Analysis of the disease burden and attributable risk factors of tuberculosis in G20 countries from 1990 to 2021

Abstract

Objective

Tuberculosis (TB) remains a leading global cause of death attributable to a single infectious agent. However, studies examining the TB burden in The Group of Twenty(G20) countries are limited. This study sought to analyze temporal trends in TB burden and identify key risk factors across G20 nations from 1990 to 2021, while projecting the risk factors-attributable disease burden through 2035 using advanced modeling techniques.

Methods

Data were obtained from the Global Burden of Disease (GBD) 2021 database. Disease burden was assessed using the disability-adjusted life years (DALYs) and the age-standardized DALYs rates (ASDR), characterizing temporal trends, regional variations, age-specific patterns, and sex disparities in TB burden across G20 countries. Joinpoint regression analysis identified periods with significant temporal changes in country-specific TB ASDR attributable to different risk factors (1990–2021). Health inequality analysis was performed to assess inequalities in TB burden attributable to risk factors relative to the socio-demographic index (SDI). Decomposition analysis was performed to investigate the drivers of changes in TB burden across G20 countries. Additionally, trends in the population attributable fractions (PAF) and summary exposure values (SEV) were analyzed for each risk factor. Finally, Bayesian age-period-cohort (BAPC) modeling was used to project DALYs and ASDR for TB related to six risk factors across the G20 countries from 2022 to 2035.

Results

From 1990 to 2021, overall TB DALYs in G20 countries decreased by 50% (95% UI: 36% – 56%). In 2021, India recorded the highest DALYs, followed by Indonesia, while South Africa exhibited the highest ASDR, with Indonesia and India ranking second and third, respectively. Smoking constituted the leading risk factor for TB ASDR in 2021, followed by high alcohol use. Over the past three decades, the PAF of TB DALYs attributable to each risk factor varied significantly across age and gender groups. Health inequality analysis revealed narrowed absolute disparities but widespread exacerbation of relative inequalities in TB burden related to risk factrs across G20 nations. Decomposition analysis demonstrated divergent proportional contributions of drivers–aging, population growth, and epidemiological changes–to the risk-attributable TB burden across G20 countries. BAPC model projections indicated persistently high TB burdens in India, Indonesia, China, and South Africa throgh 2035. Trends in TB DALYs and ASDR attributable to risk factors across G20 countries were heterogeneous: while most nations showed declining smoking-attributable burdens relative to 2021 levels, other risk factors contributed to increased burdens to varying degrees in numerous countries.

Conclusion

Despite declining TB burden across G20 nations, substantial heterogeneity persists. Smoking and high alcohol use remained the dominant risk factors contributing to TB burden, while comorbidities like diabetes and obesity warrant continued focus. The implementation of interventions specifically targeting these risk factors, combined with enhanced collaborative frameworks within the G20, is essential to further mitigate TB burden and disparities.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-025-11843-0.

Introduction

Tuberculosis(TB), a longstanding infectious disease caused by Mycobacterium tuberculosis, has remained a persistent public health concern for thousands of years [1]. It can infect multiple organs in the body but primarily affects the lungs. Approximately one-quarter of the global population is estimated to be infected with M. tuberculosis [2], with 90% – 95% of infected individuals remaining asymptomatic and non-infectious [3]. However, approximately 5%–10% of infected individuals develop symptoms and progress to active TB [1]. In 2023 [4], an estimated 10,800,000 people, 95% uncertainty interval (UI: 10,100,000–11,700,000) globally were diagnosed with TB, affecting all countries and age groups. Additionally, 1.25 million deaths, including those of HIV-positive individuals, were attributed to TB in 2023, making it the leading cause of death globally from a single infectious agent. TB is a critical global public health issue that requires urgent attention. The 2024 Global Tuberculosis Report indicated a 50% reduction in the TB incidence rate compared to 2015, and TB-related deaths could be reduced by 75% by 2025 [4]. A Global Burden of Disease (GBD) study projection aligns with this trend, indicating that global TB deaths could drop from the 14th to the 25th leading cause of death between 2022 and 2050 [5]. The reduction in TB-related deaths has enhanced human life expectancy. Given that TB is preventable, controllable, and curable, its disease burden exhibited a positive trend toward improvement. The United Nations and the World Health Organization(WHO) have established a sustainable development goal to end the TB epidemic by 2030. Despite a decline in TB-related deaths, disability-adjusted life years (DALYs), age-standardized mortality rates, and age-standardized disability rates have fallen short of the 2020 milestones [6]. Therefore, global collaboration is essential to strengthen efforts to control the TB epidemic.

The Group of Twenty (G20) is an international economic cooperation organization comprising developed and developing countries, representing the world’s major economies and holding significant influence over global economic and health policies [7]. Most annual TB cases occur in 30 high-burden countries, the top three being India, Indonesia, and China, all G20 members [4]. G20 member countries have distinct internal challenges in combating TB. For instance, HIV infection, diabetes, and smoking are among the leading risk factors [4], while in Eastern European countries, alcohol abuse, which impairs immune function, is closely associated with the development and progression of TB [8]. The incidence and mortality rates of TB vary significantly across countries. In 2019, India reported over 420,000 TB-related deaths [9, 10], while some European Union countries recorded fewer than ten new TB cases, with some even reporting zero TB deaths [11]. A negative correlation exists between the sociodemographic index (SDI) and TB burden, implying that higher socioeconomic levels are associated with a lower TB burden, while poverty exacerbates it [12]. Extensive study has demonstrated that the spatiotemporal distribution of TB exhibits complex dynamic features, with specific patterns observed at various scales [13]. Environmental factors also affect TB incidence in region-specific ways. In Brazil, climatic factors and air quality are closely related to TB incidence [14]. Although poor air quality is not a direct risk factor for TB, it may indirectly exacerbate its disease burden. Furthermore, the global proliferation of multidrug-resistant TB (MDR-TB) poses a substantial challenge to TB control initiatives. Identifying and addressing the risk factors associated with MDR-TB will mitigate the emergence of drug resistance [15], advancing the objective of eradicating the TB epidemic by 2030.

Despite ongoing global initiatives to monitor TB burden and associated risk factors, systematic analyses of major risk factors attributable to TB burden in G20 countries–and their temporal trends based on GBD data–remain limited. We extracted GBD 2021 data on TB burden and risk factor-attributable burdens for G20 countries. Joinpoint regression analysis was applied to characterize temporal trends in risk factors-related TB burden across these nations. SDI-based health inequities in risk factor-attributable TB burden were assessed. Decomposition analysis was performed to identify drivers of changes in risk factor-specific TB burden between 1990 and 2021. Finally, Bayesian age-period-cohort(BAPC) modeling projected country-level risk factor-specific TB burdens from 2022 to 2035.

Methods

Data sources

Our research data is sourced from the GBD 2021 database of the Institute for Health Metrics and Evaluation (IHME)(http://ghdx.healthdata.org/). The GBD 2021 database offers estimates of counts and age-standardized rates(ASR) for 371 diseases and injuries across 204 countries and territories globally. It includes core burden metrics: years lived with disability (YLDs), years of life lost (YLLs), disability-adjusted life years (DALYs; calculated as YLL + YLD), and healthy life expectancy [16]. Additionally, it provides epidemiological estimates for 88 risk factors and their associated health outcomes [17]. The GBD 2021 study implemented a rigorous data processing pipeline to correct for systematic biases, with a particular focus on addressing inconsistencies arising from heterogeneous data sources, case definitions, and measurement methods. This research employed advanced statistical modeling techniques to construct an analytical framework. Specifically, meta-regression Bayesian regularized trimming was used for data integration and bias adjustment, while the DisMod-MR 2.1 disease modeling software ensured comparability across three key dimensions: spatial distribution, population characteristics, and temporal trends. Through the application of iterative standardization and empirical calibration algorithms, the study significantly reduced measurement heterogeneity, thereby enhancing the reliability of comparative risk assessment results. Using the GBD 2021 database, we extracted TB DALYs and risk factor exposure levels for all G20 members: Argentina, Australia, Brazil, Canada, China, France, Germany, India, Indonesia, Italy, Japan, Mexico, Russia, Saudi Arabia, South Africa, South Korea, Turkey, the United Kingdom, the United States, and the European Union. Temporal trends from 1990 to 2021 were subsequently analyzed [18]. GBD exclusion criteria comprised country-year data flagged as unreliable in GBD metadata, subnational data, model estimates with an uncertainty interval width exceeding 100% of the point estimate, country-year estimates derived from fewer than three underlying data sources, latent tuberculosis infection, and deaths where tuberculosis was not the underlying cause. This study is a secondary analysis based on GBD 2021 database and does not include any original patient data, thus exempting it from the requirement for ethical approval. During the research process, the team carefully adhered to data use agreements, ensuring the anonymity and confidentiality of the data.

Definition of TB

The GBD 2021 database defines cases with ICD-10 codes A10-A14, A15-A18.89, A19-A19.9, B90-B90.9, K67.3, K93.0, M49.0, N74.0-N74.1, P37.0, and U84.3 as tuberculosis with no age restrictions.

