Stress Effects on Lung Function in Asthma are Mediated by Changes in Airway Inflammation
To examine the association of changes in current negative mood and long-term daily hassles with changes in lung function and airway inflammation in patients suffering from asthma and in healthy controls. Associations between psychological factors and asthma symptoms have been documented, but the relationship between airway inflammation and psychological factors has been largely unexplored. Data were analyzed from 46 asthma patients and 25 controls who completed questionnaires on current mood and daily hassles at two assessments 3 months apart. Lung function was measured by spirometry (forced expiratory volume in the first second (FEV(1))) and airway inflammation by the fraction of nitric oxide in exhaled air (FeNO). Regression analyses controlling for allergen load and air pollution (ozone) were calculated to study the association between changes in psychological factors and changes in lung function and airway inflammation, and to examine the mediational role of airway inflammation in the stress-lung function association. In patients with asthma, increases in negative affect were associated with decreases in FEV(1) and increases in FeNO. For daily hassles, a reverse pattern of associations was found, with decreases in daily hassles linked to decreases in FEV(1) and increases in FeNO. Mediation analyses showed that FeNO was a significant mediator of the association of both negative affect and daily hassles with lung function changes. No significant associations were found for healthy controls. Psychological variables are consistently associated with spirometric lung function and airway inflammation in asthma patients. For asthma patients, effects of acute negative affect must be distinguished from more chronic distress due to daily hassles.
Stress Effects on Lung Function in Asthma are Mediated by Changes in
ANTJE KULLOWATZ,PHD, DAVID ROSENFIELD,PHD, BERNHARD DAHME,PHD, HELGO MAGNUSSEN,MD,
FRANK KANNIESS,MD,AND THOMAS RITZ,PHD
Objective: To examine the association of changes in current negative mood and long-term daily hassles with changes in lung
function and airway inflammation in patients suffering from asthma and in healthy controls. Associations between psychological
factors and asthma symptoms have been documented, but the relationship between airway inflammation and psychological factors
has been largely unexplored. Method: Data were analyzed from 46 asthma patients and 25 controls who completed questionnaires
on current mood and daily hassles at two assessments 3 months apart. Lung function was measured by spirometry (forced expiratory
volume in the first second (FEV
)) and airway inflammation by the fraction of nitric oxide in exhaled air (FeNO). Regression
analyses controlling for allergen load and air pollution (ozone) were calculated to study the association between changes in
psychological factors and changes in lung function and airway inflammation, and to examine the mediational role of airway
inflammation in the stress-lung function association. Results: In patients with asthma, increases in negative affect were associated
with decreases in FEV
and increases in FeNO. For daily hassles, a reverse pattern of associations was found, with decreases in
daily hassles linked to decreases in FEV
and increases in FeNO. Mediation analyses showed that FeNO was a significant mediator
of the association of both negative affect and daily hassles with lung function changes. No significant associations were found for
healthy controls. Conclusion: Psychological variables are consistently associated with spirometric lung function and airway
inflammation in asthma patients. For asthma patients, effects of acute negative affect must be distinguished from more chronic
distress due to daily hassles. Key words: asthma, mood, daily hassles, airway inflammatory status, lung function.
⫽ forced expiratory volume in the first second; FeNO ⫽ fraction
of exhaled nitric oxide; DH ⫽ daily hassles; PANAS ⫽ Positive Affect
Negative Affect Schedule; NA ⫽ current negative affect; ATI ⫽
Asthma Trigger Inventory; ICS ⫽ inhaled corticosteroids.
ffects on physical health caused by stress, mood, or emo-
tion are of recurring interest in asthma research. Clinical
reports and observational studies have suggested a role for
emotional and behavioral states in asthma exacerbations (1,2).
Besides effects on the self-management of this chronic dis-
ease, psychological factors can also directly influence the
pathophysiological process. In carefully controlled laboratory
studies, emotions and stress have been shown to adversely
affect lung function in patients with asthma (3,4). Longitudi-
nal studies using self-monitoring of lung function have dem-
onstrated an association between mood and lung function
(5–7), and negative life events have been shown to contribute
substantially to asthma exacerbations in children (8).
It is commonly accepted that asthma is a chronic disease
that involves inflammatory processes in the airways. Airway
inflammation is viewed as one of the key determinants of lung
function decline in asthma (9). It has been established that the
fraction of nitric oxide in the exhaled air (FeNO) can serve as
a marker of eosinophilic inflammatory processes in the air-
ways (10,11). Correlational and prospective findings both
have demonstrated an association between elevated FeNO
levels and asthma (12–14) as well as reduced levels of FeNO
after corticosteroid therapy in patients with asthma (15,16).
To date, the potential link between airway inflammation
and lung function decline due to psychosocial factors is
largely unexplored. In general, immune responses to psycho-
social stress in patients with asthma are indicative of alter-
ations in the balance between cellular and humoral immunity,
with a shift toward the latter and production of Type 2 cyto-
kines that promote inflammatory processes (17,18). Recent
animal studies also suggested that acute and/or chronic envi-
ronmental stress can increase airway inflammation or proin-
flammatory cytokines in the airways (19 –21). A study with
adult asthma patients showed an exaggerated response to
allergen exposure during periods of academic examination
stress (22). Most recently, chronic stress and threat perception
have been found to mediate the association between low
socioeconomic status and enhanced inflammatory processes
marked by higher blood eosinophil counts in asthmatic chil-
dren (23). This association was in the opposite direction for
children without asthma. However, Ho¨glund et al. (24), mea-
suring the relationship between academic stress and airway
inflammation, did not observe an expected increase in FeNO
during an examination stress phase in patients with atopy;
rather, they found a small decrease in FeNO values in healthy
Thus, although some studies have provided initial evidence
that stress can lead to adverse outcomes in markers of airway
inflammation, the extent to which levels of current negative
mood and long-term daily hassles (DH) affect airway inflam-
mation in adults with asthma is unknown. Also, it is unclear
whether changes in airway inflammation can explain the lung
function decline typically observed in psychosocial stress.
