Abstract Background Duchenne muscular dystrophy (DMD) is the most common muscular dystrophy of childhood. Untreated, this incurable disease, which has an X-linked recessive inheritance, is characterised by muscle wasting and loss of walking ability, leading to complete wheelchair dependence by 13 years of age. Prolongation of walking is a major aim of treatment.

Evidence from randomised controlled trials (RCTs) indicates that corticosteroids significantly improve muscle strength and function in boys with DMD in the short term (six months), and strength at two years (two-year data on function are very limited). Corticosteroids, now part of care recommendations for DMD, are largely in routine use, although questions remain over their ability to prolong walking, when to start treatment, longer-term balance of benefits versus harms, and choice of corticosteroid or regimen. We have extended the scope of this updated review to include comparisons of different corticosteroids and dosing regimens. Objectives To assess the effects of corticosteroids on prolongation of walking ability, muscle strength, functional ability, and quality of life in DMD; to address the question of whether benefit is maintained over the longer term (more than two years); to assess adverse events; and to compare efficacy and adverse effects of different corticosteroid preparations and regimens.

Methodology and Precision Issues. Schwartz*, D. In comparison with commonly used manual methods the improvement in accuracy is significant. Registration uncertainties. Key words: MEG; MRI; Data fusion; Localization; Anatomical constraint; Registration evaluation; Multimodal registration.

Search methods On 16 February 2016 we searched the Cochrane Neuromuscular Specialised Register, CENTRAL, MEDLINE, EMBASE, CINAHL Plus, and LILACS. We wrote to authors of published studies and other experts. We checked references in identified trials, handsearched journal abstracts, and searched trials registries. Selection criteria We considered RCTs or quasi-RCTs of corticosteroids (e.g. Prednisone, prednisolone, and deflazacort) given for a minimum of three months to patients with a definite DMD diagnosis.

We considered comparisons of different corticosteroids, regimens, and corticosteroids versus placebo. Data collection and analysis The review authors followed standard Cochrane methodology. Main results We identified 12 studies (667 participants) and two new ongoing studies for inclusion. Six RCTs were newly included at this update and important non-randomised cohort studies have also been published. Some important studies remain unpublished and not all published studies provide complete outcome data.

Primary outcome measure: one two-year deflazacort RCT (n = 28) used prolongation of ambulation as an outcome measure but data were not adequate for drawing conclusions. Secondary outcome measures: meta-analyses showed that corticosteroids (0.75 mg/kg/day prednisone or prednisolone) improved muscle strength and function versus placebo over six months (moderate quality evidence from up to four RCTs).

Evidence from single trials showed 0.75 mg/kg/day superior to 0.3 mg/kg/day on most strength and function measures, with little evidence of further benefit at 1.5 mg/kg/day. Improvements were seen in time taken to rise from the floor (Gowers' time), timed walk, four-stair climbing time, ability to lift weights, leg function grade, and forced vital capacity.

One new RCT (n = 66), reported better strength, function and quality of life with daily 0.75 mg/kg/day prednisone at 12 months. One RCT (n = 28) showed that deflazacort stabilised muscle strength versus placebo at two years, but timed function test results were too imprecise for conclusions to be drawn. One double-blind RCT (n = 64), largely at low risk of bias, compared daily prednisone (0.75 mg/kg/day) with weekend-only prednisone (5 mg/kg/weekend day), finding no overall difference in muscle strength and function over 12 months (moderate to low quality evidence).

Two small RCTs (n = 52) compared daily prednisone 0.75 mg/kg/day with daily deflazacort 0.9 mg/kg/day, but study methods limited our ability to compare muscle strength or function. Adverse effects: excessive weight gain, behavioural abnormalities, cushingoid appearance, and excessive hair growth were all previously shown to be more common with corticosteroids than placebo; we assessed the quality of evidence (for behavioural changes and weight gain) as moderate. Hair growth and cushingoid features were more frequent at 0.75 mg/kg/day than 0.3 mg/kg/day prednisone. Comparing daily versus weekend-only prednisone, both groups gained weight with no clear difference in body mass index (BMI) or in behavioural changes (low quality evidence for both outcomes, one study); the weekend-only group had a greater linear increase in height. Very low quality evidence suggested less weight gain with deflazacort than with prednisone at 12 months, and no difference in behavioural abnormalities. Data are insufficient to assess the risk of fractures or cataracts for any comparison.

Non-randomised studies support RCT evidence in showing improved functional benefit from corticosteroids. These studies suggest sustained benefit for up to 66 months. Adverse effects were common, although generally manageable. According to a large comparative longitudinal study of daily or intermittent (10 days on, 10 days off) corticosteroid for a mean period of four years, a daily regimen prolongs ambulation and improves functional scores over the age of seven, but with a greater frequency of side effects than an intermittent regimen. Authors' conclusions Moderate quality evidence from RCTs indicates that corticosteroid therapy in DMD improves muscle strength and function in the short term (twelve months), and strength up to two years. On the basis of the evidence available for strength and function outcomes, our confidence in the effect estimate for the efficacy of a 0.75 mg/kg/day dose of prednisone or above is fairly secure.

There is no evidence other than from non-randomised trials to establish the effect of corticosteroids on prolongation of walking. In the short term, adverse effects were significantly more common with corticosteroids than placebo, but not clinically severe. A weekend-only prednisone regimen is as effective as daily prednisone in the short term (12 months), according to low to moderate quality evidence from a single trial, with no clear difference in BMI (low quality evidence). Very low quality evidence indicates that deflazacort causes less weight gain than prednisone after a year's treatment. We cannot evaluate long-term benefits and hazards of corticosteroid treatment or intermittent regimens from published RCTs. Non-randomised studies support the conclusions of functional benefits, but also identify clinically significant adverse effects of long-term treatment, and a possible divergence of efficacy in daily and weekend-only regimens in the longer term. These benefits and adverse effects have implications for future research and clinical practice.

Plain language summary Corticosteroid therapy for Duchenne muscular dystrophy Review question Is there new evidence for benefit from corticosteroids for prolongation of walking, and improving muscle strength and functional abilities in Duchenne muscular dystrophy (DMD), particularly over the long term (more than two years)? Are different corticosteroids, or different regimens equally effective, with similar side effect profiles? Background DMD is an incurable disease beginning in childhood that almost exclusively affects boys. Muscle wasting and loss of walking lead to wheelchair dependence and early death. Randomised controlled trials (RCTs) have shown that corticosteroids improve muscle strength and function for up to six months and strength up to two years (evidence on function at two years is limited). Data from other study types suggest that corticosteroids produce better function over a five-year period in many patients.

Overall, long-term benefit remains unclear, and has to be weighed against long-term side effects. It is also unclear whether different corticosteroids differ greatly in side effects. Earlier versions of this review found insufficient evidence to determine whether an intermittent regimen is as effective as a daily regime, or produces fewer side effects.

