Sage Journals: Discover world-class research

Abstract

Objective

Dietary fiber and physical activity are independent determinants of blood pressure. However, their interaction regarding hypertension remains poorly understood. This study examined the association between dietary fiber intake and hypertension. It also aimed to determine if this relationship is modified by physical activity in a representative sample of U.S. adults.

Methods

This cross-sectional study used data from 26,556 U.S. adults (National Health and Nutrition Examination Survey 2007–2018). Dietary fiber intake was assessed via two 24-h recalls and categorized into quartiles. Hypertension was defined by measured blood pressure, self-reported diagnosis, or antihypertensive medication use. Physical activity was quantified and divided into three levels. Missing covariate data were addressed using multiple imputations. Survey-weighted multivariable logistic regression was used to examine the association. A multiplicative interaction term was introduced to assess effect modification by physical activity. An XGBoost was also developed and interpreted using SHapley Additive exPlanation to explore nonlinear associations.

Results

In the fully adjusted model, lower dietary fiber intake was associated with increased odds of hypertension (OR for Q3 vs. Q4: 1.25; 95% CI [1.08, 1.44]). This relationship showed a potential interaction with physical activity (p for interaction = .067). Stratified analysis showed no association in the low or moderate physical activity groups. Among individuals with high physical activity, lower fiber intake was associated with significantly higher odds of hypertension (OR for Q1 vs. Q4: 1.50; 95% CI [1.10, 2.04]). The machine learning model confirmed that higher fiber intake was protectively associated with hypertension, an effect most prominent in physically active individuals.

Conclusion

The protective association of dietary fiber with hypertension was observed primarily in highly active adults, but not in less active individuals. These findings highlight the importance of the diet-activity interplay for hypertension prevention.

Keywords

Hypertension dietary fiber physical activity effect modification NHANES

Introduction

Hypertension is a leading risk factor for cardiovascular disease globally.^1–3 Dietary modification is a primary strategy for its prevention and management.^4–7 Dietary fiber, a key component of a healthy diet, has been associated with cardiovascular benefits.^8–10 The relationship between dietary fiber and hypertension may, however, be influenced by other lifestyle behaviors.^11–13 Physical activity, in particular, is a critical determinant of blood pressure.^14,15 Therefore, clarifying the association between dietary fiber and hypertension, while considering the modifying role of physical activity, is essential.

Numerous epidemiological studies and meta-analyses report an inverse association between dietary fiber intake and the risk of hypertension.^16,17 The proposed biological mechanisms for this protective effect include modulation of the gut microbiota, reduction of systemic inflammation, and improved insulin sensitivity.^17,18 However, the strength of this association may differ across populations and study designs.¹⁹ Critically, most research has investigated the effects of dietary fiber in isolation. Separately, physical activity is a well-established lifestyle factor for blood pressure control.¹⁵ Consistent evidence shows that regular aerobic exercise effectively lowers both systolic and diastolic blood pressure in adults with hypertension.²⁰

Despite the independent benefits of dietary fiber and physical activity, their potential interaction on hypertension is not well understood.^21,22 Composite dietary indices aggregate multiple nutrients. This aggregation masks the specific biological effects of individual components. Dietary fiber uniquely promotes short-chain fatty acid production via gut fermentation.^17,18 Physical activity independently enhances this microbial pathway.²³ Focusing on fiber isolates this specific mechanistic synergy. It avoids confounding from other nutrients present in broad dietary patterns. These lifestyle behaviors are often concurrent, and their combined influence on blood pressure may not be merely additive. Previous observational studies have typically treated physical activity and diet as separate exposures or adjusted for one as a confounder, rather than investigating their interplay.^24,25 Moreover, some analyses have been limited to specific populations or may have lacked the statistical power to detect meaningful interactions.^26,27 The National Health and Nutrition Examination Survey (NHANES) uses a complex sampling design to provide a nationally representative sample of the U.S. adult population. This database therefore offers a unique opportunity to address this research gap.

This study aimed to investigate the association between dietary fiber intake and hypertension, and to determine whether this association is modified by physical activity. The analysis utilized data from a large, nationally representative sample of U.S. adults from the 2007–2018 NHANES. Multivariable logistic regression was used to quantify the association and test for statistical interaction. An eXtreme Gradient Boosting (XGBoost) machine learning model was also developed to explore complex, nonlinear relationships. This dual-method approach provides a comprehensive assessment of the interplay between diet and physical activity. It was hypothesized that the association between dietary fiber intake and hypertension would differ significantly across levels of physical activity. The findings may inform more specific public health strategies for hypertension prevention.

Materials and methods

Study design and population

This cross-sectional study utilized data from the NHANES. NHANES is a continuous, cross-sectional survey designed to assess the health and nutritional status of the noninstitutionalized civilian population in the United States.^28–30 The survey is conducted by the National Center for Health Statistics (NCHS).³¹ A complex, multistage probability sampling design was employed to obtain a nationally representative sample.³² The reporting of this study conforms to the STROBE guidelines.³³

Data were collected through in-home interviews and physical examinations conducted in mobile examination centers (MECs). The NCHS Research Ethics Review Board approved all NHANES protocols. The study adhered to the Declaration of Helsinki of 1975, as revised in 2024. Written informed consent was obtained from all participants before data collection. All participant data were deidentified by the NCHS prior to public release.

