Benchmarking progress in cause-specific cancers in China: a nationwide analysis of premature mortality from 1990 to 2023

Summary

Background

The United Nations passed Sustainable Development Goal 3.4 (SDG 3.4), which aims to reduce by one-third premature mortality from noncommunicable diseases, including cancer, by 2030. However, there is still no comprehensive assessment of the historical trends and projections in premature cancer mortality nationally, regionally, and by cancer type.

Methods

We estimated premature mortality by sex, age, and 33 provinces for China using the results of the Global Burden of Disease Study 2023. The premature mortality from total cancers and 29 cancer types was estimated for ages 30–69 years, using standard life table methods. The average annual rate of change was calculated to examine the past trends in premature mortality from 1990 to 2023. We further assessed whether historical trends would be sufficient to meet the SDG 3.4 targets for reducing premature mortality by 2030.

Findings

Over the period from 1990 to 2023, the cancer-related premature mortality declined from 13.11% to 6.66% (AAPC = −1.93%, 95% CI = −2.20% to −1.66%). In particular, stomach cancer (AAPC = −3.80%, 95% CI = −4.18% to −3.42%), esophageal cancer (AAPC = −3.40%, 95% CI = −3.75% to −3.06%), nasopharynx cancer (AAPC = −3.55%, 95% CI = −4.03% to −3.07%), and Hodgkin's lymphoma (AAPC = −4.89%, 95% CI = −5.12% to −4.66%) showed higher rates of decline. However, increasing premature mortality from pancreatic, breast, and kidney cancers was documented among males in northern, northeastern, and northwestern China. The causes of greatest premature mortality were lung cancer (1.97%), stomach cancer (0.98%), colon and rectum cancer (0.65%), esophageal cancer (0.65%), and liver cancer (0.48%) throughout the study period. In 2023, cancer-related premature mortality peaked in Heilongjiang (11.44%), followed by Chongqing (11.15%), Sichuan (10.49%), Liaoning (10.26%), and Shandong (10.00%) for males. Heilongjiang (6.63%), Jilin (5.63%), Liaoning (5.44%), Xinjiang (5.37%), and Inner Mongolia (4.98%) comprised the five provinces with the highest premature cancer mortality among females. Whereas cancer-related premature mortality declined at an average annual rate of 2.05% from 1990 to 2023, meeting the SDG 3.4 targets necessitates a future decline, 1.3 times faster. Hubei (2.89%), Shanghai (2.75%), and Hong Kong (2.71%) exhibited the fastest decline rates and were projected to meet the SDG 3.4 targets based on current trends.

Interpretation

From 1990 to 2023, the probability of dying from cancers declined across all provinces in China. Cancers with primary prevention or early detection strategies in place, such as stomach cancer, colorectal cancer, and cervical cancer, were performing better in premature mortality compared with cancers that lack such strategies, such as kidney cancer, multiple myeloma and pancreatic cancer.

Funding

None.

Research in context.

Evidence before this study

We searched PubMed, Google Scholar, Web of Science, and China Academic Journals database for papers with the terms “premature mortality” “cancer” or “projection”. We identified studies reporting premature mortality reductions for non-communicable diseases. A study based on WHO Global Health Estimates examined the premature mortality from cancer for different countries, WHO regions, and World Bank income levels. The study reports that despite the declining trend in the global premature mortality from cancer, achieving the SDG 3.4 targets remains challenging, even in high-income and upper-middle-income countries. Assessing progress in cancer prevention and control requires comprehensive estimates and ongoing tracking of premature mortality rates and the trend of different cancer types. However, to the best of our knowledge, no studies have assessed premature mortality by cancer type, province, or region in China. Predicting premature mortality by cancer type and region assists in assessing the gaps between various types of cancers at different regional levels in achieving the SDG 3.4 targets.

Added value of this study

The present study provides, to our knowledge, the first estimates of premature mortality by cancer type, region and province. We examined the trends in premature mortality during the period from 1990 to 2023 to identify specific geographic and demographic populations experiencing slower progress. We estimated whether the SDG 3.4 targets of reducing premature mortality by region and cancer type are likely to be achieved. As a populous nation, the evolving cancer burden in China holds critical importance for the global attainment of SDG 3.4 targets. This study provides valuable insights for international efforts, particularly in low- and middle-income countries. It also provides a benchmark for evaluating SDG 3.4 targets and a crucial foundation for adapting prevention and control strategies in the next decade.

Implications of all the available evidence

Although the premature cancer mortality rate declined from 1990 to 2023 in China, meeting the SDG 3.4 targets necessitates a future rate of decline that is 1.3 times faster. The premature mortality of stomach cancer, esophageal cancer, nasopharynx cancer, and Hodgkin's lymphoma, showed higher rates of decline. Premature mortality declines more rapidly in the central, eastern and, southwestern regions, and sex-related disparities in premature mortality vary across regions by cancer type. These projections serve as a benchmark for measuring the impact of future interventions and identifying the need for continued efforts to implement policies based on the type of cancer in a given region.

Introduction

Cancer is a major contributor to the global burden of disease and is estimated to be the leading cause of death before the age of 70 in most countries.^1,2 According to the report of the Global Cancer Observatory 2022, approximately 9.7 million cancer-related deaths occurred in 2022 worldwide, with China accounting for 26.4% of these deaths.³ Over the past decade, cancer has consistently been the leading cause of death in China, marked by rising incidence, mortality, and disease burden.4, 5, 6 In 2022, approximately 4.82 million new cancer cases and 2.57 million cancer-related deaths were reported in China.⁷ Given the aging population and prevalent unhealthy lifestyles, cancer-related deaths are projected to continue to rise globally and in China, posing a severe threat to the lives and health of the population.^5,8,9

