Identifying and treating catatonia in children with neurodevelopmental disorders: A case series

Abstract

Catatonia is a neuropsychiatric syndrome that is an increasingly recognized cause of acute behavioural changes in children and adolescents with neurodevelopmental disorders (NDD). Literature suggests that catatonia can present differently in this population and can be missed due to diagnostic overshadowing. Catatonia is a treatable condition, and management strategies in children with NDD include benzodiazepines and electroconvulsive therapy (ECT). Untreated, it can cause significant morbidity including severe medical complications, and therefore timely recognition and management of catatonia in children and adolescents with NDD is essential. In this case series, we present three cases of children ages 7, 14, and 10, with diagnoses of autism spectrum disorder, Down syndrome, and Prader-Willi syndrome, respectively. All were admitted to a pediatric inpatient unit for acute behavioural regression. Each had symptoms consistent with catatonia, resulting in trials of benzodiazepine therapy with inadequate response, and were then treated with bilateral ECT. In all cases, marked improvement was noted after ECT, with no apparent adverse effects. The cases are used to highlight the nuances of diagnosis and management of catatonia in children and adolescents with NDD. This includes insights on how presentations of catatonia may differ in this population, challenges with the use of available diagnostic tools, and how these patients may respond differently to recommended treatments such as benzodiazepines. The case series aims to increase clinicians’ awareness of pediatric catatonia when children and adolescents with NDD present with acute behavioural changes, and to encourage consideration of the full spectrum of treatments, including bilateral ECT.

Introduction

Catatonia is a neuropsychiatric syndrome that can involve motor, speech, behavioural, affective, and autonomic dysregulation, with the potential for severe morbidity and mortality (1, 2). It can arise secondary to psychiatric conditions, such as mood or psychotic disorders, as well as underlying medical conditions (1, 3). Catatonia has been increasingly recognized in children and adolescents, specifically those with autism spectrum disorder (ASD) and other neurodevelopmental disorders (NDD) (3, 4). Gamma-aminobutyric acid (GABA) dysfunction has been a hypothesized mechanism, with evidence of a shared genetic vulnerability for both GABA dysfunction and catatonia in ASD (5–8) and Prader-Willi syndrome (9, 10). An acute onset functional regression has also been identified in Down syndrome, termed Down syndrome disintegrative or Down syndrome regression disorder (DSRD) (11). The mechanism for DSRD is poorly understood, but catatonia has been identified as a potential cause of its behavioural manifestations (11, 12).

Clinicians should have an increased index of suspicion for catatonia when children and adolescents with NDD present with motor, speech, affective, and behavioural changes, because children with NDD are at risk of a missed diagnosis (4). Many factors contribute to this risk; importantly, the behavioural features of catatonia can overlap with the symptom profiles of NDD. For example, mutism, echolalia, grimacing and stereotypies are features in the diagnostic criteria for catatonia, but also commonly occur in children with ASD (13). Without a clear understanding of the patient’s baseline behaviour, there is a risk of catatonic symptoms being explained away by the NDD (14). This concept of bias regarding co-occurring diagnoses in those with NDD is termed “diagnostic overshadowing” (15).

Presentations of catatonia in the pediatric population may appear differently. The motor symptoms of catatonia can be categorized into hyperkinetic (e.g., agitation), hypokinetic (e.g., stupor, mutism, negativism) or dyskinetic (e.g., grimacing, posturing) activity (16). Pediatric catatonia has been described as having more hyperkinetic features, such as psychomotor agitation, compared to adults (13, 17). Regressive features, such as urinary incontinence, have also been described in children (17).

Catatonia puts patients at risk for medical co-morbidities including pneumonias, malnutrition, and thrombosis, has significant impacts on function and quality-of-life, and can progress to malignant catatonia (where catatonia presents with clinically significant autonomic abnormalities) (18). This latter sequela carries a mortality rate of up to 20% in adolescents (19). It is therefore essential that providers diagnose catatonia and offer appropriate treatment. In a patient with suspected catatonia, current guidelines suggest investigating for underlying psychiatric and medical disorders, the use of catatonia rating scales, and a benzodiazepine challenge to help confirm diagnosis (20). The Bush Francis Catatonia Rating Scale (BFCRS) is the most used worldwide; this is a 23-item scale where the presence of 2 or more of the first 14 items is consistent with catatonia (20). Treatment then involves escalating doses of benzodiazepines, and if lack of response (as informed by daily administration of a catatonia rating scale), consideration for bilateral electroconvulsive therapy (ECT) (20). When these first line therapies are unsuccessful, guidelines suggest consideration of N-methyl-D-aspartate (NMDA) antagonists such as amantadine or memantine (20).

The safety and efficacy of ECT has been demonstrated in pediatric catatonia, with one review finding a 76% positive response rate (21). Case series data of patients with catatonia and both ASD and Down syndrome suggests that those with NDD may have a less robust response to benzodiazepines than neurotypical patients (12, 22). This further emphasizes the importance of considering ECT in these patients.

In this report, we present three cases of pediatric catatonia in children with co-morbid NDD. Each were noted to have improved after bilateral ECT following inadequate response to benzodiazepine therapy. In each case, the presentation of catatonia, the medical workup and diagnostic process, and the treatment plan are reviewed. Additionally, challenges in the diagnostic and treatment process are highlighted and may serve as learning points for clinicians. Case 1 demonstrates challenges in the diagnosis of catatonia when a child’s neurocognitive baseline is unclear. Case 2 demonstrates the difficulties associated with the lorazepam challenge as a diagnostic tool in a child with NDD. Case 3 demonstrates the identification of catatonia in the presence of other psychiatric symptoms, such as psychosis. Verbal consent from the patient’s guardians and assent from patients was obtained and documented, to allow for sharing of these cases.

Case 1

A seven-year-old girl with no known psychiatric history presented to hospital with sudden-onset behavioural changes, two days after the onset of an upper respiratory tract infection. Her symptoms included aimless wandering, repetitive non-purposeful movements, facial grimacing, echolalia with lack of meaningful speech, and negativism. This clinical picture was in keeping with catatonia, and she underwent a thorough evaluation for an organic etiology. This involved workup for infections, heavy metal toxicity, autoimmune disorders, and metabolic disorders. Initially, an electroencephalogram (EEG) and magnetic resonance imaging (MRI) of the brain demonstrated non-specific changes that could suggest an encephalitis, however, repeat testing was not suggestive of this. Except for an elevated immunoglobulin (IgE) antibody level, all investigations were within normal limits. Refer to Table 1 for details of the diagnostic workup. Her medical history was significant for a nuchal cord at delivery, neonatal jaundice, and five febrile seizures between nine months and five years of age. Parent and school collateral information suggested prior difficulties with reading and with waiting her turn, but no assessment for NDD had taken place.

Table 1.

Investigations undertaken to identify medical causes of behavioural change by case

Medical domain	Case 1	Case 2	Case 3
Hematology and biochemistry	CBC1 + differential Electrolytes Liver, kidney, thyroid function	Electrolytes Thyroid function	Iron studies
Infectious sources	Blood, urine, throat, stool, CSF2 cultures	Urine, CSF2 cultures	Urine culture Serum lyme
Autoimmune sources: non-specific markers	Serum IgE, IgG, IgA3, complement 3, 4, ANA4, CRP5, ESR6	Serum ANA4, CRP5, ESR6	Serum IgA3, ANA4
Autoimmune sources: specific antibodies	Anti-NMDAR7, anti-ASOT8, anti-TPO9	Anti-NMDAR7, anti-ASOT8, anti-TPO9, anti-thyroglobulin, anti-TSH receptor10	Anti-ASOT8, anti-TPO9, anti-TTG11
Metabolic sources	Serum ammonia, urine porphyrins Serum ceruloplasmin Serum lysosomal/peroxisomal storage disorders screen	Serum ammonia
Toxins	Serum lead, mercury, arsenic
Imaging and other investigations	MRI12 brain, EEG13	Sleep study	MRI12 brain, EEG13

¹Complete blood count

²Cerebrospinal fluid

³Immunoglobulins – E, G, and A

⁴Anti-nuclear antibody

⁵C-reactive protein

⁶Erythrocyte sedimentation rate

⁷N-methyl-D-aspartate receptor

⁸Anti-streptolysin O titre

⁹Thyroid peroxidase

¹⁰Thyroid stimulating hormone

¹¹Transglutiminase

¹²Magnetic resonance imaging

¹³Electroencephalograph

Table 2.

Summary of lorazepam challenge by case

Challenge details	Case 1	Case 2	Case 3
Total dose	4mg	6mg	6mg
Route	Intravenous	Intravenous	Intravenous
BFCRS1 before challenge	18	15	5
BFCRS1 after challenge	17	16	4

¹Bush Francis Catatonia Rating Scale

Table 3.

Summary of electroconvulsive therapy by case

ECT details	Case 1	Case 2	Case 3
Electrode Placement	Bilateral	Bilateral	Bilateral
Frequency of treatments/week1	2	3	2
Number of treatments in series	10	6	10
BFCRS2 at end of series	4	0	0

¹Treatments administered twice weekly were on Tuesdays/Thursdays, and treatments administered three times weekly were on Mondays/Wednesdays/Fridays

²Bush Francis Catatonia Rating Scale

A lorazepam challenge was done, where she received two doses of 1mg intravenous (IV) lorazepam spaced one minute apart. Observations included disinhibition (laughing and moving her bed up and down). She then received two further IV doses of 1mg, administered half an hour and one hour after the initial dosing. Improved verbal responsiveness was observed, and there was no sedation throughout. Her BFCRS score was 18 before the trial, and 17 after. Despite the lack of meaningful BFCRS change, clinical suspicion remained high for catatonia based on this trial. The patient received a course of steroids in the hopes of treating an immunologic etiology underlying the catatonia, despite workup not identifying a specific source. The steroid protocol was IV methylprednisolone, 30mg/kg daily for three days, then oral prednisone, 52mg daily for two days. This was followed by one treatment of intravenous immunoglobulin (IVIG) treatment. Within the same one-week timeframe as these treatments, treatment for catatonia was initiated with scheduled dosing of lorazepam, dosed at 1mg IV every four hours while awake. Over the one-week period on these treatments, the team observed increased appropriateness of verbal responses, reduced negativism (e.g., eating and drinking, more engaged in play), and decreased grimacing. Residual features included echolalia and echopraxia, stereotypies and aimless wandering. She was discharged on a lorazepam dose of 1mg orally every eight hours for ongoing management of catatonia.

One week after discharge, lorazepam was discontinued by the patient’s mother, as she described increased tearfulness and behavioural outbursts, and that the patient was hitting herself. A trial of amantadine was then initiated; she took 25mg for one week then 50mg for one week; this was then also discontinued as the patient’s family felt it was causing her to be more withdrawn. At two months post-discharge, she was started on a regimen of ibuprofen and azithromycin. This treatment choice was based on a diagnostic impression of a neuroinflammatory etiology, wherein a neurological insult such as an immune-related encephalitis (which may have gone undetected on testing) was presenting as ongoing catatonia. This trial lasted two months, and over the course of the trial the patient was found to be increasingly spontaneous and reactive. However, she had not re-gained previous skills that had regressed in the catatonic state (e.g., knowing letters of the alphabet, word-finding difficulty for previously known words). Of note, the extent to which these skills deviated from baseline was challenging to establish, as the child had moved to a new school after the admission, and collateral information from the previous school was limited.

Six months later, the patient experienced marked deterioration consistent with the initial catatonia presentation, including slurred speech, aimless wandering, slowed movements, and repetitive behaviours. A plan was made for bilateral ECT treatment of a prolonged catatonia syndrome. The family was provided with resources on ECT and met to discuss risks and benefits with the care team, which included a second consultation with a psychiatrist with expertise in ECT. The patient received ten treatments in total, administered at a frequency of twice weekly, with no apparent adverse effects. The team and family observed marked improvements including language (e.g., less word-finding difficulty, spontaneous speech about old memories), emotional state (e.g., parent reported a higher frequency of “good days”) and motivation (e.g., asking to join new activities). The BFCRS score decreased to 4 after the ten treatments. The patient was felt to be at her previous baseline except for a few specific deficits, including that she was sometimes mixing up letters while writing, and had some residual stereotypic movements.

