Summary of findings
This protocol is the first of a series of three review protocols exploring the safety profile of inhaled corticosteroids (ICS) in terms of growth in children with persistent asthma. The present review explored the dose-response effect of ICS on growth. The second review compares the long-term effects of ICS on growth (Zhang 2011), and the third examines the effects of different drugs and delivery devices on growth. For more comprehensive background data and additional references, see Zhang 2011.
Description of the condition
Asthma is defined as a chronic inflammatory disorder of the airways in which many cells and cellular elements play a role. The chronic inflammation is associated with airway hyperresponsiveness that leads to recurrent episodes of wheezing, breathlessness, chest tightness and coughing, particularly at night or in the early morning. These episodes are usually associated with widespread, but variable, airflow obstruction within the lung that is often reversible either spontaneously or with treatment (GINA 2014). In developed countries, the prevalence of childhood asthma has markedly increased over the past few decades (ISAAC 1998; Masoli 2004; Asher 2010); however, this increase has recently reached a plateau in some of these countries (Lai 2009; Asher 2010). In contrast, asthma prevalence is sharply increasing in developing countries (Africa, Central and South America, Asia and the Pacific region), probably as a result of rapid and ongoing urbanisation and westernisation (Braman 2006; Asher 2010). The global burden of childhood asthma is continuing to rise.
Description of the intervention
ICS are widely considered the first-line treatment for persistent asthma, both in adults and in children (NHLBI 2007; BTS 2012; GINA 2014; Chauhan 2012; Lougheed 2012). Studies have demonstrated the clinical benefits of ICS in controlling asthma symptoms, reducing exacerbations and hospitalisations, decreasing airway hyperresponsiveness and airway inflammation, improving pulmonary function, improving quality of life and reducing asthma-related deaths (Juniper 1990; Van Essen-Zandvliet 1992; Olivieri 1997; Van Rensen 1999; Suissa 2000; Covar 2003; Adams 2011a; Adams 2011b; Adams 2011c). Seven ICS are currently available for clinical use worldwide: beclomethasone dipropionate, budesonide, fluticasone propionate, mometasone fumarate, ciclesonide, flunisolide and triamcinolone acetate. Each inhaled corticosteroid has different pharmacokinetic and pharmacodynamic properties and biologic characteristics; however, all ICS can achieve similar therapeutic benefits when given at equipotent doses (Sobande 2008; BTS 2012; GINA 2014; Lougheed 2012).
The optimal doses of ICS for persistent childhood asthma remain unclear. The most recent asthma guidelines recommend initiating ICS at low or medium daily doses for children with mild to moderate persistent asthma; however, patients with more severe asthma and those with poor response to low to moderate doses of ICS may require higher doses (≥ 400 μg/d of hydrofluoroalkane (HFA)-beclomethasone or equivalent) to achieve satisfactory control of asthma (NHLBI 2007; BTS 2012; GINA 2014; Lougheed 2012).
Although ICS are generally considered safe treatment for children with asthma, the potential systemic adverse effects related to long-term use of these drugs have been, and continue to be, a matter of concern, especially the effects on growth (Pedersen 2001; Allen 2002). In 1998, based on a report of the panel of experts, the US Food and Drug Administration (FDA) required labels on all ICS warning of a potential reduction in growth in children (FDA 1998). Since that time, the relationship between ICS and growth impairment in children with asthma has been extensively debated in the literature and more so with the advent of new molecules with allegedly safer profiles (Witzmann 2000; Brand 2001; Creese 2001; Wolthers 2001; Carlsen 2002; Price 2002a; Sizonenko 2002; Salvatoni 2003; Allen 2006).
How the intervention might work
ICS are the most potent anti-inflammatory drugs available for long-term treatment of persistent asthma. Possible molecular mechanisms for the anti-inflammatory effects of ICS and for corticosteroid-induced growth impairment have been reviewed previously (Barnes 2003; Zhang 2011).
Why it is important to do this review
One Cochrane systematic review (Sharek 2000a) produced solid evidence supporting growth suppression estimated at 1.5 cm per year over seven to 12 months for 400 μg/d inhaled chlorofluorocarbon (CFC)-propelled beclomethasone (equivalent to 200 μg/d of HFA-propelled beclomethasone) in children with asthma. This review lately has been converted to a journal article (Sharek 2000b). However, it remains unclear whether corticosteroid-induced growth retardation is dose dependent. We therefore decided to conduct this systematic review to evaluate the relationship between dose of ICS and risk of growth impairment in children with persistent asthma.
To assess whether increasing the dose of ICS is associated with slower linear growth, weight gain and skeletal maturation in children with asthma.
Criteria for considering studies for this review
Types of studies
Parallel-group randomised controlled trials.
Types of participants
Children one to 17 years of age with the diagnosis of persistent asthma.
Types of interventions
Each treatment group should be given the same ICS at two or more different doses via the same delivery system for at least three months. ICS may be administered as monotherapy or in combination with other non-steroidal asthma drugs (e.g. long-acting beta-agonists (LABAs), leukotriene receptor antagonists (LTRAs)). In all included trials, the intervention group depicted is the lower-dose ICS and the control (comparison) group is the higher-dose ICS.
Types of outcome measures
Linear growth velocity (cm/y), obtained by measuring height at a number of time points during the study and performing linear regression of height over time (Price 2002a).
