Global trends in COPD mortality attributable to ambient O3
Figure 1 shows the variation in global COPD mortality attributable to ambient O3 from 1990 to 2021. Overall, the absolute number of deaths increased sharply from 0.235 (95% UI: 0.041 to 0.330) million in 1990 to 0.490 (95% UI: 0.092 to 0.721) million in 2021, with a predominant increase in adults aged ≥ 65 years (0.187 [95% UI: 0.041 to 0.330] million to 0.420 [95% UI: 0.092 to 0.721] million) (Fig. 1A). There was a relatively small increase in adults aged 25–64 years (0.048 [95% UI: 0.010 to 0.084] to 0.070 [95% UI: 0.015 to 0.119] million) (Fig. 1A). However, a decreasing trend in ASMR was observed in both groups. In adults aged ≥ 65 years, ASMR decreased significantly from 63.8 (95% UI: 13.9 to 112.2) per 100 000 in 1990 to 57.7 (95% UI: 12.6 to 99.1) per 100 000 in 2021 (Fig. 1B), with an AAPC of −0.138 (95% CI: −0.217 to −0.059) (Fig. 1C). In adults aged 25–64 years, ASMR also decreased significantly from 2.37 (95% UI: 0.50 to 4.14) per 100 000 to 1.71 (95% UI: 0.37 to 2.93) per 100 000 (Fig. 1B), with an AAPC of −0.020 (95% CI: −0.023 to −0.016) (Fig. 1C).
Temporal and population trends in COPD attributable to environmental O3 exposure from 1990 to 2021 in the elderly, globally. Changes in the composition of deaths (A), age-standardized mortality rate (ASMR) (B) and average annual percentage change (AAPC) (C) in the groups of adults aged 25–64 and 65 +. Changes of ASMR in the groups of adults aged 25–64 and 65 +. Number of deaths (D), ASMR (E) and AAPC (F) in adults aged 65 + by sex. Number of deaths (G) in seven age groups of adults aged 65 + and the number of deaths (H), ASMR (I) and AAPC (J) by sex
Further analysis in adults aged ≥ 65 years showed that deaths and ASMR were higher in males than in females. Specifically, the number of deaths increased from 0.102 (95% UI: 0.022 to 0.180) million to 0.232 (95% UI: 0.047 to 0.400) million for males and from 0.085 (95% UI: 0.017 to 0.150) million to 0.187 (95% UI: 0.042 to 0.322) million for females (Fig. 1D). ASMR for males decreased significantly from 87.23 (95% UI: 18.99 to 153.26) per 100 000 to 75.69 (95% UI: 15.52 to 130.48) per 100 000 (Fig. 1E), with an AAPC of −0.328 (95% CI: −0.424 to −0.231) (Fig. 1F). For females, ASMR decreased from 48.84 per 100 000 (95% UI: 9.58 to 85.84) to 44.76 per 100 000 (95% UI: 9.98 to 77.05) (Fig. 1E), but the difference was not significant (AAPC −0.066 [95% CI: −0.135 to 0.003]) (Fig. 1F). Among the elderly, the number of deaths increased over time in all seven age groups (Fig. 1G). However, significant increases in mortality were found only in the 90–94 and 95 + age groups (AAPC 0.429 [95% CI: 0.111 to 0.747] and AAPC 1.23 [95% CI: 0.857 to 1.602], respectively) (Figs. 1H, 1I), and both were driven by increased female mortality (Fig. 1J).
Age-period-cohort analysis of COPD mortality
The effects of age, period and cohort on changes in COPD mortality attributable to ambient O3 were further analyzed (Fig. 2). Age-specific mortality rates in six periods of the calendar years between 1992 and 2021 for both sexes are shown in Fig. 2A. There was an overall increasing and then decreasing trend in age-specific mortality rates for both sexes within the past 30 years. Specifically, in the first four calendar year periods, age-specific mortality rates increased rapidly from ages of 65–69 to 90–94, and then decreased in ages of 95–99. In the last two calendar year periods, however, age-specific mortality rates increased steadily (Fig. 2A). Notably, for adults aged ≥ 75 years, birth cohort mortality rates increased and then decreased for both sexes, and increased more than decreased for those aged ≥ 90 years, but for adults aged 65–74 years, birth cohort mortality rates did not increase but continued to decrease slowly (Fig. 2B).
