Chapter  44:  Management of Chronic Asthma: Evidence Report/Technology Assessment Number 44

Part 1. Long-Term Management of Asthma in Children

Part 2. Effects of Delayed ICS Use on Progression and Reversibility

Part 3. Addition of Other Long-Term Controller Medications to ICS

Part 4. Effect of Antibiotics on Acute Asthma Exacerbations

Part 5. Asthma Management Plans

Part 1. Long-Term Management of Asthma in Children

Part 2. Effects of Delayed ICS on Progression and Reversibility in Patients With Mild-to-Moderate Asthma

• The evidence is insufficient to permit conclusions on whether early intervention with long-term controller medications is superior to delayed introduction. The best available evidence does not support the assumption that mild-to-moderate asthmatics have a progressive decline in lung function that can be prevented by early initiation of ICS.

The CAMP trial is the most robust evidence to date on long-term lung function outcomes in a group of patients treated with ICS compared with a placebo-control group. Although immediate initiation and delayed initiation of ICS were not directly compared, CAMP provides the strongest prospective evidence available on the natural history of mild-to-moderate asthma managed without ICS or other long-term controller medication. The CAMP trial did not find progressive decline in lung function over a 4-year period in a population of children with mild-to-moderate asthma managed without ICS; nor was there a significant difference or change between treated and control groups in postbronchodilator FEV1.

It is possible that the findings of the CAMP study are not generalizable to patients with less intensive overall care. Also, the findings may not be generalizable over longer periods of followup, to populations newly diagnosed with asthma, to groups of patients with more severe asthma, or to a subset of patients with a more variable disease course. But, for the general group of children with mild-to-moderate asthma, there is no convincing evidence that a progressive and clinically measurable decline in lung function can be altered by early initiation of ICS.

The available evidence on immediate vs. delayed initiation of ICS is from four studies. These studies have notable limitations with respect to relevance of the population, time frames for study entry and followup, clarity of reporting, and the use of appropriate control groups. None of these studies was prospectively designed to address the key question in the specific population of interest and thus, did not provide rigorous data relevant to this particular key question. Two studies (n=52; n=102, respectively) were open-label extensions of randomized controlled trials on the efficacy of ICS, in which the patients initially assigned to the control group were subsequently treated with ICS. There were also two single-arm studies, one of adults (n=105) and one of children (n=216), in which patients were stratified by duration of asthma prior to initiating ICS treatment (with outcomes being compared across the strata).

Due to high withdrawal rates, the most relevant of the extension phase randomized trials reported only on 16 patients who received immediate corticosteroid treatment; and no data were provided to test the statistical significance of results at the final 3-year time point. The larger of the extension phase randomized trials did not report on the patient population and outcomes most relevant to the key question. Neither of the single-arm studies clearly demonstrated a relationship between asthma duration and outcomes that was consistent among all strata analyzed.

Part 3. Addition of Other Long-Term Controller Medications to ICS in Patients With Moderate Asthma

• There is a large body of evidence on the addition of long-acting beta-2 agonists to ICS, consisting of 28 studies that enrolled over 7,000 patients, with 1-year followup in the longest trials. The evidence consistently shows improvement in lung function outcomes, symptom outcomes, and supplemental beta-2 agonist use. Limited evidence shows that ICS dosage may be reduced without diminishing asthma control. However, there are only two pediatric studies that together report on 383 children, only 167 of whom were treated with the addition of long-acting beta-2 agonists.

Sixteen randomized, double-blinded trials that enrolled a total of 3,163 patients compared the addition of long-acting beta-2 agonists to a fixed dose of ICS. This evidence consistently showed improvement in lung function outcomes, symptom outcomes, and supplemental beta-2 agonist use. The combined estimate of treatment effect for FEV1 is 0.17 L (95 percent CI, 0.12-0.22) or 3.71 percent predicted (95 percent CI, 2.67-4.75), based on 14 studies with 2,781 evaluable patients. For morning, patient-measured PEF, the combined estimate of treatment effect is 24.7 L/min (95 percent CI, 17.7-31.7) or 7.3 percent predicted (95 percent CI, 5.3-9.3), based on nine studies with 1,678 evaluable patients. For supplemental beta-2 agonist use, the combined estimate of treatment effect was 1.18 fewer puffs/day (95 percent CI, −1.56 to −0.84), based on six studies with 1,142 evaluable patients.

Three crossover trials that enrolled a total of 151 patients evaluated reducing the dose of ICS after the addition of long-acting beta-2 agonists compared to placebo. The largest of these trials, which was randomized and double-blinded, reported on 84 patients treated for 6 months. All three trials demonstrated statistically significant reductions in ICS dosage for the long-acting beta-2 agonist group, ranging from 13.5 percent to 23.4 percent less than placebo. The evidence suggests that the reduction in dose is achieved without diminishment of lung function or increase in symptoms; and there is limited evidence to suggest improvement in symptoms.

Twelve randomized trials that enrolled more than 4,000 patients compared the addition of a long-acting beta-2 agonist to low or moderate dose ICS with an increased dose of ICS. All trials but one were double-blinded. This evidence consistently showed improvement in lung function outcomes, symptom outcomes, and supplemental beta-2 agonist use. The combined estimate of the magnitude of treatment effect for FEV1 is 0.11 L (95 percent CI, 0.07-0.15) or 2.32 percent predicted (95 percent CI, 1.48-3.16), based on eight studies with 2,754 evaluable patients. For morning, patient-measured PEF, the combined estimate of treatment effect is 11.6 L/min (95 percent CI, 5.2-18.0) or 3.4 percent predicted (95 percent CI, 1.5-5.3), based on 10 studies with 3,042 evaluable patients. For supplemental beta-2 agonist use, the combined estimate of treatment effect was 0.19 fewer puffs/day (95 percent CI, −0.06 to −0.31), based on three studies with 725 evaluable patients.

Data on adverse events were abstracted from clinical trials selected for inclusion in this report. In general, the adverse event profile for the addition of long-acting beta-2 agonists was similar to that for ICS alone. This analysis is limited because it examines only short-term adverse events for patients enrolled in clinical trials.

• There is a small body of evidence on the addition of theophylline, consisting of six studies that enrolled 408 patients. The evidence suggests that the addition of theophylline to ICS produces improved lung function and symptoms. However, there are only two pediatric studies available that together reported on only 47 children treated with theophylline.

Six studies that evaluated a total of 408 patients compared the addition of theophylline to ICS, with 6 months of treatment in the longest trial. Four of these compared the addition of theophylline to a fixed ICS dose, and two compared the addition of theophylline to a higher dose of ICS. The four studies on the addition of theophylline to fixed ICS dose are generally mixed in their results, but the qualitative direction of the results suggests that the addition of theophylline to a fixed ICS dose produces improved lung function and symptoms. Based on two randomized clinical trials, theophylline plus ICS vs. a higher dose of ICS appears to produce roughly equivalent improvements in lung function and symptoms.

• The evidence on the addition of leukotriene antagonists to ICS consists of five studies that enrolled a total of 1,111 patients, with 4 months of treatment in the longest trial. The evidence shows improved lung function and symptom scores when leukotriene antagonists are added to ICS. One trial showed that ICS dosage may be reduced without diminishing asthma control. There are no pediatric studies available.

Five studies enrolling 1,111 patients compared the addition of leukotriene antagonists to ICS. These studies are mostly randomized controlled trials that report on short-term outcomes. Four studies compared the addition of a leukotriene antagonist to a fixed dose ICS, and the fifth one evaluated the ability to reduce the ICS dose after starting a leukotriene antagonist. Of the four studies using a fixed dose of ICS, all showed that lung function was better when a leukotriene antagonist was added to a fixed dose of ICS. Three of these four studies also showed that symptom scores were improved. Two of the studies showed decreased use of beta-2 agonist under the combined regimen. The fifth study showed that the addition of a leukotriene antagonist allowed a greater number of patients to reduce the dosage of ICS under protocol-guided dosing guidelines.

Part 4. Effect of Antibiotics on Acute Asthma Exacerbations

• The limited available evidence suggests that there is no benefit of using antibiotics routinely, or when suspicion of bacterial infection is low. No study addressed whether using antibiotics when suspicion of bacterial infection is moderate or high improves the outcome of treatment for acute exacerbation of asthma.

The available evidence consists of two randomized, placebo-controlled trials that enrolled a total of 121 hospital admissions for acute asthma exacerbations. Both studies were relatively old, having been published in 1974 and 1982. Furthermore, they may have been underpowered to detect treatment differences. One of the studies evaluated lung function and symptom outcomes only at 24 hours after patient admission, a length of time that may be insufficient to evaluate the benefit of antibiotics. In addition, the antibiotics used in these studies did not have activity against atypical organisms, such as Mycoplasma or Chlamydia. It is not known whether antibiotics in current use that have activity against atypical organisms may improve outcomes.

The available evidence suggests that there is no benefit of using antibiotics routinely or when suspicion of bacterial infection is low. Neither study found a statistically significant benefit for antibiotics on the outcomes of lung function at time of discharge, hospital length-of-stay, or symptom scores. There were no studies that addressed the question of greatest relevance to contemporary clinical practice, whether using antibiotics when suspicion of bacterial infection is moderate or high (i.e., when signs and symptoms suggest the possibility of bacterial infection, but do not clearly indicate its presence) improve the outcomes of treatment for acute exacerbation of asthma.

Part 5. Asthma Management Plans

• The available evidence is insufficient to demonstrate that, compared to medical management alone, the use of a written asthma action plan improves outcomes. The available evidence is also insufficient to demonstrate that, compared to a written action plan based on symptoms, use of a written action plan based on peak flow monitoring improves outcomes.

A large body of literature on self-management interventions in asthma was reviewed for this report. From this literature, randomized controlled trials were selected that contained specific comparisons relevant to the key question. These trials were also largely free of contamination by interventions that were not directly relevant to the key question. Many articles were excluded due to the presence of multimodal interventions in the treatment group, particularly intensive patient education or optimization of medications, which were likely to confound results.

Nine randomized controlled trials that enrolled a total of 1,501 patients met the study selection criteria for this key question. Two of these trials included three arms: medical management alone, PFM-based written action plan, and symptom-based written action plan. This resulted in 11 comparisons among the nine studies. Seven trials (n=1,079) compared medical management with and without a written action plan (all having used a PFM-based plan). The two types of written action plans (PFM-based and symptom-based) were compared in four trials (n=393).

Of the nine trials reviewed, seven reported no significant differences in any measure of utilization, symptoms, or lung function. This included the largest of these trials (n=569), the Grampian Asthma Study of Integrated Care. However, as a group, the included trials were underpowered to detect differences in utilization outcomes (such as hospitalization and emergency room visits [which are events that occur infrequently]). Two trials reported statistically significant and striking reductions in emergency room utilization with use of a PFM-based action plan. However, both trials have serious flaws that diminish confidence in the results.

The available evidence does not demonstrate that written asthma action plans improve outcomes. Nor does this evidence refute the hypothesis that use of a written asthma plan is beneficial. If there is benefit in a written asthma action plan, it is most likely to be found in a population with severe or poorly controlled asthma leading to high utilization of in-hospital and emergency room treatment. As previously stated, two trials reported benefits from a PFM action plan, but neither trial provided a rigorous comparison with a symptom-based plan.

Burden of Illness

Asthma is estimated to affect 14 million to 15 million persons in the U.S. It is the most common chronic disease of childhood, estimated to affect 4.8 million children (National Heart, Lung, and Blood Institute, 1997). As many as 10-15 percent of boys and 7-10 percent of girls may have asthma at some time during childhood (Behrman, Kliegman, and Jensen, 2000).

There are 70,000 asthma-related hospitalizations annually, and more than 5,000 people die of asthma each year. Hospitalization rates have been highest among blacks and children, and death rates have been consistently highest among blacks aged 15 to 24 years. Rates of asthma have increased or remained stable over the past decade (National Heart, Lung, and Blood Institute, 1997).

Epidemiology and Natural History of Asthma

Asthma commonly arises in childhood, but may have its onset at any age. Prevalence among young children varies according to definition, but the 1988 Health Interview Survey indicated a prevalence of asthma of nearly 5 percent in the United States among children 10 to 17 years of age (Behrman, Kliegman, and Jensen, 2000). However, the long-term prognosis for childhood asthma is quite variable. Longitudinal studies show that about 50 percent of asthmatic children are free from symptoms within 10-20 years, but recurrences may occur. Children with severe asthma are less likely to have disease remission. Although the majority of childhood asthmatics have a good prognosis, some are at risk of impaired maturation of lung function and growth of lung tissue during childhood, lower attained level of lung function at adulthood, and decline of lung function during adulthood (Ulrik, 1999). Because of disease variability between individuals, it is difficult to predict the natural history of asthma in any given patient and difficult to know if treatment can alter the natural history of the disease.

