The Local Side Effects of Inhaled Corticosteroids

The frequent use of inhaled corticosteroids (ICSs), especially at higher doses, has been accompanied by concern about both systemic and local side effects. The systemic complications of ICSs have been extensively studied and are welldocumented in the literature. There are comparatively few studies reporting on the local complications of ICSs. Compared with systemic side effects, the local side effects of ICSs are considered to constitute infrequent and minor problems. However, while not usually serious, these local side effects are of clinical importance. They may hamper compliance with therapy and the symptoms produced may mimic more sinister pathology. This review considers the prevalence of local side effects, their clinical features, the potential causes, the role of inhaler devices, and current measures that have been suggested to avoid the problem. Key Words: asthma • corticosteroid • inhaler • laryngitis • local side effects • pharyngitis

Glucocorticoids are the most potent and reliable of the available agents among the antiinflammatory drugs, and have assumed a major role in the management of asthma.1 This has subsequently resulted in the widespread use of inhaled corticosteroids (ICSs), at higher doses and for longer periods of time.2 The mode of action of glucocorticoids in suppressing inflammation in asthma is not completely understood. It is thought that a cytoplasmic steroid receptor translocates to the cell nucleus and either induces increased synthesis of lipocortin-1, an inhibitor of phospholipase A2 and hence the arachidonic pathway of inflammatory mediator production, or reduces the transcription of cytokines such as interleukin-3 and interleukin-5, and granulocyte-macrophage colony-stimulating factor.34 The vasoconstrictor activity of corticosteroids aids to reduce bronchial mucosal edema and thickening, and ßadrenergic responsiveness is enhanced. The late response to allergen exposure is prevented by an inhibitory action on monocytes and granulocytes. The long-term administration of inhaled steroids also may reduce the number of mast cells in airway mucosa and decrease the immediate response to allergen and exercise.

The frequent use of ICSs, especially at higher doses, has been accompanied by concern about both systemic side effects and local side effects. The systemic complications of ICSs have been extensively studied and are well-documented in the literature.27 Systemic complications occur because up to 80% of the dose delivered by a conventional metered-dose inhaler (MDI) is swallowed.8 Systemic side effects are dose-dependent, and obvious differences exist between ICSs in their ability to cause systemic glucocorticoid activity. Problems include the following: calcium and phosphate metabolism with subsequent risk of osteoporosis910; adrenocortical suppression211; bruising and skin thinning1213; posterior subcapsular cataracts14; and glaucoma.15 It is surprising however, that there are comparatively few studies whose sole objective was to report the local complications of ICSs. It is well-known that hoarseness and pharyngeal discomfort are common problems in asthmatic patients using inhalers, especially among those using inhaled steroid preparations.1617 Compared with systemic side effects, the local side effects of ICSs are considered to constitute infrequent and minor problems. However, while not usually serious, these local side effects are of clinical importance. They may hamper compliance with therapy, and the symptoms produced may mimic more sinister pathology.

Reported prevalences vary significantly depending on the type of studies, length of observations, and methods for recording side effects (questionnaire or inspection). Many trials estimate these symptoms to occur in the region of 5 to 10% of the treated population, but no factual scientific evidence is given to corroborate these claims. Some studies quote figures of 55 to 58%,1618 but again these are largely symptom-based questionnaires. Dubus et al19 used both a questionnaire and a clinical examination in children. They found that > 60% of asthmatic children and infants treated with ICSs were affected by at least one local side effect in daily life. Furthermore, clear distinctions should be drawn between the clinical diagnosis of oropharyngeal candidiasis and positive cultures for Candida albicans.

