ADHD: When Storebø et al. (2015) was published, it was criticized for being "flawed" and "misleading" by numerous other scientists, so I decide to "do a deep-dive" to see if Storebø et al. (2023) have improved their "very low reporting quality".
ADHD: When Storebø et al. (2015) was published, it was criticized for being “flawed” and “misleading” by numerous other scientists, so I decide to “do a deep-dive” to see if Storebø et al. (2023) have improved their “very low reporting quality”.
In 2015 Storebø et al. (2015) was criticized for being both “flawed and undertaken without sufficient scientific justification” and “extremely dubious from an ethical perspective” and rated by Dr. Dr. Tobias Banaschewski et al. (2016) as having “underestimated efficacy, overestimated risks, and uncertainty on the part of clinicians and patients.”. Seen in that light, I decided to do as he suggested in his commentary back then – “Trust, but verify …”.
Let’s see what Storebø et al. (2023) have for us, this time …
Key messages
‐ Methylphenidate might reduce hyperactivity and impulsivity and might help children to concentrate. Methylphenidate might also help to improve general behaviour, but does not seem to affect quality of life.
‐ Methylphenidate does not seem to increase the risk of serious (life‐threatening) unwanted effects when used for periods of up to six months. However, it is associated with an increased risk of non‐serious unwanted effects like sleeping problems and decreased appetite.
‐ Future studies should focus more on reporting unwanted effects and should take place over longer periods of time.
What is attention deficit hyperactivity disorder (ADHD)?
ADHD is one of the most commonly diagnosed and treated childhood psychiatric disorders. Children with ADHD find it hard to concentrate. They are often hyperactive (fidgety, unable to sit still for long periods) and impulsive (doing things without stopping to think). ADHD can make it difficult for children to do well at school, because they find it hard to follow instructions and to concentrate. Their behavioural problems can interfere with their ability to get on well with family and friends, and they often get into more trouble than other children.
How is ADHD treated?
Methylphenidate (for example, Ritalin) is the medication most often prescribed to children and adolescents with ADHD. Methylphenidate is a stimulant that helps to increase activity in parts of the brain, such as those involved with concentration. Methylphenidate can be taken as a tablet or given as a skin patch. It can be formulated to have an immediate effect, or be delivered slowly, over a period of hours. Methylphenidate may cause unwanted effects, such as headaches, stomachaches and problems sleeping. It sometimes causes serious unwanted effects like heart problems, hallucinations, or facial ‘tics’ (twitches).
What did we want to find out?
We wanted to find out if methylphenidate improves children’s ADHD symptoms (attention, hyperactivity) based mainly on teachers’ ratings using various scales, and whether it causes serious unwanted effects, like death, hospitalisation, or disability. We were also interested in less serious unwanted effects like sleep problems and loss of appetite, and its effects on children’s general behaviour and quality of life.
What did we do?
We searched for studies that investigated the use of methylphenidate in children and adolescents with ADHD. Participants in the studies had to be aged 18 years or younger and have a diagnosis of ADHD. They could have other disorders or illnesses and be taking other medication or undergoing behavioural treatments. They had to have a normal IQ (intelligence quotient). Studies could compare methylphenidate with placebo (something designed to look and taste the same as methylphenidate but with no active ingredient) or no treatment. Participants had to be randomly chosen to receive methylphenidate or not. We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.
Just to make sure that we keep track of all the little details of this massive 900+ page report, I’ve made some summary tables to help us keep the overview
Unbeknownst to me, I somehow get different averages than Storebø et al. (2023) but that is most likely due to me using Excel to recalculate their Comparison Groups results, and since I am not a statistician, I most probably made some minor mishaps. Nonetheless, the numbers I present here are all based on Storebø et al. (2023) own data from their Cochrane publication, and they are for informational purposes only anyways.
(SMD: -0,74 = Moderate)
ADDspeaker interpretation: will most often improve ADHD symptoms, but it is varying a lot across age groups: 0-6 years: SMD: -0,50 (Moderate), 7-11 years: SMD: -1,77 (Large), 12-18 years: SMD: -0,48 (Small)
(RR: 1,98 = Large)
ADDspeaker Interpretation: will not likely cause serious adverse events, and if so, it is at a very low probability, and mainly hypertension. (MPH: 8 of 1000 vs. NoMPH: 6 of 1000)
(RR: 1,37 = Large)
ADDspeaker Interpretation: will most often cause non-serious adverse events, but AEs are mainly decreased appetite and increased sleep problems). (MPH: 53,8% vs. NoMPH: 43,7%)
(SMD: -0,59 = Moderate)
ADDspeaker Interpretation: will most likely improve general behavior, depending on which rater is used: Teachers: SMD: -0,66 (Moderate), Independent: SMD: -0,79 (Moderate), Parent: -0,55 (Moderate), Subgroup: SMD: -0,42 (Small)
(SMD: 0,42 = Small)
ADDspeaker Interpretation: – will most often improve Quality of Life, although it varies much across which rating scale used: CHQ: SMD: 0,54 (Moderate), CGAS: SMD: 0,79 (Moderate), CHIP-CE:PRF: SMD: 0,64 (Moderate), KINL-R: SMD: -0,32) (Small (negative))
What did we find?
