When Agitation Meets Intoxication: A Retrospective Comparison of Intramuscular Olanzapine and Haloperidol Among Patients With and Without Substance Use in in a Psychiatric Emergency Service

Abstract

Background:

Substance use is a major contributor to psychiatric emergencies, often complicating the management of acute agitation. Despite the high prevalence of co-occurring substance use in emergency psychiatric settings, evidence guiding pharmacologic interventions remains limited. This study examined whether active substance use, as measured by urine drug screening (UDS), predicted clinical and safety outcomes among patients requiring intramuscular (IM) medication for agitation in an emergency psychiatric setting.

Methods:

This retrospective chart review included 103 patients presenting to a Comprehensive Psychiatric Emergency Program (CPEP) between January 2020 and November 2022 who received IM haloperidol lactate (HL) or IM olanzapine (OL) and completed UDS. Patients with positive (UDS+) screens were compared to those with negative (UDS−) screens on demographic, clinical, and medication-related outcomes. Primary medication outcomes were need for repeat IM medication administration and time to next administration. Secondary measures included physiologic changes, adverse effects, and restraint/seclusion use.

Results:

Fifty-eight patients (56.3%) had a positive UDS. The UDS+ group was younger and more frequently had incarceration history, but did not differ from the UDS− group in terms of diagnoses, hospitalization history, or adverse events. UDS+ patients had a significantly longer time to next medication administration (mean 10.0 vs 6.5 hours, p=0.050). There were no differences in terms of repeat IM use, objective measures of agitation, or safety outcomes. Subanalysis comparing HL and OL among UDS+ patients was limited by sample size, but showed comparable efficacy and tolerability.

Conclusions:

Patients with active substance use exhibited similar or greater response to standard IM treatment for agitation without increased safety concerns. Findings support potential safety and efficacy of common pharmacologic approaches in this high-risk population, though a modest single-site cohort may limit generalizability. There is need for prospective, multisite replication.

Introduction:

Utilization rates of emergency psychiatric care in the United States have significantly increased over the past several decades, including visits for substance use disorders (SUDs).^1,2 One review found that substance use disorders, including alcohol, accounted for an annual average of 9.4% of all emergency department (ED) visits and 11.9% of hospitalizations, with rates significantly increasing between 2014 and 2018.³

Patients requiring emergent psychiatric care often present with acute agitation, necessitating rapid evaluation and management.⁴ Co-occurring substance use may increase agitation risk, which can cause excessive bottom-up activation, related to subcortical arousal systems, and insufficient top-down inhibition, known as prefrontal control.^5–9 For example, stimulants such as methamphetamine have been associated with decreased serotonin transporter density in the orbital frontal cortex and the anterior cingulate cortex, correlating with increased aggression.^9,10 Similarly, substances acting at the cannabinoid 1 receptor, such as cannabis or synthetic cannabinoids, can result in excessive glutamatergic and dopaminergic activity, leading to psychotic symptoms and agitation.^9,11

Despite the high frequency of agitation/violence in emergency settings, limited data exist on appropriate management. In 2012, the American Association of Emergency Psychiatry released guidelines for managing agitation, titled Project BETA (Best-practices in the Evaluation and Treatment of Agitation).¹² These guidelines highlighted the importance of non-pharmacologic behavioral interventions when possible. When necessary, the guidelines recommended pharmacologic interventions, including benzodiazepines and dopamine receptor-blocking agents (DRBAs), preferring oral medications if possible, and short-acting intramuscular (IM) administrations reserved as a last resort.¹²

While there is specific guidance in some circumstances depending on the etiology of agitation (e.g., delirium, intoxication, etc.), there is relatively little guidance regarding the pharmacologic management of agitation in the context of co-occurring substance use. For example, in patients acutely intoxicated with stimulants, benzodiazepines have been recommended as monotherapy, whereas coadministration of a DRBA is recommended if psychosis is also present.^12,13 In acute alcohol intoxication, haloperidol specifically has been recommended, whereas in alcohol withdrawal, benzodiazepine monotherapy is preferred.^12,13 Otherwise, specific guidance is limited for patients with substance use or for special populations, such as pediatric or peripartum patients.

