Two Time Point Analysis of the Change in Risk and Aging Factors for Major Cancers: A 10-Year Longitudinal Study in China

Objective To quantify the change in risk and aging factors with a two time point analysis for major cancers to assess supportive strategies. Methods The 2004 and 2015 mortality statistics in China were accessed. The standardized mortality rates of the two periods were used to calculate the ratio of change (RC) value to assess the risk of death associated with time (social development with time) for cancers. The role of age in mortality with time was evaluated by the interaction between time and age using a Poisson regression. Results In ascending order of RC, the factors were uterus; other malignant neoplasms; esophagus; stomach; skin; liver; leukemia; “lip, oral cavity, and pharynx”; bladder; “colon and rectum”; breast; prostate; lung; ovary; pancreas; “lymphoid, hematopoietic, and related tissue”; and cervix cancers. According to their location on the scatter diagram, the 17 neoplasms could be divided into three groups, comprising undeveloped cancers (including four cancers), developed cancers (including three cancers), and cancers insensitive to social development. Unexpectedly, about 60% (as assessed by type of cancer) and two-thirds (as assessed by constituent ratio of death from all cancers) of cancers did not change with time. Conclusions Most cancers may be insensitive to social development. Internal factors, including aging, may be a key factor for the occurrence of cancer.


Introduction
Cancers are complex disorders, and it is virtually impossible to prove what causes a particular cancer, because most cancers have multiple possible causes. However, we can classify cancer-related factors as risk or protective factors to assess preventative strategies. A hitherto unknown number of risk and protective factors contribute to cancer development by interacting with each other. Although cancer can develop at any age, cancer incidence increases with age [1][2][3]; therefore, it is believed that old age is the most important risk factor for cancer development and was treated as a major variable in the present study.
Worldwide, the rates of cancer are increasing, mainly related to the aging population and changes in lifestyle resulting from social development [4][5][6]. Some cancers seem to be more prevalent in people with a lower quality of life, whereas other cancers may be more prevalent in people with a higher quality of life [7]. Social development or urbanization will result in changes to cancer risk and pro-tective factors, including aging, diet and obesity, lack of physical activity, smoking, radiation, infections, stress, and environmental pollutants [8][9][10], and their impact on the occurrence of cancer [11][12][13]. Therefore, in growing economies, whether the time trend of increased social development might influence the development and burden of cancer should be explored, which will allow us to identify and implement appropriate medical strategies. To date, there have been no comprehensive and quantitative two time point analyses on cancers.
The cancer incidence (201.7/100000) in China is similar to that of the overall worldwide incidence (197.9/100000) in 2018 [14]. China's ethnic composition is relatively simple and is presently undergoing rapid urbanization [15][16][17]. China's gross domestic product was 1.955 trillion USD in 2004 and 11.016 trillion USD in 2015 ( Figure 1). Health and medical care has been markedly affected by economic growth, making China an ideal model to study disease patterns associated with modernization [18][19][20]. In this study, we analyzed the characteristics of cancers using two properties: (1) age [8] and (2) risk factors [7]. The results showed that although these two properties varied significantly among some cancers, the rates of most cancers were unchanged with time. We also categorized 17 major cancers using two time point analysis and found that about 30% of cancers were sensitive to social development, and more than 60% of cancers were insensitive to social development. An improved understanding of such associations would benefit the development of a theoretical basis for cancer prevention and control strategies.

Original Data.
Original data was obtained from the 2004 and 2015 data sets of the national disease mortality surveillance system, which was edited by the Chinese Center for Disease Control and Prevention [21,22], representing data obtained from over 73 million out of the 1.4 billion population of China between 2004 and 2015. In 2004, life expectancy at birth was 73.6 years, whereas in 2015, it was 76.1 years. The International Classification of Diseases-(ICD-) 10 codes [23] were used to classify the underlying causes of death, allowing us to determine the mortality statistics. Tables 1 and 2 show the raw data.

Two Time Point Analysis of Deaths Caused by Neoplasms.
The mortality rates in 2004 and 2015 were standardized using the 2000 Standard in China [7] and were used to calculate the ratio of change (RC) with time with respect to deaths caused by a particular cancer. The RC values were calculated using the following formula: where P 1 represents a cancer's mortality rate in 2015 and P 0 represents the cancer's mortality rate in 2004. An RC value > 1 indicated that the mortality rate had increased with time, whereas a value < 1 indicated a decreasing trend with time. To evaluate the degree of deviation from 1 (the values greater than 1 and those less than 1 were com-parable), logarithmic transformation was used to transform the RC values into logarithms, with the absolute value, lnRC, serving as the quantitative value for the two time point analysis.
A scatter diagram was plotted using the lnRC values. The degree of deviation from 1 and the positions of the various cancers on the scatter diagram were used to sort neoplasms according to the changes in the mortality rate with time.

Assessing the Role of Age on Mortality from Cancer over
Time. To evaluate the mortality risk from aging and social development, a loglinear model using the Poisson regression was employed, with covariates of time (2004 and 2015) and a categorical age of 60 years old [7]. Poisson regression is a generalized linear model form of regression analysis used to model count data and contingency tables. Poisson regression assumes the response variable Y has a Poisson distribution and assumes the logarithm of its expected value can be modeled by a linear combination of unknown parameters. From the model, the B coefficients for the interaction between time and age were obtained, in which exp ðBÞ was taken as the risk ratio. The greater the positive value of B, the greater the effect of the aging factor with time was on a particular cancer [7].
A probability of a type I error ≤ 0:05 was considered statistically significant. SPSS ver. 17.0 software for Windows (IBM, Armonk, NY, USA) was used to analyze the data statistically.