Therefore, the purpose of this study was to explore the asso-
ciation of current negative mood and long-term low-level
stressors with airway inflammation and lung function. We also
sought to examine a potential mediational role of inflamma-
From the Department of Environmental Health (A.K.), Harvard University,
Boston, MA; Department of Psychology (D.R., T.R.), Southern Methodist
University, Dallas, Texas; Department of Psychology (B.D.), University of
Hamburg, Hamburg, Germany; Pulmonary Research Institute (H.M., F.K.),
Hospital Grosshansdorf, Grosshansdorf, Germany.
Address correspondence and reprint requests to Antje Kullowatz, 401 Park
Drive, Landmark Center, Rm. 404 West, Boston, MA 02215. E-mail:
Received for publication April 20, 2007; revision received December 19,
Supported by Grant Ri 957/3 to 1 from the German Research Society,
468 Psychosomatic Medicine 70:468– 475 (2008)
Copyright © 2008 by the American Psychosomatic Society
tory changes. We expected that both current negative affect
and long-term DH would aggravate airway inflammation in
asthma patients and thereby reduce lung function.
Design and Protocol
Data for these analyses were collected as part of a larger laboratory study
on emotional effects on the airways. Patients completed questionnaires on
current mood and on DH experienced during the previous 3 months at two
assessments, which occurred 3 months apart.
Participants were recruited at a special treatment and research clinic
(Pulmonary Research Institute, Hospital Grosshansdorf) and by advertise-
ments at different departments of the University of Hamburg from May 2004
to September 2005. Inclusion criteria were the diagnosis of asthma (or having
no lung disease for healthy controls), being a nonsmoker for ⬎6 months,
having smoked ⬍10 pack-years in their lifetime, and currently being in stable
mental and physical health. Excluded were individuals with no fluent German
language skills and individuals with an intake of oral corticosteroids in the
past 3 months. The study was approved by the local Ethics Committee and
participants signed a consent form.
Eighty-two participants were invited to participate. Nine participants were
excluded; reasons for exclusion were pregnancy (n ⫽ 1), scheduling problems
after pretest (n ⫽ 1), technical problems with the instruments (n ⫽ 1), intake
of oral corticosteroids for reasons other than asthma (n ⫽ 2), incomplete data
from questionnaires (n ⫽ 4), or nonparticipation at Assessment 2 (n ⫽ 2).
Patients were requested to stay on a stable dose of their asthma medication
during the study. Before each assessment, patients were also asked to discon-
tinue short-acting bronchodilators for 6 hours, long-acting
agonists for 12
hours, long-term anticholinergics (Tiotropium) for 7 days, and leukotriene
inhibitors for 3 days.
A pulmonologist performed a diagnostic evaluation on each participant,
which included asthma history, physical examination, lung function assess-
ment by spirometry, and an allergy Skin Prick Test for sensitivity to common
airborne allergens. Mental health was ascertained by a clinically trained
psychologist using the Structured Clinical Interview for Diagnostic and Sta-
tistical Manual of Mental Disorders-IV-Text Revision (SCID-I) (25). Partic-
ipants with severe current problems or with episodes of major depression,
high likelihood of suicide, schizophrenia, or substance abuse were excluded
(n ⫽ 1).
Airway inflammation by FeNO (in ppb) was measured using the che-
moluminescence analyzer NIOX (Aerocrine, Inc., Stockholm, Sweden). The
measurement is a noninvasive procedure whereby participants exhale for 10
seconds through a mouthpiece with a valve mechanism that keeps flow
constant at 50 mL/second. The measurement of FeNO was repeated nine
times and an average score was calculated at each assessment. Missing data
(4.4% of all measurements) were caused by malfunctions of the device.
Consistent with measurement guidelines (26), ingestion of nitrate-rich food,
caffeine, and alcohol was discouraged, and exercise or heavy physical activity
had to be finished at least 1 hour before measurements were started.
Lung function was determined by spirometry (Jaeger/Toennies, Wuer-
zburg, Germany). We measured forced expiratory volume in the first second
), expressed as the percentage of the norm, from the forced expiratory
maneuver according to current guidelines (27,28).
Outdoor ozone levels taken from the nearest public monitoring station
were recorded at each assessment as an indicator of air pollution. Mean levels
of ozone (in ppb) were 58.5 ⫾ 24.6 (standard deviation) and 67.7 ⫾ 24.8, at
Assessments 1 and 2, respectively. To calculate pollen allergen load for each
participant, the respective skin test wheal size (minus negative control) was
multiplied by the pollen level on the respective assessment day. Levels of
allergen load across the whole sample were 6.00 ⫾ 10.26 and 5.03 ⫾ 9.49 for
Assessments 1 and 2, respectively. Information on each allergen pollen count
for the area of the study was downloaded daily from public access databases.
An average load score for all pollen allergens (wheal size ⫻ pollen count) was
calculated for each participant at each assessment. Both variables (ozone level
and allergen load) were used as control variables because they have been
shown to be associated with airway inflammation and/or asthma exacerba-
At the beginning of each assessment, participants completed the validated
German version of the Positive and Negative Affect Schedule (PANAS) (31).
Participants rated their current mood using 20 items with a 5-point scale (very
slightly or not at all; a little; moderately; quite a bit; extremely). A sum score
for Negative Affect (NA) was calculated across 10 items, which characterized
a general dimension of perceived distress and unpleasurable engagement.
Internal consistency for NA is
⫽ 0.86, and retest-reliability for an interval
of 1 week is low with r
⫽ 0.54, which is consistent with its conceptualization
as a state measure.
Participants also completed a questionnaire on DH—the Mainzer Hassle
Scale for Families (MHS-F) (Hundelshausen AM. [Daily hassles and well-
being in young parents]. Unpublished diploma thesis [in German], University
of Mainz, 1994) (Wellinger C. [Daily demands and their association with
critical life changes in a German speaking student population]. Unpublished
diploma thesis [in German], University of Mainz, 1993), which includes
relevant items for students as well as families. Hassles are defined as “irri-
tating, frustrating demands that occur during every day transaction with the
environment” (32). The questionnaire we used was based on English language
DH scales (33,34) and has a good reliability (0.79). Participants rated how
often (0 ⫽ not at all; 1 ⫽ seldom; 2 ⫽ often; 3 ⫽ very often) each of the 52
hassle items had occurred during the prior 3 months. A sum score across all
the items of the scale was calculated.
