Objective. In an observational long-term study, we followed 62 children (37 males, 25 females; mean age: 11.6 ± 2.9 years) with moderate-to-severe asthma for 2 years and studied the effects of fluticasone propionate (176–1320 μg/day) on the function of the hypothalamic-pituitary-adrenal axis.
Study Design. Morning cortisol levels were monitored after patients had been on fluticasone for a mean of 8.0 ± 5.2 months. Patients who had abnormal low morning cortisol levels (<5.5 μg/dL) were then switched either to lower fluticasone dosage or to other inhaled steroid formulation. Exact methods based on the binomial distribution were used to construct a 95% confidence interval for the true proportion of abnormal readings among those treated, and the Wilcoxon signed rank test was used to test for a significant difference between cortisol levels taken before and after the switch.
Results. Twenty-two patients (36%) had abnormal morning cortisol levels while on fluticasone. Of the patients on a low dose (176 μg/day), 17% had abnormal values, whereas 43% of patients on a high dose (≥880 μg/day) were abnormal. Patients with abnormal results (17/22) had their morning cortisol levels repeated 3 months after the switch. Thirteen of these patients (77%) had normal levels. A stratified analysis of the difference in morning cortisol levels before and after the switch showed significant increase in morning cortisol levels in the group receiving 440 μg/day or less of fluticasone (median difference: 5.25; confidence interval: 3.60–8.15), as well as in the group receiving 440 μg/day or more (median difference: 3.85; confidence interval: 1.00–7.60).
Conclusion. Inhaled fluticasone, even at conventional doses, may have greater effects on the adrenal function than previously recognized, but the clinical significance of this suppression still remains to be established.
Inhaled glucocorticoids have been typically considered to be free of side effects on the hypopituitary adrenal (HPA) axis function except in relatively large doses. Fluticasone propionate (FP) is a potent inhaled glucocorticoid reported to have minimal oral bioavailability. The initial clinical trials assessing the effect of inhaled FP on the HPA axis did not demonstrate any adverse effects in adult asthmatics at a daily dose of 500 μg or less.1 Long-term use of FP in adults has also been shown to have no discernible effects on the HPA axis, even when doses up to 1000 μg daily are given for 2 years.2 Long-term effects on the HPA axis in asthmatic children treated with FP have not yet been established, but short-term studies have found FP via Diskus (GlaxoSmithKline, Middlesex, UK; dry powder inhaler [DPI]) to be safe and effective.3–8 FP via Diskhaler (GlaxoSmithKline), another DPI device, has been approved by the Food and Drug Administration for children 4 years and above. Inhaled FP via the Diskhaler device up to 400 μg/day for 8 weeks has been shown to have minimal effects on the HPA axis in children with asthma.5 However, FP via the Diskhaler is less commonly used in young children than the metered dose inhaler (pMDI) with a spacer device, and FP via Diskus has not yet been approved in the United States. Despite lack of Food and Drug Administration approval for FP via pMDI in children <12 years of age, and despite the paucity of published data assessing the effects of FP via pMDI on the HPA axis in this age group,3–10 the use of FP has become a common practice among physicians caring for children with asthma. This is primarily attributable to the extrapolation of safety data from studies using FP via DPI, despite the fact that drug delivery to the lung varies greatly between the 2 devices, with the FP pMDI delivering up to twice the amount to the lungs than the FP via DPI.11 We evaluated the effects on the HPA function in children with moderate-to-severe asthma treated chronically with FP via the pMDI used in conjunction with a spacer device.
Patients’ Selection and Characteristics
Children with moderate-to-severe persistent asthma treated with inhaled FP were included in the study. The inclusion criteria were clinical symptoms and lung function suggestive of moderate-to-severe persistent asthma as defined by the National Heart, Lung, and Blood Institute. Patients with a history of cardiac, hepatic, renal, or other medical diseases were excluded. Informed parental consent was obtained for all procedures and medications. All patients were followed at the Childhood Asthma Care and Education Center, Kosair Children’s Hospital, for difficult to control asthma. All patients received inhaled FP (GlaxoSmithKline, Research Triangle Park, NC) at doses required to both control symptoms and to improve lung function. Clinical evaluation and lung function were assessed at 3-month intervals, and the dosage of inhaled steroids was adjusted as deemed necessary for control by the clinician.
