Longitudinal Symptom Burden and Pharmacologic Management of Catatonia in Autism With Intellectual Disability: An Observational Study

ABSTRACT

Catatonia is a highly morbid psychomotor and affective disorder, which can affect autistic individuals with and without intellectual disability. Catatonic symptoms are treatable with pharmacotherapy and electroconvulsive therapy, but the longitudinal effectiveness of these treatments in autistic individuals has not been described. We conducted a prospective observational cohort study of patients with autism and co‐morbid catatonia who received outpatient care in a specialized outpatient clinic from July 1, 2021 to May 31, 2024. Data investigating pharmacologic interventions, and clinical measures including the Bush Francis Catatonia Rating Scale (BFCRS), Kanner Catatonia Severity Scale (KCS), Kanner Catatonia Examination (KCE), and Clinical Global Impression—Improvement (CGI‐I) were collected. Forty‐five autistic patients with co‐morbid catatonia were treated during the study period. The mean age was 15.6 (SD = 7.9) years [Mdn = 16.0, range 6.0–31.0]. Forty‐one patients (91.1%) met criteria for autism with co‐occurring intellectual disability. All patients received pharmacotherapy. Forty‐four (97.8%) were treated with benzodiazepines with a mean maximal daily dose of 17.4 mg (SD = 15.8) lorazepam equivalents. Thirty‐five patients (77.8%) required more than one medication class for treatment. Sixteen (35.6%) patients received electroconvulsive therapy. Fourteen patients (31.1%) attempted to taper off benzodiazepines after achieving clinical improvement during the study period; of these, 5 patients (11.1%) were successfully tapered off, and the remaining 9 (17.8%) discontinued the taper due to a return of catatonic symptoms. Statistically significant improvement was observed across all clinical domains except the KCS. However, the majority remained at least partially symptomatic over the study period. Three patients (6.7%) died over the study period. Despite clinical improvements while receiving the gold standard for psychopharmacologic management of catatonia, chronic symptoms remained for the majority of catatonia patients over the study period, and few were able to taper and discontinue benzodiazepine treatment. Notably, the open label design of this study is a limiting factor when interpreting the results.

Summary.

• Catatonia is a psychomotor syndrome that can significantly impact the quality of life of autistic individuals, yet is not well understood.

• Symptoms of catatonia are responsive to treatment.

• However, over the 3‐year period of this study, most patients experienced residual symptoms, and few could discontinue psychotropic medication.

Abbreviations

ABA

applied behavioral analysis

BFCRS

Bush Francis Catatonia Rating Scale

BFCSI

Bush Francis Catatonia Screening Item

CGI‐I

clinical Global Impression‐Improvement

DSM‐5

Diagnostic and Statistical Manual, Fifth Edition

ECT

electroconvulsive therapy

KCE

Kanner Catatonia Examination

KCS

Kanner Catatonia Severity

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

1. Introduction

Autism is a heterogeneous neurodevelopmental condition characterized by difficulties in social interaction and communication along with restricted/repetitive patterns of behaviors and interests (American Psychiatric Association ²⁰²²). Autism affects 1%–2% of the population worldwide. 18%–48% of autistic individuals are also intellectually disabled (Lord et al. 2022). Those with co‐morbid intellectual disability are at an elevated risk for co‐occurring psychiatric conditions, caregiver distress, aggression, residential placement, and hospitalization (Sheehan et al. 2020; Siegel et al. ²⁰¹⁵; Ferguson et al. ²⁰²⁴).

One co‐occurring psychiatric condition, which may occur at a higher frequency in autistic individuals with co‐morbid intellectual disability is catatonia (Vaquerizo‐Serrano et al. ²⁰²²; Smith et al. ^2023a). Catatonia is a psychomotor syndrome with affective domains and distinct physical examination findings, which can affect individuals across the lifespan (Rogers et al. 2023; Smith et al. ^2024a; Hirjak et al. ^2024a). Recognition of catatonia is challenging in autistic individuals due to the baseline symptoms of autism, requiring careful review of change from baseline (Hauptman et al. 2023; Ferrafiat et al. ²⁰²⁴). A recent meta‐analysis by Vaquerizo‐Serrano et al. found that 20.2% of autistic individuals had features of catatonia, with 10.4% meeting full criteria for catatonia. The most common symptoms of catatonia in autism were new‐onset speech impairment, negativism, hyperactivity, and aggression deviating from the individual's baseline, which were not due to other underlying medical or psychiatric conditions. However, caution is warranted in interpreting the results of this meta‐analysis due to an oversampling of inpatients, likely inflating these estimates to some degree (Vaquerizo‐Serrano et al. ²⁰²²). Moreover, recent literature has suggested that recurrent self‐injury that is nonresponsive to intensive behavioral interventions can be a symptom along the catatonia spectrum for autistic individuals with co‐morbid intellectual disability (Wachtel, Shorter, and Fink ²⁰¹⁸; Wachtel ²⁰¹⁹). Although intensive behavioral interventions based on the principles of applied behavior analysis (ABA), including functional analysis and related function‐based treatments (Jessel et al. 2024), remain the gold‐standard for the assessment and treatment of self‐injury in the autistic population and often result in substantial benefit (Minshawi et al. 2014; Du, Guo, and Xu ²⁰²⁴), self‐injury due to catatonia will frequently be deemed “automatically reinforced” on a functional analysis and therefore poorly amenable to this otherwise effective treatment. As such, additional treatment modalities such as pharmacologic and neuromodulatory treatments are also needed (Wachtel, Shorter, and Fink ²⁰¹⁸), and invasive treatments are even considered in the most refractory of cases (Gorodetsky et al. 2024). While classically described features of catatonia (such as staring, posturing, and waxy flexibility) do occur in autism, externalizing (e.g., aggression, impulsivity, self‐injury) and hyperactive symptoms have been identified as being more common (Vaquerizo‐Serrano et al. ²⁰²²). Due to the nonspecific nature of many of these features, it must be confirmed (i.e., with careful history and differential treatment response, including lorazepam challenge) that such symptoms are due to catatonia directly and not more common disorders such as mood and anxiety disorders or baseline stereotypies (Hauptman et al. 2023). Overall, there may be an increased risk of a missed diagnosis given the phenotypic variability between catatonia occurring in neurotypical individuals and those with autism.

