Utah Valley, USA, Dec 1989 to Mar 199029 | Unselected, aged 9–11 (41) | PM10 195 μg.m−3 | NO2 (note: very low ozone and undetectable PSA and H2SO4) | Regression of daily group symptom prevalence (logistic model) and mean of individuals’ daily PEF deviations (linear model) | Linear trend, temperature, autocorrelation (GEE approach for logistic models) |
Utah Valley, USA, 1990 to 199130 | Symptomatic or asthmatic (39) and non-symptomatic (40), aged 9–11 | PM10 251 μg.m−3 | | | |
The Netherlands, winters of 1987–88 to 1989–9026 | Unselected, aged 7–11 (up to 390) | PM10 174 μg.m−3 | SO2 (105 μg.m−3), NO2, other aerosols including HNO2 (10.6 μg.m−3), SO42− (29.7 μg.m−3), NO3− (36.8 μg.m−3), PSA | Regression of daily group symptom incidence/prevalence (logistic model) and mean of individuals’ daily PEF z-scores (linear model) | Linear trend, temperature, autocorrelation (1st order) |
The Netherlands, winter 1990 to 199128 | Symptomatic, aged 6–12 (73) | | | | |
Mexico City, 1991 to 199240 | Asthmatics, aged 5–13 (71) | PM10 363 μg.m−3, PM2.5 177 μg.m−3 | NO2, SO2 (200 ppb), hourly maximum ozone (370 ppb) | Regression of daily group symptom prevalence/medication use (logistic model) and mean of individuals’ daily PEF z-scores (linear model) | Linear trend, temperature, autocorrelation (GEE approach for logistic models) |
Eastern Europe (Erfurt and Weimar, Germany and Sokolov, Czech Republic), winters 1990 to 199227,41,44,49 | Asthmatics, aged 6–15 (up to 163) | Sokolov: PM10 171 μg.m−3, number and mass conc. of fine (0.1–2.5 μm) and ultrafine (10–100 nm) particles | Sokolov: SO2 (492 μg.m−3), NO2, SO42− (30.2 μg.m−3), PSA (7.71 μg.m−3) | Regression of daily group mean of individuals’ deviation in log symptom scores, log medication use and PEF (linear models with weighted pooling by location) | Linear and quadratic trends, weather variables, weekend indicator, autocorrelation (1dt order) |
Six Cities, USA, 1984 to 198824,38 | Unselected, aged 6–10 (1844) | PM10 117 μg.m−3, PM2.5 86 μg.m−3 | SO2 (82 ppb), NO2 (44 ppb), ozone (8 hour mean 87 ppb), SO42− (15 μg.m−3), PSA (372 nmol.m−3) | Regression of daily group symptom incidence with weighted pooling by location (logistic model with weighted pooling by location) | Complex non-linear trends, temperature, autocorrelation (GEE approach) |
Uniontown, Pennsylvania, USA, summers 1990 and 199120,24 | Symptomatics (71) and non-symptomatics (27), aged 9–11 | PM10 83 μg.m−3, PM2.1 88 μg.m−3 | 12 hour average SO2 (44.9 ppb), ozone (87.5 ppb), SO42− (515 nmol.m−3), PSA (676 nmol.m−3) | Regression of daily group symptom prevalence (logistic model) and weighted mean of individual’s daily PEF deviations (linear model). Uniontown results weighted by time spent outdoors | Linear trend, temperature, autocorrelation (2nd order in linear models, GEE approach for logistic models) |
State College, Pennsylvania, USA, summers 1990 and 199121,24 | Symptomatics (62) and non-symptomatics (46), aged 9–11 | PM10 82.7 μg.m−3, PM2.1 85.8 μg.m−3 | 12 hour average SO2 (26.5 ppb), ozone (92.3 ppb), SO42− (481 nmol.m−3), PSA (670 nmol.m−3) | | |
Surrey, UK, summer 199423 | Unselected (9% with recent wheeze), aged 7–11 (154) | PM10 150 μg.m−3 | NOx, SO2, ozone (8 hourly mean 128 ppb) | Two-stage weighted pooling of subject specific coefficients derived from regression of each subject’s daily spirometry (linear model) | Linear trend, weather variables, autocorrelation (1st order) |
The Netherlands, summer 199519 | Chronically ill children (77% taking anti-asthma medication), aged 7–13 (61) | PM10 60.3 μg.m−3 | Ozone (8 hour mean 111 ppb) | Two-stage weighted polling of subject specific coefficients derived from regression of each subject’s daily symptom prevalence (logistic model) and PEF (linear model) | Linear trend, temperature, weekend indicator, autocorrelation (1st order) |
Vancouver, Canada, 1990 to 199232 | Asthmatic (132) and non-asthmatic (74), aged 6–13 | PM10 159 μg.m−3 and number concn for various particle fractions | | Two-stage weighted polling of subject specific coefficients derived from regression of each subject’s daily symptom prevalence (logistic model) and PEF (linear model) | Linear trend, weather variables, autocorrelation (GEE approach for logistic models) |
