Secular trends in Gastrointestinal cancers mortality across China, Japan, the U.S., and India: An Age-Period-Cohort Analysis and Join- point

Abstract

Objective: Colon cancer, esophageal cancer, and stomach cancer are the common cause of morbidity and mortality in China, Japan, the U.S. and India. The current study was aimed to assess and compare secular trends of Gastrointestinal cancers mortality during the period (1990-2017) in age-specific, time period and birth cohort effects.

Method: We used Joinpoint model to collect age-standardized mortality rates for four countries. And we designed an age-period-cohort (APC) analysis to estimate the independent effects on mortality of three cancers.

Result: Joinpoint model shows that in addition to esophageal cancer death rate in Japan, the age-standardized death rate (ASDR) of esophageal cancer and stomach cancer in other countries declined rapidly. APC analysis presented a similar pattern of age effect between four countries for colon cancer and stomach cancer, which increased from 20-89 age groups. Differently, the period effect rapidly increased for esophageal cancer and stomach cancer in U.S., the period effect of China presented a declining volatility, showing its highest value in 2007. The future forecast highest mortality trends were found in China.

Conclusion: Therefore, the obvious increase in colon cancer recommended that earlier tactics must be performed to reduce the mortality from specific cause from 2018 to 2027.

Background

Colon cancer, esophageal cancer, and stomach cancer are the most common gastrointestinal cancers ¹. Not only are they one of the most common malignant tumors, but also they are usually identified as the main cause of cancer death ². The longitudinal trends of these three cancers may be affected by many risk factors, such as dietary structure, lifestyle changes, and the implementation of screening ³. The monitoring data mortality rate was reported was 6.9/100,000 in the 2017. According to the GBD data, the DALYs caused by colon cancer in China were 4.254 million per year (doubled from 1990), accounting for 22.4% of the global burden of colon cancer in 2017 ⁴.

Although China's population accounts for nearly 20% of the world's population, it accounts for half of the world's total deaths from esophageal cancer. Esophageal cancer is the fourth most common cancer in China. However, there are about 223,000 new cases and 197,200 deaths in China each year, accounting for 52.8% of the global morbidity and 49.3% of the death rate. According to the latest data in China, the age-standardized death rate of esophageal cancer is 8.36 per 100,000 cases, which is 1.52 times the global rate ⁵. Esophageal cancer is a chronic disease with multiple factors. In developing countries, on the one hand, there is an interaction between genetic and environmental factors. On the other hand, it may be that the risk factors of esophageal cancer are related to lack of nutrition, unhealthy lifestyle and smoking. However, the exact mechanism of esophageal cancer is still unclear ⁶. Among cancers caused by alcohol, esophageal cancer accounts for the highest proportion in Japan. GBD2010 estimates the death of esophageal cancer caused by alcohol of 76700 and 1,825 000 disability-adjusted life years (DALYs) lost ⁷.Esophageal cancer is relatively rare in most Western Europe and North America, but overall, it is the sixth leading cause of death from cancer in the world ⁸. In Africa, the incidence of esophageal cancer may increase, and this increase is mainly in people with lower socioeconomic class ⁹.

Stomach cancer is a common malignant tumor, and stomach cancer deaths account for about 50% of the global stomach cancer deaths in China ¹⁰. It is estimated that, about 498 000 people in China died of stomach cancer and about 1,364 people died every day in 2015. Stomach cancer has always been one of the main causes of death for Chinese people¹¹. It is not only related to industrialization and urbanization of China, but also related to high-risk factors such as Helicobacter pylori and unhealthy lifestyles ¹².

Materials And Methods

Source of Data

The age-standardized death and population data used in this study are from the Institute for Health Metrics and Evaluation (IHME, http:/ghdx.healthdata.org/). The Global Health Data Exchange (GHDx) released by IHME provides a systematic assessment of important health issues and disease burdens in 197 countries and regions, providing a scientific basis for the formulation of relevant health policies ¹³. The data of IHME comes from surveys, censuses, vital statistics and other health-related data in various countries around the world. Meanwhile, it is also an open database with credibility worldwide. The GBD not only estimates the incidence, prevalence, mortality, YLL, YLD and DALY indicators of each disease and injury, but also it is reported by year, location, age group and sex. The purpose of GBD research is to establish comprehensive and comparable global health indicators.