This analysis included data from six consecutive NHANES cycles (2007–2018). Data collection for the subsequent cycle was disrupted by the COVID-19 pandemic. The pandemic severely altered population-level physical activity and dietary habits. Inclusion of post-2018 data introduces significant confounding bias. The 2007–2018 timeframe ensures the behavioral data represent typical, prepandemic lifestyle patterns. This restriction maintains the internal validity of the behavioral assessments. The study specifically included adults aged 18 years or older. Participants were selected using a complex multistage probability design. This method ensures representation of the noninstitutionalized U.S. civilian population. The initial sample consisted of 59,842 participants. Individuals aged under 18 years were excluded (n = 23,262). Participants with incomplete or invalid blood pressure data were subsequently removed (n = 5854). Pregnant or lactating individuals were excluded (n = 507). Participants lacking dietary interview sample weights were also omitted (n = 3663). The final analytical sample comprised 26,556 participants. A detailed selection process is presented in Figure 1.

Figure 1.

Flowchart of participant selection from the NHANES 2007–2018.

Assessment of dietary fiber intake

Dietary fiber intake was assessed using data from two 24-h dietary recall interviews. The first recall was conducted in-person at the MEC. A second recall was conducted by telephone 3–10 days later. The average daily dietary fiber intake was calculated from the two recalls. This average provided a stable estimate of usual consumption. For regression analyses, participants were categorized into quartiles based on their average fiber intake. This threshold aligns with the 2017 ACC/AHA Guideline.³⁴ The quartiles were defined as Q1 (lowest), Q2, Q3, and Q4 (highest).

Ascertainment of hypertension

Blood pressure was measured by trained technicians at the MEC. Up to three consecutive blood pressure readings were obtained for each participant. The average of available readings was used for the analysis. Participants were classified as having hypertension if they met one or more of the following conditions. The first condition was an average systolic blood pressure of 130 mmHg or higher, or an average diastolic blood pressure of 80 mmHg or higher. The second was a self-reported history of a physician's diagnosis of hypertension. The third was the current use of prescribed antihypertensive medication.

Assessment of covariates

Covariate information was obtained from participant interviews, questionnaires, and physical examinations. The primary effect modifier, physical activity, was assessed using the Physical Activity Questionnaire. Total physical activity was quantified in metabolic equivalent of task (MET) minutes per week. This value was calculated by summing the durations of work, transport, and recreational activities. Each activity was weighted by its standard MET value, corresponding to moderate (4.0 METs) or vigorous (8.0 METs) intensity. For analysis, this continuous variable was categorized into three levels: Low, Moderate, and High.^35,36 Sociodemographic covariates included age (continuous), gender (Male, Female), and race/ethnicity (Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, Other Race/Multi-Racial). Education was classified as less than high school, high school graduate, or more than high school. The family poverty income ratio (PIR) was categorized into three levels: Low, Middle, and High. Lifestyle factors included smoking status (never, former, or current smoker) and alcohol consumption. Alcohol intake was derived from the alcohol use questionnaire based on consumption frequency and quantity, and was treated as a continuous variable (grams per day). Body mass index (BMI) was also included as a continuous variable. The average daily intakes of total energy, sodium, and potassium were calculated from two 24-h recalls and included as continuous covariates. Covariates were selected a priori based on established risk factors identified in previous epidemiological literature. Variables potentially acting as mediators, such as diabetes and chronic kidney disease, were excluded to prevent overadjustment bias. Detailed definitions for all covariates are provided in Supplemental Table S1.

Statistical analysis

All statistical analyses were performed using R software (version 4.4.3). The complete analysis code is publicly available on GitHub (https://github.com/Michael1006-dev/fiber-activity-interaction-nhanes). All analyses incorporated the complex, multistage sampling design of NHANES by applying a six-cycle dietary interview weight, strata, and primary sampling units. The extent of missing data for covariates is detailed in Supplemental Table S2. Missingness was addressed using multiple imputation by chained equations (MICE) to generate five complete datasets. All final results were pooled from these five datasets according to Rubin's rules. A detailed description of the MICE procedure is provided in Section 2 of the Supplemental Materials.

Baseline characteristics of the study population were presented according to hypertension status. Continuous variables were reported as mean ± standard deviation, and categorical variables were reported as number (percentage). Survey-adjusted chi-square tests and analyses of variance were used to assess differences between groups for categorical and continuous variables, respectively.

Survey-weighted multivariable logistic regression models were used to calculate the odds ratios (ORs) and 95% confidence intervals (CIs) for the association between dietary fiber intake and hypertension. The highest quartile of dietary fiber intake (Q4) served as the reference group. In these models, all covariates with a continuous scale were treated as continuous variables. Three models were constructed with sequential adjustments. Model 1 was the crude model with no adjustments. Model 2 was adjusted for age, gender, and race/ethnicity. Model 3 was further adjusted for education, family PIR, BMI, smoking status, physical activity level, alcohol intake, and the average daily intake of total energy, sodium, and potassium. To assess for a dose–response relationship, a p for trend was calculated by treating the fiber intake quartiles as a continuous variable in the models.

The potential effect modification by physical activity was evaluated by introducing a multiplicative interaction term between dietary fiber quartiles and physical activity levels in the fully adjusted model. Stratified analyses were then conducted within each level of physical activity. A two-sided p-value less than .05 was considered statistically significant for all analyses.

Machine learning model interpretation

To complement the traditional regression analysis and explore complex, nonlinear relationships, an XGBoost model was developed to predict hypertension. The analysis was conducted on each of the five imputed datasets. To retain maximal information, all predictors with a continuous scale, including age and BMI, were utilized as continuous variables in the models. For each dataset, data were partitioned into training, validation, and testing sets. The model's hyperparameters were optimized using five-fold cross-validated grid search with an early stopping criterion. Model interpretation was performed using SHapley Additive exPlanations (SHAP) to quantify the impact of each predictor on the model output. A detailed description of the model development and SHAP analysis is provided in Section 3 of the Supplemental Materials. The code for the XGBoost model development and SHAP analysis is available at the same repository.