Use of electroconvulsive therapy (ECT) and benzodiazepines has demonstrated efficacy in the treatment of catatonia for both autistic adults and children (Espinoza and Kellner 2022; Smith et al. ²⁰²²; Smith et al. ^2024b). Some inpatient data and a small longitudinal study of only male patients is available (Luccarelli et al. ^2024a; Ohta, Kano, and Nagai ²⁰⁰⁶), and there is a need for larger longitudinal outpatient studies of catatonia in autism to define the course of the disorder (Ferrafiat et al. 2024). There is limited data available to guide the use of other pharmacologic agents beyond benzodiazepines in the population of autistic individuals with catatonia, particularly in the outpatient setting (Smith et al. ^2023b; Beach et al. ²⁰¹⁷). Thus, we identified the need for and conducted a prospective observational longitudinal study of autistic individuals. We present such a study of autistic individuals who received care in a specialized neurodevelopmental catatonia clinic. The aim of our study was to provide longitudinal diagnostic, pharmacologic, and therapeutic data to increase our knowledge of catatonia patterns in autistic individuals and to assist clinicians in providing care for this high‐need population.

2. Methods

2.1. Study Population

Utilizing Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (von Elm et al. ²⁰⁰⁸), we conducted an observational cohort study of patients with autism and co‐morbid catatonia who received outpatient care in a specialized clinic at Vanderbilt University Medical Center. Patient information was collected longitudinally from July 1, 2021 to May 31, 2024. Patients were included if they were seen at least twice in the outpatient clinic, were diagnosed with both autism and catatonia, and were followed for at least 45 days in the specialty clinic. The diagnosis of autism with or without intellectual disability was made, or confirmed if present historically, by the treating psychiatrist utilizing criteria from the Diagnostic and Statistical manual, 5th edition (DSM‐5) (American Psychiatric Association ²⁰²²). Catatonic symptom severity was assessed using the Bush Francis Catatonia Rating Scale (BFCRS) (Bush et al. 1996), Kanner Catatonia Severity Scale (KCS), and Kanner Catatonia Examination (KCE) (Carroll et al. 2008). Other diagnoses that could present in a similar fashion such as anxiety, psychotic, and mood disorders were ruled out using comprehensive psychiatric interview. No patients in this cohort received intensive inpatient behavioral therapies due to their outpatient status. The cohort did include individuals who received outpatient ABA. The clinic was staffed by child and adolescent psychiatry fellows and adult psychiatry residents. Two attending psychiatrists supervised the fellows and residents (J.R.S. and B.R.). J.R.S. is the medical director of the clinical service and thus, all catatonia assessments were discussed and reviewed by J.R.S. during each clinic day.

This study was reviewed by the Vanderbilt University IRB (#170317) with a waiver of informed consent from participants as data, which was collected as part of routine clinical activities. This study is a continuation of our previous work investigating the use of ECT for adults and children with autism and intellectual disability (Smith et al. 2022, ^2024b; Louie et al. ²⁰²⁴), and pediatric catatonia more broadly (Smith et al. ^{2023a, 2023b, 2024a}). Thus, data from some patients in this cohort has been previously reported in terms of diagnostic categorization and therapeutic applications in catatonia. However, the specific longitudinal clinic data presented here is novel and has not yet been reported.

2.2. Measures

Therapeutic data including psychopharmacologic interventions, use of ECT, and catatonia assessment measures (the BFCRS (Bush et al. 1996), KCS, and KCE) (Carroll et al. 2008), were recorded over the course of the study period. The BFCRS (Bush et al. 1996), the most cited rating scale in the field of catatonia (Weleff et al. 2023), consists of a 14‐item Bush Francis Catatonia Screening Item (BFCSI) and a full 23‐item BFCRS which measure a greater number of signs. The Kanner Catatonia Rating Scale includes the KCS severity assessment and the KCE, a standardized examination (Carroll et al. 2008). To determine if clinical improvement occurred over the course of longitudinal clinical care, total BFCRS scores were collected at the time of clinic intake (patients presenting for clinical care and receiving a diagnosis of catatonia upon intake) or, when the diagnosis of catatonia was first given (for patients receiving a diagnosis of catatonia while already followed in‐clinic). Secondarily, KCS and KCE scores were collected in the same manner. The KCS/KCE was also included as this assessment measures nudism, incontinence, and self‐injury which are not contained in the BFCRS (Carroll et al. 2008). Lastly, the Clinical Global Impression‐Improvement (CGI‐I) scale (Busner and Targum 2007) score was determined based on the patient's improvement since clinic intake and was calculated by the primary treating psychiatrist (author J.R.S.). Notably, the treating psychiatrist was not blind to the patient's overall treatment course when the CGI‐I was assigned. To investigate the impact of catatonia on families and autistic individuals of varying socioeconomic backgrounds, data regarding healthcare payers (Medicaid, private insurance, self‐pay) were also collected.

2.3. Data Collection

Demographics, co‐occurring medical and psychiatric conditions, the number of failed psychiatric medications, frequency and severity of catatonic symptoms as measured by catatonic assessment measures (BFCRS, KCE, KCS), use of ECT, use of ABA, mortality, and pharmacologic initiation and/or tapering were recorded prospectively during clinic visits. Catatonic signs and symptoms assessed at baseline (particularly those such as verbigeration and stereotypy, which could overlap with features of autism) were verified with patients' caregivers as representing a noticeable difference from the patient's typical behavior before the presumed onset of catatonic symptoms. Regarding verbigeration specifically, collateral from family was collected to delineate between baseline repetitive speech and new onset verbigeration. Specific examples include transitions from speaking in one to two‐word sentences, to screaming repetitively without the use of words. While changes in speech in autistic individuals may occur in the setting of psychotic or mood disorders, in these cases they were attributed to catatonia due to the presence of other catatonic symptoms as measured by the validated catatonia assessment tools.

Given the heterogeneity of benzodiazepines used in the treatment of catatonia in this population, dosing of benzodiazepines was converted into lorazepam equivalents, with 0.5 of clonazepam, 5 mg of diazepam, 10 mg of clobazam considered equivalent to 1 mg of lorazepam (Benzodiazepine Conversion Calculator n.d.). Based on our previous research, alternative benzodiazepines other than lorazepam were frequently utilized as they may be more efficacious for catatonia in autism (Smith et al. ^{2023a, 2023b}). Length of outpatient follow up was calculated from the date of clinic intake until either clinic discharge, the end of the study period, or death.

