American Journal of Epidemiology
Copyright O 1996 by The Johns HopWns University School of Hygiene and Public Health
All rights reserved
Vol. 143, No. 8
Printed in U.SA
Fungus Spores, Air Pollutants, and Other Determinants of Peak Expiratory
Flow Rate in Children
Lucas M. Neas,1i2p3 Douglas W. Dockery,1'3 Harriet Burge,1 Petros Koutrakis,1 and Frank E. Speizer1'3
Trie impact of summertime haze episodes on daily variations in symptoms and peak expiratory flow rates
(PEFRs) was examined in a study of 108 children living in State College, Pennsylvania, during the summer of
1991. Twice daily, each child recorded symptoms, PEFRs, and hours spent outdoors. Environmental measurements included daily 12- and 24-hour averages for meteorologic and air pollutant variables and 24-hour
average fungus spore concentrations. A10,000-spore/m 3 increment in Cladosporium spore concentration was
associated with a deficit in morning PEFR (-1.0 liters/minute, 95% confidence interval (Cl) -1.9 to -0.2). A
60-spore/m3 increment in Epicoccum spore concentration was associated with increased incidence of
morning cough (odds ratio (OR) = 1.8, 95% Cl 1.0-3.2) and a deficit in morning PEFR (-1.5 liters/minute, 95%
Cl -2.8 to -0.2). Fungi spore counts were not associated with respirable particle mass. A 125-nmol/m 3
increment in 12-hour daytime particle-strong acidity was associated with a deficit in evening PEFR (-0.5
liters/minute, 95% Cl - 1 . 2 to 0.2) and increased incidence of cold episodes that evening or the subsequent
morning (OR = 1.35, 95% Cl 1.14-1.61). A 20-/ig/m 3 increment in 24-hour respirable particles lagged by 24
hours was associated with a deficit in evening PEFR (-0.5 liters/minute, 95% Cl -1.4 to 0.4) and increased
incidence of cough episodes that evening or the subsequent morning (OR = 1.37, 95% Cl 1.13-1.66). These
results confirm the acute effects of summertime particulate air pollution and suggest that aeroallergens have
independent effects on respiratory symptoms and PEFR in children. Am J Epidemiol 1996;143:797-807.
air pollution; child; fungi; peak expiratory flow rate
Respirable particulate matter from combustion
sources has been associated with decreased level of
pulmonary function (1,2) and increases in respiratory
symptoms (3-5), hospital admissions (6), and school
absenteeism (7, 8) in previous studies of US children.
Similar results have been found in recent crosssectional studies of children in Europe (9-11) and
Canada (12).
Previous reports from the Harvard Six Cities Study
have shown that chronic exposure to respirable particulate matter was associated with increased bronchitis
and decreased pulmonary function levels among chil-
dren in six US cities (13-15). Additional follow-up of
the children in these six cities and one additional area
demonstrated similar effects for particle-strong acidity
(16). A recent investigation of children in 24 different
communities in the United States and Canada confirmed that annual levels of respirable particulate matter and particle-strong acidity were associated with
increased bronchitis (17) and decreased forced vital
capacity (18).
Summertime episodes of high particle-strong acidity
and ozone provided opportunities to supplement the
cross-sectional data with longitudinal data on shortterm exposure-response effects in children in several
of these 24 communities. A summer study of children
in Uniontown, Pennsylvania (19), found that summertime episodes of aerosol acidity and particulate sulfate
pollution were acutely associated with declines in peak
expiratory flow rates and increased incidence of cough
episodes.
In most areas in the eastern United States, the seasons of high ozone and acidic pollution also include
episodes of high exposures to biologic aerosols (including pollen and fungus spores). Lehrer et al. (20)
report a significant association between skin reactivity
to extracts of basidiospores that are common in out-
Recetved for publication March 8,1995, and in final form July 28,
1995.
Abbreviations: Cl, confidence interval; FEV, Q, forced expiratory
volume in 1 second; OR, odds ratio; PEFR, peak expiratory flow
rate; P M 2 , particulate matter with an aerodynamic diameter equal
to or less than 2.1 microns; PM 10 , particulate matter with an aerodynamic diameter equal to or less than 10 microns.
1
Environmental Epidemiology Program, Department of Environmental Health, Harvard School of Public Health, Boston, MA.
2
Department of Epidemiology, Harvard School of Public Health,
Boston, MA.
3
Channing Laboratory, Department of Medicine, Brigham and
Women's Hospital, Harvard Medical School, Boston, MA.
Reprint requests to Lucas M. Neas, Department of Environmental
Health, Harvard School of Public Hearth, 665 Huntington Ave.,
Boston, MA 02115.
797
798
Neas et al.
door air and atopy, asthma, and asthma with rhinitis,
but not rhinitis alone. The incidence of asthma has also
been associated with skin reactivity to Alternaria
(odds ratio (OR) = 5.1, 95 percent confidence interval
(CI) 2.9-8.9) (21). Exposure to Alternaria allergens
can result from both indoor and outdoor sources, and
O'Hollaren et al. (22) suggest that sensitization to
Alternaria allergens in young asthmatics is a risk
factor for respiratory arrest. Outdoor exposures to fungal spores has not been studied as a determinant of
specific pulmonary function changes in a nonclinical
population of schoolchildren.
The present investigation has replicated the previous
Uniontown summer study in a cohort of children in
State College, Pennsylvania, and has included daily
measurements of specific pollen and fungus spore
concentrations.
MATERIALS AND METHODS
Site description
State College, Pennsylvania, was selected from a
larger study of acid aerosols in 24 cities throughout the
United States and Canada (23). High levels of acid
aerosols in State College measured during previous
summers are attributable to the transport and transformation of sulfur emissions from the Ohio River Valley
(200 km to the west), Pittsburgh (180 km to the west),
and the Monongahela River Valley (240 km to the
southwest). State College is a rural community with no
major local point sources of air pollution.
