Skip to main content
NCBI home page
Springer logoLink to Springer
. 2026 Sep 26;91(1):303. doi: 10.1007/s12020-026-04810-5

Co-occurrence patterns and related risk factors of type 2 diabetes mellitus and metabolic dysfunction-associated steatotic liver disease

Jie Pan 1,#, Chengnan Guo 1,2,#, Shuzhen Zhao 2,#, Lina Zhang 1, Mengxia Chen 1, Yongping Liu 1, Fang Liu 1, Chao Zheng 1,✉
PMCID: PMC13616014  PMID: 42799817

Abstract

Introduction

Type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD) are interrelated metabolic disorders worldwide. We aimed to characterize global co-occurrence patterns of T2DM and MASLD from a spatial perspective and to identify shared risk factors.

Methods

We extracted incidence rates of T2DM and MASLD and exposure rates of 68 risk factors among individuals aged ≥ 20 years, derived from Global Burden of Disease Study 2021. Three co-occurrence areas were classified based on four quartiles of both diseases incidence. Interpretable machine learning and negative binomial linear mixed-effects model were applied to identify and quantify effects of risk factors. We calculated population-attributable fraction (PAF) to evaluate incidence burden of T2DM and MASLD attributable to identified risk factors.

Results

Fifty-nine countries were classified as consistent, 71 as T2DM-dominant, and 74 as MASLD-dominant. Three areas overlap spatially with exposure rates to metabolic, environmental, and behavioral risk factors. After feature selection, we identified nine risk factors with substantial effects on the comorbidity. Functionally, T2DM incidence was mainly attributable to metabolic risks (PAF = 35.28%) and behavioral risks (11.78%), whereas MASLD incidence was attributable to behavioral risks (28.84%), environmental risks (21.65%), and metabolic risks (14.63%). Overall, 42.9% of T2DM incidence and 52.41% of MASLD incidence were attributed to cumulative exposure to risk factors, with specific combinations and variations of these factors significantly impacting the disparate co-occurrence patterns worldwide.

Conclusion

Comorbidity of T2DM and MASLD is a prevalent global health issue. Co-management of T2DM and MASLD can proceed from interventions targeting corresponding modifiable risk factors.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s12020-026-04810-5.

Keywords: Type 2 diabetes mellitus, Metabolic dysfunction-associated steatotic liver disease, Global burden of disease, Risk factor

Introduction

Type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD) are major metabolic diseases worldwide, affecting approximately 10% and 30% of the global adult population, respectively [1, 2]. With the obesity pandemic, the incidence rates of T2DM and MASLD are predicted to rise steadily [1–4]. Extensive epidemiological evidence has indicated overlapping of both diseases, with 55.48% to 65.04% of T2DM patients comorbid with MASLD [5, 6]. The co-existence of T2DM and MASLD further increases the risk of adverse liver outcomes and extrahepatic outcomes (such as cardiovascular diseases and disability) [7], resulting in substantial public health and economic burden. Given that a considerable disease burden stems from modifiable risk factors [8, 9], studies on these factors are of vital importance for preventing both diseases.

T2DM and MASLD are associated with common risk factors due to shared pathogenic mechanisms, especially perturbations in glucose and lipid metabolism [10–12]. In addition to metabolic risk factors such as hyperglycemia and hyperlipidemia [1, 2, 7], environmental and behavioral modifiable factors also contribute substantially to the occurrence of both conditions [1, 13–15]. Elucidating the role of those factors and quantifying their impact is critical for primary prevention and clinical intervention of T2DM and MASLD. However, most studies treat both diseases as mutual risk factors or examine T2DM and MASLD separately. For example, A Korean study showed that persistent MASLD increase the risk of incident T2DM by 5.28 times [16]. A multinational meta-analysis showed that T2DM increased the risk of developing hepatic decompensation among MASLD patients [17]. High-quality global evidence on the co-occurrence of T2DM and MASLD remains unavailable because of challenges in data accessibility and analytical framework. Linking both diseases into a unified global syndemic framework will raise awareness of the magnitude and urgency of addressing these complex challenges [18].

Therefore, the objective of this study was to investigate the co-occurrence of T2DM and MASLD from a spatial perspective by analyzing data from the Global Burden of Disease Study (GBD). Through a comprehensive investigation, we aimed to reveal the co-occurrence patterns on a global scale and determine their corresponding risk factors. Precise identification of these risk factors at the national level provides guidance and evidence-based support for targeted prevention and control strategies for the co-management of T2DM and MASLD.

Methods

Data source

In our study, incidence rate data for T2DM and MASLD and exposure rate data for risk factors were sourced from the GBD database. The GBD 2021 database aggregated data on 371 diseases and injuries and 88 risk factors for 204 countries and territories. Details of the general methods for GBD 2021 have previously been described elsewhere [19, 20]. Diabetes was defined as a fasting plasma concentration of glucose of ≥ 126 mg/dL (7 mmol/L) or a report of treatment for diabetes. Since the inconsistent diagnostic criteria for T2DM in the methodological sections of studies, the incidence of T2DM was estimated indirectly by subtracting the estimation of type 1 diabetes mellitus from overall diabetes. According to the consensus statement of revising the “non-alcoholic fatty liver disease” to MASLD [21], this study renamed the total burden related to non-alcoholic fatty liver disease as MASLD. MASLD encompasses the spectrum of MASLD including fat deposition without cirrhosis and cirrhosis that can result from longstanding and progressive fat deposition and inflammation.

Incidence rates for T2DM and MASLD from 1990 to 2021 among adults aged ≥ 20 years were extracted, as well as exposure rates for all 68 most detailed risk factors, including environmental, occupational, behavioral, and metabolic factors. For instance, iron deficiency is defined as inadequate iron to meet the body’s needs. The exposure rate of iron deficiency is quantified in terms of mean hemoglobin concentration at the population level from the cumulative effect of all causes that lead to iron deficiency. In this analysis, each country was defined as a basic unit, and map information was spatially matched with disease incidence and exposure rates of risk factors.