Population attributable fraction (PAF) of DALYs

The GBD 2021 uses PAF to quantify the impact of risk factors on the burden of TB. The PAF represents the proportion by which the incidence of a target disease in a population would decrease if a specific risk factor were completely eliminated. The 2021 GBD study calculated the PAF by integrating risk functions and exposure distributions for each individual based on age, sex, location, and year. This was accomplished by combining risk exposure levels, meta-analysis-based relative risk estimates, and the theoretical minimum risk exposure level, facilitating the independent assessment of each risk factor’s contribution [19].

Risk factors and SEV

GBD 2021 identified six third-level attributable risk factors for TB, ranked by PAF: smoking, high alcohol use, high fasting plasma glucose, high body-mass index (BMI), dietary risks, and low physical activity. These risk factors were selected according to robust causal evidence, data availability, behavioral modifiability, and public health policy relevance [20]. Notably, the “tobacco” category (a third-level category encompassing only “smoking”) is referred to as “smoking” throughout this study; all diet-related risks are grouped under “dietary risks.”

Furthermore, percentage changes in the SEV were employed to describe changes in exposure levels to individual risk factors across G20 countries between 1990 and 2021. The GBD 2021 risk factor analysis evaluated 88 risk factors across countries and regions globally and calculated the SEV. The SEV quantifies the risk-weighted prevalence of exposure by combining the excess risk associated with exposure with the population distribution of exposure levels. This combined metric is then contrasted against a counterfactual scenario where all individuals experience the theoretical maximum risk level. The SEV thus represents the total exposure level for each risk factor, providing a fundamental metric for understanding the distribution of risk factors and their impact on disease burden. Higher SEV values indicate a greater risk-weighted prevalence of exposure for a given risk factor [21].

Statistical analysis

This study comprehensively analyzed the TB disease burden in G20 countries from 1990 to 2021 using multiple methodologies. The burden of TB and the burden attributable to specific risk factors were quantified using both absolute case numbers and ASDR. Uncertainty in these estimates was expressed using 95% UI.Temporal trends were assessed using the EAPC. The EAPC was calculated by fitting a log-linear regression model: ln(rate) = β₀ + β₁ × year + ε, where EAPC = [exp(β₁) − 1] × 100%. Joinpoint regression, health inequality analysis, and decomposition analysis were employed to assess temporal trends and identify potential drivers of risk factors-related TB burden across G20 countries between 1990 and 2021. The BAPC model projected future burden from 2022 to 2035. A significant trend was defined by a 95% confidence interval(CI) excluding zero [11]. All statistical analyses and visualizations were performed using R software (version 4.3.3).

Decomposition analysis

We evaluated the relative contributions of changes in population growth, population age structure, and TB prevalence of risk factors-related TB burden [22]. We employed the method developed by Das Gupta [23], which algebraically separates the standardized impacts of each contributing multiplicative factor to encapsulate the contributions of various factors to the observed changes.

Joinpoint regression analysis

We performed Joinpoint regression analysis using the Joinpoint command-line version assessed through R functions. This analysis assessed and compared the annual percentage change (APC) and average annual percentage change (AAPC) of TB burden related to six risk factors across various countries and regions from 1990 to 2021. Primarily, we applied a log-linear model (ln y = β × x + constant) to perform segmented regression for the temporal trends of TB burden in G20 countries. Thereafter, the grid search method was employed to generate all possible grid points, from which the grid point with the lowest mean squared error was selected as the joinpoint. Ultimately, a Monte Carlo permutation test was employed to determine the optimal model for joinpoint regression, with the maximum number of potential joinpoints set to 5 and the minimum set to 0. APC and AAPC obtained from the optimal model developed through these steps were utilized to quantify the trends in disease burden rates from 1990 to 2021, assuring accurate and compelling research outcomes. The formula for APC is :

APC = (e^β−1) × 100%.

The AAPC, calculated as a weighted average of segment-specific APC values based on segment duration, represents the overall trend in TB burden from 1990 to 2021. Negative APC and AAPC values indicate a declining trend in TB burden over time, while positive values signify an increasing trend [24]. CI for all Joinpoint regression outputs–including APC, AAPC, and joinpoint locations–were estimated using the empirical quantile method. Detailed methodological specifications and parameter settings align with the WHO Global Tuberculosis Report technical annex (https://www.who.int/teams/global-tuberculosis-programme/data) and the IHME trend analysis protocols (https://www.healthdata.org/methods-library).

Health inequality analysis

To measure socioeconomic inequalities in TB burden attributable to risk factors across nations, we employed WHO-recommended methods, calculating the slope index of inequality (SII) and the Concentration Index [25]. SDI data for G20 countries were extracted from the GBD 2021 database to assess socioeconomic development levels across nations [26]. This index positively correlates with national/regional economic development levels [27]. The SII, defined as the slope of the regression line between TB ASDR and countries’ weighted rankings, quantifies absolute socioeconomic health inequality. The concentration index quantified relative cross-national TB burden inequality by fitting a Lorenz concentration curve to cumulative DALYs against cumulative population. This index represents twice the area between the concentration curve and the line of equality, ranging from − 1 to 1. A negative value indicates concentrated TB burden in low-SDI populations [28].

BAPC model

This study employed the BAPC package in R to construct BAPC models using the integrated nested Laplace approximation method, projecting risk factors-related TB burden in G20 countries from 2022 to 2035 [29]. The BAPC model is based on a generalized linear model and integrates the theoretical correlation between disease burden metrics, age distribution, and population size. For outcome predictions, we used the posterior median as the point estimate for counts, rates, and ASDR, and constructed 95% CI based on the 2.5th and 97.5th percentiles of the posterior distribution to comprehensively assess prediction uncertainty.This model enables robust predictive capabilities by integrating a Bayesian framework. It comprehensively accounts for multiple factors, providing a powerful tool for our predictive analysis.

Results

TB burden in G20 countries

From 1990 to 2021, the total DALYs and ASDR for TB in G20 countries indicated a steady decline. In 2021, the total DALYs for TB in G20 countries were 21,522.27 (95%UI: 19,202.2 − 2 5,525.71), representing a 50% (95%UI: 56% – 36%) reduction compared to 1990 (42,886.36, 95%UI: 38,378.59–46,853.04). Among G20 countries, India had the highest DALYs (15,057.8, 95%UI: 13,108.01–18,233), followed by Indonesia (3,162.14, 95%UI: 2,638.16–3,774.75), while Australia had the lowest (1.69, 95%UI: 1.52–1.85). Turkey, South Korea, and China experienced the most significant declines in DALYs from 1990 to 2021, whereas South Africa experienced an 11% reduction (Fig. 1). Regarding the age distribution of TB DALYs in G20 countries, the 35–64 age group accounted for the highest number of DALYs. From 1990 to 2021, the TB burden among children(0–9 age group) and adolescents(10–19 age group) decreased significantly in the G20 and most countries, while changes in other age groups were less pronounced. However, in 2021, South Africa had a disproportionately high TB burden among children under 9 years old. In addition, in Italy, Australia, France, Japan, and South Korea, the TB burden was relatively higher among individuals aged ≥ 70 (Table S1, Figure S1).

Fig. 1.

Trends in all-age number of TB DALYs in G20 countries from 1990 to 2021 (*Y-axis scales differ across panels). TB, tuberculosis; DALYs, disability-adjusted life-years; G20, Group of Twenty

Similarly, ASDR decreased from 1,222.93 (95%UI: 1,094.05–1,338.49) in 1990 to 387.41 (95%UI: 346.51–458.83) in 2021, with the EAPC is 3.9% (95%CI: 4.08% – 3.71%) In 2021, South Africa had the highest ASDR for TB (1,855.04, 95%UI: 1,661.27–2,115.3), followed by Indonesia (1,220.1, 95%UI: 1,026.87–1,443.58) and India (1,119.71, 95%UI: 977.01–1,355.66). Australia had the lowest ASDR (4.38, 95%UI: 3.9–4.9). Among these countries, Turkey experienced the fastest decline, with an EAPC of 8.72%(95%CI: −9.22% – −8.21%), while South Africa had the slowest decline, with an EAPC of 1.86% (95%CI: 2.74% – 0.96%) (Table 1, Figure S1). In 1990 and 2021, the ASDR for males was higher than that for females in all countries except Saudi Arabia.

Table 1.