Meteorologic and pollutant measurements
Precipitation, solar radiation, and wind direction
were measured by the Environmental Protection
Agency Clean Air Status and Trends Network at the
Rock Springs site within the city limits of State College. The hourly meteorologic data were summarized
over a 24-hour period (8:00 p.m. to 8:00 p.m.). Binary
indicators were created for rain days (any day with at
least 1 mm of precipitation) and sunny days (any day
with more than 67 percent of the maximum observed
solar radiation).
Pollutants were measured from June 10 through
August 26, 1991, at the National Dry Deposition Network Site located 6 km west of the center of State
College. Particle-strong acidity and total sulfate particles (in nmol/m3) were measured by an annular
denuder/filter pack sampler over two 12-hour periods:
daytime (8:00 a.m. to 8:00 p.m.) and overnight (8:00
p.m. to 8:00 a.m.) (24, 25). Temperature and the
concentrations of the continuously monitored air pollutants are expressed as their average values over the
same 12-hour periods and were measured in the same
manner as in the previous study in Uniontown (19).
Black carbon soot (in nanograms/m3) was continuously measured using a Magee Scientific, Inc. (Berkeley, California) model AE-9 aethalometer. Respirable
particulate matter with an aerodynamic diameter less
than 2.1 microns (PM2.i in /xg/m3) and inhalable particulate matter with an aerodynamic diameter less than
10 microns (PM10 in /xg/m ) were measured over a
24-hour period using Harvard impactors with the appropriate inlets (26). A complete description of sampling and quality assurance procedures is provided
elsewhere (27).
Pollen and fungus spore measurements
Samples of airborne pollen and fungus spores were
collected from June 21 through August 29, 1991, using
a 7-day recording Burkard spore trap (Burkard Manufacturing Co., Ltd., Rickmansworth, United Kingdom) located on a roof top in the Pennsylvania State
College campus. Particles were collected at a flow rate
of 10 liters per minute on Melenex tape coated with
Lubriseal (A. H. Thomas, Swedesboro, New Jersey)
affixed to a drum that rotated at 2 mm/hour. Each
exposed tape was cut into 1-day segments (midnight to
midnight) and mounted on glass microscope slides in
hard glycerin jelly with phenosafranin stain. Pollen
grains and large fungus spores {Alternaria, Epicoccum, Drechslera, Pithomyces, and Stemphylium) were
identified by trained technicians and counted at X400
power magnification along two lengthwise transects of
each slide. Cladosporium spores, colored asymmetric
basidiospores (categorized as Coprinus, Ganoderma,
or other colored basidiospores), and colored ascospores (categorized as Xylaria or other colored ascospores) were identified and counted at X 1,000 power
magnification along a single lengthwise transect of the
slide. These counts were converted into 24-hour average concentrations in particles per cubic meter.
Children's cohort
Parents and guardians of fourth and fifth grade students completed a standard respiratory symptom questionnaire in September 1990. Children who used any
asthma medication during the previous year were excluded. Symptomatic children were defined as those
who were reported as ever wheezing or as usually
having a cough in the morning or at other times of day
for as much as 3 months during the year. Asymptomatic children had no history of wheeze or chronic
cough and no previous diagnosis of asthma. Enrollment began in June and ended on July 12, 1991.
Am J Epidemiol
Vol. 143, No. 8, 1996
Determinants of Peak Expiratory Flow Rate in Children
Pulmonary function measurements
Each child performed a pulmonary function measurement on a recording survey spirometer (Warren E.
Collins, Braintree, Massachusetts) following the same
protocol as in the previous study in Uniontown (19).
Pulmonary function tests were repeated up to four
times during the summer.
Peak expiratory flow rate measurements
Each child was instructed in the use of the MiniWright Peak Flow Meter (Armstrong Medical Industries, Lincolnshire, Illinois) and practiced the maneuver under the supervision of a technician. Three peak
expiratory flow rate (PEFR) measurements were made
in a standing position on arising in the morning and
again in the evening before bedtime. Children recorded respiratory symptoms and the number of hours
spent outdoors during the last 12 hours. PEFR and
symptom diaries were collected at approximately
3-week intervals. Children who completed fewer than
14 diary days were excluded from the analysis.
The child's mean PEFR over the entire study was
substracted from the maximum measurement for each
reporting period. The child-specific deviations were
normalized for size by dividing by the child's mean
PEFR and then multiplying by 300 liters/minute. The
mean deviation in PEFR was then calculated for all
children participating in each reporting period.
Respiratory symptom diaries
The reporting period for respiratory symptoms was
the interval since the last PEFR measurement Each
child reported up to three respiratory symptoms
(cough, wheeze, and colds) at the time of the PEFR
measurement. A symptom episode was defined by
consecutive symptomatic periods preceded by a
symptom-free period. The number of reporting periods
at risk was determined as the number of periods that
were preceded by a symptom-free period. Symptom
incidence was the number of new episodes per 100
child-periods at risk.
Statistical methods
The mean deviation in PEFR was analyzed by an
autoregressive linear regression model using the SAS
autoregressive integrated moving average procedure
(28). Symptom incidence was analyzed by an autoregressive logistic regression model using the SAS nonlinear regression procedure (29-31). Because the
number of subjects varied daily, the regression was
weighted by the reciprocal of the number of children at
each reporting period. The covariates in both models
Am J Epidemiol
Vol. 143, No. 8, 1996
799
included a linear trend, a binary indicator of morning
or evening reporting period, and the average temperature and pollutant concentration. For pollutants measured in 12-hour intervals, separate pollutant effects
were estimated for daytime and overnight exposures.
For respirable and inhalable particulate matter (PM2 ,
and PM 10 ), 24-hour concentrations ending at 8:00 a.m.
were used. The 24-hour concentration ending at midnight of the same day was used for the analysis of each
fungus taxon.
Effect estimates are scaled to reflect an approximate
interquartile change in the average concentration for
air pollutants or the difference between the minimum
and the upper decile of the 24-hour average concentration for fungal taxa. Time-weighted exposures to
pollutants were modeled by weighting the central site
concentrations by the proportion of time spent outdoors during the prior 12 hours. Child-specific estimates of associations were generated using the same
regression models and then summarized as an average
using inverse-variance weights.