Statistical analysis

To explore the co-occurrence characteristics of T2DM and MASLD, along with the differences in their global spatial distribution, we integrated incidence rates of T2DM and MASLD into a unified framework. Specifically, we first divided the incidence rates into four levels based on quartiles: low, lower-middle, upper-middle, and high. The countries or territories were then classified into consistent areas, T2DM-dominant areas, or MASLD-dominant areas based on the incidence level of two diseases. The country or territory was classified as a consistent area while incidence rate level of both diseases was identical; the country or territory was classified as a T2DM-dominant area while the T2DM incidence rate level was higher than that of MASLD; the country or territory was classified as a MASLD-dominant area while the MASLD incidence rate level was higher than that of T2DM.

To quantifying the associations of risk factors with T2DM and MASLD under stochastic theory, national-level incidence rates and exposure rates of risk factors were treated as stochastic variables. We did not rely on the existing risk factors for T2DM and MASLD as reported in previous population-level evidence, to evaluate the validity of new design. Instead, all 68 most detailed risk factors were selected through a three-step feature selection strategy. Firstly, based on expert knowledge and exploratory analysis, we excluded risk factors that showed no significant associations with both T2DM and MASLD, such as risk factors related to underage populations and occupational exposures. According to the definition, alcohol use was excluded from the potential risk factors of MASLD. Secondly, of the remaining 34 risk factors (33 for MASLD), the SHapley Additive exPlanation (SHAP) interpretable method was applied to rank the feature importance of each risk factor in a gradient-boosting machine learning model (XGBoost). SHAP dictated the estimated average marginal contribution of the individual factors for a prediction with local additivity and can be applied to any machine learning algorithm [22]. The main advantage of machine learning in design as it can address numerous risk factors and complex issues, such as collinearity and nonlinearity. We considered the top 15 variables with positive effects in further analysis for T2DM and MASLD to ensure a consistent screening criterion. Finally, the generalized linear mixed-effects model with negative binomial distribution was used to determine the risk factors for both diseases.

For identified risk factors, we applied negative binomial linear mixed-effects model to estimate the relative risk (RR) and calculated the population attributable fraction (PAF) according to the RR value. The model treat incidence count as the outcome, with risk factor exposure rates (all risk factors were standardized), age, gender, social development index, and year (1990–2021, random effect variable) as the independent variables, and the country-level population as an offset. We also applied the PAF to evaluate disease burden of T2DM and MASLD attributable to identified risk factors across different co-occurrence areas. The PAF of the  ith risk factor for the  jth ( j=1,2) disease in the  kth ( k=1,2,3) area is calculated by the following formula: [23]

graphic file with name d33e346.gif

where  x indicates exposure level of risk factors. Inline graphic is derived from the negative binomial linear mixed-effects regression and represents the RR value measuring the association of the  ith risk factor with the  jth disease at exposure level  x. Inline graphic denotes the exposure distribution of the  ith risk factor in the  kth area, while Inline graphic denotes the theoretical distribution of the minimum-risk exposure level for the  ith risk factor, estimated from exposure level distributions below the median global exposure rate across countries and territories. The joint PAF for joint exposure of risk factor clusters attributed to the  jth disease in the  kth area is calculated as follows: [23]

graphic file with name d33e404.gif

An integrated risk index of identified risk factors was constructed to reflect the overall exposure risk intensity of country or territory. Specifically, the exposure rates of risk factors were divided into four quartiles and assigned values from 0 to 3 based on the exposure level. The integrated risk index was then defined as the sum of scores for identified risk factors, with higher score representing greater overall exposure risk. We used R version 4.3.3 (R Project for Statistical Computing) for statistical analyses and ArcGIS release 10.6 (ESRI, Redlands, CA, USA) for regional mapping and spatial localization.

Results

Regional division of co-occurrence characteristics for T2DM and MASLD

Overall, the global incidence rate of MASLD was higher than that of T2DM among adults aged ≥ 20 years, with incidence rates of 789.83 per 100,000 population for MASLD and 435.14 per 100,000 population for T2DM. The spatial distribution of T2DM and MASLD incidence rates exhibited certain heterogeneity, characterized by higher incidence rates in the Eastern Mediterranean, North America, and North Africa of T2DM and higher rates in the Eastern Mediterranean, Africa, and Southeast Asia of MASLD (Fig. 1).

Fig. 1.

The image presents three world maps comparing the incidence rates of type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD) across countries. The first map shows T2DM incidence per 100,000 people, with the highest rates in the United States, parts of the Middle East, and some countries in South America and North Africa. The second map shows MASLD incidence per 100,000 people, with the highest rates in countries in the Middle East, North Africa, and parts of Southeast Asia. The third map categorizes countries by whether MASLD incidence is dominant, T2DM incidence is dominant, or if both diseases have consistent incidence levels. MASLD dominance is seen mainly in parts of Africa, the Middle East, and South America; T2DM dominance is observed in North America, parts of Europe, and some countries in Asia; consistent incidence occurs in various countries across different continents. A matrix shows the percentile distribution of MASLD and T2DM incidences, indicating how countries fall into categories of dominance or consistency based on their relative incidence percentiles.