Study characteristics We found 12 studies of corticosteroid treatment in DMD, involving a total of 667 randomised boys; two other studies are ongoing. Among the 12 completed studies, the treatments were: a corticosteroid versus inactive medicine (placebo) (in nine trials); daily versus weekend-only prednisone (in one trial); and deflazacort versus prednisone (in three trials). Some studies included more than one comparison; some were not fully reported or provided results that could not be analysed. Key results and quality of the evidence One trial, a two-year study comparing a corticosteroid (deflazacort) with placebo, assessed the effects of corticosteroids on the ability to continue walking, but the data were not suitable for analysis. Most studies did not report ability to continue walking. At the usual 0.75 mg/kg/day dose, corticosteroids improved muscle strength and function over six months compared to placebo.

These results are based on combined data (up to 152 participants) from four trials, which provided moderate quality evidence. Improvements were seen in timed tests (eg. Timed walk or run, time to stand, stair climb), ability to lift weights, a leg function grade, and a measure of the strength of muscles used in breathing. Evidence from single trials showed 0.75 mg/kg/day prednisone to be superior to 0.3 mg/kg/day on most strength and function tests, with little evidence of greater benefit at 1.5 mg/kg/day. Changes in appearance and hair growth were more common at 0.75 mg/kg/day than 0.3 mg/kg/day. One RCT (n = 66) also reported better strength, function and quality of life at 12 months with daily 0.75 mg/kg/day prednisone.

The two-year RCT, which had 28 participants, showed that deflazacort stabilised muscle strength for up to two years compared to placebo. This study did not show benefit on timed tests at two years; however, these results are imprecise and at high risk of bias, with less than half the original participants contributing data. One trial found that changes in muscle strength and function were similar with daily and weekend-only prednisone regimens over a 12-month period (low to moderate quality evidence). Adobe Acrobat Installer Http Request Failed Elder. Two small RCTs compared daily prednisone 0.75 mg/kg/day with daily deflazacort 0.9 mg/kg/day, but trial methods did not allow comparisons of muscle strength or function. Previous versions of this review have found adverse events such as excessive weight gain, abnormal behaviour, changes in appearance, and abnormal hair growth to be more common with corticosteroids than with placebo. We assessed the quality of evidence for abnormal behaviour and weight gain for this review and found it to be moderate.

The newer study of daily versus weekend-only prednisone showed that both groups gained weight. The body mass index (BMI; a measure of weight for height) did not show any clear difference between the regimens (low quality evidence). The weekend-only group had a greater increase in height.

According to very low quality evidence from two studies, deflazacort appeared to cause less weight gain at one year than prednisone, and no significant difference in numbers with behaviour change. Data were insufficient to assess the risk of fractures or cataracts. The evidence is up to date to February 2016. Summary of findings for the main comparison. Corticosteroids versus placebo for Duchenne muscular dystrophy • 1Mean of mean control group values. 2Two additional trials (6 and 12 months' duration), which could not be included in the meta-analysis, also demonstrated improvements in muscle strength over placebo with daily or intermittent prednisone 0.75 mg/kg/day - see text for details (; ). 3Single downgrading for unclear risk of allocation bias and possible publication bias. 4Mean of the mean control group values at 6 months from, and (data are not provided in report).

5Two additional trials (6 and 12 months' duration), which could not be included in the meta-analysis, also demonstrated improvements in timed walk over placebo with daily or intermittent prednisone 0.75 mg/kg/day - see text for details (; ). 6An additional 6-month trial, which could not be included in the meta-analysis, also demonstrated improvements in 4-stair climb over placebo with intermittent prednisone (0.75 mg/kg/day given for the first 10 days of every month for six months) - see text for details (; ). 7Two additional trials (6 and 12 months' duration), which could not be included in the meta-analysis, monitored weight during daily or intermittent prednisone 0.75 mg/kg/day; no clear difference was present between groups at six months (intermittent dosing) or a year; however these results were imprecise - see text for details (; ). 8For details of other dosages and the deflazacort versus placebo comparison see the review text.

Corticosteroids versus placebo for Duchenne muscular dystrophy Patient or population: patients with Duchenne muscular dystrophy Setting: outpatient Intervention: corticosteroids Comparison: placebo Outcomes Anticipated absolute effects * (95% CI) Relative effect (95% CI) Number of participants (studies) Quality of the evidence (GRADE) Comments Risk or score/value with placebo Risk or score/value with corticosteroids Prolongation of time to loss of ambulation - not reported See comment See comment Not estimable - - An outcome measure in one 2-year trial (n = 28). Summary of findings 2. Weekend-only versus daily prednisone for Duchenne muscular dystrophy • 1Trial authors defined equivalence limits (limits within which the regimens can be considered of equivalent efficacy) of ± 2 lb for QMT. We downgraded once as although serious imprecision is present, the CIs fall within the equivalence limits. 2We downgraded the quality of evidence twice for serious imprecision due to small sample size; trial authors defined equivalence limits (limits within which the regimens can be considered of equivalent efficacy) of ± 17 points and the CIs allow for the possibility of non-equivalence. 3Trial authors defined equivalence limits (limits within which the regimens can be considered of equivalent efficacy) of ± 0.4 seconds for timed tests.

We downgraded once; although serious imprecision is present, the CIs fall within the equivalence limits. 4 We downgraded the evidence for very serious imprecision (due to small sample sizes, plus the CI includes appreciable differences in favour of either intervention). Summary of findings 3. Deflazacort versus prednisone for Duchenne muscular dystrophy • 1We downgraded the quality of evidence twice for a high risk of bias in most domains. 2We downgraded the quality of evidence once for possible publication bias - a large study remains unpublished ().

3Analysis involved some statistical assumptions in calculating SD. 4CIs include the possibility of both a large effect and a clinically unimportant effect (i.e. • • Study flow diagram. Included studies Corticosteroids versus placebo See. We included data from five randomised, parallel-group, double-blind studies of corticosteroids versus placebo (;;;; ), and one randomised, placebo-controlled, cross-over trial ().

Overall, these studies comprised 332 participants. The 315 participants in the randomised parallel-group trials involved 88 in the placebo groups and 161 in the corticosteroid treatment groups (;;;; ). One hundred and one of the 218 participants in the control groups and 164 of the 197 in the corticosteroid groups were walking, either independently or with the help of long leg braces. The corticosteroid treatment groups included prednisone (n = 170), prednisolone (n = 10), and deflazacort (n = 17).

Was a cross-over study comprising 17 boys, all walking independently, who received prednisone during the six-month active treatment period. In three of the included studies, all lasting six months, treatment groups received prednisone or prednisolone in a daily dose regimen (;; ). These studies included a total of 144 participants in the treatment group and 77 in the placebo group. Prednisone is broken down in the body to prednisolone and they are equipotent in glucocorticoid effect (; )., used an intermittent regimen, prednisone 0.75 mg/kg/day given for the first 10 days of every month, in the active treatment phase. The participants were 17 independently ambulant boys.

