9

Addiction, Reward, and ADHD

There are two features of ADHD that can cause alarm. Firstly, the very nature of the symptoms seen in ADHD means that the person is more likely to be a risk taker, sensation seeker, or an experimenter in search of stimulation. Such characteristics mean there is a potential for recreational drug use, and, indeed, the epidemiological studies indicate that there is a higher incidence of substance misuse amongst those with ADHD. Secondly, and perhaps the one that produces the most angst, the pre­dominant treatment for ADHD is a controlled substance. Methylphenidate is pharmacologically similar to cocaine, a well-known addictive drug, and can be seen as a pharmakon: a substance that is both cure and poison [1020].

The question arises: “Will taking methylphenidate increase the likelihood of illegal drug consumption?” This is the essence of the so-called “gateway hypothesis” – one drug leads to another, more problematic drug. We therefore have the potential of treating ADHD, but the potential risk of causing addiction.

Coupled together with these concerns about ADHD is the more academic pursuit focusing on the notion that those with ADHD may somehow have faulty reward circuits [1021–1022].

Addiction

Addiction is an interesting problem to study in the context of ADHD: both can be called impulse control disorders (see [1023]) and both have dopaminergic imbalances [688, 1024–1026].

The DSM-IV does not use the term “addiction” but the term “substance use disorder” (SUD), plus substance abuse and substance dependence. In fact we can get into the same quandary over the diagnosis of addiction as we do with ADHD.

Addiction, dependence, abuse, are often used interchangeably by people, adding to the confusion. However, the terms can reflect different problems and levels of severity. According to Altman et al., addiction is the extreme or psychopathological state where control over drug use is lost; dependence is the state of needing a drug or to operate within normal limits; abuse is the use of drugs which leads to problems for the individual; and use/misuse is any non-medical consumption [1027].

Given that there are a number of drugs to which one can become addicted, the question of common mechanisms has been addressed.

The drugs that we choose to take, for whatever reason, differ considerably in their pharmacology, e.g. nicotine is very different from heroin. The differing pharmacology of these drugs would at first lead one to assume that there are many different mechanisms that lead to addiction.

The work of Professor Sagvolden has indicated that there is a deficient reward processing in spontaneous hypertensive rats (SHRs) [1028–1029] that is similar to what he has seen in ADHD children, where they prefer small immediate rewards [1030–1031]. The common denominator for mediating addiction and the rewarding effects of drugs has long been assumed to be DA in particular within the mesolimbic system, comprising the ventral tegmental area (VTA) and the nucleus accumbens (NAcc).

Methylphenidate acts similarly to amphetamine and cocaine, indicating a common mechanism for all three [1032].

Not only is the pharmacology of methylphenidate similar to cocaine (and amphetamine, too, is prescribed for ADHD), it also produces a conditioned place preference (CPP), in which rats prefer to spend time in an environment they associate with a drug – a measure of addiction [1033–1037]. However, in mice that have been specifically bred not to have DAT (the DAT knockout mouse), methylphenidate surprisingly still produced a CPP [1038]. Given that the main mechanism by which methylphenidate works has been genetically removed, then one must presume that methylphenidate’s actions in this paradigm are mediated elsewhere, e.g. noradrenaline, or that there are compensatory mechanisms that may develop in DAT knockout mice.

With CPP studies, the age of the animal is important, where adolescent pre-exposure to methylphenidate reduces the magnitude of subsequent CPP induced by cocaine [1039–1040] or nicotine [1036], but not morphine [1041]. Clearly this is of concern; however, the reason for this effect it is unknown. One possibility may be that it is developmentally determined, as Crawford et al. [1041] gave the methylphenidate at an earlier age than the previous studies with psychostimulants.

Despite all the negative hyperbole surrounding addiction, it has generated research that has proven useful in understanding reward in ADHD.

The textbook case for mesolimbic DA as the reward pathway now points to the involvement of the VTA and the NAcc in the learning of associations between predictive stimuli and reward, and the chance of receiving a reward. This is interesting in the context of ADHD, in which there may be a failure to learn associations between events and future reward.

Much of the work that contributes to our new understanding of dopamine and reward has been conducted on animals. The focus has been on the phasic DA¹ response linked to reward (see [1042–1044] for reviews). When an animal is presented with a primary reinforcer (reward), e.g. food or water, DA is released. The same is also true for stimuli that are associated with the reward [1045–1047]. DA neurons are depressed when a signaled reward is omitted or by stimuli predicting the absence of reward [1048]. This phasic DA responses is argued to be a teaching signal involved in the learning of associations [1042].

DA neurons project to various brain regions, including the dorsal and ventral striata (both of which have independent actions on learning with transfer of new learning in the ventral regions to automated process in the dorsal regions [1049]) and subregions of the prefrontal cortex. The ventral striatum has recently been shown to have a reduced volume in ADHD [1050]. The prefrontal neurons carry signals related to the preparation of movement and goal achievement [1051] and the motivational value of rewards [1052]. Imaging studies have indicated that there is a dysfunction in cortical regions of the brain that are linked to drug compulsion and a lack of behavioral inhibition, which in the case of addiction is not being able to stop taking the drug [1053].

An interesting feature of the phasic response to reward-associated stimuli is that it occurs when rewards are different from predictions – the reward prediction error [1046]. The phasic DA response differs if a reward is unpredicted, not available, or delayed [1054]. Thus the DA codes for: (1) an unpredicted reward, which elicits an activation – a positive prediction error; (2) a predicted reward, which elicits no response; and (3) the omission or extended delay of a predicted reward, which induces a depression [1042]. In the words of Professor Wolgang Schultz, “A ‘prediction error’ message may constitute a powerful teaching signal for behavior and learning [and] it may contribute to the self-organization of goal-directed behavior” [1055] (p. 293).

