Metabolic Multimorbidity and Risk of All-Cause and Cause-Specific Mortality: A Retrospective Cohort Study of 123,791 Chinese Adults

Introduction

The burden of metabolic diseases has continued to increase worldwide. Over the past two decades, disability-adjusted life years (DALYs) attributable to metabolic diseases have increased by nearly 50%, reaching more than 470 million in 2021.¹ Metabolic diseases such as hypertension and diabetes now rank among the top five causes of DALYs worldwide.¹ Notably, many of these diseases share common risk factors and underlying mechanisms, including chronic inflammation, metabolic inflexibility, and mitochondrial dysfunction.² They often coexist and interact with one another, synergistically contributing to cardiovascular disease (CVD) and premature mortality.^3–5 Emerging evidence from metabolomics has demonstrated shared metabolic signatures across cardiometabolic diseases and systemic inflammatory states.⁶ The concept of a “Global Metabolic Syndemic” emphasizes the need to address the collective metabolic burden rather than considering each condition in isolation.^7–9 Characterizing patterns of metabolic multimorbidity and their associations with cause-specific mortality is therefore critical for identifying high-risk populations and informing integrated prevention and management strategies.

Most previous studies have examined metabolic conditions individually or focused primarily on metabolic syndrome (MetS),^2,3 overlooking the continuous risk gradient associated with the increasing number and types of metabolic diseases. For example, recent Global Burden of Disease (GBD) studies reported the prevalence of five major metabolic diseases but did not characterize their coexistence patterns due to the lack of individual-level data.^8,9 A few studies have reported the prevalence of metabolic multimorbidity,^10–12 yet these analyses did not include non-alcoholic fatty liver disease (NAFLD), despite its rising global burden and well-documented links with both MetS and CVD.^4,13 Furthermore, few studies have examined the associations between metabolic multimorbidity and cause-specific mortality.

Therefore, using data from a large retrospective cohort in northern China, we aimed to characterize the prevalence of metabolic multimorbidity across sex and age groups and quantify its associations with all-cause and cause-specific mortality. We hypothesized that metabolic multimorbidity would vary across population subgroups and that increasing numbers of coexisting metabolic diseases would be associated with a graded increase in mortality risk.

Methods

Study Design and Participants

Routine general health examinations are widely implemented in China and are part of the government’s policy on health literacy.¹⁴ In 2021, approximately 549 million health examinations were performed nationwide,¹⁵ with most participants being employed adults, as many employers provide health checks as employee benefits.¹⁴ In this retrospective cohort study, we identified all adults aged 25–74 years who underwent routine health examinations at the China-Japan Friendship Hospital, a tertiary care medical center in Beijing, China, between January 1, 2015, and December 31, 2022. More than 90% of participants were current or retired employees from enterprises, government agencies, or public institutions in Beijing. Participants were followed from baseline until death or censored on December 31, 2024, whichever occurred first.

Among the 136,157 adults identified, 11,039 (8.1%) were excluded due to missing data on blood glucose (n = 2694), blood lipids (n = 2385), blood pressure (BP; n = 1944), body mass index (BMI; n = 2912), or abdominal B-type ultrasonography (n = 1104). In addition, participants who were lost to follow-up (n = 1327) were excluded. After exclusions, the final analytic cohort included 123,791 participants. The study flow diagram is shown in Supplementary Figure S1. The Human Research Ethics Committee of the China-Japan Friendship Hospital approved this study and waived the requirement for informed consent, given its observational design and use of fully de-identified data (Approval No. 2025-KY-109-1). The study was conducted in accordance with the Declaration of Helsinki.