Before the first assessment, all participants filled out an ad hoc question-
naire on demographics. Asthma patients completed an additional question-
naire on aspects of disease manifestation and a validated German version of
the Asthma Trigger Inventory (ATI) (35). For the purpose of this study, we
calculated the mean of the subscale for psychological triggers of the ATI,
which has a good internal consistency (
⫽ 0.88). Asthma severity was
determined according to National Heart, Lung, and Blood Institute/World
Health Organization 2006 guidelines (36).
Participants were recruited for a laboratory study with assessments at five
appointments. At two assessments scheduled 3 months apart, they completed
the DH questionnaire for the preceding 3 months and the PANAS for current
mood. The data from these two assessments was the subject of the current
investigation. Individuals also performed FeNO measurements followed by
Descriptive statistics were calculated to characterize the participants in
terms of demographics, NA, DH, basal lung function, and airway inflamma-
tion. To establish if psychological factors were related to both airway inflam-
mation and lung function, regression analyses using residualized change
scores (from the first to the second assessment) were performed. Values of
FeNO, NA, and DH were log-transformed to reduce skewness (37). Because
we expected that patients suffering from asthma might differ from healthy
controls in their reactions to NA and DH, we calculated interaction terms by
multiplying the dummy coded Group variable (asthma patients versus con-
trols) by the residualized change scores for these variables. Regression anal-
yses were performed using the following independent variables: residualized
change scores for NA, DH, allergen load, and ozone levels, plus Group (0 ⫽
We also measured total respiratory impedance at 10 Hz as a more direct
measure of airway obstruction using impulse oscillometry (Jaeger/Toenniess,
Wuerzburg, Germany), which is a variant of the forced oscillation technique
(28). However, because this direct index of the airway caliber did not yield
any significant findings, the results will not be presented in detail here.
MEDIATION EFFECT OF AIRWAY INFLAMMATION
469Psychosomatic Medicine 70:468– 475 (2008)
patients suffering from asthma; 1 ⫽ controls), the interaction between Group
and NA, and the interaction between Group and DH. The two dependent
variables for the regression analyses were the residualized change scores for
and FeNO. To be cautious about generalizing findings to both asthma
patients and controls, if there was an interaction effect at p ⬍ .10, we
calculated the relationship between the independent variable and the outcome
(the “simple slopes”) separately for each group (asthma patients and controls)
(38). Calculating the effect separately for each group protects us from incor-
rectly assuming that the finding is significant in both groups.
Because of the central role of airway inflammation in lung function
decline in asthma, we performed a mediation analysis to determine if airway
inflammation mediated the relationship between the psychological variables
and lung function. Path coefficients were calculated for the mediation model
using regression analyses (39). To test for mediation, we used the z⬘ test,
which directly tests the significance of a mediated pathway. A z⬘⬎兩0.97兩
indicates a significant mediation effect (p ⬍ .05).
Because airway inflammation is reduced by the intake of inhaled corti-
costeroids (ICS), a supplementary regression analysis explored the role of
intake of this medication (1 ⫽ yes; 0 ⫽ no) as a moderator variable of the
association between the psychological variables (NA, DH) and FEV
FeNO within the asthma patients. Intake of ICS, plus the two interaction terms
for ICS by NA and ICS by DH, were used as additional predictor variables in
this analysis. A similar analysis was performed to explore a potential mod-
erator effect of the psychological triggers subscale of the ATI, because greater
frequency of psychological triggers has been associated with stronger airway
constriction after exposure to emotional stimuli (40). These latter two regres-
sion analyses exceeded the typically recommended ratio of predictors to
subjects (37); therefore, their status was purely exploratory.
Seventy-one participants (46 asthma patients and 25
healthy controls) between the ages of 19 and 52 years (mean
age ⫽ 27.8 ⫾ 6.8 years) were included in the analyses (Table
1). Nearly half (48%) of the patients met the criteria for
intermittent asthma, 35% for mild persistent asthma, and 17%
for moderately persistent asthma. The average duration of the
disease was 14.5 years (range ⫽ 1–39 years), with onset
before the age of 18 years reported in 76% of the sample. A
family history of asthma and/or allergies was reported by
72%. Seasonal symptoms of asthma were reported by 24% in
the spring and 15% in the summer. ICS were used by 37% of
For Assessment 1, 15 participants were tested during win-
ter (January to March), 26 in spring (April to June), 11 in
summer (July to September), and 19 in fall (October to De-
cember). For Assessment 2, the number tested in each season
was 21, 17, 24, and 9 for winter, spring, summer, and fall,
Values of FeNO and FEV
were stable over time, F(1,64)
⬍1 (Table 2). There were no differences between participants
with asthma and those without asthma in FEV
, and the values
for both groups showed normal lung function. Fewer DH were
reported for the 3 months before Assessment 2 than for the 3
months before Assessment 1 (F(1,64) ⫽ 15.44; p ⬍ .001). The
correlation between NA and DH was significant at both as-
sessments (Assessment 1: r(66) ⫽ .41, p ⬍ .001; Assessment
2: r(69) ⫽ .35, p ⫽ .005).
Associations of Current Negative Affect and
Long-Term Stress With Airway Inflammation and
The regression analysis with FEV
as the dependent vari
able indicated that, in general, higher short-term NA was
significantly related to lower FEV
). However, this
main effect was qualified by a Group by NA interaction that
was almost statistically significant. Because of the clinical
relevance of this finding and because the interaction indicated
that the relationship between NA and FEV
may be different
for asthma patients and controls, we calculated simple slopes
for each group separately, computing the relationship between
NA and FEV
within each group (38). This analysis revealed
that, for asthma patients, higher NA was related to lower FEV
(b ⫽⫺15.4; t(62) ⫽⫺2.51; p ⬍ .05),
whereas for healthy
controls, no significant relationship between NA and FEV
was found (b ⫽ 2.0; t(55) ⫽ 0.30; p ⫽ NS). Thus, the
As shown by Aiken and West (38), the regression coefficient for a main
effect (like NA or DH) that is involved in an interaction with a grouping
variable (like our two groups) is the slope of that relationship for the group
that is coded 0. Thus, the regression coefficient for NA in the overall
regression is the same as the regression coefficient for NA within the groups
of participants that was coded 0, i.e., the asthmatics.
