Abstract
Objective
This study investigates the association between dietary live microbial intake and mortality among obese adults.
Methods
From the 1999–2018 cycles of the U.S. National Health and Nutrition Examination Survey (NHANES), we included 22,746 adults with obesity and categorized them into three groups according to consumption of foods with medium or high microbial content (MedHi): G1 (no intake), G2 (>0 and below the median), and G3 (at or above the median). Associations with all-cause, CVD, and cancer mortality were examined using Kaplan–Meier curves and fully adjusted Cox proportional hazards models. Additionally, dose-response and sensitivity analyses were performed to further explore these relationships.
Results
Over a median follow-up period of 9.67 years, a total of 3,792 deaths occurred, comprising 952 CVD-related and 866 cancer-related deaths. Compared to G1, participants in G3 had a 16% lower risk of all-cause mortality (hazard ratio [HR] = 0.84, 95% confidence interval [CI]: 0.77-0.92) and an 18% lower risk of CVD mortality (HR = 0.82, 95% CI: 0.69-0.98). However, no significant association was observed between dietary live microbial intake and cancer mortality. Dose–response modelling revealed an inverse approximately linear association between dietary live microbial intake and both all-cause and CVD mortality.
Conclusions
These findings suggest that higher dietary exposure to live microbes is associated with lower risks of all-cause and CVD mortality in individuals with obesity, highlighting a potential target for obesity-oriented health policies and preventive strategies.
Introduction
As chronic disease burden continues to climb, researchers and clinicians are increasingly focusing on the contribution of dietary patterns to both prevention and treatment.1,2 Among these dietary components, live microorganisms have emerged as a focus of interest due to their potential modulatory effects on host health. Research demonstrates that dietary sources of live microbes, including probiotic strains and microorganisms inherent to fermented products, may influence host health through effects on gut microbial composition, immune regulation, and metabolic homeostasis. 3 The outlined mechanisms point to a potential role for dietary live microorganisms in both preventing and managing diverse chronic conditions. Dietary live microorganisms may influence host health partly through interactions with the gut microbiota, modulating microbial diversity and metabolic activity and thereby affecting the production of bioactive metabolites such as short-chain fatty acids (SCFAs), including acetate, propionate, and butyrate.4–6 These metabolites play important roles in maintaining intestinal barrier integrity, regulating immune responses, and improving metabolic homeostasis, which may contribute to the prevention of chronic diseases.
Obesity, a prominent threat to global health, not only increases the risk of numerous chronic diseases but also substantially elevates all-cause mortality rates.7,8 Recent investigations have shown diet to be a critical driver in developing obesity and the spectrum of disorders linked to it.9,10 Given the close relationship between dietary patterns, metabolic health, and gut microbial ecology, increasing attention has been directed toward the potential role of dietary live microorganisms in influencing long-term health outcomes. Several recent studies using National Health and Nutrition Examination Survey (NHANES) data have explored the association between dietary live microbial intake and mortality outcomes. For example, analyses of NHANES data in U.S. adults have reported that higher intake of foods containing live microbes is associated with lower risks of all-cause and cardiovascular mortality.11,12 Similarly, research focusing on older populations reported inverse associations between estimated live microbe intake and mortality risk. 13 These findings provide preliminary epidemiological evidence supporting potential health benefits of dietary live microorganisms.
However, while the role of live-microbe diets in conditions such as cardiovascular disease (CVD) and metabolic syndrome has been explored, evidence specifically examining the association between dietary live microbial intake and mortality among individuals with obesity remains limited, despite the high cardiometabolic risk in this population. Therefore, the present study aimed to investigate the association between dietary live microbial intake and both all-cause and cause-specific mortality in obese adults using data from the U.S. NHANES from 1999 to 2018.
Methods
Study population
Data for the present analysis were obtained from NHANES, conducted between 1999 and 2018. Administered by the National Center for Health Statistics (NCHS), NHANES uses a stratified, multistage probability design to generate nationally representative estimates of health and nutritional status among non-institutionalized U.S. residents. The study included adults aged ≥20 years and excluded non-obese participants, those lost to follow-up, individuals with missing dietary live microbial intake data, participants with incomplete covariate data, as well as pregnant participants. The final sample comprised 22,746 obese participants (Figure 1). The NHANES study protocol was approved by the National Center for Health Statistics Research Ethics Review Board, and written informed consent was obtained from all participants at enrollment. The study was conducted in accordance with the principles of the Declaration of Helsinki (1975, revised in 2024).