This is all well and good for rats and monkeys, but does the reward prediction error occur in humans, and is it differentially affected in those with ADHD? It has been demonstrated that stimuli associated with drugs such as cocaine can increase brain activity [1056–1059]. Indeed a recent study using methylphenidate to elevate DA indicated that DA alone does not produce drug craving, but requires cues associated with the drug [1060].

The use of neuroimaging techniques has revealed that there is a case for the reward prediction error in humans (see [1061] for review). Very little research has been conducted into the reward prediction error in ADHD. Given that the dopamine system in those with the disorder is different from those without it, then it is possible that reward processing and reinforcement learning is also different in ADHD. The evidence suggests that reward and reinforcement is different in ADHD, but only one study to date has looked at the reward prediction error, which was not evident in the children with ADHD [1062]. These children were all ADHD combined type and off medication at the time of the study. In the future it would be interesting to see what the effects of ADHD treatments are on this measure and how subtypes differ.

The new direction that addiction research is taking with regard to DA will be important as it takes us away from reward per se and focuses on the learning about reward contingencies.

Addiction in ADHD

The question that is frequently reviewed in the literature is whether childhood ADHD is predictive of future substance use. It is a simple question to ask; the answer is rather more difficult to obtain.

Retrospective studies in drug users suffer from unreliability, whereas cross-sectional studies do not provide the information necessary for a conclusive answer [1063]. There are two ways to look at the problem: (1) look at drug users for symptoms of ADHD; and (2) look at drug use in those with ADHD.

Early reports suggest that many drug users seeking treatment had a previous history consistent with ADHD [1064–1075]. In one review of the evidence, an estimated 20 percent of substance abusers had symptoms consistent with ADHD [1071]. However, these studies cannot be used as conclusive evidence supporting a causal role of ADHD in substance abuse; the methodologies used do not permit that [1063, 1076]. Asking people, especially drug users, to recall past events is not a sturdy basis for any study. They do, however, add some circumstantial evidence, which will require further careful study. Retrospective studies looking at ADHD and substance abuse indicate a link with substance use [1071, 1077–1080]. However, the case for a common substrate of ADHD and alcoholism has not been supported [1081], whereas it is supported for nicotine use [1082].

Longitudinal prospective studies that follow a cohort of people with ADHD through childhood, adolescence, and into adulthood help address questions about substance use disorder (SUD). Such studies are expensive and obviously take a long time. Testing can occur throughout the period of investigation so that you can get samples or snapshots of their functioning at different ages. Such studies indicate an increased likelihood of substance abuse amongst those with ADHD [1083–1090], especially late adolescents [1091]. In a study of 1,142 cases of ADHD it was found that severity of childhood inattention symptoms predicted substance use outcomes and that childhood Oppositional Defiant Disorder/Conduct Disorder (ODD/CD) symptoms predicted drug use. Persistence of ADHD and adolescent CD were associated with increased substance use behaviors relative to controls [1092]. Again methodology gets in the way of being able to pronounce a definitive conclusion; such studies are not always representative of all those with ADHD – they are also highly likely to have received some treatment which may include methylphenidate. Treatment itself requires consideration (see below).

In a ten-year follow-up of ADHD cases, it was found that there was a greater risk of substance abuse as well as other comorbidities (e.g. anxiety) [176]. Comorbidities may also increase the vulnerability to SUD. Nicotine was found to be the main drug currently used by people with ADHD, which prompted the conclusion that smoking is a gateway drug to alcohol and illegal substances [1093].

Nicotine as a gateway drug to other substances is not unique to ADHD – it can affect all children [1094–1095]. Starting with nicotine, the young person with ADHD may move next to alcohol, then to cannabis, and finally to other illicit substances [1096]. Maintenance of smoking in those with ADHD may be increased because they experience more withdrawal symptoms than their non-ADHD counterparts [1097], suggesting that avoidance of withdrawal could maintain smoking behavior, or that they are self-medicating.

To avoid sampling bias, it has been suggested that studies should take place in representative populations and be studied for the symptoms of ADHD a priori [1076]. Studies using this methodology indicate that there is a higher rate of substance abuse in those with ADHD [1098–1100], but that is was associated more with conduct problems than inattention [1099]. Early use of alcohol was also associated with ADHD [1101].

Comorbidity is also a problem that needs to be disentangled. Substance abuse was associated more with conduct problems [1099], which indicates a possible influence of CD and ODD etiology. According to one study, the vast majority also met the criteria for CD and ODD [1102], hence it is difficult to ascribe causality to ADHD alone. Comorbid CD has been demonstrated to increase the severity of the substance abuse with a greater variety of substance [1103].

The role of CD and ODD in substance abuse may be greater for adolescents and adults as the prevalence is higher in older age groups. Some authors have stated that the role of ADHD is small [1076], whereas other do not consider this to be the case [1063]. A study of Brazilian adolescents found an association between ADHD and SUD without CD mitigating the two [1104]. When it comes to multiple comorbidities, it has been noted that those who have ADHD and substance abuse disorder are more likely to have mood and anxiety disorders as well compared to if they only had ADHD or SUD [1105].

Another variable that should be considered is peer group affiliation. It is an integral part of folklore that if you fall in with the wrong crowd, then your destiny will be influenced by the group’s norms. This is also the case for those with ADHD: deviant peer group membership mediates the relationship between ADHD and substance abuse, thus making membership of such groups a high risk for those with ADHD [1106].

Drug taking may be a form of self-medication, and we need to know what symptoms the individual might be treating. This will facilitate the understanding of the common processes in ADHD and substance abuse, and also highlight what symptoms are important to the individual and the impact they have on their quality of life.