Results of age-period-cohort (APC) analysis for O3-related COPD mortality globally, 1990–2021. Age-specific mortality rates by period for both (A), females (G) and males (M). Cohort-specific mortality rates by period for both (B), females (H) and males (N). The local drift with net drift values of ASMR for both (C), females (I) and males (O). Longitudinal age curves of ASMR for both (D), females (J) and males (P). Period effects of ASMR for both (E), females (K) and males (Q). Cohort effects of ASMR for both (F), females (L) and males (R)
The net drift for both sexes was −0.5894324% (95% CI: −0.8009684%, −0.3774453%), and the local drift values were greater than 0 only for the age groups 90–94 (0.12% [95% CI: −0.34% to 0.59%]) and 95–99 (0.74% [95% CI: −0.39% to −0.89%]) (Fig. 2C). For both sexes, there was an almost linear upward trend in the effect of age on mortality rate until the age group 90–94, after which the mortality rate decreased (Fig. 2D). Using 2002–2006 as the reference period, the rate ratio increased before 1997–2001, then decreased and increased again after 2012–2016 (Fig. 2E). The cohort effect rose sharply in the earlier birth cohorts, but fell sharply after the 1917 birth cohort (Fig. 2F), which is also consistent with Fig. 2B. The effects of age, period and cohort on mortality for males were consistent with those for both sexes (Figs. 2M-R), whereas the effects for females were different, with the effects of age showing a consistent upward trend in mortality and no decline in the 95–99 age group (Figs. 2G, 2I), consistent with the results in Figs. 1H and 1 J, and the other effects being consistent with those for both sexes (Figs. 2H, J-L).
National (territorial) trends in COPD mortality attributable to ambient O3 exposure
In 1990, deaths were clustered in certain countries with large populations in Asia, North America and Europe (Fig. 3A). The top 5 countries were, in descending order, China (99,303 [95% UI: 21,087 to 174159] cases), India (34,934 [95% UI: 6726 to 63749] cases), the United States of America (8708 [95% UI: 1918 to 15042] cases), Pakistan (4328 [95% UI: 760 to 8005] cases) and the Russian Federation (3128 [95% UI: 680 to 5477] cases). These countries account for 80.2% of all deaths worldwide (Fig. 3B). In 2021, the number of deaths increased in most countries globally, but remained largely clustered in Asia, North America and Europe. In addition, some countries in Asia with low and low-middle SDI and in Europe with high-middle SDI also experienced increases in the number of deaths (Fig. 3A). The top five countries shifted to India (197,024 [95% UI: 42,918 to 336706] cases), China (117,036 [95% UI: 24,290 to 210694] cases), Bangladesh (12,336 [95% UI: 2440 to 22339] cases), the United States of America (12,315 [95% UI: 2612 to 21485] cases) and Pakistan (10,638 [95% UI: 2144 to 19117] cases), accounting for 83.3% of all deaths (Fig. 3B).
Distribution of global O3-related COPD deaths, age-standardized mortality rate (ASMR) and average annual percentage change (AAPC) of ASMR in 1990 and 2021. Distribution map of deaths (A) and ASMR (C). Scatter plot of deaths (B) and ASMR (D) in different countries by SDI region. AAPC of ASMR in different countries by 21 super-regions (E) and 5 SDI regions (F)
In terms of ASMR, most of the countries with high ASMR in 1990 were clustered in East Asia and South Asia (Fig. 3C). Of these, Nepal had the highest ASMR (249 [95% UI: 46 to 474] per 100 000), followed by China (230 [95% UI: 48 to 403] per 100 000), Democratic People’s Republic of Korea (143 [95% UI: 26 to 303] per 100 000), India (133 [95% UI: 26 to 245] per 100 000) and Kyrgyzstan (128 [95% UI: 25 to 226] per 100 000) (Fig. 3D). The vast majority of countries worldwide, particularly those with high-middle or high SDI, showed a declining trend in ASMR in 2021. However, South Asian countries with low or low-middle SDI showed a smaller decline in ASMR (Fig. 3C). In 2021, Nepal still had the highest ASMR (315 [95% UI: 63 to 559] per 100 000), followed by India (237 [95% UI: 52 to 406] per 100 000), Democratic People’s Republic of Korea (168 [95% UI: 30 to 351] per 100 000), Pakistan (153 [95% UI: 31,275] per 100 000) and Bhutan (146 [95% UI: 28,277] per 100 000) (Fig. 3D).