Aims and Hypotheses:

In this retrospective, real-world, study of patients presenting for emergency psychiatric care, and who required emergent pharmacologic intervention for agitation, the authors hypothesized that, when comparing patients with and without co-occurring substance use, patients with substance use would be 1) more likely to require repeat medication administration, and 2) have shorter intervals between medication administrations, due to potentially increased levels of agitation. Additionally, the authors hypothesized that patients with co-occurring substance use would be more susceptible to adverse effects, such as oversedation and hypotension, due to potential interactions between substances used and medications prescribed. As a secondary aim, the authors aimed to determine if IM olanzapine is superior to IM haloperidol in these outcome measures, given its faster onset and potential superior efficacy demonstrated in prior studies.^14–16

Methods:

Participants:

This retrospective chart review was approved by the Institutional Review Board at Icahn School of Medicine at Mount Sinai (Protocol #22-01627). The sample included all adult patients who presented to the Comprehensive Psychiatric Emergency Program (CPEP) at Mount Sinai Beth Israel between 1/1/2020 and 11/30/2022, and who received either IM haloperidol lactate (HL) or IM olanzapine (OL) for the acute management of agitation. All patients who present to the CPEP are requested to complete urine drug screening (UDS), although some decline. The UDS at this program utilizes an immunoassay, and can detect: amphetamines, cocaine, opiates, phencyclidine, methadone and methadone metabolites, benzodiazepines, barbiturates, cannabinoids, fentanyl, tramadol, oxycodone, buprenorphine, and 6-acetylmorphine. Only those with completed UDS were included in this analysis. Typical agitation protocols at this institution utilize IM HL (2.5–5mg) or OL (5–10mg). Patients who received other IM medications for agitation (IM chlorpromazine, IM fluphenazine) were excluded due to the small number of cases (fewer than 10).

Source of Data:

Data were extracted from the electronic medical records by members of the study team (SL, MM, VP, CE). These included sociodemographic information (age, race, gender identity) and clinical history (prior hospitalizations, previous CPEP visits, history of violence, and incarceration history). Information from the index encounter was also collected, including mode of arrival to the CPEP (self, emergency medical services (EMS), or New York Police Department (NYPD)), reason for CPEP presentation, and primary diagnosis of the encounter. Medication-related data including which DRBA was administered, and any co-administered medications was also collected.

The primary measures of efficacy utilized in this study were (1) any administration of additional IM medications within 24 hours and (2) time to next medication administration. Secondary measures of efficacy included need for restraints or seclusion during the encounter, the change in Brøset violence checklist score, and the length of stay (LOS). The Brøset violence checklist is a clinician administered brief scale including 6 items on a Likert-style scale, with a maximum score of 6, with higher scores corresponding with increased levels of agitation and violence.¹⁷ Measures of safety and/or tolerability included changes in heart rate, changes in mean arterial pressure (MAP), and any documented adverse reaction to medication. Interrater reliability on a subset of included patients was high, over 95%. Any item for which there was disagreement or uncertainty was reviewed by the entire study team to reach consensus.

Data Analysis:

First, patients with positive urine drug screenings (UDS+) were compared to those with negative screenings (UDS−). Patients were classified as UDS+ if any substance use was detected; however, if it were for a medication prescribed to them (i.e., testing positive for buprenorphine but being prescribed buprenorphine-naloxone by a provider), they were classified as UDS−, as this would not reflect active illicit substance use. Subsequently, among UDS+ patients specifically, IM HL was compared to IM OL as an exploratory analysis. Statistical analyses were conducted with SPSS software Version 30.0.¹⁸ Descriptive statistics (mean, standard deviation, frequency) were used to characterize the study sample. Chi-square tests were used to compare groups on categorical variables, and independent t-tests were used for continuous variables. Effect sizes were calculated for comparative statistics that were statistically significant, using Cohen’s d (d) for independent samples t-tests and Cramer’s V (V) for chi-square tests. Kaplan-Meier survival analysis was used to examine differences in time to next IM medication administration.

