What would you like to find today?

A meta-analysis of short-term, placebo-controlled, randomized clinical trials (Cortese et al. 2018), looking at both efficacy and safety, supported prescribing stimulants – methylphenidate use in children and adolescents and amphetamine use in adults – as first-choice medications. 

However, these were short-term studies, and they focused on relieving ADHD symptoms. What about longer-term outcomes, especially looking more broadly at functional impairment and overall quality of life? 

Sweden has a single-payer health insurance system that encompasses virtually every resident and is linked to national registers that enable researchers to conduct nationwide population studies. 

A joint Finnish-Swedish research team used Sweden’s registers to study outcomes for all individuals of working age, 16 to 65 years old, living in Sweden who had received a diagnosis of ADHD from 2006 through 2021. The resulting study cohort encompassed 221,714 persons with ADHD. 

The team adjusted for the following confounding variables: Genetics, baseline severity of symptoms, baseline comorbidities, temporal order of treatments (which medication was used as first, second, third, and so forth, including also nonuse of ADHD medications), time since cohort entry, and time-varying use of psychotropic drugs, including antidepressants, anxiolytics, hypnotics, mood stabilizers (carbamazepine, valproic acid, and lamotrigine), lithium, antipsychotics, and drugs for addictive disorders. 

With these adjustments, they discovered that amphetamine treatment was associated with a roughly 25% reduction in psychiatric hospitalization relative to unmedicated ADHD. Lisdexamphetamine was associated with a roughly 20% reduction, dexamphetamine with a 12% reduction, and methylphenidate with a 7% reduction. All four medications are stimulants. 

None of the non-stimulant medications – atomoxetine, guanfacine, clonidine – had any significant effect on psychiatric hospitalization. Nor did modafinil a drug that is not FDA approved for ADHD but is sometimes used when other drugs fail. 

Amphetamine was also associated with the greatest reduction in suicide attempts or deaths, with a roughly 40% decline relative to unmedicated ADHD. Dexamphetamine was associated with a roughly 30% decline and lisdexamphetamine with a roughly 25% decline. The stimulant methylphenidate was only associated with an 8% reduction, and modafinil had no significant effect. 

Surprisingly, non-stimulant medications were associated with significant increases in suicide attempts or deaths: 20% for atomoxetine, 65% for guanfacine, and almost double for clonidine. 

Amphetamine and lisdexamphetamine also reduced the risk of nonpsychiatric hospitalization by more than a third compared to unmedicated ADHD. Dexamphetamine was associated with a risk reduction of more than 25%, methylphenidate with 20% lesser risk.  

The non-stimulant atomoxetine was associated with a roughly 15% reduction in risk of nonpsychiatric hospitalization. But neither guanfacine nor clonidine had any significant effect. 

Turning to work disability, atomoxetine was the only ADHD medication associated with a reduction – a roughly 10% improvement. All other medications had no significant effect. 

The team concluded, “In this cohort study of adolescents and adults with ADHD, the use of medications for ADHD, especially lisdexamphetamine and other stimulants, was associated with decreased risk of psychiatric hospitalizations, suicidal behavior, and nonpsychiatric hospitalizations during periods when they were used compared with periods when ADHD medication was not used. Non-stimulant atomoxetine use was associated with decreased risk of work disability.” 

‍

‍

Noting that “Oxidative stress disrupts the structure and function of neurons in the prefrontal lobe of the brain,” and “Structural and functional impairments in the prefrontal cortex have been shown to be highly correlated with behavioral and emotional problems of ADHD,” a Chinese team at Dalian University set out to systematically evaluate the safety and efficacy of antioxidant therapy in children and adolescents with ADHD. 

The team’s systematic search of the peer-reviewed medical literature identified a total of 48 randomized controlled trials (RCTs) or prospective studies involving 12 antioxidant agents (resveratrol, pycnogenol, omega-3, omega-6, quercetin, phosphatidylserine, almond, vitamin D, zinc, folic acid, ginkgo biloba, Acetyl-L-carnitine) that met criteria for inclusion: 

• Ages 18 or younger. 

• Clinical diagnoses of ADHD. 

• Minimum treatment duration of two weeks. 

• Experimental group received antioxidant treatment. 

• Control group received either a placebo, the stimulant medication methylphenidate, or a different antioxidant or combination of antioxidants. 

Treatment efficacy was measured through ADHD symptom scores using Conners’ parent rating scale (CPRS), Conners’ teacher rating scale (CTRS), ADHD rating scale-parent (ADHD RS-Parent), and ADHD rating scale-teacher (ADHD RS-Teacher), as well as secondary outcome indicators such as the Clinical Global Impressions scale (CGI) and Continuous Performance Test (CPT), relative to controls. 