Open in a new tab

Global distribution of the incidence of T2DM (A) and MASLD (B), and characteristics of co-occurrence in the incidence of T2DM and MASLD (C). The country or territory was classified as a consistent area if the incidence rate levels of both diseases were identical. T2DM=Type 2 diabetes mellitus. MASLD=metabolic dysfunction-associated steatotic liver disease

The co-occurrence patterns of T2DM and MASLD were depicted in Fig. 1C. The number of countries covered by T2DM-dominant (71 [34.8%]) and MASLD-dominant (74 [36.3%]) areas was approximately equal (Table S1). The consistent areas included 59 (28.9%) of 204 countries and territories. Countries with the highest incidence rates for both diseases were predominantly located in Eastern Mediterranean, Southeast Asia, and North Africa (e.g., Morocco: T2DM 994.93/100,000 population and MASLD 1341.76/100,000 population). T2DM-dominant areas were mainly located within countries in North America and Europe (e.g., USA, Canada, and Russia). In the USA, the incidence rate of T2DM was at the highest level (783.27/100,000 population), whereas the incidence rate of MASLD was at the low level (520.47/100,000 population). Meanwhile, MASLD-dominant areas were mainly located within countries in East Asia, Africa, and South America (e.g., China, Egypt, and Brazil). In China, the incidence rate of MASLD was at the upper-middle level (806.28/100,000 population), whereas the incidence rate of T2DM was at the lower-middle level (334.7/100,000 population).

Interpretable machine-learning framework for risk factor selection

The SHAP value of the XGBoost model was applied to investigate the 34 risk factors (33 for MASLD) potentially influencing the incidence of T2DM and MASLD (Fig. 2). In terms of feature selection, the top 15 risk factors with positive effects served as the candidate risk factors of T2DM and MASLD. Eight risk factors were included for T2DM: high fasting plasma glucose, high body-mass index (BMI), diet high in sodium, diet low in legumes, diet low in nuts and seeds, high low-density lipoprotein (LDL) cholesterol, low physical activity, and diet low in fruits. Nine risk factors were included for MASLD: high BMI, ambient particulate matter pollution, high temperature, low physical activity, diet low in seafood omega-3 fatty acids, diet low in calcium, iron deficiency, high fasting plasma glucose, and high LDL cholesterol. The generalized linear mixed-effects model with negative binomial distribution was then used to assess the relative risk of candidate risk factors with both diseases after controlling for covariates. Eventually, nine risk factors with substantial effects were identified, including three common risk factors for both T2DM and MASLD, one specific risk factor for T2DM, and five specific risk factors for MASLD (Fig. 3).

Fig. 2.

The plot ranks risk factors by their influence on the incidence of Type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD). For T2DM, high fasting plasma glucose has the strongest positive influence, followed by a diet high in trans fatty acids and high body-mass index. Other factors with positive influence include diets high in sodium, red meat, and processed meat, as well as low temperature and low intake of legumes. Negative influences include low bone mineral density, intimate partner violence, and smoking. For MASLD, a diet high in processed meat has the strongest positive influence, followed by diets high in sodium and high body-mass index. Environmental factors like ambient particulate matter pollution and high temperature also show positive influence. Factors with smaller positive influence include low physical activity, low intake of seafood omega-3 fatty acids, fiber, and calcium, iron deficiency, drug use, and smoking. Some factors show negative influence, such as low temperature and secondhand smoke. The SHAP value scale quantifies the strength and direction of each factor's influence on disease risk.

Open in a new tab

SHAP summary plot of T2DM (A) and MASLD (B). Plots display variables ranked by importance in terms of risk factor exposure from top to bottom. The SHAP value on the horizontal axis represents the influence of the factor on the outcome, with positive values indicating promotion and negative values indicating inhibition. The larger the absolute value of SHAP, the greater its impact on disease incidence. T2DM=Type 2 diabetes mellitus. MASLD=metabolic dysfunction-associated steatotic liver disease. LDL = low-density lipoprotein. SHAP=Shapley additive explanations

Fig. 3.

The image presents relative risk values for various risk factors affecting Type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD). For T2DM, common risk factors include high fasting plasma glucose with a relative risk of 1.462, high body-mass index at 1.067, and low physical activity at 1.018. A risk factor specific to T2DM is a diet high in sodium, with a relative risk of 1.061. For MASLD, common risk factors are high fasting plasma glucose at 1.086, high body-mass index at 1.108, and low physical activity at 1.013. MASLD-specific risk factors include ambient particulate matter pollution with a relative risk of 1.161, high temperature at 1.031, diet low in seafood omega-3 fatty acids at 1.034, diet low in calcium at 1.057, and iron deficiency with the highest relative risk of 1.236. All relative risk values are presented with 95 percent confidence intervals.

Open in a new tab

Relative risk of risk factors affecting T2DM (A) and MASLD (B). Relative risk values were obtained by generalized linear mixed-effects model with negative binomial distribution independently conducted in this study, with incidence count as the outcome, risk factor exposure rates, age, gender, social development index, and year (1990–2021, random effect variable) as the independent variables, and the country-level population as an offset in the model. T2DM=Type 2 diabetes mellitus. MASLD=metabolic dysfunction-associated steatotic liver disease. RR=relative risk

Three common risk factors were high fasting plasma glucose, high BMI, and low physical activity. High fasting plasma glucose imposed the highest effect on T2DM (RR = 1.462 [95% CI, 1.451–1.474]), indicating that for per 1 SD increase in fasting plasma glucose exposure, the incidence rate of T2DM increased by 46.2%. Similarly, high BMI significantly affected MASLD (1.108 [1.100-1.115]). The sole specific risk factor significantly affecting T2DM was diet high in sodium (1.061 [1.055–1.067]). For MASLD, the five significant factors included ambient particulate matter pollution (1.161 [1.155–1.167]), high temperature (1.031 [1.026–1.036]), diet low in seafood omega-3 fatty acids (1.034 [1.029–1.040]), diet low in calcium (1.057 [1.050–1.063]), and iron deficiency (1.236 [1.218–1.254]).

Population attribution of risk factors

The PAF of risk factors for the incidence rate of T2DM and MASLD was estimated in Table 1. Globally, 42.90% of the T2DM incidence and 52.41% of the MASLD incidence were attributed to their respective risk factors. Metabolic risks contributed the highest population attribution to T2DM (35.28%), while behavioral risks contributed the highest population attribution to MASLD (28.84%). In particular, high fasting plasma glucose contributed most to T2DM incidence, accounting for 31.10%, while iron deficiency contributed most to MASLD incidence, accounting for 20.70%. Meanwhile, the environmental risks specifically contributed a 21.65% proportion to MASLD, including ambient particulate matter pollution (18.00%) and high temperature (4.45%).

