Impact of a comprehensive healthy lifestyle on pre-diabetes risk and glycaemic control: insights from NHANES 2005–2018

Abstract

Objective

To evaluate the association between a comprehensive multidimensional healthy lifestyle and the risk of pre-diabetes and glycaemic control markers, specifically glycated haemoglobin (HbA1c) and fasting plasma glucose (FPG), and to conduct subgroup analyses to examine these associations across different populations.

Methods

A cross-sectional analysis of 9567 adults aged 20 and older was conducted using data from the National Health and Nutrition Examination Survey from 2005 to 2018 participants. A lifestyle score was created based on eight modifiable factors: diet, physical activity, smoking, alcohol consumption, sleep duration, waist circumference, depressive symptoms and marital status. Weighted logistic regression estimated ORs for pre-diabetes, and linear regression assessed associations with HbA1c and FPG.

Results

A higher healthy lifestyle score was associated with a significantly lower risk of pre-diabetes and improved glycaemic markers. Compared with participants with 0–2 healthy lifestyle factors, those with 7–8 factors exhibited a 53% lower risk of pre-diabetes (OR: 0.47, 95% CI: 0.31 to 0.69; p<0.001, P for trend <0.001). The results showed significant reductions in HbA1c by 0.08 units (95% CI: −0.16 to −0.01) and in FPG by 2.74 mg/dL (95% CI: −4.47 to −1.01). Stratified analyses showed stronger associations in women, individuals without a family history of diabetes, and those aged 40–59. Sensitivity analyses confirmed the robustness of these findings.

Conclusions

The study results highlight the value of a multidimensional, modifiable healthy lifestyle in reducing pre-diabetes risk and improving glycaemic control. These results support integrated lifestyle interventions in population-level pre-diabetes prevention strategies.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Pre-diabetes is a precursor to diabetes and is linked to increased risks of vascular complications and premature mortality. Glycaemic control is critical in the management of pre-diabetes. While specific healthy lifestyle factors are known to reduce pre-diabetes risk and improve glycaemic control, the effect of a comprehensive healthy lifestyle management remains unclear.

WHAT THIS STUDY ADDS

• A significant negative association between a comprehensive healthy lifestyle and risk of pre-diabetes, accompanied by improvements in glycaemic control. Noteworthy differences were observed across various population subgroups.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The synergistic effect of multiple healthy lifestyle factors highlights their interactions, suggesting that a comprehensive healthy lifestyle approach could be integrated into pre-diabetes care to lower disease risk and improve glycaemic control.

Introduction

Pre-diabetes is characterised by elevated blood glucose levels that do not meet the diagnostic criteria for diabetes, and represents a critical transitional stage in the progression from normoglycaemia to diabetes.¹ Without effective intervention, approximately 50% of individuals with pre-diabetes will develop type 2 diabetes within 5 years.² Additionally, the prevalence of pre-diabetes is projected to exceed 470 million globally by 2030,³ underscoring the urgent need to identify and implement effective preventive strategies.

Pre-diabetes is not only a precursor to diabetes but also associated with increased risks of macrovascular and microvascular complications, and premature mortality.⁴ Optimising glycaemic control can effectively prevent or delay the onset of microvascular complications.^{5 6} Lifestyle interventions have demonstrated the potential to reduce the relative risk of pre-diabetes by 40%–70%,⁷ underscoring the pivotal role of lifestyle modification in disease prevention.

The American Diabetes Association (ADA) emphasises that psychosocial factors are as important as conventional healthy lifestyle factors in lifestyle management.⁸ However, previous studies have primarily focused on conventional factors such as smoking, alcohol consumption, diet, physical activity and waist circumference^{9 10} often neglecting psychosocial dimensions such as depressive symptoms and marital status. Furthermore, most existing studies have examined individual lifestyle components in isolation, providing limited insight into the potential synergistic effects of a comprehensive, multidimensional healthy lifestyle on pre-diabetes risk and glycaemic control. Moreover, the extent to which these associations may vary across demographic and clinical subgroups has been insufficiently explored.

