chews/bite (Eating Speed)	n	Visceral Fat (Mean ± SD)	95% CI of mean	Test → p	Effect Size
Slow (> 20)	51	9.03 ± 7.58	7.00 to 11.06	-	-
Moderate (10–20)	208	8.59 ± 5.55	7.83 to 9.35	-	-
Fast (< 10)	206	10.37 ± 6.86	9.44 to 11.30	-	-
Total	465	9.43 ± 6.44	8.88 to 9.98	One-way ANOVA F (2,462) = 4.12, p = 0.0166	η² = 0.017 [95% CI 0.002–0.044]

Note: Data expressed as Mean ± SD with 95% Confidence Intervals (CI). One-way ANOVA was used to test differences across eating-speed categories. Effect size is reported as eta-squared (η²) with 95% CI. Statistical significance was set at p < 0.05.

Table 8.

Post-hoc tests within eating-speed groups (Visceral fat as outcome).

Comparison	Statistical Test	Mean Difference (%)	95% CI for Mean Difference	Adjusted p (Tukey)	Effect Size (Cohen’s d)	95% CI for d
Moderate vs. Slow	Tukey HSD (ANOVA Post-hoc)	−0.44	[−2.23, 1.34]	0.895	–0.07	[−0.38, 0.23]
Fast vs. Slow	Tukey HSD (ANOVA Post-hoc)	1.34	[−0.58, 3.27]	0.376	0.19	[−0.12, 0.50]
Fast vs. Moderate	Tukey HSD (ANOVA Post-hoc)	1.78	[0.39, 3.18]	0.013*	0.28	[0.09, 0.48]

Note: *Significant at p < 0.05. p values are adjusted using Tukey’s HSD post-hoc test following One-way ANOVA F (2, 462) = 4.12, p = 0.0166, η² = 0.017 [95% CI 0.002–0.044]. Effect sizes are reported as Hedges-corrected Cohen’s d with 95% confidence intervals. Significant at p < 0.05.

Comparison between eating speed and resting metabolism

A one-way ANOVA revealed no statistically significant difference in resting metabolism across the three eating-speed categories (F (2,462) = 2.73, p = 0.0657, η² = 0.012, 95% CI [0.000–0.035]). Although fast eaters had a slightly higher mean resting metabolism compared to slow and moderate eaters, these differences did not reach statistical significance. The small effect size suggests only a minor, non-significant trend toward higher metabolism with faster eating speed. (Table 9)

Table 9.

Comparison of resting metabolism fat across Eating-Speed categories using One-way ANOVA.

No. of chews/bite (Eating Speed)	n	Resting Metabolism (Mean ± SD)	95% CI of mean	Test → p	Effect Size
Slow (> 20)	51	1390.47 ± 239.96	1323 to1458	-	-
Moderate (10–20)	208	1404.60 ± 278.26	1366 to 1443	-	-
Fast (< 10)	206	1459.02 ± 256.14	1421 to 1497	-	-
Total	465	1427.16 ± 265.63	1397 to 1457	One-way ANOVA F (2, 462) = 2.73, p = 0.0657	η² = 0.012 [95% CI 0.000–0.035]

Note: Data expressed as Mean ± SD with 95% Confidence Intervals (CI). One-way ANOVA was used to test differences across eating-speed categories. Effect size is reported as eta-squared (η²) with 95% CI. Statistical significance was set at p < 0.05.

Eating speed and physical activity category

A Chi-square test of independence was conducted to examine the association between eating-speed category and physical-activity level (IPAQ classification). The association was not statistically significant (p = 0.71). These findings indicate that physical activity levels are similarly distributed across eating-speed groups and do not vary meaningfully with eating behavior. (Table 10)

Table 10.

Association between Eating-Speed categories and physical activity classification (Chi-square Test).

No. of chews/bite (Eating Speed)	Mild PA	Moderate PA	Vigorous PA
Slow (> 20)	39.21%	49.01%	11.76%
Moderate (10–20)	33.65%	57.69%	8.65%
Fast (< 10)	37.38%	55.34%	7.28%
Total	35.91%	55.70%	8.39%

Note: Data represents the percentage of participants within each physical-activity category across eating-speed groups. Association tested using Pearson’s Chi-square test (χ²). Statistical significance was set at p < 0.05. The result was non-significant (p = 0.71), indicating that physical activity does not differ across eating-speed groups.

