Intended for healthcare professionals

Clinical Review State of the Art Review

Why and how to step down chronic asthma drugs

BMJ 2017; 359 doi: (Published 16 October 2017) Cite this as: BMJ 2017;359:j4438
  1. Michael R Gionfriddo, assistant professor1,
  2. John B Hagan, associate professor2,
  3. Matthew A Rank, associate professor3
  1. 1Center for Pharmacy Innovation and Outcomes, Geisinger Health System, Forty Fort, PA, USA
  2. 2Division of Allergic Diseases, Mayo Clinic, Rochester, MN, USA
  3. 3Division of Allergy, Asthma, and Clinical Immunology, Mayo Clinic, Scottsdale, AZ, USA
  1. Correspondence to: M A Rank rank.matthew{at}


Asthma is a common chronic airways disease. The goal of asthma management is to control symptoms while minimizing the side effects of treatment. Following a period of stable asthma, clinicians should consider stepping down treatment. This approach is recommended by current guidelines. Step-down has been studied for several types of asthma drug regimens, and certain approaches may have lower risk than others. Systematic reviews of multiple trials support the following specific step-down approaches: optimizing inhaled corticosteroid dosing when stepping down oral corticosteroid, reducing inhaled corticosteroid from a higher dose, lowering inhaled corticosteroid-long acting bronchodilator (ICS-LABA) dose while adding ICS-LABA on-demand, adding leukotriene receptor antagonist (LTRA) while lowering inhaled corticosteroid dose, and using allergen immunotherapy when reducing inhaled corticosteroid from a higher dose. Systematic reviews of multiple trials support an increased risk of asthma exacerbation for patients who completely stop taking inhaled corticosteroid or long acting bronchodilator. Strategies to implement step-down in practice include the use of risk prediction as well as tools to support shared decision making and communication about risk between clinicians and patients.


Asthma can be managed by controller drugs (also called chronic or preventive drugs) and rescue drugs (intended to provide temporary relief). Patients who have frequent symptoms and/or exacerbations of asthma should be treated with controller drugs, with the dose and type of controller drug titrated upward until an acceptable level of disease control is achieved.1 Once control is achieved, drugs can be maintained at that level or attempts can be made to step down treatment over time. Step-down is also referred to as de-escalation, tapering, or discontinuation in other reports. Stepping down asthma drugs presents an opportunity to reduce the risk of side effects and the burden and costs of treatment, all of which are important to patients.2 The best approach for step-down of asthma drugs is often unclear, and step-down is seldom done in clinical practice.234 This review article intends to fill this gap between what is known about stepping down from research studies and how this can practically be applied in clinical practice.


Asthma is a common chronic respiratory condition that affects about 4% of people worldwide, with wide variation of prevalence between countries.5 Asthma causes substantial morbidity and mortality with a quarter of a million deaths worldwide,6 and in the US it accounts for nearly 25 million missed school or work days annually.7 Asthma affects people of all ages, boys more than girls and women more than men, and disproportionately affects people of specific race/ethnicities and those with lower socioeconomic status.7 Many people with asthma have symptoms or risk that call for use of asthma controller drugs (also called persistent asthma). An estimated 25% of patients with persistent asthma could be considered for step-down each year on the basis of current guideline recommendations.8

Sources and selection criteria

The data for this review came from searches of multiple databases (Medline, Embase, Central, and Cochrane Database of Systematic Reviews), with no date restrictions. We modeled the searches after previous searches we did as part of systematic reviews on specific step-down decisions.9101112 We sought out systematic reviews relating to the specific step-down questions covered in this review. We used similar keywords to identify step-down events (for example, deescalate, taper, withdraw, discontinue). For this review, we considered prospective and observational controlled studies but did not consider case series or reports. We did not limit selection of studies by language, date, size, or quality. We selected studies that were pertinent to the specific questions related to step-down of asthma drugs. We prioritized systematic reviews with meta-analyses and randomized controlled trials.

Summary of evidence for step-down of asthma controller drugs

The management of asthma is individualized initially on the basis of the patient’s symptoms and risk of exacerbation and then modified according to the level of control. Coupled with the variety of treatment options, this leads to many possible asthma drug regimens. These regimens can then be stepped down in several ways, potentially forcing the clinician to consider multiple step-down options in a single clinical encounter. Here, we will review the available evidence around step-down and provide the framework to promote evidence based conversations with patients about asthma step-down. For clinicians, it is important to keep in mind that some patients with asthma have been shown to have a higher overall risk and that extra caution may be warranted when stepping down in patients with these risk factors. Specific risk factors that should be considered include previous intensive care episodes/intubations/frequent hospital admissions/frequent oral corticosteroids,13 poor perception of dyspnea/short attack onset,14 concurrent psychiatric diagnoses,15 alternaria sensitivity,16 food allergy,17 non-white ethnicity,18 poverty,18 and non-adherence.19Table 1 provides a summary of each specific step-down option. Figure 1 shows a practical algorithm to help clinicians to apply step-down of asthma drugs in practice.

