Incidence of Neuroleptic Malignant Syndrome During Antipsychotic Treatment in Children and Youth: A National Cohort Study

Abstract

Objective:

The incidence of neuroleptic malignant syndrome (NMS), a rare, potentially fatal adverse effect of antipsychotics, among children and youth is unknown. This cohort study estimated NMS incidence in antipsychotic users age 5–24 years and described its variation according to patient and antipsychotic characteristics.

Methods:

We used national Medicaid data (2004–2013) to identify patients beginning antipsychotic treatment and calculated the incidence of NMS during antipsychotic current use. Adjusted hazard ratios (HRs) assessed the independent contribution of patient and antipsychotic characteristics to NMS risk.

Results:

The 1,032,084 patients had 131 NMS cases during 1,472,558 person-years of antipsychotic current use, or 8.9 per 100,000 person-years. The following five factors independently predicted increased incidence: age 18–24 years (HR [95% CI] = 2.45 [1.65–3.63]), schizophrenia spectrum and other psychotic disorders (HR = 5.86 [3.16–10.88]), neurodevelopmental disorders (HR = 7.11 [4.02–12.56]), antipsychotic dose >200mg chlorpromazine-equivalents (HR = 1.71 [1.15–2.54]), and first-generation antipsychotics (HR = 4.32 [2.74–6.82]). NMS incidence per 100,000 person-years increased from 1.8 (1.1–3.0) for those with none of these factors to 198.1 (132.8–295.6) for those with 4 or 5 factors. Findings were essentially unchanged in sensitivity analyses that restricted the study data to second-generation antipsychotics, children age 5–17 years, and the 5 most recent calendar years.

Conclusion:

In children and youth treated with antipsychotics, five factors independently identified patients with increased NMS incidence: age 18–24 years, schizophrenia spectrum and other psychotic disorders, neurodevelopmental disorders, first-generation drugs, and antipsychotic doses greater than 200 mg chlorpromazine-equivalents. Patients with 4 or 5 of these factors had more than 100 times the incidence of those with none. These findings could improve early identification of children and youth with elevated NMS risk, potentially leading to earlier detection and improved outcomes.

Introduction

Neuroleptic malignant syndrome (NMS) is a rare, potentially fatal neurological adverse effect of antipsychotics that presents with fever, rigidity, altered mental status, and autonomic and metabolic dysfunction (Schneider et al., 2020, Tse et al., 2015, Wijdicks, 2023). Case–control studies in adults have consistently linked higher doses (Berardi et al., 1998, Guinart et al., 2021, Keck et al., 1989, Sachdev et al., 1997, Su et al., 2014), rapid dose increase (Berardi et al., 1998, Keck et al., 1989), and first-generation antipsychotics—particularly haloperidol and other high-potency drugs (Nielsen et al., 2012, Su et al., 2014)—with increased incidence of NMS. Other markers of heightened NMS risk include prior NMS, catatonia, antipsychotic combinations, polypharmacy with lorazepam or other psychotropic medications, dehydration, and physical restraints (Berardi et al., 1998, Berman, 2011, Funayama et al., 2018, Nielsen et al., 2012, Su et al., 2014, Tse et al., 2015, Wijdicks, 2023), although some of these may be early manifestations of NMS or their treatment.

NMS has been reported for children and youth (Berloffa et al., 2021, Croarkin et al., 2008, Ghaziuddin et al., 2017, Henderson, 2011, León-Amenero and Huarcaya-Victoria, 2020, Neuhut et al., 2009, Ty and Rothner, 2001), who frequently are prescribed antipsychotics (Olfson et al., 2015) and who may be more vulnerable than adults to their adverse neurological effects (Correll et al., 2006). Case reports of NMS in children and adolescents have identified schizophrenia, neurodevelopmental disorders, and bipolar disorders as the diagnoses associated with antipsychotic treatment (Croarkin et al., 2008, Ghaziuddin et al., 2017, Henderson, 2011, León-Amenero and Huarcaya-Victoria, 2020), suggesting that, independent of antipsychotic dose and type, patients with certain major neuropsychiatric disorders might have increased risk of NMS. However, the case reports may simply reflect reporting bias, diagnosis-related differences in antipsychotic utilization, or antipsychotic treatment characteristics. Consequently, whether or not increased risk of NMS for children and youth is associated with specific antipsychotic indications, dose, or drug type is unknown.

Thus, we conducted a retrospective cohort study of NMS in a large national population of antipsychotic users 5–24 years of age (Ray et al., 2024). Our objectives were as follows: (1) to estimate the incidence of NMS in children and youth and (2) to describe variation in risk according to patient age, sex, diagnosis, and antipsychotic treatment characteristics, all of which are readily observable and potentially useful for identifying high-risk patients. To the best of our knowledge, this is the first controlled study of NMS in children and youth.

Methods

Cohort

Data

Study data came from a prior cohort study of antipsychotics and mortality in Medicaid beneficiaries 5–24 years of age (Ray et al., 2024). The data included enrollment, pharmacy, hospital, outpatient, and long-term care files from the Medicaid Analytic Extract, a national repository of data from the Medicaid and Children’s Health Insurance Programs (Supplementary Appendix SA1-a) (Leonard et al., 2017, Palmsten et al., 2013). The cohort was identified from data for 2004–2013; subsequent data availability and quality were adversely affected by a major transition in state Medicaid reporting requirements (Supplementary Appendix SA1-b) (anon, 2023a, anon, 2023b). Data came from 30 states collectively accounting for approximately 90% of Medicaid antipsychotic prescriptions for children and youth during the study period (Supplementary Appendix SA1-c, Supplementary Table SA1). Quality edits were performed to assure complete information on antipsychotic days of supply and dose (Supplementary Appendix S1-d, 1-e). Deaths were identified by linking Medicaid files with the National Death Index (NDI), a national registry of decedents (Supplementary Appendix SA1-f) (Wojcik et al., 2010). In accordance with Centers for Medicare and Medicaid Services policy, no data are reported for categories with fewer than 11 patients. The Vanderbilt University Medical Center IRB reviewed and approved the study, waiving informed consent.