To address the prevalence of abnormal physical examination findings, the frequency of catatonic symptoms as measured by the BFCRS, KCS, and KCE was also collected. For patients who were first evaluated for catatonia in the inpatient setting, the BFCRS, KCS, and KCE at the time of the first and last assessment were included in the analysis. The specific clinical signs measured by each scale are listed in Tables 1 and 2. For the KCS, positive items (scoring a 2 or more on the KCS, and a score of 1 on the KCE) were included when determining symptom or physical examination finding frequency. Notably, self‐injury on the KCS is reported as mannerism or stereotypy with accompanying self‐injury (indicated by a score of 4 or higher on each item). Thus, symptom severity was also reported in these two domains to best capture the presence of self‐injury. Lastly, given that recurrent self‐injury can occur for many reasons in autism, generally symptoms of recurrent self‐injury are attributed to catatonia only if they have failed intensive behavioral interventions. As this cohort includes only outpatients and thus, could not receive intensive behavioral interventions, recurrent self‐injury alone was not considered a diagnostic symptom of catatonia in this cohort.

TABLE 1.

Frequency of index catatonic symptoms in autism as measured by the Kanner Catatonia Severity Scale and Kanner Catatonia Examination.

	N	%
N	42
Kanner Catatonia Severity Scale
Excitement‐hyperactivity	26	61.9
Immobility	12	28.5
Stupor	4	9.5
Mutism	18	42.9
Staring	34	80.9
Posturing	13	30.9
Grimacing	23	54.8
Stereotypy	29	69.0
Stereotypy with self‐injury (score of four or higher)	8	19.0
Mannerisms	21	50.0
Mannerisms with self‐injury (score of four or higher)	5	11.9
Rigidity	18	42.9
Flaccidity	0	0
Negativism	28	66.7
Reduced food intake	5	11.9
Reduced fluid intake	5	11.9
Impulsivity	34	80.9
Nudism	3	7.1
Incontinence	18	42.9
Combativeness	24	57.1
Kanner Catatonia Examination	40
Parroting (Echolalia)	8	20.0
Gibberish (Verbigeration)	14	35.0
Perseveration	24	60.0
Waxy flexibility	2	5.0
Catalepsy	3	7.5
Echopraxia	6	15.0
Command‐verbal (Automatic obedience and ambitendency)	17	42.5
Command‐motor (Mitgehen)	5	12.5
Paratonia (Gegenhalten)	3	7.5
Grasp reflex	2	5.0
Metronome test	2	5.0
Magnetism	8	20.0

TABLE 2.

Frequency of index catatonic symptoms in autism as measured by the Bush Francis Catatonia Rating Scale.

	N	%
N	37
Excitement	21	56.8
Immobility	10	27.0
Mutism	14	37.8
Staring	29	78.4
Posturing	10	27.0
Grimacing	18	48.6
Echolalia/echopraxia	13	35.1
Stereotypy	22	58.5
Mannerisms	16	43.2
Verbigeration	10	27.0
Rigidity	17	45.9
Negativism	20	54.1
Waxy flexibility	1	2.7
Withdrawal	6	16.2
Impulsivity	28	75.7
Automatic obedience	8	21.6
Mitgehen	5	13.5
Gegenhalten	1	2.7
Ambitendency	11	29.7
Grasp reflex	3	8.1
Perseveration	16	43.2
Combativeness	14	37.8
Autonomic instability	10	27.0

The KCS and KCE report catatonic symptoms such as self‐injury and incontinence, which are not reported in the BFCRS. The BFCRS does not separately assess for reduced food and fluid intake but combines these two measures. The BFCRS also does not assess for magnetism or the metronome test on physical examination, both of which are present on the KCE. For ease of administration for children and autistic populations, the KCE combines automatic obedience and ambitendency into one physical examination finding referred to as command‐verbal. Notably, affective symptoms of catatonia are not captured by either measure.

2.4. Statistical Analyses

Statistical analyses were performed using Python (Version 3.12.4; Python Software Foundation) and R (Version 4.4.1; R Core Team, 2024). The primary statistical endpoint was the change in BFCRS total score between time of first diagnosis and time of final clinic appointment. This endpoint was chosen to permit comparison with prior research (Weleff et al. 2023; Hirjak et al. ^2024b).

To determine if improvement in BFCRS, KCS, and KCE scores occurred over the study period, the most recent BFCRS, KCS, and KCE scores were collected from the last documented clinical note of each participant within the study period. Hierarchical models were then estimated to approximate the linear trends in catatonia symptom improvement between baseline clinical assessment and the end of the follow‐up period. All available BFCRS, KCS, and KCE data from either the baseline or most recent timepoints were modeled using outcome‐specific intercept‐only hierarchical linear or ordered‐probit models, with one model constructed for each catatonia outcome measure. Hierarchical linear models (used to model BFCRS and KCS scores) were fit using restricted maximum likelihood estimation in the lme4 R package (Bates et al. 2015), whereas the single hierarchical ordered‐probit model (Bürkner and Vuorre 2019) was fit using a Laplace approximation to maximum likelihood, as implemented in the clmm function in the ordinal R package (Christensen 2024), and both model types utilized Wald approximations for fixed‐effect p values and profile likelihood for the calculation of 95% confidence intervals. Participants with a BFCRS, KCS, or KCE at only one time point still had their data included in these models, as full information maximum likelihood estimation allowed for such participants to contribute to the marginal effects of each model. Effects were quantified as the raw mean difference in BFCRS/KCS score points over the course of the study (denoted using the regression slope, β), as well as the standardized mean difference (d/d _ord ) defined as the reduction in BFCRS/KCS/KCE outcome standard deviation units (or latent probit units for d _ord ) between timepoints 1 and 2. Additionally, CGI‐I scores for all participants were compared to a baseline score of 4 (“No change”) using a one‐sample ordinal statistical test (Wilcoxon signed‐rank test; rank‐biserial correlation r _c used as a standardized effect size). A two‐tailed p value of < 0.05 was used to determine statistical significance, and Bonferroni‐Holm corrections were performed to control the familywise error rate of the four statistical tests (test of timepoint parameters for BFCRS/KCS/KCE hierarchical models, CGI‐I Wilcoxon test).

3. Results

3.1. Demographics, Autism Diagnoses, and Length of Outpatient Follow Up

Overall, observational data was collected for 45 patients with autism and catatonia (Table 3). The mean age of patients in the clinic at index visit was 15.6 (SD = 7.9) years [Mdn = 16.0, range 6.0–31.0]. With regard to sex assigned at birth in the 45 patients, 31 (68.9%) were male and 14 (31.1%) were female. Medicaid was the primary payer for 30 (66.7%) patients, while 12 (26.7%) were covered by private insurance, and 3 (6.7%) were self‐pay. Forty‐one patients (91.1%) met criteria for autism with co‐occurring intellectual disability.

