Efficacy and safety of established and off‐label ADHD drug therapies for cognitive impairment or attention‐deficit hyperactivity disorder symptoms in bipolar disorder: A systematic review by the ISBD Targeting Cognition Task Force

Abnormalities in dopamine and norepinephrine signaling are implicated in cognitive impairments in bipolar disorder (BD) and attention‐deficit hyperactivity disorder (ADHD). This systematic review by the ISBD Targeting Cognition Task Force therefore aimed to investigate the possible benefits on cognition and/or ADHD symptoms and safety of established and off‐label ADHD therapies in BD.


| INTRODUC TI ON
Bipolar disorder (BD) and attention-deficit hyperactivity disorder (ADHD) are neuropsychiatric disorders with a broad overlap in psychopathology, and possibly, pathophysiology.Comorbidity of BD and ADHD is common, with an estimated frequency of 9%-35% in adults and 38%-98% in children with juvenile BD. [1][2][3] There are important differences between the disorders, such as the episodic nature, the changing polarity in BD and the on average earlier onset in ADHD.
However, the disorders also share important features, including impulsivity, psychomotor agitation, poor sleep, and cognitive impairments, including inattention and distractibility.][8][9][10][11][12] The pattern of cognitive impairments is similar in BD and ADHD, with broad deficits across several cognitive domains that vary in their relative severity. 10,11According to meta-analyses, verbal memory, working memory, and executive function are particularly affected in BD (e.g., Ref. [3,12] although with some heterogeneity, Ref. [13,14]).In ADHD, the greatest deficits have been observed in working memory, response inhibition, and reaction time variability. 15Taken together, the symptomatic overlap, comorbidity, and similar cognitive impairments suggest common neurobiological underpinnings in BD and ADHD that may be targeted to improve cognitive outcomes and daily functioning. 16normal dopamine signaling likely plays a role in the cognitive difficulties across BD and ADHD.Both disorders have been linked to dysfunction of the dopamine transporter (DAT), a transmembrane protein that reduces dopamine signaling by driving its reuptake into the presynaptic neuron.8][19][20] DAT also serves as a key target for various stimulant and nootropic ADHD medications. 21More broadly, cognitive impairment across neuropsychiatric disorders has also been linked to the risk variant of the catechol-O-methyltransferase (COMT) gene (the Val158Met polymorphism), which mediates the degradation of dopamine in the synaptic cleft. 22In particular, the Val allele is associated with lower levels of synaptic dopamine in the prefrontal cortex, while the Met allele is related to greater prefrontal dopamine. 23In general, studies have linked low cortical dopamine tonus to cognitive impairments across patients and healthy individuals, 22,24,25 with relatively consistent evidence for poorer attention, working memory, and executive function in Val homozygotes (with low PFC DA tonus) than in Met carriers.In keeping with this, we found that Val (compared with Met) homozygous BD individuals showed reduced recruitment of dorsolateral prefrontal resources during working memory, which was accompanied by impaired task performance. 26Importantly, such task-related dorsal prefrontal hypo-frontality has been proposed as a key neurocircuitry biomarker of cognitive impairment across neuropsychiatric disorders 27 including BD 28 and ADHD. 29,30dications that increase brain dopamine and/or norepinephrine are widely used for treating ADHD to improve cognitive and daily functioning.The stimulant medication methylphenidate, which is the most used ADHD medication, increases extracellular dopamine in the striatum by blocking the DAT. 31,32Furthermore, nonstimulants like atomoxetine and antidepressants used in ADHD, such as bupropion, increase norepinephrine and dopamine levels in the prefrontal cortex. 33Finally, alpha-2-adrenergic agonists, including clonidine and guanfacine, facilitate norepinephrine neurotransmission by stimulating the norepinephrine alpha 2A receptors on prefrontal cortical neurons. 34[37] Pharmacologic upregulation of dopamine and/or noradrenaline may thus be a promising avenue to target the hypo-frontality to improve cognitive functions in BD.Nevertheless, there is a paucity of studies investigating how these medications influence cognition in BD.This is due to an ongoing concern that psychostimulant medications could trigger or aggravate manic symptoms in patients with BD. 38,39 The assumption linking ADHD hyperactivity and bipolar mania lacks empirical evidence, mainly relying on animal studies.In contrast, the "vigilance regulation model of mania" suggests that unstable wakefulness due to low dopamine levels contributes to mania symptoms in both ADHD and BD. 40Manic behavior is seen as an attempt to stabilize wakefulness through external stimulation. 40Following this, it has been suggested that medications enhancing dopamine signaling may reduce mania symptoms by improving vigilance. 40The efficacy towards pro-cognitive effects of modafinil on some cognitive domains.No increased risk of (hypo)mania was observed.Five studies had low risk of bias, eleven a moderate risk, and one a serious risk of bias.
Conclusions: Methylphenidate or mixed amphetamine salts may improve ADHD symptoms in BD.However, there is limited evidence regarding the effectiveness on cognition.The medications produced no increased mania risk when used alongside mood stabilizers.Further robust studies are needed to assess cognition in BD patients receiving psychostimulant treatment alongside mood stabilizers.