Logistic regression analysis (Table 11) identified age, male sex, faster eating speed, and higher body fat percentage as significant predictors of elevated visceral fat (≥ 10%). After adjustment for lifestyle and physiological variables, fast eaters were 2.18 times more likely to have high visceral fat compared to moderate eaters (95% CI: 1.27–3.74, p = 0.005). Physical activity level and BMR were not significant predictors. The final model demonstrated good fit (Hosmer–Lemeshow p = 0.51) and no multicollinearity among predictors (all VIF < 2.0).

Table 11.

Binary logistic regression predicting high visceral fat (≥ 10%).

Predictor	B	SE	Wald χ²	p-value	Odds Ratio (OR)	95% CI for OR	VIF
Model 1 – Demographic variables
Age (years)	0.032	0.011	8.47	0.004*	1.03	[1.01,1.05]	1.12
Sex (male)	0.614	0.227	7.32	0.007*	1.85	[1.18, 2.90]	1.08
Model 2 – Lifestyle variables
Eating speed (Fast vs. Moderate)	0.781	0.279	7.85	0.005*	2.18	[1.27, 3.74]	1.21
Eating speed (Slow vs. Moderate)	−0.191	0.298	0.41	0.523	0.83	[0.46, 1.48]	1.19
Physical activity (Low vs. Moderate)	0.258	0.248	1.08	0.299	1.29	[0.79, 2.09]	1.14
Physical activity (High vs. Moderate)	−0.394	0.263	2.24	0.135	0.67	[0.40, 1.13]	1.11
Sleep duration (hrs/day)	−0.118	0.084	1.97	0.160	0.89	[0.75, 1.05]	1.03
Model 3 – Physiological variables
Body fat (%)	0.042	0.013	10.38	0.001*	1.04	[1.02, 1.07]	1.32
Basal Metabolic Rate (kcal/day)	−0.001	0.001	1.44	0.230	0.99	[0.99, 1.00]	1.27

Note: * Significant at p < 0.05.

Hosmer–Lemeshow χ²⁸= 7.29, p = 0.51 (good fit).

Cox & Snell R² = 0.18; Nagelkerke R² = 0.24.

All VIFs < 2.0, indicating no multicollinearity.

Overall model correctly classified 73.4% of cases.

Fast eating remained a significant predictor of overweight/obesity compared with moderate eating (OR = 1.75, 95% CI 1.14–2.68, p = 0.012) after adjustment for demographics, physical activity, sleep, body fat %, and BMR. Body fat % was strongly associated with BMI category (OR = 1.07 per 1% increase, p < 0.001). Physical-activity level and BMR were not significant predictors; diagnostics indicated good model fit and no multicollinearity (Table 12).

Table 12.

Binary logistic regression predicting Overweight/Obesity (BMI ≥ 25 kg/m²).

Predictor	B	SE	Wald χ²	p-value	Odds Ratio (OR)	95% CI for OR	VIF
Model 1 – Demographic variables
Age (years)	0.026	0.010	6.57	0.010*	1.03	[1.01,1.05]	1.12
Sex (male)	0.548	0.214	6.54	0.011*	1.73	[1.14, 2.64]	1.08
Model 2 – Lifestyle variables
Eating speed (Fast vs. Moderate)	0.560	0.223	6.32	0.012*	1.75	[1.14, 2.68]	1.21
Eating speed (Slow vs. Moderate)	−0.118	0.263	0.20	0.652	0.89	[0.53, 1.49]	1.19
Physical activity (Low vs. Moderate)	0.182	0.210	0.75	0.387	1.20	[0.79, 1.83]	1.14
Physical activity (High vs. Moderate)	−0.251	0.236	1.13	0.287	0.78	[0.49, 1.24]	1.11
Sleep duration (hrs/day)	−0.091	0.077	1.39	0.238	0.91	[0.78, 1.07]	1.03
Model 3 – Physiological variables
Body fat (%)	0.067	0.012	31.35	< 0.001*	1.07	[1.05, 1.10]	1.32
Basal Metabolic Rate (kcal/day)	−0.001	0.001	2.29	0.130	0.99	[0.99, 1.00]	1.27

Note: Significant at p < 0.05.