Procedures and Measures

Routine general health check-ups are offered as comprehensive health packages that include routine screenings for multiple diseases during a single visit.¹⁴ Demographic characteristics, medical history, and medication use were recorded during the health check-up. Systolic and diastolic BP, height, weight, and abdominal B-mode ultrasonography were measured following standardized protocols. BP was measured twice in the right upper arm after 5 minutes of rest in a seated position using a standardized electronic monitor (Omron HEM-7430, Omron Corporation, Kyoto, Japan). The average of the two readings was used for analysis. Participants were instructed to wear light clothing and refrain from wearing shoes or caps when their height and weight were measured. BMI was calculated as weight in kilograms divided by height in meters squared. Fasting blood glucose (FBG), hemoglobin A1c (HbA1c), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) were measured using standardized laboratory methods and devices. More detailed materials and methods of data collection are described in Supplementary Text S1.

We focused on five common metabolic diseases: obesity, hypertension, dyslipidemia, NAFLD, and diabetes mellitus. These diseases were selected because of their high prevalence in China, established associations with cardiometabolic outcomes, and consistent inclusion in prior GBD analyses.^8,9 Hypertension was defined as systolic blood pressure (SBP) ≥140 mm Hg, diastolic blood pressure (DBP) ≥90 mm Hg, self-reported diagnosis of hypertension, or use of antihypertensive medication. Dyslipidemia was defined by TC ≥6.22 mmol/L, TG ≥2.26 mmol/L, HDL-C <1.04 mmol/L, LDL-C ≥4.14 mmol/L, self-reported diagnosis of dyslipidaemia, or use of lipid-lowering medication, in accordance with the 2023 Chinese Guidelines for Lipid Management.¹⁶ Diabetes was defined as FBG ≥7.0 mmol/L, HbA1c ≥6.5%, self-reported diagnosis of diabetes, or use of antidiabetic medication.¹⁷ According to the Chinese BMI Classification, obesity was defined by BMI ≥28 kg/m². NAFLD was defined as the presence of hepatic steatosis on ultrasonography without viral hepatitis or excessive alcohol consumption (≥30 g/day).¹⁸ Metabolic disease status was assessed at baseline only. Metabolic multimorbidity was defined as the coexistence of two or more of the five metabolic diseases.

Follow-Up and Outcome Assessment

Participants’ vital status and cause of death were collected through passive follow-up by linkage between the cohort database and the national mortality registration system, which covers all 31 mainland province in China.¹⁹ The death records in this system are reported by health care institutions almost in near real time, and subsequently verified against local residential records and health insurance records on an annual basis. CVD death (codes I00–I99) and cancer death (codes C00–C97) were defined according to International Classification of Diseases, 10th Revision (ICD-10).²⁰

Statistical Analysis

We evaluated the prevalence of metabolic multimorbidity overall and by sex and age group. Prevalence was reported according to the number of coexisting metabolic diseases. To assess whether the association between age and metabolic multimorbidity differed by sex, an interaction term between age and sex (age × sex) was included in multivariable logistic regression models. Additionally, we assessed the distribution of metabolic multimorbidity by disease type. Chi-square (χ²) tests were used to compare distributions across groups.

We calculated the cumulative incidence of death from all causes and specific causes for individuals with varying numbers of coexisting metabolic diseases, using time from baseline to death as the time-to-event variable. The cumulative incidence of death from each specific cause was estimated using the cumulative incidence function (CIF) in a competing-risk framework, treating other causes of death as competing risks.²¹ Gray’s test was used to compare cause-specific cumulative incidence curves across groups. All-cause mortality was compared across groups using the Log rank test. The number of deaths and mortality rates (per 1000 person-years) were calculated according to the number of coexisting metabolic diseases. Person-years were calculated from baseline to the date of death or censoring at the end of follow-up, whichever occurred first.

Cox proportional hazards regression models were used to examine the associations between metabolic multimorbidity and all-cause mortality. For cause-specific mortality, cause-specific Cox proportional hazards models were applied, treating deaths from other causes as censored at the time of occurrence.²¹ The number of coexisting metabolic diseases at baseline (0–5) was modelled as a categorical variable, with participants without any metabolic disease as the reference group. To assess dose–response relationships, the number of coexisting diseases was additionally modelled as a continuous variable to calculate P values for trend. Separate models were further fitted using predefined cutoffs (≥2 vs 0 and ≥3 vs 0 metabolic diseases). Given the limited availability of additional covariates, all models were adjusted only for age and sex. Adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were reported. We examined the proportional hazards assumption using the Schoenfeld residuals test and interaction terms between time and covariates. We conducted analyses using R software (version 4.2.3). A two-sided p-value of less than 0.05 was considered statistically significant.

