Pharmacological Management of Steroid-Induced Psychosis: A Review of Patient Cases

Abstract

Objective: To review the efficacy and safety of medications used in the management of steroid-induced psychosis. Data Sources: A comprehensive literature search was conducted using PubMed, MEDLINE, ProQuest, and Scopus between May and October 2020 using the following search terminology: “steroid-induced psychosis” OR “corticosteroid-induced psychosis.” Study Selection and Data Extraction: Definitive cases, as defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, were included in this review. Geriatric patients >65 years of age, those with a confounding neurological condition such as a traumatic brain or spinal cord injury, or those with active malignancy were excluded. Data Synthesis: A total of 13 patient cases were included in this review, representing 8 male patients and 5 female patients. The mean age at symptom presentation was 42.5 years. Six patients presented with delusions, 5 presented with hallucinations, and 2 presented with both manifestations; 12 patients were managed with an antipsychotic, with haloperidol being the most commonly prescribed, followed by risperidone. One patient was managed with lithium and clonazepam alone. All patients returned to their psychological baseline upon the discontinuation or decreased dose of steroids in combination with Pharmacological intervention, though the time to resolution of symptoms varied significantly. No notable adverse drug events associated with treatments were reported. Conclusions: Steroid-induced psychosis is a serious adverse effect of corticosteroid therapy; however, management strategies that combine a dose reduction or elimination of steroids, in combination with an antipsychotic medication, are effective in resolving this syndrome.

Introduction

Corticosteroids are used to manage a host of medical conditions and are widely prescribed because of their ability to suppress inflammatory processes and the immune system.¹ Although initially brought to market under a narrow scope in 1949 for the management of rheumatoid arthritis, glucocorticoids have become a mainstay of current pharmacotherapy, with more than 25 million prescriptions for prednisone written for various indications in 2017 alone.^2,3 Although clinical benefit can be derived from these agents, it is also important for clinicians to be cognizant of associated adverse effects, which are numerous and include an increased risk of infection, leukocytosis, hyperglycemia, and psychiatric manifestations such as steroid-induced psychosis.^1,4

Steroid-induced psychosis is considered a type of substance-induced psychotic disorder and is defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5).⁵ Patients may experience either delusions or hallucinations shortly after beginning a course of steroids that cannot be better explained by another condition or substance and significantly obstructs the patient’s ability to function.⁵ Additional psychiatric symptoms may develop, such as mania, depression, and anxiety.⁶ In the most severe cases, patients may experience suicidal ideation, may commit suicide, or become outwardly aggressive and violent.^7-10 In an analysis of chronic objective pulmonary disease patients, Swinburn et al¹¹ noted that euphoria developed in patients taking prednisolone 30 mg/d.

Historically, patients suffering from steroid-induced psychosis have been managed through the tapering and eventual discontinuation of the offending corticosteroid, in addition to the use of modulating agents in severely obtunded individuals. Although first-generation antipsychotic medications, such as haloperidol, have been used extensively, evidence also exists to support second-generation antipsychotics, lithium, selective serotonin reuptake inhibitors, tricyclic antidepressants (TCAs), and select antiepileptic drugs, including carbamazepine as well as valproic acid and its derivatives.^7-10 Although the use of these agents is patient specific and driven largely by prescriber preference, there is limited information to support a formal management path for affected patients.

A definitive understanding of risk factors for steroid-induced psychosis remains elusive; however, the dose of the steroid has been implicated.⁷ The Boston Collaborative Drug Surveillance Program evaluated 718 hospitalized patients on prednisone and found that 4.6% of patients receiving doses greater than 40 mg/d of prednisone presented with psychiatric symptoms. The incidence rose to 18.4% for patients receiving more than 80 mg/d.¹² A review by Lewis and Smith¹³ found similar patterns of increasing incidence with higher doses of prednisone.

The mechanism behind steroid-induced psychosis remains unproven. It is known that the administration of exogenous corticosteroids can produce stress on the hypothalamic-pituitary-adrenal (HPA) axis, which relies on a negative feedback system to regulate the release of hormones such as cortisol.¹ This stress on the HPA axis may lead to deleterious effects on cognitive function caused by a change of metabolic needs that can shrink the hippocampus.^14-16 An increase of dopamine that may be attributed to the induction of tyrosine hydroxylase by corticosteroids, which has been observed in animal models, could lend itself to the presentation of psychosis as hypothesized in patients with schizophrenia.^9,14,15,17

The management of steroid-induced psychosis begins with the attempt to taper or discontinue the offending steroid, though this option may not be feasible for all medical indications. When symptomatology persists with a reduced dosage, or the clinician deems it necessary to continue the current course of steroids, additional Pharmacological interventions may be required to manage the psychosis. The objective of this review was to review the efficacy and safety of medications used in the management of steroid-induced psychosis from a collection of case reports in a small sample of patients.

Methods

A comprehensive literature search was performed on PubMed, MEDLINE, ProQuest, and Scopus using the search terms “steroid-induced psychosis” OR “corticosteroid-induced psychosis.” The search was conducted between May and October 2020. For the search performed on ProQuest, additional filters were activated after the initial search to include articles written in English and exclude newspapers, dissertations and theses, blogs, podcasts, websites, historical newspapers, trade journals, conference papers and proceedings, and magazines.

The inclusion criteria were case reports or series, observational studies, and controlled or uncontrolled clinical trials available as full text; a definitive diagnosis of steroid-induced psychosis aligning with DSM-5; and documentation of the agent(s) used for the treatment of the psychotic symptoms, steroid(s) that induced the psychotic symptoms, and patient outcomes related to the intervention.

The exclusion criteria were literature that did not discuss the pharmacotherapeutic intervention of steroid-induced psychosis and outcomes related to those efforts; cases in which the diagnosis of steroid-induced psychosis may have been confounded by another neurological condition; patients younger than 18 years old, older than 65 years old, pregnant, or with active malignancy; literature involving in vitro or in vivo animal models; full texts that were not available in English; and literature that was not retrievable. Other neurological conditions may have included, but were not limited to, those with a spinal cord or traumatic brain injury, poststroke patients, confusion assessment method for the intensive care unit positive status, or underlying delirium of any cause. Adults older than 65 years were excluded in an effort to control results against diagnosed or undiagnosed Alzheimer’s dementia or other progressive neurological deficits that have a higher prevalence in the elderly. Similarly, patients with active malignancy were excluded in an effort to control against the high doses of corticosteroids that frequently accompany chemotherapy interventions because these corticosteroid doses could have skewed results. A schematic of the search methodology can be found in Figure 1.

Figure 1.

Search methodology.

Results

The initial search resulted in 579 pieces of literature. After screening the results for possibly relevant literature and applying filters for ProQuest, a total of 60 results remained. Duplicates were removed and the inclusion and exclusion criteria were applied to the full-text articles, yielding 12 case reports and 1 case series that provided a total of 15 patients. Two patients from the case series were excluded because of incompleteness of information as outlined in the inclusion and exclusion criteria for this evaluation, thereby leaving a total of 13 patients for inclusion in this review. Results can be viewed in Table 1.

Table 1.

Results.

