INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The benefits of inhaled glucocorticoids in the management of asthma are well established, but concern about potential systemic adverse effects, including hypothalamic-pituitary-adrenal (HPA) suppression, has focused much attention on the dose-response relationships for efficacy and side effects with different corticosteroid preparations (1). Fluticasone propionate (FP) is a new, highly potent glucocorticoid which, via pressurized metered-dose inhaler (pMDI), appears to be twice as effective (on a microgram-for-microgram basis) as beclomethasone dipropionate (BDP) (2) and budesonide (BUD) (3) in patients with moderate-to-severe asthma. Whether FP has a more favorable risk-benefit profile compared with those of BDP and BUD depends to a large extent on the corresponding potency ratio for its systemic activity, e.g., adrenal suppression and effects on bone turnover and growth retardation. Data from animal models suggest that FP has a topical-to-systemic potency ratio of 25.0 compared with 1.0 for BUD and 0.1 for BDP (1), but the systemic potency in humans, relative to BDP or BUD, has not been clearly established, mainly because of incomplete characterization of the full dose-response relationships in randomized comparative studies.

With regard to adrenal suppression, it has been suggested that FP has a better safety profile than do BDP and BUD based on early morning cortisol levels in a clinic population (5, 6), but recent controlled studies using more accurate measurements of HPA function have shown that FP has greater systemic activity than previously suspected (7), with clinically significant systemic effects at higher doses (> 1,000 µg/d) (10). Several important methodologic factors have limited the interpretation of previous studies investigating adrenal suppression with inhaled corticosteroids: for example, the timing and frequency of blood sampling can affect the conclusions (11); single-dose effects may not apply to chronic administration; and most studies, by including only one dose level for each drug, provide little or no information about dose-response relationships over the range of doses used in clinical practice.

The primary aim of this study was to characterize dose- response relationships for the effects of FP and BUD on diurnal cortisol production, and thereby obtain a quantitative estimate of the dose-potency ratio (FP:BUD) for adrenal suppression. In seeking to avoid some of the limitations of previous study designs, we assessed the short-term effects of three doses of FP and three doses of BUD on adrenal function (the integrated 24-h profile of plasma cortisol concentration) in healthy male volunteers using a randomized, placebo-controlled, crossover comparison of seven 5-d treatment periods.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Twenty-nine healthy male volunteers 18 to 35 yr of age gave written informed consent to participate in this study, which was approved by the institutional Ethics Committees of Royal Prince Alfred Hospital and St. Vincent's Hospital, Sydney. At a preliminary screening visit, each subject had a physical examination and routine blood tests, and was trained in the use of a pMDI. Any volunteer with a significant laboratory abnormality or past medical history, including asthma, those receiving regular medication of any kind, and any subject unable to correctly use a pMDI, were excluded from participation.

Study Design

This was an open-label, randomized, placebo-controlled seven-period, crossover study to compare the effects of 5 d of treatment with FP (750, 1,500, and 2,000 µg/d by pMDI) and BUD (800, 1,600, and 3,200 µg/d by pMDI) on the integrated area under the curve (AUC) of 24-h plasma cortisol profiles. Each treatment was separated by a 10-d washout period.

On the first day of each 5-d treatment period, subjects attended the research center and were required to practice their inhalation techniques using a Vitalograph compact recorder (Vitalograph Ltd, Buckingham, UK). Each treatment (BUD, FP, and placebo) was administered as a twice-daily regimen (10:00 A.M. and 10:00 P.M.) for 5 d, i.e., FP: 375, 750, and 1,000 µg twice a day versus BUD: 400, 800, and 1,600 µg twice a day versus placebo twice a day, and for the first 4 d, while volunteers were ambulant at home, compliance with each dose was confirmed by telephone contact with the investigators. Budesonide (Pulmicort^TM) and placebo pMDIs were supplied by 3M Health Care Ltd (UK) and FP (Flixotide^TM) pMDIs were manufactured by Allen and Hanburys Ltd (UK). On each dosing occasion, subjects were instructed to shake and prime the pMDI by firing one puff into the air prior to taking the requisite dose and rinsing the mouth with water after completing their inhalations.