Table 1.

Population attributable fractions for MASLD and T2DM of various risk factors

All countries and territories Consistent area T2DM-dominant area MASLD-dominant area
T2DM MASLD T2DM MASLD T2DM MASLD T2DM MASLD
Individual factors
High fasting plasma glucose 31.10(30.00,32.20) 7.32(6.82,7.89) 14.00(12.40,15.80) 10.10(9.15,11.10) 23.30(21.40,25.40) 6.61(5.65,7.51) 35.20(33.30,37.10) 0.81(0.38,1.20)
High body-mass index 6.06(5.35,6.78) 7.89(7.18,8.56) 10.00(8.61,11.20) 5.26(4.55,5.98) 8.73(7.52,9.90) 5.52(4.46,6.46) 14.00(12.60,15.40) 8.09(7.24,8.80)
Low physical activity 1.87(1.18,2.62) 1.2(0.64,1.76) 3.27(2.05,4.43) 2.15(0.26,4.13) 4.14(2.66,5.62) 3.92(3.14,4.95) 3.80(3.15,4.48) 5.16(4.41,5.74)
Diet high in sodium 10.10(9.06,10.90) NA 4.84(3.71,5.92) NA 9.05(8.17,10.10) NA 6.83(5.14,8.72) NA
Ambient particulate matter pollution NA 18.00(17.00,18.90) NA 20.10(18.80,21.60) NA 5.89(4.79,6.79) NA 14.00(12.20,15.70)
High temperature NA 4.45(3.72,5.20) NA 1.35(-1.33,3.45) NA 1.10(0.32,1.97) NA 3.14(2.51,3.83)
Diet low in seafood omega-3 fatty acids NA 4.78(4.10,5.50) NA 17.50(15.60,19.60) NA 3.51(2.45,4.45) NA 1.96(1.18,2.83)
Diet low in calcium NA 4.62(3.94,5.29) NA 7.46(6.28,8.66) NA 0.90(0.07,1.61) NA 2.62(2.09,3.06)
Iron deficiency NA 20.70(18.90,22.60) NA 26.10(21.90,30.40) NA 13.10(11.50,14.70) NA 3.17(1.32,5.41)
Risk class
Metabolic risks 35.28(33.75,36.80) 14.63(13.51,15.77) 22.6(19.94,25.23) 14.83(13.28,16.42) 30.00(27.31,32.79) 11.77(9.86,13.48) 44.27(41.70,46.79) 8.83(7.59,9.89)
Environmental risks NA 21.65(20.09,23.12) NA 21.18(17.72,24.30) NA 6.93(5.10,8.63) NA 16.70(14.40,18.93)
Behavioral risks 11.78(10.13,13.23) 28.84(25.77,31.95) 7.95(5.68,10.09) 44.79(38.38,51.00) 12.82(10.61,15.15) 20.16(16.44,23.78) 10.37(8.13,12.81) 12.33(8.73,16.01)
Common/Special factors
Common factors 36.49(34.53,38.45) 15.66(14.06,17.26) 25.13(21.58,28.54) 16.66(13.51,19.87) 32.89(29.24,36.56) 15.22(12.69,17.77) 46.39(43.54,49.17) 13.54(11.67,15.07)
Special factors 10.10(9.06,10.90) 43.57(40.30,46.74) 4.84(3.71,5.92) 55.53(49.17,61.31) 9.05(8.17,10.10) 22.66(18.13,26.73) 6.83(5.14,8.72) 22.99(18.27,27.77)
All 42.90(40.46,45.16) 52.41(48.69,55.93) 28.75(24.49,32.77) 62.94(56.04,69.00) 38.97(35.02,42.97) 34.43(28.52,39.74) 50.05(46.44,53.60) 33.42(27.81,38.65)
Open in a new tab

T2DM=Type 2 diabetes mellitus. MASLD=metabolic dysfunction-associated steatotic liver disease. NA = not applicable

While stratifying by co-occurrence area, the PAFs were 28.75% for T2DM and 62.94% for MASLD in the consistent area; PAFs were 38.97% for T2DM and 34.43% for MASLD in the T2DM-dominant area; and PAFs were 50.05% for T2DM and 33.42% for MASLD in the MASLD-dominant area. Among common factors, the PAF of T2DM was consistently higher than that of MASLD (25–46% vs. 14–17%). Among special factors, the PAF of T2DM was lower than that of MASLD (5–10% vs. 23–56%).

Integrated risk index and risk factor localization

The spatial distribution characteristics of the nine risk factors are shown in Fig. 4. Overall, the highest exposure rates of metabolic risk factors (including high fasting plasma glucose and high BMI) were mainly in North America, North Africa, and Eastern Mediterranean. The highest exposure rates of environmental risk factors (including ambient particulate matter pollution and high temperature) were mainly located near the equator. Regarding behavioral risk factors, heterogeneity existed across countries. Low physical activity was most prevalent in Western Pacific, Mediterranean, and South America. Diet low in seafood omega-3 fatty acids was prevalent in the inland regions of Asia and Africa. Diet low in calcium and iron deficiency were consistently prevalent in Africa. Diet high in sodium was prevalent in the Europe, America, and Asia.

Fig. 4.