14
The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
15
Study flow chart.
Determination of dietary live microbial intake
Dietary exposure to live microorganisms was estimated based on the food classification system proposed by Marco et al. for defining dietary intake of live microbes. 16 Dietary intake data were obtained from the 24-hour dietary recall interviews in NHANES, and all reported foods were identified using the USDA Food and Nutrient Database for Dietary Studies (FNDDS) food codes. NHANES food codes were then matched to the live-microbe classification framework developed by Marco et al. Each food item was assigned to one of three categories according to its estimated live microbial content: low (<104 CFU/g), medium (104–107 CFU/g), or high (>107 CFU/g). 17 This classification was derived from expert evaluation of food processing conditions, fermentation status, and available literature on microbial survival in foods. Foods subjected to routine heat processing, including pasteurized milk, cooked meat and poultry, cooked seafood, and sauces, were categorized as low, as thermal processing markedly reduces viable microbial counts. Fresh fruits and vegetables were generally categorized as medium, reflecting moderate levels of naturally occurring microorganisms, whereas fermented dairy products were commonly classified as high due to their substantial content of live microbes. Following previous NHANES-based studies, foods categorized as medium or high microbial content (MedHi) were considered sources of dietary live microbes. Individual intake of live-microbe–containing foods was estimated by summing the consumption of MedHi foods reported in the dietary recall. Participants were subsequently classified into three groups according to their MedHi food intake: G1 (no consumption of MedHi foods), G2 (MedHi intake >0 but below the population median), and G3 (MedHi intake ≥ the population median). 18 This approach has been widely applied in NHANES-based studies to estimate population-level dietary exposure to live microbes.
Determination of mortality
Participant mortality was tracked via the NHANES Linked Mortality dataset, with ascertainment available up to 31 December 2019. Causes of death were coded according to the 10th revision of the International Classification of Diseases (ICD-10). We analysed overall mortality as well as cause-specific deaths, focusing on cardiovascular disease (ICD-10 codes I00–I09, I11, I13, I20–I51) and malignant neoplasms (ICD-10 codes C00–C97). Survival time was calculated beginning from participants’ baseline assessment in NHANES.
Assessment of covariates
Age, gender, race, marital status, poverty income ratio (PIR), education level, Healthy Eating Index (HEI-2015), physical activity, smoking status, alcohol intake, and diagnoses of CVD, hypertension, hyperlipidemia, diabetes, and cancer were included as covariates to control for potential confounding. Definitions are detailed as follows: age entered all statistical models as a continuous covariate; gender was recorded as male or female; race was divided into two groups: non-Hispanic White and “other,” the latter comprising non-Hispanic Black, Mexican American, other Hispanic, and other races; marital status was dichotomized into married/living with a partner or unmarried/other (including widowed, divorced, or separated); PIR was grouped into low (1.00–1.30), middle (1.31–3.50), and high (>3.50) categories 19 ; education level was grouped as less than high school, high school or equivalent, and more than high school; smoking status was categorized as never smokers (fewer than 100 lifetime cigarettes), former smokers (≥100 lifetime cigarettes but not smoking at present), and current smokers (≥100 lifetime cigarettes and still smoking on some or all days) 20 ; alcohol intake was categorized as follows: never drinkers (fewer than 12 drinks over a lifetime), former drinkers (at least 12 drinks in any single year but none in the past 12 months), and current drinkers (at least 12 drinks in any year with alcohol use during the last year) 21 ; physical activity was categorized as inactive (0 MET-min/week), insufficiently active (1–599 MET-min/week), and sufficiently active (≥600 MET-min/week). 22 Hypertension was classified by any of the following: systolic pressure averaging at least 140 mm Hg, diastolic pressure averaging at least 90 mm Hg, self-reported hypertension, or current use of blood-pressure-lowering drugs. 