- Email: [email protected]
Impact of inhaled corticosteroids on cortisol suppression in adults with asthma: a quantitative review
Impact of inhaled corticosteroids on cortisol suppression in adults with asthma: a quantitative review John H. Wlodarczyk, PhD*; Peter G. Gibson, FRACP†; and Manfred Caeser, PhD‡
Background: Studies examining the effects of inhaled corticosteroids (ICSs) on cortisol suppression show inconsistent results, and there is uncertainty regarding the dose-response relationship between ICSs and cortisol suppression. Objective: To determine, using meta-analysis, the extent of cortisol suppression after administration of clinically relevant ICS doses in adults with asthma. Methods: Database searches (MEDLINE, EMBASE, and The Cochrane Library) using appropriate indexed terms were performed to identify eligible articles for review. Articles reporting the effects of ICSs on cortisol levels in asthmatic adults, measured using the cumulative serum or plasma cortisol, morning serum or plasma cortisol, or cumulative overnight urinary free cortisol method, were identified. All available cortisol measurements were extracted. Cortisol suppression was estimated, and treatment arms were grouped into low-, medium-, and high-dose ranges according to the Global Initiative for Asthma guidelines. A multivariate model was used to determine relationships between ICS dose and cortisol suppression and to explore sources of heterogeneity among trials. Results: Thirty-one studies providing information on 216 measures of cortisol suppression were included in this meta-analysis. Cortisol suppression in the low-, medium-, and high-dose groups were estimated to be 17.92% (95% confidence interval [CI], 11.08%–24.77%), 26.55% (95% CI, 17.29%–35.80%), and 36.31% (95% CI, 26.48%– 46.13%), respectively. Conclusions: Statistically significant cortisol suppression was evident at low doses of ICSs and increased with dose. These results support an impact of all ICSs on endogenous cortisol levels and underscore the importance of titrating ICS doses to the minimum required to maintain symptom control. Ann Allergy Asthma Immunol. 2008;100:23–30.
INTRODUCTION Inhaled corticosteroids (ICSs) have been established as cornerstone therapy for chronic asthma, with treatment guidelines1– 4 recommending ICS treatment for patients with persistent asthma of all severities. Although the potential for systemic effects is far lower than with a comparable dose of oral corticosteroids (OCSs), the use of ICSs can be associated with dose-related systemic adverse effects. These effects include hypothalamic-pituitary-adrenal (HPA) axis suppression, decreased bone density, growth suppression, cataracts, and skin changes.5 Clinical concerns regarding the untoward effects of ICSs on the HPA axis have been heightened by recent case reports associating the use of ICSs with acute adrenal crisis in children6,7 and with the risk of adrenal insufficiency in adults.8 Despite the heightened awareness of the potential for untoward HPA axis effects, the clinical relevance of HPA axis suppression resulting from ICS use is controversial. Central Affiliations: * John Wlodarczyk Consulting Services, New Lambton, New South Wales, Australia; † Respiratory and Sleep Medicine, John Hunter Hospital, Newcastle, New South Wales, Australia; ‡ Altana Pharma AG, Konstanz, Germany. Disclosures: Authors have nothing to disclose. Funding Sources: This study was sponsored by Altana Pharma. Received for publication September 26, 2006; Received in revised form July 16, 2007; Accepted for publication July 23, 2007.
VOLUME 100, JANUARY, 2008
to this controversy is the inconsistent reporting of the presence and extent of HPA axis suppression after ICS administration. As a result, the impact of perturbations to the HPA axis during the months and years that these agents are typically prescribed is difficult to estimate. The effects of ICSs on HPA suppression have been measured directly in individual studies and indirectly in systemic reviews and meta-analyses. Individual studies9 –16 have shown inconsistent effects of ICSs on cortisol suppression, a sensitive marker of HPA axis function. The inconsistencies seen across studies may be due to factors such as differences in dose, formulation, frequency of administration, treatment duration, previous systemic corticosteroid therapy, method of cortisol measurement used, and patient population studied.17 Systematic reviews and meta-analyses have been used in an effort to overcome the limitations associated with the individual studies.5,18 Chrousos and Harris18 performed a systematic review on the results of 50 studies and concluded that ICSs, when administered in recommended dose ranges, did not endanger HPA axis functioning. However, a caveat that the authors noted was that some patients exhibited measurable attenuations in HPA axis function, even at low doses of ICS (fluticasone propionate, 200 – 400 g/d).18 In a metaanalysis of 27 studies, Lipworth5 concluded that cortisol suppression was evident after high-dose ICS administration. The author highlighted that fluticasone propionate was asso-
23
ciated with greater dose-related suppression vs other ICSs; effects were most apparent at fluticasone propionate doses greater than 800 g/d. Although both studies examined ICS effects on cortisol suppression, many factors limit the interpretation of these results. The study by Chrousos and Harris was a qualitative study and did not provide quantification of the effects of ICSs on cortisol suppression. The Lipworth study included studies examining different patient populations (controls and adult and pediatric patients) and focused on detecting large changes in HPA axis function as opposed to the more subtle effects of commonly used doses. In addition, neither study used the therapeutic equivalence of the ICSs to combine information from different drugs. Although there is general agreement that ICSs affect HPA axis function, studies detailing the extent of this response have been inconclusive. Careful quantification of the effect of ICSs on cortisol suppression, therefore, represents an important step in elucidating the clinical relevance of HPA suppression after ICS administration. The aim of this study was to determine, using meta-analysis, the extent of cortisol suppression after administration of clinically relevant ICS doses in adults with asthma. To combine data extracted from studies that used different ICSs and formulations, we used the classification system published in the Global Initiative for Asthma (GINA) guidelines.2 This classification system facilitated the investigation of ICS effects on cortisol suppression by grouping ICSs with different potencies and formulations into approximately equipotent dose categories (low, medium, and high doses). METHODS Search Strategy Database searches (MEDLINE, EMBASE, and The Cochrane Library) for articles published between January 1, 1966, and March 31, 2005, were performed to identify eligible articles for review. All searches included the terms inhaled corticosteroid and asthma and 1 of the following terms: adrenal function, adrenal suppression, or cortisol. The only limits used on the searches were for human subjects. Studies were included in the meta-analysis if they met the following criteria: used 1 or more of the ICSs outlined in the GINA guidelines (beclomethasone, budesonide, flunisolide,
fluticasone propionate, triamcinolone acetonide, and mometasone furoate); defined the number of patients in each study arm; included only adults (ⱖ18 years or older); provided defined inclusion and exclusion criteria for study participants (studies in nonasthmatic patients were excluded); defined the treatment protocol (ie, medications used); defined the ICS dose for each treatment group; included a minimum 3-day washout period from previous ICS use, or at least a 1-month washout from OCSs; presented primary results (ie, no review articles and meta-analyses were included); and included 1 or more of the following outcomes: cumulative overnight urinary free cortisol (UFC), cumulative serum or plasma cortisol, and morning serum or plasma cortisol levels. There were no minimum treatment period requirements. Case reports, small-scale trials (⬍5 patients), letters to the editor, reviews, trials on topical corticosteroids using nebulizers, trials on children or healthy patients, and trials with insufficient information to estimate a drug effect were excluded from the meta-analysis. Each article was reviewed by 2 independent reviewers. Any discrepancies in the reason for inclusion or exclusion or in data abstraction were discussed, and disagreement regarding eligibility was resolved by consensus. The references of all the articles that were evaluated for inclusion were also appraised for potential inclusion in the meta-analysis. Statistical Methods Articles meeting the eligibility criteria were abstracted into data tables by 2 independent reviewers. Each dose of ICS was classified according to GINA guidelines, where daily doses of ICSs are defined as low, medium, and high (Table 1).2 All available cortisol measurements were extracted. The primary outcome measure used in this study was cortisol suppression expressed as the percentage change in cortisol from baseline, which was calculated as follows: (end point cortisol ⫺ baseline cortisol)/baseline cortisol. In some studies, only end point cortisol measurements were reported, and, hence, the percentage change from baseline could not be calculated for each arm. If the studies included a placebo arm, an estimate of cortisol suppression was calculated using the difference between end point measurements for active treatment compared with placebo. Several studies also measured
Table 1. GINA Inhaled Corticosteroid Dose Groups GINA dose group, g/d
Drug Fluticasone propionate Budesonide dry powder inhaler Triamcinolone acetonide Flunisolide Beclomethasone dipropionate chlorofluorocarbon Beclomethasone dipropionate hydrofluoroalkane Mometasone furoate
Low
Medium
High
100–250 200–600 400–1,000 500–1,000 200–500 100–250 200–400
⬎250–500 ⬎600–1,000 ⬎1,000–2,000 ⬎1,000–2,000 ⬎500–1,000 ⬎250–500 ⬎400–800
⬎500 ⬎1,000 ⬎2,000 ⬎2,000 ⬎1,000 ⬎500 ⬎800
Abbreviation: GINA, Global Initiative for Asthma.