The image presents global maps showing the incidence rates of nine health risk factors across countries, categorized into four levels: lowest, lower middle, upper middle, and highest. For high fasting plasma glucose, the highest incidence is seen in the United States, parts of the Middle East, and Australia. High body-mass index is most prevalent in the United States, parts of Europe, and Australia. Ambient particulate matter pollution is highest in South Asia, parts of East Asia, and some African regions. High temperature incidence is concentrated in northern Africa and the Middle East. Low physical activity rates are highest in South America, parts of Africa, and Australia. Diet low in seafood omega-3 fatty acids is most common in parts of Africa, South Asia, and the Middle East. Diet low in calcium shows highest incidence in parts of Africa and Southeast Asia. Iron deficiency is most prevalent in central and eastern Africa and parts of South Asia. Diet high in sodium has the highest incidence in North America, Europe, and East Asia. Each map highlights distinct geographic patterns of these risk factors without relying on color alone, using labeled incidence categories to convey the data.

Open in a new tab

Global distribution of nine identified risk factors

To reflect the overall exposure risk intensity of country or territory, the integrated risk index of identified risk factors was constructed, ranging from 0 to 27 (Fig. 5 and Table S2). A higher value indicated more severe combined exposure and higher risks of T2DM and MASLD. Regions with the highest integrated risk index were relatively concentrated, mainly in North Africa (e.g., Algeria and Morocco) and Eastern Mediterranean (e.g., Iraq and Palestine). Regions with the lowest risk index were also relatively concentrated, mainly in Europe (e.g., Sweden and Denmark) and central Africa (e.g., Kenya and Burundi). This distribution pattern of identified risk factors matched the distribution of T2DM and MASLD closely, indicating that the incidence rates of T2DM and MASLD were related to the integrated effects of multiple risk factors.

Fig. 5.

The image presents two key pieces of information about a composite risk index for nine risk factors. The first part is a world map categorizing countries into four risk index ranges: 6 to 11, 12 to 13, 14 to 15, and 16 to 22. Countries in North America, Northern Europe, Australia, and parts of Africa mostly fall into the lowest risk range of 6 to 11. Countries in Central and South America, parts of Africa, and Asia are distributed across the middle ranges of 12 to 13 and 14 to 15. The highest risk range of 16 to 22 is concentrated in parts of Central America, northern Africa, the Middle East, and South Asia. The second part is a histogram showing the frequency distribution of the risk index values. The risk index values range from about 6 to 22, with the highest frequency counts occurring between 11 and 15. The distribution is roughly symmetrical, with fewer countries at the extreme low and high ends of the risk index.

Open in a new tab

Global distribution of quartile levels (A) and frequency distribution (B) of the composite risk index for nine risk factors

Discussion

Global co-occurrence patterns of T2DM and MASLD from a spatial perspective were proposed in this study. Candidate risk factors were functionally distinguished and quantitatively assessed in each co-occurrence area of T2DM and MASLD. The findings demonstrated that the geographic distribution of three co-occurrence areas overlapped with exposure to metabolic, behavioral, and environmental risk factors. Furthermore, nine identified risk factors of T2DM and MASLD were classified into common and specific risk factors. Overall, 42.9% of the T2DM incidence and 52.41% of the MASLD incidence were attributed to cumulative exposure to their corresponding risk factors. The global disparities of T2DM and MASLD incidence were greatly influenced by the specific combinations and variations of identified risk factors.

T2DM and MASLD tend to co-exist and share specific underlying mechanisms, and it is common to encounter patients with both conditions in clinical practice [7, 24, 25]. Moreover, the shared pathogenic mechanisms contribute to the development of cardiovascular diseases, adverse liver outcomes, disability, and premature death [7]. Our results reported certain heterogeneity of global geographical distribution and socioeconomic status on T2DM and MASLD, with higher T2DM incidence in the Eastern Mediterranean, North America, and North Africa and higher MASLD incidence in the Eastern Mediterranean, Africa, and Southeast Asia. The disparities shed light on priority areas for public health measures, where a one-size-fits-all approach is unlikely to effectively address important health determinants at both the individual and population level [18]. Identifying of common drivers of T2DM and MASLD within populations is critical to implementing precise strategies to reduce these entrenched contributors. Therefore, we further explore the co-occurrence patterns of T2DM and MASLD. T2DM-dominated areas were mainly high-income countries, whereas MASLD-dominated areas were mainly low- and middle-income countries. The above phenomenon indicates that economic development is the primary strategic role in the prevention of metabolic diseases.

Our findings indicated that two environmental risk factors, including ambient particulate matter pollution and high temperature, were specifically related to the occurrence of MASLD. Previous cohort studies established that higher PM2.5 and PM10 exposure contribute to the morbidity of MASLD [13, 26–28]. A recent Mendelian randomization study in the UK further validated the causal relationship between PM2.5 and MASLD [14]. Our results further showed that ambient particulate matter pollution accounted for a relatively high proportion (18%) of MASLD incidence from a global population perspective. Regarding the association between high temperature and MASLD, several plausible mechanisms may be involved [29, 30]. First, elevated ambient temperatures may discourage outdoor physical activity, leading to reduced energy expenditure and increased sedentary behavior, which in turn promotes hepatic fat accumulation. Second, under warmer conditions, the metabolic demand for converting food energy into body heat is diminished, potentially tipping the energy balance toward storage rather than dissipation, thereby facilitating lipid deposition in the liver. Third, heat exposure may induce low-grade systemic inflammation and oxidative stress, which could exacerbate insulin resistance and hepatic steatosis. These pathways, either independently or synergistically, may explain the observed positive association between high temperature and MASLD incidence. Although a recent study has suggested a potential association between non-optimal temperature exposure and MASLD burden [31], our study complements this finding by quantifying the PAFs and integrating temperature as part of a broader risk factor framework within the co-occurrence context of T2DM and MASLD. The findings further showed that among 50 countries with the most severe exposure to high temperatures (e.g., Egypt and Libya), the average incidence rates of MASLD ranked at a high or middle level. The establishment of environmental early warning systems (e.g., smart heat-health warning systems) and improvements of green infrastructure in urban areas can address the dual objectives of promoting global health and environmental sustainability [32, 33].