Temporal changes in age-standardized TB DALY rates: overall and attributable to six risk factors across G20 countries from 1990 to 2021. TB, tuberculosis; DALYs, disability-adjusted life years; G20, group of Twenty

Location		Age − standardised rate of DALYs for TB
		Attributable risk factors
		Overall	High alcohol use	Dietary risks	High body-mass index	High fasting plasma glucose	Low physical activity	Smoking
G20	1990	1222.93(1094.05,1338.49)	108.75(−55.61,413.44)	11.64(2.35,21.47)	22.45(4.65,62.04)	62.33(44.57,83.18)	5.53(1.63,10.5)	203.52(160.39,251.22)
	2021	387.41(346.51,458.83)	43.92(−28.25,162.14)	4.94(1.09,8.97)	16.9(3.61,43.95)	33.19(23.81,44.74)	2.07(0.68,3.82)	53.13(40.56,69.45)
	EAPC(95%CI)	−3.9 (−4.08,−3.71)	−3.02 (−3.18,−2.85)	−2.98 (−3.15,−2.81)	−0.98 (−1.21,−0.75)	−2.22 (−2.38,−2.06)	−3.43 (−3.6,−3.25)	−4.4 (−4.59,−4.21)
Argentina	1990	196.17(186.81,205.79)	46.74(−58.92,142.19)	3.49(0.85,6.53)	15.12(2.6,41.92)	8.91(6.19,11.83)	0.73(0.23,1.53)	44.69(35.66,53.82)
	2021	52.52(48.91,56.32)	11.78(−12.77,38.38)	1.7(0.5,3.14)	7.24(1.52,17.6)	4.74(3.41,6.18)	0.33(0.1,0.62)	10.04(7.65,12.56)
	EAPC(95%CI)	−4.18 (−4.55,−3.8)	−4.29 (−4.67,−3.92)	−2.23 (−2.51,−1.94)	−2.3 (−2.53,−2.07)	−2.23 (−2.56,−1.9)	−2.57 (−2.82,−2.31)	−4.72 (−5.01,−4.44)
Australia	1990	13.96(12.93,15.25)	2.95(−4.09,9.67)	0.55(0.13,1.02)	1.66(0.35,3.94)	0.86(0.61,1.17)	0.21(0.06,0.4)	2.73(2.08,3.43)
	2021	4.38(3.9,4.9)	1.02(−1.29,3.25)	0.26(0.06,0.48)	0.82(0.21,1.69)	0.42(0.29,0.55)	0.09(0.03,0.17)	0.53(0.39,0.68)
	EAPC(95%CI)	−3.91 (−4.23,−3.58)	−3.54 (−3.82,−3.26)	−2.51 (−2.85,−2.17)	−2.5 (−2.79,−2.2)	−2.59 (−2.92,−2.25)	−2.8 (−3.1,−2.5)	−5.37 (−5.71,−5.04)
Brazil	1990	328.64(317.46,341.19)	52.68(−45.67,184.31)	3.86(0.87,7.18)	18.46(3.47,53.55)	19.04(13.86,25.37)	1.86(0.6,3.59)	76.41(61.32,91.25)
	2021	90.42(85.49,95.28)	18.25(−16.85,62.4)	2.42(0.63,4.37)	11.44(2.39,27.36)	7.99(5.89,10.22)	0.93(0.28,1.72)	12.9(9.67,16.42)
	EAPC(95%CI)	−4.51 (−4.7,−4.32)	−3.82 (−4,−3.65)	−1.65 (−1.78,−1.53)	−1.9 (−2.04,−1.76)	−2.97 (−3.12,−2.83)	−2.38 (−2.49,−2.28)	−6.29 (−6.53,−6.05)
Canada	1990	19.93(18.7,21.23)	3.82(−4.05,12.95)	0.3(0.07,0.57)	1.62(0.24,5.13)	1(0.7,1.35)	0.07(0.02,0.14)	5.7(4.5,6.89)
	2021	4.72(4.25,5.22)	1.05(−1.21,3.44)	0.1(0.02,0.18)	0.48(0.09,1.31)	0.55(0.4,0.72)	0.02(0.01,0.04)	0.83(0.61,1.05)
	EAPC(95%CI)	−5.01 (−5.49,−4.52)	−4.38 (−4.85,−3.91)	−4.16 (−4.55,−3.77)	−4.42 (−4.82,−4.02)	−2.35 (−2.92,−1.78)	−4.54 (−4.9,−4.19)	−6.62 (−7.18,−6.05)
China	1990	719.42(610.63,837.38)	89.11(−76.5,315.98)	7.82(1.53,14.58)	15.89(3.8,45.45)	44.33(30.58,61.21)	3.86(1.22,7.43)	158.48(114.49,203.94)
	2021	76.22(62.59,94.45)	13.55(−12.69,47.83)	1.45(0.33,2.76)	5.47(1.13,14.58)	6.74(4.63,9.45)	0.6(0.19,1.16)	19.15(13.85,26.7)
	EAPC(95%CI)	−7.49 (−7.7,−7.27)	−6.21 (−6.36,−6.06)	−5.44 (−5.61,−5.27)	−3.63 (−3.82,−3.44)	−5.87 (−6.16,−5.57)	−6.14 (−6.34,−5.93)	−6.84 (−6.98,−6.71)
European Union	1990	79.71(76.84,82.74)	19.95(−28.78,60.71)	1.8(0.43,3.24)	6.86(1.26,19.06)	5.07(3.66,6.62)	0.5(0.16,0.93)	25.69(21.02,30.03)
	2021	15.41(14.3,16.56)	3.98(−5.81,11.99)	0.57(0.15,1.06)	2.02(0.43,4.99)	1.51(1.11,1.93)	0.13(0.04,0.24)	3.91(3.1,4.68)
	EAPC(95%CI)	−5.74 (−6.02,−5.46)	−5.66 (−5.95,−5.36)	−3.88 (−4.11,−3.64)	−4.44 (−4.78,−4.1)	−4.27 (−4.49,−4.05)	−4.7 (−4.93,−4.46)	−6.58 (−6.91,−6.25)
France	1990	64.98(60.54,68.93)	16.86(−25.11,51.28)	1.84(0.41,3.46)	4.73(0.91,12.97)	3.17(2.22,4.42)	0.58(0.17,1.13)	15.62(12.29,19.26)
	2021	9.87(8.93,10.78)	2.43(−3.09,7.57)	0.36(0.08,0.65)	1.11(0.21,2.87)	0.7(0.48,0.97)	0.1(0.03,0.2)	1.76(1.34,2.18)
	EAPC(95%CI)	−6.71 (−6.99,−6.44)	−6.85 (−7.11,−6.58)	−5.87 (−6.13,−5.61)	−5.4 (−5.7,−5.1)	−5.72 (−6.1,−5.35)	−6.35 (−6.69,−6.02)	−7.49 (−7.74,−7.24)
Germany	1990	41.81(39.14,44.27)	11.68(−18.25,34.87)	1.03(0.25,1.86)	3.06(0.56,9.44)	4.06(2.74,5.52)	0.23(0.07,0.43)	12.91(10.22,15.62)
	2021	6.59(5.9,7.4)	1.76(−2.51,5.27)	0.19(0.04,0.34)	0.51(0.1,1.48)	0.95(0.68,1.23)	0.04(0.01,0.07)	1.35(1.05,1.67)
	EAPC(95%CI)	−6.02 (−6.55,−5.49)	−6.14 (−6.68,−5.59)	−5.45 (−5.85,−5.04)	−5.9 (−6.4,−5.4)	−4.88 (−5.38,−4.39)	−6.11 (−6.59,−5.63)	−7.25 (−7.85,−6.65)
India	1990	3910.27(3490.87,4366.49)	328.82(−77.6,1344.27)	43(9.22,80.66)	60.09(12.63,160.54)	266.43(188.2,356.05)	7.01(2.31,13.41)	646.03(498.62,814.53)
	2021	1119.71(977.01,1355.66)	131.94(−71.43,503.98)	16.49(3.31,30.06)	45.11(10,111.67)	129.11(92.05,178.66)	22.19(6.49,43.5)	145.39(106.96,199.62)
	EAPC(95%CI)	−4.22 (−4.39,−4.06)	−2.9 (−3.02,−2.78)	−3.23 (−3.39,−3.07)	−0.93 (−1.14,−0.71)	−2.65 (−2.83,−2.47)	−4.03 (−4.25,−3.82)	−4.87 (−5.04,−4.69)
Indonesia	1990	3709.07(2833.4,4285.24)	83.24(−25.35,441.18)	20.09(3.57,38.14)	41.21(8.67,125.46)	166.11(103.94,228.99)	19.98(5.67,39.05)	563.92(351.1,757.04)
	2021	1220.1(1026.87,1443.58)	27.48(−12.28,141.75)	8.9(1.67,16.85)	36.27(7.52,105.68)	132.84(90.22,181.17)	9.24(2.86,17.13)	230.27(168.08,308.09)
	EAPC(95%CI)	−3.45 (−3.68,−3.21)	−3.83 (−4,−3.66)	−2.67 (−2.79,−2.54)	−0.16 (−0.5,0.19)	−0.66 (−0.91,−0.41)	−2.47 (−2.63,−2.3)	−2.72 (−3,−2.43)
Italy	1990	28.53(26.89,30.25)	7.45(−10.77,22.61)	0.77(0.19,1.38)	2.2(0.44,6.32)	2.14(1.56,2.81)	0.29(0.09,0.56)	8.36(6.65,9.97)
	2021	5.45(4.92,6.02)	1.25(−1.58,4.01)	0.17(0.04,0.32)	0.54(0.12,1.42)	0.52(0.38,0.67)	0.06(0.02,0.11)	0.94(0.73,1.16)
	EAPC(95%CI)	−5.44 (−5.76,−5.12)	−5.86 (−6.19,−5.52)	−4.98 (−5.23,−4.73)	−4.75 (−5.04,−4.46)	−4.92 (−5.18,−4.65)	−5.35 (−5.61,−5.1)	−6.98 (−7.37,−6.6)
Japan	1990	73.88(70.03,77.54)	17.65(−20.98,57.62)	3.1(0.73,5.41)	5(1.3,12.38)	8.95(6.65,11.55)	0.28(0.09,0.53)	23.41(19.06,27.54)
	2021	12.31(10.83,13.43)	2.51(−2.57,8.61)	0.53(0.13,0.92)	1.03(0.27,2.46)	1.85(1.34,2.47)	1.54(0.49,2.8)	1.87(1.45,2.37)
	EAPC(95%CI)	−6.36 (−6.6,−6.11)	−6.97 (−7.25,−6.69)	−6.39 (−6.77,−6.01)	−5.82 (−6.11,−5.54)	−5.94 (−6.28,−5.6)	−6.2 (−6.5,−5.9)	−8.85 (−9.18,−8.52)
Mexico	1990	465.02(452.6,478.48)	75.45(−71.92,252.88)	12.07(3.2,21.13)	50.87(12.25,122.82)	64.07(48.71,80.23)	3.7(1.13,6.83)	74.2(57.94,89.98)
	2021	74.22(65.95,83.6)	14.86(−14.51,48.36)	2.28(0.63,4.24)	12.09(3.09,25.93)	9.64(7.18,12.4)	0.61(0.18,1.16)	6.51(4.78,8.33)
	EAPC(95%CI)	−5.96 (−6.46,−5.45)	−5.27 (−5.83,−4.71)	−5.4 (−5.85,−4.94)	−4.71 (−5.18,−4.24)	−6.31 (−6.72,−5.91)	−5.91 (−6.49,−5.32)	−7.95 (−8.44,−7.45)
Republic of Korea	1990	783.08(690.59,919.15)	178.78(−177.75,594.28)	13.32(2.9,23.67)	31.97(7.29,82.4)	66.64(44.3,95.92)	8.65(2.5,16.53)	240.42(189.01,319.04)
	2021	58.02(50.24,67.53)	12.6(−11.02,41.79)	1.87(0.45,3.34)	4.75(1.16,11.43)	8.43(6,11.72)	1.03(0.33,1.93)	11.83(8.77,15.56)
	EAPC(95%CI)	−8.28 (−8.52,−8.04)	−8.5 (−8.73,−8.27)	−6.25 (−6.53,−5.97)	−6.13 (−6.51,−5.75)	−6.59 (−6.79,−6.39)	−6.79 (−7.18,−6.39)	−9.72 (−9.94,−9.5)
Russian Federation	1990	274.1(263,286.14)	69.46(−92.43,209.21)	8.72(1.55,17.45)	13.34(2.05,42.02)	8.13(5.7,10.95)	0.52(0.16,1.01)	89.9(74.16,104.56)
	2021	128.7(118.65,138.75)	33.27(−46.86,101.55)	2.46(0.55,4.48)	12.41(2.13,33.79)	6.22(4.4,8.32)	0.35(0.11,0.69)	44.71(36.28,53.53)
	EAPC(95%CI)	−3.15 (−4.7,−1.57)	−3.23 (−4.92,−1.51)	−5.52 (−7.02,−3.99)	−0.8 (−2.31,0.73)	−1.46 (−2.82,−0.09)	−1.78 (−2.97,−0.57)	−3.06 (−4.74,−1.36)
Saudi Arabia	1990	794.48(551.98,1053.6)	9.65(−2.96,56.46)	10.57(2.55,19.01)	76.13(14.35,208.16)	97.62(60.35,143.35)	7.39(2.16,13.93)	73.56(42.15,119.22)
	2021	152.4(116.83,222.38)	1.42(−0.54,9.4)	4.93(1.41,9.54)	32.31(8,69.85)	28.3(18.92,43.55)	2.7(0.76,5.33)	18.14(11.16,31.4)
	EAPC(95%CI)	−5.25 (−5.37,−5.13)	−4.66 (−5.25,−4.07)	−2.89 (−3.13,−2.64)	−3.05 (−3.22,−2.88)	−4.2 (−4.43,−3.97)	−3.66 (−3.88,−3.43)	−4.33 (−4.52,−4.14)
South Africa	1990	3420.54(2935.5,4192.02)	651.17(−393.09,2164.99)	19.69(4.17,36.79)	142.43(20.7,446.59)	118.36(80.23,164.54)	9.44(3.04,17.16)	659.04(484.55,858.72)
	2021	1855.04(1661.27,2115.3)	358.49(−227.57,1168.71)	19.03(4.28,35.03)	148.59(26.84,419.51)	156.61(112.55,208.24)	10.93(3.36,21.74)	212.92(155.94,273.88)
	EAPC(95%CI)	−1.86 (−2.74,−0.96)	−1.94 (−2.79,−1.08)	0.28 (−0.63,1.2)	0.24 (−0.74,1.24)	1.38 (0.5,2.27)	−0.35 (−1.24,0.54)	−3.58 (−4.25,−2.9)
Turkey	1990	421.54(295.85,580.04)	26.03(−13.26,113.96)	3.26(0.79,6.25)	31.97(5.88,86.19)	19.94(11.87,31.68)	2.51(0.71,4.81)	93.59(52.86,145.96)
	2021	32.9(27.28,39.19)	2.52(−1.26,10.45)	0.57(0.17,1.06)	5.23(1.3,11.55)	4.6(3.21,6.17)	0.4(0.12,0.77)	6.47(4.65,8.62)
	EAPC(95%CI)	−8.72 (−9.22,−8.21)	−8.29 (−8.75,−7.83)	−6.27 (−6.8,−5.73)	−6.6 (−7.05,−6.16)	−5.02 (−5.67,−4.37)	−6.59 (−7.03,−6.14)	−9.25 (−9.7,−8.79)
United Kingdom	1990	23.97(22.7,25.33)	5.56(−7.45,17.57)	0.47(0.1,0.85)	1.44(0.26,4.54)	1.81(1.27,2.41)	0.16(0.05,0.29)	7.16(5.79,8.51)
	2021	6.95(6.39,7.53)	1.66(−2.27,5.19)	0.18(0.04,0.32)	0.54(0.1,1.55)	0.84(0.62,1.08)	0.05(0.02,0.09)	1.21(0.92,1.52)
	EAPC(95%CI)	−3.73 (−3.92,−3.55)	−3.76 (−3.96,−3.55)	−2.94 (−3.15,−2.74)	−3.1 (−3.31,−2.89)	−2.41 (−2.57,−2.25)	−3.49 (−3.67,−3.31)	−5.7 (−5.84,−5.56)
United States of America	1990	22.16(21.3,23.02)	4.41(−4.79,14.2)	1.11(0.26,1.97)	3.15(0.67,7.32)	1.76(1.3,2.26)	0.22(0.07,0.42)	6.67(5.28,7.97)
	2021	5.01(4.65,5.41)	1.11(−1.21,3.58)	0.39(0.1,0.69)	0.93(0.24,1.93)	0.82(0.62,1.01)	0.06(0.02,0.11)	0.98(0.75,1.2)
	EAPC(95%CI)	−5.14 (−5.7,−4.58)	−4.73 (−5.33,−4.13)	−3.96 (−4.38,−3.53)	−4.55 (−5.08,−4.01)	−2.69 (−3.01,−2.36)	−4.94 (−5.55,−4.34)	−6.57 (−7.24,−5.89)