None of the antioxidant therapies were significantly better than placebo.

One limitation is that no effort was made to assess publication bias. 

These results indicate that antioxidants should not be used for treating ADHD.

‍

A 2021 consensus statement by an international group of scientists and clinicians (Bauer et al.) recommended that pregnant individuals “forego [acetaminophen] unless its use is medically indicated,” due to the potential risk of developmental disorders such as autism and attention-deficit/hyperactivity disorder (ADHD). 

A mostly Swedish research team, collaborating with a U.S. researcher, nevertheless noted that previous studies have been limited by: 

• Confounding by indication, because acetaminophen is taken for infection, fever, and pain (including pain from autoimmune disease), which are themselves risk factors for neurodevelopmental disorders such as ADHD. 

• Confounding by parental health and genetics, because neurodevelopmental disorders are highly heritable. 

• Small sample sizes. 

Sweden has a single-payer health insurance system that includes virtually its entire population, and national registers that enable tracking the health history of mothers and their children, including their children’s siblings. 

The team used the Swedish registers to identify the roughly two-and-a-half million children born in Sweden from mid-1995 through 2019. They were also able to identify all siblings to be able to control for otherwise unmeasured familial and genetic confounding. 

Almost 186,000 of these children were exposed to acetaminophen during pregnancy.  

After adjusting for available known confounders, including (but not limited to) child sex and birthdate, mother’s age and medical history, use of any other painkillers, use of any psychoactive medications, country of birth, residential region, smoking status, highest household education, and disposable income, children exposed to acetaminophen during pregnancy were 7% more likely to be diagnosed with ADHD subsequently than those who were not exposed. 

However, roughly the same results were found for other painkillers, including aspirin, non-aspirin nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, and antimigraine medication.   High doses of acetaminophen did not produce any stronger association with subsequent ADHD than low dosage. 

Moreover, when confining results to siblings – 8,526 children who were exposed versus 87,679 who were unexposed – the association between acetaminophen use during pregnancy and subsequent offspring ADHD vanished altogether (and, again, at every dose level). The associations similarly vanished with every other painkiller medication.  

The Swedish team concluded, “Acetaminophen use during pregnancy was not associated with children’s risk of autism, ADHD, or intellectual disability in sibling control analyses. This suggests that associations observed in models without sibling control may have been attributable to confounding.” 

‍

Antipsychotic medications are used to treat a variety of psychiatric disorders, including schizophrenia, bipolar disorder, sleeping problems, major depression, and severe anxiety. 

Untreated maternal mental illness is associated with poor health outcomes for both mothers and their offspring. On the other hand, one must guard against any potential direct harms of medications on development – including neurological development – of the fetus.  

Because prenatal use of antipsychotics is infrequent, previous observational studies have suffered from small sample sizes that have not enabled precise and reliable assessment of risk. The clinical decision about whether to continue antipsychotic treatment in patients who become pregnant has therefore remained inconclusive. 

In search of more reliable guidance, an international study team conducted a systematic search of the peer-reviewed medical literature to perform the first meta-analysis on this topic.  

They evaluated study quality and only included studies rated “good” or better. 

Identification of ADHD was determined by clinical diagnosis. 

Meta-analysis of four studies encompassing over eight million participants found a slight association. Children exposed to maternal antipsychotics during pregnancy were 11% more likely to be diagnosed with ADHD subsequently.  

But even in observational studies with millions of participants, such associations – especially when slight to begin with – could be due to unmeasured confounders. 

The team therefore compared children with gestational exposure to siblings from the same mother who were not exposed, to address shared genetic and social factors at the family level. 

Meta-analysis of two population-based sibling-matched studies with a combined total of over 4.6 million participants in Denmark, Norway, Sweden, Finland, Iceland, and Hong Kong found no significant association between gestational exposure to antipsychotic medications and subsequent diagnosis of ADHD. 

The team concluded, “Our systematic review and meta-analysis of observational studies indicates that the heightened risks of ADHD and ASD observed in children gestationally exposed to antipsychotics appear to be attributable to maternal characteristics, rather than having a causal relation to the antipsychotic itself.” 

‍

Most previous studies of suicide and self-harm risk among persons with ADHD have focused on adolescents and adults. They’ve also tended to be cross-sectional, analyzing data from a population at a specific point in time. 

An Australian study team took a different approach, conducting a before-and-after study through the birth cohort of the Longitudinal Study of Australian Children (LSAC), comprising 5,107  children who have been followed up every two years since birth. 

