INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Despite controversy about the use of steroids in the emergency room (ER) (1, 2), most available data support a corticosteroid benefit. This fact has been confirmed by a meta-analysis (3) in which more than 700 articles were reviewed to identify 30 randomized, controlled trials appropriate for analysis. The investigators concluded that steroids significantly reduce the rates of admission and the number of future relapses, as well as improve pulmonary function, albeit slowly (after 6 to 12 h). An abrupt increase in lung function early in the course of treatment has never been found with any dose of steroids despite being repetitively sought (4). However, inhaled steroids are rarely mentioned, and aerosol glucocorticoid treatment has been considered as not effective in acute severe asthma (5).

We designed a randomized, placebo-controlled study to determine the benefit of adding high and cumulative doses of flunisolide, administered by metered-dose inhaler with spacer, to salbutamol in the treatment of patients with acute severe asthma. Flunisolide is a halogenated glucocorticoid with relatively strong topical anti-inflammatory effects. This drug was selected for the study because it possesses a shorter half-life (1.6 h) and a higher degree of first-pass metabolism (8).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We evaluated all adult patients with acute asthma who were seen in the ER of the Military Hospital in Montevideo, Uruguay, over a 6-mo period. The inclusion criteria for patients were: (1) age between 18 and 50 years, (2) FEV₁ and peak expiratory flow rate (PEFR) less than 50% of predicted value, (3) patients with a history of chronic cough, cardiac, hepatic, renal, or other medical diseases, or pregnancy, were excluded, and (4) an expressed willingness to participate in the study, with written informed consent obtained. The study was approved by the Hospital Ethics Committee.

Subjects were entered, in a double-blind, randomized manner, into one of two groups. Both groups received salbutamol (Ventolín; Glaxo-Welcome, London, UK) delivered by a metered-dose inhaler into a spacer devise (Volumatic; Allen & Hanburys Ltd, Greenford, UK) in a dose of four puffs at 10-min intervals (100 µg per actuation). In one group of patients (the flunisolide group), flunisolide (Flunitec; Boehringer Ingelheim KG, Ingelheim am Rhein, Germany) delivered by metered-dose inhaler into a spacer in four puffs at 10-min intervals (250 µg per actuation) was administered after salbutamol treatment. The second group (placebo group) received four puffs from an identical and specially prepared metered-dose inhaler containing only propellant at 10-min intervals. Each puff was followed by a deep inhalation from the spacer. The protocol involved 3 h of this treatment (2,400 µg of salbutamol and 6 mg of flunisolide each hour). The randomization list was kept concealed from the researchers, assistants, and emergency physicians, and all measures were made by investigators unaware of the patients' group assignment. Intravenously administered aminophylline was excluded in all patients. Additionally, each patient was given O₂ at a rate of 4 L/min.

The following variables were measured in each patient immediately before starting treatment and at 30-min intervals for 3 h after presentation: FEV₁, PEFR, respiratory rate, heart rate, accessory-muscle use, dyspnea, and wheezing. Additionally, serum theophylline concentrations were determined from all subjects' pretreatment. PEFR was measured with a mini-Wright peak-flow meter (Armstrong Industries, Inc., Northbrook, IL). The highest of three values was recorded. FEV₁ was measured using a Vitalograph Compact spirometer (Vitalograph Ltd, Maids Moreton House, Buckingham, UK). Three successive maximal expiratory curves were recorded at each assessment, and the highest value was selected according to the criteria of the American Thoracic Society (9). Heart rate was measured on a continuous electrocardiogram. Accessory-muscle use was defined as visible retraction of the sternocleidomastoid muscles (10). Dyspnea was defined as the patient's own assessment of breathlessness. Wheezing was defined as musical or whistling breath sounds heard with a stethoscope during expiration. These clinical factors were graded in a scale from 0 to 3 in which 0 denoted absent, 1 mild, 2 moderate, and 3 severe. We defined a clinical index as the average of the three measures at 30-min intervals (range, 0 to 3). At the end of the therapy, the patient was asked to indicate the presence or absence of each of five symptoms (nausea, palpitations, tremor, anxiety, and headache). To compare side effects in each group, we compiled a composite symptom score for each subject by arbitrarily assigning each symptom a value of 1 if present and 0 if absent. The decision to discharge or admit a patient was made by senior ER staff without knowledge of previous patient group allocation. Patients with an incomplete response (persistent wheezing or dyspnea, FEV₁ > 40% and < 70% of predicted) required continued treatment in the ER. On the other hand, patients with poor response after 3 h of treatment (marked wheezing and shortness of breath, accessory-muscle use, and a FEV₁  40% of predicted) were admitted. The physicians prescribed oral prednisone (40 mg for 7 d) for all discharged patients, or beta-agonists and intravenous steroids for those who were admitted. Also, to determine previous duration of attack, patients were asked to indicate the onset of wheeze, cough, shortness of breath, or some combination of these symptoms; also, a decline in the PEFR, if available, was considered. When it was possible, patient's relatives were asked to confirm patient's information.