The association of weather with fungal spore counts
was examined using a second-order autoregressive
model with wind direction and second-order polynomial distributed lags for a 3-day history of sunny days
and rain days. Based on these distributed lag models,
first-order autoregressive models were developed with
wind direction and indicators for sunshine and rainfall
of the current and the prior day.
RESULTS
Meteorology and air pollutant concentrations
Temperatures were mild throughout the summer
months with 12-hour average daytime 21-29°C temperatures on 90 percent of the days (table 1). Precipitation of at least 1 mm occurred on 18 days, but only
6 days had more than 10 mm of precipitation. After
controlling for diurnal variation, temperature was associated with ozone (Pearson correlation coefficient
(r) = 0.64), total sulfate particles (r = 0.60), and
particle-strong acidity (r = 0.51). Particle-strong acidity was correlated with total sulfate particles (r =
0.92), sulfur dioxide (r = 0.57), black carbon soot
(r = 0.48), and ozone (r = 0.61). The 24-hour average
concentration of respirable particulate matter (PM2 ,)
was highly correlated with total surf ate particles (r =
0.98), inhalable particulate matter (PM10) (r = 0.97),
and particle-strong acidity (r = 0.94) averaged over
the same 24-hour period.
The only major episode of acid aerosol pollution
occurred during the first 4 weeks of the study (figure
1). Daytime 12-hour averages of particle-strong acidity reached 550 nmol/m3 on July 17th, reached a high
800
Neas et al.
TABLE 1. Unlvariate statistics on temperature and pollutant concentrations, State College, PA, June 29
through August 20,1991
Mean
Upper
dede
Maximum
28.0
75.4
14.1
1,230.9
282.6
265.5
29.5
92.3
26.5
1,672.7
480.7
669.5
43.3
51.5
85.8
82.7
Interquartile
range
12-hour daytime average (8:00 a.m. to 8:00 p.m.)
Temperature (°C)
Ozone (ppb)
Sulfur dioxide (ppb)
Carbon soot (ng/m3)
Total sulfate particles (nmol/m3)
Particle-strong acidity (nmol/m3)
24.2
54.9
6.2
707.2
150.5
119.7
3.6
22.4
5.6
535.7
112.6
103.5
24-hour average (8:00 a.m. to 8:00 a.m.)
23.5
31.9
Resplrable particles (PM Z1 ») (ug/m3)
InhalaWe particles (PM10») (ug/mS)
16.7
17.7
, , , respirable particulate matter with an aerodynamic diameter less than 2.1 microns; PM 10 , Inhalable
particulate matter with an aerodynamic diameter less than 10 microns.
of 670 nmol/m3 on July 20th, and remained above
nmol/m 3 through July 22nd. The second 4 weeks
mild in comparison, with only a single daytime
hour average of particle-strong acidity above
nmol/m3.
260
was
12250
Pollen and fungus spore concentrations
Pollen counts were uniformly low throughout the
study period. The study began after the tree pollen
season and near the end of the grass pollen season. The
ragweed pollen season began during the last week of
the study; however, the concentrations in the study
period were low compared with concentrations after
the study period. Minor species (lamb's quarter, pigweed, aster, plaintain, and nettle) reached maximum
concentrations less than 10 grains per cubic meter
throughout the study period.
Fungus spore counts varied considerably over the
study (table 2) with the highest levels for Cladosporium, Epicoccum, and Coprinus during the first 10
study days (figure 1). Cladosporium concentrations
were correlated with those of Epicoccum (r = 0.64)
Days with precipitation over 1mm
Rain
CladosDOfium
Total sutfate particles (nmoJe/m3)
"I
Julyi
July 16
August 1
I111
August 16
Jutyl
July 16
August 1
August 16
FIGURE 1. Daily rainfall, 12-hour average daytime temperatures, ambient pollutant concentrations (8:00 a.m. to 8:00 p.m.), and 24-hour
average fungus spore concentrations by date: State College, Pennsylvania, June 29 through August 20, 1991.
Am J Epidemiol
Vol. 143, No. 8, 1996
Determinants of Peak Expiratory Row Rate in Children
801
TABLE 2. Unlvarlate statistics on 24-hour average concentrations (count per cubic meter) of fungal
spores by taxon, State College, PA, June 29 through August 20,1991
Oadosporium
Epicoccum
Basidiospores
Ganoderma
Coprinus
Other colored basidiospores
Ascospores
Xylaria
Other colored ascospores
Mean
Minimum
Maximum
Upper
decile
7,196
35
1,300
5
42,250
130
12,655
64
1,590
116
597
420
10
130
2,605
850
1,270
2,225
185
1,130
20
710
0
75
85
2,715
40
1,610
but only weakly correlated with Coprinus (r = 0.41).
Concentrations of Coprinus and Ganoderma were not
correlated; however, the concentration of other colored basidiospores was correlated with both Coprinus
(r = 0.57) and Ganoderma (r = 0.45). Cladosporium
counts were associated with southwesterly winds and
the first nonrain day following a rain day (model r2 =
0.34). Counts of other colored ascospores were associated with periods of rain and little sunshine (model
r2 = 0.68) and showed weak inverse correlations with
ozone (r = —0.44) and particle-strong acidity (r =
—0.22). Ganoderma counts were associated with
southwesterly winds and sunny days (model r2 =
0.53) and were weakly correlated with ozone (r =
0.46) and particle-strong acidity (r = 0.32). None of
the other fungus taxa were correlated with these pollutants (r < 0.32).
Number of children by prior symptom status
100-i OSymptomatic
^•Asymptomatic
Mean deviation in peak expiratory flow rate
Composition of children's cohort
A total of 137 children were enrolled in the diary
study. Restrictions excluded days when the child reported being outside the community for more than 2
days, 22 children with fewer than 14 days of observation, and seven children whose diaries indicated a
departure from proper reporting of flow measurements. The final cohort of 108 children were under
observation for a total of 2,504 child-days and completed diaries on 69 percent of the potential observation days. The number of days of observation for each
child varied from 14 to 52.
