- Top of page
Background and objective: The harmful effect of dust storm on lung health is controversial. This study aimed to assess any associations between dust storms and emergency hospital admissions due to respiratory disease in Hong Kong.
Methods: Data on daily emergency admissions for respiratory diseases to major hospitals in Hong Kong, and indices of air pollutants and meteorological variables from January 1998 to December 2002 were obtained from several government departments. We identified five dust storm days during the study period. Independent t-tests were used to compare the mean daily number of admissions on dust storm and non-dust storm days. Case-crossover analysis using the Poisson regression was used to examine the effects of PM10 to emergency hospital admissions due to respiratory diseases.
Results: Significant increases in emergency hospital admission due to COPD were found 2 days after dust storm episode. The relative risk of PM10 for lag 2 days was 1.05 (95% CI: 1.01–1.09) per 10 µg/m3.
Conclusions: Dust storms have an adverse effect on emergency hospital admission for COPD in Hong Kong. This also suggests the adverse effect of coarse particles on lung health.
- Top of page
A dust storm is a meteorological phenomenon that arises when a gust front blows loose sand and dust off the surface of an arid or semiarid landscape. Soil particles are mass-transported by saltation and suspension, causing erosion from one place and deposition in another.1 The deserts of northern China are widely considered to be a major source of Asian dust,2 and dust storms originating from these regions have a far-reaching impact that has been reported not only within China3 but also in neighbouring countries such as Japan and Korea.4,5 Under certain weather conditions, wind-blown dust has travelled south towards Taiwan,6 and with a prevailing easterly wind, dust clouds can subsequently reach Hong Kong.7 Occasionally, dust storms may even travel across the Pacific Ocean to North America.8
Dust particles vary greatly in size. The largest are coarse particles, defined here as particulate matter ranging between 2.5 and 10 microns in aerodynamic diameter and commonly denoted by PM2.5–10. Fine particles (PM2.5) with aerodynamic diameter less than 2.5 µm are considered to be much more harmful. PM2.5 has frequently been the focus of air pollution and health studies.9,10 Smaller still are the ultrafine particles (PM0.1), with an aerodynamic diameter less than 0.1 µm. In a dust storm, the predominant fraction tends to consist of coarse particles,11 or the coarse fraction of PM10 (expressed as PM10–PM2.5). The adverse effects of coarse particles on morbidity and mortality have also been documented in several time series studies.12–14
Besides lowering visibility,15 dust storms can produce harmful effects on respiratory and cardiovascular health, as has been shown in several Asian studies.16–18 However, such an association has not been examined in Hong Kong, a subtropical metropolis in southern China. The aim of this study is to determine whether the risk of emergency hospital admissions due to respiratory diseases would be increased during and after dust storm episodes. We hypothesized that the mean number of hospital admissions for respiratory diseases would be higher during dust storms compared with that in the absence of dust storms.
- Top of page
There were a total of 462 308 emergency hospital admissions due to respiratory diseases over the study period. Among them, 111 419 admissions (24.1%) were due to COPD and 74 513 admissions (16.1%) were due to pneumonia and influenza.
Table 1 shows the means and standard deviations of daily admissions, pollutant concentrations and meteorological variables by dust storm (index) and non-dust storm days (non-index). Compared with all non-dust storm days in the 5-year period, the mean concentrations of NO2, O3, PM10 and PM2.5 were all significantly higher on dust storm days, as were the number of hospital admissions for all respiratory diseases, COPD, and pneumonia and influenza. When the comparison was limited to non-dust storm days in March and April during the study period, the mean concentrations of NO2, O3, PM10 and PM2.5 were again significantly higher on dust storm days, as were the hospital admissions for all respiratory diseases and for COPD. By contrast, the humidity was significantly lower on dust storm days. For PM10, the mean concentration on index days (134.3 µg/m3) was more than 2.5 times greater than on non-dust storm days in March and April (53.2 µg/m3), and around 2.2 times greater when compared with the control days (60.0 µg/m3).