Several studies have looked at local infective complications of steroid inhaler usage, such as oropharyngeal and laryngeal candidiasis2717 (Fig 1 ), and dysphonic problems without candidiasis manifested as bowing of the vocal folds on phonation.18 To date, however, there have been no studies to evaluate the cause or nature of the inflammatory response at a local level (ie, pharyngeal and laryngeal) produced by the action of ICSs (Fig 2 ). Little is known about the very action that causes the inflammation. Some have speculated that a "residue" from the inhaled substance irritates the pharyngolaryngeal mucosa. Indeed, both the propellant and lubricant components of MDI preparations will have a proinflammatory local effect. This may partly explain the difference in frequency of local side effects between steroid pressurized MDIs (pMDIs) and high-tomedium-high-resistance dry-powder inhalers (DPIs). Low-resistance steroid DPIs (eg, Rotahaler, Diskhaler, and Diskus/Accuhaler; GlaxoSmithKline; Research Triangle Park, NC) are associated with higher frequencies of local side effects because of the greater oropharyngeal deposition, compared with DPIs, with a higher inbuilt resistance. Lactose, as a component of lactose-based DPIs, also may have an irritating role on the pharyngolaryngeal mucosa. But why would an antiinflammatory steroid preparation cause inflammation in the upper airway? The problem is probably multifactorial, depending on the following factors: The steroid (eg, preparation, carrier substance, dose of steroid, and regime); The manner in which it is propelled into the airways (ie, the inhaler device); Intrinsic inflammation of the upper airway in asthmatic patients; Mechanical irritation because of cough; Intercurrent inflammatory disease (eg, rhinitis and postnasal catarrh); and Intercurrent inflammatory stimuli (eg, smoking and noxious agentsin the workplace).

There is a range of local side effects that includes perioral dermatitis,20 tongue hypertrophy,21 oral and oropharyngeal candidiasis,2717 pharyngeal inflammation, laryngeal disorders18 (Fig 3 ), cough during inhalation, and a sensation of thirst.