Analysis was restricted to the period from June 29th
through August 20th (figure 2). At least 39 children
(22 symptomatic) were participating each day. The
average enrollment was 66 children, and at least 55
Cough episodes per 100 child-days
Cold episodes per 100 child-days
I
Jutyl
July 16
August 1
August 16
July 1
July 16
August 1
August 16
FIGURE 2. Number of children by prior symptom status, mean deviation in peak expiratory flow rate, and cough and cold episodes per 100
child-days by date: State College, Pennsylvania, June 29 through August 20, 1991.
Am J Epidemiol
Vol. 143, No. 8, 1996
802
Neas et al.
children reported on 46 of the 53 study days (87
percent).
The symptomatic (n = 62) and asymptomatic (n =
46) children had similar demographic and household
characteristics, although the symptomatic children's
homes were more likely to have reports of molds or
mildew on the questionnaire (table 3). The symptomatic children had a higher lifetime cumulative incidence of a physician's diagnosis of allergies and more
chronic runny nose in the previous year.
Children spent an average of 4.9 hours outdoors
during the 12 hours preceding their evening peak
expiratory flow measurement but only 0.8 hours outdoors preceding their morning measurement. Each
child's time-weighted exposure was estimated as the
pollutant concentration times the proportion of hours
spent outdoors by that child during the prior 12 hours.
though no diurnal pattern was observed for the incidence of cold episodes.
Mean deviation in PEFR measurements
The average child-specific PEFR measurements
over the study period varied from a low of 241 liters/
minute to a high of 452 liters/minute with an overall
mean of 341 liters/minute (standard deviation = 43
liters/minute). Children's average self-reported PEFR
measurements were correlated with their average
FE^i.0 (r = 0.66) and forced vital capacity (r = 0.55)
as measured by the spirometry examinations.
The adjusted deviation in PEFR showed a trend of
0.86 liters/minute/week (95 percent CI 0.52-1.20) or a
0.29 percent increase per week in the adjusted mean
PEFR of 300 liters/minute (figure 2). The trend in
PEFR was similar to the trends in forced vital capacity
(0.27 percent/week ) and in FEV 1 0 (0.34 percent/
week). Morning measurements were 3.4 liters/minute
lower than evening measurements (95 percent CI 2.04.8), controlling for second-order autocorrelation. The
adjusted mean deviation in PEFR was not associated
with day of the week (p > 0.74). PEFR increased by
0.8 liters/minute with a 5°C increase in the average
temperature for the 12-hour period preceding the mea-
Respiratory symptom episodes
The average incidence rate was highest for cough
(2.7 episodes per 100 children at risk), and lower for
colds (1.4 episodes per 100 children at risk) and
wheeze (0.6 episodes per 100 children at risk) (table
2). The incidence of symptom episodes was more
common in the evening for cough and wheeze, al-
TABLE 3. Parental response* to prior respiratory symptom questionnaire by study participation and
symptom status, State College, PA, 1990
Participants In summer study
Demographic characteristics
Boys
Age £10 years
Parental college education
Household characteristics
Furry or feathered animals
Gas cooking stove
Air conditioning
Molds or mildew
Environmental tobacco smoke
Lifetime medical history
Severe illnesses before age 2 years
Chest ever sounded wheezy
Asthma*
Allergy*
Illness reported for prior year
Chronic runny nose
Chronic cough
Chronic phlegm
Wheeze apart from colds
Attacks of shortness of breath
Exercise-induced wheezing
Required asthma medication
1
Symptomatic
children
Asymptomatic
children
Children not in
summer study
(n»62)
(n = 46)
(n = 601)
37
55
95
41
70
91
51
57
66
13
67
11
39
61
10
35
37
20
44
19
37
87
24
9
50
17
13
23
16
27
31
16
6
21
6
10
0
11
0
0
0
0
0
47
55
15
86
0
0
21
7
3
10
7
8
11
0
Parental report of a physician's diagnosis.
Am J Epidemiol
Vol. 143, No. 8, 1996
Determinants of Peak Expiratory Flow Rate in Children
surement. Rain days (precipitation greater than 1 mm)
were associated with a —0.96-liter/minute change in
PEFR (95 percent CI -1.84 to -0.08), which was
independent of fungus spore concentrations.
Effect of fungus spore concentrations
Deficits in the adjusted mean deviation in morning
PEFR were associated with Cladosporium, Epicoccum, and Coprinus after adjustment for 12-hour average temperature, trend, evening measurement, and
second-order (24-hour lag) autocorrelation (table 4).
Adjustment for rain days did not reduce the magnitude
of these associations. In this population-average analysis, Ganoderma spore counts were associated with an
improvement in the adjusted mean deviation in PEFR.
The adjusted mean deviation in evening PEFR was not
associated with the concentrations of any spore type.
Subject-specific analysis revealed a more complex
pattern among 97 children with sufficient data for
individual regressions. Seven children showed statistically significant deficits in morning PEFR associated
with Ganoderma concentrations, with five children
having a greater than 5 percent deficit in PEFR.
Cladosporium and Epicoccum were associated with
a more general tendency toward a deficit in childspecific morning PEFR, with an average change of
—0.88 liters/minute associated with Cladosporium (95
percent CI -1.62 to -0.15) and -2.04 liters/minute
associated with Epicoccum (95 percent CI —3.11 to
—0.97). These results are similar to those in the
population-average analysis.
Epicoccum spore concentrations were associated
with the incidence of morning cough episodes among
children who reported no cough the previous evening
(OR = 1.83, 95 percent CI 1.04-3.22, for an increment of 60 spores/m3). The spore concentrations of the
other fungus spore types were not associated with the
incidence of cough.
Effect of air pollution concentrations
Over the entire study period, the adjusted mean
deviation in evening PEFR showed weak associations
with various measures of particle acidity and particle
mass (table 5). When the pollutant concentrations
were weighted by the proportion of hours spent outdoors, the magnitude of the associations was increased; however, the stability of these associations as
indicated by the t statistic was not increased. For
example, a 12-hour time-weighted exposure to a 125nmol/m3 increment in daytime particle-strong acidity
was associated with a PEFR change of —1.2 liters/
minute (95 percent CI - 2 . 8 to 0.4); however, the t
statistic was —1.47 for both measures.