24
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
cortisol levels in the same individuals at different dose levels. The percentage of cortisol suppression for each dose level was included in the analysis as a separate observation. Other variables extracted included trial duration (in days), trial design (parallel group vs crossover), and previous OCS use. The SAS (SAS version 9.1; SAS Institute Inc, Cary, North Carolina) was used to perform statistical analyses. A generalized estimating equation model with an exchangeable correlation structure was used to determine the relationships between ICS dose groups and cortisol suppression after adjusting for multiple measurements in some of the studies. Weighting was by the number of patients in each treatment arm. All available measures of cortisol were retained in the model, and indicator variables were used to evaluate the impact of different measurement methods (ie, cumulative overnight UFC, cumulative serum or plasma cortisol, and morning serum or plasma cortisol levels). The heterogeneity of cortisol suppression was explored using bubble plots. Exploratory modeling showed that trial duration, design, and previous OCS use did not significantly affect cortisol
suppression. The effect of previous ICS washout duration on cortisol suppression was modeled, and there was no significant difference on cortisol suppression if those studies with washout periods shorter than 7 days were eliminated or retained in the analysis. Modeling the individual measurement methods produced results consistent with the combined model. A reduced model containing only those variables that were significant at the P ⬍ .05 level is presented. RESULTS We reviewed 385 articles from MEDLINE, 90 from EMBASE, and 198 from The Cochrane Library. The overall search strategy, the number of articles retrieved, and the reasons for article exclusion are depicted in Figure 1. A total of 28 articles were retained11,13,19 – 44; one article34 reported on 2 studies and another article42 reported on 3 studies. Thus, 31 studies were included in the meta-analysis. Twenty-four of the included studies were randomized. Of these, 14 were double masked, 6 were single masked, and 4 did not specify the type of masking. Eleven of the random-
Figure 1. Search strategy, including the number of articles retrieved and the reasons for inclusion or exclusion. A total of 28 articles were retained; one article reported on 2 studies and another article reported on 3 studies. Thus, 31 studies were included in the meta-analysis.
VOLUME 100, JANUARY, 2008
25
ized studies used a parallel group design, 9 used a crossover design, and 4 were open label. Of the 7 studies that did not specify randomization, 3 were double masked, 1 was single masked, and 3 did not specify the type of masking; 1 used a parallel group design, 5 used a crossover design, and 1 was open label. Most included studies were conducted in patients with mild to moderate asthma based on a percentage of the predicted forced expiratory volume or peak expiratory function and on patient-defined symptoms. Only 1 study30 reported on patients with severe asthma, and 2 studies33,36 reported on patients with moderate to severe asthma. All included studies had a minimum 1-month washout period from previous OCS use, with most studies requiring a 6-month washout period. Several of the included studies had multiple treatment arms and used more than 1 method to measure cortisol levels. In studies using multiple methods to measure cortisol levels,22,23,30,37,39,42,44 no qualitative differences in the direction of drug effect on cortisol suppression among the various modes of measurement were observed. Overall, 216 estimates of cortisol suppression were extracted from the 31 studies and were included in the analysis. The greatest numbers of individual cortisol measurements were obtained from fluticasone propionate and budesonide studies (Table 2). There was considerable heterogeneity between studies and study arms. Cortisol suppression varied from ⫺18.8% to 91.5%, with an interquartile range of 7.2% to 30.75%. The interquartile ranges for the low, medium, and high GINA dose groups were 5.0% to 14.3%, 12.9% to 35.7%, and 24.3% to 47.1%, respectively. Allocation of the cortisol measurements into the appropriate GINA dose groups revealed that the number of cortisol measurements across the 3 dose ranges were similar (Table 3). Likewise, the types of cortisol measurements were evenly distributed across the dose ranges except for morning serum or plasma cortisol levels, where the largest proportion of measurements (41%) was in the high-dose group.
Statistically significant cortisol suppression after ICS treatment was evident in all 3 GINA dose groups; the extent of suppression was dose dependent and increased as the dose increased (P ⬍ .001). Based on a combined model (in which all measures of cortisol were correlated with dose groups), suppression in the low, medium, and high GINA dose groups relative to baseline was estimated to be 17.92% (95% confidence interval; [CI], 11.08%–24.77%), 26.55% (95% CI, 17.29%–35.80%), and 36.31% (95% CI, 26.48%– 46.13%), respectively. Modeling the individual measurement methods produced similar results as the combined model and provided additional information regarding the measurement methods (Fig 2). Cumulative serum or plasma cortisol and cumulative overnight UFC measures were more sensitive to cortisol changes, although there was greater variance associated with these measures. Compared with cumulative measures (combined cumulative serum or plasma cortisol and cumulative overnight UFC), estimates of cortisol suppression using morning serum or plasma cortisol levels indicated, on average, 10.91% less suppression (95% CI, 4.30 –17.51). DISCUSSION This meta-analysis provides clear evidence regarding the extent of cortisol suppression after ICS administration. We demonstrated that low doses of commonly used ICSs produce statistically significant effects on cortisol suppression in adults with asthma. A statistically significant dose-response relationship was evident, with the extent of cortisol suppression becoming more marked at higher ICS doses. By showing that cortisol suppression was evident, even at low ICS doses, these findings represent an important step in elucidating the clinical relevance of HPA axis suppression after ICS administration. The dose-response estimates presented in this study provide a basis on which to calculate the average systemic impact of currently available ICSs in adults with asthma. Furthermore, the impact of low-dose ICS use on cortisol suppression as reported herein reinforces the need for titration
Table 2. Cortisol Measurements Available for Each Study Drug Cumulative Overnight UFC
Drug
Fluticasone propionate Budesonide Triamcinolone acetonide Flunisolide Beclomethasone dipropionate Mometasone furoate Total
Serum or plasma cortisol
Morning serum or plasma cortisol
Treatment arms, No.a
Studies (patients), No.
Treatment arms, No.a
Studies (patients), No.
Treatment arms, No.a
Studies (patients), No.
30 18 16 9 7 3 83
8 (528) 6 (318) 5 (274) 3 (152) 2 (123) 1 (63)
25 6 8 7 7 0 53
8 (521) 2 (142) 3 (173) 3 (154) 2 (135) 0
30 21 12 2 4 11 80
14 (608) 8 (367) 4 (190) 2 (49) 2 (99) 4 (175)
Abbreviation: UFC, urinary free cortisol. Estimates of percentage change from baseline in cortisol levels might include several arms from a single study.
a
26
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
Table 3. Cortisol Measurements Available for Each GINA Dose Group Measurements, No. GINA dose groupa
Low Medium High Total
Cumulative overnight UFC
Cumulative serum or plasma cortisol
Morning serum or plasma cortisol
Total
27 27 29 83
21 15 17 53
24 23 33 80
72 65 79 216
Abbreviations: GINA, Global Initiative for Asthma; UFC, urinary free cortisol. Fluticasone propionate– equivalent doses: low, 100 to 250 g/d; medium, greater than 250 to 500 g/d; and high, greater than 500 to 1,000 g/d.
a
Figure 2. Mean percentage of cortisol suppression for each cortisol measurement method by Global Initiative for Asthma (GINA) dose group. Error bars represent 95% confidence intervals. UFC indicates urinary free cortisol.
