The inflammation produced by corticosteroid inhalers in the
ORIGINAL ARTICLE
pharynx in asthmatics
Bhalla, R.K., Taylor, W., Jones, A.S. & Roland, N.J.
Department of Otolaryngology ⁄ Head & Neck Surgery, University Hospital Aintree, Liverpool, UK
Accepted for publication 29 September 2008
Clin. Otolaryngol. 2008, 33, 581–586
Objectives: To investigate inflammation of the pharyn-
geal mucosa caused by inhaled corticosteroids.
Design: Prospective, cross-sectional, single-blinded
study.
Setting: University Hospital Aintree, Liverpool, UK.
Patients: Fifty adults were recruited from two local gen-
eral practices and from general ENT clinics at our Uni-
versity hospital. Patients were allocated to one of four
groups according to use of inhaled corticosteroids and
the presence of adverse local side effects.
Main outcome measures: Scores achieved on a respira-
tory symptom questionnaire. Histological markers of
inflammation and their correlation with pharyngitis. Sta-
tistical modelling included univariate and multivariate
analyses, which included multiple linear and logistic
regression, and discriminant analysis.
Results: The regular use of inhaled corticosteroids pre-
disposed subjects to hoarseness, weakness of voice, sore
throat and throat irritation (P < 0.0001). Pharyngitis was
significantly different between the groups (P < 0.0001).
Furthermore, those not using an inhaled corticosteroid
regularly had little or no clinically apparent pharyngitis,
whereas those using an inhaled corticosteroid regularly
Introduction
The pharynx is susceptible to inflammatory changes as a
consequence of inhaled corticosteroid use. This may be
because up to 80% of the inhaled dose is deposited on
the pharyngeal mucosa prior to being swallowed.1 An
accurate explanation is difficult however, as corticoster-
oids are powerful anti-inflammatory agents.
Glucocorticoids are thought to suppress inflammation
in asthma by increasing the synthesis of lipocortin-1 or
Correspondence: R.K. Bhalla, Consultant ENT ⁄ Head & Neck Surgeon,
Department of Otorhinolaryngology, Manchester Royal Infirmary,
Oxford Road, Manchester M13 9WL, UK. Tel.: +44 161 276 4426; fax:
+44 161 276 5003; e-mail: rajiv.bhalla@cmmc.nhs.uk
 2008 The Authors
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 581–586
had significantly higher pharyngitis scores (P = 0.0204).
Similarly, weakness of voice (P = 0.0234), hoarseness
(P < 0.001) and sore throat (P < 0.001) were also more
common in those patients that used an inhaled cortico-
steroid on a regular basis. To our surprise, however, cel-
lular markers of inflammation did not corroborate the
appearances of clinical examination. We found that the
five most important discriminators, between those that
were using inhaled corticosteroid therapy regularly and
those that were not, to be intra-epithelial inflammatory
cells (scdf )1.2939); age (scdf 0.8389); use of a spacer
device (scdf 0.5456); sore throat (scdf 0.4230) and throat
irritation (scdf 0.4015). The groups were significantly
different (P < 0.0001). The statistical model used, classi-
fied 68% of the cases correctly into their respective
groups.
Conclusions: Inhaled corticosteroids predispose to phar-
yngitis and an inflammatory infiltrate. However, the clini-
cal diagnosis of pharyngitis does not correlate well with
cellular inflammatory infiltrate and is therefore, not a
reliable measure of underlying inflammation. We advo-
cate caution in the use of pharyngeal erythema as a mea-
sure of underlying inflammation.
by reducing the transcription of interleukins 3 and 5. The
consummate effect is to reduce the production of inflam-
matory mediators.2,3 Vasconstrictive effects reduce bron-
chial mucosa oedema and thickening and beta-adrenergic
responsiveness is enhanced. The long-term use of inhaled
corticosteroids may decrease the immediate response to
allergen and exercise by reducing the number of mast
cells in the airway.4 The benefits of inhaled corticoster-
oids are unquestionable. However, adverse mucosal effects
causing inflammation may result in pain, a chronic sore
throat, hoarseness and odynophagia.5,6 Subsequent com-
pliance with therapy can be hampered by the local side
effects.
The complications of inhaled corticosteroid therapy are
not uncommon. More than 60% of one sample group of
581
582 R.K. Bhalla et al.
children using inhaled corticosteroids were afflicted by at
least one local side effect in daily life.7 Several studies