The question of which symptoms and subtype are more likely to give rise to substance abuse needs addressing. For example, the ADHD-H subtype emphasizes impulsivity; substance abuse is also regarded as an impulse control problem [1057, 1107]. The ADHD-H subtype has been shown to increase the likelihood of substance use, especially if combined with aggression [1108]. Hyperactive problems predicted both occasional and frequent drunkenness [1109], and cognitive functioning mediates the connection between ratings of hyperactivity and drinking habits in non-ADHD individuals [1110]. Good performance on tests of executive function were associated with positive outcomes [1111], whereas those performing poorly were at a high risk of substance use [1112].

Are ADHD Subtypes Linked to Any Particular Drugs?

ADHD symptoms were linked to alcohol and marijuana use with the ADHD-I subtype, but not with the ADHD-H subtype, which was associated with marijuana and nicotine dependence [1113–1114]. Other studies have found many different drugs been consumed [1115].

When it comes to smoking cigarettes, the same was true of the ADHD-I symptoms and ADHD-H [1114, 1116]. In other studies, by contrast, it was the ADHD-H set of symptoms that was predictive of the initiation of drug taking [1117]; of smoking [1118]; and of substance abuse in general [1119]. The picture of which symptoms are linked to later substance use has been shown in those with nicotine dependence (ND), where “ADHD-I symptoms were associated with ND symptoms acceleration in adolescence, but slowing acceleration in young adulthood, whereas ADHD-H symptoms were associated with ND symptoms acceleration in young adulthood” [1120] (p. 563). The severity of ADHD symptoms was also associated with participants choosing methylphenidate over placebo, but this was linked more to self-medication rather than substance abuse [1121].

One possible hypothesis is that those with ADHD are actually self-medicating their symptoms with nicotine, cocaine, and other drugs. The self-medication hypothesis sees the person as not trying to get high, but rather trying to avoid unpleasant experiences [1122]. Anecdotal evidence often supports a self-medication view; after all, cocaine is pharmacologically similar to methylphenidate. Even nicotine has been evaluated as an ADHD treatment. Thus, those with ADHD may use some drugs to calm down; to be able to think more clearly and to concentrate; to reduce anxiety or depression; and to relieve boredom. Even adults with ADHD will choose methylphenidate over placebo to control their symptoms, and this is independent from other measures of abuse [1121]. Although internet videogaming is not a drug, a South Korean study has suggested that it, too, may be used as a form of self-medication in ADHD [1123], which is tentatively supported by the increase of DA in the striatum seen during gaming [1124].

The drug of choice amongst those with ADHD appears to be nicotine, with 36 percent using drugs for self-medication purposes and 25 percent to get high [1125]. However, no difference emerged when data were analyzed between ADHD and control groups, and the symptoms did not differ between those self-medicating and those getting high [1125].

Using the theory of BI as the cause of ADHD (e.g. [412]), is impulsivity being self-medicated? Some have proposed a role in ADHD for acetylcholine, which is the neurotransmitter system with which nicotine interacts [857, 1126–1127]. The effects of nicotine have been shown to mitigate some of the deficits of BI [857]. One has to remember to separate the positive effects of nicotine from the negative effects of smoking. Experimental studies do not use smoking as a route of delivery; injections or patches are more routine.

The use of nicotine in disorders such as ADHD has been suggested [1128], and an early report has shown some benefit from a nicotine patch [859]; however, children had difficulty with some of the unpleasant side-effects of nicotine. If some people with ADHD are self-medicating with nicotine, what is the next stage in the process? Do they self-medicate with alcohol? Why drink alcohol? The pharmacology would not suggest this is a useful medication; alcohol has many targets in the brain. Of course people with ADHD may consume alcohol for exactly the same reasons as everyone else. The answers to why they progress from smoking may be no different to the general population, but the many comorbidities make it difficult to assess what symptoms could possibly be self-medicated.

Amphetamines, Methylphenidate, and Addiction

The simple equation “amphetamine and cocaine are bad; methylphenidate is like cocaine, therefore methylphenidate is bad” has some basis in the scientific literature. Add to this equation the gateway hypothesis of addiction, and before you know it Ritalin is up there with crack cocaine and heroin.

Amphetamines and cocaine have a long history of use/abuse, and readers who want a historical account of amphetamines and methylphenidate are recommended the book by Professor Leslie Iversen called Speed, Ecstasy, Ritalin [1129].

Let us assume that methylphenidate is addictive and warrants a status similar to cocaine. If this is the case, we should see methylphenidate being diverted from clinical use and being used illicitly. Methylphenidate seizures do not get the same media attention as a truck-load of cocaine, but the drug does have a market. In the USA, methylphenidate was diverted by theft etc. for illicit use, and in nearly two years 700,000 doses were stolen [1130]. Methylphenidate is difficult to make in back-street labs (other drugs such as amphetamine are by comparison easy to make), and its appearance on the black market is by diversion. Diversion is the norm for the illegal supply of methylphenidate [1131–1132]. Not all is stolen; there are also “Attention Deficit Scams,” where a parent or other adult takes a child who allegedly has ADHD to a number of medics to obtain methylphenidate prescriptions for the adult to use, sell, or trade. In fact most diversion is via family or friends [1133].

The view that children are trading their prescription medications in the playground has some support in the USA, where adolescent schoolchildren are selling their methylphenidate medication to friends and classmates, who are crushing the tablets and snorting the powder like cocaine [1130]. Clearly this can be avoided with various preparations that make it difficult to extract the pure methylphenidate. In the USA, a study of school-aged children indicated that 16 percent of those prescribed methylphenidate were asked to trade, sell, or give away their drugs [1134]. There was a 4 percent prevalence of illicit methylphenidate use in adolescent schoolchildren in the USA [1135], and a range of 3 to 43 percent amongst undergraduates [1136, 1137]. About 10 percent of methylphenidate misusers met the criteria for dependence [1138]. Prescription medication misuse was associated in females with the appetite-suppressant properties of amphetamine, whereas methylphenidate was commonly misused by males [1139].