Throughout the study, all patients used their pMDI with a spacer device with inspiratory flow signal (Aerochamber, Monagham Medical Corporation, Pattsburgh, NY). All patients were given education by a respiratory therapist before the start of the study on the technique to use the spacer device, and the technique was reinforced every 3 months. Standard spacer technique included slow inhalation with 10-seconds breath-hold before exhalation and 4 to 6 regular tidal breaths per activation. All patients understood and adequately reproduced the technique. None of the patients developed hoarseness or oral candidiasis during the study. Patients were instructed to wash and clean their spacer once a week.
HPA Axis Assessment
Patients were assessed for HPA axis function when their dose of inhaled steroid had been constant for at least 3 months. All cortisol levels were measured between 7 am and 9 am. A cortisol level was considered abnormal if the morning cortisol level was <5.5 μg/dL (<140 nmol/L). Patients with an abnormal morning cortisol level were switched to half the dose of their current inhaled steroid or to another inhaled steroid formulation while maintaining asthma control. The decision to either reduce or switch steroids was based on the prescription formulary of each patient’ insurance. A follow-up morning cortisol was drawn after at least 3 months on the new dose or treatment. The incidence of abnormal HPA axis function requiring a treatment change was assessed and related to the initial FP dose. Institutional review board approval for chart review was obtained before the study.
Exact methods based on the binomial distribution were used to construct a 95% confidence interval for the true proportion of abnormal readings among those identified.12 The Wilcoxon signed rank test was used to test for a significant difference between cortisol levels taken before and after switching to a lower dose of FP or to another inhaled steroid formulation, and the Hodges-Lehman method was used to find an exact 95% confidence interval for the true median difference in cortisol levels before and after the switch. Results for continuous variables were summarized as mean ± standard deviation.
Sixty-two children (37 males, 25 females) with a mean age of 11.6 ± 2.9 years were included for review for whom all baseline and follow-up data on FP dose and HPA axis assessments were available. Of these 62 children, 49 (79%) were white and 12 (21%) were black. Data on ethnic origin were missing for 1 participant. Fifty-two of these children (84%) had previously used inhaled steroids, with a mean duration of 25.7 ± 18.4 months. Treatment before FP was primarily flunisolide, triamcinolone acetonide, and beclomethasone dipropionate, with a few children having used cromolyn sodium or nedocromil before being placed on FP. Morning cortisol levels were first measured after a mean duration of 8.0 ± 5.2 months of treatment with FP as the sole inhaled steroid. Oral steroids had not been used by any child for at least 5 months at the time of the first adrenal assessment.
Twenty-two patients (36%) had abnormal initial cortisol levels while on FP. Seventeen percent of patients on the lowest dose (176 μg/day) had abnormal values, whereas 43% of patients on the highest doses (≥880 μg/day) were abnormal (Table 1). Although there was a wide range of responses (Fig 1), the incidence of adrenal suppression by FP was significantly dose-related.
Table 2 shows the results of the stratified analysis of the differences in cortisol levels before and after the switch was conducted as described above in 19 of the 22 patients with abnormal initial cortisol levels. In the group receiving 440 μg/day or less, there was a statistically significant increase in cortisol levels after the switch (P = .001). There was also a significant increase in the group receiving >440 μg/day (P = .006). When all patients were combined, there was a statistically significant increase in the morning cortisol level after reducing the FP dose or switching to another inhaled steroid (P < .001).
Figure 2 shows the individual data for change in morning cortisol levels on repeat testing. Of the 22 patients with an initial abnormal morning cortisol, 17 patients had a second morning cortisol level drawn after the treatment was changed. Thirteen (77%) of 17 patients had a normal morning cortisol level on the second test. Six of these patients were placed on lower FP doses, and 6 were switched to another steroid formulation (3 budesonide DPI, 2 beclomethasone pMDI, and 1 flunisolide pMDI). All 4 patients with an abnormal morning cortisol level on repeat testing were using 440 μg/day of inhaled FP at time of second assessment.