Results

Baseline characteristics of the study population

The analysis included 26,556 participants. Of these, 13,204 individuals had hypertension and 13,352 did not. The baseline characteristics of the study population, stratified by hypertension status, are presented in Table 1. The mean dietary fiber intake was 16.72 g/day for participants with hypertension and 17.35 g/day for those without hypertension.

Table 1.

Baseline characteristics of the study population according to quartiles of dietary fiber intake.

Variables	Level	Hypertension	No hypertension	p	SMD
N		13,352	13,204
Age (%)	18–39	7548 (55.6)	1901 (16.5)	<.001	0.994
	40–59	4001 (32.9)	4543 (41.1)
	≥60	1803 (11.6)	6760 (42.3)
Gender (%)	Male	6264 (46.6)	6945 (52.6)	<.001	0.119
Gender (%)	Female	7088 (53.4)	6259 (47.4)
Race (%)	Mexican American	2368 (10.8)	1654 (6.5)	<.001	0.21
	Other Hispanic	1533 (6.6)	1251 (4.5)
	Non-Hispanic White	5387 (64.7)	5669 (68.6)
	Non-Hispanic Black	2292 (9.6)	3353 (13.2)
	Other Race/Multi-Racial	1772 (8.3)	1277 (7.1)
Education (%)	Less than high school	2957 (14.4)	3416 (16.4)	<.001	0.087
	High school graduate/GED	3031 (22.3)	3231 (24.5)
	More than high school	7364 (63.3)	6557 (59.1)
PIR (%)	Low (<1.3)	4504 (23.7)	4124 (20.8)	.001	0.085
	Middle (1.3–3.5)	4826 (33.5)	5151 (36.9)
	High (≥3.5)	4022 (42.8)	3929 (42.3)
BMI (%)	Normal （<25 kg/m²）	5133 (39)	2647 (19.6)	<.001	0.502
	Overweight (25–29.9 kg/m²)	4378 (33.1)	4298 (32)
	Obese (≥30 kg/m²)	3841 (27.9)	6259 (48.3)
Smoking_status (%)	Never smoker	8090 (60.3)	6880 (51.7)	<.001	0.262
	Former smoker	2331 (19.2)	3894 (30.4)
	Current smoker	2931 (20.6)	2430 (17.9)
Met_total_Q3 (%)	Low (<600 MET-min/week)	4267 (28.3)	6057 (41.3)	<.001	0.279
	Moderate (600–3000 MET-min/week)	4242 (34.3)	3623 (29.7)
	High (≥3000 MET-min/week)	4843 (37.4)	3524 (29)
Dietary_fiber (g, mean (SD))		17.35 (9.58)	16.72 (8.51)	.002	0.07
Potassium (mg, mean (SD))		2668.11 (1122.55)	2650.30 (1038.47)	.439	0.016
Sodium (mg, mean (SD))		3517.52 (1489.93)	3427.53 (1452.55)	.003	0.061
Energy (kcal, mean (SD))		2119.94 (802.57)	2060.33 (794.45)	<.001	0.075
Daily_alcohol (g, mean (SD))		9.69 (150.39)	13.33 (202.56)	.152	0.02

Note: Descriptive statistics were generated from the first of five multiply imputed datasets. Data are presented as mean (SD) for normally distributed continuous variables and n (%) for categorical variables. All means, standard deviations, and percentages are weighted to be representative of the U.S. population. p-values were calculated using weighted linear regression for continuous variables and weighted chi-square tests for categorical variables to test for differences across dietary fiber intake quartiles. SD: standard deviation; BMI: body mass index; PIR: poverty income ratio; MET: metabolic equivalent of task.

Significant differences were observed between the two groups for most baseline characteristics. Compared to individuals without hypertension, participants with hypertension were older, with 42.3% aged 60 years or older versus 11.6% in the nonhypertensive group. The hypertension group had a higher proportion of males (52.6% vs. 46.6%) and Non-Hispanic Black individuals (13.2% vs. 9.6%). Participants with hypertension also had a significantly higher prevalence of obesity (48.3% vs. 27.9%). Furthermore, the hypertension group had a larger proportion of former smokers, individuals with lower education levels, and lower levels of physical activity. Dietary intakes of fiber, sodium, and total energy were also significantly lower in the group with hypertension. No significant between-group differences were found for potassium or daily alcohol intake.

Association between dietary fiber intake and hypertension

Table 2 shows the results from the multivariable logistic regression analyses. In the crude model, participants in the lowest dietary fiber intake quartile (Q1) had higher odds of hypertension compared to those in the highest quartile (Q4) (OR: 1.16; 95% CI [1.03, 1.30]). This association became stronger after adjustment for age, gender, and race (Model 2) (OR: 1.49; 95% CI [1.29, 1.73]).

Table 2.

Logistic regression analysis of the association between dietary fiber intake and hypertension.

Dietary_fiber	Total No.	No. of cases (%)	OR [95% CI]
Dietary_fiber	Total No.	No. of cases (%)	Model 1	Model 2	Model 3
Q4 (highest)	6382	2991 (46.87)	Ref	Ref	Ref
Q1 (lowest)	7247	3698 (51.03)	1.16 [1.03–1.30]	1.49 (1.29–1.73)	1.21 (1.00–1.47)
Q2	6622	3316 (50.08)	1.14 [1.03–1.27]	1.29 (1.14–1.46)	1.08 (0.92–1.28)
Q3	6305	3199 (50.74)	1.25 [1.12–1.39]	1.35 (1.18–1.55)	1.25 (1.08–1.44)
p for trend			0.068	p < .001	.225

Note: Results are pooled ORs and 95% CIs from five multiply imputed datasets. Model 1 is the crude, unadjusted model. Model 2 was adjusted for age, gender, and race. Model 3 was adjusted for the variables in Model 2 plus education level, poverty income ratio, body mass index, smoking status, physical activity level, total energy intake, sodium intake, potassium intake, and daily alcohol intake. The p for trend was calculated by treating the dietary fiber quartiles as a continuous variable in each respective model. OR: odds ratio; CI: confidence interval.