When the analysis was restricted to the first 4 weeks
of the study, an unweighted 125-nmol/m3 increment in
daytime particle-strong acidity was associated with a
deficit in the adjusted mean deviation in evening
PEFR (change = - 1 . 0 liters/minute, 95 percent CI
— 1.9 to —0.2). The symptomatic children showed an
association between the deviation in evening PEFR
and a 125-nmol/m3 increment in daytime particlestrong acidity (change = —0.8 liters/minute, 95 percent CI —2.0 to 0.4), whereas the asymptomatic children showed no association (change = 0.2 liters/
minute, 95 percent CI - 0 . 5 to 0.9).
The preceding 12-hour average daytime concentrations of total sulfate particles and particle-strong acid
were associated with the incidence of new cough and
cold symptoms that evening or on the subsequent
morning among children who reported no symptoms at
the previous reporting period (table 6). All of the
24-hour average concentrations of particulate matter
TABLE 4. Adjusted change In mean deviation In morning peak expiratory flow and 95 percent
confidence intervals (CIs) associated with 24-hour average fungus spore concentration among 108
children, State College, PA, June 29 through August 20,1991
Mean deviation tn peak expiratory flow
Spores/m*
Change*
(Kara/minute)
as%
-
u
t
Time-weighted 12-hour daytime average (8:00 a.m. to 8:00 p.m.)
Cladosporium spores
Epicoccum spores
Basidiospores
Ganoderma
Coprinus
Other colored basidiospores
Ascospores
Xylaria
Other colored ascospores
10,000
60
-1.03
-1.50
-1.86 to-0.20
-2.83 to-0.18
-2.45
-2.22
2,000
170
1,000
1.53
-0.78
-0.07
-0.36 to 3.43
-1.49 to-0.07
-1.74 to 1.60
1.59
-2.15
-0.08
40
1,500
-0.49
-0.45
-1.87 to 0.90
-1.91 to 1.01
-0.69
-0.61
' Adjusted for 12-nour average temperature, trend, afternoon measurement, and first-order autocorrelation.
Am J Epidemiol
Vol. 143, No. 8, 1996
803
804
Neas et al.
TABLE 5. Adjusted change in mean deviation in evening peak expiratory flow and 95 percent
confidence intervals (CIs) associated with pollutant concentration among 108 children, State College,
PA, June 29 through August 20,1991
Mean deviation in peak expiratory flow
Change*
(Were/minute)
95% Cl
t
12-hour daytime average (8:00 a.m. to
i 8:00 p.m.)
Ozone (30 ppb)
Sulfur dioxide (10 ppb)
Black carbon soot (500 ng/m3)
Total sulfate particles (125 nmol/m3)
Particle-strong acidity (125 nmol/m3)
0.62
-0.26
-0.02
-0.53
-0.51
-1.41
-1.51
-1.04
-1.47
-1.20
to
to
to
to
to
2.66
0.99
1.00
0.40
0.17
0.60
-0.41
-0.04
-1.11
-1.47
-1.38 to 0.31
-1.38 to 0.50
-1.25
-0.91
24-hour average (8:00 a.m. to 8:00 a.m.)
Particle-strong acidity (125 nmol/m3)
ResplraUe particles ( P M ^ t ) (20 ng/m3)
-0.54
-0.44
24-hour average (8:00 a.m. to 8:00 a.m.) lagged by 24 hours
Particle-strong acidity (125 nmol/m3)
ResplraHe particles (PM a i t) (20 ng/m3)
-0.37
-0.51
-1.19 to 0.45
-1.38 to 0.35
-0.88
-1.16
• Adjusted for 12-hour average daytime temperature, trend, Cladospoiium spore count, Eplcoccum spore
count, Ganodetma spore count, Coprfnus spore count, and first-order autocorrelation.
t P M , V respirable particulate matter with an aerodynamic diameter less than 2.1 microns; PM 10 , inhalable
paniculate matter with an aerodynamic diameter less than 10 microns.
were associated with the incidence of new cold episodes that evening or on the subsequent morning.
DISCUSSION
The effects of time of day, growth, and temperature
among the State College children were equivalent to
those observed during the previous summer among a
similar cohort of children in Uniontown, Pennsylvania
(19). The average morning/evening PEFR difference of
1.1 percent was the same in both studies and was comparable to the 1.5 percent difference among nonasthmatic
children reported by Quackenboss et al. (32). The trend
in the PEFR measurements among State College children
(0.86 liters/minute/week) was nearly equal to the trend
among the Uniontown children (0.93 liters/minute/week)
and was similar to trends in forced vital capacity and
FEV1-0. The temperature effect of 0.8 liters/minute/5°C
was equal to that observed in Uniontown, and effects of
a similar magnitude have been found in previous studies
(33, 34).
Airborne spore concentrations of Cladosporium,
Epicoccum, and Coprinus were associated with deficits in the mean deviation of morning PEFR for the
entire cohort. Fungal spores are known to be related to
the incidence of asthma and to produce IgE-mediated
hypersensitivity (20, 21). The fungal cell wall, including that surrounding spores, contains l,3-/J-r>glucans;
and exposure to glucans has been reported to cause
decrements in pulmonary function (35). Fungal spores
also contain mycotoxins, which have been hypothesized to affect pulmonary function. A previous clinical
report cited a few anecdotal cases in which individual
deficits in PEFR measurements were apparently associated with daily spore concentrations of Cladosporium and colored basidiospores (36). An Austrian
study of 40 asthmatic children found an association
between 10-day average Cladosporium concentrations
and a 10-day average symptom score, but no significant daily relations (37). In the population-average
analysis, Ganoderma spore concentrations were somewhat associated with an improvement in the mean
deviation of morning PEFR. Conversely in the individual analyses, five children showed significant
child-specific declines in morning PEFR with increased Ganoderma spore concentrations. Basidiospores such as mose produced by Ganoderma have
been recognized as clinically important aeroallergens
in New Orleans (38) and the Pacific Northwest (39).
