Corticosteroids in the Emergency Department Therapy of Acute Adult Asthma

Corticosteroids in the Emergency Department Therapy of Acute Adult Asthma

Corticosteroids in the Emergency Department Therapy of Acute Adult Asthma* An Evidence-Based Evaluation Gustavo Rodrigo, MD; and Carlos Rodrigo, MD O...

287KB Sizes 0 Downloads 16 Views

Corticosteroids in the Emergency Department Therapy of Acute Adult Asthma* An Evidence-Based Evaluation Gustavo Rodrigo, MD; and Carlos Rodrigo, MD

Objective: To review the literature to determine the benefits of corticosteroids (CCSs) (oral, IM, IV, or inhaled) in the treatment of adult patients with acute asthma presenting at an acute-care setting. Search strategy: A MEDLINE search was conducted using the following terms: (1) Asthma OR Wheez*, AND (2) Glucocorticoids OR Steroids, AND (3) Acute* OR Emerg. Other sources were the CURRENT CONTENTS database, review articles, reference sections of located studies, and a manual search of the top 15 journals for respiratory and emergency medicine. Selection criteria: Patients were selected for the study by the following criteria: (1) English language; (2) adult patients with asthma whose acute exacerbations were the primary reason for assessment; (3) involvement in randomized, controlled trials conducted in an emergency care setting; (4) patients had participated in a study investigating a primary research question involving treatment with CCSs; and (5) outcomes based on results of pulmonary function tests and on hospital admission rates. Results: At the 3-h assessment, only high doses of inhaled CCSs significantly improved pulmonary function compared with placebo (effect size [ES], 0.56; 95% confidence interval [CI], 0.15 to 0.97). On the other hand, after receiving IV CCSs, patients required at least 6 to 24 h to show moderate but nonsignificant improvements of pulmonary function (6-h ES, 0.44 [95% CI, 20.01 to 0.89]; 12-h ES, 0.54 [95% CI, 20.08 to 1.17]; and 24-h ES, 0.53 [95% CI, 20.39 to 1.45]). The data from the six studies that we used to pool information on admission rate outcome showed a 32% reduction in favor of the use of IV CCSs (relative risk [RR], 0.68 [95% CI, 0.47 to 0.99]; number needed to treat, 12.5 [95% CI, 7.1 to 50]). However, the pooled effect of the three high-quality studies showed no difference between groups (RR, 1.21; 95% CI, 0.67 to 2.18). Oral CCSs provided a similarly beneficial effect on pulmonary function when compared with parenteral administration (ES, 20.14; 95% CI, 20.82 to 0.31. Finally, the results showed a nonsignificant favorable trend toward improved outcome with medium or high doses of CCSs. Conclusions: This evidence-based evaluation suggests that the administration of parenteral CCSs to the patient on arrival at the emergency department (ED) neither improves airflow obstruction nor reduces the need for hospitalization. Parenteral CCSs probably require > 6 to 24 h to begin to act. Comprehensible conclusions about admission rates in the ED setting are difficult to make. At the 3-h assessment, only high doses of inhaled CCSs (in one study) significantly improved pulmonary function compared with placebo. IV and oral CCSs appear to have equivalent effects, and there is a tendency toward improvement in pulmonary function with medium or high doses. (CHEST 1999; 116:285–295) Key words: acute asthma; corticosteroids; emergency department; evidence-based evaluation; therapeutics Abbreviations: CCS 5 corticosteroid; CF 5 confidence interval; df 5 degrees of freedom; ED 5 emergency department; ES 5 effect size; NNT 5 number of patients needed to treat; PEFR 5 peak expiratory flow rate; RCT 5 randomized controlled trial; RR 5 relative risk

*From the Departamento de Emergencia (Dr. G. Rodrigo), Hospital Central de las FF.AA., Montevideo, Uruguay; and Centro de Tratamiento Intensivo (Dr. C. Rodrigo), Asociacio´n Espan˜ola Primera de Socorros Mutuos, Montevideo, Uruguay. Manuscript received January 8, 1999; revision accepted March 23, 1999. Correspondence to: Carlos Rodrigo, MD, Centro de Tratamiento Intensivo, Asociacio´n Espan˜ola Primera de Socorros Mutuos. Bulevar Artigas 1465, Montevideo 11300, Uruguay; e-mail: [email protected]

chronic airway inflammation appears to be S ince central to the pathogenesis of asthma, it is logical 1

For editorial comment see page 273 to use agents that suppress this process, such as corticosteroids (CCSs). These agents have been used in the treatment of asthma since 1950.2 However, CHEST / 116 / 2 / AUGUST, 1999