K E Y W O R D S
attention-deficit hyperactivity disorder, cognitive impairment, bipolar disorder, ISBD task force, medication, recommendations, systematic review of methylphenidate in acute mania remains controversial, but it does not appear to worsen mania symptoms. 41Corroborating this finding, a large-scale registry-based study found that methylphenidate was safe and not associated with manic switch in BD patients who received concomitant mood-stabilizing medication, although mania risk did increase in people not taking mood stabilizers. 42Similarly, armodafinil was also found to be well tolerated in bipolar depression, in terms of no increase in mania risk. 43,44An open-label study of lisdexamfetamine dimesylate likewise showed that the drug was well tolerated in BD with no increases in manic symptoms. 45ven the similar cognitive impairments in BD and ADHD, overlapping neurobiological underpinnings, and pressing need to identify effective pro-cognitive treatments in BD, 46 it seems timely to reexamine the potential benefits on cognition and ADHD symptoms and the safety of ADHD medications in BD.We did not conduct a quantitative meta-analysis of the available studies due to the significant heterogeneity in terms of intervention characteristics (pharmacological compounds, dose, and combination with other drugs), and population characteristics (child and adolescent versus adult populations, BD with comorbid ADHD versus BD only).Instead, this systematic review by the ISBD Targeting Cognition Task Force aimed to: (I) examine the evidence in the field for potential benefits of medications that are used to treat ADHD, either established or off-label drug therapies, on cognition and/or ADHD symptoms in patients with BD with or without ADHD comorbidity, and (II) investigate the side effects of these medications in BD, with a particular focus on its potential to trigger and/or aggravate mania symptoms.

| ME THOD
This review was conducted in accordance with the PRISMA 2020 statement.The systematic review protocol was registered in the online PROSPERO database (registration number: CRD-42023385497).

| Review question
The PICO framework (Population, Intervention, Comparison, Outcome) was used to structure the review question.This review aimed to critically examine data from randomized and non-randomized controlled trials investigating (a) the efficacy of established or off-label ADHD drug therapies to treat cognitive impairments and/or ADHD symptoms in BD patients with or without comorbid ADHD (BD-cADHD), (b) the side effects of these interventions with a focus on the risk of inducing hypo/manic symptoms.

| Search strategy
We performed systematic computerized literature searches using the databases PubMed, Embase, and PsycINFO up until June 2023.
The search term "bipolar disorder" and the database-specific variant terms were combined with search terms for several classes of psychotropic drugs used for the treatment of ADHD.These classes include medical psychostimulants (methylphenidate and amphetamine derivatives), alpha-2-adrenergic agonists (e.g., clonidine and guanfacine), norepinephrine reuptake inhibitors (e.g., atomoxetine and viloxazine), and pharmacological compounds used off-label in the treatment of ADHD (e.g., bupropion, modafinil, armodafinil).
For the drugs approved for ADHD treatment, the database-specific keywords were used in the search strings.For the most common drugs, popular brand names were added (Ritalin, Concerta, Vivanse, Dexedrine).See Appendix S1 for details.
The included off-label compounds sometimes used to treat ADHD were included because we judged there is enough experimental evidence of the possible efficacy in ADHD treatment to warrant inclusion in the review.This is especially because modafinil and armodafinil are both psychostimulant compounds classified as wakefulness-promoting drugs that increase dopamine in the brain 47 like approved ADHD psychostimulant drug therapies.Modafinil had enough evidence backing it for FDA approval as a drug to manage ADHD in children and adolescents but was rejected due to concerns about dermatological side effects. 48There is also experimental evidence that modafinil could be effective in treating ADHD in adults. 49While there, to our knowledge, are no studies investigating armodafinil, it is a longer lasting isomer of modafinil, 47 which makes it pharmacologically similar enough to modafinil to warrant inclusion in the present review.Like atomoxetine which is approved for ADHD treatment, bupropion also increases the availability of dopamine and norepinephrine, albeit by a different mechanism of action than atomoxetine.There is experimental evidence that bupropion is effective in treating ADHD. 50The combination of the similar neuropharmacological effects of bupropion to other approved ADHD drug therapies, and the experimental evidence of its treatment efficacy for ADHD led us to include bupropion in our review.As for other drug classes, such as dopamine agonists and COMT inhibitors that increase dopamine and/or norepinephrine in the brain, we are unaware of any evidence in human participants, experimental or otherwise, of efficacy in treating ADHD.Furthermore, a recent study of the dopamine agonist pramipexole suggested that the compound had no pro-cognitive efficacy in BD patients. 51For these reasons, dopamine agonists and COMT inhibitors were not included in the present review.
The relatively broad search strategy was chosen for two reasons.
First, because of the suspected paucity of studies investigating possible pro-cognitive effects of medications that increase dopamine and/or norepinephrine in BD.Second, reviewing possible side effects of these compounds in addition to their cognitive benefits was considered important because of a frequent concern that the compounds could induce manic/hypomanic symptoms.
Two authors (RG and ZO) independently conducted a primary title and abstract screening to identify potentially eligible articles, and a subsequent full-text screening.The systematic literature searches were supplemented using Google Scholar's 'Related Articles' function, as well as manual citation searches using the included articles.Conflicts were resolved through discussion, and if needed, through further discussion with two other authors (KWM, GH).
Screening, selection, data extraction, and risk-of-bias assessments were conducted using the Covidence platform. 52roughout the screening phases, the eligibility of the articles was evaluated based on their accordance with our inclusion/exclusion criteria.