20 Hyperlipidemia was defined as meeting at least one of the following conditions, including treatment with lipid-modifying medication, triglyceride levels ≥150 mg/dL, or dyslipidemia indicated by total cholesterol ≥200 mg/dL, high-density lipoprotein cholesterol (HDL-C) <40 mg/dL, or low-density lipoprotein cholesterol (LDL-C) ≥130 mg/dL. 23 Diabetes was defined by any of the following criteria: a documented physician diagnosis, random plasma glucose or 2-hour glucose during an oral glucose tolerance test (OGTT) ≥11.1 mmol/L, glycated hemoglobin (HbA1c) ≥6.5%, fasting plasma glucose ≥7.0 mmol/L, or current use of insulin or other glucose-lowering medications. 24 Cancer was defined by a “yes” response to the questionnaire item: “Have you ever been told by a doctor or other health professional that you have cancer or a malignancy of any kind?”. 22
Statistical analysis
Since the sample size was entirely based on available data, no a priori power calculation was conducted. Statistical analyses were conducted with R (The R Foundation for Statistical Computing; https://www.r-project.org/) and Free Statistics software version 2.1. Statistical significance was set at a two-tailed
Kaplan–Meier curves and multivariable Cox proportional hazards models were used to examine differences in all-cause and cause-specific mortality across the three MedHi food intake groups. Multivariable Cox proportional hazards regression models were employed to assess the associations between the three MedHi food intake groups and all-cause mortality, CVD mortality, and cancer-specific mortality. Variables entered the multivariable model if they were clinically relevant, showed
Subgroup analyses were performed by age (<65 years, ≥65 years), gender (male, female), race (White, other), marital status (married, unmarried), CVD, hypertension, hyperlipidemia, and diabetes. Subgroup analyses were conducted using Cox proportional hazards regression models, and potential interactions were assessed by including multiplicative interaction terms, with significance evaluated using likelihood-ratio tests. Robustness was assessed through a series of sequential sensitivity analyses. First, deaths occurring within the first two years of follow-up were excluded to minimize potential reverse causation. Second, individuals with pre-existing CVD were excluded. Third, participants with a history of cancer at baseline were removed. Fourth, survey-weighted analyses were performed to account for the complex multistage sampling design of NHANES. Because ten survey cycles (1999–2000 through 2017–2018) were combined in this study, sampling weights were recalculated according to NHANES analytic guidelines. Specifically, for participants from the 1999–2002 cycles, the 4-year MEC examination weights (WTMEC4YR) were multiplied by 2/10, whereas for participants from subsequent cycles (2003–2004 through 2017–2018), the 2-year MEC examination weights (WTMEC2YR) were divided by 10. Fifth, additional models were fitted with further adjustment for total energy intake to evaluate potential confounding by overall dietary intake. For each sensitivity analysis, the models examining the association between MedHi food intake group classification and mortality risk were re-fitted. These procedures were undertaken to evaluate the stability and reliability of the primary findings.
Results
Participant characteristics
Characteristics of US obese adults by three levels of MedHi food intake.
Abbreviations: HEI-2015, Healthy Eating Index-2015; PIR, poverty income ratio; CVD, cardiovascular disease.
Association between MedHi food intake and mortality
During a median follow-up period of 9.67 years (interquartile range: 5.33–14.50 years), there were 3,792 deaths, including 952 CVD deaths and 866 cancer deaths. In Figure 2, Kaplan–Meier analysis showed markedly reduced all-cause, CVD, and cancer mortality rates in Groups 2 and 3 relative to Group 1 (all Kaplan–Meier survival curves for all-cause mortality (a), CVD mortality; (b), and cancer mortality; (c) according to three levels of MedHi food intake among obese adults.
Adjusted hazard ratios of three levels of MedHi food intake with risk of all-cause mortality and cause-specific mortality.
Model 1 was adjusted for age and gender. Model 2 was additionally adjusted for race, marital status, PIR group, educational level, HEI-2015, physical activity, smoking status, and alcohol intake. Model 3 was additionally adjusted for CVD, hypertension, hyperlipidemia, diabetes, and cancer.
Abbreviations: HEI-2015, Healthy Eating Index-2015; PIR, poverty income ratio; CVD, cardiovascular disease.
Dose–response associations between MedHi food intake and all-cause mortality (a) and CVD mortality (b) among obese adults.