Pharyngeal disease tends to present with pain, irritation, orsoreness in the throat. The pain may be aggravated by swallowing (odynophagia), and, on occasions, patients may present with dysphagia. By far the most frequent local side effect is hoarseness of the voice (dysphonia) due to the action of the steroid inhaler on the larynx. Cough can be troublesome for this same reason.Cough during inspiration is usually only associated with pMDIs with or without a large-volume spacer. DPIs, with a larger proportion of fine particles, are practically devoid of this side effect (eg, reservoir inhalers, Turbohaler [AstraZeneca; Geneva, Switzerland], and Twisthaler [ScheringPlough; Kenilworth, NJ]). However, cough may still be seen as a side effect with DPIs containing large amounts of lactose (eg, capsule-based, Diskhaler, and Diskus/Accuhaler). Dubus et al19 have studied the local side effects of ICSs in asthmatic children, but most studies have concentrated on adults. Patients with pharyngeal and laryngeal disease form a significant part of the workload in ear-nose-throat surgery clinics. Such disease can be due to a number of different causes, including infection and neoplasia. The fact that the cardinal symptoms of throat cancer are persistent soreness in the throat and hoarseness of the voice underpins the importance of excluding other causes sooner rather than later. Dysphonia Dysphonia has been reported in 5 to 50% of patients using inhaled steroids.222 The wide range in this prevalence is a reflection of the means by which this data are calculated (ie, as a coincidental finding in many studies that have ultimately set out to investigate a different, although associated, problem). It is also interesting that many studies19 use the terms dysphonia and hoarseness as different phenomena when, in fact, the difference is very subtle. Furthermore, it is apparent that dysphonia (or hoarseness) usually has been assessed only by questionnaires rather than by any clinical measurement. A dose-dependent hoarseness has been reported in 34% of patients treated with beclomethasone dipropionate (BDP) or budesonide (BUD) when both ICSs were administered via pMDIs.16 Other studies23242526 have reported an increased risk of hoarseness with use of fluticasone propionate compared to BDP, and with pMDIs compared to DPIs. It has been suggested that the etiology of dysphonia in some cases is due to a steroid myopathy affecting the vocal cord muscles. Consequently, there is bilateral adductor vocal fold deformity with bowing of the folds on phonation.18 This is thought to be an extremely rare condition, but, in the authors’ opinion, a closer examination using flexible laryngoscopy and videostroboscopy reveals varying degrees of myopathy in symptomatic patients. This problem can, however, be reversed when therapy with the inhaled steroid is stopped. In contrast, Shaw and Edmunds27 found dysphonia not to be a problem when they examined 129 children in the age range 6 to 15 years using regular inhaled BDP, 100 to 1,500 µg per day, although no objective measure of dysphonia was used in this young age group. Similarly, Agertoft et al28 showed a comparable incidence of hoarseness in healthy control subjects and children receiving long-term BUD therapy via Turbuhaler. Oropharyngeal Candidiasis Oropharyngeal candidiasis has a reported incidence of 0 to 70% in ICS users.29303132 The disparity here is once again probably related to the differences in diagnostic criteria. For example, in the study by Dubus et al19 no cultures were made of Candida, but the diagnosis was based on clinical findings. This side effect may be due to a decreased local immunity or to an increase in salivary glucose levels, which stimulates C albicans growth.3233 In some cases, the patient is notably asymptomatic. More significantly, perhaps, is that a small proportion of patients being treated with BDP but without clinical evidence of candidiasis complained of sore throat and hoarseness.29 No explanation for this association was given in this report. The association between ICS use and oropharyngeal candidiasis is not clear, since there is no appreciable enhancement of C albicans pseudomycelial growth in vitro by BDP pure substance.29 In all cases, however, candidiasis is seen where the ICS was likely to have contacted the oral mucosa.3034353637383940414243 It has been suggested that it is perhaps a consequence of immunosuppression at the oral mucosal surface.2 With current instructions to rinse the mouth after inhalation, however, this is a much rarer complication and is usually only seen in patients with poor mouth hygiene or in conjunction with other diseases/therapies (eg, diabetes, immunosuppression, or concomitant oral corticosteroid treatment). Cough Cough is an inherent symptom of asthmatic disease and has been correlated with worse control.44 The occurrence of cough during inhalation has been observed in more than one third of the children treated with ICSs.19 It has been proposed4445 that this side effect occurs in concurrence with asthma, as a result of a toxic role of inhaled excipients (oleic acid) from pMDIs, and from nonspecific irritant effects of ICSs. Perioral Dermatitis Dubus et al19 found perioral dermatitis to be infrequent in children, and dependent on whether they used a spacer device equipped with facemask (5%), a nebulizer with facemask, or nebulizer without a mouthpiece (14%). Dermatitis around the mouth is thought to be due to a direct local effect of ICSs on the facial skin. There is evidence that BUD may have an effect on collagen synthesis in the skin.46 Held et al20 recommended the use of topical erythromycin or metronidazole in severe cases. Thirst A thirsty feeling after ICS delivery may be caused by throat irritation or as a manifestation of oral candidiasis. Dubus et al19 found that > 20% of children taking an ICS complained of thirst. The only risk factor was a combined treatment of ICSs with a long-acting ß2-agonist. Tongue Hypertrophy This is a rarely reported problem. Tongue hypertrophy has been described in infants treated with nebulized BDP for bronchopulmonary dysplasia21 and in asthmatic children treated with nebulized BUD.19 It is thought to be due to a direct effect of the ICS causing tongue muscle hypertrophy and local fat accumulation. The condition resolves after cessation of the steroid treatment.

Without a specific diagnosis, it is difficult to comment on the most appropriate therapy to treat these problems. No study has been performed to evaluate the value of any specific treatment. Most patients are advised to rinse their mouths and oropharynx by gargling with water immediately after using the inhaler. Provision of a spacer device is an attempt to minimize laryngeal and pharyngeal deposition of the inhaled material. In one study,24 this has been shown to be of some benefit, but, in contrast, another study19 found that cough was a spacer device-dependent side effect. Increased dose frequency is known to positively correlate with the incidence of local side effects. Toogood et al47 found that twice-daily regimens reduced the risk of dysphonia and candidiasis compared with administration four times per day. Once-daily use of BUD delivered via a Turbuhaler is practically free from local side effects in patients starting to receive this treatment.24 Previous treatment with other ICSs and devices resulting in local side effects may result in carryover effects.