Hosmer–Lemeshow χ²⁸= 6.81, p = 0.56 (good fit).

Cox & Snell R² = 0.20; Nagelkerke R² = 0.26.

All VIFs < 2.0 (no multicollinearity).

Overall model classification accuracy = 75.2%.

Discussion

This study aimed to examine the association between eating speed, body composition, and physical activity among individuals in Gujarat, India. The results demonstrated that a rapid eating pace was significantly associated with increased BMI and visceral fat, partially validating our hypothesis. No significant associations were identified for body fat percentage or resting metabolism. Similar findings were observed in a study conducted by Yuka Hamada et al. on objective and subjective eating speed and its relationship with body composition in female college students, where only 9 participants were found to eat slowly out of 84 total participants²⁴. The primary reason for such an outcome may be attributed to the fast-modern lifestyle due to which people tend to eat faster. It was found from our study that the frequency of eating fried food was substantially higher, with 141 participants eating it 3–4 times/week and 57 of them eating it daily, which may lead to being overweight. In contrast, junk food was rarely consumed by the participants, although the worldwide consumption of junk food is higher. This may be primarily due to the cultural eating habits and preferences.

The study conducted by Yumi Hurst et al. examined the effects of changes in eating speed on obesity among Japanese patients with diabetes, using data from available insurance claims and health check-up records from 2013 to 2016, and found that slow and normal eaters are less likely to have a higher BMI than fast eaters²⁶. Additionally, T. Ohkuma et al. conducted a systematic review and meta-analysis of 23 published epidemiological studies that reported on obesity in relation to eating speed, concluding that faster eating speeds are associated with higher BMI. This can be explained by the fact that fast eaters tend to overeat before the brain recognizes the satiety signal, which is triggered by nutrient intake, distension of gastric muscles, and the release of gut hormones like cholecystokinin²⁷. A study conducted by Takahiro Iwasaki et al. among 398 Japanese adults at Asahi University Hospital in 2016 also suggested that eating quickly was positively correlated with the visceral fat percentage²⁸. However, the body fat percentage and resting metabolism showed no significant association with the eating speed. A similar study conducted by Yuka Hamada et al. in Japanese university students also illustrated that the total number of chews and total meal duration were found to have no relation with body fat%, while the eating speed was significantly associated with the BMI²⁴.

The logistic regression model indicated that eating speed remained a strong and independent predictor of visceral adiposity, even after adjusting for age, sex, and physical activity. The relatively small but consistent effect sizes reinforce that rapid eating contributes to excess central fat accumulation irrespective of metabolic rate or physical activity level. In other words, the association between faster eating and higher BMI or visceral fat persisted across varying activity intensities. While regular physical activity is well established as a protective factor against obesity, these findings imply that rapid eating behaviour independently promotes adiposity. Nevertheless, habitual physical activity may mitigate the long-term metabolic consequences of fast eating through improved energy balance and enhanced fat oxidation. Future longitudinal and interventional studies are warranted to explore whether increasing activity levels can offset the adverse metabolic effects of rapid eating.

Employing BMI, body fat percentage, visceral fat percentage, and resting metabolism provides a more thorough comprehension of obesity-related body composition than relying solely on weight or waist measurement. BMI indicates overall body fat, but visceral fat and resting metabolism offer insights into metabolic risk and energy expenditure, respectively. Thus, this multi-parameter methodology enhances the understanding of the influence of eating speed on obesity. Dietary practices, such as the consumption of fried foods and vegetarian or non-vegetarian choices, additionally affect body composition. In our study, participants who regularly consumed fried foods or adhered to non-vegetarian diets exhibited elevated BMI and visceral fat, consistent with existing data suggesting that high-fat, energy-dense diets increase obesity risk. No substantial relationship was identified between eating velocity and levels of physical activity. This indicates that even moderate to high levels of physical activity may not completely mitigate the consequences of quick eating, emphasizing that behavioural changes such as mindful eating are still crucial for effective weight management.