Antipsychotics

Study antipsychotics were oral and intramuscular first- and second-generation drugs, including long-acting injectable antipsychotics and combinations with other medications (Supplementary Appendix SA2, Supplementary Table SA2). Antipsychotic prescriptions were identified from pharmacy claims, as well as outpatient claims, for administered antipsychotic injections. The prescription date was either that of dispensing (pharmacy) or administration (outpatient).

Cohort eligibility

The cohort included children and young adults (youth) 5 to 24 years of age. The lower age limit is that of initial school attendance for many children, with the consequent social and behavioral demands. The upper limit is that for the World Health Organization’s definition of youth (Bobo et al., 2013), which corresponds closely to definitions of emerging adulthood (Bennett et al., 2018, Kuwabara et al., 2007) and is consistent with other studies of psychotropic drugs in younger populations (Bobo et al., 2013, Cooper et al., 2011, Friedman and Leon, 2007, Olfson et al., 2012).

Cohort members were Medicaid beneficiaries in the 30 study states linked to the NDI with at least one antipsychotic prescription between January 1, 2004, and September 29, 2013, who met age and other cohort eligibility criteria (Supplementary Table SA3) on the prescription date. The patient had to have Medicaid enrollment on the qualifying prescription date and for the 4 preceding months and could not have Medicare or other insurance coverage. Because NMS risk is thought to be greatest early in treatment (Tse et al., 2015), there could be no antipsychotic prescription in the prior 120 days. There could be no inpatient, outpatient, or long-term care diagnosis of NMS in the past 120 days.

Follow-up

Follow-up began on t₀, the date of the initial qualifying antipsychotic prescription. Follow-up ended on t₁, the first of: September 29, 2013, the day preceding death or NMS occurrence, or 2 days prior to the 25th birthday. The analysis included all antipsychotic prescriptions with a date between t₀ and t₁ (Supplementary Appendix SA3, Supplementary Fig. SA1).

Neuroleptic malignant syndrome

NMS cases were identified from inpatient claims with an ICD-9-CM diagnosis code of 333.92 (neuroleptic malignant syndrome), which in one study had a positive predictive value of 82.9% (Lao et al., 2020), or an ICD-10 underlying cause of death code of G21.0 (malignant neuroleptic syndrome). The case definition did not consider emergency department or outpatient visits because it is unlikely that this life-threatening syndrome would be treated outside of inpatient settings. Secondary inpatient diagnoses were accepted because review of such claims indicated that the primary diagnosis typically was an NMS symptom (e.g., altered mental status), consequence (e.g., rhabdomyolysis), or psychiatric comorbidity (e.g., psychosis). The date of NMS generally was that of hospital admission, but was set to the prior day if there was an emergency department visit with an NMS diagnosis.

The case hospitalization was characterized according to length of stay and diagnoses and procedures during the stay. The stay included the initial hospitalization, as well as subsequent transfers. Because some lengthy hospital stays could include long-term care, the maximum length of stay was 30 days. Case fatality was defined as any death within 365 days of the NMS date.

Analysis

Antipsychotic prescription episode

Because antipsychotic use often is episodic and, over the 10-year study period, both patient and antipsychotic characteristics could change substantially, each prescription during follow-up defined an antipsychotic prescription episode (Supplementary Appendix SA4, Supplementary Fig. SA2). The definition of NMS generally requires recent antipsychotic exposure (Schneider et al., 2020); thus, analysis of person-time consisted of the current drug use defined by each episode, which began the day following the prescription date and ended with the first of the next prescription date, 30 days after the end of the days of supply (allowed for drug half-life, delayed outcome detection, nonadherence), or the day after t₁.

Each antipsychotic prescription episode was characterized according to daily dose in chlorpromazine-equivalents (Supplementary Table SA2) (Lehman et al., 1998, Leucht et al., 2014, Meltzer et al., 2008, Woods, 2003), recent dose increase, antipsychotic type, and individual drug. If there were multiple prescriptions on a single date, the doses were summed and combinations with both first- and second-generation drugs were considered first generation. Recent dose increase was defined as either no antipsychotic prescription in the past 120 days or a dose greater than twice that of the prior prescription.

Diagnoses

Antipsychotic prescription episodes were classified into mutually exclusive diagnostic categories. The major neuropsychiatric disorders, for which there is an established clinical rationale for antipsychotic treatment in children and youth (DeBattista and Hawkins, 2009, Hirsch, 2018, Hodgins et al., 2022), included schizophrenia spectrum and other psychotic disorders, neurodevelopment disorders—autism spectrum disorders and intellectual disability, bipolar or major depressive disorder, and multiple major neuropsychiatric disorders (Supplementary Table SA4-a). The other diagnoses were defined hierarchically as other mood disorders, conduct or impulse control disorder, attention-deficit/hyperactivity disorder (ADHD), anxiety, panic or obsessive–compulsive disorder, stress, other psychiatric diagnosis, and no recorded psychiatric diagnosis (Supplementary Table SA4-b).

Diagnoses were identified from inpatient, outpatient, and long-term care claims during the 120 days preceding through the prescription date (Supplementary Fig. SA2). The primary analysis required only a single claim, which reduced misclassification of the “other diagnosis” category by inclusion of patients with major neuropsychiatric disorders. A sensitivity analysis required one inpatient or emergency department claim or at least two other outpatient claims on separate dates, similar to the claims-based studies that validated these diagnostic categories (Hartung et al., 2013, Lurie et al., 1992, Shea et al., 2024, Straub et al., 2021).

Statistical analyses

The unit of analysis was the individual antipsychotic prescription episode. Independence assumptions were not violated because the episodes were nonoverlapping and the outcome occurred only once (Arnold, 1990).

NMS incidence was calculated by dividing the number of events by person-years of current antipsychotic use; variances were estimated assuming the Poisson distribution. To determine the independent association of patient and treatment characteristics with NMS, a proportional hazards regression was performed with a model that included patient age, sex, prescription year, diagnosis, and antipsychotic treatment characteristics. The antipsychotic dose term indicated those in the upper quartile. The analysis of individual antipsychotics included those with at least 11 cases of NMS; the reference category was risperidone. All statistical analyses were performed with SAS version 9.4. Statistical significance was defined as a 95% confidence interval (CI) for the hazard ratio (HR) that excluded 1.