Statistical Analysis

Age-period-cohort model

The mortality of colon cancer, esophageal cancer, stomach cancer stratified by countries (China/Japan/the United States/India), gender (male/female), and age groups (20−90 by 5 years) were calculated. These death rates were smoothed with 5-year moving averages to reveal secular trends. For the APC analysis, the colon cancer, esophageal cancer, stomach cancer data in this study were collected in five successive five-year periods from 1992-1997 to 2012-2017, and 14 five-year age groups, ranging from 20–24 years to 85-89 years, and 10 five-year birth cohorts, ranging from 1903-1907 to 1993-1997.

The Age-Period-Cohort Model (APC model) can simultaneously evaluate the contribution of age effect, period effect and the corresponding birth effect to morbidity and mortality. The APC model represents an epidemiological methods, which can be used to extract information about historical changes in mortality and morbidity risks from cross-sectional data, it also widely used in the fields of sociology, demography and epidemiology.

Because the cohort=period-age. This means that there is multicollinearity between variables. Multicollinearity is a common problem in the application of APC models. Yang first proposed the Intrinsic Estimator (IE) to solve the multicollinearity problem in the application of the APC model ^14-16. In this study, the model is expressed as: 

Y = log(M) = μ + αage_i + βperiod_j + γcohort_k + ε. 

M is defined as the death rate; α, β, and γ refer to three coefficients (α refers to the age effect, β represents the period effect, and γ is the cohort effect, μ and ε are defined as the intercept and random error. The APC analysis was performed using Stata 12.0 software. In addition, we considered the result p<0.05 as statistically significant. We use Akaike Information Criteria (AIC), Bayes Information Criteria (BIC), Standard error (SE) and risk ratios to evaluate the degree of model fitting.

Joinpoint Regression Analysis

Research on trend changes is an important issue for analyzing cancer mortality and morbidity data. We use Joinpoint to estimate the time trends of three cancers' ASMR, and determine the changes in China, Japan,the United States and India in different years based on the turning points, and estimate the annual percentage change (APC), average annual percentage change (AAPC) and 95% of each trend. In other words, this model allows us to see the trend of the death of three cancers, and intuitively find the point of change in each trend. The model uses the standardized rate of cancer as the dependent variable and the year as the independent variable to establish a log-linear model. The model is implemented using the Joinpoint regression program version 4.7.0.0 of the Statistical Research and Application Department of the National Cancer Institute's Monitoring Research Program.

Fitting Holt exponential smoothing model

Holt model uses double exponential smoothing ^17,18. The two smoothing components are adjusted according to the level and trend of the data. The Holt model is used to predict the death of three cancers in Japan, Japan, the United States, and India from 2018 to 2027.

Results

Descriptive Analysis of Death rate Trends 

Figure 1 shows the age-standardized death rate of three cancers in China (A), Japan (B), the United States(C), and India (D) from 1990 to 2017. In China, all three cancers death rate show a slight upward trend and then decline. The ASDR of stomach cancer in China is particularly significantly higher than that of colon cancer. In 2017, the death rate of colon cancer in China was very close to that of esophageal cancer. However, the ASDR is lowest in esophageal cancer in Japan, compared with stomach cancer and colon cancer. The ASDR of Japan and India are almost at the same level among esophageal cancer, lower than that of stomach cancer and colon cancer. Besides, the trend of ASDR in the United States is opposite to that in China, and stomach cancer has a lower death. The standardized death rate of esophageal cancer in the United States is very high, and it has dropped from 20.53 per 100,000 in 1990 to 11.25 per 100,000 in 2017.