Iron deficiency was also identified as an important risk factor for MASLD incidence in this study. Previous experimental studies have shown that iron distribution disorders contribute to the occurrence of MASLD, which were characterized by iron-deficient hepatocytes and iron overload in hepatic stellate cells [34]. This study further demonstrated the relationship from a global perspective. Moreover, iron deficiency is more common in obesity populations, suggesting that prevention and control of the alarming increase in MASLD burden over the past two decades needs to consider the potential synergistic effects of obesity and iron deficiency [35, 36].

This study further revealed high fasting plasma glucose, high BMI, low physical activity, and unhealthy dietary habits as joint promoting factors for T2DM and MASLD. Both obesity and hyperglycemia are recognized as key drivers of T2DM and MASLD [1, 2, 7], indicating the urgency of integrating collective metabolic risk factors into a unified global syndemic framework and prioritizing upstream solutions to mitigate the holistic metabolic environment of patients [18]. The global shift in lifestyle practices, including low physical activity and changes in dietary habits, forecast a large metabolic wave, which will pose greater challenges to global healthcare systems. The important stakeholders in policy-making should propose target prevention strategies of upstream factors that mitigate the overall metabolic milieu, such as weight reduction, increased physical activity, and ideal dietary habits. Our study indicated that diets low in seafood omega-3 fatty acids and calcium were specific risk factors for MASLD, and diets high in sodium were a specific risk factor for T2DM. Countries dominated by MASLD had the lowest exposure rates to seafood omega-3 fatty acids and calcium, whereas T2DM-dominated countries had highest exposure rates to sodium. Currently, the Mediterranean diet, which is rich in omega-3 fatty acids and low in sodium intake, is highly recommended for its cholesterol-lowering, anti-inflammatory, and blood pressure-lowering effects [37]. Health policy makers are encouraged to prioritize the development of easy-to-implement and culturally appropriate dietary guidelines, particularly in low- and middle-income countries where dietary guidelines are lacking. When sustained weight reduction through lifestyle interventions is not effective, hypoglycemic medications, emerging therapeutic drugs, and bariatric surgery are viable clinical alternatives for early treatment of T2DM and MASLD [7].

This study applied a spatial perspective to comprehensively evaluate the co-occurrence patterns of T2DM and MASLD on a global scale. The framework allowed us to address the limitations of existing clinical and epidemiological investigations effectively, while preserving the explanatory capacity of epidemiological studies. Moreover, we treated the national-level incidence rates and summary exposure rates of risk factors as stochastic variables, considering their global variations as random variation. Under this assumption, we developed an independent modeling framework that enabled integrated quantitative analyses, including machine learning and classical statistical approaches. This methodology differs from previous GBD studies, which predominantly rely on existing disease burden estimates for secondary descriptive or predictive analyses [3, 18]. Instead, our approach systematically extracts insights from random data level, providing theoretical support for subsequent quantitative analyses. The methodology allows us to identify the precise targets for prevention and control strategies, informing global public health policy formulation of T2DM and MASLD.

Our study has several limitations. Firstly, the candidate risk factor set is limited to the 68 risk factors evaluated by GBD 2021. Risk factors related to T2DM and MASLD co-occurrence from other perspectives could not be included owing to the limitations of GBD data. Secondly, T2DM and MASLD are the consequences resulting from the long-term accumulation of various risk factors. Although we applied a generalized linear mixed-effects model to estimate the longitudinal effects of risk factors from 1990 to 2021, the potential lagging effect should be adequately assessed in the future. Third, although this study applied a longitudinal analysis, we could only identify associations rather than establish causal relationships.

Conclusion

Global management of T2DM and MASLD should adopt a unified global syndemic framework, treating both individual diseases and comorbidities while recognizing their unique properties. High fasting plasma glucose, high BMI, low physical activity, and unhealthy dietary habits serve as joint promoting factors for the co-occurrence of T2DM and MASLD. Meanwhile, environmental risk factors, including ambient particulate matter pollution and high temperature, and iron deficiency were specific risk factors for incidence of MASLD.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (23.9KB, docx)

Acknowledgements

We appreciate the works of the Global Burden of Disease study 2021 collaborators.

Author contributions

Concept and design: JP, CG, CZ. Data collection: CG. Statistical analysis: CG, SZ. Results interpretations: All authors. Manuscript writing: JP, CG. Manuscript proofing: CZ. Obtained funding: JP, CG. Supervision: CZ. Prof. Chao Zheng is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Funding

This research was supported by the Huadong Medicine Joint Funds of the Zhejiang Provincial Natural Science Foundation of China (No. LHDMZ24H070001) and the China Association for Science and Technology Youth Talent Support Program for Doctoral Students (156-O-330-0000096-8).

Data availability

The data involved in this study are available on the Global Health Data Exchange GBD 2021 website (https://ghdx.healthdata.org/gbd-2021).

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Jie Pan, Chengnan Guo and Shuzhen Zhao contributed equally to this work.