Results

Characteristics of the Study Population

The demographic and clinical characteristics of the study population are shown in Table 1. Among the 123,791 participants, 62,852 (50.8%) were men, and the mean (SD) age was 41.3 (11.9) years. Among the five metabolic diseases, NAFLD was the most prevalent (37.0%), followed by dyslipidemia (33.8%), hypertension (17.2%), obesity (13.1%), and diabetes (5.3%). A total of 28,803 participants (23.3%) had one metabolic disease, 20,782 (16.8%) had two, 12,280 (9.9%) had three, 4975 (4.0%) had four, and 908 (0.7%) had all five metabolic diseases.

Table 1 Demographic and Clinical Characteristics of the Study Population

Prevalence of Metabolic Multimorbidity by Sex and Age

The prevalence of metabolic multimorbidity was significantly higher in men than in women (46.1% vs 16.4%; P <0.001). The age-related pattern differed significantly by sex (P for interaction <0.001) (Supplementary Table S1). Among men, the prevalence was already high in early adulthood and increased steadily with age (25–29 years: 27.9%; 70–74 years: 61.2%). In contrast, women had a low prevalence in early adulthood (25–29 years: 3.8%), followed by a marked increase after midlife, reaching levels comparable to men by ages 70–74 (58.0%) (Figure 1).

Figure 1 Sex-specific distribution of the number of metabolic diseases across age groups at baseline. The stacked area plots show the proportion of participants with 1, 2, 3, 4, or 5 metabolic diseases across age groups, separately for men (left panel) and women (right panel). Metabolic diseases include obesity, diabetes, hypertension, dyslipidemia, and non-alcoholic fatty liver disease (NAFLD). Each color represents the cumulative proportion of participants with a given number of metabolic diseases within each age group. Percentages are calculated within sex- and age-specific strata.

Prevalence of Metabolic Multimorbidity by Disease Type

By disease type, metabolic multimorbidity was most prevalent among participants with diabetes (96.5%) and obesity (93.8%). Over 70% of those with hypertension, dyslipidemia, or NAFLD also had metabolic multimorbidity. These patterns were consistent across sexes, with men showing a higher prevalence for each disease type (Figure 2).

Figure 2 Distribution of the number of coexisting metabolic diseases by index metabolic disease and sex. Horizontal stacked bar charts display the proportion of participants with 1, 2, 3, 4, or 5 metabolic diseases (including the index disease itself) among individuals with each index metabolic disease, presented overall and stratified by sex. Metabolic diseases include obesity, diabetes, hypertension, dyslipidemia, and non-alcoholic fatty liver disease (NAFLD). Participants with a value of 1 have the index disease only, whereas higher numbers indicate increasing metabolic multimorbidity. Percentages sum to 100% within each subgroup. Only categories accounting for ≥3% are labeled for clarity.

Metabolic Multimorbidity and Mortality

During a median follow-up of 6.1 years (IQR, 4.2–7.6), 724 deaths (0.6%) occurred, including 216 (0.2%) from CVD and 239 (0.2%) from cancer (Table 2). Cumulative all-cause mortality increased progressively with the number of coexisting metabolic diseases (log-rank P <0.001). Similar graded patterns were observed for CVD and cancer mortality (Gray’s test, P <0.001 for both) (Figure 3).

Table 2 All-Cause and Cause-Specific Mortality During Follow-Up Overall and by the Number of Coexisting Metabolic Diseases

Figure 3 Cumulative incidence of all-cause, cardiovascular disease (CVD), and cancer mortality by number of coexisting metabolic diseases. Cumulative incidence curves show the risk of death from all causes (top panel), cardiovascular disease (middle panel), and cancer (bottom panel) according to the total number of coexisting metabolic diseases present at baseline (0–5). Metabolic diseases include obesity, diabetes, hypertension, dyslipidemia, and non-alcoholic fatty liver disease (NAFLD). Cumulative incidence was estimated using competing-risk methods, with deaths from other causes treated as competing events for cause-specific mortality. Numbers indicate participants contributing to each curve at baseline: 0 diseases, 56,043; 1disease, 28,803; 2 diseases, 20,782; 3 diseases, 12,280; 4 diseases, 4975; 5 diseases, 908.