Because of its simplicity and safety, early morning serum cortisol level has been accepted as a reliable test for HPA axis function and has been adopted in large multicenter, clinical studies.13 Because of the diurnal nature of cortisol secretion, production is maximal between 4:00 am and 6:00 am. Basal serum levels obtained at 8:00 am are adequate in most cases to screen for adrenal dysfunction. Levels <5 μg/dL, (<140 nmol/L) give a presumptive diagnosis of adrenal insufficiency, and levels <10 μg/dL, (<275 nmol/L) are strongly suggestive. Levels above 300 nmol/L almost exclude insufficiency. Therefore, basal morning cortisol levels can be used as the first test in the evaluation of HPA insufficiency and in most will obviate more sophisticated and expensive testing. In a subset of cases, a patient may exhibit normal basal levels, but an inadequate response to stress, which would not be detectable without adrenocorticotropic hormone (ACTH) stimulation testing.14–16 Cortisol secretion shows a distinct diurnal variation after the age of 6 months. With the usual sleep-awake cycle, hormone levels rise after midnight to reach peak levels at approximately 6 am. With an intact hypothalamic-pituitary-adrenal axis, levels at 8 am should fall between 5 and 25 μg g/dL. In addition, cortisol secretion increases rapidly under conditions of stress, including phlebotomy. This acute response occurs within minutes, therefore, the finding of depressed morning cortisol levels is important from a screening perspective and should alert clinicians to the possibility of secondary adrenal dysfunction.
The effects of FP on the HPA axis in children has been primarily limited to studies assessing low-to-medium doses (200–400 μg daily) delivered via the Diskus or Diskhaler, 2 somewhat different DPI devices.3–8 These data have been reassuring in that minimal adrenal suppression has been shown at these doses. However, data in older patients suggest that lung deposition via the FP pMDI is greater (up to twice the percent nominal dose) than lung deposition via the DPIs.11 A recent study by Edsbacker et al11 showed that the same nominal dose of FP produced greater systemic exposure and significantly more adrenal suppression when given by pMDI rather than Diskus either as a single or in repeated doses in normal adult patients.
A recent study by Bisgaard et al17 evaluated symptoms’ improvement in young children using FP via pMDI in conjunction with the Babyhaler spacer device (GlaxoSmithKline), but did not include HPA axis assessment.17 However, 2 studies have assessed the dose-response relationship of FP via pMDI used with a spacer device on adrenal suppression in asthmatic children.9,10 Clark et al9 assessed the effect of single doses of FP ranging from 400 to 1250 μg delivered via Volumatic (GlaxoSmithKline, Middlesex, UK) or Nebuhaler spacer devices (AstraZeneca, Södertälje, Sweden). Overnight urinary cortisol/creatinine excretion showed suppression with all doses of FP compared with placebo, and the degree of suppression was significantly dose-related. Lipworth et al10 assessed FP at 200 and 400 μg per day via the large Volumatic spacer for 4 days in asthmatic children. This study showed no significant suppression of overnight urinary cortisol/creatinine excretion at these low-to-moderate doses.
Recently, individual case reports documenting HPA axis inhibition with the use of FP have been described. Zimmerman et al18 reported 2 patients who developed adrenal suppression on low doses of FP; 1 of these 2 patients was an 8-year-old girl who was taking 250 μg/day of FP. Todd et al19 reported acute adrenal insufficiency in a 6-year-old girl with asthma treated with a prolonged course of 1000 μg/day of FP. The patient exhibited all cardinal symptoms of adrenal insufficiency, mainly continuous nausea, vomiting, severe fatigue, and abdominal pain a few days after her FP was abruptly switched to budesonide.19 The authors postulated that the systemic effects of 800 μg of budesonide delivered by the Turbohaler device (AstraZeneca) was less than that of FP 1000 μg/day delivered via pMDI so that an acute adrenal crisis was precipitated by the switch. The same authors had previously reported adrenal suppression in 6 children receiving high doses of FP (>1000 μg/day) via Diskhaler.20 It should be pointed out, however, that in the last 2 reports, toxicity occurred at much higher doses than usually recommended in this age group. More recently, Taylor et al21 described iatrogenic adrenal suppression and exogenous glucocorticoid excess in a 9-year-old child receiving 550 μg of FP daily for 6 months. The patient had evidence of Cushing syndrome, which required a long taper of inhaled FP to resolve.21
In our study, a population of children with moderate-to-severe asthma using inhaled FP as a controller medication was identified as having a dose-dependent decrease of morning cortisol levels. We also observed a much larger proportion with low morning cortisol levels than previously recognized, even in children treated with relatively low doses of FP. This effect on the HPA axis was normalized in 75% of patients when FP was either lowered or switched to another inhaled steroid preparation. Our study is the first to document low morning cortisol levels in children using a wide dosing range of FP via pMDI plus the Aerochamber in an actual clinical situation. Despite the large extent of the HPA axis dysfunction as assessed by the morning serum cortisol levels, none of the children developed clinical correlates of adrenal suppression. Therefore, the clinical relevance of this biochemical disturbance remains primarily unknown. However, as suggested by Taylor et al, this phenomenon may portend decreased adrenal response to stress and infections, and should be closely monitored.21 Two study limitations need to be recognized: first, this was an observational, uncontrolled study; and second, patients were taking other inhaled steroids formulation before FP therapy; however, the changes observed after the switch would most likely implicate FP therapy as a causal agent of the observed effects. In addition, and despite the fact that inhaled steroids doses were adjusted as deemed necessary for asthma control by the clinician based on clinical evaluation and lung function, it is possible that some children may have received doses above the minimal required doses.