TABLE 2. Mean and Standard Deviation Values at
Assessment 1 Assessment 2
FeNO 50.1 ⫾ 28.2 24.2 ⫾ 23.1 47.2 ⫾ 34.8 21.9 ⫾ 13.6
96.4 ⫾ 16.0 103.6 ⫾ 11.0 96.8 ⫾ 14.9 102.9 ⫾ 9.1
NA 12.4 ⫾ 2.7 12.9 ⫾ 2.7 11.8 ⫾ 2.6 12.5 ⫾ 3.5
DH 47.7 ⫾ 17.0 48.9 ⫾ 15.7 42.1 ⫾ 16.8 39.9 ⫾ 18.5
FeNO ⫽ fraction of exhaled nitric oxide; FEV
⫽ forced expiratory volume
in the first second, expressed as the percentage of the norm; NA ⫽ current
negative affect; DH ⫽ frequency of daily hassles.
TABLE 1. Demographics, Lung Function Measurements, and Skin
Test Results for Asthma Patients and Healthy Controls
Mean ⫾ SD
Mean ⫾ SD
Gender, women (%) 69.6 64.0
Age, years 27.8 ⫾ 7.2 28.04 ⫾ 6.2
Education, A-levels or degree (%) 84.8 80.0
Marital status: single (%) 73.9 76.0
Lung function (FEV
99.2 ⫾ 14.5 104.1 ⫾ 12.8
Airway inflammation (FeNO) 48.6 ⫾ 35.2 23.2 ⫾ 25.1
Pollen 80.4 44.0
Trees 82.6 48.0
Animals 93.5 60.0
Mites 89.4 56.0
SD ⫽ standard deviation; FEV
⫽ forced expiratory volume in the first
second, expressed as the percentage of the norm; FeNO ⫽ fraction of exhaled
A. KULLOWATZ et al.
470 Psychosomatic Medicine 70:468– 475 (2008)
relationship between NA and FEV
was significant only in
In this regression analysis, there was also a significant
Group by DH interaction, indicating that for asthma patients,
higher DH tended to lead to higher FEV
(b ⫽ 6.0; t(55) ⫽
1.71; p ⫽ .09), whereas for healthy controls, higher DH
tended to be associated with lower FEV
(b ⫽⫺5.38; t(55) ⫽
⫺1.80; p ⫽ .09).
Repeating the regression analysis employing FeNO as the
dependent variable, we found that higher DH was significantly
related to lower FeNO, but this relationship was also effected
by a borderline significant interaction between Group and DH
(Table 3). Again, to avoid overgeneralizing this effect and
assuming that the relationship was significant in both subject
populations, simple slopes were calculated separately for each
group. This analysis showed that, for asthma patients, higher
DH was related to lower FeNO (b ⫽⫺0.8; t(54) ⫽⫺2.81;
p ⬍ .01), but for healthy controls there was no significant
association between DH and FeNO (b ⫽⫺0.1; t(54) ⫽
⫺0.38; p ⫽ NS). The only other relationship of note from this
regression analysis was a tendency for higher NA to be related
to higher FeNO (b ⫽ 0.8; t(54) ⫽ 1.69; p ⬍ .10).
Airway Inflammation as a Mediator of the Effects of
Psychological Variables on Lung Function
As expected, changes in FeNO were negatively associated
with changes in FEV
(r(43) ⫽⫺0.44, p ⬍ .005), indicating
that lung function declined with higher airway inflammation.
We hypothesized that FeNO would partially mediate the re-
lationship between the psychological variables and FEV
path diagram for this mediation model is shown in Figure 1,
and the path coefficients were calculated using multiple re-
gression analysis (41). This analysis was performed using just
the asthma patients because no relationships between the
psychological variables and lung function were found for
control subjects. When FeNO was included in the multiple
regression analysis along with the psychological predictors of
(NA, DH), the previously significant associations of NA
and DH with FEV
became nonsignificant (b ⫽⫺9.0, t(35) ⫽
⫺1.32, p ⫽ NS, and b ⫽ 1.1, t(35) ⫽ 0.28, p ⫽ NS,
respectively), indicating that the psychological variables no
longer have a significant effect on FEV
when their relation
ship with FeNO was controlled. Further, we directly tested the
significance of the mediated pathways from the NA and DH
through FeNO to FEV
using the z⬘ test. Results demonstrated
that both mediated pathways were significant, z⬘⫽⫺1.57,
p ⬍ .05 for the NA to FEV
mediated pathway, and z⬘⫽1.96,
p ⬍ .05 for the DH to FEV
These results indicated that a significant portion of the
effects of both short-term NA and long-term low-level stress
on lung function were mediated by changes in airway inflam-
mation. Using the procedure suggested by Shrout and Bolger
(42), we found that 37% of the effect of NA on FEV
mediated by FeNO and that 81% of the effect of DH on FEV
was mediated by FeNO.
Because recent evidence also suggests that low positive affect may play a
role in health outcomes (41), we recalculated the analyses replacing the NA
subscale with the Positive Affect subscale of the PANAS. No significant
effects were obtained with this measure.
β = .31*
β = .05
β = -.48***
β = -.43**
β = -.23
Figure 1. Path diagram representing the mediation model for patients suf-
fering from asthma. FeNO ⫽ fraction of exhaled nitric oxide; FEV
expiratory volume in the first second, expressed as the percentage of the norm.
*p ⬍ .05; **p ⬍ .01; ***p ⬍ .001.
