Professor Samuele Cortese Hello, everyone, my  name is Samuele Cortese. I’m an NIHR Research   Professor, Professor of Child Adolescent  Psychiatry at the University of Southampton in   the UK, also, Professor of Child Neuropsychiatry  at the University of Bari in Italy, and Adjunct   Professor of Child Psychiatry with NYU in the  United States. My talk today will focus on the   “Evidence on the Pharmacological Treatment  for ADHD.” Before starting, these are my   disclosures of possible conflicts of interest,  and so, let’s get started with the evidence. And, first of all, let’s have a look at  the current recommendations from the NICE   guidelines on the management of ADHD. The most  recent ones were published in 2018/19 and,   since then, the evidence has  not changed significantly. So,   when we consider the recommendation in  school-age children and adolescents with ADHD,   NICE recommends first group-based ADHD-focused  support, which basically is psychoeducation,   and then the recommendation is to go straight to  medication for ADHD with a specific hierarchy. So, NICE recommends to start first with  methylphenidate and, if this is not effective,   or it is effective but not well tolerated, then  to move to lisdexamfetamine, or an alternative   formulation of amphetamine, like dexamphetamine,  if lisdexamfetamine is not well tolerated or not   effective. And if stimulants, so  methylphenidate or amphetamines,   are not effective or not well tolerated, then  the recommendation is to move to a non-stimulant,   namely atomoxetine or guanfacine. And in  another video I will address the evidence on the   non-pharmacological treatments for ADHD, so  I will skip this part for the time being. Right, so, why did NICE recommend this  hierarchy? Well, actually, as you may know,   NICE is very rigorous in examining all the  existing evidence and complementing the evidence   with expert recommendation when the evidence is  not enough, or is lacking. So, at the time when   NICE was completing these recommendations,  we, as the European ADHD Guidelines Group,   published a very large network meta-analysis of  randomised controlled trials of ADHD medications. So, I’m sure some of you, at least, will  be familiar with network meta-analysis. So,   basically, a network meta-analysis is a specific  type of meta-analysis, where we can compare   two or more treatments, even if they have not  been compared head-to-head in the individual   trials that are included in the meta-analysis.  And the way a network meta-analysis result is   represented is, as you can see here, via nodes,  which represents the treatment, and via lines,   which connects the node. Every time there  is a line connecting two nodes, it means   that there is at least one randomised trials  comparing head-to-head those two treatments. So, if we look at the plot representing  the results of our network meta-analysis,   on the left hand side, you can see the results  for children and adolescents. And you can see   here that for some type of treatments there  are actually trials comparing them directly,   but, for instance, if we want to know  how atomoxetine compares to guanfacine,   there are not direct trials. And so,  we need to rely on the results of the   network meta-analysis, which under certain  methodological assumptions are valid. So, what we found – actually,  this is a very complex analysis,   so I’m including in the slide just a  figure summarising some of the results. So,   let me guide you through this figure. On the  Y axis, you have the tolerability, which is   the number of individuals who drop out from the  trials due to side effects. And on the X axis,   you have the efficacy, which is the reduction of  the symptomatology, the severity of ADHD symptoms,   measured typically with one of the rating  scales we also use in clinical practice. So, each dot represent the effect, the effect size  for that particular medication, you see here the   legend, and each effect size is – counts alongside  its confidence interval, because, of course,   when we – in a meta-analysis, we give the  measure of an effect size, we also calculate the   confidence interval which is, as you may remember  from statistic, that range of value within which   95% of the time that the result may fall. So, if we look, for instance, at the effect   of the amphetamines in terms of efficacy, this is  here, and you can see really well on the scale,   this is around one, which is very high  effect size, and this is compared to placebo,   which is set as zero. So, the more on the  left side the effect size are for efficacy,   the better it is, because it means that the  medication is more efficacious than placebo. And, actually, you can see here that the  medication with the highest effect size   in times of efficacy is actually amphetamine.  