Introduction

Attention deficit/hyperactivity disorder (ADHD) is one of the most commonly diagnosed diagnoses in children and adolescents, affecting about 5 % of children and adolescents worldwide [1]. Those with ADHD who are un- or undertreated are at heightened risk for problems in school and work, relationships, and overall functioning [2•]. Psychopharmacology for ADHD has been shown to be quite effective both alone and in combination with behavioral treatments [3, 4•, 5]. Still, possibly driven by media queries of “over-medication” of children or case reports of medication-related adverse events, some still view treatment as controversial [6]. Clinicians must be aware of the significance of adverse events and approaches to manage risk to best counsel patients and families.

Cardiovascular

Concern over cardiovascular side effects has prompted rigorous debate and study. In examining cardiovascular risk for the purposes of this review, we will examine three salient questions. First, what are the known/confirmed cardiovascular manifestations of ADHD medications? Second, do ADHD medications increase the risk of sudden cardiac death? Finally, what is the appropriate recommended workup prior to initiation of ADHD medications?

Known/Confirmed Cardiovascular Manifestations

We will begin with a look at the stimulants, as these are most commonly prescribed. Since the 1970s, we have known that stimulants have modest cardiovascular effects, such as increasing systolic blood pressure (3-8 mmHg), diastolic blood pressure (2-14 mmHg), and pulse (3-10 beats per minute) [7, 8•, 9, 10]. These changes are minor in all but a small group of patients where changes of 5-15 % were observed [8•].

Both clonidine and guanfacine have original indications for use as antihypertensive agents. And both have shown asymptomatic decreases in both blood pressure and pulse in long-term studies. And like stimulants, there have been case reports of sudden deaths in children on clonidine; however, analysis reveals these events occurred in the context of polypharmacy and/or congenital heart malformation [11]. The Daviss study showed more incidents of bradycardia in those treated with clonidine compared to placebo, but no group differences were found on electrocardiograms or other cardiovascular outcomes. Of note, this held true for patients in the study on clonidine and mehtylphenidate [11]. Similarly, no serious electrocardiogram abnormalities have been seen in guanfacine studies.

Combination pharmacology is increasingly common in the treatment of ADHD, such as with a stimulant plus an alpha agonist. In one controlled trial of guanfacine ER plus psychostimulants, a small mean decrease in pulse, systolic, and diastolic blood pressure was noted in those receiving the alpha agonist plus stimulant vs. placebo plus stimulant, and no new safety problems emerged [12•]. In an open-label, randomized study, it was shown that combining guanfacine extended release and methylphenidate did not result in significant drug-drug interactions, and no unique adverse events were observed [13•]. Although more monitoring is required, as of now, there is no evidence to suggest altering pre-treatment recommendations from a cardiac standpoint prior to initiation of combination treatment.

Do ADHD Medications Increase Risk of Sudden Cardiac Death?

Concern has remained, fueled by intermittent reports of sudden death in patients taking ADHD medications, that stimulants may be a contributing factor in unexpected cardiac death. To this end, officials in Canada at one point removed Adderall XR from their market (although this was ultimately reversed). Clinical medication trials, often too small and of too short duration, have not been able to assess long-term cardiac risk well. Here, large-scale epidemiological studies have helped with clarification. In a large cohort analysis of over 170,000 youths aged 6 to 21 years old without known cardiovascular risk factors, diagnoses of cardiovascular events and symptoms were not associated with stimulant use [14•]. In another study of over 440,000 young and middle-aged adults, there was no increased risk of serious cardiovascular events associated with current or remote use of ADHD medications as compared to non-use [15]. Despite reports of unexpected deaths of children receiving stimulants, large cohort studies have not shown a link between stimulants and serious cardiovascular adverse events [16•, 17•, 18]. Indeed, after considerable study, there is general consensus that the frequency of sudden death in patients taking stimulant medications is the same as for the general population.

A recent, longer term cohort study did find that cardiovascular events were indeed rare, but twice as likely to occur in stimulant users as in non-users [19•]. This study defined cardiovascular event as the child having any hospital contact (inpatient, outpatient, emergency room) with a diagnosis of cardiovascular disease, defined as an I00-I99 diagnosis. Of note, among children with ADHD (n = 8,300), 111 cardiovascular events were identified. Hypertension, heart disease NOS, and cardiovascular disease NOS comprised 62 %, arrhythmias comprised 23 %, while cardiac arrest accounted for less than 1 % of events. This long-term cohort study highlights the ongoing need for continued monitoring of patients treated with stimulants [19•].