have assessed the frequency of oropharyngeal and laryn-
geal candidiasis.8,9 Dysphonia usually resolves after stop-
ping inhaled corticosteroid therapy.10
Little is known about the very action that causes
inflammation and indeed, whether there is an underlying
susceptibility in the pharyngeal mucosa of asthmatics.
Furthermore, no cellular assessment of inflammation has
been made in asthmatics. This study has provided a base-
line comparison of normal and asthmatic pharyngeal
mucosa. It also examines whether pharyngitis is a reliable
indicator of underlying cellular inflammatory infiltrate.
Methods
The study design was prospective, cross-sectional and sin-
gle-blinded.
Recruitment
Fifty volunteers were recruited from two local general
practices and from general ENT clinics at our University
hospital. An assessment of inhaler technique was made at
the point of recruitment, and guidance provided if the
technique of delivery was found to be sub-optimal, with
special reference to sufficient shaking of the canister prior
to inhalation, breathing and timing of delivery with inspi-
ration. Patients were also educated regarding the function
of their corticosteroid inhaler versus the short or long-act-
ing b2-agonist and the benefits of compliant use at the
prescribed regimen.
Ethical considerations
The local research and ethics committee approved a pro-
tocol for the study, patient information sheet and patient
consent form. Patients were allocated a study number for
anonymity.
Inclusion criteria
Individuals over the age of 18 and able to partake in
informed consent. The healthy controls (group A) did
not have documented respiratory disease and were not
using inhaled corticosteroids, short or long-acting b2-
agonists or topical intranasal steroids. Asthmatics (groups
B–D) had a baseline predicted FEV1 of >60% and <90%.
An increase in FEV1 of >12% after reversibility testing
had been established during the preceding 6 months.
Subjects using inhaled corticosteroids were required to
have done so for at least 30 consecutive days leading into
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 581–586
the study. There were no patients using nebulised prepa-
rations. Patients were advised to stop over-the-counter or
prescription, topical or systemic antihistamines at least
7 days prior to enrolment in the study. None of the
patients were taking leukotriene inhibitors.
Allocation to study groups
Subjects were allocated to one of four groups on the basis
of inhaled corticosteroid use and adverse symptoms at
the outset of the study (Table 1). Asthmatic patients were
deemed symptomatic if they used an inhaled corticoste-
roid and had hoarseness, pharyngeal discomfort or sore
throat of more than one-month duration. Other causes of
these symptoms were excluded by history and thorough
ENT examination. Seasonal asthmatics gave a history of
using their inhalers infrequently, for example ‘once a
month’ or ‘over the summer months’, but had been diag-
nosed as asthmatic by their general practitioner because
of variability in peak flows. These individuals were either
not or very infrequently using their inhaled corticosteroid
during the period of the study.
Exclusion criteria
Subjects were excluded if other causes of pharyngeal or
laryngeal disease were identified at the initial assessment.
These included a history of reflux, malignancy, vocal fold
lesions or neurological causes of dysphonia. Those using
topical intranasal steroids to treat co-existent inflamma-
tory or allergic nasal disease were excluded. Those with
an upper respiratory tract infection in the preceding
4 weeks, who were medically unfit because of intercurrent
disease, or had chronic obstructive pulmonary disease,
were excluded. Poorly controlled asthmatics requiring
Table 1. Description of study groups
Group Description
A Normal healthy volunteers. (4)
B Seasonal asthmatics: patients using inhalers
(rescue and ⁄ or preventer medications) infrequently,
and asymptomatic. (12)
C Asthmatic patients using inhaled corticosteroids
regularly, and were symptomatic. (25)
D Asthmatic patients using inhaled corticosteroids
regularly, and were asymptomatic. (9)
Symptomatic asthmatics were defined as those using an
inhaled corticosteroid with adverse symptoms of hoarseness,
pharyngeal discomfort and ⁄ or sore throat of more than one-
month duration.
Numbers in each group are given in parentheses.
 2008 The Authors