- Change in growth velocity standard deviation (SD), defined as the difference between an individual's growth velocity and predicted growth velocity divided by the predicted growth velocity SD for individuals of the same age and sex (and ethnicity if available) (Pedersen 2001).
- Change in absolute height (cm) over time.
- Change in weight (kg or z-score) over time.
- Change in body mass index (added post hoc).
- Change in skeletal maturation (added post hoc).
We did not intend to include lower leg length measured by knemometry as the outcome because this measurement correlates poorly with statural height and tends to overestimate potential effects of ICS on growth (Efthimiou 1998; Allen 1999).
Search methods for identification of studies
We identified trials from the Cochrane Airways Group Specialised Register of Trials (CAGR), which were derived through systematic searches of bibliographic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, CINAHL, AMED and PsycINFO, and through handsearching of respiratory journals and meeting abstracts (see Appendix 1 for further details). All records in the CAGR coded as 'asthma' were searched using the following terms.
(((steroid* or corticosteroid* or glucocorticoid* ) and inhal*) or budesonide or Pulmicort or fluticasone or Flixotide or Flovent or ciclesonide or Alvesco or triamcinolone or Kenalog or beclomethasone or beclometasone or Becotide or Becloforte or Becodisk or QVAR or Flunisolide or AeroBid or mometasone or Asmanex or Symbicort or Advair or Inuvair) AND (grow* or height* or SDS) AND (child* or paediat* or pediat* or adolesc* or teen* or prepubertal* or pre-pubertal* or puberty or pubertal* or infan* or toddler* or bab* or young*) AND (dose* or dosage* or delivery* or administ* or response* or high* or low*)
We also conducted a search of the ClinicalTrials.gov website. All databases were searched from their inception until March 2014 with no restriction on language of publication.
Searching other resources
We checked the reference lists of all primary studies and review articles for additional references. We also searched manufacturers' clinical trial databases for potentially relevant unpublished studies, if needed.
Data collection and analysis
Selection of studies
Two review authors (AP and LZ or SP) independently assessed the titles and abstracts of all potential studies for inclusion identified by the search strategy. Full-text articles were retrieved when they appeared to meet the inclusion criteria or when data in the title and abstract were insufficient to permit a clear decision regarding their inclusion. We resolved disagreements through discussion, or, if required, we consulted the third review author.
Data extraction and management
Two review authors (AP and BC) independently extracted data from the included trials using specially designed and pilot-tested data extraction forms. For trials with multiple reports, we extracted data from each report separately and combined information across multiple data collection forms afterwards. We resolved disagreements by discussion and entered the extracted data into RevMan version 5.1 (Review Manager 5).
We extracted the following data.
- Study characteristics: year of publication, name of the first author, setting and source of funding/sponsorship.
- Methods: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, completeness of outcome data, selective reporting and other sources of bias.
- Participants: sample size, demographics, inclusion and exclusion criteria.
- Intervention: type of ICS, dosage, frequency of administration, inhalation device, treatment duration and adherence to treatment, if available.
- Comparator: the same corticosteroid given at different dosage regimens (the same details as for intervention).
- Co-interventions: type, dosage regimen and duration.
- Results: mean value of the outcome measures in each group, SD or other metrics for uncertainty (standard errors (SEs), confidence intervals (CIs), P values for differences in means) of outcome measurements in each group, number of participants who underwent randomisation, number of participants on whom outcomes were measured in each group.
Assessment of risk of bias in included studies
Two review authors independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). Disagreements were resolved by discussion or by involving the third review author. We assessed the risk of bias according to the following domains.
- Allocation sequence generation.
- Concealment of allocation.
- Blinding of participants and investigators.
- Incomplete outcome data.
- Selective outcome reporting.
- Other risk of bias.
We noted other sources of bias. We graded each potential source of bias as low, high or unclear risk. Studies were deemed to be of high methodological quality if information on randomisation generation, blinding and incomplete outcome data was available, indicating a low risk of bias.
Measures of treatment effect
Measurements of growth were continuous outcomes, so we used mean difference (MD) and 95% CI as the metrics for treatment effects, as appropriate.
Unit of analysis issues
We considered each individual comparison as the unit of analysis. We used analysed participants as sample size rather than the number of participants randomly assigned in the included studies. We had planned three pair-wise comparisons of ICS doses in HFA-beclomethasone or equivalent: low (≤ 200 μg) versus medium (201 to 400 μg) versus high dose (> 400 μg) and low (≤ 200 μg) versus high (> 400 μg) dose (Lougheed 2012). The ICS dose equivalence used for this review was based on Canadian Asthma Guidelines (Lougheed 2012), which are based on a combination of the dose equivalency mentioned in GINA 2014 and reported safety and efficacy data: 1 μg fluticasone = 1 μg mometasone = 1 μg ciclesonide = 1 μg of hydrofluoroalkane HFA-beclomethasone = 2 μg budesonide = 2 μg CFC-BDP = 4 μg flunisolide = 4 μg triamcinolone acetate.
Dealing with missing data
We contacted investigators or study sponsors to verify key study characteristics and to obtain missing numerical outcome data when possible.
Assessment of heterogeneity
We used the I
Assessment of reporting biases
We planned to contact study authors to ask them to provide missing outcome data if we suspected reporting bias. When this was not possible, and when the missing data were thought to introduce serious bias, we planned to explore the impact of excluding such studies on the overall assessment of results by performing a sensitivity analysis.