Pre-Prescription Workup

Although some clinicians obtain electrocardiograms prior to initiating treatment, there are no compelling data to suggest this approach. Indeed, of note, there is no empirical evidence to suggest that stimulant use causes clinically significant changes in EKG parameters (PR, QRS, and QT intervals) [8•]. The American Academy of Pediatrics (AAP) and the American Heart Association (AHD) released a joint advisory statement in 2008 regarding the evaluation of children who are to receive medications for ADHD. In it, they recommend an approach of careful assessment to detect cardac disease based on history and physical examiniation. As well, they state that acquiring an electrocardiogram is reasonable for a physician to consider, but is not mandatory to obtain [20]. Indeed, the recommended treatment approach of the American Academy of Child and Adolescent Psychiatry (AACAP) is for thorough history taking and examination, which, if not suggestive of cardiac disease, may be followed by pharmacotherapy without additional evaluation. Providers should take a careful personal and family cardiovascular history as well as cardiac examinations, deferring a more comprehensive assessment to an appropriate specialist for those deemed at risk [21, 22].

Negative family and personal cardiovascular histories coupled with a recent cardiovascular physical examination should allow for prompt initiation of treatment without further workup. When risk factors are present, one can consider ordering an EKG and referral to an appropriate specialist.

Before starting a medication, collect the personal and family cardiovascular history including history of hypertension, tachycardia, palpitations, fainting, exercise intolerance, murmurs, or history of family members dying of unexplained cardiovascular death before the age of 40. Monitor other medications, especially those with known cardiovascular effects. Measure baseline pulse and blood pressure, and monitor them every 3-6 months [23•]. It is not necessary to order routine EKG unless there are concerning findings in either the history or physical examination. Refer to a cardiologist if there is clinical concern.

Appetite and Growth

Loss of appetite is a well-known adverse effect of stimulant use. A meta-analysis published in 2001 including selected RCTs of MPH-IR in children and adolescents found a significant difference in the frequency of appetite loss (mean intergroup difference in prevalence: 30.3 %, 95 % CI: 18.0-42.6). The mean doses varied, but generally remained low. The studies averaged only 3 weeks, making it difficult to comment on possible attenuation over time [24].

Although not as common, a meta-analysis of appetite decrease with atomoxetine use showed a significant difference of 15.4 % versus 4.1 % of those treated with placebo [25]. In one randomized double-blind placebo-controlled study comparing atomoxetine and methylphenidate in children, atomoxetine was significantly more associated with anorexia, nausea, and vomiting, although it should be noted that the atomoxetine dose was relatively high compared to the methylphenidate dose [26]. There is not enough significant evidence of the effects of alpha-agonists on appetite and growth to make a comment.

While it is well established that stimulants can suppress appetite, there is ongoing debate as to whether or not they inhibit growth in children and adolescents. The effect on growth could be explained by either poor nutrition, due to lack of appetite, or the inhibitory effects of increased synaptic dopamine on growth hormones. A 2010 prospective study of height and weight in children with ADHD found no evidence of an association of ADHD with height trajectories over time and found no evidence that stimulant use was associated with differences in growth over a 10 − 11-year follow-up [27]. Other studies have suggested a tendency for the deficit to attenuate over time after stimulant discontinuation [28].

The European ADHD Guidelines Group suggests that the clinician monitor height, weight, BMI, and appetite every 6 months of treatment. They also suggest plotting height and weight on standardized national charts. The American Academy of Child and Adolescent Psychiatry (AACAP) recommends serial plotting of height and weight on growth charts 1-2 times per year. If the patient has a change on either scale that crosses two percentile lines for length for age or weight for age on the CDC growth chart, it could suggest an aberrant growth trajectory. If this happens, a drug holiday during weekends and/or summers should be considered, as should switching the patient to an alternative ADHD agent. Cyproheptadine may also be helpful with stimulant-induced anorexia, as one chart review demonstrated [29]. As with any medication decision, the risks and benefits of treatment should be considered, especially given that the evidence currently does not show a significant reduction in ultimate adult height.