An improvement of 0.25 L in FEV₁ was considered to be clinically important. In a previous study (11), we could estimate the mean (± SD) final FEV₁ value (expressed in liters) to be expected at 3 h was 1.52 ± 0.48 L. Hence, to detect a difference of 0.2 L in FEV₁ between the two groups with a two-sided significance level of 0.05 and a power of 80% a total of 91 subjects would be needed. All data were analyzed with a "SPSS PC plus" software package (SPSS Inc., Chicago, IL). Changes in PEFR, FEV₁, heart rate, and clinical index were evaluated using repeated measures analysis of variance, with one between-subject factor (placebo-flunisolide) and one within-subject factor (time). One-way repeated measures analysis of variance (ANOVA) was used to compare baseline values for each variable. After a significant ANOVA, the Newman-Keul multiple-comparison method was used to test equality of means. Baseline data of the two treatments were compared using Student's t test for normally distributed independent samples or the Mann-Whitney U test for nonnormally distributed continuous variables. Chi-square with Yate's correction was used for categorical variables (12). A p value of less than 0.05, using a two-tailed test, was taken as being of significance for all statistical tests. Means are reported with standard error of the mean (SEM) in the text.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One hundred twenty-four patients who met the criteria for asthma of the American Thoracic Society (13) were assessed in the ER. Of these, 30 did not fit the inclusion criteria for the study because they did not meet the age requirement (12 patients) or FEV₁ requirement (10 patients), had cardiac, hepatic, or renal disease (six patients), or were suspected of being pregnant (two patients). Of the remaining 94 patients (mean age ± SEM, 32.4 ± 0.98), 47 were randomly assigned to the flunisolide group and 47 to the placebo group. All patients in both groups completed the protocol. It can be seen in Table 1 that all baseline demographic and clinical variables were evenly distributed between the two groups. There were no significant differences between the groups for the characteristics examined.

The relationships between the cumulative doses of salbutamol and the change in the PEFR and FEV₁ were analyzed (Figures 1 and 2).

View larger version (46K): [in this window] [in a new window]   Figure 1.   PEFR values (percent of predicted) after administration of placebo or flunisolide. Data points are mean values, and the brackets represent 1 SEM: *p = 0.02, **p = 0.003, ***p = 0.004.

View larger version (43K): [in this window] [in a new window]   Figure 2.   FEV1 values (percent of predicted) after administration of placebo or flunisolide. Data points are mean values, and the brackets represent 1 SEM: *p = 0.02, **p = 0.007, ***p = 0.002.

