The burden of kidney cancer and its attributable risk factors in 195 countries and territories, 1990–2017

Kidney cancer globally accounts for more than 131,000 deaths each year and has been found to place a large economic burden on society. However, there are no recent articles on the burden of kidney cancer across the world. The aim of this study was to present a status report on the incidence, mortality and disability-adjusted life years (DALYs) associated with kidney cancer in 195 countries, from 1990 to 2017. Vital registration and cancer registry data (total of 23,660 site-years) were used to generate the estimates. Mortality was estimated first and the incidence and DALYs were calculated based on the estimated mortality values. All estimates were presented as counts and age-standardised rates per 100,000 population. The estimated rates were calculated by age, sex and according to the Socio-Demographic Index (SDI). In 2017, kidney cancer accounted for 393.0 thousand (95% UI: 371.0–404.6) incident cases, 138.5 thousand (95% UI: 128.7–142.5) deaths and 3.3 million (95% UI: 3.1–3.4) DALYs globally. The global age-standardised rates for the incidence, deaths and DALY were 4.9 (95% UI: 4.7–5.1), 1.7 (95% UI: 1.6–1.8) and 41.1 (95% UI: 38.7–42.5), respectively. Uruguay [15.8 (95% UI: 13.6–19.0)] and Bangladesh [1.5 (95% UI: 1.0–1.8)] had highest and lowest age-standardised incidence rates, respectively. The age-standardised death rates varied substantially from 0.47 (95% UI: 0.34–0.58) in Bangladesh to 5.6 (95% UI: 4.6–6.1) in the Czech Republic. Incidence and mortality rates were higher among males, than females, across all age groups, with the highest rates for both sexes being observed in the 95+ age group. Generally, positive associations were found between each country’s age-standardised DALY rate and their corresponding SDI. The considerable burden of kidney cancer was attributable to high body mass index (18.5%) and smoking (16.6%) in both sexes. There are large inter-country differences in the burden of kidney cancer and it is generally higher in countries with a high SDI. The findings from this study provide much needed information for those in each country that are making health-related decisions about priority areas, resource allocation, and the effectiveness of prevention programmes. The results of our study also highlight the need for renewed efforts to reduce exposure to the kidney cancer risk factors and to improve the prevention and the early detection of this disease.

Regional level. At the regional-level, we found that High-income North-America [ [0.85 (95% UI: 0.68-1.1)] had the lowest age-standardised death rates. The age-standardised incidence and death rates were higher for males in all of the GBD regions, although this difference was not statistically significant in all regions (Fig. 1a, b).
Most regions experienced an increase in age-standardised incidence rates, with South Asia [48% (95% UI: 22- (Table 1). However, it is important to again note that some of these increases or reductions were not statistically significant.
Age and sex patterns. Sex differences in the incident rates first appeared in the 35-39 age group and increased up to the oldest age group (95+). The number of incidents was also higher in males, from the 30-34 age group up to the 85-89 age group, with a peak being seen in the 65-69 age group (Fig. 6). The death rate was also higher in males, than in females, in all age groups. The number of deaths was also higher in males in all age groups, except the 1-4 and 95+ age groups (Online Appendix Fig. 1). However, the pattern for DALY rates was slightly different, such that the trend started declining after 80-84 for males and 85-89 for females. The number of DALYs was also higher in males, in most of the age groups, except the 5-9, 10-14, 90-94 and 95+ age groups. The number of DALYs peaked in the 60-64 age group (Online Appendix Fig. 2). The YLL rate peaked in the 80-84 age group, which comprised a large proportion of the DALYs (Online Appendix Fig. 3). a b Figure 2. The percentage change in the age-standardised incidence (a) and death (b) rates of kidney cancer from 1990 to 2017 for the 21 GBD regions by sex.
Scientific RepoRtS | (2020) 10:13862 | https://doi.org/10.1038/s41598-020-70840-2 www.nature.com/scientificreports/ Burden of kidney cancer by SDI. At the regional-level, the age-standardised DALY rate increased up to an SDI of approximately 0.74 and then decreased with increasing SDI values (Fig. 7). The global age-standardised DALY rate was initially higher than expected, but then the rate fell below the expected level for the last 14 years. High-income Asia-Pacific was the only region in the GBD high-income super regions that had a lower than expected DALY rate across the entire measurement period. For the GBD super-regions of Central Europe, Eastern Europe and Central Asia, only Eastern Europe had a higher than expected DALY rate for the entire measurement period, while Central Europe had a higher than expected level for the last 13 years. From the Latin America and Caribbean super-region, only Southern and Central Latin America had higher than expected DALY rates across the entire measurement period. In the Sub-Saharan Africa super-region, only Southern Sub-Saharan Africa was found to have a lower than expected DALY rate for the entire measurement period. In the Southeast Asia, East Asia and Oceania super-regions, only Southeast Asia and East Asia had a lower than expected DALY rate from 1990 to 2017, but Oceania was lower than expected during the last 3 years of the measurement period. The South Asia region had a lower than expected DALY rate for the entire measurement period, while the North Africa and Middle East region was lower for most of the measurement period (from 1992 to 2017) (Fig. 7). Figure 8 presents the country-level age-standardised DALY rates and its expected relationship with SDI. The expected patterns were non-linear in nature, peaking at an SDI of 0.84. However, there were large national differences in age-standardised DALY rates. Uruguay, the Czech Republic, Lithuania, Ukraine, Iceland, Greenland www.nature.com/scientificreports/ and many other countries showed higher than their expected DALY rates, whereas Singapore, Kuwait, China, Algeria, Morocco and many other countries had much lower than expected DALY rates, based only on their SDI.