| Selection criteria
Studies were included if they (a) examined changes in neurocognitive function or ADHD symptoms as a primary, secondary, or tertiary The following inclusion criteria for the studies were applied: (a) participants had BD with or without ADHD comorbidity in any phase of their illness (depressive, manic, remitted, or mixed); (b) the study design involved a control group of participants who received mood stabilizing drugs, placebo, or another psychopharmacological intervention; (c) the study findings were reported in peer-reviewed articles, as defined either at the journal website or noted in the article with information on when it was received, revised, and accepted.
In addition, relevant studies by ISBD Targeting Cognition Task Force members that were under review at the time of writing were also included; (d) the articles were written in English.Studies were excluded if they examined mixed samples with several diagnoses, unless data for BD with or without comorbid ADHD were reported separately, and if they were case reports or series.

| Data extraction
RG and ZO extracted the following predefined data items of interest: author, title, year of publication, participant details (including sample size, diagnosis, psychiatric comorbidities, and medication status), intervention details (e.g., dose, duration, frequency), cognitive measures (scores in various cognitive tasks or ADHD symptoms), as well as functional and mood scale measures, and side effects.

| Quality assessment
Included studies were assessed for risk-of-bias using either the Revised Cochrane Risk-of-Bias Tool for Randomized Trials (RoB2), 53 the RoB2 crossover version 53 or the Risk-of-Bias in Non-Randomized Studies-of Interventions (ROBIN-S) assessment tool. 54RG and ZO independently evaluated included studies using the appropriate, above-described tools and subsequently reached consensus following the Covidence protocol.Only outcomes which correspond to this review's primary or secondary outcome were assessed for bias, regardless of the respective study's priority of outcomes.To aid RoB assessments, trial protocols, statistical analysis plans, and non-commercial trial registry records were obtained whenever available.All RoB assessments were of the intent-to-treat (ITT) analyses of the effects.

| RE SULTS
After the removal of duplicate hits, the literature searches yielded 1957 articles.These articles were subjected to a primary title and abstract screening (primary screening), which excluded 1927 articles, leaving 30 studies for full-text review (secondary screening) (see PRISMA flowchart in Figure 1).One study was included via other methods, namely a study by a task force member that was under review at the time of writing.Seventeen studies met our inclusion criteria and were included in the review (see Tables 1 and 2).

| Methylphenidate
In the methylphenidate studies, a total of 66 participants received the active immediate release compound, while 54 received placebo.This includes two crossover studies in children/adolescents, 57,60 where 38 participants went through both active treatment and placebo conditions.Dosages of methylphenidate ranged from five to 40 mg/day.In all studies, methylphenidate was given twice a day, generally in the morning and at midday.None of the methylphenidate studies were sponsored by the pharmaceutical industry.In two cases, 41,57 the study medication was donated by pharmaceutical companies (Abbott Laboratories, Chicago, IL, USA; Medice, Iserlohn, Germany).
One study examined the effects of three weeks of methylphenidate administration in combination with mood stabilizing medications, mostly a combination of lithium and divalproex sodium, given for at least five days prior to study start. 57The study was conducted in a sample of 16 euthymic children and adolescents with BD and ADHD.Change in ADHD symptoms was defined as the primary outcome with change in affective symptoms being the secondary outcome.Participants were allowed clonidine as a soporific agent at the discretion of research clinicians but not tricyclic antidepressants or antipsychotic medication.Participants received five mg twice daily, then 10 mg twice daily, and then 15 mg twice, each for 1 week, over a course of three weeks.The study was crossover-and placebocontrolled and double-blinded.Methylphenidate was significantly more efficacious than placebo in reducing ADHD symptoms, with a large effect size (d = 0.9), measured with the ADHD Rating Scale A second study examined methylphenidate compared with placebo in a double-blind design for treating 42 adult patients with acute mania. 41Severity of manic symptoms was defined as the primary outcome, while cognitive change was a secondary outcome.None of the crossover patients achieved antidepressant response.
Including the crossover patients, the overall efficacy was 50% and 55% for desipramine and bupropion, respectively.
In the acute phase 30% of subjects receiving desipramine switched into (hypo)mania, while 11% of subjects receiving buproprion did so.Over the entire study period the switch rates for desipramine were 50% (N = 5), and 11% for bupropion (N = 1).In sum the antidepressant efficacy of bupropion and desipramine did not differ significantly.However, the findings suggest that bupropion is less likely to induce (hypo)mania than desipramine. Abbreviations