To fill this knowledge gap, our study comprehensively evaluates the association of a multidimensional lifestyle—encompassing sleep duration, depressive symptoms, marital status, diet quality, physical activity, smoking status and alcohol consumption—and pre-diabetes risk as the primary outcome. Associations with glycaemic markers, including glycated haemoglobin (HbA1c) and fasting plasma glucose (FPG), are assessed as secondary outcomes. Furthermore, we explore whether these associations vary across population subgroups, thereby providing insight into the demographic specificity of lifestyle-based prevention strategies.

Materials and methods

Study population

This cross-sectional analysis uses data from the National Health and Nutrition Examination Survey (NHANES), a nationally representative survey operated by the National Center for Health Statistics.¹¹ NHANES uses a complex, multistage probability sampling design to assess the health and nutritional status of the US civilian, non-institutionalised population.¹²

We selected 39 749 participants aged 20 years and older from the 2005–2018 NHANES cycles. Participants were excluded if they had missing information on HbA1c and PFG (n=17 138), had a diagnosis of diabetes based on HbA1c or PFG (n=1903), missing lifestyle information (n=4968) and covariates information (n=4890), a PHQ-9 depression score of less than 10 while taking antidepressants (n=798), or normal blood glucose but were receiving diabetes treatment (n=398) to prevent the normalisation of blood glucose levels due to pharmacological interventions. This resulted in a final analytic sample of 9567 participants. A flowchart illustrating the inclusion and exclusion process is provided in online supplemental figure A1. The NHANES protocol was approved by the NCHS Research Ethics Review Board, and all participants provided written informed consent.¹³

Definition of healthy lifestyle factors

To generate a comprehensive lifestyle score, we assessed eight modifiable lifestyle factors: diet, physical activity, smoking, alcohol consumption, sleep duration, waist circumference, depression status and marital status. Details of the assessment of each lifestyle factor are provided in the supplementary material. Non-current smoking was categorised as healthy. Healthy alcohol consumption was defined as no more than one drink per day for women, two drinks per day for men,¹⁴ never drinking alcohol and currently not drinking alcohol. Because the 2005–2006 physical activity questionnaire lacked data on the number of days and minutes of physical activity, participants indicating vigorous or moderate physical activity were categorised as healthy. From 2007 to 2018, the number of days and minutes of activity were recorded, defining at least 150 min of moderate-intensity activity or 75 min of vigorous-intensity activity per week,¹⁵ or an equivalent combination of the two, considered healthy. Sensitivity analyses were conducted using the 2007–2018 physical activity data to ensure accuracy. Diet quality was assessed using the Healthy Eating Index (HEI)−2015, with scores ranging from 0 to 100. A score of 60 or above indicates a healthy diet.¹⁶ A healthy circumference was defined as less than 102 cm for men and less than 88 cm for women.¹⁷ Adequate sleep duration was defined as 7–9 hours per day.¹⁸ In the NHANES 2017–2018 cycle, sleep duration was collected separately for weekdays and weekends. Given potential differences between these two periods, weekday sleep duration was used in the main analysis to represent habitual sleep behaviour. Psychological health status is assessed by analysing depressive symptoms.^{19 20} Depression status was assessed using the PHQ-9 questionnaire; a total score exceeding 10 indicates the presence of depressive symptoms. For marital status, being married or cohabiting was considered healthy.²¹

Each lifestyle characteristic was assigned points, with a score of 1 representing that the participant was healthy and 0 otherwise. Adding up all of the eight factors constituted the final healthy lifestyle score, ranging from 0 to 8. Given the few participants with scores of 0, 1 and 8, scores of 0–2 were grouped into one category and scores of 7–8 into another to enhance the statistical analysis’ power and verify the reliability of the results.

