Authors: James Luccarelli (1Harvard Medical School, Boston, MA, USA; 2Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA), Thomas H. McCoy, Jr (1Harvard Medical School, Boston, MA, USA; 2Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA), Tasia York (3Division of Child and Adolescent Psychiatry, Department of Psychiatry and Behavioral Sciences; Vanderbilt University Medical Center at Village of Vanderbilt, 1500 21st Avenue South, Suite 2200, Nashville, Tennessee, 37212), Isaac Baldwin (rd), Gregory Fricchione (1Harvard Medical School, Boston, MA, USA; 2Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA), Catherine Fuchs (3Division of Child and Adolescent Psychiatry, Department of Psychiatry and Behavioral Sciences; Vanderbilt University Medical Center at Village of Vanderbilt, 1500 21st Avenue South, Suite 2200, Nashville, Tennessee, 37212), Joshua R. Smith (3Division of Child and Adolescent Psychiatry, Department of Psychiatry and Behavioral Sciences; Vanderbilt University Medical Center at Village of Vanderbilt, 1500 21st Avenue South, Suite 2200, Nashville, Tennessee, 37212; 4Vanderbilt Kennedy Center, Vanderbilt University; 110 Magnolia Circle, Nashville, TN, 37203; rd)
Categories: Article, catatonia, cohort studies, diagnosis, child psychiatry, neurodevelopmental disorders, lorazepam
Source: Schizophrenia research
Authors: James Luccarelli, Thomas H. McCoy, Tasia York, Isaac Baldwin, Gregory Fricchione, Catherine Fuchs, Joshua R. Smith
Catatonia is a neuropsychiatric disorder associated with changes in behavior and affect. In adults, catatonia can respond rapidly to treatment with benzodiazepines as part of the “lorazepam challenge test.” The acute effectiveness of benzodiazepine treatment in pediatric catatonia, however, has received less study. This study reports catatonia severity as measured by the Bush Francis Catatonia Rating Scale (BFCRS) in pediatric patients before and after treatment with lorazepam.
Multicenter retrospective cohort study from 1/1/2018 to 6/1/2023 of patients aged 18 and younger with a clinical diagnosis of catatonia and assessment using the Bush Francis Catatonia Rating Scale (BFCRS) before and after treatment with lorazepam.
Among 54 patients, median age was 16, and 26 (48.1%) were female. Neurodevelopmental disabilities were present in 24 (44.4%) of patients. Prior to treatment, patients had a mean BFCRS score of 16.6 ± 6.1, which significantly reduced to 9.5 ± 5.3 following treatment with lorazepam (mean paired difference 7.1; t=9.0, df=53, p<0.001), representing a large effect size (Hedges’s g = 1.20; 95% CI: 0.85 to 1.55). No significant association was found between lorazepam dose or route of administration and clinical response, nor were age, sex, study site, the presence of a neurodevelopmental disorder, the presence of hyperactive catatonic features, or the time between treatment and reassessment associated with post-treatment BFCRS.
Lorazepam resulted in a rapid improvement in BFCRS score in pediatric patients, with a large effect size. Further research is needed into optimal dosing and route of administration of the lorazepam challenge test in pediatric patients.
Catatonia is a neuropsychiatric disorder which is characterized by changes in psychomotor function and affect. Catatonia presents in patients of any age and may be caused by a myriad of neurodevelopmental, genetic, medical, or psychiatric conditions across the lifecycle. (Fink, 2013; Luccarelli et al., 2022a, 2022b; Oldham, 2024; Patterson et al., 2023; Raffin et al., 2018; Smith et al., 2023a) Clinically, catatonia can vary between hyperactive and hypoactive states within the same episode of illness, complicating diagnosis and treatment. Moreover, children experience unique signs of catatonia including a loss of previously acquired skills / communicative abilities, urinary incontinence, acrocyanosis, and schizophasia.(Benarous et al., 2018, 2016; Rogers et al., 2023) In neurodiverse individuals, externalizing signs including aggression and self-injury are more common, increasing the difficulty of clinical assessment. (Smith et al., 2023a; Smith et al., 2023b; Vaquerizo-Serrano et al., 2022) Thus, catatonia can be difficult to identify and is frequently underdiagnosed in pediatric and neurodiverse patients. (Luccarelli et al., 2022a; Vaquerizo-Serrano et al., 2022) This is of significant clinical concern as misdiagnosis or delayed identification of catatonia may result in progression to malignant catatonia, a condition associated with autonomic instability, and mortality rates as high as 10%-20% if left untreated.(Walther et al., 2019) In children specifically, pediatric catatonia is associated with significant morbidity,(Walther et al., 2019) a sixty-fold higher risk of death for pediatric catatonia patients relative to the general population,(Cornic et al., 2009) and strain on hospital systems. (Becker et al., 2020; Hickox et al., 2023) In light of these findings, there is an urgent need to better assess the treatment response in pediatric catatonia patients.