of the ICS dose to the minimum required to maintain symptom control. Until the clinical relevance of HPA suppression is better understood, physicians can help enhance the benefitrisk profile associated with ICS use in asthma by minimizing the risk of cortisol suppression. Inconsistent reporting of the extent of HPA suppression after ICS administration in individual studies has contributed to the controversy surrounding the clinical relevance of HPA suppression. Previous studies5,18 have used systematic review and meta-analysis techniques to objectively examine the available data by aggregating the results of individual studies to determine the impact of ICS therapy on cortisol suppression. However, in contrast to these earlier studies, the present study eligibility criteria were more stringent and controlled for factors that are now widely recognized as confounding factors in studies of cortisol suppression. These factors include the type of patients studied (based on age, severity of
VOLUME 100, JANUARY, 2008
airflow obstruction, and previous corticosteroid exposure) and the type of drug used (based on dose, formulation type, and drug potency). The present patient population was restricted to adults with asthma. Patient age may affect cortisol levels; compared with adults, children treated with ICSs may experience a delay in the time to onset of peak cortisol levels.45 Recent studies36,46,47 have also demonstrated that the extent of airflow obstruction is inversely associated with the systemic availability of ICSs; this, in turn, can have a marked effect on the extent of cortisol suppression. This relationship means that a patient with asthma may absorb less of an ICS dose into their lungs than a control and, consequently, may have less drug available for systemic absorption. Conversely, for the same ICS dose, healthy individuals with no airway obstruction may absorb more of the drug into the lungs, which may result in greater systemic absorption. In the present study, we did not include
27
data from controls, and most of the studies (28 of 31) were conducted in patients with mild to moderate asthma. Researchers have also shown that previous exposure to corticosteroids may affect not only the extent of cortisol suppression after ICS administration but also the baseline cortisol values.48 This was an important consideration when designing the inclusion criteria in the present study, as the primary outcome, cortisol suppression is expressed as percentage change from baseline cortisol. Any factor that affected the baseline cortisol value could confound the interpretation of these results. To minimize the impact of this potentially confounding variable, we included only studies of patients with previous corticosteroid exposure if the design included a minimum 3-day washout period from previous ICS use. A subgroup analysis excluding studies with a washout period of less than 7 days (6 studies) produced results consistent with the main analysis. All included studies had a minimum 1-month washout period from previous OCS use, with most studies requiring a 6-month washout period. We also considered the potentially confounding factors associated with directly comparing the cortisol suppressive effects between different ICSs when designing this study. To minimize the effects of the differences in ICS dose, formulation type, device, and potency used across studies, we used the classification system published in the GINA guidelines,2 which facilitates the investigation of ICS effects on cortisol suppression by grouping ICSs with different potencies and formulations into approximately equipotent dose categories (low, medium, and high). We extracted data from studies that used different ICSs and formulations and then used the GINA classification system to make meaningful comparisons of the effects of low-, medium-, and high-dose ICSs, as a class, on cortisol suppression. These results align well with findings from previous systematic reviews and meta-analyses,2,5,18 which showed that high-dose ICS use can cause cortisol suppression. However, this study extends these earlier findings by demonstrating that cortisol suppression is evident even at low ICS doses. This additional finding is arguably of equal or greater clinical importance than showing that high-dose ICS use causes cortisol suppression. The impact of low-dose ICS use on cortisol suppression has direct implications for the safety profile associated with ICS use: patients with asthma who use low doses of ICSs are not immune from the risk of cortisol suppression. Perhaps of greater concern is the uncertainty of the effect that these perturbations in cortisol secretion will have across the months or years that these agents are usually prescribed. Because ethical constraints prevent long-term, placebo-controlled trials that examine the safety and efficacy of ICSs in patients with asthma, modeling offers an alternative means of evaluating the long-term clinical and economic consequences of the effect of ICSs on cortisol suppression. The data from this study may prove useful in modeling the long-term systemic effects of ICSs. There are many different measures of cortisol suppression, all of which are surrogate measures for assessing HPA axis
28
function. The 2 broad categories of measurement methods that use cortisol levels to evaluate HPA axis function include measures of reserve function and measures of basal function. There is no universal agreement about which of these measures is the most clinically relevant. Measures of reserve function include the insulin tolerance test and cosyntropin stimulation. Although these measures evaluate dynamic HPA axis function, they have been criticized for lack of sensitivity in detecting low levels of cortisol suppression.48 Study outcomes included in this meta-analysis evaluated basal HPA axis function. Measures of basal function, including cumulative overnight UFC, cumulative serum or plasma cortisol, and morning serum or plasma cortisol concentrations, are more sensitive than measures of reserve function for detecting cortisol suppression.48 These practical and reproducible screening measurements are frequently used in clinical trials as a surrogate marker for the systemic bioavailability and safety of ICSs.45,48 Sensitive measures are indicative of systemic activity; however, the clinical relevance of these measures is unclear. Thus, caution must be used when interpreting results of studies that used sensitive measures. However, there is a close correlation between cortisol levels measured by UFC and cortisol levels measured by the insulin tolerance test, the gold standard measure of HPA axis reserve function.45 Previous researchers49,50 have criticized the use of morning serum cortisol concentration owing to its lack of sensitivity for detecting adrenal suppression. The present results support this criticism. Although Martin and colleagues41 reported that only the cumulative serum or plasma cortisol method gave consistent dose-response measurements, we found that cumulative serum or plasma cortisol concentrations and cumulative overnight UFC excretion were sensitive to alterations in cortisol levels and produced consistent results across the dose ranges examined. The strengths and limitations of this study should be considered when evaluating these results. We examined ICSs that are commonly used in current clinical practice, and we used stringent inclusion and exclusion criteria for study selection to control for variables such as the degree of airflow obstruction and drug formulation that could confound the interpretation of the results. Despite the use of stringent eligibility criteria, there may be unknown factors that could affect cortisol suppression that we have not accounted for in this study. In fact, there was considerable heterogeneity among studies and study arms. Despite this variability, we detected a significant effect of ICSs on cortisol suppression. Furthermore, the consistency in these results and compared with previously published studies reinforces the comparability of the studies included and the validity of these conclusions. We also appreciate that other, newer ICS compounds, including ciclesonide, have become available, and further study is warranted to determine whether the current findings extend to those newer agents. The clinical relevance of cortisol suppression is unclear. Thus, it is difficult to predict whether patients with cortisol
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
suppression may be at risk for acute adrenal crisis. Further investigation is warranted to determine whether measurements of basal and reserve function are indicative of clinically relevant suppression of cortisol levels. In the meantime, physicians could consider additional investigations for patients who receive ICSs and develop marked cortisol suppression. These additional investigations could include dynamic stimulation testing (if impaired cortisol reserve is suspected) or evaluations of bone density, bruising, and blood pressure (if exogenous Cushing disease is suspected). In conclusion, this meta-analysis demonstrates that a statistically significant effect of cortisol suppression is evident even at low doses of ICSs and increases with dose. The results provide evidence to support an impact of all ICSs on endogenous cortisol levels and underscore the importance of titrating ICS doses to the minimum required to maintain symptom control to maximize the balance between efficacy and safety.
13. 14. 15. 16. 17. 18. 19.
20.
ACKNOWLEDGMENT We acknowledge the independent medical writing assistance provided by ProScribe Medical Communications. ProScribe’s services complied with international guidelines for Good Publication Practice.
21.
22.
REFERENCES 1. British Thoracic Society; Scottish Intercollegiate Guidelines Network. British guideline on the management of asthma. Thorax. 2003;58(suppl 1):i1–i94. 2. Global Initiative for Asthma. Global Strategy for Asthma Management and Prevention. Bethesda, MD: National Heart, Lung, and Blood Institute; 2002. NIH publication 02–3659. 3. National Asthma Education and Prevention Program. Expert Panel Report II: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Heart, Lung, and Blood Institute, National Institutes of Health; July 1997. NIH publication 97– 4051. 4. National Asthma Education and Prevention Program. Guidelines for the Diagnosis and Management of Asthma: Update on Selected Topics 2002. Bethesda, MD: National Heart, Lung, and Blood Institute, National Institutes of Health; 2002. NIH publication 02–5075. 5. Lipworth BJ. Systemic adverse effects of inhaled corticosteroid therapy: a systematic review and meta-analysis. Arch Intern Med. 1999;159: 941–955. 6. Macdessi JS, Randell TL, Donaghue KC, Ambler GR, van Asperen PP, Mellis CM. Adrenal crises in children treated with high-dose inhaled corticosteroids for asthma. Med J Aust. 2003;178:214 –216. 7. Todd GR, Acerini CL, Ross-Russell R, Zahra S, Warner JT, McCance D. Survey of adrenal crisis associated with inhaled corticosteroids in the United Kingdom. Arch Dis Child. 2002;87:457– 461. 8. Greenfield JR, Samaras K. Suppression of HPA axis in adults taking inhaled corticosteroids. Thorax. 2006;61:272–273. 9. Barnes NC, Hallett C, Harris TA. Clinical experience with fluticasone propionate in asthma: a meta-analysis of efficacy and systemic activity compared with budesonide and beclomethasone dipropionate at half the microgram dose or less. Respir Med. 1998;92:95–104. 10. Bisgaard H, Damkjaer Nielsen M, Andersen B, et al. Adrenal function in children with bronchial asthma treated with beclomethasone dipropionate or budesonide. J Allergy Clin Immunol. 1988;81:1088 –1095. 11. Clark DJ, Grove A, Cargill RI, Lipworth BJ. Comparative adrenal suppression with inhaled budesonide and fluticasone propionate in adult asthmatic patients. Thorax. 1996;51:262–266. 12. Kamada AK, Szefler SJ, Martin RJ, et al. Issues in the use of inhaled
VOLUME 100, JANUARY, 2008
23. 24. 25. 26. 27. 28.
29. 30.
31.