Table 1 shows the annual percentage change (APC) and average annual percent change (AAPC of three cancers death in China, Japan, the United States and India from 1990-2017. Joinpoint model showed that Esophageal cancer and stomach cancer death rate decreased rapidly from 2004 to 2007 This trend can be seen for stomach cancer, with overall AAPC values of -2.2%(-2.5,-2.0) in China, -2.8%(-3.1,-2.5) in Japan, -2.0%(-2.2,-1.8) in the U.S. and -1.4%(-1.7,-1.0) in India

The age, period, and cohort effects on colon cancer mortality

We analyzed the age, period, and cohort effects of colon cancer death rate in China, Japan, the United States, and India from 1992 to 2017 based on the APC model, and calculated RR values of age, period, and cohort. Finally, the results are shown in Figure 2 and Appendix 1. Regarding the age effect of colon cancer mortality, we set the age RR of colon cancer mortality in the 20-24 age group as 1. The RR of China is 63.92 in the 85-89 age group, and the RR of 85-89 age group in Japan is 180.17, and the RR value of the US in the 85-89 age group is 149.19, while the RR value of India in the 85-89 age group is 79.07. The RR values of the death rate of colon cancer in the four countries all increased monotonously with age. But generally speaking, the RR values of colon cancer mortality in Japan and the United States are higher than those of China and India in all age groups.

For the period RR of colon cancer mortality, we set the RR to 1 of colon cancer mortality in 1992. In 2017, the RR of China was 1.54, the RR of Japan was 1.70, the RR of United States was 1.73, and the RR of India was 2.23. From 1992 to 2017, the RR value of death rate in the four countries showed an overall upward trend, while the RR of colon cancer mortality in India showed a tortuous upward trend. For the cohort effect of colon cancer death rate, we set the RR of colon cancer to 1 in the birth cohort 1903-1907, the RR of China in 1993-2007 of the birth cohort was 0.34, and the RR of the United States was 0.14. The RR for India is 0.07. The RR of colon cancer death rate in the four countries showed a downward trend with the cohort, and China had a slight upward trend in 1948-1952.

The age, period, and cohort effects on esophageal mortality.

We analyzed esophageal death rate in China, Japan, the U.S., and India from 1992 to 2017 based on the APC model, and calculated RR values of the results are shown in Figure 3 and Appendix 2. The age effect in Japan increased approximately linearly with increasing age from 30 to 75 years and peaking at approximately 75 year. After 75 years of age, the age effect continued to decrease. The RR of the United States has the similar trend to Japan. Other two countries age RR increased all the time from 20 to 90 age group. The period effect in China decreased from 1997 to 2002, and thereafter generally increased until the year 2007. After 2008, the period effect decreased again. However, the period RRs in Japan, the United States and India has an upward trend from 1992 to 2017.

The cohort effect in China, Japan, the United States and India generally decreased from the 1910s to the 1990s. The cohort effect maintained a relatively stable trend before 1937 in China, Japan and the United States, which was followed by a decline. However, the cohort effect declined at an accelerated rate since the 1910s in Japan.

3.4  The age, period, and cohort effects on stomach mortality 

We analyzed stomach death rate in China, Japan, the U.S., and India from 1992 to 2017 based on the APC model, and calculated RR values of the results are presented in Figure 4 and Appendix 3. The age effects for China, Japan, the United States and India showed that the death rate of stomach cancer had consistent increasing trends with age. In China and India, fluctuating period trends were presented in stomach cancer which showed a decreasing trend first during the period of 1992-2002 to 2007-2017; then the period effects increased rapidly from 2002 to 2007. However, stomach cancer as a whole showed an upward change in the United States. In Japan, the trend shows a decreased trend first from 1992 to 2012 and then increased from 2012 to 2017. The cohort effect death rate patterns of stomach cancer in four countries showing a decreasing trend (Figures 3-3). The cohort effect of stomach cancer peaked at around the year 1903 to 1907. The risk of stomach cancer related death was highest for those born in earlier birth cohort, but the later birth cohort of RR was lowest for four countries.

Forecast of the three cancers mortality trends using holt model

Figure 5 depicted the forecasts of the three cancers by Holt models during the period 1990–2027 for four countries. Overall trend showed an increase in colon cancer trends for the period from 2018 to 2027 in China, the U.S. and India. Whereas, the forecast trend of stomach cancer in China, Japan and India showed a slightly declined trend. However, it was also observed that the confidence interval of colon cancer for China was wider than Japan. This indicated more uncertainty during that period.