Premature mortality is independent of the age composition of the population and can be compared across different periods and regions. Therefore, the World Health Organization (WHO) recommends it as a critical indicator for evaluating the control of chronic diseases in countries.^10,11 In 2012, the 65th World Health Assembly proposed for the first time that premature mortality from four types of chronic diseases be reduced by 25% by 2025 compared with 2010.¹² Additionally, China has made great efforts in the prevention and control of cancer.^9,13 The Healthy China 2030 Plan Outline, issued in 2016, emphasizes that the National Health and Family Planning Commission of the People's Republic of China has expressed strong support for these goals and plans to reduce the absence of any intervention from major chronic diseases by 30% by 2030 compared to 2015.¹⁴ Meanwhile, the United Nations General Assembly released the 2030 Sustainable Development Goals (SDGs) agenda, which set a new target of reducing premature mortality from non-communicable diseases by at least a 33% reduction from 2015 levels.¹⁵

In order to achieve the SDG 3.4 targets, analyzing the changes in premature deaths from different types of cancer is essential, especially among different regions and provinces. Currently, premature mortality and life expectancy varies across regions and provinces, mainly due to differences in demographic risk factors resulting from socio-economic changes in China.^16,17 Multiple published studies have examined progress toward achieving the 2030 SDG of reducing premature mortality by 30% by 2030 relative to 2015 levels.18, 19, 20 One study investigated premature mortality from cancer in China by 2030 under different risk factor control scenarios.²¹ A recent study quantified the premature mortality from cancer across different countries and WHO regions and evaluated whether each region is progressing towards reaching SDG 3.4 targets by 2030.²² However, there is limited evidence on premature cancer mortality by region, province, and cancer type in China, which is crucial not only for understanding the current trends in different cancers across regions but also for further identifying areas for intervention to reduce cancer mortality in the population.

Therefore, we conducted this study to (1) quantify changes in cancer-specific premature mortality in China and its provinces from 1990 to 2023; (2) compare differences in premature mortality by region, province, and cancer type; (3) predict the expected premature mortality for individual cancer types and combined for 2030; and (4) assess the progress of each province toward the SDG 3.4 targets, to eventually provide a basis for the government to optimize cancer prevention and control in the context of limited health resources.

Methods

Data sources

We used China results from the Global Burden of Disease Study (GBD) 2023 as empirical data for projection. Age-, sex-, and year-specific deaths and population size for cancers between 1990 and 2023 were extracted. The Burden of Disease Study for China's Provinces was part of GBD 2023, a collaboration between the Chinese Center for Disease Control and Prevention (CDC) and the Center for Health Metrics and Evaluation at the University of Washington. GBD 2023 utilized the analytical and comparable framework from the GBD study to comprehensively estimate the population, morbidity, and mortality for 204 countries and territories. Standard GBD data processing procedures, the standard methods of mortality estimation, and special considerations for China have been described in detail in other publications.^5,23, 24, 25 The GBD 2023 estimates of China were obtained from several data sources: National Mortality Surveillance System (NMSS), Maternal and Child Surveillance data, Vital Registration, cancer registries, censuses, cause-of-death data from Macao Special Administration Region (SAR) and Hong Kong SAR, as well as other published studies or reports.^5,26,27

This analysis included all 33 provinces, comprising 31 mainland provinces, autonomous regions, municipalities, and the two special administrative regions of Hong Kong and Macao in China, excluding Taiwan Province. These province-level units are unified as provinces throughout this study. The 33 provinces were divided into seven regions according to standard classifications: East, North, South, Central, Northeast, Southwest, and Northwest, as detailed in Figure S1. Previous studies have also described.¹⁶

Mortality estimates

The GBD 2023 collaborators estimated incidence and mortality rates for both males and females across all age groups using modeling techniques. The cause-specific mortality for each combination of sex, age, location, and year was estimated by using the cause-of-death ensemble model (CODEm). CODEm incorporated four models to examine the relationships between the estimates (including linear mixed-effects regression models of the natural logarithm of cause-specific death rates, the logit of the causal fraction, spatiotemporal Gaussian process regression models of the natural logarithm of cause-specific death rates, as well as the logit of the causal fraction). An ensemble model was then constructed by weighting each model according to its predictive validity ranking. The mortality rate was calculated using the Mortality-to-Incidence Ratio (MIR) model, based on incidence data, in areas where mortality data is unavailable. For cancer estimates, 95% uncertainty intervals (UIs) were reported. The computation process stores the distribution of each step in 250 draws, which were then utilized in subsequent steps. These distributions were derived from the sampling error of data inputs, the uncertainty associated with model coefficients, and the uncertainty of severity distributions and disability weights. Uncertainty was propagated through each step of the estimation process, with UIs representing the 2.5th and 97.5th percentiles of the distribution of 250 draws at each step. The current GBD 2023 cancer mortality and MIR estimates have also been described in the GBD 2023 Causes of Death Collaborators publication.^28,29 We selected all cancers and 29 cancer groups based on the GBD 2023. The malignant neoplasms in this study were classified and coded using the International Classification of Diseases (ICD-10), with malignant neoplasms being coded as C00–C97.

Premature mortality

According to the World Health Organization's (WHO) definition, premature death was considered as death between the ages of 30 and 70, excluding the age of 70. Premature cancer mortality refers to the unconditional probability of dying from cancer between the ages of 30 and 70. We used the life table method recommended by the WHO to calculate the annual age-specific premature mortality in 5-year age groups for people aged 30–70 years.³⁰ The calculation steps are shown below. The first step was to calculate the mortality rate for cancer in the 5-year age groups:

$${}_5{}^{\ast }{M}_x=\frac{Totaldeathsfromcancercausesbetweenexactagexandexactagex+5}{Totalpopulationbetweenexactagexandexactagex+5}$$

The second step calculated the probability of mortality from cancer for each 5-year age group. ${}_5{}^{\ast }{q}_x$ represents the probability that a cohort born at the same time will die in the age group from x to (x + 5) years. Based on the relationship between ${}_5{}^{\ast }{q}_x$ and ${}_5{}^{\ast }{M}_x$, the probability of death for the 5-year age group was calculated according to the following formula:

$${}_5{}^{\ast }{q}_x=\frac{{}_5{}^{\ast }{M}_x\ast 5}{1+{}_5{}^{\ast }{M}_x\ast 2.5}$$

Third, the probability of death from age 30–70, independent of other causes of death can be calculated as:

$${}_{40}{}^{\ast }{q}_{\mathrm{30}}=1-\prod _{x=30}^{65}\left(1-{}_5{}^{\ast }{q}_x\right)$$

Projected premature mortality from cancer by 2030

To assess whether historical trends over the period 1990–2023 would meet the SDG 3.4 targets, we calculated the premature mortality reduction target values and rates of change needed to meet the targets for 29 cancer types and for all cancers combined. We hypothesized that the simplest scenario, absent any intervention, would be to reduce mortality from each cancer and the other major NCDs at the same rate.