We performed the meta-analyses using the Cochrane statistical package RevMan 5 (Review Manager 5). We used the fixed-effect model unless statistical heterogeneity was found, in which case we used the random-effects model.
Subgroup analysis and investigation of heterogeneity
We planned to carry out the following subgroup analyses for the primary outcome, measured at various points in time.
- Participant age: preschoolers (two to five years), prepubertal children (> five to 12 years), adolescents (> 12 to 18 years).
- Asthma severity: mild versus moderate versus severe.
- ICS molecule: beclomethasone, budesonide, fluticasone, mometasone, ciclesonide, flunisolide, triamcinolone.
- Concomitant use of non-steroidal antiasthmatic drugs: ICS alone, ICS combined with non-steroidal drugs, such as LABAs and LTRAs.
- Dose difference of ICS in HFA-beclomethasone or equivalent (added as post hoc analysis).
Sensitivity analysis was used to assess the potential impact of particular decisions or missing information on the findings of the review (Higgins 2008). We planned to carry out the following sensitivity analyses with regards to primary outcome by excluding from the analysis trials with the following.
- High risk of bias owing to missing data or unbinding, or both.
- Rate of adherence to ICS lower than 75% or lack of available data regarding adherence to treatment.
- Pharmaceutical industry sponsorship.
Description of studies
Results of the search
The literature search conducted until March 2014 identified a total of 406 citations and abstracts (Figure 1). Of these, 71 potential full texts were reviewed thoroughly for inclusion criteria. Twenty-two trials, including 34 comparisons (Characteristics of included studies), were eligible for inclusion. Of these, 12 trials (17 comparisons) contributed no usable data to this review; four trials (five comparisons) either presented data in a different format than was specified in the protocol or reported incomplete data (Jonasson 2000; Chen 2001; Teper 2004; Gelfand 2006; Gelfand 2006 b); seven trials (11 comparisons) did not measure children's growth as an outcome (Jonasson 1998; Giorgi 1998; Peden 1998; Peden 1998 b; Baker 1999; Baker 1999 b; Kemp 1999; Kemp 1999 b; Doniec 2004; Kerwin 2008; Kerwin 2008 b) and one trial was published as an abstract (Lemanske 2004). Consequently, 10 trials (17 comparisons) published as full text contributed at least one outcome to the meta-analysis.
|Figure 1. Flow diagram of screening of trials.|
Ten trials, reporting 17 comparisons (Allen 1998; Shapiro 1998; Shapiro 1998 b; Shapiro 1998 c; Shapiro 1998 d; Verberne 1998; Verberne 1998 b; Wasserman 2006; Sorkness 2007; Skoner 2008; Pedersen 2010; Pedersen 2010 b; Vaessen-Verberne 2010; Brand 2011; Brand 2011 b; Skoner 2011; Skoner 2011 b) and enrolling 3394 children with confirmed persistent asthma, contributed data to the review. The following information pertains only to the 17 comparisons (from 10 included trials) contributing data to this review (Characteristics of included studies). The FDA has produced a guideline on evaluation of the effects of orally inhaled and intranasal corticosteroids, specific to placebo-controlled trials in children (US FDA 2007); although some criteria were not relevant for dose-response studies, we ascertained the compliance status to these guidelines of trials that contributed data to the meta-analysis ( Table 1; Table 2; Table 3).
All trials used a parallel-group design.
Three comparisons involved children two to five years of age (Wasserman 2006; Brand 2011; Brand 2011 b), six comparisons involved prepubertal children, five to 12 years of age (Allen 1998; Skoner 2008; Pedersen 2010; Pedersen 2010 b; Skoner 2011; Skoner 2011 b), and eight comparisons involved prepubertal and pubertal children (Shapiro 1998; Shapiro 1998 b; Shapiro 1998 c; Shapiro 1998 d; Verberne 1998; Verberne 1998 b; Sorkness 2007; Vaessen-Verberne 2010). Most trials described a gender ratio hovering around 65% male participants. With regards to asthma severity, one comparison (Skoner 2008) focused on asthmatic individuals with mild airway obstruction, two comparisons (Verberne 1998; Verberne 1998 b) focused on asthmatic individuals with mild to moderate airway obstruction, four comparisons (Shapiro 1998; Shapiro 1998 b; Shapiro 1998 c; Shapiro 1998 d) focused on asthmatic individuals with moderate to severe airway obstruction and the remaining six comparisons (Allen 1998; Wasserman 2006; Pedersen 2010; Pedersen 2010 b; Skoner 2011; Skoner 2011 b) failed to report the severity of baseline airway obstruction. Two comparisons (Brand 2011; Brand 2011 b) pertained to preschool children with recurrent wheezing and a positive asthma predictive index or a positive screening test for atopy. Asthma triggers were seldom reported.
The duration of intervention varied from 12 weeks (seven comparisons; Shapiro 1998; Shapiro 1998 b; Shapiro 1998 c; Shapiro 1998 d; Wasserman 2006; Pedersen 2010; Pedersen 2010 b) to 24 weeks (two comparisons; Brand 2011; Brand 2011 b) to 26 weeks (one comparison; Vaessen-Verberne 2010) to 52 weeks (seven comparisons; Allen 1998; Verberne 1998; Verberne 1998 b; Sorkness 2007; Skoner 2008; Skoner 2011; Skoner 2011 b).