285

the favorable effects seen with these drugs in the chronic and subacute phases of asthma have produced expectations about the use of CCSs that have been inappropriately extended to emergency situations. Despite almost 50 years of clinical use and intense research, the resolution of fundamental issues such as the quantity of a CCS that is required to induce a rapid remission, the time it takes for CCSs to act on the patient, the route of administration, the existence of dose-response effects, and the patient populations likely to require, or respond to, CCS therapy in emergency situations remain uncertain.3 In fact, several articles have questioned the role of CCSs in early emergency department (ED) treatment of acute exacerbations.4 – 6 Despite the fact that the benefits of CCSs were confirmed by meta-analysis several years ago,7 that systematic review included a small number of trials involving children and adults and quasi-experimental designs. As the results of several new, well-designed controlled trials have been published since 1992, the conclusions of the original review may need revision. Recently, we reported for the first time that inhaled CCSs speed the resolution of acute airflow limitation.8 The pooling of information from a larger number of trials may provide greater power for detecting group differences, and also a better insight into the influence of trial characteristics and treatment modalities on the results. The objective of this review was to reevaluate the literature on the effectiveness of CCS administration in the treatment of adult patients with acute asthma presenting to an acute-care setting (ie, usually the ED). This review addresses specifically the following questions. (1) Do CCSs improve the results of pulmonary function tests over the first 24 h of treatment in cases of acute asthma? (2) Do CCSs decrease the hospital admission rate? (3) Does the parenteral (IV/IM) route of CCS administration improve outcomes when compared with oral or inhaled CCS administration of a dose of similar potency? (4) What is the relative efficacy of different CCS doses (“high dose” vs “low dose”)? Materials and Methods Search Strategy To identify all relevant studies, a computerized MEDLINE search was conducted. We searched for studies published in the English language for the years 1966 to October 1998. The following MeSH terms were used in the search: (1) Asthma OR Wheez*, AND (2) Glucocorticoids OR Steroids, AND3 Acute* OR Emerg. The apply limits were: (1) language, English; (2) target population, adults . 18 years old; and (3) publications type, clinical trials. Other sources of relevant articles were the CURRENT CONTENTS database, a previous meta-analysis,7 286

review articles,1,9 –17 and the reference sections of located studies. Finally, a manual search of the top 15 journals in respiratory care and emergency medicine was performed. Study Selection The reference list was reviewed independently by the two researchers (G.R. and C.R.); from the full text, using specific criteria, trials were selected for inclusion. Disagreements were resolved by consensus. Agreement was measured using k-statistics. The following inclusion criteria were used to select studies: (1) randomized, controlled trial (RCT) conducted in an emergency-care setting; (2) inclusion of patients with asthma whose acute exacerbations were the primary reason for assessment and exclusion of patients with chronic airflow limitation; (3) studies in which the primary research question involved treatment with parenteral (IV, IM), oral, or inhaled administration of CCSs; and (4) outcomes based on pulmonary function and hospital admission rates. The level of evidence for the selected studies ranged among the following18: (1) level I, RCT with definitive results (that is, a result with confidence intervals [CIs] that do not overlap the threshold clinically significant effect); and (2) level II, RCT with nondefinitive results (that is, a point estimate that suggests a clinically significant effect but with CIs overlapping the threshold for this effect). Methodological Quality Specific validity criteria were applied by the authors to all trials (Table 1). The methodological quality of each trial was assessed focusing on the following features: (1) randomization method; (2) demographic characteristics of the sample; (3) inclusion/exclusion criteria; (4) asthma definition; (5) sample size calculation; and (6) withdrawals. The mean score of the two evaluations was divided by the total possible score of 12 and was expressed as a

Table 1—Methodological Scoring* Randomization method Specified (2) Not specified (1) Demographic characteristics of the sample Detailed (2) Partial (1) None (0) Inclusion/exclusion criteria specified Detailed (2) Partial (1) None (0) Asthma defined by ATS criteria Yes (2) Another definition (1) No definition (0) Sample size calculation Detailed (2) Partial (1) None (0) Withdrawals None or specified # 20% of total subjects (2) Mentioned no. of withdrawals but not reasons (1) . 20% of subjects (0) *Numbers in parentheses are points contributing to the total score (total points could range from 1 [all studies received a point for randomization] to 12). ATS 5 American Thoracic Society. Clinical Investigations

value between 0.08 (1/12 because all studies were randomized) and 1.0. Trials with a score of . 0.7 were considered to be good quality. Inter-rater reliability was measured by using the k-statistic. Data Analysis Data were pooled for the following group comparisons: (1) any CCS vs placebo; (2) parenteral vs oral administration of CCS; and (3) high or moderate dose vs low dose. Doses were converted to the hydrocortisone pharmacologic equivalents using the tables provided by Siegel.19 Outcomes were classified based on pulmonary function testing or hospital admission rates. For pulmonary function testing, percent of predicted FEV1 was most often reported. If not available, absolute FEV1, percentage change in FEV1, or percent of predicted peak expiratory flow rate (PEFR) was used. The effect of treatment in each trial was computed using the standardized mean difference, reported in SD units.20 For our purposes, the effect size (ES) is the ratio of the mean difference and the pooled SD. Thus, the absolute difference between the two group means is divided by the pooled SD.21 If the SD was not available, the standard error was used to compute the SD by multiplying it by the square root of the treatment sample size. In some cases, values were measured from graphic displays. An adjustment was necessary to account for differences in the treatment and control groups at baseline. When pooled statistics were calculated, studies were weighted by the reciprocal of the variance; in case of heterogeneity, we used the random effect model.22 We interpreted the ES using guidelines of Cohen,23 in which a small ES is 0.2 SD units, a medium ES is 0.5 SD units, and a large ES is 0.8 SD units or greater. The effect sizes are presented with 95% CIs. Dichotomous outcome (admission rate) was reported as pooled relative risk (RR) with the fixed-effects model using the Mantel-Haenszel technique24 or, in case of heterogeneity, the DerSimonian and Laird randomeffects model.22 Homogeneity of effect sizes were tested with p 5 0.1 as the cutoff point for significance.25 When heterogeneity was found, subgroup analyses were carried out in an attempt to explain the findings. Sensitivity analysis was performed to examine the effect on results of methodological quality or the administration method of CCSs. To detect possible biases, funnel plot symmetry was examined for trials contributing data to admission rates.26 The number of patients needed to treat (NNT) to prevent one complication (admission) was calculated using the following formula: NNT 5 1/absolute risk reduction; the 95% CI for the NNT was obtained by taking reciprocals of the values defining 95% CI for the absolute risk reduction.27