Risk factors.
Globally, 18% of kidney cancer DALYs was attributable to high BMI in both sexes (Male: 16.5%; Female: 22.1%). The proportion of kidney cancer DALYs there were attributable to high BMI ranged from 7.1% in Eastern Sub-Saharan Africa to 29.2% in High-income North America. Furthermore, 16.6% of kidney cancer DALYs was attributable to smoking in both sexes, but this burden was higher in males (21.6%) than females (7.3%). The smoking-attributable burden also differed across GBD regions, ranging from 3.9% in Western Sub-Saharan Africa to 22.9% in Eastern Europe. Finally, the burden of kidney cancer attributable to occupational exposure to trichloroethylene was negligible (Fig. 9).

Discussion
This study reported the incidence, mortality, and DALYs for kidney cancer in 195 countries from 1990 to 2017. Globally, the age-standardised incidence and death rates have increased while the DALY rates have declined, although neither of these changes were statistically significant. Our results show that there has been little or no progress in reducing the burden of Kidney cancer over the past 28 years and we call for renewed efforts to reduce the burden of this disease. GLOBOCAN 9 . There are several reasons for these differences, which may related to the data sources and/or different methodological approaches.
The GBD methodology considers all causes of deaths in each run, whereas GLOBOCAN only provides cancer mortality. Previous research has found that more developed countries have a higher incidence of kidney cancer than less developed countries 8 , which was also confirmed in the present study. The association that development level has with kidney cancer incidence and mortality has only been investigated in a small number of studies and these studies only used data from selected countries, meaning that their results must be interpreted with caution 8,17 . There are also a number of other problems with the previous research on this topic. For example, previous  www.nature.com/scientificreports/ research found that the incidence rate of kidney cancer was twice as high in developed countries as in developing countries 8 . However, this research only compared two states of development (developed vs. developing) and they did not measure each country's level of development using the SDI. In fact, using only two categories of development leads to information loss, which means that accurate and precise patterns may not be produced. Secondly, previous research determined development status using the Human Development Index (HDI), which is problematic as one of the HDI's components (life expectancy) is associated with health. Therefore, the association that development status has with kidney cancer may have previously been over-estimated 8 . In order to address this issue, in GBD 2017 we used the SDI, which does not include any health-related components. Thirdly, considering the association between the variables to be linear may be inaccurate 8 . Hence, we examined the non-linear association between SDI and kidney cancer burden, in order to determine the shape of the association. Finally,  www.nature.com/scientificreports/ previous research examined the association between the HDI of a specific year (e.g., 2000) with the incidence and mortality rates of a different year (e.g., 2012) 8 . In this study, we examined the same years. There are a number of possible reasons for the higher burden of kidney cancer in developed countries. Firstly, the prevalence of risk factors, such as smoking, high BMI and low physical activity and hypertension may be higher in developed countries than in developing regions 8 . Secondly, the increases in the incidence of kidney cancer could also be partly due to improvements in the early detection of cancer using imaging procedures, such as ultrasonography, computed tomography, and magnetic resonance imaging in high income countries 3 . Perhaps the increases in the incidences of kidney cancer may be due to exposure to occupational and environmental risk factors, such as trichloroethylene, cadmium, arsenic, radon and nitrate. Although we know exposure to these risk factors have declined in the developed world, there is no evidence to suggest this same pattern has been replicated in the developing world 3,5,18 .
Although a number of risk factors have found to be associated with kidney cancer, the attributable burden was only calculated for those risk factors that had robust evidence of their relationship with kidney cancer 10 . Therefore, the attributable burden was calculated for two life style risk factors (smoking and high body mass index) and one environmental and occupational risk factor (occupational exposure to trichloroethylene).
High BMI (overweight/obesity) is one of the important risk factors, contributing 18.5% to the burden of kidney cancer in the population. Previous research has found that the prevalence of obesity has continuously increased in most countries during the period 1990-2015 and has doubled in more than 70 countries 19 . Moreover, in many countries the rate of increase in childhood obesity has been greater than the rate of increase among adults 19 . There are a number of approaches that can be taken to reduce the prevalence of being overweight and obese and to thereby reduce the burden of this disease. These measures should include a ban on advertising unhealthy foods, improving school meals, taxation, subsidies, and incentives to increase the production of healthy foods 20 .