systemic steroid or antibiotic therapy in the preceding
4 weeks and those who had been hospitalised or suffered
an acute asthma attack in the preceding 3 months were
excluded. Patients using Seretide (Allen & Hanburys Ltd,
Uxbridge, England) were excluded. Seretide is a combina-
tion of fluticasone propionate and salmeterol, a long-
acting b2-agonist. Inclusion of this group would potentially
have confounded the results, as it would have been
impossible to attribute inflammation solely to the inhaled
corticosteroid.
Data collection
Subjects withheld their short-acting b2-agonist for 6 h
and long-acting b2-agonist for 12 h prior to assessment.
An asthma and allergy history determined eligibility for
inclusion. A respiratory symptom questionnaire was com-
pleted. This tool had been designed, piloted and used in
a prior prevalence study. Pharyngeal discomfort at the
time of assessment was scored subjectively using a
ten-centimetre visual analogue scale. A thorough ENT
examination was performed. The lead investigator alone,
utilising his clinical acumen as a practicing ENT physi-
cian, graded pharyngeal inflammation as 0 (none), 1
(mild), 2 (moderate) or 3 (severe). Pharyngeal assess-
ment was controlled with regard to headlight and room
lighting. Candida infection was determined by skin
prick test and identification of fungal hyphae in the his-
tological specimens. A standardised biopsy of the pos-
terior pharyngeal wall was taken. Histological specimens
were examined without prior knowledge of the donor
(investigator-blind).
Pharyngeal biopsy technique
After spraying the posterior pharyngeal wall mucosa with
4% Lidocaine to provide surface anaesthesia, a biopsy
was taken using micro-cupped laryngeal forceps. The pos-
terior pharyngeal wall, for the purpose of this study, was
the mucosal surface of the throat that extended from the
soft palate in a mouth-breathing patient to the level of
the depressed tongue. The most inflamed area was
selected as the biopsy site. The specimen was fixed in
Formalin and processed for histological examination.
Histopathological examination
One investigator examined histological specimens blindly.
Specimens were stained with haematoxylin and eosin.
The squamous epithelium and underlying connective tis-
sue were assessed. Markers of acute and chronic inflam-
mation were graded. Cells typical of acute inflammation
 2008 The Authors
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 581–586
The inflammation produced by corticosteroid inhalers 583
were neutrophils. Those characteristic of chronic inflam-
mation were lymphocytes and plasma cells. The presence
of eosinophils and macrophages was also recorded. Vas-
cular markers of inflammation included oedema, vessel
dilation and congestion. The degree of penetration of
connective tissue papillae into the overlying squamous
epithelium was gauged. Epithelial thickness, basal cell
hyperplasia and parakeratosis were recorded. Histopatho-
logical findings were graded as present or absent, or as 1
(normal), 2 (mild inflammation) or 3 (marked inflamma-
tion).
Statistical analysis
For comparisons between groups, the Kruskal–Wallis H
test was applied. For comparisons between the two active
groups, C and D, the Mann–Whitney U test was utilised.
For all calculations, the Bonferoni correction was used
because of the large number of variables studied. The
multivariate analysis included multiple linear and logistic
regression, and discriminant analysis. The discriminant
analysis technique was used for categorising data and had
similarities with multiple logistic regression, but with the
advantage of being able to study more than two catego-
ries. The variables assessed included age, smoking, dura-
tion of asthma, duration of inhaled corticosteroid use,
use of a spacer device, onset of adverse side effects after
commencing inhaled corticosteroid therapy, concurrent
b2-agonist use, symptom reporting (cough, sore throat,
throat irritation, hoarseness, weakness of voice, aphonia),
pharyngitis and histological markers of inflammation
(epithelial thickness, basal cell hyperplasia, parakeratosis,
extension of connective tissue papillae into the overlying
squamous epithelium and vascular markers such as
oedema, dilation or congestion).
Results
The fifty recruited volunteers were allocated to their
respective groups according to the definitions provided
earlier. The age range was 20–81 years, median 59 years.
There were 26 males. There were seven smokers: one in
the normal group and two each in the other three groups.
In this sample, smoking did not appear to predispose to
statistically significant adverse symptoms. Skin prick test-
ing excluded Candida as a cause for pharyngitis in the
entire sample. Furthermore, histological examination did
not detect fungal hyphae in any of the sample.
Pharyngitis was graded in all fifty individuals (Fig. 1).
The median and mean discomfort scores in inhaled corti-
costeroid users were 3.7 and 3.45 respectively. On this
scale, 0 indicated no discomfort caused by the inhaled
584 R.K. Bhalla et al.

