Catatonia Under-Diagnosis in the General Hospital

An initial study by our group²³ detected 54 cases of inpatients who were diagnosed with catatonia by the treating clinicians and met DSM-5 criteria for catatonia retrospectively at the University of Chicago between 2011 and 2013. After institutional review board approval was obtained, another retrospective chart review was conducted of all adult inpatients at the University of Chicago between 2011 and 2013. In order to obtain charts with potentially under-diagnosed catatonia, we ideated a list of 38 keywords describing catatonia-related signs and used the electronic medical record to select the cases for review.

The presence of three or more distinct keywords describing catatonic-related signs in the progress notes was used to identify cases for review. Charts were reviewed for evidence of catatonia during hospital admission by use, retrospectively, of DSM-5 criteria. Case subjects with a documented clinical diagnosis of catatonia were not reviewed again, since they were already included in the diagnosed group.

Initially, authors J.R.L. and M.M. reviewed separately all the notes of the selected charts to see if the case was a possible episode of catatonia as described by the DSM-5. During the initial chart review, the words echolalia and echopraxia were also searched in every chart, since they define DSM-5 criteria and had not been included in the initial list of catatonia-related keywords. All cases in which at least one reviewer identified possible catatonia were reviewed again independently by authors J.R.L. and M.M. in order to detail the DSM-5 criteria met during the episode. Only case subjects with three or more documented DSM-5 criteria were included in the study.

Each reviewer used a standardized sheet and only attributed the clinical signs to catatonia when there was no alternative better explanation. For example, the presence of grimacing documented to be related to pain, mutism due to aphasia or intubation, stupor due to coma, and decerebrate or ictal posturing were not considered to be signs of catatonia. Sociodemographic and clinical data, including the total number of Bush Francis Catatonia Screening Instrument criteria met, were obtained for all case subjects who met inclusion criteria.

This study is the first, to our knowledge, to estimate the proportion of under-diagnosed catatonia in the general hospital and to find potential factors contributing to the under-diagnoses. Of the 133 case subjects who retrospectively met three or more DSM-5 criteria for catatonia in the general hospital over a 3-year period, the majority (N=79, 59%) were not diagnosed with catatonia during admission. Under-diagnosed catatonia was more frequent in the presence of agitation, echolalia, or grimacing; conversely, completion of a psychiatric consultation during the episode substantially decreased the likelihood of under-diagnosis of catatonia.

Case subjects meeting DSM-5 criteria for catatonia retrospectively who received a psychiatry consultation during admission had greater than 44 times the odds of being diagnosed with catatonia compared with those who did not receive psychiatric consultation. This finding supports the crucial role of psychiatry consultation teams in detecting catatonia in the general hospital. However, more than one-third (37%) of under-diagnosed case subjects in our sample did receive a psychiatric consult, which suggests, in agreement with previous literature,^7,16 that under-recognition was frequent among case subjects evaluated by the psychiatry team and supports the need for greater recognition of catatonia across disciplines.

Our findings also suggest that under-recognition of catatonia may relate to physician unawareness of the heterogeneous cluster of signs and symptoms that constitute this syndrome. In particular, this study is the first to suggest that clinicians might often not consider agitation, grimacing, and echolalia to be attributable to catatonia. While these results warrant replication in other studies, there is indirect partial support for our findings from prior literature. A study of drug-naïve patients with first-episode psychosis found that agitation and grimacing were the signs of catatonia least correlated to other catatonic signs,⁹ and case reports have shown under-detection of catatonia in agitated patients.^33,34 Thus, clinicians should be aware of the possibility of unexplained agitation being secondary to an underlying catatonic state.

The prevalence of grimacing and echolalia also independently predicted under-diagnosis of catatonia. One explanation for these results is that these clinical signs may be more likely than other signs of catatonia to be attributed to other conditions. For example, grimacing is a common clinical sign for pain, and treating physicians may attribute it to other medical conditions and not recognize it as a symptom of catatonia in their differential diagnoses. Echolalia is found in multiple neuropsychiatric disorders, including autism, dementia, and Tourette’s syndrome.

Of note, one study³⁵ reviewing the clinical characteristics of nonconvulsive status epilepticus (NCSE) presentations found that echolalia, mutism (defined as aphasia/interrupted speech in the study), automatisms, catalepsy, staring, and increased tone are frequent features in patients with NCSE. Thus, NCSE seems to share clinical features with catatonia, in that it presents frequently with delirium in the elderly and tends to respond to treatment with GABA A receptor agonists such as lorazepam or midazolam. Indeed, the association of NCSE with catatonic presentations has been suggested by prior literature.^36,37 It has also been argued that, in comparison to classic psychiatric literature, modern research on catatonia has focused more on motor rather than verbal signs of catatonia, including echolalia and verbigeration.³⁸ Our study suggests that there is a need for increased awareness of both verbal and nonverbal signs of catatonia, and that excited catatonia might be more frequently missed than retarded catatonia, given that agitation and grimacing are more frequently associated with excited catatonia.¹⁵

Interestingly, there were no significant group differences in variables like mean number of symptoms or distribution of number of symptoms, suggesting that it is the patterning of specific symptoms and not the overall number of symptoms that predicts under-diagnosis. There were no differences in rates of suspected delirium or NMS between groups, which suggests that these might not be factors driving the under-diagnosis. Of note, a high proportion of abnormal EEG findings were found in our sample, as have been reported in patients with catatonia regardless of an underlying medical or psychiatric etiology.³⁹ Our findings support prior literature showing that an abnormal EEG does not preclude the diagnosis of catatonia.

We also found limited evidence for demographic differences between groups. There was no difference between diagnosed and under-diagnosed groups in gender or past history of psychiatric diagnosis. Although age and race were initially significant in the univariate analyses, follow-up analyses (not shown) revealed that this was due to confounds with psychiatric consultations. Specifically, younger patients and African American patients were more likely to receive a psychiatric consultation. Similarly, these follow-up analyses revealed that neurology team evaluation and EEG tests were less common when a psychiatry consultation was requested, and all but two of the emergency department cases were seen by psychiatry consultants. Thus, when all of the factors were considered in the multiple regression simultaneously, only psychiatric consultation and the three clinical signs had significant, independent effects on under-diagnosis.

Finally, we note that there were some clinically relevant differences between groups in treatment received and clinical outcomes. First, a similar proportion of diagnosed and under-diagnosed cases received treatment with lorazepam during the episode, likely because lorazepam is widely used in our hospital for agitation management. However, under-diagnosis was related to lower daily doses of lorazepam received and lower rate of standing lorazepam treatment. Lorazepam is an effective treatment for catatonia, but high doses of lorazepam for several days or longer might be needed to treat catatonia effectively.^5,28,40 Thus, our study suggests that under-diagnosed cases received suboptimal treatment.

Moreover, even in diagnosed cases, the lorazepam treatment rates, standing lorazepam treatment rates, and average doses were relatively small, suggesting suboptimal treatment even when the syndrome is recognized. Suspected delirium comorbidity might explain suboptimal treatment among diagnosed cases, since in a secondary analysis of our data we found that treated case subjects with suspected comorbid delirium received significantly fewer total lorazepam doses than those without delirium only in the diagnosed group (N=17, mean=0.7, SD=0.7 mg/day versus N=12, mean=2, SD=1.6 mg/day, p=0.006). This finding suggests that clinicians might be reluctant to give optimal lorazepam doses in patients with catatonia and delirium, and that more evidence is needed on how to manage catatonia in the presence of delirium.

Although the data suggest that under-diagnosed catatonia case subjects have increased mortality during admission and increased length of stay, this was not statistically significant. Our study might have been underpowered to detect differences in mortality, given that less than 10% of the sample died during admission. Regarding length of stay, differences should be interpreted with more caution, since we did not compare differences in specific disposition among case subjects who did not die, such as a transfer to psychiatric hospital for catatonia management. However, we also suspect that our study might have been underpowered to detect differences in length of stay related to under-diagnosis of catatonia.

Overall, our findings suggest that under-diagnosis of catatonia might have played a role in increasing likelihood of poorer hospital outcomes. Future studies with larger samples could examine the effects of under-diagnosis on clinical outcomes.

A significant number of cases of catatonia in the general hospital might not be properly diagnosed and might be suboptimally treated. Consultation-liaison psychiatrists have a crucial role in the diagnosis and treatment of catatonia in the general hospital, and should implement liaison and educational strategies to increase the knowledge about catatonic signs, such as unexplained agitation, grimacing, or echolalia. Improving detection and treatment of catatonia could help improve clinical outcomes of patients with this reversible syndrome in the general hospital.