We found 212 studies with 16,302 children or adolescents with ADHD. Most of the trials compared methylphenidate with placebo. Most studies were small with around 70 children, with an average age of 10 years (ages ranged from 3 to 18 years). Most studies were short, lasting an average of around a month; the shortest study lasted just one day and the longest 425 days. Most studies were in the USA.
Based on teachers’ ratings, compared with placebo or no treatment, methylphenidate:
‐ may improve ADHD symptoms (21 studies, 1728 children)
‐ may make no difference to serious unwanted effects (26 studies, 3673 participants)
‐ may cause more non‐serious unwanted effects (35 studies, 5342 participants)
‐ may improve general behaviour (7 trials 792 participants)
‐ may not affect quality of life (4 trials, 608 participants)
Our confidence in the results of the review is limited for several reasons. It was often possible for people in the studies to know which treatment the children were taking, which could influence the results. The reporting of the results was not complete in many studies and for some outcomes the results varied across studies. Studies were small and they used different scales for measuring symptoms. And most of the studies only lasted for a short period of time, making it impossible to assess the long‐term effects of methylphenidate. Around 41% of studies were funded or partly funded by the pharmaceutical industry.
How up to date is this evidence?
This is an update of a review conducted in 2015. The evidence is current to March 2022.
Methylphenidate may improve attention deficit hyperactivity disorder (ADHD) symptoms and general behaviour in children and adolescents with ADHD aged 18 years and younger. We rated the evidence to be of very low certainty and, as a result, we cannot be certain about the magnitude of the effects from the meta‐analyses. The evidence is limited by the serious risk of bias in the included trials, underreporting of relevant outcome data, and a high level of statistical variation between the results of the trials. There is also very low‐certainty evidence that methylphenidate causes numerous adverse events. The risk of serious adverse events seems low, but data were only available from 43 of the 212 included trials. It is also problematic that only 93 of the 212 included trials reported on specific and overall non‐serious adverse events. Accordingly, we cannot rule out the possibility that non‐serious harms are more prevalent than reported in our review.
If methylphenidate treatment is considered, clinicians might need to use it for short periods, with careful monitoring of both benefits and harms, and cease its use if no evidence of clear improvement of symptoms is noted, or if harmful effects appear. A problem is that clinicians very often rely on their assessment of methylphenidate in their clinical evaluation. Arguments like “I know that this medication helps” can be problematic when they are based on anecdotal evidence and case reports. A new review found that clinicians had difficulties in assessing benefits or harms following treatment, with inaccuracies in both directions (Hoffmann 2017). Clinicians mostly underestimated instead of overestimated harms and overestimated rather than underestimated benefits. Inaccurate perceptions of the benefits and harms of treatments are likely to result in uncertain clinical management choices (Hoffmann 2017).
This review highlights the urgent need for long‐term, high‐quality, and large randomised clinical trials (RCTs), at low risk of bias, to investigate the benefits and harms of methylphenidate treatment versus placebo in children and adolescents with ADHD. Such trials ought to be designed according to the SPIRIT (Standard Protocol Items: Recommendations for Intervention Trials) guidelines (Chan 2013), and reported in keeping with the CONSORT (Consolidated Standards of Reporting Trials) standards Moher 2010). Pre‐published protocols could help reduce the inconsistent measurement of benefits and harms caused by the use of many different rating scales and by lack of assessment of adverse events.
The important issue of protecting blinding of these trials needs to be addressed urgently. Immediate measures could be implemented to improve blinding. Having independent, blinded assessors monitor adverse effects, whilst separate, independent blinded assessors measure efficacy, is likely to reduce the risk of unblinding due to adverse effects. Active placebos need to be sought and are likely to be important in the future, but their development is still at the very early stages. Research in this field should be strongly supported, but it is likely to take many years before such substances can be used safely and ethically in research with children and adolescents. The prevalent use of cross‐over trials needs to be reconsidered as they usually only provide short‐term interventions, which can limit the assessment of benefits and harms. However, we were not able to identify major differences when comparing parallel‐group trials with cross‐over trials.