For the last 24 h of each 5-d treatment period (i.e., 10:00 P.M. on Day 4 to 10:00 P.M. on Day 5), subjects were admitted to the research center for blood sampling and direct supervision of the last two doses of medication. On each occasion, an indwelling cannula was inserted into a forearm vein, and 5-ml blood samples were collected at 2-h intervals for 24 h. The samples at 10:00 P.M. on Day 4 and at 10:00 A.M. on Day 5 were collected immediately prior to administration of the scheduled dose. The blood was centrifuged within 30 min of collection, and the plasma samples were stored at 80° C until assayed.

Measurement of Plasma Cortisol

Plasma cortisol concentrations were measured using an automated system based on a solid-phase chemiluminescent enzyme immunoassay (IMMULITE System; Diagnostic Products Corp., San Diego, CA). All the samples for each subject were measured in duplicate in the same assay run, and the lower limit of detection was 5.5 nmol/L. In instances where the measurement was below the limit of detection, a value of 2.75 nmol/L (i.e., half the detection limit) was entered as the result. The intra-assay and interassay coefficients of variation were 6.0 and 8.6%, respectively, in the range of 300 to 500 nmol/L, and there was negligible cross-reactivity between the IMMULITE antibody and the study drugs.

Statistical Analysis

The primary efficacy variables were: (1) the integrated AUC of plasma cortisol levels during the last 24 h of each treatment period (AUC_24h) calculated using the trapezoidal rule, and (2) the plasma cortisol concentration at 8:00 A.M. on Day 5 of each treatment period. These variables were analyzed using ANOVA, with subject, treatment, and period as factors. A multiplicative model was used, i.e., data were log-transformed before analysis. The possibility of carryover effects was explored by including an additional factor for the effect of treatment in the previous period. The results are presented as the ratio of the geometric means, together with 95% confidence intervals (95% CI) and the p values for each comparison. Statistical significance was defined as p < 0.05.

Estimation of Relative Systemic Dose Potency

Dose-response relationships for the placebo-subtracted effects of FP and BUD on each of the primary efficacy variables, i.e., AUC_24h and 8:00 A.M. plasma cortisol concentration, were derived by fitting a nonlinear (sigmoidal) relationship to the individual subject data on a logarithmic scale using the following equation (12):

where response is the cortisol value as a percentage of placebo. The dose potency ratio for FP relative to BUD is a model-derived estimate of the degree of leftward displacement of the fitted FP log-dose- response curve relative to the BUD log-dose-response curve. A nonlinear mixed effect model (13) was used to derive the dose potency ratio (), as described previously with inhaled corticosteroids (14), with an assumption that the parameters Emax (placebo level), b and a are individual values normally distributed in the population. Mathematically,  is defined as the exponential of the difference in parameter a obtained with FP and BUD. However, because  is a model-derived mean parameter for the group, it is expressed with 95% CI as an estimate of relative dose potency.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Twenty-nine subjects were randomized, and data for 28 were included in the final analysis (mean age, 24 ± 4 yr and BMI, 23 ± 2 kg/m²). One subject withdrew during the first treatment period for reasons that were unrelated to the study. Inhalation technique was verified at the start of each treatment period using a Vitalograph recorder, which confirmed good intrasubject reproducibility for use of the pMDI. There were no carryover effects identified in any of the analyses.

24-Hour Profiles of Plasma Cortisol Concentration

Compared with placebo, BUD and FP caused dose-dependent suppression of plasma cortisol levels at all time points throughout the 24-h period at the end of each 5-d treatment cycle, with the two highest doses of FP (1,500 and 2,000 µg/d) causing the greatest overall reduction in AUC_24h (Table 1). BUD 800 µg/d was the only treatment not associated with a statistically significant reduction in AUC_24h compared with placebo (Table 1).