The magnitude of PEFR improvements over baseline values in both the placebo and the flunisolide groups were significant at all times of treatment (p < 0.01 by one-way repeated measures ANOVA). The two-way repeated measures ANOVA showed a significant difference between groups (p = 0.01); compared with the placebo group, the flunisolide group had better PEFR at 90 (p = 0.02), 120 (p = 0.003), 150 (p = 0.004), and 180 (p = 0.002) min (all by the Newman-Keul multiple comparison method). Also, the two-way ANOVA suggested that difference between groups increased with time (p = 0.0001). The mean PEFR values at 30, 60, 90, 120, 150, and 180 min were 42.2 ± 1.91% (210 ± 10.1 L/min), 46.8 ± 2.08% (232.8 ± 10.8 L/min), 49.0 ± 2.24% (243.9 ± 12.0 L/min), 49.8 ± 2.15% (249.7 ± 12.0 L/min), 51.7 ± 2.27% (258.9 ± 11.6 L/min), and 52.5 ± 2.44% (264.5 ± 12.4 L/min), respectively, in the placebo group, and 46.2 ± 1.95% (231.4 ± 9.67 L/min), 52.1 ± 2.51% (261.7 ± 12.2 L/min), 56.5 ± 2.42% (282.5 ± 12.1 L/ min), 59.9 ± 2.58% (299.6 ± 13.2 L/min), 61.6 ± 2.48% (306.8 ± 12.7 L/min), and 64.0 ± 2.73% (315.7 ± 13.1 L/min) in the flunisolide group. The same pattern held for changes in FEV₁. The improvement over baseline values was significant in both groups (p < 0.01). The repeated measures ANOVA presented a significant difference between groups (p = 0.04); the flunisolide group showed higher rates of improvement than did the placebo group at 120 (p = 0.02), 150 (p = 0.007), and 180 (p = 0.002) min (all by the Newman-Keul multiple comparison method). The mean percent predicted FEV₁ values at 30, 60, 90, 120, 150, and 180 min were 38.8 ± 2.26% (1.32 ± 0.06 L), 43.4 ± 2.51% (1.36 ± 0.07 L), 44.6 ± 2.75% (1.37 ± 0.07 L), 44.9 ± 2.78% (1.36 ± 0.07 L), 45.9 ± 2.68% (1.38 ± 0.07 L), 46.6 ± 2.80% (1.41 ± 0.07 L), respectively, in the placebo group, and 41.5 ± 2.00% (1.26 ± 0.05 L), 46.0 ± 2.16% (1.41 ± 0.06 L), 50.9 ± 2.54% (1.54 ± 0.07 L), 53.2 ± 2.44% (1.62 ± 0.07 L), 56.4 ± 2.68% (1.70 ± 0.08 L), and 58.7 ± 2.78% (1.77 ± 0.08 L) in the flunisolide group. Also, difference between groups increased with time (p = 0.0001).

Patients were divided into four subgroups according to the duration of their asthma attacks prior to ER presentation (patients with  24 h versus patients with < 24 h): placebo < 24 h, placebo  24, flunisolide < 24 h, and flunisolide  24 h. The four subgroups were equally distributed in age, weight, height, respiratory and heart rates, FEV₁, index symptoms, and serum theophylline levels at entry into the study (by one-way ANOVA). Furthermore, there were no significant differences between the subgroups concerning previous asthma medications usage (within 24 h of the ER visit) (Table 2).

The two-way repeated measures ANOVA showed FEV₁ values (as percent of predicted) significantly different between groups (p = 0.038) (Figure 3) at 120, 150, and 180 min (p < 0.05); the placebo  24 h subgroup produced a significantly lower FEV₁ than did the remaining subgroups (p = 0.041 by the Newman-Keul multiple comparison method).

View larger version (17K): [in this window] [in a new window]   Figure 3.   FEV1 (percent of predicted) according to duration of attack before ER presentation. *p < 0.05.

The repeated measures ANOVA showed a significant clinical index group effect (p = 0.04). When the index was studied in the four subgroups of patients divided according to the duration of their attacks prior to the ER visit, once more the placebo  24 h subgroup presented a significantly higher score than the other subgroups at 120, 150, and 180 min (p = 0.04 by the Newman-Keul multiple comparison method) (Figure 4).

View larger version (16K): [in this window] [in a new window]   Figure 4.   Clinical index according to duration of attack before ER presentation. *p < 0.05.

There was no significant difference between placebo and flunisolide groups in heart rate (p = 0.4 by repeated measures ANOVA). The baseline, 1-h, and 3-h mean heart rates were 103.4 ± 2.43, 100.3 ± 2.26, and 104.2 ± 2.43 in the placebo group and 101.6 ± 2.58, 100.9 ± 2.27, and 103.5 ± 2.51 in the flunisolide group. At the end of treatment, patients in the placebo group had a mean increase in heart rate of 1.12 ± 1.92 and those in the flunisolide group had a mean increase of 1.81 ± 1.91 from baseline values. Despite the continuous ECG recording, there were no signs of arrhythmia.