One study (n = 28) used deflazacort (2 mg/kg body weight on alternate days for two years) in the treatment group (). This was the only study to address the primary outcome measure of prolongation of walking. Was a placebo-controlled study of 66 independently ambulant boys and the only placebo-controlled corticosteroid study published since the previous update of this review. The intervention was prednisone 0.75 mg/kg/day given daily for a year. The secondary outcome measures of this review were assessed by different parameters and assessment tools in the five prednisolone/prednisone studies that were published in full and which provided data for our secondary outcomes. However, and, the two studies that comprised 80% of the participants for all the four included and analysed studies, used the same outcome measures, as described in (see ).,, and, the other three published studies of prednisolone or prednisone, also used some of these outcome measures., a cross-over trial, reported efficacy as the numbers improving (improved or unchanged across two-thirds or more of the tested measures) and numbers deteriorating during treatment with prednisolone (0.35 mg/kg/day) or placebo. The participants were 37 boys with Duchenne muscular dystrophy (DMD) (22 of whom were ambulant) and four with Becker muscular dystrophy.

Although outcome data were not adequate for inclusion in the review, the trial provided some adverse event data. Was a 20-month study involving 34 boys, who received prednisone 2 mg/kg alternate days (high dose) or placebo. No results have been published. Was a 12-month randomised, double-blind comparative trial with four arms: placebo, prednisone 0.75 mg/kg, deflazacort 0.9 mg/kg, and deflazacort 1.2 mg/kg, published in abstract form. We presume, although the abstract does not specify, that these were daily doses.

After three months, participants in the placebo group were randomised to one of the three active treatment arms and followed up for a further nine months. The abstract presents data for the average muscle score (the method of measurement is not defined), and for weight gain as a percentage of baseline weight. This study was large, at 196 randomised participants, but we were unable to obtain data to allow for any analysis. Weekend-only versus daily prednisone Only one randomised study provided data for a comparison of different prednisone dosing regimes. Performed a double-blind placebo-controlled randomised study comparing daily prednisone 0.75 mg/kg/day (and weekend placebo) with weekend-only prednisone 5 mg/kg/weekend day (and daily placebo), taken over a 12-month period.

The study comprised 64 eligible participants with a mean age of 7.3 years (range 4 to 10), all of whom were ambulant at the start. The study did not measure the primary outcome of this review, prolongation of time to loss of ambulation, but assessed secondary outcomes at 12 months using multiple measures, including the change from baseline of quantitative muscle testing (QMT) arm and leg scores, and mean body mass index (BMI). Deflazacort versus prednisone, a double-blind study, randomised 18 participants to treatment with 0.75 mg/kg/day prednisone (mean age 7.5 years, range, 5.1 to 10) or 0.9 mg/kg/day deflazacort (mean age 8.6 years, range 5.3 to 14.6) for 12 months. Investigators assessed muscle strength and function using a summed Medical Research Council (MRC) score of four muscles (two right upper limb, two right lower limb) and a summed functional score comprising several timed assessments including a 10-metre walk, rise from chair and floor, and four-stair climbing. Mean weight increase after 12 months was expressed as a percentage of initial weight. The trial authors presented outcome data for MRC scores, functional scores, and weight increase graphically, limiting full analysis. They tabulated adverse events.

We requested further data but received no response from study authors. Initially randomised 34 participants to either prednisone 0.75 mg/kg/day or deflazacort 0.9 mg/kg/day. Eight early dropouts occurred, but the trial continued for 12 months with 14 participants in the deflazacort group (mean age 7.1 years, range 3.2 to 10.5) and 12 in the prednisone group (mean age 7.37 years, range 6 to 10). In total, the study ran for 18 months, but at 12 months a further four participants were excluded from the prednisone group due to unacceptable weight gain; these four also had poor motor function scores.

The report presented limited outcome data at 12 and 18 months. We contacted study authors for more data but did not receive a reply. As noted above, also studied this comparison but did not provide data for analysis. Excluded studies Non-randomised excluded studies See. We excluded three RCTs. These were: a comparison of prednisone and azathioprine with no placebo group (), a study of deflazacort versus prednisolone in which the high dropout rate invalidated results (), and a study of ayurvedic medicine, prednisone and placebo in which investigators modified the design mid-trial (). We also listed non-randomised studies in the excluded studies.

Thirty-eight of these were fully published (;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ). Eight non-randomised studies were published in abstract format only (;;;;;;; ). One paper was a discussion of corticosteroid use (). We identified and excluded six review articles reporting the various studies (;;;;; ).

Ongoing studies See. Is a large ongoing randomised double-blind study taking place at 40 centres throughout the US, UK, Canada, Germany, Italy, and Spain. This study is comparing three corticosteroid regimens for efficacy and adverse events: prednisone 0.75 mg/kg/day, prednisone 0.75 mg/kg/day switching between 10 days on and 10 days off treatment, and deflazacort 0.9 mg/kg/day daily.

The planned follow-up period is three to five years. No outcome data are yet available. Risk of bias in included studies illustrates the review authors' 'Risk of bias' assessments of included studies. • • Risk of bias summary: review authors' judgements about each risk of bias item for each included study. Allocation Although all the studies were described as randomised, most reports did not provide enough detail to assess whether the method of randomisation was adequate. We were able to determine that three studies were at low risk of bias (;; ), and one was at high risk of bias ().

For allocation concealment, we assessed three trials at low risk of bias on the basis of information provided by the trial authors (;; ), eight studies at unclear risk of bias, as the reports provided no information, and one study at high risk of bias () (see ). Blinding Trial authors described eight of the 12 studies as double blind, but we considered only six of them at low risk of both performance and detection bias (;;;;; ).

Blinded participants but it is unclear whether blinding of outcome assessors or investigators was attempted. Three studies provided too little information to form a judgement (;; ). We judged both, which used a vitamin control intervention, and, a single-blind study, at high risk of bias. Incomplete outcome data Six of the 12 studies described withdrawals and dropouts and we judged these studies at low risk of bias (;;;;; ). Trialists described dropouts from in a subsequent sequential study (). Reported one dropout and described it in response to the Cochrane authors' request. The risk of attrition bias was unclear in four studies (;;; )..

Was at high risk of attrition bias, as dropouts were those with worse outcomes. Most two-year analyses in included fewer than 50% of the randomised participants and we judged it at high risk of attrition bias.

Selective reporting Reporting bias was difficult to assess as trial registration records and protocols are not available for earlier trials, and outcomes were rarely fully defined in methods. Our assessment of bias was 'high' for six trials and 'unclear' for two.

Only,,, and had a 'low risk' assessment. Other potential sources of bias and were reported in abstracts and provided no information to assess the presence of other bias. Our assessment was 'unclear' for these trials and low risk for others.

Effects of interventions See:;; Corticosteroids versus placebo Six studies provided data for this comparison (;;;;; )., a cross-over trial, reported efficacy as the numbers improving (improved or unchanged across two-thirds or more of the tested measures) and numbers deteriorating. Although these data were not adequate for inclusion in the comparison of outcome measures, the trial did provide adverse event data. The two studies reported in abstract form provided only limited information: reported no usable results and provided some numerical data, but with insufficient detail for analysis. Primary outcome measure: prolongation of time to loss of ambulation Only (n = 28), a two-year study, used prolongation of time to loss of ambulation as an outcome measure. The other studies were of short duration (six months or one year), and not designed to demonstrate prolongation of walking. Reported that deflazacort (2 mg/kg on alternate days) prolonged ambulation by 13 months, but the statistical technique used to infer this result was not appropriate. Four of the 17 participants in the deflazacort group became wheelchair dependent, at a mean interval of 33.2 months after randomisation, versus six of 11 placebo participants, at a mean interval of 20.5 months.