Table 1. Mean (SD) number of daily hospital admissions, pollutant concentrations and metrological variables from January 1998 to December 2002, sorted by dust storm and non-dust storm days
| ||All respiratory diseases||COPD||Pneumonia and influenza||Temperature (°C)||Humidity (%)||NO2 (µg/m3)||O3 (8-hour) (µg/m3)||PM10 (µg/m3)||PM2.5 (µg/m3)||SO2 (µg/m3)||Coarse fraction† (µg/m3)|
|Dust storm days (n = 5)||328.2||80.2||55.4||21.2||71.8||79.6||62.5||134.3||59.9||19.2||74.4|
|All non-dust storm days (n = 1822)||253.0||61.0||40.8||23.7||78.0||57.7||37.0||49.9||35.2||17.2||14.9|
|Non-dust storm days in March and April (n = 301)||285.4||67.7||50.6||22.0||82.1||60.2||35.1||53.2||35.3||16.8||18.2|
Table 2 shows the adjusted relative risks (RR) of hospital admissions for all respiratory diseases and COPD associated with PM10 on dust storm days, in comparison with control days. The relative risks of emergency hospital admissions for COPD at lag 0, 1, 2 and 3 days were all greater than 1. The relative risk was 1.05 for 10 µg/m3 increase of PM10 at lag day 2 and was statistically significant (95% CI: 1.01–1.09). The relative risks of emergency hospital admissions for all respiratory diseases at lag 0, 1, 2 and 3 days were greater than 1 but were not significant. The relative risks for pneumonia and influenza were close to unity and not significant.
Table 2. Adjusted relative risks (RR)† of hospital admissions for a 10 µg/m3 increase in PM10
| ||Lag 0||Lag 1||Lag 2||Lag 3|
|No. of hospital admissions on dust storm days||80.2 (7.6)||85.6 (12.3)||85.4 (14.5)||73.0 (19.1)|
|No. of hospital admissions on control days||73.9 (15.7)||73.9 (14.3)||69.7 (11.8)||71.5 (13.5)|
|Relative risk of hospital admissions||1.06 (1.00, 1.12)||1.02 (0.97, 1.07)||1.05* (1.01, 1.09)||1.00 (0.95, 1.06)|
|Pneumonia and influenza|
|No. of hospital admissions on dust storm days||52.4 (9.4)||58.2 (6.1)||57.2 (55.8)||55.8 (7.8)|
|No. of hospital admissions on control days||55.0 (9.3)||53.7 (6.1)||58.5 (8.8)||54.4 (7.8)|
|Relative risk of hospital admissions||0.99 (0.93, 1.05)||1.02 (0.99, 1.06)||0.96 (0.92, 1.00)||1.02 (0.99, 1.05)|
|All respiratory diseases|
|No. of hospital admissions on dust storm days||328.2 (27.7)||338.6 (32.4)||328.0 (58.6)||320.6 (56.8)|
|No. of hospital admissions on control days||314.1 (45.4)||307.8 (36.5)||305.6 (33.2)||313.2 (41.9)|
|Relative risk of hospital admissions||1.03 (0.99, 1.08)||1.02 (0.99, 1.06)||1.02 (0.99, 1.06)||1.01 (0.97, 1.05)|
- Top of page
Dust storms have been reported in ancient literature23,24 and have long been suspected of producing adverse effects on health.25 This study examined whether dust storms affected emergency hospital admissions due to respiratory diseases in Hong Kong. We have shown that the number of daily emergency hospital admissions due to COPD at 1 and 2 days after the index days were significantly higher than when compared with those on control days.
While fine particles (PM2.5) usually dominate the airborne particulates count on normal days, the predominant particulates during dust storms tend to be coarse particles.11 Dust storms thus provide good opportunities for us to examine the effect of coarse particles on lung health.