Future trials ought to publish depersonalised individual participant data and should report all outcomes, including adverse events, to ensure that future systematic reviews and meta‐analyses can access and use individual participant data. Only through meta‐analyses will we be able to assess differences between intervention effects according to age, sex, comorbidity, ADHD subtype, and dose. Reviews show that many different rating scales are used for children with ADHD. Consistent use of well‐validated scales is needed, as is a country‐wide adverse events reporting system, such as the Food and Drug Administration, to increase awareness of adverse events. In addition, the findings in this review clearly show the urgent need for large RCTs to investigate the efficacy of non‐pharmacological treatments. As with RCTs, systematic reviews of such trials assess average effects in groups of individuals. Such average effects may comprise strong benefits for a single participant or a few participants and no effect or negative effects for others. Despite more than 50 years of research in this field, we have no knowledge on how to identify patients who may obtain more benefits than harms. Individual patient data meta‐analyses are needed to identify such patient characteristics. Therefore, it would be extremely helpful for review authors to gain full access to anonymised individual participant data for inclusion in meta‐analyses examining these data (Gluud 2015). Patient subgroups may benefit from intervention if those with reduced rates of adverse events can be identified. This personalised medicine approach can be used for discovering predictors and moderators for treatment response. The use of biomarkers for both more precise diagnoses and for more precise assessment of treatment response is necessary in future RCTs. The use of enrichment designs will improve statistical power for biomarker analyses (Buitelaar 2022).
We considered 34 trials to be at low risk of bias for random sequence generation, four trials to be at high risk of bias (Connor 2000; Green 2011; Heriot 2008; Tannock 2018), and 18 at unclear risk of bias.
We considered 26 trials to be at low risk of bias for allocation concealment (often because medications and packaging were identical in appearance for blinding purposes) and two trials to be at high risk of bias (Barragán 2017; Green 2011). Twenty‐eight trials did not report allocation concealment in sufficient detail to allow us to make a judgement, so were at unclear risk.
We considered 64 trials to be at low risk for random sequence generation, nine trials at high risk of bias (Carlson 1995; Fitzpatrick 1992a; Kaplan 1990; Kelly 1989; Manos 1999; McBride 1988a; Szobot 2008; Tirosh 1993b; Wigal 2014), and 84 trials at unclear risk of bias.
We considered 57 trials to be at low risk of bias for allocation concealment (often because medications and packaging were identical in appearance for blinding purposes), five trials at high risk of bias (Carlson 1995; Fitzpatrick 1992a; Szobot 2008; Ullmann 1985; Wigal 2014), and 95 trials did not sufficiently report allocation concealment so we judged them at unclear risk of bias.
We considered that 40 trials adequately described their method of blinding of participants and personnel so we judged them to be at low risk of bias. Ten trials gave insufficient information about their methods so we judged them to be at unclear risk of bias (Biederman 2003b; Childress 2009; Childress 2017; Childress 2020b; Childress 2020c; Findling 2010; Greenhill 2006; Lin 2014; Tucker 2009; Wigal 2017). Six trials were not blinded (Barragán 2017; Brown 1985; Duric 2012; Jensen 1999 (MTA); NCT00409708; Perez‐Alvarez 2009), so we judged them to be at high risk of bias.
We considered that 34 trials adequately described their method of blinding of outcome assessment so we judged them to be at low risk of bias. Seventeen trials gave insufficient information about their methods and were considered to be at unclear risk of bias. Six trials did not include blinded outcome assessors (Barragán 2017; Brown 1985; Duric 2012; Jensen 1999 (MTA); NCT00409708; Perez‐Alvarez 2009), so we judged them to be at high risk of bias.
We judged that 118 trials adequately described their method of blinding of participants and personnel so we therefore judged them at low risk of bias. Thirty‐one trials were considered at unclear risk of bias. Eight trials were not blinded (Lopez 2003; Manos 1999; Pearson 2013; Pelham 2014; Ramtvedt 2013; Stein 2003; Ullmann 1985; Wigal 2013), so we judged them to be at high risk of bias.
We judged that 86 trials adequately described their method of blinding of outcome assessment and were therefore considered at low risk of bias, and 66 trials were considered at unclear risk of bias. Five trials did not include blinded outcome assessors (Cox 2006; Douglas 1986; Manos 1999; Wigal 2013; Wodrich 1998), so we judged them to be at high risk of bias.
Thirty trials adequately addressed incomplete data and were considered at low risk of bias. Twenty trials did not so we judged them to be at high risk of bias. Six trials gave insufficient information for us to assess whether the method they used to handle missing data was likely to bias the estimate of effect (Childress 2020b; Coghill 2013; Firestone 1981; NCT02293655; Palumbo 2008; Wigal 2004), and we therefore considered them at unclear risk of bias.
Eighty‐five trials adequately addressed incomplete data so we judged them to be at low risk of bias. Forty‐eight trials gave insufficient information to assess whether the method they used to handle missing data was likely to bias the estimate of effect so we judged them to be at unclear risk of bias. Twenty‐four trials had incomplete outcome data and were therefore considered at high risk of bias.
Thirty‐four trials reported all pre‐defined or otherwise expected outcomes so we judged them at low risk of bias. Five trials did not (Childress 2020b; Greenhill 2006; Lin 2014; NCT02293655; Wigal 2015), and these were considered at high risk of bias. In 17 trials it was unclear whether trial authors reported all pre‐defined or otherwise expected outcomes (Arnold 2004; Barragán 2017; Biederman 2003b; Brown 1985; Butter 1983; Findling 2006; Firestone 1981; Greenhill 2002; Heriot 2008; Horn 1991; Ialongo 1994; Schachar 1997a; Szobot 2004; Tannock 2018; Tucker 2009; Wigal 2004; Wolraich 2001), so we judged them at unclear risk of bias.