Subgroup analysis
As shown in Figure 4, subgroup analyses indicated that MedHi food intake exerted a similar protective effect on all-cause mortality across every subgroup examined, including age, gender, race, marital status, CVD, hypertension, hyperlipidemia, and diabetes. Specifically, in all subgroups, the group with high live microbial intake (G3) exhibited a significantly lower risk of all-cause mortality. Age significantly modified the observed association. Subgroup analyses of the associations between three levels of MedHi food intake and all-cause mortality among obese adults.
Sensitivity analysis
Sensitivity analyses of the association between MedHi food intake and all-cause mortality among obese adults.
Model 1 was adjusted for age and gender. Model 2 was additionally adjusted for race, marital status, PIR group, educational level, HEI-2015, physical activity, smoking status, and alcohol intake. Model 3 was additionally adjusted for CVD, hypertension, hyperlipidemia, diabetes, and cancer.
Abbreviations: HEI-2015, Healthy Eating Index-2015; PIR, poverty income ratio; CVD, cardiovascular disease.
Discussion
This study systematically investigated the association between dietary live microbial intake and both all-cause and cause-specific mortality in an obese population. The results indicate that higher intake of foods with medium or high microbial content was significantly associated with reduced all-cause mortality and CVD mortality among obese individuals. However, no significant association was found between dietary live microbial intake and cancer mortality. Notably, the dose-response analysis revealed a linear negative correlation between MedHi food intake and both all-cause and CVD mortality. These findings provide new evidence for the potential role of dietary live microorganisms as part of prevention and management strategies in obese populations, suggesting that they may play a critical role in improving obesity-related health outcomes.
The health benefits of dietary live microorganisms have gained widespread attention in recent years. Previous studies have demonstrated that MedHi food intake is significantly associated with improvements in cardiovascular health, metabolic syndrome, and immune function. 25 Our study extends these findings by showing that high MedHi food intake is associated with lower risks of all-cause mortality and CVD mortality in an obese population. This finding is consistent with previous studies reporting that higher intake of foods containing live microorganisms is associated with improved cardiovascular health in the general population. 26 However, similar to findings from prior research, our study did not observe a significant association between MedHi food intake and cancer mortality, suggesting that cancer mortality may be influenced by a variety of complex factors, including cancer type, treatment regimens, and other unmeasured confounders.
Previous studies have also supported the beneficial effects of dietary live microorganisms on metabolic health. For example, higher MedHi food intake has been reported to be inversely associated with the prevalence of non-alcoholic fatty liver disease (NAFLD), further supporting the potential role of dietary live microorganisms in managing obesity-related metabolic disorders. 27 In addition, higher consumption of MedHi foods has been associated with a lower risk of abdominal aortic calcification, suggesting broader protective effects on cardiovascular health. 28 Moreover, evidence from NHANES-based analyses indicates that consumption of live microbe–rich foods, particularly fermented foods, may improve cardiometabolic health indicators, including reductions in blood pressure, blood glucose, lipid levels, and waist circumference. 29 Although causality cannot be established, these findings provide additional evidence supporting the link between dietary live microbial intake and reduced mortality.
Our subgroup analysis revealed a consistent association between MedHi food intake and mortality across various subgroups defined by age, gender, race, and other variables. Notably, a significant interaction was observed between age and high MedHi food intake in relation to all-cause and CVD mortality. As individuals age, dietary habits tend to change due to declining dental and salivary function, lifestyle changes, and reduced digestive and absorptive capacities. For middle-aged and elderly individuals, especially those with reduced dietary diversity and elevated levels of inflammatory cytokines, consuming foods with moderate to high live microbial content may offer significant health benefits. 30 These findings provide a scientific basis for the development of personalized dietary recommendations and suggest that future research should explore the mechanisms underlying these differences to optimize dietary interventions for different populations.