TA B L E 2 (Continued)
eight weeks of open-label treatment (with no placebo control) with either lithium, carbamazepine, or divalproex sodium after which those presenting with predominating ADHD symptoms were given adjunctive methylphenidate and, in some cases, a second mood stabilizer in a 16-week open-label continuation phase. 58These participants were given 5-10 mg of methylphenidate twice daily (titration procedure not specified).The treatment was found to be efficacious in reducing ADHD symptoms measured using the Clinical Global Impressions-Improvement scale, with an overall reported response rate of 92% (n = 12/13 of participants receiving adjunctive methylphenidate) defined as a score of 1 or 2 (no further statistics were specified regarding ADHD symptoms).
The fourth study examined low-dose methylphenidate adjunctive to aripiprazole for euthymic child-and adolescent participants with BD and comorbid ADHD over a 4-week period. 60e study included 14 participants and utilized a placebo In summary, two studies showed that methylphenidate treatment adjunctive to mood stabilizers improved ADHD symptoms in pediatric BD populations.In contrast, one study found no effect on ADHD symptoms of methylphenidate adjunctive to aripiprazole in a pediatric population.Finally, one study examined the antimanic and cognitive effects of short-term methylphenidate for adults with acute mania but found no treatment effects on cognition of mania symptoms.For details on these studies, see Table 1.

| Other psychostimulants
One study to date investigated the effects of other psychostimulants on ADHD symptom change in BD.This placebo-controlled, crossover study investigated the effects of four weeks of mixed amphetamine salts (5 mg/daily) adjunctive to divalproex sodium on ADHD symptoms and manic symptoms in 30 pediatric patients with BD and comorbid ADHD. 59Patients were eligible for inclusion if they were in a (hypo)manic or mixed affective state at the start of the first study phase where divalproex sodium was given as monotherapy.Participants who were responsive to treatment, defined as 50% or greater reduction in symptoms measured with the YMRS, were subsequently crossover into the phase where they received mixed amphetamine salts in addition to divalproex sodium.Patients were required to not have taken any other psychotropic drugs for two weeks or fluoxetine for four weeks prior to study start.The study found that mixed amphetamine salts were significantly more effective in reducing ADHD symptoms than placebo as measured with the Clinical Global Impressions-Improvement scale (CGI-I).The authors reported that the improvement from divalproex sodium plus mixed amphetamine salts was 1.9 points greater on the CGI-I than when patients were taking divalproex sodium plus placebo.The difference was significant (p =< 0.0001).The reported mean CGI-I score while patients were taking mixed amphetamine salts was 1.8 (SD = 0.6) while the mean score when taking placebo was 3.7 (SD = 1.0).There were no significant changes in manic symptoms associated with mixed amphetamine salts compared to placebo.
Another study of the effect of other psychostimulants exam- erally safe in terms of (hypo)mania risk, and side effects.However, the authors noted that some people may not tolerate modafinil well, as two people discontinued the drug during the trial.Additionally, the study found near significant positive effects on some domains of cognitive function, specifically processing speed and verbal learning, as well as increased daytime wakefulness.However, a negative effect of modafinil was found on sleep quality.

| Interim summary of cognitive or ADHD symptom change
Three out of four studies examining change in ADHD symptoms showed positive effects of ADHD drug therapies.These three positive studies were all conducted in children or adolescents; two studies were of methylphenidate, while one was of mixed amphetamine salts.Of the three studies investigating objective neurocognitive change, one found no effect of short-term (2.5 days) methylphenidate in acute mania on the SCIP, another found no effects of clonidine on the MMSE, while a third found marginally significant pro-cognitive effects on the MCCB processing speed and verbal learning domains.
These findings indicate that psychostimulant treatment with methylphenidate or mixed amphetamine salts adjunctive to mood stabilizers may lead to ADHD symptom improvements, whereas there is insufficient evidence to make any conclusions regarding efficacy on cognition.However, caution should be taken in interpreting the results as it is a very small number of studies.

| Psychostimulants
Twelve studies examined the side effects of psychostimulants given either as an add-on to mood stabilizing medication to a total of 1775 participants 43,[56][57][58][59][60][61][62][63]65,66 or as monotherapy for 42 participants in one study. 41 In11 of 12 studies, the control condition was placebo under double-blind conditions, while baseline measures and alternative non-ADHD pharmacotherapy served as the control in a final study.58 In all 12 studies, psychostimulant treatments were found to be well tolerated, with no study reporting a statistically significantly increased risk of (hypo)manic symptoms compared to the control condition (see Table 3 for details).Side effects of special interest for the review purposes were presence of/worsening of hypomania or mania.In total, 11 patients (1.3%) in the treatment groups versus eight (0.9%) experienced hypomania across the studies. Elven (1.3%) in the intervention groups versus nine (1%) in placebo groups became manic.
Overall, side effects of psychostimulants were similar across studies (see Table 3).11 of 12 studies reported side effect data, while one study merely stated that the active treatment (methylphenidate) was well tolerated. 58Another study of adjunctive methylphenidate did not report side effects in such a way where it is possible to determine how many participants and what percentage experienced the measured side effects. 60Yet, another study reported side effects in a way where it was not possible to determine which condition (active/control) the side effect occurred in. 59These three studies are thus not represented in the numbers in Table 3, but all showed that there were no significant differences in side effects between the active and control conditions.
Transient headache was the most frequently reported side effect experienced by 110 (14.6%) participants out of the total number of 751 participants receiving the active psychostimulant treatments across the nine studies detailed in Table 3.Of the 850 participants receiving placebo, 80 (9.4%) also reported experiencing transient headache.
Nausea, insomnia, and diarrhea were the next most reported side effects in decreasing order of frequency, with some studies finding that one or more of these were more common in the active conditions versus placebo (see Table 3 for details).However, only two studies employed a psychometrically validated measure of side effects. 57,60