Authors, year	Study design	Patients who met criteria (sex/age in years)	Indication for steroids	Steroids (prednisone equivalent in average mg/d)4	Duration of steroid therapy	Onset of psychosis from first dose of steroids	Steroid management following diagnosis of psychosis	Pharmacotherapeutic intervention	Outcome
Ahmad and Rasul, 199920	Case report	1 (F/55)	Dermatological reaction	Prednisone (60)	5 Days followed by taper	10 Days	Discontinued	Haloperidol 5 mg IV followed by 2 mg po bid	Resolved
Fischer and Kim, 201921	Case report	1 (M/49)	Chronic low back pain	Dexamethasone (66.67)	Single injection	3 Days	N/A	Lorazepam, quetiapine	Resolved
Hong et al, 200622	Case report	1 (F/48)	Sheehan syndrome	Prednisolone (15)	5 Days	<24 Hours	Tapered	Haloperidol 5 mg, lorazepam 4 mg	Resolved
Jacob et al, 200223	Case report	1 (M/26)	Crohn disease	Hydrocortisone (100), prednisolone (30)	12 Days	10 Days	Tapered	Haloperidol 2.5 mg tid, lorazepam 2 mg PRN	Resolved
Koh et al, 200224	Case report	1 (M/40)	Asthma	Methylprednisolone (625), prednisolone (40)	10 Days	10 Days	Discontinued	Lorazepam, risperidone	Resolved
López-Medrano et al, 200225	Case series	1 (F/20)	Lupus nephritis	Prednisone (60)	3 Days	3 Days	Tapered	Clonazepam 0.5 mg po bid, risperidone 2 mg po bid	Resolved
Marian et al, 201026	Case report	1 (F/24)	Systemic lupus erythematosus	Prednisolone (60)	A few days	3 Days	Tapered	Risperidone 2 mg	Resolved
Milanlıoğlu and Güleç, 201127	Case report	1 (M/46)	Optic neuritis	Methylprednisolone (1250)	2 Days	2 Days	Discontinued	Risperidone 2 mg	Resolved
Mulky and Debbarma, 201828	Case report	1 (M/38)	Muscle growth	Dexamethasone (40)	6 Months	6 Months	Discontinued	Diazepam 5 mg, olanzapine 10 mg IV	Resolved
Olson and Lewis, 201729	Case report	1 (F/46)	Chronic pancreatitis	Triamcinolone (100)	Once every 4 months for 5 sessions	After fifth session	Discontinued	Clonazepam, lithium	Resolved
Sato et al, 199818	Case report	1 (M/54)	Pemphigus vulgaris	Betamethasone (23.33)	2 Years	2 Years	Discontinued	Flunitrazepam, haloperidol, levomepromazine maleate	Improved
Silva and Tolstunov, 199519	Case report	1 (M/45)	Facial reconstruction	Dexamethasone (100)	4 Days	4 Days	N/A	Haloperidol 1 mg IM every hour of sleep and PRN confusion; perphenazine initially 4 mg po every 2 hours PRN anxiety, then increased to 16 mg po qid	Resolved
Ward and George, 201630	Case report	1 (M/62)	Rheumatoid arthritis exacerbation	Methylprednisolone (100)	Had been on prednisolone for 40 years	Within 1 week of methyl prednisolone injection for exacerbation	Tapered	Valproic acid, zuclopenthixol depot injection	Resolved

Abbreviations: F, female; IM, intramuscular; IV, intravenously; M, male; N/A, not applicable; PRN, as needed.

Five patients (38.46%) were female; the mean age was 42.54 years; and reasons for starting corticosteroid treatment varied significantly. Psychotic symptoms of delusion occurred in 6 patients (46.15%), hallucinations occurred in 5 (38.46%), and 2 presented with both delusions and hallucinations (15.38%). Other reported psychiatric symptoms were insomnia, agitation, irritability, combativeness, confusion, cognitive impairment, mania, depression, and suicidal ideation. Four patients were reported to have had a history of psychiatric conditions, including anxiety (n = 2) and depression (n = 2). A total of 11 patients received greater than 40 mg of prednisone or its equivalent per day, whereas 6 patients (46.15%) eventually eclipsed 80 mg of prednisone or its equivalent.

The onset of symptoms after the administration of steroids was not reported consistently, but the earliest occurred less than 24 hours after the first dose and the latest occurred after 2 years of chronic steroid administration. The rest of the patients’ symptoms occurred within 2 days to up to 2 weeks.

Treatment modalities also varied widely. The typical antipsychotics that were used in some cases include haloperidol (n = 5), levomepromazine (n = 1), perphenazine (n = 1), and zuclopenthixol (n = 1). The atypical antipsychotics used were risperidone (n = 4), olanzapine (n = 1), and quetiapine (n = 1). Lithium (n = 1) and valproic acid (n = 1) were also used for mood stabilization, whereas benzodiazepines such as clonazepam, diazepam, lorazepam, and flunitrazepam were largely reserved to manage insomnia or anxiety.

All patients evaluated in this review demonstrated sufficient improvement to the point that they could be discharged from the institution or setting in which they were managed in stable mental and physical condition or completely recovered after the management of their steroid-induced psychosis. Onset of relief from steroid-induced psychosis varied, with improvement happening as soon as within 24 hours. Typical recovery occurred within a month, with the longest reported time being 8 weeks. Patients who were able to discontinue their steroid, as opposed to tapering or continuing their steroid, seemingly experienced a quicker resolution of symptoms.

Discussion

The few patient encounters that met inclusion criteria limit the generalizability of the findings of this review; however, it is abundantly clear from this evaluation that steroid-induced psychosis often presents with adjunctive symptomatology beyond the traditional delusions and hallucinations previously mentioned.^18-30 In addition to psychosis, most of the patients included in the review developed other psychiatric symptoms, a finding consistent with other reports in the literature discussing the prevalence of psychiatric symptoms during corticosteroid therapy.^13,31 Although it is not within the scope of this review to quantify definitive risk factors, it is important to note that many of the patients in this review did not have a history of psychiatric illness. Regardless, the significance of having a psychiatric disorder does not appear to be linked to the likelihood of developing steroid-induced psychosis.^13,32 Although the majority of the patients in this review were male, evidence suggests that biological female sex is a risk factor for developing steroid-induced psychiatric symptoms, even when controlling for female-predominated disease states such as systemic lupus erythematosus.^13,33

Although variable in the small sample size included in this review, the onset of symptoms aligned with the notion that steroid-induced psychosis is more prevalent immediately following steroid initiation. Previous literature has demonstrated the mean time to symptom onset to be approximately 11.5 days.¹³ Observations about the relationship between doses and risks of developing steroid-induced psychosis are congruent with previously published literature, though it is imperative to recognize that the dose of steroid has not been related to the type of presentation, duration, or onset.¹³ Finally, all the patients who were reviewed either improved or had complete resolution of their psychotic conditions, which is likewise synonymous with previous literature.¹³

Regarding the pharmacotherapeutic interventions for psychosis in the cases reviewed, the most used agents were antipsychotics, particularly the first-generation drug haloperidol and second-generation drug risperidone. These medications, in addition to others in their respective classes, mediate the action of dopamine, specifically at the D₂ receptor in the central nervous system.³⁴ The improvement in symptoms demonstrated by these and similar agents correlates with the proposed mechanism of how corticosteroids may induce psychosis.³⁴

Second-generation antipsychotics are commonly favored for their relatively safer adverse effect profile when compared with first-generation agents. Although certainly not benign, first-generation antipsychotics are known to cause cardiac dysrhythmias via QTc prolongation and extrapyramidal effects, whereas the newer second-generation agents are more commonly associated with metabolic alterations such as hyperglycemia and hypertriglyceridemia.³⁴ The second-generation antipsychotics evaluated in this review were risperidone, olanzapine, and quetiapine. Risperidone and olanzapine were found to be more effective typical antipsychotics for treating schizophrenia; quetiapine’s effectiveness was not shown to be significantly different.^35,36 For the treatment of positive symptoms in schizophrenia that are also present in steroid-induced psychosis, such as hallucinations and delusions, risperidone and olanzapine were more effective compared with other antipsychotics.³⁶ The favoritism shown toward using atypical antipsychotics over the typical antipsychotics is possibly attributed to its reduced risk of dystonia, tardive dyskinesia, and sedation.^6,34