First, we measure the magnitude of change in the premature mortality from cancer from 1990 to 2023 using the average annual rate of change, which is calculated as:

$$Averageannualrate=\sqrt[n]{\frac{a_0}{a_n}}-1$$

where $a_0$ is the premature mortality in the base period; $a_n$ is the premature mortality in the reporting period. Then, the target for 2030 was set at two-thirds of the 2015 premature mortality, representing the required one-third reduction.¹⁶ Finally, we projected the premature mortality in 2030 on the basis of the average rate of change from 1990 to 2023.

To quantify the additional efforts required to achieve SDG 3.4 targets, we calculated the following ratio: the required average annual rate of change in premature cancer mortality for 2015–2030 divided by the average annual rate of change based on historical trends for 1990–2023. If the rate ratio is greater than 0 and less than 1, it indicates that the current downward trend is likely to meet the SDG 3.4 targets. We further calculate the expected years to achieve the goal of reducing cancer-related premature mortality. Based on the continuation of the trend from 1990 to 2023, we calculated the expected year by which premature cancer mortality would achieve a one-third reduction compared with 2015 levels.

Statistical analysis

Joinpoint Regression analysis was used to described the direction and magnitude of general trend. The Monte Carlo permutation test was used to select the optimal number of joinpoints. In our analysis, we used the number of connection points recommended by the software for analysis. By selecting the log-linear model: lny = xb, we conducted Joinpoint Regression analysis to calculate the annual percent change (APC) with its 95% confidence interval (CI) to depict the trend over the delineated timeframe. Then, the average annual percent change (AAPC) and 95% confidence intervals (CIs) calculated based on the trend of the APCs.³¹ All analyses were performed using R (v4.4.0, https://www.r-project.org/) and Joinpoint Regression Program (Version 4.9.0.1, National Cancer Institute, Rockville, MD, US).

Ethical approval

As this study utilized publicly available data containing no confidential or personally identifiable patient information, no ethical approval was required.

Role of the funding source

This study received no external funding.

Results

Premature mortality from cancer in China, 1990–2023

Between 1990 and 2023, the number of total cancers premature deaths increased by 14.77% to 1.26 million in both sexes. Cancer deaths and proportion of premature deaths in males were consistently higher than that in females, both showing a downward trend from 1990 to 2023 (Figure S2). The number of premature deaths and premature mortality in the cancer types from 1990 to 2023 were shown in Table 1. In 2023, approximately 0.35 million premature deaths were caused by lung cancer, with premature mortality of 1.97% in China. Lung cancer has the highest premature mortality, followed by stomach cancer (0.98%), Colon and rectum cancer (0.65%) and esophageal cancer (0.65%) in both sexes. From 1990 to 2023, premature mortality significantly declined for most cancer subcategories. The most pronounced decreases were observed in Hodgkin's lymphoma (AAPC = −4.89%, 95% CI = −5.12% to −4.66%), followed by stomach cancer (AAPC = −3.80%, 95% CI = −4.18% to −3.42%) and nasopharynx cancer (AAPC = −3.55%, 95% CI = −4.03% to −3.07%). Conversely, a few cancers, such as multiple myeloma (AAPC = 2.44%, 95% CI = 1.38%–3.51%) showed an increasing trend in premature mortality (Table 1).

Table 1.

Deaths number, premature deaths number and premature mortality (%) for cancer groups for both sexes in China, 1990–2023.

