| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Chemical Pathology Department, West Middlesex University Hospital/Quest Diagnostics, Inc., Isleworth, Middlesex, United Kingdom TW7 6AF; Clinical Pharmacology, GlaxoSmithKline Research and Development, Greenford, Middlesex, United Kingdom UB6 0HE; Respiratory Therapeutic Group, GlaxoSmithKline, Uxbridge, United Kingdom UB11 1BT; and Department of Chemical Pathology, University College London Hospitals, London, United Kingdom W1T 4JF
Address all correspondence and requests for reprints to: Dr. R. Fink, West Middlesex University Hospital/Quest Diagnostics, Inc., Chemical Pathology Department, Isleworth, Middlesex, United Kingdom TW7 6AF. E-mail: richard.s.fink{at}questdiagnostics.com.
Abstract
Free cortisol in the urine (UFC) is frequently measured in clinical research to assess whether inhaled corticosteroids (ICS) cause suppression of the hypothalamic-pituitary-adrenal axis. Thirteen healthy male subjects received single inhaled doses (of molar equivalence) of fluticasone propionate (FP), triamcinolone acetonide (TAA), budesonide (BUD), and placebo in this single blind, randomized, cross-over study. UFC output was measured using four commercial immunoassays in samples collected in 12-h aliquots over 24 h. The cortisol production rate was assessed from the outputs of cortisol metabolites. UFC showed a 100% increase over placebo levels in the Abbott TDX assay after the administration of BUD. The other assays detected variable suppression (ranging from 29–61% suppression for FP, 30–62% suppression for TAA, and 25% suppression to 100% stimulation for BUD). Suppression was more pronounced in the first 12 h after TAA and in the second 12 h after FP. Similar suppression was found in each 12-h period after BUD. UFC estimation based on immunoassays after ICS may be an unreliable surrogate marker of adrenal suppression. Many of the published studies describing or comparing the safety of different ICS should be reevaluated, and some should be interpreted with caution.
THE POTENTIAL FOR systemic effects on the hypothalamic-pituitary-adrenal (HPA) axis after exogenous steroids, including intranasal or inhaled corticosteroids (ICS), is assessed frequently in both the clinical setting and research studies. Physicians concerned about adrenal suppression in asthmatic patients receiving ICS have recourse to a number of diagnostic procedures. These include the measurement of plasma cortisol at 0800 h, integrated plasma cortisol concentrations over 24 h, 17-hydroxycorticosteroids (17-OHCS), the measurement of saliva or urinary free cortisol (UFC), and the plasma cortisol response to ACTH (1, 2, 3, 4, 5).
Of these options urinary measurements are the most commonly used because of the noninvasive nature and convenience of sample collection. Although determination of UFC concentrations using automated immunoassays may be convenient, some systems may be subject to interferences from compounds produced by the metabolism of oral steroids or ICS (6, 7, 8, 9, 10). Theoretically, this cross-reactivity could result in falsely elevated results, thereby masking endogenous cortisol suppression. A second drawback is the question of sensitivity, because commercially available immunoassays for urine samples were originally introduced for the detection of cortisol excess (Cushing’s syndrome) and not for the detection of adrenal suppression. Under normal circumstances serum cortisol-binding globulin is saturated and binds most of the circulating cortisol. The fraction of free cortisol filtered by the kidney and excreted in the urine is thus small, and UFC concentrations are inherently low. A small degree of suppression may therefore result in even lower UFC concentrations (<50 nmol/liter urine), at which point the precision of many commercial immunoassays may be unacceptable [for diagnostic purposes a precision <10% is recommended (10)].
Despite these considerations there is no published work in which the effect of commonly used ICS on standard UFC assays has been studied rigorously. Therefore, we have investigated the effects of single doses (near the highest recommended doses) of three inhaled corticosteroids, budesonide (BUD), triamcinolone acetonide (TAA), and fluticasone propionate (FP), on the HPA axis as measured by four different UFC assays that are used frequently in routine and research practice.