32. 33.
glucocorticoids: the Asthma Clinical Research Network. Am J Respir Crit Care Med. 1996;153:1739 –1748. McCubbin MM, Milavetz G, Grandgeorge S, et al. A bioassay for topical and systemic effect of three inhaled corticosteroids. Clin Pharmacol Ther. 1995;57:455– 460. Pedersen S, Fuglsang G. Urine cortisol excretion in children treated with high doses of inhaled corticosteroids: a comparison of budesonide and beclomethasone. Eur Respir J. 1988;1:433– 435. Prahl P, Jensen T. Decreased adreno-cortical suppression utilizing the Nebuhaler for inhalation of steroid aerosols. Clin Allergy. 1987;17: 393–398. Prahl P. Adrenocortical suppression following treatment with beclomethasone and budesonide. Clin Exp Allergy. 1991;21:145–146. Dluhy RG. Clinical relevance of inhaled corticosteroids and HPA axis suppression. J Allergy Clin Immunol. 1998;101:S447–S450. Chrousos GP, Harris AG. Hypothalamic-pituitary-adrenal axis suppression and inhaled corticosteroid therapy, 2: review of the literature. Neuroimmunomodulation. 1998;5:288 –308. Aaronson D, Kaiser H, Dockhorn R, et al. Effects of budesonide by means of the Turbuhaler on the hypothalamic-pituitary-adrenal axis in asthmatic subjects: a dose-response study. J Allergy Clin Immunol. 1998;101:312–319. Altman LC, Findlay SR, Lopez M, et al. Adrenal function in adult asthmatics during long-term daily treatment with 800, 1,200, and 1,600 micrograms triamcinolone acetonide: multicenter study. Chest. 1992; 101:1250 –1256. Bootsma GP, Dekhuijzen PN, Festen J, Mulder PG, van Herwaarden CL. Comparison of fluticasone propionate and beclomethasone dipropionate on direct and indirect measurements of bronchial hyperresponsiveness in patients with stable asthma. Thorax. 1995;50:1044 –1050. Casale TB, Nelson HS, Stricker WE, Raff H, Newman KB. Suppression of hypothalamic-pituitary-adrenal axis activity with inhaled flunisolide and fluticasone propionate in adult asthma patients. Ann Allergy Asthma Immunol. 2001;87:379 –385. Clark DJ, Lipworth BJ. Adrenal suppression with chronic dosing of fluticasone propionate compared with budesonide in adult asthmatic patients. Thorax. 1997;52:55–58. Clark DJ, Lipworth BJ. Evaluation of corticotropin releasing factor stimulation and basal markers of hypothalamic-pituitary-adrenal axis suppression in asthmatic patients. Chest. 1997;112:1248 –1252. Derom E, Van Schoor J, Verhaeghe W, Vincken W, Pauwels R. Systemic effects of inhaled fluticasone propionate and budesonide in adult patients with asthma. Am J Respir Crit Care Med. 1999;160:157–161. Fardon TC, Lee DK, Haggart K, McFarlane LC, Lipworth BJ. Adrenal suppression with dry powder formulations of fluticasone propionate and mometasone furoate. Am J Respir Crit Care Med. 2004;170:960 –966. Hauache OM, Amarante EC, Vieira JG, et al. Evaluation of bone metabolism after the use of an inhaled glucocorticoid (flunisolide) in patients with moderate asthma. Clin Endocrinol. 1999;51:35–39. Kannisto S, Laatikainen A, Taivainen A, Savolainen K, Tukiainen H, Voutilainen R. Serum dehydroepiandrosterone sulfate concentration as an indicator of adrenocortical suppression during inhaled steroid therapy in adult asthmatic patients. Eur J Endocrinol. 2004;150:687– 690. Lawrence M, Wolfe J, Webb DR, et al. Efficacy of inhaled fluticasone propionate in asthma results from topical and not from systemic activity. Am J Respir Crit Care Med. 1997;156:744 –751. Lee DK, Bates CE, Currie GP, Cowan LM, McFarlane LC, Lipworth BJ. Effects of high-dose inhaled fluticasone propionate on the hypothalamicpituitary-adrenal axis in asthmatic patients with severely impaired lung function. Ann Allergy Asthma Immunol. 2004;93:253–258. Li JT, Ford LB, Chervinsky P, et al. Fluticasone propionate powder and lack of clinically significant effects on hypothalamic-pituitary-adrenal axis and bone mineral density over 2 years in adults with mild asthma. J Allergy Clin Immunol. 1999;103:1062–1068. Li JT, Goldstein MF, Gross GN, et al. Effects of fluticasone propionate, triamcinolone acetonide, prednisone, and placebo on the hypothalamicpituitary-adrenal axis. J Allergy Clin Immunol. 1999;103:622– 629. Molema J, Lammers JW, van Herwaarden CL, Folgering HT. Effects of
29
34.
35. 36. 37.
38. 39.
40. 41. 42.
30
inhaled beclomethasone dipropionate on 2-receptor function in the airways and adrenal responsiveness in bronchial asthma. Eur J Clin Pharmacol. 1988;34:577–583. Sorkness CA, LaForce C, Storms W, Lincourt WR, Edwards L, Rogenes PR. Effects of the inhaled corticosteroids fluticasone propionate, triamcinolone acetonide, and flunisolide and oral prednisone on the hypothalamic-pituitary-adrenal axis in adult patients with asthma. Clin Ther. 1999;21:353–367. Szefler SJ, Martin RJ, King TS, et al. Significant variability in response to inhaled corticosteroids for persistent asthma. J Allergy Clin Immunol. 2002;109:410 – 418. Weiner P, Berar-Yanay N, Davidovich A, Magadle R. Nocturnal cortisol secretion in asthmatic patients after inhalation of fluticasone propionate. Chest. 1999;116:931–934. Wilson AM, McFarlane LC, Lipworth BJ. Dose-response effect for adrenal suppression with repeated twice daily inhaled fluticasone propionate and triamcinolone acetonide in adult asthmatics. Am J Respir Crit Care Med. 1997;156:1274 –1277. Wilson AM, Lipworth BJ. Dose-response evaluation of the therapeutic index for inhaled budesonide in patients with mild-to-moderate asthma. Am J Med. 2000;108:269 –275. Wilson AM, Clark DJ, Devlin MM, McFarlane LC, Lipworth BJ. Adrenocortical activity with repeated administration of once-daily inhaled fluticasone propionate and budesonide in asthmatic adults. Eur J Clin Pharmacol. 1998;53:317–320. Wilson AM, Sims EJ, Struthers AD, Lipworth BJ. Inhaled corticosteroid therapy reduces the early morning peak in cortisol and aldosterone. Clin Sci. 1998;95:513–517. Martin RJ, Szefler SJ, Chinchilli VM, et al. Systemic effect comparisons of six inhaled corticosteroid preparations. Am J Respir Crit Care Med. 2002;165:1377–1383. Affrime MB, Kosoglou T, Thonoor CM, Flannery BE, Herron JM. Mometasone furoate has minimal effects on the hypothalamic-pituitary-
adrenal axis when delivered at high doses. Chest. 2000;118:1538 –1546. 43. Currie GP, Fowler AM, Wilson EJ, Sims L, Orr C, Lipworth BJ. Airway and systemic effects of hydrofluoroalkane fluticasone and beclomethasone in patients with asthma. Thorax. 2002;57:865– 868. 44. Wilson AM, Brewster HJ, Lipworth BJ. Dose-response comparison of systemic bioactivity with inhaled budesonide and triamcinolone acetonide in asthmatic adults. J Allergy Clin Immunol. 1998;102:751–756. 45. Boner AL. Effects of intranasal corticosteroids on the hypothalamicpituitary-adrenal axis in children. J Allergy Clin Immunol. 2001;108: S32–S39. 46. Brutsche MH, Brutsche IC, Munawar M, et al. Comparison of pharmacokinetics and systemic effects of inhaled fluticasone propionate in patients with asthma and healthy volunteers: a randomised crossover study. Lancet. 2000;356:556 –561. 47. Harrison TW, Tattersfield AE. Plasma concentrations of fluticasone propionate and budesonide following inhalation from dry powder inhalers by healthy and asthmatic subjects. Thorax. 2003;58:258 –260. 48. Kelly HW, Nelson HS. Potential adverse effects of the inhaled corticosteroids. J Allergy Clin Immunol. 2003;112:469 – 478. 49. Krasner AS. Glucocorticoid-induced adrenal insufficiency. JAMA. 1999; 282:671– 676. 50. Henzen C, Suter A, Lerch E, Urbinelli R, Schorno XH, Briner VA. Suppression and recovery of adrenal response after short-term, highdose glucocorticoid treatment. Lancet. 2000;355:542–545.