Results Computerized Search A total of 87 articles were identified in the initial search. Of these, 21 RCTs were selected for inclusion.2,4 – 6,8,28 – 43 The k-agreement was 0.81 for this assessment (very good agreement). Reasons for exclusion were the following: non-RCT (40/67, 60%); outpatient studies (12/67, 18%); nonoriginal data (12/67, 18%); and study involved children (3/67, 4%). Following validity assessment, two studies were excluded because they did not fit any of the categories for analysis. One study examined the efficacy of two methylprednisolone prodrugs,38 while the other examined the effectiveness of inhaled steroids in con-

junction with IV steroids.32 Also, three additional studies were excluded, the first due to inadequate data reporting,28 the second because results were reported using a symptom measure,2 and the third because insufficient data were presented.39 Finally, the statistical analysis was performed with the remaining 16 RCTs. Eight studies were from North America, two were from Uruguay, one each was from England, Denmark, France, Iceland, Spain, and Australia. The studies were published between 1983 and 1998. Although all eligible reports were described as RCTs, only seven studies specified the randomization method,4 – 6,30,31,42,43 and three included detailed sample size calculations.4,8,43 Four studies4,6,34,37 included PEFR as the only measure of pulmonary function. The mean sample size was 61 patients (range, 18 to 150 patients), and the mean age was 32.3 years. The methodological scores ranged from 0.30 to 0.92 (mean, 0.55). Five studies were of high quality.4 – 6,8,43 The agreement among investigators for quality score was 0.89. Five studies reported side effects.5,6,8,35,43 There were no significant differences in cardiovascular, gastric, and metabolic effects between groups. A marked increase in WBC count was observed after a high dose of CCSs.43 Pulmonary Function Over the First 24 h The effectiveness of CCSs as measured by pulmonary function in the first 24 h of ED treatment has been reported in seven RCTs (Table 2).4 – 6,8,29 –31 Equivalent hydrocortisone doses administered ranged between 8.3 and 300 mg/kg. Only one study administered high doses of inhaled flunisolide (6 mg/h).8 The severity of the asthmatic exacerbation was determined from these studies. Mean pretreatment FEV1 levels were reported to be 25 to 32%,29 40 to 42%,30 0.55 to 0.57 L,31 27 to 29%,5 and 26 to 28%.8 Stein and Cole4 reported a mean pretreatment PEFR of 246 to 264 L/min, and Lin et al6 reported 28 to 30%. Pooled ES for pulmonary function was calculated for 1-, 3-, 6-, 12-, and 24-h assessment times (Table 3). There was statistically significant heterogeneity at the 1-h assessment (x2, 8.91; degrees of freedom [df], 2; p , 0.02). Sensitivity analysis explained the heterogeneity on the basis of the CCS administration method (parenteral vs inhaled). At the 3-h assessment, we obtained the same pattern (x2, 8.42; df, 2; p , 0.02). At the 6-h assessment, two trials reported a nonsignificant difference (ES, 0.44; 95% CI, 20.01 to 0.89) favoring CCS use. Finally, there was significant heterogeneity at the 12- and 24-h assessments, and sensitivity analysis was inadequate to explain the heterogeneity on the basis of methodological quality or dose. At the 1- and 3-h CHEST / 116 / 2 / AUGUST, 1999

287

288

Clinical Investigations

47

INH Placebo

49 23 22 47

10.4

IV Placebo IV MTP (125 mg)

26 49

47 29

b2-agonist (Am)

b2-agonist (Am)

b2-agonist (Am)

b2-agonist (Am)

b2-agonist (Am)

b2-agonist (Am)

9 48 49 44

b2-agonist (Am)

11

0.75

0.83

0.88

0.54

0.71

0.4

0.50

Cointerventions

N

IV Placebo INH FLU (6 mg/h)

8.3

IV Placebo IV HYD (500 mg)

300.0

10.4

IV Placebo IV MTP (125 mg)

IV Placebo IV MTP (60 mg/kg/4 h)

10.4

14.0

IV HYD (2 mg/kg 1 0.5 mg/kg/h) IV Placebo IV MTP (125 mg)

Treatment

Quality Score

0.21 (20.20–0.72)

20.81 (21.60–20.22)

NA

20.44 (20.96–0.08)

NA

NA

NA

1-h

0.56 (0.15–0.97)

NA

20.09 (20.48–0.30)

20.19 (20.72–0.34)

NA

NA

NA

3-h

NA

NA

NA

0.38 (20.14–0.90)

NA

NA

0.63 (20.25–1.51)

6-h

ES†

NA

NA

NA

NA

0.12 (20.29–0.53)

0.31 (20.08–0.70)

1.57 (0.69–2.45)

12-h

NA

NA

NA

0.12 (20.40–0.65)

NA

NA

1.08 (0.20–1.96)

24-h

*Calculations are based on a dose given to a 60-kg adult. HYD 5 hydrocortisone; MTP 5 methylprednisolone; FLU 5 flunisolide; INH 5 inhaled; N 5 No. of subjects; Am 5 aminophylline; NA 5 not available. †Values given as ES (95% CI).