The Netherlands, winters 1992–93 to 1994–9525,43 | Symptomatics (396) and non-symptomatics (399), aged 7–11, with and without bronchial hyperreactivity or increased IgE levels | PM10 112 μg.m−3 urban and 242 μg.m−3 rural | SO2 (urban 152 μg.m−3, rural 43 μg.m−3), NO2, SO42− (urban 24 μg.m−3, rural 23 μg.m−3) | Regression of daily group symptom prevalence (logistic model), mean of individuals’ PEF deviations (linear model) and daily group prevalence of 10% and 20% falls in PEF (logistic model) | Linear and other trends, temperature, weekend indicator, autocorrelation (1st order) |
California, USA, Autumn 199534 | Asthmatics, aged 9–17 (25) | PM10 54 μg.m−3 | Ozone (8 hour mean 110 ppb) | Two-stage weighted polling of subject specific coefficients derived from regression of each subject’s daily symptom prevalence/severity (logistic model) and PEF (linear model) | Linear trend, temperature, autocorrelation (1st order, GEE approach for logistic models) |
Paris, France, 199231 | Asthmatics, aged 7–15 (84), stratified by inhaled corticosteroid dose | PM13 95 μg.m−3 | SO2 (83.8 μg.m−3), NO2 (note ozone excluded as very low levels only) | Regression of group daily symptom incidence/prevalence (logistic model) and mean of individuals’ % change in PEF z-scores (linear model) | Linear trend, weather variables, autocorrelation (GEE approach for logistic models) |
Mexico City, 1991 (3 periods of 59 days)35 | Symptomatics (21) and non-symptomatics (19), aged 11 | PM10 87 μg.m−3, PM2.5 69 μg.m−3 | Ozone (8 hour mean 103 ppb) | Regression of group daily symptom counts (Poisson model) and mean of individuals’ PEF deviations (linear model) | Linear and other trends, weather variables, autocorrelation (1st order) |
Summer Camps in Philadelphia, USA, summer 199322 | Unselected (15% asthmatic), aged 6–11 | PM10, PM2.5 (50 μg.m−3) | Ozone (12 hour mean 106 ppb), SO42−, PSA (1.9 μg.m−3) | Regression of mean individuals’ % change in PEF x300 (linear model) | Linear trend, temperature, autocorrelation |
Sydney, Australia, 199436,50 | Symptomatics, aged 7–10 (148) | 15 hour average (day time) PM10 123 μg.m−3 (peaked at 210 μg.m−3 during bushfires) | 15 hour average (day time) NO2 (47 ppb), ozone (43 ppb) | Regression of mean individuals’ deviation from personal mean PEF x300 (linear model) | Linear, quadratic and seasonal trends, weather variables, autocorrelation (1st order, GEE approach) |
14 centres in Europe (i.e. 28 panels–PEACE study), winter 1993–943,45,51,52 | Symptomatics, aged 7–11 (up to 75 in each panel –2010 in total) | PM10 peaks at each location varied from 29 to 242 μg.m−3 | SO2, NO2 | Regression of group daily symptom incidence/prevalence (logistic model) and mean of individuals’ % deviation from personal mean PEF (linear model) | Linear and other trends, temperature, weekend indicator, autocorrelation (1st order) |
Subset of 17 PEACE panels42 | | Particulate silicon and iron mass conc. | | | |
Bangkok, Thailand, Nov 1995 to May 199633 | Unselected, aged 8–12 (79) | PM10⩽242 μg.m3, PM2.5⩽122 μg.m3 | | Regression of group daily symptom incidence/prevalence (logistic model) | Loess smoothing to model trend, weather variables, autocorrelation |
California, USA, August to October 199337 | Treated asthmatics, aged 8–13 (138) | PM10 119 μg.m3, PM2.5 (as 12 hour average) 209 μg.m3 | Ozone (8 hour average 134 ppb), NO2 (as hourly maximum) | Regression of group daily symptom incidence/prevalence (logistic model) | Linear trend, weather effects, weekend indicator, autocorrelation (GEE approach) |
West Midlands, UK, winter and summer 199739 | Unselected, aged 9–10 (162, of which 39 symptomatic) | PM10 46 μg.m3, PM2.5 37 μg.m3 | Ozone 41 ppb, NO2 35 ppb, SO2 18 ppb, SO42− 14.9 μg.m3, NO3− 29.9 μg.m3, HNO3 3.8 μg.m3, HCl 1.7 μg.m3, NH4+ 15.5 μg.m3, H+ 12.7 ng.m3 | Regression of group daily symptom incidence/prevalence (logistic model) and mean of individuals’ % deviation from personal mean PEF (linear model) | Linear trend, weather variables, weekend indicator, autocorrelation |