Requests for reprints should be addressed to: John Wlodarczyk, BEc Dip Med Stats, PhD John Wlodarczyk Consulting Services 10 Mirral Crescent New Lambton, NSW 2305, Australia E-mail: [email protected]
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
Background: Studies examining the effects of inhaled corticosteroids (ICSs) on cortisol suppression show inconsistent results, and there is uncertainty regarding the dose-response relationship between ICSs and cortisol suppression. Objective: To determine, using meta-analysis, the extent of cortisol suppression after administration of clinically relevant ICS doses in adults with asthma. Methods: Database searches (MEDLINE, EMBASE, and The Cochrane Library) using appropriate indexed terms were performed to identify eligible articles for review. Articles reporting the effects of ICSs on cortisol levels in asthmatic adults, measured using the cumulative serum or plasma cortisol, morning serum or plasma cortisol, or cumulative overnight urinary free cortisol method, were identified. All available cortisol measurements were extracted. Cortisol suppression was estimated, and treatment arms were grouped into low-, medium-, and high-dose ranges according to the Global Initiative for Asthma guidelines. A multivariate model was used to determine relationships between ICS dose and cortisol suppression and to explore sources of heterogeneity among trials. Results: Thirty-one studies providing information on 216 measures of cortisol suppression were included in this meta-analysis. Cortisol suppression in the low-, medium-, and high-dose groups were estimated to be 17.92% (95% confidence interval [CI], 11.08%–24.77%), 26.55% (95% CI, 17.29%–35.80%), and 36.31% (95% CI, 26.48%– 46.13%), respectively. Conclusions: Statistically significant cortisol suppression was evident at low doses of ICSs and increased with dose. These results support an impact of all ICSs on endogenous cortisol levels and underscore the importance of titrating ICS doses to the minimum required to maintain symptom control. Ann Allergy Asthma Immunol. 2008;100:23–30.
INTRODUCTION Inhaled corticosteroids (ICSs) have been established as cornerstone therapy for chronic asthma, with treatment guidelines1– 4 recommending ICS treatment for patients with persistent asthma of all severities. Although the potential for systemic effects is far lower than with a comparable dose of oral corticosteroids (OCSs), the use of ICSs can be associated with dose-related systemic adverse effects. These effects include hypothalamic-pituitary-adrenal (HPA) axis suppression, decreased bone density, growth suppression, cataracts, and skin changes.5 Clinical concerns regarding the untoward effects of ICSs on the HPA axis have been heightened by recent case reports associating the use of ICSs with acute adrenal crisis in children6,7 and with the risk of adrenal insufficiency in adults.8 Despite the heightened awareness of the potential for untoward HPA axis effects, the clinical relevance of HPA axis suppression resulting from ICS use is controversial. Central Affiliations: * John Wlodarczyk Consulting Services, New Lambton, New South Wales, Australia; † Respiratory and Sleep Medicine, John Hunter Hospital, Newcastle, New South Wales, Australia; ‡ Altana Pharma AG, Konstanz, Germany. Disclosures: Authors have nothing to disclose. Funding Sources: This study was sponsored by Altana Pharma. Received for publication September 26, 2006; Received in revised form July 16, 2007; Accepted for publication July 23, 2007.
VOLUME 100, JANUARY, 2008
to this controversy is the inconsistent reporting of the presence and extent of HPA axis suppression after ICS administration. As a result, the impact of perturbations to the HPA axis during the months and years that these agents are typically prescribed is difficult to estimate. The effects of ICSs on HPA suppression have been measured directly in individual studies and indirectly in systemic reviews and meta-analyses. Individual studies9 –16 have shown inconsistent effects of ICSs on cortisol suppression, a sensitive marker of HPA axis function. The inconsistencies seen across studies may be due to factors such as differences in dose, formulation, frequency of administration, treatment duration, previous systemic corticosteroid therapy, method of cortisol measurement used, and patient population studied.17 Systematic reviews and meta-analyses have been used in an effort to overcome the limitations associated with the individual studies.5,18 Chrousos and Harris18 performed a systematic review on the results of 50 studies and concluded that ICSs, when administered in recommended dose ranges, did not endanger HPA axis functioning. However, a caveat that the authors noted was that some patients exhibited measurable attenuations in HPA axis function, even at low doses of ICS (fluticasone propionate, 200 – 400 g/d).18 In a metaanalysis of 27 studies, Lipworth5 concluded that cortisol suppression was evident after high-dose ICS administration. The author highlighted that fluticasone propionate was asso-
23
ciated with greater dose-related suppression vs other ICSs; effects were most apparent at fluticasone propionate doses greater than 800 g/d. Although both studies examined ICS effects on cortisol suppression, many factors limit the interpretation of these results. The study by Chrousos and Harris was a qualitative study and did not provide quantification of the effects of ICSs on cortisol suppression. The Lipworth study included studies examining different patient populations (controls and adult and pediatric patients) and focused on detecting large changes in HPA axis function as opposed to the more subtle effects of commonly used doses. In addition, neither study used the therapeutic equivalence of the ICSs to combine information from different drugs. Although there is general agreement that ICSs affect HPA axis function, studies detailing the extent of this response have been inconclusive. Careful quantification of the effect of ICSs on cortisol suppression, therefore, represents an important step in elucidating the clinical relevance of HPA suppression after ICS administration. The aim of this study was to determine, using meta-analysis, the extent of cortisol suppression after administration of clinically relevant ICS doses in adults with asthma. To combine data extracted from studies that used different ICSs and formulations, we used the classification system published in the Global Initiative for Asthma (GINA) guidelines.2 This classification system facilitated the investigation of ICS effects on cortisol suppression by grouping ICSs with different potencies and formulations into approximately equipotent dose categories (low, medium, and high doses). METHODS Search Strategy Database searches (MEDLINE, EMBASE, and The Cochrane Library) for articles published between January 1, 1966, and March 31, 2005, were performed to identify eligible articles for review. All searches included the terms inhaled corticosteroid and asthma and 1 of the following terms: adrenal function, adrenal suppression, or cortisol. The only limits used on the searches were for human subjects. Studies were included in the meta-analysis if they met the following criteria: used 1 or more of the ICSs outlined in the GINA guidelines (beclomethasone, budesonide, flunisolide,
fluticasone propionate, triamcinolone acetonide, and mometasone furoate); defined the number of patients in each study arm; included only adults (ⱖ18 years or older); provided defined inclusion and exclusion criteria for study participants (studies in nonasthmatic patients were excluded); defined the treatment protocol (ie, medications used); defined the ICS dose for each treatment group; included a minimum 3-day washout period from previous ICS use, or at least a 1-month washout from OCSs; presented primary results (ie, no review articles and meta-analyses were included); and included 1 or more of the following outcomes: cumulative overnight urinary free cortisol (UFC), cumulative serum or plasma cortisol, and morning serum or plasma cortisol levels. There were no minimum treatment period requirements. Case reports, small-scale trials (⬍5 patients), letters to the editor, reviews, trials on topical corticosteroids using nebulizers, trials on children or healthy patients, and trials with insufficient information to estimate a drug effect were excluded from the meta-analysis. Each article was reviewed by 2 independent reviewers. Any discrepancies in the reason for inclusion or exclusion or in data abstraction were discussed, and disagreement regarding eligibility was resolved by consensus. The references of all the articles that were evaluated for inclusion were also appraised for potential inclusion in the meta-analysis. Statistical Methods Articles meeting the eligibility criteria were abstracted into data tables by 2 independent reviewers. Each dose of ICS was classified according to GINA guidelines, where daily doses of ICSs are defined as low, medium, and high (Table 1).2 All available cortisol measurements were extracted. The primary outcome measure used in this study was cortisol suppression expressed as the percentage change in cortisol from baseline, which was calculated as follows: (end point cortisol ⫺ baseline cortisol)/baseline cortisol. In some studies, only end point cortisol measurements were reported, and, hence, the percentage change from baseline could not be calculated for each arm. If the studies included a placebo arm, an estimate of cortisol suppression was calculated using the difference between end point measurements for active treatment compared with placebo. Several studies also measured
Table 1. GINA Inhaled Corticosteroid Dose Groups GINA dose group, g/d
Drug Fluticasone propionate Budesonide dry powder inhaler Triamcinolone acetonide Flunisolide Beclomethasone dipropionate chlorofluorocarbon Beclomethasone dipropionate hydrofluoroalkane Mometasone furoate
Low
Medium
High
100–250 200–600 400–1,000 500–1,000 200–500 100–250 200–400
⬎250–500 ⬎600–1,000 ⬎1,000–2,000 ⬎1,000–2,000 ⬎500–1,000 ⬎250–500 ⬎400–800
⬎500 ⬎1,000 ⬎2,000 ⬎2,000 ⬎1,000 ⬎500 ⬎800
Abbreviation: GINA, Global Initiative for Asthma.