Results:

Study Sample Characteristics:

A total of 247 patients received IM HL or OL during the study period. Of these, 103 completed UDS and were included in the following analyses. The study sample was approximately 37 years in age, over half male, with a wide racial distribution. The sample had on average 2.6 prior psychiatric hospitalizations, and 0.7 CPEP visits (not leading to admission) within the past year. About a third had a history of violence, while about 15% had a history of incarceration (Table 1).

Table 1. Sociodemographic and Clinical Differences Between UDS+ and UDS−.

Variable	Total Sample (N=103) (N, %) or (Mean, SD)	Positive UDS (N=58) (N, %) or (Mean, SD)	Negative UDS (N=45) (N, %) or (Mean, SD)	Statistics (test statistic, degrees of freedom, significance, effect size)
Sociodemographic Characteristics
Male gender	60 (58.3%)	37 (63.8%)	23 (51.1%)	χ2=1.676 df=1 p=0.195
Age (in years)	37.2 (13.7)	34.1 (11.4)	41.1 (15.5)	t=2.656 df=101 p=0.012 d=0.528
Race				χ2=10.188 df=4 p=0.037 V=0.315
Asian or South Asian	8 (7.8%)	3 (5.2%)	5 (11.1%)
Black	35 (34.0%)	20 (34.5%)	15 (33.3%)
Latinx	22 (21.4%)	11 (19.0%)	11 (24.4%)
White	24 (23.3%)	11 (19.0%)	13 (28.9%)
Other or Unknown	14 (13.5%)	13 (22.4%)	1 (2.2%)
Clinical History
Number of prior psychiatric hospitalizations	2.6 (3.2)	2.6 (3.5)	2.6 (2.8)	t=0.111 df=101 p=0.912
Number of prior CPEP visits (not leading to admission) within past year	0.7 (1.8)	0.8 (1.9)	0.7 (1.5)	t=−0.359 df=101 p=0.720
History of violence	34 (33.0%)	21 (36.2%)	13 (28.9%)	χ2=0.614 df=1 p=0.433
History of incarceration	16 (15.5%)	13 (22.4%)	3 (6.7%)	χ2=4.789 df=1 p=0.029 V=0.216

Over half of the patients presented via EMS, approximately one-quarter via NYPD, and the remainder self-presented. The most common reasons for presentation were agitation (36.9%) and psychosis (35.0%), followed by suicidality (10.7%), mania (8.7%), and “other” reasons (8.7%). Over half of the patients were diagnosed with a psychotic disorder, with roughly equal splits between bipolar and unipolar mood disorders. Three quarters of the patients received IM HL, with the remainder received IM OL. Over half the patients were co-administered lorazepam, about a quarter were co-administered midazolam, and about a third were co-administered diphenhydramine (Table 2).

Table 2. Differences in Clinical Presentation and Outcomes Between UDS+ and UDS−.

Variable	Total Sample (N=103)	Positive UDS (N=58)	Negative UDS (N=45)	Statistics
Current Clinical Encounter
Modality of presentation to CPEP				χ2=3.979 df=2 p=0.137
EMS	57 (55.3%)	29 (50.0%)	28 (62.2%)
NYPD	26 (25.2%)	19 (32.8%)	7 (15.6%)
Self	20 (19.4%)	10 (17.2%)	10 (22.2%)
Reason for presentation				χ2=8.020 df=4 p=0.091
Agitation	38 (36.9%)	22 (37.9%)	16 (35.6%)
Mania	9 (8.7%)	3 (5.2%)	6 (13.3%)
Psychosis	36 (35.0%)	19 (32.8%)	17 (37.8%)
Suicidality	11 (10.7%)	10 (17.2%)	1 (2.2%)
Other	9 (8.7%)	4 (6.9%)	5 (11.1%)
Billing diagnosis				χ2=2.263 df=3 p=0.520
Bipolar spectrum illness	22 (21.4%)	12 (20.7%)	10 (22.2%)
Unipolar mood disorder (depression, adjustment)	24 (23.3%)	16 (27.6%)	8 (17.8%)
Psychotic disorder	53 (51.5%)	27 (46.6%)	26 (57.8%)
Other	4 (3.9%)	3 (5.2%)	1 (2.2%)
Medications Administered
Dopamine receptor-blocking agent used				χ2=1.982 df=1 p=0.159
Haloperidol	80 (77.7%)	48 (82.8%)	32 (71.1%)
Olanzapine	23 (22.3%)	10 (17.2%)	13 (28.9%)
Co-administered with lorazepam	57 (55.3%)	31 (53.4%)	26 (57.8%)	χ2=0.192 df=1 p=0.661
Co-administered with midazolam	24 (23.3%)	17 (29.3%)	7 (15.6%)	χ2=2.683 df=1 p=0.101
Co-administered with diphenhydramine	36 (35.0%)	20 (34.5%)	16 (35.6%)	χ2=0.013 df=1 p=0.910
Clinical Outcomes
Restraints or Seclusion within 24 Hours in CPEP	9 (8.7%)	4 (6.9%)	5 (11.1%)	χ2=0.564 df=1 p=0.452
Change in Brøset Violence Score (N=31)	−2.1 (1.9)	−2.1 (2.0)	−2.0 (2.0)	t=0.154 df=29 p=0.879
Required additional intramuscular medication	35 (34.0%)	21 (36.2%)	14 (31.1%)	χ2=0.293 df=1 p=0.588
Number of additional administrations within 24 hours	0.6 (1.2)	0.6 (1.1)	0.6 (1.2)	t=0.081 df=101 p=0.935
Length of stay in CPEP (in hours)	13.5 (9.3)	13.7 (9.7)	13.1 (10.0)	t=−0.326 df=101 p=0.745
Admitted to inpatient psychiatry	74 (71.8%)	39 (67.2%)	35 (77.8%)	χ2=1.391 df=1 p=0.238
Safety Outcomes
Documented adverse reaction to medication	2 (1.9%)	1 (1.7%)	1 (2.2%)	χ2=0.033 df=1 p=0.856
Change in mean arterial pressure (in mmHg) (N=95)	−12.4 (14.7)	−13.1 (15.4)	−11.7 (14.0)	t=0.462 df=93 p=0.645
Change in heart rate (in bpm) (N=95)	−7.9 (16.4)	−10.0 (16.6)	−5.1 (15.7)	t=1.469 df=93 p=0.145

Fewer than 10% of patients required restraints or seclusion in CPEP within 24 hours of medication administration. About a third required at least one additional administration of IM medications, with an average number of repeat administrations of 0.6. Time to next medication administration was on average 8.6 hours. Only 2 patients had any adverse reaction. About three quarters of patients were admitted to inpatient psychiatry following their CPEP stay (Table 2).

Comparison by Urine Drug Screen Results:

Of the 103 patients included, 58 (56.3%) had a positive UDS. The most commonly detected substance was cannabis (79.3%), followed by amphetamines (24.1%), cocaine (17.2%), fentanyl (13.8%), phencyclidine (PCP) (5.2%), benzodiazepines (3.4%), and other opioids (3.4%). Thirty-seven patients (63.8%) were positive for only one substance, predominantly cannabis (32 patients) while 21 individuals (36.2%) tested positive for one or more.

Patients with a positive UDS (UDS+) were on average about 7 years younger than those with a negative UDS (UDS−). There were significant racial differences, with the UDS+ group having a higher proportion of “Other or Unknown” patients, and the UDS− group having a higher proportion of Asian or South Asian patients and White patients, with a moderate effect size (Table 1). No significant difference regarding gender was found.

The UDS+ group had a higher proportion of patients with history of incarceration, though the effect size was small. There were no statistical differences in prior psychiatric hospitalizations, prior emergency psychiatric visits, or history of violence (Table 1).

No differences emerged between UDS groups in encounter-level clinical characteristics, including mode of arrival, encounter diagnosis, or co-administered medications. The UDS+ group showed a non-significant trend toward higher rates of suicidality, whereas mania and psychosis were more common in the UDS− group (Table 2).

Regarding clinical outcomes, there were no significant differences between groups in need for repeat medication administration, with about a third of each group receiving at least one additional administration. Change in objective measures of violence were similar. There were no differences in length of CPEP stay, nor in rates of restraints or seclusion. There were similarly low rates of adverse reactions to medication, with only two patients having any documented negative effect, one from each UDS group. Both patients had received HL co-administered with a benzodiazepine, with one patient experiencing acute dystonia that resolved with benztropine administration, and the other patient having a fall. Changes in both mean arterial pressure and heart rate were comparable between groups (Table 2).

The UDS+ group demonstrated a longer median time to next medication administration compared to the UDS− group (8.5 hours, 95% CI 4.0–13.0 vs 5.5 hours, 95% CI 2.8–8.2). Mean time to next medication administration was also longer in the UDS+ group (10.0 hours, 95% CI 7.1–12.9) than in the UDS− group (6.5 hours, 95% CI 3.9–9.2). Kaplan-Meier survival analysis stratified by UDS status using the log-rank test demonstrated a statistically significant difference between groups (χ²=3.838, df=1, p=0.050). (Figure 1).

Figure 1. Kaplan-Meier Analysis of Time to Next Medication Administration Comparing UDS+ to UDS−.

Log Rank test: χ²=3.838, df=1, p=0.050

Olanzapine vs. Haloperidol lactate in patients with positive UDS:

Amont the 58 UDS+ patients, 10 received IM OL and 48 received IM HL. The average dose of OL was 8.5mg (SD 2.4mg) while the average of HL was 4.9mg (SD 0.4mg). No differences were found in terms of need for repeat medication administration (p=1.000), change in Brøset Violence Score (p=0.661), or restraint/seclusion utilization (p=1.000). Similarly, no differences in terms of adverse reactions (p=1.000), change in MAP (p=0.344), or change in HR (p=0.482) were found. The OL group appeared to have a slightly shorter mean time to next medication administration than the haloperidol lactate group (6.8 hours vs 10.8 hours); however, this did not reach statistical significance using Kaplan-Meier analysis (Log-rank test: χ²=2.414, df=1, p=0.120).

Excluded Patients:

Sociodemographic and clinical characteristics of the 144 patients without UDS testing are summarized in Supplemental Table 1, alongside the 103 included patients for reference. One patient was excluded due to missing outcome data. Patients without UDS results were more frequently male. They had fewer prior psychiatric hospitalizations, but a greater number of prior CPEP visits within the past year, and were more likely to present with NYPD and to present for agitation specifically. They were more likely to have a primary unipolar diagnosis (such as adjustment disorder). They had lower rates of repeat medication administration, a shorter CPEP LOS, and were less frequently admitted to inpatient psychiatry (Supplemental Table 1).

Discussion:

This real-world study examined 103 patients in a busy metropolitan CPEP who completed UDS and aimed to characterize if UDS predicted clinical outcomes. The sample overall reflected significant psychiatric morbidity, with high rates of prior hospitalizations and CPEP visits, and agitation or psychosis as common presenting complaints. Patients with positive UDS were younger, more likely to have a history of incarceration, and demonstrated longer intervals between IM medication administrations compared to those with negative UDS. Differences between OL and HL among the UDS+ were small, though interpretation was limited by sample size. Findings overall were consistent with existing recommendations from Project BETA.

Comparison by Urine Drug Screen Results:

Over half of patients in this sample who completed UDS were positive for one or more substances, which is higher than the typically reported rates of 30% among patients seeking emergency psychiatric services.^19–21 This may indicate that patients with co-occurring substance use comprise a disproportionately larger share of those who receive emergency medication for agitation, consistent with prior findings linking substance use to increased agitation risk.^7,8 The elevated proportion may also reflect sampling bias, as providers may have been more likely to encourage patients to complete UDS when substance use was suspected.