Sensitivity analyses

Sensitivity analyses focused on the following patient groups of particular interest: those receiving second-generation antipsychotics, children age 5–17 years, those with antipsychotic prescriptions in the 5 most recent study years, and patients without severe somatic illness (e.g., cerebral palsy, Supplementary Table SA5). They also assessed methodologic alternatives as follows: modeling antipsychotic dose as a continuous variable to control for residual confounding, a more restrictive definition of major neuropsychiatric diagnoses, and increasing the length of the look-back window for eligibility criteria to 365 days.

Results

Cohort

The study included 1,032,084 Medicaid patients age 5–24 years who began antipsychotic treatment between 1st January 2004 and 29th September 2013 (Supplementary Fig. SA3). Cohort members had 15,881,698 antipsychotic prescription episodes during follow-up (Table 1), of which 81.9% were for children age 5–17 years, 65.7% were for males, and 64.3% had a prescription date between 2009 and 2013. Major neuropsychiatric disorders were associated with 43.7% of prescription episodes, including schizophrenia spectrum and other psychotic disorders (5.1%), neurodevelopmental disorders (10.3%), bipolar or major depressive disorders (21.8%), and multiple major neuropsychiatric disorders (6.5%). The most frequent other diagnoses were other mood disorders (13.3%), conduct or impulse control disorders (12.3%), and ADHD (17.0%). The median (interquartile range [IQR]) antipsychotic daily dose was 100 (50–200) mg chlorpromazine-equivalents, 19.0% had a recent dose increase, and 97.3% of the prescription episodes were for second-generation drugs. The most frequently prescribed individual antipsychotics (Supplementary Table SA6) were risperidone (40.5%), aripiprazole (23.1%), quetiapine (20.4%), and ziprasidone (4.6%). Haloperidol was the most frequently prescribed first-generation drug (1.0%). Long-acting injectable antipsychotics accounted for 0.8% of prescription episodes (Supplementary Table SA6).

Table 1.

Characteristics of Antipsychotic Prescription Episodes During Cohort Follow-up

Antipsychotic prescription episodes, % (N)	100.0	(15,881,698)
Age^a, years, % (N)
5–17	81.9	(13,005,290)
18–24	18.1	(2,876,408)
Sex, % (N)
Female	34.3	(5,440,513)
Male	65.7	(10,441,185)
Prescription year, % (N)
2004–2008	35.7	(5,668,258)
2009–2013	64.3	(10,213,440)
Diagnoses associated with antipsychotic treatment, % (N)
Major neuropsychiatric disorders	43.7	(6,938,280)
Schizophrenia spectrum and otherpsychotic disorders	5.1	(808,269)
Neurodevelopmental disorders	10.3	(1,639,341)
Bipolar/major depressive disorder	21.8	(3,465,373)
More than one major neuropsychiatric disorder	6.5	(1,025,297)
Other diagnoses^b	56.3	(8,943,418)
Other mood disorders	13.3	(2,113,038)
Conduct or impulse control disorder	12.3	(1,956,748)
Attention-deficit/hyperactivity disorder	17.0	(2,692,909)
Anxiety, panic, or obsessive–compulsive disorder	1.4	(218,897)
Stress	1.0	(151,887)
Other psychiatric diagnoses	2.3	(372,611)
No recorded psychiatric diagnosis	9.1	(1,437,328)
Antipsychotic daily dose, mg chlorpromazine equivalents, median (IQR)	100	(50–200)
Recent antipsychotic dose increase^c, % (N)
No	81.0	(12,863,255)
Yes	19.0	(3,018,443)
Antipsychotic type, % (N)
Second generation	97.3	(15,453,296)
First generation	2.7	(428,402)

On the prescription date.

Categories defined hierarchically. For example, the “Conduct or impulse control disorder” does not include episodes with other mood disorders.

No antipsychotic prescription past 120 days or dose greater than twice that of prior prescription.

IQR, interquartile range.

NMS cases

Cohort members had 1,472,558 person-years of current antipsychotic use, during which there were 131 NMS cases, or 8.9 per 100,000 person-years. The median (IQR) length of the case inpatient stay was 7 (5–19) days (Table 2). During the hospital stay, 54.2% had an intensive care unit stay, 48.1% had CT or MRI imaging of the brain or head, and 21.4% had a lumbar puncture. Frequent diagnoses were altered mental status (46.6%), autonomic dysregulation (46.6%), fever (36.6%), rhabdomyolysis (35.9%), and metabolic disorders (19.1%). Case fatality could not be reported because of the small number of deaths.

Table 2.

Neuroleptic Malignant Syndrome Case Characteristics

Cases, N	131
Length of stay, days, median (IQR)	7 (5–19)

	N	%
Intensive care unit	71	54.2%
CT scan or MRI imaging brain or head	63	48.1%
Altered mental status	61	46.6%
Autonomic dysregulation^a	61	46.6%
Fever	48	36.6%
Rhabdomyolysis	47	35.9%
Seizure or convulsion	33	25.2%
Extrapyramidal symptoms	30	22.9%
Lumbar puncture	28	21.4%
Respiratory symptoms, severe^b	25	19.1%
Respiratory symptoms, other^c	27	20.6%
Metabolic disorders^d	25	19.1%
Dehydration	20	15.3%
Sepsis	13	9.9%
Death within 365 days	^*	^*

Hypertension/hypotension, arrhythmia, dysphagia.

Pulmonary collapse, acute or chronic respiratory failure, respiratory arrest, other pulmonary insufficiency, tracheostomy/intubation, mechanical ventilation, cardiopulmonary respiration.

Dyspnea or other respiratory symptoms.

Hyperosmolality/hypernatremia, hypoosmolality/hyponatremia, acidosis/alkalosis, mixed acid–base balance disorder, hyperpotassemia, hypopotassemia, other electrolyte and fluid disorders.

‘*’

indicates fewer than 11 cases. Characteristics determined from diagnoses and procedures during the inpatient stay for NMS.

IQR, interquartile range.