24
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
cortisol levels in the same individuals at different dose levels. The percentage of cortisol suppression for each dose level was included in the analysis as a separate observation. Other variables extracted included trial duration (in days), trial design (parallel group vs crossover), and previous OCS use. The SAS (SAS version 9.1; SAS Institute Inc, Cary, North Carolina) was used to perform statistical analyses. A generalized estimating equation model with an exchangeable correlation structure was used to determine the relationships between ICS dose groups and cortisol suppression after adjusting for multiple measurements in some of the studies. Weighting was by the number of patients in each treatment arm. All available measures of cortisol were retained in the model, and indicator variables were used to evaluate the impact of different measurement methods (ie, cumulative overnight UFC, cumulative serum or plasma cortisol, and morning serum or plasma cortisol levels). The heterogeneity of cortisol suppression was explored using bubble plots. Exploratory modeling showed that trial duration, design, and previous OCS use did not significantly affect cortisol
suppression. The effect of previous ICS washout duration on cortisol suppression was modeled, and there was no significant difference on cortisol suppression if those studies with washout periods shorter than 7 days were eliminated or retained in the analysis. Modeling the individual measurement methods produced results consistent with the combined model. A reduced model containing only those variables that were significant at the P ⬍ .05 level is presented. RESULTS We reviewed 385 articles from MEDLINE, 90 from EMBASE, and 198 from The Cochrane Library. The overall search strategy, the number of articles retrieved, and the reasons for article exclusion are depicted in Figure 1. A total of 28 articles were retained11,13,19 – 44; one article34 reported on 2 studies and another article42 reported on 3 studies. Thus, 31 studies were included in the meta-analysis. Twenty-four of the included studies were randomized. Of these, 14 were double masked, 6 were single masked, and 4 did not specify the type of masking. Eleven of the random-
Figure 1. Search strategy, including the number of articles retrieved and the reasons for inclusion or exclusion. A total of 28 articles were retained; one article reported on 2 studies and another article reported on 3 studies. Thus, 31 studies were included in the meta-analysis.
VOLUME 100, JANUARY, 2008
25
ized studies used a parallel group design, 9 used a crossover design, and 4 were open label. Of the 7 studies that did not specify randomization, 3 were double masked, 1 was single masked, and 3 did not specify the type of masking; 1 used a parallel group design, 5 used a crossover design, and 1 was open label. Most included studies were conducted in patients with mild to moderate asthma based on a percentage of the predicted forced expiratory volume or peak expiratory function and on patient-defined symptoms. Only 1 study30 reported on patients with severe asthma, and 2 studies33,36 reported on patients with moderate to severe asthma. All included studies had a minimum 1-month washout period from previous OCS use, with most studies requiring a 6-month washout period. Several of the included studies had multiple treatment arms and used more than 1 method to measure cortisol levels. In studies using multiple methods to measure cortisol levels,22,23,30,37,39,42,44 no qualitative differences in the direction of drug effect on cortisol suppression among the various modes of measurement were observed. Overall, 216 estimates of cortisol suppression were extracted from the 31 studies and were included in the analysis. The greatest numbers of individual cortisol measurements were obtained from fluticasone propionate and budesonide studies (Table 2). There was considerable heterogeneity between studies and study arms. Cortisol suppression varied from ⫺18.8% to 91.5%, with an interquartile range of 7.2% to 30.75%. The interquartile ranges for the low, medium, and high GINA dose groups were 5.0% to 14.3%, 12.9% to 35.7%, and 24.3% to 47.1%, respectively. Allocation of the cortisol measurements into the appropriate GINA dose groups revealed that the number of cortisol measurements across the 3 dose ranges were similar (Table 3). Likewise, the types of cortisol measurements were evenly distributed across the dose ranges except for morning serum or plasma cortisol levels, where the largest proportion of measurements (41%) was in the high-dose group.
Statistically significant cortisol suppression after ICS treatment was evident in all 3 GINA dose groups; the extent of suppression was dose dependent and increased as the dose increased (P ⬍ .001). Based on a combined model (in which all measures of cortisol were correlated with dose groups), suppression in the low, medium, and high GINA dose groups relative to baseline was estimated to be 17.92% (95% confidence interval; [CI], 11.08%–24.77%), 26.55% (95% CI, 17.29%–35.80%), and 36.31% (95% CI, 26.48%– 46.13%), respectively. Modeling the individual measurement methods produced similar results as the combined model and provided additional information regarding the measurement methods (Fig 2). Cumulative serum or plasma cortisol and cumulative overnight UFC measures were more sensitive to cortisol changes, although there was greater variance associated with these measures. Compared with cumulative measures (combined cumulative serum or plasma cortisol and cumulative overnight UFC), estimates of cortisol suppression using morning serum or plasma cortisol levels indicated, on average, 10.91% less suppression (95% CI, 4.30 –17.51). DISCUSSION This meta-analysis provides clear evidence regarding the extent of cortisol suppression after ICS administration. We demonstrated that low doses of commonly used ICSs produce statistically significant effects on cortisol suppression in adults with asthma. A statistically significant dose-response relationship was evident, with the extent of cortisol suppression becoming more marked at higher ICS doses. By showing that cortisol suppression was evident, even at low ICS doses, these findings represent an important step in elucidating the clinical relevance of HPA axis suppression after ICS administration. The dose-response estimates presented in this study provide a basis on which to calculate the average systemic impact of currently available ICSs in adults with asthma. Furthermore, the impact of low-dose ICS use on cortisol suppression as reported herein reinforces the need for titration
Table 2. Cortisol Measurements Available for Each Study Drug Cumulative Overnight UFC
Drug
Fluticasone propionate Budesonide Triamcinolone acetonide Flunisolide Beclomethasone dipropionate Mometasone furoate Total
Serum or plasma cortisol
Morning serum or plasma cortisol
Treatment arms, No.a
Studies (patients), No.
Treatment arms, No.a
Studies (patients), No.
Treatment arms, No.a
Studies (patients), No.
30 18 16 9 7 3 83
8 (528) 6 (318) 5 (274) 3 (152) 2 (123) 1 (63)
25 6 8 7 7 0 53
8 (521) 2 (142) 3 (173) 3 (154) 2 (135) 0
30 21 12 2 4 11 80
14 (608) 8 (367) 4 (190) 2 (49) 2 (99) 4 (175)
Abbreviation: UFC, urinary free cortisol. Estimates of percentage change from baseline in cortisol levels might include several arms from a single study.
a
26
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
Table 3. Cortisol Measurements Available for Each GINA Dose Group Measurements, No. GINA dose groupa
Low Medium High Total
Cumulative overnight UFC
Cumulative serum or plasma cortisol
Morning serum or plasma cortisol
Total
27 27 29 83
21 15 17 53
24 23 33 80
72 65 79 216
Abbreviations: GINA, Global Initiative for Asthma; UFC, urinary free cortisol. Fluticasone propionate– equivalent doses: low, 100 to 250 g/d; medium, greater than 250 to 500 g/d; and high, greater than 500 to 1,000 g/d.
a
Figure 2. Mean percentage of cortisol suppression for each cortisol measurement method by Global Initiative for Asthma (GINA) dose group. Error bars represent 95% confidence intervals. UFC indicates urinary free cortisol.