Forty‐seven trials reported all pre‐defined or otherwise expected outcomes so we judged them at to be at low risk of bias. Thirteen trials did not so these were considered at high risk of bias (Castellanos 1997; Chacko 2005; CRIT124US02; Froehlich 2018; Gonzalez‐Heydrich 2010; Gorman 2006; Hawk 2018; Huang 2021; McInnes 2007; NCT02536105; Stein 2003; Taylor 1993; Wallace 1994). In 97 trials it was unclear whether trial authors reported all pre‐defined or otherwise expected outcomes so we judged them at unclear risk of bias.
We identified no other potential sources of bias for either parallel or cross‐over trials.
The reviewers calculated a standardized mean difference. These are hard to interpret clinically but rules of thumb in their interpretation suggest that
Storebø et al. (2023)
0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect.
Researchers analyzed trials according to the person assessing the ADHD symptoms (teacher, independent assessor, or parent).
Children/adolescents had lower symptom scores with methylphenidate than with placebo/no treatment, regardless of who undertook the assessment.
Subgroup analyses by medication status (naive or not), duration of treatment (≤ 6 months, > 6 months), or dose (low, high, unknown) showed similar results to the main analysis.
We were able to combine data on ADHD symptoms from 47 parallel‐group trials and 82 cross‐over trials, of which five also provided first‐period data.
Parallel‐group trials and cross‐over trials (end of first‐period data only)
A meta‐analysis showed a difference in effects between methylphenidate and placebo on teacher‐rated ADHD symptoms favouring methylphenidate (SMD −0.74, 95% CI −0.88 to −0.61; I² = 38%; 21 trials, 1728 participants; Analysis 1.1). The SMD of −0.74 for ADHD symptoms corresponds to a mean difference (MD) of −10.58 points (95% CI −12.58 to −8.72) on the ADHD Rating Scale (DuPaul 1991a). This is an effect above the minimal relevant difference (MIREDIF; Zhang 2005).
Most independent assessors were clinicians.Parallel‐group trials and cross‐over trials (end first‐period data only)
A meta‐analysis suggested there was a difference in effect between methylphenidate and placebo on independent assessor‐rated ADHD symptoms favouring methylphenidate (SMD −1.10, 95% CI −1.44 to −0.77; I² = 95%; 22 trials, 3724 participants; Analysis 2.1). The SMD effect of −1.10 for ADHD symptoms corresponds to an MD of −15.7 points (95% CI −14.7 to −7.9) on the ADHD‐RS (DuPaul 1991a). This is a clinical effect above the MIREDIF (Zhang 2005). Five trials reported change from baseline scores (Findling 2008; McCracken 2016; Newcorn 2008; Newcorn 2017a (flexible dose); Newcorn 2017b (forced dose)), but removing these trials did not significantly change the estimate. Two trials were outliers as they reported unrealistically high effect sizes. These were: Kollins 2021 and Wigal 2017. Removing these trials showed a SMD effect of −0.62 (95% CI −0.79 to −0.46). The SMD effect of −0.62 for ADHD symptoms corresponds to an MD of −8.86 points (95% CI −11.3 to −6.6) on the ADHD‐RS (DuPaul 1991a), which is a clinical effect above the MIREDIF (Zhang 2005).
Parallel‐group trials and cross‐over trials (end of first‐period data only)
A meta‐analysis suggested there is a difference in effects between methylphenidate and placebo in parent‐rated ADHD symptoms favouring methylphenidate (SMD −0.63, 95% CI −0.76 to −0.50; I² = 58%; 27 trials, 2927 participants; Analysis 3.1). The SMD effect of −0.63 for ADHD symptoms corresponds to an MD of −9.0 points (95% CI −10.9 to −7.0) on the ADHD‐RS (DuPaul 1991a). This is a clinical effect above the MIREDIF (Zhang 2005). Three trials in the meta‐analysis reported change from baseline scores (Carlson 2007; Newcorn 2008; Tucker 2009), but removing these trials did not significantly change the estimate.
Cross‐over trials (end of last period data)
A meta‐analysis suggested a difference in effect between methylphenidate and placebo in parent‐rated ADHD symptoms favouring methylphenidate (SMD −0.70, 95% CI −0.86 to −0.55; I² = 84%; 45 trials, 4971 participants; Analysis 3.9).
Parallel‐group trials and cross‐over trials (end of first period) and cross‐over trials (end last of last period)
When combining data from parallel‐group trials with endpoint data from cross‐over trials our meta‐analysis suggested a difference in effects between methylphenidate and placebo on reduced parent‐rated ADHD symptoms favouring methylphenidate (SMD −0.67, 95% CI −0.78 to −0.56; I² = 79%; 69 trials, 7838 participants; Analysis 3.11).