In addition, a gas chromatography-mass spectrometric (GC-MS) technique (11) was used as a reference method for estimating adrenal activity as total cortisol metabolites (TCM). TCM includes most cortisol metabolites, approximates to cortisol production rates closer than those achieved through measurement of 17-OHCS. The concentration of TCM in urine is relatively high and can therefore be measured accurately and precisely even in patients with adrenal suppression (12, 13). A second advantage of TCM is that this analysis is not subject to interference from ICS and their metabolites.
The doses of ICS administered were clinically relevant and anticipated to achieve systemic exposure that was likely to have a measurable effect on the HPA axis in healthy subjects. For each ICS, a similar molar dose was administered. These doses were therefore not therapeutically equivalent (relative potency of FP/BUD/TAA, 1:0.5:0.25) (14).
Subjects and Methods
Thirteen healthy male subjects were enrolled in the study with an average age of 24.6 yr (range, 20–34 yr) and body mass index of 25 kg/m2. All subjects gave written informed consent. Each subject had a physical examination and routine hematology, blood chemistry, urinalysis, and drug screen on entry into the study. Adverse events were recorded throughout the study period. One subject withdrew from the study due to personal reasons after TAA and BUD.
The study was approved by the research ethics committee of the South-Eastern Sydney Area Health Service, Australia, and was conducted at the James Lance GSK Medicines Research Unit (Sydney, Australia).
This was a single center, randomized, single blind, four-way cross-over study with washout periods of 1 wk between treatments. Subjects attended the unit on the evening before each treatment and remained in the unit until the end of each treatment period. The baseline 24-h urine collection commenced at approximately 0730 h. The urine was collected in two 12-h aliquots. All urine collections were kept refrigerated between voids. The subjects received single inhaled doses of BUD, TAA, FP, or placebo at approximately 0730 h in random order under supervision. All urine was then collected over the subsequent 24 h. Urine volumes were calculated from the urine weight, and 10-ml aliquots were taken from each 12-h collection, frozen immediately, and stored at -70 C until assayed.
All study drugs were administered according to recommendations via metered dose inhaler (TAA with an integrated spacer). BUD (Pulmicort, AstraZeneca, Wilmington, DE) was administered as 8 inhalations delivering 200 µg ex-valve (172 µg ex-actuator)/actuation. TAA (Azmacort, Aventis Pasteur, Lyon, France) was administered as 16 inhalations delivering 200 µg ex-valve (100 µg ex-actuator)/actuation. FP (Flixotide, GlaxoSmithKline, Middlesex, UK) was administered as 7 inhalations delivering 250 µg ex-valve (220 µg ex-actuator)/actuation. The ex-actuator doses delivered were 1376, 1600, and 1540 µg for BUD, TAA, and FP, respectively. All doses are within the maximum recommended daily dose. These doses are approximately equivalent on a molar basis (3.2–3.7 µmol).
Methods
UFC was determined by four immunoassay methods using protocols recommended by the manufacturers. In each case cortisol was extracted into dichloromethane, and extracts were dried before reconstitution in buffer or zero standards before the assay. UFC was measured by RIA (Corti-Cote, Product 06B-256440, ICN Biomedicals, Inc., Costa Mesa, CA) and automated immunoassays (DPC Immulite, Product LKC05, Diagnostic Products, Los Angeles CA; Abbott TDX, Abbott Laboratories, Abbott Park, IL; Bayer ACS 180, Chiron Corp., Norwood, MA). Assay sensitivities were 4.5, 5.5, 7, and 14 nmol/liter, respectively. Assay performances were continuously checked with quality control of assayed urine control samples (Products AU 2353 and AU 2352, Randox Laboratories Ltd., Ardmore, Crumlin, UK) at two levels (Table 1
). The majority of results from the study were between the quality control levels tested. All samples from each subject were analyzed in one assay batch to minimize subject variations. None of the commercial kits had been tested for cross-reaction with the agents administered. (Conversion from Systeme International units to conventional units is based on the formula: 1 nmol/liter cortisol = 36.1 ng/dl cortisol.)