| Nonstimulant ADHD drug therapies
Side effects of nonstimulant ADHD drug therapies are summarized in Table 4.The five studies of side effects of nonstimulant ADHD drug therapies, all suggest that these treatments are well tolerated.
Specifically, adjunctive clonidine, and adjunctive bupropion, as well as bupropion monotherapy did not increase the risk (hypo)mania in the BD cohorts when compared to the studies' respective control conditions.
In the single study of adjunctive clonidine in 70 acutely manic patients, the authors did not report side effects explicitly. 55Thus, the study is only relevant to the side effects of special interest insofar as it showed that clonidine was associated with a greater decrease in manic symptoms when compared to placebo, with a medium effect size.
[69] One study with 14 participants did not report explicitly on side effects other than stating that, overall, both bupropion and idazoxan were well tolerated and that there were no distinguishable differences between them in terms of side effects. 64However, the authors state that a psychometrically validated side effect measure was used.
Another study with 36 participants reported specific side effects (see Table 4). 67In this study, comparing bupropion to topiramate, no patients experienced a manic switch, and difficulty sleeping was the only side effect that was more frequent in the bupropion group. 67psychometric measure of side effects was also used in this study.
A third study with 184 participants reported on the risk of affective switch into (hypo)mania using two different measures with different thresholds for determining when a switch had occurred. 68ing the YMRS, a manic switch was defined as a score of ≥13, while it required a score of ≥3 using the CGI-BP.Using the YMRS criterion, 4% of participants receiving bupropion switched into mania or hypomania, compared to 7% and 15% for sertraline and venlafaxine, TA B L E 3 Psychostimulant side effects.respectively. 68Using either the YMRS or the CGI-BP criterion, the percentage of participants who met one of the thresholds for a manic switch were 14%, 16%, and 20%, respectively.Post hoc analyses demonstrated a lesser risk of hypo/mania with bupropion than venlafaxine, with no differences between sertraline and venlafaxine or bupropion and sertraline. 68fourth study with 15 participants reported on the risk of affective switch into (hypo)mania using the YMRS, but without specifying the threshold scores they used to determine (hypo)mania. 69The study was designed such that patients underwent an 8-week acute treatment phase followed by a continuation phase that lasted up to 1 year, or until patients met DSM-III-R criteria for hypomania, mania, or depression.The study found that 30% (N = 3) of those receiving desipramine and only 11% (N = 1) of those receiving bupropion developed mania during the 8-week acute treatment phase.During the continuation phase, two desipramine-treated and no bupropiontreated patients experienced hypomania or mania, making the incidence of affective switch over the entire study 50% (N = 5) and 11%

Study
(N = 1) for the desipramine and bupropion patients, respectively.

| Summary of side effects
Overall, the studies suggest that psychostimulant treatments and alternative ADHD drug therapies such as clonidine and bupropion are well tolerated, especially when given as adjunctive interventions to mood stabilizing medications, and that they do not impose any increased risk of inducing (hypo)mania in BD patients when compared to placebo or other control conditions.Notably, this evidence comes from a vast majority of studies in which these medications were given in combination with mood stabilizing medications (n = 895 participants) but also from a small sample of patients who received them as monotherapy (n = 56).Hence, it can be concluded with greatest confidence that the compounds are safe regarding induction of mania when used in combination with mood stabilizers.
Regarding non-psychiatric side effects such as transient headache and diarrhea, it is uncertain how most of the studies monitored these side effects, which makes the validity of the reported differences between treatment and placebo conditions less certain.

| Risk-of-bias assessments
The risk-of-bias (RoB) assessment was conducted with three different tools: the RoB2, RoB2 crossover version, and the ROBINS-I for the RCTs, the crossover RCTs, and the non-randomized studies, respectively.The results of these assessments are displayed in Figure 2A-C, respectively.Five of the 17 studies (29%) were rated as being at low risk of bias, 55,57,60,66,68 while 11 studies (65%) [64][65]67,69 The one study that was rated as being at serious risk of bias was the non-randomized study of methylphenidate 58 (note that "serious" is the second worst rating in ROBINS-I with "critical" being the worst).
[64][65]67,69 However, we identified no baseline imbalances between intervention and control groups in these studies that could have suggested problems with the randomization process.The only other bias domain among the RCTs that was rated "some concerns" was the measurement domain in a study of methylphenidate in patients with acute mania 41 because it was considered that two days was too short a time for measurable cognitive change to occur.In the non-randomized study of methylphenidate, 58 the greatest source of bias was confounding bias, which limited the ability to determine whether the observed effects were due to the intervention or other factors that were not controlled.