The ICS molecule used was beclomethasone dipropionate (BDP) (two comparisons; Verberne 1998; Verberne 1998 b), budesonide (BUD) (four comparisons; Shapiro 1998; Shapiro 1998 b; Shapiro 1998 c; Shapiro 1998 d), ciclesonide (CIC) (five comparisons; Skoner 2008; Pedersen 2010; Pedersen 2010 b; Brand 2011; Brand 2011 b), fluticasone propionate (FP) (four comparisons; Allen 1998; Wasserman 2006; Sorkness 2007; Vaessen-Verberne 2010) or mometasone fumarate (MF) (two comparisons; Skoner 2011; Skoner 2011 b). The difference in the dose of ICS between two comparison groups (reported in HFA-beclomethasone equivalent) varied by ≤ 150 μg in most trials. Most compared 100 μg (low dose) versus 200 μg (the cutoff limit between low and medium doses of ICS); in only four comparisons (Shapiro 1998 b; Shapiro 1998 d; Verberne 1998; Vaessen-Verberne 2010 ) was the difference in the dose of ICS between groups ≥ 400 μg. Different devices were used, including aerochamber, diskhaler, dry powder inhaler, metered-dose inhaler with or without spacer, nebuliser and turbohaler (further details are available in the Characteristics of included studies table). Yet all trials used the same inhalation device in within-trial group comparisons. Adherence rate to ICS was reported by three of 10 trials; when reported, adherence was at or above 80%. All trials but one (Sorkness 2007) were funded by the pharmaceutical industry.
Three comparisons (Verberne 1998; Pedersen 2010; Pedersen 2010 b) enrolled only participants receiving ICS as monotherapy. Eleven comparisons (Allen 1998; Shapiro 1998; Shapiro 1998 b; Shapiro 1998 c; Shapiro 1998 d; Wasserman 2006; Skoner 2008; Brand 2011; Brand 2011 b; Skoner 2011; Skoner 2011 b) reported accepting participants who were using co-interventions with additional antiasthmatic drugs such as LABAs, antileukotrienes or theophylline. Three comparisons (Verberne 1998 b; Sorkness 2007; Vaessen-Verberne 2010) specifically compared ICS alone versus ICS + LABA, without other co-interventions.
The primary outcome was linear growth velocity (zero to 12 months), which was documented in four comparisons involving prepubescent children (Allen 1998; Skoner 2008; Skoner 2011; Skoner 2011 b); in all cases, linear growth was analysed in three or more height measurements by regression analysis, with adjustment for co-variates in all but one trial (Allen 1998). Secondary outcomes included change in height, growth velocity, weight, body mass index and skeletal maturation.
Of 406 citations searched, 384 (94%) were excluded for the following exclusive reasons (Figure 1): (1) duplicate references (N = 11), (2) not a randomised controlled trial (N = 76), (3) not a parallel-group study (N = 84), (4) participants aged < one year or ≥ 18 years (N = 33), (5) participants not asthmatic (or participants with asthma selected for another co-morbidity, e.g. hypertension, diabetes) (N = 16), (6) participants with episodic asthma (N = 2), (7) acute and emergency care settings (N = 13), (8) no daily ICS stable dose in all participants in one of the comparison groups (N = 86), (9) not testing an additional ICS dose using the same molecule in all participants of the other comparison group (N = 50), (10) co-interventions with oral corticosteroids (N = 3), and (11) treatment administered for less than 12 weeks (N = 10). Reasons for exclusion are provided in the Characteristics of excluded studiestable.
Risk of bias in included studies
Details on risk of bias for each included trial are presented in the Characteristics of included studies tables. A graphical summary of risk of bias judgements is presented in Figure 2. Although all trials were randomised, only 14 comparisons (41%) reported the method of randomisation.
|Figure 2. Risk of bias summary: review authors' judgements about each risk of bias item for each included study.|
26 comparisons did not mention the method of concealment of treatment, and eight comparisons (23.5%) reported use of an appropriate concealment technique.
31 comparisons (90%) reported double-blinding with convincing details, two comparisons (Chen 2001; Doniec 2004) did not report sufficient information to allow the review authors to ascertain blinding and one comparison (Giorgi 1998) used an open-label study design.
Incomplete outcome data
31 comparisons (91%) reported all data with balanced numbers in both groups, and data from three comparisons (Giorgi 1998; Chen 2001; Lemanske 2004) were unclear. All trials reported numbers of and reasons for withdrawals in both comparison groups. The proportion of overall withdrawals was variable between studies (10% to 30%), with a balance in withdrawal rates noted between groups given different ICS doses.
33 comparisons (97%) reported all outcomes mentioned in the methods section, with no apparent bias, and one comparison (Chen 2001) was unclear.
Other potential sources of bias
In 31 comparisons, we encountered no other significant sources of bias, two comparisons (Chen 2001; Doniec 2004) were unclear and one comparison (Giorgi 1998) was an open-label study for which the primary outcome was not specified clearly.
Except for three trials, all eligible trials contributing data were of high methodological quality. Two of four comparisons contributing to the primary outcome (Allen 1998; Skoner 2008) were of high methodological quality.