The trial authors reported the difference of 13 months between these two sets of participants who lost walking ability as 'mean prolongation of walking', ignoring the 13 participants in the deflazacort group and five in the placebo group who were still walking at the end of the study. The trialists did not report the age of boys who remained ambulant at the end of the study and this information was not available on contacting the lead investigator. We therefore were not able to construct Kaplan-Meier survival curves for evaluating prolongation of walking as an outcome measure. Secondary outcome measures (1) Muscle strength (a) Average muscle score, and reported muscle strength as an average muscle score (as described in and ).

The two large studies had one placebo arm and two treatment arms (; ). Studied two prednisone dose regimens (0.75 mg/kg/day and 1.5 mg/kg/day), comparing them with a placebo group. Compared 0.3 mg/kg/day and 0.75 mg/kg/day prednisone regimens with placebo. Studied a 0.75 mg/kg daily prednisone regimen, reporting scores for lower limb muscle strength (right hip flexion and right knee extension) according to the MRC scale expanded to a 10-point scale, at six and 12 months.

These data were not suitable for meta-analysis with, and. Evaluated muscle strength in three ways: (a) average muscle strength from 26 muscle groups on the MRC zero to five grading system and the performance scores were added and divided by the number of muscle groups to get the average muscle strength; (b) isometric muscle strength, measured in 24 muscle groups with a Penny and Giles myometer; and (c) hand-grip strength measured bilaterally with a strain gauge. The publication did not report data, nor could the review authors obtain data from the surviving study author. Measured muscle strength in two ways: (a) MRC index calculated by assessing four limb muscle groups using the MRC scale; and (b) myometry (but the number of muscle groups tested and the myometer used were not described). Assessed changes in muscle force in nine muscle groups with hand-held dynamometry (; ). Was a four-way comparison of two doses of deflazacort (0.9 mg/kg and 1.2 mg/kg), prednisone (0.75 mg/kg), and placebo, which reported average change in muscle strength at three months ('based on a standardised method used in several previous trials').

Analysis of pooled data from three trials (n = 147) demonstrated a statistically significant improvement in average muscle score in the prednisone 0.75 mg/kg/day group versus placebo, with a mean difference (MD) of 0.52 (95% confidence interval (CI) 0.33 to 0.71) after six months of treatment; moderate quality evidence (;; ) (see;; ). Removal of, the trial at high risk of bias, had no substantial effect on the result (MD 0.47, 95% CI 0.32 to 0.63). • • Forest plot of comparison: 1 Glucocorticoid corticosteroids versus placebo, outcome: 1.1 MRC - Average muscle score after 6 months of treatment - prednisone. (n = 61) also compared prednisone 0.3 mg/kg/day with placebo and after six months of treatment there was statistically significant improvement in average muscle score in favour of the prednisone group, with a MD of 0.34 (95% CI 0.17 to 0.51) (see ). (n = 65) also compared prednisone 1.5 mg/kg/day with placebo and after six months of treatment there was a statistically significant improvement in average muscle score in the prednisone group, with a MD of 0.45 (95% CI 0.23 to 0.67) (see ).

(n = 16), a cross-over study, compared an intermittent regimen of prednisone (0.75 mg/kg/day given for the first 10 days of every month for six months) with placebo. There was a statistically significant difference in the muscle force during the prednisone phase compared to the placebo phase.

Using the standard errors (SEs) inferred from the quoted P values, the RevMan GIV facility gave a difference in favour of prednisone of 99.2 N (95% CI 15.63 to 182.81) (see ). (n = 196) reported an average change in muscle strength ('based on a standardised method used in several previous trials') after three months. Reported changes were -0.1 with placebo, +0.27 with prednisone 0.75 mg/kg, +0.8 with deflazacort 0.9 mg/kg, and +0.26 with deflazacort 1.2 mg/kg.

For all comparisons versus placebo, P. • • Forest plot of comparison: 1 Glucocorticoid corticosteroids versus placebo, outcome: 1.7 Nine-metre walking/running time after 6 months of treatment - prednisone. Also compared prednisone 0.3 mg/kg/day with placebo. The GIV analysis showed that after six months of treatment, there was a difference in favour of prednisone of -1.18 (95% CI -2.65 to 0.29, n = 40) (see ). Also compared prednisone 1.5 mg/kg/day with placebo.

The GIV analysis showed that after six months of treatment, there was a difference in favour of prednisone of -2.64 seconds (95% CI -4.45 to -0.83, n = 57) (see ). Reported time to walk 10 metres, which prevented inclusion of data in the meta-analysis. The trial reported a post treatment MD (seconds) in favour of daily prednisone 0.75 mg/kg/day over placebo; at six months, the MD was -0.94 (95% CI -1.73 to -0.15, n = 63), and at one year -1.71 seconds (95% CI -2.74 to -0.68, n = 58) (see;. Reported change in timed walk (we assume in seconds); the MD favoured deflazacort (2 mg/kg alternate days) at six months (MD -3.01, 95% CI -4.76 to -1.26, n = 23), but no clear difference was present at 24 months (MD -0.67, 95% CI -2.37 to 1.03, n = 12) (see; ). (c) Four-stair climbing time Five studies reported the time taken to climb four standardised stairs (as described in ) at six months (;;;; ). A decrease in four-stair climbing time indicates ability to ascend stairs faster, representing improvement.

Compared an intermittent regimen of prednisone (0.75 mg/kg/day given for the first 10 days of every month for six months) with placebo, in a cross-over design study. There was a statistically significant difference in four-stair climbing time during the prednisone phase compared to the placebo phase. Using the SEs inferred from the quoted P values the RevMan GIV facility gave a difference in favour of prednisone of -1.93 seconds (95% CI -3.56 to -0.30, n = 16) at 6 months (see ).

Analysis of pooled data from,,, and demonstrated a statistically significant benefit over placebo in four-stair climbing time in the prednisone 0.75 mg/kg/day group after six months of treatment. Using the SEs inferred from the quoted P values for the older studies, and using the RevMan calculator tool to derive SE from the (final values) data, the RevMan GIV facility gave a difference in favour of prednisone of -3.09 seconds (95% CI -4.33 to -1.85, n = 135; moderate quality evidence) (see; ). Removal of had no substantial effect on the result (MD -2.98, 95% CI -4.43 to -1.53), and increased heterogeneity (I 2 = 66%).

Also compared prednisone 0.3 mg/kg/day with placebo. The GIV analysis showed that after six months of treatment, there was a difference in favour of prednisone of -2.68 seconds (95% CI -4.06 to -1.30, n = 32) (see; ). • • Forest plot of comparison: 1 Glucocorticoid corticosteroids versus placebo, outcome: 1.11 Four-stair climbing time after 6 months of treatment - prednisone. Compared prednisone 1.5 mg/kg/day with placebo.