Following treatment of catatonia, further testing was completed for an underlying NDD. This was supported by teachers’ reports from her new school that she still struggled with inattention, impulsivity, and poor social skills. A neuropsychiatric evaluation was conducted by a neuropsychologist, including an Autism Diagnostic Observation Schedule and Autism Spectrum Rating Scales. The patient was diagnosed with ASD with associated cognitive impairment; her academic and adaptive functioning was estimated to be at a level expected for a six-year-old child (the patient was nine at age of testing). She was also diagnosed with attention deficit/hyperactivity disorder (ADHD), both of which then became targets for treatment.

Case 2

A 14-year-old boy with a history of Down syndrome and severe intellectual disability was admitted to hospital for functional regression. The changes were first noticed by family nine months prior, and at that time presented as periods of inattention and decreased engagement. Over the nine months, this was followed by decline in ability to conduct activities of daily living (e.g. needing prompting to void, change clothes), and decreased speech content.

On admission, several symptoms consistent with catatonia were noted, including generalized motor slowing, difficulty initiating actions, holding sustained postures, and non-purposeful movements (grimacing, wringing hands, and finger flicking). Investigations were completed to assess for an organic etiology for the regression. This included infectious workup and autoimmune markers. Given a history of numerous otolaryngology surgeries, a sleep study was conducted to rule out obstructive sleep apnea. All investigations were within normal limits. Refer to Table 1 for details of the diagnostic workup.

A lorazepam challenge was conducted to confirm the diagnosis of catatonia. He received 1mg IV lorazepam, followed by 1mg IV one minute later. This was repeated three times within a day for a total of 6mg. Following each set of two doses, the patient was found to be increasingly sedated. After the third administration, disinhibition including tearfulness and urinary incontinence occurred. No improvement in catatonic symptoms was noted. His BFCRS score was 15 before the trial and increased to 16 after. However, catatonia remained the diagnostic impression, given the high likelihood of atypical responses in children with NDD, and reported lack of response to benzodiazepines in chronic catatonia (22). Over the next three months the family noted some improvements but estimated that the patient was only functioning at 50% of his baseline. They decided to move forward with recommended ECT treatment (to which they had been initially hesitant). After the first session of ECT, a drastic improvement was noted with marked increased spontaneity and interactions with others. He received a total of six bilateral ECT treatments, administered at a frequency of three times a week. This was tolerated without adverse effects and was associated with a full remission of catatonia; BCFRS score was 0 after the six treatments. He was seen in follow-up three months later and had experienced some return of symptoms including whispering rather than speaking, facial grimacing, loss of fine motor skill (buttoning pants), and decreased interaction with others. He received another two sessions of bilateral ECT, again followed by marked resolution of symptoms, which did not return in the next six years of follow-up.

Case 3

A ten-year-old boy with Prader-Willi syndrome (unknown variant) developed recurrent behavioural “spells” of symptoms consistent with catatonia. The family reported two distinct symptom clusters. The first involved more hypokinetic motor changes, including staring, mutism, and negativism where he would refuse food and fluids. He also demonstrated aimless wandering, grimacing and pursing of lips, and posturing of hands. These episodes lasted 1–3 days. The second cluster of symptoms had more hyperkinetic features, where he would laugh, talk to himself, walk repetitively in a circle, and demonstrate new aggression towards his mother. These episodes only lasted a few hours to a single day and did not necessarily follow or precede the episodes of hypokinetic symptoms. On a few other occasions he had short periods (hours) of hysterically crying, saying that people were monitoring him and wanted to kill him, and that he was unsafe. In addition, he also had 1–4-day periods of being almost completely back to his happy, chatty, reality-intact baseline. These episodes increased in frequency and duration over the six-month period between initial symptom onset and consultation to psychiatry. At the time of consultation, parents estimated the child spending about 80% of time in the hypokinetic-type catatonia phase, 5% in either the hyperkinetic or paranoid phase and 15% at his normal baseline.

Prior to these episodes, his care team had described him as functioning well with appropriate supports. He had mild hyperphagia associated with Prader-Willi, and some developmental delays. Specifically, neuropsychological testing at age eight had identified low cognitive ability (functioning at the level of a five-year-old child), and delays in fine motor and speech milestones. He was receiving support from occupational therapy and speech language pathology and was working with a full-time educational assistant in his school.