The undergraduate use of illicit methylphenidate was dependent on a number of factors such as sex and ethnicity, with higher rates in males, white groups, and members of fraternities (peer group affiliation). Rates were also higher at colleges with more competitive admission standards and with individuals with lower grades scores [1140]. Clearly the illicit use of methylphenidate is associated with cognitive enhancement rather than getting high. Recent studies have indicated that illicit methylphenidate is used to aid concentration and help study, or to increase alertness [1141] or general productivity [1133].

The illicit use of methylphenidate for cognitive enhancing effects should be distinguished from the question of its addictiveness – there is a greater similarity with caffeine rather than cocaine. Here we have the use of methyl­phenidate for a specific purpose and not to get high – and, quite frankly, if you want to get high, there are better drugs than methylphenidate.

Amphetamine is a known addictive drug; humans and animals will take it (see [1142]). Self-administration of drugs in animals is an important step in identifying its abuse potential. In animal studies, the animal, usually a rat, is trained to press a lever in a box in order to receive an injection of a drug (see [1143] for a recent review). Virtually all known addictive drugs are self-administered in the rat. Methylphenidate is no exception. Rats, and other animals, will readily press levers for methylphenidate [1144–1148].

Humans were also seen to work in return for methylphenidate [1149, 1150]. The mechanism by which methylphenidate achieves this status is via the DA system [1144, 1151]. In doing so, methylphenidate is similar in self-administration procedures to all addictive drugs (see [1142]). A word of caution needs to be exercised on the numerous studies that have looked at methylphenidate and addiction. The studies mentioned use normal animals or people without ADHD – such data may have a limited application to ADHD and substance abuse [1152]. The feelings of being high are dose-dependent and it is to be noted that an immediate-release methylphenidate is used which is more closely linked with the feelings of a high [659].

The speed of entry to the brain and the rate of metabolism are important factors in methylphenidate’s abuse likelihood [659, 1151, 1153]. The ability of methylphenidate to treat the symptoms of ADHD and its abuse potential can be differentiated pharmacologically [1153–1154]. To have an abuse potential similar to that of cocaine requires the methylphenidate to enter the brain rapidly. Once in the brain, methylphenidate blocks the DAT, thereby increasing DA levels within minutes. The speed on entry and time taken to clear the brain of methylphenidate differentiates it from cocaine despite a similar action at the DAT. The feelings of being high decreased with the decline of cocaine [659]. Long-lasting effects of methylphenidate are not associated with the high [659]. Therefore, the speed by which a drug acts is as important as what it does at the target. Not surprisingly, slow-release methylphenidate was less likely to lead to abuse than IR methyl­phenidate [1155].

The literature identifies four variables that can influence abuse and clinical efficacy: (1) the dose of methylphenidate needs to reach a threshold that increases the levels of DA such as to be perceived as reinforcing and also to produce therapeutic effects; (2) the reinforcing effects of methylphenidate are associated with rapid increases in DA, whereas the therapeutic effects are associated with a slow clearance of DA; (3) sensitivity to methyl­phenidate varies across individuals and thus sets an different thresholds for the levels of methylphenidate required to be reinforcing and therapeutic; and (4) the effects of methylphenidate are modulated by different environmental contexts, e.g. for abuse there are the rituals of drug taking and for clinical efficacy the symptom management [1153].

The use of different delivery mechanisms for methylphenidate can minimize its abuse potential. Slow-release systems have become the mainstay of treatment. In a double-blind randomized study, a clinically therapeutic dose of oral osmotic-controlled extended-release methylphenidate did not differ from placebo on abuse potential, and in general slow release was associated with less high than IR methylphenidate in non-ADHD volunteers [1156–1157]. The message is that methylphenidate is addictive but not normally in the preparations frequently prescribed.

Methylphenidate Treatment and Addiction

We have seen nicotine as a gateway drug and that those with ADHD are more likely to smoke. However, the chance that nicotine is the first psychostimulant that has been taken is low; after all, methylphenidate is a psychostimulant – arguably a more powerful one than nicotine! Thus the question remains: is methylphenidate a gateway drug and does it increase other drug use?

In chapter 1 we saw how Courtney Love blamed methylphenidate for Kurt Cobain’s later addictions – that early exposure left a void that needed to be filled in adult life. Love is quoted as saying about methylphenidate, “It was euphoric when you were a child – isn’t that memory going to stick with you?” [11] (p. 20). When Cobain was prescribed there was only IR methylphenidate available. Cobain was a smoker, a heroin user, and a user of countless other drugs; he was also a very troubled individual.

In non-ADHD smokers, methylphenidate was shown to increase the number of cigarettes smoked and other smoking-related measures, e.g. number of puffs [1158–1159]. A similar effect was found with amphetamine [1160–1164] and cocaine [1165]. What we have to remember from studies that have looked at methylphenidate increasing the risk of subsequent substance use is that they were conducted in non-ADHD individuals or normal animals. We are dealing with a very different biological system in the person with ADHD or the animal model of ADHD.

However, the very important question remains whether methylphenidate is a risk factor for later addiction in those using it to treat ADHD. The evidence is not conclusive on this matter, and methodological problems make strong conclusions difficult.