In the fully adjusted model (Model 3), the associations were attenuated. The odds of hypertension were significantly higher for participants in the third quartile (OR: 1.25; 95% CI [1.08, 1.44]) relative to the reference group. The association for the lowest quartile was at the threshold of statistical significance (OR: 1.21; 95% CI [1.00, 1.47]). No significant association was observed for the second quartile. A statistically significant linear trend was present in Model 2 (p for trend < .001), but this trend was not significant in the fully adjusted model (p for trend = .225).

Effect modification by physical activity

The analysis tested for an interaction between dietary fiber intake and physical activity on the odds of hypertension. A potential interaction was observed (p for interaction = .067). Stratified analyses were subsequently performed across low, moderate, and high physical activity levels. Table 3 presents the detailed results, which are visualized in Figure 2.

Figure 2.

Association between dietary fiber intake and hypertension stratified by physical activity level. Points represent the odds ratios derived from survey-weighted logistic regression models. Error bars indicate the 95% confidence intervals. The fourth quartile of dietary fiber intake serves as the reference group. The model adjusts for age, gender, race, education, poverty income ratio, body mass index, smoking status, alcohol consumption, and dietary intakes of energy, sodium, and potassium. Exact numerical values are provided in Table 3.

Table 3.

Stratified analysis of the association between dietary fiber intake and hypertension by physical activity level.

Dietary fiber intake	Physical activity level
	Low	Moderate	High
	OR [95% CI]	OR [95% CI]	OR [95% CI]
Q4 (highest)	Ref	Ref	Ref
Q1 (lowest)	1.03 [0.74, 1.45]	1.06 [0.73, 1.56]	1.50 [1.10, 2.04]
Q2	0.94 [0.72, 1.21]	0.96 [0.69, 1.34]	1.31 [1.01, 1.70]
Q3	1.06 [0.83, 1.35]	1.21 [0.92, 1.61]	1.44 [1.12, 1.85]
p for trend	.919	.860	.039
p for interaction	.067

Note: Data are presented as ORs and 95% CIs. All models were adjusted for age, gender, race, education level, poverty income ratio, BMI, smoking status, average daily intake of potassium, sodium, and energy, and daily alcohol consumption. Dietary fiber intake was categorized into quartiles, with Quartile 4 (the highest intake) serving as the reference group. Physical activity was categorized into tertiles (Low, Moderate, High). The p for interaction was derived from a likelihood ratio test comparing the fully adjusted model with and without the multiplicative interaction term between dietary fiber (as a continuous variable) and physical activity (as a categorical variable). A single p-value is reported for the overall interaction. OR: odds ratio; CI: confidence interval; BMI: body mass index.

No significant association between dietary fiber intake and hypertension was found in the low or moderate physical activity groups. Dose–response analysis within the low physical activity stratum revealed no significant trend (p for trend = .919). Even participants in the highest fiber quartile (Q4) showed no reduction in hypertension odds compared to the lowest quartile (Q1). In contrast, a significant association was evident in the high physical activity group. Within this stratum, participants in the lowest (Q1), second (Q2), and third (Q3) quartiles of fiber intake had higher odds of hypertension compared to those in the highest quartile (Q4). The corresponding ORs were 1.50 (95% CI [1.10, 2.04], 1.31 (95% CI [1.01, 1.70]), and 1.44 (95% CI [1.12, 1.85]). A significant dose–response trend was also observed exclusively in the high physical activity stratum (p for trend = .039). Sensitivity analysis using the traditional 140/90 mmHg threshold showed an attenuation of statistical significance (p for interaction = .303), likely due to the reduced number of cases. However, the direction of the association in the high physical activity group remained consistent (Supplemental Table S4). To address potential reverse causation, a sensitivity analysis excluded participants with a self-reported diagnosis or medication use. The analysis was restricted to individuals with undiagnosed hypertension and normotensive controls. The interaction between dietary fiber and physical activity remained statistically significant. The protective association in the high physical activity group was consistent with the primary findings (Supplemental Table S5).

Machine learning-based interpretation of predictors for hypertension

The XGBoost model demonstrated good predictive performance for hypertension, with a mean area under the receiver operating characteristic curve of 0.82 (Supplemental Table S3). Figure 3 displays the SHAP summary plot, which ranks predictors based on their overall impact on the model's output. Age and BMI were the two most important predictors. Higher values for both variables were associated with an increased likelihood of hypertension. Dietary fiber intake (Fiber_avg) and total physical activity (Met_total) were also identified as important predictors. Higher fiber intake was consistently associated with a decreased likelihood of hypertension. The effect of total physical activity was more complex. While low physical activity had a minimal impact on the prediction, high physical activity was associated with a decreased likelihood of hypertension for many individuals, though not universally.

Figure 4 presents SHAP dependence plots to further explore the complex relationships of key predictors with hypertension. Figure 4(a) demonstrates a strong, positive linear relationship between age and its corresponding SHAP value. This relationship was modified by gender. Before approximately 50 years of age, males generally had higher SHAP values than females. This pattern inverted among older participants, where females exhibited higher SHAP values than males of the same age.