References

  • 1.E. Ahmad, S. Lim, R. Lamptey, D.R. Webb, M.J. Davies, Type 2 diabetes. Lancet. 400, 1803–1820 (2022). 10.1016/S0140-6736(22)01655-5 [DOI] [PubMed] [Google Scholar]
  • 2.D.Q. Huang, V.W.S. Wong, M.E. Rinella, J. Boursier, J.V. Lazarus, H. Yki-Järvinen et al., Metabolic dysfunction-associated steatotic liver disease in adults. Nat. Rev. Dis. Primers. 11, 14 (2025). 10.1038/s41572-025-00599-1 [DOI] [PubMed] [Google Scholar]
  • 3.J. Xie, M. Wang, Z. Long, H. Ning, J. Li, Y. Cao et al., Global burden of type 2 diabetes in adolescents and young adults, 1990–2019: systematic analysis of the Global Burden of Disease Study 2019. BMJ. 379, e072385 (2022). 10.1136/bmj-2022-072385 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Z. Zarkasi, R.N. Mudaliar, W.A. Majeed, A.A. Syed, Obesity and metabolic syndrome: Current insights and future directions. Br. J. Hosp. Med. (Lond). 86, 1–20 (2025). 10.12968/hmed.2025.0127 [DOI] [PubMed] [Google Scholar]
  • 5.E. En Li Cho, C.Z. Ang, J. Quek, C.E. Fu, L.K.E. Lim, Z.E.Q. Heng et al., Global prevalence of non-alcoholic fatty liver disease in type 2 diabetes mellitus: an updated systematic review and meta-analysis. Gut. 72, 2138–2148 (2023). 10.1136/gutjnl-2023-330110 [DOI] [PubMed] [Google Scholar]
  • 6.Z.M. Younossi, P. Golabi, de L. Avila, J.M. Paik, M. Srishord, N. Fukui et al., The global epidemiology of NAFLD and NASH in patients with type 2 diabetes: A systematic review and meta-analysis. J. Hepatol. 71, 793–801 (2019). 10.1016/j.jhep.2019.06.021 [DOI] [PubMed] [Google Scholar]
  • 7.X. Qi, J. Li, C. Caussy, G.-J. Teng, R. Loomba, Epidemiology, screening, and co-management of type 2 diabetes mellitus and metabolic dysfunction–associated steatotic liver disease. Hepatology. 83, 661–678 (2026). 10.1097/HEP.0000000000000913 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.N.A. ElSayed, G. Aleppo, V.R. Aroda, R.R. Bannuru, F.M. Brown, D. Bruemmer et al., 5. Facilitating Positive Health Behaviors and Well-being to Improve Health Outcomes: Standards of Care in Diabetes-2023. Diabetes Care. 46(Supple 1), S68–96 (2023). 10.2337/dc23-S005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.M.E. Rinella, B.A. Neuschwander-Tetri, M.S. Siddiqui, M.F. Abdelmalek, S. Caldwell, D. Barb et al., AASLD Practice Guidance on the clinical assessment and management of nonalcoholic fatty liver disease. Hepatology. 77, 1797–1835 (2023). 10.1097/HEP.0000000000000323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.G. Targher, K.E. Corey, C.D. Byrne, M. Roden, The complex link between NAFLD and type 2 diabetes mellitus - mechanisms and treatments. Nat. Rev. Gastroenterol. Hepatol. 18, 599–612 (2021). 10.1038/s41575-021-00448-y [DOI] [PubMed] [Google Scholar]
  • 11.R.J. Perry, V.T. Samuel, K.F. Petersen, G.I. Shulman, The role of hepatic lipids in hepatic insulin resistance and type 2 diabetes. Nature. 510, 84–91 (2014). 10.1038/nature13478 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.M.J. Watt, P.M. Miotto, De W. Nardo, M.K. Montgomery, The Liver as an Endocrine Organ-Linking NAFLD and Insulin Resistance. Endocr. Rev. 40, 1367–1393 (2019). 10.1210/er.2019-00034 [DOI] [PubMed] [Google Scholar]
  • 13.P. Deng, H. Tang, L. Zhu, J. Duan, F. Li, Y. Li et al., Association of long-term ambient fine particulate matter (PM(2.5)) and incident non-alcoholic fatty liver disease in Chinese adults. Environ. Pollut. 329, 121666 (2023). 10.1016/j.envpol.2023.121666 [DOI] [PubMed] [Google Scholar]
  • 14.S. Ran, J. Zhang, F. Tian, Z.M. Qian, S. Wei, Y. Wang et al., Association of metabolic signatures of air pollution with MASLD: Observational and Mendelian randomization study. J. Hepatol. 82, 560–570 (2025). 10.1016/j.jhep.2024.09.033 [DOI] [PubMed] [Google Scholar]
  • 15.M.S. Kim, I. Shim, A.C. Fahed, R. Do, W.-Y. Park, P. Natarajan et al., Association of genetic risk, lifestyle, and their interaction with obesity and obesity-related morbidities. Cell. Metab. 36, 1494–1503e3 (2024). 10.1016/j.cmet.2024.06.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.E. Han, K.-D. Han, Y.-H. Lee, K.-S. Kim, S. Hong, J.H. Park et al., Association of temporal MASLD with type 2 diabetes, cardiovascular disease and mortality. Cardiovasc. Diabetol. 24, 289 (2025). 10.1186/s12933-025-02824-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.D.Q. Huang, N. Noureddin, V. Ajmera, M. Amangurbanova, R. Bettencourt, E. Truong et al., Type 2 diabetes, hepatic decompensation, and hepatocellular carcinoma in patients with non-alcoholic fatty liver disease: an individual participant-level data meta-analysis. Lancet Gastroenterol. Hepatol. 8, 829–836 (2023). 10.1016/S2468-1253(23)00157-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.N.W.S. Chew, C.H. Ng, D.J.H. Tan, G. Kong, C. Lin, Y.H. Chin et al., Cell. Metab. 35, 414–428e3 (2023). The global burden of metabolic disease: Data from 2000 to 2019 10.1016/j.cmet.2023.02.003 [DOI] [PubMed] [Google Scholar]