Methodological limitations are evident in the widely publicized MTA study conducted across the USA (see chapter 7). The three-year follow-up indicated that ADHD groups had a greater chance of substance abuse and those receiving behavioral therapy fared best [801]. Owing to the size of the study and the weight it appears to carry, it might appear to provide damning evidence against the use of methylphenidate. Previous reports on the progress of the MTA study had painted a positive picture of medication effects in ADHD [110, 792, 795, 797, 1166]. At the 24-month update of the MTA study it was reported that those receiving intensive behavioral therapy had less substance use than other groups, including methylphenidate treatment [801]. This is in contrast to at 14 months, when a clear benefit of methylphenidate could be seen. Surprisingly age was not a factor in substance use outcomes. There are some methodological issues regarding such studies that need to be addressed before we give too much weight to the MTA results. In a companion paper the limitations with the methodology of the MTA were noted. These included the stability of the groups to which children were assigned. Unlike a randomized-controlled clinical trial, there was movement in the groups over the three years, e.g. 26 percent of the behavioral therapy group went on to take medication and about 87 percent of the medication groups adhered to their treatment [798]. Thus, those who did well in medication groups may have discontinued and those who who not do well with behavioral therapy went on to take medication to manage their ADHD – ultimately the groups at the start of the study were different to those three years later. The MTA study is large, if nothing else, and looks at long-term effects of various interventions. However, it did not have a placebo comparison group and the maintenance of randomization was not going to be ethically achievable; after all, no one is going to want to wait three years on a treatment that is not effective just for the sake of scientific rigor. Moreover, the methylphenidate was not a modified slow-release preparation.

Many other studies have found a beneficial effect of early treatment with methylphenidate with regard to substance abuse.

An early study that looked at the effect of methylphenidate on alcohol intake stated that there was a non-significant trend for those who had been treated with the drug to consume more beer and wine [1167]. Since then the late Professor Nadine Lambert was successful in bringing claims of methylphenidate’s gateway to addiction to the attention of clinicians [1098]. A follow-up study of charting the progress of 5,212 from children into adulthood supported her initial claim and extended it to suggest that ADHD behaviors did not increase the odds of drug use whereas the treatment with methylphenidate did [1168]. Lambert took her position on ADHD and subsequent addiction from the incentive-salience theory of addiction (addressed later; [1169]). Lambert’s studies are in contrast to many others, and in her defense Brian Kean alludes to the involvement of the pharmaceutical industry in the funding of individuals and projects finding the opposite – protective – effect of methylphenidate, although he is careful to avoid litigation [1170]. Thus your theoretical perspective as well as your funding can determine how you approach research.

Others have declared a beneficial effect of methylphenidate on latter substance use. People who respond well to methylphenidate tended to drink less than those who did not respond well [1103]. Studies that record positive outcomes of treatment have looked at the chances of developing SUD if they have ADHD. Professor Timothy Wilens and colleagues have performed meta-analyses on the growing body of research and provide us with the odds (odds ratio) of developing SUD in those with ADHD who are treated vs those who are not treated [1171–1172]. An odds ratio is a way of assessing the chance of something happening.² We can operationalize the odds ratio as follows: an odds ratio of 1 indicates that the condition or event under study is equally likely in both groups (i.e. no effect); an odds ratio greater than 1 indicates that the condition or event is more likely in the first group rather than the second; an odds ratio of less than 1 indicates that the condition or event is less likely in the first group, but more likely in the second. In the studies by Wilens et al. [1171–1172], the odds ratio estimates the increase in the odds of not developing SUD (i.e. a protective effect) among those individuals previously treated pharmacologically compared to individuals with ADHD who were not treated pharmacologically.

Wilens has led a number of studies that indicate a protective effect of ADHD medication on substance abuse. Using meta-analysis on six published studies, a pooled odds ratio of 1.9 was obtained, conferring a protective effect of treatment [1172], although others did indicate a risk factor (e.g. [1098]). A study published by Barkley et al. [1173] looking at 147 ADHD children over 13 years indicated that medication in childhood or adolescence did not have an impact on later substance misuse. More recent studies have supported this notion, but the effect might not be as great as early studies indicated [1174–1177]. Using retrospective reports of adults with either a current or past history of treatment for ADHD did not reveal a risk for later substance abuse, and furthermore there was no effect of the timing of onset of treatment or its duration of use on subsequent drug use [1178]. A similar effect was found in adolescents [1174, 1176]. However, one study failed to find a protective effect of methylphenidate in adults, leading the authors to postulate that during adolescence the drug might delay an inevitable onset of SUD [1175]. Pelham has suggested that this may be so because the untreated group were on average two years older than the treated group, which means the treated group have yet to start their drug-taking careers (Pelham cited in [1179]).

At the heart of the concern over methylphenidate is the fact that we are giving these powerful drugs during a child’s development. The timing of the intervention may be predictive of potential substance misuse [1180]. Methylphenidate was not associated with a substance use problem when treatment commenced in childhood [1176], whereas it was associated with substance use when treatment commenced in adolescence and adulthood [1181]. The effects of methylphenidate need to be understood in terms of neural development. The protective effect may not be conveyed once the brain has undergone maturation.

