Background
Epidemiological evidence has shown that both short-term and long-term exposure to air pollution is associated with kidney outcomes [
1‐
7]. A few time-series studies reported positive associations between short-term exposure to air pollution and emergency department visits or hospitalizations for kidney diseases in North American and East Asian populations, and the risk ratios ranged from 1.010 to 1.034 [
1‐
3]. The associations of long-term exposure to air pollution with incident kidney diseases and kidney function decline were examined by several large national cohort studies from North America and China, and the magnitude of effect sizes was larger: the hazard ratios or odds ratios ranged from 1.07 to 1.39. Although these previous studies have reported the positive associations between air pollution and renal outcomes, the relationship between short-term air pollution exposure and mortality of kidney disease has not yet been investigated.
Air pollution, especially those with smaller aerodynamic diameter, may be inhaled into the alveoli, penetrate through biological membranes, and enter the blood stream [
8]. People with kidney diseases may be particularly vulnerable to these hazardous pollutants since kidneys filter around 20% of cardiac output, and a decline in kidney function may result in retainment of environmental toxins in the blood [
9]. Animal-based experimental studies proposed several biological mechanisms that support the detrimental effects of air pollution exposure on the increased risk of kidney dysfunction and mortality, including inflammation and oxidative stress [
10‐
14]. These biological mechanisms, supported by several large epidemiological studies that also report the risk of increased mortality [
15‐
18], justify the plausibility of testing the short-term effects of air pollution on death from kidney diseases in humans.
Time-stratified case-crossover study design has been widely used to estimate acute effects in air pollution epidemiology [
17‐
19]. Controls in this design were simulated on multiple days prior to and after each case day at the same location. Each case serves as its own control to reduce individual-level time-invariant confounding; environmental variables were the factors that vary over time for each person. Although most kidney diseases are chronic in nature, death is typically acute. Thus, a time-stratified case-crossover design is appropriate in estimating the triggering effects of short-term exposure to air pollution on death due to kidney diseases.
In addition to China having the largest number of chronic kidney disease cases (132.3 million) in the world in 2017 [
20], the pollution level is well above the reference guidelines proposed by the World Health Organization (WHO) [
21‐
23]. Not only does investigating the relationship between short-term air pollution exposure and death from kidney diseases in China examine this untested hypothesis, but it also offers insights on the burden of death in a country where both factors are extremely prevalent. In this study, we collected a national sample of 101,919 deaths due to kidney diseases in China from 2015 to 2019. We aim to quantify the association between short-term exposure to air pollution and mortality of kidney diseases using a time-stratified case-crossover study design.
Discussion
In this nationally representative sample of deaths from kidney diseases in China spanning five years, our time-stratified case-crossover study suggests that short-term exposures to PM1, PM2.5, PM10, NO2, and SO2 were associated with a significantly elevated risk of death, exhibited by a nonlinear concave-down pattern. The results were generally consistent when the exposure was measured at different lag periods and in two-pollutant models.
This was the first nationwide study to examine the relationship between short-term exposure to air pollution and mortality from kidney diseases in China, where both large burden of kidney diseases and air pollution exist. Although air pollution in China has experienced a substantial and steady decline in recent years [
45‐
47], the current level of particulate matter air pollution is still above the 2021 WHO Air Quality Guideline [
23]. For example, the medians of particulate matter air pollutants on case days in this study were 34.86 µg/m
3 for PM
2.5 and 60.49 µg/m
3 for PM
10, both substantially higher than the WHO guideline, in which 24-h pollutants were recommended to be lower than 15 µg/m
3 and 45 µg/m
3 respectively. The findings of this study deepen our understanding that elevated air pollution is associated with a small but significant excess in burden of death from kidney disease in China.