An ideal inhaler device should deliver a predetermined dose of drug to the lungs, in an easy-to-use, reproducible, and cost-effective manner, with minimal deposition of drug in other sites.22 Both patient factors and the inhaler device itself can affect drug delivery. Age, physical disability, or cognitive disability may render a patient unable to use certain devices. Three main methods of dispersing medication into an aerosol will be described, as follows: pMDI; DPI; or nebulizer.22 A pMDI may be used with a spacer device. pMDIs In a pMDI, the drug is dissolved or suspended in a propellant under pressure, and, when activated, a valve system releases a metered dose of the drug and propellant. The propellant provides the force to propel and disaggregate particles. pMDIs may be manually actuated or breath-actuated. They can be used alone or in combination with various devices or adaptations (eg, spacers or extended mouthpieces) designed to slow the aerosol cloud, reduce oropharyngeal deposition, and promote ease of use.22 The inhaler must emit the drug in a particle size that can reach the lungs and deposit in the airways. Airway deposition is probably maximal with a particle size diameter of 1 to 3 µm. Most therapeutic aerosols are formulated to produce particles with a diameter of 1 to 5 µm. Particles with a diameter of 10 µm deposit mainly in the mouth and throat, or do not enter the upper airway due to abrupt changes in airflow or the cough reflex. The following two types of pMDIs are currently in use: manually actuated pMDIs; and breath-actuated MDIs. The former type is familiar to many patients as they have been available for 40 years. They are convenient to carry, quick to actuate, and generally inexpensive. For effective drug delivery, however, they require good coordination and psychomotor skills to ensure that actuation, inhalation, and breathholding occur in precise sequence. Common failings are not shaking the canister before use, inhaling too rapidly or "jerkily," or not holding the breath long enough at the end of inspiration. The "cold Freon effect," in which high-velocity aerosol hits the back of the throat, also causes patients to stop inhaling prematurely. Drug delivery varies from 7 to 20%, depending on the patient’s technique,4849 and, again, as much as 80% of the dose deposits in the oropharynx.50 pMDIs alone are therefore not suitable for physically or cognitively impaired adults, 5152 or for most children under the age of 12 years,53 and this is the reason why a spacer device (see below) is specifically recommended for patients in these two age groups. Breath-actuated MDIs are activated at an inhalatory flow rate of about 30 L/min. This reduces the need for coordination of actuation and inhalation, making the device easier to use for elderly, physically impaired patients. 51 The use of these devices should be reserved for adults and older children. There have been no efficacy trials assessing the delivery of corticosteroids using these devices. The devices are bulkier and less portable than conventional MDIs. The cold Freon effect is sometimes a problem, and oropharyngeal deposition of corticosteroids is high.8 All pMDIs contain chlorofluorocarbon propellants, but, because chlorofluorocarbons damage the ozone layer of the earth,54 their use will be phased out over the next 3 years or so. Currently, therefore, there is a transition period to non-ozone-depleting propellants such as hydrofluorocarbons.5556 PMDIs With Spacer Devices Spacer devices are used with pMDIs and are of the following two broad types: holding chambers; and extension devices. Holding chambers provide a reservoir of drug from which the patient breathes, and are easier for older, frailer patients52 and children53 to use. An extension device increases the distance that the aerosolized drug has to travel before it is inhaled. This has the effect of slowing the aerosol and allowing the propellant to evaporate. This reduces the size of the aerosol droplets and traps large (nonrespirable) particles within the spacer, thereby reducing oropharyngeal impaction of the drug. This has been shown to be of some benefit,24 since as much as 80% of an inhaled dose of drug can be deposited on the mucosa of the pharynx and larynx.50 Drugs should be administered as single actuations into the spacer and inhaled with minimum delay after each puff, repeating these actions until the entire prescribed dose has been given.57 The canister should be shaken between actuations. Spacers have been shown to decrease the oropharyngeal deposition of inhaled isotope-labeled aerosols and to increase their intrapulmonary deposition.5859 Hence, a large-volume spacer reduces the local unwanted effects of ICSs by decreasing oropharyngeal deposition: effects such as oropharyngeal candidiasis and a reduction in the amount of absorption from the alimentary tract.5057 They should also be of use in patients with dose-limiting oropharyngeal complications, and as a means of reducing drug costs by effectively delivering the same concentration of drug without increasing the number of puffs required per day for effective asthma control. 60 A large-volume spacer is recommended for administering ICSs via an MDI in children, 61 or for giving high doses in adults.62 Static charge accumulates on the walls of plastic and polycarbonate spacers, attracting drug particles and, hence, reducing the output of medication in vitro. 63 Washing the spacer in washing-up liquid, and allowing it to air dry, before its first use and after monthly intervals, reduces the static charge and increases the delivery of salbutamol to the lungs. 64 Dry Powder Inhalers DPIs do not require propellants but rely on the patient’s inspiratory effort to disperse the drug into small particles and deliver it to the lungs. An inspiratory flow rate of 30 L/min is needed to work the most efficient DPIs, and nearly all adults can achieve this (adult average, 60 L/min), even when wheezy.65 Oropharyngeal deposition in these is 60% of the delivered dose, and patients should still be advised to rinse their mouth with water after inhaling from a DPI to minimize local side effects.2450 The possible protective effect of DPIs might be related to the position of the vocal cords, which are open during inhalation against resistance. Inhaler technique needs to be monitored, and difficulties need to be identified and corrected early. Acceptability of the prescribed device is likely to influence adherence to treatment.22 Common problems include the inability to coordinate actuation and inspiration precisely enough to use a pMDI, or the inability to inhale forcefully enough when wheezy to use a DPI.