Trends of TB burden attributable to risk factors in G20 countries

The GBD 2021 study identified six level–3 risk factors associated with TB: smoking, high alcohol use, dietary risks, high BMI, high fasting plasma glucose, and low physical activity. South Africa, Indonesia, and India consistently exhibited the highest ASDR for TB associated with six risk factors, while Australia had the lowest (Table 1). In 1990 and 2021, smoking remained the leading exposure factor regarding PAF of ASDR among G20 countries, although its value range decreased from 0.10 to 0.22 to 0.06–0.10. Similarly, the PAF of ASDR values for all six risk factors declined; however, their rankings remained consistent in the following order: smoking, high alcohol use, high fasting plasma glucose, high BMI, dietary risks, and low physical activity (Fig. 2). In 1990, smoking ranked first in 15 member countries; however, by 2021, it was the leading risk factor in only five member nations: China, Indonesia, India, the Russian Federation, and Turkey, which ranked second in 12 member countries. In 2021, high alcohol use was the primary risk factor for TB in 14 member countries. In Saudi Arabia, high alcohol use ranked sixth among TB risk factors. The rankings for high fasting plasma glucose and high BMI were relatively stable, predominantly occupying the third and fourth positions. In 2021, high BMI became the predominant risk factor for TB in Saudi Arabia and ranked second in Australia, Mexico, and Turkey. Moreover, dietary risks and low physical activity rankings were consistent, with both factors predominantly occupying the fifth and sixth positions in the PAF of ASDR for TB across G20 member countries in 1990 and 2021(Fig. 2).

Fig. 2.

Rankings of age-standardized DALY rates attributable to six TB risk factors across G20 countries: 1990(A) versus 2021(B). The colors in the figure represent rankings from high (red) to low (blue), with the numbers representing the specific rankings of six risk factors. DALYs, disability-adjusted life years; TB, tuberculosis; G20, The Group of Twenty

Over the past 30 years, the PAF of DALYs for TB associated with each risk factor exhibited significant variability across age groups and genders. Smoking was the predominant risk factor for TB DALYs among individuals aged 30 to 64 years, with its contribution to DALYs increasing with age, peaking in the 55–59 age group and subsequently declining in the 60–64 age group. In the 65–69 age group, high fasting plasma glucose exceeded smoking as the predominant risk factor for TB DALYs. The contribution increased with age until the 80–84 age group, after which it began to decline. However, high fasting plasma glucose remained the leading risk factor for TB DALYs until the age of 95. A similar pattern was observed in males. For females, high BMI was the predominant risk factor for TB DALYs in individuals under 50 years; however, high fasting plasma glucose became the leading risk factor for those aged 50 and above. Significant gender and age differences were observed in the contributions of major risk factors to TB DALYs(Fig. 3).