TABLE 3. Regression Analyses Using Residualized Change Scores, for All Participants
Lung Function (FEV
Airway Inflammation (FeNO)
⌬ NA ⫺15.44 ⫺0.46 .015 0.80 0.31 .097
⌬ DH 6.01 0.34 .093 ⫺0.76 ⫺0.56 .007
⌬ Ozone level ⫺0.03 ⫺0.13 .366 0.004 0.26 .064
⌬ Pollen allergic load ⫺0.01 ⫺0.02 .881 0.004 0.08 .572
0.24 0.02 .898 ⫺0.02 ⫺0.02 .876
Group by ⌬ NA interaction 17.95 0.34 .061 ⫺0.93 ⫺0.23 .204
Group by ⌬ DH interaction ⫺11.46 ⫺0.45 .021 0.71 0.36 .063
⌬⫽residualized change score between Assessments 1 and 2; FeNO ⫽ fraction of exhaled nitric oxide; FEV
(%) ⫽ forced expiratory volume in the first second,
expressed as the percentage of the norm; NA ⫽ current negative affect; DH ⫽ frequency of daily hassles; ozone level ⫽ concentration of ozone in air one hour
before assessment started; pollen allergic load ⫽ interaction of each participant’s sensitivity to the respective pollen allergen and the actual levels of the respective
pollen allergen in the air.
Patients suffering from asthma were coded 0; controls were coded 1.
MEDIATION EFFECT OF AIRWAY INFLAMMATION
471Psychosomatic Medicine 70:468– 475 (2008)
Role of Inhaled Corticosteroids and Psychological
To investigate whether ICS intake moderated the relation-
ship between the psychological variables and lung function,
we added ICS intake and the interactions of ICS with NA and
DH as additional predictors of lung function and airway
inflammation in an exploratory analysis. A significant inter-
action would indicate that ICS did moderate the relationship
between the psychological variables and lung function. This
regression analysis, using FEV
as the dependent variable,
yielded a significant ICS by DH interaction (b ⫽⫺24.9; t(33) ⫽
⫺3.50; p ⬍ .001). The simple slopes indicated that, for asthma
patients without ICS intake, higher DH was significantly related
to better lung function, b ⫽ 17.0, t(33) ⫽ 3.67, p ⬍ .001, whereas
for patients who took ICS, this relationship was not significant
(b ⫽⫺3.9; t(33) ⫽⫺0.70; p ⫽ NS).
We repeated this regression analysis using FeNO as the
dependent variable. This analysis yielded only a significant
main effect for DH (b ⫽⫺0.8; t(34) ⫽⫺2.28; p ⬍ .05). No
significant interactions with ICS treatment were observed.
One final exploratory regression analysis was performed to
investigate the possibility that ATI psychological triggers
moderated the relationship between the psychological vari-
ables and lung function. This analysis showed no significant
effects for psychological triggers.
In this study, we found evidence for an association between
psychological factors and both airway inflammation and lung
function, but only in patients with asthma. Our mediational
analysis suggested that, in these asthma patients, effects of
psychological factors on lung function were dependent on
changes in airway inflammation. An unexpected finding of
our study was the divergent effects of short-term negative
mood from long-term low-level stress measured by DH.
Whereas increases in current negative mood were associ-
ated with a decline in lung function, increases in DH tended
to be related to an improvement in lung function. Similarly,
more negative mood tended to be associated with greater
airway inflammation and more DH were related to lower
airway inflammation. Although some of the associations
were only marginally significant, the compatibility of the
patterns of association was compelling and in line with the
expectation that stronger airway inflammation is the basis
of compromised lung function (9).
Our results on negative mood for patients with asthma are
consistent with prior studies that suggested a decline in lung
function with NA induced in the laboratory or observed in
daily life (5,7). Mechanisms underlying short-term changes in
lung function are still poorly understood; however, potential
explanatory pathways include autonomic excitation, espe-
cially from a vagal origin (43), and/or ventilatory adjustment
(44). Our findings also direct attention to the role of airway
inflammation as a mediator of such changes. The observed
increases in airway inflammation are in line with the general
expectation that higher stress levels would increase inflamma-
tory processes in asthma and allergy (17,18,22). Both activa-
tion of the hypothalamic-pituitary-adrenal (HPA) axis and the
sympathetic nervous system through stress could lead to im-
munoregulatory changes, with a shift from a Type 1 cytokine-
based cellular immunity response to antigen challenge toward
a Type 2 cytokine response supporting a humoral response
(18,45). Cytokines are produced by a variety of cells, in
particular, subpopulations of CD4⫹helper T cells; and Type 2
cytokines are known to be involved in inflammatory processes
in the airways. However, it is not known whether the time
frames for this pathway of inflammatory exacerbation would
be compatible with the changes in short-term NA we ob-
served. A recently reported peak-flow diary study with asth-
matic children suggested that negative life events typically
require a few days to unfold their adverse impact on asthma
status (8). Protocols with repeated measurements of NA and
inflammatory and endocrine markers over an extended period
of time would be needed to further elucidate the time course
and mechanisms underlying asthma exacerbations induced by
inflammatory responses to negative psychosocial events.
The association of low-level long-term daily stress with
better lung function and reduced airway inflammation in pa-
tients with asthma was unexpected. These findings echo the-
oretical considerations regarding the beneficial effects of
stress on asthma pathophysiology (46). Both sympathetically
mediated bronchodilation and anti-inflammation effects of
HPA axis activation have been discussed, with the paradoxical
expectation that stressful life episodes would reduce rather
than exacerbate asthma symptoms. However, experimental
studies have rarely found evidence for improvement in lung
function when short-term stress was induced (47,48). But,
such theoretical predictions regarding potential benefits of
stress are at variance with the empirical results just cited,
which suggest that inflammation is exacerbated by stress, and
they cannot be invoked in explaining associations with short-
term NA found in our study.
One solution to these contradictions could lie in the role of
perceived control of stress. Diverging patterns of immune
response to stressors have been observed for stressors per-
ceived as controllable or uncontrollable by individuals (49).