Discussion

Cancer burden has been alarming around the globe, magnified by aging process and increasing population density ¹⁹. Irregularities in various lifestyle related factors like smoking, alcohol consumption, being obese/overweight, reduced physical activity, less intake of vegetable and fruits along with bacterial/viral infections can lead to carcinogenesis ²⁰. However, it is not possible to quantify the impact of all the established risk factors due to data constrains ²¹. Some other factors are categorized as modifiable risk factors and are not decisively recognized yet such as intake of processed or red meat, vegetables, fruits, dietary calcium, fiber and vitamins ²². 

In the current study we have compared the trends of colon, esophageal and stomach cancer death in China, Japan, the U.S., and India from 1990 to 2017 using an age-period-cohort analysis. The age and birth year of cases and years or time during which they have lived, are the three foremost time‐reliant factors which are capable of independently impacting the time trends of cancer incidence and mortality. During carcinogenesis under the multistage paradigms, each of these variables is deemed to have a definite biological consequence / significance. For example, carcinogens exposure which triggers the primary tumor development meant to generate cohort effects, however cancer promoters which are involved in the later stages of disease development meant to create period effects. Conversely, variable control measures like screening programs, better diagnostic and treatment facilities in the target population can also generate period effects in the mortality trends. APC model agonizes with the identification problem therefore, to overcome this problem and provide better insights into disease death rates we have used intrinsic estimator technique. 

In China, gastric/stomach ²³, colorectal ²⁴ and esophageal ²⁵ cancer ranked second, third and fifth among all the malignancies respectively. Cancer statistics has shown roughly diagnosis of 1.2 million new gastric cases all over the world and among which around 40% came from China representing its high incidence and mortality among Chinese²³. Colorectal cancer is also a leading cause of cancer related deaths in China and risk of disease increases with age particularly after 35 years of age and attains the peak among individuals aged 80-84 years ²⁶. Esophageal cancer is the fourth leading cause of cancer related mortality, its incidence is more common in rural areas and among males with greater incidence above the age of forty years ^25-27. We have found that death rate of all the three cancer types showed slightly increasing trend and declined at the end of the period possibly paving towards the better treatment and healthcare facilities in China. Following the similar pattern of cancer incidence ^23,24, ASDR of stomach cancer is significantly higher than that of colon cancer. Interestingly, the death rate of colon and esophageal cancer are nearly equal in 2017 representing lesser rate of early cancer diagnosis and sub-standard treatment approaches offered by diverse regions in China. The death rate of colon cancer peaks at the age of 80 to 90 years highlighting age as a confounding factor in cancer incidence and mortality consistent with the previously reported literature ^28,29. It has been suggested to offer regular colon cancer screening to the residents over 60 years of age being high-risk population because of limited treatment related resources in China ³⁰. Additionally, this trend of death rate keeps on increasing with the time period / every year might be due to the adoption of westernized lifestyle and less treatment facilities in some of the regions. As some of the regions in China are more to develop certain type of cancers due to their local environmental and lifestyle behaviors or traditions therefore, elucidating the role these factors may help in controlling or decreasing the cancer burden ³¹. 

Additionally, we have used age-period-cohort analysis to explore the cancer trends and found that colon and esophageal cancer has caused much damage in China. All the cancer types have shown an increasing trend with increasing age possibly due to environmental and genetic changes. Death rate of stomach cancer began to rise at the age of 35 years among Chinese population. Increased prevalence of Helicobacter pylori, improper sanitation and unhygienic storage of food and its consumption are the basic reasons behind increased stomach cancer incidence and mortality in younger population ³². It has been observed that stomach cancer trend first rises and then declines with the period showing betterment in the healthcare and treatment facilities. Consistent with the results of other people’s studies, the uptrend in the ASMRs may be caused by the period RRs of Esophageal cancer and stomach cancer from 2004 to 2007 in China ³³.  