Factors associated with NMS incidence

Both patient and antipsychotic treatment characteristics were independently associated with the incidence of NMS (Table 3). Risk increased for patients age 18–24 years (HR [95% CI] = 2.45 [1.65–3.63]), but was not increased for males (HR = 0.94 [0.66–1.35]). When patients with a single major neuropsychiatric disorder were compared to those with other diagnoses, the incidence was significantly increased for schizophrenia spectrum and other psychotic disorders (HR = 5.86 [3.16–10.88]) and neurodevelopmental disorders (7.11 [4.02–12.56]), but was not significantly increased for bipolar or major depressive disorders (HR = 1.20 [0.61–2.37]). The incidence of NMS for prescriptions with daily doses >200mg chlorpromazine-equivalents was greater than that for lower doses (HR = 1.71 [1.15–2.54]). However, a recent dose increase was not associated with elevated risk (HR = 1.00 [0.68–1.47]).

Table 3.

Neuroleptic Malignant Syndrome Occurrence According to Patient and Antipsychotic Treatment Characteristics

	Person-years	NMS cases	Rate/10⁵ PY	Adjusted HR (95% CI)
Age, years
5–17	1,211,696	54	4.5	Reference
18–24	260,862	77	29.5	2.45 (1.65–3.63)
Sex
Female	505,500	51	10.1	Reference
Male	967,058	80	8.3	0.94 (0.66–1.35)
Prescription year
2004–2008	538,781	44	8.2	Reference
2009–2013	933,777	87	9.3	1.02 (0.71–1.47)
Diagnosis
Schizophrenia spectrum and other psychotic disorders	68,810	24	34.9	5.86 (3.16–10.88)
Neurodevelopmental disorders	141,491	26	18.4	7.11 (4.02–12.56)
Bipolar/major depressive disorder	319,733	14	4.4	1.20 (0.61–2.37)
More than one major neuropsychiatric disorder	80,340	44	54.8	11.44 (6.62–19.77)
Other diagnoses	862,183	23	2.7	Reference
Dose greater than 200 mg chlorpromazine-equivalents
No	1,212,398	73	6.0	Reference
Yes	260,160	58	22.3	1.71 (1.15–2.54)
Recent antipsychotic dose increase^a
No	1,175,757	94	8.0	Reference
Yes	296,802	37	12.5	1.00 (0.68–1.47)
Antipsychotic type
Second generation	1,439,767	100	6.9	Reference
First generation	32,791	31	94.5	4.32 (2.74–6.82)

No antipsychotic prescription in past 120 days or dose greater than twice that of prior prescription.

HRs are adjusted for the variables shown in table 1

CI, confidence interval; HR, Hazard ratio; NMS, Neuroleptic Malignant Syndrome.

The incidence of NMS for first-generation antipsychotics was greater than that for second-generation drugs (Table 3, HR = 4.32 [2.74–6.82]). When individual antipsychotics were compared with risperidone (Supplementary Table SA7), the HR was highest for haloperidol (7.10 [3.51–14.36]). The HRs for aripiprazole (0.92 [0.51–1.66]) and quetiapine (0.96 [0.51–1.80]) were not significantly increased, but that for ziprasidone (2.66 [1.34–5.29]) was. Patients administered long-acting injectable antipsychotics had too few cases of NMS for analysis.

We calculated the incidence of NMS according to the number of the following five factors each independently associated with increased risk: age 18–24 years, schizophrenia spectrum and other psychotic disorders, neurodevelopmental disorders, first-generation antipsychotic, and daily dose >200 mg chlorpromazine-equivalents. The incidence increased markedly with the number of these risk factors (Fig. 1). For those with none, the incidence was 1.8 (1.1–3.0) NMS cases per 100,000 person-years, whereas for those with 4 or 5 factors, the incidence was 198.1 (132.8–295.6) cases per 100,000 person-years.

FIG. 1.

Incidence of NMS in children and youth according to N of risk factors: age 18–24 years, schizophrenia spectrum or other psychotic disorders, neurodevelopmental disorders, antipsychotic daily dose >200 mg chlorpromazine-equivalents, or first-generation antipsychotic. Whiskers represent the 95% CI. CI, confidence intervals; NMS, neuroleptic malignant syndrome.

Sensitivity analyses

The same five factors were independently associated with increased NMS incidence in each of the sensitivity analyses (Table 4). The 95% confidence intervals for the point estimates of the HRs in the sensitivity analyses included that of the primary analysis.

Table 4.

Sensitivity Analyses

	Primary analysis	Second-generation antipsychotics	Children age5–17 years	Calendar year 2009–2013	No severe somatic illness^a	Antipsychotic dose continuous^b	Alternate neuropsychiatric diagnosis definition^c	365-day look-back window^d
Person-years	1,472,558	1,439,767	1,211,696	933,777	1,423,602	1,472,558	1,472,558	1,382,297
Neuroleptic m alignant syndrome cases	131	100	54	87	115	131	131	123
Incidence/10⁵ person-years	8.9	6.9	4.5	9.3	8.1	8.9	8.9	8.9
Adjusted HR (95% CI)
Age 18–24 years	2.45 (1.65–3.63)	2.89 (1.88–4.44)	n/a	2.70 (1.66–4.42)	2.39 (1.57–3.63)	2.32 (1.56–3.44)	2.52 (1.70–3.73)	2.46 (1.64–3.69)
Male sex	0.94 (0.66–1.35)	1.14 (0.75–1.73)	0.82 (0.47–1.43)	n/a	0.98 (0.67–1.44)	0.93 (0.65–1.33)	0.93 (0.65–1.34)	0.87 (0.60–1.25)
Prescription year 2009–2013	1.02 (0.71–1.47)	1.05 (0.69–1.58)	0.91 (0.53–1.58)	1.02 (0.65–1.58)	1.09 (0.73–1.61)	1.02 (0.71–1.47)	1.03 (0.71–1.48)	0.95 (0.65–1.38)
Diagnosis
Schizophrenia spectrum and other psychotic disorders	5.86 (3.16–10.88)	5.08 (2.45–10.56)	9.17 (3.78–22.27)	5.61 (2.62–12.00)	7.01 (3.69–13.34)	5.54 (3.00–10.22)	5.63 (3.15–10.08)	5.62 (3.01–10.52)
Neurodevelopmental disorders	7.11 (4.02–12.56)	6.56 (3.50–12.30)	6.44 (2.91–14.28)	5.83 (2.85–11.90)	6.56 (3.45–12.49)	7.06 (3.97–12.55)	7.97 (4.75–13.36)	6.80 (3.81–12.15)
Bipolar/unipolar depressive disorder	1.20 (0.61–2.37)	1.43 (0.71–2.87)	1.78 (0.76–4.16)	1.04 (0.44–2.49)	1.30 (0.64–2.66)	1.14 (0.58–2.25)	1.38 (0.74–2.58)	0.94 (0.45–1.95)
Multiple major neuropsychiatric disorders	11.44 (6.62–19.77)	13.49 (7.47–24.36)	11.05 (4.94–24.74)	11.83 (6.08–23.00)	13.04 (7.29–23.32)	10.46 (6.04–18.11)	11.78 (6.87–20.20)	10.83 (6.22–18.85)
Antipsychotic daily dose >200 mg chlorpromazine equivalents	1.71 (1.15–2.54)	1.65 (1.07–2.55)	2.87 (1.59–5.20)	1.70 (1.05–2.76)	1.75 (1.15–2.66)	1.25 (1.09–1.43)	1.73 (1.16–2.57)	1.56 (1.03–2.36)
Recent antipsychotic dose increase^e	1.00 (0.68–1.47)	0.72 (0.44–1.18)	1.31 (0.74–2.35)	0.94 (0.58–1.54)	0.89 (0.58–1.35)	0.99 (0.67–1.46)	1.04 (0.70–1.53)	1.31 (0.88–1.96)
First-generation antipsychotic	4.32 (2.74–6.82)	n/a	6.44 (2.96–13.99)	3.75 (2.14–6.55)	3.97 (2.42–6.51)	4.06 (2.54–6.48)	4.31 (2.73–6.82)	4.75 (2.98–7.57)

See Supplementary Table S5.

Dose continuous variable. Units log(base 2) because change in risk hypothesized to be proportional to change in dose. The point estimate is the increase in HR for dose doubling.

Requires a single inpatient or emergency department diagnosis or outpatient diagnoses on at least two separate dates.

The look-back period is the interval preceding the prescription date to determine if cohort eligibility criteria are met.

No antipsychotic prescription in past 120 days or dose greater than twice that of prior prescription.

CI, confidence interval; HR, Hazard ratio.

Discussion

Neuroleptic malignant syndrome is one of the most serious adverse drug effects child and adolescent psychiatrists encounter in clinical practice. However, because this complication is rare, detection may be delayed or the syndrome misidentified, thus potentially leading to increased morbidity, prolonged inpatient stay, or death. This national cohort study of more than one million children and youth treated with antipsychotics during a 10-year period found that the following five factors were independently associated with significantly increased risk for NMS: age 18–24 years, schizophrenia spectrum and other psychotic disorders, neurodevelopmental disorders, antipsychotic dose greater than 200 mg chlorpromazine equivalents, and first-generation drugs. Incidence increased with the number of risk factors present; for patients with 4 or 5 the incidence was more than 100 times that for patients with none. These data thus provide a basis for early identification of children and youth at elevated risk for NMS for whom monitoring for signs of this life-threatening syndrome could be increased, potentially leading to earlier detection and improved outcomes.

The focus of prior NMS research has been adults with schizophrenia spectrum and other psychotic disorders (Berardi et al., 1998, Guinart et al., 2021, Keck et al., 1989, Nielsen et al., 2012, Sachdev et al., 1997, Su et al., 2014). However, we found that in the study cohort of children and youth there were more cases of NMS in patients with autism spectrum disorders or intellectual disability than for any other single study neuropsychiatric disorder. Furthermore, the point estimate of the HR for the study of neurodevelopmental disorders was greater than that for schizophrenia spectrum and other psychotic disorders. These findings underscore the need for practitioners who care for children and youth with neurodevelopmental disorders to monitor patients treated with antipsychotics for signs of NMS.

Several possible mechanisms may contribute to the strong association of schizophrenia spectrum and other psychotic disorders and neurodevelopmental disorders with increased incidence of NMS. In both disorders, the prevalence of the catatonia syndrome (Heckers and Walther, 2023) is elevated (Hauptman et al., 2023, Rogers et al., 2023, Walther et al., 2019), and in patients treated with antipsychotics, catatonia is associated with increased risk of NMS (Berardi et al., 1998, Funayama et al., 2018, Rogers et al., 2023). Disturbances in early brain development, thought to partially explain the psychopathology of both schizophrenia and neurodevelopmental disorders (Owen and O’Donovan, 2017, Rees et al., 2021), might also increase vulnerability to NMS. Other patient or treatment characteristics may be involved. Regardless of the mechanisms, our data indicate that these disorders identify children and youth for whom careful monitoring for signs of NMS should be considered.

Unlike some case–control studies in adults (Berardi et al., 1998, Keck et al., 1989), recent dose increase was not associated with increased risk for NMS. Differences in the study populations may explain this finding. The adult studies were set in inpatient psychiatric facilities where patients were predominantly treated with high doses of first-generation drugs, and rapid, substantial dose increases may have been more common than in outpatient settings. In contrast, the population in this study of children and youth was not restricted to inpatients, more than 97% of prescriptions were for second-generation antipsychotics, and dose escalation may have been more gradual.

Haloperidol had the highest incidence of NMS, which is consistent with the findings of one adult study (Su et al., 2014). Although haloperidol and other first-generation antipsychotics have more frequent extrapyramidal adverse effects than the second-generation drugs, haloperidol continues to be prescribed in emergency department and inpatient settings for children with agitation or delirium (Nash et al., 2023). Our findings suggest that younger patients in such settings who receive haloperidol, particularly those with schizophrenia spectrum and other psychotic disorders, or neurodevelopmental disorders, should be monitored for signs of NMS.

A case–control study in adults reported that aripiprazole was associated with increased risk of NMS (Su et al., 2014). In contrast, we found no significant differences in the incidence of NMS for risperidone, aripiprazole, and quetiapine, the most frequently prescribed antipsychotics in the study cohort. However, the incidence was significantly increased for ziprasidone; further research is needed to confirm this finding.

Limitations

There were several limitations. First, as in other studies (Guinart et al., 2021, Lao et al., 2020, Modi et al., 2016), NMS cases were identified retrospectively from inpatient discharge or underlying cause of death diagnostic codes and thus may be misclassified (Schneider et al., 2020, Wijdicks, 2023). Several lines of evidence suggest that the false positive rate for study cases was low. A chart-review study reported a positive predictive value of 82.9% for inpatient stays with an NMS discharge diagnosis (Lao et al., 2020), although whether this finding can be applied to the Medicaid database has not been demonstrated. The lengthy case inpatient stays indicated a severe neurological illness with diagnostic procedures consistent with NMS, and the incidence of complications such as rhabdomyolysis was similar to that reported in a national inpatient sample study of this syndrome (Modi et al., 2016). Finally, every case occurred during antipsychotic use, satisfying the criterion for recent exposure to a dopamine antagonist (Wijdicks, 2023). Nevertheless, residual misclassification, if nondifferential, would attenuate the association of patient and treatment characteristics with this syndrome.

However, cases of NMS may have been misdiagnosed and thus not included in the analysis. Although NMS is a severe disorder that is likely to lead to hospitalization, diagnosis is clinically based and the syndrome may be mistaken for many drug-related and other conditions such as malignant catatonia, malignant hyperthermia, serotonin syndrome, seizures, or infections (Berloffa et al., 2021, Duhan et al., 2023, Wijdicks, 2023). Consequently, the study estimates of NMS incidence in children and youth may be too low and thus understate the clinical impact of this antipsychotic adverse effect.

Second, the study’s neuropsychiatric disorders were defined from Medicaid claims. Although claims data can reliably identify the broad study diagnostic categories (Hartung et al., 2013, Lurie et al., 1992, Shea et al., 2024, Straub et al., 2021), their capacity to provide additional detail on important characteristics of the disorders that potentially could better identify patients with high risk of NMS—such as subtype, etiology, and severity—is uncertain.

Third, the number of NMS cases was limited. Thus, we could not perform more detailed analyses of various clinical subgroups of interest, particularly since privacy safeguards prohibited reporting of table cells with fewer than 11 cases. Similarly, we could not conduct exploratory analyses of several other potential correlates of NMS suggested by adult studies (Berardi et al., 1998, Berman, 2011, Funayama et al., 2018, Nielsen et al., 2012, Su et al., 2014, Wijdicks, 2023). Consequently, we focused on patient demographics, neuropsychiatric diagnoses, and antipsychotic treatment characteristics that are readily available to practitioners and plausibly could identify patients with increased NMS risk for whom more intense monitoring could be appropriate.

Fourth, the most recent study data were from 2013, as the availability and quality of more recent data were influenced by a major transition in state Medicaid reporting requirements (anon, 2023a, anon, 2023b). However, although antipsychotic prescribing for younger populations is decreasing (Bushnell et al., 2023), risperidone, aripiprazole, and quetiapine still are the most frequently prescribed drugs for children and youth (Hsueh et al., 2023), and haloperidol continues to be prescribed in pediatric patients (Nash et al., 2023). Furthermore, findings did not change in a sensitivity analysis that was restricted to the 5 most recent calendar years. Thus, the study findings should be applicable to contemporary practice.

Fifth, cohort members were all Medicaid beneficiaries, which limits generalizability. Nevertheless, the national Medicaid population is important per se, as this program provides health insurance coverage for an estimated 36% of U.S. children (Keisler-Starkey and Bunch, 2022), who have elevated prevalence of antipsychotic use (Edelsohn et al., 2017). Because study data were from prior to the Affordable Care Act Medicaid expansions in 2014 (Fung et al., 2022), Medicaid eligibility criteria were more restrictive for cohort members 18–24 years than for children, which would further limit generalizability for young adults.

Conclusion

In children and youth treated with antipsychotics, the following five factors independently identified patients with increased NMS incidence: age 18–24 years, schizophrenia spectrum and other psychotic disorders, neurodevelopmental disorders, first-generation drugs, and antipsychotic doses greater than 200 mg chlorpromazine-equivalents. Patients with 4 or 5 of these factors had more than 100 times the incidence of those with none. These findings could improve early identification of children and youth with elevated NMS risk, potentially leading to earlier detection and improved outcomes.

Footnotes

Clinical Significance

Practitioners who prescribe antipsychotics for children and youth, particularly for those with schizophrenia spectrum and other psychotic disorders, autism spectrum disorders, and intellectual disability, should be alert for signs and symptoms of neuroleptic malignant syndrome. Study findings strengthen recommendations to avoid limit or avoid, when possible, treatment of children and youth with FGAs, particularly haloperidol.

Authors’ Contributions

Concept and design: Ray, Fuchs, Olfson, Cooper. Acquisition, analysis, or interpretation of data: Ray, Fuchs, Olfson, Stein, Murray, Daugherty, Cooper. Drafting of the manuscript: Ray, Fuchs, Olfson, Daugherty. Critical review of the manuscript for important intellectual content: Ray, Fuchs, Olfson, Stein, Murray, Cooper. Statistical analysis: Ray. Obtained funding: Ray, Murray. Administrative, technical, or material support: Daugherty, Cooper. Supervision: Ray.

Disclosures

The authors disclose that the only financial affiliation is with Grant HD097344 from the National Institute for Child Health and Human Development, which funded the study. The sponsor had no role in study design or conduct. The authors have no other relevant commercial or financial relationships.

Data Availability Statement

The Medicaid Analytical eXtract files, the source of the study data, are housed in the Center for Medicare and Medicaid Studies Chronic Condition Warehouse and are available on negotiation of a Data Use Agreement ().

Supplementary Material

Supplementary Figure SA1

Supplementary Figure SA2

Supplementary Figure SA3

Supplementary Appendix S1

Supplementary Appendix SA1

Supplementary Appendix SA2

Supplementary Appendix SA3

Supplementary Appendix SA4

Supplementary Table SA1

Supplementary Table SA2

Supplementary Table SA3

Supplementary Table SA4

Supplementary Table SA5

Supplementary Table SA6

Supplementary Table SA7

References

anon. Available from: https://www.cms.gov/Research-Statistics-Data-and-Systems/Computer-Data-and-Systems/MedicaidDataSourcesGenInfo/MAXGeneralInformation 2023a

anon. Available from: https://www.macpac.gov/publication/update-on-transformed-medicaid-statistical-information-system-t-msis/ 2023b

Arnold

. Mathematical statistics. Prentice-Hall, 1990:157.