The research contributes to the limited Indian data linking dietary habits with objective evaluations of body composition. It used approved tools, like the Omron BIA analyzer, and established measurement methods (ISAK), which made the data more reliable. The addition of the IPAQ to the evaluation of physical activity gives a full picture of lifestyle and obesity factors. The Omron HBF-375 has shown a moderate correlation with DEXA-based and CT-based assessments, but BIA-derived visceral fat is still just an estimate. So, the results should be considered indicative rather than absolute measures of adiposity and metabolism. For more accurate quantification, future studies should use gold-standard techniques like DEXA or CT imaging. The study didn’t take into consideration the caloric intake of individuals. Another limitation relates to the method used to measure eating speed. The data may be biased due to the subjective nature of individuals’ recollections regarding the frequency of chewing and the duration of meals. The categorization of participants into slow (> 20 chews/bite), moderate (10–20 chews/bite), and fast (< 10 chews/bite) eaters was based on methodologies utilized in studies by Hamada et al. (2017)²⁴and Woodward et al. (2020)¹⁵, which demonstrated adequate alignment between self-reported and observed eating rates in significant samples. Nevertheless, these classifications were not empirically validated in the present study. Future research should employ direct observational or video-based timing methodologies, or innovative digital sensors and wearable chewing monitors, to obtain objective, standardized evaluations of eating rate. Adding both self-reported and objective data would help make the connections between eating habits and body composition more reliable and less biased.

The study can be done by objectifying the eating speed and taking the full-day caloric intake into consideration. Participants were recruited by convenience from a tertiary hospital, a college, and nearby community settings, which may over-represent health-conscious or help-seeking adults and under-represent lower-engagement groups. As such, prevalence estimates and associations may not generalize to the broader Gujarat or Indian adult population. Future studies should consider using probability-based sampling strategies and recruiting from diverse community settings to enhance representativeness and external validity. Although standard measuring tape was used for height and waist circumference, it may not be as precise as a stadiometer or anthropometer. Future studies could adopt these instruments for enhanced measurement reliability. Additionally, eating speed was assessed through self-reported questionnaires, which may introduce subjective bias. The pre-assessment nutritional state of participants was not fully standardized. Although participants were advised to avoid heavy meals before the BIA assessment, fasting status could not be strictly verified. Variations in recent food intake or hydration may have introduced modest measurement variability, which should be considered when interpreting the BIA-derived estimates. The eating speed assessment relied entirely on self-reported number of chews per bite, which introduces measurement error. Participants may inaccurately estimate their chewing behaviour, and social desirability bias may lead to over-reporting of slower eating. The cut-off points used to categorize eating speed (< 10, 10–20, and > 20 chews per bite) were not validated against objective measures such as video observation, bite monitoring devices, or standardized meal tests, which may have introduced misclassification bias. This measurement limitation likely attenuated the true associations and may partially explain the modest effect sizes observed. Consequently, the self-report nature of eating speed assessment represents a significant methodological constraint that limits the precision and generalizability of the findings. Future studies should employ objective eating speed assessment methods, including video-based bite counting, wearable sensors that detect chewing patterns, standardized test meals with timed consumption, or validated eating behavior questionnaires (e.g., the Eating Rate Questionnaire). Future studies should employ standardized fasting protocols to improve measurement precision. Although sex was included as a covariate in the logistic regression models, we could not conduct additional sensitivity analyses such as gender-stratified models or exclusion of statistical outliers.

Conclusion

The investigation reveals that a rapid eating pace is independently associated with increased BMI and visceral fat, even when accounting for physical activity. Altering eating pace, together with promoting physical activity and nutritious dietary selections, can effectively modify behavior to reduce the risk of obesity. These findings underscore the imperative for public health initiatives that highlight mindful eating behaviors as essential elements of obesity prevention methods.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AG contributed to study conception, design, and supervision of data collection and analysis. AR contributed to literature review, data validation, and initial drafting of the manuscript. PK coordinated the research process, performed statistical analysis, and led manuscript writing. He is the corresponding author. SB managed participant recruitment, data collection, and reviewed the manuscript critically for content. All authors (AG, AR, PK, SB) reviewed and approved the final manuscript. They declare no conflicts of interest and take full responsibility for the integrity and accuracy of the work.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

Ethical clearance was obtained from Institutional Ethics Committee – 2, HM Patel Centre for Medical Care and Education, Karamsad (IEC/HMPCMCE/93/Faculty/2/84/18) before the conduct of the study. The Participants Information sheet was provided, and the informed consent form was also signed by the participants.