This retrospective analysis of PAD incidence, mortality and DALYs made use of data collected for the GBD study. Estimates for these factors were extracted from online data repositories; data from 1990 to 2021 were analyzed. These methods have been published previously (7). The incidence data in the GBD database are not directly collected but estimated through the integration of multiple data sources and the application of specific models and methods. The data sources of the GBD database include death registration, disease surveillance, population censuses, household surveys, clinical records, health service data, etc., and also involve innovative data sources such as satellite remote sensing and social media. The research team uses Bayesian meta - regression tools like DisMod - MR to conduct comprehensive analysis and modeling of these data, thereby estimating the incidence rates of different diseases in different regions, among different age groups, genders, and at different time points. The GBD database indeed provides incidence and prevalence data at different time points. It covers the time span from 1990 to the present. Users can select specific parameters such as time points, geographical scope, age, and gender according to their needs through platforms like GBD Compare to query and download the corresponding incidence and prevalence data. The GBD database does not directly collect data on DALYs. Instead, it integrates multiple data sources and applies specific calculation methods to estimate DALYs. The details are as follows: Data sources: The GBD study utilizes a vast array of diverse data sources, including vital registration systems, verbal autopsies, censuses, household surveys, disease—specific registries, and health service contact data. These data sources cover numerous countries and regions across the globe. Calculation method: DALYs consists of two components: Years of Life Lost (YLL) due to premature death and Years Lost due to Disability (YLD) caused by disease—related disabilities. That is, DALYs = YLL + YLD. YLL is calculated by multiplying the number of deaths by specific cause—age—sex—location—year by the standard life expectancy at the age of death. YLD is obtained by multiplying the prevalence of sequelae by specific cause—age—sex—location—year by their respective disability weights. The disability weights, determined through expert reviews, are authoritative and range from 0 (perfect health) to 1 (death). Ethics approval and informed consent were not required for this study because of public accessibility to the data. All methods in this paper were performed following the relevant guidelines and regulations.

Our analysis made use of the GBD project's definition of PAD. Information regarding the three main clinical manifestations of PAD was collected from both males and females, across all age groups and regional groups covering 21 countries with similar epidemiological characteristics and in close geographical proximity. To differentiate in detail the burden of PAD as a function of age, patients were categorized in terms of age into subgroups of <70 years, 70–74 years, 75–79 years, and 80 + years. We also analyzed the disease burden of PAD patients aged over 70 and under 70. The socio-demographic index (SDI) is calculated by GBD 2021 for each country. This is a composite indicator comprising social and economic factors known to affect health outcomes in each place (8). The SDI is the normalized geometric mean of indices for the total fertility rate among people under the age of 25 years, the average years spent in education among people over 15 years of age, and the lagged per capita income, with 0 representing the combination of the highest fertility rate, the least time spent in education, and the lowest per capita income (9). Five quintiles are used to classify the SDI: low, low-middle, middle, high-middle, and high. Incidence, mortality, and DALYs are taken from GBD 2021, and each rate is shown per 100,000. The 2.5th and 97.5th values of the ordinal 1000 estimates calculated using the GBD are taken as the 95% uncertainty interval (UI) boundaries.

Age-standardized incidence rate (ASIR), age-standardized mortality rate (ASMR), and age-standardized disability-adjusted life years (ASDALYs) were calculated per 100,000 in terms of sex, geographic location, and SDI. The estimated annual percentage change (EAPC) of these metrics was calculated to quantify the evolution of early onset PAD burden over time. The natural logarithm of the rates was fitted via linear regression, denoted as y = α + βx + ε, where y = ln (ASR), and x = calendar year. The EAPC is then 100 × [exp(β)−1], and the linear regression model provided the corresponding 95% confidence interval (CI).

Furthermore, research indicates that the Bayesian age-period-cohort (BAPC) framework can enhance the precision of statistical predictions with respect to alternative methods including generalized additive, joinpoint, and Poisson regression analyses (10). Therefore, this methodology, integrated with nested Laplace approximation techniques, was used to project PAD-related ASIR, ASMR and ASDALYs for the years 2022–2035. The R software package (version 4.2.3) and JD_GBDR (V2.36.2, Jingding Medical Technology Co., Ltd.) were used to draw the figures.

Globally, the ASIR of PAD has declined from 1990 to 2021 (EAPC, −0.4; 95% CI, −0.4 to −0.4) (Table 1); in 1990 the ASIR per 100 000 was 130.3 (95% UI, 112.8–149.4), dropping to 115.4 (95% UI, 100.0–132.7) in 2021. Both the ASMR (EAPC, −1.7; 95% CI, −1.8 to −1.5) and ASDALYs (EAPC, −1.4; 95% CI, −1.5 to −1.3) exhibited a similar downward trend. The ASMR reduced from 1.3 (95% UI, 1.2–1.4) in 1990 to 0.8 (95% UI, 0.7–0.9) in 2021. The ASDALYs declined from 26.6 (95% UI, 22.3–33.8) in 1990 to 18.6 (95% UI, 15.2–24.2) in 2021.

The EAPC of ASMR map of 204 countries and regions shows that the three countries with the highest EAPC in 2021 were Georgia (16.7), Afghanistan (10.3) and Sudan (9.7) (Figure 1A). The corresponding countries in terms of the EAPC of ASDALYs (Figure 1B) and ASIR (Figure 1C) were Georgia (5.5), Bahrain (3.4) and Mongolia (3.1), and Lebanon (1), Afghanistan (0.9) and Egypt (0.8), respectively.