Fig. 3.

The PAFs of TB DALYs attributable to six risk factors, stratified by age group and sex across G20 countries from 1990 to 2021. PAFs, population attributable fractions; TB, tuberculosis; DALYs, disability-adjusted life years

Smoking

From 1990 to 2021, the overall and country-specific ASDR attributable to smoking in G20 countries exhibited a declining trend. Compared to 1990, the ASDR attributable to smoking in G20 countries in 2021 was 53.13 (95%UI: 40.56–69.45), with an AAPC of −4.82%, (95%CI: −4.32% – −4.24%). Among the member countries, Indonesia (230.27, 95% UI: 168.08–308.09), South Africa (212.92, 95%UI: 155.94–273.88), and India (145.39, 95%UI: 106.96–199.62) had the highest ASDR attributable to smoking. The fastest declines were observed in South Korea (AAPC = 9.31%, 95%CI: −9.47% – 9.2%) and Turkey (AAPC=−8.28%, 95%CI: −8.37% – −8.21%) while the slowest declines were seen in Russia (AAPC=−2.24%, 95%CI: −2.57% – −1.9%) and Indonesia (AAPC=−2.85%, 95%CI: −2.88% – −2.82%). Over the three decades, the APC of ASDR mostly exhibited a downward trend, except for South Africa during 2001–2006(APC = 2.46%, 95%CI: −7.84% – 4.15%), Russia during 1990–1994(APC = 19.02%, 95%CI: 14.73% – 25.8%), 1994–2004(APC = 5.05%, 95%CI: 3.73% – 6.26%), the United States during 2014–2021(APC = 1.56%, 95%CI: 0.41% – 3.27%), and Italy during 2013–2017(APC = 0.04%, 95%CI: −2.05% – 3.39%) (Fig. 4A).

Fig. 4.

Temporal trends in age-standardized DALY rates of TB attributable to smoking (A), high alcohol use (B), high fasting plasma glucose (C), high BMI (D), dietary risks (E), and low physical activity (F) across G20 countries from 1990 to 2021 based on joinpoint regression analysis (*p < 0.05; *Y-axis scales differ across panels).TB, tuberculosis; DALYs, disability-adjusted life years; BMI, body mass index; G20, The Group of Twenty

As illustrated in the Fig. 5A, changes in TB disease burden were associated with variations in SEV of smoking in certain countries and regions. Brazil, Mexico, South Africa, the United Kingdom, Canada, and India experienced the most significant reductions in smoking exposure. However, Russia and Indonesia were the countries where SEV of smoking increased over the past three decades, which may explain the slower decline of ASDR from 1990 to 2021(Fig. 5A).

Fig. 5.

Percentage change in SEV levels for six TB risk factors across G20 countries from 1990 to 2021: smoking (A), high alcohol use (B), high fasting plasma glucose (C), high BMI (D), dietary risks (E), and low physical activity (F). TB, tuberculosis; SEV, summary exposure value; G20, Group of Twenty. BMI, body mass index

The inequality analysis demons that variations in SDI levels across regions resulted in absolute and relative disparities in TB disease burden. The absolute health inequality index decreased from − 709.53(95%CI: −929.20 – −489.86) in 1990 to −72.7(95%CI: −111.13 – −34.27) in 2021(Fig. 6A), signifying that narrowing socioeconomic development gaps were associated with a reduction in the TB disease burden in G20 countries. The absolute value of the concentration index reduce from 0.51(95%CI: −0.65 – −0.37) to 0.50(95%CI: −0.72 – −0.29) (Fig. 7A).

Fig. 6.

Health inequality regression curves of TB age-standardized DALY rates attributable to six risk factors across G20 countries, 1990 versus 2021: smoking (A), high alcohol use (B), high fasting plasma glucose (C), high BMI (D), dietary risks (E), and low physical activity (F). Panels show SII relationships between SDI and age-standardized DALY rates. Point size = population. (*Y-axis scales differ across panels). DALYs, disability-adjusted life years; TB, tuberculosis; G20, Group of Twenty. BMI, body mass index; SII, slope index of inequality; SDI, sociodemographic index

Fig. 7.

Health concentration curves of TB age-standardized DALY rates attributable to six risk factors across G20 countries, 1990 versus 2021: smoking (A), high alcohol use (B), high fasting plasma glucose (C), high BMI (D), dietary risks (E), and low physical activity (F). Panels present the concentration index, which quantifies relative inequalities by integrating the area under the Lorenz curve, aligning age-standardized DALY rates distribution with population distribution by SDI. (*Y-axis scales differ across panels). DALYs, disability-adjusted life years; TB, tuberculosis; G20, Group of Twenty. BMI, body mass index; SDI, sociodemographic index

Decomposition analysis revealed that the reduction in TB burden associated with smoking in G20 and most member nations was closely correlated with the decreasing trend in smoking prevalence. Population aging contributed only marginally to the increase in TB disease burden, and in some countries, it reduced the burden. However, changes in national population size were the primary factor attributable to the increase in DALYs associated with smoking, with Saudi Arabia experiencing the largest impact, resulting in a 215.53% increase(Fig. 8A).

Fig. 8.

Population-level determinant changes (aging, population growth, and epidemiological changes) for six risk factors-related TB DALYs across G20 countries from 1990 to 2021: smoking (A), high alcohol use (B), high fasting plasma glucose (C), high BMI (D), dietary risks (E), and low physical activity (F). Black dots represent the total change contributed by all four components. A positive value for each component indicates a corresponding positive contribution in DALYs, and a negative value indicates a corresponding negative contribution in DALYs. TB, tuberculosis; DALYs, disability-adjusted life years; G20, Group of Twenty. BMI, body mass index

The BAPC analysis indicated that by 2035, most G20 countries will experience a decline in TB DALYs related to smoking compared to 2021, except for Canada, Mexico, South Korea, South Africa, and the United States. India, Indonesia, and China are projected to have the highest TB burden in 2035, with DALYs estimated at 1,511,921.016(95%CI: 638,196.66–2,385,645.37), 486,332.53(95%CI: 268,295.93–704,369.12), and 265,944.07(95%CI: 0–549,915.09), respectively. However, Australia (170.42, 95%CI: 2.57–365.37) is expected to have the lowest DALYs. Additionally, except for Australia, males are projected to more DALYs attributable to smoking than females in 2035(Fig. 9). Regarding ASDR, the top three countries are South Africa(436.12, 95%CI: 0–1,215.54), Indonesia(279.04, 95%CI: 153.57–404.51), and India(173.73, 95%CI: 72.80–274.65). While Canada and the United States are expected to see an increase in ASDR attributable to smoking, all other countries are projected to experience a decline.

Fig. 9.

Sex-stratified trends in DALYs counts and age-standardized DALY rates for smoking-related TB across G20 countries: observed (1990–2021, solid lines) and BAPC-projected (2022–2035, dashed lines) (*Y-axis scales differ across panels). TB, tuberculosis; DALYs, disability-adjusted life years; G20, Group of Twenty; BAPC, Bayesian age-period-cohort

High alcohol use

From 1990 to 2021, the overall and country-specific ASDR attributable to high alcohol use in G20 countries exhibited a declining trend. Compared to 1990, the ASDR attributable to high alcohol use in G20 countries in 2021 was 43.92 (95%UI: −28.25–162.14), with an AAPC of −2.91%(95%CI: −2.94% – −2.88%). Among the member countries, South Africa (358.49, 95%UI: 227.57–1168.71) exhibited the highest ASDR attributable to high alcohol use, while Australia remained the lowest at 1.02 (95UI: −1.29–3.25). The countries with the smallest AAPC values were South Korea (AAPC=−8.24%, 95%CI: −8.39% – −8.14%) and Turkey (AAPC=−7.36%, 95%CI: −7.44% – −7.27%). The slowest declines were observed in South Africa (AAPC=−1.93%, 95%CI: −2.16% – −1.73%) and Russia (AAPC=−2.34%, 95%CI: −2.65% – −2.02%)(Fig. 4B).

As illustrated in the Fig. 5B, changes in TB disease burden were associated with variations in SVE of high alcohol use in certain countries and regions. India, Brazil, China, Mexico, Canada, South Korea, and the United Kingdom were among the countries where SVE of high alcohol use increased(Fig. 5B). This may explain why India ranked second regarding DALYs attributable to high alcohol use.

Health inequality analysis revealed that the SII associated with TB attributable to high alcohol use improved significantly from − 183.00(95%CI: −264.69 – −101.32) in 1990 to −37.37(95%CI: −52.20 – −22.55) in 2021. The SII and concentration index were negative, suggesting a heavier TB burden among populations in lower SDI countries(Fig. 6B). The absolute value of the concentration index increased from 0.45(95%CI: −0.63 – −0.26) to 0.55(95%CI: −0.8 – −0.3) (Fig. 7B), suggesting a worsening of inequity in TB distribution among various socioeconomic groups, with a relatively heavier burden among lower socioeconomic populations in 2021.