Effects of interventions
Linear growth velocity (cm/y)
A statistically significant group difference in linear growth (cm/y) over 12 months was noted between intervention (lower ICS dose) and control (higher ICS dose) groups (four comparisons; N = 728 children; MD 0.20 cm/y, 95% CI 0.02 to 0.39; Figure 3); no heterogeneity was apparent. The different molecules used (mometasone, ciclesonide and fluticasone) did not seem to influence the magnitude of effect: χ
|Figure 3. Forest plot of comparison: 1 Inhaled corticosteroids dose-response effect, outcome: 1.1 Growth velocity (cm/y) by stadiometry from 0-12 months.|
|Figure 4. Forest plot of comparison: 1 Inhaled corticosteroids dose-response effect, outcome: 1.2 Subgroup analysis on the ICS molecules: growth velocity by stadiometry from 0-12 months.|
We could not perform subgroup analysis on age, severity and ICS dose, as all trials contributing data to the primary outcome had similar characteristics in that they enrolled prepubertal children with mild or unknown severity of airway obstruction, used similarly low ICS doses and did not report or failed to specify the use of co-interventions. Of note, in all four comparisons contributing data, the ICS dose difference between the two groups was less than or equal to 150 μg of HFA-beclomethasone.
As all trials contributing data to the primary outcome were published in full text with high methodological quality and were sponsored by the pharmaceutical industry, we could not perform sensitivity analyses to assess bias due to publication status, poor methodology or funding status. As the adherence rate for ICS was seldom or incompletely reported, sensitivity analysis was not performed on this criterion.
No statistically significant group differences in linear growth (standardised in cm/y) were seen over the first three months (six comparisons; N = 1114 children; MD -0.12, 95% CI -0.51 to 0.27; Analysis 1.3) and no heterogeneity was apparent. Only two comparisons from the same trial provided data on growth velocity from zero to six months ( Analysis 1.4) and from three to six months ( Analysis 1.5); in both cases, a statistically significant group difference was not reported.
Change in growth velocity (cm/y)
Only one trial reported change in growth velocity from zero to 12 months with no statistically significant group difference (one comparison; N = 181 children; MD 0.06 cm/y, 95% CI -0.43 to 0.55; Analysis 1.6).
Change in height (cm)
This outcome reflects the net change between final and initial height, without linear regression or adjustment for important co-variates such as age, sex, puberty and baseline height. A statistically significant group difference was noted in the change in height from zero to three months in favour of the higher ICS dose (nine comparisons; N = 944 children; MD -0.15 cm, 95% CI -0.28 to -0.02; Analysis 1.7); children were described as having mild to moderate to severe asthma, and the ICS used were cicleconide, budesonide and fluticasone. However, the group difference was not statistically significant over longer or subsequent periods, that is, from zero to six months (three comparisons; N = 211 children; MD -0.03, 95% CI -0.33 to 0.27) ( Analysis 1.8), from three to six months (two comparisons; N = 58 children; MD -0.01, 95% CI 0.74 to 0.71) ( Analysis 1.9) and from zero to 12 months (four comparisons; N = 548 children; MD 0.25, 95% CI -0.04 to 0.54; Analysis 1.10).
Change in standard deviation score (SDS) (height)
No statistically significant group difference in change in SDS (height) from zero to 12 months was reported (three comparisons; N = 328 children; MD 0.08, 95% CI -0.03 to 0.20; Analysis 1.11).
Change in weight (kg)
Change in body mass index (BMI) (kg/m
Change in skeletal maturation
Only one trial reported change in skeletal maturation, with a statistically significant group difference from zero to 12 months in favour of a lower ICS dose (one comparison; N = 181 children; MD 0.18, 95% CI 0.02 to 0.34; Analysis 1.17).
This meta-analysis aggregated data from 10 paediatric trials, providing 17 comparisons, as several studies tested more than two different doses of ICS or provided additional data subgrouped by age. In the four trials reporting the main outcome, a statistically significant group difference was seen in linear growth velocity measured by stadiometry over 12 months in prepubertal school-aged children treated with low doses (i.e. 50 to 100 μg) versus low to medium doses (i.e. 200 μg of fluticasone, mometasone and ciclesonide). Of note, the statistically significant group difference was observed despite the small ICS dose difference between compared groups, varying between 100 and 150 μg/d (although most vary by 100 μg/d) of HFA-propelled beclomethasone or equivalent in the four studies pooled. Of interest, a change in height between zero and three months showed a significant decrease of 0.15 cm in the opposite direction, that is, in disfavour of a lower ICS dose, underlying the impact of neglecting important co-variates influencing growth (e.g. sex). This also raised the possibility of a beneficial effect of rapidly achieving asthma control (although this was not measured) and the impact of the timing of measurement of effect size, as this unadjusted group difference was not observed over subsequent and longer time periods. No statistically significant change from baseline in linear growth velocity, weight and body mass index was noted over zero to 12 months of ICS therapy in children. Our findings suggest a clear, yet small, dose-dependent effect on growth when ICS are used at 200 μg/d or less—the cutoff for low to medium doses of ICS in children.