Although this epidemiological case-crossover study does not shed mechanistic insights on the link between short-term air pollution exposure and mortality among people with kidney disease, several postulated biological mechanisms may help explain the observed associations in this study. The most widely cited mechanisms are inflammation and oxidative stress, suggested by laboratory experiments and population-based epidemiological evidence [
12‐
14]. Elevation of short-term air pollution may increase pulmonary inflammation and cause damage to distant organs including the kidney, and further trigger death among a vulnerable population [
9]. These biological mechanisms need further studies supported by animal-based laboratory evidence or epidemiological studies that delve deeper into the chemical components or different sources of air pollution [
9,
22].
Compared to previous cohort studies using long-term exposure to air pollution [
48] or aggregated time-series studies [
15], the effect estimates in this study were relatively small (< 3% per 10 µg/m
3 increase); these small estimates are within our expectations given the time-stratified case-crossover design. For example, a large time-stratified case-crossover study in the US Medicare population showed that each 10 µg/m
3 increase in PM
2.5 was associated with a 2.3% (95% CI: 1.7% to 3.0%) relative increase in the risk of acute and unspecified renal failure hospital admission [
3]. Other time-series studies that investigated associations between short-term exposure to air pollution and emergency department visits for kidney diseases in North American and East Asian population reported relative percent changes ranging from 10% to 3.4% [
1,
2]. Another time-stratified case-crossover study in Hubei province of China estimated that each 10 µg/m
3 increase in PM
2.5, PM
10, and NO
2 was associated with a 4.14% (95% CI: 1.25% to 7.12%), 2.67% (95% CI: 0.80% to 4.57%), and 1.46% (95% CI: 0.76% to 2.17%) increase in odds of myocardial infarction mortality [
17]. The magnitude of these effect estimates in previous papers utilizing the same time-stratified case-crossover or time-series study design was comparable to the findings in our study. In addition, we found that the relative percentage increases per 10 µg/m
3 increment for particulate matter air pollution with smaller particle sizes were greater than those with larger particle sizes were also consistent with findings in several other studies on air pollution and mortality [
33,
49].
This study has several strengths. A large nationwide sample of deaths collected from over 600 surveillance sites in China spanning the past five years yields nationally representative estimates on the association between short-term exposure to air pollution and risk of death from kidney diseases. The levels of air pollution, notably particulate matter air pollution, were considerably higher than the WHO-recommended levels, providing us an opportunity to investigate our hypothesis in an ideal setting. Since the time-stratified case-crossover design uses the individual as its own control, it eliminates the possibility of unknown individual-level time-invariant residual confounding including access to renal replacement therapy and genetic risk factors [
50,
51]. Compared to other time-series studies that measured air pollution at city levels, our study ascertained air pollution at the individual level using residential latitudes and longitudes, producing more accurate estimates of air pollution.
This study has several limitations. Air pollution was estimated only at residential addresses and did not account for place of employment, and thus, we were unable to obtain a fully comprehensive measurement of air pollution for each case. This may lead to potential exposure misclassification bias, but this is likely a minor issue since the physical activity of patients with kidney problems is limited, and it is unlikely that many individuals were far from their residential addresses for very long. The current CHAP air pollutant data were estimated at 10*10 km grids and had relatively low spatial resolution, this may lead to exposure misclassification and the estimates may be biased towards null. Indoor air pollution subject to cooking, solid fuel use, and smoking is another important source of air pollution that may contribute the death from kidney diseases [
52], but this information was unavailable in the study. The exposure source data only measured air pollution as a whole and did not account for the chemical composition of air pollution, and future studies should take advantage of advances in remote sensing technologies and evaluate the effects of chemical components of air pollution [
53]. The data collection system included prespecified choices for categorical variables, and some detailed sub-categories may not be collected in the original data (for example, illiteracy for education) [
54]. Although the time-stratified case-crossover design eliminates time-invariant factors and we accounted for meteorological variables, it may still be subject to residual confounding caused by unmeasured time-varying factors. The case-crossover design generally has lower statistical power compared to a time-series study design [
38], and we were underpowered to detect statistically significant heterogeneity across different subgroups. The findings based on the Chinese population may not be generalizable to other countries due to population dissimilarity and differing levels of air pollution.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.