Thus, although life events and other intense stressors may
constitute instances of uncontrollable or overwhelming stress,
DH may still be experienced as controllable challenges by the
patient, or at least they may not induce feelings of being
overwhelmed or helpless. Hassles are also described as being
ubiquitous and most people perceive hassles as an unavoid-
able part of life, an experience from which they expect to
recover relatively quickly (50). It is possible that, because of
the repeated occurrence of these DH, people habituate to them
or learn to cope with them, and thus they would have minimal
impact on people’s level of psychological stress, or even
strengthen coping resources that promote better health. Thus,
on an immunoregulatory level, inflammatory processes could
be attenuated with more frequent DH due to the reduction in
stress hormone response to individual hassles. In line with
that, longitudinal studies demonstrating adverse effects of
A. KULLOWATZ et al.
472 Psychosomatic Medicine 70:468– 475 (2008)
stress in asthma have mostly focused on distinct life events
that arouse stronger NA or distress (8) and cortisol levels are
often reported to be low in chronic versus acute stress (51). On
the other hand, the type of chronic stress implemented in
animal studies that have shown adverse effects on airway
inflammation is typically of an uncontrollable character (19 –
21). Challenges that require passive coping or withdrawal
have also been thought to favor parasympathetic activity,
although demonstrating this effect in asthma has remained
unsuccessful (48). More refined conceptualizations of stress
and coping should be invoked in future studies of psychoso-
cial factors and asthma. It should also be noted that other studies
of stress effects on immune parameters have found similar di-
verging effects of acute, short-term stress versus chronic long-
term stress (52), but the extent to which the observed regulatory
processes can be compared with those underlying inflammatory
asthma exacerbation remains unclear.
Although the validity of elevated FeNO as a marker of the
extent of inflammation of the airways is firmly established, the
actual function of nitric oxide in asthma has been a subject of
debate (53,54). Nitric oxide is a key biological messenger, play-
ing a role in a variety of biological processes, such as blood
vessel dilatation and neurotransmission. Based on its various
bioactive forms and depending on a wide local concentration
range, it could have either protective or deleterious effects during
states of airway inflammation, damage, and repair (54). For
example, with regard to potentially protective effects, a recent
animal study showed that endogeneous nitric oxide dampens the
airway constrictor effects of cholinergic excitation (55). Thus,
care must be taken in the interpretation of changes in FeNO and
their potential relevance for behavioral states in health and
In this study, we made an effort to control for possible
mediators and moderators of the relationship between affect or
stress and airway pathophysiology. Although this constitutes
an improvement over most prior research attempting to study
such associations, our operationalizations of allergen load and
ozone exposure were global, using overall regional pollen
counts and air pollution reports. Much greater methodological
sophistication would be needed to monitor individual patterns
of exposure, which are not only dependent on patients’ be-
havior (including indoor versus outdoor self-exposure), but
also on factors, such as regional variations in air pollution
levels and spatial variability in deposition of pollen (56).
Nevertheless, ozone tended to be associated with airway in-
flammation, although controlling for it did not abolish the
observed effects of stress on the airways. The comparably
weak association between ozone levels and FeNO may be
explained by activation of different types of inflammatory cells.
Neutrophils typically infiltrate the airways in response to ozone
(30), whereas eosinophils are activated by allergen exposure and
are one of the major sources of nitric oxide in the exhaled breath.
Exploratory analysis of differences between patients taking
or not taking ICS did suggest that this factor influenced only
the association between DH and mechanical lung function,
with corticosteroids dampening the association. Although not
spectacular, it demonstrates a need to control for patients’
medication regimen in this research. This analysis was limited
by not including dosage information for ICS medication.
Given the well-known problems with medication adherence in
asthma (57), electronic monitoring of inhaler acuations would
be needed to ensure the validity of more refined information
on medication use. We also explored the role of patients’ prior
adverse experience with psychological triggers as a potential
moderator variable in the relationship between psychological
and pathophysiological variables because we had previously
found an association between psychological trigger reports
and airway responses to emotional stimulation (35). However,
those patients reporting to be more susceptible to psycholog-
ical triggers did not show a different pattern of associations
than those reporting not to be susceptible.
Beyond the limitations in operationalizing some of the
control variables, our study was also limited due to the small
sample size. This made it difficult to test the role of all
moderators and mediators in one model, reduced some of the
analyses to a more exploratory status, and increased the pos-
sibility of the findings being spurious. Also, the sample of
patients consisted mainly of mild-to-moderate cases of asthma
whose lung function was clearly in the normal range. Al-
though FeNO values were indicative of disease-typical airway
inflammation (confirming dissociations between indicators of
pathophysiology described in mild asthma before) (58), our
study cannot contribute to knowledge about patients suffering
from more severe asthma. But, other studies have demon-
strated the importance of psychosocial factors for patients
with severe asthma (59). Another shortcoming of the study is
the retrospective nature of the DH assessment, which could be
associated with a recall bias. Daily monitoring of mood and
DH would yield more valid information on actual levels of
hassles patients were confronted with. A fourth limitation is
that, due to the design of our study, causal conclusions are not
possible. The assumption of our mediation analysis, with
airway inflammation acting as the mediator of lung function
change, was largely guided by current evidence and plausibil-
ity, which suggests that inflammation leads to bronchocon-
striction (9,12) rather than vice versa. Also, there is currently
substantial evidence indicating a modulatory role of psycho-
logical factors on immune function (52), rather than influences
of altered immune function on psychological states.
We conclude that, for patients with asthma, psychological
variables are associated with airway inflammation and lung
function changes, with airway inflammation acting as a me-
diator. The observed divergent associations of current NA and
DH indicate that, with regard to airway pathophysiology, care
must be taken to distinguish between different forms of psy-
chosocial challenge and stress. Recording the mood and stress
levels of patients suffering from asthma could be beneficial to
optimize management of the disease. However, given the
limitations of the present study, replication of our findings is
clearly warranted. More studies are needed with assessments
of mood, DH, lung function, and airway inflammation over
longer periods of time.
MEDIATION EFFECT OF AIRWAY INFLAMMATION
473Psychosomatic Medicine 70:468– 475 (2008)
We are indebted to the staff of the Pulmonary Research Institute,
Hospital Grosshansdorf, Germany for their help.
1. Lehrer PM, Isenberg S, Hochron SM. Asthma and emotion: a review.
J Asthma 1993;30:5–21.