Epicoccum is a fungus that is abundant in agricultural
settings, and both spore and metabolic extracts produce positive immediate skin tests at a frequency equal
to or exceeding those produced by Alternaria extracts
(40). Indoor exposure to Epicoccum spores has been
associated with hay fever symptoms in children (41).
Precipitation affected both the mean deviation in
PEFR and the concentrations of airborne contaminants. Days with precipitation greater than 1 mm were
associated with a — 0.96-liters/minute change in PEFR
Am J Epidemiol
Vol. 143, No. 8, 1996
Determinants of Peak Expiratory Row Rate in Children
805
TABLE 6. Ad]usted odds ratios (ORs) and 95 percent confidence intervals (CIs) for the association between average pollutant
concentration and Incidence of symptom episodes that evening or subsequent morning among 108 children, State College, PA,
June 29 through August 20,1991
Cough
Wheeze
Adjusted
OR*
Adjusted
OR*
95% Cl
CoU
95% Cl
Adjusted
OR*
95% Cl
0.90-2.24
0.75-1.46
0.82-1.39
1.00-1.60
1.01-1.39
1.62
1.13
1.35
1.53
1.35
0.95-2.78
0.77-1.67
1.01-1.82
1.17-2.00
1.14-1.61
0.99-1.44
0.93-1.47
1.41
1.47
1.15-1.73
1.16-1.86
1.10
1.13
0.85-1.41
0.87-1.46
12-hour daytime average (8:00 a.m. to 8:00 p.m.)
Ozone (30 ppb)
Sulfur dioxide (10 ppb)
Black carbon soot (500 ng/m3)
Total sulfate particles (125 nmol/m3)
Particle-strong acidity (125 nmol/m3)
0.47
0.70
1.37
1.30
1.08
0.17-1.31
0.31-1.58
0.81-2.34
0.79-2.15
0.75-1.56
1.42
1.05
1.07
1.26
1.19
24-hour average (8:00 a.m. to 8:00 a.m.)
3
Particle-strong acidity (125 nmol/m )
Resplrable particles (PM z 1 t) (20 lig/m3)
1.03
1.11
0.65-1.64
0.66-1.88
1.19
1.17
24-hour average (8:00 a.m. to 8:00 a.m.) lagged by 24 hours
3
Particle-strong acidity (125 nmol/m )
RespiraMe particles (PM z 1 t) (20 ng/m3)
1.33
1.45
0.89-1.99
0.95-2.21
1.34
1.37
1.13-1.61
1.13-1.66
* Adjusted for 12-hour average daytime temperature, trend, Cladosporlum spore count, Epicoccum spore count Ganoderma spore count,
Coprinus spore count, and first-order autocorrelation.
t PM^v resplrable particulate matter with an aerodynamic diameter less than 2.1 microns; PM 10 , inhalable particulate matter with an
aerodynamic diameter less than 10 microns.
among the State College children. Similar effects have
been found in previous studies (42,43). Cladosporium
levels were highest on days following episodes of
precipitation. Release of Cladosporium spores from
conidiophores has been associated with movements of
the conidiophore during drying. Colored ascospore
levels were highest on days with precipitation. The
spore release mechanism for many ascomycetes depends on an increase in available water. Release of
basidiospores, including those of Ganoderma, is also
related to changes in water availability. Most basidiospores (especially those produced by mushrooms, including Coprinus) show strong diurnal variation patterns, with most releases occurring during the predawn
hours. Specific factors controlling release of Ganoderma spores, which apparently do not show this diurnal pattern, are unknown; however, heating by sunlight may pjay a role. The major episode of particulate
air pollution occurred in mid-July during a period with
no precipitation. The inclusion in the model of an
indicator for rainy days did not alter the effect estimates for the airborne contaminants.
In this study, a 12-hour daytime time-weighted exposure to an additional 125-nmol/m3 particle-strong
acidity was associated with a mean 1.2-liters/minute
decrease in evening PEFR. This is less than the —1.4to — 1.9-liters/minute change in PEFR observed by
Raizenne et al. (44) and about half the —2.4-liters/
minute change in PEFR observed among the Uniontown children associated with the same exposure. InAm J Epidemiol
Vol. 143, No. 8, 1996
verse variance weights were used to combine the
results of the comparable studies in State College and
Uniontown. A 12-hour daytime time-weighted exposure to an additional 125-nmol/m3 particle-strong
acidity was associated with a -1.8-liters/minute
change in evening PEFR (95 percent Cl —2.9 to
-0.7). Cough episodes were associated with 12-hour
daytime exposures to particle-strong acidity and total
sulfate particles in both the State College and Uniontown cohorts. Aside from a weak association with
Ganoderma, fungus spore concentrations were not associated with the levels of particulate air pollutants.
Potential limitations
The subjects in this study were selected from a
cross-sectional survey of all public elementary schoolchildren on the basis of their reported chronic respiratory symptoms without knowledge of the timing of
these symptoms during the prior year. State College
was selected because of prior observations of high
particle-strong acidity in central Pennsylvania. Subsequent analyses of the 24 communities indicated that
State College did not have an exceptionally high prevalence of a questionnaire report of bronchitis (17) or
low level of forced vital capacity from spirometry (18)
compared with other communities.
The PEFR measurements were supported by complete spirometry. The child-specific mean PEFRs were
correlated with their mean FEV, 0 values (r = 0.66)
806
Neas et al.
and had similar trends over the study period. The
misclassification of respiratory symptoms could not be
assessed inasmuch as the children were the sole source
for symptom information. The associations with cough
symptoms were probably not due to reporting bias
because the periods of increased particle-strong acidity
were not associated with increased wheeze symptoms
or colds.
The population-average model assumed that the
children in this study had similar sensitivity to the
effects of the ambient pollutants and that failures to
report were not associated with the ambient pollutant
levels. Child-specific regressions using the same
model as the population-average analysis showed similar associations between PEFR and particle-strong
acidity. The children completed their diaries on 69
percent of the potential observation days, and missing
reporting periods did not appear to covary with the
ambient pollutant levels (figures 1 and 2).