| DISCUSS ION
This systematic review by the ISBD Targeting Cognition Task Force investigated the possible benefits on cognition or ADHD symptoms and studies, of which two were negative and one identified marginally significant pro-cognitive effects, investigated cognition.Regarding Aim (II), the use of ADHD medications was safe in BD regarding mania symptoms, with none of the studies indicating any risk of mania switch in these patients who mostly received concomitant mood stabilizing treatment.The studies were generally characterized by low or moderate risk of bias, of which the latter was mainly due to a lack of details on the randomization procedure.
A common concern in the clinic is that stimulant medications carry a risk of manic switch or worsening of mania symptoms in patients with BD.Therefore, BD patients who previously received ADHD medications often have this medication removed when they get their BD diagnosis.This clinical practice contrasts with the observed lack of evidence for mania-inducing effects of ADHD medications in the present and previous systematic reviews. 42,43portantly, most of the studies included in the present review involved patients who received concomitant mood stabilizing medications.Therefore, based on the lack of evidence for mania risk in these 12 studies, we can only conclude that ADHD medication does not seem to be counter indicated in BD patients when they receive concomitant mood stabilizing mediations, since no increased mania risk was identified in these cases.However, it must be remembered that the data with regards to manic switch risk with commonly used ADHD medications, such as methylphenidate and amphetamine salts, is limited.Of other side effects reported in the reviewed studies of ADHD medications, transient headache was most common and mainly related to armodafinil treatment.Other common side effects were nausea, insomnia, and diarrhea.Sleep problems may be important because sleep and circadian rhythms are a common symptom in BD. 70 When considering ADHD medications for BD patients, careful consideration should also be given to irritability, a mixed/unspecific symptom in BD that may be enhanced by stimulants.Furthermore, anxiety symptoms, which are common in BD patients, could also be worsened by stimulants. 71Future studies of ADHD medications in BD should therefore include acceptability/ tolerability aspects such as worsening of headaches, insomnia, irritability, and anxiety.Notwithstanding these considerations, adjunctive ADHD medications to mood stabilizers seems to be safe in BD patients with ADHD comorbidity based on the reviewed studies.
While administration of ADHD medications is a viable option from a safety perspective, there is a paucity of studies into its benefits for treating concomitant ADHD symptoms and cognitive impairments in BD.Notably, three of the four studies of ADHD symptoms found remarkable benefits with either a large effect size or a large response rate (92%) of methylphenidate or amphetamine salt treatment in pediatric or adolescent populations with BD. 57,58,60 We therefore consider add-on of these ADHD medications to mood stabilizers as a viable option for BD patients with comorbid ADHD symptoms based on their beneficial effects on ADHD symptoms, good safety profile (i.e., no increase in mania risk) when taken together with mood stabilizing medication, and low risk of bias in these studies (low: k = 2, moderate: k = 1).However, we recommend that more research is conducted in adult BD patients with ADHD comorbidity given that the extant studies were mostly conducted with pediatric or adolescent populations.
In contrast, there is insufficient evidence at this stage for recommending ADHD medications to treat cognitive impairments in BD.Indeed, two of the three studies investigating cognitive change with objective neuropsychological measures showed no benefits of methylphenidate on the SCIP 41 or of clonidine on the MMSE performance, 55 whereas a study of modafinil showed only marginally significant procognitive effects on MCCB processing speed and verbal learning domains. 56Nevertheless, major methodological limitations may have obscured possible cognitive benefits.The initial methylphenidate study, focusing on cognitive change in SCIP among manic patients over a 2.5-day period, 41 presents potential confounding factors arising from acute mania symptoms. 46,72Moreover, this timeframe is likely insufficient for observing treatment-related cognitive improvements, as suggested by the ISBD Targeting Cognition Task Force recommendations. 72The absence of specified timing for SCIP assessments also raises concerns about potential diurnal cognitive effects.Additionally, the limited sensitivity of SCIP to executive dysfunction, that is common in BD and ADHD, further complicates interpretation.In the second study involving clonidine, 55 the use of the MMSE dementia screening tool is a limitation.While suitable for older populations when dementia is suspected, such tools prove suboptimal in younger, higherfunctioning patients due to ceiling effects. 73Moreover, the focus on the memory domain neglects executive dysfunction in BD and ADHD.
The third study that demonstrated borderline significant pro-cognitive effects of modafinil, involved a very small sample (N = 12), making the findings hypothesis-generating in nature. 56ven these limitations, we encourage further studies of the possible cognitive benefits of ADHD medications in BD, which employ more comprehensive cognitive assessments and larger samples.We also suggest caution when interpreting the findings regarding possible pro-cognitive effects of the medications given the highly preliminary state of the evidence.Based on the extant evidence, our suggested algorithm in the clinical treatment of BD patients is therefore as follows: When mood stabilization for a patient has been obtained with mood stabilizers, add-on treatment with an ADHD medication in those with comorbid ADHD symptoms may be considered to target these symptoms.In contrast, for BD patients without comorbid ADHD symptoms, more research into the possible cognitive benefits is needed before such medications can be recommended.
There is a clear impetus for more research into the cognitive benefits of ADHD medications as add-on to mood stabilizing treatment in symptomatically stable patients with BD who experience cognitive impairments.Indeed, there is a clear knowledge gap with only two extant studies on the topic, and-in contrast with the common concern of mania risk-these medications were found to have a good safety profile when given as add-on to mood stabilizing medication.
We therefore encourage studies of the pro-cognitive potential of these medications in BD given the pressing need for pro-cognitive treatments in the 40%-70% of BD patients with persistent cognitive impairments despite symptomatic remission.For the design of such studies, researchers are recommended to follow the recently updated methodological recommendations for cognition trials that include suggestions for sample size estimations, pre-screening of cognitive impairment, selection of outcome measures, and strategies to aid transfer to daily functioning. 46Among the key recommendations are (i) to pre-screen participants for objective cognitive impairments with a brief cognition screening battery, (ii) to generally include partially or fully remitted patients in trials to minimize "pseudospecificity" issues, (iii) to preselect one (broad) cognition measure as the primary outcome, such as a composite measure of working memory, sustained attention and executive function in trials investigating ADHD medications, (iv) in general to administer medications for several weeks, and (v) to use an add-on design if the candidate medication does not have a mood stabilizing effect for ethical reasons and to ensure generalizability. 46 conclusion, we identified preliminary evidence for efficacy of adjunctive ADHD medications in BD patients with comorbid ADHD symptoms, with robust beneficial effects of methylphenidate and one of amphetamine salts in three of the four studies.Importantly, none of the identified studies reported any heightened risk of mania switch in these BD patients who mostly received concomitant mood stabilizing treatment, although the study samples sizes were small.This indicates that adjunctive ADHD medications may be a viable option for some patients with BD with comorbid ADHD.Given this good safety profile of the medications and paucity of research into their possible cognitive benefits, we encourage further research into the pro-cognitive potential of these medications in BD.