Insomnia

Insomnia is one of the more common side effects of stimulant medications. One meta-analysis on MPH-IR did show an average group difference in insomnia versus placebo of 17 % [24]. Most of the studies were brief, making it difficult to comment on possible attenuation of sleep disturbance over time. Several studies did show that most children do not have difficulty with sleep initiation with a third late afternoon dose of MPH-IR, although individual cases may differ [30]. Another study showed no difference in sleep difficulty between MPH-OROS, three-times daily MPH-IR, and placebo [30].

Atomoxetine and alpha-agonists generally have the side effect of fatigue rather than insomnia. There are few studies examining the relationship between sleep and non-stimulant treatment, although there are reports of its occurrence [25].

It is often difficult to tell whether insomnia is due to an extended stimulant effect or a rebound of ADHD symptoms when the stimulant wears off; the management would differ based on the root cause. Adding a late dose of IR stimulant can help if the emergence of ADHD symptoms is contributing to sleep disturbance [31]. Alternatively, if insomnia is thought to be due to the long effects of the stimulant, in some studies melatonin in 3-mg doses has been shown to be effective in improving sleep quality [32]. An evaluation of sleep quality should be included in each visit, and education about proper sleep hygiene should be the first intervention. If sleep quality cannot be improved with this alone, one should consider trying an alternate stimulant, utilizing short-acting formulations rather than long-acting, or using a non-stimulant agent. Low doses of trazodone, clonidine, or an antihistamine can also be helpful for insomnia. The clinician should monitor for priapism in patients treated with trazodone [33].

The monitoring of sleep disturbance should include a thorough evaluation of sleep at baseline and routine questioning about sleep quality at follow-up evaluations. At times, specific questionnaires such as the Children’s Sleep Habit’s Questionnaire (CSHQ) as well as sleep diaries may be helpful. Other causes of sleep disturbance such as restless legs syndrome, which is possibly associated with ADHD, or sleep apnea should be further investigated if suspected [34].

Tics

Tics commonly co-occur with ADHD. As dopamine is involved in the pathophysiology of tics, there is concern that stimulants may precipitate new-onset tics in patients without comorbid tic disorders or exacerbate tics in patients who already have tic disorders. There is controversy in the somewhat limited scientific literature regarding the onset of tics following treatment with stimulants. Some studies suggest that tic development or increased severity of tics may occur with treatment with stimulants [35]. By contrast, a meta-analysis of studies with high methodological quality of first-onset tics during stimulant treatment found no evidence of elevated risk of first-onset tics with stimulant treatment [36]. Likely this remains a case-by-case evaluation, and although initiation of tics is rare, it may indeed occur in individual cases.

As ADHD and tics co-occur so commonly, an equally salient question is whether or not stimulants are likely to exacerbate pre-existing tic disorders. A large meta-analysis of relevant studies of co-occurring ADHD and tic disorders was conducted including 477 subjects. Clinical efficacy of dextroamphetamine, methylphenidate, clonidine and guanfacine, desipramine, atomoxetine and deprenyl was assessed. Methylphenidate did not worsen tic severity, although supratherapeutic doses of dextroamphetamine did. Additionally, alpha-2 agonists and atomoxetine improved tics [37]. Although exacerbation of tics is not as common as previously thought, worsening can occur in some cases.

As is so often true in clinical management, prescribers and patients are encouraged to create a hierarchy of impairing problems. In those situations in which ADHD is more impairing than tics, it may indeed be reasonable to try a stimulant. Patients should be monitored for new-onset tics or tic exacerbation when using medications for ADHD. If tolerable, given tics’ tendency to wax and wane during their naturalistic course, observe tic intensity over 3 months without medication change. If tics worsen during this period, or fail to be alleviated, one may then consider dose reduction or cessation of the stimulant in lieu of an alternative agent. If stimulant treatment is necessary to keep ADHD symptoms under control, providers may add an antipsychotic, alpha agonist, or another tic treatment in addition to the stimulant. Providers may consider atomoxetine or alpha agonists if tic exacerbation is a concern.

Substance Use, Abuse, Misuse, and Diversion

The relationship between ADHD, ADHD medications, and other substances of abuse is an often discussed topic. It is known that substance use and abuse are more common in adolescents and adults with ADHD than in age-matched peers [38, 39, 40•], but clinicians are frequently asked about the potential relationship between stimulants and substance use, abuse, misuse, and diversion. These terms are often conflated, so clarifying the terms is useful.

  • Drug use refers to taking a medication as prescribed, but can also refer to “recreational” use or use for other reasons such as performance enhancement or intoxication.