We were able to include in our analyses data on general behaviour from 13 parallel‐group trials and from 21 cross‐over trials.
Parallel‐group trials and cross‐over trials (end‐of‐first‐period data only)
A meta‐analysis suggested a difference in effect between methylphenidate and placebo in teacher‐rated general behaviour favouring methylphenidate (SMD −0.62, 95% CI −0.91 to −0.33; I² = 68%; 7 trials, 792 participants; Analysis 9.1). The SMD effect of −0.62 for general behaviour corresponds to an MD of −3.58 points (95% CI −5.26 to −1.91) on the CGI (Conners 1998a). Due to a lack of MIREDIF, we do not know whether this is a clinical relevant difference.
We assessed the evidence to be of very low certainty (see GRADE assessment below). Therefore we are uncertain that the estimated effect accurately reflects the true effect and the addition of more data could change the findings.
Cross‐over trials (endpoint data)
Meta‐analysis suggested a difference in effects between methylphenidate and placebo in reduced teacher‐rated general behaviour favouring methylphenidate (SMD −0.75, 95% CI −0.87 to −0.63; I² = 5%; 16 trials, 1302 participants; Analysis 9.6).
Parallel‐group trials and cross‐over trials (endpoint data)
When combining data from parallel‐group trials with endpoint data from cross‐over trials our meta‐analysis similarly suggested a difference in effects between methylphenidate and placebo in reduced teacher‐rated general behaviour favouring methylphenidate (SMD −0.72, 95% CI −0.84 to −0.60; I²= 37%; 23 trials, 2094 participants; Analysis 9.8). The intervention effect did not vary according to high or low risk of vested interest (Analysis 9.9).
Parallel‐group trials and cross‐over trials (first‐period data only)
Meta‐analysis suggested a difference in effects between methylphenidate and placebo in reduced independent assessor‐rated general behaviour favouring methylphenidate (MD 1.10, 95% CI −1.01 to 3.21; 1 trial, 94 participants; Analysis 10.1).
Cross‐over trials (endpoint data)
Meta‐analysis suggested a difference in effects between methylphenidate and placebo in reduced independent assessor‐rated general behaviour favouring methylphenidate (SMD −0.98, 95% CI −1.39 to −0.57; I² = 87%; 9 trials, 987 participants; Analysis 10.2).
Parallel‐group trials and cross‐over trials (endpoint data)
When combining data from parallel‐group trials with endpoint data from cross‐over trials our meta‐analysis similarly suggested a difference in effects between methylphenidate and placebo in reduced independent assessor‐rated general behaviour favouring methylphenidate (SMD −0.86, 95% CI −1.27 to −0.46; I² = 89; 10 trials, 1081 participants; Analysis 10.4).
Parallel‐group trials and cross‐over trials (first‐period data only)
Meta‐analysis suggested a difference in effects between methylphenidate and placebo in reduced parent‐rated general behaviour favouring methylphenidate (SMD −0.42, 95% CI −0.62 to −0.23; I² = 57%; 10 trials, 1376 participants; Analysis 11.1).
Cross‐over trials (endpoint data)
Meta‐analysis suggested a difference in effects between methylphenidate and placebo in reduced parent‐rated general behaviour favouring methylphenidate (SMD −0.84, 95% CI −1.05 to −0.63; I² = 0%; 6 trials, 384 participants; Analysis 11.6)
Parallel‐group trials and cross‐over trials (endpoint data)
When combining data from parallel‐group trials with endpoint data from cross‐over trials our meta‐analysis similarly suggested a difference in effects between methylphenidate and placebo in reduced parent‐rated general behaviour favouring methylphenidate (SMD −0.56, 95% CI −0.74 to −0.39; I² = 59%; 16 trials, 1760 participants; Analysis 11.8).
We could include data on quality of life from only four parallel‐group trials in our analyses. We assessed the evidence to be of very low certainty (see GRADE assessment below).
There was no difference in effects between methylphenidate versus placebo in quality of life at end of treatment (SMD 0.40, 95% CI −0.03 to 0.83; I²= 81%; 4 trials, 608 participants; Analysis 13.1). The SMD of 0.40 for quality of life corresponds to an MD of 4.94 (95% CI −0.37 to 10.25) on the Child Health Questionnaire (CHQ; Landgraf 1998), which ranges from 0 to 100 points. This is below the MIREDIF of 7.0 points on CHQ (Rentz 2005).