Assessment of covariates

Covariates were obtained through questionnaires, laboratory tests and physical examinations, and included age, sex, race, education level and annual household income, Body Mass Index (BMI), high-density lipoprotein (HDL), total cholesterol (TC), triglycerides, TC/HDL-C ratio, estimated glomerular filtration rate (EGFR), serum creatinine (SCr), urinary creatinine (UCr) and blood urea nitrogen (BUN), hypertension, anaemia, arthritis, stroke, liver disease, thyroid diseases, cancer and the use of medications (eg, antihypertensive drugs, lipid-lowering drugs, aspirin, antidepressants and antitumour drugs). The low-density lipoprotein (LDL) was not adjusted as a covariate due to an excessive amount of missing data. Sensitivity analyses were performed at follow-up, with the addition of LDL as a covariate to test the robustness of the results.

Definition of pre-diabetes

Our study based on the ADA diagnostic criteria for pre-diabetes²² defines pre-diabetes as meeting one of the following conditions: (a) the fasting blood glucose level of 100–125 mg/dL (5.6–6.7 mmol/L), (b) an HbA1c level of 5.7%–6.4% or (c) the participant confirmed in the questionnaire that they had been told by a healthcare professional that they have pre-diabetes, abnormal fasting blood glucose, impaired glucose tolerance, borderline diabetes or higher-than-normal blood glucose but do not meet the criteria for a diabetes diagnosis.

Statistical analysis

Data from seven NHANES cycles (2005–2018) were combined and analysed with sample weights to maintain population representativeness. Continuous variables were presented as mean±SD or median (IQR), and categorical variables were shown as frequencies and percentages. Group comparisons were performed using ANOVA or the Kruskal-Wallis test for continuous variables and the Rao-Scott χ² test for categorical variables.

We employed weighted logistic regression to estimate the ORs and 95% CIs for pre-diabetes risk in relation to the comprehensive lifestyle score. Weighted linear regression was used to examine associations with continuous glycaemic outcomes (HbA1c and FPG). Further linear trend tests were conducted to evaluate the association between increasing lifestyle scores and outcomes. Subgroup analyses by sex, family history of diabetes, age, BMI, education level and income were conducted, and interaction effects were tested.

Four models were developed to control for potential confounders: Model 1 adjusted for demographics (age, sex, race, education, income). Model 2 further adjusted for lipid profile (HDL, total cholesterol, triglycerides, TC/HDL-C ratio), renal function (EGFR, SCr, UCr) and BUN. Model 3 additionally adjusted for comorbidities (hypertension, anaemia, arthritis, cardiovascular disease, stroke, liver disease, thyroid disease, cancer) and medication use.

Several sensitivity analyses were performed to test the robustness of our findings. First, data from 2005 to 2006 were excluded due to potential bias from changes in the physical activity questionnaire, and only data from 2007 to 2018 were included in the study. Second, low-density lipoprotein cholesterol (LDL-C) was included as an adjustment variable. Third, an analysis without weighting the data was conducted to further assess the robustness of the findings. Fourth, missing covariate data were handled using multiple imputations with the ‘mice’ package in R to evaluate the potential impact of missing values. Fifth, to assess the potential influence of weekend sleep, we replaced weekday sleep duration with weekend sleep duration in the construction of the lifestyle score.

All statistical analyses were performed in Stata statistical software, version 18.0 (StataMP), and R (4.4.3). Two-sided p<0.05 was considered statistically significant.

Results

Baseline characteristics

A total of 9567 individuals were enrolled in the analysis, with a mean±SD age of 46.7±16.9 years and 49.4% were male (online supplemental table A1).