Our study also found that smoking contributes 16.6% to the burden of kidney cancer (both sexes) and reducing exposure to this risk factor could play an important role in decreasing the burden of this disease. A study of the global progress in reducing the prevalence of smoking has reported heterogeneous findings, according to country, development status, and sex. Globally, the age-standardised prevalence rate of daily smoking declined by 28.4% and 34.4%, respectively, among men and women from 1990 to 2015 21 . There is a need to achieve greater success in the control of tobacco smoking through the use of effective, comprehensive, and adequately implemented and enforced policies.
The third risk factor assessed in our study was exposure to Trichloroethylene, which is usually used as a metal cleaner and degreaser 15 . The burden of kidney cancer that was attributable to occupational exposure to trichloroethylene was found to be 0.1%. The attributable burden of this risk factor is negligible, as the populations' exposure is very low. A meta-analysis found that occupational exposure to trichloroethylene increased the risk of kidney cancer by 32% 22 .
Physical activity and alcohol consumption have also been considered to be lifestyle risk factors. A metaanalysis found a negative association between physical activity and kidney cancer 23 , which was also confirmed in a pooled analysis of cohort studies 24 . However, research has also shown that prolonged sitting does not increase kidney cancer among men and women 25 . Therefore, it is not entirely clear how physical activity changes the risk www.nature.com/scientificreports/ of kidney cancer and its association with kidney cancer has not been examined independently from high body mass index and hypertension 5 . Alcohol consumption is another potential risk factor that has been extensively studied. Two meta-analyses 26,27 have found there to be an inverse relationship between alcohol consumption and the risk of kidney cancer, and these findings have been confirmed by large scale prospective studies 28,29 . All of the aforementioned studies reported a lower risk for drinkers, compared to non-drinkers or light drinkers. GBD 2017 did not calculate the burden of kidney cancer attributable to alcohol consumption, but this addition has been suggested for future GBD cycles, as there is sufficient evidence of the association between these two variables.
Although the association between diet and kidney cancer has been examined in previous research, the evidence is not robust 5 . Most studies have reported no association between fruit and vegetable intake and the risk of kidney cancer, and nutrient specific associations have not been reported 5 .
Medical history, including hypertension, chronic kidney disease, kidney stones and diabetes mellitus have also been found to be associated with kidney cancer 5,6 . Several studies have reported that hypertension increases the risk of kidney cancer and have reported dose-response relationships between blood pressure and kidney cancer risk 30,31 . This risk factor has also been suggested for inclusion in the next cycle of the GBD study. Chronic kidney disease has also been found to be a risk factor for kidney cancer [32][33][34] . Furthermore, previous large scale studies 35,36 and one meta-analysis 37 have found a higher risk of kidney cancer to be associated with kidney stones. Finally, previous large scale studies 38,39 and a meta-analysis 40 have found diabetes mellitus to be associated with a higher risk of kidney cancer.
Considering the evidence reported above, there is sufficient evidence to suggest that the risk of kidney cancer is associated with: alcohol consumption, hypertension, chronic kidney diseases, kidney stones and diabetes mellitus. These risk factors should all be included into the next cycle of the GBD project to inform public policy and health policy makers how much of the kidney cancer burden could be attributed to each of these risk factors. However, the evidence for physical activity, diet and several other risk factors are not compelling and further research is needed.
The present study provides important information on the proportion of the kidney cancer burden that is attributable to modifiable risk factors, such as smoking, high body mass index and occupational exposure to trichloroethylene, which can be used for primary prevention purposes. However, the role of other risk factors, such as hypertension and diabetes mellitus, should be calculated in future iterations of the GBD project. In addition, improvements in diagnostic measures are needed, with the identification of blood-and urine-based markers being one approach with considerable merit 5 . These improvements are needed to allow earlier detection of kidney cancer and thereby a better prognosis for patients.

Strengths and limitations
The present research had a number of limitations. Firstly, it is possible that in some countries the rate of cancer detection is low and hence the incidence is lower than reported here. Secondly, some countries do not have the vital statistics to capture the causes of death. GBD methodology adjusts for these biases and provides uncertainty intervals for all estimates. conclusions There has been little or no improvement in the burden of kidney cancer over the last 28 years. Our study provides much needed information about the burden of kidney cancer in each country, to enable countries to better plan to address their burden and to allocate their limited resources more appropriately. Our results highlight the need for renewed efforts to reduce exposure to risk factors and to improve the prevention and early detection of this disease.