Study region (time period) | Study design | Results |
Asthma | ||
This study Taipei (1995–2002) | Poisson time-series modelling. Sandstorms modelled through multiple indicators | Associations with L1, L2 and L03 O3 and L03 PM10 |
Taipei (1996–2001)14 | Dust storm event defined as PM10 >125 μg/m3 at Yangmingshan station. Compared admissions on storm days to those a week before and week after. Considered lags of 0, 1, 2 and 3 days | Risk of admissions 8% higher two days after storm; results not statistically significant |
Taiwan (1998–2001)23 | Auto-regressive integrated moving average time-series approach | Seasonal trends in admissions associated with monthly mean PM10, SO2, CO and NO2 |
Kaohsiung (1996–2003)24 | Case-crossover design. Compared air pollution on date of admission with pollutant levels one week before and after. Used cumulative lag of same day and up to two days previous | PM10, SO2, NO2, CO and O3 associated with admissions on cool days (<25°C). All but SO2 also associated on warm days (⩾25°C). Largest association for NO2 on cool days |
Taipei (1 year)25 | Cross-sectional study based on monthly hospital admissions for respiratory illness | NO2 and PM10 levels significantly correlated with admissions. No consistent association with O3 or SO2 |
Pneumonia | ||
This study Taipei (1995–2002) | Poisson time-series modelling. Sandstorms modelled through multiple indicators | No statistically significant associations |
Cerebrovascular | ||
This study Taipei (1995–2002) | Poisson time-series modelling. Sandstorms modelled through multiple indicators | Associations with PM10 and CO at three day lags |
Kaohsiung (1997–2000)26 | Outcome was stroke (ICD-9 430 to 438) | PM10, NO2, SO2, CO, CO and O3 associated with intracerebral haemorrhage (ICD-9 431 and 432) and ischaemic stroke admissions (ICD-9 433 and 435) on warm days (⩾20°C). CO associated with ischemic stroke admissions on cool days (<20°C) |
Case-crossover approach. Compared air pollution on date of admissions to days a week before and week after. Considered lags up to 2 days | ||
Taipei (1996–2001)17 | Outcome was stroke (ICD-9 430 to 438) | Non-statistically significant results for risk of stroke admissions for dust storms compared to control days at a three-day lag. Significant associations for primary intracerebral hemorrhagic stroke (ICD-9 431 and 432) |
Dust storm event defined as PM10 >125 μg/m3 at Yangmingshan station. Compared admissions on storm days to those a week before and week after. Considered lags of 0, 1, 2 and 3 days | ||
Taipei (12 April 1997–31 December 2002)27 | Outcome was emergency admissions for those >50 years (ICD-9 430–437) | Effects for O3 at lag 0, CO at lag 2, PM2.5 at lag 3, and PM10 at lag 3 days |
Poisson time-series model | ||
Cardiovascular | ||
This study Taipei (1995–2002) | Outcome was ischemic heart disease (ICD-9 410, 411, and 414). Poisson time-series modelling. Sandstorms modelled through multiple indicators | Associations with same day NO2 and CO, and with L03 PM2.5. Also associated with presence of sandstorms as indicated by high PM10 levels overall and at a background station (L3), coarse PM (L0, L1, and L03), and the ratio of PM10 to PM2.5 (L3) |
Kaohsiung (1997–2000)28 | Outcome was cardiovascular disease (ICD-9 410–429) | PM10, NO2, CO and O3 associated with admissions on warm days (⩾25°C). All but O3 also associated on cool days (<25°C) |
Case-crossover approach. Compared air pollution on date of admissions to days a week before and week after. Considered lags up to 2 days | ||
Taipei (1997–2001)29 | Outcome was cardiovascular disease (ICD-9 410–429) | PM10, NO2, CO and O3 associated with admissions on warm days (⩾20°C). PM10, NO2 and O3 associated on cool days (<20°C) |
Case-crossover approach. Compared day of admissions to days one week before and 1 week after. Considered cumulative lag up to two days previous | ||
Taipei (1996–2001)16 | Outcome was cardiovascular disease (ICD-9 410–429) | Risk of admissions increased 3.65% on days following a storm; results not statistically significant |
Dust storm event based on PM10 levels at Yangmingshan station. Compared admissions on storm days to those a week before and week after. Considered lags of up to 3 days |
Associations are statistically significant, unless otherwise noted. ICD-9 codes match those used in this study, except where otherwise specified.