We included 55 parallel‐group trials and 156 cross‐over trials in this review and one trial with a parallel‐phase (114 participants randomised) and a cross‐over phase (165 participants randomised). Altogether, these trials randomised more than 16,000 participants and were reported in 614 publications. The majority compared methylphenidate with placebo in short‐term trials of less than six months’ duration. The average trial duration in the 56 parallel trials was 67.1 days (range 1 to 425 days). Most were conducted in outpatient clinics in high‐income countries, particularly the USA. Participants’ ages ranged from 3 to 18 years across most studies; in two studies ages ranged from 3 to 21 years (Green 2011; Szobot 2008). Both boys and girls were recruited, in a ratio of 3:1.
We considered 22 trials (9 parallel‐group trials and 13 cross‐over trials, including the two phases in Kollins 2006 (PATS)) to have an overall assessment at low risk of bias. We considered 191 trials to have an overall assessment at high risk of bias. We considered all trials to be of high risk of bias due to the risk of deblinding described elsewhere. This raises important concerns, which are discussed after a summary of the results.
A meta‐analysis of data from parallel‐group trials combined with data from the first period of cross‐over trials suggests that methylphenidate may improve ADHD symptoms as reported by teachers. The SMD calculated modest improvement in ADHD symptoms on the ADHD‐RS scale (DuPaul 1991a), however, we judged the certainty of the evidence to be ‘very low’ (see Quality of the evidence).
We found that the types of scales used influenced the intervention effect of methylphenidate. The differences between scales ranged from 0.5 SMD to 2.0 SMD. We found lower effects of methylphenidate in long‐term trials compared to short‐term trials. There was no difference between the subgroups of trials using placebo compared to the trials using no intervention in the control group.
Methylphenidate does not appear to be associated with an increased occurrence of serious adverse events. However, data for this outcome were only available in 42 of the 212 included trials (20%) and we judged the certainty of the evidence to be ‘very low’ (see Quality of the evidence).
Amongst those in the methylphenidate‐exposed groups, 538 per 1000 experienced non‐serious adverse events, compared with 437 per 1000 in the control group. The most common non‐serious adverse events were sleep problems and decreased appetite.
We also judged the overall certainty of the evidence for this outcome to be ‘very low’, and as a result, we are uncertain of the magnitude of the harmful effects. Furthermore, for methodological reasons, we used only dichotomous outcomes reflecting the number of participants affected by the event per the total number of participants. As most participants reported more than one adverse event, the actual increase in risk of non‐serious adverse events may very well be higher than the 23% calculated.
Meta‐analyses of data from only seven parallel‐group trials indicated that methylphenidate was associated with an improvement in children’s general behaviour, as reported by teachers. We cannot state anything for sure about the clinical importance of this SMD value. Comparable findings emerged from meta‐analyses of cross‐over trials (endpoint data) as reported by teachers, and from meta‐analyses of nine cross‐over trials (endpoint data) as rated by independent assessors. We also judged this evidence to be of ‘very low’ certainty (see Quality of the evidence).
Meta‐analyses of data from only four parallel‐group trials indicated that methylphenidate was associated with no improvement in children’s quality of life as reported by parents and clinicians. We judged the certainty of the evidence to be ‘very low’ (see Quality of the evidence).
The very low certainty of the evidence, as assessed using the GRADE approach, undermines the confidence that can be placed in the magnitude of any effect. In particular, the prevalence of non‐serious adverse events raises questions about the effectiveness of blinding in these trials. If blinding was broken in just 20% or 30% of participants given methylphenidate, the resulting bias might well account for the small but statistically significant findings concerning the possible benefits of methylphenidate (Coghill 2021; Storebø 2015a).
This review highlights two major issues concerning the overall completeness and applicability of the evidence of the benefits and harms of methylphenidate for children with ADHD: the dearth of trials conducted in children and adolescents in low‐ and middle‐income countries, and the lack of follow‐up beyond six months. Here, we focus on the impact on the applicability of findings of decisions taken as part of this review (choice of rater for assessing change in ADHD symptoms and quality of life, choice of dose and diagnosis), together with issues relating to rating scales, diagnostic criteria, choice of comparators and adverse events.
We chose to use teacher‐rated outcomes as the primary measure for both ADHD symptoms and general behaviour, although a number of trials used or relied on parent reports. Some researchers have argued that parent evaluations of ADHD symptoms may not be as reliable as those of other raters such as teachers of pre‐school children (Murray 2007), or college students (Lavigne 2012). For example, Caye 2017 suggests inconsistency in ratings between parents, and in the MTA trial (MTA 1999b), information provided by parents was not always thought to be strong (Efstratopoulou 2013). We tested the robustness of our decision by conducting subgroup analyses and found no significant differences between this score and those of other raters.
Importantly, we do not really know what a lower score on an ADHD symptom scale (like that reported in this review) means for a child’s quality of life and ability to live, learn and function with other people.