Although most asthma arises during childhood, the annual incidence of asthma after the age of 20 is about 100 per 100,000 for the rest of the life span (Reed, 1999). An adult with asthma may have had the symptoms consistently since childhood, have recovered from the disease in childhood only to relapse, or have acquired the disease later in life. Unlike asthma in childhood, family history has not been demonstrated for adult-onset asthma. Occupational and environmental exposures, respiratory infections, and smoking have been implicated in the etiology of adult-onset asthma. Complicating the understanding of asthma in adults is the existence of other common conditions such as chronic obstructive pulmonary disease and chronic bronchitis, which may coexist in certain individuals. In persons who exhibit characteristics of asthma and either chronic obstructive pulmonary disease or chronic bronchitis, it may be difficult or impossible to determine which condition is the cause or result of worsening lung function over time. Thus, in both adults and children the natural history of asthma is poorly understood, and the effect of long-term control medication on long-term prognosis needs to be assessed in light of this context.

Pathophysiology and Pathology of Asthma

The important pathophysiologic feature of asthma that causes its classic clinical symptoms is an exaggerated bronchoconstrictor response. Wheezing, shortness of breath, and declines in lung function tests occur in response to exposure to allergens, environmental irritants, viral infections, cold air, exercise, or other poorly defined factors. Current research has demonstrated that inflammation is a critical process involved in the pathophysiology of asthma. Airway markers of inflammation correlate with measures of bronchial hyperresponsiveness. Secondly, treatment of asthma with certain anti-inflammatory medications reduces inflammation and diminishes airway hyperresponsiveness.

In addition to acute bronchoconstriction, inflammation is thought to contribute to acute asthma symptoms by contributing to airway edema and by causing formation of chronic mucous plugs.

Although the changes in lung function induced by inflammation are largely reversible with appropriate treatment, inflammatory cell infiltration coupled with airway smooth muscle spasm, mucosal edema, and chronic mucous plugging of smaller airways can result in airway remodeling. In some patients, this leads to airflow limitations that are only partially reversible. A pathologic feature of asthma is an alternation in the amount and composition of extracellular material in the airway wall, which may cause permanent changes in airway anatomy. Although airway remodeling is not fully understood, the irreversible changes in lung function in some asthma patients suggests that long-term anti-inflammatory therapy which is started early and continued according to a management plan may modify the long-term course of the disease.

Asthma Severity

Because asthma is a chronic disease characterized by differing frequencies of daily symptoms, exacerbations and remissions, characterizing the severity of the disease is problematic. Level of symptoms may be confounded by current treatment, which may be adequate or inadequate. Tests of lung function may not correlate with the severity and frequency of symptoms. However, in order to facilitate useful clinical guidelines for management and to classify patients consistently to evaluate outcomes, the National Heart, Lung, and Blood Institute (NHLBI) currently classifies asthma into four levels of severity based on pretreatment measures and symptoms: mild intermittent, mild persistent, moderate persistent, and severe persistent (National Heart, Lung, and Blood Institute, 1997). The classification system uses symptom frequency, exacerbation severity and frequency, frequency of nighttime symptoms, and lung function (as assessed by forced expiratory volume in one second [FEV1] or peak expiratory flow [PEF] variability). The presence of the most severe features in any category places a patient in that category of asthma severity.

The current classification system differs from the prior NHLBI Expert Panel report on asthma (National Asthma Education and Prevention Program, 1992). Most published reports of clinical research on asthma do not use the NHLBI classification system. The classification schemes in most studies are quite variable. Comparisons of patients outside of or between clinical trials can be problematic. In addition, other aspects of asthma not captured by the NHLBI classification scheme are likely to be associated with many outcome measures. Duration of asthma diagnosis, prior treatment with corticosteroids or other long-term medication, or presence of atopy are just a few possible characteristics that may be unmeasured in a research study.

Objectives of Asthma Treatment

Since asthma is a chronic condition for which no treatment has been proven to be curative (although remissions are common), the overall goal of treatment is to control asthma symptoms as much as possible with as few adverse effects of treatment as possible. The NHLBI defines control of asthma as: 1) prevention of chronic and troublesome symptoms, 2) maintenance of (near) "normal" pulmonary function, 3) maintenance of normal activity levels, 4) prevention of recurrent exacerbations of asthma and minimization of emergency department visits, 5) minimal or no adverse effects from pharmacotherapy, and 6) fulfillment of patients' and families' expectations of asthma care.

Recognition of the need to balance control of asthma against possible adverse effects of treatment is reflected in the NHLBI treatment guidelines, which are stratified by severity of asthma. For the moderate and severe categories of asthma, different regimens are recommended for control of a potentially life-threatening condition. However, for the milder categories of asthma, and particularly the mild persistent category, controversy exists about the strategies that will best control asthma at the lowest risk of adverse effects. Both medications and dosages can be varied to obtain a certain level of control, and different patients may have different opinions regarding benefits and risks of treatment.

Another aspect of treatment addressed relates to strategies patients use to self-manage fluctuations in asthma symptoms -- asthma management plans. In addition to controlling asthma with specific medications, there are different ways to have patients manage their own asthma to control fluctuations in asthma symptoms, which may reduce exacerbations and risk of hospitalization. The objective is to have patients optimally manage fluctuations in their disease to minimize asthma morbidity.

A final pertinent objective of treatment, for the purposes of this report, is prevention of progression of asthma, in addition to concurrent control of asthma. Although patients may be well controlled on certain regimens of medications, over time asthma can worsen, as manifested by increasing doses of medications needed for the same degree of control, worsening lung function over time, increasing frequency of exacerbations, or evidence of permanent lung dysfunction. If treatment with a long-term anti-inflammatory medication were proven to prevent progression of asthma in addition to providing improved control of acute exacerbations, then it would be useful even for asthmatics with milder classes of disease, for whom other medications or combinations of medications could provide equivalent control of symptoms.

Lung Function Measures

Several measures of lung function as assessed with spirometry are commonly used in asthma studies. Instrumentation and techniques for obtaining standard lung function values have been well standardized by the American Thoracic Society, so that values obtained between studies are likely to be comparable (American Thoracic Society, 1995). However, careful attention to proper technique is critical, and full effort and cooperation of subjects is essential. Spirometry can be used to generate several possible parameters of lung function, but the most common one employed for assessing asthma is the forced expiratory volume in 1 second, FEV1. FEV1 values vary by gender, age, height, and ethnicity. In studies of children and in studies assessing changes over long periods of time, FEV1 values are usually transformed into normalized values (expressed as percent predicted) based on comparisons to large reference populations.

It is critical to note under which conditions a particular FEV1 value is obtained in order to make the correct inference from comparisons. The baseline value of FEV1 in many clinical trials is usually assessed after a washout period of several weeks, during which time patients are taken off all medications to assess the FEV1 in the absence of any treatment. It is common to obtain values at baseline and during the treatment period at regular intervals. The prebronchodilator FEV1 is measured during treatment, several hours after any daily medication that is being taken. The postbronchodilator FEV1 is measured right after taking a dose of bronchodilator administered during the testing session.

FEV1 is the most common lung function assessed in clinical trials of asthma treatment carried out in the United States. In most studies examining the benefits of asthma medication on control of asthma, baseline FEV1 is compared to prebronchodilator FEV1. In studies examining the effect of asthma medication on changes in long-term lung function, serial postbronchodilator FEV1 measurements provide the best measure of long-term changes in lung function. Serial FEV1 measurements over a relatively long time span are also a reasonable surrogate for lung growth in children as it correlates well with lung volumes, but is easier to measure (Childhood Asthma Management Program Research Group, 1999).

Bronchial hyperresponsiveness, a key element of the definition of asthma, can be directly assessed by inducing bronchial constriction with inhalations of methacholine or histamine aerosol solutions. Two common ways that the patient's bronchial hyperresponsiveness is measured are 1) the amount or concentration of solution needed to induce a 20 percent fall in FEV1 (i.e., PC 20) or 2), whether the patient has a 20 percent fall in FEV1 at a particular dose of solution. Changes in bronchial hyperresponsiveness between baseline and during treatment are thought to correlate with improvements in asthma control. However, the relationship between bronchial hyperresponsiveness and clinical asthma is complicated. Although hyperresponsiveness is modestly predictive of concurrent or future asthma, many studies show poor correlation between hyperresponsiveness and severity of asthma and adequacy of treatment.

Peak flow meters (PFMs) are used to measure the PEF. Peak flow meters are small portable devices much simpler to use than spirometers and can be used by patients at home to monitor asthma. However, there is no standardization of these devices, and values from different manufacturers' devices are not comparable. For research purposes, it is critical that a reference set of values be available to adequately assess changes over time as patients age and children grow. Over short periods of time, changes in PEF can be assessed simply as changes compared to baseline or to another prior value. Some studies rely on patients' self-report of daily or periodic PEFs as recorded at home at specific times.

Symptoms

Several methods are used to assess symptoms in patients over time.

Health Care Utilization

Certain measures of health care utilization such as emergency room visits, unscheduled physician visits, and hospitalizations represent failures of long term and acute medical interventions and other maneuvers that patients can initiate themselves. These events are relatively rare, and thus most clinical trials of small or medium size are statistically underpowered to detect differences in these outcomes, particularly in trials among patients with mild-to-moderate asthma. However, in studies examining the role of asthma self-management programs, these measures provide one of the direct intended outcomes of these interventions.

Beta-2 Adrenergic Agonists

Corticosteroids (Oral and Inhaled)

Leukotriene Modifiers

Mast-Cell Stabilizing Agents

Theophylline

Growth

There are several methods to assess growth over varying intervals of time, but each method has problems in relation to evaluating the potential effect of ICS. Over the very short term (i.e., less than 100 days), lower leg growth can be assessed using knemometry. However, knemometry does not correlate well with linear growth rate. Over periods of 6 months or more, height measured with stadiometry can provide a sufficiently precise measure called growth velocity. Finally, measurement of final adult height could be considered a gold standard -- however, other factors which may influence final adult height, especially height of parents, needs to be considered. Most studies of adult attained height use the height of both parents plus a correction factor for gender to calculate the predicted adult height. All studies that have used final adult height as an outcome are observational, and thus may have problems such as high loss to follow up, confounding by severity of asthma, and unmeasured oral corticosteroid use.

The most rigorous evidence on the effect of ICS on growth velocity comes from randomized clinical trials that used sufficiently precise growth measurement techniques and followed patients for sufficiently long periods of time (i.e., at least 1 year). A meta-analysis of such randomized clinical trials was recently published by Sharek and Bergman (2000) and will be reviewed in this evidence report in lieu of an original review of these data.

It should be noted that another meta-analysis by Allen, Mullen, and Mullen (1994) is commonly cited as evidence that attainment of expected adult height is not affected by ICS (National Heart, Blood, and Lung Institute, 1997). Examination of this analysis reveals that it does not assess adult height in most of the included studies, but measures attained height as compared to predicted height for a given age, mostly different childhood ages.

Bone Mineral Density

There are many potential laboratory measurements that could be used to assess different aspects of bone metabolism, but BMD would appear to be the best, because it can be quickly and reliably measured and is a very strong risk factor for the ultimate outcome of interest, fractures.

The problems of studying a potential adverse effect of ICS on BMD are very similar to those of studying the effect of ICS on growth. The duration of therapy and the interval for assessing BMD in clinical trials is usually very short. Evaluation of BMD in this setting results in differences in BMD that might be attributed to ICS, but may not be clinically relevant. Finally, physical activity might be affected by improved asthma treatment, which in turn has an effect on long-term changes in BMD.

Additionally complicating the matter is that BMD is constantly changing throughout life, increasing over time until early adulthood and then gradually declining, with accelerating decline in older ages. The effect of ICS on BMD may be different when the body is building up BMD at younger ages versus when BMD is generally declining at older ages. A subtle effect on BMD would not become clinically relevant until it was additive to other risk factors, such as age. Any effect possibly attributable to childhood use of ICS might be confounded by continued use of ICS into adulthood.

Ocular Toxicity

Oral corticosteroids are known to be associated with subcapsular cataracts. In children with chronic diseases taking high doses of oral corticosteroids, the prevalence of subcapsular cataracts has been noted to be between 21 and 38 percent (Tripathi, Kipp, Tripathi et al., 1992; Limaye, Pillai, and Tina, 1988). ICS could conceivably also cause cataracts either by systemic absorption or by direct contact during inhalation that might occur due to incorrect use of metered dose inhalers.

Prior research has suggested an association between ICS and cataracts among adults (Garbe, Suissa, and LeLorier, 1998, Cumming, Mitchell, and Leeder, 1997). In both studies, most cataracts occurred among the much older subjects, who could not have taken ICS as children because the drugs did not yet exist.

Thus, the question of the effect of childhood use of ICS on cataracts must necessarily be limited to effects on cataracts that occur during childhood up to early adulthood. Since the incidence of childhood cataracts is extremely low, and the expected effect of ICS is expected to be much lower than for oral corticosteroids, studies must be carefully evaluated for sufficient statistical power to study this outcome.