Cancer type	1990			2023			AAPC in Premature mortality, 1990–2023 (95% CI)
	Deaths, thousands (95% UI)	Premature deaths, thousands (95% UI)	Premature mortality (%)	Deaths, thousands (95% UI)	Premature deaths, thousands (95% UI)	Premature mortality (%)
Total cancers	1708.42 (1529.52, 1984.88)	1094.06 (941.38, 1297.50)	13.11 (11.4, 15.36)	2563.97 (2227.03, 2876.75)	1255.65 (1009.21, 1537.37)	6.66 (5.38, 8.09)	−1.93 (−2.20, −1.66)a
Tracheal, bronchus, and lung cancer	308.91 (262.23, 397.22)	200.02 (164.47, 260.02)	2.66 (2.20, 3.47)	766.18 (659.96, 887.56)	352.60 (271.83, 443.83)	1.97 (1.51, 2.46)	−0.81 (−1.16, −0.45)a
Stomach cancer	428.07 (348.68, 530.54)	280.63 (218.19, 351.34)	3.63 (2.85, 4.52)	381.50 (301.10, 479.84)	178.49 (130.47, 246.72)	0.98 (0.72, 1.36)	−3.80 (−4.18, −3.42)a
Colon and rectum cancer	133.56 (113.85, 155.44)	83.17 (67.96, 99.55)	1.03 (0.85, 1.23)	258.09 (225.36, 305.79)	119.12 (93.75, 149.55)	0.65 (0.51, 0.81)	−1.28 (−1.78, −0.77)a
Esophageal cancer	237.74 (170.01, 279.10)	154.94 (109.61, 187.67)	2.10 (1.50, 2.53)	254.46 (217.04, 293.12)	115.93 (89.25, 149.31)	0.65 (0.50, 0.83)	−3.40 (−3.75, −3.06)a
Liver cancer	92.33 (80.18, 108.03)	69.28 (57.30, 84.65)	0.82 (0.68, 1.00)	143.50 (126.37, 163.70)	91.71 (70.66, 117.39)	0.48 (0.37, 0.61)	−1.52 (−2.18, −0.86)a
Pancreatic cancer	42.73 (36.79, 49.93)	28.53 (23.74, 34.35)	0.37 (0.31, 0.44)	115.40 (101.50, 132.27)	59.48 (46.96, 73.23)	0.33 (0.26, 0.40)	−0.27 (−0.95, 0.42)
Breast cancer	50.96 (41.92, 63.60)	40.74 (31.39, 52.62)	0.45 (0.35, 0.58)	80.22 (67.69, 94.92)	54.80 (40.30, 73.17)	0.28 (0.21, 0.38)	−1.32 (−1.73, −0.91)a
Brain and central nervous system cancer	40.65 (28.80, 50.17)	21.68 (15.24, 27.64)	0.25 (0.18, 0.32)	62.88 (49.43, 83.39)	35.20 (26.51, 48.88)	0.19 (0.14, 0.26)	−0.68 (−0.83, −0.53)a
Cervical cancer	41.08 (30.15, 59.13)	31.44 (22.18, 46.43)	0.36 (0.25, 0.52)	47.87 (30.05, 62.79)	31.91 (18.16, 46.40)	0.16 (0.09, 0.24)	−2.25 (−2.55, −1.95)a
Leukemia	77.70 (61.80, 90.77)	29.71 (23.41, 36.72)	0.31 (0.25, 0.39)	56.03 (49.15, 67.98)	29.30 (22.83, 38.43)	0.15 (0.12, 0.20)	−1.98 (−2.10, −1.86)a
Non-Hodgkin lymphoma	25.92 (21.87, 30.85)	15.17 (12.42, 18.54)	0.18 (0.15, 0.22)	39.90 (33.97, 46.64)	21.76 (17.03, 28.11)	0.12 (0.09, 0.15)	−1.24 (−1.43, −1.05)a
Nasopharynx cancer	38.01 (30.47, 43.82)	28.48 (22.35, 34.40)	0.33 (0.26, 0.40)	27.11 (23.76, 32.18)	18.87 (14.46, 24.38)	0.10 (0.08, 0.13)	−3.55 (−4.03, −3.07)a
Other malignant neoplasms	38.94 (32.21, 46.42)	20.53 (16.08, 25.46)	0.25 (0.20, 0.32)	36.02 (28.91, 43.64)	16.78 (12.62, 21.94)	0.09 (0.07, 0.12)	−2.93 (−3.39, −2.47)a
Gallbladder and biliary tract cancer	18.50 (14.57, 24.14)	10.83 (8.47, 14.13)	0.15 (0.11, 0.19)	33.95 (27.35, 44.92)	14.42 (10.74, 20.05)	0.08 (0.06, 0.11)	−1.71 (−2.15,−1.26)a
Ovarian cancer	14.00 (10.41, 18.31)	10.40 (7.41, 13.88)	0.12 (0.09, 0.16)	23.71 (19.08, 28.80)	15.76 (11.00, 21.19)	0.08 (0.06, 0.11)	−1.14 (−1.56, −0.71)a
Bladder cancer	23.48 (19.40, 28.23)	11.43 (9.11, 14.25)	0.16 (0.13, 0.19)	44.80 (38.37, 50.82)	13.23 (10.05, 17.03)	0.08 (0.06, 0.10)	−2.10 (−2.48, −1.72)a
Lip and oral cavity cancer	11.57 (9.75, 14.04)	7.67 (6.19, 9.51)	0.10 (0.08, 0.12)	21.48 (18.06, 25.87)	11.44 (8.61, 14.88)	0.06 (0.05, 0.08)	−1.27 (−1.70, −0.83)a
Kidney cancer	8.66 (6.73, 11.44)	4.97 (3.71, 7.18)	0.06 (0.05, 0.09)	21.88 (17.00, 29.27)	10.79 (7.75, 15.07)	0.06 (0.04, 0.08)	−0.06 (−0.33, 0.21)
Larynx cancer	13.09 (9.66, 16.66)	8.44 (5.68, 11.11)	0.11 (0.08, 0.15)	20.44 (15.80, 25.05)	10.79 (7.58, 14.70)	0.06 (0.04, 0.08)	−1.91 (−2.48, −1.33)a
Uterine cancer	13.94 (8.37, 18.26)	10.38 (6.07, 14.30)	0.12 (0.07, 0.17)	12.99 (9.77, 19.90)	8.10 (5.23, 12.99)	0.04 (0.03, 0.07)	−3.09 (−3.36, −2.83)a
Prostate cancer	12.24 (9.24, 16.79)	4.49 (3.19, 6.24)	0.07 (0.05, 0.09)	36.82 (25.59, 46.00)	7.09 (4.90, 9.73)	0.04 (0.03, 0.06)	−1.31 (−1.81, −0.82)a
Multiple myeloma	2.27 (1.48, 4.02)	1.44 (0.91, 2.64)	0.02 (0.01, 0.03)	12.00 (9.09, 14.78)	6.61 (4.56, 8.60)	0.04 (0.03, 0.05)	2.44 (1.38, 3.51)a
Non-melanoma skin cancer	5.59 (4.22, 7.27)	3.14 (2.31, 4.16)	0.04 (0.03, 0.06)	14.96 (10.84, 19.26)	5.40 (3.74, 7.32)	0.03 (0.02, 0.04)	−0.89 (−1.32, −0.46)a
Other pharynx cancer	4.50 (2.90, 5.90)	3.21 (1.94, 4.33)	0.04 (0.03, 0.06)	5.82 (4.73, 7.05)	3.42 (2.50, 4.59)	0.018 (0.013, 0.025)	−2.42 (−2.96, −1.88)a
Malignant skin melanoma	2.83 (1.85, 3.86)	1.88 (1.20, 2.60)	0.02 (0.01, 0.03)	5.85 (4.21, 8.50)	3.06 (2.06, 4.50)	0.016 (0.010, 0.023)	−0.82 (−1.34, −0.29)a
Thyroid cancer	4.00 (3.27, 4.79)	2.24 (1.81, 2.79)	0.028 (0.022, 0.035)	6.44 (4.98, 7.71)	2.63 (1.93, 3.45)	0.014 (0.010, 0.020)	−1.99 (−2.56, −1.42)a
Mesothelioma	1.14 (0.86, 1.52)	0.78 (0.58, 1.08)	0.009 (0.007, 0.013)	2.79 (2.08, 3.52)	1.63 (1.16, 2.22)	0.008 (0.005, 0.010)	−0.17 (−0.60, 0.27)
Hodgkin lymphoma	5.09 (2.33, 6.96)	2.89 (1.26, 4.11)	0.03 (0.02, 0.05)	2.24 (1.49, 2.96)	1.20 (0.79, 1.69)	0.006 (0.004, 0.009)	−4.89 (−5.12, −4.66)a
Testicular cancer	1.23 (0.95, 1.52)	0.63 (0.47, 0.81)	0.006 (0.004, 0.008)	1.07 (0.85, 1.33)	0.53 (0.38, 0.72)	0.003 (0.001, 0.004)	−2.33 (−2.87, −1.79)a