BUD 800, 1,600, and 3,200 µg/d caused percentage reductions in AUC_24h (relative to placebo) of 23% (p = 0.21), 41% (p < 0.01), and 69% (p < 0.0001), respectively, whereas the corresponding doses of FP (750, 1,500, and 2,000 µg/d) produced reductions of 46% (p < 0.004), 85% (p < 0.0001), and 93% (p < 0.0001) (Figure 1). At higher doses, FP 1,500 µg/d produced a significantly greater reduction in AUC_24h compared with more than twice the dose of BUD (3,200 µg/d, p < 0.001) (Table 1).

View larger version (24K): [in this window] [in a new window]   Figure 1.   AUC24h values for each treatment expressed as a percentage relative to AUC24h (placebo); p values for between treatment comparisons are shown in Table 1.

The fitted mean dose-response curves for the (placebo-subtracted) effects of BUD and FP on AUC_24h are shown in Figure 2, top panel. There were large intersubject variations in the effects of each drug, especially with the two higher doses of FP, but the model-derived parameters showed that FP was 2.9 times more potent than BUD in suppressing plasma cortisol concentrations over a 24-h period (95% CI, 2.5 to 3.5).

View larger version (10K): [in this window] [in a new window]   Figure 2.   Mean fitted dose-response curves for the effects of BUD (solid line) and FP (dashed line) on AUC24h (upper panel ) and 8:00 A.M. plasma cortisol concentration (lower panel ).

8:00 A.M. Plasma Cortisol Concentrations

There was a clear trend towards dose-dependent reductions in 8:00 A.M. plasma cortisol concentrations on Day 5 of each treatment period (Figure 3), but only the highest dose of BUD and the two higher doses of FP (1,500 and 2,000 µg/d) produced significant reductions in early morning cortisol levels compared with placebo (p < 0.05). For example, the three doses of BUD (800, 1,600, and 3,200 µg/d) produced average reductions of 10, 30, and 66%, respectively, whereas the corresponding dose-dependent effects of FP were 38, 88, and 94%.

View larger version (14K): [in this window] [in a new window]   Figure 3.   Individual values and geometric mean (n = 28) for 8:00 A.M. plasma cortisol concentrations on Day 5 of each treatment period. Horizontal dotted line shows the lower limit of the normal range for 8:00 A.M. plasma cortisol (200 nmol/L); mean values for each dose are joined with a solid line, and the standard deviation is expressed as a rectangular box.

The fitted mean dose-response curves for the effects of BUD and FP on 8:00 A.M. plasma cortisol concentrations are shown in Figure 2, bottom panel. The intersubject variation in early morning cortisol levels was greater than for measurements of AUC_24h, but the model-derived parameters gave a similar estimate of 3.1 (95% CI, 2.4 to 4.0) for the dose potency of FP, relative to BUD, in suppressing 8:00 A.M. plasma cortisol concentrations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has shown that in normal healthy volunteers both BUD and FP cause dose-dependent reductions in AUC for the plasma cortisol concentration-time profile during the last 24 h of a 5-d treatment period. In contrast to some previous studies that have relied on acute dosing and/or single (early morning) blood samples (5, 6, 9, 15), the dose-response relationships characterized in this study reflect steady-state drug administration and take full account of placebo effects and diurnal variations in cortisol production. We deliberately chose relatively high doses of FP and BUD so that the dose-response information would be especially relevant to patients with moderate to severe asthma in whom maximal doses of inhaled corticosteroids are often required (16) and risk-benefit considerations are most contentious (1). Only one of the six treatments (BUD, 3,200 µg/d) exceeded the recommended dose range for clinical practice, but the higher range of doses chosen for this study has usefully extended the dose-response relationships shown in previous studies (14, 17) and confirmed that dose-dependent effects on adrenal suppression show no sign of approaching a plateau (i.e., Emax), even with the maximal recommended dose of FP (2,000 µg/d) or a supramaximal dose of BUD (3,200 µg/d is higher than the maximum recommended dose of BUD).