  • Drug abuse refers to use in situations that create risk for harm, failure to meet personal obligations, symptoms of tolerance, withdrawal, or dependence [41]

  • Drug misuse refers to the taking of medication by someone for whom it is not prescribed or in a way in which it was not prescribed

  • Drug Diversion – refers to the transfer of medication to someone for whom it is not prescribed

In systematic reviews of the relavent literature, there is no evidence that taking stimulants increases the risk for patients with ADHD to develop substance use disorders (SUDs), and perhaps there may be reduced risk [42]. In a case-controlled, longitudinal, prospective study, researchers found that patients with ADHD who were not treated had a 2.6 times higher risk to develop any SUDs at follow-up as compared to healthy controls and a 2 times higher risk than those with ADHD who did receive treatment [43•]. However, these findings may not be consistent as follow-up data from the Multimodal Treatment Study of Attention-Deficit/Hyperactivity Disorder failed to show protective effects of stimulant treatment; stimulants neither protected from nor contributed to the risk of substance use or substance use disorders at follow-up [40•].

To the best of our knowledge, after careful consideration, stimulants do not seem to contribute to future substance use or abuse. However, the mixed data on any potential protective effects suggest the need for further research as well as identification of alternative or adjunct approaches to substance abuse prevention in the ADHD population already known to be at heightened risk.

These medications again are found to be relatively safe when used as intended, but it is important to keep in mind that these medications can be used for alternate purposes or used by people for whom they are not prescribed. In a systematic review of the literature, rates of past-year nonprescribed stimulant use ranged from 5-9 % for grade school-age and high school-age children and from 5-35 % in college-age individuals. Of the individuals prescribed stimulants, 16-29 % were asked to give, sell, or trade their medications [44]. Researchers have attempted to identify those who are at higher risk for misuse or diversion. For example, in samples of college-age students, misuse occurs more frequently among students who are white, in fraternities or sororities, and have lower grades [45]. Also, students likely to misuse also misuse or abuse alcohol and other illicit substances.

Regarding managing risk, clinicians should be aware that their patients may very well be enticed to give away or sell their medications and should be counseled about safeguards. Current evidence is assuring that stimulants will not cause later substance abuse, although careful monitoring is still required. Clinicians should be mindful of high-risk patients and should consider abuse, misuse, and diversion. Psychoeducation is of utmost importance; counsel patients about the need to safeguard prescriptions. When SUDs do occur, treatment of abuse/dependence requires treatment before or along with treatment of ADHD. Consider medications with lower abuse potential, like non-stimulant medications, extended-release stimulants, or lisdexamphetamine in patients with histories of or co-occuring substance use disorders. Finally, physicians may choose non-stimulant treatments when the risk for abuse and misuse exceeds tolerable levels.

Psychiatric Adverse Events

Psychiatric adverse events have been reported in association with both stimulant and non-stimulant ADHD treatments, ranging from affective effects (irritability, sadness, crying, and withdrawal) to psychosis and mania.

Anecdotal clinical evidence and case reports exist demonstrating irritability and depressive symptom onset coinciding with initiation of stimulant medication [46]. How prevalent this side effect may be is not well known, although it is thought to be less common than appetite suppression and insomnia [47]. Some evidence suggests sadness and irritability are more common in preschool children [48, 49]. Other studies suggest long-term stimulant medications may actually improve emotional lability in patients with ADHD [50•].

Mania and psychosis are thought to be even rarer. In a meta-analysis of 49 randomized, controlled clinical trials, the rate of psychosis or mania was 1.48 per 100 person-years in the pooled active drug group. Put another way, in the 49 randomized clinical trials reviewed, a total of 11 psychosis/mania evets were found during 743 person-years of exposure to drug treatment compared to no psychosis/mania events in placebo groups, putting the number needed to harm (NNH) at 526 [51]. A more recent look at the relevant data regarding psychostimulant and antidepressant use in children also found psychiatric adverse events to be rare. Risk of anti-depressant-induced mania was also generally low (less than 2 %) but thought to be greater in patients with some symptoms of mania or family history of bipolar disorder [52].