Crude mortality rates (per 1000 person-years) increased progressively with the number of coexisting metabolic diseases. After adjustment for age and sex, compared with participants without metabolic diseases, HRs for all-cause mortality increased with the number of coexisting diseases, ranging from 1.38 (95% CI, 1.09–1.76) for one disease to 2.92 (95% CI, 1.82–4.68) for five diseases. Similar graded associations were observed for CVD mortality (HRs, 1.78 [95% CI, 1.06–2.99] for one disease to 5.13 [95% CI, 2.25–11.7] for five diseases) and cancer mortality (HRs, 1.36 [95% CI, 0.91–2.03] for one disease to 3.84 [95% CI, 1.90–7.78] for five diseases) (Table 3).

Table 3 Associations of the Number of Coexisting Metabolic Diseases with All-Cause and Cause-Specific Mortality

To further explore the heterogeneity underlying this graded association, we examined prevalent combinations of metabolic diseases (prevalence ≥0.5%). The most common combination was dyslipidemia plus NAFLD, accounting for 8.8% of the study population. The combination of all five metabolic diseases had the highest all-cause mortality risk compared with those without any metabolic disease (HR 3.08, 95% CI 1.92–4.95). Elevated risks were also observed for combinations of diabetes, dyslipidemia, obesity, and NAFLD (HR 3.03, 95% CI 1.73–5.31) and hypertension, obesity, and NAFLD (HR 3.00, 95% CI 1.96–4.60) (Supplementary Table S2).

Discussion

In this large cohort of Chinese adults, more than 31% of participants had metabolic multimorbidity. The prevalence was substantially higher in men than in women and exhibited distinct age-related patterns: men exhibited a higher prevalence at younger ages, whereas women showed a lower prevalence in early adulthood followed by a marked increase after midlife. Metabolic multimorbidity was particularly common among individuals with diabetes or obesity, affecting more than 90% of these groups. Increasing numbers of coexisting metabolic diseases were associated with progressively greater risks of all-cause, CVD, and cancer mortality. These findings may reflect shared underlying pathophysiological mechanisms and support moving beyond disease-specific management toward integrated, sex-specific strategies to address overall metabolic risk.

Despite the well-documented interrelationships among metabolic diseases, most prior studies have examined individual conditions in isolation or focused only on the coexistence of two diseases.^22,23 Recent GBD analyses have examined five common metabolic diseases and introduced the concept of a “Global Metabolic Syndemic”, emphasizing the need to address their collective burden.^8,9 In the present study, we assessed the multimorbidity pattens of these five diseases and found that more than 31% of Chinese adults had at least two diseases, and 14% had three or more. These estimates align with earlier reports. A study examining four metabolic risk factors reported that 34% of adults in South and Southeast Asia had two or more, and 12% had three or more, of these conditions.¹⁰ Another analysis of five cardiovascular risk factors found that 31% of Chinese adults had two or more, and 15% had three or more.²⁴