Table 1

Summary of step-down choices

View this table:

Fig 1 Practical algorithm for stepping down chronic asthma drugs. *This is one example of a simple rule that can be used to help in predicting risk of exacerbation after stepping down asthma drugs6343940

Step-down of oral corticosteroids

Chronic use of oral corticosteroids is low (<5%) among patients with asthma.8 The long term adverse effects of chronic oral corticosteroid use, which include weight gain, hypertension, and osteoporosis, are a compelling reason to limit their use and attempt step-down if appropriate.41 Evidence suggests that patients on long term oral corticosteroid treatment may achieve similar levels of control if they are appropriately dosed on inhaled corticosteroids,2042 and such a change would theoretically limit systemic corticosteroid exposure and reduce adverse effects. However, in studies comparing oral corticosteroid with inhaled corticosteroid, adverse effects were reported too variably to allow for confident comparisons.20212223 If the patient is concurrently taking inhaled corticosteroid with oral corticosteroid, care should be taken to ensure optimal inhaler technique as well as adherence to therapy, as these have been shown to be consistently poor and will affect the effectiveness of these drugs.434445464748

Asthma biologics, a class of drugs that target specific immune pathways in asthma, present another option for facilitating step-down from oral corticosteroid. Omalizumab was not found to reduce oral corticosteroid use (n=95 patients; odds ratio 0.99, 95% confidence interval 0.44 to 2.24) in a systematic review and meta-analysis but was found to reduce inhaled corticosteroid dosage (n=1188; mean difference −118 (−154 to −84) μg beclomethasone equivalents).49 Single trials suggest a reduced dose of oral corticosteroid when using mepolizumab (n=135; odds ratio 2.39, 1.25 to 4.56) and benralizumab (n=220; 4.09, 2.22 to 7.57 for every four week dosing).2425 Approaches to help in tailoring step-down from chronic oral corticosteroid include the use of biomarkers such as sputum eosinophils and fractional exhaled nitric oxide, although the evidence is limited and is not focused specifically on step-down from oral corticosteroid,505152 and an internet facilitated tapering strategy, which proved successful in a single prospective randomized trial of 95 patients (absolute median reduction in prednisone 205 (45 to 365) mg, over a six month period).53

A practical note for patients with chronic oral corticosteroid use is that a comprehensive approach to identifying relevant comorbidities in severe asthma should also be pursued as part of any strategy for stepping down oral corticosteroid in severe asthma.54 In summary, evidence for step-down from oral corticosteroid is based on systematic reviews of multiple trials for using inhaled corticosteroids,20212223 and more data are needed for comparisons of adverse effects between oral and inhaled corticosteroids. Evidence from single trials support the use of mepolizumab and benralizumab for facilitating oral corticosteroid step-down.

Step-down of asthma biologics

Three biologics are clinically available for asthma, one targeting immunoglobulin E (omalizumab) and two targeting interleukin 5 (mepolizumab and reslizumab). Benralizumab is not available outside of clinical trials. All three available biologics were studied in patients whose moderate to severe asthma was not controlled using inhaled corticosteroid. For initiation, immunoglobulin E and specific allergy testing or blood eosinophil concentrations are used to help to select patients more likely to respond to the specific treatment. Data on stepping down, however, are limited. For the interleukin 5 drugs, no data have been published on step-down; a clinical trial examining step-down from mepolizumab is expected to be completed in 2019.55

Data on step-down do exist, however, for omalizumab. The Evaluating Xolair Persistency of Response After Long Term Therapy (XPORT) trial was a double blind, placebo controlled, randomized clinical trial of 152 patients treated with omalizumab for more than five years.26 Patients were randomized to continue or stop taking omalizumab and were followed for one year. The odds of not having an asthma exacerbation after adjustment for several prognostic factors was 0.44 (0.23 to 0.82), favoring continuation of omalizumab.

Observational data on step-down of omalizumab are mixed, with some studies showing sustained control of asthma,56575859 whereas another indicates worsened control on discontinuation.60 Additionally, a randomized controlled trial (RCT) examining physiologic correlations with immunoglobulin E changes and outcomes found that discontinuation would likely worsen asthma control.61 These studies are limited by their observational design, small sample sizes, and lack of control groups.

In summary, the evidence for decisions to step down omalizumab is based on one trial,26 and evidence for mepolizumab and reslizumab is insufficient. As most patients who start taking biologics are at high risk, extra caution including intensified monitoring is needed if discontinuation is considered.

Inhaled corticosteroid combined with long acting bronchodilator to corticosteroid alone

A systematic review and meta-analysis of 2781 summarized the data for stepping down from inhaled corticosteroid combined with long acting bronchodilator (ICS-LABA) to inhaled corticosteroid alone.27 Discontinuation of long acting bronchodilator was associated with reduced quality of life scores (odds ratio 0.36, 0.15 to 0.57) and asthma control questionnaire scores (0.24, 0.13 to 0.35). The clinical relevance of mean decreases of 0.36 and 0.24 in quality of life and asthma control scores, respectively, is less certain given that the minimal clinically important difference for these measures is 0.5. Risk of needing oral corticosteroid was not significantly elevated despite an odds ratio that approached 2 (1.74, 0.83 to 3.65), and the risk of hospital admission was not estimable owing to no events occurring in the 1342 participants included from five trials. No intubations or deaths were reported.

A subsequent RCT of 11 679 patients compared ICS-LABA with inhaled corticosteroid alone and included a subgroup analysis of patients well controlled on ICS-LABA who were randomized to continue or to change to inhaled corticosteroid alone.62 The risk of asthma exacerbation was significantly reduced in patients who continued ICS-LABA (hazard ratio 0.76, 0.65 to 0.91). Of note, the results from these studies are derived primarily from adults.28 Clinicians attempting step-down from ICS-LABA to inhaled corticosteroid alone should consider the use of asthma control questionnaires and spirometry to help to determine the risk of exacerbation.39 The evidence supporting an increased risk of loss of asthma control following step-down from ICS-LABA to inhaled corticosteroid alone is based on systematic review of multiple trials.27

Higher fixed dose ICS-LABA to lower dose ICS-LABA

One approach to reduce overall inhaled corticosteroid use is to include inhaled corticosteroid with a fast acting LABA for on-demand use with a lower background maintenance dose. This strategy has been called single inhaler maintenance and reliever therapy (SMART) or single inhaler therapy (SiT). A systematic review in 2013 identified four trials (two open label) in 9130 patients, which together strongly suggested improved efficacy compared with higher fixed dose therapy with overall lower inhaled corticosteroid dose (odds ratio for emergency department visit or hospital admission for asthma 0.72, 0.57 to 0.90),29 with an additional RCT of 1701 patients reported supporting efficacy of SMART since the systematic review was completed (hazard ratio for asthma exacerbation 0.64, 0.49 to 0.82).30

The SMART strategy has also been tried in patients with more severe asthma, using higher baseline inhaled corticosteroid dosing, with lower overall asthma exacerbation risk and roughly equivalent corticosteroid dosing (n=303; relative risk 0.54, 0.36 to 0.82).63 The most recent international guidelines list the SMART approach as an option for managing asthma but do not specifically list this is an endorsed strategy for stepping down asthma drugs.1 The evidence for stepping down from higher fixed dose ICS-LABA to lower dose ICS-LABA plus ICS-LABA on demand is based on systematic review of multiple trials.

A second approach to stepping down from ICS-LABA to a lower dose ICS-LABA is to do so without an on-demand ICS-LABA dose. A systematic review and meta-analysis in adults identified two RCTs relevant to this question.316465 In the first, a medium dose ICS-LABA was reduced to a low dose ICS-LABA in 476 patients.64 In the second, medium high doses of inhaled corticosteroid were reduced in 259 patients, but only 36.9% of patients assigned to the step-down intervention and 30.2% of controls were using concurrent long acting bronchodilator.65 Comparing those who stepped down with controls, no significant differences were found for several outcomes including needing oral corticosteroid (odds ratio 1.31, 0.82 to 2.08), hospital admission (4.06, 0.45 to 36.9), or all cause serious adverse events (0.60, 0.11 to 3.33).31 The odds ratio of 4 for hospital admission may be clinically relevant; however, owing to the limited number of events and high level of imprecision, this result should be interpreted with caution.

Thus, the evidence for reducing inhaled corticosteroid dose when using ICS-LABA combination therapy is based on a systematic review of two trials. However, caution is needed when interpreting these findings given that in one of the studies more than half of the participants were not using long acting bronchodilator.65 Additional caution is warranted when extrapolating results to children because data in this population are lacking.

50% or greater dose reduction of inhaled corticosteroid but not complete discontinuation

For patients using inhaled corticosteroid, clinicians can consider reducing the dosage. A systematic review assessed the effect of reducing the dosage of inhaled corticosteroid by 50% or more.11 It identified six RCTs with a combined population of nearly 900 patients. Compared with continuing inhaled corticosteroid, reducing the dose by 50% or more increased the risk of exacerbation by 25% (relative risk 1.25, 0.96 to 1.62). Patients in some of these studies were required to have only four weeks of stability, less than the minimum three months recommended by current guidelines.1 Therefore, it is possible, and supported by some observational data,34 that if the patients had greater lengths of stability before stepping down the risk of exacerbation may have been lower.

The evidence that stepping down inhaled corticosteroid by 50% or more does not significantly increase the risk of future asthma exacerbations in properly selected patients is supported by systematic review of multiple RCTs.

Daily inhaled corticosteroid to on-demand inhaled corticosteroid

Anecdotal experience and empirical data indicate that some patients take their inhaled corticosteroid on an on-demand basis rather than as scheduled.66 In patients with poor asthma control, this behavior is concerning and may contribute to the lack of control. For patients who are controlled, however, a systematic review has assessed stepping down from a daily dose to an on-demand dose of inhaled corticosteroid.9

The review identified two RCTs in 377 patients and suggested that stepping down from daily to on-demand inhaled corticosteroid was not associated with an increase in asthma exacerbations (relative risk 1.32, 0.81 to 2.16) or changes in lung function (0.11, −3.81 to 4.03) but that patients in the daily inhaled corticosteroids group had a higher number of symptom-free days (standardized mean difference 0.26, 0.02 to 0.49). These data are limited owing to a low number of events, leading to imprecision, as well as heterogeneity in that one study was conducted in adults and the other in children. Other systematic reviews have been conducted but are limited because they also included patients with wheezing and did not require patients to have stable asthma.6768 Despite the limitations in the evidence, potential benefits of this approach include lower drug costs and lower exposure to inhaled corticosteroid. This approach may serve as a middle ground for patients in whom the necessity for treatment with inhaled corticosteroid is questionable but the risk of completely stopping inhaled corticosteroid is high.

The evidence that stepping down from daily to on-demand inhaled corticosteroid does not lead to an increased risk of asthma exacerbation is based on a systematic review of two RCTs and tempered by evidence that asthma symptoms may increase.

Substitution of leukotriene receptor antagonists for inhaled corticosteroid

Poor inhaler technique contributes to poor asthma control.43 Furthermore, some patients may prefer to take oral drugs rather than inhalers.69 Leukotriene receptor antagonists (LTRAs) provide an option for patients whose asthma is controlled with inhaled corticosteroid and who wish to stop using their inhaler, but for whom stopping altogether is deemed too risky.

A systematic review examined either substituting LTRAs for inhaled corticosteroid or supplementing dose reduction of inhaled corticosteroid with LTRA.10 It identified one RCT examining substitution of inhaled corticosteroid for a LTRA.1070 Fluticasone propionate 100 μg twice daily was continued in 169 patients and discontinued with substitution of montelukast in 166 patients. Over a 16 week period, treatment failure was seen in approximately 20% of patients who continued inhaled corticosteroid compared with 30% in whom inhaled corticosteroid was stopped and montelukast begun (hazard ratio 1.6, 1.1 to 2.6).70 A subsequent RCT of 40 patients found similar results with substitution of inhaled corticosteroid with pranlukast (72% controlled in pranlukast group compared with 90% controlled in inhaled corticosteroid group).71

One observational study suggests that sputum eosinophil counts may help to guide step-down with LTRA.7273 Therefore, although substitution of inhaled corticosteroid for LTRAs seems to increase the risk of exacerbation, the use of biomarkers may help to target step-down to those patients who are most likely to be successful, as may the use of a predictive model using multiple variables.74 In summary, evidence from two trials shows that substitution of LTRA for inhaled corticosteroid leads to an increased risk of asthma exacerbation.7071

Using long acting bronchodilator or LTRA to facilitate inhaled corticosteroid step-down

A systematic review and meta-analysis of using long acting bronchodilator to facilitate inhaled corticosteroid step-down identified two RCTs (n=484 patients) and found no increased risk of asthma exacerbations (relative risk 1.0, 0.76 to 1.32) or adverse events (0.97, 0.74 to 1.28) when comparing continuation of inhaled corticosteroid at current dose with a lower dose ICS-LABA.32 A systematic review assessed supplementation of inhaled corticosteroid dose reduction with the addition of LTRAs to reduce the risk of exacerbation that may occur due to dose reduction of inhaled corticosteroid.10 Data from three RCTs indicated that supplementing inhaled corticosteroid dose reduction with an LTRA reduces the risk of exacerbation compared with not supplementing the dose reduction (relative risk 0.57, 0.36 to 0.90).

In summary, in patients whose asthma is controlled and for whom inhaled corticosteroid step-down is being considered and a high risk of or concern about exacerbation exists, three RCTs suggest that supplementing the inhaled corticosteroid reduction with LTRA reduces the risk of an exacerbation. Two trials using long acting bronchodilator found similar results.

Conversion from twice daily dosing to once daily dosing

One year adherence to asthma drugs, as measured by mean proportion of days covered, was recently estimated to be 21% for inhaled corticosteroid, 25% for ICS-LABA, and 30% for LTRA.75 Data indicate that for drugs generally, the frequency of administration is a predictor of adherence.76 This pattern seems to hold for asthma drugs as well.7778 One study reported an increased adherence to asthma controller drugs of 20% for once daily dosing compared with multiple daily dosing of asthma controller drugs.79 This means that a “step-down” of 50% in overall recommended dose may end up being a lesser overall reduction in dose if adherence is higher for once daily than twice daily dosing. Stepping down from twice daily to once daily dosing may also reduce costs,8081 although the effect on other outcomes is less clear.78

Use of allergen immunotherapy to facilitate step-down

Allergen immunotherapy, unlike other treatments for asthma, has disease modifying effects yielding long term changes in allergen tolerance and symptom control even after treatment is stopped, making it an attractive option for many patients with allergic diseases. Using allergen immunotherapy in children sensitized to allergens may even prevent asthma from developing.82 A systematic review assessed the role of allergen immunotherapy in facilitating step-down of inhaled corticosteroid.33 It identified six trials of allergen immunotherapy for asthma—five of house dust mite allergen immunotherapy and one for seasonal asthma associated with birch tree pollen. It found that for patients at Global Initiative for Asthma (GINA) levels 3-4 (typically taking a medium dose of inhaled corticosteroid), reduction of inhaled corticosteroid dose by approximately 50% was more likely to be successfully accomplished with allergen immunotherapy compared with placebo.

This review supports an approach of using allergen immunotherapy for patients allergic to house dust mites who are using medium dose inhaled corticosteroid. More data are needed for other forms of allergen immunotherapy and for patients at different asthma treatment levels.

Seasonal step-down

For many patients with asthma, the burden of disease varies by season. Guidelines recommend avoiding step-down during periods when patients are at a high risk for exacerbation.1 A high risk period may be during a patient’s allergy season or during times of the year when viral upper respiratory infections are more likely (for example, September for schoolchildren).

In an observational study of stepping down asthma treatment in a guideline based pediatric asthma management program, season of step-down (not stepping down during fall season) was the strongest predictor of successful step-down.3 Another study found a summer decline in both inhaled corticosteroid and short acting bronchodilator refills in children, which suggests that patients and their families self select to reduce asthma drugs during periods when the risk of exacerbation is lower.35 A study of a large claims database also found that time to failure of step-down was shortest during the fall season.34 We did not identify any controlled, prospective trials examining the timing of seasonal step-down. The evidence for seasonal step-down of asthma drugs is insufficient. Although not directly applicable, clinicians may consider the data from studies of step-down from scheduled to on-demand regimens informative,9 as discussed in a previous section in this review. Clinicians may consider offering an on-demand regimen during low risk seasons and suggest a scheduled approach to treatment during high risk seasons.

Discontinuation of inhaled corticosteroids

A systematic review and meta-analysis of seven trials and 1040 patients assessed the evidence for discontinuing low doses of inhaled corticosteroid.12 The relative risk of asthma exacerbation for patients stopping inhaled corticosteroid compared with those continuing was 2.35 (1.88 to 2.92). Therefore, clinicians should be cautious before completely discontinuing inhaled corticosteroid. Additional data suggest that caution should be used when stopping inhaled corticosteroids, even after the three months of asthma stability suggested in guidelines have elapsed. Observational data suggest that longer periods of asthma stability before stepping down predict a greater likelihood of successful step-down (n=26 292; risk of asthma exacerbation with at least 12 months of stability before stepping down 21% compared with 44% if four months or less).34

Despite the evidence and concerns regarding discontinuation of inhaled corticosteroid, many patients who discontinue inhaled corticosteroid do so through lack of adherence to recommended therapy.244 The evidence for an increased risk of asthma exacerbations after discontinuing inhaled corticosteroid is based on a systematic review of multiple trials, and clinicians should therefore proactively discuss the possible harms of discontinuing inhaled corticosteroid (absolute risk of asthma exacerbation 0.38 for stopping inhaled corticosteroid and 0.16 for continuing the same dose over a six month period; number needed to harm=5) and acknowledge the potential benefits (lower costs and fewer potential adverse effects). Another compelling reason to consider stepping down inhaled corticosteroid is for diagnostic clarity. Recent data indicate that up to 30% of patients with asthma may be misdiagnosed.83848586

Discontinuation of LTRAs

We did not identify any systematic reviews on the risk of exacerbation following discontinuation of LTRAs. We identified three studies (one RCT and two observational studies) pertinent to this question in our search.363738 These studies all measured fractional exhaled nitric oxide as a measure of asthma activity and found that following discontinuation of the LTRA, exhaled nitric oxide increased, which suggests that asthma activity is increased. The effect of discontinuation on exacerbations is unclear, although one study indicated a reduction in lung function after discontinuation.36 Although one of these studies was a randomized trial,36 the data from it are from a comparison not related to randomization, so all data come from small observational studies; the total population studied was 58 children.363738 In patients with controlled asthma who are taking only LTRAs, insufficient data exist to guide step-down. Step-down may be considered, as patients may do it anyway through non-adherence244; however, a study of 20 patients found that two had an asthma exacerbation during montelukast washout.38 Therefore, clinicians should exercise caution during step-down and carefully monitor the patient’s status. An additional reason for discontinuation of LTRA, as with discontinuation of inhaled corticosteroid, is for diagnostic clarity when reasonable suspicion exists that asthma is an incorrect diagnosis.

Practical considerations for step-down of asthma drugs

As highlighted in the previous sections, step-down is a safe option for many patients and is recommended by current guidelines.1 Despite this guidance, stepping down is not routinely done in clinical practice.34 In this section, we will review barriers to and facilitators of step-down.

Barriers to step-down

The failure to translate guideline recommendations into practice is influenced by a variety of factors including those related to systems, clinicians, and patients. From a systems perspective, incentives are generally aligned around step-up rather than step-down of treatment.4 Additionally, healthcare systems are often fragmented, resulting in patients seeing clinicians who may not be in the same system or otherwise do not communicate well or have access to the same electronic health records. This fragmentation contributes to barriers identified by clinicians.

Although the literature on barriers to step-down in asthma is sparse, barriers to step-down for prescribers in the setting of potentially inappropriate drugs has recently been reviewed.87 The review identified four domains of barriers including awareness, inertia, self efficacy, and feasibility. Prescribers are often unaware of their own prescribing practices, and this lack of awareness led to discordance between what clinicians thought (for example, potential for step-down should be assessed regularly) and what they did (for example, not assessing potential for step-down). Even when clinicians were aware of their prescribing practices, inertia led them to continue treatment although they knew it should be stepped down. This inertia was often driven by fear of what would happen if the treatment was stepped down (for example, exacerbation after stepping down inhaled corticosteroid therapy). For some clinicians, inertia is also driven by a perception of treatment that tends to maximize potential benefits and minimize potential harms/costs. Finally, inertia may also be driven by concerns about who is responsible for step-down (should the primary care clinician be responsible for stepping down asthma drugs or is it the job of the asthma specialist?).

The third theme of self efficacy reflects prescribers’ perceived ability (or inability) to step down therapy. Contributing factors to this theme were a lack of knowledge and skills compounded by the lack of evidence and clear, practical guidance on step-down. Clinicians desired practical guidance, as many thought that implementing step-down was not feasible. Factors affecting the feasibility of step-down include patients’ resistance to change, limited time, and reimbursement for step-down.

Patients’ attitudes or beliefs may be a barrier to step-down. Patient related barriers to step-down in general have been recently reviewed.88 Barriers included the appropriateness of cessation, fear, and the process of cessation. The effect of these factors on a patient’s willingness to step down treatment can be influenced by a variety of things including their clinician, their family, the media, and their own past experiences. The appropriateness of cessation may be an important theme to consider in asthma, as the treatment directly affects symptoms. Many patients feel better when they take their medicine, and fear of having symptoms or an exacerbation may improve adherence.89 Conversely, many patients dislike drugs and would welcome step-down.8890 Therefore, stepping down therapy may be intimidating for some patients but not others, and clinicians should carefully consider patients’ attitudes toward their treatment and asthma when assessing the potential for step-down.

Similar to clinicians, patients reported that they did not have many opportunities to discuss step-down with their clinician.88 For example, a study from Sweden found that more than a third of patients had not seen an asthma nurse or physician for several years.91 Finally, patients also reported a lack of support following step-down,88 which may stem from clinicians’ lack of self efficacy regarding step-down.87 Some of these barriers can be overcome through structural changes (increased length of visit or incentives for assessing potential for step-down), but others will need to be tackled by training and feedback regarding not only prescribing habits and the appropriateness of step-down but also communication between clinicians and patients. We used these data, along with feedback received from patient advisory groups, to construct a figure representing a patient centric model of stepping down asthma drugs (fig 2).


Fig 2 Patient centric model of asthma step-down

Approaches to implementing step-down

Firstly, an assessment should be done to provide a clear picture of the patient’s current asthma status. Adherence to treatment, device technique, and identification of triggers should all be taken into account. A review of previous step-up and step-down drug steps and an attempt to judge response may further inform the current evidence framework presented in table 1. Reassessment two to six weeks after step-down and regularly thereafter is important, as patients are still at risk of exacerbation (see figure 1 for how these factors are integrated into a practical and algorithmic approach to stepping down asthma controller drugs).

Strategies for implementing step-down in asthma have not been well studied. One potential approach could include the use of asthma action plans. Having an asthma action plan may be an important factor that increases the likelihood of step-down.2 Additional approaches can be adapted from the general literature on step-down. As highlighted above, fear, knowledge, and skills are all barriers that can affect both the patient’s and the clinician’s willingness to step down.

Shared decision making

One approach to help to manage these barriers is through improved communication, such as shared decision making. Shared decision making has been advocated as an approach to step-down in older adults.92 Shared decision making is a patient centered approach to communication in which the patient and clinician work together in conversation to construct what is best, based on the interplay between the evidence, the patient’s values, and the patient’s context.93 The implementation of shared decision making through the use of decision aids may improve knowledge, and patients report being better informed and being more clear about their personal values, are more active in decision making, and have more accurate risk perceptions.94

Shared decision making in asthma has been studied, but research has been primarily focused on step-up therapies.9596979899100 An RCT in 612 patients compared a shared decision making approach with usual care and found that a shared decision making approach improved drug adherence (continuous adherence 0.67 in the intervention group compared with 0.46 in usual care) and other outcomes such as lung function (percentage predicted forced expiratory volume in one second (FEV1) 76.5% for intervention versus 73.1% for usual care) and asthma symptoms (odds ratio for not reporting any asthma symptoms 1.9, 1.3 to 2.9).95 Although not specifically focused on step-down, the authors lay out a shared decision making approach in which patients’ goals and priorities were specifically collected from the patients and treatment options were discussed specifically considering this information. Greater detail on the specifics of their intervention can be found in the online supplement to their study. A second study used a participatory research approach andlooked at the implementation of shared decision making in primary care on the basis of the model used in the study above.100 It found that shared decision making could be successfully implemented into primary care practices, with 100% of practices having maintained the intervention and 90% of patients reporting a shared decision making experience. The study prompted the development of a shared decision making toolkit and implementation resources including training resources that can be accessed at

A recently published RCT in 51 patients used a decision aid developed in a previous study.97 The decision aid was structured using the framework provided by the Ottawa Personal Decision Guide,40 with the aim of facilitating conversations around inhaled corticosteroid use with or without long acting bronchodilator.96 The Ottawa Personal Decision Guide has four steps.40 (1) Clarify the decision (what is asthma, what are the options, and how do they work to control asthma?). (2) Explore the decision (describe and compare the pros and cons of the available options). (3) Identify decision making needs (are patients sure of the best choice for them, do they understand the risks and benefits, are they clear about what matters to them, and do they have enough support to make a choice?).40101 (4) Plan the next steps (make or defer a decision).40 Comparing this approach with education alone, the researchers found no significant differences between groups over the two month study period, but knowledge (mean difference 3.51, 1.74 to 5.3) and asthma control (8.47, 2.51 to 14.4) increased significantly from baseline, the proportion of patients using their asthma drugs appropriately increased non-significantly from baseline (1.35, 0.80 to 2.27), and decisional conflict decreased significantly from baseline (−1.25, −2.97 to −1.90).96

A practice based RCT of 60 families over six months developed a patient portal for asthma (“MyAsthma”) and found that 92% of parents who used the portal reported that they had better communication with the office, were better able to manage their child’s asthma, and were more aware of the importance of treatment. In addition, families in the intervention group had fewer missed days of work due to asthma (−1.1 v 0.7; difference in difference of −1.8) and their children had asthma flares less frequently (−3.3 v −1.3, difference in difference of −2.0).98 Features of the portal include goal identification and tracking, tracking of symptoms and side effects, education for patients, and an asthma action plan. In addition, a review by Butz et al provides guidance on practicing shared decision making in children with asthma including using visual aids, turn taking, role playing, and teach-back.99

Risk prediction

A complementary approach to facilitating step-down is risk prediction. Using predictive risk modeling can help to make the risks of step-down not only more transparent but more personalized as well. Successful step-down can be facilitated by considering factors such as guideline eligibility for step-down, age, sex, race, comorbidities, healthcare utilization, and length of stability.83439102

A retrospective study examined 126 Japanese patients whose asthma had been controlled for three months and who had stepped down according to GINA guidelines, examining factors that predicted successful step-down.102 Twenty nine patients failed their step-down attempt; only comorbidity with rhinitis/rhinosinusitis and phlegm grade predicted failure (odds ratio 3.8, 1.0 to 13.3).

In a prospective observational study of 56 patients in Finland,39 control was defined by no courses of oral corticosteroids related to asthma for one year and an asthma control questionnaire score of less than 0.75. All patients were taking combination therapy with ICS-LABA. The step-down protocol involved discontinuing the long acting bronchodilator and after six weeks either directly discontinuing the inhaled corticosteroid if the dose was 400 μg or less of budesonide or equivalent or reducing the dose by half for six weeks before complete discontinuation. Patients were followed until exacerbation or stability off ICS-LABA for six weeks. Approximately half of the patients successfully stepped down from ICS-LABA therapy. The largest proportion of failures occurred after inhaled corticosteroids were stopped completely. Better lung function (percentage predicted forced expiratory volume in one second) and level of control (according to the asthma control questionnaire) were predictive of success.

Finally, asthma step-down was reviewed within two large secondary datasets (the Medicare Expenditure Panel Survey and the Optum Labs Data Warehouse).834 Together, these studies included more than 30 000 US patients; definitions of step-down were based on GINA guidelines but tailored to fit the structure of the data. The review concluded that step-down eligibility, duration of stability, depression, age, sex, race, comorbidity, and healthcare utilization could all be used to predict the success of step-down.

These studies highlight factors that clinicians can use to guide clinical decision making. However, existing studies are limited by small sample sizes used to derive the predictive factors,39102 or they have yet be to be validated in independent samples and prospectively tested.834 Studies of step-down in which asthma endotypes are carefully classified were not identified but are likely to improve predictive models. Despite these limitations, combining the use of predictive modeling with shared decision making may help to facilitate the implementation of step-down into practice by overcoming existing barriers, such as uncertainty around the patient’s risk.

The conversation around step-down could occur as early as at the time of diagnosis. At this time, the clinician could describe asthma and its causes and consequences, as well as the general treatment and management approach (including the potential for step-down). Many patients are concerned about cost or have trouble affording their asthma drugs. Clinicians should discuss the affordability of the prescription for the patient and work with the patient to develop a treatment plan that is affordable or connect them with resources that may cover the costs of the prescriptions.103104105106 Then, using predictive risk modeling, the clinician could describe what the patient’s individualized risk of exacerbation (or other adverse consequences of asthma) is if they adhere to the treatment or if they are non-adherent. The clinician should also encourage open communication between the patient and the healthcare team so that triggers and other factors can be identified and managed. If step-down is forecast at diagnosis or at step-up, future conversations around step-down are likely to be more readily accepted.


The most recent international guideline on asthma, the 2017 GINA guideline, lists several recommendations for stepping down asthma drugs that are generally consistent with the recommendations made in this review.1 Firstly, step-down should not occur until the patient has been stable for at least three months and should be timed to avoid potentially high risk time periods such as respiratory infection, traveling, and pregnancy. Our review identified evidence suggesting that for many patients, stability longer than three months is likely to lead to safer step-down outcomes.834 Secondly, a monitoring and re-evaluation plan should be agreed on by the patient and healthcare professional, which may include a written asthma action plan. The findings from this review, highlighted in the patient centric model of stepping down asthma treatment, support the importance of anticipatory guidance and ongoing monitoring. Thirdly, several different specific drug step-down options are available (for example, ICS-LABA to inhaled corticosteroid alone). Varying levels of evidence, as described in this review, support different step-down choices.

Future directions

For clinicians to implement step-down more confidently in practice, certain gaps in the evidence need to be addressed. Although some evidence around the risk of step-down in asthma exists, many of the approaches have evidence that is limited by study design (lack of randomized trials or problems with risk of bias), study size (thus limiting the number of events and precision around the estimate of effect), heterogeneity (potential problems combining studies of children and adults and studies in which asthma endotypes are not sorted), or simply a lack of evidence.

Large, high quality randomized trials are needed to build a more robust evidence base around the risk of step-down, especially in settings outside of developed countries where step-down choices may vary considerably. Trial designs should be attentive to the risks of adverse drug effects, costs, treatment burden, and additional treatment step-down questions (for example, step-down from theophylline or long acting muscarinic antagonists).

To make the evidence from randomized trials more applicable to the individual patient, further studies are needed to identify and validate predictors of successful step-down. Although some of these can be identified in randomized trials, these trials often lack the power to robustly assess whether a subgroup effect exists. Therefore, large observational studies, ideally combining electronic health record data with claims data, are needed to better identify patients who are most likely to benefit from step-down in real world clinical settings.

Finally, perhaps the largest gap in evidence is on how to successfully implement step-down in practice. Future studies should examine the effect of components such as education of patients or clinicians around the appropriateness of step-down, the influence of incentives for clinicians or patients on the likelihood of step-down, and personalized approaches such as risk prediction and shared decision making, as well as other system or structural changes to facilitate step-down.


Before any attempt to step down chronic asthma drug, careful assessment of current and past asthma stability is important. It is helpful to consider drug adherence, device technique, and seasonal and situational asthma triggers when determining the best time for stepping down asthma controller drugs. Being mindful of factors at the level of the patient, clinician, and system will help to overcome barriers to stepping down chronic asthma drugs. Many specific options for stepping down asthma controller drugs are supported by evidence that can help clinicians to estimate outcomes when conversing with patients. Systematic reviews of multiple trials support the safety of the following specific step-down approaches: optimizing inhaled corticosteroid dosing when stepping down oral corticosteroids, reducing inhaled corticosteroid from a higher dose, lowering ICS-LABA dose from a higher dose while adding ICS-LABA on-demand, adding LTRA while lowering inhaled corticosteroid dose, and using allergen immunotherapy when reducing inhaled corticosteroid from a higher dose. Systematic reviews of multiple trials support an increased risk of asthma exacerbation for patients who completely stop taking inhaled corticosteroid or long acting bronchodilator. Finally, it is important for clinicians to regularly reassess patients who step down asthma controller drugs.

Future research questions

  1. What is the risk of step-down for different levels of treatment?

  2. Which patients have the highest/lowest likelihood of successful step-down?

  3. How can step-down be routinely implemented into practice?

Patient involvement

To gain insight on the patient’s perspective around step-down of asthma drugs, we spoke with the members of two patient advisory councils. Conversations with members of the advisory councils helped to inform the development of a patient centric model of step-down. Many patients’ experience of asthma, especially initially, is one of fear. Over time, through education and experience, fear of asthma may be replaced by understanding, acceptance, or comfort. This change may help to facilitate step-down. Our patient partners also emphasized the need for close collaboration between patients and clinicians, as well as the need for the process of step-down to include education, monitoring, and follow-up to ensure that it is done as safely as possible.

Key points on stepping down chronic asthma treatment

  • Step-down is recommended by guidelines

  • Step-down is not routinely integrated into most practices

  • Step-down may allow for reassessment of an asthma diagnosis

  • Step-down may reduce costs, risk of adverse effects of drugs, and treatment burden

  • Consider adherence to treatment, device technique, and asthma triggers in step-down decisions

  • Personalize the risk estimation for patients who are considering step-down

  • Use shared decision making approaches to facilitate step-down

  • The risk of stepping down seems to be highest when inhaled corticosteroid or long acting bronchodilator is completely stopped

  • Monitor patients after step-down, making a clear plan for when reassessment occurs


  • Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors

  • We acknowledge the contributions from the Community Engaged Research Advisory Board (Rochester, MN) and the Mayo Clinic Family Medicine Advisory Board (Phoenix, AZ) in the development of the patient centric model of stepping down asthma drugs.

  • Contributors: MRG and JBH contributed equally to the article. All authors substantially contributed to the design of the work, interpretation of the data, and drafting of the manuscript. All authors revised the manuscript critically for important intellectual content and approved the final version. All three authors are guarantors.

  • Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: MRG has received research support from AstraZeneca, has sat on advisory boards sponsored by Pfizer in the area of shared decision making, and has had travel and an honorarium paid by the Pharmaceutical Research and Manufacturers of America Foundation to Geisinger for speaking on drug adherence at a symposium; JBH has had research support paid to his institution by GlaxoSmithKline, Astra-Zeneca, Medimmune, Sanofi, and the NIH.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract