• A-Z
  • Directory
  • myUVM
  • Loading search...

Home – Child Mental Health Blog

Very High Rates of Psychotic Symptoms Linked to Stimulant Medications

Posted: December 28th, 2015 by David Rettew

(Note: portions of this review were recently published in an article in Clinical Psychiatry News)

While it is known that stimulant medications used to treat ADHD might occasionally to lead to psychotic symptoms, this side effect is considered to be rare and usually associated with either medication abuse or very high dosages.  A report just released today in the journal Pediatrics, however, now calls into question how rare psychotic symptoms from stimulants might actually be.

The data for this study come from the Families Overcoming  Risks and Building Opportunities for Wellbeing study.  The sample size was small but thoroughly evaluated. A total of  141 children in Ritalinwhich at least one of the parents was diagnosed with a major mental illness, mainly major depressive disorder and bipolar disorder but also schizophrenia (5%).  The mean age of the children was about 12 years old.  Of these children, 24 (or 17%) had taken stimulant medication.  Psychotic symptoms or psychotic like experiences were assessed through four different methods including an instrument called the “Funny Feelings” interview with psychotic symptoms counted as present any of their four ascertainment methods yielded a positive result.  The main analysis was a logistic regression examining the association between lifetime use of stimulant and lifetime use of a psychotic or psychotic like experience.   The analyses controlled for age, gender, and family history.

The major finding of the study was that psychotic symptoms were quite common, present in one-third of this sample overall, and were found to be more likely among those children taking stimulants.  A total of 62.5% of children taking stimulants versus 27.4% of children who had not, reported some kind of psychotic symptoms.  After account for some potential confounding variables, the odds ratio related to taking stimulants was 4.41 which was found to be statistically significant.  The odds ratio related to ADHD itself and psychosis was 1.98 but this was not significant due to the small sample size.  When looking at current stimulant use, the rate of current psychotic symptoms was 27%. The type of psychotic symptoms ranged widely and included a number of children with auditory and visual hallucinations.

The authors concluded that psychotic symptoms may be more likely to occur among children with a family history of mental illness, particularly when taking stimulant medications.  They urge close monitoring of these symptoms during treatment.

It is likely that many practicing clinicians are going to be quite stunned to see rates of psychotic symptoms this high.  While the study will surely activate the current anti-medication fervor common among those who don’t see patients, many practicing physicians will likely find that these rates simply don’t fit with clinical experience.  Indeed, there may be some reason to be cautious in interpreting these results.  As mentioned, the sample size was quite small, and psychotic symptoms were almost exhaustively assessed.  Yet strangely, association between psychosis and stimulant dose were not examined.  Perhaps most limiting, overall level of psychiatric symptoms in children was not controlled for in the analysis. As a result, it is hard to know whether the group taking stimulants was more impaired from the start.

Nevertheless, this study suggests that more vigilance and perhaps more rigor is needed when querying for possible psychotic symptoms among children who take stimulant medications, particularly among youth who have a family history of major mental illness.

Ironically, while the intent of the study certainly is to advise caution in the use of psychiatric medications, the article could lead to the opposite, if more antipsychotic medications are used to treat these more subtle symptoms.  This would be unfortunate, in my view, until these results are more thoroughly replicated using a randomized design.  Studies like this remind us to maximize many non-pharmacological treatments that can be effective in ADHD.


MacKenzie LE, et al.  Stimulant Medication and Psychotic Symptoms in Offspring of Parents With Mental Illness.  Pediatrics 2015; epub ahead of print.

Melatonin: What Do We Really Know?

Posted: December 11th, 2015 by David Rettew

It is probably safe to say that melatonin has become the go-to agent for the treatment of children who struggle to fall asleep. According to a recent article in the Wall Street Journal, melatonin is now a 260 million dollar market in this country where it is sold over the counter, unlike some European countries where a prescription is still required.

Melatonin is a hormone that is synthesized from the amino acid tryptophan in the pineal gland. Its release is controlled by the suprachiasmic nucleus of the hypothalamus with the level peaking in the evening.  The fact that our body makes melatonin naturally has  been one factor related to many primary care physicians and psychiatrists alike giving the hormone a “pass” when it comes to

Image by phanlop88 and freedigitalphotos.net

Image by phanlop88 and freedigitalphotos.net

scrutiny, but questions about both is efficacy and safety remain.

Recently, a selective review (which reads more like a commentary) about melatonin use in youth was published in the Journal of Paediatrics and Child Health.  While the article certainly seems to be trying to persuade as much as inform, the author makes some important and surprising points.  One is that animal studies have shown a range of effects on the reproductive system for melatonin that have not been well understood. Indeed, there still exists a registered use for melatonin in animals to increase fertility.  These effects on reproduction or puberty in humans have yet to be demonstrated, although the few studies that have examined this issue have been small and contain significant limitations.  It is also worth noting that melatonin is metabolized in the liver by the enzymes CYP1A2 and CYP2C19 which can be inhibited by antidepressants such as fluvoxamine and citalopram, respectively.  As melatonin is often used in conjunction with psychiatric medications, these potential interactions need to be considered.

With regard to efficacy, studies seem to support its usefulness, although perhaps not to the extent one might expect.  Two recent controlled trials in children found that melatonin increased sleep onset by about 30 to 45 minutes but it does not generally increase total amount of sleep.

The author concluded that the research on melatonin in children is inadequate, especially given the effects on the reproductive system documented in animal studies. He urged much more research before governing bodies decree melatonin as safe and effective.

The review was later criticized in a letter to the editor for omitting studies from melatonin prescription products such as Circadin which is approved in Europe (although not for kids).  In the USA, there is also the melatonin receptor agonist remelton (Rozerum) which also has some data but, again, is not approved for pediatric ages.

Counter-balancing the concern of melatonin is the large body of literature that demonstrates the negative effects of poor and inadequate sleep.   In trying to balance these concerns, I’m not sure I’m ready to take melatonin out of my bag (particularly in comparison to other sleep medications that are used).  At the same time, however, it is important not to use melatonin or any other sleep-aid as a shortcut to addressing behavioral measures.   Many children benefit from consistent routines that gradually wind down stimulation and prepare a child to settle down for the night, and families that try to circumvent this process often run into problems.  Exercise (or lack thereof) can also be a major player in why some kids just don’t feel tired at night.  Nevertheless, it is also certainly true that many children will struggle with sleep no matter how good the bedtime routine is.


Kennaway D.  Potential safety issues in the use of the hormone melatonin in paediatrics.  J Paediatrics Child Health.  2015: 51(6):584-589.



Primary Care – Got Child Psychiatry Consult?

Posted: December 9th, 2015 by David Rettew

The Vermont Center for Children, Youth and Families at FAHC/UVM, in partnership with VCHIP, is entering our fourth year of the Child Psychiatric Phone and Email Consultation Program. The goal of this program is to provide healthcare providers with free curbside phone and email consultation to assist in the management of emotional and behavioral problems in primary care settings. Examples of these consultations include assisting in assessment, diagnosis, medication management and Family Wellness recommendations.

Results taken from a questionnaire given before and after PCP’s have accessed our consultation service show that users confirm an increased ability to quickly access expert psychiatric consultation as well as experience an increase in confidence in diagnosing and treating children’s mental health issues.

Whether you are already familiar with our consultation services, or hearing about them for the first time, please review the below commitment.

We will:

  • Designate a child psychiatric specialist (MD or NP) to return your phone call/e-mail within 24 hours during the regular work week (weekend coverage is already available through the on call system).
  • Provide you and your staff with phone and E-mail access to our experienced Family Wellness Coach, Eliza Pillard, LICSW, who can help guide your practice in the search for evidence based interventions for emotional and behavioral problems on behalf of your patients.
  • Connect you to our VCCYF Primary Care blog http://blog.uvm.edu/drettew/ and VFBA (Vermont Family Based Approach) blog http://blog.uvm.edu/vfba/ which deliver regular postings on topics that may be relevant for your patients, such as updates on child mental health and family wellness research and interventions.
  • Update all participants with news of access to our services (e.g., our Autism Assessment Clinic), educational opportunities, and advances in our field.

Please note that this program will not serve as a rapid conduit for assessment and consultation at the VCCYF. As you know, we have a long waiting list and feel obligated to serve families on a first-come-first-serve basis.

If you are interested in accessing a consult, please contact Eliza at eliza.pillard@uvmhealth.org or (802) 847-9759. We look forward to hearing from you.



Jim Hudziak, MD, Director                                                                                                      Eliza Pillard, LICSW

Vermont Center for Children, Youth, and Families                                                            Family Wellness Coach

Study Explores Humans’ “Natural” Sleep Patterns

Posted: December 3rd, 2015 by David Rettew

Despite the fact the we spend nearly a third of our life asleep, many fundamental questions, such as how much sleep we really need, remain debated.   One popular view is that most of us living in modern cultures are chronically sleep deprived and this may be one of the causes of many health problems, from obesity to depression.

To try and get a better handle on natural sleep habits, some researchers recently published a study that examined the sleep patterns of three pre-industrial societies in Tanzania, Bolivia, and SleepNamibia that were mainly hunter-gatherer or hunter-horticulturalist groups. None of the three groups used modern conveniences to alter light or temperature.  Of chief interest were the basic statistics regarding sleep onset and duration and their relations to environmental variables such as light and temperature. Sleep was measured using watch like devices worn on the wrists for 6 to 28 days.

Interestingly (and fortunately for the researchers trying to synthesize their data), there were many similarities between the three societies with regards to sleep habits. The total sleep duration was found to range between 5.7 to 7.1 hours per day, with about 1 hour more of sleep occurring during the winter.   Initiation of sleep tended to occur about 3 hours after sunset with awakening occurring generally before sunrise.  Somewhat unexpected was that these patterns seemed to be as much related to temperature as light, with sleep onset occurring during the period of rapid temperature decline and awakening occurring around the time temperature is at its lowest.

Contrary to some data suggesting that our ancestors slept in two waves with a period of quiet wakefulness in between, extended periods of nocturnal awakening were rare. Napping was also not particularly common but did occur in about 7% of days during the winter and 22% of days in the summer (this was an estimate and may have counted periods of rest without sleeping). Interviews with subjects revealed that, despite none of the three cultures having a word for insomnia, the problem was reported by some, albeit at levels typically lower than that of industrialized societies.

The authors concluded that their data suggest, contrary to some popular opinions, that sleep in industrial societies has not been artificially reduced below natural levels due to the pressures of modern culture.  At the same time, however, there was evidence that modern sleep schedules have veered in some different ways, possibly due to all the artificial light and temperature controls we now enjoy.  

While definitely an interesting study, the clinical takeaway is not immediately clear.  One possible implication might be to relax our pushing of people to get at least 8 hours of sleep per night.  There could also be benefit for people with sleep problems to try and shift their schedules somewhat or create light and temperature variations that more closely mirror the natural world.


Yetish G, et al.  Natural Sleep and Its Seasonal Variations in Three Pre-industrial Societies.  Current Biology 2015; 25:1-7.


Bias Found for Depression Psychotherapy Trials

Posted: November 19th, 2015 by David Rettew

The medical community was disappointed, but perhaps not shocked, several years ago when it was revealed that the efficacy of antidepressant medications in the treatment of adolescent depression might not be as strong as we thought.  This bias was due to the fact that the clinical trials, most of which were funded by pharmaceutic companies, that showed a positive effect for theCounseling medication were published prominently in journals while negative trials got much less publicity and often were not published at all.   A deliberate effort to downplay unfavorable studies of antidepressants was generally suspected, although some people have argued that other factors may also be at work, such as journals being less enthusiastic to publish negative findings and researchers not wanting to devote lots of time and energy describing hypotheses that did not pan out.

With regard to depression, the publication bias for antidepressants made some clinicians think more strongly about recommending psychotherapy.  Now, however, a new study recently published in PLOS One, examines whether or not a similar bias may be present in psychotherapy studies of depression.

The authors looked back at grants awarded by the NIH for controlled psychotherapy studies for depression conducted between the years 1972 to 2008.  They found the publications for these studies and examined the effect size by which the therapy improved symptoms relative to placebo.  If they found a funded grant but no publication, the authors contacted the investigators and calculated the effect size for those studies.   Studies of children and adolescents, unfortunately, were excluded for these analyses.

Surprisingly, nearly a quarter (23.6%) of the 55 funded studies did not result in publication and two studies were never started at all. Most of the unpublished studies were never submitted for publication in the first place (rather than it being written up and rejected over and over by different journals).  Most notably, the effect size of unpublished trials was less than half that of published trials. When these unpublished studies were added to the published ones, the effect size for psychotherapy dropped by 25%.

The authors concluded that there is strong evidence to conclude that a publication bias exists for psychotherapy trials for depression and that our understanding of how effective psychotherapy is, at least for depression, may need to be adjusted somewhat.  That said, the evidence continues to show that psychotherapy generally works for depression and there is no reason to alter the general recommendation of psychotherapy as a first line treatment.

This study is useful in reminding us that bias can exist in many domains and for many reasons.  Most of us are now quite primed to recognize potential bias and conflict when it comes to the pharmaceutical industry, but we need to be vigilant about some of the less obvious sources as well.


Driessen E, et al.  Does Publication Bias Inflate the Apparent Efficacy of Psychological Treatment for Major Depressive Disorder? A Systematic Review and Meta-Analysis of US National Institutes of Health-Funded Trials.  PLOS One, epub ahead of print, 2015.

An Update on Concussion

Posted: November 4th, 2015 by David Rettew

(Editor’s Note:  I’m pleased to share this guest post by one of our first year child psychiatry fellows, Dr. Winston Chung.)

More recently, one of the hot topics in Pediatrics, Neurology, and Psychiatry is the impact of concussion(s) on the developing mind. It came into the public consciousness due to the highly publicized death of beloved NFL football players like Junior Seau, and the tragic accident of 13yo Zackery Lystedt who collapsed after a football game and later required neurosurgery to alleviate the pressure produced by the swelling in his brain.   Increasingly, we’re forced to confront the difficult decision of whether we should allow our kids to play football, hockey, or soccer, and the morality of supporting billion dollar industries that produced a product where men launch themselves at one another destroying their body and mind.

The American Academy of Neurology defines concussion as a biomechanically induced clinical syndrome leading to alteration of brain function, affecting memory, and orientation, and may involve loss of consciousness. Though definitions vary, many clinicians define concussion as a subset of mild Traumatic Brain Injury which is defined as Glasgow Coma Scale score of 13-15, loss of consciousness of less than 30 minutes (if present), and posttraumatic amnesia of 24 hours (if present). There are between 1.5-3.8 million sport-related TBIs in the US each year, the majority of which are mild TBI and concussions. Concussions lead to 100,000 Emergency Room visits each year for school-aged children and represents 9% of all high school sports injuries.

Concussions are produced by acceleration and deceleration forces on the brain, whether linear or rotational. When these forces act on the brain, it stretches and deforms, doing the same to the individual components that make up the brain such as neurons, and blood vessels. The axons that make up the white matter tracts are especially vulnerable to injury given their length. The disruption of the neuronal membranes leads to disruption of neurotransmitters and electrolytes that are neatly sequestered, causing the brain to go into a hypermetabolic state (thus requiring glucose) to return it to its former equilibrium. This leads to a widened gap between the increased energy needs of the neurons and the decreased blood flow to the brain, resulting in an energy crisis.

The result of this disruption are headaches, fatigue, dizziness, slowed mentation, drowsiness, difficulty concentrating, nausea, light sensitivity, noise sensitivity, forgetfulness, blurry vision, sleep disturbance, irritability, depression, vomiting, and tinnitus. Usually the symptoms are self-limiting and they resolve after 7-10 days. However, symptoms in children may resolve more slowly. The mean duration of post-concussion symptoms among patients age 8-23 years who were referred to a concussion clinic was 43+/- 55 days. A minority of the patients reports post-concussion symptoms that persist for months, sometimes years. Repeated head trauma appears to lead to one of two presentations. Mood and behavioral symptoms usually appearing in third decade of life, and cognitive impairment and memory loss developing in fifth decade of life. The majority of subjects who present initially with behavior or mood symptoms progress to have cognitive symptoms before passing away, whereas those who present with cognitive impairment less commonly develop mood symptoms.

One topic that hasn’t been discussed enough in the press is the potential health effects of subconcussive injuries. Subconcussive impact is defined as biomechanically induced injuries to the brain that do not result in clinical symptoms, but may have a cumulative effect leading to chronic traumatic brain injury in some adult patients. Recently, a small study showed white matter microstructure changes in professional soccer players who did not have a history of concussions when compared to age-matched swimmers. And it’s been reported that an offensive lineman in American football can experience over 1,000 subconcussive hits over the level of 10g in the course of a single collegiate season.

(Of note, for further information, please see an excellent presentation on the topic by Dr. Jim Hudziak who discussed concussions at a recent Community Medical School talk at the University of Vermont College of Medicine.  The video will soon be available and can be accessed here.)

Anxiety Risk with Stimulants Overstated

Posted: November 4th, 2015 by David Rettew

There is a widely held concern that psychostimulants when used to treat ADHD can induce or exacerbate anxiety.  This belief has led many clinicians to hesitate about using stimulants in patients who already suffer from high level of anxiety, prompting some prescribers to try, as first-line agents, nonstimulant medications that supposedly have less risk but may be less effective against core ADHD symptoms.  The actual data about stimulants and anxiety, are more mixed.  Recently, a meta-analysis was published with the hopes of coming to a more definitive answer on this questions.Stimulants and anxiety

The study reviewed 23 different placebo-controlled trails and encompassed nearly 3000 children and adolescents. Clinician rated side effect measures of anxiety, when available, was utilized to assess patient anxiety. Subgroup analyses and meta-regression techniques were carried out to examine differences between medication types and doses.

In terms of results, there was overall a decreased risk of anxiety in patients treated with stimulants versus controls.  This global finding, when analyzed more closely, held particularly with methylphenidate preparations (e.g. Ritalin, Concerta, etc) while amphetamine derivatives (e.g. Adderall) did not separate statistically from placebo. There was also some indication that short versus long-acting stimulants and higher versus lower doses were associated with this reduced anxiety risk. No evidence of publication bias was found, although a number of stimulant trials did not meet the authors’ inclusion criteria.

The authors concluded that, contrary to conventional wisdom and some FDA labelling, the risk of anxiety is actually lower compared to placebo in patients treated with stimulant medications.

In the Discussion section, the authors commented about the possible mechanism of their main finding and hypothesized that the reduced anxiety risk was likely not a direct antianxiety effect of the stimulant rather an indirect effect of improving ADHD symptoms.

While the study certainly doesn’t exclude the possibility of some patients experiencing worsening anxiety with stimulant treatment, the article should be somewhat reassuring to those who might otherwise be reluctant to prescribe stimulants to the large number of youth with ADHD who also experience significant anxiety.


Coughlin CG, Cohen SC, et al. Meta-Analysis: Reduced Risk of Anxiety with Psychostimulant Treatment in Children with Attention-Deficit/Hyperactivity Disorder. J Chi Adolesc Psychopharm. 2015;25:611-617.

Autism Diagnosis Made Later For Children with ADHD

Posted: October 21st, 2015 by David Rettew

Autism and ADHD can share many common features, and both disorders are being diagnosed more frequently compared to previous decades.  Making matters more complicated is the accumulating evidence that the two disorders may share some underlying genetics and neurobiology.  While autism is optimally diagnosed when children are around two years of age, many are not diagnosed until much later which represents a lost opportunity for early intervention.  Some preliminary evidence suggests that one factor that is related to a delayed diagnosis of autism is the presence of other psychiatric and developmental problems.  This study, recently published in Pediatrics, examines specifically the association between the age of autism diagnosis and presence of an ADHD diagnosis.

The data come from the National Survey of Children’s Health in the years 2011-2012.  This survey was conducted by randomly calling households in all 50 states. The response rate was 23%. From Autism age of DXan initial calling of over 85,000 households, a final sample of nearly 1500 autistic children between the ages of 2 and 17 was identified.  The main variable of interest was whether or not a child had also been given a diagnosis of ADHD and the specific ages that the two diagnoses were given.  Other factors such as illness severity were recorded an controlled for in some of the analyses.

A total of 42.9% of the autistic sample had also received an ADHD diagnosis, with slightly less than half of that number receiving the ADHD diagnosis first. In those cases in which ADHD was diagnosed first, the autism diagnosis was found to occur roughly 3 years later compared to children in which the ADHD diagnosis came at the same time or after the autism diagnosis, controlling for other variables. Examined another way, children in whom the ADHD diagnosis came first were found to be nearly 30 times more likely to have their autism diagnosis come after the age of 6. While an older age of diagnosis was found to be related to less severe autism, the delay in autism diagnosis among children first diagnosed with ADHD was found regardless of the severity of the autism or ADHD symptoms.

The authors concluded that the presence of ADHD symptoms could cause a delay in an autism diagnosis. They advise clinicians to consider autism in young children who present with ADHD symptoms.

This study raises an important issue in clinical practice, and supports the idea that clinicians may not be as vigilant about autism when a child has symptoms of ADHD.  I’ve tried to train myself (and teach others) that seeing a young child with OCD behaviors should trigger some additional investigation about autism, but thinking the same way for ADHD behaviors is new.

There was one aspect of the study that was puzzling to me and not well discussed.  From the Discussion section, one gets the impression is that the way this all works is that once a child receives a diagnosis of ADHD then subsequent behaviors are viewed through an “ADHD lens” which in turn causes less diligence to search for other potential diagnoses such as autism.  However, in reading the fine print (specifically Table 1) it appears that the mean age of a child receiving a diagnosis of ADHD was about 6, which is well beyond the period that autism screening and assessment is recommended.  Thus, it does not appear that the presence of the ADHD diagnosis itself is clouding judgements about an autism diagnosis but rather something else.  What could that be?  The article doesn’t really say.  Perhaps there is something about early emerging ADHD behaviors that are particularly salient and that may mask more subtle autistic symptoms.  More work needs to be done to figure out what really is going on here.


Miodovnik A, et al.  Timing of the Diagnosis of Attention-Deficit/Hyperactivity Disorder and Autism Spectrum Disorder.  Pediatrics 2015; epub ahead of print.

Antidepressants and Violent Crime in Youth

Posted: September 30th, 2015 by David Rettew

Antidepressants have carried a black box warning for years related to new or worsening suicidal behavior in children and young adults. A possible link, however, between SSRIs and other types of violent behavior has been more difficult to identify with studies finding inconsistent results. A recent large Swedish study, recently published in the journal PLOS One, now jumps into the debate.

The study used a somewhat novel approach in employing a “within subject” design by comparing individuals between periods someone was and was not taking a medication.  Records between

photo by patrisyu and freedigitalphotos.net

photo by patrisyu and freedigitalphotos.net

2006 and 2009 from a large national registry were used for this study and over 850,000 individuals were identified who were prescribed an SSRI medication (subjects who likely only took the medication very briefly were excluded). Another national registry was also used to identify individuals convicted of violent crimes, although other types of crimes were also investigated as a secondary outcome. The authors also tried to quantify the cumulative SSRI dose and divided subjects into groups of low, moderate, and high.

A total of 10.8% of the sample had been prescribed an SSRI with citalopram and sertraline being the most common. The main finding of the study, and the one that received the most press, was that for both males and females between the ages of 15 and 24 only, there was a statistically significant increase in the rate of violent crime during the period someone was taking the medication compared to the intervals that they weren’t. The “hazard ratio” for the 15-24 age group was 1.4 which roughly translates into a 40% increased likelihood. Described in the original study but not well reported by many press articles, however, was the important fact that this risk was increased only among youth with low SSRI doses and not those with moderate or high overall SSRI exposure. Some significant associations were also found between SSRI use and some non-violent crimes as well as non-fatal accidents. Regarding other types of antidepressants, a link was also found between violent crime and the antidepressant venlafaxine while, interestingly, the antidepressant mirtazapine was found to be related to a reduced risk of violent crime.

Sorting out what all of this means is challenging.  While it might be easy to jump to the conclusion that SSRIs cause violent behavior (and many have), the data can’t really support that claim.  In fact, the authors state in their discussion that one possibility is almost the opposite, namely that the finding of the link with violent crime only among those taking subtherapeutic doses suggests that undertreatment may be the mechanism behind this link.  That said, if antidepressants really worked wonders for young people, we should see that therapeutic SSRI usage was associated with a reduced risk of violent behavior, which it wasn’t.  These kinds of finer points have unfortunately been missing in much of the media coverage of this study.

While the within-subjects design is a clever way to reduce some potential confounding factors, it is important to remember also that the study is not randomized.   Therefore, another complicating issue is that individuals were probably more likely to be prescribed medications during times when they were feeling more depressed, anxious, and angry to start with, and thus at higher risk of acting violently. Another fact to keep in mind is that while a 40% increase in violent crime sounds scary, the absolute numbers remain low.  When taking SSRI medications, the conviction rate of violent crimes was 1.0% rate compared to 0.6% without medications.

It is certainly true that young people may become quite agitated when given SSRI antidepressants and it can be a tough call for clinicians to be able to distinguish between because a medication problem (requiring a drop or discontinuation of medication) and  a worsening of the primary condition (which might prompt the opposite response).

In the end, we need to be cautious when using SSRIs, like any class of prescription medication, and be open to the possibility that sometimes the medication is the problem and not the solution.  At the same time, it would be a shame if this study raises unnecessary panic that possibly could end up making the situation worse.


Molero Y, et al.  Selective Serotonin Reuptake Inhibitors and Violent Crime: A Cohort Study.  PLOS One.  Sept 2015, epub ahead of print.


Extreme Picky Eating Linked to Mental Health Problems

Posted: September 17th, 2015 by David Rettew

A parent’s concern about their young child’s picky eating is one of the most common presenting complaints to primary care clinicians.  Generally, if a child is growing and otherwise healthy, the most common response is reassurance and some helpful strategies for how to help kids slowly expand their food repertoire.  This strategy usually works well, and most kids do go on to eat a wider variety of foods with the tireless encouragement and cajoling from their parents.

For some, however, it is not so simple.  The current list of psychiatric disorders in DSM-5 now includes the term Avoidant/Restrictive Food Intake Disorder (ARFID). The diagnosis replaced the term Feeding Disorder of Infancy and Early Childhood which was rarely used and not well researched.  The definition includes the requirement that there is a “persistent failure to meet appropriate nutritional and/or energy needs.” Nevertheless, many medical professionals may be wondering whether ARFID represents a real problem worthy of clinical intervention or is an Picky eater 2example of the DSM over-pathologizing typical and transient child behavior.

To look at this question in more detail, researchers from the Duke Preschool Anxiety Study examined over 900 kids who on average were about 4 years old. They assessed the degree of selective eating through an interview and classified children into groups of “normal,” “moderate,” and “severe” levels of selective eating. Emotional-behavioral problems were was assessed with rating scales and a structured diagnostic interview to see if the children actually met criteria for certain psychiatric disorders.

Overall, at least moderate selective eating was present in 20% of the sample, while 3% were found to be in the severe range. Furthermore, severe selective eating was associated with higher rates of anxiety and depression both with regard to quantitative levels of symptoms and rates of some specific diagnoses (depressive disorder and social anxiety disorder). Investigating other domains, the authors also found that children with selective eating were also much more likely to be hypersensitive to smells, textures, or visual stimuli.

A subset of the sample was followed over time, and the high levels of anxiety were found to continue for many children when they were up to 8 years old.

The authors concluded that especially more severe levels of selective eating were related to other types of emotional-behavioral problems. Their hypothesis was that that the link was not causal (i.e. that selective eating caused anxiety and mood problems) but rather than an enhanced sensory sensitivity may underlie both the food selectivity and some of the associated emotional behavioral problems.

This study has some significant limitations.  While the authors do suggest that their data are relevant to the new ARFID diagnosis, they don’t directly assess ARFID in their study. Weight and weight trajectory were also not rigorously measured and indeed, the number of children with weight loss was not found to differ between the three groups of children (although 45% of the severe SE group had low growth). Finally, the number of 4-year-old children meeting DSM criteria for psychiatric illness in this study will strike many people as quite high. For example, 6% and 33% of the severe SE group met criteria for a depressive disorder and social anxiety disorder, respectively.  Of note, none of the subjects met criteria for autism.

Another practical issue is that an effective and well researched treatment for ARFID has yet to be determined, although it isn’t hard to find a multitude of recommendations for what to do about moderate and more severe selective eaters alike.

The take-away from this study is not to suddenly start treating every instance of picky eating as a mental health crisis, but perhaps to gather some additional information regarding the severity of the selective eating and the presence of other types of behavioral problems or sensory hypersensitivies that may deserve additional follow-up.


Zucker N.  Psychological and Psychosocial Impairment in Preschoolers With Selective Eating.  Pediatrics 2015; August, epub ahead of print.

Contact Us ©2010 The University of Vermont – Burlington, VT 05405 – (802) 656-3131
Skip to toolbar