Global burden of prostate cancer attributable to smoking among males in 204 countries and territories, 1990–2019

Annual number of prostate cancer-related deaths, DALYs, age-standardized mortality rate (ASMR), and age-standardized DALY rate (ASDR) attributable to smoking from 1990 to 2019, by year, age, region, and country were extracted from the Global Health Data Exchange (GHDx) query tool [12]. Data from a total of 204 countries and territories were available. These countries and territories were then divided into 5 levels (low, low-middle, middle, high-middle, and high) in terms of the socio-demographic indices (SDI). Besides, the world is separated into 21 regions according to epidemiological similarities and geographical proximity. Because only one in 10,000 men under age 40 will be diagnosed, we extracted 12 age categories by five-year age groups within the ages of 40–94, and ≥ 95 years old to investigate the age patterns of mortality and DALYs. Data on patients in 4 age groups, 15–74 years and ≥ 75 years, were also included, as previous literature mentioned patients with early-onset prostate cancer continued to increase over the years and that these patients with a greater genetic component were inclined to die from the cancer. We wonder the distribution of mortality and DALYs across age groups in patients exposed to smoking, and whether they would be increased in earlier age groups as predicted. We also collected human development index (HDI) data at the national level from the World Bank.

In GBD 2019, prostate cancer was defined by the International Classification of Diseases (ICD) code 9^th and 10^th Revision (185–185.9, V10.46, V16.42, V76.44, and C61-C61.9, Z12.5, Z80.42, Z85.46, respectively [13]. Exposure to smoking is defined as the prevalence of current use of any smoked tobacco product and prevalence of former use of any smoked tobacco product. Among current smokers, it indicates cigarette equivalents smoked per smoker per day and cumulative pack-years of exposure, while among former smokers, the distribution of the number of years since cessation is estimated [5].

The socio-demographic index (SDI) estimated by GBD researchers is expressed on a scale of 0 to 1, and estimated based on lag distributed income (LDI) per capita, mean education for those ages 15 and older (EDU15 +), and total fertility rate under the age of 25 (TFU25) [3, 6, 14]. The DALY is a summary measure that quantifies the overall disease burden. It represents the sum of years of life lost due to premature death and years lived with disability [5, 14, 15]. One DALY can be regarded as the loss of 1 year in full health [2]. GBD 2019 modeling strategies for estimating DALYs have been described in detail elsewhere [14, 15].

We computed the number of deaths, DALYs, ASMR, and ASDR with 95% confidence intervals (CIs) (generated using the 25th and 97.5th values of the ordered 1,000 estimates) to quantify the burden of prostate cancer attributable to smoking [16]. Age-standardized rates were calculated by standardization to the global age structure from GBD and the attributable proportions of ASMR and ASDR due to smoking were measured using population attributable fractions, which represent the ASMR and ASDR that could have been avoided if the exposure to smoking was reduced to an alternative ideal exposure scenario. Population attributable fractions were estimated using the GBD 2019 comparative risk assessment approach [5]. The process of age-standardization of rates is a classic epidemiological method that removes the confounding effect of differences in age structure between the populations being compared, and the formula is as follows:

The ASR (per 100,000 population) is equal to the sum of the product of the specific age ratio (a_i, where i denotes the i^th age class) and the number of persons (or weight) (w_i) in the same age subgroup i of the chosen reference standard population, then dividing the sum of standard population weights.

An estimated annual percentage change (EAPC) was calculated to quantify the secular trends of ASR from 1990 to 2019 through the following formula: ln(y) = α + βx + ε, where y = ln(ASR), x = calendar year, and ε is the error term. EAPC could be calculated as 100 × (exp (β)-1), and its 95% confidence interval (CI) was obtained from the linear regression model. The ASR was considered to be in an upward trend if the EAPC estimation and its lower 95% CI were both > 0. Conversely, the ASR was in a downward trend if the EAPC estimation and its upper 95% CI were both < 0. Otherwise, the ASR was considered to be stable over time. In addition, to explore the influential factors for the EAPC, we assessed the association between EAPCs and ASRs (1990), and HDI (2019) at the national level using the pearson test. At last, a hierarchy cluster analysis was conducted to cluster the countries and territories into 5 categories (a: remained stable; b: minor increase; c: significant increase; d: significant decrease; e: minor decrease) according to their temporal trends in ASMR and ASDR. In addition, we first used the decomposition methodology of Das Gupta [17] to decompose global prostate cancer death and DALYs due to smoking by population age structure, population growth, and epidemiologic changes. All statistics and visualization were performed using R program (Version 4.1.2).

Of all prostate cancer deaths and DALYs globally in 2019, 6% and 6.6% were attributable to smoking (Fig. 1), respectively; resulting in 29,298 (12,789 to 46,609) deaths and 571,590 (253,490 to 917,820) DALYs due to prostate cancer attributable to smoking among males in 2019. The number of smoking-related prostate cancer deaths and DALYs showed an upward trend from 1990 to 2019, an increase of approximately 50% compared to those of 1990. However, the age-standardized mortality rates were 0.87 per 100,000 people in 2019 and demonstrated a downtrend with an EAPC of -1.83 (95% CI -1.94,-1.72) between 1990 and 2019. By contrast, the age-standardized DALYs rates were 15.58 per 100,000 people and demonstrated a stable trend with the EAPC of -0.8 (95% CI -2.35, 0.78) (Table 1; Fig. 2). Among the 4 age subgroups, the distribution of deaths and DALYs across age groups remained stable over time (Additional file 1).