A FFI LI ATI O N S
outcome and/or (b) investigated side effects of the interventions in BD.Cognitive change was operationalized as changes in objective neurocognitive test performance, while ADHD symptom change was measured with ADHD symptom rating scales (see Appendix S1, for details on the represented ADHD symptom measures).

(
ARS-IV) (primary outcome).Paired sample t test also revealed significant reductions compared to placebo on the Conners Parent Rating Scale 48 (CPRS-48) Conduct Problems subscale (p = 0.05), Impulsivity-Hyperactivity subscale (p = 0.02), and Hyperactivity Index subscale (p = 0.02) (secondary outcomes).There were no significant differences between methylphenidate and placebo on mood symptoms measured with the Young Mania Rating Scale (YMRS) or the Children's Depression Rating Scale Revised (CDRS-R).

Patients received 15
mg of methylphenidate at 10 a.m. and 3 p.m. on Day 1, 20 mg at 9 a.m. and 3 p.m. on Day 2, and 20 mg at 9 a.m. on Day 3. The timing of the cognitive assessments was not specified.91% of participants received concomitant lithium, valproate, carbamazepine, antipsychotics, or benzodiazepines.Participants were excluded if they had taken MAO-inhibitors within 14 days prior to study start, fluoxetine within six weeks, or other classes of primarily psychotropic drugs within the prior week. 41Cognitive change was measured with the Screen for Cognitive Impairment in Psychiatry (SCIP) from the baseline assessment to on the final of treatment after the initiation of methylphenidate/placebo treatment.No differences between methylphenidate and placebo groups were found in cognitive change or in mania symptoms, and no significant side effects were detected.In a third study of combination pharmacotherapy, 35 child and adolescent participants with BD and comorbid ADHD underwent F I G U R E 1 PRISMA flowchart.Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al.The PRISMA 2020 statement: an updated guideline for reporting systematic reviews.BMJ 2021;372:n71.doi: 10.1136/bmj.n71.For more information, visit: http:// www.prism a-state ment.org/ .TA B L E 1 Studies with stimulants or stimulant-like substances.
-and crossover-controlled, double-blind design.Primary outcomes were the change in ADHD symptoms and change in manic symptoms.Participants were excluded if they had received any other psychotropics within 10 weeks prior to the study.Patients treated with aripiprazole and low-dose methylphenidate (0.3-0.7 mg/kg/ day) showed no significant reduction ADHD symptoms, measured using the Swanson, Nolan and Pelham Questionnaire (SNAP IV), compared to aripiprazole and placebo.No significant changes in manic symptoms were detected.
the safety of 100-200 mg/day of modafinil adjunctive to mood stabilizing medication, with cognitive change and sleep measures as secondary outcomes. 56This 8-week, randomized, double-blind, placebo-controlled trial was conducted with an adult sample of 12 euthymic patients with BD who also presented with subjective sleep problems and/or clinically significant cognitive impairments.Patients were not allowed to take more than three psychotropic medications, or any drug known to interact with modafinil.Patients were also not allowed to take any medications with known adverse cognitive effects (topiramate, tricyclic antidepressants, and anticholinergics), compounds that may enhance cognition (amphetamine and dopamine agonists), or benzodiazepines.Cognitive change was measured with the MATRICS Consensus Cognitive Battery (MCCB), with the primary cognitive outcome being the MCCB composite score from baseline through Week 4, and follow-up at Week 8.The secondary cognitive outcomes were the MCCB domain scores, including processing speed, attention/vigilance, working memory, verbal learning, visual learning, reasoning, and problem solving, and social cognition.The study found that adjunctive modafinil was gen- One double-blinded, placebo-controlled study investigated the use of clonidine (0. 2-0.6 mg/day divided in two daily doses) or placebo in combination with mood stabilizing medication (lithium or valproate sodium) for cognitive change as measured with the dementia screening tool, the Mini-Mental State Examination (MMSE) for 70 individuals with acute mania.55In addition to cognitive change, the study examined the interventions effects on manic symptoms and sleep disturbances, with none of the outcomes prioritized over others.In total, 36 manic participants received clonidine, while 34 received a placebo.Participants were allowed to also receive concomitant antipsychotics, but not anticholinergic drugs or tricyclic antidepressants.No differences in cognitive change were found between the clonidine and the placebo-treated patients.However, significant reductions in the severity of manic symptoms and sleep disturbances were observed, both with medium effect sizes.No eligible studies were found that examined the effectiveness of nonstimulant medications used for ADHD treatment in participants with BD.