The present systematic review has many strengths. We developed a protocol for this review according to instructions provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022a), and this protocol was published before we embarked on the review itself. We conducted extensive searches of relevant databases, and we requested published and unpublished data from pharmaceutical companies manufacturing methylphenidate, including Takeda Pharmaceuticals, Medice (represented in Denmark by HB Pharma), Janssen‐Cilag, Novartis, Rhodes Pharmaceuticals, Ironshore Pharmaceuticals and Pfiizer. Two review authors, working independently, selected trials for inclusion and extracted data. We resolved disagreements by discussion with team members. We assessed risk of bias in all trials according to the recommendations provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We conducted Trial Sequential Analyses to control the risk of type I errors and type II errors and to estimate how far we were from obtaining the DARIS to detect or reject a certain plausible intervention effect (CTU 2022). In the meta‐analyses on non‐serious adverse events, the Trial Sequential Analysis showed that observed intervention effects were not likely to be due to type I error and confirmed that sufficient data had been obtained.
Although we added new search terms to the strategy, we limited the search to the period since the previous search (2015 onwards), so it is possible that we did not capture pre‐2015 records containing the new search terms. However, we believe that we are unlikely to have missed any important trials because of our supplementary searches, which included identifying studies through reference checks of relevant reviews, and contacting pharmaceutical companies.
We excluded 126 trials described in 144 reports, which assessed the effects of methylphenidate on specialised outcomes (e.g. experimental, neurocognitive or functional outcomes) in children or adolescents with ADHD (see Characteristics of excluded studies). This raises the issue of bias in our review process as we did not write to these authors asking whether they collected data on other outcomes. This potential bias, however, is not likely to change our conclusions.
Over the past 20 years, several published systematic reviews and narrative reviews have examined the efficacy of methylphenidate for ADHD (with or without meta‐analysis). All of them described methylphenidate as being very helpful to children and adolescents with ADHD. However, they each had methodological shortcomings.
The 2015 version of this review contradicted earlier published reviews as we reported that methylphenidate may improve teacher‐reported ADHD symptoms, teacher‐reported general behaviour, and parent‐reported quality of life among children and adolescents diagnosed with ADHD, but the low quality of the evidence meant that we could not be certain of the true magnitude of these effects (Storebø 2015a). The 2015 version of this review provoked many critical responses, published in articles, letters to editors and blogs. We responded to all of the comments; for more details please see the section “Why it is important to do this review“. An overview article found 24 eligible systematic reviews and meta‐analyses published after our 2015 review (Ribeiro 2021). The results from the overview also showed that the evidence was uncertain due to its low quality. Additionally, this overview highlighted the underreporting of adverse events in RCTs, and concluded that evidence supporting methylphenidate being beneficial in the treatment of children and adolescents with ADHD remains uncertain (Ribeiro 2021).
Our current updated review confirms our 2015 findings with 29 additional RCTs included (Storebø 2015a). These results differ from the large network meta‐analysis by Cortese and colleagues (Cortese 2018), in which the use of methylphenidate in children and adolescents was strongly supported by the evidence where they compared the efficacy and tolerability of methylphenidate for ADHD to placebo alongside other medications (Cortese 2018). We published a letter to the editor of the Lancet in which we highlighted several problems with their review, namely, the exclusion of many relevant trials in order to fulfil statistical and methodological assumptions that they made. While the pooled comparison for clinician‐rated effects of methylphenidate versus placebo for children was rated as “moderate quality of evidence”, they also assessed all of the indirect comparisons as being of “low to very low quality of evidence”. Indirect evidence differentiates network meta‐analyses from conventional meta‐analyses. Given the decreased interpretability of the indirect comparisons, there are no novel findings in this network meta‐analysis (Faltinsen 2018a; Storebø 2018a). In sum, the part of the network meta‐analysis that is different from our review published in 2015 (Storebø 2015a), consists of evidence of low to very low certainty.
In a response to our letter, the authors confirmed that they had excluded 65% of the trials that we had included in our 2015 review. They excluded 51 trials that had under seven days of treatment, 38 cross‐over trials without pre‐crossover data or a washout, 18 trials with responders to previous treatment, and 14 trials where treatment was not monotherapy (Cipriani 2018). They did this because including these trials would have been a clear violation of their published protocol and would have compromised the transitivity of the network meta‐analyses (Cipriani 2018).
Catalá‐López and colleagues published a large systematic review with network meta‐analyses in 2017 (Catala‐Lopez 2017). They included 190 RCTs with a total of 26,114 children and adolescents with ADHD, and found that stimulant monotherapy was significantly more efficacious than placebo; however, all analyses were assessed in GRADE at “low or very low certainty”. The authors of the review concluded that stimulants may improve the symptoms of ADHD, but the strength of the underlying evidence remains uncertain (Catala‐Lopez 2017).