Wiz Khalifa See You Again Lyrics on this page. The GIV analysis showed that after six months of treatment, there was a difference in favour of prednisone of -3.05 seconds (95% CI - 4.41 to -1.69, n = 42) (see ). Compared daily prednisone 0.75 mg/kg/day with placebo. After a year of treatment, the mean difference in four-stair climb time favoured prednisone, at -1.63 seconds (95% CI -3.07 to -0.19, n = 52) (see ). Reported change in 'time, stairs' (not further specified), comparing deflazacort (2 mg/kg alternate days) and placebo. The results (we assume in seconds) were imprecise, allowing for the possibility of effects in either direction, MD -2.96, 95% CI -7.02 to 1.10, n = 23 at six months, and MD 0.63, 95% CI -4.29 to 5.55, n = 11 at 24 months (see; ).

(d) Leg function grade Leg function grade (as described in and ) was assessed in two studies (; ). The leg function grade is assessed on a 10-point scale: grade 1 representing ability to walk and climb stairs without assistance; and grade 10 representing confinement to bed. Analysis of pooled data from these studies demonstrated a statistically significant improvement in the prednisone 0.75 mg/kg/day group versus placebo after six months of treatment, with a MD of -0.41 points (95% CI -0.73 to -0.09, n = 129) (see ). Also compared prednisone 0.3 mg/kg/day with placebo. After six months of treatment the mean improvement in leg function grade was 0.39 points (95% CI 0.01 to 0.79, n = 58) less than in the placebo group (see ).

Also compared prednisone 1.5 mg/kg/day with placebo and after six months of treatment the mean improvement in the prednisone group was 0.49 points (95% CI 0.05 to 0.93, n = 68) less than in the placebo group (see ). (3) Pulmonary function - forced vital capacity (FVC) FVC (as described in ) was measured in two studies (; ). Analysis of pooled data from these studies demonstrated a mean improvement in FVC in the prednisone 0.75 mg/kg/day group, after six months of treatment of 0.17 L more than in the placebo group (95% CI 0.10 to 0.24, n = 127) (see ). Also compared prednisone 0.3 mg/kg/day with placebo. After six months of treatment the improvement in FVC in the prednisone group was 0.16 L (95% CI 0.05 to 0.27, n = 59) more than in the placebo group (see ).

Also compared prednisone 1.5 mg/kg/day with placebo. After six months of treatment the mean improvement in FVC in the prednisone group was 0.14 L (95% CI 0.05 to 0.23, n = 62) more than in the placebo group (see ). (4) Quality of life (QoL) Measured in and. Measured QoL with the DUX-25 at the start and end of both six-month trial periods. This questionnaire covers four domains: physical, emotional, social, and home functioning.

The items are scored using a five-point scale. The raw data or statistical analysis of QoL were not available. The QoL did not change significantly during the prednisone period. With every new measurement, however, participants reported a slightly higher QoL, irrespective of the medication given, resulting in a significant improvement in the last measurement on two scales (emotional functioning and the total scale); Beenakker et al considered this to be possibly related to the attention of being involved in a trial.

Assessed child self reported and parent proxy reported quality of life using the Chinese version of PedsQTLM 3.0 NMM (total score). Items are rated on a five-point scale, and transformed linearly to a zero to 100 scale. 'Scores were computed as the sum of items divided by the number of items answered.' Higher scores indicated better quality of life. Twenty-nine boys were too young to complete the questionnaire at baseline, being under seven years old. Clear differences in favour of prednisone 0.75 mg/kg/day were present at six and 12 months in self reported and proxy reported quality of life.

At six months, the MD for the self reported questionnaire was 10.87, 95% CI 0.64 to 21.10, n = 38 and 9.97, 95% CI 1.96 to 17.98, n = 63, for the proxy reported measure. At 12 months, corresponding values were MD 16.05, 6.46 to 25.64, n = 41 and MD 14.42, 95% CI 5.85 to 22.99, n = 58 (; ). (5) Adverse events Adverse events were evaluated by the different investigators as follows. Examined the participants for adverse effects in an area separate from that of clinical evaluation at baseline and at one, two, three, and six months after the start of prednisone treatment. Trialists reported data for both treatment and placebo groups. Examined the participants and interviewed the parents for adverse effects at baseline and at one, two, three, and six months of treatment.

Trialists reported data for both treatment and placebo groups. Did not report adverse effect data. Measured and reported body weight, height, body mass index (BMI) and diastolic blood pressure in prednisone (0.75 mg/kg/day) and control groups at six and 12 months. The report did not provide data on the incidence of other adverse effects for the placebo group; adverse effects occurred in 16 of the 31 children receiving prednisone who completed the 12-month study. Monitored the participants every two months of the study for adverse effects. Trialists reported weight gain data for treatment (deflazacort) and placebo groups, but incidence of the other adverse effects only for the deflazacort group. Asked the parents of participants at the end of the study to report any signs or symptoms that could possibly be related to the treatment.

Evaluated the adverse effects at each visit by using a standard list that described the corticosteroid-related adverse effects. This included patient and parent interview for symptoms and examination for physical signs relating to adverse effects. (a) Weight gain and reported this adverse event as per cent weight gain at last visit above baseline (first visit), on the presumption of six months of treatment. As per cent weight gain was only available as the number of participants in each of a set of intervals on the per cent weight gain scale, we derived the mean and SD for each group assuming each individual had the mid-value of the interval in which they fell. The review authors did not use Sheppard's correction for bias in variances obtained using grouped data because the interval widths were variable and the magnitude of the correction for bias in the SDs was found to be less than 2%.

Analysis of pooled data from and demonstrated a statistically significant weight gain in the prednisone 0.75 mg/kg/day group after six months of treatment as compared to placebo, with a MD of 9.27% (95% CI 6.87% to 11.68%, n = 126; moderate quality evidence) (see; ). Reported that one participant in the prednisone (0.75 mg/kg/day) group showed obvious weight gain at 12 months (placebo group not reported).

However, no clear difference in weight was present between the prednisone and placebo groups at six and 12 months, with wide CIs (; ). Similarly, BMI (kg/m 2) showed no clear difference at six or 12 months, with wide CIs (; ). Compared an intermittent regimen of prednisone (0.75 mg/kg/day given for the first 10 days of every month, for six months) with placebo, in a cross-over design study. They reported the difference and P value in weight as daily rate of change, obtained from a regression using data from a six-month follow-up period. We scaled up the response to 24 weeks (six months) equivalent, and deduced the standard error (SE) from the P values.

The mean weight gain during the prednisone phase (2.37 kg) was greater than in the placebo phase (1.47 kg), but the analysis using the SEs inferred from the quoted P values and the RevMan GIV facility showed that the difference, 0.84 kg (95% CI -0.04 to 1.72), did not quite reach statistical significance (see ). Compared deflazacort with placebo and presented weight gain data for 11 deflazacort and five placebo patients as per cent weight change.