Drugs, Cannabis, and Psychiatry

Can the use of illicit drugs mimic or indeed precipitate ADHD? Amphetamine can induce symptoms similar to schizophrenia (see [1182]), and many drugs can have a major effect on the cognitive functioning of the individual [1183]. Cannabis use requires special mention due to the ongoing debate in the UK about its classification as a controlled substance and its role in the onset of psychiatric disorders. Much has been said about the connections between schizophrenia and cannabis use [1184–1185], but does the same hold for ADHD? Some of the effects of cannabis include cognitive changes seen in ADHD [1186]. The interference of non-medical drug use was considered sufficiently problematic to prompt a letter to the American Journal of Psychiatry which stated: “Only prolonged abstinence from nonmedical drug use can allow for a diagnosis of ADHD to be made with confidence” [1187] (p. 973). Although some of those who have been diagnosed with ADHD use cannabis, the link is not at all clear. Cannabis has been considered a gateway drug to other addictions, but cannabis use is not as prevalent as cigarette smoking in ADHD [1125]. Hollis and colleagues [1188] were unable to find a link between adolescent cannabis use and ADHD. Furthermore, a New Zealand study looked at 25-year-olds with ADHD and found that smoking cannabis was associated with the symptoms of ADHD, but that the association was mediated by other drugs such as MDMA and amphetamine, thus drug taking in general was linked to ADHD and not a specific drug as an exacerbating factor [1189]. Similarly, cannabis, along with smoking and alcohol consumption, was related to ADHD symptoms [1190]. A very specific effect of cannabis was found in a subgroup of ADHD. In a study of 916 members of the general population, males who had symptoms along the ADHD-H subscale were more likely to take cannabis and other drugs [1191]. Thus whilst cannabis is a drug that is used early in the substance abuser’s drug career, it does not appear to be a drug precipitating ADHD itself.

Methylphenidate: Slow Release and the Treatment of Cocaine and Amphetamine Addiction

Paradoxically, methylphenidate has been proposed as a treatment for psychostimulant addiction, and case studies have been supportive of this use [1192]. The rationale for methylphenidate use in psychostimulant addiction is similar to that of methadone treatment in heroin addicts or nicotine replacement in smokers. If a drug user is prevented from taking their drug, a constellation of effects known as withdrawal symptoms emerge. Withdrawal symptoms are extremely unpleasant and can often lead to relapse because the quickest way to alleviate them is to take the drug again. How to prevent relapse is an important clinical question, and in heroin users this is achieved by providing methadone. Methadone is similar to heroin, but unlike heroin it takes a lot longer to get into the brain and then a long time to be removed. Thus, methadone provides a background amount of heroin-like activity which avoids the worst of the withdrawal effects seen after abrupt cessation, thereby increasing the chances of rehabilitation. Of course there is more to treating addiction than withdrawal management – if only it were that simple!

Using this logic, methylphenidate has been suggested as a possible intervention for cocaine-induced withdrawal symptoms [1193–1196]. Thus, methylphenidate can occupy the DAT and increase background (tonic) levels of DA in a gradual manner that will hopefully alleviate cocaine withdrawal symptoms.

The work of Professor Nora Volkow and colleagues indicates that slow-release (SR) methylphenidate has clinical utility to treat ADHD with minimal abuse potential [1153–1154] and is effective in treating comorbid ADHD and cocaine dependence [1197–1198]. Despite the fact that methyl­phenidate and cocaine are pharmacologically similar [659], one of the key issues in their abuse liability is the speed at which they enter the brain. As a rule of thumb, the quicker the drug can get to the brain and increase DA levels, the more likely it is to be rewarding and therefore addictive.

In Volkow’s neuroimaging studies it has been demonstrated that the high obtained from intravenous cocaine or methylphenidate correlated with the concentration of the drug [659]. Intravenous administration of drugs is an extremely effective route to the brain; oral administration is slower and at the mercy of digestive processes. In the baboon, it was shown that intravenous administration of methylphenidate results in a large and rapid increase of DA in the striatum, whereas a slower rate to the same levels was found with oral administration [1199]. In humans, oral methyl­phenidate was not linked to getting high [1151, 1200–1203]. However, the high obtained from intravenous methylphenidate was not entirely linked with blockade of the DAT [1204], and occupancy of the DAT may need to be above 80 percent to get a high [1205].

The studies using methylphenidate demonstrate that it has addictive qualities if given intravenously; the oral route which is used in treatment is less likely to be associated with getting high. Many studies have compared IR methylphenidate with the various forms of SR methylphenidate and found a reduced abuse potential with the latter [1156–1157]. The SR methylphenidate should have less of an effect on highs and abuse than the immediate-release (IR) methylphenidate. Spencer and coworkers [1155] looked at the feeling obtained from methylphenidate via the two modes of delivery. They found that an effect was more readily detected and liked with IR methylphenidate.

The evidence raises the possibility of a beneficial effect of methylphenidate on cocaine abuse. However, a review of the studies to date looking at methylphenidate as a potential candidate for the treatment of cocaine dependence has not supported the earlier assumptions of its usefulness [1206].

New developments in drug treatment emphasize a minimized abuse potential. Part of the rationale for the continuing modification of the drug delivery system is to avoid the large transient peak in DA that can be obtained from methylphenidate. The pharmaceutical companies are utilizing new technology to deliver the methylphenidate slowly to the brain and making its extraction near to impossible for abuse purposes. New developments such as prodrugs and patches are going some way to achieve this aim.

Methylphenidate and Long-Term Neural Changes

Methylphenidate interacts with and changes the DAT [1207]. Long-term changes to the DA system have been the focus of one of the most studied theories in addiction – sensitization. We are more familiar with the concept of tolerance (e.g. the effect gets smaller or you need more of the drug to get an effect), but sensitization is when the drug has a bigger and bigger effect after successive administrations. It is the opposite of tolerance. A drug can induce both tolerance and sensitization; it is the behaviors that are measured that respond differently to the drug, e.g. tolerance occurs to the analgesic effects of morphine and sensitization to the stimulatory effects.

Sensitization is seen in many addictive drugs and can be measured pharmacologically, physiologically, and behaviorally.