14 1.3621). Between groups C and D, hoarseness and phar-

yngitis were highly correlated (P = 0.0015, OR = 12.3204
12 None & P = 0.0049, OR = 8.7481).
Mild
Moderate
The discriminant analysis technique was used for cate-
10
Severe gorising data and had similarities with multiple logistic
regression, but with the advantage of being able to study
8
more than two categories. For the purpose of this analy-
sis, the normal and seasonal asthmatics were considered
6
as one group because of the paucity of adverse symptoms
in the latter group. Therefore, with reference to three
4
study groups (normals ⁄ seasonal asthmatics versus asymp-
2
tomatic asthmatics versus symptomatic asthmatics), we

0 (a)
A B C D
Severity of pharyngitis at clinical examination

Fig. 1. Bar chart to illustrate the severity of pharyngitis at clini-

cal examination. Group A were healthy volunteers; Group B
were seasonal asthmatics; Group C were symptomatic asthmatics
using inhaled corticosteroids regularly; and Group D were
asymptomatic asthmatics, also using inhaled corticosteroids reg-
ularly (see Table 1).

corticosteroid and 10 indicated the most severe discomfort

imaginable.

Biopsy data

All fifty biopsies were adequate for assessment of inflam-

matory infiltrate. Lymphocytes and plasma cells, markers
of chronic inflammation, predominated over other cell (b)
types in the connective tissue, with greater subsequent
migration of lymphocytes into the squamous epithelium
(Fig. 2). Although pharyngitis was commoner in asthmat-
ics (P < 0.0001), this was not reflected in either the
inflammatory infiltrate or the other markers of inflamma-
tion (epithelial thickness, basal cell hyperplasia, parakera-
tosis, extension of connective tissue papillae into the
overlying squamous epithelium and vascular markers
such as oedema, dilation or congestion).

Statistical modelling

The multivariate analysis included multiple linear and

logistic regression, and discriminant analysis. For the
multiple linear and logistic regressions, the stepwise pro-
cedure was used and all variables were included in the
model. As the model was refined, the non-significant
variables were removed leaving only those that were sig-
nificant. Consequently, both weakness of voice and phar-
yngitis were positively correlated from group A to D
(P = 0.0234, Odds Ratio [OR] = 1.227 & P = 0.0395, OR
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 581–586
Fig. 2. Photomicrographs showing: (a) A chronic inflammatory
infiltrate of lymphocytes and plasma cells in the squamous epi-
thelium and underlying connective tissue, (b) A high-powered
view with marked oedema of the connective tissue and associ-
ated blood vessel dilatation.
 2008 The Authors
 The inflammation produced by corticosteroid inhalers 585

were interested to know if a model built of the variables
could then successfully classify the three disease groups,
i.e. to tell us if the variables in the model were important
and, of course, different between the disease categories.
The results of this analysis showed a large Eigenvalue of
3.2465 with a P-value of P = 0.0003. This indicated a
large difference between the three groups. Based on a cut-
off of 0.40, the five most important discriminators (in
order of importance) with their standardised coefficients
of the discriminant functions (scdf) were: the presence of
intra-epithelial inflammatory cells (scdf )1.2939); age
(scdf 0.8389); use of a spacer device (scdf )0.5456); sore
throat (scdf )0.4230) and throat irritation (scdf 0.4015).
A negative sign denoted an inverse effect. Assessment of
the mahalanobis distance between the groups with signifi-
cance values showed that the three test groups (nor-
mal ⁄ seasonal asthmatics versus asymptomatic asthmatics
versus symptomatic asthmatics) were all very significantly
different and, in particular, the asymptomatic and symp-
tomatic were different based on the variables we have.
The model used classified 68% of the cases correctly into
the three groups.
Discussion
But why would an anti-inflammatory steroid preparation
cause inflammation in the upper airway? This observation
is very likely to be multifactorial and may depend on
several factors (Table 2). Snoring was not evaluated or
controlled for as a potential risk factor within the context
of this study, but the investigators accept that mouth-
breathing in general may well have contributed to phar-
yngitis. However, as discussed in the methodology,
patients did not have significant sinonasal disease that
warranted topical intranasal steroid use. Further research
is undoubtedly required to evaluate a potential associa-
tion between mouth-breathing ⁄ snoring and pharyngitis.
Current measures to reduce the risk of oropharyngeal
inflammation
Unfortunately, the management of unwanted side effects
is wholly unsatisfactory. Sufferers are advised to gargle
after using the inhaler and commence using a spacer
device.13 However, despite stringent compliance to these
measures, there remains a group of asthmatics that con-
tinue to be affected by corticosteroid local side effects.
This is not surprising as up to 80% of an inhaled dose
may deposit in the oropharynx.14