As per cent weight change was only reported as the number of participants in each of a set of intervals on the per cent weight gain scale, the mean and SD for each group were derived as described above for and. After two years of treatment, the degree of weight gain in the deflazacort group was slightly greater than that in the placebo group, but as CIs include the possibility of large effects in either direction (MD 1.09%, 95% CI -13.92 to 16.10, n = 16) (see ), we can draw no conclusions. (b) Behavioural changes Three studies reported the number of patients with behavioural changes in treatment and placebo groups (;; ). Analysis of pooled data from these studies demonstrated a statistically non-significant risk of behavioural changes in the prednisone 0.75 mg/kg/day group after six months of treatment as compared to placebo, with a RR of 1.39 (95% CI 0.94 to 2.06; moderate quality evidence) (see; ). Also compared prednisone 0.3 mg/kg/day with placebo. After six months of treatment there was no statistically significant difference in behavioural changes in the prednisone and placebo groups, with a RR of 1.02 (95% CI 0.67 to 1.56) (see ). Compared prednisone 1.5 mg/kg/day with placebo.

After six months of treatment there was a trend to increased risk of behavioural changes in the prednisone group but this was not statistically significant, with a RR of 1.43 (95% CI 0.92 to 2.24) (see ). Compared an intermittent regimen of prednisone (0.75 mg/kg/day given for the first 10 days of every month, for six months) with placebo, in a cross-over design study. The study reported the number of patients with behavioural side effects (hyperactivity, irritability, euphoria) in prednisone-treated and placebo- treated participants, but data on occurrence of these adverse effects during all four phases of the cross-over trial were not presented or available, and because of this, the review authors could not undertake appropriate statistical analysis. Reported behavioural changes in six of 11 participants in the deflazacort group at six months but did not report the data for the placebo group. (c) Cushingoid appearance Three studies reported the number of participants with cushingoid appearance in treatment and placebo groups (;; ). Compared an intermittent regimen of prednisone (0.75 mg/kg/day given for the first 10 days of every month, for six months) with placebo, in a cross-over design study.

Four participants were noted to have cushingoid appearance during the prednisone treatment period as compared to one in the placebo period. Data on occurrence of this adverse effect during all four phases of the cross-over trial were not presented or available, and because of this, we could not undertake appropriate statistical analysis. Analysis of pooled data from and demonstrated a significant risk of cushingoid appearance in the prednisone 0.75 mg/kg/day group after six months of treatment as compared to placebo, with a RR of 2.37 (95% CI 1.53 to 3.67) (see ). Also compared prednisone 0.3 mg/kg/day with placebo.

After six months of treatment there was no significant difference in cushingoid appearance in the prednisone and placebo groups, with a RR of 1.15 (95% CI 0.60 to 2.17) (see ). Also compared prednisone 1.5 mg/kg/day with placebo, and after six months of treatment there was a significant risk of cushingoid appearance in the prednisone group, with a RR of 4.36 (95% CI 2.04 to 9.33) (see ). Reported a cushingoid appearance in two of 11 participants in the deflazacort group at six months but did not report data for the placebo group. Reported a cushingoid appearance in three of 31 participants in the prednisone (0.75 mg/kg/day) group at 12 months but did not report data for the placebo group. (d) Excessive hair growth (hirsutism) Two studies reported the number of participants with excessive hair growth in treatment and placebo groups (; ).

Analysis of pooled data from these studies demonstrated a statistically significant risk of excessive hair growth in the prednisone 0.75 mg/kg/day group after six months of treatment as compared to the placebo group, with a RR of 2.60 (95% CI 1.47 to 4.60) (see ). Also compared prednisone 0.3 mg/kg/day with placebo.

After six months of treatment there was no significant difference in excessive hair growth in the prednisone and placebo groups, with a RR of 0.73 (95% CI 0.18 to 3.0) (see ). Also compared prednisone 1.5 mg/kg/day with placebo, and after six months of treatment there was a significant increase in the number of boys with excessive hair growth in the prednisone group, with a RR of 2.32 (95% CI 1.16 to 4.64) (see ). Reported excessive hair growth in none of the 11 participants at six months and in three out of eight patients at two years in the deflazacort group, but did not report data for the placebo group.

Reported hair growth in two of 31 participants in the prednisone group at 12 months, but did not report data for the placebo group. (e) Acne Two studies reported the number of participants with acne in treatment and placebo groups (; ). Analysis of pooled data from these studies demonstrated a trend to develop acne in the prednisone 0.75 mg/kg/day group after six months of treatment as compared to placebo but this was not statistically significant, with a RR of 1.78 (95% CI 0.96 to 3.32) (see ). Also compared prednisone 0.3 mg/kg/day with placebo and after six months of treatment there was no significant difference in acne in the prednisone and placebo groups, with a RR of 0.73 (95% CI 0.18 to 3.0) (see ). Compared prednisone 1.5 mg/kg/day with placebo and after six months of treatment there was a trend to develop acne in the prednisone group, but this was not statistically significant, with a RR of 1.77 (95% CI 0.84 to 3.73) (see ). Reported acne in two of 31 participants in the prednisone group at 12 months, but did not report data for the placebo group. (f) Osteoporosis, fractures None of the included studies performed bone densitometry studies.

Two studies instructed the participants in the study to take 0.3 g calcium carbonate with each meal (; ). Two of the included studies commented upon fractures (; ). Reported pathological fracture of the tibia in one participant in the deflazacort-treated group. There was no description of the timing of the fracture in relation to duration of deflazacort treatment, circumstances leading to the fracture, or results of any bone density studies. One participant, randomised to the placebo group in the first phase of, developed a traumatic fracture of the femur 10 days into the study and dropped out. One participant in the placebo treatment group in dropped out of the study because of an arm fracture; subsequently reported this incident.

(g) Hyperglycemia/glycosuria and checked blood glucose, and another two studies checked urine dipstix (; ). Reported glycosuria in one participant, who was on prednisone 0.75 mg/kg/day. The report did not state the severity of glycosuria and its impact. (h) Hypokalemia Only and performed blood tests for electrolyte surveillance. Reported 'mild hypokalemia' in three of 11 deflazacort-treated participants but this was 'easily correctable' with oral potassium supplements. (i) Hypertension reported hypertension with a blood pressure of 130/110 in one participant taking prednisone 0.75 mg/kg/day.

Did not report any hypertension. Monitoring of diastolic blood pressure revealed no statistically significant differences at six or 12 months between the group treated with prednisone 0.75 mg/kg/day and the group receiving placebo. (j) Gastrointestinal side effects Gastrointestinal side effects were defined differently and inconsistently in the included studies. Grouped increased appetite, nausea and stomach discomfort under the umbrella of gastrointestinal symptoms; these, as a whole, were not significantly different between the placebo and prednisone treatment groups. Reported increased appetite as a separate side effect and this was significantly more frequent in the prednisone 0.75 mg/kg/day group as compared to the placebo group (P = 0.02). Reported that in their two-year study, none of the participants developed gastrointestinal disturbances on deflazacort 2 mg/kg on alternate days; they had, however, treated all the children with antacids (drug name not specified). Parents of the participants in, the study of prednisolone 0.35 mg/kg/day, did not report gastrointestinal side effects.