We further observed a graded association between the number of coexisting metabolic diseases and risks of mortality, consistent with previous evidence showing that the coexistence of hypertension and diabetes confers greater mortality risk than either condition alone.²² Several mechanisms may explain this cumulative effect. First, metabolic diseases share common pathophysiological pathways, including insulin resistance, chronic low-grade inflammation, and mitochondrial dysfunction, which may act synergistically to impair cardiovascular and metabolic homeostasis.^2,25 Metabolomics studies in cardiometabolic disease, systemic inflammation, and critical illness have further identified perturbations in energy, amino acid, and lipid metabolism associated with adverse outcomes.^6,26 Second, recent evidence suggests that metabolic multimorbidity may represent an accelerated biological aging phenotype, contributing to excess mortality risk.^27–29 Together, these findings support a shift from disease-specific management toward integrated strategies targeting shared pathophysiological pathways. For example, lifestyle interventions address shared upstream mechanisms, and pharmacologic agents with pleiotropic metabolic effects, such as glucagon-like peptide-1 receptor agonists, may provide benefits across multiple metabolic conditions.³⁰ In addition, future research incorporating omics-informed risk stratification may help identify high-risk individuals with metabolic multimorbidity.⁶ Implementation could involve coordinated management across primary, endocrine, and cardiology services, risk-based screening programs, and multidisciplinary lifestyle intervention platforms, providing continuous, patient-centered care.

Consistent with previous reports,²⁴ our findings showed a significantly higher prevalence of all five metabolic diseases in men than in women. We further demonstrate that men also exhibited a substantially greater burden of metabolic multimorbidity. Notably, men bear a considerable burden of metabolic multimorbidity from early adulthood, markedly exceeding that of women in the same age group. These patterns were also reflected in data from Chinese children and adolescents, which show significantly higher rates of hypertension and obesity among boys than girls.^31,32 In contrast, women exhibited relatively low prevalence of metabolic multimorbidity in early adulthood, followed by a marked increase around the menopausal transition. Similar trends have been reported in cohort studies of Chinese and Dutch populations, which showed the most significant lipid changes occurring in the age group of 40–49 years.^33,34 Together, these findings suggest that sex differences in metabolic health may emerge early in life, and may help inform the timing and focus of preventive efforts, with earlier attention among men and heightened monitoring during midlife among women, particularly around the menopausal transition.

Furthermore, we observed that more than 90% of individuals with obesity had metabolic multimorbidity. A nationwide analysis of approximately 2.22 million obese adults in China reported that 82% had NAFLD and 42% had dyslipidemia.²³ Although observations that some individuals with obesity remain free of metabolic abnormalities led to the concept of metabolically healthy obesity (MHO),³⁵ our results suggest that this phenotype is uncommon. In line with this, emerging evidence suggests that MHO represents a transient phenotype, reflecting its underlying molecular and metabolic instability.^35–37 Over 30% of Chinese adults with MHO transitioned to a metabolically unhealthy state within just 2 years.^38,39 In the Nurses’ Health Study, the vast majority of women with MHO showed conversion to metabolically unhealthy obesity over 30 years.⁴⁰ Moreover, current evidence indicates that complex cardiometabolic multimorbidity and cardiovascular–kidney–metabolic syndrome commonly originates from excess and dysfunctional adipose tissue.⁴¹ These findings highlight the important of targeting upstream determinants and addressing excess weight at early points in the life course.

This study has several strengths. First, it used data from a large retrospective cohort with accurate and reliable measurements of metabolic diseases obtained through standardized health examinations.⁸ Second, mortality outcomes were ascertained through linkage with the national surveillance system, which provides complete, medically verified outcome data with standardized ICD coding and a low loss-to-follow-up rate (<1%). In addition, the study simultaneously included five common metabolic diseases, allowing for a detailed characterization of their coexistence patterns and associated mortality risk.

Our study has several limitations. First, the observational design precludes causal inference, and the findings should be interpreted as associations with increased risk. Second, our cohort consisted of relatively young individuals undergoing health examinations at a single center in Beijing, which may limit generalizability and likely contributes to lower absolute risks compared with the general Chinese population.⁴² Third, adjustment was limited to age and sex because data on lifestyle and socioeconomic factors were unavailable, which may have affected the precision of effect estimates.

In summary, this study found that metabolic multimorbidity was highly prevalent and exhibited distinct age- and sex-specific patterns, with a graded association with mortality. These findings support the need to move beyond a single-disease approach toward integrated strategies, including systematic screening and early intervention address shared upstream mechanisms. Future research leveraging omics-informed risk stratification, including metabolomics, may help identify high-risk populations and guide timely interventions for metabolic multimorbidity.