Most studies on air pollution and health focus on fine particulates of transport-related origin. Their small size implies deep penetration into the lung parenchyma. Furthermore, their complex chemical compositions, which may include a surface of toxic metals and polycyclic aromatic hydrocarbons, have been incriminated in the pathogenesis of several respiratory and cardiovascular diseases.26–28 Coarse particles, on the other hand, are mostly composed of crustal elements and considered to be less toxic. Nevertheless, they have been shown to trigger inflammatory processes, such as inducing cytokine release both in vitro and in animals.29–31 Becker et al. reported that coarse particles had larger effects than fine and ultrafine particles in inducing inflammatory mediators; they suggested that the effects were linked with the presence of microbial cell structures and endotoxins.32 Experiments on rats showed significant suppression of alveolar macrophage function and increased epithelial permeability after inhalation of coarse particles produced by re-suspending road dusts.33 The triggering of immunological response in the respiratory tract by high concentrations of coarse particles is a plausible mechanism that could explain the increase in hospital admissions for COPD patients during dust storms. A recent study in Korea has shown that Asian dust storm particles can induce toxicological effects on human skin through the activation of cellular detoxification system and production of various cytokines.34 Although many experimental studies had shown adverse effects of coarse particles on health, limited epidemiological studies are available in the literature.13,35
Studies on dust storms have focused on their adverse effects on health. Hefflin and colleagues reported a 4.5% increase in the number of emergency room visits 2 days after the PM10 levels exceeded 150 µg/m3.36 Recently, a significant association between Asian dust storms and hospital admissions due to pneumonia after one lag day was reported in Taipei,37 while the concentration of ambient influenza A virus was significantly higher during the dust storm days than during the background days in Kaohsiung City, Taiwan.38 Positive but insignificant associations between Asian dust storms and hospital admissions for COPD,18 asthma16 and cardiovascular diseases39 have been reported in Taiwan. However, time series in Hong Kong demonstrated significant and positive associations between PM10, and the risk of hospital admissions for COPD.40,41 Heavy dust events have recently been reported to be associated with an increased risk of asthmatic hospitalizations in children in Toyama, Japan.42
Our study provides further evidence of the adverse effects of dust storms or, more specifically, coarse particles on health. Having adjusted for other pollutants and confounding variables, there was a 5% increase in daily emergency hospital admissions per 10µg/m3 increase in PM10 due to COPD 2 days after a dust storm. The mean concentration of PM10 on dust storm days was 134.3 µg/m3—much higher than the mean concentration on non-dust storm days (53.2 µg/m3) in March and April. As coarse particles dominate PM10, this result can be interpreted as a 5% increase in daily COPD admissions per 10 µg/m3 increase in the concentration of coarse particles.
Meng and Zhang43 suggested that particulates from dust storms could cause oxidative damage in rat internal organs, although their study focused mainly on fine particulates. Schwartz et al.35 were dismissive of the role coarse particles may play in affecting human mortality—a conclusion partly supported by Brunekreef and Forsberg in their review,44 which reported that coarse particulates were not significantly associated with long- or short-term mortality in most urban areas, with only a few exceptions. However, the same review also mentioned that coarse particulate matter was associated with a strong short-term effect on respiratory health, including asthma and COPD, which can lead to increased hospital admissions. This conclusion is supported by our findings. Epidemiological studies by Ostro et al.45,46 have also found associations between PM10 and coarse particulates, and cardiovascular mortality.
Because the concentration of coarse particulates during a dust storm can reach levels several times higher than normal, the impact on community morbidity and the demand on hospital beds can be considerable. A better system for warning the public of impending dust storms is warranted. Specifically, those with chronic lung diseases should be advised to reduce outdoor exposure.
One advantage of studies on dust storms is that the high wind speed during such storms could diminish the concentration of other combustion-related pollutants, thus allowing us to differentiate more clearly the effect of the coast fraction of PM10 from that of traffic-generated PM2.5. We have also adjusted for the differences in the concentrations of PM2.5 between dust storm days and non-dust storm days. Hence, we can attribute the effect on hospital admissions for COPD to the coarse fraction of PM10 unambiguously. As for the other air pollutants, no positive significant effects were observed for hospital admissions due to ‘any respiratory diseases’, and ‘pneumonia and influenza’. Significant positive effects were observed on lag 3 days for COPD admissions with O3 (RR = 1.07; 95% CI: 1.02–1.12) and SO2 (RR = 1.15; 95% CI: 1.04–1.27). PM10 had no significant effect on lag 3 days.
One major limitation of our study is that dust storms episodes during the study period were rare; hence, statistical significance could not be reached even when all RRs were positive for any respiratory diseases and COPD. For comparison, we also conducted similar analyses on hospital admissions due to ‘any cardiovascular diseases’ (ICD: 390–459), ischaemic heart disease (ICD: 410–414), heart failure (ICD: 428) and cerebrovascular disease (ICD: 430–438). No significant differences were found in the number of hospital admissions with these diagnoses between dust storm days and non-dust storm days.
In conclusion, our study has demonstrated that dust storms affecting Hong Kong can cause a significant increase in local emergency hospital admissions due to COPD 2 days after the event. This has added new evidence to the adverse effects of coarse particles on pulmonary health. Timely health warning to those with chronic lung diseases to reduce outdoor activities during dust storm days may reduce their need for hospital admissions.