  • 19.GBD 2021 Risk Factors Collaborators, Global burden and strength of evidence for 88 risk factors in 204 countries and 811 subnational locations, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 403, 2162–2203 (2024). 10.1016/S0140-6736(24)00933-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.GBD 2021 Diseases and Injuries Collaborators, Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 403, 2133–2161 (2024). 10.1016/S0140-6736(24)00757-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.M.E. Rinella, J.V. Lazarus, V. Ratziu, S.M. Francque, A.J. Sanyal, F. Kanwal et al., A multisociety Delphi consensus statement on new fatty liver disease nomenclature. J. Hepatol. 79, 1542–1556 (2023). 10.1016/j.jhep.2023.06.003 [DOI] [PubMed] [Google Scholar]
  • 22.C. Guo, Z. Liu, H. Fan, H. Wang, X. Zhang, S. Zhao et al., Machine-learning–based plasma metabolomic profiles for predicting long-term complications of cirrhosis. Hepatology. 81, 168–180 (2025). 10.1097/HEP.0000000000000879 [DOI] [PubMed] [Google Scholar]
  • 23.M. Ezzati, S.V. Hoorn, A. Rodgers, A.D. Lopez, C.D. Mathers, C.J.L. Murray, Estimates of global and regional potential health gains from reducing multiple major risk factors. Lancet. 362, 271–280 (2003). 10.1016/s0140-6736(03)13968-2 [DOI] [PubMed] [Google Scholar]
  • 24.N. Stefan, K. Cusi, A global view of the interplay between non-alcoholic fatty liver disease and diabetes. Lancet Diabetes Endocrinol. 10, 284–296 (2022). 10.1016/S2213-8587(22)00003-1 [DOI] [PubMed] [Google Scholar]
  • 25.H. Zhang, X.-D. Zhou, M.D. Shapiro, G.Y.H. Lip, H. Tilg, L. Valenti et al., Global burden of metabolic diseases, 1990–2021. Metabolism. 160, 155999 (2024). 10.1016/j.metabol.2024.155999 [DOI] [PubMed] [Google Scholar]
  • 26.F.-R. Li, J. Liao, B. Zhu, X. Li, Z. Cheng, C. Jin et al., Long-term exposure to air pollution and incident non-alcoholic fatty liver disease and cirrhosis: A cohort study. Liver Int. 43, 299–307 (2023). 10.1111/liv.15416 [DOI] [PubMed] [Google Scholar]
  • 27.S. Sun, Q. Yang, Q. Zhou, W. Cao, S. Yu, S. Zhan et al., Long-term exposure to air pollution, habitual physical activity and risk of non-alcoholic fatty liver disease: A prospective cohort study. Ecotoxicol. Environ. Saf. 235, 113440 (2022). 10.1016/j.ecoenv.2022.113440 [DOI] [PubMed] [Google Scholar]
  • 28.T. VoPham, N.J. Kim, K. Berry, J.A. Mendoza, J.D. Kaufman, G.N. Ioannou, PM(2.5) air pollution exposure and nonalcoholic fatty liver disease in the Nationwide Inpatient Sample. Environ. Res. 213, 113611 (2022). 10.1016/j.envres.2022.113611 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.L. Wei, H. Zhang, L. Chen, X. Chen, D. Sun, P. Pei et al., Short-term effects of temperature on physical activity and sedentary behavior in Chinese adults: An accelerometer-based analysis. Environ. Int. 210, 110203 (2026). 10.1016/j.envint.2026.110203 [DOI] [PubMed] [Google Scholar]
  • 30.S. Henkel, P. Frings-Meuthen, C. Diekmann, M. Coenen, B. Stoffel-Wagner, R. Németh et al., Influence of ambient temperature on resting energy expenditure in metabolically healthy males and females. J. Nutr. 155, 862–870 (2025). 10.1016/j.tjnut.2025.01.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Y. Wang, X. Gao, Q. Wang, Y. Li, Q. Li, Y. Xing et al., Nonlinear climate–liver gradients: Non-optimal temperature and variability increase risk of metabolic dysfunction-associated steatotic liver disease. Global Transitions. 8, 259–268 (2026). 10.1016/j.glt.2026.02.003 [DOI] [Google Scholar]
  • 32.P. Kumar, A. Druckman, J. Gallagher, B. Gatersleben, S. Allison, T.S. Eisenman et al., The nexus between air pollution, green infrastructure and human health. Environ. Int. (2019). 133 Pt A:105181. 10.1016/j.envint.2019.105181 [DOI] [PubMed] [Google Scholar]
  • 33.T. Li, C. Chen, W. Cai, The global need for smart heat-health warning systems. Lancet. 400, 1511–1512 (2022). 10.1016/S0140-6736(22)01974-2 [DOI] [PubMed] [Google Scholar]
  • 34.H. Gao, Z. Jin, G. Bandyopadhyay, G. Wang, D. Zhang, K.C.E. Rocha et al., Aberrant iron distribution via hepatocyte-stellate cell axis drives liver lipogenesis and fibrosis. Cell. Metab. 34, 1201–1213e5 (2022). 10.1016/j.cmet.2022.07.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.A. Siddique, J.E. Nelson, B. Aouizerat, M.M. Yeh, K.V. Kowdley, Iron deficiency in patients with nonalcoholic Fatty liver disease is associated with obesity, female gender, and low serum hepcidin. Clin. Gastroenterol. Hepatol. 12, 1170–1178 (2014). 10.1016/j.cgh.2013.11.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.G.E.H. Lim, A. Tang, C.H. Ng, Y.H. Chin, W.H. Lim, D.J.H. Tan et al., An Observational Data Meta-analysis on the Differences in Prevalence and Risk Factors Between MAFLD vs NAFLD. Clin. Gastroenterol. Hepatol. 21, 619–629e7 (2023). 10.1016/j.cgh.2021.11.038 [DOI] [PubMed] [Google Scholar]
  • 37.J. Delgado-Lista, J.F. Alcala-Diaz, J.D. Torres-Peña, G.M. Quintana-Navarro, F. Fuentes, A. Garcia-Rios et al., Long-term secondary prevention of cardiovascular disease with a Mediterranean diet and a low-fat diet (CORDIOPREV): a randomised controlled trial. Lancet. 399, 1876–1885 (2022). 10.1016/S0140-6736(22)00122-2 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (23.9KB, docx)

Data Availability Statement

The data involved in this study are available on the Global Health Data Exchange GBD 2021 website (https://ghdx.healthdata.org/gbd-2021).

Articles from Endocrine are provided here courtesy of Springer

RESOURCES