Overall, the ASIR, ASMR and ASDALYs for both males and females declined from 1990 to 2021 (Table 1). The EAPCs of ASIR were the same for males and females (EAPC, −0.4; 95% CI, −0.4 to −0.4), and the EAPCs of ASMR were similar for males (EAPC, −1.7; 95% CI, −1.8 to −1.5) and females (EAPC, −1.7; 95% CI, −1.9 to −1.5). Finally, the EAPC of ASDALYs in males was −1.5 (95% CI, −1.6 to −1.4) and that in females was −1.3 (95% CI, −1.4 to −1.2). In comparison, the ASIR (96.9; 95% UI, 84.0–111.4), ASMR (1.5; 95% UI, 1.4–1.7) and ASDALYs (29.5; 95% UI, 25.5–34.6) of males in 1990 were higher than those in 2021 (86.6; 95% UI, 75.4–99.6), (1.0; 95% UI, 0.9–1.1) and (20.0; 95% UI, 17.3–24.2), respectively. A similar trend was observed in females: ASIR (158.4; 95% UI, 137.6–182.1), ASMR (1.1; 95% UI, 1.0–1.3), and ASDALYs (24.0; 95% UI, 19.1–32.3) in 1990; (141.2; 95% UI, 122.0–162.2), (0.7; 95% UI, 0.6–0.8), and (17.2; 95% UI, 13.0–24.3) in 2021, respectively.

The SDI in terms of ASIR, ASMR, and ASDALYs has evolved over the past three decades (Table 1). Among these factors, the ASIR (EAPC, 0.2; 95% CI, 0.2–0.2), ASMR (EAPC of low SDI, 1.3; 95% CI, 1.1–1.5; EAPC of middle-low SDI, 1.6; 95% CI, 1.5–1.7) and ASDALYs (EAPC of low SDI, 0.8; 95% CI, 0.7–0.9; EAPC of middle-low SDI, 0.7; 95% CI, 0.7–0.9) in low and middle-low SDI areas all exhibited an upward trend. In contrast, the incidence, mortality and DALYs in high, high-middle and middle SDI areas declined. In both 1990 and 2021, it was observed that the SDI was positively correlated with the incidence rate, mortality rate and DALYs, excluding low SDI areas.

Analysis of the trends of ASIR, ASMR, and ASDALYs in 21 regions globally revealed a correlation between socio-economic development and the rates of ASIR, ASMR, and ASDALYs. For instance, North America had the highest ASIR (229.8; 95% UI, 199.7–265.0; 207.2; 95% UI, 184.8–231.5.0) in both 1990 and 2021. Fortunately, over the past 30 years, these overall metrics in economically and socially developed regions have declined, including the high-income Asia Pacific region (EAPC of ASIR, −1.5; 95% CI, −1.5 to −1.4), Australasia (EAPC of ASIR, −1.2; 95% CI, −1.3 to −1.1), and Western Europe (EAPC of ASIR, −1.1; 95% CI, −1.1 to −1.0) etc.

The incidence, mortality rate and DALYs of PAD in the global population aged under 70 years old were all less than 10% across the time period studied (Figure 2). No region-related differences were found in terms of age across the 21 GBD regions. Among the three groups of people aged 70–74, 75–79 and 80+, the incidence rates were almost equal (Figure 2A), with people aged 80+ accounting for over 50% of PAD deaths (Figure 2B). The DALYs were positively correlated with the age of the PAD population (Figure 2C).

Given the apparent significant epidemiological differences between PAD patients aged <70 years and those aged ≥70 years, an analysis was conducted of the temporal trends of the incidence, mortality, and DALYs across these two groups (Figure 3). All of these metrics exhibited an upward trend for under 70s, while the opposite was true for people over 70. In other words, PAD was increasingly affecting younger people.

Smoking was the greatest risk factor for DALYs of PAD, both across the globe and across all levels of SDI (Figure 4). The next most significant risk factors globally were diets consisting of high proportions of processed meat and red meat, respectively. Similar results were observed in high and high-middle SDI regions, while in medium SDI regions, the second and third most significant risk factors were diets low in whole grains and high in red meat. Finally, in low-middle and low SDI areas, these factors were a diet low in whole grains, and lead exposure.

The temporal trends of risk factors between 1990 and 2021 were also analyzed. Smoking remains the biggest risk factor globally, and for high, high-middle and middle SDI areas, although it has shown a downward trend (Figure 5). However, in the low-middle and low SDI areas, the trend of this risk factor fluctuated but generally showed an upward trend.

The ASIR, ASMR and ASDALYs of PAD are predicted to show a downward trend (Figure 6). Specifically, the ASIR is expected to fall from 115.2 (95% CI, 114.8–116.9) in 2022 to 110.5 (95% CI, 107.0–114.1) in 2035; ASMR from 0.8 (95% CI, 0.8–0.9) in 2022 to 0.7 (95% CI, 0.6–0.8) in 2035; and ASDALYs from 18.5 (95% CI, 17.9–19.4) in 2022 to 16.5 (95% CI, 14.2–18.7) in 2035. In the low-middle and low SDI regions, both the incidence rate, mortality rate and DALYs showed an upward trend, while in the high, high-middle and middle SDI regions, they showed a downward trend (Supplementary Figure S1–S3).