The present study has a limited ability to detect
associations of respiratory symptoms and PEFR with
potentially important environmental risk factors. Although the temperature effect of PEFR was consistent
with that observed in Uniontown, the precision of the
effect estimate was reduced by the limited exposure
gradient. A similar problem arose for particle-strong
acidity. The results for ozone were affected by the
limited exposure gradient and by the colinearity of
ozone levels with beneficial factors such as temperature. The lack of an exposure gradient precluded an
analysis of pollen and some fungi species. Other fungi
species, especially indoor fungi, may have important
health effects in addition to those species measured in
this study.
The models assumed that the effects of trend and
temperature would be linear over the observed range
of exposure. Nonlinear PEFR trends in the first 7 or 14
days were examined by exclusion of these days, by
modeling with separate intercepts for the first 7 or 14
days, and by an indicator for linear trend over the first
14 days. None of these changes substantially altered
the results. The coherence of the daily trend in PEFR
with the trends in repeated spirometry suggests that
these changes are due to the growth of the children
over the observation period rather than to a training
effect. The 12-hour average daytime temperatures
ranged from 18 to 30°C (64-86°F) during the study
period (figure 1), and the weak effect of temperature
over this range is not Likely to be significantly nonlinear.
Conclusions
The principal determinants of the day-to-day variation in PEFR among children include time of obser-
vation, growth, temperature, precipitation, air pollutants, and airborne aeroallergens. The effects of time of
observation, growth, and temperature are very consistent across two studies, with similar measurements for
comparable children's cohorts. In this study, airborne
spore concentrations of Cladosporium, Epicoccum,
and Coprinus were associated with deficits in the
mean deviation of morning PEFR for die entire cohort;
and five children showed significant child-specific
declines in morning PEFR with increased Ganoderma
spore concentrations. The effect of particle-strong
acidity on the mean deviation in evening PEFR was
somewhat less in this study. These data are not incompatible with the previous finding of an association of
PEFR with aerosol acidity among the Uniontown children. These results are also compatible with a respirable particle hypothesis. Daytime exposures to ozone
and particulate air pollutants were associated with the
increased incidence of summertime cough and cold
symptoms that evening or on the subsequent morning.
Precipitation days and fungal spore concentrations did
not confound die association of PEFR widi the levels
of particulate air pollutants. Fungus spore concentrations affected the mean deviation in morning PEFR,
and air contaminants affected the mean deviation in
evening PEFR. We conclude that summer episodes of
excessive aerosol acidity and particulate pollution are
acutely associated with declines in peak expiratory
flow rates and increased incidence of cough and cold
episodes in children. Similar, if not stronger, effects
were associated with variation in fungus spore counts,
which suggests that additional exploration of the influence of aerobiologic contaminants in conjunction
with summertime haze episodes is warranted.
ACKNOWLEDGMENTS
This research was supported by Electric Power Research
Institute contract RP-1001 and National Institute of Environmental Health Sciences grants ES-00002 and ES-04595.
Dr. Neas was supported in part by National Institute of
Health National Research Service award HL-07427 and US
Environmental Protection Agency cooperative agreement
CR-821762.
The authors thank David Tollerud, Martha Fay, George
Allen, Andrew Damokosh, Joan Cunningham, Lynn
McClelland, and Barbara Phelps for collecting the data and
assisting in the preparation of the datasets.
REFERENCES
1. Shy CM, Hasselblad V, Burton RM, et al. Air pollution effects
on ventilatory function of US schoolchildren: results of studies
in Cincinnati, Chattanooga, and New York. Arch Environ
Am J Epidemiol
Vol. 143, No. 8, 1996
Determinants of Peak Expiratory Row Rate in Children
Health 1973;27:124-8.
2. Mostardi RA, Wocbkenberg NR, Ely DL, et al. The University
of Akron study of air pollution and human health effects. II.
Effects on acute respiratory illness. Arch Environ Health
1981;36:250-5.
3. French JG, Lowrimore G, Nelson WC, et al. The effect of
sulfur dioxide and suspended sulfates on acute respiratory
disease. Arch Environ Health 1973;27:129-33.
4. Love GJ, Lan SP, Shy CM, et al. The incidence and severity
of acute respiratory illness in families exposed to different
levels of air pollution. New York Metropolitan Area,
1971-72. Arch Environ Health 1981;36:66-74.
5. Dodge R. The respiratory health and lung function of AngloAmerican children in a smelter town. Am Rev Respir Dis
1983;127:158-61.
6. Richards W, Azen SP, Weiss J, et al. Los Angeles air pollution
and asthma in children. Ann Allergy 1981;47:348-54.
7. Ponka A. Absenteeism and respiratory disease among children
and adults in Helsinki in relation to low-level air pollution and
temperature. Environ Res 1990;52:34-46.
8. Ransom MR, Pope CA. HI. Elementary school absences and
PM 10 pollution in Utah Valley. Environ Res 1992;58:204-19.
9. Schwartz J, Spix C, Wichmann HE, et al. Air pollution and
acute respiratory illness in five German communities. Environ
Res 1991;56:1-14.
10. Jaakkola JJK, Paunio M, Virtanen M, et al. Low-level air
pollution and upper respiratory infections in children. Am J
Public Health 1991;81:1060-3.
11. Jedrychoski W, Krzyzanowski M. Ventilatory lung function
and chronic chest symptoms among the inhabitants of urban
areas with various levels of acid aerosols: prospective study in
Cracow. Environ Health Perspect 1989;79:101-7.
12. Stern B, Jones L, Raizenne M, et al. Respiratory health effects
associated with ambient sulfates and ozone in two rural Canadian cities. Environ Res 1989;49:20-39.
13. Ferris BG Jr, Speizer FE, Spengler JD, et al. Effects of sulfur
oxides and respirable particles on human health: methodology
and demography of populations in study. Am Rev Respir Dis
1979;120:767-79.
14. Ware JH, Ferris BG Jr, Dockery DW, et al. Effects of ambient
sulfur oxides and suspended particles on respiratory health of
preadolescent children. Am Rev Respir Dis 1986;133:
834-42.
15. Dockery DW, Speizer FE, Stram DO, et al. Effects of inhalable particles on respiratory health of children. Am Rev Respir
Dis 1989; 139:587-94.
16. Damokosh Al, Spengler JD, Dockery DW, et al. Effects of
acidic particles on respiratory symptoms in seven US communities. (Abstract). Am Rev Respir Dis 1993;147:A632.
17. Dockery DW, Cunningham J, Damokosh Al, et al. Health
effects of acid aerosols on North American children: respiratory symptoms and illness. Environ Health Perspect (in press).
18. Raizenne M, Neas LM, Damokosh Al, et al. Health effects of
acid aerosols on North American children: pulmonary function. Environ Health Perspect (in press).
19. Neas LM, Dockery DW, Koutrakis P, et al. The association of
ambient air pollution with twice daily peak expiratory flow
rate measurements in children. Am J Epidemiol 1995; 141:
111-22..
20. Lehrer SB, Hughes JM, Altaian LC, et al. Relavence of
basidiomycete allergy in the USA and Europe and its relationship to allergic respiratory symptoms. Allergy 1994;49:
460-5.
21. Gergen PJ, Turkeltaub PC. The association of individual allergen reactivity with respiratory disease in a national sample:
data from the second National Health and Nutrition Examination Survey, 1976-80 (NHANES II). J Allergy Clin Immunol
1992;90:579-88.
22. O'Hollaren MT, Yunginger JW, Offord KP, et al. Exposure to
an aeroallergen as a possible precipitating factor in respiratory
arrest in young patients with asthma. N Engl J Med 1991 ;324:
359-63.
Am J Epidemiol
Vol. 143, No. 8, 1996
807
23. Speizer, FE. Studies of acid aerosols in six cities and in a new
multi-city investigation: design issues. Environ Health Perspect 1989;79:61-7.
24. Koutrakis P, Wolfson JM, Slater JL, et al. Evaluation of an
annular denuder/filter pack system to collect acidic aerosols
and gases. Environ Sci Technol 1988;22:1463-8.
25. Koutrakis P, Wolfson JM, Spengler JD. An improved method
for measuring aerosol strong acidity: results from a nine
month study in St Louis, Missouri, and Kingston, Tennessee.
Atmos Environ 1988;22:157-62.
26. Marple V, Rubow KL, Turner W, et al. Low flow rate sharp
cut impactors for indoor air sampling: design and calibration.
J Air Pollution Control Assoc 1987;37:1303-7.
27. Allen GA. Atmospheric chemistry and personal exposure to
acid aerosols in Uniontown, Pennsylvania, June-August,
1990: method descriptions and data summaries. Boston, MA:
Exposure Assessment and Engineering Program, Department
of Environmental Health, Harvard University School of Public
Health, 1991.
28. SAS Institute, Inc. SAS/ETS user's guide, version 6, ed 1.
Cary, NC: SAS Institute, Inc., 1988.
29. Schwartz J, Zeger S. Passive smoking, air pollution, and acute
respiratory symptoms in a diary study of student nurses. Am
Rev Respir Dis 1990; 141:62-7.
30. Zeger SL, Liang KY, Albert PS. Models for longitudinal data:
a generalized estimating equation approach. Biometrics 1988;
44:1049-60.
31. SAS Institute, Inc. SAS/STAT user's guide, version 6, ed 4,
vol 2. Cary, NC: SAS Institute, Inc., 1989:1135-94.
32. Quackenboss JJ, Lebowitz MD, Krzyzanowski M. The normal
range of diurnal changes in peak expiratory flow rates: relationship to symptoms and respiratory disease. Am Rev Respir
Dis 1991;143:323-30.
33. Higgins ITT, D'Arcy JB, Gibbons DI, et al. Effect of exposures to ambient ozone on ventilatory lung function in children. Am Rev Respir Dis 1990;141:l 136-46.
34. Hoek G, Brunekreef B, Kosterink P, et al. Effect of ozone on
peak expiratory flow of exercising children in the Netherlands.
Archiv Environ Health 1993,48:27-32.
35. Rylander R, Sorenson S, Goptoo H, et al. The Importance of
endotoxin and glucan for symptoms in sick buildings. In:
Bieva CJ, Courtois Y, Govaerts M, eds. Present and future of
indoor air quality. New York, NY: Elsevier, 1989.
36. Brown HM. The relationship of aerobiological data to seasonal allergic symptoms: a review of 27 years' experience.
Immunol Allergy Prac 1992;14:318-29.
37. Klabuschnigg A, Gotz M, Horak F, et al. Influence of aerobiology and weather on symptoms in children with asthma.
Respiration 1981;42:52-60.
38. Butcher BT, O'Neil CE, Reed MA, et al. Basidiomycete
allergy: measurement of spore specific IgE antibodies. J Allergy Clin Immunol 1987;80:803-9.
39. Sprenger JD, Altaian LC, O'Neil CE, et al. Prevalence of
basidiospore allergy in the Pacific Northwest. J Allergy Clin
Immunol 1988;82:1076-80.
40. Portnoy J, Chapman J, Burge H, et al. Epicoccum allergy: skin
reaction patterns and spore/mycelium disparities recognized
by IgG and IgE ELISA inhibition. Ann Allergy 1987;59:
39-43.
41. Su HJ, Burge HA, Spengler JD. Indoor saprophytic aerosols
and respiratory health. (Abstract). J Allergy Clin Immunol
1990;85(pt 2):248.
42. Lippmann M, Lioy PJ, Leikauf G, et al. Effects of ozone on
the pulmonary function of children. Adv Environ Toxicol
1983;5:423-46.
43. Mann M, Patel K, Reardon JZ, et al. The influence of spring
and summer New England meteorologic conditions on the
respiratory status of patients with chronic lung disease. Chest
1993;103:1369-74.
44. Raizenne ME, Burnett RT, Stern B, et al. Acute lung function
responses to ambient acid aerosol exposures in children. Environ Health Perspect 1989;79:179-85.