We have determined the trends of colon, esophageal and stomach cancer death rates in Japan. Following the similar pattern like China, age standardized colon cancer death rate remains at the highest position in Japan followed by stomach and esophageal cancer mortal lifestyle and changes in dietary patterns in Japan since World War II, like obesity, increased uptake of alcohol and beverages with less to no physical activity ³⁴. In Japan, National Health and Nutrition Survey exposed a vibrant rise in the energy obtained from fat and showed that it rises from 7.0% to 26.6% in 1946 to 2000 indicating the establishment of westernized diet among Japanese population with time ³⁵. We have observed that Japan has rapid colon cancer death, peaking at the age of 75 to 80 years and this death rate increases with every calendar year. An upward trend of colon cancer has been found with age and period, but it declines in the cohort. In 1992, Japan has introduced colorectal screening programs using immunochemical faucal occult blood testing for residents of age forty years or above to control the disease spread and those who are tested positive on initial screening are meant to opt for sigmoidoscopy or colonoscopy for confirmation ³⁶. Besides all these preventive measures, it is still hard to control the disease spread might be due to poor dietary patterns. 

Although esophageal cancer is much lower than colon cancer in Japan but follows the same trend in terms of APC analysis. The death rate of esophageal cancer generally increases with age and period whereas a decline has been observed in the cohort effect since 1910. Better socioeconomic conditions along with the introduction of screening programs might partially explain these results. One of the studies has analyzed esophageal mortality trends from 1960-2000 in Japan and has observed a U shaped curve among the age group of 45 to 79 years with higher in males for mortality trends using the WHO data ³⁷. Although Japan has introduced screening programs for precursor lesions and early-stage detection in some populations, prevention persists to be the appropriate approach to lessen esophageal cancer burden most probably by modifying the possible risk factors like smoking and alcohol consumption ³⁸. The cohort studies showed that compared with people who never drink alcohol, people who drink 100 grams of pure alcohol a day have 11.71 times the risk of esophageal cancer in Japan ³⁹. In Japan ASDR of stomach cancer increases with the age particularly after the age of 40 years. While analyzing the period, decreasing trend has been observed first from 1992 to 2012 and then it dramatically increased from 2012 to 2017. Whereas in the birth cohort a declining trend was observed, and people born in later years have less death rate as compared to the one born in earlier times. One of the prospective cohort studies in Scotland established that men with adverse socioeconomic status in their childhood had greater stomach cancer mortality rates which is independent of their lifestyles in the adulthood ⁴⁰. Researchers have thought that there might be two hypotheses for the mechanisms underlying this type of correlation including greater biological vulnerability to ailment acquired unswervingly by poor resource settings in childhood and unhealthy lifestyles attained/activated through childhood environments and upheld throughout their adulthood. Poor dietary habits that are not meeting the nutritional standards, smoking and alcohol consumption are possible causes of stomach cancer and aforementioned mechanisms ⁴¹. 

Following the similar pattern like China and Japan, ASDR of colon cancer remains at the highest position in United States peaking at the age of 70-80 years followed by stomach and esophageal cancer mortality rates. Interestingly the death rate of colon cancer decreases over the course of period possible due to better treatment facilities. Overall, this death rate is still much higher than the other selected countries like China and Japan. However, in the U.S., the death rate of esophageal cancer generally increased with age and period and significantly pronounced in earlier age groups. Whereas stomach cancer rises in the earlier stages and then declines but the death rate was higher in elderly people. One of the probable justifications for the detected decrease in cohort effect, which is strictly linked to childhood socioeconomic status and strongly based on proof from epidemiological, biological and clinical studies, is the declining prevalence of Helicobacter pylori infection in younger cohorts. Another potential reason in decrease cohort trend is possibly decreased disease incidence due to better health and environmental facilities ^41,42. 