The AAPC analysis further confirmed that seven of the 21 global regions experienced an increase in ASMR from 1990 to 2021. Of these, South Asia experienced a large increase in ASMR (AAPC 3.051 [95% CI: 2.732 to 3.369]), while Southeast Asia, Eastern Sub-Saharan Africa, Southern Sub-Saharan Africa, Western Sub-Saharan Africa, Tropical Latin America and Andean Latin America experienced smaller increases (Fig. 3E). In addition, 10 regions showed a decrease in ASMR, with the largest decrease in East Asia (AAPC −4.604 [95% CI: −4.912 to −4.296]) and relatively smaller decreases in Central Latin America, Eastern Europe, Central Europe, Central Asia, North Africa and Middle East, high-income North America, Western Europe, Caribbean and Oceania (Fig. 3E). Regarding the distribution of countries with different SDIs, the higher the SDI classification, the higher the proportion of countries with decreasing ASMR. Conversely, the proportion of countries with increasing ASMR was greater in countries with lower SDI classifications. The top five countries with the largest increases were India (AAPC 3.397 [95% CI: 2.786 to 4.007]), Nepal (AAPC 2.297 [95% CI: 1.956 to 2.637]), Pakistan (AAPC 1.421 [95% CI: 1.117 to 1.724]), Bangladesh (AAPC 1.175 [95% CI: 0.972 to 1.378]) and Lesotho (AAPC 0.988 [95% CI: 0.808 to 1.168]) (Fig. 3F).
Trends of COPD mortality in various regions and countries (territories) based on SDI
From 1990 to 2021, the ASMR of COPD attributable to ambient O3 exposure was negatively correlated with the SDI in different regions (R = −0.1872, p = 1.065e-06 < 0.01) (Fig. 4A). Specifically, as the SDI value increased, the ASMR of the different regions first increased and then reached an inflection point around the SDI value of 0.456, after which it continued to decrease. Interestingly, around the inflection point, many regions showed very different changes in ASMRs than expected. For example, ASMRs in South Asia and East Asia were significantly higher than expected, but South Asia showed a steady increase while East Asia continued to decline (Fig. 4A). At the national level, ASMR was not associated with SDI across countries in 1990 (R = −0.0656, p = 0.3511 > 0.05) (Fig. 4B). In 2021, ASMR was negatively correlated with SDI across countries (R = −0.4314, p = 1.726e-10 < 0.001). Similar to the results in Fig. 4A, ASMR of countries initially increased slightly, with an inflection point around 0.456, after which the ASMR decreased with increasing SDI values. Around the inflection point, ASMR changed very differently than expected in a few countries. For example, ASMRs in Nepal, India, the Democratic People’s Republic of Korea, Pakistan and Bhutan were significantly higher than expected (Fig. 4C).