Except for South Africa, where DALYs attributable to high alcohol use increased by 6.45%, other G20 countries experienced a reduction, with males in South Africa being the predominant contributors. Decomposition analysis indicated that the reduction in TB burden attributable solely to high alcohol use exhibited a negative impact on the overall TB burden in G20 countries and member states. Population aging exacerbated the disease burden in Japan, while population growth worsened the burden in Saudi Arabia and Indi (Fig. 8B).

To project TB burden and trends attributable to high alcohol use up to 2035, BAPC analysis revealed that compared to 2021, the DALYs attributable to high alcohol use in G20 countries will differ by 2035. Increases are expected in India, South Africa, South Korea, the United States, France, Turkey, Canada, and Australia. By 2035, India is projected to have the highest DALYs attributable to high alcohol use at 2,060,213.85(95%CI: 23,975.94–212,813.78), followed by South Africa(68,454.40, 95%CI: 0–253,584.13) and China(12,243.74, 95%CI: 0–25,541.70). However, Saudi Arabia is expected to have 381.03(95%CI: 0–1,241.75) DALYs. Males are projected to contribute more to TB DALYs than females in 2035. Regarding ASDR, South Africa, India, and Russia are the top three countries(Figure S2).

High fasting plasma glucose

From 1990 to 2021, the overall and country-specific ASDR attributable to elevated high fasting plasma glucose in G20 countries, except for South Africa, exhibited a downward trend. In 2021, the ASDR attributable to high fasting plasma glucose in G20 countries was 33.19 (95%UI: 23.81–44.74), with an AAPC of −2%(95%CI: −2.09% – −1.92%) compared to 1990. Among the member nations, South Africa (156.61, 95%UI: 112.55–208.24) had the highest ASDR attributable to high fasting plasma glucose, followed by Indonesia and India. Australia remained the lowest at 0.42 (95%UI: 0.29–0.55). The nations with the smallest AAPC values were South Korea (AAPC=−6.51%, 95%CI: −6.61% – −6.42%) and China (AAPC=−5.86%, 95%CI: −5.99% – −5.75%). South Africa exhibited an upward trend at 0.82% (AAPC = 95%CI: 0.69% – 0.95%); however, Russia exhibited the slowest decline with an AAPC of −0.73% (95%CI: −1.05% – −0.4%). Throughout the past three decades, the APC of ASDR predominantly demonstrated a downward trend; however, some countries exhibited positive values during specific periods(Fig. 4C).

India exhibited the highest percent change in SEV of high fasting plasma glucose, indicating the most significant increase in exposure over the three decades. In addition, during this period, India, Canada, Brazil, China, and Mexico exhibited increasing trends in SEV of high fasting plasma glucose(Fig. 5C).

The SII for TB decreased from − 273.71(95%CI: −363.61 – −183.81) in 1990 to −45.15(95%CI: −75.03 – −15.27) in 2021(Fig. 5C), signifying a significant decrease in the absolute gap in TB burden attributable to high fasting plasma glucose among various socioeconomic groups. The concentration index slightly decreased from − 0.58(95%CI: −0.75 – −0.4) in 1990 to −0.6(95%CI: −0.8 – −0.39) in 2021(Fig. 7C), indicating a minor deterioration in health inequity of TB attributable to high fasting plasma glucose during this period.

Decomposition analysis indicated that, in 2021, the total DALYs attributable to high fasting plasma glucose across G20 countries increased by 10.41% compared to 1990. This change was influenced by population growth, aging, and factors beyond high fasting plasma glucose alone, while the risk associated with high fasting plasma glucose had a negative impact. Saudi Arabia was an exception, where population aging contributed to decreased DALYs. Furthermore, India, Indonesia, Brazil, South Africa, Saudi Arabia, Canada, and Australia exhibited varying degrees of DALY increase(Fig. 8C). The increases in India, Canada, and Brazil correlated with rising SEV of high fasting plasma glucose.

By 2035, India(2,414,381.06, 95%CI: 555,601.26–4,454,594.62), Indonesia(456,613.72, 95%CI: 11,528.59–1,088,744.93), and South Africa(144,689.75, 95%CI: 0–453,983.93) will have the highest TB DALYs related to high fasting plasma glucose. Regarding ASDR, the rankings are expected to be South Africa(368.97, 95%CI: 0–1,161.14), India(247.84, 95%CI: 56.56–458.03), and Indonesia(243.06, 95%CI: 6.35–577.75). Compared to 2021, 15 countries, Argentina, Australia, Canada, France, Germany, India, Indonesia, Japan, Mexico, South Korea, Saudi Arabia, South Africa, Turkey, and the United States, are projected to experience an increase in DALYs by 2035. (Figure S3).

High BMI

From 1990 to 2021, the overall and country-specific ASDR attributable to high BMI in G20 countries, except for South Africa(0.1, 95%CI: −0.12% – 0.30%), exhibited a declining trend. The TB ASDR attributable to high BMI in G20 countries in 2021 was 16.9 (95%UI: 3.61–43.95), with an AAPC of −0.93% (95%CI: −0.96% – −0.90%) compared to 1990. The APC values were positive during specific intervals: 0.26%(95%CI: −0.10% – 0.37%) from 1990 to 2000 and 0.90% (95%CI: −1.00% – 1.16%) from 2000 to 2003. In 2021, among the member countries, South Africa (148.59, 95%CI: 26.84–419.51) had the highest ASDR related to high BMI, followed by India (45.11, 95%CI: 10.00–111.67) and Indonesia (36.27, 95%CI: 7.52–105.68). Canada had the lowest ASDR at 0.48 (95%CI: 0.09–1.31). The countries with the fastest declines in AAPC were South Korea (AAPC=−6.10%, 95%CI: −6.19% – −6.01%), while Russia (AAPC=−0.16%, 95% CI: −0.51% – 0.23%) exhibited the slowest decline. Over the past three decades, the ASDR predominantly exhibited a downward trend; however, some countries fluctuated between positive and negative values during specific periods(Fig. 4D).

Health inequality analysis revealed that TB ASDR attributable to high BMI decreased with higher SDI. The SII decreased from − 64.15(95%CI: −89.59 – −38.70) in 1990 to −42.79(95%CI: −61.61 – −21.98) in 2021(Fig. 6D), signifying a decrease in absolute health inequity associated with TB attributable to high BMI. However, the improvement was less significant compared to other risk factors. The concentration index is −0.33(95%CI: −0.50 – −0.61) in 1990 and − 0.45(95%CI: −0.65 – −0.25) in 2021, which increasingly approached 1(Fig. 7D), indicating that expanding social disparities have significantly exacerbated the TB burden related to high BMI.

Decomposition analysis demonstrated that, in 2021, the DALYs related to high BMI in G20 countries and South Africa, Indonesia, Saudi Arabia, India, Brazil, Australia, and Russia increased compared to those in 1990. Non-high BMI but related factors and population changes were the primary drivers of this increase. In Saudi Arabia, population growth considerably influenced the increase in DALYs, while in other countries, factors other than high BMI played a significant role(Fig. 8D).

The percent change in SEV of high BMI was positive across all G20 countries, indicating an increase in high BMI exposure, differentiating it from other risk factors. China, India, Indonesia, and Saudi Arabia were the top four countries with the highest percent change in SEV of high BMI. Except for China, the DALYs related to high BMI in the remaining three countries increased in 2021, indicating a possible correlation between national SEV of high BMI and rising DALYs (Fig. 5D).

Furthermore, the BAPC model projected that by 2035, India(830,045.92, 95%CI: 368,106.92–1291,984.92), South Africa(190,593.05, 95% CI: 0–598,418.34), and China(159,550.39, 95% CI: 0–347,739.55) would have the highest DALYs attributable to high BMI. Except for Italy, Japan, and the United Kingdom, all other countries are expected to have higher DALYs attributable to high BMI in 2035 compared to 2021 (Figure S4).

Dietary risks

From 1990 to 2021, the overall and nation-specific ASDR associated with dietary risks in G20 countries demonstrated a declining trend. Compared to 1990, the ASDR attributable to dietary risks in G20 countries in 2021 was 4.94 (95%UI: 1.09–8.97), with an AAPC of −2.73%(95%CI: −2.77% – −2.69%). In 2021, among the member countries, South Africa (19.03, 95%UI: 4.28–35.03) had the highest ASDR attributable to dietary risks, followed by India (16.49, 95%UI: 3.31–30.06) and Indonesia (8.90, 95%UI: 1.67–16.85). Canada had the lowest ASDR at 0.1 (95%UI: 0.02–0.18). South Korea(AAPC=−6.27%, 95%CI: −6.35% – −6.17%), Japan(AAPC=−5.49%, 95%CI: −5.62% – −5.35%), and Turkey(AAPC=−5.49%, 95%CI: −5.57% – −5.41%) experienced the fastest declines in AAPC. Over the three decades, the APC of ASDR predominantly demonstrated a downward trend; however, South Africa and Russia experienced significant fluctuations in APC values, with their TB burden related to dietary risks indicating an increasing trend during certain intervals(Fig. 4E).