UI: uncertainty interval.

^aThe AAPC is significantly different from zero (P < 0.05).

There were also significant differences in the deaths and premature deaths of common cancers between males and females. In males, lung cancer recorded the highest premature mortality (2.88%) with 0.26 million premature deaths, accounting for 30.31% of all male cancer-related premature deaths. This was followed by stomach cancer (1.52%) with 0.14 million premature deaths, accounting for 16.22% of all male cancer-related premature deaths, and esophageal cancer (1.11%) with 0.10 million premature deaths, which accounted for 11.78% of total fatalities in 2023 (Table S1 and Figure S3a). In females, lung cancer also had the highest premature mortality (1.05%) with 0.09 million premature deaths, accounting for 23.40% of all female cancer-related premature deaths, followed by breast cancer (0.54%) with 0.05 million premature deaths, accounting for 12.92% of cancer-related premature deaths, and stomach cancer (0.45%) with 0.04 million premature deaths, which accounted for 10.01% of total premature deaths associated with cancer in females. The premature mortality trends of pancreatic cancer, mesothelioma and breast cancer were inversely correlated between sexes. The premature mortality of mesothelioma (AAPC = −0.56%) and breast cancer (AAPC = −1.46%) showed significant decreases in females (Table S2 and Figure S3b).

Premature cancer mortality by region

The number of deaths, premature deaths and premature mortality of the cancers in seven regions were analyzed. In 2023, cancer caused approximately 2.56 million deaths, including 1.26 million premature deaths. The premature mortality due to cancer was 6.66%, with males (8.95%) having a higher premature mortality compared to females (4.32%). The eastern of China recorded the highest number of premature deaths from cancer, totaling around 0.37 million deaths, accounting for 29.60% of all premature deaths in China. The provinces with higher premature mortality were mainly located in the northeast and north of China (Table 2). For males, the five provinces with the highest premature mortality were Heilongjiang (11.44%), Chongqing (11.15%), Sichuan (10.49%), Liaoning (10.26%) and Shandong (10.00%), each exceeding 10.00%. For females, the premature mortality ranged from 2.75% in Hong Kong to 6.63% in Heilongjiang, with the top five provinces being Heilongjiang (6.63%), Jilin (5.63%), Liaoning (5.44%), Xinjiang (5.37%) and Inner Mongolia (4.98%) (Fig. 1).

Table 2.

Deaths number, premature deaths number and premature mortality of cancer in China by region, 2023.

Region	Both			Male			Female
	Deaths, thousands (95% UI)	Premature deaths, thousands (95% UI)	Premature mortality (%)	Deaths, thousands (95% UI)	Premature deaths, thousands (95% UI)	Premature mortality (%)	Deaths, thousands (95% UI)	Premature deaths, thousands (95% UI)	Premature mortality (%)
North	329.83 (281.03, 389.83)	161.25 (118.09, 213.78)	6.81 (5.03, 8.93)	207.87 (170.85, 253.25)	106.31 (69.86, 154.6)	8.95 (5.98, 12.75)	121.96 (97.61, 154.86)	54.94 (36.07, 81.10)	4.65 (3.07, 6.79)
Northeast	257.63 (215.36, 303.99)	136.47 (99.65, 182.75)	8.18 (6.05, 10.81)	155.59 (122.24, 193.69)	87.36 (57.06, 127.31)	10.55 (7.03, 15.00)	102.04 (82.63, 126.16)	49.11 (31.89, 71.93)	5.86 (3.85, 8.47)
East	839.16 (704.51, 968.43)	371.62 (270.10, 498.23)	6.50 (4.77, 8.62)	554.25 (453.65, 670.28)	258.81 (169.30, 375.5)	8.99 (5.97, 12.75)	284.90 (225.83, 343.47)	112.81 (73.26, 163.43)	3.97 (2.59, 5.70)
South	245.04 (215.39, 281.9)	139.36 (102.39, 187.72)	6.19 (4.58, 8.27)	165.77 (139.10, 194.86)	98.27 (65.66, 141.63)	8.58 (5.83, 12.13)	79.28 (64.40, 96.17)	41.10 (26.95, 59.45)	3.70 (2.43, 5.30)
Central	383.74 (327.15, 451.43)	187.26 (136.77, 251.50)	6.37 (4.69, 8.46)	246.74 (201.54, 297.58)	125.17 (81.44, 183.78)	8.53 (5.64, 12.27)	137.01 (112.47, 173.36)	62.09 (40.54, 90.76)	4.21 (2.77, 6.09)
Northwest	144.20 (123.16, 170.04)	76.56 (55.39, 103.34)	5.92 (4.30, 7.91)	89.61 (72.95, 110.91)	48.58 (31.89, 72.28)	7.44 (4.95, 10.87)	54.59 (43.82, 68.24)	27.98 (18.03, 41.12)	4.34 (2.82, 6.31)
Southwest	364.37 (307.18, 431.79)	183.13 (133.20, 247.08)	6.80 (4.98, 9.07)	239.89 (189.77, 294.23)	125.73 (82.53, 183.84)	9.22 (6.14, 13.21)	124.48 (99.79, 156.59)	57.40 (37.67, 84.47)	4.29 (2.84, 6.25)
China	2563.97 (2227.03, 2876.75)	1255.65 (1009.21, 1537.37)	6.66 (5.38, 8.09)	1659.71 (1453.24, 1905.46)	850.22 (636.89, 1097.85)	8.95 (6.78, 11.39)	904.27 (750.28, 1062.12)	405.43 (303.60, 532.02)	4.32 (3.25, 5.63)

UI: uncertainty interval.