Dietary live microorganisms may influence survival in obese populations through various mechanisms. First, dietary live microorganisms can modulate gut microbiota, reduce systemic inflammation, and improve metabolic function, which is particularly important for obese individuals. Research has shown that the regulation of gut microbiota can enhance host immune function and reduce the risk of CVD and metabolic syndrome. 31 Additionally, metabolites produced by dietary live microorganisms, such as short-chain fatty acids, may improve gut barrier function, reduce the absorption of harmful metabolites, and subsequently lower the risk of CVD. 32 The regulation of the gut-brain axis may also indirectly reduce obesity-related mortality by improving mental health.33,34
This study utilized data from the 1999-2018 NHANES, a nationally representative sample, which enhances the external validity of the findings. Furthermore, the study adjusted for a wide range of potential confounders, including age, gender, race, education level, and health behaviors, which enhanced the robustness of the findings. The stability of the results was further supported by multiple sensitivity analyses, including survey-weighted models and additional adjustment for total energy intake. By employing Kaplan-Meier survival analysis and multivariable Cox proportional hazards regression models, we were able to comprehensively assess the association between dietary live microbial intake and mortality, providing new evidence for dietary management strategies in obese populations. Despite its strengths, this study also has several limitations. First, although prospective cohort data were used, baseline dietary intake data were collected through cross-sectional surveys, limiting the ability to infer causality. Future studies should consider longitudinal designs to more accurately assess the long-term health impacts of dietary live microbial intake. Second, dietary intake data were self-reported based on 24-hour dietary recall interviews in NHANES, which reflect short-term dietary consumption and may not fully represent participants’ habitual long-term dietary patterns. This approach may introduce recall bias or measurement error, potentially affecting the precision of the results. Third, survival analyses were conducted using Cox proportional hazards models, and competing risks were not explicitly accounted for. In the presence of competing events, such as deaths from other causes, the estimated associations may be subject to bias. Future studies using competing risk models, such as the Fine–Gray subdistribution hazards model, may provide more accurate estimates. Finally, the estimation of dietary live microbial intake was based on food coding and did not capture specific microbial species or strains present in foods. As a result, the differential effects of individual microorganisms on health outcomes could not be evaluated. In addition, the NHANES dataset used in this study does not include gut microbiome sequencing data, preventing direct assessment of the relationship between dietary live microbial intake and gut microbiota composition or function. Moreover, cardiometabolic and inflammatory biomarkers, such as C-reactive protein (CRP), glycated hemoglobin (HbA1c), and blood lipids, may represent potential biological pathways linking dietary live microbial intake to mortality outcomes, but formal mediation analyses were beyond the scope of the present study. Future research integrating dietary data, microbiome profiling, and biomarker-based mediation analyses may help clarify the underlying biological mechanisms.
Conclusion
This study suggests that higher MedHi food intake is associated with reduced all-cause and CVD mortality in obese individuals, although no significant association was observed for cancer mortality. These findings provide additional epidemiological evidence regarding the potential role of dietary live microorganisms in obesity-related health outcomes and may inform future research and public health strategies.
Supplemental material
Supplemental material - Association between dietary live microbial intake and all-cause and cause-specific mortality in obese adults: A prospective cohort study
Supplemental material for Association between dietary live microbial intake and all-cause and cause-specific mortality in obese adults: A prospective cohort study by Yaqiong Zhuang, Xiaobo Liu, Yingxuan Huang, Chanchan Lin, Yingyi Li, Yisen Huang, Yubin Wang, Xiaoqiang Liu in Science Progress
Footnotes
Acknowledgements
We are grateful to Huanxian Liu (Department of Neurology, Chinese PLA General Hospital, Beijing) for his insightful feedback on both the study design and the manuscript.
Ethical considerations
The NHANES study protocol was approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided written informed consent.
Author contributions
Y.Z. and X.L. devised the study design, carried out data extraction and analysis, and drafted the manuscript. Y.H. and C.L. collected the data. Y.L. and Y.H. participated in the design of research schemes. Y.W. and X.L. reviewed the manuscript. All authors contributed to the article and approved the submitted version. All authors read and approved the final manuscript.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by China Scholarship Council (Grant No. 201906070289), Startup Fund for Scientific Research, Fujian Medical University (Grant No. 2022QH1268). The funding bodies were not involved in designing the study, analysing the data, deciding to publish, or drafting the manuscript.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The datasets analyzed in this study are publicly available from the National Health and Nutrition Examination Survey (NHANES) website.
Supplemental material
Supplemental material for this article is available online.