The main outcome, growth velocity, that is, the pattern of growth measured repeatedly over time and adjusted for relevant co-variates (in all individual trials but one (Allen 1998)), was measured in prepubertal school-aged children (< 12 years) treated with fluticasone propionate, ciclesonide and mometasone for 52 weeks. Of the 10 trials contributing data, only three trials (four comparisons) contributed data to the primary outcome (i.e. growth velocity (cm/y)) from zero to 12 months; all performed repeated height measurements using a stadiometer, were funded by pharmaceutical companies and were of high methodological quality. Trials used either a dry powder inhaler or a metered-dose inhaler with spacer to deliver these three molecules with lower systemic bioavailability than budesonide and beclomethasone. Because of trial homogeneity, it was not possible to explore a possible modifier effect of age, severity of airway obstruction, asthma control, use of co-interventions and ICS dose difference on growth velocity. Indeed, trials contributing data to this outcome predominantly compared low ICS doses versus low to medium doses, with a dose difference of 100 to 150 μg/d of HFA-beclomethasone equivalent (GINA 2014); higher doses of ICS theoretically offer greater potential for growth suppression (NHLBI Expert Panel Report 2012).
No effect of the choice of molecules within those tested was apparent. Indeed, several placebo-controlled trials and Cochrane reviews have documented molecule dependency of growth suppression of ICS. Zhang and colleagues (Zhang 2011) are evaluating the growth-suppressive effect of several ICS molecules compared with placebo, reporting minimal and less effect of fluticasone, mometasone and ciclesonide compared with budesonide and beclomethasone. Trials aggregated in this latter review had independently documented a growth-suppressive effect at equivalent ICS doses of between 1.1 and 1.2 cm/y (CAMP Research Group 2000; CAMP Research Group 2012) with budesonide, 0.7 cm/y with mometasone (Skonner 2011), a non-significant group difference of 0.43 cm/y with fluticasone (Sharek 2000b) and none with ciclesonide (Skoner 2008) in prepubertal school-aged children, suggesting molecule dependence of the impact of ICS on growth. This finding is consistent with that of a previous Cochrane systematic review (Sharek 2000a), which had produced solid evidence supporting the growth suppression of 400 μg of inhaled CFC-propelled beclomethasone (equivalent to 200 HFA-BDP) estimated at 1.54 cm/y over seven to 12 months in children with mild to moderate asthma. Current findings provide a clear indication that the use of ICS molecules believed to have no or little suppressive effect does have a minor, yet statistically significant, effect on growth when used at the lowest cutoff of the medium dosage compared with a lower dose.
In this review, the observed group difference of 0.2 cm in growth velocity over the first year of treatment (with an upper confidence interval limit of 0.4 cm/y), associated with an ICS dose higher by 100 to 150 μg, represents less than half the observed effect with similar doses compared with placebo (CAMP Research Group 2000; Sharek 2000a; Sharek 2000b; Skonner 2011; CAMP Research Group 2012). It is consistent with a very small dose-response effect and arguably is impossible to detect on a standard growth curve. One must recognise that the small observed group difference with the use of most recent molecules (fluticasone, mometasone and ciclesonide) might be much higher with a higher ICS dose and/or with older molecules (budesonide and beclomethasone), which have well-documented growth-suppressing effects.
The two included trials (Shapiro 1998 b; Verberne 1998) that compared low doses versus higher doses of ICS (800 HFA-BDP equivalent) contributed between 3% and 30% of the weight in only a few outcomes (1.7, 1.8, 1.10, 1.11 and 1.12), such that we cannot adequately explore the possibility of a differential effect on growth of a high versus low ICS dose. Although poorly controlled asthma may delay growth in children (NHLBI Expert Panel Report 2012), evidence to support this statement is weak. Yet we cannot rule out the possibility of a growth-suppressive effect of poorly controlled asthma in children receiving a lower ICS dose, which could counterbalance the growth suppression associated with a higher ICS dose. If disease-associated growth suppression was indeed possible, even in children with mild to moderate asthma, the design of this review is adequate, as we are interested in the net growth-suppressive effect of ICS therapy in children with asthma. In the absence of a placebo-controlled group, we cannot rule out the unlikely hypothesis that most growth retardation may occur at a very low dose of ICS therapy, which could explain the clinically small group difference between different ICS doses. The systemic availability of ICS is directly related to cortisol suppression and growth suppression, especially in children. The particle size of the drug molecule and use of different devices influence systemic availabilities (Martin 2002; Agertoft 2003; Agertoft 2003a). The third of this series of Cochrane reviews will examine the effects of different devices on the growth of asthmatic children.
As trials contributing data lasted a maximum of one year, the long-term impact of different ICS doses on growth velocity beyond one year could not be explored. Our observations complement those of several placebo-controlled studies, suggesting that the growth-suppressive effect of ICS is non-cumulative (Simons 1997) and may be associated with partial catch-up (Guilbert 2006a), as a growth deficit may be sustained until adulthood (CAMP Research Group 2012).
Of interest, the significant group difference in the 'unadjusted' change in height between zero and three months suggests a favourable effect of ICS on growth in the first three months of use, perhaps via improved asthma control. Of note, 54% of the weight of this analysis is derived from a single trial testing various doses of ciclesonide (with a molecule with no demonstrated suppressive effect on growth) in children with partially or poorly controlled asthma (Pedersen 2010; Pedersen 2010 b). However, this hypothesis is weakened by the absence of any statistically significant effect observed between three and six months and between zero and six months, suggesting a transient beneficial effect on growth, insufficient power or a type 1 error, that is, falsely identifying a significant effect when one does not exist. Of importance, the absence of adjustment for important confounders decreases the quality of the evidence derived from this outcome.