The diagnosis of ADHD was based on parents reporting that their child had received a diagnosis of ADHD at or before age ten.  

Suicide and self-harm were defined as children’s self-report at age 14 of any thought or attempt of suicide and self-harm respectively over the past year. 

The team adjusted for the following confounders: socioeconomic status, birth weight, ADHD medication history, maternal education level, maternal age at birth, experience in bullying victimization at age 12, and depression score based on Short Mood and Feelings Questionnaire (SMFQ). 

Of the 5,107 participants, 3,696 had all the valid data required for analysis and were included in the final cohort. Of these, 3.6% were diagnosed with ADHD by age 10. 

With a diagnosis of ADHD at age 10 and all other factors held constant: 

• The odds of suicidal thoughts, plans, or attempts at age 14 increased elevenfold. This was twice as pronounced among boys as among girls. 

• The odds of self-harm at age 14 increased 25-fold. This was more than three times as pronounced among boys as among girls. 

Both depression and exposure to bullying were statistically significant mediators for the relationship. Nevertheless, depression and exposure to bullying each accounted for well under 10% of the overall effect. 

Neither socioeconomic status nor maternal factors had any significant mediating effect on outcomes. 

The authors concluded, “This study provides compelling evidence that children diagnosed with ADHD at the age of 10 years face significantly elevated risks of experiencing suicidal thoughts, planning, or attempts, as well as self-harm, by the age of 14 years, which underscores the critical importance of recognizing and addressing these heightened risks in children with ADHD.” 

‍

Inflammatory bowel disease (IBD) consists of 2 main subtypes: Crohn’s disease and ulcerative colitis. Typical symptoms include abdominal pain, diarrhea, and rectal bleeding. Both are incurable, increase the risk of colorectal cancer, and often affect other organs as well. 

A single earlier study suggested a weak link between childhood-onset IBD and ADHD. 

A Danish research team used its country’s national registers – based on a single-payer national health insurance system that encompasses virtually the entire population – to include all 3,559 patients diagnosed with pediatric-onset IBD from 1998 through 2018.  

The team then matched these individuals five-to-one on age, age of diagnosis, year of diagnosis, sex, municipality of residence, and time period, with 17,795 individuals from the same pool who were free of IBD. 

ADHD was identified based on two criteria: clinical diagnoses in patient records, and methylphenidate stimulant prescriptions in the medications register. 

Overall, the team found no significant association between pediatric-onset IBD and ADHD. The same was true for both Crohn’s disease and ulcerative colitis. 

There were no differences in outcomes for boys or girls. 

There was also no significant association found using only ADHD diagnoses or only methylphenidate prescriptions.  

Among children and adolescents with IBD onset under age 14, there was a borderline significant association, but it was a negative one: They were less likely to subsequently be clinically diagnosed with ADHD or to receive prescriptions for methylphenidate.  

The team concluded, “Remarkably, we found a reduced risk of receiving methylphenidate and being diagnosed with ADHD, which merits further investigation.” 

‍

The first-line treatment for ADHD in both adults and children is stimulant medication such as methylphenidate or amphetamines. These medications function by increasing bioavailability of the neurotransmitters dopamine and norepinephrine within the brain. Some animal studies have suggested these medications could impact gonadal function, and more specifically testosterone production. 

A U.S. study team accessed electronic medical records (diagnoses, procedures, medications, and laboratory values), as well as insurance claims for about 108 million patients from 76 healthcare organizations. They used these to assess the risk of long-term ADHD stimulant medication on developing a diagnosis of testosterone deficiency in males above the age of puberty. 

They compared 20-40-year-old men with a clinical diagnosis of ADHD and long-term exposure to ADHD stimulant medications – including methylphenidate, dextroamphetamine, lisdexamphetamine, amphetamine, and dexmethylphenidate – with ADHD patients who did not receive any medication. 

After adjusting for confounding factors, they compared 17,224 men with a diagnosis of ADHD who had received at least 36 prescriptions of ADHD stimulant medications with an equal number with a diagnosis of ADHD who never received any ADHD medications. 

ADHD patients on long-term stimulant medication had a roughly 1.75 times higher rate of subsequently being diagnosed with low testosterone levels within five years than unmedicated ADHD patients. 

The team also compared 17,217 men with a diagnosis of ADHD who had received at least 36 prescriptions of ADHD stimulant medications with an equal number of men without a diagnosis of ADHD.  

Again, patients on long-term stimulant medication had a 75% higher rate of subsequently being diagnosed with low testosterone levels within five years than matched individuals without an ADHD diagnosis. 