Premature mortality is the probability of dying between ages 30 and 69 years from cancer that was calculated using life table method.

North: Beijing, Tianjin, Hebei, Inner Mongolia, Shanxi; Northeast: Heilongjiang, Jilin, Liaoning; East: Anhui, Fujian, Jiangsu, Jiangxi, Shandong, Shanghai, Zhejiang; South: Guangdong, Guangxi, Hainan, Hong Kong, Macao; Central: Hubei, Hunan, Henan; Northwest: Ningxia, Qinghai, Shaanxi, Xinjiang, Gansu; Southwest: Chongqing, Guizhou, Sichuan, Yunnan, Tibet; excluding Taiwan.

Fig. 1.

Rank of premature mortality for cancer groups in male (a) and female (b) by province in China, 2023.

The ranking of cancer types by premature mortality in China and its 33 provinces was also analyzed. In 2023, the three cancers with the highest premature mortality for males were lung cancer, stomach cancer, and esophageal cancer (Fig. 1a). For females, lung cancer, breast cancer, and stomach cancer were the top three cancers that caused premature death (Fig. 1b). Lung cancer continued to be the leading cause of premature mortality in most provinces for both sexes. Stomach cancer was the foremost cause of premature cancer-related mortality for men in several northwestern provinces, such as Qinghai, Gansu, Ningxia, and Tibet. The highest premature mortality in Tibet, Gansu and Qinghai Province was stomach cancer among both males and females.

Annual rate of change in premature cancer mortality varied considerably across regions. Premature mortality of total cancer rates declined in China and across all regions from 1990 to 2023 (Fig. 2). The largest decreases of premature mortality were observed in the eastern, central, and southwestern regions. For 27 of the 29 cancers, premature mortality was declining in all regions. However, multiple myeloma had rising premature mortality rates in all regions. The premature mortality of mesothelioma has been trending downwards across regions, except in the northern China. Stomach cancer, esophageal cancer, nasopharynx cancer, uterine cancer, liver cancer and Hodgkin's lymphoma exhibited significant declining trends across all regions in China.

Fig. 2.

Annual rate of change in premature mortality for cancer by region, 1990–2023.

Sex differences were observed in the annual rate of change for premature cancer mortality. Premature mortality of pancreatic, breast and kidney cancers had opposite trends in males and females. The premature mortality from breast cancer showed an increasing trend between 1990 and 2023 in males, except in the southwestern region. For pancreatic cancer, the premature mortality among males in the north, northeast, and northwest regions showed an increasing trend. The premature mortality of kidney cancer was regionally increasing among males, mainly concentrated in the northern, northeastern and southwestern regions of China (Figures S4 and S5).

Predicting the premature mortality for cancers by 2030

Progress towards SDG 3.4 targets were shown in Figure S6. The premature mortality is projected to reach the target around 2035, at the average annual rate of decline from 1990 to 2023. The southwestern, eastern and central regions are projected to reach the target earlier than other regions, by approximately 2035. Fig. 3 shows the year when a 30% reduction, relative to 2015 levels, is expected to be achieved if the average rate of decline from 1990 to 2023 continues, by province and sex. Based on the current downward trend, females are more likely than males to achieve the SDG 3.4 targets (Fig. 3b).

Fig. 3.

Expected years of achieving targets for reducing cancer-related premature mortality. (a) Male and female. (b) Male. (c) Female. The color scale indicates the years in which provinces are expected to achieve one-third reduction (relative to 2015 levels, if trends from 1990 to 2023 continue) in the premature mortality from cancer. The study was conducted in mainland China. Taiwan was not included. Base map from the National Platform for Common Geospatial Information Services (https://www.tianditu.gov.cn/).

Premature mortality from cancer is projected to trend downward until 2030. For China to attain SDG 3.4 targets, the annual rate of reduction in premature cancer mortality must be 1.3 times faster than the historical trend. In the absence of intervention, the premature mortality from all cancers in Hong Kong, Hubei, and Shanghai is likely to reach the SDG 3.4 targets at the previous rate of decline. For specific cancer types, the premature mortality of stomach cancer, esophageal cancer, nasopharynx cancer and uterine cancer can achieve the goals in China. Cervical cancer is on track to reach the targets in southwestern regions, including four provinces. For liver cancer, only four provinces will reach 2030 targets. To meet targets for liver cancer, the annual decline rate of premature cancer mortality must reach 1.64 times the historical trend (Fig. 4).

Fig. 4.

The additional multiplier in cancer-related premature mortality reduction required beyond current trends to achieve SDG 3.4 targets, by province and region in China. The additional effort required to achieve SDG 3.4 targets was quantified by calculating the ratio of the target annual rate of change in premature cancer mortality (2015–2030) to the annual rate of change based on historical trends (1990–2023). Red: shows an increase in premature mortality by 2030. Green: suggests the additional reduction in premature cancer mortality required to achieve the targets (relative to current trends). Blue: means that the current downward trend is likely to meet the targets (rate ratio is greater than 0 and less than 1).

Achieving the SDG 3.4 target for premature cancer mortality in men requires an acceleration of the annual decline rate to 1.56 times that of the historical trend. It is projected that the premature mortality from stomach, esophageal, and nasopharynx cancer among men in China will meet the SDG 3.4 targets. At the current rate of decline, Shanghai is likely to achieve this target among males. Meeting the SDG 3.4 targets for premature lung cancer mortality in male will require an annual rate of decline 3.74 times faster than the historical trend. The premature mortality of from pancreatic, kidney, and breast cancer, and Multiple myeloma among males in northern, northeastern, and northwestern China will increase by 2030. To achieve SDG 3.4 for premature cancer mortality in women, the annual rate of decline must reach 1.02 times the historical trend. This includes eight provinces that are likely to meet the target if current trends are maintained. Among females, the premature mortality of stomach cancer, esophageal cancer, nasopharynx cancer, liver cancer, bladder cancer, leukemia, and uterine cancer is on track to reach the targets by 2030. However, the premature mortality from pancreatic cancer, breast cancer, lung cancer, kidney cancer, and Multiple myeloma are expected to increase in some provinces by 2030. Notably, the mortality multiple myeloma is projected to increase by 2030 in both sexes (Figures S7 and S8).