TABLE 3.

Demographics, genetic and medical co‐morbidity for patients with autism and catatonia.

	N	%
N	45
Sex
Male	31	68.9
Female	14	31.1
Age (Mean, SD)	15.6 (SD = 7.9)
Autism
Autism with intellectual disability	41	91.1
Autism without intellectual disability	4	8.9
Referral base
Recruited from CL psychiatry after initial catatonia diagnosis	25	55.6
Outpatient referral	20	44.4
Healthcare payer
Medicaid	30	66.7
Private insurance	12	26.7
Self‐pay	3	6.7
Race
Asian	3	6.7
Black	7	15.6
Indian	1	2.2
White	34	75.6
Ethnicity
Hispanic	1	2.2
Non‐Hispanic	44	97.8
Medical co‐morbidity
Seizure history	20	44.4
Epilepsy	7	15.6
Delirium	2	4.4
Anti‐NMDA receptor encephalitis	2	4.4
Genetic Co‐morbidity
16p12.2 microdeletion syndrome	2	4.4
17q12 microdeletion syndrome	1	2.2
Neurobeachin variant	1	2.2
GLUT1 deficiency syndrome	1	2.2
SYNGAP‐1 related intellectual disability	1	2.2
Phelan‐McDermid syndrome	1	2.2

Twenty‐five patients (55.6%) were first diagnosed and treated for catatonia in the inpatient setting by the consult‐liaison psychiatry service and referred to the clinic for continued outpatient care. Twenty patients (44.4%) were referred to the clinic from outpatient referrals. Patients were followed for a mean of 409.6 days (SD = 256.0), with a median of 295 days [range 58–910]. There were 477 total outpatient appointments during the study period, of which 320 (67.1%) were conducted via telemedicine. The mean number of total appointments for each patient was 11 (SD = 7.2), with a median of 10.

3.2. Co‐Occurring Medical, Genetic, and Psychiatric Conditions

Co‐occurring medical, genetic, and psychiatric conditions were common in this cohort (Tables 3 and 4). Twenty (44.4%) had a history of seizures, seven (15.6%) were diagnosed with epilepsy, and two (4.4%) patients each had prior diagnoses of delirium and Anti‐N‐methyl‐D‐aspartate (NMDA) receptor encephalitis. Additional psychiatric diagnoses were common: 35 patients (77.8%) were diagnosed with at least one psychiatric condition besides catatonia and autism. The most common of these was ADHD (n = 25; 55.6%).

TABLE 4.

Psychiatric co‐morbidity and clinical care for patients with autism and catatonia.

	N	%
N	45
Psychiatric co‐morbidity
Attention‐deficit hyperactivity disorder	25	55.6
Unspecified mood disorder	7	15.6
Obsessive compulsive disorder	7	15.6
Bipolar disorder	4	8.9
Unspecified psychosis	2	4.4
Major depressive disorder	2	4.4
Impulse control disorder	2	4.4
Disruptive mood dysregulation disorder	2	4.4
Conduct disorder	1	2.2
Post‐traumatic stress disorder	1	2.2
Generalized anxiety disorder	1	2.2
Total number of patients with history of psychiatric medication use	38	84.4
Number of failed psychiatric medications
Mean	4.8 (SD = 3.6)
Median	4
Range	1–16

3.3. Frequency of Index Catatonic Symptoms and Abnormal Physical Examination Findings

Tables 1 and 2 contain data reporting the frequency of index catatonic symptoms and physical examination findings as reported by the KCS/KCE and BFCRS, respectively. Forty‐two (93.3%) patients had a completed index KCS/KCE and thirty‐seven (82.2%) had a completed index BFCRS. The three most common symptoms documented on the KCS included impulsivity (n = 34, 80.9%), staring (n = 34, 80.9%), and stereotypies (n = 29, 69.0%). When explicitly measuring self‐injurious mannerisms and/or stereotypies (indicated by a KCS score of 4 or more for either item), 13 patients (28.9%) presented with this symptom, with 8 (19%) patients experiencing self‐injury due to stereotypies and 5 (11.9%) due to mannerisms. The three most common abnormal physical examination findings identified via the KCE were perseveration (n = 24, 60%), command‐verbal (n = 17, 42.5%), and gibberish or verbigeration (n = 14, 35%). The three most common catatonic symptom categories on the BFCRS included staring (n = 29, 78.4%), impulsivity (n = 28, 75.7%), and stereotypies (n = 22, 58.5%).

3.4. History of Prior Psychopharmacologic Treatment

As per Table 4, 38 (84.4%) patients presenting for longitudinal psychopharmacologic care had a history of previous psychiatric medication trials. The mean number of failed medications was 4.8 (SD = 3.6), with a median of 4 [range 1–16]. Table S1 reports medication trials by psychiatric medication class.

3.5. Use of Alternative benzodiazepines in the Treatment of Catatonia in Autism

Overall, 44 (97.8%) patients received benzodiazepines over the course of treatment (Smith et al. ^{2023a, 2023b}). The one patient (2.2%) who did not receive benzodiazepines was treated with memantine alone. Over half of patients (n = 24; 55.5%) were prescribed more than one benzodiazepine over the course of treatment. As seen in Table 5, over the course of the study, 33 (73.3%) patients were treated with clonazepam due to the long half‐life and previously reported efficacy of clonazepam in this patient population (Smith et al. ^{2023a, 2023b}). Thirty (66.7%) were treated with lorazepam, five (11.1%) with diazepam, and three (6.7%) with clobazam. Two patients (4.4%) were prescribed both clonazepam and clobazam. In these patients, clobazam was continued due to established efficacy and historic success in controlling seizures for these patients.

TABLE 5.

Management of benzodiazepines, electroconvulsive therapy, and antipsychotic tapering for catatonia in autism.