In order to establish more effective treatment of local side effects, they first need to be recognized. Local side effects are mostly minor and are of little consequence to the overall health of the patient. However, patients tend to be afraid of the side effects of ICSs, and local problems may have a deleterious effect on compliance. Some of the local side effects may be dose-dependent, and this emphasizes the need to find the lowest effective dose of an ICS. More importantly, however, the literature shows that many of the local side effects are device-dependent, and a change of the inhalation device should be considered. Precautions such as rinsing the mouth, gargling, and washing the face after inhalation also should be recommended.

Last, it is evident from this review that more observational and randomized controlled trials are necessary to examine the existing inhalers, and, specifically, to investigate how and why they cause local side effects. A prospective strobovideolaryngoscopic study to visualize and record on video vocal cord fold motion, position, and mucosal waves during ICS therapy would be of great value and interest. A relevant study would include a series in which strobovideolaryngoscopic documentation was performed pretherapy as a baseline and then repeated in every ICS patient who developed hoarseness. Successful recruitment and analysis would, however, depend on cooperation and collaboration among pulmonologists, voice care specialists, and other laryngologists in the care of asthmatic patients requiring ICS therapy.

Abbreviations: BDP = beclomethasone dipropionate; BUD = budesonide; DPI = dry powder inhaler; ICS = inhaled corticosteroid; MDI = metered-dose inhaler; pMDI = pressurized metered-dose inhaler N.J. Roland and R.K. Bhalla wrote the paper. R.K. Bhalla was investigating the clinical impact of this problem at the time of writing. J. Earis was supervising R.K. Bhalla in aspects of his research. N.J. Roland accepts full responsibility for the integrity of the article. Received for publication May 14, 2002. Accepted for publication October 2, 2003.