DA reward pathways become over-active in response to the drug, and this activity is central to the theory presented by Professors Terry Robinson and Kent Berridge. Their theory argues that motivation for drug consumption is increased as a result of changes to the brain that occur as a result of the drugs [1169, 1208–1209]. Why is this important to methylphenidate and ADHD? If the theory is correct, then methylphenidate will increase sensitization in the regions of the brain that mediate addiction. Worryingly, methylphenidate will, over a period of time, produce long-term neural changes that maintain addictive behaviors. Once established, sensitization is regarded to be permanent; the animal data and anecdotal accounts of human addiction support this notion.

Within the theoretical context of sensitization there is reason to be wary of methylphenidate treatment. Not surprisingly, there are many studies that have demonstrated sensitization to methylphenidate in animals [1039, 1210–1222].

One of the general criticisms of the sensitization theory has been that it is not readily testable in humans, hence the reliance on animals. A review of the clinical literature found little evidence of sensitization in adults who had received methylphenidate as children, though this study did not assess sensitization directly but rather looked at the prevalence of substance abuse [1223]. Other studies of sensitization in methylphenidate-treated children found tentative evidence for sensitization, but conclusions are limited by their methodology [1224].

The sensitization theory purports that once sensitization is established to a drug such as methylphenidate, then cross-sensitization can occur. Cross-sensitization is when there is a heightened response to a different drug to that which produced the sensitization in the first place. We might predict on this basis that sensitization induced by methylphenidate can lead to sensitization to other drugs such as cocaine, amphetamine, MDMA, and nicotine, and enhance their abuse liability. As the theory suggests, methylphenidate does indeed induce cross-sensitization to other psychostimulants [1219–1221, 1225–1226]. However, looking at the molecular changes that occur in response to long-term drug use has revealed that methylphenidate, whilst sharing the same pharmacology as cocaine, did not have the same effect on neuroadaptations [1227]. This is good news!

There are some factors that determine sensitization that are relevant to methylphenidate and ADHD, e.g. the genetics of the animal – some are more susceptible [1228] and the question of a shared genetic heritage needs addressing. More importantly, the age at which the methylphenidate is given is critical, with adolescent rats not showing methylphenidate-induced sensitization [1219, 1228–1232]. One study using rats has indicated that early treatment with methylphenidate lasted into adulthood; those rats treated early with methylphenidate showed less responsiveness to cocaine reward in adulthood [1233]. Again, understanding the action of methylphenidate during developmental changes will be critical to our understanding of its long-term effects [330].

All the studies on sensitization have in the main taken place in normal rats and not animal models of ADHD. There is a key difference between the two. In the normal rat we are elevating the levels of DA far above the normal with methylphenidate – even if it is a clinical dose. However, the DA levels in someone with ADHD are far from normal and are most likely under-active. The question remains as to whether methylphenidate induces sensitization in an underactive DA system. I am aware of only one study using the SHR model of ADHD and sensitization in which adolescent SHRs were exposed to methylphenidate and followed through to adulthood. These rats were found to have less sensitivity to cocaine in adulthood; the neural changes in DA in the NAcc were not evident in these rats [1040]. This raises the important question as to how methylphenidate is operating in models of ADHD and ADHD itself. Drugs such as cocaine increase DA in the NAcc [1234], but perhaps in ADHD the levels of DA are not sufficient for enhanced reward and addiction.

If methylphenidate brings the levels of DA up in ADHD, this may not reach the levels required for sensitization.

The mechanism(s) by which methylphenidate-induced sensitization occurs remains unclear; what we do know is that distant regions of the brain are involved and these may be affected differently depending on the age at which administration is given. These are academic questions; whether they are important clinically remains to be seen. Little work has been done in animal models. In the world of addiction, the reversal of sensitization may be crucial, and as yet there are only a few accounts that the process can be reversed (e.g. [1235]).

Reward Deficiency in ADHD

Whilst the sensitization theory of addiction has implications regarding treatment of ADHD, the Reward Deficiency hypothesis has a direct application to the disorder. Reward deficiency is where there is a reduced efficacy of the reward pathways in the brain. The reward pathways ensure our survival as individuals and a species. In ADHD and addiction they are argued not to be optimal. Thus addiction and ADHD have some underlying similarities. Some have even argued that the genetic basis of reward systems in ADHD renders individuals vulnerable to obesity (food is a reinforcer much the same as a drug) now that our world is more abundant in cheap fat-rich foods [577].

Taylor and Jentsch [1236] have reviewed the data of reward function in ADHD and suggest that the reward system is indeed dysfunctional and treatment with methylphenidate corrects this dysfunction. Such notions of faulty reward systems have been incorporated in theoretical accounts of ADHD [484, 911, 1021–1022, 1030, 1237–1239]. In general, children with ADHD have been shown to choose small immediate rewards over larger delayed rewards, which are chosen by healthy controls [1240–1246]. Children with ADHD want frequent rewards immediately, even if it means they get a smaller reward [1247]. If you are a carer of a child with ADHD. you will no doubt be aware that for any behavioral strategy to be effective it needs to involve immediate and frequent rewards; children with ADHD are not good at waiting, and rewards at the end of the day are not effective. Intolerance to waiting for rewards has been implicated in the tendency for those with ADHD to escape or avoid delay [1243–1244], presumably because the frustration experienced is to great [1246].

The Reward Deficiency syndrome, as described by Blum and colleagues [1021], postulates that a common mechanism in ADHD, and other disorders, is a reduction in D2 receptors. Some support exists for this in addiction with imaging studies showing a reduction of D2 receptors [1248–1250]. As a rule of thumb, drugs that stimulate neural activity produce a compensatory reduction in receptors and their sensitivity. Studies in monkeys have shown that those who have lower D2 receptors are more likely to self-administer cocaine [1251–1252]. In unmedicated adults with ADHD, the striatal D2 and D3 receptors were fewer than healthy control adults, prompting the authors to suggest that reduced DA was associated with inattention and a vulnerability to substance abuse [688]. On the whole, genetic studies on the DA D2 receptor point to a dysfunction in ADHD [634–635, 1253–1255] but there are some that do find a link [1256]. However, genetics studies tell us of a difference in genes, not how this is translated into the brain and then behavior.