The overall symptom score in the placebo group patients (0.84 ± 0.13) did not differ from the score in the flunisolide group patients (0.96 ± 0.12) (p = 0.5 by chi-square test). Tremor was the most frequently observed effect (53.1% of patients in the placebo group and 37.5% in the flunisolide group). Placebo group patients reported headache (18.7%), anxiety (12.5%), palpitations (9.37%), and nausea (3.12%). The rates for these side effects in the flunisolide group were 21.8%, 3.12%, 12.5%, and 6.25%, respectively. No patient in the flunisolide group had clinical deterioration in the ER, and all subjects in both groups completed the protocol.

After the end of protocol (3 h) 25.5% (n = 12) of patients in the placebo group and 8.51% (n = 4) of patients in the flunisolide group were admitted (p = 0.09 by chi-square test). When admissions were considered in relation to the duration of attack prior to ER treatment, the placebo  24 h showed the higher rate: placebo < 24 h, 17% (n = 2); placebo  24 h, 58% (n = 7); flunisolide < 24 h, 8% (n = 1); and flunisolide  24 h, 17% (n = 2) (p = 0.001 by chi-square test) (Figure 5). The placebo  24 h group produced significantly higher admissions than the flunisolide  24 h (p = 0.005). On the contrary, the placebo < 24 h group did not differ from the flunisolide < 24 h group (p = 0.68).

View larger version (26K): [in this window] [in a new window]   Figure 5.   Percent of admissions according to duration of attack before ER presentation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this randomized, controlled trial, we demonstrated a significant advantage in the overall maximal bronchodilator response when high doses of flunisolide and salbutamol were combined in the emergency treatment of acute severe asthma. The improvement associated with flunisolide was reflected in higher bronchodilator responses, lower clinical ratings, a lower rate of hospital admissions, and minimal side effects. Previous research found that oral or parenteral steroids given in moderate or high doses reduce admissions and improve pulmonary function (3, 4). However, these effects require at least 6 to 12 h to occur. On the contrary, our findings showed that inhaled flunisolide administered in high doses produces therapeutic effects as soon as 2 h after ER presentation. Although a comparison between inhaled and systemic steroids needs to be made, research has shown that a proper evaluation of the value of steroid therapy on pulmonary function necessitates an observation period sufficiently long to ensure the appearance of an effect. In a recent meta-analysis (3), only one article reported significant FEV₁ improvements 6, 12, and 24 h (differences of 10, 16, and 9 in percent predicted FEV₁). On the other hand, we obtained significant FEV₁ differences at 90, 120, 150, and 180 min of 6, 9, 11, and 12 in percent predicted, respectively. Furthermore, we obtained greater differences when patients were divided into four groups.

When patients were divided according to the duration of their attacks prior to presentation, data showed that only in the  24 h subgroup did the flunisolide action produce advantageous results. Because this subgroup probably represented patients with a high level of airway inflammation, we conclude that high doses of inhaled flunisolide can be very useful, specifically in patients with acute asthma and a prolonged duration of symptoms before the ER presentation. Although it is difficult to ascertain by history the exact onset of an asthmatic attack, previous duration of attacks was similar to those obtained in our prior research with the same population, indicating data accuracy (14, 15).

In terms of molecular mechanisms, glucocorticoids are thought to act by binding to a glucocorticoid receptor presentation in the cytoplasm, regulating as well the production of RNA and proteins (16, 17). In view of these basic mechanisms, the slowness of their therapeutic actions is not surprising (18). However, the rapid response found in our trial suggests that there is another mechanism. Because of their topical anti- inflammatory effect, corticosteroids can reduce the airway mucosal edema and thickening that occur in asthma. Cutaneous blanching has been used to measure the corticosteroid activity of topical preparations, and more recently glucocorticoid resistance (19). This blanching is seen within 2 h and is often measured at 6 to 7 h. Similarly, Gibson and colleagues (20) found at 6 h after a single dose of 2.4 mg of inhaled budesonide, a significant change in the airway responsiveness (PD₂₀ budesonide 3.0 ml versus PD₂₀ placebo 1.35 ml), a difference in airway eosinophils of 12%, and an increase of lung function (a 4% improvement in FEV₁).