|
Statistical analysis
Cortisol measurements for each 12-h urine aliquot were summed to give the full 24-h measurement pre- and posttreatment. Results are presented as the geometric mean with 95% confidence intervals (CI). Treatment-related suppression of absolute urinary cortisol or cortisol metabolite excretion relative to placebo was estimated for each ICS according to assay with pairwise comparisons of ICS and placebo. The data were analyzed using two linear models. Absolute UFC and TCM measurements were log transformed and fitted to a linear model, including terms for subject, period, treatment, assay, and treatment by assay interaction. Baseline data were included in the model as a covariate. No adjustments were made for multiple comparisons. All statistical tests were two-sided alternatives with a 5% level of significance. It is common practice to present 24-h UFC as a ratio of creatinine excretion. As the subjects were supervised and resident within a clinical research unit, a complete 24-h urine collection was achieved in this study. However, when this correction was made, no difference in the result was detected for creatinine-corrected data. These data are therefore not presented.
All statistical analyses were carried out using SAS package (version 6.12, SAS Institute, Inc., Cary, NC).
Results
There were no serious adverse events reported during this study, and no subject withdrew due to adverse events.
24-h UFC output before ICS administration
In general, all four UFC assays gave similar UFC per 24 h results at baseline, although the Immulite gave consistently higher values than the Abbott TDX, Bayer ACS 180, and Corti-Cote (Table 2). The measurements were all within the reported normal ranges for UFC (50–300 nmol/24 h) (1). When comparing the two 12-h aliquots, UFC excretion was greater in the samples collected during the day (0730–1930 h) than those collected overnight. This was evident for all four immunoassays and GC-MS (Table 2).
|
No subject had a value below 50 nmol/24 h after placebo as measured by all four immunoassays (Fig. 1). A number of subjects showed UFC values less than 50 nmol/24 h with each ICS, but not with each immunoassay. For example, after treatment no subject showed a value below 50 nmol/24 h with the Immulite system. In contrast, posttreatment UFC fell below this value, although it was still measurable by the Abbott TDX, Bayer ACS, and Corti-Cote methods. With regard to treatment type, subjects receiving FP had UFC values below 50 nmol/24 h in one, six, and two cases (Abbott TDX, Bayer ACS, and Corti-Cote, respectively). After TAA and BUD, this was observed in one, eight, and one subjects and in one, three, and two subjects, respectively (Fig. 1).
|
In addition to the absolute UFC output, the change in 24-h UFC after each ICS was compared with the placebo values. UFC excretion reported after placebo showed no differences from baseline for any of the immunoassays. In keeping with the absolute UFC values, changes in UFC varied depending on the assay method used (Table 3). Most strikingly, after BUD, UFC excretion appeared to increase by 100% compared with placebo in those samples assayed by the Abbott TDX system. This was significantly different from the results of the other three immunoassays (P < 0.001). Furthermore, the anticipated suppression of cortisol excretion after a high dose of BUD was clearly observed with the other immunoassays and varied from -25% to -39%. These were significantly different from placebo (P < 0.01). A significant reduction of UFC was detected after both FP (P 
0.003) and TAA (P 0.002) for all assays used. However, there was considerable variation in the degree of suppression observed, ranging from 29–61% after FP and from 30–62% after TAA depending on the immunoassay (Fig. 2). These differences reached statistical significance in many cases despite the small number of subjects in this study (Fig. 2).
|
|
0.001; Table 3
and Fig. 2). However, this was generally less than that found with the immunologically based assays. Influence of sample collection time (day and overnight) on apparent cortisol suppression
In addition to the 24-h UFC data, cortisol excretion over the 12 h during the day (0730–1930 h) and the 12 h overnight (1930–0730 h) was investigated. The apparent significant increase in cortisol excretion after BUD seen in the Abbott TDX system was most prominent in the first 0–12 h of collection. For the other three immunoassays, there was no obvious difference or pattern in cortisol suppression detected in the two 12-h aliquots. There was a tendency for greater effect on TCM excretion in the overnight collection compared with the first 12 h, although this was not statistically significant (Fig. 3 and Table 4).
|
|
). This was significant for the Bayer ACS 180 and GC-MS methods (Fig. 3). The overnight collection tended to suggest greater suppression than the 24-h collection with the Abbott TDX, Bayer ACS, and GC-MS methods, although this was only significant with the Bayer ACS.