Ocular hypertension or glaucoma is another recognized adverse effect of topical ophthalmic corticosteroids. It occurs in 1 to 2 percent of persons above age 60. The condition in its early stages is asymptomatic and detected only by measuring intraocular pressure. An effect of sufficient size of ICS on intraocular pressure could be assessed in randomized clinical trials, because intraocular pressure can be easily and precisely measured as an additional outcome. However, if the effect is small or rare, then it would only be detectable in observational studies with sufficient statistical power such as a large case-control or cross-section study.

Two studies in older adults indicate a possible association between ICS and glaucoma (Garbe, LeLorier, Boivin et al., 1997; Mitchell, Cumming, and Mackey, 1999). These studies on adults may or may not be generalizable to children, because the baseline risk of glaucoma is very low in children.

Suppression of Adrenal/Pituitary Axis

Corticosteroids can cause hypothalamic-pituitary-adrenal (HPA)-axis suppression by reducing adrenocorticotropin (ACTH) production, via negative feedback, which in turn leads to a reduced cortisol secretion by the adrenal gland (Cave, Arlett, and Lee, 1999). Following prolonged systemic therapy, atrophy of the adrenal gland may result, resulting in subnormal response of the adrenal gland to ACTH stimulation. If severe enough, clinical symptoms can result when patients undergo stress or when corticosteroid doses are reduced.

There are three related but different phenomena that are potentially caused by ICS, that may occur independently or together, and in different degrees. Patients can show evidence of hypercortisolism or iatrogenic Cushing's, in which systemic absorption of the inhaled corticosteroid causes clinical symptoms associated with high cortisol levels, the excess being caused by the inhaled corticosteroid. Second, patients can show laboratory evidence of low cortisol levels, measured in various ways, due to the suppression of ACTH by systemic absorption of inhaled corticosteroid. Third, patients can show laboratory evidence of a subnormal response to ACTH stimulation, in that cortisol levels will not rise as high as they would normally.

Further complicating this area is the variety of tests that can be used to detect low cortisol or subnormal stimulation and the methods of analyzing the data. Cortisol can be measured by a single plasma level, multiple plasma levels over a day, or by measuring urinary cortisol over a day. Because "normal" values vary widely for all these tests, using the occurrence of abnormal values for analysis is an insensitive measure of function. Measuring change from baseline between groups is more sensitive, but then the clinical significance of any statistically significant difference is uncertain. The same problems hold for the stimulation tests. Varying the dose of the agent used to stimulate the adrenal gland can vary the sensitivity of the test, and different methods of analyzing the data can detect statistically significant differences of uncertain clinical significance.

Thus it is difficult to determine the clinical significance of any abnormal cortisol level or stimulation test without knowledge of clinical correlates of these abnormalities. The clinical presentation of iatrogenic Cushing's may or may not be accompanied by abnormal cortisol levels or abnormal stimulation tests.

Question 2: For patients with mild-moderate asthma, does early initiation of long-term controller therapy prevent progression of asthma?

Although the effectiveness of ICS in controlling symptoms of asthma and improving pulmonary function is unquestioned, an important question is whether ICS may modify the natural history of the disease. If the causal chain of events for some asthmatic patients is from inflammation to airway remodeling to irreversible airway obstruction, then anti-inflammatory medications may be able to prevent permanent declines in lung function. However, the question is whether there is evidence of a long-term benefit in terms of minimizing loss of lung function over time.

Except for the recently published Childhood Asthma Management Program (CAMP; Childhood Asthma Management Program Research Group, 2000a) study, none of the current studies cited to support this hypothesis were designed specifically to test this hypothesis. The logistics of carrying out clinical trials over sufficiently long periods of time to detect differences in asthma progression are extremely difficult. For observational cohort studies, assessments of asthma would be necessary at equivalent points in the patients' course of disease in order to adequately adjust for severity of asthma. Finally, appropriate measurements that assess changes in lung function or severity of asthma need to be obtained, versus measurements that merely assess the improvement in lung function noted when patients are initially started on ICS therapy.

Question 3: In patients with moderate asthma who are receiving ICS, does adding another long-term control agent improve outcomes?

Although ICS are recommended for patients with moderate and severe asthma, the desire to minimize corticosteroid dosage has led to consideration of use of combinations of drugs to treat these patients. Several classes of drugs can potentially be used in combination with ICS, including leukotriene antagonists, long-acting beta-2 agonists, and theophylline preparations. Patients who are successfully controlled with ICS alone may have equally effective control with a lower dose of ICS in combination with another medication.

Alternatively, patients who are inadequately controlled on a certain dose of ICS may achieve good control with the addition of another medication, rather than increasing the dose of corticosteroid. A limitation of the available literature is that studies of patients inadequately controlled on ICS did not demonstrate that patients prior to entering the study were taking the ICS consistently. Numerous studies have documented that adherence to ICS averages around 50 percent of doses prescribed (Kelloway, Wyatt, and Adlis, 1994; Sherman, Hutson, Baumstein et al., 2000) Nonetheless, ICS may not be uniformly effective in all patients; and among patients who are adherent and not well controlled on ICS, addition of other agents may increase symptom control and further reduce inflammation.

Question 4: Does addition of antibiotics to standard care improve the outcomes of treatment for acute exacerbation of asthma?

As stated in the NHLBI guidelines, antibiotics are not recommended for treatment of asthma exacerbations, but may be necessary for comorbid conditions (National Heart, Lung, and Blood Institute, 1997). According to the guidelines, bacterial respiratory tract infections are thought to contribute only infrequently to exacerbations of asthma, and antibiotics should be reserved for those patients with evidence of pneumonia, fever and purulent sputum, or evidence of bacterial sinusitis.

Although there is scant recent literature reporting the prevalence of use of antibiotics for asthma exacerbations, recent surveys of physicians seem to indicate that antibiotics may be overused in their management. A survey by Connolly, Murthy, Prescott et al. (1991) of Scottish physicians revealed that 43 to 83 percent of survey respondents believed that the risk of bacterial infection in asthma exacerbations was greater than 20 percent. The majority of respondents felt they frequently prescribed antibiotics for treatment of asthma. In a cross-national survey of physician practice, Lagerlov, Veninga, Muskova et al. (2000) report data that indicates that large proportions of physicians believe that asthma exacerbations are commonly associated with bacterial infection.

The role of bacterial infections in asthma etiology, severity, and exacerbation is currently not completely understood. Theories of bacterial involvement in asthma may be roughly differentiated by whether bacterial infections are a cause of asthma or a trigger for asthma exacerbations. According to some theories, infections such as chronic sinusitis and chlamydial infection may cause asthma and/or contribute to its severity (Cypcar, Stark, and Lemanske Jr., 1992). This aspect of bacterial infections and asthma will not be specifically addressed as part of this evidence report.

Others have investigated the role of bacterial infections not as causative agents of asthma, but as triggers for exacerbations among those with established asthma. This small body of research compares the frequency of bacterial and viral infections in acute exacerbations of asthma (MacIntosh, Ellis, Hoffman et al., 1973; Hudgel 1979; Nicholson, Kent, and Ireland, 1993; Johnston, Pattemore, Sanderson et al., 1996). These studies have reported that viral infection can be documented in up to half or more of acute asthma exacerbations, although some of these studies (Johnston, Pattemore, Sanderson et al., 1996; Nicholson, Kent, and Ireland, 1993) have included Chlamydia as a viral infection. Bacterial infections are less commonly associated with acute exacerbations as compared to viral infections in these studies, especially in children (Johnston, Pattemore, Sanderson et al., 1996). However, these studies do not rule out the possibility that a substantial proportion of asthma exacerbations are triggered by bacterial infections.

Even if bacterial infections, however defined, were demonstrated to be a trigger for asthma exacerbation in some cases, whether a routine diagnostic workup for bacterial infection should be done for all exacerbations has not been determined. It is unknown whether certain clinical criteria such as purulent or discolored sputum should be considered as presumptive evidence of bacterial infection, or whether additional diagnostic tests (e.g., cultures, gram stains) should be used to detect the presence of bacterial infection to direct antibiotic treatment.

Question 5a: Compared to medical management alone, does the use of a written asthma action plan improve outcomes, and Question 5b: Compared to a written action plan based on symptoms, does use of a written action plan based on peak flow monitoring improve outcomes?

The goal of self-management in asthma is to reduce morbidity through early recognition of signs and symptoms indicative of worsening of disease, and appropriate modification of asthma treatment to prevent and/or reverse deterioration in respiratory status. Self-management interventions have been widely promoted by health plans and specialty societies with the expectation that they will improve care. The 1997 NHLBI guidelines on treatment of asthma emphasize self-management activities as a crucial component of asthma care (National Heart, Lung, and Blood Institute, 1997). Within these recommendations, self-management programs can vary considerably. For example, among educational interventions, there is wide variability in the objectives of the programs, the duration and number of sessions, who delivers the educational sessions, the setting in which education is delivered, and the tools that are used for training (Sudre, Jacquemet, Uldry et al., 1999).

A relatively large body of literature has accumulated focusing on primarily the effect of asthma education interventions. Several literature syntheses of primary studies (Devine, 1996; Bernard-Bonnin, Stachenko, Bonin et al., 1995; Gibson, Coughlan, Wilson et al., 2000a) have been performed. These reviews have revealed mixed results. The Devine (1996) meta-analysis of 31 studies showed beneficial effects of such programs on a variety of outcomes, such as frequency of asthma attacks, PEF, functional status, adherence, use of as-needed medication, and utilization of medical services. The analysis by Bernard-Bonnin, Stachenko, Bonin et al. (1995), evaluating 23 randomized controlled trials on asthma self-management teaching programs showed little effect on the outcomes of asthma attacks, hospitalizations, emergency room (ER) visits, or school absenteeism. The analysis by Gibson, Coughlan, Wilson et al., (2000a) reviewed 11 such studies and found no effect for these limited education programs on hospitalizations, doctor visits, lung function, or medication use.

From this body of research, it is not possible to determine the impact of education alone on outcomes, nor is it possible to determine whether the impact depends on the type of educational program delivered, the type of recipient, or other specific factors. However, it is likely that educational interventions alone do not produce a large and consistent improvement in outcomes for the general population of asthmatics.

Other research studies have focused on the effect of interventions which include additional self-management activities in addition to education. This body of research shows a more consistent positive effect of self-management interventions on outcomes. A Cochrane collaboration systematic review of the effects of regular practitioner review and self-management education was completed by Gibson, Coughlan, Wilson et al. (2000b). A total of 24 controlled clinical trials were evaluated, and such programs were associated with a reduced rate of hospitalization (Odds Ratio [OR] 0.58, 95 percent Confidence Interval [CI] 0.38-0.88) and ER visits (OR 0.71, 95 percent CI 0.57-0.90).

Due to the multidimensional nature of asthma self-management interventions, it is not clear which specific components may contribute to improved outcomes. This evidence review will focus on two components, PFMs and written action plans, and whether they contribute to improved outcomes.

Peak flow meters are most commonly used in conjunction with a written action plan, indicating peak flow levels at which changes in medications and/or other interventions are triggered. Related to this, written action plans can be constructed such that triggers for change in management are either based on peak flow readings or on symptoms. Previous studies that have attempted to evaluate the utility of PFMs have not consistently demonstrated improvements in association with PFM use (Partridge, 1994; Ruffin and Pierce, 1994; Chee, 1998; National Heart, Lung, and Blood Institute, 1997). Furthermore, compliance with PFM use may be problematic. In a study of inner city asthmatics (Redline, Wright, Kattan et al., 1996), by the third week of PFM use in their study, readings were not being correctly recorded on more than 50 percent of days.

In attempting to determine the impact of specific components of multimodal disease management interventions, several challenges are present. First, isolating the specific effect of each component, such as PFM use, from the effect of other interventional components is difficult as disease management interventions are intended to be multifactorial. Second, relative to other interventions, such as optimization of medication regimens, the effect of self-management interventions is likely to be relatively small, particularly on outcomes such as lung function or symptom control. Therefore, trials of self-management interventions are prone to being underpowered, unless the researchers perform careful power calculations that include an accurate determination of the base rate for the outcomes and the expected impact of the intervention.

Study Selection Criteria

Criteria that were specific to each key question were developed for selecting studies for inclusion in this review. Following is a summary of the criteria used for defining the types of participants, types of interventions and types of studies. In general, this systematic review was limited to comparative intervention trials that used a concurrent control group. However, as described in the following sections, observational studies were included for two questions: (1) immediate versus delayed corticosteroid use; and (2) adverse effects.

Types of Participants

All patients included in this systematic review had persistent asthma requiring treatment. Where key questions addressed a population of a specified level of severity, judgements of severity level for study populations were based on the 1997 NHLBI guidelines (National Heart, Lung, and Blood Institute, 1997).

However, few if any studies used selection criteria that exactly matched specific NHLBI severity classifications. Many studies included lung function eligibility parameters that spanned two or more levels of severity. Few studies included symptom frequency, a major determinant of severity by the NHLBI system, as an eligibility criteria. As a result, the NHLBI criteria were grouped and/or modified in order to classify study populations into the following defined categories:

Types of Interventions

Except for investigation of long-term adverse events of ICS therapy, studies that compared the outcomes of managing asthma with the treatment of interest to an identified standard were required.