The ASDR of India is similar to Japan where esophageal cancer ASDR is lower than that of stomach and colon cancer concordant with the finding of another study from India highlighting increasing colorectal cancer mortality rates ⁴³. Death rate of all the cancer types increases with increased age in India, for colorectal cancer it was peaked at the age group of 80 to 90 years. But this rate was lower than United States and Japan possibly due to less westernized diet in some of the regions in India. Death rate of stomach and esophageal cancer was started at the age group of 35 years among Indians. Poor and unhygienic feeding habits might be reason behind early onset of these cancers in India ⁴⁴. Studies from India are limited but have shown certain reasons for stomach cancer development such as H. pylori infection, pickled food, spicy food, smoking and alcohol consumption ⁴⁵. Whereas declining trend was observed for the cohort effect in all three cancer types. 

Limitation

Firstly, the data in the study was based on GBD 2017. Our study is provided by local governments of worldwide countries, which may have very different systems for collecting vital statistics and methods for confirming the cause of death. In addition, although the IE method has unbiased and larger estimation, studies with more risk factors are needed to explain more accurate results in the future. However, GBD provides us with an important platform for international comparison.

Conclusion

In conclusion, age standardized death rate of colon cancer ranked top among all the selected countries i.e., China, Japan, United States and India. The trend of death rates among colon, stomach and esophageal cancers has been increased in all countries with declining cohort trends. Excessive struggle is required not only to progress public awareness and endorse early detection of cancer but also to deliver more manageable health services, adequate finance, and suitable cancer care related infrastructure around the globe.

Declarations

Availability of data and materials

The list of data sources used is publicly available at the Global Health Data Exchange website (http://ghdx.health data.org/gbd-results-tool), no additional data available

Authors’ contributions

Y C: design this context, acquisition of data or analysis and drafting the article. YY: have access to the data and con-trolled the decision to publish. GC: Revise the article. GT: Revise the article. SH: Revise the article. 

Funding

This work was funded by the National Natural Science Foundations (NSFC) of China [grant number 81973153, 81673276, and 81373101].

Conflict of Interest 

The authors declare that there is no conflict of interest regarding the publication of this paper.

Acknowledgments

We would like to thank the Institute for Health Metrics and Evaluation for the data.