Synopsis of key findings Clinical applicability of this study

Inhaled corticosteroids, when used for any length of time, Prior to this investigation, it has been assumed that phar-
may initiate pharyngeal and laryngeal complications. yngitis and cellular inflammatory infiltrate must be posi-
These include throat irritation and soreness, hoarseness tively correlated. Although we have shown that asthmatics
and voice weakness. Oral and oropharyngeal opportunis- using corticosteroid inhalers regularly had a significantly
tic infections, particularly Candida albicans, are frequently increased risk of hoarseness, voice weakness, soreness and
encountered.8 Other complications include perioral der-
matitis and tongue hypertrophy.11,12 We found that the
most troublesome functional complications were cough
during inhalation and dysphonia. Frequent complaints Table 2. Factors that may contribute towards inflammation of
the airways
were of a dry sensation in the throat, of the voice repeat-
edly breaking up, inability to raise the voice effectively Possible irritant Mechanism of irritation
and a lack of confidence in the speaking voice. Steroid Preparation
The implications of these unwanted side effects are Carrier substance
far reaching and often hamper compliance. Any patient Dose of steroid
presenting with the symptoms described, above all else, Dosing regimen
requires investigation to exclude an underlying neo- Propulsion Inhaler device
plasm. If no malignancy is found, the patient can be Propellant
reassured that the symptoms are attributable to the Intrinsic inflammation Irritable airways of
inhaled corticosteroid. asthmatics
Mechanical irritation Cough
Inter-current Rhinosinusitis
Suggested mechanisms of pharyngeal inflammation in inflammatory disease Post-nasal catarrh
asthmatics Inter-current Smoking
inflammatory stimuli Environmental pollutants
Little is actually known about the mechanism by which Noxious workplace agents
inhalers cause inflammation of the pharyngeal mucosa.
 2008 The Authors
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 581–586
586 R.K. Bhalla et al.
throat irritation compared with normals and seasonal
asthmatics, there was no relationship between pharyngeal
appearance and inflammatory infiltrate. Therefore, a clini-
cally red pharynx may have had very little in the way of
inflammatory infiltrate and conversely, one that appeared
relatively normal may in fact have had a substantial infil-
trate. The clinical inference was that a normal or near-
normal pharynx on inspection did not exclude an asth-
matic from having a substantial mucosal inflammatory
reaction, with the ensuing side effects already alluded to.
Conflict of interest
None to declare.
Declaration
Mr RK Bhalla wrote the paper. Dr W Taylor assessed the
histological specimens. Professor AS Jones provided sta-
tistical analysis. Mr NJ Roland supervised the study and
also provided editorial advice. Mr RK Bhalla accepts full
responsibility for the integrity of the article.
References
1 Inhaler devices for asthma. Drug Ther. Bull. 38, 9–14
2 Schleimer R.P. (1990) Effects of glucocorticoids on inflamma-
tory cells relevant to their therapeutic applications in asthma.
Am. Rev. Respir. Dis. 141, S59–S69
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 581–586
3 Taylor I.K. & Shaw R.J. (1993) The mechanism of action of cor-
ticosteroids in asthma. Respir. Med. 87, 261–277
4 Barnes P.J. (1993) Anti-inflammatory therapy for asthma. Annu.
Rev. Med. 44, 229–242
5 Babu S. & Samuel P. (1988) The effect of inhaled steroids on
the upper respiratory tract. J. Laryngol. Otol. 102, 592–594
6 Williamson I.J., Matusiewicz S.P., Brown P.H. et al. (1995) Fre-
quency of voice problems and cough in patients using pressur-
ised aerosol steroid preparations. Eur. Respir. J. 8, 590–592
7 Dubus J.C., Marguet C., Deschildre A. et al. (2001) Local side
effects of inhaled corticosteroids in asthmatic children: influence
of drug, dose, age and device. Allergy 56, 944–948
8 Hanania N.A., Chapman K.R. & Kesten S. (1995) Adverse effects
of inhaled corticosteroids. Am. J. Med. 98, 196–208
9 Barnes P.J., Pedersen S. & Busse W.W. (1998) Efficacy and
safety of inhaled corticosteroids. Am. J. Respir. Crit. Care Med.
157, S1–S53
10 DelGaudio J.M. (2002) Steroid inhaler laryngitis: dysphonia
caused by inhaled fluticasone therapy. Arch. Otolaryngol. Head
Neck Surg. 128, 677–681
11 Held E., Ottervanger V., Petersen S. et al. (1997) Perioral der-
matitis in children under steroid inhalation therapy. Ugeskr. Lae-
ger 157, 7002–7003
12 Linder N., Kuint J., German B. et al. (1995) Hypertrophy of the
tongue associated with inhaled steroid therapy in premature
infants. J. Paediatr. 127, 651–653
13 Selroos O., Backman R., Forsen K.O. et al. (1994) Local side
effects during 4-year treatment with inhaled corticosteroids –
a comparison between pressurised metered-dose inhalers and
Turbohaler. Allergy 49, 888–890
14 Lipworth B.J. (1995) New perspectives on inhaled drug delivery
and systemic bioactivity. Thorax 50, 105–110
 2008 The Authors