safety of ADHD medications in patients with BD.In total, 17 studies were identified (N = 2136) that investigated armodafinil (k = 4), methylphenidate (k = 4), bupropion (k = 4), clonidine (k = 1), lisdexamphetamine (k = 1), mixed amphetamine salts (k = 1), or modafinil (k = 2).Of these, four studies investigated the effects on ADHD symptoms and three studies the effects on cognition.The remaining 11 studies investigated the safety of the medications regarding possible mania induction.Three of the four studies on ADHD symptoms were conducted in pediatric and adolescent populations and one in an adult population.The three studies of cognitive change effects were conducted in adult populations.Studies of safety only were conducted in adult populations.Regarding aim (I), we identified preliminary evidence for efficacy on ADHD symptoms, with beneficial effects on three of the four studies: Two studies of methylphenidate and one of amphetamine salts, all conducted in pediatric or adolescent populations.Only two F I G U R E 2 Risk-of-bias evaluations.[Correction added on 06 May 2024, after first online publication: Figure 2 has been updated in color.](A) Risk of bias: ROBINS-I Unique ID D1 D2 D3 D4 D5 D6 D7 Overall Ketter et al. (2016) Kowatch et al. (2003) D1 = Bias due to confounding Low = D2 = Bias in selection of participants Moderate = D3 = Bias in classification of interventions Serious = D4 = Bias due to deviations from intended interventions Critical = D5 = Bias due to missing data D6 = Bias in measurement of outcomes D7 = Bias in selection of the reported result (B) Risk of bias: RoB2 Unique ID D1 D2 D3 D4 D5 Overall Ahmadpanah et al (2022) Calabrese et al. (2010) Calabrese et al. (2014) Frye et al. (2007) Frye et al (2015) Grossman et al. (1999) Hegerl et al. (2018) Ketter et al. (2015) Lipschitz et al. (2023) McElroy et al. (2015) McIntyre et al. (2002) Post et al. (2006) Sachs et al. (1994) D1 = Bias arising from the randomization process Low = D2 = Bias due to deviations from the intended interventions Some concerns = D3 = Bias due to missing outcome data High = D4 = Bias in measurement of outcomes D5 = Bias in selection of the reported results (C) Risk of bias: RoB2 Crossover version Unique ID D1 DS D2 D3 D4 D5 Overall Findling et al. (2007) Scheffer et al. (2005) Zeni et al. (2009) D1 = Bias arising from the randomization process Low = DS = Bias arising from period and carryover effects Some concerns = D2 = Bias due to deviations from the intended interventions High = D3 = Bias due to missing outcome data D4 = Bias in measurement of outcomes D5 = Bias in selection of the reported results

Reference Population Intervention Comparator Outcome Design Main findings
Studies of nonstimulant ADHD drug therapies.
Placebo pills Change in depressive symptoms measured with the IDS-C30 (primary outcome), MADRS, HARS, CGI-BP, QIDS-SR16, Q-LES-Q-SF (secondary outcomes) tolerability, and adverse events (tertiary outcome) and 15 mg/day Placebo pills Total score on ARS-IV (primary outcome), CPRS-48, CDRS-R, YMRS, CGI-S, and SEBMS (secondary outcomes) IDS-C30 (primary outcome), clinical response (defined as a 50% reduction in IDS-C30 scores), remission (defined as IDS-C30 score <12), CGI-BP, YMRS and side effects (secondary outcomes) 6-week, double-blinded, randomized, placebocontrolled trial Adjunctive modafinil improved symptoms of depression over placebo.The treatment was safe and well tolerated, with only minor side effects Frye et al 2015 Adult patients with Bipolar I and a current depressive episode unresponsive to mood stabilizers.N = 399 Adjunctive armodafinil 150 mg/day Placebo pills Change in depressive symptoms measured with IDS-C30 (primary outcome), CGI-S, GAF, YMRS, C-SSRS-SLV, HAM-A, ISI, and side effects (secondary outcomes) Adjunctive armodafinil had no effect on depressive symptoms in patients with bipolar I disorder, when compared to placebo.However, secondary measures of illness severity and functioning did improve more in patients receiving armodafinil.study Methylphenidate was well tolerated and safe in patients with acute mania.However, it did not reduce manic symptoms or alter cognitive function TA B L E 1 (Continued) TA B L E 2