2. Wright RJ, Rodriguez M, Cohen S. Review of psychosocial stress and
asthma: an integrated biopsychosocial approach. Thorax 1998;53:
3. Ritz T, Kullowatz A. Effects of emotion and stress on lung function in
health and asthma. Curr Respir Med Rev 2005;1:209 –18.
4. Isenberg SA, Lehrer PM, Hochron S. The effects of suggestion and
emotional arousal on pulmonary function in asthma: a review and a
hypothesis regarding vagal mediation. Psychosom Med 1992;54:
5. Ritz T, Steptoe A, DeWilde S, Costa M. Emotions and stress increase
respiratory resistance in asthma. Psychosom Med 2000;62:401–12.
6. Affleck G, Apter A, Tennen H, Reisine S, Barrows E, Willard A, Unger
J, ZuWallack R. Mood states associated with transitory changes in asthma
symptoms and peak expiratory flow. Psychosom Med 2000;62:61– 8.
7. Ritz T, Steptoe A. Emotion and pulmonary function in asthma: reactivity
in the field and relationship with laboratory induction of emotion. Psy-
chosom Med 2000;62:808 –15.
8. Sandberg S, Jarvenpaa S, Penttinen A, Paton JY, McCann DC. Asthma
exacerbations in children immediately following stressful life events: a
Cox’s hierarchical regression. Thorax 2004;59:1046–51.
9. Rosias PP, Dompeling E, Dentener MA, Pennings HJ, Hendriks HJ, Van
Iersel MP, Jobsis Q. Childhood asthma: exhaled markers of airway
inflammation, asthma control score, and lung function tests. Pediatr
10. Kharitonov SA, Barnes PJ. Exhaled markers of inflammation. Curr Opin
Allergy Clin Immunol 2001;1:217–24.
11. Battaglia S, den Hertog H, Timmers MC, Lazeroms SP, Vignola AM,
Rabe KF, Bellia V, Hiemstra PS, Sterk PJ. Small airways function and
molecular markers in exhaled air in mild asthma. Thorax 2005;60:
639 – 44.
12. Harkins MS, Fiato KL, Iwamoto GK. Exhaled nitric oxide predicts
asthma exacerbation. J Asthma 2004;41:471–6.
13. Pijnenburg MW, Bakker EM, Lever S, Hop WC, De Jongste JC. High
fractional concentration of nitric oxide in exhaled air despite steroid
treatment in asthmatic children. Clin Exp Allergy 2005;35:920 –5.
14. Malmberg LP, Pelkonen AS, Haahtela T, Turpeinen M. Exhaled nitric
oxide rather than lung function distinguishes preschool children with
probable asthma. Thorax 2003;58:494 –9.
15. Kanniess F, Richter K, Bo¨hme S, Jo¨rres RA, Magnussen H. Montelukast
versus fluticasone: effects on lung function, airway responsiveness and
inflammation in moderate asthma. Eur Respir J 2002;20:853– 8.
16. Kharitonov SA, Yates DH, Barnes PJ. Inhaled glucocorticoids decrease
nitric oxide in exhaled air of asthmatic patients. Eur Respir J 1997;10:
17. Kang DH, Coe CL, McCarthy DO, Jarjour NN, Kelly EA, Rodriguez RR.
Cytokine profiles of stimulated blood lymphocytes in asthmatic and
healthy adolescents across the school year. J Interferon Cytokine Res
18. Marshall GD, Agarwal SK. Stress, immune regulation, and immunity:
applications for asthma. Allergy Asthma Proc 2000;21:241–6.
19. Datti F, Datti M, Antunes E, Teixeira NA. Influence of chronic unpre-
dictable stress on the allergic responses in rats. Physiol Behav 2002;77:
79 – 83.
20. Joachim RA, Quarcoo D, Arck PC, Herz U, Renz H, Klapp BF. Stress
enhances airway reactivity and airway inflammation in an animal model
of allergic bronchial asthma. Psychosom Med 2003;65:811–5.
21. Forsythe P, Ebeling C, Gordon JR, Dean Befus A, Vliagoftis H. Oppos-
ing effects of short-term and long-term stress on airway inflammation.
Am J Respir Crit Care Med 2004;169:220– 6.
22. Liu LY, Coe CL, Swenson CA, Kelly EA, Kita H, Busse WW. School
examinations enhance airway inflammation to antigen challenge. Am J
Respir Crit Care Med 2002;165:1062–7.
23. Chen E, Fisher EB, Bacharier LB, Strunk RC. Socioeconomic status,
stress, and immune markers in adolescents with asthma. Psychosom Med
24. Ho¨glund CO, Axen J, Kemi C, Jernelov S, Grunewald J, Muller-Suur C,
Smith Y, Gronneberg R, Eklund A, Stierna P, Lekander M. Changes in
immune regulation in response to examination stress in atopic and healthy
individuals. Clin Exp Allergy 2006;36:982–92.
25. Wittchen H-U, Zaudig M, Fydrich T. SKID-I. Strukturiertes Klinisches
Interview fuer DSM-IV [Structured clinical interview for DSM-IV].
Goettingen: Hogrefe; 1997.
26. American Thoracic Society; European Respiratory Society. ATS/ERS
recommendations for standardized procedures for the online and offline
measurement of exhaled lower respiratory nitric oxide and nasal nitric
oxide, 2005. Am J Respir Crit Care Med 2005;171:912–30.
27. American Thoracic Society. Standardization of spirometry. 1994 update.
Am J Respir Crit Care Med 1995;152:1107–36.
28. Ritz T, Dahme B, Dubois AB, Folgering H, Fritz GK, Harver A, Kotses
H, Lehrer PM, Ring C, Steptoe A, Van de Woestijne KP. Guidelines for
mechanical lung function measurements in psychophysiology. Psycho-
physiology 2002;39:546 – 67.
29. Baraldi E, Carra S, Dario C, Azzolin N, Ongaro R, Marcer G. Effect of
natural grass pollen exposure on exhaled nitric oxide in asthmatic chil-
dren. Am J Respir Crit Care Med 2005;171:1065– 6.