The effect is here, as I said, it’s around one,   but the confidence interval means that the  real effect may be somewhere here or here. So,   amphetamines are the most efficacious  according to our network meta-analysis,   which, as you can see, includes data from more  than 10,000 individuals, children with ADHD. However, when we assess a medication,  we don’t assess just the efficacy,   we look also at the tolerability. And, in that  case, as you can see here, all – many medications   are less well tolerated than placebo because  their effect size and the confidence interval   is below this line of reference of placebo. And  certainly this is the case for amphetamines, but,   also, this is the case, for instance,  for guanfacine, you see here all the   confidence interval is below the line of zero,  in times of tolerability compared to placebo. However, you see here that  methylphenidate, you see here,   its confidence interval in terms of tolerabilita  – tolerability, crosses the line of zero. So,   this means that, actually, when we  compare methylphenidate and placebo,   in terms of tolerability in children,  there is no significant difference. So, to sum up, we have a situation whereby if we  look at the efficacy, amphetamines are the most   efficacious, but they are also less well tolerated  than placebo. While if we look at methylphenidate,   which is in green here, its efficacy is  slightly lower compared to amphetamines,   as you can see here, we are around 0.8 here, in  terms of efficacy, but, interestingly, at least   in children, that tolerability, the  tolerability of methylphenidate is not   significantly different compared to placebo.  So, when we balance efficacy and tolerability,   we may consider that methylphenidate is actually  – should actually be the first choice, because,   as I said, it’s just slightly less efficacious  than amphetamines, but it is better tolerated. Now, in this slide, I have reported  the effect size which are measured in   meta-analysis with a measure which is  called “standardised mean difference,”   and you may remember from statistics  that values of effect size between   zero and point – 2.3 highlights just a  very small effect. So, statistically,   the treatment is better than the control,  but, clinically, the effect is quite small. Values of the effect size between 0.4 and  0.6 generally are considered moderate,   and values higher than .7, .8, are considered high  effect size, so, clinically very tangible. And,   as you can see here, once again, in terms of  efficacy, amphetamines are the highest one,   as I mentioned, around one, followed  by methylphenidate. The one with the   lowest effect size, in terms of efficacy, is  atomoxetine, 0.56, which is still within the   range of moderate. So this is comparable, for  instance, to the effect size of antidepressants. And I would like just to highlight the fact that  actually when we compare the effect size of these   medications for ADHD, in particular of stimulants,  to the effect size of other medications used in   general medicine, we actually find out that  the effect size of stimulants is quite good,   not only in psychiatry, but in general medicine.  Here, and in this figure, you have the effect size   in psychiatric drugs in blue, and at the top,  actually, we have amphetamines, around one,   as I mentioned. And here in white we have the  effect size of medications used in other areas   in medicine, and with the exception of a few  medications here at the top, you can see that   stimulants are quite high compared, also,  to many other medications used in medicine. So, this means that, at least in  the short-term, because this data   importantly derive from short-term randomised  controlled trials, and when I say short-terms,   we are talking about a few weeks, a few months,  at least in the short-term, these medications   are the most effective in psychiatry and,  also, among the most effective in medicine. Do they work just for ADHD core symptoms? The  result – the values I presented on efficacy   referred to the efficacy, the reduction  in the severity of ADHD core symptoms,   so, inattention hyperactivity, impulsivity. The  question is, do they work also on other outcomes,   on other aspects? For instance, we may wonder, do  stimulants, methylphenidate in particular, work   well also in terms of improving neuropsychological  function, so, for instance, executive dysfunction? So, the answer, according to  this meta-analysis is, yes,   but, as you can see here, the effect size, the  magnitude of the effect, is quite lower compared   to the effect on the core symptoms of ADHD,  which means that actually if we address these   issues just with medication, we may still  have residual problems. So, for instance,   as you can see, the effect of methylphenidate on  working memory is quite low, It is significant,   statistically significant, it is better than  the control condition, but it is not high. A similar conclusion can be applied to the effect  of these medications on the quality of life.   This is interesting, because, of course, more  and more, we are considering measures other   than the core symptoms of ADHD. Core symptoms  are important, but, of course, in the life of a   child or an individual with ADHD, there are many  other aspects. An important aspect, of course,   to consider quite crucial is the quality of  life. So the question is, do these medication   improve also globally the quality of life of that  person and not just the severity of the symptoms? We conducted recently a meta-analysis to answer  this question, and what we found is that, yes,   medications for ADHD, both stimulants,  as you can see here, and non-stimulants,   in particular, atomoxetine, do improve  quality of life, But, once again,   the effect is lower, and you can see here the  effect size, compared to the effect on ADHD   symptom severity. And, interestingly, there  is not much difference between stimulants   and non-stimulants, in regard to  their impact on quality of life. As I mentioned, all the trials that we  have included in the analysis that I   showed you so far are a short-term trial and,  of course, it is difficult, if not impossible,   to conduct longer randomised trials.  So a trial lasting years or decades,   because of economic reasons, of course,  but, also because of ethical reasons.   