Bennett

, Miller

, Stollon

, et al. Autism spectrum disorder and transition-aged youth. Curr Psychiatry Rep, 2018; 20(11):103; doi: 10.1007/s11920-018-0967-y

Berardi

, Amore

, Keck

Jr. , et al. Clinical and pharmacologic risk factors for neuroleptic malignant syndrome: A Case-Control Study. Biol Psychiatry, 1998; 44(8):748–754; doi: 10.1016/s0006-3223(97)00530-1

Berloffa

, Dosi

, Tascini

, et al. Neuroleptic malignant syndrome in children with autism spectrum disorder (ASD): A case report and brief review of recent literature. Children (Basel), 2021; 8(12); doi: 10.3390/children8121201

Berman

. Neuroleptic malignant syndrome: A review for neurohospitalists. Neurohospitalist, 2011; 1(1):41–47; doi: 10.1177/1941875210386491

Bobo

, Cooper

, Stein

, et al. Antipsychotics and the risk of type 2 diabetes mellitus in children and youth. JAMA Psychiatry, 2013; 70(10):1067–1075.

Bushnell

, Lloyd

, Olfson

, et al. Antipsychotic medication use in Medicaid-insured children decreased substantially between 2008 and 2016. Health Aff, 2023; 42(7):937–980.

10.

Cooper

, Habel

, Sox

, et al. ADHD drugs and serious cardiovascular events in children and young adults. N Engl J Med, 2011; 365(20):1896–1904.

11.

Correll

, Penzner

, Parikh

, et al. Recognizing and monitoring adverse events of second-generation antipsychotics in children and adolescents. Child Adolesc Psychiatr Clin N Am, 2006; 15(1):177–206; doi: 10.1016/j.chc.2005.08.007

12.

Croarkin

, Emslie

, Mayes

. Neuroleptic malignant syndrome associated with atypical antipsychotics in pediatric patients: A review of published cases. J Clin Psychiatry, 2008; 69(7):1157–1165; doi: 10.4088/jcp.v69n0716

13.

DeBattista

, Hawkins

. Utility of atypical antipsychotics in the treatment of resistant unipolar depression. CNS Drugs, 2009; 23(5):369–377; doi: 10.2165/00023210-200923050-00002

14.

Duhan

, Keisham

, Jain

. An unusual cause of persistent tachycardia: Atypical neuroleptic malignant syndrome. Cureus, 2023; 15(3):e36428; doi: 10.7759/cureus.36428

15.

Edelsohn

, Karpov

, Parthasarathy

, et al. Trends in antipsychotic prescribing in medicaid-eligible youth. J Am Acad Child Adolesc Psychiatry, 2017; 56(1):59–66; doi: 10.1016/j.jaac.2016.10.005

16.

Friedman

, Leon

. Expanding the black box—depression, antidepressants, and the risk of suicide. N Engl J Med, 2007; 356(23):2343–2346; doi: 10.1056/NEJMp078015

17.

Funayama

, Takata

, Koreki

, et al. Catatonic stupor in schizophrenic disorders and subsequent medical complications and mortality. Psychosom Med, 2018; 80(4):370–376; doi: 10.1097/psy.0000000000000574

18.

Fung

, Yang

, Cook

, et al. Changes in insurance coverage continuity after affordable care act expansion of medicaid eligibility for young adults with low income in massachusetts. JAMA Health Forum, 2022; 3(7):e221996; doi: 10.1001/jamahealthforum.2022.1996

19.

Ghaziuddin

, Hendriks

, Patel

, et al. Neuroleptic malignant syndrome/malignant catatonia in child psychiatry: Literature review and a case series. J Child Adolesc Psychopharmacol, 2017; 27(4):359–365; doi: 10.1089/cap.2016.0180

20.

Guinart

, Taipale

, Rubio

, et al. Risk factors, incidence, and outcomes of neuroleptic malignant syndrome on long-acting injectable vs oral antipsychotics in a nationwide schizophrenia cohort. Schizophr Bull, 2021; 47(6):1621–1630; doi: 10.1093/schbul/sbab062

21.

Hartung

, Middleton

, McFarland

, et al. Use of administrative data to identify off-label use of second-generation antipsychotics in a Medicaid population. Psychiatr Serv, 2013; 64(12):1236–1242; doi: 10.1176/appi.ps.005482012

22.

Hauptman

, Cohen

, Dhossche

, et al. Catatonia in neurodevelopmental disorders: Assessing catatonic deterioration from baseline. Lancet Psychiatry, 2023; 10(3):228–234; doi: 10.1016/s2215-0366(22)00436-9

23.

Heckers

, Walther

. Catatonia. N Engl J Med, 2023; 389(19):1797–1802; doi: 10.1056/NEJMra2116304

24.

Henderson

. Neuroleptic malignant syndrome in adolescents: Four probable cases in the Western Cape. S Afr Med J, 2011; 101(6):405–407.

25.

Hirsch

. Dosing and Monitoring: Children and adolescents. Psychopharmacol Bull, 2018; 48(2):34–92.

26.

Hodgins

, Winsor

, Barnhill

. Pharmacotherapy of disruptive behaviors in children with intellectual disabilities. Paediatr Drugs, 2022; 24(5):465–482; doi: 10.1007/s40272-022-00517-y

27.

Hsueh

, Iturralde

, Slama

, et al. Cardiometabolic monitoring and sociodemographic and clinical characteristics of youths prescribed antipsychotic medications. Psychiatr Serv, 2023; 74(8):801–808; doi: 10.1176/appi.ps.20220151

28.

Keck

Jr. , Pope

Jr. , Cohen

, et al. Risk factors for neuroleptic malignant syndrome. A case-control study. Arch Gen Psychiatry, 1989; 46(10):914–918; doi: 10.1001/archpsyc.1989.01810100056011

29.