The available data suggest that while it is possible that stimualnts or antidepressants used to treat ADHD may increase the risk of psychiatric adverse events, this risk is extremely small. If mania or psychosis is present prior to initation of medications for ADHD, it is appropriate to treat these impairing conditions first. In the rare cases when psychosis or mania does manifest after initation of treatment, the stimulant or antidepressant dose should be lowered or stopped. Similarly, clinicians should monitor for irritability or depression when starting or titrating medications. Proper psychoeducation is the most important risk-management strategy, and should adverse events occur, clinicians should reduce the dose or discontinue medication. If psychiatric symptoms resolve, clinicians may consider rechallenging with ADHD medications in the future.

Seizure

Patients with ADHD are 2-3 times more likely to experience seizures than age-matched peers [53]. Similarly, ADHD is the most frequent comorbidity in patients diagnosed with epilepsy. Therefore, a discussion about the risks of treating both conditions pharmacologically is prudent. One retrospective study aimed to establish the efficacy and safety of MPH treatment in children with concomitant ADHD and difficult-to-treat epilepsy. There was no evidence of a deterioration of epilepsy in those treated with MPH [54]. A pilot study of the impact of MPH treatment on the frequency and severity in children with ADHD and epilepsy showed that 73 % of children treated with MPH no longer met the DSM-IV criteria for ADHD. Seizure severity mostly improved, although 18 % showed increased seizure frequency [55•]. Several other studies show no increased risk of seizure in non-epileptic patients treated with MPH [23•]. The effect of ATX and alpha-agonists on seizure risk is assumed to be low, but more studies are needed.

When treating patients with ADHD and no history of epilepsy with MPH, there seems to be no increased risk of seizure; therefore, there is no need for a baseline EEG. Given the high concomitance of both disorders, epileptic patients with ADHD tend to be under-medicated for their ADHD symptoms. Treatment is effective in this population, although close monitoring of seizure frequency should occur. The treating physician should also be in close contact with the patient’s neurologist for the management of all medications.

Priapism

Priapism, a prolonged and painful erection, has been reported in patients treated with methylphenidate. Untreated priapism can cause permanent penile damage if symptoms last longer than 4 h. Although thought to be quite rare, there is little evidence to predict the incidence of priapism with stimulant treatment. A MEDLINE search revealed three case reports about priapism with methylphenidate published since 2006, and one report published in 2004 about stuttering priapism with withdrawal from sustained-release methylphenidate [56–59]. There is no published evidence of an increased risk of priapism with atomoxetine or alpha-agonists.

In 2013, the FDA updated the drug labels and patient Medication Guides to inform patients about this rare, but serious risk. Practitioners should educate their patients about the signs of priapism and when to seek medical treatment. Younger, prepubescent males may need more detailed education, as they may not recognize the problem or may be embarrassed to tell an adult about the symptoms.

Suicide

In 2005, the FDA ordered the maker of Strattera (atomoxetine), Eli Lilly & Co., to place a black box warning on its product label based on previous studies showing a mild increase in suicidal ideation in patients treated with this medication. Recent studies have continued to show that the risk is low. A retrospective study of suicide-related events of 1,024 patients enrolled in comparative, placebo-controlled studies showed only 5/1,024 had increased SI. There were no suicide attempts or completed suicides in any of the studies. A systematic review of the efficacy and safety of atomoxetine from 2009-2011 showed that suicide-related events are not a significant clinical risk and similar to the risk with methylphenidate treatment [60]. Currently, there are no significant data to allow commenting on the risk of suicide-related events with stimulant or alpha-agonist treatment.

Patients with ADHD are at an increased risk of suicide-related events, as are patients with comorbid conditions such as depressive and conduct disorders [61]. Therefore, the benefits and risks of treatment with stimulants and non-stimulants should be weighed carefully. In general, suicide-related events with ADHD treatment are rare and unlikely to be different from the rate in the general population [23•]. A thorough evaluation of mood symptoms, including suicidal ideation, should be made at the initial evaluation. Routine check-ins about SI should occur throughout treatment, and certain measures, such as the Suicidal Ideation Questionnaire (SIQ), may be helpful.

Conclusions

Patients and parents often worry about the side effect profiles of various ADHD treatments. Stimulants have been prescribed since the 1960s, providing physicians with robust data about their safety. Overall, the various agents used to treat ADHD are quite safe, and the assessment of potential risks to individual patients mostly relies on a detailed medical history. More recent risk reports, such as suicidal ideation and priapism, should be considered, although the data supporting these risks are limited. A prescriber should always weigh the risks and benefits of treating ADHD and evaluate each individual case thoroughly. Consulting with other practitioners, such as cardiologists and pediatricians, should always be considered in complex cases.