This study analyzed data from the 2021 Global Burden of Disease (GBD) study, accessible via the Global Health Data Exchange (GHDx) at https://ghdx.healthdata.org. GBD 2021 provides the latest dataset on the global burden of ADOD across 204 countries and regions from 1990 to 2021. This cycle integrates new data sources and improved methods for more accurate estimates (8). Since 1990, GBD has consistently applied standard methodologies to estimate mortality and health losses for various diseases, using the Cause of Death Ensemble Model (CODEm) to estimate mortality data. This framework combines statistical models to validate cause-specific mortality (9), with results re-estimated every 2–3 years (10, 11). Mortality and disability-adjusted life years (DALYs) are derived from the average of 1,000 draws, with 95% uncertainty intervals (UI) calculated to estimate disease burden. UIs account for both parameter estimation variance and uncertainties in data collection and model selection.

In GBD 2021, the theoretical minimum risk exposure level for HFPG is 4.8–5.4 mmol/L (12). This is calculated as the person-year weighted average FPG level associated with the lowest mortality risk in meta-analyses of prospective cohort studies (13). According to the GBD framework, HFPG disease burden is only observed in individuals over 25.

For data analysis, the 2021 GBD study obtained the disease burden for 204 countries and regions, dividing the world into 21 regions based on epidemiological similarity and geographic proximity. Additionally, the Socio-demographic index (SDI), a composite indicator of social and economic conditions affecting health outcomes, is provided. Countries and regions are classified into SDI quintiles (high, high-middle, middle, low-middle, and low development levels). Age-standardized rates (ASR), mortality numbers and percentages, and DALYs were extracted from GBD 2021. EAPC of ASMR was calculated using a generalized linear model with Gaussian distribution to quantify the temporal trend of HFPG-induced ADOD. A higher EAPC indicates a heavier disease burden. If both the EAPC and its 95% confidence interval (CI) lower limit are greater than 0, the disease burden is increasing; otherwise, it indicates a declining trend. Decomposition analysis at the regional level assessed the contribution of population growth, aging, and epidemiological changes to ASDR in different SDI regions. We then used Spearman correlation analysis to examine the relationship between ASDR and SDI. Frontier analysis identified leading countries or regions (at the frontier and driving progress) with the lowest HFPG-induced ADOD burden at their SDI level. The distance from the frontier, termed “efficiency gap,” represents the gap between observed burden and potential achievable burden under its SDI; this gap may be reduced or eliminated based on the country’s or region’s socio-demographic resources. Development of Prediction Model The BAPC model assumes similar effects of age, period, and cohort in adjacent time intervals. In the BAPC model, all unknown parameters are considered random with appropriate prior distributions. The excellent predictive performance of the BAPC model has been previously validated (14, 15). The BAPC method allows us to review historical data and provides forward-looking insights for predicting future trends, aiding in public health strategy formulation. This study uses BAPC to predict HFPG-induced ADOD ASMR through 2050. HFPG-ADOD Platform Development To comprehensively assess the disease burden and BAPC predictions for 204 countries and territories, this study developed a free platform using the Shiny package (16).

Data analysis was conducted using R software version 4.3.2. R packages used in this study include “BAPC” (17), “tidyverse” (18), “factoextra” (19), and “ggplot2” (20). Statistical significance was considered at p < 0.05*.

In 2021, approximately 15% of all ADOD DALYs were due to HFPG. The number of DALYs increased by 271.05%, from 1.44 million in 1990 to 5.35 million in 2021, doubling over the period. The incidence rate per 100,000 rose from 47.07 (2.72, 126.46) to 66.42 (3.83, 178.85), indicating ineffective control over the past three decades. The estimated annual percentage change (EAPC) was 1.175 (1.087, 1.264), reflecting a continued global increase in HFPG-ADOD DALYs, consistent with previous findings (Table 1). Similarly, nearly 15% of all ADOD-related deaths globally were attributed to HFPG, up from 10% in 1990. Deaths due to HFPG-ADOD increased by 305.81% over the past 30 years, with an EAPC of 1.198 (1.103, 1.293), indicating a significant global rise in mortality (Table 2).

The DALYs and mortality burden of HFPG-ADOD have generally increased globally, with regional differences. By SDI classification, both High SDI [DALYs EAPC: 1.175 (1.087, 1.264), ASMR EAPC: 1.198 (1.103, 1.293)] and Low SDI [DALYs EAPC: 1.197 (1.16, 1.234), ASMR EAPC: 1.335 (1.275, 1.396)] regions showed similar trends (Tables 1, 2). Further gender-based analysis revealed consistent ASMR EAPC trends across genders, with regional variations such as in North Africa (Figure 1), warranting further stratified analysis.

Over the past 30 years, the DALYs and mortality burden for both male and female patients have increased. Notably, ASDR and ASMR were similar across genders (Figure 2A). Given the global variability in HFPG-ADOD ASDR and ASMR burdens, we identified regions and genders with the highest growth and best mitigation. For ASDR growth, Greenland’s female population observed the highest increase [EAPC: 3.66 (3.14, 4.18)], followed by females in Georgia [EAPC: 3.52 (3.21, 3.83)], and males in Egypt [EAPC: 3.44 (3.16, 3.72)] and females in Uzbekistan [EAPC: 3.43 (3.19, 3.68)], with top regions having a 2021 SDI of 0.40–0.88 (Figure 2B). Some regions achieved negative ASDR growth, including females in Cyprus [EAPC: −0.52 (−0.56, −0.48)], males in Mexico [EAPC: 0.26 (−0.3, −0.23)], and males in Cyprus [EAPC: −0.2 (−0.29, −0.12); (Figure 2C], with top regions having a 2021 SDI of 0.62–0.91, higher than regions with ASDR growth. In ASMR growth rankings, Greenland’s female population observed the highest increase [EAPC: 3.90 (3.32, 4.47)], followed by females in Georgia [EAPC: 3.61 (3.28, 3.94)] and males in Egypt [EAPC: 3.45 (3.16, 3.74)], consistent with ASDR rankings. Top regions had a 2021 SDI of 0.47–0.88 (Figure 2D). Some regions achieved negative ASDR growth, including females in Cyprus [EAPC: −0.68 (−0.75, −0.61)], males in Mexico [EAPC: 0.29 (−0.33, −0.25)], and males in Cyprus [EAPC: −0.22 (−0.33, −0.12); Figure 2E], with top regions having a 2021 SDI of 0.63–0.91, higher than regions with ASDR growth, consistent with ASDR trends.

In this study, the DALYs burden for HFPG-ADOD in 1990 was primarily in the 80–84 age group for both males and females. By 2021, this age group remained the main burden, with a decrease on either side (Figure 3A). The mortality burden shifted to the 85–89 age group, older than the 80–84 group in 1990, with significant increases in DALYs and deaths (Figure 3A). Age-adjusted ASDR and ASMR also showed a rising trend (Figure 3B). We further analyzed mortality burden by SDI region. Over the past 30 years, ASDR increased in High, High-middle, Middle, and Low-middle SDI regions, except Low SDI (Figure 3C). ASMR trends rose in High-middle, Middle, Low-middle, and Low SDI regions, except High SDI (Figure 3D). Age, Period, and Cohort Effects on ASDR The APC model estimated the effects of age, period, and cohort by SDI quintiles, evaluating their contribution to ASMR in different SDI regions. Age effects are shown by longitudinal age curves, describing ASMR’s natural history. Period effects indicate changing relative risks over time, tracking progress. Cohort effects show relative risk changes in birth cohorts. Overall, age, period, and cohort effects showed consistent trends globally and across SDI regions. The period-cohort model indicated a global similarity in birth cohort effects, generally declining (Figure 4A). Period effects were more pronounced, with ASDR gradually decreasing over 30 years without clear shifts (Figure 4A). The age-period cohort model (Figure 4B) confirmed that higher age correlates with higher ASDR, as shown in the age-period model (Figure 4C), highlighting age as a key factor in HFPG-induced ADOD mortality burden. Notably, in the past 5 years, ASDR showed the highest ASMR globally and across SDI regions (Figure 4C), suggesting recent medical advancements have not been effectively realized.

To further validate the correlation between ASMR and SDI, Spearman correlation analysis showed a positive relationship globally (cor = −0.142, p = 0.000 < 0.001; Figure 5A). Using a Joinpoint AAPC model, ASMR has shown a continuous upward trend over the past 30 years, consistent across both male (Figure 5B) and female (Figure 5C) populations.

Overall, as SDI increases, the efficiency gap tends to widen, indicating less effective ASMR control in high SDI regions (Figure 6A). The top five countries with the largest efficiency gaps compared to the frontier include Morocco, Qatar, Afghanistan, Tokelau, and the Marshall Islands (range: 5.21–7.40), showing significantly higher ASMR at similar SDI levels. Conversely, the top five countries with the lowest ASDR and smallest efficiency gaps at similar SDI levels (range: 0.42–0.62) are Madagascar, Belarus, Togo, Mongolia, and Somalia (Figure 6B).

Globally, population growth is the main contributor to ASMR, followed by epidemiological changes and aging. These factors show regional differences: in Low and Low-middle SDI areas, population is the largest contributor; starting from Middle SDI, aging becomes more significant. In High SDI regions, aging’s impact is close to population, with epidemiological changes surpassing population growth (Figure 7).

Finally, using BAPC, we predicted global HFPG-ADOD ASMR from 2022 to 2050 and conducted gender subgroup analysis. By 2050, the absolute number of cases is projected to rise to 1,003,018.047 (26,124.40377, 12,114,480.49), with males at 345,342.5738 (1,431.57781, 6,022,119.213) and females at 657,675.4731 (24,692.82596, 6,092,361.28). ASMR is projected to rise to 4.40 (0.09, 84.53), with a gradual increase, while male ASMR is 4.53 (0.94, 8.14), showing a moderate trend. It indicates that females will contribute most to the global ASDR increase over the next 30 years (Figure 8).

Using the Shiny platform, we developed the HFPG-ADOD DATABASE to visualize disease burden and predictions for 204 countries and regions, accessible at http://116.196.73.86:3838/GBD/HFPG-ADOD/. For example, in China, EAPC analysis shows an annual increase in ASDR [EAPC: 0.57 (0.41, 0.73); (Figure 9A], and ASMR also shows an increasing trend [EAPC: 0.37 (0.21, 0.53); Figure 9B]. Based on BAPC predictions, by 2050, HFPG-ADOD deaths will rise from 13,039.81 (5,592.12, 109,990.79) to 224,964.46 (0, 2,257,566.33), but ASMR is expected to decrease [ASMR: 3.96 (0, 82.31); Figure 9C].