After TAA, in contrast to FP, there was a tendency for less effect on cortisol excretion in the overnight collection compared with the first 12-h collection (although this was not statistically significant). This was particularly apparent with the Abbott TDX and Immulite systems, although the latter was associated with large variability (Fig. 2).
Discussion
The results of this study demonstrate that the assessment of adrenal suppression by measuring UFC is more problematic than previously realized and is markedly influenced by the assay methods used as well as the protocol for urine collection.
One of the most striking observations reported here is the dramatic interference seen with BUD and the Abbott TDX system. With this assay, BUD appeared to cause an increase in cortisol excretion of 100%. This result is likely to reflect cross-reactivity of BUD metabolites with the anticortisol antibody used in the assay. The main BUD metabolites formed in man are 6ß-hydroxybudesonide and 16
-hydroxyprednisolone (15), one or both of which may be detected by the assay. This potential problem of exogenous oral corticosteroid and/or cortisol metabolite interference has been recognized previously (7, 8, 9, 10), and immunoassay kit manufacturers recommend extraction procedures to reduce this problem. Despite pretreatment with dichloromethane, the Abbott assay in particular was subject to interference generated by BUD or its metabolites. Although others have recognized cross-reactions with immunoassays in patients receiving oral prednisolone (7, 8), it is not acceptable to adjust the measured cortisol for the presumed cross reactivity (16). This is the first report showing that a commonly used ICS can perturb assays used in the routine clinical laboratory. The marked interference seen with BUD was not apparent with the other assays used, although it may have been present to a lesser extent. Of the ICS tested, BUD is chemically the closest to prednisolone. Reports in which the effects of BUD on the HPA axis have been evaluated using the Abbott system should be reviewed with caution (17), although the importance of this in serum or plasma has not been evaluated in the present study.
In addition to the interference seen with BUD and the Abbott TDX, this study has clearly demonstrated that the extent of cortisol suppression after ICS administration is influenced by the analytical method. The same urine sample could yield up to a 2-fold difference in observed HPA axis suppression when comparing the immunoassays and up to a 5-fold difference when compared using the reference GC-MS method. Thus, depending on the assay used, entirely different conclusions may be drawn on the potential for and magnitude of HPA effects induced by these ICS. Importantly, this variability was only seen in the presence of ICS. After placebo, no significant difference was detected between the immunoassays. Lack of specificity and poor precision at low concentrations are the likely explanations for this difference. Further, although significant suppression was detected for BUD, FP, and TAA by TCM measurement, the reductions observed were generally lower than those seen with the immunoassays. These data may be more representative of the actual HPA axis effects, as the aforementioned caveats are not applicable.
Comparison of daytime urine collections with overnight collections demonstrates that the timing of urine collection can significantly influence the cortisol suppression detected. The pharmacology and metabolism of the drugs may explain these findings. The interference after BUD in the Abbott TDX system was most prominent in the first 12 h posttreatment urine collection. This supports the conclusion that the apparent increase in UFC is due to metabolite cross-reactivity and is consistent with the short half-life of BUD and its metabolites (18).
TAA produced marked suppression in the first 12-h period, however, particularly when using the Abbott TDX or Diagnostic Products Immulite, but no significant effect was seen in the overnight collection. This is also likely to reflect the 2.5-h half-life of TAA (19), with approximately 97% of the dose being eliminated over the first 12-h collection period. Therefore, an overnight collection after a single morning dose of TAA would be inappropriate when assessing the HPA effects of this ICS. On the other hand, FP has a plasma half-life of 10 h from the metered dose inhaler (20), and in some assays demonstrated more suppression in the 12- to 24-h collection than in the 0- to 12-h collection. Despite these considerations, 12-h overnight collections are frequently used in clinical medicine and clinical research to assess cortisol production. In view of the number of available ICS and their respective pharmacokinetic profiles, collection periods should be chosen to obtain meaningful information in all cases.