Correlation analysis of COPD deaths and ASMR attributable to ambient O3 exposure in different super-regions and countries based on SDI. A Correlation of SDI values with ASMR by super-region from 1990 to 2021.Correlation of ASMR with SDI values by country for 1990 (B) and 2021 (C)
Global trends in COPD mortality attributable to O3 by sex, age subgroup and SDI
Decomposition analysis was used to assess the impact of ageing, population growth and epidemiological changes on the epidemiology of O3-related COPD from1990 to 2021 (Fig. 5). The overall difference in ASMR showed an increasing trend globally and in all SDI regions, with the largest increases in low-middle SDI regions. Globally, ageing and population growth contributed 8.4% and 107.16% respectively to the increase in disease burden, while epidemiological changes contributed 15.55% to the decrease. In particular, the impact of epidemiological changes varied across SDI regions, with 34.2% and 37.08% in low and low-middle SDI regions, and −136.13%, −363.53% and −62.59% in middle, high-middle and high SDI regions, respectively (Fig. 5A). The impact of ageing was the largest contributor to the increase in middle SDI region (66.89%), followed by high-middle (23.84%), low-middle (9.08%) and low (1.03%) SDI regions, while it was the contributor to the decrease in high SDI region (−16.16%). Changes in population growth increased the burden of disease in different SDI regions, particularly in high-middle SDI region (439.69%), followed by high (178.75%), middle (169.24%), low (64.77%) and low-middle SDI regions (53.84%) (Fig. 5A). The overall difference increased in 18 of the 21 global regions, particularly in South Asia and East Asia, while it decreased in the other three regions, Eastern Europe, Central Europe and Central Asia. Specifically, the increase in burden was driven by ageing (18.02%), population growth (50.97%) and epidemiological changes (31.01%) in South Asia. In East Asia, ageing (289.05%) and population growth (538.35%) were responsible for increasing burden, whereas epidemiological changes (−727.4%) were responsible for decreasing burden. In Eastern Europe, Central Europe and Central Asia, active epidemiological changes were mainly responsible for the decreasing overall difference (Fig. 5B).
Population-level determinant changes in aging, population growth, and epidemiological changes for O3-related COPD deaths globally and in various SDI regions from 1990 to 2021. Black dots represent the total change contributed by all three components. A positive value for each component indicates a corresponding positive contribution in deaths, and a negative value indicates a corresponding negative contribution in deaths
Cross-country inequality analyses further revealed that there were clear absolute and relative inequalities associated with SDI, with a disproportionate burden falling on countries and territories with lower SDIs (Fig. 6). As shown by the SII, the gap in ASMRs between the highest and lowest SDI countries and territories increased from −2.08 (95% CI: −8.92 to 4.76) in 1990 to −19.45 (95% CI: −24.81 to −14.10) in 2021 (Fig. 6A). The CI was −0.11 (95% CI: −0.21 to −0.01) in 1990, but it increased to −0.27 (95% CI: −0.34 to −0.19) in 2021 (Fig. 6B). The results analyzed by sex were consistent with those for both sexes, with the range of variation in SII and CI being greater for males than females (Figs. 6C-F). These results suggest that inequalities in the global burden of O3-related COPD had been progressively worsen between 1990 and 2021, particularly among males. To further explore the potential for improvement in the burden of O3-related COPD according to the level of development of countries and territories, frontier analyses were also performed (Fig. 7). The 15 countries and territories with the largest actual differences in potential improvement were Nepal, India, the Democratic People’s Republic of Korea, Pakistan, Bhutan, Bangladesh, the Central African Republic, Myanmar, China, Bahrain, Kiribati, Afghanistan, the Democratic Republic of Congo, Lesotho and the Maldives (Fig. 7A-B). However, there were slight differences between females and males. For females, Lesotho was replaced by the United Arab Emirates among the 15 countries with the largest actual difference in potential improvement (Fig. 7C-D). For males, Afghanistan, the Democratic Republic of Congo and the Maldives were replaced by Namibia, Turkey and Kyrgyzstan (Fig. 7E-F).
Health inequality regression curves and concentration curves for O3-related COPD ASMR globally, 1990 and 2021. Panels A, C and E show the slope index of inequality (SII), depicting the relationship between SDI and ASMR for each condition, with points representing individual countries and territories sized by population. Panels B, D and F present the concentration index (CI), which quantifies relative inequalities by integrating the area under the Lorenz curve, matching the distribution of deaths to the distribution of population by SDI. Blue represents data from 1990, and red represents data from 2021
Frontier analysis exploring the relationship between SDI and O3-related COPD ASMR in 204 countries and territories. The color change from light blue (1990) to dark blue (2021) represents the change in years (A, C and E). Each point represents a specific country or territory in 2021 (B, D and F), the border line is shown in black and the 15 countries and territories with the largest differences from the border line are also shown in black. The direction of change in ASMR from 1990 to 2021 is indicated by the color of the points, with red points representing decreases and blue points representing increases
link