	N	%
N	45
Number of patients treated with a benzodiazepine	44	97.8
Number of patients treated with clonazepam	33	73.3
Number of patients treated with lorazepam	30	66.7
Number of patients treated with diazepam	5	11.1
Number of patients treated with clobazam	3	6.7
Number of patients who were either switched from one benzodiazepine to another or received two benzodiazepines simultaneously over the course of treatment	24	55.5
Benzodiazepine(s) prescribed at the end of the study period
Clonazepam	22	48.9
Lorazepam	12	26.7
Diazepam	4	8.9
Clobazam	1	2.2
Clonazepam+Clobazam	2	4.4
Patients never prescribed a benzodiazepine	1	2.2
Highest daily dosage of benzodiazepine in lorazepam equivalents
Mean, Median	17.4 mg (SD = 15.8), 10.3 mg
Range	1–64 mg
Received electroconvulsive therapy	16	35.6
Number of patients who underwent antipsychotic tapering	14	31.1
Number of patients who were tapered from more than one antipsychotic	3	6.7
Patients tapered from risperidone	5	11.1
Patients tapered from aripiprazole	4	8.9
Patients tapered from olanzapine	3	6.7
Patients tapered from quetiapine	2	4.4
Patients tapered from ziprasidone	2	4.4
Patients tapered from clozapine	1	2.2

3.6. Benzodiazepine Dosing in Lorazepam Equivalents and Tapering of Antipsychotics and Benzodiazepines

Over the course of the study period, the mean (within‐individual) highest daily dose of benzodiazepines in lorazepam equivalents for the treatment of catatonia was 17.4 mg (SD = 15.8) [Mdn = 10.3 mg, range = 1–64 mg]. Despite the high dosage of benzodiazepines, no severe adverse effects were reported, including respiratory suppression. Adverse effects were determined based on patient or informant report. Any aspect of the treatment of catatonia which would warrant acute medical care and/or hospitalization were considered severe adverse effects. Given the questionable efficacy of antipsychotics in catatonia (with the potential exception of clozapine) (Saini et al. 2024; Thom et al. ²⁰²³), 14 (31.1%) patients had antipsychotic medications reduced or discontinued after the identification of catatonia. Three (6.7%) were tapered from more than one antipsychotic.

By the end of the study period, 39 patients (86.7%) remained on at least one benzodiazepine (Table 5). At the end of the study period, 22 (48.9%) were prescribed clonazepam, 12 (26.7%) lorazepam, 4 (8.9%) diazepam, 1 (2.2%) clobazam, and 2 (4.4%) were prescribed both clonazepam and clobazam. After experiencing significant improvement in catatonic symptoms, 14 patients (31.1%) attempted to taper off benzodiazepines during the study period; of these, 5 patients (35.7%; 11.1% of the full sample) were successfully tapered off, and the remaining 9 (64.2%; 17.8% of the full sample) discontinued the taper or resumed benzodiazepine therapy due to a return of catatonic symptoms (Table 6).

TABLE 6.

Alternative pharmacologic management and benzodiazepine tapering for catatonia in autism.

	N	%
N	45
Number of pharmacologic agents at the end of study period
Patients on one pharmacologic agent	10	22.2
Patients on two pharmacologic agents	20	44.4
Patients on three pharmacologic agents	11	24.4
Patients on four pharmacologic agents	4	8.9
Patients on five pharmacologic agents	1	2.2
Number of patients treated with memantine	21	46.7
Mean memantine dosage	22.6 mg (SD = 13.1)
Number of patients treated with guanfacine	18	40.0
Mean guanfacine dosage	2.4 mg (SD = 1.3)
Number of patients treated with aripiprazole	11	24.4
Mean aripiprazole dosage	10.9 mg (SD = 6.2)
Number of patients treated with clozapine	6	13.3
Mean clozapine dosage	160.4 mg (SD = 136.1)
Number of patients treated with quetiapine	2	4.4
Mean quetiapine dosage	350 mg (SD = 212.1)
Number of patients treated with olanzapine	1	2.2
Mean olanzapine dosage	20 mg
Number of patients treated with valproic acid	9	20.0
Mean valproic acid dosage	861.1 mg (SD = 397.5)
Number of patients treated with lithium	6	13.3
Mean lithium dosage	725 mg (SD = 220.8)
Number of patients who underwent a benzodiazepine taper	14	31.1
Taper discontinued due to return of catatonic symptomology	8	17.8
Number of patients successfully tapered off benzodiazepines	5	11.1

3.7. Use of Nonbenzodiazepine Pharmacologic Agents in the Treatment of Catatonia in Autism

In addition to benzodiazepines, the use of other pharmacologic agents was common, with benzodiazepines, atypical antipsychotics, mood stabilizers, NMDA receptor antagonists, and alpha agonists used for the treatment of catatonia. Figure 1 provides a visualization of the number of psychopharmacologic medication classes used for treatment of catatonia in this cohort.

FIGURE 1.

Number of medication classes for catatonia in autism. Specific medication classes included atypical antipsychotics, mood stabilizers, NMDA receptor antagonists, and alpha agonists.

Of the specific nonbenzodiazepine medications used to treat catatonia in this cohort, the most frequently used were memantine (21 patients [46.7%]), guanfacine (18 patients [40%]), and aripiprazole (11 patients [24.4%]). Pharmacologic agents and doses are listed in Table 6.

3.8. Patterns of ECT Use for Patients With Catatonia and Autism

Due to partial or nonresponse to pharmacotherapy to the degree that the individual or family's quality of life was significantly impacted, ECT was recommended for 22 patients (48.9%). However, ECT was not pursued for six of these patients due to the following: parental preference (n = 2, 33.3%), transportation challenges (n = 2, 33.3%), difficulties in obtaining insurance coverage (n = 1, 16.7%), and underlying medical concerns (n = 1, 16.7%). Overall, 16 (35.6%) patients were treated with ECT due to lack of response to pharmacologic treatments alone. Of the 16 patients who received ECT, 7 (n = 43.8%) received ECT prior to clinic intake, and 1 patient (n = 6.3%) started ECT at an outside facility prior to clinic intake. For the remaining eight patients, the mean number of days from intake until the first ECT procedure was 181.6 days (SD = 184.9). The mean dose of benzodiazepine in lorazepam equivalents at the time of the first ECT procedure was 20.7 mg (SD = 15.3).

3.9. ABA For Autistic Individuals With Catatonia

For patients in our sample, 35 (77.8%) of patients received ABA services at some point across their lives. Over the study period, 19 patients (42.2%) previously received ABA services but did not while experiencing catatonia. ABA was discontinued/suspended for these patients for the following reasons: severe aggression (n = 5, 26.3%), parental preference to discontinue ABA services (n = 4, 21.1%), geographic location (n = 2, 10.5%), the COVID‐19 pandemic (n = 2, 10.5%), and loss of insurance coverage (n = 2, 10.5%). Sixteen patients (n = 45.7%) continued to receive ABA services while experiencing catatonic symptoms.