Our results have implications beyond patient management. Immunoassays are used extensively in clinical trials to assess and compare the safety of ICS, and a huge body of literature has evolved over the past 10 yr. When evaluating these reports, the reader should consider whether potential cross-reacting substances were effectively removed from the urine, and which immunoassay was selected for the ICS under study. For example, based on our data, estimated 56% and 79% reductions in overnight UFC excretion would be observed in a study in which TAA and FP are compared with samples measured on the Bayer ACS. In contrast, a study using 24-h collections and the same analytical platform would demonstrate reductions of 62% and 61%. The reference assay of TCM indicated suppressions of 34% and 27% for TAA and FP, respectively, in 24-h collections. When examining UFC values below the reference range (50 nmol/24 h used here and other researchers have used similar limits) (21, 22, 23), discrepant results would also be observed. For example, the combination of FP and the Bayer ACS would show UFC suppression in 6 of 12 subjects. BUD and the Bayer ACS would show suppression in 3 of 13. However, with the Corti-Cote system, these identical samples would show suppression in 2 of 12 and 2 of 13 cases, respectively. Clearly, the selected assay significantly influences the outcome and conclusions in such studies. The findings of our study may go some way to explain the discrepancies in exogenous steroid (inhaled and intranasal) HPA axis effects described in the literature, although unfortunately the details of assays and sample treatment are often omitted in published papers on this topic.
In summary, our study shows that UFC results in subjects receiving ICS can be determined by variables within the methodology. These variables exert such powerful effects that suppression or, indeed, normal function can be diagnosed with equal facility in the same individual depending on the assay and collection protocol used. This has critical implications for patient management as well as for clinical research, where conclusions about ICS safety may need to be reevaluated because of issues raised in this study.
Acknowledgments
Richard Hodkinson, Nina Joseph, and Manesha Patel performed the steroid assays.
Footnotes
Abbreviations: BUD, Budesonide; CI, confidence interval; FP, fluticasone propionate; GC-MS, gas chromatography-mass spectrometry; HPA, hypothalamic-pituitary-adrenal; ICS, inhaled corticosteroids; 17-OHCS, 17-hydroxcorticosteroids; TAA, triamcinolone acetonide; TCM, total cortisol metabolites; UFC, urinary free cortisol.
Received February 22, 2002.
Accepted July 19, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  M. Masoli, M. Weatherall, S. Holt, P. Shirtcliffe, and R. Beasley Inhaled fluticasone propionate and adrenal effects in adult asthma: systematic review and meta-analysis Eur. Respir. J., November 1, 2006; 28(5): 960 - 967. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. K. Jerjes, N. F. Taylor, T. J. Peters, S. Wessely, and A. J. Cleare Urinary Cortisol and Cortisol Metabolite Excretion in Chronic Fatigue Syndrome Psychosom Med, July 1, 2006; 68(4): 578 - 582. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. P. Chrousos, L. Ghaly, A. Shedden, D. G. Iezzoni, and A. G. Harris Effects of Mometasone Furoate Dry Powder Inhaler and Beclomethasone Dipropionate Hydrofluoroalkane and Chlorofluorocarbon on the Hypothalamic-Pituitary-Adrenal Axis in Asthmatic Subjects Chest, July 1, 2005; 128(1): 70 - 77. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Pescollderungg, D. G. Peroni, A. Pietrobelli, and G. Radetti Inhaled Corticosteroids and Urinary Free Cortisol Pediatrics, December 1, 2003; 112(6): 1464 - 1465. [Full Text] [PDF]
|
|
|
|
 |
|
 L Pescollderungg, G Radetti, E Gottardi, D G Peroni, A Pietrobelli, and A L Boner Systemic activity of inhaled corticosteroid treatment in asthmatic children: corticotrophin releasing hormone test Thorax, March 1, 2003; 58(3): 227 - 230. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