For SDI regions, high SDI region had the highest number of smoking-related prostate cancer deaths (9 thousand) and DALYs (174 thousand) in 2019, both accounting for over 30% worldwide. Additionally, high SDI region also carried the highest ASMR and ASDR. It was worth noting that the ASMR decreased across the 5 SDI regions from 1990 to 2019, among which high-middle and high SDI regions had a faster decrease compared to low and low-middle SDI regions. By contrast, the ASDR remain stable in low, low-middle, middle and high-middle SDI regions, but high SDI region had a rapid decrease (EAPC -2.17, 95% CI: -4.06, -0.25) (Table 1).

For geographical regions, the heaviest burden had occurred in Western Europe and East Asia for over 30 years, accounting for almost half deaths in the world. However, the highest ASMR and ASDR occurred in Caribbean. As for ASMR, the most significant decrease was detected in Australasia and High-income North America from 1990 to 2019, with EAPCs all over 3, whereas Eastern Europe had the largest increase. Concurrently, high-income North America and East Asia had the fastest decrease in ASDR, with EAPC -2.42 (95% CI: -4.46,-0.35) and -2^.08 (95% CI: -3.73,-0.4), respectively (Table 1).

In 1990, there was an approximately fourfold difference between the regions with the highest and lowest percentage of prostate cancer deaths and DALYs attributable to smoking, with the highest percentage in Eastern and Central Europe, East and Southeast Asia, and High − income Asia Pacific and the lowest percentage in Andean Latin America, Central Sub − Saharan Africa, and Western Sub − Saharan Africa. The proportion attributable to smoking in 2019 had a similar regional distribution, except that the percentage in Central Europe and High − income Asia Pacific didn’t rank among the highest. Besides, it was worth noting that in all GBD regions, the contribution of smoking to the total number of deaths and DALYs due to prostate cancer decreased between 1990 and 2019 (Fig. 1).

At the country level, China ranked first in the number of prostate cancer deaths and DALYs attributable to smoking in 2019, followed by the U.S. (Additional file 2: Table S1-S2). Seychelles, Dominica, and Zimbabwe were the top three in ASMR and ASDR in 2019 (Fig. 3A and B; Additional file 2: Table S3-S4). However, the fastest increase in ASMR and ASDR occurred in Niger, with EAPCs 2.96 (95% CI: 2.78, 3.13) and 2.77 (95% CI: 2.61, 2.94) in ASMR and ASDR, respectively, and the most rapid decreased in ASMR and ASDR occurred in Canada, with EAPCs -4.15 (95% CI: -4.38, -3.92) and -4.01 (95% CI: -4.27, -3.75) in ASMR and ASDR (Fig. 3C and D; Additional file 2: Table S5-S6). Percentage change across countries of the fraction of all prostate cancer deaths and DALYs that are attributable to smoking was shown in Additional files 3 and 4.

According to the results derived from cluster analysis, 79 countries (or territories) were categorized into ‘‘remained stable” group, including China, Ukraine, Iraq, and Cuba. Thirty countries (or territories) were categorized into ‘‘minor increase” group, including Afghanistan, Egypt, and Vietnam. Sixteen countries (or territories) were categorized into ‘‘significant increase” group, including Russia and Ivory Coast. 22 countries (or territories) were categorized into ‘‘significant decrease” group, including the U.S., the UK, France, and Canada. The remaining 47 countries (or territories) were categorized into ‘‘minor decrease” group, including India, Turkey, Brazil, and Japan (Additional file 5).

In 2019, the number of prostate cancer deaths attributable to smoking first increased and then decrease with age. Most deaths occurred in ages 70–84 years old, with the peak at the age group 75–79, and more age-specific deaths occurred in high and high-middle SDI regions compared to that in low and low-middle regions (Fig. 4A). Correspondingly, the age-specific mortality rate kept increasing from 40 to 95 years old. A similar pattern to that of deaths, with most DALYs occurring for 65–74 year olds with a peak at 70–74. Correspondingly, age-standardized DALY rates gradually decreased after reaching their highest value in the age range of 85–89 years old (Fig. 4B).

Globally, the age-specific mortality rate has decreased among all ages group from 1990 to 2019, with the fastest decrease occurring in 75–79 years old. Separately, the age-specific mortality rates have decreased in low-middle, middle, high-middle, and high SDI regions from 1990 to 2019, among which high-middle and high SDI regions had higher EAPCs in mortality rates in each age group compared to low and low-middle SDI regions. In low SDI region, the age-specific mortality rates have increased in 40–49 years old and increased over 90–95 years old (Fig. 5A). The EAPCs in age-specific DALY rate showed the same pattern as that in age-specific mortality rate (Fig. 5B).

Overall, the ASMR and ASDR across regions had an M-shaped association with SDI in 2019, with the inflection point around 0.55 and 0.75 (Fig. 6). Moreover, the ASMR and ASDR association with SDI across countries has been visualized. Across countries, as SDI increased, ASMR or ASDR increased until SDI was about 0.75, and then decreased with higher SDI. Based solely on SDI, the ASMR or ASDR was much higher than expected in Zimbabwe, Seychelles, and Dominica (Additional files 6 and 7). The EAPC in ASMR was highly negatively associated with HDI in 2019 (ρ = -0.46, P < 0.001), especially with HDI greater than 0.8 (Fig. 7). However, no significant associations were observed between EAPC in ASMR and ASMR in 1990 across different countries (Additional file 7: Fig S6). The same patterns were also observed between EAPC in ASDR and HDI in 2019, and ASDR in 1990 (Additional files 8 and 10).