Obesity is a multifaceted health problem characterised by excessive fat accumulation and energy imbalance. It arises from a combination of genetic, environmental and lifestyle factors(Reference Mohamed, Ibrahim and Elkhayat1). Lifestyle alterations, particularly the widespread use of highly processed foods rich in calories and refined carbohydrates but lacking in nutrients and fibres, are increasing in prevalence globally(Reference Popkin and Ng2,Reference Clemente-Suárez, Beltrán-Velasco and Redondo-Flórez3) . These changes contribute to overeating and excess weight accumulation. In this context, dietary fibre has emerged as a significant and modifiable dietary component that may play a crucial role in obesity prevention and management.
Dietary fibre comprises a diverse range of plant-derived substances that resist digestion and absorption within the small intestine(Reference Rezende, Lima and Naves4). Despite varying definitions, consensus exists on the health benefits of fibre-rich diets, including improved intestinal function and reduced risks of chronic diseases(Reference Barber, Kabisch and Pfeiffer5). Certain fibres, such as resistant oligosaccharides, exhibit probiotic properties, influencing intestinal bacteria and promoting the production of beneficial SCFA(Reference Rezende, Lima and Naves4).
While the benefits of fibre intake in addressing obesity are recognised(Reference Dayib, Larson and Slavin6), a more nuanced exploration is necessary, based on the sex and age of populations. Biological and behavioural differences between men and women, as well as various age groups, may influence the impact of dietary fibre on weight-related outcomes. Previous studies have reported conflicting findings regarding the association between dietary fibre intake and obesity when stratified by sex, with some studies suggesting a significant inverse relationship between higher fibre intake and lower obesity risk in men(Reference Kim, Hong and Suh7), whereas others have reported opposite results(Reference Howarth, Huang and Roberts8). These disparities may arise from variations in study populations and methodologies, as well as differences between men and women in terms of body composition, hormones and dietary habits. Interestingly, research has shown that gut microbiota differ between men and women after puberty, which is modulated by sex hormones, diet, and metabolic and inflammatory states(Reference Beale, Kaye and Marques9). The gut microbiota influences weight loss, representing a potential factor contributing to gender variations in diet-related outcomes.
Given conflicting findings of previous studies, this study aimed to investigate the association between dietary fibre intake and obesity, stratified by sex and age, in a population of Japanese outpatients with type 2 diabetes. Additionally, we explored lifestyle and dietary factors that may contribute to this association.
Methods
Study population and design
This cross-sectional study was conducted among outpatients with type 2 diabetes who received treatment at clinics in Japan that participated in the Japan Diabetes Clinical Data Management Study Group (JDDM) from December 2014 to December 2019. Details regarding the JDDM and data collection software have been described previously(Reference Kobayashi, Yamazaki and Hirao10). The cohort consisted of 1565 Japanese outpatients aged 30–89 years (mean age: 62·3 (sd 11·6) years) with type 2 diabetes, of whom 987 (63·1 %) were men.
Anthropometric, dietary and lifestyle assessments
The participating clinics administered a lifestyle questionnaire, developed by JDDM, to willing outpatients. Participants self-reported their height and weight. Dietary habits were evaluated using a validated self-administered FFQ(Reference Takahashi, Yoshimura and Kaimoto11), and nutrient and food intakes were calculated using a standardised software (Eiyo-kun; Kenpakusha Co., Ltd)(Reference Horikawa, Yoshimura and Kamada12). The FFQ was validated in a Japanese population of sixty-six adults aged 19–60 years by comparison with 7-d weighed dietary records, showing an average ratio of 104 % for nutrient intake estimates between the two methods. Energy intake of ≤ 600 or ≥ 4000 kcal per d was excluded from the analysis due to concerns about potential underreporting or overreporting. Supplement intake was not recorded. Physical activity was determined using a self-administered short version of the Japanese version of the International Physical Activity Questionnaire (IPAQ)(Reference Murase13,Reference Craig, Marshall and Sjöström14) . IPAQ data are summed for each category (vigorous intensity, moderate intensity and walking) to estimate weekly physical activity time. Total weekly physical activity is calculated by multiplying the reported time for each activity by its corresponding metabolic equivalent of task (MET). The resulting MET values were converted from minutes to hours.