All bloodwork was negative, including workup for infectious, autoimmune, and metabolic etiologies, and a brain MRI was normal. An EEG during one of the spells of hypokinetic catatonic symptoms showed mild focal epileptiform discharges from the left occipital region. It was not felt that this represented seizure activity or required any further workup or anti-seizure medications, as it had been present since birth. Refer to Table 1 for a summary of the diagnostic workup.

Lorazepam was trialed as treatment for presumed catatonia; the patient trialed 1mg orally at home on two occasions. His parents reported minimal change in the first hour after taking it, but that he then became extremely fearful and agitated for several hours; the trial was then discontinued. Valproic acid then was started at a dose of 250mg daily and increased in increments of 250mg every two weeks to a total dose of 500 mg twice daily. This medication was chosen as it has effects to increase GABA and decrease glutamate (a dysfunction of which is implicated in catatonia) (9). The team also considered that the catatonia could be resulting from an underlying rapid-cycling bipolar disorder, which could also be treated with valproic acid. Antipsychotics were not felt to be indicated due to the risk of neuroleptic malignant syndrome with comorbid catatonia. The patient continued at the dose of 500mg twice daily for six weeks, and this was associated with a resolution of the catatonia phases with hyperkinetic behaviour (agitation). However, parents reported that the episodes with hypokinetic features became more frequent and of longer duration. He was then admitted for an IV lorazepam trial for diagnostic confirmation. He received 1mg lorazepam IV every 30 minutes over a period of three hours, resulting in a total of 6mg. After receiving the fourth dose, a marked change was observed where the patient sang a song under his breath, spoke a full sentence, and asked for something to eat. No sedation was observed throughout. His BFCRS score was 5 before the trial and 4 after. Despite the lack of meaningful BFCRS change, clinical suspicion remained high for catatonia based on this trial. A decision was made to move to bilateral ECT, given the lack of tolerability and/or improvement observed with previous pharmacotherapy (lorazepam and valproic acid). Informed consent was obtained and he received ten treatments, administered twice weekly, with no apparent adverse effects. He was estimated by family to have improved by 80–90% at the end of the ten treatments. His BFCRS score after the ten treatments was 0. Ongoing maintenance ECT treatment was offered, but this was declined by the family.

Discussion

These three cases share several features, including regression from baseline in children/adolescents with NDD, features consistent with catatonia, and positive change following bilateral ECT.

In case 1, a complicating factor was that the patient did not have a known pre-existing NDD diagnosis at the time of presentation. She had a reported history of delays in reading but did not have a previous psychoeducational assessment describing her baseline functioning. In this case, gathering a thorough developmental history despite limited documentation was essential to understanding the behavioural change, and making the catatonia diagnosis. Had the diagnosis of autism been pre-established, the team may have also been cued towards catatonia by the hyperkinetic symptom (agitation) predominant in her presentation, which is more common in children with NDD (17).

The diagnosis was also complicated by the symptom presentation following an infection, causing clinical suspicion for a neuroinflammatory process. Catatonia can arise secondary to autoimmune conditions, and concurrent management of both conditions is required (23, 24). Proposed indicators of an immunologic source of catatonia include the presence of atypical psychiatric features (e.g. acute regression, lack of response to usual treatments), inflammatory markers, and response to therapy for the suspected autoimmune condition (25). This patient had evidence of elevated IgE and a response to antibiotic/anti-inflammatory treatment, in addition to her neuropsychological symptoms. In summary, she demonstrated a combination of acute neuropsychiatric changes, clinical concern for NDD, and features of an inflammatory process. All three of these factors should increase the index of suspicion for catatonia.

Case 2 highlights an important consideration in catatonia diagnosis and management, regarding the limitations of the lorazepam challenge as a diagnostic tool. In this adolescent, a significant improvement in symptoms was not noted in response to lorazepam; rather, sedation and disinhibition were noted and the BFCRS score increased. However, treatment with ECT was still pursued based on the clinical picture of catatonia. Benzodiazepines are most efficacious when used early in the course of catatonia (22). In this case, the patient presented to hospital nine months after symptom onset which may have rendered their catatonia more treatment-resistant. Due to the issue of diagnostic overshadowing in patients with NDD, these children and adolescents may be unwell for disproportionate lengths of time relative to neurotypical peers. Case series data also suggest greater response to ECT compared to benzodiazepines in youth with Down syndrome and catatonia (12). This case highlights the importance of continued consideration of catatonia, even when the lorazepam challenge is not supportive of the diagnosis.