Health inequality analysis revealed that ASDR related to dietary risks diminished with higher SDI. The SII decreased from − 32.97(95%CI: −45.52 – −20.42) in 1990 to −9.71(95%CI: −15.40 – −4.02) in 2021(Fig. 6E), signifying an enhancement in relative health inequity associated with TB attributable to dietary risks compared to 1990. However, the absolute value of the concentration index slightly increased between 1990 and 2021, indicating a minor worsening of health inequity associated with TB, although the difference was non-significant(Fig. 7E).

During this period, Saudi Arabia, South Africa, Brazil, Indonesia, and Australia experienced increased TB DALYs related to dietary risks. Additionally, the percentage change in risk diminished the TB burden attributable to dietary risks, while the percentage change in population and aging increased DALYs attributable to dietary risks, with population growth being a major driver of changes in TB burden in most countries(Fig. 8E).

By 2035, India(244,237.93, 95%CI: 91,327.94–397,147.91), China(27,447.66, 95%CI: 793.03–56,103.41), and Indonesia(25,612.47, 95%CI: 2,647.01–50,167.59) are projected to exhibit the highest DALYs related to dietary risks. Regarding ASR, South Africa(50.69, 95%CI: 0–173.30), India(24.19, 95%CI: 9.02–39.36), and Indonesia(12.83, 95%CI: 1.34–25.12) are expected to have the heaviest burden. Compared to 2021, China, Germany, Italy, Japan, and the United Kingdom are projected to experience a reduction in DALYs related to dietary risks by 2035, while all other countries are expected to experience an increase. Regarding ASDR, Brazil, China, France, Germany, India, Indonesia, Italy, Japan, South Korea, Saudi Arabia, and the United Kingdom are projected to exhibit a declining trend. Countries with inconsistent changes in DALYs and ASDR include Brazil, France, India, Indonesia, South Korea, and Saudi Arabia (Figure S5).

Low physical activity

From 1990 to 2021, the overall and country-specific ASDR attributable to low physical activity in G20 countries exhibited a downward trend. The ASDR attributable to low physical activity in G20 countries in 2021 was 2.07 (95%UI: 0.68–3.82), with an AAPC of −3.13% (95%CI: −3.19% – −3.06%) over the three decades. In 2021, among the member countries, South Africa (9.44, 95%CI: 3.04–17.16) had the highest ASDR attributable to low physical activity, followed by Indonesia (9.24, 95%CI: 2.86–17.13) and India (7.01, 95%CI: 2.31–13.41). Canada had the lowest ASDR at 0.02 (95%CI: 0.01–0.04). The countries with the fastest declines in AAPC were South Korea (AAPC=−6.67%, 95%CI: −6.68% – −6.65%), while South Africa (AAPC=−0.47%, 95%CI: −0.68% – −0.32%) demonstrated the slowest decline. Over the three decades, APC predominantly exhibited a downward trend, although South Africa and Russia experienced significant fluctuations in APC values, alternating between positive and negative values(Fig. 4F). From 1990 to 2021, most G20 countries remained in a state of high SEV of low physical activity, with India and Germany exhibiting a reduction in this percentage. Saudi Arabia, Australia, and Brazil exhibited the greatest regarding SEV for low physical activity(Fig. 5F).

Health inequality analysis revealed that ASDR related to low physical activity diminished with higher SDI. The SII decreased from − 25.52(95%CI: −45.52 – −20.42) in 1990 to −3.13(95%CI: −45.52 – −20.42) in 2021(Fig. 6F), signifying an enhancement in relative health inequity related to TB attributable to low physical activity compared to 1990, as socioeconomic disparities narrowed and reduced the TB burden. The concentration index changed from − 0.52(95%CI: −0.65 – −0.38) to −0.5(95%CI: −0.67 – −0.33)(Fig. 7F), with its absolute value approaching 0, indicating improved health inequity.

Decomposition analysis revealed that non-physical activities and associated factors exacerbated the TB burden attributable to low physical activity in most countries. Besides, the percent change in risk reduced the TB burden attributable to low physical activity; however, the percent change in population and aging increased DALYs attributable to dietary risks, with population growth being a major driver of changes in TB burden attributable to low physical activity in most countries. However, in countries such as Japan, population aging contributed more significantly to the TB burden(Fig. 8F).

The BAPC analysis projected that by 2035, India(101,604.37, 95%CI: 26,177.41 − 177,031.32), Canada(31,722.12, 95%CI: 0–9,001,899.86), and Indonesia(25,037.69, 95%CI: 6,825.57–43,249.80) will exhibit the highest DALYs attributable to TB, while the United Kingdom(49.41, 95%CI: 0–131.20) will record the lowest. Regarding ASDR, Canada(169.28, 95%CI: 0–48051.85), South Africa(22.99, 95%CI: 0–76.94), and Indonesia(12.54, 95%CI: 3.43–21.64) are expected to rank the highest, while Germany(0.05, 95%CI: 0–0.17) is expected to rank the lowest. By 2035, Australia, Canada, Mexico, Russia, South Africa, and the United States are projected to have an increase in ASDR compared to 2021. All other countries are expected to experience a decline in ASDR (Figure S6).

Discussion

This study comprehensively assessed TB burden across G20 countries, examining historical trends, regional variations, age-specific patterns, DALYs, and ASDR. We quantified correlations between six major risk factors and TB burden using: regional heatmap ranking, Joinpoint regression, SEV change assessment, health inequality metrics, and decomposition analysis. We projected DALYs and ASDR trends attributable to six TB risk factors through 2035. These findings underscore TB’s persistent health threat in G20 countries and provide evidence-based guidance for targeted policy formulation and resource allocation.

This study demonstrated a significant 50% reduction in TB DALYs across G20 countries from 1990 to 2021, indicating diminished overall health impact. This trend aligns with GBD study findings [11]. This improvement can be attributed primarily to public health interventions, advances in medical technology, policy improvements, increased funding, socioeconomic progress, and strengthened international cooperation and surveillance among G20 countries. India (15,057.8), Indonesia (3,162.14), and China (1,375.51) ranked highest in TB DALYs burden among G20 countries. These countries have large populations, collectively accounting for over 40% of the global population, which considerably influences the overall burden of TB. After age standardization, which removes the confounding effect of population age composition, the data more precisely reflect TB’s impact on the working-age population. In 2021, South Africa recorded the highest ASDR for TB (1,855.04), followed by Indonesia (1,220.1) and India (1,119.71). South Africa’s leading position may be partly due to its highest global HIV prevalence [30], HIV/AIDS co-infection increases the likelihood of developing active TB. Although South Africa has fewer total TB cases than China and Indonesia, the higher TB prevalence among its younger population and uneven distribution of healthcare resources result in a greater disease burden after age standardization compared with these countries. China’s relatively lower disease burden ranking suggests successful TB control measures despite an aging population.

In G20 countries, the 35–64 years age group carried the highest TB DALYs burden, largely attributable to its substantial population size. This working-age cohort demonstrates elevated TB susceptibility and experiences treatment delays due to workforce participation constraints and postponed care-seeking behavior, which may compromise immune function. By 2021, TB burden among children and adolescents had decreased significantly across the G20 and most member states. However, South Africa exhibited a disproportionately elevated TB burden in children under 9 years old, a phenomenon strongly associated with prevalent HIV-TB coinfection in the country [31]. However, countries such as Italy, Australia, France, Japan, and South Korea exhibited higher DALYs among individuals aged ≥ 70 years, reflecting their aging populations. Age-related immunosenescence elevates progression risk from latent to active TB. Older adults frequently experience comorbidities and malnutrition, predisposing them to atypical presentations, complications, and drug-resistant TB, significantly hindering diagnosis and treatment. These findings suggest population-specific interventions may enhance TB control efficacy. Implementing mandatory annual medical examinations in high-risk industries and establishing a national TB data platform integrated with corporate health systems could improve TB detection rates among working-age populations, particularly in high-burden countries such as Indonesia and India. Integrating comprehensive household contact investigation with community-based strategies remains critical for reducing the pediatric TB burden [32, 33], particularly in regions like South Africa. Among aging G20 countries, active screening and treatment of latent TB in older adults, development of shorter and less-toxic treatment regimens, and facilitation of early diagnosis are critical for achieving the 2025 targets of the WHO End TB Strategy [34, 35].

Regarding the gender disparity in TB disease burden, the ASDR was consistently higher among males than females in both 1990 and 2021, with Saudi Arabia being the sole exception. Studies suggest this pattern may be linked to immunosuppressive effects of testosterone–particularly the suppression of Th1-type immune responses [36]. Additionally, males exhibit greater occupational exposure to confined/dust-laden environments and higher prevalence of smoking and alcohol use, further contributing to their elevated TB burden.

The ranking of six risk factors attributable to TB burden, ordered PAF from highest to lowest was: smoking, high alcohol use, high fasting plasma glucose, high BMI, dietary risks, and low physical activity. South Africa, Indonesia, and India ranked among the top three in ASDR of TB across all risk factors except high alcohol use, suggesting socioeconomic development levels and healthcare infrastructure influence TB burden. Results indicate that despite declining values between 1990 and 2021, smoking remained the leading risk factor-related TB ASDR in G20 countries, closely followed by high alcohol use. Alcohol misuse emerged as the primary TB risk factor in 14 countries. Evidence demonstrates alcohol use disorder increases TB susceptibility, enhances infectiousness, accelerates disease progression, and elevates risks of adverse treatment outcomes [37]. Moreover, daily alcohol intake exhibits a dose-dependent positive association with pulmonary TB risk [38]. Mechanistic studies demonstrate that alcohol abuse induces glutathione (GSH) depletion in alveolar cells, amplifying oxidative stress while reducing the peripheral abundance and functionality of antigen-presenting cells (APCs) [39, 40]. These pathophysiological alterations may: increase alveolar epithelial permeability; suppress surfactant synthesis; impair macrophage phagocytic efficiency, and cause ciliary dysfunction through desensitization and motility impairment [41]. Consequently, these effects collectively facilitate Mycobacterium tuberculosis infection by compromising host defenses and enhancing pulmonary TB susceptibility.