of the ICS dose to the minimum required to maintain symptom control. Until the clinical relevance of HPA suppression is better understood, physicians can help enhance the benefitrisk profile associated with ICS use in asthma by minimizing the risk of cortisol suppression. Inconsistent reporting of the extent of HPA suppression after ICS administration in individual studies has contributed to the controversy surrounding the clinical relevance of HPA suppression. Previous studies5,18 have used systematic review and meta-analysis techniques to objectively examine the available data by aggregating the results of individual studies to determine the impact of ICS therapy on cortisol suppression. However, in contrast to these earlier studies, the present study eligibility criteria were more stringent and controlled for factors that are now widely recognized as confounding factors in studies of cortisol suppression. These factors include the type of patients studied (based on age, severity of
VOLUME 100, JANUARY, 2008
airflow obstruction, and previous corticosteroid exposure) and the type of drug used (based on dose, formulation type, and drug potency). The present patient population was restricted to adults with asthma. Patient age may affect cortisol levels; compared with adults, children treated with ICSs may experience a delay in the time to onset of peak cortisol levels.45 Recent studies36,46,47 have also demonstrated that the extent of airflow obstruction is inversely associated with the systemic availability of ICSs; this, in turn, can have a marked effect on the extent of cortisol suppression. This relationship means that a patient with asthma may absorb less of an ICS dose into their lungs than a control and, consequently, may have less drug available for systemic absorption. Conversely, for the same ICS dose, healthy individuals with no airway obstruction may absorb more of the drug into the lungs, which may result in greater systemic absorption. In the present study, we did not include
27
data from controls, and most of the studies (28 of 31) were conducted in patients with mild to moderate asthma. Researchers have also shown that previous exposure to corticosteroids may affect not only the extent of cortisol suppression after ICS administration but also the baseline cortisol values.48 This was an important consideration when designing the inclusion criteria in the present study, as the primary outcome, cortisol suppression is expressed as percentage change from baseline cortisol. Any factor that affected the baseline cortisol value could confound the interpretation of these results. To minimize the impact of this potentially confounding variable, we included only studies of patients with previous corticosteroid exposure if the design included a minimum 3-day washout period from previous ICS use. A subgroup analysis excluding studies with a washout period of less than 7 days (6 studies) produced results consistent with the main analysis. All included studies had a minimum 1-month washout period from previous OCS use, with most studies requiring a 6-month washout period. We also considered the potentially confounding factors associated with directly comparing the cortisol suppressive effects between different ICSs when designing this study. To minimize the effects of the differences in ICS dose, formulation type, device, and potency used across studies, we used the classification system published in the GINA guidelines,2 which facilitates the investigation of ICS effects on cortisol suppression by grouping ICSs with different potencies and formulations into approximately equipotent dose categories (low, medium, and high). We extracted data from studies that used different ICSs and formulations and then used the GINA classification system to make meaningful comparisons of the effects of low-, medium-, and high-dose ICSs, as a class, on cortisol suppression. These results align well with findings from previous systematic reviews and meta-analyses,2,5,18 which showed that high-dose ICS use can cause cortisol suppression. However, this study extends these earlier findings by demonstrating that cortisol suppression is evident even at low ICS doses. This additional finding is arguably of equal or greater clinical importance than showing that high-dose ICS use causes cortisol suppression. The impact of low-dose ICS use on cortisol suppression has direct implications for the safety profile associated with ICS use: patients with asthma who use low doses of ICSs are not immune from the risk of cortisol suppression. Perhaps of greater concern is the uncertainty of the effect that these perturbations in cortisol secretion will have across the months or years that these agents are usually prescribed. Because ethical constraints prevent long-term, placebo-controlled trials that examine the safety and efficacy of ICSs in patients with asthma, modeling offers an alternative means of evaluating the long-term clinical and economic consequences of the effect of ICSs on cortisol suppression. The data from this study may prove useful in modeling the long-term systemic effects of ICSs. There are many different measures of cortisol suppression, all of which are surrogate measures for assessing HPA axis
28
function. The 2 broad categories of measurement methods that use cortisol levels to evaluate HPA axis function include measures of reserve function and measures of basal function. There is no universal agreement about which of these measures is the most clinically relevant. Measures of reserve function include the insulin tolerance test and cosyntropin stimulation. Although these measures evaluate dynamic HPA axis function, they have been criticized for lack of sensitivity in detecting low levels of cortisol suppression.48 Study outcomes included in this meta-analysis evaluated basal HPA axis function. Measures of basal function, including cumulative overnight UFC, cumulative serum or plasma cortisol, and morning serum or plasma cortisol concentrations, are more sensitive than measures of reserve function for detecting cortisol suppression.48 These practical and reproducible screening measurements are frequently used in clinical trials as a surrogate marker for the systemic bioavailability and safety of ICSs.45,48 Sensitive measures are indicative of systemic activity; however, the clinical relevance of these measures is unclear. Thus, caution must be used when interpreting results of studies that used sensitive measures. However, there is a close correlation between cortisol levels measured by UFC and cortisol levels measured by the insulin tolerance test, the gold standard measure of HPA axis reserve function.45 Previous researchers49,50 have criticized the use of morning serum cortisol concentration owing to its lack of sensitivity for detecting adrenal suppression. The present results support this criticism. Although Martin and colleagues41 reported that only the cumulative serum or plasma cortisol method gave consistent dose-response measurements, we found that cumulative serum or plasma cortisol concentrations and cumulative overnight UFC excretion were sensitive to alterations in cortisol levels and produced consistent results across the dose ranges examined. The strengths and limitations of this study should be considered when evaluating these results. We examined ICSs that are commonly used in current clinical practice, and we used stringent inclusion and exclusion criteria for study selection to control for variables such as the degree of airflow obstruction and drug formulation that could confound the interpretation of the results. Despite the use of stringent eligibility criteria, there may be unknown factors that could affect cortisol suppression that we have not accounted for in this study. In fact, there was considerable heterogeneity among studies and study arms. Despite this variability, we detected a significant effect of ICSs on cortisol suppression. Furthermore, the consistency in these results and compared with previously published studies reinforces the comparability of the studies included and the validity of these conclusions. We also appreciate that other, newer ICS compounds, including ciclesonide, have become available, and further study is warranted to determine whether the current findings extend to those newer agents. The clinical relevance of cortisol suppression is unclear. Thus, it is difficult to predict whether patients with cortisol
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
suppression may be at risk for acute adrenal crisis. Further investigation is warranted to determine whether measurements of basal and reserve function are indicative of clinically relevant suppression of cortisol levels. In the meantime, physicians could consider additional investigations for patients who receive ICSs and develop marked cortisol suppression. These additional investigations could include dynamic stimulation testing (if impaired cortisol reserve is suspected) or evaluations of bone density, bruising, and blood pressure (if exogenous Cushing disease is suspected). In conclusion, this meta-analysis demonstrates that a statistically significant effect of cortisol suppression is evident even at low doses of ICSs and increases with dose. The results provide evidence to support an impact of all ICSs on endogenous cortisol levels and underscore the importance of titrating ICS doses to the minimum required to maintain symptom control to maximize the balance between efficacy and safety.
13. 14. 15. 16. 17. 18. 19.
20.
ACKNOWLEDGMENT We acknowledge the independent medical writing assistance provided by ProScribe Medical Communications. ProScribe’s services complied with international guidelines for Good Publication Practice.
21.
22.
REFERENCES 1. British Thoracic Society; Scottish Intercollegiate Guidelines Network. British guideline on the management of asthma. Thorax. 2003;58(suppl 1):i1–i94. 2. Global Initiative for Asthma. Global Strategy for Asthma Management and Prevention. Bethesda, MD: National Heart, Lung, and Blood Institute; 2002. NIH publication 02–3659. 3. National Asthma Education and Prevention Program. Expert Panel Report II: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Heart, Lung, and Blood Institute, National Institutes of Health; July 1997. NIH publication 97– 4051. 4. National Asthma Education and Prevention Program. Guidelines for the Diagnosis and Management of Asthma: Update on Selected Topics 2002. Bethesda, MD: National Heart, Lung, and Blood Institute, National Institutes of Health; 2002. NIH publication 02–5075. 5. Lipworth BJ. Systemic adverse effects of inhaled corticosteroid therapy: a systematic review and meta-analysis. Arch Intern Med. 1999;159: 941–955. 6. Macdessi JS, Randell TL, Donaghue KC, Ambler GR, van Asperen PP, Mellis CM. Adrenal crises in children treated with high-dose inhaled corticosteroids for asthma. Med J Aust. 2003;178:214 –216. 7. Todd GR, Acerini CL, Ross-Russell R, Zahra S, Warner JT, McCance D. Survey of adrenal crisis associated with inhaled corticosteroids in the United Kingdom. Arch Dis Child. 2002;87:457– 461. 8. Greenfield JR, Samaras K. Suppression of HPA axis in adults taking inhaled corticosteroids. Thorax. 2006;61:272–273. 9. Barnes NC, Hallett C, Harris TA. Clinical experience with fluticasone propionate in asthma: a meta-analysis of efficacy and systemic activity compared with budesonide and beclomethasone dipropionate at half the microgram dose or less. Respir Med. 1998;92:95–104. 10. Bisgaard H, Damkjaer Nielsen M, Andersen B, et al. Adrenal function in children with bronchial asthma treated with beclomethasone dipropionate or budesonide. J Allergy Clin Immunol. 1988;81:1088 –1095. 11. Clark DJ, Grove A, Cargill RI, Lipworth BJ. Comparative adrenal suppression with inhaled budesonide and fluticasone propionate in adult asthmatic patients. Thorax. 1996;51:262–266. 12. Kamada AK, Szefler SJ, Martin RJ, et al. Issues in the use of inhaled
VOLUME 100, JANUARY, 2008
23. 24. 25. 26. 27. 28.
29. 30.
31.