Discussion

In this study, we assessed premature mortality by cancer type, region, and province in China. Lung cancer remained the leading cause of premature deaths among all cancers. The findings suggest that the current trend of premature mortality from cancer is expected to continue to decline through 2030, this pace of reduction will be insufficient to meet the SDG 3.4 targets, necessitating an acceleration to 1.3 times the historical trend. Premature mortality from stomach cancer, esophageal cancer, nasopharynx cancer, and uterine cancer is projected to meet the targets in both sexes. There are still discrepancies in premature cancer mortality across various regions, with higher premature mortality in the northeast, northern and the southwest.

While the number of premature deaths from cancer increased between 1990 and 2023, the premature mortality was on a downward trend in China. The premature mortality from cancer was declining globally, especially in countries in high-income regions.²² An Australian study expects the observed downward trend in age-standardized premature mortality rates to continue over the period 2020–2044.³² Recent studies have also suggested that the age-standardized incidence of cancer in China has been on the rise in the last decade or so, while the age-standardized mortality rate has declined.³³ The results indicate that China has achieved notable progress in the early diagnosis and treatment of cancer over the past decade. The decreasing premature mortality can be also explained by the changes in risk factors. China still faces high smoking rates. One study reported that between 2007 and 2018, smoking prevalence among men decreased from 58.4% to 50.8% and among women from 2.2% to 1.9% in China.³⁴ Tobacco control can prevent cancer-related deaths effectively, however, despite some progress in tobacco control efforts.35, 36, 37

In recent years, the Chinese government has issued a series of comprehensive guidelines and actions in an effort to reduce deaths caused by risk factors.38, 39, 40, 41, 42 Moreover, developments in early screening, molecular diagnostic techniques, and therapeutic modalities have contributed to the reduction in premature death. The overall decline in premature cancer mortality can be attributed in part to significant reductions in stomach and esophageal cancers. Recently, the incidence and mortality rates showed a noticeable decline in stomach cancer worldwide.⁴³ The mortality rates have declined as a result of Helicobacter pylori eradication treatment, early screening, and improved food hygiene.^44,45 These measures have significantly contributed to reducing mortality from stomach cancer. We observed a downward trend in the premature mortality for breast or cervical cancer in females, which may be attributed to the national “two cancers” screening program.^46,47 The full coverage of hepatitis B vaccination played an important role in reducing the liver cancer burden.⁴⁸ It is worth noting that, despite observing a decreasing trend in the proportion of premature deaths, this trend was intimately linked to the aging of the population and the rise in life expectancy, rather than being a direct consequence of enhanced health.

An upwards trend in premature mortality for some cancer types was observed in our study. Among the top 10 tumors with the highest premature mortality, pancreatic cancer showed an increasing trend in males. Due to the lack of typical clinical manifestations and effective diagnostic methods, most of pancreatic cancer patients are in the advanced stage when they are diagnosed, with poor therapeutic effects and a survival rate of only about 10%.⁴⁹ In recent years, the incidence of pancreatic cancer has been rising worldwide, particularly among younger individuals.^3,50 Furthermore, several neoplasms have a low premature mortality, which should not be ignored. For instance, multiple myeloma, mesothelioma and kidney cancer are projected to increase by 2030. The increasing prevalence of obesity, diabetes, and alcohol consumption has been suggested as potential contributing factors to the rise in young-onset pancreatic cancer.⁵¹ To achieve the SDG 3.4 target for premature cancer mortality, the intensive interventions must be tailored to different cancers.

Despite the fact that a series of preventive and control measures could yield reductions for premature mortality, especially in the eastern region, we observe there would still exist a gap in premature mortality among different geographic groups due to geographic differences in environmental factors and lifestyle habits. Our study shows that the provinces with the highest premature mortality were predominately located in the northeast and north, while the provinces with the lowest premature mortality were concentrated in the more economically developed regions in the eastern and southern regions. For example, the province with the highest premature mortality (Heilongjiang) was 1.98 times higher than the lowest province (Hong Kong). The disparity was even greater among women, with a ratio of 2.41 between the highest and lowest premature mortality. Previous studies have shown that the provinces with lower life expectancy are mainly in the southwest and northwest regions, where economic development is less developed.⁵² For instance, low-dose CT lung screening for high-risk individuals reduces lung cancer mortality⁵³; however, screening rates may be lower in economically underdeveloped areas due to high economic costs, which explains the faster decline in premature mortality from lung cancer in the eastern region in China. Lower socioeconomic status and poorer health literacy were also associated with lower levels of fruit consumption and higher prevalence of H. pylori and hepatitis B virus in China.^54,55 Efforts to strengthen health systems are focused on addressing the social determinants of health access to and utilization of cancer services and the capacity of health systems. Addressing health inequalities facilitates the prioritization of allocation and guaranteeing equitable and effective healthcare services to people.

Sex is a key determinant of health behaviors and outcomes, resulting in disparities in premature mortality between males and females. Women are projected to meet the SDG 3.4 targets for reduced cancer premature mortality earlier than men. Several studies have shown that the mortality rates were higher in male than in female.^7,56 This may be explained by the higher prevalence of smoking, harmful drinking, obesity, and other risk factors in males than in females.^9,57, 58, 59 Meanwhile, the premature mortality from breast cancer, kidney cancer, and pancreatic cancer shows opposite trends in males and females. A study reporting on urban-rural differences in the burden of cancer in China from 2008 to 2020 showed that the burden of thyroid, kidney, and bladder cancers was higher in rural areas than in urban areas for male.³³ In highly developed countries in Europe and the United States, a downward trend in breast cancer mortality was observed from 2000.⁶⁰ Notably, there was an increase in the premature mortality from breast cancer among male. Since multiple unsolved questions remain regarding the causes, treatment, and optimal care of male breast cancer, further research is necessary to elucidate the biological basis, to assess the risks of specific treatments, and to address the need to improve the quality of life for men with breast cancer. These findings suggest that there were some public health issues that need continued or additional attention, especially in terms of risk factor interventions and resource allocation needs to be based on different cancer risk populations.

The total deaths from non-communicable diseases in the population aged 30–69 years reached 12.9 million in 2017.⁶¹ Cancer remains the leading cause of premature death.⁶¹ Without scaling up efforts to prevent and control non-communicable diseases, most countries globally will likely fail to achieve the SDG targets.⁶² The Global Burden of Disease (GBD) 2023 study indicated that the probability of dying due to cancer among individuals aged 30–70 years worldwide is expected to decline by 6.5% between 2015 and 2030.²⁵ In scenarios where targets for certain risk factor interventions are met, the premature deaths in China could be significantly reduced by 2030, meeting the target of reducing premature mortality from non-communicable diseases by one-third, although the reduction in the premature death from cancer is still far from the targets.⁶³ A study based on GBD 2019 reported that China could avoid 710,400 cancer deaths in 2030 if all risk factors (smoking, physical inactivity, high BMI, PM2.5 exposure, and high-salt diet) were controlled to meet the targets of the WHO's Global Surveillance Framework for NCDs program.²¹ Future studies should incorporate risk factors for cancers such as inadequate intake of fruits and vegetables, alcohol consumption, infectious agent (H. pylori, Hepatitis B virus, Hepatitis C virus, Epstein–Barr virus), as well as cancer screening and vaccination, to further explore premature cancer mortality in different regions of China under various risk factor control scenarios. Population growth and aging are also posing huge challenges to achieve the SDG 3.4 targets. Changes in the definition of premature death (30–69) as expectations lengthen may affect how we measure health progress and policy effects. China's health goals may need to be adjusted in response to demographic changes. For example, while declining premature mortality indicates improved health, the burden of chronic disease will become more pronounced with the advent of an aging society. We therefore recommend that future research continue to focus on deaths in the 70 and over group and take this into account in target setting and policy implementation.

Nonetheless, based on past rates of progress, Hong Kong, Shanghai and Hubei are currently on track to achieve the SDG 3.4 targets for total cancers, which are expected to be reached by 2030. Our study, which is grounded in historical rates of change, offers a mechanism to assess the ambitiousness of targets. It is recognized that the premature mortality from cancer without intervention is still short of the target. To attain the 2030 target benchmarks, most cancer types will need to achieve a rate of change that exceeds that of the past. This is particularly true for pancreatic cancer, kidney cancer, multiple myeloma and mesothelioma, which are still on the rise in most provinces. To bring such goals closer to reality for all populations, our government and other health institutions will need to provide technical leadership and financial support, especially for northwest, north and southwest regions with the lowest performance on goals.

This study predicts premature mortality from cancer in China for 2030 based on GBD 2023, but there are still some limitations. First, the accuracy of the GBD estimates depends to a large extent on the quality and quantity of the data used, and mortality rates for several cancers in the provinces may be underestimated due to under-reporting and estimation from the 24% of the population under surveillance. Adjusted mortality rates were reported in the regular surveys conducted by the national monitoring system to minimize under-reporting bias. Also, the iterative multilevel stratified approach was used to ensure the representativeness of the data at the national and provincial levels. We observed some differences between the GBD and GLOBOCAN estimates. There were even significant discrepancies between the same cancer in the same country and between different databases. For example, according to GLOBOCAN estimates, there were 866,136 new cases of liver cancer and 758,725 deaths of liver in 2022.³ However, there were 529,202 and 483,875 new and death cases in 2021, respectively, based on GBD 2021.⁶⁴ Differences in the methods of estimating cancer incidence and mortality between the two databases may have contributed to these differences. In GLOBOCAN, cancer incidence was first estimated, followed by mortality estimation for each country through survival modeling. For countries with national mortality data but no national incidence data, cancer incidence was estimated from national mortality using statistical models. The mortality to incidence ratios (MIR) were derived from cancer registry data within the country. Some studies also have emphasized that there were discrepancies between the GLOBOCAN and GBD estimates in certain countries and regions.65, 66, 67 The GBD study was advantageous because it provides a comprehensive, longitudinal trend analysis of global health, allowing for comparisons across countries and regions. Of course, the interpretation of the results should be done with caution. Second, there were variations in the quality of coding across provinces, with relatively poorer coding quality in areas with less economic development. We redistributed garbage codes to maximize the accuracy and completeness of data on cancer-related causes of death. Thirdly, the trend of increasing cancer incidence and deaths was also associated with a significant increase in cancer registrations compared to the past, leading to a misinterpretation of this trend.

In conclusion, we assessed the premature mortality from cancer in China by sex, province, region, and cancer type. Our findings underscored the need to prioritize more effective programs, interventions, and policies tailored to specific cancer types and regions for policymakers, governments, and public health professionals. The Chinese government should promote cost-effective cancer screening and immunization programs, strengthen the management of risk factors, and commit to raising awareness of healthy lifestyles through education to reduce premature mortality rates nationwide.

Contributors

HS and MZ conceived the ideas for this research and gave final approval of the version to be published. BZ did statistical analyses and drafted the first version of the paper. PY, LW, JL, LL and JY accessed and verified the data. MZ and ZZ revised the manuscript critically. All authors accessed and verified the underlying data, contributed to data interpretation, and approved the final manuscript.

Data sharing statement

Data for this study were from the GBD, to download the data please visit the Global Health Data Exchange GBD 2023 website. Province-level data are not publicly available due to data sharing restrictions imposed by the China CDC. However, these data may be obtained from the MZ, upon reasonable request.

Editor note

The Lancet Group takes a neutral position with respect to territorial claims in published maps and institutional affiliations. Map sources are provided in the figure legends.

Declaration of interests

We declare no competing interests.