The marked influence of inhaled fluticasone on the function of the HPA axis seems to be mediated by the enhanced absorption of the drug via the respiratory tract and the enhanced binding of fluticasone to the glucocorticoid receptors,20–25 as well as perhaps to its prolonged plasma half-life.23 Ironically, FP has been considered safer than other inhaled glucocorticoids because of its complete first-pass hepatic inactivation; however, the drug is readily absorbed in the lung.24 And similar to other potent inhaled steroids with high degree of lipophilicity, it is the lung rather than the gut that determines the final systemic bioavailability of the drug.25 Therefore, one might expect that lung absorption of a corticosteroid with enhanced potency would produce greater systemic activity on the steep part of the dose-response curve. Our results suggest that a dose-response relationship exists for adverse effects on the HPA axis in children treated chronically with FP. Additional studies are urgently needed to identify those most at risk for this phenomenon.
It is important to note, however, that although low early morning levels of cortisol are suspicious for depressed adrenal responsiveness, especially in the context of a physical stressor (ie, phlebotomy) and exogenous glucocorticoid use. Other tests for adrenal axis function such as a simultaneous ACTH level, ACTH stimulation test, or 24-hours collection of urinary cortisol should be performed to confirm HPA inhibition. Although these data are not available because of the nature of our observational study, a prospective study using ACTH levels, stimulation testing, or urinary cortisol secretion would be beneficial to substantiate our current findings.
- Received April 19, 2001.
- Accepted September 21, 2001.
- Reprint requests to (N.E.) Department of Pediatrics, 571 South Floyd St, Suite 414, Louisville, KY 40202. E-mail:
- ↵Hoekx JCM, Hedlin G, Pedersen W, Sorva R, Hollingworth K, Efthimiou J. Fluticasone propionate compared with budesonide: a double-blind trial in asthmatic children using powder devices at a dosage of 400 μg.day. Eur Respir J.1996;9 :2263– 2272
- ↵Clark DJ, Lipworth BJ. Adrenal suppression with inhaled budesonide and fluticasone propionate given by large volume spacer to asthmatic children. Thorax.1996;51 :941– 943
- ↵Lipworth BJ, Clark DJ, McFarlane LC. Adrenocortical activity with repeated twice daily dosing of fluticasone propionate and budesonide given via a large volume spacer to asthmatic school children. Thorax.1997;52 :686– 689
- ↵Edsbäcker S, Källén A. Differences in bioavailability of fluticasone propionate via Diskus and pMDI. Am J Respir Crit Care Med.1999;159 :A118
- ↵Conover WJ. Practical Nonparametric Statistics. 3rd ed.1999:131– 132
- ↵Lipworth BJ, Jackson CM. Safety of inhaled and intranasal corticosteroids, lessons for the new millennium. Drug Safety.2000;1 :11– 33
- Brattsand R, Axelsson B. New inhaled glucocorticoids. In: Barnes PJ, ed. New Drugs for Asthma. Vol. 2. London, United Kingdom: IBC Technical Services Ltd;1992:192– 207
- ↵Harding SM. The human pharmacology of fluticasone dipropionate. Respir Med.1990;84(suppl) :25– 29
- ↵Lipworth BJ. Pharmacokinetics of inhaled drugs. Br J Clin Pharmacology.1996;42 :697– 705
- Copyright © 2002 by the American Academy of Pediatrics