32. 33.
glucocorticoids: the Asthma Clinical Research Network. Am J Respir Crit Care Med. 1996;153:1739 –1748. McCubbin MM, Milavetz G, Grandgeorge S, et al. A bioassay for topical and systemic effect of three inhaled corticosteroids. Clin Pharmacol Ther. 1995;57:455– 460. Pedersen S, Fuglsang G. Urine cortisol excretion in children treated with high doses of inhaled corticosteroids: a comparison of budesonide and beclomethasone. Eur Respir J. 1988;1:433– 435. Prahl P, Jensen T. Decreased adreno-cortical suppression utilizing the Nebuhaler for inhalation of steroid aerosols. Clin Allergy. 1987;17: 393–398. Prahl P. Adrenocortical suppression following treatment with beclomethasone and budesonide. Clin Exp Allergy. 1991;21:145–146. Dluhy RG. Clinical relevance of inhaled corticosteroids and HPA axis suppression. J Allergy Clin Immunol. 1998;101:S447–S450. Chrousos GP, Harris AG. Hypothalamic-pituitary-adrenal axis suppression and inhaled corticosteroid therapy, 2: review of the literature. Neuroimmunomodulation. 1998;5:288 –308. Aaronson D, Kaiser H, Dockhorn R, et al. Effects of budesonide by means of the Turbuhaler on the hypothalamic-pituitary-adrenal axis in asthmatic subjects: a dose-response study. J Allergy Clin Immunol. 1998;101:312–319. Altman LC, Findlay SR, Lopez M, et al. Adrenal function in adult asthmatics during long-term daily treatment with 800, 1,200, and 1,600 micrograms triamcinolone acetonide: multicenter study. Chest. 1992; 101:1250 –1256. Bootsma GP, Dekhuijzen PN, Festen J, Mulder PG, van Herwaarden CL. Comparison of fluticasone propionate and beclomethasone dipropionate on direct and indirect measurements of bronchial hyperresponsiveness in patients with stable asthma. Thorax. 1995;50:1044 –1050. Casale TB, Nelson HS, Stricker WE, Raff H, Newman KB. Suppression of hypothalamic-pituitary-adrenal axis activity with inhaled flunisolide and fluticasone propionate in adult asthma patients. Ann Allergy Asthma Immunol. 2001;87:379 –385. Clark DJ, Lipworth BJ. Adrenal suppression with chronic dosing of fluticasone propionate compared with budesonide in adult asthmatic patients. Thorax. 1997;52:55–58. Clark DJ, Lipworth BJ. Evaluation of corticotropin releasing factor stimulation and basal markers of hypothalamic-pituitary-adrenal axis suppression in asthmatic patients. Chest. 1997;112:1248 –1252. Derom E, Van Schoor J, Verhaeghe W, Vincken W, Pauwels R. Systemic effects of inhaled fluticasone propionate and budesonide in adult patients with asthma. Am J Respir Crit Care Med. 1999;160:157–161. Fardon TC, Lee DK, Haggart K, McFarlane LC, Lipworth BJ. Adrenal suppression with dry powder formulations of fluticasone propionate and mometasone furoate. Am J Respir Crit Care Med. 2004;170:960 –966. Hauache OM, Amarante EC, Vieira JG, et al. Evaluation of bone metabolism after the use of an inhaled glucocorticoid (flunisolide) in patients with moderate asthma. Clin Endocrinol. 1999;51:35–39. Kannisto S, Laatikainen A, Taivainen A, Savolainen K, Tukiainen H, Voutilainen R. Serum dehydroepiandrosterone sulfate concentration as an indicator of adrenocortical suppression during inhaled steroid therapy in adult asthmatic patients. Eur J Endocrinol. 2004;150:687– 690. Lawrence M, Wolfe J, Webb DR, et al. Efficacy of inhaled fluticasone propionate in asthma results from topical and not from systemic activity. Am J Respir Crit Care Med. 1997;156:744 –751. Lee DK, Bates CE, Currie GP, Cowan LM, McFarlane LC, Lipworth BJ. Effects of high-dose inhaled fluticasone propionate on the hypothalamicpituitary-adrenal axis in asthmatic patients with severely impaired lung function. Ann Allergy Asthma Immunol. 2004;93:253–258. Li JT, Ford LB, Chervinsky P, et al. Fluticasone propionate powder and lack of clinically significant effects on hypothalamic-pituitary-adrenal axis and bone mineral density over 2 years in adults with mild asthma. J Allergy Clin Immunol. 1999;103:1062–1068. Li JT, Goldstein MF, Gross GN, et al. Effects of fluticasone propionate, triamcinolone acetonide, prednisone, and placebo on the hypothalamicpituitary-adrenal axis. J Allergy Clin Immunol. 1999;103:622– 629. Molema J, Lammers JW, van Herwaarden CL, Folgering HT. Effects of
29
34.
35. 36. 37.
38. 39.
40. 41. 42.
30
inhaled beclomethasone dipropionate on 2-receptor function in the airways and adrenal responsiveness in bronchial asthma. Eur J Clin Pharmacol. 1988;34:577–583. Sorkness CA, LaForce C, Storms W, Lincourt WR, Edwards L, Rogenes PR. Effects of the inhaled corticosteroids fluticasone propionate, triamcinolone acetonide, and flunisolide and oral prednisone on the hypothalamic-pituitary-adrenal axis in adult patients with asthma. Clin Ther. 1999;21:353–367. Szefler SJ, Martin RJ, King TS, et al. Significant variability in response to inhaled corticosteroids for persistent asthma. J Allergy Clin Immunol. 2002;109:410 – 418. Weiner P, Berar-Yanay N, Davidovich A, Magadle R. Nocturnal cortisol secretion in asthmatic patients after inhalation of fluticasone propionate. Chest. 1999;116:931–934. Wilson AM, McFarlane LC, Lipworth BJ. Dose-response effect for adrenal suppression with repeated twice daily inhaled fluticasone propionate and triamcinolone acetonide in adult asthmatics. Am J Respir Crit Care Med. 1997;156:1274 –1277. Wilson AM, Lipworth BJ. Dose-response evaluation of the therapeutic index for inhaled budesonide in patients with mild-to-moderate asthma. Am J Med. 2000;108:269 –275. Wilson AM, Clark DJ, Devlin MM, McFarlane LC, Lipworth BJ. Adrenocortical activity with repeated administration of once-daily inhaled fluticasone propionate and budesonide in asthmatic adults. Eur J Clin Pharmacol. 1998;53:317–320. Wilson AM, Sims EJ, Struthers AD, Lipworth BJ. Inhaled corticosteroid therapy reduces the early morning peak in cortisol and aldosterone. Clin Sci. 1998;95:513–517. Martin RJ, Szefler SJ, Chinchilli VM, et al. Systemic effect comparisons of six inhaled corticosteroid preparations. Am J Respir Crit Care Med. 2002;165:1377–1383. Affrime MB, Kosoglou T, Thonoor CM, Flannery BE, Herron JM. Mometasone furoate has minimal effects on the hypothalamic-pituitary-
adrenal axis when delivered at high doses. Chest. 2000;118:1538 –1546. 43. Currie GP, Fowler AM, Wilson EJ, Sims L, Orr C, Lipworth BJ. Airway and systemic effects of hydrofluoroalkane fluticasone and beclomethasone in patients with asthma. Thorax. 2002;57:865– 868. 44. Wilson AM, Brewster HJ, Lipworth BJ. Dose-response comparison of systemic bioactivity with inhaled budesonide and triamcinolone acetonide in asthmatic adults. J Allergy Clin Immunol. 1998;102:751–756. 45. Boner AL. Effects of intranasal corticosteroids on the hypothalamicpituitary-adrenal axis in children. J Allergy Clin Immunol. 2001;108: S32–S39. 46. Brutsche MH, Brutsche IC, Munawar M, et al. Comparison of pharmacokinetics and systemic effects of inhaled fluticasone propionate in patients with asthma and healthy volunteers: a randomised crossover study. Lancet. 2000;356:556 –561. 47. Harrison TW, Tattersfield AE. Plasma concentrations of fluticasone propionate and budesonide following inhalation from dry powder inhalers by healthy and asthmatic subjects. Thorax. 2003;58:258 –260. 48. Kelly HW, Nelson HS. Potential adverse effects of the inhaled corticosteroids. J Allergy Clin Immunol. 2003;112:469 – 478. 49. Krasner AS. Glucocorticoid-induced adrenal insufficiency. JAMA. 1999; 282:671– 676. 50. Henzen C, Suter A, Lerch E, Urbinelli R, Schorno XH, Briner VA. Suppression and recovery of adrenal response after short-term, highdose glucocorticoid treatment. Lancet. 2000;355:542–545.
Requests for reprints should be addressed to: John Wlodarczyk, BEc Dip Med Stats, PhD John Wlodarczyk Consulting Services 10 Mirral Crescent New Lambton, NSW 2305, Australia E-mail: [email protected]
ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY