Systematic Literature Review of Systemic Corticosteroid Use for Asthma Management

Abstract

Systemic corticosteroid use to manage uncontrolled asthma and its associated healthcare burden may account for important health-related adverse effects. We conducted a systematic literature review to investigate the real-world extent and burden of systemic corticosteroid use in asthma. We searched MEDLINE and Embase databases to identify English-language articles published in 2010–2017, using search terms for asthma with keywords for oral corticosteroids and systemic corticosteroids. Observational studies, prescription database analyses, economic analyses, and surveys on oral/systemic corticosteroid use in children (>5 yr old), adolescents (12–17 yr old), and adults with asthma were included. We identified and reviewed 387 full-text articles, and our review included data from 139 studies. The included studies were conducted in Europe, North America, and Asia. Overall, oral/systemic corticosteroids were commonly used for asthma management and were more frequently used in patients with severe asthma than in those with milder disease. Long-term oral/systemic corticosteroid use was, in general, less frequent than short-term use. Compared with no use, long-term and repeated short-term oral/systemic corticosteroid use were associated with an increased risk of acute and chronic adverse events, even when doses were comparatively low. Greater oral/systemic corticosteroid exposure was also associated with increased costs and healthcare resource use. This review provides a comprehensive overview of oral/systemic corticosteroid use and associated adverse events for patients with all degrees of asthma severity and exposure duration. We report that oral/systemic corticosteroid use is prevalent in asthma management, and the risks of acute and chronic complications increase with the cumulative oral corticosteroid dosage.

Contents

Methods

 Search Strategy and Selection Criteria

 Data Analysis

Results

 Manuscript Identification

 Study Populations

 Patterns of CS Use

 Clinical Burden of SCS

 Healthcare Resource Use and Economic Impact of SCS Use

Discussion

Systemic corticosteroids (SCS) became available in 1956, and their introduction provided effective treatment for the control of asthma symptoms and exacerbations (1). However, their widespread use quickly led to the recognition that long-term SCS use is associated with significant adverse events (AEs) (2). Inhaled CS (ICS), which have a reduced risk of AEs but are as effective as SCS for most patients, were introduced in 1972 as maintenance treatment for patients with asthma (1, 3). However, SCS, usually in the form of oral CS (OCS) and occasionally as injectable CS, remained the mainstay treatment for asthma exacerbations and severe disease over the next four decades (1, 4).

Today, ICS are the primary therapeutic intervention for persistent asthma along with other controller therapies, including predominantly long-acting β₂-agonists (LABAs) and leukotriene receptor antagonists, as additional treatments to reduce ICS dosages, control asthma symptoms, and decrease exacerbation risk for patients with asthma (5). Add-on treatments, traditionally long-acting muscarinic antagonists or low-dosage OCS (before the introduction of targeted biologics), are recommended for patients with asthma that is not controlled by medium- to high-dosage ICS plus controller medications (5). In 2003, omalizumab (anti-IgE therapy), the first targeted biologic therapy for asthma, was approved by the U.S. Food and Drug Administration for add-on maintenance treatment (6). Subsequently, the anti–IL-5 treatments mepolizumab and reslizumab and the anti–IL-5 receptor α–directed cytolytic therapy benralizumab were approved for the treatment of severe, eosinophilic asthma (6, 7). More recently, the anti–IL-4 and anti–IL-13 therapy dupilumab was approved for the treatment of moderate-to-severe eosinophilic or OCS-dependent asthma (8). These targeted biologic treatments have demonstrated greater specificity for achieving disease control by reducing the risk of exacerbations and requirements for rescue medication and OCS use in their respective target patient populations, with limited AEs (7–17). Biologics are now also recommended in guidelines for the treatment of appropriate patients with severe asthma (5). For patients with severe asthma who are not eligible for the currently available biologic treatments, the 2019 Global Initiative for Asthma (GINA) guidelines recommend that several other strategies be considered before maintenance OCS/SCS (5).

Despite the availability of these new well-tolerated, effective, targeted biologic add-on treatments and the well-recognized AEs associated with SCS use, OCS are still treatment options in the current asthma treatment guidelines. The 2019 GINA guidelines recommend that if OCS are prescribed, they should be prescribed at lesser dosages. Prescribers should be fully aware of and monitor for AEs. OCS should be considered, along with targeted biologic treatments, for add-on treatment for patients with uncontrolled asthma despite the use of high-dosage ICS therapy (GINA step 5 treatment) (5). The GINA guidelines also recommend short-term OCS use for patients experiencing a severe exacerbation who do not respond to treatment (5). However, this guidance is relatively unspecific. Patients can repeatedly receive prescriptions for both short- and long-term OCS and ultimately become OCS dependent (18). The continued inclusion of OCS in guidelines, together with their worldwide easy accessibility, familiarity of use, and low acquisition costs compared with newer targeted treatments, contributes to the ongoing use of SCS for patients with severe asthma. In addition, it is likely that some patients fail to benefit significantly from the OCS-sparing effect of biologic treatments because of differences in their susceptibility to OCS-related adverse effects, or because of their unwillingness to initiate new treatment options and reduce OCS use.

Studies from France and the United Kingdom provide evidence that overall OCS use has increased over the last decade and continues to increase (19, 20). Respiratory disease is the most frequently recorded indication for OCS treatment, accounting for approximately 40% of prescriptions (20, 21). OCS use has declined in recent decades for other indications for which biologics have been available for more than 20 years. In rheumatology, for example, the introduction of biologics has significantly reduced OCS use, and this change has been cost-effective in the treatment of rheumatoid arthritis (22, 23).

Responses to SCS and the risk of AEs vary considerably among patients (24). Many patients with persistent SCS use demonstrate a relative resistance to treatment (24), but they continue to receive SCS prescriptions and increasing dosages, which further contributes to the prevalence of long-term use. In recent studies in which responses to SCS were comprehensively characterized, 25–35% of patients with asthma did not respond to OCS treatment by exhibiting evidence of a reduction in type 2 biomarkers, suggesting SCS resistance (25, 26). However, despite these findings and the extensive variation in prescription durations and dosages used in clinical practice, the optimal duration and dosage of SCS maintenance and short-term treatment have not been studied extensively (27). A recent systematic literature review evaluated the long-term use of OCS for patients with asthma and reported that the risk of developing OCS-related complications, including infections, diabetes, osteoporosis, and psychiatric disorders, was greater for patients with long-term OCS exposure compared with control groups, even for those receiving dosages below 5 mg/d (28). However, these findings were limited to adult patients with severe asthma and long-term OCS use and were based on just nine publications (seven large datasets).

In this systematic literature review, we investigated the extent and nature of real-world SCS use (oral and parenteral CS) for the treatment of asthma in children (>5 yr old), adolescents (12–17 yr old), and adults with asthma of any severity. We also examined available evidence regarding the clinical and economic impacts of sporadic, repeated short- and long-term SCS use in the general asthma population and by the degree of asthma severity. The findings of this review provide a comprehensive overview to improve our understanding of the risk of health-related adverse effects associated with SCS use for patients with asthma and the extent of SCS use for patients with all degrees of asthma severity.

Methods

Search Strategy and Selection Criteria

A literature search was conducted in MEDLINE and Embase databases via Ovid to identify English-language articles published between January 1, 2007, and December 4, 2017, using the following search terms: asthma AND ([corticosteroid OR corticosteroids OR glucocorticoid OR glucocorticoids OR prednisone OR prednisolone OR dexamethasone OR methylprednisolone OR hydrocortisone] AND [oral OR orally OR systemic OR parenteral OR intravenous OR intramuscular]) OR [SCS OR OCS]). Filters were applied to exclude reviews, letters, and information presented at scientific conferences. Full search details are provided in Table E1 in the online supplement. Searches were initially restricted to English-language articles published since 2007. However, because of the large number of relevant articles identified, and to ensure that the most up-to-date data relevant to current clinical practices were evaluated, the search was subsequently limited to articles published since January 1, 2010.

Abstracts and full-text articles were screened to determine their eligibility for inclusion according to prespecified patient-, intervention-, comparison-, outcome-, and study-design–related selection criteria (Table 1). To summarize, the systematic literature review included observational studies that reported information on frequency of SCS use and treatment patterns, as well as clinical complications and the economic burden associated with SCS use for children (>5 yr old), adolescents (12–17 yr old), and adults with asthma of any degree of severity. Studies that included <100 patients with asthma were excluded, as this sample size was considered too small to provide results likely to be representative of the overall asthma population.

Table 1.

Study Selection Criteria for the Systematic Literature Review on the Use of OCS and SCS for Treatment of Asthma

Category	Inclusion Criteria	Exclusion Criteria
Population	• Pediatric, adolescent, and adult patients (≥5 yr old) with asthma of any severity and all degrees of disease control • Mixed study populations or subpopulations in which ≥85% of patients met the above criteria (apart from age criteria, which had to be met by all patients)	• Children <5 yr old with asthma • Patients who did not have asthma • Patients with ACOS • Patients with asthma during pregnancy • Mixed-study populations or subpopulations in which <85% of patients met the inclusion criteria • Studies containing <100 patients with asthma • Studies containing <500 patients with asthma, except for those reporting long-term OCS/SCS use or burden of OCS/SCS use
Intervention/comparators	• OCS, parenteral CS, or SCS	• No mention of OCS, parenteral CS, or SCS use • CS treatment in emergency department and/or in hospitalized patients with asthma only
Outcomes	• Frequency/patterns of OCS and SCS use (patient-level data) • Long-term OCS and SCS use, including Length of time (average duration) receiving long-term OCS Dosage and frequency of changing dosage	• Outcome measures not listed in the inclusion criteria • Frequency/patterns of OCS and SCS use among physicians
	• Clinical burden of OCS and SCS use: Asthma clinical features   Exacerbation history   Previous hospitalizations   Degree of asthma control (based on ACQ or ACT)   Phenotypes (eosinophilic, allergic, etc.) Comorbidities and complications for OCS and SCS users:    Diabetes/metabolic    Bone (e.g., osteoporosis)    Cardiovascular disease    Psychiatric (e.g., depression)    Others as available
	• Economic burden of OCS and SCS use:  Healthcare resource use, including hospitalizations and doctor and emergency department visits  Costs (including direct and indirect costs), both asthma-related and all costs  Cost consequences of prolonged OCS and SCS use
Study design	• Observational studies, including prospective and retrospective cohort studies, and cross-sectional analyses • Prescription database analyses • Economic analyses • Patient, parent/guardian, and HCP surveys	• Systematic and narrative reviews • Case reports and case series • Comments, letters, and editorials • Animal/in vitro studies • Meta-analyses/pooled analyses • Clinical trials • Asthma phenotyping analyses • Treatment guidelines • Patient education studies • Studies of healthcare management (e.g., physician prescribing preferences, physician/hospital management approaches, and care quality evaluation)
Time period	• January 1, 2007, to December 4, 2017	• Studies published before January 1, 2007, or after December 4, 2017
Other criteria	• Studies published in English • Limited to humans	• Non–English language publications • Conference abstracts

Definition of abbreviations: ACOS = asthma–chronic obstructive pulmonary disease overlap syndrome; ACQ = Asthma Control Questionnaire; ACT = Asthma Control Test; HCP = healthcare professional; OCS = oral corticosteroids; SCS = systemic corticosteroids.

Data Analysis

Data from the included articles were extracted into a Microsoft Excel spreadsheet. At extraction, studies were separated into the following categories of use: oral (OCS), parenteral, and systemic (both oral and parenteral; SCS). The evidence identified was summarized qualitatively, and studies were separated according to whether they reported long- or short-term use.

Results

Manuscript Identification

The search identified 2,277 unique references across the databases. Abstract screening identified 387 references that required further review of full-text articles. Of those, 149 publications met the criteria for inclusion (Figure 1). During data extraction, related publications reporting on the same patient populations or updates from the same studies were identified. To avoid study duplication, data from these publications were only included once in the analysis, resulting in 139 studies included in the final analyses. The most common reason for exclusion at the full-text level was no relevant data.

Figure 1.

Flow diagram of the article screening and evaluation process. *“Other” included asthma–chronic obstructive pulmonary disease overlap syndrome (n = 5), letter (n = 4), erratum (n = 2), not English language (n = 2), and retracted publication (n = 1). ^†“Other” included cost-effectiveness modeling, letter, not patients with asthma, not English language, asthma phenotyping analysis, controlled trial, and retracted publication (each n = 1). ED = emergency department.

Study Populations

The included studies were conducted in Europe, North America, and Asia (Table E2). However, 64% of the studies presented data from centers or databases in the United States (n = 55) or the United Kingdom (n = 34). UK studies accounted for 52% (n = 34/65) of all studies from Europe. Studies reporting data for adults (n = 48/139) or combined adult/adolescent populations (n = 44/139) accounted for approximately two-thirds of the studies (Table E2). Seventeen studies reported data for individuals ≥5 years of age. Age limits were undefined in 14 studies, but mean or median ages of ≥40 years were reported, suggesting that the studies consisted of largely adult populations.

Overall, 41% (n = 57) of the studies involved patients with any degree of disease severity; 23% (n = 32) included data for patients with GINA step 2 or greater treatment, and 19% (n = 26) involved only patients with severe asthma (GINA step 4 or 5). Studies of patients with severe asthma were smaller in size than those that covered a wider degree of disease severity, with 22 of 26 studies including fewer than 1,000 patients. The majority of studies reported data for OCS use (87%; n = 121), although some institutions, including centers participating in the Severe Asthma Research Program, exclusively recommended SCS use. However, because the results reported for SCS use were similar to those reported for OCS use, the rest of this review will focus on OCS use. The results for SCS use are reported in the relevant corresponding tables and online supplement.

Patterns of CS Use

Patterns of short-term CS use

Short-term use was defined by terms such as “acute,” “burst,” and “for asthma exacerbations,” or when OCS/SCS were used as reliever medications or had a defined short-term exposure. Overall, 58 studies reported short-term OCS/SCS use for patients with asthma (Table E3). Short-term OCS/SCS use in studies of patients with any degree of asthma severity ranged from 3.6% (underweight patients before OCS/SCS exposure in a U.S. study of children with asthma living in the vicinity of wildfires) (29) to 62.0% (U.S. observational study of primary care patients with asthma) (30). Short-term OCS/SCS use in studies of patients receiving GINA step 2 or greater treatment ranged from 2.1% (international study of SCS use for patients with asthma ≥12 yr old) (31) to 41.9% (fluticasone propionate/salmeterol fixed-dose combination inhaler use in a U.S. retrospective study) (32). Short-term OCS/SCS use in studies of patients with severe or difficult-to-treat asthma ranged from 23.2% (patients who experienced exacerbations within the prior 3 mo in a U.S. study of severe refractory asthma) (33) to 92.6% (UK study of OCS use for patients with severe asthma [age unspecified]) (34). The percentage of patients receiving short-term OCS/SCS therapy increased with increasing disease severity.

Patterns of short-term CS use by disease severity

Five retrospective cohort studies of adult patients with asthma of any degree of severity reported that approximately a quarter of patients required at least one short-term OCS course during a 1-year period (29, 35–38). Studies involving patients of all ages reported incidences of 16.2–30.9% for 1-year use of short-term OCS (39–41).

For patients with severe or difficult-to-treat asthma, short-term OCS use increased to 46.3–92.6% over a 1-year period (33, 34, 42–46). In one cross-sectional study of patients with uncontrolled asthma, 60% of whom had severe asthma, 24.4% overall were reported to have received greater than or equal to 3 short-term OCS courses in the previous year (42).

Short-term CS prescription use

The mean number of OCS short-term prescriptions ranged from 0.1 to 2.16 prescriptions per year in the nine studies that reported these data for patients with any degree of asthma severity or those using controller therapy (GINA step 2 or greater treatment) (32, 47–54). Studies involving patients with severe or uncontrolled asthma reported greater prescription rates compared with patients with less severe disease. A study that examined prescription rates in relation to treatment steps (according to the British Thoracic Society guidelines) reported that the mean number of OCS courses per year ranged from 1.2 to 2.1 at steps 1–4 and 5.3 at step 5 (maintenance OCS therapy) (47).

Patterns of long-term CS use

A total of 62 studies evaluated long-term OCS (54 studies) and SCS (eight studies) use for patients with asthma, defined as “daily” or “continuous” OCS/SCS use, or described OCS/SCS use as “chronic,” “maintenance,” or “controller medication,” or specified durations of long-term OCS/SCS exposure. For patients with any degree of asthma severity, long-term OCS/SCS use ranged from 1.2% (UK retrospective study of asthma therapy) (47) to 30.9% (patients with asthma and vitamin D insufficiency in a German study) (55) (Table 2 [extended version: Table E4]). For patients receiving GINA step 2 or greater treatment, long-term OCS/SCS use ranged from 0% (patients with nonsevere asthma in a U.S. study) to 100% (UK study of patients with severe asthma) (54). In general, long-term OCS/SCS use was less frequent than short-term use.

Table 2.

Frequency of Long-Term Use of OCS and SCS for Patients with Asthma, Categorized by Disease Severity

Source	N	Long-Term OCS/SCS Definition	OCS/SCS Use at Follow-Up/Postindex, % (n/N)	OCS/SCS Use at Baseline/Preindex, % (n/N)
Any degree of asthma severity
Allen-Ramey et al., 2013 (30) (U.S.)	21,199	An order quantity ≥30 with one or more refills	OCS	—
			8.9 (1,883/21,199)
Arellano et al., 2011 (79) (U.S.)	6–18 yr: 659,169	Continuous OCS use >15 d	Addition of long-term OCS in first year in Tx-naive pts initiating Tx with:	—
	6–11 yr: 348,991		SABA (n = 309,947): 22.8%
	Tx-naive (6–18 yr): 595,619		ICS (n = 16,783): 18.5%
			ICS/LABA (n = 13,980): 15.1%
Bottero et al., 2014 (125) (Italy)	159	Continuous or near continuous (≥50% of yr) oral prednisone use	OCS	—
			9.4 (15/159)
			HLA-DRB4 positive: 16.7 (10/60)
			HLA-DRB4 negative: 5.1 (5/99)
Broder et al., 2010 (56) (U.S.)	18,343 (uncontrolled asthma)	Total supply of ≥60 d in a 6-mo period	—	OCS Preindex (EPR3 step 6): 0.6 (105/18,343)
Covvey et al., 2013 (47) (UK)	12,319	>14-d supply with no titration schedule (BTS/SIGN step 5)	1.2 (149/12,319)	—
Dalal et al., 2016 (103) (U.S.)	603,147; SCS users: 12,697 Nonusers: 590,450	SCS user: ≥6 mo of continuous long-term SCS use Nonuser: never exposed to SCS	—	Long-term SCS use (overall pop): 2.1 (12,697/603,147)
Dodd and Mazurek, 2018* (59) (U.S.)	14,915	Controller therapy	OCS	—
			% (95% CI)
			WRA: 5.5 (3.8–7.3)
			Possible WRA: 3.0 (1.8–4.2)
			Non-WRA: 2.5 (1.7–3.4)
Fardet 2011 (57) (UK)	4,518,753 (total) 167,886 (long-term OCS)	Tx lasting ≥3 mo	OCS	—
			Prevalence, % (95% CI): 1.3 (1.1–1.4)†
Ferguson et al., 2014‡ (60) (U.S.)	812	Long-term OCS	OCS	—
			9 (70/812)
Hasegawa et al., 2012‡§ (61) (Japan)	1998: 3,347	Controller medication	OCS	—
			1998: 18.8
	2000: 3,069		2000: 12.3
	2002: 2,593		2002: 10.4
	2004: 2,865		2004: 7.4
	2006: 3,066		2006: 7.8
	2008: 3,146		2008: 5.2
			P < 0.001 over study
Korn 2013§|| (55) (Germany)	280	Daily OCS maintenance	OCS	—
			26.8 (75/280)
			Vitamin D concentrations
			25(OH)D <30 ng/ml (insufficiency):
			30.9 (58/188)
			25(OH)D ≥30 ng/ml:
			18.5 (17/92)
			P = 0.031
Lee et al., 2013 (98) (South Korea)	TB cases: 4,136 Matched control subjects: 20,538¶	OCS user: cumulative dosage ≥1,680 mg of hydrocortisone equivalents during 1 yr before index date	—	OCS
				Asthma pts
				TB cases: 11.2 (54/484)
				Control subjects: 4.8 (117/2,420)
Lefebvre et al., 2017** (106) (U.S.)	SCS users: 3,628 Nonusers: 26,987	Daily doses ≥5 mg of prednisone equivalent with no gap of ≥14 d between 2 SCS claims	—	Overall population: 11.9 (3,628/30,615)
Luskin et al., 2016†† (49) (U.S.)	3,604 (high OCS use)	High OCS use: pts who had a ≥30-d supply of OCS in each study year	OCS	—
			5.3 (3,604/67,860)
Papaioannou et al., 2016 (62) (Greece)	171	Regular/continuous SCS	OCS	—
			20.5 (35/171)
Price et al., 2015 (35) (UK)	2,042‡‡	BTS step 5	—	OCS
				0.5 (10/2,042)
Price et al., 2016 (58) (UK)	130,547	BTS step 5	—	OCS
				0.8 (1,080/130,547)
Reddy et al., 2011§§ (63) (U.S.)	(257‖‖)	Regular OCS	OCS	OCS
			5.8 (15/257)	10.9 (28/257)
			P = 0.055 vs. baseline	—
Sato et al., 2017 (64) (Japan)	114	Regular use of OCS	OCS	—
			4 (5/114)
Shigemura et al., 2012‡ (65) (Japan)	126	Regular OCS	—	OCS
				Baseline: 15.4 (18/117)
Tattersall et al., 2015|| (66) (U.S.)	667; Intermittent: 511 Persistent: 156	Controller medication	—	OCS
				Overall: 4.8 (32/667)
				Persistent: 20.5 (32/156)
				Intermittent: NA
Zeiger et al., 2017 (94) (U.S.)	9,546	Long-term OCS: Average daily dosage ≥2.5 mg in 2010 Short-term OCS: average daily dosage <2.5 mg or no OCS in 2010	Long-term OCS: 8.2 (782/9,546)	—
	Long-term OCS: 782 Short-term OCS: 8,764
GINA step 2 or greater treatment
Bengtson et al., 2017 (4) (U.S.)	Escalation: 5,044 Unchanged: 21,967	≥90 consecutive days of OCS coverage	Unchanged: 0.2 (52/21,967) Escalation:	—
			Before: 0.1 (6/5,044) After: 0.1 (7/5,044)
Hawcutt et al., 2015¶¶ (126) (UK)	525	Regular maintenance	10.8 (47/435)	—
Barry 2017†† (96) (UK)	7,195; Severe: 808 Mild/moderate: 3,975 Nonasthma: 2,412	Severe asthma: regular OCS use***	Severe: 100 (808/808) Mild/moderate: 25 (995/3,975) Nonasthma: 0 (0/2,412)	—
Broder 2017 (92) (U.S.)	3,355	High OCS users: ≥1 OCS fill with ≥30 d of supply or ≥6 bursts of OCS	High OCS use: 15.4 (517/3,355)	—
Chipps et al., 2017‡ (33) (U.S.)	341†††	Long-term SCS	11.2 (37/331)	—
Daugherty et al., 2017 (104) (UK)	60,418; SCS nonuser: 24,994 SCS user: 35,444	SCS user: SCS use at baseline and observation periods‡‡‡	58.6 (35,444/60,418)	25.6 (15,490/60,418)
		SCS nonuser: no SCS use at baseline or observation periods‡‡‡	(SCS use during baseline and observation periods)	—
Denlinger et al., 2017‡ (42) (U.S.)	709	Daily OCS	11.0 (78/709)	—
Gibeon et al., 2015|| (43) (UK)	346	Maintenance OCS	42.6 (123/289)	41.2 (119/289)
Lefebvre et al., 2015** (105) (U.S.)	3,628; SCS exposure (mg/d): Low (≤6): 368 Medium (>6–12): 1,630 High (>12): 1,630	Daily SCS dosage ≥5 mg of prednisone equivalent with no gap of ≥14 d between two SCS claims	At index date§§§	—
			Low SCS: 10 (368/3,628) Medium SCS: 45 (1,630/3,628) High SCS: 45 (1,630/3,628)
Maio et al., 2017|| (46) (Italy)	493	Long-term OCS use	16.0 (78/488)	—
Moore 2011|| (119) (U.S.)	339; Nonsevere: 196 Severe: 102 Very severe: 41	OCS ≥20 mg/d for ≥50% of year	—	Baseline: %
				Nonsevere: 1 (2/196)
				Severe: 21 (21/102)
				Very severe: 80 (33/41)
				P < 0.0001
O’Neill et al., 2015 (34) (UK)	596; Severe: 516 Nonsevere: 80	Maintenance OCS	—	Overall: 34 (201/596)
				Severe: 38 (196/516)
				Nonsevere: 5 (6/80)
Phipatanakul et al., 2017|| (25) (U.S.)	6–17 yr: 188; Nonsevere: 77 Severe: 111	≥3 mo with OCS use in past year	—	6–17 yr
				Nonsevere: 1.3 (1/77)
				Severe: 9.9 (11/111) P < 0.05
	Adult (≥18 yr): 526; Nonsevere: 213 Severe: 313			Adult
				Nonsevere: 0 (0/213)
				Severe: 22.4 (70/313) P < 0.01
Reddy et al., 2014§§ (67) (U.S.)	228; Current (2003–2007): 65 Historic (1993–1997): 163	Daily OCS use	Current: 28 (11/41) Historic: 51 (28/55) P = 0.002	—
Rijssenbeek-Nouwens et al., 2012|| (127) (the Netherlands)	137‖‖‖	Daily OCS maintenance	Overall: 29.9 (41/137)	Overall: 51.1 (70/137)
	HDM: 68		HDM: 22 (15/68); P < 0.001 vs. baseline	HDM: 43 (29/68)
	Non-HDM: 69		Non-HDM: 38 (26/69); P < 0.001 vs. baseline	Non-HDM: 59 (41/69)
	SEN: 92		SEN: 29 (27/92); P < 0.001 vs. baseline	SEN: 49 (45/92)
	Non-SEN: 45		Non-SEN: 31 (14/45); P < 0.001 vs. baseline	Non-SEN: 56 (25/45)
Schleich et al., 2014‡ (128) (Belgium)	350	Daily maintenance SCS	24 (84/350)	—
Shaw et al., 2015‡|| (75) (Europe)	209; Severe nonsmoker: 311 Sever current/ex-smoker: 110 Mild/moderate: 88	Daily OCS	—	Severe all: 45.5 (181/398)
				Severe nonsmoker: 45.8 (135/295)
				Severe ex-smoker: 44.7 (46/103)
				Mild/moderate: 0 (0/88)
Sweeney et al., 2012 (76) (UK)	349	Maintenance OCS	57 (199/349)	42 (146/349)
Sweeney et al., 2016 (102) (UK)	770¶¶¶	Daily SCS	57.1 (442/770)	—
Tay et al., 2017‡|| (129) (Singapore)	423	Maintenance OCS	—	1.4 (6/423)
				Severe: 4.1 (2/49)
				Nonsevere: 1.1 (4/374)
Westerhof et al., 2016‡ (95) (the Netherlands)	153; Current/ex-smoker: 83 Never-smoker: 70	Long-term OCS use >50% past yr	—	Current/ex-smoker: 28 (23/83)
				Never-smoker: 29 (20/70)

Definition of abbreviations: 25(OH)D = 25-hydroxyvitamin D; BTS = British Thoracic Society; CI = confidence interval; EPR = expert panel report; GINA = Global Initiative for Asthma; HCP = healthcare professional; HDM = house dust mite; ICS = inhaled corticosteroids; LABA = long-acting β₂-agonists; NA = not applicable; OCS = oral corticosteroids; pts = patients; SABA = short-acting β₂-agonists; SCS = systemic corticosteroids; SEN = sensitized; SIGN = Scottish Intercollegiate Guidelines Network; TB = tuberculosis; Tx = treatment; WRA = work-related asthma.

Studies used a retrospective cohort study design unless otherwise stated.

^*Patient survey.

^†Prevalence was determined by dividing the number of person-years with asthma receiving long-term OCS therapy by the total number of person-years with asthma.

^‡Cross-sectional study.

^§Patient/HCP survey.

^||Prospective study.

^¶Cases of TB identified after the treatment initiation date were matched with up to five control individuals without TB for age, sex, diagnosis of asthma or chronic obstructive pulmonary disease, and initiation date.

^**Longitudinal open cohort.

^††Retrospective matched cohort.

^‡‡Patients who received tiotropium add-on therapy postindex (dry powder inhaler or soft mist inhaler).

^§§Retrospective cross-sectional and historic cohort.

^‖‖Patients with asthma who underwent bariatric surgery, consented to and had reached 1 yr of follow-up (n = 606), and returned for a follow-up survey (n = 257).

^¶¶Prospective/retrospective cohort.

^***GINA Step 5 treatment and four or more OCS prescriptions per year for each of the two consecutive study years.

^†††Data from TENOR (The Epidemiology and Natural History of Asthma. Outcomes and Treatment Regimens) II, a 10-year follow-up assessment of patients from TENOR I; age criteria provided for inclusion in TENOR I study.

^‡‡‡Baseline period: 6 months before index date (date the patient was identified as having severe asthma [GINA Step 4/5]); observation period: follow-up after index date.

^§§§Index date was defined as the first day with a daily dosage of ≥5 mg of prednisone or equivalent after the first 6 months of long-term SCS use (baseline period).

^‖‖‖High-altitude therapy for patients with severe asthma and without HDM sensitization (HDM and non-HDM groups) and patients with and without any allergic sensitization (sensitized and nonsensitized groups).

^¶¶¶Patients with severe asthma at registry baseline assessment. Patients were divided into two groups: those who required daily SCS therapy to maintain asthma control and those who did not require maintenance SCS but required frequent rescue SCS courses.

Patterns of long-term CS use by disease severity

In general, <12% of patients with any degree of asthma severity received long-term OCS therapy (although values varied from 0.5% to 26.8% in different studies) (35, 55). This degree of variation could partly be related to how the studies defined long-term use, as well as differences in study inclusion criteria, geographic region, and availability of OCS-sparing treatments. Rates in 11 studies that applied precise definitions ranged from 0.5% to 9.4%, with five studies reporting rates of <3% (35, 47, 56–58). Studies that used more generalized terms, such as “regular OCS” and “controller medication,” reported rates of 4–20.5% (59–66).

Long-term OCS use was consistently reported as low across all age groups for patients with nonsevere disease (1.3% and 0.0% of children/adolescents and adults, respectively) (25). Among patients with severe or uncontrolled asthma, 20–60% were reported to have received long-term OCS therapy.

Long-term CS dosage

The dosage of long-term OCS was reported in 23 studies, most of which reported results for patients with severe disease. The mean daily OCS dosage (expressed as prednisolone-equivalent dosages) ranged from 4.0 to 21.4 mg (Table E5) (67, 68). In 12 of these studies, the mean daily dosage was 10–22 mg (34, 43, 55, 68–76).

Patterns of general CS use

Studies that did not specify short- or long-term use were classified as reporting general OCS use. These were studies that described any OCS/SCS use, or those in which OCS/SCS use was undefined and/or described with general terms, such as “OCS prescriptions,” “OCS use,” and “OCS claims.”

General CS use by disease severity

Forty-one studies reported general SCS/OCS use. Overall, OCS/SCS were reported to be commonly used for asthma management (Table 3 [extended version: Table E6]). Use was more frequent for patients with severe asthma than for those with milder disease. Reported OCS/SCS use varied considerably across a variety of patient populations, ranging from 2.8% (UK study of OCS use in patients 13–65 yr old with GINA step 2 or greater treatment) (77) to 93.5% (U.S. study of OCS use for patients of all ages with uncontrolled asthma) (78). Variations were likely related to differences in inclusion criteria, geography, and availability of OCS-sparing treatments. No clear geographic or age-related variations in general OCS use were observed in the identified studies (Tables 3 and E7). OCS use was typically reported for approximately 20–48% of patients in the general asthma population over a 1-year study period (Table 3) (79–83).

Table 3.

General OCS and SCS Use for Patients with Asthma, Categorized by Disease Severity

Source	Sample Size	OCS or SCS Use at Follow-Up/Postindex, % (n/N)	OCS or SCS Use at Baseline/Preindex, % (n/N)
Any degree of asthma severity
Afshar et al., 2017 (130) (U.S.)	14,012 (all individuals)	OCS	—
	Asthma prevalence: 7.4%	8.8 (95% CI, 6.2–11.3)
Arellano et al., 2011 (79) (U.S.)	6–18 yr: 659,169	OCS	—
	6–11 yr: 348,991	6–18 yr: 25.2 (165,783/659,169)
		6–11 yr: 27.6 (103,292/374,068)
Black et al., 2012 (80) (U.S.)	74,057	—	OCS
			31.2
			(23,115/74,057)
Björnsdóttir et al., 2014 (131) (Iceland)	6,142	—	OCS
			12.7 (783/6,142)
Butler et al., 2016 (132) (U.S.)	123,868	OCS	—
		Jan 2005: 7.3%
Choi et al., 2017 (133) (South Korea)	831,613	OCS	—
		40.61
Cooper et al., 2015 (134) (UK)	2,624	OCS	—
		12.0 (314/2,624)
Delate et al., 2017 (135) (U.S.)	2,360	SCS	SCS
		30.6 (723/2,360)	35.8 (845/2,360)
		P < 0.001 vs. preintervention	—
Farber et al., 2017 (81) (U.S.)	5–8 yr: 20,645	OCS	—
	9–12 yr: 14,716	Rx (2,015)
	13–17 yr: 11,142	5–8 yr (N = 20,645)
		1: 31.9 (6,580)
		2: 6.6 (1,371)
		≥3: 3.5 (720)
		9–12 yr (N = 14,716)
		1: 28.2 (4,143)
		2: 5.3 (781)
		≥3: 2.2 (317)
		13–17 yr (N = 11,142)
		1: 30.1 (3,357)
		2: 4.3 (484)
		≥3: 2.4 (269)
Iribarren et al., 2012 (82) (U.S.)	203,595	—	OCS
			20%
Laforest et al., 2015 (136) (France)	UK: 38,637	SCS	—
	Non: 6,996 Low: 14,903 High: 16,738	Non/low/high ICS*
		UK: 6.2%/22.5%/12.7%
	France: 4,587	P < 0.0001
	Non: 1,176 Low: 1,358 High: 2,053	France: 21.9%/36.1%/30.5%
		P < 0.0001
Lee et al., 2014 (137) (South Korea)	736	SCS	—
		71 (523/736)
Lin et al., 2016 (107) (Taiwan)	24,109	—	SCS
			3 mo: 8.0 (1,926/24,109)
			12 mo: 22.3 (5,378/24,109)
Luskin et al., 2016 (49) (U.S.)	67,860	Any OCS use	—
		66.0 (44,764/67,860)
Walters et al., 2011 (101) (UK)	3,320	—	OCS
	Cases: 1,660		Cases: 57.4%
	Control subjects: 1,660†		Control subjects: 42.6%†P < 0.001
Windt and Glaeske, 2010 (138) (Germany)	DMP: 317	OCS	OCS
	Not DMP (control): 317‡	DMP: 26.5 (84/317)	DMP: 25.9 (82/317)
	Not DMP (all): 20,566	Control: 24.3 (77/317)§	Control: 20.5 (65/317)§P = 0.002
			Not DMP: 25.9 (5,320/20,566)
Wong et al., 2010 (83) (U.S.)	1,835	OCS	OCS
		48.2 (884/1,835)	48.1 (882/1,835)
GINA step 2 or greater treatment
Ali et al., 2015 (77) (UK)	51,103	—	OCS
	ICS: 46,928		Overall: 5.8 (2,976/51,103)
	LABA: 714		By postindex Tx
	ICS/LABA: 3,461		ICS: 5.7 (2,673/46,928)
			LABA: 6.4 (46/714)
			ICS/LABA: 7.4 (257/3,461)
Bengtson et al., 2017 (4) (U.S.)	Escalation: 5,044|| Unchanged: 21,967||	OCS	—
		Unchanged||
		31.9 (7,002/21,967)
		Escalation||
		Pre: 29.8 (1,501/5,044)
		Post: 30.7 (1,548/5,044)
Corrao et al., 2016 (84) (Italy)	2,335	OCS	—
		18.3 (428/2,335)
Hagiwara et al., 2010 (139) (U.S.)	FP: 469	SCS	—
	FSC: 3,881	Matched¶:
		FP: 32 (143/447)
		FSC: 24 (106/447) P = 0.006
Hagiwara et al., 2013 (140) (U.S.)	18,283	SCS	—
	FSC: 14,044	Matched**:
	MF: 4,239	FSC: 18.1 (688/3,799)
		MF: 20.5 (780/3,799) P < 0.001
Hagiwara et al., 2014 (141) (U.S.)	7,779	—	SCS claims and procedures
	FP: 2,010		FP: 69 (1,385/2,010)
	FSC: 5,769		FSC: 73 (4,232/5,769) P < 0.001
Laforest et al., 2014 (85) (France)	919	—	OCS
			Overall: 46.4 (394/849)
Laforest et al., 2014 (86) (France)	2,162	OCS	OCS
	ICS:	2008, % (n)	2007, % (n)
	1,757	ICS (N = 1,757)	ICS (N = 1,757)
	ICS+LTRA: 1,826	0 units: 57.4 (1,009)	0: 52.8 (928)
		1 unit: 22.1 (388)	1: 24.4 (428)
		2 units: 10.6 (187)	2: 11.5 (202)
		≥3 units: 9.8 (173)	≥3: 11.3 (199)
		ICS+LTRA (N = 1,826)	ICS+LTRA
		0 units: 57.7 (1,053)	(N = 1,826)
		1 unit: 22.1 (403)	0 units: 53.1 (970)
		2 units: 10.7 (195)	1 unit: 24.3 (443)
		≥3 units: 9.6 (175)	2 units: 11.4 (209)
			≥3 units: 11.2 (204)
Moderate-to-severe or severe asthma
Broder et al., 2011 (87) (U.S.)	2003: 302	OCS	—
	2004: 970	2007 cohort: 33.2 (459/1,382)††
	2005: 1,301
	2006: 1,361
	2007: 1,382
Bruno et al., 2014 (88) (France/Italy)	102	OCS	—
		64.7 (66/102)
DiSantostefano and Davis, 2011 (142) (UK)	1,233	SCS	SCS
		1–6 mo: 11%	7–12 mo: 9%
		7–12 mo: 10%	1–6 mo: 19%
Eisner et al., 2012 (89) (U.S.)	2,878	—	OCS
			51.2 (1,473/2,878)
Lafeuille et al., 2013 (90) (U.S.)	3,044	—	OCS
			49 (1,479/3,044)
Lafeuille et al., 2012 (78) (U.S.)	644	—	93.5 (602/644)
			(OCS)
Sposato et al., 2017 (143) (Italy)	340	OCS	—
	OMB Tx duration	≤12 mo: 13%
	≤12 mo: 39	12–≤24 mo: 9%
	12–≤24 mo: 94	24–≤60 mo: 6%
	24–≤60 mo: 171	>60 mo: 3%‡‡
	>60 mo: 36	‡‡P = 0.044 vs. ≤12 mo
Sullivan et al., 2015 (91) (U.S.)	25,297	OCS	OCS
	HDICS: 11,445	HDICS: 34%	HDICS: 35%
	HICS: 6,926	HICS: 65%	HICS: 53%
	OMB: 856	OMB: 52%	OMB: 63%
Sweeney et al., 2014 (144) (UK)	2,670 pts with new ICS/LABA Rx and no prior ICS Rx	—	OCS
			5 (132/2,670)
Turner et al., 2017 (145) (UK)	2,660	OCS	—
		FDC ICS/LABA: 6.5%
		ICS+LABA:
		8.8% P = 0.084

Definition of abbreviations: CI = confidence interval; DMP = disease management program; FDC = fixed-dose combination; FP = fluticasone propionate; FSC = FP/SAL fixed-dose combination inhaler; GINA = Global Initiative for Asthma; HDICS = high-dosage ICS; HICS = high-intensity corticosteroids; ICS = inhaled corticosteroids; LABA = long-acting β₂-agonists; LTRA = leukotriene receptor antagonists; MF = mometasone furoate; OCS = oral corticosteroids; OMB = omalizumab; Rx = prescription; SABA = short-acting β₂-agonists; SCS = systemic corticosteroids; Tx = treatment.

^*Groups were defined according to the value of the ICS to total asthma medication ratio in 2008: R = 0% (non-ICS users), 0% < R < 50% (low ICS ratio group), and R ≥ 50% (high ICS ratio group). The ratio constituted the proportion of prescribed units of ICS out of the overall number of respiratory medication units prescribed during 2008.

^†Patients with asthma and a diagnosis of depression during the study period (cases) were matched to patients with asthma without depression (control subjects) according to the date of asthma diagnosis.

^‡Occasional, intermittent, or continuous ICS/LABA use in baseline period.

^§Patients in the DMP group were propensity matched with non-DMP control subjects based on a range of variables, including demographics, asthma care/therapy, and comorbidities.

^||Escalation group: ICS or ICS-containing therapy dosage increase; a switch between ICS, LABA, or LTRA, or add-on of another controller within 12 months after the index date. Unchanged group: patients with ≥1 additional fill indicating continuation of index treatment regimen within 12 months after the index date.

^¶Patients receiving FSC who stepped down to FSC at a smaller dosage of FP or switched to FP only at the same dosage. Patients in the FSC group were matched to those in the FP group through propensity score matching.

^**Each patient in the FSC group was matched to one patient in the MF group through propensity score techniques.

^††Based on OCS use in the top 10 medication patterns.

^‡‡Pharmacist intervention to reduce SABA over-dispensing.

One study reported OCS use by 66% of patients over a 2-year period (U.S. study of adult patients) (49).

Similar incidences of general OCS use (range 18.3–47.2%) were observed in four of five studies involving patients receiving GINA step 2 or greater treatment across all age categories (Table 3) (4, 84–86). The exception was a UK-based retrospective cohort study (adolescent/adult patients) in which general OCS use was reported by 5.7%, 6.4%, and 7.4% of patients in the year before they initiated ICS, LABA, or ICS/LABA treatment, respectively (77).

General OCS use in studies involving patients with moderate or severe asthma was generally greater than that reported in studies involving a broader asthma population, although considerable interstudy variation existed (Table 3). OCS use ranged from 33.2% to 65% in five studies of patients with moderate-to-severe or severe asthma (6- to 12-mo assessment period) (87–91). A small cross-sectional study conducted in Italy and France reported general OCS use by 64.7% of patients with severe asthma in the previous year (88).

Clinical characteristics of patients prescribed CS

Overall, OCS users tended to be older than nonusers (Table E7) (30, 92–94). A U.S.-based prospective cohort study reported that for patients with severe asthma, long-term OCS use was more common among adults than among children/adolescents (22.4% and 9.9%, respectively, had at least 3 mo of OCS use in the previous year) (25). There was a greater percentage of female patients among OCS users in three U.S.-based studies reporting on use for patients with asthma overall, persistent asthma, and moderate-to-severe asthma (Table E7) (92–94). The percentage of patients receiving long-term OCS was greatest in the subgroups of patients who reported the greatest number of exacerbations (42, 46, 58, 62, 95). Disease severity and relative resistance to standard therapy (high-dosage ICS, LABA, and other controllers) are among the potential explanations for this finding. In addition, studies suggested that general and short-term OCS use was greater among obese patients than among nonobese patients. However, this may reflect the relatively poorer asthma control in this patient population (80, 88). A U.S.-based study reported that compared with nonusers or patients receiving short-term OCS therapy, adult patients who had uncontrolled asthma and were receiving long-term OCS were more likely to have received specialist asthma care (18.9% vs. 52.8%, respectively) and to have experienced more asthma exacerbations (rate: 0.3 vs. 1.6, respectively) (94). In all studies that reported on disease severity, for both long-term and short-term use, there was a strong correlation between increasing OCS use and increasing disease severity.

Clinical Burden of SCS

Comorbidities and complications associated with OCS/SCS use for patients with asthma were reported in 17 studies. Eleven of those studies reported OCS use (60, 72, 82, 93, 94, 96–101) and six reported SCS use (102–107).

Studies reported that the use of both short- and long-term OCS was consistently associated with a greater risk of acute and chronic CS-related complications compared with no OCS use. The risk of complications increased with increasing exposure (Table E8). A study that assessed the risk of any CS-related AEs reported a 1.3-fold increase in risk with long-term OCS use compared with no OCS use (108).

Acute complications

Identified studies reported acute OCS-associated complications that included infections and gastrointestinal events (Table E8). In a U.S.-based matched cohort study, approximately three times as many OCS users (≥30 d/yr) had pneumonia (28.4%) and opportunistic infections (1.5%) compared with non-OCS users (10.9% and 0.4%, respectively) (93). Four of five studies that reported gastrointestinal complications with OCS use found an increased risk associated with long-term use versus no use (odds ratio [OR], 1.035 [ulcers/bleeds] to 2.89 [gastroesophageal reflux]) (93, 94, 96, 97, 100).

Chronic complications

Identified studies reported chronic OCS-associated complications that included metabolic, bone-related, and cardiovascular events (Table E9). Most studies (U.S. and UK) reported increased risks of comorbid diabetes and obesity for patients with severe asthma and greater/long-term OCS/SCS use compared with patients with milder disease and less/no OCS use (93, 96, 100). However, one U.S.-based retrospective cohort study of adults with persistent asthma reported a similar prevalence of diabetes between patients with long- and short-term OCS use (94).

Bone- and muscle-related complications

Four studies reported an increased risk of bone- and muscle-related complications with long-term OCS use versus no use (Table E8) (93, 94, 96, 100). For example, one study reported that current OCS users with asthma (≥4 prescriptions/yr) had ORs of 1.44 and 1.21 for osteoporosis and bone fracture, respectively, compared with nonusers (100).

Cardiovascular complications

Identified studies consistently reported an association between long-term OCS use and increased risk of cardiovascular complications, hypertension, and hypercholesterolemia compared with no use or short-term use (Table E9) (60, 82, 93, 94, 96, 100). For example, in a UK-based matched cohort study of adults and adolescents, comorbid hypertension was more common with long-term OCS use among patients with severe asthma (34%) than with less OCS use among patients with mild-to-moderate asthma (29%) or without asthma (25%) (96). In a U.S.-based cross-sectional study of patients with asthma, long-term OCS use was significantly associated with systemic hypertension using univariate analysis. The association was no longer significant after adjustment for demographic and asthma-related variables and presence of obstructive sleep apnea (60). Importantly, in a U.S.-based retrospective cohort study, long-term OCS use was also associated with risk of coronary heart disease (hazard ratio, 2.59) and heart failure (hazard ratio, 3.48) (82).

Psychiatric complications

Identified studies reported mixed results for the association between OCS use and psychiatric complications (Table E9). A UK-based matched cohort study found that greater OCS use among patients with severe asthma was associated with a significant increase in the prevalence of comorbid psychiatric conditions compared with less OCS use (38% and 31%, respectively, vs. 25% for non–OCS use) (96). In a U.S.-based retrospective study, adults with persistent asthma reported an increased prevalence of anxiety (16.2% vs. 10.6%), but not depression (13.0% vs. 12.3%), with longer OCS use compared with short-term use (94). A database study of patients in the Netherlands with difficult-to-treat asthma reported an association between long-term OCS use and anxiety or depression (OR, 1.38) (97).

Ocular complications

Long-term OCS use was associated with an increased risk of cataracts, regardless of whether the risk was compared with short-term use or no exposure (Table E9) (93, 94, 96, 100). The OR for risk of cataracts in current versus nonusers of OCS was reported as 1.26 in one U.S.-based study (93). In a UK-based matched cohort study, cataracts were more common with long-term OCS use among patients with severe asthma (9%) than among those with less OCS use and mild-to-moderate asthma (5%) or without asthma (4%) (96). No studies reported an association between OCS use and increased risk of glaucoma.

Other complications

Included studies provided limited data regarding the effects of OCS therapy on the risk of other complications. Individual studies reported associations between long-term OCS use and asthma-related bronchiectasis (72), chronic kidney disease (96), and sleep disorders (96).

Healthcare Resource Use and Economic Impact of SCS Use

The impact of OCS/SCS use on healthcare resource use and costs for patients with asthma was reported in 12 studies (Tables E10 and E11). Across eight studies that evaluated the economic burden of OCS use for patients with asthma, long-term OCS and SCS use was consistently associated with increased healthcare costs compared with no or short-term use (34, 49, 92, 96, 103, 105, 106, 109). However, short-term use was also associated with increased costs and healthcare resource use compared with no OCS use.

Healthcare costs were reported to increase with greater OCS exposure. Data from a UK database estimated that the mean annual total costs of high, low, and no OCS use were £2,603, £978, and £560, respectively (96). Another UK-based database analysis reported that nonasthma-related medication costs were 58% greater for patients receiving long-term OCS than for nonusers (costs included prescriptions for proton pump inhibitors and bisphosphonates, treatments used to manage OCS-related AEs) (34). Nonmedication costs were 19% greater and total healthcare expenditures were 43% greater (34).

A U.S. claims-based analysis of adult patients with asthma receiving high-dosage OCS therapy reported that annual total healthcare costs were significantly greater for patients with OCS-related complications ($25,168) than for those without such complications ($21,882); however, asthma-related costs were comparable between the groups ($4,213 and $3,952, respectively) (49).

This association between healthcare costs and OCS exposure might reflect OCS use itself or disease severity. Patients with more severe asthma are likely to have had greater OCS exposure because of the need for treatment to control their symptoms. They are also likely to have incurred additional healthcare costs associated with increased clinic visits and hospitalizations for exacerbations. Currently, the literature does not appear to address indirect costs, such as lost productivity, associated with long-term OCS use. A link between treatment-related AEs and increased healthcare costs and resource use was identified in five articles that assessed the comorbidities and AEs associated with long-term OCS use (94, 96, 103, 105, 106). Furthermore, evaluation of data from the UK general practice database found that relative to patients with no OCS exposure, the additional costs for nonasthma-related medication use were notably greater for patients with greater OCS exposure (£772) than for those with less OCS exposure (£112) (96).

Discussion

This systematic literature review identified a large body of evidence from observational studies documenting real-world clinical practice regarding the use of OCS/SCS therapy for the general asthma population and especially for patients with more severe disease. Identified studies reported that short- and long-term OCS/SCS therapy is widely used to treat patients with asthma, particularly those with severe disease. Overall, OCS use was typically reported in approximately half of patients in the general asthma population over a 1-year period, and short-term use was reported in up to 36% of patients (35–38, 110). Guidelines recommend OCS for short-term treatment of serious exacerbations or as add-on maintenance therapy for patients with severe disease that is not controlled by high-dosage maintenance treatment (5). These recommendations restrict OCS use to an estimated 10% of patients with severe disease (111) and approximately 10% who experience an exacerbation in any given year (112). Thus, our findings indicate that in many instances, >10% of patients are receiving OCS treatment, suggesting that OCS may be overused for asthma management.

Long-term OCS use was consistently infrequent across all age groups for patients with nonsevere disease (<2%). However, up to 60% of patients with severe or uncontrolled asthma were reported to have received long-term OCS therapy. The extent of long-term OCS/SCS use has not been reported in other analyses. Furthermore, identified studies reported that OCS are used as long-term therapy at dosages up to 22 mg/d (67, 68), which is greater than the prednisone-equivalent dosage (≤7.5 mg/d) recommended in GINA guidelines as add-on therapy for patients with asthma that is not controlled by high-dosage therapy (GINA step 4 treatment) (5). These results suggest that patients are repeatedly treated with OCS at increasingly greater dosages, possibly despite their lack of response to OCS therapy or because of the nontargeted action of OCS. This may be because OCS are considered generally effective for all patients with asthma and are frequently prescribed in the absence of defined disease markers and objective response monitoring, such as lung function.

Notably, the identified studies predominantly reported on patients from Europe and North America, and therefore these findings may be biased toward approaches in these regions. It is possible that OCS are used more widely in low-income countries than in high-income ones because they are inexpensive. Indeed, until 2010, the World Health Organization recommended OCS as an essential medication to treat asthma (113). Moreover, payer criteria for biologics may impact OCS prescribing. For example, several reports from the United Kingdom (where four or more courses of OCS are required for eligibility for biologics) indicate that more than half of patients with severe asthma were prescribed four or more courses of OCS in 1 year. However, the percentages are lower for countries where there are no such requirements or fewer courses are required. In Australia, the payer requirement is two or more courses, and exemptions are available for OCS toxicity.

Studies have found that OCS are not equally effective for all patients and that most patients with severe asthma are steroid resistant to varying extents, by definition of their general lack of response to ICS (24, 25). Clinical characteristics of asthma, including eosinophilic airway inflammation, have been identified as being associated with response to OCS (1). Genetic factors have been identified that could also account for some of this patient variability in OCS response (24, 114–118). For example, genetic polymorphisms in the glucocorticoid-induced transcript gene GLCCI1 have been reported to account for 6.6% of the overall variability in clinical responses to ICS (117, 118). However, responses to both ICS and OCS are likely to be affected by several genetic variations, some of which have yet to be identified (24).

Identified studies reported that the use of both short- and long-term OCS/SCS is associated with an increased risk of acute and chronic complications, and this risk increases with greater exposure. The risk of any steroid-related AE was found to be up to 3.6-fold greater with long-term OCS use than with no use (103, 106, 108). Patients will become aware of any acute complications, if they occur, after treatment initiation. These complications can often have serious short-term consequences. However, patients may be less aware of the chronic complications associated with OCS/SCS use. In contrast, OCS/SCS-associated chronic conditions are often anticipated and monitored by physicians. Conversely, acute conditions, namely infections and gastrointestinal events, are frequently not considered by prescribers, although many OCS users experience acute complications, and relatively fewer studies have reported on these types of AEs. For example, pneumonia was the most frequently identified comorbidity associated with severe asthma in the Severe Asthma Research Program cohort of U.S. and UK patients with severe asthma (119). Even patients with low OCS exposure were reported to have an increased risk of infection (103). These findings are consistent with recent publications that were not included in our review (2, 120). One publication reported that AEs for patients with asthma receiving SCS began at cumulative exposures of 1.0 to <2.5 g (vs. >0 to <0.5 g reference), suggesting a relationship between cumulative SCS exposure and risk of AEs (2). Another publication reported that the 15-year cumulative incidence of type 2 diabetes was 9.5% for OCS/SCS users versus 5.6% for nonusers (120). However, the risk of type 2 diabetes began with a cumulative exposure of 0.5 to <1 g, which is equivalent to four lifetime OCS/SCS courses. This suggests that the incidence of comorbid type 2 diabetes is not necessarily exclusively influenced by the cumulative dosage of OCS (2). However, cumulative OCS/SCS exposure may not be an ideal measure because of possible variations among patients with regard to factors such as disease duration and severity. It is potentially an important metric for the medical community to understand and adopt in a clinical setting because it provides a means of assessing long-term exposure and associated adverse effects of OCS/SCS use. As with the response to OCS/SCS treatment, the risk of any patient experiencing an AE is likely to be influenced by underlying genetic factors, many of which have yet to be identified (24, 121).

The burden of comorbidities associated with both long-term and repeated short-term OCS/SCS use adds to the asthma burden, leading to increased risks of hospitalizations and emergency department visits, and corresponding increased healthcare resource use. We found reports of 43% greater overall healthcare expenditures for patients receiving long-term OCS therapy compared with nonusers, and 58% greater nonasthma-related costs (including costs of treatments used to manage OCS-related AEs) (108). Thus, although a prescriber’s decision to use OCS/SCS therapy instead of targeted treatments can be influenced by the initial unit price, incurred healthcare costs associated with OCS/SCS use may bring the validity of this choice into question. However, not all of the increased costs for patients receiving OCS are potentially directly attributable to OCS-related AEs: the direct costs arising from the management of severe asthma, such as treatment and hospitalization costs, are also likely to contribute to this increase. In an analysis of patients with asthma who were receiving intermittent or long-term SCS matched with patients who were not, 42% greater overall costs were reported for those who were receiving SCS (120). The associated average annual costs for adverse outcomes and asthma were £1,483 and £403, respectively, for patients receiving SCS, compared with £1,165 and £166, respectively, for nonusers. The individual contribution of OCS sparing to patients’ health-related quality of life is also difficult to elucidate. In future studies, the use of a global approach that includes generic and disease-related patient health-related quality of life questionnaires could help investigators assess the effects of OCS reduction and distinguish between improvements that result from reduced OCS use and those that are attributable to improved asthma control. It is also difficult to directly attribute the occurrence of some common comorbidities in the general population to OCS use. For example, gastroesophageal reflux disease is particularly common among patients with asthma. This condition is worsened by OCS use but can be improved by greater asthma control and increased patient activity (122).

Our review has several advantages over similar publications. This review included a search of both MEDLINE and Embase databases. We also used a strong methodology, with a study design that allowed the inclusion of a wide range of ages (children, adolescents, and adults), the full spectrum of disease, and the prevalence of both short- and long-term use. This resulted in the identification and reporting of results from 139 studies that included populations with varying degrees of asthma severity. Our findings are robust compared with recent reviews that identified fewer publications (<50 studies) and reported only on long-term OCS use among adult patients with severe asthma or AEs and their corresponding economic burden (28, 123, 124).

The limitations of our review include the fact that our search was conducted in December 2017, and congress abstracts were not included. Therefore, some more recent studies may have been omitted from our review. We also used the definitions of long-term and short-term use provided in the identified literature, but interpretations varied among the studies. Another potential limitation is our inclusion of 13 studies that reported results of patient surveys, as the data from such surveys are considered to be weaker than those obtained from cohort or database studies. Furthermore, some studies from which data were obtained were not designed to directly identify AEs associated with OCS/SCS use. Therefore, AEs may have been underreported in those studies. In addition, data on dose-dependent effects and the frequency of OCS/SCS use associated with AEs were not available for all studies and are not summarized here. The information provided on asthma severity varied among the identified studies, which prevented us from obtaining a more generalized synthesis of the data by GINA treatment step. Overall, a considerable amount of available data was obtained from retrospective reports from patients and physicians, and as such, was subject to recall bias. These data suggest that use remains a frequent component of treatment for patients with asthma, primarily related to the lack of control with other medications. They also suggest that OCS/SCS use is widespread and associated with significant adverse effects. Quantitation of these data would be helpful, but this was beyond the scope of this review. Future longitudinal studies using objective methods to collect data regarding the use of OCS and its associated clinical burden are needed.

Overall, this review demonstrates that OCS and SCS, including long-term OCS, continue to be commonly used and overused for the management of asthma across the disease spectrum and particularly for severe asthma. This use is associated with both acute and chronic complications. Importantly, patients receiving repeated short-term, high OCS dosages may incur a greater risk of AEs than those receiving long-term, low dosages, as the risk of AEs increases with the cumulative OCS dosage. The introduction of biologics has led to a reduction in OCS use in other disease areas, but our review shows that a similar change has not yet occurred for asthma (18). This is because biologics have only recently been approved for the treatment of severe asthma. Omalizumab was the only add-on biologic therapy available for the treatment of severe asthma during the time periods covered in many of the identified studies (18). The recent approval of further biologic treatments that reduce asthma symptoms and exacerbation risks and allow OCS tapering for OCS-dependent patients has the potential to reduce future OCS use for patients with asthma (10–14, 16, 17).