Rodrigo and Rodrigo8

Lin et al6

Rodrigo and Rodrigo5

Morell et al31

Stein and Cole4

Littenberg and Gluck30

Fanta et al29

Study

HYD Equivalent, mg/kg/24 h

Table 2—Time Course of Pulmonary Function in the First 24 h*

Table 3—Pooled Standardized Mean Differences in Lung Function for Each Assessment Time Study Parenteral CCSs Morell et al31 Weight, % ES 95% CI Lin et al6 Weight, % ES 95% CI Rodrigo and Rodrigo5 Weight, % ES 95% CI Fanta et al29 Weight, % ES 95% CI Littenberg and Gluck30 Weight, % ES 95% CI Stein and Cole4 Weight, % ES 95% CI Totals ES 95% CI x2 df p Value Inhaled CCSs Rodrigo and Rodrigo8 ES 95%

1h (n 5 194)

3h (n 5 153)

6h (n 5 75)

55 20.44 20.96–0.08

36 20.19 20.72–0.34

73 0.38 20.01–0.89

12 h (n 5 208)

24 h (n 5 75)

73 0.12 20.40–0.65

45 20.81 21.40–20.22 64 20.09 20.48–0.30 27 0.63 20.25–1.51

9 1.57 0.69–2.45

29 1.08 0.20–1.96

47 0.31 0.08–0.70 44 0.12 20.29–0.53 20.60 20.99–20.21 0.01 1 . 0.8

20.11 20.42–0.20 0.3 1 . 0.5

0.21 20.20–0.62

0.56 0.15–0.97

assessments, only inhaled CCSs showed small and moderate ESs favoring CCSs. At the 6-, 12-, and 24-h assessments, parenteral CCSs showed moderate benefits. ED Admission Rates Six studies, totaling 480 patients, examined the effect of early administration of CCSs in preventing hospital admissions (Table 4).4 – 6,8,30,33 Medications were administered within 30 min of arrival in the ED. Admissions were assessed at approximately 2 to 6 h, depending on the study. Using sensitivity analysis based on the CCS administration method (parenteral vs inhaled) and methodological score (good vs poor quality), pooled studies did not show heterogeneity (Table 4). Low-quality trials with parenteral administration of CCSs significantly reduced admission rates (RR, 0.48; 95% CI, 0.29 to 0.81) Given a baseline admission risk of 25%, we calculated that

0.44 20.01–0.89 0.23 1 . 0.5

0.54 20.08–1.17 8.64 2 , 0.01

0.53 20.39–1.45 3.35 1 , 0.1

7.7 patients (95% CI, 4.0 to 100) would need to be treated with CCSs early in their ED stay to prevent one hospitalization. On the other hand, data suggest a possible benefit of inhaled CCSs. To the contrary, the pooled RRs of the three high-quality studies (score, . 0.7) presented a nonsignificant difference between groups (RR, 1.21; 95% CI, 0.67 to 2.18). Pooling of the six trials contributing data to this outcome showed a 32% reduction in hospital admissions in favor of CCSs (RR, 0.68 [95% CI, 0.47 to 0.99]; NNT, 12.5 [95% CI, 7.1 to 50]). There was no evidence of systematic bias identified by the measure of funnel plot (intercept, 4.49; 95% CI, 21.31 to 10.3). Parenteral vs Oral CCSs Four studies totaling 157 patients compared parenteral and oral administration of CCSs in acute asthma treatment.34 –37 Outcome assessment of paCHEST / 116 / 2 / AUGUST, 1999

289

290

Clinical Investigations

47 47

INH placebo

27

48 49 27

INH FLU (6 mg/h)

IV placebo

150.0

10.4

22

49 23

47 49

44

N

b2-agonist (Am)

b2-agonist (Am)

' 6 h spirometric and symptom assessment

3 h spirometric and symptom assessment

b2-agonist (Am)

b2-agonist (Am)

b2-agonist (Am)

b2-agonist (Am)

Cointervention

' 4 h NA

' 5.5 h clinical and laboratory assessment

' 2.2 h symptom assessment

' 6.5 h symptom assessment

Clinical Assessment Timing

*Calculations are based on a dose given to a 60-kg adult. See Table 2 for definitions of abbreviations. †x2, 0.07; df, 2; p . 0.9. ‡x2, 0.005; df, 1; p . 0.9. §x2, 6.27; df, 5; p . 0.2.

Subtotal Total§

Subtotal‡ Inhaled CCS Rodrigo and Rodrigo8

Schneider et al33

IV MTP (125 mg) IV placebo IV MTP (30 mg/kg)

IV placebo

10.4

IV placebo IV MTP (125 mg)

Lin et al6

Subtotal† Low quality Littenberg and Gluck30

8.3

IV placebo IV HYD (500 mg)

Rodrigo and Rodrigo5

HYD Equivalent mg/kg/24 h

10.4

Treatment

IV MTP (125 mg)

Parenteral CCS High quality Stein and Cole4

Study

Table 4 —Early Administration of CCSs and Admission Rates*

0.75

0.33

0.42

0.83

0.88

0.71

Quality Score

25.5

8.5

44.0

18.8 46.9 19.0

18.1

8.1 30.4

13.0 10.2

18.0

Admission Rate, %

0.45 (0.19–1.08) 0.68 (0.47–0.99)

0.48 (0.29–0.81)

0.49 (0.23–1.08)

0.47 (0.26–0.84)

1.21 (0.67–2.18)

1.35 (0.76–2.39)

1.12 (0.60–2.09)

1.22 (0.73–2.04)

RR (95% CI)

35

65

28

37

35

Weight, %

tients was made at 24 h using FEV1 (two studies)35,36 or PEFR (two studies)34,37 (Table 5). All trials were placebo controlled. The pooled ES was 20.14 (95% CI, 20.82 to 0.31), favoring the oral administration; this combined result did not demonstrate heterogeneity. One SD was approximately equivalent to 22% of predicted, suggesting a nonsignificant mean difference of 3%. There was no evidence of systematic bias (intercept, 0.73; 95% CI, 20.32 to 1.78). Oral administration of CCSs provided a similarly beneficial effect on pulmonary function when compared to parenteral administration. Different CCS doses Five studies compared different CCS doses in acute asthma patients (Table 6).33,40 – 43 Two comparisons of low dose to medium dose, and five comparisons of low dose to high dose were available for analysis. Outcome assessments of patients were made between 3 and 72 h. The equivalent administered hydrocortisone doses ranged between 3.3 and 300 mg/kg/24 h. The pooled ES for all studies showed a nonsignificant favorable trend toward improved outcome with medium or high doses (ES, 0.12; 95% CI, 20.08 to 0.32). Similar results were obtained when low doses were compared with medium and high doses. Again, there was no evidence of bias (intercept, 1.18; 95% CI, 20.78 to 3.16). No significant heterogeneity was found in these results. Discussion The purpose of this review was to perform an evidence-based evaluation of the effectiveness of CCS administration in the treatment of adult pa-

tients with acute exacerbations of asthma. This study met most of the methodological criteria that have been suggested as guidelines for scientific reviews,44 and all included trials were randomized and placebo controlled. In addition, this review has identified five trials published between 1994 and 1998, totaling 434 new subjects, that were not included in a previous meta-analysis.7 The funnel plot statistical assessment indicated the absence of bias associated with sample size. However, there is a possibility of publication bias or study selection bias in this review: unpublished and foreign language papers have not been included, and the number of studies being pooled remains small. Certainly, the present conclusions may be seriously modified by the results of large trials.45 The first analysis was performed to examine the time course of pulmonary function between CCSs and placebo. There was significant heterogeneity at the 1-h assessment; however, sensitivity analysis explained this heterogeneity on the basis of the CCS administration method. Only one study,8 which examined the administration of high doses of inhaled CCSs, produced a nonsignificant small pulmonary function improvement (ES, 0.21; 95% CI, 20.20 to 0.62). At the 3-h assessment, we found the same pattern. In this case, inhaled CCSs significantly improved pulmonary function compared with placebo (ES, 0.56; 95% CI, 0.15 to 0.97). In the remaining assessment times (6, 12, and 24 h), IV CCSs produced a nonsignificant but moderate improvement of pulmonary function compared with placebo. Even though more trials with adequate statistical power are needed to confirm these findings, available data support the current knowledge that CCSs required a minimum of 6 to 12 h to

Table 5—Parenteral vs Oral CCSs in the Treatment of Acute Adult Asthma*

Study Ratto et al35

Jo´nsson et al36 Engel et al37 Harrison et al34

Total‡

Treatment IV MTP (500–1,000 mg/24 h) po MTP (160–320 mg/24 h) IV MTP (80 mg/24 h) po MTP (80 mg/24 h) IV MTP (1,000 mg) po PRED (50 mg) IV HYD (12 mg/kg) 1 po PRED (75 mg) po PRED (75 mg)

HYD Equivalent mg/kg/24 h

N

Quality Score

Type of Pulmonary Function Test

Assessment Time, h

Final Pulmonary Function, %

ES (95% CI)

41.6–83.3

38

0.50

FEV1

24

55 (18)†

20.14 (20.61–0.33)

13.3–26.6

40

6.6 6.6 83.3 4.2 18.3

11 11 8 10 8

6.2

10

58 (24) 0.33

FEV1

24

0.42

PEFR

24

0.30

PEFR

24

72 (15) 65 (15) 79 (18) 78 (18) 52 (21)

0.06 (20.39–1.29) 0.05 (20.89–0.99) 20.25 (20.82–0.31)

58 (25) 20.14 (20.8–0.31)

*Calculations are based on a dose given to a 60-kg adult. PRED 5 prednisolone. See Table 2 for definitions of other abbreviations. †Values in parentheses are percent predicted for pulmonary function test. ‡x2, 0.33; df, 3; p . 0.9. CHEST / 116 / 2 / AUGUST, 1999

291

292

Clinical Investigations

LD: IV MTP (15 mg/6 h) MD: IV MTP (40 mg/6 h) HD: IV MTP (125 mg/6 h) 3 3 d LD: IV HYD (50 mg/6 h) MD: IV HYD (100 mg/6 h) HD: IV HYD (500 mg/6 h) 3 2 d LD: IV MTP (2 mg/kg/4 h) HD: IV MTP (10 mg/kg/4 h) IV placebo LD: IV MTP (100 mg) HD: IV MTP (500 mg) LD: IV MTP (1 mg/kg/24 h) HD: IV MTP (6 mg/kg/24 h) 3 48 h