(k) Increased appetite Two studies reported the number of participants with increased appetite in treatment and placebo groups or phases (; ). Compared an intermittent regimen of prednisone (0.75 mg/kg/day given for the first 10 days of every month, for six months) with placebo, in a cross-over design study.

Four of the 16 participants were noted to have increased appetite during the prednisone treatment period as compared to one out of 16 in the placebo period. Data on occurrence of this adverse effect during all four phases of the cross-over trial were not presented or available, and because of this, review authors could not undertake appropriate statistical analysis. Compared prednisone 0.3 mg/kg/day with placebo. After six months of treatment, there was no significant difference in the prednisone and placebo groups with a RR of 1.54 (95% CI 0.90 to 2.62) (see ). Also compared prednisone 0.75 mg/kg/day with placebo. After six months of treatment, there was no significant difference in the prednisone and placebo groups, with a RR of 1.80 (95% CI 1.09 to 2.99) (see ). Reported increased appetite in six of 31 participants in the prednisone (0.75 mg/kg/day) group at 12 months, but did not report data for the placebo group.

(l) Cataracts The participants were evaluated for cataracts in four of the six included studies (;;; ), but the studies did not describe the precise examination (slit lamp or red reflex) performed for detection of cataracts. No cataracts were reported. (m) Death reported two deaths during the study.

A 16-year-old boy died of pneumonia and a four-year-old died during an appendectomy. The authors did not report whether the deaths occurred during the prednisolone or the placebo phases. (n) Life-threatening infections Two studies described specific monitoring to document episodes of intercurrent infection (; ). None of the studies described the treatment strategy for exposure to chicken pox (varicella zoster). Apart from the 16-year-old boy who died of pneumonia described above (), the trials reported no other episodes of infection. (o) Height restriction and stated that they measured height, but presented no data., and did not describe height measurement.

Measured height (cm) at 6 and 12 months, reporting no clear difference in height between prednisone-treated and placebo-treated boys, although the results were imprecise, and allowed for effects in either direction (MD -0.88, 95 CI -6.89 to 5.13, n = 63 at 6 months and MD -2.62, 95% CI -8.66 to 3.42; n = 58, at 12 months) (see; ). Monitored height every 2 months.

By two years, growth was 11.4 ± 2.7 cm in the treated group and 11.2 ± 2.2 cm in the placebo group; however, the report does not state the numbers of boys measured at this time point. Observations on prednisone dose-response relationship and adverse events A full investigation of the prednisone dose-response relationship to identify the optimum dose would need individual patient data within-study analyses, and the included studies reported no such analyses. We consider this further in the.

Two studies made direct comparisons of prednisone doses (; ). • • Forest plot of comparison: 3 Deflazacort versus prednisone, outcome: 3.1 Weight gain (%). The abstract reported weight gain as a percentage of baseline weight at 12 months as follows: deflazacort 0.9 mg/kg = 16.8%, deflazacort 1.2 mg/kg = 18.3%, and prednisone = 26.7%; P.

Implications for practice Randomised controlled trials (RCTs) provide moderate quality evidence that treatment with corticosteroids in Duchenne muscular dystrophy (DMD) compared with placebo improves muscle strength and function, including respiratory muscle strength and function, for six months. There is evidence of continuing benefit on muscle strength and function at one year. On the basis of the evidence available, our confidence in the effect estimate for the efficacy of a 0.75 mg/kg/day dose of prednisolone or above is fairly secure.

Little RCT evidence is available on longer-term effects of corticosteroids versus placebo; one small longer-term RCT found an improvement in muscle strength at two years with deflazacort, with imprecise results on function at two years. Not enough data were available to adequately compare the efficacy of prednisone and deflazacort, although there is very low quality data favouring deflazacort for less weight gain. In the short term (12 months), a weekend-only prednisolone regimen is as effective as daily prednisolone according to low to moderate quality evidence from a single trial. Low quality evidence did not show a difference between the regimens on change in body mass index (BMI). A greater increase in linear height occurred in the weekend-only regimen, but no appreciable difference in other side effects. The long-term benefits and harms of daily corticosteroids or daily versus intermittent regimens are not clear.

Non-randomised studies suggest that clinically significant prolongation of time to loss of ambulation is possible with daily corticosteroids, though potential harms, including weight gain, behavioural changes, vertebral fractures, and cataracts, are significant. Non-randomised studies also suggest there may be a divergence in efficacy between daily and intermittent prednisolone regimens beyond the age of seven years, with greater side effects from daily regimens in the longer term. Implications for research Many issues, including the ideal age or functional stage for initiation of treatment, the optimal corticosteroid type, regimen and dose, strategies for prevention of osteoporosis, and the age for discontinuation of corticosteroid treatment still need to be clarified with RCTs. This will require national and international collaboration, standardised and comparable protocols of assessment, timely publication of studies and the facility of sharing anonymised individual patient data.

While previous studies have focused mainly on muscle strength, walking, and motor aspects, studies are now beginning to address respiratory, cardiac, and quality of life issues; this review or separate Cochrane reviews will examine these outcomes in future. The Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society emphasised the need for studies to be long term to evaluate the effect of corticosteroids on ambulation, respiratory function, cardiac function, and quality of life. There is a need to identify and evaluate strategies to prevent the predictable adverse effects of long-term corticosteroid treatment, particularly excessive weight gain, osteoporosis, and growth retardation. The incorporation of patient and caregiver evaluations of the beneficial and adverse effects of treatment, as additional outcome measures, should be considered. The impact of corticosteroid therapy on quality of life of the patient and the family, in relation both to benefits and adverse effects, should also be evaluated. Acknowledgements We are grateful to Professor C Angelini, Dr MM Rahman, and Professor AB Taly for provision of data and Dr SM Todorovic, Dr DM Escolar, Professor E Svanborg, Dr M Brooke and Dr EAC Beenakker for their communications.

The authors would like to thank Miss Tracey Elliot who was our statistician co-author on the protocol for the review. Help from Kate Jewitt, then Cochrane Neuromuscular Disease Review Group Co-ordinator, and Professor Richard Hughes, then Co-ordinating Editor, was instrumental in completing the first version of the review (). We are grateful to the Association Francaise contre les Myopathies (AFM) for a bursary to the Cochrane Neuromuscular Disease Group to support a previous update of this review (). Editorial support from the Cochrane Neuromuscular Disease Group for the 2008 update was funded by the TREAT NMD Network European Union Grant 036825. This work was supported by the National Institute for Health Research (NIHR) under its Cochrane Incentive Award scheme (award number 14/175/34 to Cochrane Neuromuscular). This project was also supported by NIHR via Cochrane Infrastructure funding to Cochrane Neuromuscular. The views and opinions expressed herein are those of the review authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, National Health Service (NHS), or the Department of Health.

Cochrane Neuromuscular is also supported by the Medical Research Council (MRC) Centre for Neuromuscular Diseases. Data and analyses. Comparison 1. Corticosteroids versus placebo Outcome or subgroup title No. Of studies No.