The typical antipsychotics evaluated in this review were haloperidol, perphenazine, levomepromazine, and zuclopenthixol; the latter two are not available in the United States. Haloperidol’s use to treat steroid-induced psychiatric symptoms is well documented.^13,37-39 Haloperidol is also known to have significantly higher effectiveness compared with other antipsychotics for controlling the positive symptoms of schizophrenia.³⁶ Despite a proven record of success, haloperidol should not be universally or arbitrarily prescribed in the management of steroid-induced psychosis, because of its known adverse event profile, as documented in the case reported by Silva and Tolstunov.¹⁹

The paucity of data available regarding definitive management strategies for steroid psychosis have led to conflicting data, such as with the following case. A 45-year-old male patient was initiated on 2 antipsychotics to treat his psychosis: haloperidol and perphenazine.¹⁹ According to the patient’s medication history, he was already taking the TCA amitriptyline for depression and lorazepam for anxiety. The patient’s symptoms persisted and worsened even after initiating both antipsychotics and discontinuing amitriptyline. Eventually, the patient recovered and was discharged 28 days later after increasing the dose of perphenazine. Whether his home amitriptyline interfered with the capabilities of haloperidol in the timely control of his symptoms is unclear; however, evidence from previous studies suggest that TCA agents can cause worsening agitation and anxiety in patients with psychiatric disorders induced by steroids and should be avoided in these instances.⁴⁰ Interestingly, TCA agents may be potentially beneficial in the treatment of symptoms arising from corticosteroid withdrawal.⁴¹

Despite establishing the effectiveness of atypical antipsychotics for steroid-induced psychosis, the only case included in this review that did not utilize an antipsychotic involved a female patient with chronic pancreatitis who underwent several endoscopic ultrasound-guided celiac plexus blockade sessions using triamcinolone 80-mg injections for severe pain every 4 months.²⁹ She experienced delusions with mania and severe depression on the fifth session. It is likely that the clinicians deemed that using clonazepam and lithium for her mood disturbances would be adequate to control her anxiety, insomnia, and mood disturbances; however, the case does not provide an explanation for why an antipsychotic was not initiated. Nonetheless, the patient completely recovered after several weeks and represents a limited look into the management of steroid-induced psychosis without using antipsychotics.

Limitations of this evaluation include the small and varied sample size that was derived from case reports and a single case series. This, combined with the variances in describing treatment methods within the reports, makes definitive results elusive. Furthermore, steroid-induced psychosis may be underrepresented in the literature because it may be underdiagnosed and mistaken or masked by other conditions or substances. As with any literature review, there exists an inherent possibility of inadvertently missing primary literature that would have been suited for inclusion. The authors attempted to combat the low level of evidence in this report by utilizing robust inclusion and exclusion criteria. The exclusion of certain patient populations, including those older than 65 years and those with active malignancy, was done so as to limit any potential confounding variables. Limiting the inclusion criteria to the definitive diagnosis of steroid-induced psychosis aligning with DSM-5 criteria proved to be essential; the use of “steroid psychosis” or “steroid-induced psychosis” in some texts presented cases of patients exhibiting mood disturbances without overt psychosis.

Conclusions

Steroid-induced psychosis is a severe adverse effect that can occur shortly after administering high doses of glucocorticoids. Although steroid-induced psychosis can typically be managed by tapering and discontinuing the offending steroid, patients may still experience persistent mood disturbances and psychosis. There is evidence from a small number of patient cases and studies to support the effectiveness of antipsychotics such as haloperidol and risperidone in adult patients experiencing delusions or hallucinations after administering corticosteroids. Fortunately, the great majority of patients completely recover from their acute psychosis within an average span of 2 weeks after treatment initiation. Clinicians must be judicious in the prescribing of corticosteroids in an effort to minimize the risk of this preventable adverse drug event.