30. Joerres RA, Holz O, Zachgo W, Timm P, Koschyk S, Muller B, Grimminger
F, Seeger W, Kelly FJ, Dunster C, Frischer T, Lubec G, Waschewski M,
Niendorf A, Magnussen H. The effect of repeated ozone exposures on
inflammatory markers in bronchoalveolar lavage fluid and mucosal biopsies.
Am J Respir Crit Care Med 2000;161:1855– 61.
31. Krohne HW, Egloff B, Kohlmann C-W, Tausch A. Untersuchungen mit
einer deutschen Version der “Positive and Negative Affect Schedule”
(PANAS) [Studies with a German version of the “Positive and Negative
Affect Schedule” (PANAS)]. Diagnostica 1996;139–56.
32. Lazarus RS, DeLongis A. Psychological stress and coping in aging. Am
33. Kanner AD, Coyne JC, Schaefer C, Lazarus RS. Comparison of two
modes of stress measurement: daily hassles and uplifts versus major life
events. J Behav Med 1981;4:1–39.
34. Kohn PM, Lafreniere K, Gurevich M. The inventory of college students’
recent life experiences: a decontaminated hassles scale for a special
population. J Behav Med 1990;13:619–30.
35. Ritz T, Kullowatz A, Kanniess F, de Vries U, Dahme B, Magnussen H.
Perceived triggers of asthma: evaluation of a German version of the
Asthma Trigger Inventory. Respir Med 2008;102:390 – 8.
36. National Heart, Lung, and Blood Institute/World Health Organization.
NHLBI/WHO workshop report: Global strategy for asthma management
and prevention. Bethesda, MD: National Institutes of Health, 2006.
Available at http://www.ginasthma.com.
37. Tabachnick BG, Fidell LS. Using Multivariate Statistics. 5th ed. Boston,
MA: Allyn & Bacon; 2007.
38. Aiken LS, West PE. Multiple Regression: Testing and Interpreting Inter-
actions. Newbury Park: Sage Publications; 1992.
39. MacKinnon DP, Lockwood CM, Hoffman JM, West SG, Sheets V. A
comparison of methods to test mediation and other intervening variable
effects. Psychological Methods 2002;7:83–104.
40. Ritz T, Steptoe A, Bobb C, Harris C, Edwards M. The Asthma Trigger
Inventory: validation of a questionnaire for perceived triggers of asthma.
Psychosom Med 2006;68:956 – 65.
41. Pressman SD, Cohen S. Does positive affect influence health? Psychol
42. Shrout PE, Bolger N. Mediation in experimental and nonexperimental
studies: new procedures and recommendations. Psychological Methods
43. Kullowatz A, Smith H-J, Kanniess F, Magnussen H, Dahme B, Ritz T.
Testing the vagal pathway for emotion-induced airway obstruction in
asthma: findings with pharmacologic blockade. Psychosom Med 2006;
44. Clarke PS. Emotional exacerbations in asthma caused by overbreathing.
J Asthma 1982;19:249 –51.
45. Agarwal SK, Marshall GD Jr. Glucocorticoid-induced type 1/type 2
cytokine alterations in humans: a model for stress-related immune dys-
function. J Interferon Cytokine Res 1998;18:1059 – 68.
46. Steptoe A. Psychological aspects of bronchial asthma. In: Rachman F,
editor. Contributions to Medical Psychology. Vol 3. Oxford, UK: Perga-
mon Press; 1984.
47. Laube BL, Curbow BA, Fitzgerald ST, Spratt K. Early pulmonary re-
sponse to allergen is attenuated during acute emotional stress in females
with asthma. Eur Respir J 2003;22:613– 8.
48. Lehrer PM, Hochron S, Carr R, Edelberg R, Hamer R, Jackson A, Porges
S. Behavioral task-induced bronchodilation in asthma during active and
A. KULLOWATZ et al.
474 Psychosomatic Medicine 70:468– 475 (2008)
passive tasks: a possible cholinergic link to psychologically induced
airway changes. Psychosom Med 1996;58:413–22.
49. Brosschot JF, Godaert GL, Benschop RJ, Olff M, Ballieux RE, Heijnen
CJ. Experimental stress and immunological reactivity: a closer look at
perceived uncontrollability. Psychosom Med 1998;60:359– 61.
50. Meyer IH. Prejudice as stress: conceptual and measurement problems.
Commentary. Am J Public Health 2003;93:262–5.
51. Miller GE, Chen E, Zhou ES. If it goes up, must it come down? Chronic
stress and the hypothalamic-pituitary-adrenocortical axis in humans. Psy-
chol Bull 2007;133:25– 45.
52. Segerstrom SC, Miller GE. Psychological stress and the human immune system:
a meta-analytic study of 30 years of inquiry. Psychol Bull 2004;130:601–30.
53. Ricciardolo FLM, Sterk PJ, Gaston B, Folkerts G. Nitric oxide in health
and disease of the respiratory system. Physiol Rev 2004;84:731– 65.
54. Sanders SP. Nitric oxide in asthma. Pathogenic, therapeutic, or diagnos-
tic? Am J Respir Cell Mol Biol 1999;21:147–9.
55. Khssawneh MY, Dreshaj IA, Liu S, Chang CH, Haxhiu MA, Martin
RJ. Endogenous nitric oxide modulates responses of tissue and airway
resistance to vagal stimulation in piglets. J Appl Physiol 2002;93:
56. Frenz DA. Interpreting atmospheric pollen counts for use in clinical
allergy: spatial variability. Ann Allergy Asthma Immunol 2000;84:
57. Rau JL. Determinants of patient adherence to an aerosol regimen. Respir
Care 2005;50:1346 –59.
58. Silvestri M, Spallarossa D, Battistini E, Brusasco V, Rossi GA. Disso-
ciation between exhaled nitric oxide and hyperresponsiveness in children
with mild intermittent asthma. Thorax 2000;55:484 – 8.
59. ten Brinke A, Ouwerkerk ME, Zwinderman AH, Spinhoven P, Bel EH.
Psychopathology in patients with severe asthma is associated with in-
creased health care utilization. Am J Respir Crit Care Med 2001;163:
MEDIATION EFFECT OF AIRWAY INFLAMMATION
475Psychosomatic Medicine 70:468– 475 (2008)