The team concluded, “Long-term ADHD stimulant medication use in men was found to be associated with a significant increase in relative risk for a subsequent testicular hypofunction diagnosis. This difference was found when compared to both those with ADHD not using pharmaceutical therapy and those without ADHD. These results indicate that impaired gonadal function is a potential side effect of stimulant medications.” 

Like other observational studies, this work provides an important signal that must be replicated and validated with other methods, especially those that rule out other sources of confounding not measured in this study.  Moreover, diagnoses of testosterone hypofunction in this study were relatively rare to begin with. The measured 0.5% increase in testicular hypofunction diagnosis for those on long-term stimulant medication versus those not on stimulant medication would only affect roughly one in two hundred of those on stimulant medication. This small increase in risk must be weighed against the well-documented benefits of these medications. 

‍

ADHD is associated with deficits in cognitive functions. These include such executive functions as reaction time, motor and interference inhibition, sustained attention, and working memory. 

To what extent can ADHD medications compensate for such deficits? A recent meta-analysis by a European study team has explored this question. It suggests that while medication cannot completely reverse deficits in executive functions, it can lead to significant improvements. 

Based on consistent evidence from many randomized double-blind controlled trials (RCTs) measuring behavioral improvements, first line treatment for ADHD is with stimulant medication while second-line treatment (for stimulant non-responders, or poor tolerability) is with non-stimulant medication (atomoxetine, viloxazine, guanfacine and clonidine). 

This systematic literature search yielded eighteen RCTs, not all of which covered the same executive functions or medicines. 

Meta-analyses yielded the following results: 

Reaction Time 

Eleven RCTs, encompassing 925 participants, found a small-to-medium effect size improvement with methylphenidate. Variation (heterogeneity) among these studies was moderate, and there was no sign of publication bias. 

Four RCTs with a total of 286 participants similarly reported a small-to-medium effect size improvement with atomoxetine. Again, heterogeneity was moderate, with no indication of publication bias. 

Attention 

Sixteen RCTs, with a combined 1,335 participants, found a medium effect size improvement with methylphenidate. Heterogeneity was moderate, and there was some indication of publication bias. No effort was made to correct for publication bias. 

Three RCTs, encompassing 254 persons, found a medium effect size improvement with atomoxetine. Heterogeneity was moderate, with no evidence of publication bias. 

Inhibition 

Thirteen RCTs, with a total of 1,201 participants, found a small-to-medium effect size improvement with methylphenidate. Heterogeneity was moderate, with marginal indication of publication bias. 

Six RCTs with a combined 753 individuals, reported a medium effect size improvement with atomoxetine. Heterogeneity was high, but there was no evidence of publication bias. 

Working Memory 

Nine RCTs, with a total of 1,025 participants, found a small-to-medium effect size improvement with methylphenidate. Heterogeneity was moderate, with no indication of publication bias. 

Three RCTs with a combined 132 individuals, reported a statistically nonsignificant small-to-medium effect size improvement with atomoxetine. Heterogeneity was moderate, with no indication of publication bias. The nonsignificant outcome may have been due to the much smaller number of participants. 

‍

The team concluded, “these meta-analyses of chronic effects of stimulants and non-stimulants on executive functions in ADHD showed significant improvements with both methylphenidate and with atomoxetine in all cognitive domains tested with relatively similar effect sizes, and no statistical differences between them. The findings hence suggest comparable positive effects of both ADHD medication types on the most relevant executive functions in ADHD, suggesting for the first time that stimulant and non-stimulant ADHD medications, when taking [sic] longer-term, not only improve behavioural symptoms of ADHD, but also improve executive function performance, and to a similar degree.” 

‍

A placebo is a pill that does not contain any active medication.  It is given to patients who form the control group in clinical trials.  Comparing the effects of a treatment with placebo is essential because some patients will improve with the passage of time and some will get better due to the expectation of benefit they have from being enrolled in a clinical trial.

In studies of psychiatric conditions, patients in placebo groups typically show improvement. This can be induced by combinations of hope, suggestion, expectation, and consumption of what are presented as medications. It is reinforced by the context of receiving compassionate care from others, with supportive conversations. 

A 2005 study found that placebo response is unequally distributed across psychiatric disorders, but did not address several disorders (including bipolar disorder) examined in the present meta-analysis conducted by a German research team. 

Using only high-quality randomized clinical trials (RCTs) across major psychiatric diagnoses, the team quantified differences in the change of disorder symptoms within placebo groups.  

They selected nine common and clinically significant psychiatric conditions: major depressive disorder (MDD), mania (bipolar disorder), schizophrenia, obsessive-compulsive disorder (OCD), attention-deficit/hyperactivity disorder (ADHD), generalized anxiety disorder (GAD), panic disorder, posttraumatic stress disorder (PTSD), and social phobia. For each of these, they selected the ten most recent high-quality RCTs of medicationsfor meta-analysis. 