It is not possible to randomise  somebody for years to a placebo,   when the active treatment turns out to be better,  much better, than the placebo in the short-term. So, how can we address the question as to whether  these medications are efficacious and effective,   also in the long-term? Because this is what  many patients and their family, rightly so,   wants to know. Well, there is a specific type  of design of trial which is called “withdrawal   randomised controlled trial,” or “discontinuation  trial,” whereby individuals who have been on the   same medication for months or years, they  are randomised to continue that treatment   or they are randomised to go to placebo  control, in order to see if there is still   a separation after years of treatment between  the active treatment and the placebo control. There are a few of these trials in the field  of ADHD medications, and these slides report   one of these studies, conducted in the  Netherlands. And to cut a long story short,   you can see on the right hand side here that  during this withdrawal trial, there was still a   significant separation between the methylphenidate  in this case and the placebo. So those who were   continued to methylphenidate had an improvement,  an advantage, compared to those on placebo. But, once again, the size of the effect is  lower compared to the effect size we have in   the short-term, and there may be a number of  methodological issues that may explain these   results. So, I will say that,  in relation to the question,   what kind of evidence do we have in relation to  the long-term effect of medications? I will say,   first of all, it is difficult  to answer this question. Second,   we do have some studies that show that,  at least in part, medications for ADHD   maintained their effect also in the longer-term,  but we don’t know exactly the magnitude of this. So far, I have reported evidence  from a randomised controlled trial,   which is the core, start and design that we  need to assess the efficacy and tolerability   of a treatment. But the randomised controlled  trials do have limitations. Usually they don’t   include typically the – all the patients we  see in real life, daily clinical activity,   and they have limitations in terms of the type  of outcomes that are included. For instance,   there are important outcomes, and we may wonder,  what is the effect of medications, for instance,   on injuries, traumas? When you say trauma,  it’s physical trauma, a motor vehicle accident,   criminality, suicidality, substance use disorder,  bipolar symptoms, psychosis or seizures. Usually, at least some of these outcomes  are not routinely included in randomised   controlled trials, so we need to look at  evidence, observational evidence. However,   of course, in observational studies,  meaning without randomina – randomisation,   the problem is that the lack of randomisations  does not allow us to infer cause-effect   relationship. We can see just a correlation,  but we cannot be sure that the effect we   see are due to the treatment. They may also  happen because of other confounding factors. So, I’m going to present, however, here, in the  next slide, two type of studies that tackle or,   at least in part, address this issue of lack  of randomisation. The first is the so-called   “within-individual design observational study.”  So, basically in this observational study we   follow and we measure the severity of the  outcome within the same individual when   they are and when they are not on medication. So  we can see – we can appreciate the difference. As you can see in this graph, and this is the line  of reference, every time the effect size is on the   left hand side, and the confidence interval, it  means that this particular outcome is better when   the person is on medication, compared to when they  are not on medication. So, as you can see here,   we can conclude that based on the analysis of  these within-individual observational studies,   when somebody is on medications for ADHD, in  particular stimulants, there is a reduction   on the frequency and the severity of physical  injuries and traumas, motor vehicle accident,   at least, in females – in males, sorry, and  criminality, clear reduction, suicidality,   at least according to one study, reduction in  substance use disorder, reduction in depression. There is a worsening of bipolar and  manic symptoms without mood stabilisers,   but when the individual is taking mood  stabiliser, then adding a stimulant leads   to a further decrease in the severity of manic  symptoms. And there is no impact on the rate of   psychosis, and there is a reduction,  interestingly, in terms of seizures. The other type of design which is helpful to  address an important outcome in the absence   of a proper randomised controlled  trial is what is called – and it’s   quite a recent advance in the field,  “emulated – target emulated trial.”   So this is an observational study which  allow us to simulate in a way a trial. We published recently with some  colleagues from Sweden an important   study showing that when somebody  is treated with ADHD medications,   in particular stimulants and methylphenidate,  for a couple of years, we can notice a decrease   in the rate of all-cause mortality,  and also unnatural-cause mortality,   while there is no effect and this is quite  understandable, on natural-cause mortality. So, I think this is a very important piece  of evidence showing the impact of these   medications in terms of mortality, hence, a very  important outcome in terms of public health. I   think that was pretty much everything and,  of course, there is a large body of evidence   on ADHD medications, but I tried to select the  most important information for you. Thank you.