Figure 3.

SHAP summary plot of predictor importance for hypertension. This plot summarizes the SHAP values for all predictors of hypertension. Each point corresponds to a single individual for a specific feature. The x-axis represents the SHAP value, where positive values indicate a higher predicted risk of hypertension. Features are ordered on the y-axis by their mean absolute SHAP value. The color scale shows the original feature value, from low (blue) to high (red).

Figure 4.

SHAP dependence plots for key predictors of hypertension. Each plot illustrates how a feature's value (x-axis) influences its impact on the predicted risk of hypertension (SHAP value, y-axis). Each point represents a single individual, and the color is determined by a secondary, interacting feature's value. The plots display the dependence for age with interaction from male gender (a), BMI with interaction from age (b), daily fiber intake with interaction from total physical activity (c), and total physical activity with interaction from age (d).

A strong positive association was also observed between BMI and its SHAP value (Figure 4(b)). This relationship was interactive with age. For individuals with a BMI below approximately 30 kg/m², older age was associated with a greater contribution to hypertension risk. Conversely, in the obesity range (BMI > 30 kg/m²), younger age was associated with higher SHAP values for a given BMI.

The dependence plot for dietary fiber shows an overall negative association with SHAP values, indicating a protective effect (Figure 4(c)). This effect appeared more pronounced at intakes above a range of 15–20 g/day. The plot also reveals that for a given level of fiber intake, particularly at higher levels, greater physical activity was associated with more negative SHAP values.

The relationship between total physical activity and hypertension risk was highly nonlinear (Figure 4(d)). SHAP values were highest among individuals with sedentary-to-low activity levels. The values then generally decreased as physical activity increased to approximately 10,000 MET-min/week. Beyond this level, an interaction with age was observed. At very high levels of physical activity, older individuals had lower SHAP values than younger individuals.

Discussion

In this study, the association between dietary fiber intake and hypertension was significantly modified by physical activity. A protective association of higher fiber intake was evident only among individuals with high levels of physical activity. This relationship was absent in participants with low or moderate physical activity levels. These findings from traditional regression models were supported by a complementary machine learning analysis. An XGBoost model identified both dietary fiber and physical activity as important protective predictors. Furthermore, model interpretation using SHAP demonstrated that the protective effect of dietary fiber was most pronounced in the context of greater physical activity, suggesting a potential interplay.

The primary finding of this study suggests that physical activity may modify the association between dietary fiber intake and hypertension. The protective effect of high dietary fiber intake was only observed among individuals with high levels of physical activity. This suggests a potential interplay between these two lifestyle factors. Prior research has established that healthy dietary patterns are associated with a reduced risk of hypertension.³⁷ Many studies, however, have examined dietary factors in isolation. They often neglect the potential interactions with other lifestyle components, such as physical activity. The current study addresses this gap by demonstrating a significant effect modification in a large, nationally representative population. This finding may help explain inconsistencies in previous literature, as the background physical activity levels of study populations could influence the observed diet–hypertension associations. The concept of lifestyle synergy is supported by other research exploring the joint effects of diet and physical activity on cardiovascular endpoints.^38,39 These findings have specific implications for sedentary populations. Dietary fiber modification alone appears insufficient for hypertension prevention in the absence of physical activity. Physical activity may be a physiological prerequisite to unlock the cardiovascular benefits of fiber. Consequently, clinical advice for sedentary patients should emphasize that dietary improvements may yield limited results without concurrent increases in activity. Conversely, highly active individuals maintain a residual risk of hypertension due to age or genetics. Increasing fiber intake provides an additive benefit for this group, further optimizing their cardiovascular profile.

The synergistic effect of dietary fiber and physical activity on blood pressure may be explained by several biological mechanisms. This synergy is visually supported by the SHAP analysis, which showed that for any given fiber intake, higher physical activity corresponded to a greater reduction in predicted hypertension risk. First, both factors beneficially impact the gut microbiome. Dietary fiber fermentation produces short-chain fatty acids, which have antiinflammatory and blood pressure-regulating properties.⁴⁰ Independently, physical activity also promotes a healthier gut microbiota composition.^23,41 The combination may therefore amplify improvements in the gut microbial environment and reduce systemic inflammation. Second, both interventions are known to improve insulin sensitivity and vascular endothelial function.^42–44 Enhanced insulin sensitivity and improved nitric oxide bioavailability are critical for blood pressure control. The concurrent improvement of these pathways through both diet and exercise could produce a more potent antihypertensive effect than either factor alone.⁴²

The importance of this stratified analysis is underscored when examining the main effect of dietary fiber across the entire population. In the fully adjusted analysis, the inverse association between total dietary fiber intake and hypertension was attenuated. A significant dose–response trend was not observed. This finding differs from some epidemiological studies and meta-analyses that report a more definitive protective effect of dietary fiber against hypertension.^7,45 One explanation for this attenuated effect is the extensive covariate adjustment in the final model. The model controlled for strong potential mediators, including BMI, total energy intake, sodium, and potassium. The independent effect of fiber is reduced when the influence of these pathways is accounted for. Residual confounding from unmeasured variables may also contribute an inherent limitation of observational research. However, the primary explanation is the significant effect heterogeneity demonstrated in this study. The analysis of the total population aggregates individuals with distinct physical activity levels. This aggregation can mask or dilute a true association that exists only within a specific subgroup. The clear, significant association observed exclusively in the high physical activity stratum supports this conclusion.

The analysis was limited by the use of total dietary fiber intake. This measure does not distinguish between different fiber types or their food sources. Different fiber types, specifically soluble and insoluble fiber, have distinct physiological properties.⁴⁶ The health effects of fiber may also vary depending on its origin, such as cereals, fruits, or vegetables. For example, fiber from whole grains has been independently associated with a lower risk of hypertension.^19,47 Studies have also shown that fiber from fruits and legumes is inversely related to metabolic syndrome risk factors.^46,48 The synergistic effect between fiber and physical activity may therefore be stronger for specific fiber types. Future studies should investigate whether the observed interaction is driven by fiber from particular food sources.

Beyond testing the primary hypothesis, this study employed an XGBoost model with SHAP interpretation. This approach was used to explore nonlinear relationships and complex interactions among hypertension predictors. The machine learning methods provided insights beyond the constraints of traditional linear regression models.^49,50 The model identified a significant interaction between age and gender. It captured the well-documented crossover phenomenon in hypertension risk around the age of 50. This pattern is consistent with the loss of estrogen's protective cardiovascular effects in women during the menopausal transition.^51,52 Before this age, men typically exhibit higher hypertension risk, a trend that reverses in older age groups.^53,54 This concordance with established pathophysiology validates the biological plausibility of the model. It also underscores the necessity of considering the age–gender interaction in cardiovascular risk assessment.

The analysis also revealed novel interactions between predictors. A key finding involved the relationship between BMI and age. Among individuals with obesity (BMI > 30 kg/m²), younger age was associated with a greater contribution to hypertension risk. This finding suggests that early-onset obesity may have a more pronounced adverse cardiovascular impact.⁵⁵ The marginal risk contribution of obesity may be attenuated in older individuals who already have a high baseline risk from ageing. In contrast, obesity in younger individuals may signify a more severe underlying pathophysiological state.^56,57 A distinct interaction was also observed between physical activity and age. At very high levels of physical activity, older individuals demonstrated a lower hypertension risk contribution compared to younger individuals. This implies that high-intensity physical activity may provide a substantial risk-offsetting benefit in older adults. The relative protective value of such activity appears to be greater than in the young.

A sensitivity analysis using the 140/90 mmHg threshold revealed that the interaction was no longer statistically significant (Supplemental Table S4). This attenuation suggests that the synergistic protection of dietary fiber and physical activity is most pronounced in preventing Stage 1 Hypertension (130–139/80–89 mmHg). This finding is clinically relevant. The 2017 ACC/AHA guidelines specifically prioritize lifestyle modifications, rather than pharmacotherapy, for managing this early hypertensive stage. Therefore, the observed diet–activity interplay may be a critical strategy specifically for early intervention. Reverse causation is a concern in cross-sectional studies. Individuals with known hypertension may adopt healthier behaviors. A sensitivity analysis excluded participants with diagnosed hypertension or medication use. The results confirmed the primary findings. The interplay between fiber and physical activity persisted in the population with undiagnosed hypertension. This supports the robustness of the observed association.

This study has several strengths. The analysis was based on a large, nationally representative sample. It controlled for a comprehensive set of confounding variables. The use of stratified analysis identified a significant interaction between dietary fiber and physical activity. These findings suggest that public health recommendations for hypertension prevention should consider the synergy between diet and exercise. The study also has limitations. The cross-sectional design prevents the determination of a causal relationship between dietary fiber intake and hypertension. Despite extensive adjustments, the potential for residual confounding from unmeasured variables remains. Reliance on self-reported data constitutes a limitation. Dietary recalls involve measurement error. Questionnaires lack the objectivity of device-based assessments. These measurement characteristics potentially influence the statistical precision of the interaction term. Future research should include prospective cohort studies and randomized controlled trials to confirm these findings. Such studies could also investigate the effects of different fiber types and specific domains of physical activity on blood pressure.

Conclusion

This cross-sectional study investigated the association between dietary fiber intake and hypertension. The analysis utilized a nationally representative sample of U.S. adults. Higher dietary fiber intake was associated with lower odds of hypertension. However, this association was observed only among adults with high levels of physical activity. No association was found in individuals with low or moderate physical activity. These findings suggest that physical activity may modify the protective potential of dietary fiber. This relationship appears absent in less active adults. Public health strategies might consider the potential interplay between diet and exercise. Future prospective studies are required to confirm these observations.

Supplemental Material

sj-docx-1-sci-10.1177_00368504261435091 - Supplemental material for Physical activity modifies the association between dietary fiber and hypertension: A cross-sectional study of U.S. adults from NHANES 2007–2018

Supplemental material, sj-docx-1-sci-10.1177_00368504261435091 for Physical activity modifies the association between dietary fiber and hypertension: A cross-sectional study of U.S. adults from NHANES 2007–2018 by Zihao Li, Xiangyu Wang and Ke He in Science Progress

Footnotes

Abbreviations

Ethics approval and consent to participate

This study was based on publicly available data from the NHANES. The original survey protocols were approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided written informed consent. Therefore, no further ethics approval was required for this analysis.

Author contributions

Z.L. conceptualized the study, conducted the formal analysis, and wrote the original manuscript draft; K.H. supervised the project and critically reviewed and revised the manuscript; and X.W. contributed to the interpretation of data and the revision of the manuscript. All authors read and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Availability of data and materials

This study utilized data from the NHANES for the 2007–2018 cycles. The datasets are publicly available for download from the National Center for Health Statistics website. Detailed descriptions of the variables used in this analysis are provided in Supplemental Table S1. The complete R and Python code for all data processing and statistical analyses is available in a public GitHub repository ().

Supplemental material

Supplemental material for this article is available online.

ORCID iDs

Zihao Li

Xiangyu Wang

Ke He

References

Faraci

Scheer

. Hypertension: causes and consequences of circadian rhythms in blood pressure. CircRes 2024; 134: 810–832.

Hazuková

Zadák

Pleskot

, et al. Oxidative DNA damage and arterial hypertension in light of current ESC guidelines. Int J Mol Sci 2024; 25: 16.

Landsbergis

Gilbert-Ouimet

Trudel

, et al. Prevention of hypertension due to long working hours and other work hazards is needed to reduce the risk of cardiovascular disease. Scand J Work Environ Health 2025; 51: 48–52.

Ribeiro

Teixeira

Alves

, et al. Lifestyle medicine as a treatment for resistant hypertension. Curr Hypertens Rep 2023; 25: 313–328.

Vemuganti

Arumugam

. Much ado about eating: intermittent fasting and post-stroke neuroprotection. J Cereb Blood Flow Metab 2021; 41: 1791–1793.

Sriperumbuduri

Welling

Ruzicka

, et al. Potassium and hypertension: a state-of-the-art review. Am J Hypertens 2024; 37: 91–100.

Nepali

Suresh

Pikale

, et al. Hypertension and the role of dietary fiber. Curr Probl Cardiol 2022; 47: 101203.

Chen

Tao

Wang

. Association of dietary fiber intake with all-cause and cardiovascular mortality in diabetes and prediabetes. Diabetol Metab Syndr 2025; 17: 9.

Han

Jiao

Zhang

, et al. Dietary fiber prevents obesity-related liver lipotoxicity by modulating sterol-regulatory element binding protein pathway in C57BL/6J mice fed a high-fat/cholesterol diet. Sci Rep 2015; 5: 10.

10.

Surendiran

Alsaif

Kapourchali

, et al. Nutritional constituents and health benefits of wild rice (Zizania spp. Nutr Rev 2014; 72: 227–236.

11.

Zhang

Yang

Tian

, et al. Dietary fiber and all-cause and cardiovascular mortality in older adults with hypertension: a cohort study of NHANES. J Nutr Health Aging 2022; 26: 407–414.

12.

Zhang

Chen

Zeng

, et al. Associations between consumption of dietary fibers and the risk of type 2 diabetes, hypertension, obesity, cardiovascular diseases, and mortality in Chinese adults: longitudinal analyses from the China health and nutrition survey. Nutrients 2022; 14: 20220627.

13.

Ballesteros-Vasquez

Cabrera-Pacheco

Saucedo-Tamayo

, et al. Dietary fiber, energy, sodium, potassium and calcium intake and its relationship to blood pressure in normotensive male adults. Salud Publica Mexico 1998; 40: 241–247.

14.

Suriyawong

Kao

T-SA

Robbins

, et al. Examining the relationship among social cognitive theory-based constructs, physical activity, and blood pressure in young- and middle-aged adults living with hypertension. West J Nurs Res 2025; 47: 563–572.

15.

Currie

Schultz

Millar

, et al. The role of exercise blood pressure in hypertension: measurement, mechanisms, and management. Med Sci Sports Exercise 2025; 57: 425–433.

16.

Reynolds

Akerman

Kumar

, et al. Dietary fibre in hypertension and cardiovascular disease management: systematic review and meta-analyses. BMC Med 2022; 20: 139.

17.

Jama

Snelson

Schutte

, et al. Recommendations for the use of dietary fiber to improve blood pressure control. Hypertension 2024; 81: 1450–1459.

18.

Chen

Shang

Xin

, et al. Beneficial effects of psyllium on the prevention and treatment of cardiometabolic diseases. Food Funct 2022; 13: 7473–7486.

19.

Luo

Wang

, et al. Intake of dietary fiber from grains and the risk of hypertension in late midlife women: results from the SWAN study. Front Nutr 2021; 8: 730205.

20.

Malem

Ristiani

Puteh

. Brisk walking exercise has benefits of lowering blood pressure in hypertension sufferers: a systematic review and meta-analysis. Iran J Public Health 2024; 53: 774–784.

21.

Qian

Liu

Choi

, et al. Exploring the causal effects of physical activity, diet, and nutrition on hypertension and hyperlipidemia: a multivariable Mendelian randomization analysis. BMC Cardiovasc Disord 2025; 25: 196.

22.

Guarneiri

Kirkpatrick

Maki

. Protein, fiber, and exercise: a narrative review of their roles in weight management and cardiometabolic health. Lipids Health Dis 2025; 24: 237.

23.

Zheng

Liu

, et al. Liuzijue training improves hypertension and modulates gut microbiota profile. Front Cardiovasc Med 2023; 10: 1075084.

24.

Shi

Zhu

Liao

, et al. The association between dietary fiber intake and the concentrations of aldehydes in serum. Environ Sci Pollut Res 2022; 29: 25790–25798.

25.

Wang

Shen

Wang

, et al. Association between constipation and major depression in adult Americans: evidence from NHANES 2005-2010. Front Psychiatry 2023; 14: 1152435.

26.

Bae

Park

. Physical activity and eating habits are related to chronic disease in the basic livelihood security program. Nutrients 2025; 17: 62.

27.

Bernardi

Fedullo

Bernardi

, et al. Diet in neurogenic bowel management: a viewpoint on spinal cord injury. World J Gastroenterol 2020; 26: 2479–2497.

28.

Kim

Iskander

Singh

. Examining modifiable risk factors for colorectal cancer: insights from the national health and nutrition examination survey (NHANES) data. Gastroenterology 2025; 169: 2.

29.

Liu

Dai

Heianza

, et al. Oral microbiome diversity and premature mortality in US adults: findings from the national health and nutrition examination survey (NHANES) 2009–2012. Circulation 2025; 151: 2.

30.

Cheng

Thorndike

. Sodium misestimation in the national health and nutrition examination survey (NHANES): implications for surveillance. Circulation 2025; 151: 2.

31.

Murphy

Flamme

Losonczy

, et al. Hearing threshold quartiles from the 1999-2006 national health and nutrition examination surveys. J Acoust Soc Am 2025; 157: 940–946.

32.

Jiang

Huang

, et al. Antioxidants and the risk of sleep disorders: results from NHANES and two-sample Mendelian randomization study. Front Nutr 2024; 11: 9.

33.

von Elm

Altman

Egger

, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007; 370: 1453–1457.

34.

Carey

Whelton

2017 ACC/AHA Hypertension Guideline Writing Committee. Prevention, detection, evaluation, and management of high blood pressure in adults: synopsis of the 2017 American College of Cardiology/American Heart Association hypertension guideline. Ann Intern Med 2018; 168: 351–358.

35.

Zhu

Wang

. Association between joint physical activity and healthy dietary patterns and hypertension in US adults: cross-sectional NHANES study. BMC Public Health 2024; 24: 55.

36.

Pereira-Payo

Mendoza-Muñoz

Denche-Zamorano

, et al. Physical activity is associated with the incidence of depression in United States adults from the NHANES 2013-18: a cross-sectional study. Healthcare (Basel 2024; 12: 20240227.

37.

Margetts

Little

Warm

. Interaction between physical activity and diet: implications for blood pressure management in primary care. Public Health Nutr 1999; 2: 377–382.

38.

Wang

Brage

Sharp

, et al. Associations of genetic susceptibility and healthy lifestyle with incidence of coronary heart disease and stroke in individuals with hypertension. Eur J Prev Cardiol 2022; 29: 2101–2110.

39.

Gerber

Cody

Beck

, et al. Cardiorespiratory fitness and cardiovascular risk among in-patients with depression compared to healthy controls. Front Psychiatry 2023; 14: 1193004.

40.

Antza

Stabouli

Kotsis

. Gut microbiota in kidney disease and hypertension. Pharmacol Res 2018; 130: 198–203.

41.

Sun

Lei

Nie

, et al. Effects of low-carbohydrate diet and exercise training on gut microbiota. Front Nutr 2022; 9: 884550.

42.

Liu

Aylor

Liu

. Liraglutide and exercise synergistically attenuate vascular inflammation and enhance metabolic insulin action in early diet-induced obesity. Diabetes 2023; 72: 918–931.

43.

Martinez

McAngus

Dickerson

, et al. Effects of 12 weeks of chromium, Phyllanthus emblica fruit extract, and shilajit supplementation on markers of cardiometabolic health, fitness, and weight loss in men and women with risk factors to metabolic syndrome initiating an exercise and diet intervention: a randomized double-blind, placebo-controlled trial. Nutrients 2025; 17: 20250619.

44.

Lupoli

Calcaterra

Ambrosino

, et al. Effects of Mediterranean diet on endothelial reactivity in individuals with high cardiometabolic risk: a randomized controlled parallel-group preliminary trial. Biomedicines 2024; 12: 20241113.

45.

Kaye

Shihata

Jama

, et al. Deficiency of prebiotic fiber and insufficient signaling through gut metabolite-sensing receptors leads to cardiovascular disease. Circulation 2020; 141: 1393–1403.

46.

Hosseinpour-Niazi

Mirmiran

Sohrab

, et al. Inverse association between fruit, legume, and cereal fiber and the risk of metabolic syndrome: Tehran lipid and glucose study. Diabetes Res Clin Pract 2011; 94: 276–283.

47.

McKeown

Meigs

Liu

, et al. Whole-grain intake is favorably associated with metabolic risk factors for type 2 diabetes and cardiovascular disease in the Framingham Offspring Study. Am J Clin Nutr 2002; 76: 390–398.

48.

Sajjadi

Gharipour

Mohammadifard

, et al. Relationship between legumes consumption and metabolic syndrome: findings of the Isfahan Healthy Heart Program. ARYA Atheroscler 2014; 10: 18–24.

49.

Layton

. AI, machine learning, and ChatGPT in hypertension. Hypertension 2024; 81: 709–716.

50.

Banerjee

Dashtban

Chen

, et al. Identifying subtypes of heart failure from three electronic health record sources with machine learning: an external, prognostic, and genetic validation study. Lancet Digit Health 2023; 5: e370–e379.

51.

Dela Justina

Miguez

JSG

Priviero

, et al. Sex differences in molecular mechanisms of cardiovascular aging. Front Aging 2021; 2: 725884.

52.

Vatner

Zhang

Vyzas

, et al. Vascular stiffness in aging and disease. Front Physiol 2021; 12: 762437.

53.

Ghazi

Bello

. Hypertension in women across the lifespan. Curr Atheroscler Rep 2021; 23: 43.

54.

Osude

Durazo-Arvizu

Markossian

, et al. Age and sex disparities in hypertension control: the multi-ethnic study of atherosclerosis (MESA). Am J Prev Cardiol 2021; 8: 100230.

55.

Bendor

Bardugo

Pinhas-Hamiel

, et al. Cardiovascular morbidity, diabetes and cancer risk among children and adolescents with severe obesity. Cardiovasc Diabetol 2020; 19: 79.

56.

Bukhari

Yaghi

Bashir

. Stroke in young adults. J Clin Med 2023; 12: 4999.

57.

Aggarwal

Yeh

Joynt Maddox

, et al. Cardiovascular risk factor prevalence, treatment, and control in US adults aged 20 to 44 years, 2009 to march 2020. Jama 2023; 329: 899–909.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.58 MB