References

1. Shao, Y. et al. Time Trends of Gastrointestinal Cancers Incidence and Mortality in Yangzhong From 1991 to 2015: An Updated Age-Period-Cohort Analysis. Frontiers in oncology8, 638, doi:10.3389/fonc.2018.00638 (2018).
2. Chen, W. et al. Cancer statistics in China, 2015. CA: a cancer journal for clinicians66, 115-132, doi:10.3322/caac.21338 (2016).
3. Li, M. et al. Time trends of esophageal and gastric cancer mortality in China, 1991-2009: an age-period-cohort analysis. Sci Rep7, 6797-6797, doi:10.1038/s41598-017-07071-5 (2017).
4. Wang, H. et al. [Disease burden of colorectal cancer in China: any changes in recent years?]. Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi41, 1633-1642, doi:10.3760/cma.j.cn112338-20200306-00273 (2020).
5. Zheng, R. S. et al. [Report of cancer epidemiology in China, 2015]. Zhonghua zhong liu za zhi [Chinese journal of oncology]41, 19-28, doi:10.3760/cma.j.issn.0253-3766.2019.01.005 (2019).
6. He, Y. et al. Incidence and mortality rate of esophageal cancer has decreased during past 40 years in Hebei Province, China. Chin J Cancer Res27, 562-571, doi:10.3978/j.issn.1000-9604.2015.10.06 (2015).
7. McGuire & S. Geneva, Switzerland: World Health Organization, International Agency for Research on Cancer, WHO Press, 2015. Advances in Nutrition An International Review Journal7, 418-419 (2016).
8. Parkin, D. M., Bray, F., Ferlay, J. & Pisani, P. Estimating the world cancer burden: Globocan 2000. International journal of cancer94, 153-156, doi:10.1002/ijc.1440 (2001).
9. Day, N. E. The geographic pathology of cancer of the oesophagus. British medical bulletin40, 329-334, doi:10.1093/oxfordjournals.bmb.a071999 (1984).
10. Ferlay, J. et al. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [Internet]. Lyon, France: International Agency for Research on Cancer. (2013).
11. Bode, A. M., Dong, Z. & Wang, H. Cancer prevention and control: alarming challenges in China. National science review3, 117-127, doi:10.1093/nsr/nwv054 (2016).
12. Zong, L., Abe, M., Seto, Y. & Ji, J. The challenge of screening for early gastric cancer in China. Lancet (London, England)388, 2606, doi:10.1016/s0140-6736(16)32226-7 (2016).
13. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet (London, England)392, 1736-1788, doi:10.1016/s0140-6736(18)32203-7 (2018).
14. Yang, Y., Fu, W. J. & Land, K. C. The Intrinsic Estimator for Age-Period-Cohort Analysis: What It Is and How to Use It1. American Journal of Sociology113, 1697-1736 (2008).
15. Luo, L. Assessing Validity and Application Scope of the Intrinsic Estimator Approach to the Age-Period-Cohort Problem. Demography50, 1945-1967 (2013).
16. Yang, Y., Fu, W. J. & Land, K. C. A Methodological Comparison of Age‐Period‐Cohort Models: The Intrinsic Estimator and Conventional Generalized Linear Models. Sociological Methodology34, 75-110 (2010).
17. Hyndman, R. J., Koehler, A. B., Ord, K. J. & Snyder, R. D. Forecasting with exponential smoothing: the state space approach. Springer Berlin Heidelberg (2008).
18. Holt, C. C. Forecasting seasonals and trends by exponentially weighted moving averages. International Journal of Forecasting20, 5-10 (2004).
19. Mubarik, S. et al. A Hierarchical Age-Period-Cohort Analysis of Breast Cancer Mortality and Disability Adjusted Life Years (1990-2015) Attributable to Modified Risk Factors among Chinese Women. International journal of environmental research and public health17, doi:10.3390/ijerph17041367 (2020).
20. Liu, X. et al. Secular Trend of Cancer Death and Incidence in 29 Cancer Groups in China, 1990-2017: A Joinpoint and Age-Period-Cohort Analysis. Cancer management and research12, 6221-6238, doi:10.2147/cmar.s247648 (2020).
21. Malik, S. S. & Masood, N. Essentials of Cancer Genomic, Computational Approaches and Precision Medicine. (Essentials of Cancer Genomic, Computational Approaches and Precision Medicine, 2020).
22. Masood, N., Mubashar, A. & Yasmin, A. Epidemiological factors related to GSTM1 and GSTT1 genes deletion in colon and rectum cancers: A case-control study. Cancer biomarkers : section A of Disease markers15, 583-589, doi:10.3233/cbm-150498 (2015).
23. National Health Commission Of The People's Republic Of, C. Chinese guidelines for diagnosis and treatment of gastric cancer 2018 (English version). Chinese journal of cancer research = Chung-kuo yen cheng yen chiu31, 707-737, doi:10.21147/j.issn.1000-9604.2019.05.01 (2019).