Results
The two time point analysis of deaths caused by neoplasms is summarized and quantified in Table 3. The contributions of time (social development) to different cancers varied: The lnRC value for the 17 neoplasms ranged from −1.04 to 0.38. The greater the lnRC, the greater the disease risk over time.
In ascending order of lnRC were uterus; other malignant neoplasms; esophagus; stomach; skin; liver; leukemia; "lip, oral cavity, and pharynx"; bladder; "colon and rectum"; breast; prostate; lung; ovary; pancreas; "lymphoid, hematopoietic, and related tissue"; and cervix cancers.  BioMed Research International Table 4 shows the results of the Poisson regression analysis to evaluate the mortality risk of aging with time (social development) for major cancers. We observed variations in the quantitative values of the effect of age with time on diseases. For the 17 cancers, the B value (the interaction between time and age) ranged from −0.169 to 0.435. A higher value of B indicated a stronger effect on cancer of aging with time (social development). The effect of aging with time was significantly reduced for cancers of the uterus, liver, pancreas, and cervix (p < 0:05) and was significantly increased for cancers of the esophagus and leukemia (p < 0:05).
According to their positions on the scatter diagram, the 17 neoplasms could be suggested to divide into three groups ( Figure 1). Cancers located in the upper right were termed developed cancers (more prevalent in people with a higher quality of life) and comprised three cancers. Cancers in the lower left part were termed undeveloped cancers (more prevalent in people with a lower quality of life) and comprised four cancers. The third group, positioned between the developed and undeveloped groups, was termed as cancers insensitive to social development and comprised ten cancers (58.8%). Figure 2 illustrates the detailed changes with social development for the three groups of cancers.
The constituent ratios of death from the three groups of cancers are shown in

Discussion
Life expectancy at birth in 2004 was 73.6 years, and in 2015, it was 76.1 years. Therefore, our data for the two time point analysis could reveal the contribution of social development to the development of cancer. The results revealed that these contributions varied without consideration of age. In the 17 cancers studied, the lnRC value ranged from −1.04 to 0.38 (Table 3), indicating that the contributions of social development to different cancers varied.
The greater the lnRC, the greater the risk of disease risk with time (social development). The lnRC data showed that among the 17 cancers studied, uterus cancer had the lowest disease risk with time (social development) and cervix cancer had the highest risk. The 17 neoplasms could be divided into developed cancers (for which social development might be a risk factor), undeveloped cancers (for which lack of social development might be a risk factor), and cancers insensitive to social development, according to their position on the scatter diagram ( Figure 2). The constituent ratio of death from all cancers increased from 4.9% in 2004 to 7.33% in 2015 for developed cancers and decreased from 34.78% in 2004 to 25.06% in 2015 for undeveloped cancers (Figure 3).  [21].
Aging is a significant and uncontrollable risk factor for death [1,24,25]; therefore, determining the effects of aging on cancers with social development is important. The different contributions of age to mortality from cancers between 2004 and 2015 could also be considered as the change in  Our results showed that the B value (representing the interaction between time and age) for the 17 cancers ranged from −0.169 to 0.435. The effect of aging with social development was significantly reduced for cancers of the uterus, liver, pancreas, and cervix (p < 0:05) and was significantly increased for cancers of the esophagus and leukemia (p < 0:05), suggesting that the structure of risk factors had changed with increasing social development for these cancers. Moreover, for most of the cancers categorized as insensitive to social development, the level of risk and the structure of the risk factors had not changed with time ( Figure 3).
The identification of a group of developed cancers, including cervix, pancreas, and "lymphoid, hematopoietic, and related tissue" cancers, showed that risk factors from social development have contributed to increases in the rates of some cancers. The risk factors might have a more important role in the relatively young population for cervix and pancreas cancers. Accordingly, an emphasis on avoiding risk factors from social development or urbanization might contribute to disease prevention for these developed cancers.
The presence of a group of undeveloped cancers, including uterus, esophagus, stomach, and other malignant neoplasms, showed that protective factors from social development have contributed to decreases in the rates of some cancers. The protective factors might have a more important role in the relatively young population for esophagus cancer and in the relatively old population for uterus cancer. Accordingly, disease prevention might be aided by an emphasis on public health and increasing basal health status for undeveloped cancers.
One limitation of this study is that the social development was not described with more indicators. Time and social development may not be equal. However, China's gross domestic product was 1.955 trillion USD in 2004 and 11.016 trillion USD in 2015; in 2004, life expectancy at birth was 73.6 years, whereas in 2015, it was 76.1 years. Therefore, time and social development may be parallel during 2004-2015. Increase of gross domestic product and rapid urbanization would change lifestyle; cancer-related risk would also be changed with this processes. Therefore, time could represent social development in this period. The findings also suggested that only about 30% of cancers are sensitive to changes in socioenvironmental factors over time. Analyzing and controlling external risk factors might lead to 30% of cancers being prevented, because the external factors leading to cancers are relatively controllable. Solutions to developed urbanized diseases mainly involve avoiding risk factors, which include tobacco use, overweight, physical inactivity, inadequate diet, sexually transmitted infections, alcohol consumption, and air pollution. General protective interventions for life quality, including nutritional status, lifestyle, and quality of social life, could potentially delay the aging process. Thus, the primary prevention of cancers will be aided by increasing our understanding of the categories of cancer affected by social development. Further studies are warranted to observe correlation between specific risk factors with social development and cancers for developing more available preventative strategies.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The author declares that he has no conflicts of interest.