Statistical analysis
Categorical variables were compared using the χ 2 test and expressed as numbers with percentages. Continuous variables were compared across age tertiles using the Jonckheere–Terpstra test and between sexes using the Mann–Whitney U test, with results presented as means and standard deviations. For MET, the median was reported due to its highly skewed distribution. Binary logistic regression analysis was used to assess tertiles of dietary fibre intake and obesity, which was defined as BMI ≥ 25 kg/m2 according to the Japan Society for the Study of Obesity(15). Residual methods were used to obtain energy-adjusted values for fibre and macronutrient intakes (protein, fat and carbohydrates). Correlations between fibre intake, food groups and micronutrients were evaluated using the Pearson correlation coefficient. Demographic characteristics, lifestyle factors, and macronutrient intakes were analysed as potential confounders. Stratification was performed based on sex and age. Age was categorised as 30–58 (first tertile), 59–68 (second tertile) and 69–89 (third tertile) years. Statistical analyses were performed using SPSS software (version 27.0; IBM Corp.). P < 0·05 indicated statistically significant differences.
Results
Characteristics of study participants according to sex and age
Table 1 compares the lifestyle behaviours and intake of nutrients and food groups according to sex and age tertiles. Women had a higher daily fibre intake (12·7 g/d) compared with men (11·6 g/d). Intakes of total vegetables, fruits and soyabeans/soya products were higher in women, whereas the intake of grains was higher in men. With regard to lifestyle habits, alcohol consumption, smoking rates and physical activity were higher in men than in women. Fibre and Na intake increased with age. Older individuals exhibited higher consumption of total vegetables, fruits and soyabeans/soya products accompanied with reduced grain intake. Alcohol consumption and smoking rates decreased with age, whereas physical activity was higher in the older age tertiles.
Table 1. Basic characteristics of participants according to sex and tertiles of age
Significant differences are highlighted in bold (P < 0·05).
T1, T2 and T3, first, second and third tertiles; N/A, not applicable; OHA, oral antihyperglycemic agents; GLP, glucagon-like peptide; MET, metabolic equivalent of task.
Data are presented as mean (sd) and n (%). Physical activity (metabolic equivalent of task (MET) hours/week) as median (interquartile range).
Lifestyle behaviours according to tertiles of dietary fibre intake by sex and tertiles of age
Table 2 presents the differences in lifestyle behaviours according to tertiles of dietary fibre intake stratified by sex and age. Among men, higher fibre intake was associated with healthier lifestyle behaviours, characterised by significantly lower smoking rates (P < 0·001) and higher levels of physical activity (P < 0·001). Among women, higher fibre intake was associated with increased physical activity (P = 0·022), but not with smoking rate (P = 0·068). Individuals with higher fibre intake exhibited higher physical activity across all age groups, with the strongest association observed in the second tertile (59–68 years) (P < 0·001). Lower smoking rates were observed among those with higher fibre intake in the first (30–58 years) (P = 0·005) and second (59–68 years) (P < 0·001) tertiles. However, no significant associations were found in the highest tertile (69–89 years) (P = 0·210). Dietary fibre intake was not significantly associated with alcohol consumption in any sex or age tertile.
Table 2. Lifestyle factors according to tertiles of dietary fibre intake stratified by sex and tertiles of age
Significant differences are highlighted in bold (P < 0·05).
MET, metabolic equivalent of task; T1, T2 and T3, first, second and third tertiles.
Data are presented as median (interquartile range) and n (%).
Multivariate analysis of associations between dietary fibre intake and obesity
Table 3 presents the multivariate analysis of associations between tertiles of dietary fibre intake in all participants and stratified by sex and tertiles of age. In model 1, without any adjustments, a higher fibre intake was associated with a significantly lower risk of obesity in all participants (95 % CI: 0·337, 0·554, P trend < 0·001). This relationship remained significant in model 2, which accounted for demographic and lifestyle factors (95 % CI: 0·400, 0·695, P trend < 0·001). Even after further adjusting for macronutrient intakes in model 3, the association persisted (95 % CI: 0·439, 0·795, P trend = 0·002). Stratified analysis showed a significant inverse association in men (P trend = 0·002) and in the older age tertiles, including 59–68 years (P trend = 0·038) and 69–89 years (P trend = 0·057). No significant association was found for women (P trend = 0·338) or the youngest age tertile of 30–58 years (P trend = 0·366).
Table 3. Binary regression analysis of tertiles of dietary fibre intake and obesity in all participants and groups stratified by sex and tertiles of age
Significant differences and trends are highlighted in bold (P < 0·05).
OHA, oral antihyperglycemic agents; GLP, glucagon-like peptide; T1, T2 and T3, first, second and third tertiles.
Model 1: Unadjusted.
Model 2: Adjusted for sex (except in the analysis by sex), age (except in the analysis by age), smoking status, alcohol consumption status, insulin treatment, OHA or GLP use, physical activity (metabolic equivalent of task).
Model 3: Adjusted by model 2 plus macronutrient intakes (protein, fat and carbohydrates).
Correlations between dietary fibre intake and food groups
Online supplementary material, Supplemental Table 1, presents the Pearson correlation coefficients between dietary fibre intake and food group intake in all participants. In all participants, the top three food groups that were most strongly correlated with the fibre intake were vegetables (r = 0·814), fruits (r = 0·512) and soyabeans/soya products (r = 0·502). On the other hand, grain intake was not strongly correlated with fibre intake (r = 0·211).
Correlations between dietary fibre and micronutrient intake
Online supplementary material, Supplemental Table 2, presents the Pearson correlation coefficients between dietary fibre intake and micronutrient intake in all participants. The vitamins and minerals most strongly correlated with dietary fibre intake were folate (r = 0·931), potassium (r = 0·879) and vitamin C (r = 0·866).
Discussion
In our study of 1565 Japanese individuals with type 2 diabetes, those in the top tertile for dietary fibre intake (mean: 16·6 g/d) showed a significantly lower prevalence of obesity, after adjustments for demographic variables, lifestyle and macronutrient intakes. In the stratified analysis by sex and age, higher fibre intake demonstrated an inverse association with obesity in men and older age groups (59–89 years), but not in women or the youngest group (30–58 years). In our prior study of the same cohort, we demonstrated an inverse association between the intake of vegetables and obesity(Reference Hatta, Horikawa and Takeda16). Our current study found that vegetables had the strongest correlation with fibre among all food groups. Additionally, we explored the correlation between fibre and micronutrients, finding the strongest associations with folate, vitamin C and potassium.
Considering that vegetables were the food group most strongly correlated with fibre in our study, a closer examination of the interplay between various dietary components is required. The observed benefits of dietary fibre may be intricately linked with the overall nutrient profile of vegetables. Most vegetables are considered very-low-to-low energy-dense, have a low glycemic level and little dietary fat content, are good sources of fibre and are an essential source of vitamins, phytochemicals and minerals(Reference Dreher and Ford17). Consequently, disentangling the specific contributions of fibre and other nutrients within vegetables may provide a more nuanced understanding of their impact on obesity in individuals with type 2 diabetes. We previously demonstrated that, among twenty-one food groups, vegetable consumption demonstrated the strongest inverse association with obesity(Reference Hatta, Horikawa and Takeda16). In contrast, a prospective study conducted in a European population reported that cereals, but not vegetables or fruits, were linked to subsequent weight changes(Reference Du, Boshuizen and Forouhi18). This discrepancy may arise from differences in the types of vegetables and cereals consumed by Asian and European populations. Additionally, in our study, vegetables and fruits were the primary fibre source, while cereals were the primary source in the European study. These differences highlight the influence of regional dietary variations on weight-related outcomes.
Examining the sex-specific nuances within our study population revealed disparities in dietary patterns. Our Japanese study revealed that women exhibited a higher intake of vegetables compared with men, in line with studies conducted in other populations that showed elevated dietary fibre intake in women, particularly from vegetables(Reference Alharbi and Alarifi19–Reference Lee, Moore and Park22). This finding suggests that sex-specific dietary habits transcend geographical and cultural boundaries. Despite this difference in vegetable consumption, we found that the inverse association between fibre intake and obesity was significant only in men. A possible explanation is that, in our study population, higher fibre intake was more significantly associated with healthier lifestyle habits, including higher physical activity and non-smoking in men. Another plausible explanation underlying these differences may be distinct physiological characteristics in both genders; men mobilise more intra-abdominal fat than women during weight loss(Reference Gasteyger, Larsen and Vercruysse23). Our results align with those of a study conducted on Korean individuals with type 2 diabetes, which reported that higher dietary fibre intake in men but not women was associated with lower odds of developing obesity(Reference Kim, Hong and Suh7). The investigators suggested that this gender-specific association observed may be due to the higher fibre density in women compared with men. In other words, the fibre density in foods consumed by women may be sufficient, potentially explaining the lack of a discernible difference in fibre intake among women with and without obesity. Conversely, a study conducted in adults aged 20–59 years from the USA found that fibre intake was inversely associated with BMI only in women but not in men(Reference Howarth, Huang and Roberts8). The proposed explanations for these findings were insufficient fibre intake among men and a potentially enhanced effectiveness of fibres in women, attributed to their higher body fat in proportion to the BMI.
We observed a positive association between age and fibre intake, and the inverse association between fibre consumption and BMI was stronger in the higher age tertiles. Ageing is associated with metabolic changes, including a decline in the BMR and alterations in body composition, such as reduced muscle and fat mass. Fibre intake may counteract age-related weight gain. Dietary fibre plays a crucial role in various age-related diseases. Lower fibre intake has been associated with cardiovascular mortality, cancer incidence and impaired cognitive function and physical performance(Reference Niero, Bartoli and De Colle24).
Among all micronutrients, folate had the strongest correlation with fibre intake, followed by potassium and vitamin C. The essential micronutrients, including folate, potassium and vitamin C, are abundant in plant-based foods and play roles in cellular function, electrolyte balance and antioxidant defence, respectively. They have been inversely associated with obesity(Reference Mazaheri-Tehrani, Yazdi and Heidari-Beni25–Reference Navarrete-Muñoz, Vioque and Toledo27). Folate, a B vitamin, is involved in the remethylation of homocysteine to methionine, which is essential for maintaining normal homocysteine levels. Elevated homocysteine levels may contribute to CHD through mechanisms involving vascular endothelial cell injury, inhibition of endothelial progenitor cells, smooth muscle cell proliferation, lipid metabolism disorders, platelet adhesion, activation of inflammatory pathways and induction of oxidative stress(Reference Wang, Mo and Wu28). Moreover, folic acid supplementation reduces the BMI in individuals with higher homocysteine levels, indicating an interplay between folate, homocysteine and obesity(Reference Jafari, Gholizadeh and Sadrmanesh29).
This study had several strengths, such as the inclusion of both men and women of different ages and adjustments for lifestyle and dietary factors. It also collected data on food groups. Furthermore, the participants who volunteered were recruited from clinics located throughout Japan, resulting in a diverse and heterogeneous sample. To ensure robustness, a sensitivity analysis excluding seventy-eight outliers in the 95th percentile of MET values confirmed that the results were consistent with the primary analysis, increasing confidence in our findings. However, certain limitations should be acknowledged. The study’s cross-sectional design restricts its ability to establish causal relationships between variables and outcomes. A limitation of the FFQ is energy underreporting, which can introduce measurement errors, but it still provides valuable dietary insights(Reference Subar, Freedman and Tooze30). Although the current FFQ was validated in a general Japanese adult population (19–60 years), our study included participants with diabetes and elderly individuals. However, the FFQ has shown reliability across diverse groups and has been used in studies involving Japanese elderly individuals with diabetes(Reference Yamaoka, Araki and Tamura31–Reference Horikawa, Aida and Tanaka32). Additionally, this study did not collect information on other sociodemographic variables, including education level and socio-economic status. It also did not account for the prevalence of chronic diseases or health behaviours, such as sleep duration and eating patterns, including late dinner and eating speed, which are associated with obesity.
In conclusion, our study of Japanese individuals with type 2 diabetes found an association between higher dietary fibre intake and lower risk of obesity. A gender difference was observed, with a significant association between higher fibre intake and lower odds of obesity only in men. This may be attributed to the stronger association between increased fibre intake and healthier lifestyle in men. Furthermore, the association was significant in the older age groups, but not in the younger group. This emphasises the need for targeted public health initiatives promoting diverse fibre-rich foods to effectively manage obesity. However, these findings are specific to a Japanese population with type 2 diabetes and may not be generalisable to other populations. Further research is needed to understand sex- and age-specific factors influencing the fibre–obesity relationship in diverse populations.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S136898002500014X
Acknowledgements
The authors thank the physicians, staff and participants of JDDM for their generous contributions to this study.
Authorship
Conceptualisation: E.d’A.F., M.H., and H.S.; Methodology: E.d’A.F., C.H., M.H., Y.T., S.Y.M., and H.S.; Validation: M.H. and H.S.; Formal analysis: E.d’A.F., M.H., K.L., I.I., and H.S.; Investigation: E.d’A.F., M.H., Y.T., S.Y.M., M.T., K.F., N.K., H.M., and H.S.; Resources: S.Y.M., N.K., H.M., and H.S.; Data curation: E.d’A.F., M.H., Y.T., S.Y.M., M.T., K.F., N.K., H.M., and H.S.; Writing – original draft: E.d’A.F., M.H., K.L., I.I., K.F., and H.S.; Writing – review and editing: C.H., Y.T., M.T., K.F., N.K., H.M., and H.S.; Visualisation: C.H. and H.S.; Supervision: H.S.; Project administration: H.S.; Funding acquisition: H.S.
Financial support
This study was supported by the Japan Society for the Promotion of Science (22H03529).
Competing interests
The authors declare that they have no conflict of interest.
Ethics of human subject participation
This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the Niigata University and Health Research Involving Human Subjects in Japan (approval no.: 1927). Informed consent was obtained from all study participants.
Open access
