Psychosis-Like Homicidality Associated With Corticosteroid-Induced Anxiety Disorder


Mr. X, an 82-year-old man of Caucasian descent, had a medical history of hypertension and no previous psychiatric history and a mother with a history of unspecified anxiety. He had worked as a high-level executive until retirement at the age of 80. At the beginning of the year, he sought care from a rheumatologist for bilateral shoulder pain and was diagnosed as having polymyalgia rheumatica. He was started on prednisone titrated up to 15 mg daily, resulting in favorable symptom resolution.


Three months after the patient started prednisone, his wife expressed concerns to his primary care physician (PCP) about his increasing irritability and anxiety. She noted that her own worsening neuropathy and increasing dependence on him were possibly contributing to his anxiety. During the PCP visit, Mr. X affirmed increased overall anxiety, which began after the initiation of his steroid treatment. Because of these factors, as well as a normal Mini-Cog screen, a neurocognitive disorder was felt to be unlikely during this assessment, and it was decided to taper prednisone to 5 mg to reduce potential medication-induced anxiety.


Two months after the PCP visit, Mr. X received care in our emergency department (ED) and was placed under involuntary detention because of danger to self and others, with a preliminary diagnosis of corticosteroid-induced psychosis. This diagnosis was initially suspected in the ED because of the uncharacteristic nature of the patient’s actions. That night, his wife had called 911 after he swung a hammer at her, intending to end both of their lives, stating, “I can’t take care of you anymore.” In a clinical interview, his wife revealed that in the weeks after the PCP visit, he had developed worsening anxiety associated with fatigue, insomnia, and restlessness despite a decrease in his steroid regimen.


On the first full day of inpatient psychiatric hospitalization after transfer from the ED, Mr. X showed a depressed affect and mild distractibility but was otherwise cognitively intact without evidence of suicidal ideation, homicidal ideation, or psychotic symptoms. He fully recalled the situation and expressed extreme guilt and deep remorse for his actions. He explained how he had lost his sense of purpose and identity after retirement and decided that ending both their lives was the only solution to resolving their distress. Results of a further laboratory workup (which included a comprehensive metabolic panel, complete blood count, thyroid-stimulating hormone test, C-reactive protein test, folate test, vitamin B12 test, and urinalysis) were unremarkable. Urine toxicology was negative for any substances, and brain MRI findings obtained several days after admission were likewise unremarkable.


The differential diagnosis included corticosteroid-induced psychosis, corticosteroid-induced anxiety disorder, major depressive disorder with psychotic features, adjustment disorder with anxiety, brief psychotic disorder, and exacerbation of unspecified anxiety disorder. Mr. X was started on olanzapine 2.5 mg daily, and prednisone was discontinued. However, subsequent collateral information and patient monitoring revealed no evidence of mania, psychosis, catatonia, or delirium contributing to the patient’s acute presentation, leading to a change in treatment from olanzapine to mirtazapine. A depressive disorder was considered as a cause for the patient’s presentation. However, many of his depressive symptoms (such as dysthymic affect, extreme guilt, and suicidality) were likely exacerbated during the acute situation, and it was determined that the long-term predisposing causes of his distress were more likely to be due to his stated anxiety.


The patient was discharged on the third day of hospitalization, with a prescription of mirtazapine 7.5 mg nightly, showing no further evidence of being a danger to self or others. He continued to report extreme guilt over his actions but reported a moderate improvement of his anxiety symptoms while taking mirtazapine. At his 1-month follow-up visit, he reported feeling better overall with minimal symptoms of anxiety. The decision was made, in conjunction with his rheumatologist, to resume prednisone 5 mg daily, given his improved mental health, with close monitoring of symptoms. The final diagnosis was medication-induced anxiety disorder, with onset after medication use.


Corticosteroids have long been implicated in drug-induced psychiatric disorders and exacerbations of previous diagnoses, including manic or depressive states, anxiety, psychosis, cognitive impairment, and delirium (4, 5). The mechanisms underlying these psychiatric changes remain unknown; however, several mechanisms have been implicated, including dysregulation of native cortisol pathways and neuromodulatory effects on the hippocampus and amygdala (6). The onset and duration of this phenomenon have been challenging to capture, with reports of patients showing psychiatric symptoms both at the start and end of corticosteroid therapy (3). Nonetheless, it is known that the dose of prescribed steroids is positively associated with the incidence of psychiatric symptoms, which are more likely to occur at the beginning of treatment (7).

The management of corticosteroid-induced psychiatric symptoms may vary because of the wide presentation of symptoms, duration, and onset. The consensus is to discontinue the suspected medication once such symptoms are present. In situations where discontinuation is not possible, tapering the dose to as low as possible is advised (3, 4). Continual assessment of psychological well-being among patients taking corticosteroids is crucial because these symptoms can occur at any time during the treatment course.

Corticosteroid-induced anxiety disorders are a rare but known phenomenon, and clinicians should be aware of such disorders. Bolanos et al. (8) reported that 20% of patients receiving long-term low-dose corticosteroid therapy developed prednisone-induced mood disorders, and 5% developed prednisone-induced anxiety disorders. These findings were seen in a similar study by Bender and colleagues (9) that reported elevated levels of depression and anxiety among children receiving corticosteroid therapy. For patients with persistent affective syndromes who are receiving corticosteroid therapy long term, the recommendation is that the patient should be maintained at the lowest effective corticosteroid dose and treated with a selective serotonin reuptake inhibitor for managing depression and anxiety symptoms (3).


A challenge in the diagnosis of anxiety disorders is the wide range of neuropsychiatric and physical symptoms that can overlap with those of other disorders, such as major depressive disorder, schizophrenia, or bipolar disorder (10). Even among individuals with the same diagnosis, symptoms can vary widely because of high levels of comorbidity among anxiety disorders such as panic disorder. Anxiety disorders have also been shown to increase the risk for suicidality and homicidality, even when studies were controlled for comorbid psychiatric disorders typically associated with suicidal ideation (such as major depressive disorder) (11, 12). Therefore, clinicians must carefully assess the severity of a patient’s anxiety to determine whether inpatient hospitalization is warranted to prevent danger to the patient or others.


Anxiety may be exacerbated by the radical changes that often occur as a person transitions from one stage of life to the next. The transition from working life to retirement is often marred by the stressors of identity disruption, loss of societal engagement, changes in social relationships, and end-of-life anxiety (13). These substantial changes can predispose elderly patients to develop new anxiety symptoms or exacerbate preexisting anxiety disorders (14). This scenario was evident for the patient in this report, who attributed much of his anxiety to the significant changes in his life and his perceived inability to care for himself and his wife. However, it is likely that corticosteroid use was the primary factor contributing to his acute suicidality and homicidality because his behavior deviated significantly from what is typically observed in retirement.


This case highlights the importance of recognizing the diverse neuropsychiatric manifestations of corticosteroids, with a particular focus on anxiety. Although the patient was given a preliminary diagnosis of corticosteroid-induced psychosis because of his acute suicidal and homicidal ideation, a thorough reassessment and collection of collateral information led to a revised diagnosis of corticosteroid-induced anxiety disorder. Clinicians should be aware of the broad spectrum of psychiatric adverse effects associated with corticosteroid use and consider the full clinical picture, such as psychosocial stressors, to ensure an accurate diagnosis and treatment plan.


The authors thank Drs. Reza Farokhpay and Bavani Nadeswaran for their mentorship and editorial support.