No statistically significant group difference was observed in other aggregated parameters, namely, change from baseline in weight, change in SD scores (height) and body mass index. A significant group difference in skeletal maturation of a quarter of a year was observed, in disfavour of a higher dose (200 μg/d), with an ICS group difference of 100 μg/d of HFA-propelled beclomethasone or equivalent (Allen 1998). Given that children with asthma may have delayed puberty (boys more than girls) (NHLBI Expert Panel Report 2012), whether the delayed maturation is due to poorer asthma control or is associated with greater use of ICS, or both, remains to be determined. Nevertheless, the observation on skeletal maturation, derived from a single study, requires replication.
Summary of main results
Three industry-funded trials with high methodological quality (resulting in four dose comparisons) contributed data to the main outcome, that is, growth velocity; they measured 728 school-aged children with mild to moderate asthma and used one of three molecules (fluticasone, ciclesonide or mometasone) to compare groups with a dose difference ≤ 150 μg over 52 weeks. A significant group difference in linear growth was observed over 12 months, indicating lower growth velocity in the higher ICS dose group (mean difference 0.20 cm/y, 95% CI 0.02 to 0.39); no heterogeneity was apparent. Within aggregated trials, the different ICS molecules did not significantly influence the magnitude of effect (P value 0.33), but no trial contributing data to the main outcome used budesonide or beclomethasone.
Overall completeness and applicability of evidence
This review summarises the best evidence available until March 2014 as derived from 10 trials (resulting in 17 comparisons) aggregating 3394 children with mild to moderate persistent asthma. Most trials were of high methodological quality. The systematic search, which identified eligible trials from published and unpublished reports (406 citations) and used selection and data extraction by two independent review authors, minimised the risk of inclusion bias. The outstanding collaboration of study authors and pharmaceutical groups from six trials (resulting in eight comparisons) allowed us to obtain additional unpublished data and to confirm methodological quality, both of which strengthened the meta-analysis. Because of the paucity of trials reporting these data, four of 15 secondary outcomes could not be aggregated. The long-term impact of low versus high ICS dose on growth velocity, weight, skeletal maturation and body mass index in children using the same and older ICS molecules beyond one year of follow-up remains to be addressed. Sensitivity analysis could not be performed, as all trials were at low risk of bias, the adherence rate of ICS was seldom reported and all included trials contributing data to the main outcome were funded by the pharmaceutical industry and published as full text. In real life, most physicians would adjust downward or upward the dose of ICS needed to maintain control; we acknowledge that the artificially fixed dose for one to four years would overestimate growth suppression when compared with the recommended practice of decreasing to the minimal effective dose, yet this is a basic requirement of FDA guidelines for assessment of the effects of ICS on growth. Our study results support the Global Initiative for Asthma (GINA) guideline recommendations and serve as a reminder that physicians should strive to adjust to the minimal effective ICS dose, irrespective of the ICS molecule selected.
Quality of the evidence
The quality of evidence of growth velocity was high, but for outcomes reflecting change in height from baseline between treatment groups, the quality of evidence was downgraded to moderate owing to possible prognostic imbalance from the use of unadjusted data in the analysis. We downgraded the quality of evidence to low for BMI, weight and skeletal maturation due to imprecision (See Summary of findings for the main comparison).
Potential biases in the review process
Some bias may or may not have affected the magnitude of effect. All trials contributing data to the main outcome used a stadiometer to measure growth; this enhances the internal validity of the findings. As each trial compared different doses using the same device, we could not explore the possibility that the magnitude of effect may be associated with the choice of inhalation device; however a linked Cochrane review is addressing this point (Zhang 2011).
Agreements and disagreements with other studies or reviews
To our knowledge, no prior systematic review has evaluated the relationship between dose of ICS and risk of growth impairment in children with persistent asthma.
Implications for practice
In prepubescent school-aged children with mild to moderate persistent asthma, a very small but statistically significant difference in linear growth over 12 months was observed between groups using ICS, with a dose difference ≤ 150 μg HFA-beclomethasone equivalent over 52 weeks. A group difference of 0.2 cm was observed, favouring higher growth velocity with the lower ICS dose of fluticasone, mometasone or ciclesonide. As ICS doses most often were in the low range or at the limit of low and medium doses (200 μg), data were insufficient to allow exploration of a potential dose-response relationship between ICS for a difference greater than 150 μg. We are unable to comment on the possible effects on growth of different ICS molecules, although fluticasone, mometasone and ciclesonide at doses of 200 μg/d or less did not appear to explain any variation in the size of effect across the studies. In view of prevailing parents’ and physicians’ concerns about the growth-suppressive effect of ICS, lack or inadequate reporting of growth measurements in more than 86% (19/22) of eligible paediatric trials is a matter of concern and should call for systematic reporting of growth in all ICS paediatric trials. Until more data on low versus moderate and higher ICS doses are available, we recommend that ICS should be used at the lowest effective dose with the safest ICS molecules, and that children's growth should be systematically monitored during any ICS treatment.
Implications for research
Long-term (longer than one year) trials of high methodological quality with adequate documentation of linear growth velocity in children with asthma treated with ICS are needed to provide a fair comparison of the safety of different ICS dose options. Future trials should aim for the following design characteristics.