Types of Studies

Types of Outcome Measures

Trials were included if they reported at least one of the following outcomes, each of which were compared and analyzed separately:

1. Lung function measures:

   • FEV1

   • PEF

   • Bronchial hyperresponsiveness

2. Patient (or family) reported symptom-based measures

   • frequency of symptoms (symptom-free days, percent of days with symptoms, percent of nights with symptoms)

   • symptom scores

   • frequency of acute exacerbations

   • frequency of nocturnal awakenings

   • overall or asthma-specific quality of life

3. Utilization parameters

   • hospitalizations

   • intensive care unit admissions

   • ER and urgent care visits

   • missed work and school days

4. Medication use outcomes

   • oral corticosteroid use

   • short-acting beta-2 agonist use

5. Treatment-related morbidity in children and adolescents

   • vertical growth

   • effect on bone mineralization and osteoporosis
     -BMD
     -fractures

   • suppression of HPA axis
     -cortisol levels
     -ACTH stimulation testing (i.e., cosyntropin stimulation)

   • ocular toxicity
     -cataracts
     -glaucoma

6. Treatment-related morbidity outcomes for studies that evaluated the addition of other medications to ICS:

   • headache

   • central nervous system (CNS) morbidity (e.g., seizures) and tremors

   • cardiac dysfunction

   • gastrointestinal (GI) dysfunction: dyspepsia, nausea, vomiting, diarrhea

   • upper respiratory infections and sinusitis

   • throat irritation, hoarseness, unpleasant taste

   • sleep disorders

   • hepatic toxicity

Adverse Events

Data on adverse events were abstracted only from included studies that reported long-term adverse events for ICS use in children (Results and Conclusions, Part 1) and from included studies of the addition of treatment medication to continuing ICS therapy (Results and Conclusions, Part 3).

For each study arm, the numbers of enrolled patients experiencing specific adverse events were abstracted exactly as reported by study authors. No attempt was made to stratify according to severity, since few studies presented information on severity. If studies reported the total number of patients experiencing any adverse event, or experiencing asthma treatment-related adverse events, these data were also abstracted. Finally, the number of patients who dropped out of the study due to adverse events were abstracted separately from patients who dropped out due to disease progression or acute exacerbation. Where reported, results for these three parameters were compared between treatment arms within each study by chi-square or Fisher's exact test.

Individual adverse events were categorized in groups as described in the preceding section. Numbers of patients reported as experiencing individual adverse events were summed within each category. Because a single patient could be represented more than once in this summation, these results were only compared qualitatively for obvious differences between study arms.

Abstraction and analysis of data on adverse events present particular difficulties. The difficulties encountered in this project are representative of the general problem of the limitations of clinical trials as a source of data on adverse events. One well-recognized problem is that some adverse events may be so infrequent that clinical trials are not large enough to capture events that may be of concern when the treatment is used in the general population of patients. A second problem is inconsistency in how adverse events are reported and measured. Efforts to improve standards in reporting of randomized trials have emphasized the need for more thorough and systematic reporting of the spectrum of adverse effects for an intervention (McPeek, Gilbert, and Mosteller, 1980).

Determining Study Eligibility

Study selection was a two-stage process. All abstracts were initially reviewed by one member of the study team. Any excluded abstracts were reviewed by a second member of the study team. If the second reviewer agreed that the abstract should be excluded, then the citation was excluded. If either reviewer indicated that the abstract should be included, the article was retrieved for full review against the formal study selection criteria.

The full-length journal articles for all included abstracts were reviewed independently by two researchers using the full-length journal article selection criteria for all topics of interest. Included articles were assigned to a topic(s) of interest; excluded articles were assigned a coded reason for exclusion. Where the two reviewers were in disagreement, a third review was performed by one of the authors of this report. If no consensus was reached following the third review, the article was discussed by the entire asthma study team and a consensus decision was reached. If substantial disagreement remained after review by the entire study team, the article was brought to the TAG, and consensus reached after consultation with TAG members. The resulting bibliography of included studies was circulated to the TAG for review for possible omissions.

Reports Published in Languages Other than English

The literature search identified 343 titles and/or abstracts of reports that had an English abstract but were published in languages other than English. Abstracts were reviewed according to the abstract selection criteria. From these, 21 full-length journal articles were retrieved for review. Publication languages included Japanese, Spanish, Danish, Polish, German, French, and Chinese. A translator was identified for each language, and the article was reviewed against the inclusion criteria by the translator with the assistance of one of the study reviewers. Of the 21 articles, two met study selection criteria and were included in this systematic review.

Data Abstraction

Two reviewers independently abstracted data from each eligible study, recording it with electronic database software (Microsoft^® Access 97). The data elements that were abstracted are listed in the data abstraction forms. Data elements were grouped into the following broad categories: trial identifiers; study design and methods (including enrollment and withdrawal numbers); patient characteristics; lung function outcomes; symptom outcomes, medication outcomes, utilization outcomes, and adverse events. If an article did not report exact numerical values for one or more of the data elements, the reviewers estimated them from figures if they were available in the published reports.

Detailed printed directions for consistent data abstraction were provided to all reviewers. Initially, all reviewers abstracted a test set of three articles and reported and discussed results in detail with supervising staff. Reviewers were then divided into pairs and assigned papers for specific topics. After each pair of reviewers completed data abstraction, their databases were compared electronically. Because electronic comparison of the two databases revealed both substantive and nonsubstantive differences, it was not possible to quantify only the substantive differences. Nonsubstantive differences included differences in spelling, capitalization, wording, spacing, minor differences in estimation from graphs, and other discrepancies that were easily resolved. Substantive differences included errors in abstraction and differences in interpretation that were discussed and, in most cases, resolved by consensus of the two reviewers. In rare cases, discrepancies were resolved by a third reviewer. Frequent staff meetings allowed for discussion of common problems and further directions for consistent abstraction.

Quality Assessment for Sensitivity Analysis

The objective of quality assessment for this systematic review was to identify a group of higher quality trials, for purposes of sensitivity analysis. The meta-analysis included a quantitative sensitivity analysis, and throughout this systematic review, qualitative sensitivity analyses have been included in study conclusion summaries. The sensitivity analyses compared the results reported and conclusions reached from all included studies to results and conclusions drawn by examining the outcomes of only higher-quality studies.

Sensitivity analysis based on study quality is useful because trials of lower quality generally overestimate the effectiveness of an intervention compared to higher quality trials. Approximately two decades ago, Chalmers and coworkers showed that randomized trials report smaller treatment effects than nonrandomized studies (Chalmers, Smith, Blackburn et al., 1981). Subsequently, many methodologists have attempted to identify the characteristics that define the quality of randomized trials, and to test whether such characteristics have an effect on study results (Schulz, Chalmers, Hayes et al., 1995). Recent analyses suggest that well-designed observational studies (using either a cohort or case-control design) may produce estimates of effectiveness that are comparable to randomized controlled trials (Concato, Shah, and Horwitz, 2000; Benson and Hartz, 2000). Nonetheless, experimental design using randomized controls remains the gold standard for studies of efficacy (Pocock and Elbourne, 2000).

Although many quality scales have been used to assess the quality of randomized controlled trials, there is a dearth of empirical evidence to validate such scales. Indeed, Juni and colleagues recently illustrated the hazards of using summary quality scores to select or pool studies for meta-analysis (Juni, Witschi, Bloch et al., 1999). They identified 25 different quality scales, which they tested for a meta-analysis of 17 trials comparing low molecular weight and standard heparin. No significant association between summary quality scores and treatment effects was found; and the results of different quality scales yielded different conclusions concerning which treatment was superior.

Although the use of quality summary scores is problematic, there are three domains of study quality that have been tested in empirical studies. These are: concealment of treatment allocation during randomization; double-blinding; and handling of withdrawals and exclusions. While there is evidence suggesting that these quality domains are associated with more valid estimates of treatment effects, not all domains have been reported as significant in all studies (Mulrow and Oxman, 1997; Schulz, Chalmers, Hayes et al., 1995; Juni, Witschi, Bloch et al., 1999; Moher, Pham, Jones et al., 1998). In an editorial accompanying the Juni study, Berlin and Rennie (1999) suggested that, to be clinically relevant, quality assessment of trials should focus on key aspects of research design relative to the outcomes of interest. Thus, where an outcome requires subjective judgement, for example, assessment of asthma symptoms, double-blinding may be of paramount importance. However, double-blinding may matter less for outcomes where there is little discretion regarding assessment or interpretation.

Moreover, assessment of study quality generally depends on information reported in journal articles, and the absence of such information may reflect incomplete reporting rather than flawed study design. This point is especially germane to studies published prior to the CONSORT (Consolidated Standards of Reporting Trials) statement, which was published in the Journal of the American Medical Association in 1996, in order to disseminate a standard for completeness of reporting in journal articles (Begg, Cho, Eastwood et al., 1996). As was the case in the prior evidence reports for the Agency for Healthcare Research and Quality (AHRQ) performed by this Evidence-based Practice Center, information on concealment of allocation was reported infrequently (Aronson, Seidenfeld, Samson et al., 1999; Aronson, Seidenfeld, Piper, et al., in press). This was found to be true of recent trials, and not confined to those papers published prior to the CONSORT statement.

To supplement the general study quality characteristics that have been validated in the literature, six asthma-specific quality indicators were also developed for purposes of sensitivity analysis. These were primarily study design features to control for confounders of treatment effect relevant to the clinical setting of asthma. These included: establishing reversibility of airway obstruction, controlling for other medication use, reporting compliance, and addressing seasonality. In addition, a priori reporting of power calculations and accounting for exclusions and withdrawals were judged to be study quality characteristics pertinent to this body of evidence. A limitation of the asthma-specific quality indicators is that they have not been validated. These indicators are based on the judgement of the authors of this evidence report in consultation with the TAG.

Study Sponsorship

As a separate issue from assessment of study quality, the included trials were classified by source of research support. Using the acknowledgements of support or provision of study drug in published papers from each study and/or institutional affiliation of authors, the trials were categorized as having been funded by one of the following:

• research grants from government or other nonprofit agencies only;

• research grants from pharmaceutical manufacturers only;

• supplies from pharmaceutical manufacturers only;

• pharmaceutical manufacturers and nonprofit agency support; or

• no sponsorship reported.

Expression of Outcomes

Test for Homogeneity

For each meta-analysis, a test for homogeneity was carried out according to DerSimonian and Laird (1986).

Estimation of a Correction Factor for Studies that Reported Only Pre- and Post-Treatment Means and Standard Deviations or Standard Errors

In some cases, outcomes were reported only as mean and SD or standard error (SE) of baseline and final values, with no p-value specified. For these studies, direct calculation of effect size would be inaccurate due to an overestimate of the variances, since the pre- and post-treatment values are related, but the correlation coefficient is unknown. A correction factor for FEV1 pre- and post-treatment SDs or SEs was estimated as follows:

The true SE was estimated from Boyd (1995), in two ways. From the confidence interval of the treatment effect, SE = 0.080. From the F-value, SE = 0.090. Thus, on average, SE=0.085. The value calculated from pre- and post-treatment means and variances is 0.146. Thus SE or SD from pre- and post-treatment means and variances must be multiplied by 0.58 to get the correct value.

From the van der Molen, Postma, Turner et al., (1997), study, the true SE estimated from the confidence interval of the treatment effect, is 0.050. The value calculated from pre- and post-treatment means and variances is 0.113. Thus, SE or SD from pre- and post-treatment means must be multiplied by 0.44 to get the correct value.

The correction factor was averaged to approximately 0.5 and applied to the SDs or SEs of those studies for which effect size was calculated using pre- and post-treatment data (FitzGerald, Chapman, Della Cioppa et al., 1999; Boulet, Cartier, and Milot, 1998; Grutters, Brinkman, and Aslander, 1999; Li, Ward, Thien et al., 1999; McIvor, Pizzichini, Turner et al., 1998; Pauwels, Lofdahl, Postma et al., 1997; Bouros, Bachlitzanakis, Kottakis et al., 1999; Kips, O'Connor, Inman et al., 2000).

Similar calculations were carried out for PEF, using data from studies by Boyd (1995), and Bouros, Bachlitzanakis, Kottakis et al. (1999). Estimates of the correction factor were 0.37 and 0.26, respectively. A conservative SD/SE correction factor of 0.5 was applied to calculations of effect size from two small studies by Li, Ward, Thien et al. (1999) and McIvor, Pizzichini, Turner et al. (1998).

Calculation of Average Standard Deviation to Convert Effect Size to Clinically Meaningful Units

Effect sizes are unitless and not clinically meaningful. However, they can be converted to clinical units by multiplying by a pooled SD of the change from baseline parameter reported in the desired units. Pooled SD values were calculated from hseveral studies that reported applicable data for both liters and percent predicted for FEV1, and for liters per minute and percent predicted for PEF as follows:

• Pooled SD for FEV1 in liters calculated from Kavuru, Melamed, Gross et al. (2000), Condemi, Goldstein, Kalberg et al. (1999), and Murray, Church, Anderson et al. (1999) was 0.521.

• Pooled SD for FEV1 as percent predicted calculated from Cluzel, Bousquet, Daures et al. (1990), Grossman (1988), and Baraniuk, Murray, Nathan et al. (1999) was 11.1.

• Pooled SD for PEF in liters per minute calculated from Aubier, Pieters, Schlosser et al. (1999), Nielsen, Pedersen, Faurschou et al. (1999), Pearlman, Stricker, Weinstein et al. (1999), Greening, Ind, Northfield et al. (1994), Kelsen, Church, Gillman et al. (1999), and Tamaoki, Kondo, Sakai et al. (1997) was 42.5.

• Pooled SD for PEF as percent predicted calculated from Shapiro, Lumry, Wolfe et al. (2000) and Baraniuk, Murray, Nathan et al. (1999) was 12.5.

Patient Populations

In seven of the 10 studies, the mean age was similar, in the range of 9-12 years, with -1SD approximately 7 years of age in most studies. For two of the studies of ICS versus no ICS (Storr, Lenney, and Lenney, 1986; Connett, Warde, Wooler et al., 1993), enrollment was restricted to patients younger than 5 years of age, with mean ages of 3.5 and 1.8 years, respectively. The results of these studies in very young children will be reported separately. In the tenth trial (Agertoft and Pedersen, 1994), the mean age was approximately 6 years (range 3-11 years), but results of older and younger children were not reported separately. This study is the only one that appears to overlap the categories for children older and younger than 5 years of age.

The study eligibility criteria varied, with various combinations of lung function, symptom-based and utilization-based eligibility criteria. Eight trials included symptom-based eligibility criteria, six trials had lung function measures as eligibility criteria, and three included utilization based measures. Specific criteria within these broad categories varied as well. For example, among the studies using lung function eligibility, five of six used FEV1, with the minimum FEV1 ranging between 50 and 75 percent. One of the six trials (Childhood Asthma Management Program Research Group, 2000a) used only bronchial hyperreactivity as a lung function eligibility criterion.

Severity of illness at the time of enrollment was estimated using the NHLBI classification system to the extent possible given the information contained in the reports. Only one study was judged to be restricted to patients with mild asthma (Jonasson, Carlsen, Blomqvist et al., 1998). Four of the studies included a population predominantly in the mild-moderate range (Childhood Asthma Management Program Research Group, 2000a; Simons, 1997; Hoekstra, Grol, Hovenga et al., 1998; Verberne, Frost, Roorda et al., 1997), while three studies included patients with severity ranging from mild to severe (Agertoft and Pedersen, 1994; van Essen-Zandvliet, Hughes, Waalkens et al., 1992; Tinkelman, Reed, Nelson et al., 1993). In the two studies of very young children (Storr, Lenney, and Lenney, 1986; Connett, Warde, Wooler et al., 1993), severity could not be estimated due to a lack of sufficient data on lung function and/or symptom levels. However, both of these trials selected patients whose symptoms were judged to be inadequately controlled, making it likely that these patients were representative of the more severe end of the disease spectrum.

Baseline mean FEV1 was reported for the eight studies that enrolled children over 5 years of age, ranging from a mean of 74.1 to 105 percent of predicted. In four studies, it was not specifically stated whether this was a pre- or postbronchodilator measure; in most cases it was probably a prebronchodilator measure. In the remaining four studies (Childhood Asthma Management Program Research Group, 2000a; van Essen-Zandvliet, Hughes, Waalkens et al., 1992; Verberne, Frost, Roorda et al., 1997; Tinkelman, Reed, Nelson et al., 1993) both pre- and postbronchodilator mean baseline FEV1 values were reported. The difference between pre- and postbronchodilator measures ranged from 8.9 to 19.1 percent predicted, which in several cases could change the classification of severity, depending on which measure was used. For the purpose of classifying severity in this evidence report, the prebronchodilator measure was used where both were reported, since this was used most consistently and allowed better comparison of severity level across studies.

Six of the trials reported baseline symptom scores and/or symptom frequency measures. Because of differences in units and type of reporting, these measures were not helpful in comparing severity levels across studies.

Interventions

All 10 trials included treatment with ICS in at least one study arm. The control arms of these studies all included treatment with short-acting beta-2 agonists on an as-needed basis, thus making the comparison primarily ICS versus as-needed beta-2 agonists alone. Three different ICS agents were employed: budesonide in five studies, beclomethasone in four studies, and fluticasone in one study.

Using current classification schemes for ICS dose level (NHLBI), 6 of the 10 studies used doses within the medium dose range. One study of children older than 5 years with three treatment groups (Jonasson, Carlsen, Blomqvist et al., 1998) used dosages of budesonide within the low range (100-200 mcg/d). A second study, in children younger than 5 years, used a low dose of beclomethasone (330 mcg/d). Two studies, both in children older than 5 years, used ICS dosages in the high range. Agertoft and Pedersen (1994) treated patients with 800 mcg/d of budesonide, while van Essen-Zandvliet, Hughes, Waalkens et al. (1992) treated patients with 600 mcg/d of budesonide. In the two studies that used salmeterol in one of the study arms, the dose was 50 mcg twice per day. Nedocromil was administered at a dose of 8 mg twice per day in the CAMP study (Childhood Asthma Management Program Research Group, 2000a), and the theophylline dosage was titrated to blood levels by Tinkelman, Reed, Nelson et al. (1993).

Outcomes

• Lung function outcomes were reported in 8 of the 10 studies. The two studies that did not report lung function outcomes (Storr, Lenney, and Lenney, 1986; Connett, Warde, Wooler et al., 1993) were the studies with patients younger than 5 years of age, in which performing lung function measures is not feasible. All eight of these studies reported FEV1 outcomes and all but one (Agertoft and Pedersen, 1994) reported PEF outcomes. The units of reporting varied (percent predicted, absolute value in liters or L/min) for FEV1 and PEF, although the majority of studies reported FEV1 as percent predicted and PEF as L/min. Five studies reported bronchial hyperresponsiveness outcomes in various ways (e.g., mg of medication required for PC20, doubling dose).

• Nine of the 10 studies reported on symptom outcomes, either as symptom scores or symptom frequencies. Eight studies reported some measure of symptom frequency, either as the percentage of days and/or nights with symptoms, the percentage of days needing rescue medication, or the percentage of days with a symptom score greater than a threshold level. Five studies reported daytime symptom scores and three reported nighttime symptom scores. The units of the symptom scores varied considerably, with one study using a 0-2 scale, two studies using a 0-3 scale, one study using a 0-6 scale, and one study using a 0-32 scale.

• Medication use outcomes were reported in 8 of the 10 studies. Six of these reported oral corticosteroid usage and five reported beta-2 agonist usage.

• Utilization outcomes were reported in 5 of the 10 studies. The utilization outcomes reported were either hospitalizations (reported as number of patients with event, number of events/person, total number of events over study), or missed days of work/school (reported as percent of patients with any missed days, or total number of missed days over entire study).

Study Quality

Quality of study design and conduct were assessed as described in the "Methodology" chapter. The objective was to identify a group of higher quality trials for purposes of sensitivity analysis. The definition for higher quality studies is applicable only to randomized controlled trials and excluded nonrandomized controlled trials and single arm studies. It includes general quality indicators that have been shown to be associated with a bias in magnitude of effect, and asthma specific study features that control for potential confounders of outcomes.

To be defined as a higher quality study, a trial needed to meet three general quality indicators: (1) double blinding; (2) appropriate handling of exclusions and withdrawals as demonstrated by percentage of excluded patients below threshold or results analyzed by intent-to-treat analysis; and (3) concealment of treatment allocation.

In addition, the presence of six features specific to the setting of asthma was assessed. The first was that power calculations for primary outcomes were specified prospectively. The second criterion was whether the study accounted for the reasons that patients withdrew from the study, particularly regarding the number of patients that were withdrawn due to lack of efficacy. Next, the presence of specific study features designed to control for potential confounders of outcome was assessed. These were: (1) whether reversibility of lung obstruction was established at study entry; (2) whether use of asthma medications other than the study medication was controlled for; (3) whether measures of patient compliance were reported; (4) and whether the influence of seasonal differences on outcomes was addressed.

Table 7. Assessment of study quality

Table 7. Assessment of study quality

	General Quality Indicators				Asthma-Specific Quality Indicators
Citation	Blinding	Percentage of excluded subjects below specified threshold?	Intent-to-treat analysis?	Allocation concealed? (NS=not specified)	Power calculations?	Accounted for excluded patients?	Revers-ibility estab-lished?	Controlled for other medication use?	Reported compliance?	Addressed season-ality?
Childhood Asthma Management Program Research Group, 2000a	Yes	Yes	Yes	Yes	Yes	NAa	Yes	Yes	Yes	No
Jonasson, Carlsen, Blomqvist et al., 1998	Yes	Yes	Yes	NS	Yes	Yes	No	NR	No	No
Simons, 1997	Yes	No	No	NS	No	No	Yes	No	Yes	No
Hoekstra, Grol, Hovenga, et al., 1998	Yes	No	No	NS	No	Yes	Yes	Yes	Yes	No
van Essen-Zandvliet, Hughes, Waalkens, et al., 1992	Yes	No	No	Yes	No	Yes	Yes	Yes	No	No
Storr, Lenney and Lenney, 1986	Yes	No	No	NS	Yes	Yes	NAb	NR	No	Yes
Connett, Warde, Wooler, et al., 1993	Yes	No	No	NS	Yes	Yes	NAb	NR	No	No
Verberne, Frost, Roorda et al., 1997	Yes	No	No	Yes	Yes	Yes	Yes	Yes	Yes	No
Tinkelman, Reed, Nelson, et al., 1993	Yes	No	No	NS	Yes	Yes	Yes	Yes	No	No

a

No patients were excluded from analysis.

b

Studies in children under 5, where lung function assessment is infeasible.

Of the eight other randomized controlled trials, all were double-blinded but only the trial by Jonasson, Carlsen, Blomqvist et al. (1998) met the criterion for percentage of subjects excluded from analysis below the specified threshold. This study also analyzed results by intent-to-treat analysis, but did not specify whether allocation to treatment arm was concealed. Jonasson, Carlsen, Blomqvist et al. (1998) also reported power calculations and accounted for excluded patients, but did not fulfill any of our other asthma specific study quality indicators.

A high number of patients excluded from the analysis of results was typical for this group of studies. With one exception (Simons, 1997), all gave an accounting of the reasons patients were excluded from the analysis. Overall, the preponderance of exclusions from analysis were patients withdrawn for the placebo arm due to lack of treatment effect (i.e., symptoms, exacerbations). Thus, the relatively high withdrawal rates from the placebo arm are an indicator that ICS treatment is more effective in controlling the symptoms of asthma. Only the CAMP (Childhood Asthma Management Program Research Group, 2000a) and Jonasson, Carlsen, Blomqvist et al. (1998) studies used an intent-to-treat analysis to control for bias related to withdrawals. Several other studies stated that intent-to-treat analysis was used, but it was evident from close review of the results sections of these papers that this was not the case.

Two trials other than CAMP reported on allocation concealment (van Essen-Zandvliet, Hughes, Waalkens et al., 1992; Verberne, Frost, Roorda et al., 1997); overall, six of the nine trials did not specify whether allocation to treatment arm was concealed. Verberne, Frost, Roorda et al. (1997) also met all the asthma-specific quality indicators, except addressing the effects of seasonality. van Essen-Zandvliet, Hughes, Waalkens et al. (1992) fulfilled three of the six asthma-specific indicators: accounting for excluded patients, establishing reversibility of lung obstruction and controlling for other medication use.

Five trials met no general quality criteria other than double-blinding (Simons, 1997; Hoekstra, Grol, Hovenga et al., 1998; Storr, Lenney, and Lenney, 1986; Connett, Warde, Wooler et al., 1993; Tinkelman, Reed, Nelson et al., 1993), and three of the five reported power calculations. Tinkelman, Reed, Nelson et al. (1993) also established reversibility, accounted for excluded patients, and controlled for other medication use. The trials by Storr, Lenney, and Lenney (1986) and Connett, Warde, Wooler et al. (1993) were of younger children, where lung function tests cannot be preformed to establish reversibility. Both accounted for excluded patients; Storr, Lenney, and Lenney (1986) also addressed seasonality. Neither study controlled for other medications or reported compliance. Storr, Lenney, and Lenney (1986) is the only trial of the 10 included in this key question that addressed seasonality. Of the two remaining trials, Hoekstra, Grol, Hovenga et al. (1998) and Simons (1997) both established reversibility and reported on compliance. But only Hoekstra, Grol, Hovenga et al. (1998) accounted for excluded patients and controlled for other medication use.

Trials Comparing ICS to "As-Needed" Beta-2 Agonists: Children Older Than 5 Years

There were six trials in this category, enrolling a total of 790 patients treated with ICS and 652 controls; 40 percent of ICS patients (n=311) and 64 percent of controls (n=418) were contributed by the CAMP study (Childhood Asthma Management Program Research Group, 2000a). Except for Agertoft and Pedersen (1994), all trials were randomized, double-blinded, and placebo-controlled. Agertoft and Pedersen (1994) enrolled 216 ICS patients but only 62 controls; which comprises 27 percent and 10 percent, respectively, of the total population of the included studies. Asthma severity in these studies was generally mild to moderate. Three studies had a population that was estimated to be confined to mild-to-moderate patients (Childhood Asthma Management Program Research Group, 2000a; Simons, 1997; Hoekstra, Grol, Hovenga et al., 1998); together, these studies contributed 50 percent of ICS patients and 75 percent of controls. One study, which contributed 16 percent of ICS patients and 6 percent of controls, had a population that was clearly limited to mild asthma (Jonasson, Carlsen, Blomqvist et al., 1998). Two studies, enrolling 35 percent of ICS patients and 18 percent of controls, had populations spanning the range of severity from mild to severe (Agertoft and Pedersen, 1994; van Essen-Zandvliet, Hughes, Waalkens et al., 1992). The range of mean baseline FEV1 was 75.7 to 105 percent predicted.

Most patients (estimated 90 percent) were followed for a year or longer. Three trials reported followup greater than 1 year; 224 weeks in CAMP (Childhood Asthma Management Program Research Group, 2000a), mean of 192.4 weeks in the ICS arm of Agertoft and Pedersen (1994), and median of 95.3 weeks in van Essen-Zandvliet, Hughes, Waalkens et al. (1992). Simons (1997) reported 52 weeks of followup; the trials by Jonasson, Carlsen, Blomqvist et al. (1998) and Hoekstra, Grol, Hovenga et al. (1998) were each only 12 weeks in duration.

Trials Comparing ICS to "As-Needed" Beta-2 Agonists: Children Younger Than 5 Years

Two studies (n=69) compared ICS to placebo in children less than 5 years of age, with treatment duration of 26 weeks (Storr, Lenney, and Lenney, 1986; Connett, Warde, Wooler et al., 1993). The main outcomes reported in these trials of very young children were symptom-based outcomes. Both reported symptom score and symptom frequency outcomes as recorded by the parents or caretakers. Both trials also reported beta-2 agonist use. Connett, Warde, Wooler et al. (1993) also reported oral corticosteroid usage and the total number of hospital visits. Lung function outcomes were not reported, as these are infeasible to measure in children younger than 5 years of age.

Both trials reported statistically significant differences favoring ICS in some symptom subscores. Storr, Lenney, and Lenney (1986) measured three symptom subscores: daytime wheezing, nighttime wheezing, and cough, on a 0-3 scale. Significant differences were found in favor of ICS on the final scores for daytime wheezing (0.26 vs. 0.33, p<0.05) and nighttime wheezing (0.26 vs. 0.35, p<0.05). Connett, Warde, Wooler et al. (1993) reported on five symptom score subscales: daytime cough, nighttime cough, daytime wheeze, nighttime wheeze, and days of limited activity, each measured on a 0-2 scale. Significant differences were found in favor of ICS on the change in two of these subscales, daytime cough (−0.5 vs. 0.05, p<0.03) and nighttime cough (−0.4 vs. 0.07, p<0.05). No differences were found on daytime or nighttime wheeze or on days of limited activity.

Other measures favored ICS use, but were not consistently statistically significant in these two small trials. Storr, Lenney, and Lenney (1986) reported significantly less beta-2 agonist use for the ICS group as compared to placebo (0.52 puffs/day vs. 0.98 puffs/day), but the difference was not significant in the trial by Connett, Warde, Wooler et al. (1993). Storr, Lenney, and Lenney (1986) found no differences in the percentage of symptom-free days or the percentage of symptom-free nights. But Connett, Warde, Wooler et al. (1993) found that the ICS group had a significantly greater percentage of symptom-free days as compared to placebo (54 percent vs. 31 percent, p<0.0001). The Connett, Warde, Wooler et al. (1993) study found less oral corticosteroid use and fewer hospitalizations in the ICS group, but the differences were not statistically significant.

Trials Comparing ICS with Salmeterol

Two randomized and double-blinded trials compared ICS to salmeterol in children (Verberne, Frost, Roorda et al., 1997; Simons, 1997). One of these (Verberne, Frost, Roorda et al., 1997) was designed as a direct comparison between the two agents. The second trial (Simons, 1997) was a three-arm study in which both ICS and salmeterol were compared with placebo, but for most outcomes, direct statistical comparisons were not reported between ICS and salmeterol. An indirect comparison of the two agents can be made by comparing the relative efficacy of each with placebo and by examining the magnitude of difference in outcomes between the two medications. Both of these trials included patient severity levels in the mild-to-moderate range and were of approximately 1-year duration. The total number of patients enrolled was 308 (including 80 patients in the placebo arm of Simons, 1997), with 237 patients evaluable. Both reported lung function outcomes, symptom frequency outcomes, and medication use outcomes. Simons (1997) also reported on the percentage of patients with missed school days.

Verberne, Frost, Roorda et al. (1997) found a significant difference in favor of ICS on the change in FEV1 over the course of the study. This study reported both pre and postbronchodilator FEV1 values and found a significant difference on both measures. There was a difference of 12 percent in the final prebronchodilator FEV1 in favor of the ICS group compared to salmeterol (95 percent vs. 83 percent, p<0.0001). A significant difference in postbronchodilator FEV1 was also reported, although this difference of 5.0 percent was a smaller absolute benefit in favor of the ICS group (102.5 percent vs. 97.5 percent, p=0.007). In contrast, the second study (Simons, 1997) reported an identical change in FEV1 (+10 percent predicted) for the ICS and salmeterol groups, and both were statistically significant compared to compared to placebo (+5 percent, p=0.001). Simons (1997) did not state whether the measure was pre or postbronchodilator.

Both trials reported PEF outcomes. Verberne, Frost, Roorda et al. (1997) reported a rise in PEF of 60.9 L/min in the ICS group as compared to 48.8 L/min in the salmeterol group, a difference that was not statistically significant. Simons (1997) reported a rise in PEF of 35 L/min for the ICS group and a slightly higher rise of 41 L/min for the salmeterol group, both significant compared to placebo. For PC20 outcomes, Verberne, Frost, Roorda et al. (1997) reported a significant difference in favor of the ICS group (increase of 2.02 doubling doses, as compared to a decrease of 0.73 doubling doses, p<0.0001). Simons (1997) reported change in mean mg of methacholine required for PC20, and reported a statistical comparison between ICS and salmeterol for this outcome in favor of ICS (increase of 1.37 mg vs. increase of 0.84 mg, p=0.01).

Symptom frequency measures were reported by Verberne, Frost, Roorda et al. (1997) as days/week with symptoms, nights/week with symptoms, and the percentage of patients with no symptoms over a 2-week period. There were no differences between groups in days/week or nights/week with symptoms. The percentage of patients with no symptoms over a 2-week period was 55 percent for the ICS group and 36 percent for the salmeterol group. This comparison was reported as "only significant at some time points," but not at others. Simons (1997) reported the percent of days in which beta-2 agonist was not required as rescue medication: 92 percent for ICS and 88 percent for salmeterol. The difference between the ICS group and the placebo group (92 versus 83 percent, p<0.001) was statistically significant, while the difference between the salmeterol group and the placebo group was not (88 percent vs. 83 percent, p=NS). There was no statistical comparison reported between the ICS and salmeterol groups.

In both studies, overall withdrawals and withdrawals for exacerbation were higher in the salmeterol group than the ICS group. Verberne, Frost, Roorda et al. (1997) reported that 10 patients withdrew from the trial. Of the seven patients who withdrew because of exacerbations, six were in the salmeterol group. Simons (1997) reported that overall withdrawals were 17 percent in the ICS group, 28 percent in the salmeterol group, and 31 percent in the placebo group (ICS vs. placebo, p=0.03). The percent of patients for whom the primary reason for withdrawal was asthma exacerbation was 5 percent in the ICS group, 15 percent in the salmeterol group, and 15 percent in the placebo group.

Beta-2 agonist use, as measured by the median number of puffs per day during the treatment period, was reduced in the ICS group as compared to salmeterol in the Verberne, Frost, Roorda et al. (1997) study (0.07 puffs/day vs. 0.44 puffs/day, p=0.0001). Simons (1997) reported that median percent albuterol-free days was significantly lower for ICS compared to placebo (92 vs. 83, p<0.001), but the difference between salmeterol and placebo (88 vs. 83) was not significant. Oral corticosteroid usage was reported in both studies. Both studies reported fewer courses of oral corticosteroids in the ICS group (2 versus 17 courses in Verberne, Frost, Roorda et al., 1997, 10 vs. 15 courses in Simons, 1997), but neither study reported a statistical comparison. Simons (1997) reported on the percentage of patients who did not miss any school days due to asthma, 88 percent in the ICS group vs. 81 percent in the salmeterol group and 66 percent in the placebo group. No statistical comparisons were reported for this outcome.

Trials Comparing ICS with Theophylline

One trial compared ICS use to theophylline (Tinkelman, Reed, Nelson et al., 1993). This was a 36-week trial enrolling 195 patients whose asthma severity ranged from mild to severe. Outcomes reported included pre- and postbronchodilator FEV1, PEF, bronchial hyperreactivity, symptom scores, symptom frequencies, oral corticosteroid usage, and ER visits. There was a high dropout rate in both arms of this trial: 25 percent of the patients in the ICS arm and 26 percent in the theophylline arm were not included in the final analysis.

There were no statistically significant differences found between groups on the majority of the outcome measures abstracted. Prebronchodilator FEV1 increased to a similar degree in both groups, while post bronchodilator FEV1 remained largely unchanged in both groups. PEF improved by 6 percent predicted in the ICS group compared with 2 percent predicted in the theophylline group (p=NS). The PC20 also showed a larger increase in the ICS group (9.04 mg methacholine vs 3.7 mg methacholine) but this difference did not reach statistical significance either. Baseline and final symptom scores were virtually identical between groups. Similarly, there were no significant differences in symptom frequencies, or ER visits. The ICS group had less oral corticosteroid use as compared to the theophylline group, with 81.4 percent of patients in the ICS group not requiring oral corticosteroids, as compared to 63.4 percent of patients in the theophylline group (p=0.007).

Trials Comparing ICS with Nedocromil/Cromolyn

The third arm of the CAMP (Childhood Asthma Management Program Research Group, 2000a) study compared nedocromil to placebo, enrolling 312 patients in the nedocromil arm and comparing outcomes to the 418 patients in the placebo group. As described previously, this was a population of mild-to-moderate asthmatics followed for over 4 years. No direct comparisons of the ICS arm with the nedocromil arm were reported. However, examination of the comparison to placebo in each of the two treatment arms allows an indirect determination of the relative efficacy of ICS vs. nedocromil

The majority of comparisons of nedocromil vs. placebo were not significantly different. These included pre- and postbronchodilator FEV1, PEF, PC20, symptom scores, symptom frequencies, and beta-2 agonist use. There were two outcome measures that showed a significant difference in favor of the nedocromil group. The amount of oral corticosteroid use was less in the nedocromil group (102 courses/100 patient-years vs. 122 courses/100 patient-years, p=0.01). The frequency of ER use was also lower in the nedocromil group as compared to placebo (16 visits/100 patient-years vs. 22 visits/100 patient-years, p=0.02).

Inhaled Corticosteroids Compared to As-Needed Beta-2 Agonists

Inhaled Corticosteroids Compared to Alternative Long-Term Control Medications

Effect on Vertical Growth

Evidence was found on three measures of vertical growth in children: short term growth velocity measured over a period of 1 year or less; growth velocity and change in height measured over longer duration (4-6 years), and final attained adult height. The evidence on short-term growth velocity is from a published meta-analysis that pooled data from five randomized controlled trials representing 855 subjects, with a mean age of 9.5 years (Sharek and Bergman, 2000). Evidence on growth velocity and height over longer duration is from the CAMP trial (Childhood Asthma Management Program Research Group, 2000a), a randomized trial comparing ICS, nedocromil, and placebo in 1,041 children with mild-to-moderate asthma who were followed for 4 to 6 years. For final attained adult height, evidence is from three retrospective cohort studies that adjusted for the potential confounding factor of parental height (Agertoft and Pedersen, 2000; Silverstein, Yunginger, Reed et al., 1997; Van Bever, Desager, Lijssens et al., 1999). Together, these three studies included a total of 243 asthmatics treated with ICS, 154 asthmatics who had not been treated with ICS, and 204 nonasthmatic controls.

Conclusions on Vertical Growth

The body of research regarding the effect of ICS on growth velocity over a relatively short period of 1 year is consistent in showing a difference of average height of 1cm/year over a period of about 1 year. This difference is demonstrated in several studies using the most rigorous study design, i.e., randomized clinical trials. In the one clinical trial extending beyond 1 year (Childhood Asthma Management Program Research Group, 2000a), a difference consistent with this magnitude also occurred in the first year of the study. However, in subsequent long-term followup, the difference in growth velocity was not maintained, and by the end of the 4- to 6-year observation period, there was still an approximately 1 cm difference in cumulative growth between the study groups.

Table 8. Difference between adult-target height

Table 8. Difference between adult-target height

Study	Group (n) Comparison	Difference in (Adult-Target) Height (cm) a
Silverstein, Yunginger, Reed et al. (1997)	All asthmatics (n=153) vs. nonasthmatics (n=153)	0.2
	All corticosteroid users (n=58) vs. noncorticosteroid asthmatics (n=95)	−1.2
	Males: All corticosteroid users (n=30) vs. noncorticosteroid asthmatics (n=45)	−1.8
	Females: All corticosteroid users (n=28) vs. noncorticosteroid asthmatics (n=50)	−0.8
	Oral corticosteroid users (n=40) vs. never used corticosteroids (n=95)	−1.4
	Inhaled corticosteroid users (n=18) vs. never used corticosteroids (n=95)	−0.9
Van Bever, Desager, Lijssens et al. (1999)	All inhaled corticosteroid users (n=43) vs. never used corticosteroids (n=42)	−2.54 b
	Males: inhaled corticosteroid users (n=23) vs. never used corticosteroids (n=26)	−3.09 b
	Females: inhaled corticosteroid users (n=20) vs. never used corticosteroids (n=16)	−1.99
Agertoft and Pedersen (2000)	All inhaled corticosteroid users (n=142) vs. noncorticosteroid using asthmatics (n=18)	+0.5
	All inhaled corticosteroid users (n=142) vs. healthy sibling control group (n=51)	−0.6
	Males: All inhaled corticosteroid users (n=86) vs. healthy sibling control group (n=24)	−0.6
	Females: All inhaled corticosteroid users (n=56) vs. healthy sibling control group (n=27)	−0.8

a

A negative number indicates that corticosteroid users had lower attained adult height than the comparison group, controlling for parental height.

b

p<0.05

The evidence regarding final adult height appears to be fairly consistent as well. However, this evidence is based on retrospective cohort studies, where the confounding effects of severity of asthma cannot be adjusted for. Selection biases cannot be assessed in retrospective studies, because only subjects who are available at the end of the study can be assessed. More severe asthma patients are more likely to use ICS. Finally, the small sample size of ICS users in the Silverstein, Yunginger, Reed et al. (1997) study make the conclusions of a negative study less firm. The studies of Silverstein, Yunginger, Reed et al., (1997) and Agertoft and Pedersen (2000) showed no difference in final attained adult height, whereas the study of Van Bever, Desager, Lijssens et al. (1999) showed a difference between ICS users and nonusers. However, the difference in adult height in the Van Bever, Desager, Lijssens et al. (1999) study, 2.54 cm, is much less than would be expected if the 1 cm/year growth velocity difference demonstrated in the short-term studies was maintained over several years (Table 8).

The differences between the studies on short-term growth velocity and final attained adult height appear to be explained by the fact that the initial difference in growth velocity is not sustained over longer periods of time. Although there is no evidence for an initial compensatory acceleration of growth rate (i.e., "catch-up" growth), after the first year, there appears to be resumption of normal growth rates.

Effect on Bone Mineral Density

Selection criteria for this outcome required that BMD alone be considered the appropriate outcome, and that studies have sufficient size (n=25) in each group for sufficient comparison. Studies that evaluated adults were included if it was felt that the outcomes largely reflected treatment in childhood or young adulthood. Studies of adults were thus included if the mean age of the population was less than 40 years and if the duration of asthma and/or ICS treatment was sufficiently long enough to conclude that ICS exposure had primarily occurred during childhood or early adulthood. Two of the studies are cross-sectional studies that evaluate BMD at a single time, and one of the studies is a randomized clinical trial which assesses changes in BMD over a period of 4-6 years.

Agertoft and Pedersen (1998) compared 157 asthmatic children treated with inhaled budesonide at a mean daily dose of 504 mcg to 111 age-matched children also suffering from asthma but who had never been treated with oral corticosteroids for more than 14 days. The children receiving ICS had been taking medication for 3 to 6 years. The mean age of the ICS users was 10.3 years. There was no difference in mean total body BMD between children taking ICS and the control group (0.92 g/cm² vs. 0.92 g/cm²). In addition, there were no significant differences in bone mineral capacity and total bone calcium.

Ip, Lam, Yam et al. (1994) studied 30 young adults with a mean duration of asthma of 14 years who had been on ICS an average of 40 months (range: 3-180 months). They were compared to a control group without asthma matched on sex, age, body mass index and menopausal status. BMD measured at the spine, femoral neck, and hip of the ICS users were all significantly lower than that of the control patients. When the subjects were stratified by sex, only females showed a significant difference in all BMD measurements. Among the female patient group, there was a significant correlation between the average daily dose of ICS and BMD of the lumbar spine (r= −0.46, p=0.054) and BMD of the femoral trochanter (r= −0.47, p=0.047).

The one randomized clinical trial, the CAMP study (Childhood Asthma Management Program Research Group, 2000a) evaluated 1,041 children who received budesonide, nedocromil, or placebo. Mean age was 9.5 years and BMD of the spine was assessed at baseline and the end of followup at 4 to 6 years. None of the patients groups had significantly different changes in spinal BMD (budesonide 0.17 g/cm², nedocromil 0.17 g/cm², placebo 0.17 g/cm²).

Conclusions on Bone Mineral Density

Table 9. Effects of ICS on bone mineral density

Table 9. Effects of ICS on bone mineral density

Citation	Treatment Arm	N Enrolled	N Evaluable	Treatment Duration (years)	Bone Density Result	p Value	Comment
Agertoft, Larsen and Pedersen, 1998	Budesonide 504 mcg per day	157	157	3.0 (minimum)	Total body BMD: 0.92 g/cm2		No significant difference between groups or between boys and girls in bone mineral capacity, or total bone calcium
	Nonsteroid asthma therapies	111	111	3.0 (minimum)	Total body BMD: 0.92 g/cm2	NS	Mean treatment time 4.4 (3-6) yrs
Ip, Lam, Yam, et al., 1994	Beclomethasone or budesonide	30	30	3.3	Spine: Femur neck: Trochanter: Ward's triangle: 0.944 0.769 0.676 0.729	0.041 0.007 0.034 0.016	Stratified by sex, all differences significant for females, but not for males
	Normal control subjects, matched by sex, age, BMI, menopausal status	30	30	NA	Spine: Femur neck: Trochanter: Ward's triangle: 1.011 0.835 0.724 0.839
Childhood Asthma Management Program Research Group, 2000a	Budesonide 400 mcg/day	311	311	4-6 yr	Change in spine BMD: 0.17 g/cm2	0.53 vs. placebo
	Nedocromil 16 mg/day	312	312	4-6 yr	Change in spine BMD: 0.17 g/cm2	0.15 vs. placebo
	Placebo	418	418	4-6 yr	Change in spine BMD: 0.18 g/cm2

The CAMP study (Childhood Asthma Management Program Research Group, 2000a), with large numbers, randomization, assessment of longitudinal changes, provides very strong evidence that there is no effect of ICS on BMD in the doses given in that study (Table 9). The other studies are subject to potential confounding because of unmeasured differences between groups that are risk factors for low BMD. In the Ip, Lam, Yam et al. (1994) study, patients' BMD was measured during adulthood, making it uncertain as to whether the effect of ICS was due to an effect during childhood (i.e., slowing the increase in BMD during growth) or adulthood (i.e., accelerating the loss of BMD due to aging). In addition, the clinical significance of the differences in BMD are unknown. Subtle differences in BMD would not have clinical impact until additive to other risk factors such as aging. It is uncertain whether differences observed during young adulthood would persist to old age, or whether small changes in childhood would lead to clinically meaningful differences in fracture risk at older ages.

Effects on Posterior Subcapsular Cataract and Glaucoma

A total of seven studies enrolling approximately 1,000 asthmatic subjects receiving ICS therapy examined occurrence of subcapsular cataracts. Because the incidence and prevalence of cataracts is expected to be zero in children and young adults, clinical trials, comparative cross-sectional studies, and single-group cross-sectional studies were all selected for review. Three studies were randomized clinical trials that examined subjects for incidence of cataracts during the trial period. Two studies were cross-sectional studies examining the prevalence of cataracts in groups already being treated with ICS compared with control groups never treated with ICS. Two studies were cross-sectional studies examining the prevalence of cataracts only among subjects already taking ICS.

Several of the clinical trials that evaluated cataracts were of relatively short duration. The randomized clinical trials by Allen, Bronsky, LaForce et al. (1998) and Tinkelman, Reed, Nelson et al. (1993) only treated and followed patients for 1 year. The CAMP study (Childhood Asthma Management Program Research Group, 2000a) assessed the incidence of cataracts over 4 to 6 years of treatment with ICS. The four cross-sectional studies evaluating subjects who had already been treated with ICS reported a mean prior treatment duration between 2.1 and 6.7 years.

Occurrence of cataract was a rare outcome in all studies. Three of the seven studies evaluating a total of 360 subjects taking ICS reported no cataracts among users (Agertoft Larsen, and Pedersen, 1998; Tinkelman, Reed, Nelson et al., 1993; Simons, Persaud, Gillespie et al., 1993). Four of the studies reported the occurrence of a single cataract that could possibly be attributed to ICS. In the randomized clinical trial by Allen, Bronsky, LaForce et al. (1998), a single patient out of 219 treated with fluticasone developed a trace subcapsular cataract at the 24th week of the study. Before enrolling in the study, the patient had been treated with ICS for 2 years. In the cross-sectional study of Nassif, Weinberger, Sherman et al. (1987), a single subcapsular cataract was noted in one patient out of 31 taking ICS. In the study of Abuikteish, Kirkpatrick, and Russell (1995), one cataract in one patient out of 140 patients taking ICS was identified, but this patient had a past history of oral corticosteroid use. In the CAMP study (Childhood Asthma Management Program Research Group, 2000a), no patients had cataracts according to lens-photography criteria, but a single patient out of 311 on ICS had a posterior subcapsular cataract detected in an ophthalmologic exam conducted 5 months after the photographs were taken. The patient had been receiving budesonide and beclomethasone, and had received 38 days of oral prednisone during the study.

Thus, the studies appear to rule out a large effect of ICS on the short-term incidence of cataract, but are insufficient to rule out an increased risk of a small absolute magnitude.

Two of these studies also reported findings on measurements of ocular pressure. In the randomized, controlled trial of Tinkelman, Reed, Nelson et al. (1993), no cases of glaucoma were detected. In the cross-sectional study by Nassif, Weinberger, Sherman et al. (1987), ocular pressures were normal in all cases and mean ocular pressures did not differ between patients taking ICS and patients not on ICS (14.0 vs. 14.0).

Effect on Hypothalamic-Pituitary-Adrenal Axis Function

Two types of evidence on the effects of ICS on HPA axis function were found. The first type of evidence consists of case reports of iatrogenic Cushing's syndrome, possibly related to ICS. The second type consists of six studies evaluating 413 patients treated with ICS where HPA axis function was followed for or assessed at least 1 year after initiation of treatment. Three studies are randomized clinical trials (or extensions or subsets of randomized clinical trials), two are cross-sectional studies, and one is a single-arm pre-post study. Each study evaluates from one to three different measures of HPA function. The three randomized clinical trials and the one pre-post study assessed HPA function over the 1-year period of trial. The two cross-sectional studies assessed HPA function at 1.4 and 2.1 mean years of treatment with ICS.

Conclusions on Effect of Inhaled Corticosteroids on HPA Axis

The findings of these studies, although varying widely as to whether a statistically significant effect of ICS on adrenal function exists, could be explained by differences in the sensitivity of the different tests used to evaluate adrenal function and the different aspects of adrenal function that are being evaluated. Measures of low cortisol, either serum or urinary, reflect diminished cortisol excretion due to the effect of exogenous steroids on the feedback mechanism which regulates cortisol excretion. Stimulation tests, on the other hand, reflect the ability of the adrenal gland to respond to stimulation by increasing cortisol excretion.

Table 10. Effects of ICS on HPA function

Table 10. Effects of ICS on HPA function

Citation	Treatment Arms	Measure of HPA Axis Function	Results	p Value	Comments
Randomized Clinical Trials
Scott and Skoner, 1999	BUD 500 mcg/day (n=132) vs. conventional treatment (n=57)	Serum cortisol at baseline and 12 mos.	BUD (0, 12 mos.): 320, 300 Conventional (0, 12 mos.): 250, 315	"No significant differences"	Subset of full trial
		ACTH-stimulated cortisol at baseline and 12 mos.	BUD (0, 12 mos.): 695, 655 Conventional (0, 12 mo): 690, 720	"No significant differences"	Subset of full trial
		% patients from normal to abnormal stimulation test between baseline and 12 mos.	BUD: 24% Conventional: 21%	"Not different"
Price, Russell, Hindmarsh, et al., 1997	FP 50 mcg/day (n=36) vs cromolyn 20 mg/day (n=27)	Urinary cortisol geometric mean ratio between patient groups at 6 and 12 mos.	Ratio of urinary cortisol at 6 mos.: 0.85 Ratio of urinary cortisol at 12 mos.: 0.96	NS: 95% CI includes 1 NS: 95% CI includes 1
Tinkelman, Reed, Nelson, et al., 1993	BDP 84 mcg/day (n=102) vs theophylline (n=93)	Serum cortisol at baseline, 6, and 12 mos.	BDP 84 mcg/day (0, 6, 12 mos.): 328, 306, 309 Theophylline (0, 6, 12 mos.): 309, 322, 334	Not stated: "similar"
		ACTH-stimulated cortisol at baseline, 6, and 12 mos.	BDP 84 mcg/day (baseline): 726, (6, 12 mos. NA) Theophylline (baseline): 723 (6, 12 mos. NA)	Not stated: "almost identical"
Cross-section studies
Gonzalez Perez-Yarza, Mintegui, Garmendia, et al., 1996	budesonide or beclomethasone, mean dose 676 +/− 280 mcg/day (range, 226-1800) (n=250) vs. normal controls (n=108)	Urinary cortisol	BUD/BDP: 58.69 nmol/m2/day Control: 81.98 nmol/m2/day	p<0.05
		No. of abnormal ACTH stimulation tests in subset with urinary cortisols below 1 standard deviation	BUD/BDP group: 2 abnormal tests (3.1%) control group: Not done	Not applicable	1 of the 2 patients with abnormal test had chronic oral corticosteroids
Nassif, Weinberger, Sherman, et al., 1987	beclomethasone 358 mcg/day (n=17) vs. beclomethasone 726 mcg/day (n=14) vs. asthmatic control group (n= 50) and normal control groups (n=215)	Serum cortisol	BDP 358 mcg/day: BDP 726 mcg/day: asthmatic controls: normal controls: 403 353 353 367	Not specifically stated: presumed NOT statistically significant
		Urinary cortisol	BDP 358 mcg/day: BDP 726 mcg/day: asthmatic controls: normal controls: 22 mcg/g creatinine 16.5 mcg/g creatinine 43 mcg/g creatinine 29.5 mcg/g creatinine	Text: "Statistically significant" from controls
Single-arm pre-post study
Ribeiro, 1993	budesonide 200 mcg/day (n=47)	Serum cortisol at baseline and 12 months	basal cortisol (0, 12 mos.): 497, 497	Not stated, presumed not statistically significant
		ACTH-stimulated cortisol at baseline and 12 months	4-hr stimulated cortisol (0, 12 mos.): 1,104, 1,131, 5-hr stimulated cortisol (0, 12 mos.): 1,242, 1,380	p=0.02 for increase from baseline, both tests

All the studies assessing adrenal function using serum cortisol level or conventional dose ACTH stimulation tests showed no effect of ICS (Table 10). However, when using more sensitive tests of such as 24-hour urinary cortisol, two out of three studies showed a statistically significant effect of ICS. It should be noted that these statistically significant results occur as comparisons of mean values between groups. Few or no patients in most studies have laboratory values out of the "normal" range. However, the clinical significance of these more sensitive indicators of adrenal function is unknown.

The question, then, is how to reconcile the case reports with the results of clinical trials and cohort studies. The case reports appear to be reasonably causally attributable to ICS based on clinical presentation, consistency with laboratory findings, and clinical response to reduction or withdrawal of treatment. Although the studies show that, on average, persons may have only clinically insignificant effects of ICS on the HPA axis, there may be individuals acutely susceptible to their effects. The relatively short duration of the reviewed studies precludes any conclusion regarding the effects of years of ICS use on HPA function.

Vertical Growth

Evidence on three measures of vertical growth in children was found: short-term growth velocity measured over a period of 1 year or less; growth velocity and change in height measured over longer duration (4-6 years), and final attained adult height. The evidence on short-term growth velocity is from a published meta-analysis which pooled data from five randomized controlled trials representing 855 subjects, with a mean age of 9.5 years. Evidence on growth velocity and height over longer duration is from the CAMP trial (Childhood Asthma Management Program Research Group, 2000a) randomized trial comparing ICS, nedocromil, and placebo in 1,041 children with mild-to-moderate asthma followed for 4 to 6 years. For final attained adult height, evidence is from three retrospective cohort studies that adjusted for the potential confounding factor of parental height. Together, these three studies included a total of 243 asthmatics treated with ICS, 154 asthmatics who had not been treated with ICS, and 204 non-asthmatic controls.

Evidence on growth velocity over 1 year is consistent in showing a difference of average height of 1 cm/year between children treated with ICS and controls. In the only trial extending beyond 1 year (Childhood Asthma Management Program Research Group, 2000a), a difference consistent with this magnitude also occurred in the first year of the study. However, in subsequent long term followup, the difference in growth velocity was not maintained. At the end of the 4 to 6 year observation period there was still an approximately 1 cm difference in cumulative growth between the study groups.

The evidence on final adult height appears to be fairly consistent, as well. However, this evidence is based on retrospective cohort studies, which are subject to selection bias and the confounding effects of severity of asthma cannot be adjusted for. Some comparisons in these studies were also limited by small sample size. One study showed a difference in final attained adult height between ICS users and nonusers. However, the difference is much less than would be expected than if a 1 cm/year growth velocity difference was maintained over several years.

Bone Mineral Density

The CAMP study (Childhood Asthma Management Program Research Group, 2000a) followed a population of mild to moderate asthmatics, mean age approximately 9 years treated for 4 years with ICS. This study, with large numbers, randomization and assessment of longitudinal changes, provides very strong evidence that there is no effect of ICS on BMD and in the doses given and time duration in that study. One retrospective study of 30 young adults found a significant correlation between BMD and ICS dosage among female patients. Such studies are subject to potential confounding because of unmeasured differences between groups that are risk factors for low BMD. In addition, the clinical significance of any observed differences in BMD are unknown. Subtle differences in BMD would not have clinical impact until additive to other risk factors such as aging, and it is uncertain whether differences observed during young adulthood would persist to old age. Alternatively, it is possible that subtle changes during critical periods of bone mineral accretion that occur in childhood could magnify the risk of osteoporotic fracture in later life.

Posterior Subcapsular Cataract and Glaucoma

Studies that report the occurrence of posterior subcapsular cataracts consist mostly of small cohorts and cross-sectional studies, with the exception of the CAMP study (Childhood Asthma Management Program Research Group, 2000a). The expected incidence rate of subcapsular cataract in any population of normal young children and adults is zero. These studies are sufficient to rule out a large effect of ICS on short-term incidence of cataract, but are not capable of detecting a small increase in risk of an event which has a baseline risk of essentially zero. Also, several of the clinical trials that evaluated development of cataracts were of relatively short duration.

Two of these studies also reported on measurements of ocular pressure. The very limited data available show no relationship between glaucoma or raised intraocular pressure and ICS.

Effect on Hypothalamic-Pituitary-Adrenal Axis Function

Two types of evidence on the effects of ICS on HPA axis function were found. These were case reports of iatrogenic Cushing's syndrome related to ICS and six studies (n=413 treated with ICS) regarding HPA axis function. Each study evaluates from one to three different measures of HPA function, with followup for at least 1 year after initiation of treatment.

The case reports show that systemic effects can occur in clinically detectable ways in individuals, with a strong case for causality in these individual patients by the accompanying laboratory tests and response when ICS were withdrawn. In the controlled clinical studies, when using more sensitive tests of cortisol such as 24-hour urinary cortisol, two out of three studies of HPA axis function showed a statistically significant effect of ICS. It should be noted that these statistically significant results occur as comparisons of mean values between groups. Few or no patients in most studies have laboratory values out of the "normal" range. However, the clinical significance of these more sensitive indicators of adrenal function is unknown.

The case reports appear to be reasonably causally attributable to ICS based on clinical presentation, consistency with laboratory findings, and clinical response to reduction or withdrawal of treatment. Although the studies show that, on average, persons may have only clinically insignificant effects of ICS on the HPA axis, there may be individuals acutely susceptible to their effects.

Randomized Controlled Trials

Single-Arm Studies

A major limitation of the single-arm studies is that patients enter the study at varying time points in the history of their disease; it is then impossible to compare outcome data at a uniform time point. For example, suppose a study starts with two patient groups; one with a prior duration of asthma of 1 year and another with a prior duration of 5 years and treats both groups with ICS for 1 year. With this design, final outcome data can only be compared 2 years from the time of diagnosis in the first group and 6 years from the time of diagnosis in the second group. For the group with the shorter duration, it cannot be determined what their outcome data will be like at the 6-year time point. A second major limitation in studies of this type is the high potential for selection bias. It is likely that patients with a longer duration of asthma will have more severe disease, both because of disease progression and because asthma is more likely to remit in milder cases. This also will limit the comparison of outcomes between patients with varying duration of disease.

Implications of the CAMP Trial

The CAMP study (Childhood Asthma Management Program Research Group, 2000a) was a three-arm randomized controlled trial evaluating the outcome effects of ICS or nedocromil sodium compared to placebo in 1,041 children over a mean followup of 4.3 years. The primary outcome measure was postbronchodilator FEV1. Although the design of this study does not directly address the question of immediate versus delayed ICS treatment, the CAMP study provides the most robust evidence to date on long term lung function outcomes in a group of patients treated with ICS and a placebo control group (Childhood Asthma Management Program Research Group, 2000a). As such, the results provide evidence on changes in lung function that occurs over 4 years in children with mild to moderate asthma, and whether treatment with ICS or nedocromil prevents a decline in lung function.

CAMP found no significant difference between treatment and control groups for change in postbronchodilator FEV1, which was the primary outcome (Childhood Asthma Management Program Research Group, 2000a). Postbronchodilator FEV1 was used as the preferred indicator of disease progression to minimize effects of reversible airway constriction and individual variability over time that are observed with prebronchodilator FEV1. The baseline postbronchodilator FEV1 was 103 (+/− 12) percent predicted for the placebo group, 103 (+/− 13) percent predicted for the ICS group, and 102 (+/− 12) percent predicted for the nedocromil group, indicating normal or near normal lung function at the outset of the trial. Overall, changes in postbronchodilator FEV1 were minimal: there was a mean decline of 0.1 percent predicted in the placebo group and 0.5 percent in the nedocromil group, while there was a 0.6 percent increase in the ICS group. These means are adjusted for characteristics at study entry, including baseline value, severity and duration of asthma, age, sex, race, and ethnicity (SD=9.6, estimated by regression model).

The finding of no difference in postbronchodilator FEV1 and minimal change overall in lung function for the entire study population does not support the hypothesis that treatment with ICS can prevent a long-term decline in lung function in mild to moderate asthmatics. CAMP did not find progressive decline in lung function in the placebo group, or significant improvement in the treatment groups (Childhood Asthma Management Program Research Group, 2000a). Had the study reported significant differences in lung function outcomes, then the data would not have allowed a determination of whether early initiation of ICS was superior to delayed initiation. A direct comparison of early versus delayed ICS initiation would then be necessary to evaluate whether irreversible decline in lung function occurs when ICS treatment is delayed.

Similar to the case with lung function outcomes, there was not a progressive decline in symptoms for the placebo group. Symptom scores improved over the course of the study in both the ICS and placebo group, with a greater improvement reported for the ICS group (−0.44 vs. −0.37 on a 0-3 scale, p<0.005). The number of night awakenings also improved in both groups, with no difference between ICS and placebo (−0.7 vs. −0.6 awakenings per month, p=NS). These symptom scores indicate that ICS lead to better symptom control in this population. However, the data do not indicate that there is a progression in severity of symptoms for patients with mild to moderate asthma in the absence of treatment with ICS.

In summary, the CAMP trial did not find progressive decline in lung function over a 4-year period in a population of children with mild-to-moderate asthma managed without ICS; nor was there a significant difference between treated and control groups in change in postbronchodilator FEV1. However, several possibilities limit the generalizability of the results of the CAMP trial. First, it is possible that this length of followup was not adequate to detect significant changes. A trial with longer followup might detect a preventable decline, although the trajectory of changes in lung function for the CAMP population does not suggest that this is likely to be of a large magnitude. A second possibility, as suggested by the CAMP investigators, is that a decline in lung function occurs early in the course of disease and that the patients in the CAMP study, with a prior duration of asthma averaging approximately 5 years, were too far along in their disease course to prevent such an early decline.

Third, it is possible that progressive decline in lung function occurs in patients with more severe disease, but not in the mild to moderate asthmatic population selected for the CAMP trial. Finally, it is possible that the lack of decline in the lung function parameters is due to the care given in the trial apart from ICS. The trial employed an intensive intervention and followup in all groups. This high level of care may have maintained optimal lung function and symptom control to a greater extent than can be achieved under the conditions of usual care.