40

8.3 41.6 5.0 30.0

5.0 13.3 41.7 3.3 6.7 33.3 60 300

8 8 8 22 20 24 27 29 26 74 76 23 24

N

Quality Score 0.38

0.50

0.54

0.46 0.92

Cointervention b2-agonist (Am) b2-agonist (Am) b2-agonist (Am) b2-agonist (Am) b2-agonist (Am)

FEV1/24 h

FEV1/3 h

FEV1/24 h

FEV1/24 h

FEV1/72 h

Type of Pulmonary Function Test/Time 55 (23)† 64 (11) 66 (11) 62 (22) 62 (23) 65 (28) 0.98 (0.35) 0.88 (0.34) 0.84 (0.35) 56.3 (18) 56.0 (19) 42 (14) 52 (26)

Final Pulmonary Function, %

ES (95% CI)

0.12 (20.08–0.32)‡ LD vs MD: 0.15 (20.36–0.66)§ LD vs HD: 0.12 (20.09–0.33)\

LD vs HD: 0.50 (20.06–1.06)

LD vs HD: 0.09 (20.22–0.40)

LD vs HD: 0.28 (20.80–0.25)

LD vs MD: 0.00 (20.31–0.31) LD vs HD: 0.11 (20.46–0.68)

LD vs MD: 0.53 (20.45–1.51) LD vs HD: 0.65 (20.33–1.63)

*Calculations are based on a dose given to a 60-kg adult. LD 5 low dose; MD 5 medium dose; HD 5 high dose. See Table 2 for definitions of other abbreviations. †Values in parentheses are percent predicted for pulmonary function test. ‡x2, 4.91; df, 4; p . 0.2. §x2, 0.81; df, 1; p . 0.3. \x2, 4.07; df, 4; p . 0.3.

All studies

Marquette et al43

Emerman and Cydulka42

Morell et al31

Bowler et al41

Haskell et al

Treatment

Study

HYD Equivalent mg/kg/24 h

Table 6 —Studies Comparing Different Dosing Regimens*

improve pulmonary function in patients. On the other hand, high doses of inhaled CCSs produced early therapeutic effects (1 to 3 h). The fact that two studies4,6 included PEFR as a pulmonary function measure could make PEFR a factor that affects the homogeneity of results. Although PEFR is an adequate measure, it has flaws: it is operator dependent and requires patient effort. The second analysis examined the ED use of CCSs and hospital admission rates. Overall, the six studies examined showed a 32% reduction in hospital admission rates in favor of CCSs (RR, 0.68; 95% CI, 0.47 to 0.99). Using sensitivity analysis based on the CCS administration method and quality score, pooled studies continue to be homogeneous. Only early administration of parenteral CCSs in low-quality trials4 – 6,30 –33 leads to reduced admission rates (RR, 0.48; 95% CI, 0.29 to 0.81). However, the admission rates of the control groups of both trials30,33 were surprisingly high (47% and 44%, respectively), raising doubts about the applicability of these findings to other emergency settings. On the contrary, the three high-quality studies presented a nonsignificant difference between groups (RR, 1.21; 95% CI, 0.67 to 2.18). Finally, inhaled CCSs produced a nonsignificant reduction of admission rate (RR, 0.45; 95% CI, 0.19 to 1.08). Consequently, clear conclusions are difficult to make. There are different reasons for this discrepancy. The decision to admit or discharge patients is based on multiple factors such as history of asthma, current history of acute asthma, results of pulmonary function tests on ED arrival and after treatment, and clinical judgment. Clearly, important variations in these factors could explain the analysis differences. The administration of IV CCSs has become routine for severely acute asthmatic patients. Nevertheless, analysis of the results of the comparison of parenteral vs oral CCSs reveal no evidence to suggest that one route of administration improves pulmonary function more than another (ES, 20.14; 95% CI, 20.45 to 0.17). However, no studies involved assessment time periods . 24 h. Thus, administration of 6.6 to 83.3 mg/kg IV hydrocortisone equivalents produces the same quantitative improvements in pulmonary function as a dose of 4.2 to 26.6 mg/kg po. With minor complications of IV therapy, these results support the use of oral forms of CCSs. The final analysis was performed to examine studies comparing different dosing regimens of CCSs in patients with acute asthma. The pooled effect for all studies presented a tendency toward improvement in pulmonary function with medium and high doses. This analysis suggests that doses of hydrocortisone or its equivalent of , 13 mg/kg/24 h are subtherapeutical. This dose can be accomplished with the admin-

istration of approximately 800 mg hydrocortisone, or 160 mg methylprednisolone per day. Similar results were obtained when low doses were compared with medium and high doses. Therefore, the evidence does not support the concept that very large doses of CCSs are more efficacious than smaller ones. It is probably the case that very high doses add little to the outcomes of these patients, as the plateau in the dose-response curve may be reached at lower doses. Because a recent meta-analysis addressed the effectiveness of CCSs on relapse following ED discharge, we did not include that topic in the review. Rowe et al46 selected seven articles (including pediatric and adult asthma patients) for their analysis. They concluded that a short course of CCSs following ED discharge significantly reduces the number of relapses requiring additional care and decreases b-agonists use without an apparent increase in side effects. Finally, IM CCSs appear to be as effective as oral agents.

Conclusion With reference to the early administration of CCSs in the ED treatment of adult patients with acute asthma, this evidence-based evaluation suggests the following: (1) parenteral administration probably requires . 6 to 24 h to improve pulmonary function in the patient (on the contrary, high doses of inhaled CCSs [one study]8 increase pulmonary function 1 to 3 h after therapy); (2) comprehensible conclusions about admission rates in the ED setting are difficult to make (overall, there is evidence about a reduction in hospital admissions, but a separate analysis, including only high-quality studies, showed no benefit); (3) IV and oral CCSs appear to have equivalent effects on pulmonary function; and (4) there is a tendency toward improvement in pulmonary function with medium and high doses, although the evidence does not support the use of very high doses. Previous research has shown two different therapeutic response patterns to inhaled b-agonists in adult patients with acute asthma47,48: (1) a good response pattern (almost 70% of patients), characterized by subjects with a quick response (two-thirds obtain the discharge threshold after 1 h of treatment); and (2) a relatively large proportion of asthmatics (30%) with a poor response to therapy. These patients are characterized by more severe disease, as judged by previous b-agonist use, larger duration of attack before an ED visit, and a more severe obstruction at presentation. However, what ultimately determines whether admission is necessary is the early response to treatment.49 These findings agree with CHEST / 116 / 2 / AUGUST, 1999

293

the biphasic time course of response to therapy in asthma, with fast and slow components.50 The more slowly resolving stage probably represents the effects of mucosal edema and impaired mucociliary transport mechanisms. When this is the major component (poor response pattern), bronchodilator therapy tends to be ineffective and the patient will require a longer period of intense treatment with CCSs to terminate the episode and prevent relapses. Is it reasonable to observe a patient in the ED to await the effect of parenteral CCSs? Patients with good responses do not require parenteral CCSs, because they do not accelerate improvement in these patients who are highly responsive to bronchodilators. On the other hand, new evidence suggests the use of high doses of inhaled CCSs in patients with a prolonged duration of symptoms before the ED presentation and a poor response to treatment. Thus, the early response of the patient to b-agonist therapy then would determine whether waiting for the onset of action of CCSs would be of value. When the PEFR is , 40% of predicted and PEFR variation over baseline is , 60 L/min, both measured at 30 min of therapy,49 we can wait for the onset of action of inhaled CCSs for 2 or 3 additional hours (by topical anti-inflammatory effect3 or by restoring airway b2adrenoceptor responsiveness51). If the PEFR remains , 40% after that time, these patients will require hospitalization and parenteral CCS therapy, since improvement from discharge from the hospital takes 4 days.52 References 1 National Asthma Education and Prevention Program. Expert panel report 2: guidelines for the diagnosis and management of asthma. Bethesda, MD: National Institutes of Health, April 1997; Publication No. 97– 405 2 Medical Research Council. Subcommittee on clinical trials in asthma: controlled trial of effects of cortisone acetate in status asthmaticus. Lancet 1956; 1:803– 806 3 McFadden ER. Inhaled glucocorticoids and acute asthma: therapeutic breakthrough or nonspecific effect? Am J Respir Crit Care Med 1998; 157:677– 678 4 Stein LM, Cole RP. Early administration of corticosteroids in emergency room treatment of acute asthma. Ann Intern Med 1990; 112:822– 827 5 Rodrigo C, Rodrigo G. Early administration of hydrocortisone in the emergency room treatment of acute asthma: a controlled clinical trial. Respir Med 1994; 88:755–761 6 Lin RY, Pesola GR, Westfal RE, et al. Early parenteral corticosteroid administration in acute asthma. Am J Emerg Med 1998; 15:621– 625 7 Rowe BH, Keller JL, Oxman AD. Effectiveness of steroid therapy in acute asthma: a meta-analysis. Am J Emerg Med 1992; 10:301–310 8 Rodrigo G, Rodrigo C. Inhaled flunisolide for acute severe asthma. Am J Respir Crit Care Med 1998; 157:698 –703 9 Hall JB, Wood LDH. Management of the critically ill asthmatic patients. Med Clin North Am 1990; 74:779 –796 294

10 Engel T, Heinig JH. Glucocorticosteroid therapy in acute severe asthma: a critical review. Eur Respir J 1991; 4:881– 889 11 McFadden ER Jr. Dosages of corticosteroids in asthma. Am Rev Respir Dis 1993; 147:1306 –13103 12 Leatherman J. Life-threatening asthma. Clin Chest Med 1994; 15:453– 479 13 Corbridge TC, Hall JB. The assessment and management of adults with status asthmaticus. Am J Respir Crit Care Med 1995; 151:1296 –1316 14 McFadden ER, Hejal R. Asthma. Lancet 1995; 345:1215– 1219 15 Beveridge RC, Grunfeld AF, Hodder RV, et al. Guidelines for the emergency management of asthma in adults. Can Med Assoc J 1996; 155:25–37 16 British asthma guidelines coordinating committee. British guidelines on asthma management: 1995-review and position statement. Thorax 1997; 52(suppl):S1-S24 17 Lipworth BJ. Treatment of acute asthma. Lancet 1997; 350(suppl II):18 –23 18 Guyatt GH, Sackett DL, Sinclair JC, et al. User’s guides to the medical literature: IX. A method for grading health care recommendations. JAMA 1995; 274:1800 –1804 19 Siegel SC. Overview of corticosteroid therapy. J Allergy Clin Immunol 1985; 76:312–320 20 Glass GV. Meta-analysis in social research. Beverly Hills, CA: Sage, 1981 21 Hedges LV, Olkin I. Statistical methods for meta-analysis. San Diego, CA: Academic Press, 1985 22 DerSimonian R, Laird N. Meta-analysis in clinical trial. Control Clin Trials 1986; 7:177–188 23 Cohen J. Statistical power analysis for the behavioral sciences. Orlando, FL: Academic Press, 1977 24 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22:719 –748 25 Fleiss JL. The statistical basis of meta-analysis. Stat Methods Med Res 1993; 2:121–145 26 Egger M, Davey Smith G, Scneider M, et al. Bias in meta-analysis detected by a simple graphical test. BMJ 1997; 315:629 – 634 27 Altman DG. Confidence intervals for the number to treat. BMJ 1998; 317:1309 –1312 28 McFadden ER, Kiser R, deGroot WJ, et al. A controlled study of the effects of single doses of hydrocortisone on the resolution of acute attacks of asthma. Am J Med 1976; 60:52–59 29 Fanta CH, Rossing TH, McFadden ER Jr. Glucocorticoids in acute asthma: a critical controlled trial. Am J Med 1983; 74:845– 851 30 Littenberg B, Gluck EH. A controlled trial of methylprednisolone in the emergency treatment of acute asthma. N Engl J Med 1986; 314:150 –152 31 Morell F, Orriols R, de Gracia J, et al. Controlled trial of intravenous corticosteroids in severe acute asthma. Thorax 1992; 47:588 –591 32 Guttman A, Afilalo M, Colacone A, et al. The effects of combined intravenous and inhaled steroids (beclomethasone dipropionate) for the emergency treatment of acute asthma: the Asthma ED Study Group. Acad Emerg Med 1997; 4:100 –106 33 Schneider SM, Pipher A, Britton HL, et al. High-dose methylprednisolone as initial therapy in patients with acute bronchospasm. J Asthma 1988; 25:189 –193 34 Harrison BDW, Hart GJ, Ali NJ, et al. Need for intravenous hydrocortisone in addition to oral prednisolone in patients admitted to hospital with severe asthma without ventilatory failure. Lancet 1986; 1:181–184 Clinical Investigations

35 Ratto D, Alfaro C, Sipsey J, et al. Are intravenous corticosteroids required in status asthmaticus? JAMA 1988; 260:527– 529 36 Jo´nsson S, Kjartansson G, Gislason D, et al. Comparison of the oral and intravenous routes for treating asthma with methylprednisolone and theophylline. Chest 1988; 94:723– 726 37 Engel T, Dirksen A, Frolund L, et al. Methyprednisolone pulse therapy in acute severe asthma: a randomized, doubleblind study. Allergy 1990; 45:224 –230 38 Muir JF, Godard PH, Verhaert P, et al. Seventy-two hour comparison of methyprednisolone suleptanate and methylprednisolone sodium succinatein patients with acute asthma. Br J Clin Pract 1996; 50:440 – 445 39 Tanaka RM, Santiago SM, Kuhn GJ, et al. Intravenous methyprednisolone in adults in status asthmaticus. Chest 1982; 82:438 – 440 40 Haskell RJ, Wong BM, Hansen JE. A double-blind, randomized clinical trial of methylprednisolone in status asthmaticus. Arch Intern Med 1983; 143:1324 –1327 41 Bowler SD, Mitchell CA, Armstrong JG. Corticosteroids in acute severe asthma: effectiveness of low doses. Thorax 1992; 47:584 –587 42 Emerman ChL, Cydulka RK. A randomized comparison of 100-mg vs 500-mg dose of methylprednisolone in the treatment of acute asthma. Chest 1995; 107:1559 –1563 43 Marquette CH, Stach B, Cardot E, et al. High-dose and low dose systemic corticosteroids are equally efficient in acute severe asthma. Eur Respir J 1995; 8:22–27 44 Oxman AD, Cook DJ, Guyatt GH. User’s guides to the

45 46

47

48

49

50

51

52

medical literature: how to use an overview. JAMA 1994; 272:1367–1371 Sim I, Hlatky MA. Growing pains of meta-analysis. BMJ 1996; 313:702–703 Rowe BH, Spooner CH, Ducharme FM, et al. The effectiveness of corticosteroids in the treatment of acute exacerbations of asthma: a meta-analysis of their effect on relapse following acute assessment. The Cochrane Database of Systematic Reviews (updated 1998). Available at: http://www.update.usa. com/clibip/clib.htm Strauss L, Hejal R, Galan G, et al. Observations on the effects of aerosolized albuterol in acute asthma. Am J Respir Crit Care Med 1997; 155:454 – 458 Rodrigo C, Rodrigo G. Therapeutic response patterns to high and cumulative doses of salbutamol in acute severe asthma. Chest 1998; 113:593–598 Rodrigo G, Rodrigo C. Early prediction of poor response in acute asthma patients in the emergency department. Chest 1998; 114:1016 –1021 McFadden ER, Kiser R, DeGroot WJ. Acute bronchial asthma: relations between clinical and physiological manifestations. N Engl J Med 1973; 288:221–225 Tan KS, Grove A, McLean A, et al. Systemic corticosteroid rapidly reverses bronchodilator subsensitivity induced by formoterol in asthmatic patients. Am J Respir Crit Care Med 1997; 156:28 –35 McFadden ER, Elsanadi N, Dixon L, et al. Protocol therapy for acute asthma: therapeutic benefits and cost savings. Am J Med 1995; 99:651– 661

CHEST / 116 / 2 / AUGUST, 1999

295