Clinical and laboratory studies comparing the anti-inflammatory effects of different corticosteroids have shown that on a microgram-for-microgram basis FP is approximately twice as effective as BUD and BDP (2, 18), but the potency ratio of FP:BUD with respect to systemic effect on cortisol secretion requires further investigation.

Data from animal models have shown that FP has a topical-to-systemic potency ratio 25 times greater than BUD (19, 20), but the results of this study in normal volunteers confirm and extend recent reports showing that the systemic potency of FP has probably been underestimated (7). Individual dose- response data were fitted to a mixed effect model to quantitatively assess the relative dose potency of the two drugs. The results were similar for each of the primary efficacy variables, AUC_24h and 8:00 A.M. plasma cortisol, and showed that FP, relative to BUD, is 2.9 times more potent in suppressing AUC_24h and 3.1 times more potent in lowering 8:00 A.M. plasma cortisol levels after 5 d. Comparable dose potency ratios of 3.7:1 and 5.2:1 for AUC_24h and early morning cortisol measurements, respectively, were reported in a similar study using lower doses of FP and BUD for 4 d (14). With data for 28 subjects and a placebo-controlled crossover design, the present study was able to calculate the dose potency ratios with greater precision.

Pharmacokinetic as well as pharmacodynamic factors might contribute to the results of this study. For drugs such as FP and BUD, which are subject to extensive first-pass hepatic metabolism, lung deposition and pulmonary absorption are the primary determinants of total systemic bioavailability (21). A number of factors affect lung deposition of inhaled steroids, including the type of inhaler device (e.g., dry powder inhaler versus pMDI) (22) and inhaler technique. The present study used similar pMDIs to deliver both drugs, and volunteers employed mouth rinsing to minimize absorption from the oropharynx and gastrointestinal tract, but we did not measure lung deposition.

From this study in which subjects were treated for a short period of 5 d, it is difficult to predict the long-term effects of FP and BUD on plasma cortisol concentrations. However, with the highest inhaled doses there was profound suppression of the AUC_24h in several subjects, suggesting that if these daily doses were continued for prolonged periods there would be failure of HPA function for considerable time after discontinuing therapy.

It is unclear whether the results of this study in normal volunteers are applicable to patients with asthma. On theoretical grounds, lung deposition, and hence lung bioavailability, might be less in patients with airflow obstruction and consequently limited drug delivery to the pulmonary vascular bed. Indeed, Melchor and coworkers (23) showed that lung deposition using a pMDI is 1.5-fold higher in normal subjects than in asthmatic patients. Although the systemic activities of FP and BUD are clearly dependent upon lung bioavailability, there is nothing to suggest that in asthmatic subjects pulmonary deposition of either drug would be differentially affected. Thus, although the absolute effects on cortisol secretion might be less in asthmatic patients, it seems unlikely that the dose potency ratios (FP:BUD) derived in this study would be any different. Indeed, the results of this study are broadly similar to those recently reported in adults (8, 17) and children (24) with asthma.

The effects of inhaled corticosteroids on plasma cortisol are mediated via steroid receptors in the hypothalamus and pituitary gland. The present study has determined a potency ratio for inhaled BUD and FP for effects mediated via these receptors. It is not known whether the same potency ratio pertains to other side effects of higher inhaled doses such as changes in bone metabolism. Initial studies, using sensitive plasma markers of collagen turnover, have indicated that inhaled BUD and FP paradoxically reduced bone resorption (25). Further studies will be necessary to determine potency ratios of inhaled corticosteroids on potentially important target tissues such as bone.

In summary, this dose-response analysis of 28 subjects who inhaled three different doses of BUD and FP on a twice-daily regimen for 5-d via pMDIs showed that FP was 2.9 times more potent than BUD in reducing AUC_24h and 3.1 times more potent in reducing 8:00 A.M. plasma cortisol concentrations. Thus, on a microgram notional dose basis, the systemic effects of a given dose of inhaled FP on plasma cortisol measurements were equivalent to the effects of three times the dose of inhaled BUD.