Given the paucity of paediatric trials reporting growth, growth measurements should be a requirement for all ICS drug trials whether funded by pharmaceutical companies or national granting agencies.
We are indebted to the following individuals, who replied to our request for confirmation of methodology and additional data in the requested format when possible: Dr. Paul LP Brand, Dr. AA Vaessen-Verberne.
A special thanks to the following pharmaceutical groups, which replied to our request for confirmation of methodology and additional data in the requested format when possible: GlaxoSmithKline Inc, Takeda Global Research & Development Centre (Europe) Ltd and AstraZeneca R&D, Mölndal, Sweden.
We are indebted to the Cochrane Airways Review Group, namely, Dr Emma Welsh, Elizabeth Stovold and Emma Jackson, for assistance with the literature search and ongoing support. A special thanks to Taixiang Wu from the Cochrane Review Group for assistance in translating three Chinese references.
We are thankful to Inge Axelsson for providing inputs to drafting of the protocol.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Appendix 1. Sources and search methods for the Cochrane Airways Group Specialised Register (CAGR)
Electronic searches: core databases
Handsearches: core respiratory conference abstracts
MEDLINE search strategy used to identify trials for the CAGR
1. exp Asthma/
3. (antiasthma$ or anti-asthma$).mp.
4. Respiratory Sounds/
6. Bronchial Spasm/
8. (bronch$ adj3 spasm$).mp.
10. exp Bronchoconstriction/
11. (bronch$ adj3 constrict$).mp.
12. Bronchial Hyperreactivity/
13. Respiratory Hypersensitivity/
14. ((bronchial$ or respiratory or airway$ or lung$) adj3 (hypersensitiv$ or hyperreactiv$ or allerg$ or insufficiency)).mp.
15. ((dust or mite$) adj3 (allerg$ or hypersensitiv$)).mp.
Filter to identify RCTs
1. exp "clinical trial [publication type]"/
2. (randomised or randomised).ab,ti.
11. 9 not (9 and 10)
12. 8 not 11
The MEDLINE strategy and RCT filter are adapted to identify trials in other electronic databases.
Last assessed as up-to-date: 5 March 2014.
Contributions of authors
Aniela Ignea Pruteanu reviewed the literature search conducted until March2014, identified and reviewed all citations for relevance, reviewed all included trials for methodology and data extraction, verified all references, described the studies and performed data entry, analysed and interpreted results of the meta-analysis, wrote the first draft of the manuscript and approved the final version.
Bhupendrasinh Chauhan reviewed all included trials for methodology and data extraction, verified the description of studies and data entry, contributed to analysis and interpretation of data, revised all drafts of the manuscript, prepared responses to editorial comments and approved the final version.
Linjie Zhang wrote the review protocol, reviewed the literature search conducted until March2014, identified and reviewed half of the citations for relevance and approved the final version of the review.
Sílvio OM Prietsch provided input to drafting of the protocol, reviewed the literature search conducted until March 2014 and identified and reviewed half of the citations for relevance.
Prof Francine Ducharme revised and approved the protocol, requested the literature search, identified and contacted corresponding authors and/or pharmaceutical companies to solicit their collaboration in this review and in identifying other possibly relevant trials, corresponded with authors or pharmaceutical companies to verify methodology and data extraction, verified all references, described studies and performed data entry, analysed and interpreted results and approved the final version of the meta-analysis.
Declarations of interest
Aniela Ignea Pruteanu, Bhupendrasinh Chauhan, Linjie Zhang and Sílvio OM Prietsch: none known.
Prof. Francine Ducharme has received travel support, research funds and fees for speaking from Glaxo SmithKline, Novartis, Nycomed and/or Merck Frosst Inc.
Sources of support
- None, Other.
- No sources of support supplied
Differences between protocol and review
The review is different from the protocol in the following ways.
- Limited lower age to one year instead of 'up to 18 years.'
- Defined which other interventions were accepted: other non-steroidal asthma drugs (e.g. long-acting beta-agonists or leukotriene receptor antagonists).
- Added post hoc secondary outcomes (change in body mass index; change in skeletal maturation).
- Removed subgroup analyses as they were included as different outcomes: time points of outcome measurements.
- Added post hoc analysis: ICS dose difference (in μg of HFA-beclomethasone or equivalent) between groups.
- Added two outcomes: change in body mass index and change in skeletal maturation.
- Following recommendations of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008), the fixed effect model was used for the data analysis if the heterogeneity of pooled trials is less than 50%; otherwise the random effects model was used, despite the use of random effect models was proposed for all data analysis in the protocol.
- Several included trials contributed more than one comparison and one group compared with two or more groups. So the individual comparison was used as the unit of analysis in place of individual trial.
Medical Subject Headings (MeSH)
Administration, Inhalation; Adrenal Cortex Hormones [*administration & dosage; adverse effects]; Androstadienes [administration & dosage; adverse effects]; Anti-Asthmatic Agents [*administration & dosage; adverse effects]; Asthma [*drug therapy]; Beclomethasone [administration & dosage; adverse effects]; Budesonide [administration & dosage; adverse effects]; Dose-Response Relationship, Drug; Fluticasone; Growth [*drug effects]; Growth Disorders [*chemically induced]; Mometasone Furoate; Pregnadienediols [administration & dosage; adverse effects]; Pregnenediones [administration & dosage; adverse effects]; Randomized Controlled Trials as Topic
MeSH check words