Of the ninety included RCTs, the team only looked at placebo groups. Because RCTs for the different diagnoses used differing established psychopathology rating scales, standardized pre-post effect sizes were used to compare outcomes across diagnoses. 

Meta-analysis of the ten ADHD RCTs with a combined total of 1,189 participants reported large effect size improvements in symptoms, with no variation (heterogeneity) across RCTs and no sign of publication bias. 

By contrast, the placebo effect size improvements in symptoms of major depressive disorder (10 RCTs, 1,598 participants) and generalized anxiety disorder (10 RCTs, 1,457 participants) were very large, well above those for ADHD, and with no overlap of 95% confidence intervals. 

At the other end of the spectrum, the placebo effect size improvements in symptoms of schizophrenia (10 RCTs, 888 participants) were moderate, well below those for ADHD, and with no overlap of 95% confidence intervals. 

There were absolutely no indications of publication bias. 

The team noted, “In all diagnoses, there were improvements in symptom severity during placebo treatment (ie, the lower limit of the 95% CIs of the pooled pre-post placebo effect sizes were >0).” Although they stated, “The large and robust improvements observed in ADHD studies have not been reported to our knowledge.”  they seemed to have missed this article by me and my colleagues:  https://pubmed.ncbi.nlm.nih.gov/34232582/. 

They also concluded, “Comparing the courses of different disorders under placebo indirectly may assist in understanding disease etiology, possibly providing insights into the proportionate influence of organic and psychogenic factors. Conditions with presumed substantial hereditary and biological components, such as schizophrenia, exhibited modest placebo responses in our analysis. Conversely, disorders with potentially less biological contribution, eg, depression and GAD, showed stronger responses. Our study may serve as an initial framework for incorporating the comprehensive insights derived from placebo groups of controlled trials into the etiopathogenetic exploration of mental illnesses.”

‍

Quality of life (QoL) is defined as a person’s satisfaction with their life, measured across several dimensions including psychological, social, health, biological, and economic well-being. For adults, these are usually self-reported. QoL for children and adolescents is usually reported by parents.  

Medications for ADHD include stimulants (methylphenidate and amphetamines) and non-stimulants (e.g., atomoxetine, clonidine, guanfacine, viloxazine). As QoL is related to ADHD symptoms’ severity, management of ADHD via medication could improve not only core symptoms but also QoL in people with ADHD.  

Noting the absence of meta-analytic evidence on the effects of ADHD medications on QoL, an international research team conducted a systematic review and meta-analysis of parallel or cross-over randomized clinical trials (RCTs) to estimate the effects of ADHD medication on QoL. They also performed secondary analyses to see if these effects differed in children and adolescents versus adults, as well as by class of medications, and if they were moderated by length of treatment. 

Meta-analyses of four RCTs with a combined total of 950 participants with ADHD (45% adults) found a medium effect size improvement among those receiving amphetamines by comparison with those receiving placebo. There was no sign of publication bias, but there was wide variation (heterogeneity) in effect size estimated among the studies.  

Meta-analysis of four RCTs with a combined total of 1,094 participants with ADHD (57% adults) found a small-to-medium effect size improvement among those receiving methylphenidate by comparison with those receiving placebo. Again, there was no sign of publication bias, but wide variation in effect sizes among the studies. 

Due to lack of sufficient data, the team could not explore whether length of treatment affected the results, or if there were differences between children/adolescents and adults. 

Finally, a meta-analysis of eleven RCTs with a combined total of 3,344 participants with ADHD (63% adults) likewise found a small effect size improvement among those taking atomoxetine compared with those receiving placebo. Once again, there was no sign of publication bias, but wide variation in effect sizes among the studies. 

The team was able to establish that for atomoxetine treatment, length of intervention – the studies ranged from 6 to 24 weeks – had no significant moderating effect. Similarly, they found no significant differences in effect on children and adolescents versus adults.  

A single RCT evaluating modafinil treatment in adults found no improvements at any dose, whereas a single RCT testing non-stimulant guanfacine reported a medium effect size improvement in QoL.  Modafinil is not FDA approved for ADHD but is sometimes used as a last resort if other treatments fail.

The team concluded that the FDA approved medications for ADHD were significantly more efficacious than placebo in improving QoL in people with ADHD.  The improvements in Q0L were, however, smaller than what has been found for improvements is the symptoms of ADHD (inattention, hyperactivity, impulsivity).  More work is needed to detect differences by age and length of treatment. 

‍