Case 3 demonstrates a cyclical course of symptoms, including brief recurrent episodes of psychotic symptoms. The phenomenon of cycloid psychosis is described in Prader-Willi syndrome patients, wherein patients have episodes of psychotic symptoms (often paranoia and auditory hallucinations) with full recovery between episodes, on a background of mood instability (9). Episodic psychosis can also be seen in bipolar disorder in patients with Prader-Willi, which was considered in this case, and catatonia can be seen in both conditions (9). It is essential to recognize in this case that there were also episodes of hypokinetic and episodes of hyperkinetic motor function. These motor features are not captured by a diagnosis of cycloid psychosis or bipolar disorder alone and are best explained by catatonia. This is essential to recognize because catatonia requires specific treatment. While psychotic symptoms in Prader-Willi syndrome may respond to atypical antipsychotics and lithium (26), in this case the child appeared to be initially responsive to valproate, then to benzodiazepines and ECT. Similar to the proposed mechanism of catatonia in autism, catatonia in Prader-Willi syndrome has also been postulated to relate to GABA-a receptor dysfunction (9), providing a mechanism for the effect of treatments such as valproate, benzodiazepines, and ECT.

Measurement challenges may be considered one limitation within our case series. As per institutional practices, the treating team used the BFCRS to assess symptoms. The scale was used before and after the lorazepam challenge, and after ECT. Subjectively, we found this measure to have some clinical limitations. In cases 1 and 3, the clinical team and family felt the change in presentation following the lorazepam challenge was significant. However, there was only a decrease in BFCRS score of 1 point after the challenge (resolution of mutism in both cases), and a typical positive challenge would entail a 50% decrease in symptoms on standardized measures (20). Changes in the BCFRS following ECT appeared to better correlate with clinical observations, though in case 3 the BCFRS score decreased to 0 after ECT but the patient only was estimated to be at 80–90% of their baseline. Subjectively, the most significant challenges identified by the treatment team were concern for over-reporting symptoms that may have been present at the patient’s baseline (e.g., impulsivity), and concern for missing symptoms not captured by the BFCRS items (e.g., loss of ability to perform activities of daily living, urinary incontinence). From these observations, we suggest that the BFCRS serve as an aspect of catatonia evaluation in children and adolescents with NDD, but that results should be correlated with qualitative clinical descriptors.

Alternatives or additions to the BFCRS may also be considered for children and those with NDD. This includes the Pediatric Catatonia Rating Scale (PCRS), a measure that draws from the BFCRS (27). The PCRS includes additional items characterizing regressive features, including items for refusal to eat/drink and for urinary incontinence (27). To specifically capture the profile of catatonia in those with NDD, Wing and Shah developed a set of criteria specific to catatonia in ASD (28), and more recently, Breen et al. developed the Attenuated Behavioural Questionnaire (29). Caroll et al. (30) identified that due to catatonia presenting in a wide variety of psychiatric and medical illnesses, the need for rating scales relevant to specific populations is needed. They propose the Katatonia Autism Neuropsychiatric and Neuromovement Examination Rating (KANNER) scale to be used in specific contexts of ASD and pervasive developmental disorder, though this is not validated in children (30). Our reporting of these cases would have benefitted from consideration of these scales, as well as more frequent reporting of rating scales throughout the clinical course of each case.

Another limitation is in the reporting of adverse effects of ECT. The three patients had close clinical monitoring in which no adverse effects were identified, however there is a lack of standardized assessment and reporting of adverse outcomes. Additionally, the clinicians assessing clinical outcomes and adverse effects of ECT were not blind to the phase of treatment, which may introduce bias in reporting of clinical outcomes.

These cases demonstrate the importance of the recognition and treatment of catatonia in pediatric patients who present with behavioural changes. Specifically, when there is an underlying NDD, care should be taken to identify changes from baseline, especially regressions and new excitatory behaviour. When diagnosed, catatonia should be treated using established catatonia guidelines (20). We would advocate for further development of algorithms specific to children and adolescents with NDD, considering that there may be limited response to benzodiazepines in this population, and that bilateral ECT may need to be promptly considered.