3.10. Longitudinal Clinical Outcomes for Catatonia in Autism

Of the 45 patients included in the study, 33 (73.83%) had a BFCRS, KCS, and KCE completed at the time of catatonia diagnosis or clinic intake if the patient was first diagnosed and treated for catatonia in the inpatient setting (n = 18, 51.4%).

Over the course of the study period, the marginal mean (±SEM) BFCRS score in the sample was reduced from 13.2 ± 1.0 to 8.7 ± 1.0 (β = −4.5, CI_95% [−6.5, −2.4], p _adj < 0.001, d = −0.70, CI_95% [−1.03, −0.38]). The marginal mean KCS reduced from 22.9 ± 1.9 to 19.5 ± 2.1 over the course of treatment, a difference that was not statistically significant (β = −3.4, CI_95% [−8.4, 1.6], p _adj = 0.178, d = −0.27, CI_95% [−0.66, 0.12]). The estimated marginal mean on the KCE latent trait score was reduced from −0.49 ± 0.21 (near the estimated threshold between a “1” and “2” score on the KCE) to −1.38 ± 0.27 (near the estimated threshold between a “0” and a “1” score on the KCE), a difference that was statistically significant (d _ord = −0.88, CI_95% [−1.41, −0.35], p _adj = 0.002). Lastly, the CGI‐I scores (Mdn = 2, range 1–3) were found to be significantly improved relative to a baseline (Figure S1) score of 4 (i.e., compared with “no improvement”; W = 0, p _adj < 0.001, r _c = −1, CI_95% [−1.0, −1.0]). Reductions in BFCRS, KCS, and KCE are plotted in Figure 2.

FIGURE 2.

Clinical outcomes for catatonia in autism. Specific clinical measures include the Bush Francis Catatonia Rating Scale (BFCRS), Kanner Catatonia Severity Scale (KCS), and Kanner Catatonia Examination (KCE). Assessment at timepoint 1 occurred upon clinic intake. Assessment at timepoint 2 was the most recent clinical assessment, which occurred over the study period.

Of note, 36 (80%) patients were followed for 6 months or longer. Similar to the full sample data reported in this manuscript, statistically significant reductions were observed in the BFCRS (β = −4.88, CI_95% [−7.15, −2.63], p _adj < 0.001), KCE (d _ord = −0.88, CI_95% [−1.41, −0.23], p _adj = 0.013). A nonstatistically significant reduction was noted in the KCS (β = −4.05, CI_95% [−9.48, 1.32], p _adj = 0.152). Reductions in the BFCRS, KCE, and KCS in this subgroup are plotted in Figure S2.

3.11. Mortality

There were three deaths during the study period (patient ages 17, 16, and 31 years at baseline; followed for 349, 206, and 263 days, respectively). All patients were assigned male at birth and were diagnosed with autism and co‐morbid intellectual disability. In addition, in all three patients, catatonia had improved slightly or moderately but not remitted by the time of death (catatonia‐specific CGI‐I ratings of 3, 2, and 2, respectively). A sensitivity analysis of CGI‐I scores was performed in which death (despite being ostensibly unrelated to catatonia or its treatment in these cases) was imputed as a score of 7 (“very much worse”) on the CGI‐I for these three patients. In that case, the result of the CGI‐I analysis was similar, albeit with a slightly smaller and less precise effect size (W = 126, p _adj < 0.001, r _c = −0.757, CI_95% [−1.0, −0.493]).

One patient died after choking on an excessive amount of food. This patient was receiving maintenance ECT, but their death was accidental and unrelated to ECT. Another died from recurrent pneumonia as discussed in a previous manuscript. This second patient had also received ECT, but their most recent procedure occurred 3 months before their death (Smith et al. ^2024b). For the final patient, the cause of death is unknown. The patient had a history of underlying cardiac pathology and developed a fever with lethargy the day before his death.

4. Discussion

To our knowledge, this study is the first observational longitudinal study to assess management and outcomes of catatonia in autism with and without intellectual disability. Patients in this cohort were followed for a median of 295 days and received extensive pharmacologic and ECT treatment (Rogers et al. 2023; Hirjak et al. ^2024a; Smith et al. ^2023b; Beach et al. ²⁰¹⁷). Despite this, most patients remained symptomatic to varying degrees over the course of the study period. Three patients died, one from pneumonia which has been historically associated with catatonia. The current connection between pneumonia and catatonia is unclear, but may be due to poor nutritional status, psychomotor changes, deconditioning, or high dosages of benzodiazepines needed for treatment (Meng et al. 2024). Given the young overall age of this cohort relative to pediatric catatonia cases from other conditions (Srinivasan, Baldwin, and Smith ²⁰²⁴; De Stefano, Palffy, and Ghaziuddin ²⁰²⁴), the high degree of morbidity and mortality provides further evidence of the substantial clinical impact of catatonia in autistic individuals (Cornic et al. 2009; Hirvikoski et al. ²⁰¹⁶). These findings reinforce the need for high level and specialized care for this patient population, along with high‐quality studies exploring further treatment options.

Like previous descriptive work in the field (Vaquerizo‐Serrano et al. ²⁰²²), patients in this cohort often presented with externalizing symptoms above baseline including impulsivity, combativeness, hyperactivity, and manneristic/stereotypic self‐injury. Overall, we found that over the course of clinical care there were improvements in the BFCRS, KCE, and CGI‐I, all with large effect sizes. As our primary end point, the statistically significant reduction in BFCRS scores represents a quantifiable change of specific catatonia symptoms and physical examination findings. In this population, assessing for the change in BFCRS scores is critical as BFCRS can be artificially elevated at baseline due to phenotypic overlap of autism and catatonia broadly (Vaquerizo‐Serrano et al. ²⁰²²; Smith et al. ^2024a). Although scores on the KCS improved in the same direction as the other measures over the course of longitudinal treatment, this change did not reach statistical significance. While numerous catatonia rating scales exist, the optimal assessment of catatonia in autistic patients remains unclear. Both the BFCRS and KCE assess for abnormal physical exam findings, while the KCS does not. The lack of improvement reported via the KCS suggests that changes in symptom burden in autistic patients with catatonia may be represented by improvement in abnormal physical examination findings associated with catatonia rather than symptom description (largely captured through an unstructured caregiver interview). Many symptoms captured in the KCS, including hyperactivity, negativism, mannerisms, stereotypies, impulsivity, incontinence, and combativeness, are present at higher rates in autistic patients with co‐morbid intellectual disability at baseline compared to autistic individuals without intellectual disability (Vaquerizo‐Serrano et al. ²⁰²²), likely due to factors other than catatonia. Whether specific combinations of catatonic signs or symptoms are best suited to capture improvement in catatonic features within the autistic population remains an active area of research.

The KCS, despite its creation as a developmentally sensitive measure of catatonic features, has not been rigorously psychometrically validated in the autistic population and may include items that function differentially for autistic children and adults with co‐morbid intellectual disability. This is especially relevant given the remarkably high prevalence of autism with intellectual disability (n = 41 [91.1%]) in this cohort. Consistent with previous research in pediatric catatonia, co‐occurring medical and psychiatric conditions were common in our cohort (Hirjak et al. ^2024a; Luccarelli et al. ^2024b).

The prevalence of autism with intellectual disability in this catatonia cohort is similar to those previously reported in the literature. Specifically, previous meta‐analytic work from Vaquerizo‐Serrano et al. (2022) suggested that the rate of intellectual disability for autistic individuals with catatonia is between 5.7% and 81.6%, but reaches nearly 100% in clinic‐based studies like the data reported in this study. However, it is notable that some studies included in this meta‐analysis may be over representative of individuals with autism and co‐morbid intellectual disability, especially in the inpatient setting (Wachtel 2019; Ohta, Kano, and Nagai ²⁰⁰⁶; Wing and Shah ^{2000, 2006}). Previous expert opinion work by Shah has also speculated that social–emotional impairment may be the driving force behind the development of catatonia in autism (Shah 2019). Given the high degree of correlation between social–emotional impairment and intellectual disability in autism (Bernard Paulais et al. 2019), it is difficult to fully differentiate these co‐occurring symptoms to determine which may be the primary risk factor for catatonia in autism. It has also been speculated that genetic syndromes (Raffin et al. 2018; Moore et al. ²⁰²²; Shillington et al. ²⁰²³) may increase one's risk for catatonia (Fink and Taylor 2006; Dhossche, Ross, and Stoppelbein ²⁰¹²). In our study, we identified 7 (15.5%) patients with co‐morbid genetic syndromes. While this study does not allow us to firmly state conclusions regarding the interplay between autistic symptomology, genetic syndromes, intellectual disability, and catatonia, it suggests there may be a cumulative effect of such factors and the overall risk of developing catatonic symptoms (Waizbard‐Bartov et al. ²⁰²³). Further research is urgently needed to investigate the relationship between catatonia, intellectual disability, and autism (Stedman et al. 2019; Russell ²⁰¹⁹).

All but one patient was treated with a benzodiazepine (97.8%). Patients were most often treated initially with clonazepam or lorazepam and over half were managed with more than one benzodiazepine. The mean highest daily dose of benzodiazepine required for treatment over the study period was 17.4 mg (SD = 15.8) lorazepam equivalents; the highest daily dose recorded was 64 mg. While there is minimal literature which directly addresses benzodiazepine dosing longitudinally, this number is substantially higher than the 6 mg reported in a longitudinal hospitalization study of pediatric catatonia from our group (Luccarelli et al. ^2024a). In autism populations specifically, doses as high as 27 mg have been reported. In these cases, symptoms of catatonia were refractory to high dose benzodiazepines but did respond to ECT (Wachtel 2019).

All benzodiazepines are positive allosteric modulators of the GABA‐A receptor (Griffin et al. 2013). However, differences in pharmacokinetic properties (e.g., absorption, half‐life) as well as pharmacodynamic properties (GABA‐A subunit specificities, potency) are present between individual agents. Thus, benzodiazepines may be expected to have varying efficacies in the treatment of catatonia. In our prior research, we found that a significant number of autistic children with catatonia, who demonstrated only a partial response to lorazepam, exhibited increased responsiveness to longer‐acting benzodiazepines, particularly clonazepam (Smith et al. ^2023b). However, despite these preliminary findings, no controlled comparison trials have been completed. Without this, firm conclusions regarding the comparative effectiveness of benzodiazepines cannot be established. This is an area which should be addressed in future research.

Over the study period, 14 patients (31.1% of the sample) experienced clinical stability over a period of time and underwent a benzodiazepine taper. Nearly two thirds of those who attempted the taper (9/14 [64.2%]) resumed benzodiazepine treatment due to a return of catatonic symptoms. Thus, only five patients in our sample (11.1% of the full cohort) were successfully tapered off benzodiazepines without a recurrence of catatonic symptoms. Given these challenges, further data is needed to explore the implications and possible benefits (as well as risk–benefit ratio) of long‐term benzodiazepine use for patients with autism and chronic catatonia. To our knowledge, this is the first study to document attempts at benzodiazepine tapering following stabilization of catatonic symptoms.

Along with benzodiazepines, ECT has long been considered the gold‐standard for treatment of catatonia across the lifespan and in autism spectrum disorders (Rogers et al. 2023; Hirjak et al. ^2024a; Smith et al. ^{2022, 2024b}; Wachtel et al. ²⁰²⁴). However, access is often limited by provider availability, stigma, and state‐dependent restrictive legislation which often specifically limits access for children (Espinoza and Kellner 2022; Ong et al. ²⁰²³). Previous research has called into question the clinical efficacy of benzodiazepines for catatonia in autism, focusing more on ECT as the primary treatment (Vaquerizo‐Serrano et al. ²⁰²²; Wachtel ²⁰¹⁹). In our study, 16 patients (35.6%) were treated with ECT due to partial response from pharmacologic approaches. ECT was recommended for another six patients, but not pursued due to family preference, transportation challenges, difficulties in obtaining insurance coverage, or underlying medical concerns. Overall, in our study the majority of patients did not receive treatment with ECT (n = 29, 64.4%). Thus, it remains imperative that future research investigate the utility of ECT for the treatment of medication partial responders or nonresponders. Data which addresses when to begin ECT in the course of treatment, as well as remission rate, recurrence risk, and long‐term outcomes (including both benefits and harms) of this treatment modality remains a greatly understudied but critical component of catatonia treatment.

The majority of patients in our cohort received ABA over the course of their lives (77.8%). However, only 45.7% of patients received ABA while experiencing catatonia. Given the efficacy of ABA for self‐injurious behaviors and social relatedness (Minshawi et al. 2014; Du, Guo, and Xu ²⁰²⁴), autistic individuals with catatonia may benefit from co‐occurring ABA while receiving pharmacologic and neuromodulatory interventions. However, it should be noted that severe symptoms of catatonia, such as aggression in our sample, may impair a patient's ability to participate in ABA services. Regarding other approaches, experts in the field have proposed a unique psycho‐ecological approach to behavioral interventions for catatonia in autism (Shah 2019). While such an approach has promise, further research is urgently needed to determine best practices for behavioral interventions for this patient population.

Regarding pharmacologic interventions beyond the use of benzodiazepines, this study adds to the existing literature which address the use of NMDA receptor antagonists, mood stabilizers, and some antipsychotics (traditionally aripiprazole and clozapine derivatives) (Smith et al. ^2023b; Beach et al. ²⁰¹⁷; Thom et al. ²⁰²³), as well as the likely utility of using multiple pharmacologic agents in the treatment of catatonia in autism. Over 75% of the patients in this cohort required multiple medications from various pharmacologic classes for stabilization. Notably, nearly one third of the cohort was tapered from antipsychotics for the treatment of catatonia. The association of antipsychotics and catatonia is another point that is challenging to fully address. There are concerns that use of antipsychotics in the presence of catatonia may ultimately result in neuroleptic malignant syndrome (NMS), a condition associated with a high degree of morbidity and mortality (Connell et al. 2023). Therefore, careful consideration, as well as prior initiation of benzodiazepine treatment is warranted before initiating antipsychotics when catatonia is on the differential. However, some antipsychotics (clozapine in particular) (Saini et al. 2024) have proven effective in managing catatonia, especially when the underlying cause of catatonia is thought to be related to a psychotic disorder (Patterson et al. 2023), though evaluation of psychosis can be challenging for individuals with autism and intellectual disability. Overall, careful monitoring for worsening of catatonic symptoms is warranted when any antipsychotic is used in treatment.

Strengths of this study include the large sample size and longer follow‐up period relative to other studies examining catatonia in autism. This study is also the first to present longitudinal outpatient pharmacologic and clinical outcome data in this population using standardized catatonia rating scales. One limitation in interpreting this data is that more than half of the patients in this cohort were diagnosed with catatonia and treatment was initiated in the inpatient setting prior to entering longitudinal outpatient care. Reductions in the catatonia assessment measures are likely greatly influenced by this, as symptoms of catatonia have been shown to improve rapidly with treatment of benzodiazepines (Luccarelli et al. ^2024c; Suchandra et al. ²⁰²¹). Another limiting factor is the open label nature of the study and that one author (J.R.S.) is the medical director of this outpatient service, and thus is responsible for reviewing and interpreting clinical data. This introduces potential for bias and should be considered when reviewing the results of this study. Lastly, given the nature of our prospective observational data, we are unable to fully determine the impact ECT made on the clinical improvement of these patients. We have attempted to address this by providing additional information regarding the time from intake to ECT procedure and dose of benzodiazepine when ECT was initiated. We hope to conduct future studies which will investigate the impact of ECT on this population broadly and when compared to catatonic individuals without autism.

The use of telemedicine in the evaluation and clinical care of these patients is also a limitation given the relevance of the physical exam to the assessment of catatonic features. At present, there are no studies which assess the sensitivity and specificity of catatonia examinations that occur over telemedicine. However, with careful involvement of family and other caregivers, telemedicine may be a viable option for catatonia assessment (Luccarelli et al. 2022). Given the high frequency of aggression, self‐injury, negativism, and impulsivity in this patient population, the use of telemedicine in evaluation may improve patient safety and increase inclusion of severely impacted individuals in future longitudinal catatonia clinics, and this should be studied in future prospective research. In addition, the presence of profound hyperactivity, aggression, and externalizing symptoms of catatonia may increase the difficulty in conducting a full catatonia examination. Previous research has shown that the BFCRS, KCS, and KCE can be successfully used in pediatric populations (Smith et al. ^2024a; Luccarelli et al. ^2024a; Smith et al. ^2023b). Still, the KCE, KCS, and BFCRS are not validated in this specific patient population (Erdoğan et al. 2022; Wilson et al. ²⁰¹⁵). Another limitation is the lack of specific intelligence quotient scores for patients, which could be utilized to further investigate the link between intellectual disability and catatonia. Finally, treatments were provided as part of routine clinical care and not as part of a prespecified clinical protocol, with most patients receiving multiple treatments simultaneously. As a result, this study cannot provide direct evidence for individual treatments' effectiveness in this population.

Overall, this study emphasizes the need for discovery regarding the intersection of catatonia in autism and provides further evidence for the substantial clinical impact of catatonia on autistic individuals, particularly those with co‐morbid intellectual disability, and demonstrates the feasibility of outpatient management of catatonic symptoms. While outpatient treatment resulted in large improvements in catatonic symptoms as measured using the BFCRS, KCE, and CGI, residual symptoms remained in most patients despite receiving the gold standard of psychopharmacologic care for catatonia. Notably, for patients who were first identified in the inpatient setting, they may have experienced significant clinical improvement prior to their first outpatient evaluation. Overall, ongoing research is needed to develop improved diagnostic tools to assess catatonia in individuals with autism and to develop more effective treatments.

Ethics Statement

This study was reviewed by the Vanderbilt University IRB (#170317) with a waiver of informed consent from participants as data which was collected was based on routine clinical activities.

Conflicts of Interest

J.R.S. receives funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, Axial Therapeutics, and Roche. ZJW has received consulting fees from Roche and Autism Speaks. He also serves as the Vice‐chair of Autistic and Neurodivergent Scholars Working for Equity in Research (ANSWER), a division of the Autism Intervention Research Network on Physical Health (AIR‐P). J.W. receives support from the Department of Veterans Affairs, Geriatric, Research, Education and Clinical Center (GRECC) at the Tennessee Valley Healthcare System in Nashville, TN. J.L. receives funding from Harvard Medical School, the Rappaport Foundation, the American Academy of Child and Adolescent Psychiatry, and the Foundation for Prader‐Willi Research. He holds equity and has received consulting income from Revival Therapeutics Inc.

Supporting information

Data S1. Supporting Information.

Figure S1. Clinical Global Impression‐Improvement Scores.

Figure S2. Clinical Outcomes for Catatonia in Autistic Patients Followed For 6 Months or Longer. Specific clinical measures include the Bush Francis Catatonia Rating Scale (BFCRS), Kanner Catatonia Severity Scale (KCS), and Kanner Catatonia Examination (KCE).

Table S1. Number of failed psychiatric medications per class.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.