24. Xu, R. et al. Disease characteristics and treatment patterns of Chinese patients with metastatic colorectal cancer: a retrospective study using medical records from China. BMC cancer20, 131, doi:10.1186/s12885-020-6557-5 (2020).
25. Liang, H., Fan, J. H. & Qiao, Y. L. Epidemiology, etiology, and prevention of esophageal squamous cell carcinoma in China. Cancer biology & medicine14, 33-41, doi:10.20892/j.issn.2095-3941.2016.0093 (2017).
26. Du, L. B. et al. [Report of colorectal cancer incidence and mortality in China, 2013]. Zhonghua zhong liu za zhi [Chinese journal of oncology]39, 701-706, doi:10.3760/cma.j.issn.0253-3766.2017.09.012 (2017).
27. Tuo, J. Y. et al. [Report of incidence and mortality of gallbladder cancer in China, 2014]. Zhonghua zhong liu za zhi [Chinese journal of oncology]40, 894-899, doi:10.3760/cma.j.issn.0253-3766.2018.12.004 (2018).
28. Chen, Q., Liu, Z. C. & Cheng, L. P. An Analysis of Incidence and Mortality of Colorectal Cancer in China,2003～2007. China Cancer (2012).
29. Liu, S., Zheng, R., Zhang, M., Zhang, S. & Chen, W. Incidence and mortality of colorectal cancer in China, 2011. Chinese journal of cancer research = Chung-kuo yen cheng yen chiu27, 22 (2015).
30. Zhu, J. et al. Epidemiological Trends in Colorectal Cancer in China: An Ecological Study. Digestive diseases and sciences62, 235-243, doi:10.1007/s10620-016-4362-4 (2017).
31. Feng, R. M., Zong, Y. N., Cao, S. M. & Xu, R. H. Current cancer situation in China:good or bad news from the 2018 Global Cancer Statistics? 癌症通讯(英文)39, 12.
32. Yang, L. et al. Incidence and mortality of stomach cancer in China, 2014. Chinese journal of cancer research = Chung-kuo yen cheng yen chiu30, 291-298, doi:10.21147/j.issn.1000-9604.2018.03.01 (2018).
33. Gao, X. et al. Trends of Esophageal Cancer Mortality in Rural China from 1989 to 2013: An Age-Period-Cohort Analysis. Int J Environ Res Public Health14, 218, doi:10.3390/ijerph14030218 (2017).
34. Nakagawa, H. et al. Changes in trends in colorectal cancer incidence rate by anatomic site between 1978 and 2004 in Japan. European journal of cancer prevention : the official journal of the European Cancer Prevention Organisation (ECP)26, 269-276, doi:10.1097/cej.0000000000000255 (2017).
35. Tada, N. et al. Japanese dietary lifestyle and cardiovascular disease. Journal of atherosclerosis and thrombosis18, 723-734, doi:10.5551/jat.8193 (2011).
36. Ransohoff, D. F. Colon cancer screening in 2005: status and challenges. Gastroenterology128, 1685-1695, doi:10.1053/j.gastro.2005.04.005 (2005).
37. Qiu, D. & Kaneko, S. Comparison of esophageal cancer mortality in five countries: France, Italy, Japan, UK and USA from the WHO mortality database (1960-2000). Japanese journal of clinical oncology35, 564-567, doi:10.1093/jjco/hyi159 (2005).
38. Lin, Y. et al. Epidemiology of esophageal cancer in Japan and China. Journal of epidemiology23, 233-242, doi:10.2188/jea.je20120162 (2013).
39. Roerecke, M. et al. Estimates of alcohol-related oesophageal cancer burden in Japan: systematic review and meta-analyses. Bull World Health Organ93, 329-338C, doi:10.2471/BLT.14.142141 (2015).
40. Smith, G. D., Hart, C., Blane, D. & Hole, D. Adverse socioeconomic conditions in childhood and cause specific adult mortality: prospective observational study. BMJ (Clinical research ed.)316, 1631-1635, doi:10.1136/bmj.316.7145.1631 (1998).
41. Tanaka, M. et al. Trends of stomach cancer mortality in Eastern Asia in 1950-2004: comparative study of Japan, Hong Kong and Singapore using age, period and cohort analysis. International journal of cancer130, 930-936, doi:10.1002/ijc.26080 (2012).
42. Newton, R. U. & Galvão, D. A. Exercise in prevention and management of cancer. Current treatment options in oncology9, 135-146, doi:10.1007/s11864-008-0065-1 (2008).
43. Thomas, V. M., Baby, B., Wang, K., Lei, F. & Mathew, A. Trends in colorectal cancer incidence in India. Journal of Clinical Oncology38, e16084-e16084 (2020).
44. Soldati, L. et al. The influence of diet on anti-cancer immune responsiveness. Journal of translational medicine16, 75, doi:10.1186/s12967-018-1448-0 (2018).
45. Dikshit, R. P., Mathur, G., Mhatre, S. & Yeole, B. B. Epidemiological review of gastric cancer in India. Indian journal of medical and paediatric oncology : official journal of Indian Society of Medical & Paediatric Oncology32, 3-11, doi:10.4103/0971-5851.81883 (2011).

Tables

Note: * APC, annual percentage change; AAPC, average annual percent change; CI, confidence interval; *Significantly different from 0 at alpha = 0.05 (p < 0.05).