Participants with a lower lifestyle score tended to be younger, less educated, more likely to have lower income and were more often widowed, divorced, separated or never married. They exhibited higher rates of smoking, excessive alcohol consumption, sleep duration outside the 7–9 hours range, insufficient physical activity, larger waist circumference, severe depressive symptoms (PHQ-9 score ≥10) and lower HEI-2015 scores. Baseline characteristics of participants excluded due to missing data indicated they were generally older, more likely to be female, less educated and exhibited poorer health behaviours, including higher BMI and a family history of various chronic conditions (online supplemental table A2).

Associations between the comprehensive healthy lifestyle score and pre-diabetes

A significant inverse association was observed between the comprehensive lifestyle score and pre-diabetes risk (figure 1). Compared with participants with scores of 0–2, those scoring 5 or higher showed a significant reduction in pre-diabetes risk across all models. In the fully adjusted model (Model 3), the ORs for pre-diabetes were 0.67 (95% CI: 0.47 to 0.96) for those with 5 factors, 0.57 (95% CI: 0.39 to 0.82) for 6 factors and 0.47 (95% CI: 0.31 to 0.69) for 7–8 factors, relative to the reference group with 0–2 factors (online supplemental table A3). When analysed as continuous variables, each additional one-point increment in the comprehensive lifestyle score was associated with a 12% lower risk of pre-diabetes. The trend test (P for trend <0.001) demonstrated a significant linear decline in pre-diabetes risk with increasing lifestyle scores.

Figure 1. Association between pre-diabetes and the comprehensive healthy lifestyle scores. NA, not available. ^aModel 1, adjusted for age (years), gender (male and female), race (Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, other race including multiracial), education (less than high school, high school graduate, some college and college graduate or more), income ($0–$19 999, $20 000–$44 999, $45 000–$99 999, $45 000–$99 999). College more income ($0–$19 999, $20 000–$44 999, $45 000–$99 999 and ≥$100 000). ^bModel 2 included further adjustments for high-density lipoprotein (HDL), total cholesterol, triglycerides, total cholesterol (TC)/HDL-C ratio, estimated glomerular filtration rate (EGFR), serum creatinine (SCr), urinary creatinine ratio (UCr) and blood urea nitrogen (BUN). ^cModel 3 included additional adjustments for hypertension (yes or no), anaemia (yes or no), arthritis (yes or no), cardiovascular disease (yes or no), stroke (yes or no), liver disease (yes or no), thyroid disease (yes or no), cancer (yes or no) and medication use (antihypertensive drugs, lipid-lowering drugs, aspirin, antidepressants, antitumour drugs).

Associations of the comprehensive healthy lifestyle score with HbA1c

Higher lifestyle scores were consistently associated with lower HbA1c levels across all models (figure 2; online supplemental table A4). Participants with a lifestyle score of 6 or higher showed significant reductions in HbA1c levels compared with those with scores of 0–2. After multivariable adjustment (Model 3), HbA1c levels were reduced by 0.07 units (95% CI: −0.15 to −0.01) for participants with six healthy factors and by 0.08 units (95% CI: −0.16 to −0.01) for those with 7–8 factors. The trend test confirmed a significant linear relationship between higher lifestyle scores and lower HbA1c levels (P for trend <0.001) with each additional point in the lifestyle score associated with a 0.02-unit reduction in HbA1c (95% CI: −0.03 to −0.01).

Figure 2. Association between glycated haemoglobin (HbA1c) levels and the comprehensive healthy lifestyle score. NA, not available; β, regression coefficient. ^aModel 1, adjusted for age (years), gender (male and female), race (Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, other race including multiracial), education (less than high school, high school graduate, some college and college graduate or more), income ($0–$19 999, $20 000–$44 999, $45 000–$99 999, $45 000–$99 999). College more income ($0–$19 999, $20 000–$44 999, $45 000–$99 999 and ≥$100 000). ^bModel 2 included further adjustments for high-density lipoprotein (HDL), total cholesterol, triglycerides, total cholesterol (TC)/HDL-C ratio, estimated glomerular filtration rate (EGFR), serum creatinine (SCr), urinary creatinine ratio (UCr) and blood urea nitrogen (BUN). ^cModel 3 included additional adjustments for hypertension (yes or no), anaemia (yes or no), arthritis (yes or no), cardiovascular disease (yes or no), stroke (yes or no), liver disease (yes or no), thyroid disease (yes or no), cancer (yes or no) and medication use (antihypertensive drugs, lipid-lowering drugs, aspirin, antidepressants, antitumour drugs).

Associations of the comprehensive healthy lifestyle score with FPG

Similar results were observed for the association between the comprehensive healthy lifestyle score and FPG levels (figure 3; online supplemental table A5). In all models, with scores of 6 or more, the association reached statistical significance. In the fully adjusted model (Model 3), FPG was reduced by 2.31 mg/dL (95% CI: −3.85 to −0.77) in participants with a score of 6 and by 2.74 mg/dL (95% CI: −4.47 to −1.01) for those with 7–8 factors, compared with the reference group with 0–2 factors. The trend test (P for trend <0.001) confirmed a significant linear decline in FPG with increasing lifestyle scores, with each additional point associated with a 0.52 mg/dL decrease in FPG (95% CI: −0.80 to −0.24).

Figure 3. Association between fasting plasma glucose (FPG) levels and the comprehensive healthy lifestyle score. NA, not available; β, regression coefficient. ^aModel 1, adjusted for age (years), gender (male and female), race (Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, other race including multiracial), education (less than high school, high school graduate, some college and college graduate or more), income ($0–$19 999, $20 000–$44 999, $45 000–$99 999, $45 000–$99 999). College more income ($0–$19 999, $20 000–$44 999, $45 000–$99 999 and ≥$100 000). ^bModel 2 included further adjustments for high-density lipoprotein (HDL), total cholesterol, triglycerides, total cholesterol (TC)/HDL-C ratio, estimated glomerular filtration rate (EGFR), serum creatinine (SCr), urinary creatinine ratio (UCr) and blood urea nitrogen (BUN). ^cModel 3 included additional adjustments for hypertension (yes or no), anaemia (yes or no), arthritis (yes or no), cardiovascular disease (yes or no), stroke (yes or no), liver disease (yes or no), thyroid disease (yes or no), cancer (yes or no) and medication use (antihypertensive drugs, lipid-lowering drugs, aspirin, antidepressants, antitumour drugs).

Subgroup analyses

Subgroup analyses by family history of diabetes, sex, age, BMI, income and education (figure 4) revealed significant interaction effects between the comprehensive lifestyle score and pre-diabetes across all stratified groups (p<0.05). Individuals without a family history of diabetes demonstrated a stronger association, with OR of 0.86 (95% CI: 0.78 to 0.94) compared with 0.90 (95% CI, 0.84 to 0.97) for those with a family history of diabetes. Sex stratification revealed a significant association among females (p<0.001) but not among males (p=0.087). Age stratification showed no significant association for participants aged ≥80 years (p=0.48), but significant associations were observed in younger age groups (p<0.05), particularly in the 40–59 age group (0.85; 95% CI, 0.77 to 0.94). Regarding BMI, no significant association was found in the normal weight group (p=0.991). However, significant associations were observed in the overweight and obese groups, with ORs of 0.85 (95% CI, 0.78 to 0.95) and 0.88 (95% CI, 0.79 to 0.99), respectively. Associations were stronger among individuals with a high school education or higher, particularly college graduates or more (OR: 0.79, 95% CI: 0.72 to 0.87), and those earning $45 000–$99 999 per year. After Bonferroni adjustment, all interaction P-values remained statistically significant, and the observed associations were predominantly evident among younger age groups.

Figure 4. Subgroup analyses of the association between the pre-diabetes and comprehensive healthy lifestyle score. BMI, Body Mass Index (calculated as weight in kilograms divided by height in metres squared); *represents a p value of less than 0.05; **represents a p value of less than 0.01;***represents a p value of less than 0.001. ^aOR (95% CIs) were calculated in weighted logistic regression model after adjusting for age (years), gender (male and female), race (Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, other race including multiracial), education (less than high school, high school graduate, some college and college graduate or more), income ($0–$19 999, $20 000–$44 999, $45 000–$99 999, $45 000–$99 999). College more income ($0–$19 999, $20 000–$44 999, $45 000–$99 999 and ≥$100 000), high-density lipoprotein (HDL), total cholesterol, triglycerides, total cholesterol (TC)/HDL-C ratio, estimated glomerular filtration rate (EGFR), serum creatinine (SCr), urinary creatinine ratio (UCr), blood urea nitrogen (BUN), hypertension (yes or no), anaemia (yes or no), arthritis (yes or no), cardiovascular disease (yes or no), stroke (yes or no), liver disease (yes or no), thyroid disease (yes or no), cancer (yes or no) and medication use (antihypertensive drugs, lipid-lowering drugs, aspirin, antidepressants, antitumour drugs). All p values for interaction were significant after Bonferroni correction except BMI group.

Sensitivity analyses

In sensitivity analyses, the associations between the comprehensive healthy lifestyle score and pre-diabetes, HbA1c and FPG remained materially unchanged after excluding data from 2005 to 2006 (online supplemental table A6), adjusting for LDL-C (online supplemental table A7), conducting analyses without weighting methods (online supplemental table A8), using chained equations for multiple imputation of missing covariate data (online supplemental table A9) and replacing weekday sleep duration with weekend sleep duration (online supplemental table A10).

Discussion

In this large nationally representative study, we observed that adherence to a healthier multidimensional lifestyle was significantly associated with a lower risk of pre-diabetes and improved glycaemic control. These associations were consistently observed across multiple statistical models, supporting the robustness and reliability of the findings. Compared with individuals who reported 0–2 healthy lifestyle factors, those with 7–8 factors had a 53% lower odds of pre-diabetes. Similar trends were observed for two key glycaemic biomarkers. Each one-point increase in the lifestyle score was associated with a 12% reduction in pre-diabetes risk, as well as reductions in HbA1c (β=–0.02) and FPG (β=–0.53 mg/dL), suggesting a clear dose–response relationship. Stratified analyses further revealed population-specific differences in these associations. These findings support the implementation of tailored, lifestyle-based prevention strategies for high-risk populations.

Our findings align with prior research that individual lifestyle factors, such as diet, physical activity and non-smoking, are associated with reduced diabetes risk and improved metabolic health.^{23 24} Additionally, psychosocial factors like marital status and depressive symptoms have been connected to glycaemic control. Individuals who were married or living with a partner demonstrated a 0.21% (95% CI, –0.31% to –0.10%) lower HbA1c level.²¹ Ulaş discovered that individuals with pre-diabetes had a higher mean score of depressive symptoms compared with those with normoglycaemia.²⁵ There was a positive correlation with abnormal fasting blood glucose levels. Sleep deprivation and unhealthy waist circumference have also been associated with metabolic disorders and poor glycaemic control.^{26 27} These studies primarily focused on individual factors, often overlooking the potential synergistic effects of multiple lifestyle behaviours and the additional influence of psychosocial health. However, this study uniquely contributes to the literature by evaluating the combined effects of multiple lifestyle factors on pre-diabetes and glycaemic outcomes. Our findings highlight that cumulative healthy lifestyle changes offer more robust protection against pre-diabetes than individual factors alone, suggesting an additive and possibly synergistic benefit.

The observed associations may be explained by several underlying mechanisms. Healthy dietary habits, regular physical activity and sufficient sleep are known to improve insulin sensitivity, reduce inflammation and maintain energy balance—key factors in glycaemic control.⁷ Additionally, psychosocial factors such as stable marital relationships and lower levels of depressive symptoms may facilitate stress reduction, emotional regulation and engagement in healthier behaviours, thereby reinforcing metabolic resilience. These interconnected effects suggest that lifestyle components may act synergistically rather than independently. Taken together, our findings support the potential utility of a holistic, multidimensional approach to lifestyle modification for preventing pre-diabetes and improving early metabolic outcomes.

Subgroup analyses revealed stronger associations between lifestyle factors and pre-diabetes risk among females, individuals without a family history of diabetes and those aged 40–59. These findings imply that certain demographic groups may benefit more substantially from lifestyle interventions. For example, women may experience greater metabolic benefits from lifestyle changes due to hormonal, metabolic or behavioural differences.²⁸ Moreover, individuals without a hereditary predisposition to diabetes appear more responsive to lifestyle modifications, as their risk factors are likely more modifiable. Additionally, the age-specific effects indicate that early lifestyle interventions in middle-aged individuals may prevent diabetes progression later in life. The benefits of healthy lifestyles for overweight and obese individuals were also validated, which was consistent with the recognised relationship between obesity and an increased risk of diabetes^{29 30。}Similar to previous research,³¹ socioeconomic and education disparities were associated with higher pre-diabetes risk, likely mediated through poorer lifestyle behaviours and limited access to preventive health resources. These findings underscore the importance of tailoring lifestyle-based prevention strategies to specific population subgroups based on biological, behavioural and structural characteristics.

These findings have substantial implications for both clinical practice and public health strategies. Incorporating a comprehensive set of healthy lifestyle factors, including psychosocial aspects, into preventive health strategies could enhance the effectiveness of diabetes prevention programmes. Public health campaigns and clinical guidelines could emphasise the importance of adopting multiple, co-occurring lifestyle modifications rather than single-behaviour changes, particularly for high-risk groups identified in this study. Furthermore, the inclusion of psychosocial factors highlights the value of a holistic, person-centred approach that addresses both physical and mental health in the management of metabolic risk. Designing tailored interventions that reflect the multidimensional nature of lifestyle and the diverse needs of specific populations may contribute to more equitable and effective chronic disease prevention.

Strengths and limitations

This study has several strengths. First, we used nationally representative NHANES data and performed multiple sensitivity analyses and imputations to ensure robustness and generalisability. Second, we constructed a comprehensive healthy lifestyle score both conventional and psychosocial factors, enabling assessment of cumulative and potentially synergistic effects. Third, we evaluated both pre-diabetes status and continuous glycaemic biomarkers (HbA1c and FPG), allowing for a more nuanced understanding of lifestyle effects across the glycaemic spectrum. Lastly, stratified analyses further identified population-specific associations, informing more tailored prevention strategies.

Our study has limitations. First, the cross-sectional design could not confirm causality and temporal relationships. Second, relying on self-reported lifestyle data may add bias and a one-time assessment may not fully capture dynamic changes in behaviour. Third, analyses were restricted to participants with complete data, which may lead to selection bias; although multiple imputation was conducted, the exclusion of incomplete cases may have affected the representativeness of the sample. Fourth, given the nature of post hoc subgroup analyses, where sample sizes for each subgroup were not calculated before data collection, the number of participants and events may have been insufficient, and the results should be interpreted with caution. Finally, we allocated equivalent scores to all healthy lifestyles, which may not accurately reflect the relative impact of different factors on pre-diabetes risk. Future analyses could consider assigning different weights to each lifestyle factor to more accurately reflect their relative contributions.

Conclusions

In conclusion, adherence to a comprehensive healthy lifestyle—encompassing both conventional and psychosocial factors—is associated with a significantly lower risk of pre-diabetes and better glycaemic control. These findings underscore the importance of addressing both physical and mental health in lifestyle-based public health strategies aimed at preventing early metabolic deterioration.

Data availability statement

Data are available upon reasonable request.