Padilha and colleagues published a network meta‐analysis investigating the benefits and harms of different types of ADHD medication (including methylphenidate) for children and adolescents with ADHD, including 48 trials with 4169 participants (Padilha 2018). The review found that there were beneficial effects of methylphenidate on the Clinical Global Impressions Improvement scale (CGI‐I) and that methylphenidate was more effective than the non‐stimulants atomoxetine and guanfacine (Padilha 2018). There are several methodological problems with this review which we commented on in a letter to the editor (Faltinsen 2019). The issues we raised were focused on selection bias (as the authors had excluded placebo‐controlled trials), the fact that authors judged the methodological quality of the included trials to be good, asserting that they were well designed, reported, and conducted, even when this clearly was not the case and that they did not include an overall assessment of certainty such as the GRADE system. They also included cross‐over trials without reporting the method as to how they pooled this data with that from parallel‐group trials and they failed to discuss other possible issues, such as carry‐over and period effects (Faltinsen 2019). Furthermore, they did not assess the transitivity assumption in their network meta‐analyses (Faltinsen 2018b; Faltinsen 2019).
A review by Cerrillo‐Urbina and colleagues investigating the benefits and harms of stimulants and non‐stimulants included 15 RCTs, with 4648 children or adolescents, or both, from 6 to 17 years of age diagnosed with ADHD (Cerrillo‐Urbina 2018). Only four trials assessed methylphenidate, all of which were conducted before 2013. The GRADE assessment of the evidence concerning the total score of ADHD symptoms was assessed to be “moderately high quality of evidence” for both stimulant and non‐stimulant medications. They downgraded the quality of evidence by one level due to a high degree of heterogeneity in the pooled results (I² > 75%) but did not downgrade it further for risk of bias or publication bias, even though they found that there was significant publication bias for all outcomes. It is striking that this review only included four trials on methylphenidate, whereas we found 184 trials in our 2015 review covering the same period (Storebø 2015a).
A network meta‐analysis by Li and colleagues found that methylphenidate was beneficial in the treatment of ADHD in children and adolescents (Li 2017). Methylphenidate was considered the second safest treatment compared to the other ADHD medications. The review included 62 trials in a meta‐analysis, which included 12,930 participants. They did not make any attempt to evaluate the risk of bias or the certainty of evidence, which significantly lowers the robustness and validity of this review.
The NICE guideline recommends methylphenidate as the first‐line pharmacological treatment for children over five and adolescents, “1.7.7: Offer methylphenidate (either short or long acting) as the first‐line pharmacological treatment for children aged 5 years and over and young people with ADHD” (NICE 2018). The NICE guideline committee concluded that methylphenidate and lisdexamphetamine provide clinically important benefits to patients with ADHD as compared to placebo and other drugs (NICE 2018). We found several methodological problems in the NICE ADHD guidelines as we believe they conducted an erroneous assessment of the certainty of the included studies. They assessed the quality of meta‐analysis to be “ high quality”, when it could be strongly argued that it was, in fact, “low quality”. In their assessment of the effect of methylphenidate, they included only 16 trials that focused solely on immediate and osmotic‐release methylphenidate in children and adolescents. We included 185 trials (175 of which were placebo‐controlled) in our 2015 review (Storebø 2015a). NICE did not adjust for multiple comparisons and did not discuss the concern that all the data arose from short‐term follow‐up (NICE 2018).
The American Academy of Pediatrics guideline was updated in 2019 based on patients’ age (Wolraich 2019). With regard to preschool children, the guideline recommends evidence‐based behavioural interventions (behavioural parent training or behavioural classroom interventions, or both) as the first‐choice treatment. Methylphenidate may be considered when a child has moderate to severe problems with functioning and if the behavioural treatment does not provide the necessary improvements. With regard to school children, the guideline strongly recommends pharmaceutical treatments (FDA‐approved medications for ADHD) together with behavioural interventions. Regarding adolescents, the guideline strongly recommends pharmaceutical treatment and if possible, evidence‐based behavioural interventions. The guideline states that there is a strong effect observed in the trials investigating the effects of stimulant medications (Wolraich 2019). For the comparison of pharmacological treatments versus placebo or usual care, the review only identified eight articles representing seven studies. The review concluded that there was limited additional evidence concerning FDA‐approved ADHD medications compared with placebo or usual care across all outcomes in this updated systematic evidence review. The conclusions regarding methylphenidate, therefore, seem overly positive. The risk of harm is considered as low and the benefits, in general, are described as outweighing the risks.
Our current updated review is in line with two recent Cochrane systematic reviews of methylphenidate in adults. These two reviews found low‐ or very low‐certainty evidence that methylphenidate, compared with placebo, improved ADHD symptoms (Boesen 2022; Candido 2021).
Storebø, O. J., Storm, M. R. O., Pereira Ribeiro, J., Skoog, M., Groth, C., Callesen, H. E., Schaug, J. P., Darling Rasmussen, P., Huus, C. L., Zwi, M., Kirubakaran, R., Simonsen, E., & Gluud, C. (2023). Methylphenidate for children and adolescents with attention deficit hyperactivity disorder (ADHD). The Cochrane database of systematic reviews, 3(3), CD009885. https://doi.org/10.1002/14651858.CD009885.pub3
This website uses cookies.
Tilføj din kommentar her - Feedback er altid velkomment!