Kelley and her colleagues [1061] have argued that the evidence base for ADHD and reward deficiency is not on a solid foundation; there are numerous facets of reward – e.g. size, speed of delivery, and chance of getting the reward – that need to be looked at in ADHD before strong conclusions can be made.

Some studies have used psychophysiological measures such as heart rate and skin conductance to assess arousal during reward. One study suggests that during decision making on a gambling task, children with ADHD may be sensitive to the frequency of reward alone but not the size [1257]. Thus the size of reward appears to be less important to the child compared to the amount of rewards. Luman et al. [1258] have also demonstrated that children with ADHD were not abnormally affected by the attractiveness of reinforcement. The probability of receiving a reward has not been extensively evaluated in ADHD. Studies in adolescents using fMRI show a reduced striatal response during reward anticipation indicative of learning about rewards [1259–1260].

In order for behavioral reinforcement strategies to work, the child must realize what they have to do in order to obtain reward. Controllability is considered very important in motivation [1261]. Little has been done on the perceptions of control in ADHD, but some studies have utilized the learned helplessness paradigm, in which one learns that no matter what you do it is pointless [1262]. Using solvable and unsolvable puzzles. Milich et al. found that boys with ADHD solved fewer puzzles and gave up on more puzzles – especially after the unsolvable condition – compared to controls, which was reversed by methylphenidate [1263]. Furthermore, ADHD boys reported being more frustrated by the task than did control boys [1264]. The boys were more likely to make external attributions for failure (i.e. blame others or the drug) and internal attributions for success (take the glory) [1263, 1265, 1266].

How important delay aversion is to the clinical picture in ADHD has been questioned. Delay aversion has been argued to only be moderately associated with ADHD and perhaps only to a subgroup [1267]. It is also important to note that delay aversion experiments have been used to assess impulsivity, and we know how difficult impulsivity is to pin down in ADHD.

Such notions of immediate gratification and impulsivity have ramifications on treatment options. Without medication, behavioral therapies may not be as effective as they could be. This is an important aspect for educators and parents alike to realize – anecdotal accounts often state that the child is somehow reward-immune. This is not necessarily the case, but rather the child needs immediate reward, and token economies that work in non-ADHD populations may be less effective. The best option may be to provide concomitant pharmacotherapy and behavioral management.

Impulsivity in SUD and ADHD (Revisited)

Theoretical accounts of ADHD have described the disorder as primarily a problem of impulsivity and a failure to inhibit responses as measured in laboratory tasks (see chapter 4). One of the criticisms of the account was the inability of tests of BI to differentiate between various disorders, and for the purpose of the present chapter SUD was one of them. Impulsivity or poor BI has been argued to represent an endophenotype for both ADHD and SUD, with genes and associated neurochemical dysregulation leading to one or the other disorders or indeed both disorders [1268]. Clearly, there is comorbidity between ADHD and SUD, but how they are connected is a question that remains open – it could be reward systems or it could be impulsivity. The impulsivity of one disorder could lead to another, but it is not necessarily the case that all cases of SUD have ADHD or even ADHD-like symptoms.

Impulsivity has been linked to addiction, and in particular to the orbito­frontal cortex and cingulate [1269–1271]. Changes in these areas have also been found in ADHD [660, 719, 1272–1275] and are active in reward anticipation in males with ADHD [1276], although the orbitofrontal cortex has recently been linked specifically with Conduct Disorder and not ADHD [1277]. Animal studies of impulsivity point to differential actions of the orbitofrontal cortex depending on the task used [1278].

Given that we can identify structures that are different in ADHD and that these regions are also implicated in addiction and impulse control, do we have a common mechanism at play? The psychological substrate is impulsivity and the neuroanatomical one comprises the frontal cortex, with subregions such as the cingulate and the orbitofrontal cortices being involved. The brain undergoes changes right up until early adulthood [649–650]. Those neural changes are thought to be critical in addiction and are vulnerable to drug use [827, 1279–1281]. Not surprisingly, adolescence is a period of development characterized by over-emotional responses that have not been thought about in a rational way – after all, the rational part of the brain is yet to be fully developed. Add to the mix that substance misuse starts during adolescence and we have a cocktail that makes non-ADHD adolescents vulnerable to drug use let alone those with the additional risk factor of ADHD.

Summary

Understandably, treatment with amphetamine and a drug like cocaine is worrying. With all of this information we fall between two stools: on the one hand, we have the knowledge that untreated ADHD is a risk factor for adolescent substance misuse, whilst, on the other, we have the knowledge that the brain is changing and that even therapeutic drugs, such as methyl­phenidate, may detrimentally change the neural wiring. Which stool should we go for? The answer is not simple and may depend on many factors. Arguably, we may be dealing the lesser of two evils. Alternatively, we may conclude that changing the brain during adolescence in ADHD just might be of long-term benefit. Such questions remain unanswered.

Research in this area has highlighted that those with ADHD have deficits in reward processing. Work on the mechanisms by which this occurs, and on addiction science in general, is developing at a fast pace, and we are beginning to get a greater understanding of reward and learning in normal and pathological states. Such an understanding of reward processes and motivations is important academically, but in ADHD it can go some way to informing behavioral therapies of the specific requirements necessary for treatment efficacy of this group. After all, behavioral therapies are based on reinforcement and reward principles.

Notes

1 See chapter 7 for phasic DA release.

2 For more on odds ratios, go to http://www.jr2.ox.ac.uk/bandolier/band25/b25-6.html.