Comparison 2. Weekend-only versus daily prednisone Outcome or subgroup title No. Of studies No. Comparison 3. Deflazacort versus prednisone Outcome or subgroup title No. Of studies No. Date Event Description 16 February 2016 New citation required and conclusions have changed Review authors expanded the scope of the review at this update to include comparisons of corticosteroids and dosing regimens. We included three trials comparing different corticosteroids or dosing regimens and one new published trial comparing corticosteroid and placebo.

We included two previously excluded abstracts that met selection criteria, as this is current practice. 16 February 2016 New search has been performed Search updated to February 2016. Tony Swan and Mike Pike withdrew from authorship; Ruth Brassington joined the authors. Review authors updated the methodology and assessed all studies using the current Cochrane 'Risk of bias' tool. We added 'Summary of findings' tables. 26 May 2008 Amended Converted to new review format.

14 November 2007 New citation required and conclusions have changed We updated the searches of the Neuromuscular Disease Trials Register (August 2006), MEDLINE (July 2007), EMBASE (August 2006), CINAHL (August 2006) and LILACS (August 2006). We identified one randomised controlled trial which fulfilled the inclusion criteria.

Another new randomised controlled trial was identified, but did not meet the inclusion criteria, and is described in this update. Twelve new non-randomised studies were identified, and are tabulated and discussed in this update. Contributions of authors AM wrote the first draft of the original review, selected studies, assessed methodological quality and extracted the data, which the Review Group Co-ordinator checked. TK selected studies and assessed their quality. AS gave statistical advice and helped with inference of data.

All four authors (AM, TK, MP, AS) approved the final text. For this update EM, AM and TK selected new studies. EM and RB assessed risk of bias, extracted data and drafted additional sections of the review.

RB entered outcome data into RevMan, which EM checked. FJ provided statistical advice. TK and AM provided advice and commented on the draft.

Declarations of interest Dr Emma Matthews has no conflicts of interest. Dr Ruth Brassington is Managing Editor of Cochrane Neuromuscular, of which The National Institute for Health Research (NIHR) is the largest single funder. The NIHR provided an incentive award to Cochrane Neuromuscular for the updating of this review (see Acknowledgements).

A grant from the Motor Neurone Disease Association to Cochrane Neuromuscular contributed to her salary in 2011-2015. She has no financial conflicts of interest. She withdrew from the later stages of the editorial process of this review. Dr Thierry Kuntzer has no conflicts of interest. Fatima Jichi has no known conflicts of interest. Dr Adnan Y Manzur, at the time of preparation and submission of the protocol for this review was the principal investigator of a proposed UK multicentre trial of prednisolone in Duchenne muscular dystrophy.

However, this trial was not funded. Currently, Dr Manzur is the lead clinician of the UK North Star Clinical Network for Neuromuscular Disorders. The clinicians on this clinical network have a consensus on approach to use of corticosteroids (prednisolone) and plans for future collaboration to audit and modify clinical practice in line with available evidence. Sources of support Internal sources • Ruth Brassington, UK. Employed as Managing Editor of Cochrane Neuromuscular by University College London Hospitals (UCLH) NHS Foundation Trust. Her work on this review was supported by NIHR under its Cochrane Incentive Award scheme (award number 13/175/49) and through Cochrane Review Group Infrastructure funding to Cochrane Neuromuscular • Fatima Jichi, UK.

UCL School of Life & Medical Sciences, Joint Research Office, University College London, London, UK External sources • Emma Matthews, UK. This work was supported by NIHR under its Cochrane Incentive Award scheme (award number 13/175/49) Differences between protocol and review Tony Swan and Mike Pike withdrew from authorship at this 2016 update; Ruth Brassington joined as an author. At this update, we extended the scope of the review to include comparisons of corticosteroids and of dosing regimens. We added quality of life and pulmonary function as outcome measures at a previous update and updated the methods in this version of the review accordingly. We revised the objectives to reflect this change and to better reflect specified outcomes. We added additional adverse events to those specifically listed in the Types of outcomes. We updated the methods section according to Methodological Expectations of Cochrane Intervention Reviews (MECIR) guidance.

We used the current Cochrane 'Risk of bias' tool and included 'Summary of findings' tables. We extended the searches to clinical trials registries. We used a random-effects meta-analysis throughout, regardless of the presence of heterogeneity.

Characteristics of studies Characteristics of included studies [ordered by study ID]. Angelini 1994 Methods Randomised double-blind trial. Randomisation followed 2:1 scheme Participants 28 boys with DMD, all ambulant at entry into the trial DMD proven by dystrophin or DNA studies Mean age: • deflazacort 98.65 ± 13.70 months • placebo 96.55 ± 15.96 months Interventions Deflazacort 2 mg/kg on alternate days for 2 years (n = 17) or placebo (n = 11) Outcomes Age at loss of ambulation, age at loss of ability to rise from floor, MRC index from 4 muscles Monitoring of: weight and height every 2 months; blood pressure; WBC; RBC and haematocrit; plasma glucose; CPK and ions. ECG and x-rays of chest and hand for bone age at beginning and end of treatment. Assessment for cataracts every 2 years Declarations of interest Not stated Funding sources A grant from Telethon, Italy Notes Dates: not reported Location: Italy Ethical approval and consent procedures not described Risk of bias Bias Authors' judgement Support for judgement Random sequence generation (selection bias) Low risk Stated to be randomised “randomization followed a 2:1 scheme. At the beginning of the trial the patients in each arm of the study, both in the drug and placebo group, were similar for motor function. At the beginning of trial, the two groups had the same age, MRC index, and functional grades” Allocation concealment (selection bias) Unclear risk No information provided Blinding of participants and personnel (performance bias) All outcomes Low risk Placbo-controlled On balance, judged to be of low risk although “Blinding and maintenance of blinding during trial was possible since only the coordinator, but not the examiner, had the key of randomization.

It is possible that, during prolonged treatment, blinding was destroyed by the appearance of side effects of the drug” Blinding of outcome assessment (detection bias) All outcomes Low risk “Blinding and maintenance of blinding during trial was possible since only the coordinator, but not the examiner, had the key of randomization. It is possible that, during prolonged treatment, blinding was destroyed by the appearance of side effects of the drug” Incomplete outcome data (attrition bias) All outcomes High risk “During the 4 years of our study (1 year of natural history and 3 years of actual drug administration trial) lack of compliance was seen in 5 placebo and 4 drug-treated patients” and 11 deflazacort and 6 placebo participants dropped out for other reasons. Authors state “lack of significance in some tests may be due to dropout of DMD patients” Selective reporting (reporting bias) Unclear risk Data reporting comprehensive in tables at 6, 12 and 24 months, but not specified in detail in methods Other bias Low risk None identified. Bonifati 2000 Methods Double-blind, randomised, multicentre, equivalence study Participants 19 boys with DMD (1 not included in evaluations as he received both drugs, each for 6 months) Inclusion criteria: diagnosis confirmed by dystrophin immunohistochemistry, age over 5 years, preserved ability to ambulate independently, and no previous steroid therapy. Griggs 1991 Methods Randomised, double-blind trial with 2 treatment groups and 1 placebo group Participants 99 boys with DMD, age range 5 to 15 years.