A 2021 meta-analysis demonstrated that individuals with Hemoglobin A1c(HbA1c) levels ≥ 7.0% had a 2.05-fold higher risk of developing pulmonary TB, with significantly higher mean HbA1c levels observed in diabetes-TB co-morbidity groups versus diabetes-only cohorts [42]. From 1990 to 2021, the AAPC in ASDR attributable to high fasting plasma glucose showed a declining trend. However, DALYs increased by 10.41% during this period. Decomposition analysis identified population aging as a positive contributor to TB DALYs. Among G20 countries, South Africa, Indonesia, and India exhibited the heaviest TB burden attributable to high fasting plasma glucose. Consequently, prioritizing glycemic control in older adults, particularly in lower-resource G20 countries, may reduce the overall TB burden.

Sedentary behavior and energy-dense diets are key modifiable factors linking elevated BMI to increased TB burden. Notably, studies confirm that low BMI exacerbates TB severity and mortality [43], while BMI < 24 kg/m² with concurrent diabetes significantly elevates TB risk [44]. Aligning with GBD 2021 evidence on BMI as a TB risk factor, our analysis reveals rising DALYs across G20 countries and seven high-burden members since 1990. The observed divergence in BMI-TB associations between population-based studies and clinical cohorts stems from methodological differences. Clinical investigations predominantly capture hospital-diagnosed TB cases exhibiting disease-associated wasting, whereas GDB estimates integrate nationally representative surveillance data. Critically, the GBD study documents an association between elevated BMI and TB risk, implicating metabolic comorbidities (e.g., diabetes and obesity) through immunometabolic dysregulation and adipose tissue inflammation––well-established biological mechanisms underlying immune compromise and heightened TB susceptibility. Consequently, improving dietary quality, enhancing physical activity, and reducing adiposity represent viable interventions for mitigating TB burden.

Projections of future TB DALYs indicate that, in most G20 countries, the burden of TB attributable to smoking will decrease compared to 2021. However, the burden associated with other risk factors is expected to increase in multiple countries: nine nations will see an increase due to high alcohol use, 15 due to high fasting plasma glucose, 16 due to high BMI, and 14 due to dietary risks and physical inactivity. By 2035, India, Indonesia, China, and South Africa will continue to face a high TB burden. Based on the findings of this study, targeted public health policies and interventions are essential for TB prevention and control.

These findings critically inform TB control strategies across the heterogeneous healthcare systems of G20 nations. At the policy level, they enable evidence-based prioritization of interventions targeting high-impact modifiable risks, including: Strategic resource allocation to high-burden populations; development of program monitoring metrics; and implementation of risk-stratified screening approaches. Clinically, the substantial disease burden attributable to comorbidities (e.g., diabetes) and lifestyle factors–particularly tobacco and alcohol use–necessitates integrated management protocols that concurrently address these conditions during TB treatment. Comprehensive interventions targeting primary risk factors (smoking and high alcohol use) should encompass: (1) Legislative measures: comprehensive indoor smoking bans, alcohol advertising restrictions, and taxation policies [45]; (2) Nationwide cessation support integrated into TB programmes; (3) Health education emphasizing causal links between substance use, poor treatment adherence, and adverse TB outcomes. Establishing multinational surveillance systems to correlate substance consumption with TB epidemiological indicators is essential. Concurrently, cross-sectoral collaboration among health, finance, and justice sectors may optimize cost-effectiveness in TB and noncommunicable disease co-management strategies [46]. Moving forward, G20 countries should enhance collaborative frameworks to improve population health outcomes.

Limitations

Substantial underreporting of TB cases and variable data quality from health statistical systems with limited capacity in low- and middle-income countries likely contribute to burden underestimation. Additional limitations include: (1) Incomplete adjustment for confounding factors and comorbidities (e.g., HIV/AIDS, diabetes), potentially biasing burden estimates and obscuring disease interactions; (2) Observational designs establishing associations but precluding causal inference; (3) Absence of GBD data on treatment settings, adherence, outcomes, and disease severity, restricting healthcare quality assessments and DALY severity spectrum evaluations; (4) Limited stratification beyond country/sex/age dimensions, impeding high-risk subpopulation identification despite analysis of core social determinants of health; and (5) Fundamental dependence on GBD modeling methodologies without independent external validation, limiting generalizability beyond this framework. Consequently, addressing these methodological constraints and data limitations requires urgent prioritization in future research to enable robust TB burden estimation and effective public health responses.

Conclusion

This study provided a comprehensive analysis of the burden of TB in G20 countries. From 1990 to 2021, TB DALYs and ASDR in G20 countries exhibited a declining trend, although significant variations were observed among countries. The TB burden was most severe among middle-aged and older working-age populations, with significant disparities between males and females. In addition, countries with higher SDI levels generally exhibited a lower TB burden. Six risk factors, all related to individual behaviors or lifestyles, were associated with TB ASDR. Smoking remained the leading factor attributable to the high TB burden; however, the contributions of the other five risk factors to DALYs and ASDR were significant. By 2035, the burden of TB associated with different risk factors is anticipated to differ, with most countries experiencing a reduction in TB burden due to smoking, while DALYs attributable to other risk factors are expected to increase in many countries. Therefore, public health interventions, including smoking cessation, alcohol moderation, dietary improvements, and increased physical activity, can effectively reduce exposure to these risk factors and lower the risk of TB incidence. Similarly, G20 countries must strengthen collaboration, increase investments in healthcare resources, and enhance the scope of preventive measures. These efforts will significantly alleviate the TB burden across countries and narrow the disparities in TB burden among G20 countries.

Abbreviations

AAPC

Average annual percentage change

APC

Annual percentage change

ASDR

Age-standardized DALY rates

ASR

Age-standardized rates

BAPC

Bayesian age-period-cohort

BMI

Body mass index

CI

Confidence interval

DALYs

Disability-adjusted life years

EAPC

Eatimated annual pecent change

G20

The Group of Twenty

GBD

Global Burden of Disease

IHME

Institute for Health Metrics and Evaluation

MDR-TB

Multidrug-resistant TB

PAF

Population attributable fraction

SDI

Sociodemographic index

SEV

Summary exposure value

SII

Slope index of inequality

TB

Tuberculosis

UI

Uncertainty interval

WHO

World Health Organization

Authors’ contributions

All authors have made significant contributions to this research. YK is responsible for formal analysis, creating visualizations, and drafting original manuscripts. SW has contributed to data organization, investigation, analysis, review, and manuscript editing. JS and XT participated in project supervision and manuscript review. RZ has made contributions to formal analysis, resources, and manuscript editing. RL is responsible for reviewing and editing grassland drafts. XL is dedicated to creating visualizations and has made contributions to supervision, methodology, project management, manuscript review, and editing. All authors have read and approved the final manuscript.

Funding

This study was supported by:

1. Young Scientists Program of the Natural Science Foundation of Shaanxi Province [No. 2025JC-YBQN-1124].

2. Shaanxi Provincial Natural Science Basic Research Program [No. 2024JC-YBMS-729].

3. General Research Program of Shaanxi Provincial Department of Education [No. 24JK0399].

4. Undergraduate Training Program for Innovation and Entrepreneurship [No. S202410716126].

5. Public Health Research and Innovation Team Project of Shaanxi Provincial Bureau of Disease Prevention and Control [No. 202512].

6. Innovative Research Team on the Etiology and Mechanisms of Chronic Non-communicable Diseases at Shaanxi University of Chinese Medicine.

7. The Key Laboratory Scientific Research Program of the Education Department of Shaanxi Provincial Government [No. 24JS009].

8. The Key Laboratory of Environment-related Diseases and TCM Prevention and Control in Universities of Shaanxi Province, Xianyang City excellent innovation team plan [NO. L2024-CXNL-KJRCTD-KJ TD-0014].

9. Project of Administration of Traditional Chinese Medicine of Shaanxi Province [NO. SZY-KJCYC-2023–059].

10. Basic Research Project of Traditional Chinese Medicine of Shaanxi Provincial Administration of Traditional Chinese Medicine [No. 2021-GJ-JC012].

11. General Project of Shaanxi Provincial Natural Science Basic Research Program [No. 2021JM-473].

Data availability

The data supporting this study are derived from the GBD 2021, which offers comprehensive estimates of health metrics, including tuberculosis, across 204 countries and territories. These estimates are publicly accessible through the Global Health Data Exchange (GHDx) at https://vizhub.healthdata.org/gbd-results/.

Declarations

Ethics approval and consent to participate

This study is a secondary analysis of publicly available data from the GBD 2021 database. As the data is anonymized and does not involve direct interaction with human participants, ethical approval and informed consent to participate were not required for this study. Clinical trial number: not applicable.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.