To address these challenges, systematic clinical assessment tools for catatonia are often used which involve a clinical rating scale that grades each of the numerous possible catatonic signs.(Sienaert et al., 2011) The most utilized of these is the Bush Francis Catatonia Rating Scale (BFCRS),(Bush et al., 1996) which consists of a 14-item Bush Francis Catatonia Screening Item assessing for the presence of absence of signs, (BFCSI) and the full 23-item BFCRS which grades signs on a scale from 0 to 3. Once catatonia is identified on systematic assessment, the clinical diagnosis can be confirmed and treatment initiated using the lorazepam challenge test (LCT). While the precise details of the LCT vary between sources, broadly it involves the administration of a benzodiazepine (typically lorazepam) by oral (PO), intravenous (IV) or intramuscular (IM) routes at a clinically determined dose, with the goal of reducing the catatonia rating scale score by a significant amount.(Fricchione et al., 1983; Hirjak et al., 2023; Rogers et al., 2023) In adults, the LCT is associated with rapid relief of catatonic signs; a prospective trial by Suchandra and colleagues reported a reduction in BFCRS scores from 14.6±5.5 to 7.2±5.4 five minutes after IV administration of 2 mg or 4 mg of lorazepam, without statistically significant differences between responses at the two doses.(Suchandra et al., 2020) The utility of the LCT has been reported in case reports in pediatric patients.(Ridgeway et al., 2021; Sharma et al., 2014) However, some studies have called the utility of lorazepam in pediatric and neurodiverse catatonia into question,(Smith et al., 2023a; Smith et al., 2023b; Termini et al., 2023; Vaquerizo-Serrano et al., 2022; Wachtel et al., 2018) and the use of the LCT in catatonic children has not been explored in larger datasets. This study assesses the efficacy of the LCT in a large multicenter retrospective cohort of youth with a diagnosis of catatonia and assessment using the BFCRS, and explores the association between baseline demographic factors, lorazepam dosing and route of administration, and baseline catatonic signs on the BFCRS score following treatment.
The clinical records of two large health systems were queried for patients with a discharge ICD-10 diagnostic code for catatonia (F06.1 or F20.2) between 1/1/2018 and 6/1/2023. Patients were included in the study if they were aged 18 or younger, had a clinical diagnosis of catatonia with at least two catatonic signs on the BFCSI, were treated with a benzodiazepine, and had full BFCRS documented before and after LCT. The BFCRS was the catatonia rating scale used at both study sites,(Smith et al., 2024) with exams performed by resident or fellow psychiatrists under the supervision of attending pediatric consultation-liaison psychiatrists. The BFCRS has demonstrated high inter-rater reliability in previous studies.(Bush et al., 1996; Sarkar et al., 2016; Zingela et al., 2021) Primary discharge diagnosis, demographics, benzodiazepine dose, and BFCRS scores before and after treatment were then extracted from documented catatonia exams in the clinical records. Data on sex, race, and ethnicity were extracted from the clinical record, and were based on patient or family self-report. This study was approved by the Institutional Review Board of each study site (Vanderbilt University IRB: 230097; Mass General Brigham IRB: 2022P000811) with a waiver of informed consent from participants.
Catatonic features were defined from the BFCRS,(Bush et al., 1996) a 23-item scale ranging from 0 to 69, with higher scores indicating more severe symptomatology based on severity, duration, or frequency of the individual item measured. Patients were considered to have hyperactive features of catatonia if they demonstrated signs of excitement, impulsivity, or combativeness on initial BFCRS.
Demographics, BFCSI signs, and BFCRS scores before and after LCT are presented using descriptive statistics. Differences in BFCRS before and after administration of a benzodiazepine were compared using paired t tests, with effect sizes described using Hedges’s g. The association between the administered lorazepam dose and the change in BFCRS was assessed using Pearson correlation. In the primary statistical analysis, the BFCRS score following LCT was analyzed in a linear model with age, sex, study site, the presence of a neurodevelopmental disorder (NDD), pre-treatment BFCRS, the presence of hyperactive features of catatonia, benzodiazepine dose (in mg/kg of lorazepam equivalents), route of benzodiazepine administration (PO, IM, IV), and time between treatment and reassessment (minutes) as descriptor variables. Statistical analyses were done using SPSS (Version 29.0. Armonk, NY: IBM Corp).
A total of 54 patients met inclusion criteria (Table 1), of whom 26 (48.1%) were female. Mean age was 14.1 ± 3.7, with a median of 16 (IQR: 13.75 to 16.25); the youngest patient was 3 years old. All were inpatients at the time of initial assessment. Neurodevelopmental disorders (NDDs) were present in 24 (44.4%) of patients, with 16 patients with autism spectrum disorder (ASD) and 18 with intellectual disability (ID), including 13 patients with both ASD and ID. Diagnostically, psychotic disorders were the most common disorder associated with catatonia, with 17 patients (31.5%) diagnosed with a psychotic disorder. NDDs were the primary associated diagnosis for 10 patients (18.5%), mood disorders for 6 (11.1%), and medical disorders for 10 (18.5%). An additional 11 patients (20.4%) were diagnosed with unspecified catatonia (Table 2).
At the time of initial catatonia diagnosis, patients presented with a mean of 6.2 ± 1.7 signs on the BFCSI, with a median of 6 (IQR: 5 to 7). The initial BFCRS score was a mean of 16.6 ± 6.1. In total, 28 patients (51.9%) endorsed hyperactive features of excitement, impulsivity, or combativeness. All patients received treatment with lorazepam, with treatment doses ranging from 0.5 to 2 mg; this corresponds to a median weight-based dose of 0.029mg/kg of lorazepam (IQR: 0.020 to 0.038 mg/kg). Lorazepam was administered intravenously for 33 (61.1%), orally for 15 (27.8%), and intramuscularly for 6 (11.1%).
Patients were re-assessed at a median of 60 minutes following treatment (IQR: 30 to 130 minutes). Following LCT, BFCRS was significantly reduced to 9.5 ± 5.3 (mean paired difference 7.1; t=9.0, df=53, p<0.01), representing a large effect size (Hedges’s g = 1.20; 95% CI: 0.85 to 1.55) (Figure 1; Figure S1). In total, 20 patients (37.0%) had BFCRS scores decrease by >50% from baseline. There was not a univariate association between the dose of lorazepam administered and a change in BFCRS score following LCT (Pearson correlation 0.04; 95% CI: −0.24 to 0.31; p= 0.78) (Figure 2).
Improvements were seen in all items of the BFCSI and BFCRS with the LCT. Following LCT, lower percentages of patients endorsed each item of the BFCSI. Among individual catatonic signs, the largest absolute difference in the number of patients no longer expressing an individual sign on the BFCSI is for posturing/catalepsy, a sign present for 30 patients pre-LCT reducing to 14 patients post-LCT (Figure 3). Mean severity scores were also reduced for each item of the BFCRS (Figure S2). Among BFCRS items, the mean severity of mutism decreased by the largest amount after LCT, from a mean severity of 1.19 to a mean severity of 0.63 (Figure S2).
In order to explore demographic, diagnostic, and treatment factors associated with LCT treatment response, the post-treatment BFCRS was analyzed in a linear model (Table 3). In this model, higher pre-treatment BFCRS was associated with higher BFCRS post-LCT (β = 0.40; 95% CI: 0.09 to 0.71; p < 0.01), while sex, age, study site, NDD diagnosis, the presence of hyperactive features, lorazepam dose, route of lorazepam administration, and time between lorazepam administration and re-assessment were not significantly associated with final BFCRS.
This multi-site sample of 54 pediatric catatonia patients, treatment with lorazepam as part of the LCT resulted in a rapid and substantial reduction in catatonia severity as measured by the BFCRS. The magnitude of improvement (Hedges’s g = 1.20) compares favorably to that of other psychiatric treatments. For instance, meta-analytic estimates of antidepressant efficacy relative to placebo for pediatric patients found smaller effect sizes for anxiety disorders (g = 0.56), depressive disorders (g = 0.20), and OCD (g = 0.39).(Locher et al., 2017) Pediatric patients with lorazepam displayed mean improvement in all BFCSI and BFCRS items, and effectiveness of LCT did not differ based on a range of demographic and diagnostic factors. Both hyperactive and hypoactive catatonia patients responded to treatment with lorazepam, with no difference in post-LCT BFCRS among hyperactive and hypoactive patients in an adjusted model. Moreover, results did not differ among the study sites, which enhances the generalizability of results.
These retrospective results are consistent with the magnitude of improvement demonstrated using the LCT in a prospective trial in adults (BFCRS reduced from 14.6±5.5 to 7.2±5.4; N = 57).(Suchandra et al., 2020) In that trial, however, 59.6% of patients had a >50% reduction in catatonia severity following LCT, while the corresponding proportion of pediatric patients was 37.0% in this study. As 50% improvement in BFCRS score has previously been suggested as a cut-off score for a successful LCT, our results suggest that this threshold on the BFCRS may not be applicable in pediatrics.(Rogers et al., 2023) Regarding symptomology in catatonia which is specific to pediatric populations, mutism on the BFCRS was associated with the greatest reduction in severity pre- and post-LCT. Mutism on the BFCRS is the lone item that firmly represents a regression/loss of previously acquired skills, a common finding in pediatric catatonia.(Benarous et al., 2018, 2016; Rogers et al., 2023)
In this cohort the baseline prevalence of NDDs was high (44.4%), but response to LCT did not differ among patients with and without NDDs in an adjusted model. Such patients can present a diagnostic challenge, as features such as social-emotional impairment, restricted and repetitive interests, cognitive impairment, or impulsivity overlap substantially with features of catatonia measured by the BFCRS.(Smith et al., 2023a; Vaquerizo-Serrano et al., 2022) In addition, signs of catatonia including affective fluctuations and recurrent self-injury occur in neurodiverse populations but are not captured by the BFCRS.(Carroll et al., 2008; Wachtel, 2019) Regardless, the results of our study suggest that despite these challenges, the timely identification of catatonia in NDD patients is critical as they may derive the same magnitude of improvement as neurotypical patients.
Regarding dosages for the LCT, our research did not find an association between lorazepam dosing and symptom improvement as measured by the BFCRS. This is consistent with a previous prospective investigation in adults,(Suchandra et al., 2020) which did not demonstrate a differential response with lorazepam doses of 2 mg or 4 mg. These findings suggest that the optimal dosing of the LCT in pediatric catatonia remains to be determined, particularly among pediatric patients who may present with wide variability in weight and for whom weight-based dosing may be preferable to a fixed dose regimen. Moreover, post-LCT BFCRS score did not vary based on route of administration in the adjusted model, which is consistent with the excellent bioavailability of lorazepam by enteral and parenteral routes. These results do imply that oral dosing for the LCT can result in equivalent response as PO and IV formulation, which may be particularly helpful for inpatient psychiatric facilities where access to parenteral forms of administration may be challenging. The retrospective nature of this data does not allow for determination of speed of response, however, and so parenteral (particularly IV) lorazepam may result in greater speed of response, and this should be explored in future prospective studies. Overall, our data continues to support the use of benzodiazepines in the treatment of pediatric catatonia, which is of significant importance given the challenges around pediatric access to electroconvulsive therapy, the other best-studied treatment of catatonia in youth.(Espinoza and Kellner, 2022; Miller et al., 2022; Ong et al., 2023)
Strengths of this study include its large sample size, inclusion of multiple sites, and broad inclusion criteria encompassing pediatric patients with a range of medical, neurodevelopmental, and psychiatric disorders. Limitations are inherent in the retrospective nature of cohort generation, relying on documentation produced as part of routine clinical care. To be included in the sample, all patients required treating clinicians to assess for catatonia, interpret results of the BFCRS, provide treatment with a benzodiazepine, and document a repeat BFCRS following treatment. While the BFCRS is the standard catatonia assessment at both study sites, it is not performed in all patients, and so less-severe catatonic presentations may not have been appropriately identified. Moreover, there may be systematic biases in recording of post-LCT BFCRS, with patients demonstrating a greater response more likely to have a repeat assessment documented, which may skew results. Reassuringly, however, results in this sample match closely with those from a prospective trial in adults,(Suchandra et al., 2020) which gives greater confidence in the results presented here. However, the BFCRS has not been specifically validated for use in pediatric or neurodiverse patients, which should be considered when comparing data across studies. Additionally, results from these academic health systems may not translate to other healthcare settings or to populations of different sociodemographics. Moreover, the results presented here are after initial treatment with lorazepam, with response measured at a mean of 60 minutes. It is unclear for how long treatment response may be sustained, and if additional benefit may derive from subsequent doses of lorazepam. We are also unable to assess for duration of catatonia prior to treatment with lorazepam, and whether this variable may modulate the effectiveness of the LCT. Finally, this retrospective sample all received treatment with lorazepam, with no control group receiving placebo, so it is unclear to what extent non-specific effects of treatment may have driven the observed change in BFCRS. Further prospective studies are required to definitively assess the effectiveness of benzodiazepines compared to other treatments for the acute treatment of catatonia.
In conclusion, in a multi-site retrospective cohort of 54 pediatric catatonia patients, treatment with lorazepam according to the LCT resulted in a significant reduction in BFCRS total score, with a large effect size. There was not a significant association between lorazepam dose or route of administration and clinical response, nor were age, sex, study site, the presence of a neurodevelopmental disorder, the presence of hyperactive catatonic features, or the time between treatment reassessment associated with post-LCT BFCRS. Further research is needed into optimal lorazepam dosing and route of administration in pediatric patients with catatonia.