The demonstration that flunisolide has a therapeutic effect at 2 h in adult patients with acute asthma is consistent with studies in animals (21, 22) and in humans. Recently, in a double-blind, placebo-controlled trial, Scarfone and coworkers (23) showed that inhaled dexamethasone was as effective as oral prednisone in the ER treatment of children with acute asthma, and it was associated with a significantly greater proportion of discharges within 2 h. Husby and coworkers (24) observed a significant improvement 2 h after administering 2 mg of nebulized budesonide to children hospitalized for croup. Additionally, two recent studies (25, 26) concluded that 2 mg of nebulized budesonide leads to prompt (2-h) and important clinical improvement in children with mild-to-moderate croup who come to the ER. On the other hand, Guttman and coworkers (7) showed that inhaled beclomethasone added to the standard regimen of intravenously administered methylprednisolone and beta-agonists did not further improve flow rates or dyspnea scores measured for as long as 12 h after presentation to the ER (7). An important difference with these studies was that we administered flunisolide repeatedly and frequently during the management of acute asthma, rather than once, as in the above studies. Although very large doses of glucocorticoids have not shown to be more efficacious than the smaller ones (4), this conclusion arises from studies where steroids were administered parenterally. Therefore, a topical dose-response effect cannot be rejected.

Because the flunisolide doses were so large, it could be possible that the response was due to systemically absorbed drug rather to than a topical effect. Nevertheless, serum flunisolide levels after the administration of 2 mg through a metered-dose inhaler ranged between 0.9 and 1.7 µg/ml (personal communication, Boehringer, Ingelheim). Assuming a linear relationship between dose and serum levels, the final flunisolide level after 3 h of treatment would be approximately 12 µg/ml.

The rationale that we used for choosing a very high dose of flunisolide was previous data showing the efficiency of deposition of radiolabeled aerosols to the lower respiratory tract; in nonintubated patients it has been reported to range from 11 to 15% with metered-dose inhaler and spacer (27). Higher doses may be important because the unpredictability of delivery systems, low tidal volumes, variable flow rates, increased frequency of breathing, peripheral dispersion of the inhaled medication, and narrowed airways (28). Additionally, there is evidence that patients respond to increasing doses of inhaled steroids (29).

The recovery pattern of the placebo group was quite similar to the one found by investigators in other research, using an identical protocol for 2-agonists (11). On the other hand, the flunisolide group obtained 12 percentage points of predicted above the placebo one in PEFR and FEV₁, suggesting the therapeutic effect of high doses of inhaled corticosteroids in patients with acute asthma.

There was a reduction in the hospitalizations for asthma from 25.5% in the placebo group to 8.51% in the flunisolide group (p = 0.09). Typical admission rates for acute asthma reported by other investigators range from 12 to 25% (16), in concordance with ours. When admissions were considered in relation to the duration of attack prior to ER treatment, the rate was statistically different only in patients with an exacerbation lasting more than 24 h: the placebo group produced significantly higher admissions than did the flunisolide group, suggesting that high doses of inhaled corticosteroids are an effective therapy, in particular for patients with a long duration of symptoms before ER presentation.

The simultaneous use of large doses of beta-agonists and corticosteroids raises the possibility of a drug interaction, with biochemical disturbances that result in adverse cardiovascular events. Recently, Lin and coworkers (30) suggested a possible enhanced -responsive state associated with inhaled corticosteroid use in patients with acute asthma. On the other hand, in this trial, both groups presented similar heart rate patterns during the 3 h of protocol. There were no differences between groups at any time point studied. The heart rate response in both groups was the characteristic of patients with acute asthma response to beta-agonists treatment (2, 11). Additionally, there were no arrhythmias detected by the continuous ECG monitoring. Likewise, the overall symptom scores in patients treated with flunisolide did not differ from the placebo group score.

In summary, our data support the theory that high and cumulative doses of inhaled flunisolide administered by metered-dose inhaler with spacer added to salbutamol are an effective therapy for patients with acute asthma and a long duration of symptoms before the ER presentation. New studies are needed to compare the effectiveness of different doses and methods of administration.