The UDS+ group was significantly younger than the UDS− group, consistent with reports of higher substance use rates among younger adults.^22–24 This also may suggest an earlier onset or a greater illness severity among patients with co-occurring substance use. Fewer Asian/South Asian patients were present in the UDS+ group, aligning with literature indicating lower SUD prevalence among Asian Americans, though variability across subgroups and nativity is generally substantial and was not able to be accounted for in this study.^22,25,26 Given the small number of Asian-identifying patients in this sample (n=8), this finding should be interpreted cautiously.

The two groups did not differ in prior hospitalizations, encounter diagnosis, or reasons for presentation. This is surprising, as one might expect substance-related presentations to be associated with fewer underlying psychiatric diagnoses. Combined with the younger age of the UDS+ group, this could indicate earlier progression to severe psychiatric illness among those with co-occurring substance use. The higher rates of incarceration among UDS+ patients is consistent with prior research.^27,28

While there were no differences in the rates of repeat IM medication administration, there were significant differences in time to next administration. The UDS+ group had significantly longer intervals between medication administrations compared to UDS−. This finding was somewhat unexpected, as co-occurring substance use was anticipated to increase agitation or tolerance to sedating medications, leading to higher or more frequent dosing needs. One possible explanation is that the IM medications may have had additive or synergistic effects with substances used prior to CPEP presentation. Another possibility is that the UDS+ group may have included more patients presenting with suicidality rather than mania or marked agitation, suggesting more frequent clinical profiles that may require less intensive or repeated medication intervention.

Although these differences in presenting complaint did not reach statistical significance, subtle variations in symptom distribution may have contributed to this observed difference in time between administrations. It is unlikely that this finding is entirely attributable to rapid substance clearance and discharge, as billing diagnoses and prior psychiatric hospitalizations were nearly identical between groups; over two thirds of patients in both groups had underlying serious mental illness, suggesting the UDS+ group represented individuals with both active psychiatric symptoms and substance use, rather than substance use alone.

No group differences emerged in secondary measures, such as Brøset score or use of seclusion or restraints, consistent with lack of difference in rates of repeat IM administration. There was also no difference in length of stay in CPEP nor in rates of psychiatric admission. This is somewhat inconsistent with prior reports, which have shown that patients presenting with substance use may be more likely to have longer CPEP stays, but reduced psychiatric admissions, after allowing time for substances to metabolize.²⁹

Safety rates were similar as well, with relatively few adverse effects across groups. This is somewhat surprising, as patients with substances in their system may be more susceptible to adverse effects of medication administration due to drug-drug interactions and oversedation. This finding provides additional reassurance regarding the safety of guideline-concordant antipsychotic use in individuals with co-occurring substance use, a population in whom clinicians may be particularly concerned about oversedation or drug-drug interactions.

Olanzapine vs. Haloperidol lactate in patients with positive UDS:

Haloperidol was more frequently used in this sample, which, while consistent with prior literature, may not have a strong evidence base in terms of superior efficacy.^14–16 Based on prior reports, OL would have been expected to have greater efficacy than HL.^12,13,30,31 However, on examination of rates of repeat medication administration, this was not found in this study, as rates were similar between groups.

OL demonstrated a shorter interval between administrations, which was also unexpected given the pharmacokinetic differences between the two agents. OL, when administered intramuscularly, demonstrates rapid absorption, with peak plasma concentrations occurring within 15–45 minutes, and a half-life on average of 30 hours^32,33, while HL’s onset is more frequently 30–60 minutes after administration, and has a half-life of about 21 hours^33–35. In this sample, the number of patients who received additional medications was small and likely insufficient to power analysis of differences in time between administrations.

Overall, these results are insufficiently powered to make clear recommendations on specific agents to use in psychiatric emergencies among patients with co-occurring substance use, but are suggestive that both HL and OL may be safe and effective.

Strengths & Limitations:

This study possesses several notable strengths and limitations. It represents one of the few real-world examinations of pharmacologic agitation management among patients with and without active substance use in an emergency psychiatric setting, which is an area with limited empirical guidance despite its substantial clinical significance. Multiple objective outcome measures were incorporated, including time-based indicators of medication efficacy and physiologic parameters, and the sample was both demographically and diagnostically diverse, increasing generalizability.

Several limitations warrant consideration, including a modest sample size, particularly within medication subgroups and among patients who received additional IM medications after the index administration. Statistical power was limited to detect smaller effects or to examine outcome differences between DRBAs or specific substances of use, and as such, non-significant group difference should not be interpreted as evidence of equivalence. As a single-site retrospective chart review, the study is subject to potential information and selection biases, and provider-level factors influencing medication choice could not be fully accounted for, introducing potential non-random drug selection leading to confounding by indication. This is particularly important as less than half of all patients with agitation had any UDS results, which could introduce non-random selection bias, as UDS testing may have been preferentially obtained in patients perceived to be at higher risk for substance-related agitation.

Consistent with this concern, patients without UDS testing differed meaningfully from included patients, demonstrating higher rates of law enforcement presentation and agitation, fewer prior psychiatric hospitalizations but more frequent recent CPEP visits, lower rates of repeat intramuscular medication administration, shorter CPEP length of stay, and lower rates of inpatient psychiatric admission. These differences suggest that UDS testing was not applied uniformly and that excluded patients may represent a clinically distinct subset of agitated presentations, which may limit the generalizability of findings to all patients receiving emergency agitation management.

The urine drug screen did not include certain substances commonly associated with agitation, such as alcohol or synthetic cannabinoids. Additionally, UDS can suggest recent substance use, but does not indicate current intoxication or withdrawal status, nor timing of use. UDS was coded as a binary variable in this study, and analyses of individual substances were not performed, preventing examination of differences across substance classes. In a busy emergency setting, mild or moderate adverse effects may be underrecognized or under-documented. In addition, the use of a convenience sample precluded formal power analysis, and multiple comparisons may have increased the risk of Type I error; however, the inclusion of effect sizes and covariate-adjusted regression models provides context for interpretation.

Despite these limitations, this study contributes novel data on an underexamined clinical population and supports the possible safety and apparent efficacy of widely used intramuscular antipsychotic agents among patients with co-occurring substance use. Results of this study complement guidance from Project BETA, and these findings offer a foundation for future prospective, adequately powered, multicenter investigations aimed at optimizing evidence-based management of agitation in this high-risk population.

Conclusion:

Patients presenting with active/recent substance use may have key sociodemographic differences from those without substance use. Those with a positive urine drug screen may be younger and more frequently have experience with carceral settings, however, they may also respond more robustly to agitation management with intramuscular dopamine receptor-blocking agents. Widely used intramuscular medications, haloperidol lactate and olanzapine, appear to be safe and effective for both positive and negative urine drug screen groups. These findings are broadly consistent with existing recommendations from Project BETA, which endorse the use of these agents in the management of acute agitation in patients with suspected psychosis or substance use.

Further research should prioritize prospective, multi-site studies with standardized urine drug screening, including blood alcohol levels, and consistent use of validated agitation assessment tools, in order to better characterize intoxication vs non-intoxication states. Randomized comparisons of commonly used agents such as HL and OL will be essential to refine evidence-based pharmacologic management of agitation in emergency psychiatric settings. Larger sample sizes should be obtained in order to power further analyses, and should include clinician-level data.

Highlights:

• Substance use did not increase medication needs for agitation care

• Patients with substance use showed no added safety risks or side effects

• Haloperidol and olanzapine worked equally well for substance use patients

• Findings support standard agitation treatment in emergency psychiatry

Data availability:

Data to be made available by the corresponding author upon reasonable request