Keisler-Starkey

, Bunch

. Current Population Reports, P60-278, Health Insurance Coverage in the United States: 2021. In: (Bureau USC ed.) U.S. Government Publishing Office: Washington, D.C., 2022.

30.

Kuwabara

, Van Voorhees

, Gollan

, et al. A qualitative exploration of depression in emerging adulthood: Disorder, development, and social context. Gen Hosp Psychiatry, 2007; 29(4):317–324; doi: 10.1016/j.genhosppsych.2007.04.001

31.

Lao

KSJ

, Zhao

, Blais

, et al. Antipsychotics and risk of neuroleptic malignant syndrome: A Population-Based Cohort and Case-Crossover Study. CNS Drugs, 2020; 34(11):1165–1175; doi: 10.1007/s40263-020-00767-9

32.

Lehman

, Steinwachs

, Projec

SC-I

. Patterns of usual care for schizophrenia: Initial results from the schizophrenia patient outcomes research team (PORT) client survey. Schizophr Bull, 1998; 24(1):11–20.

33.

León-Amenero

, Huarcaya-Victoria

. neuroleptic malignant syndrome in children and adolescents: Systematic review of case reports. Rev Colomb Psiquiatr (Engl Ed), 2020; doi: 10.1016/j.rcp.2019.10.003

34.

Leonard

, Brensinger

, Nam

, et al. The quality of medicaid and medicare data obtained from CMS and its contractors: Implications for pharmacoepidemiology. BMC Health Serv Res, 2017; 17(1):304; doi: 10.1186/s12913-017-2247-7

35.

Leucht

, Samara

, Heres

, et al. Dose equivalents for second-generation antipsychotics: The minimum effective dose method. Schizophr Bull, 2014; 40(2):314–326; doi: 10.1093/schbul/sbu001

36.

Lurie

, Popkin

, Dysken

, et al. Accuracy of diagnoses of schizophrenia in Medicaid claims. Hosp Community Psychiatry, 1992; 43(1):69–71; doi: 10.1176/ps.43.1.69

37.

Meltzer

, Bobo

, Nuamah

, et al. Efficacy and tolerability of oral paliperidone extended-release tablets in the treatment of acute schizophrenia: Pooled data from three 6-week, placebo-controlled studies. J Clin Psychiatry, 2008; 69(5):817–829.

38.

Modi

, Dharaiya

, Schultz

, et al. Neuroleptic malignant syndrome: Complications, outcomes, and mortality. Neurocrit Care, 2016; 24(1):97–103; doi: 10.1007/s12028-015-0162-5

39.

Nash

, Olfson

, Rothenberg

, et al. Psychotropic medication use in United States pediatric emergency department visits. Acad Pediatr, 2023; 23(5):971–979; doi: 10.1016/j.acap.2022.11.016

40.

Neuhut

, Lindenmayer

, Silva

. Neuroleptic malignant syndrome in children and adolescents on atypical antipsychotic medication: A review. J Child Adolesc Psychopharmacol, 2009; 19(4):415–422; doi: 10.1089/cap.2008.0130

41.

Nielsen

, Wallenstein Jensen

, Nielsen

. Neuroleptic malignant syndrome-an 11-year longitudinal case-control study. Can J Psychiatry, 2012; 57(8):512–518; doi: 10.1177/070674371205700810

42.

Olfson

, Blanco

, Liu

, et al. National trends in the office-based treatment of children, adolescents, and adults with antipsychotics. Arch Gen Psychiatry, 2012; 69(12):1247–1256; doi: 10.1001/archgenpsychiatry.2012.647

43.

Olfson

, King

, Schoenbaum

. Treatment of young people with antipsychotic medications in the United States. JAMA Psychiatry, 2015; 72(9):867–874; doi: 10.1001/jamapsychiatry.2015.0500

44.

Owen

, O’Donovan

. Schizophrenia and the neurodevelopmental continuum: Evidence from genomics. World Psychiatry, 2017; 16(3):227–235; doi: 10.1002/wps.20440

45.

Palmsten

, Huybrechts

, Mogun

, et al. Harnessing the medicaid analytic eXtract (MAX) to evaluate medications in pregnancy: Design considerations. PLoS One, 2013; 8(6):e67405; doi: 10.1371/journal.pone.0067405

46.

Ray

, Fuchs

, Olfson

, et al. Antipsychotic medications and mortality in children and young adults. JAMA Psychiatry, 2024; 81(3):260–269; doi: 10.1001/jamapsychiatry.2023.4573

47.

Rees

, Creeth

HDJ

, Hwu

, et al. Schizophrenia, autism spectrum disorders and developmental disorders share specific disruptive coding mutations. Nat Commun, 2021; 12(1):5353; doi: 10.1038/s41467-021-25532-4

48.

Rogers

, Oldham

, Fricchione

, et al. Evidence-based consensus guidelines for the management of catatonia: Recommendations from the british association for psychopharmacology. J Psychopharmacol, 2023; 37(4):327–369; doi: 10.1177/02698811231158232

49.

Sachdev

, Mason

, Hadzi-Pavlovic

. Case-control study of neuroleptic malignant syndrome. Am J Psychiatry, 1997; 154(8):1156–1158; doi: 10.1176/ajp.154.8.1156

50.

Schneider

, Regente

, Greiner

, et al. Neuroleptic malignant syndrome: Evaluation of drug safety data from the AMSP program during 1993-2015. Eur Arch Psychiatry Clin Neurosci, 2020; 270(1):23–33; doi: 10.1007/s00406-018-0959-2

51.

Shea

, Villodas

, Ventimiglia

, et al. foster care involvement among youth with intellectual and developmental disabilities. JAMA Pediatr, 2024; 178(4):384–390; doi: 10.1001/jamapediatrics.2023.6580

52.

Straub

, Huybrechts

, Mogun

, et al. Association of Neuraxial Labor Analgesia for Vaginal Childbirth With Risk of Autism Spectrum Disorder. JAMA Netw Open, 2021; 4(12):e2140458; doi: 10.1001/jamanetworkopen.2021.40458

53.

, Chang

, Hayes

, et al. Retrospective chart review on exposure to psychotropic medications associated with neuroleptic malignant syndrome. Acta Psychiatr Scand, 2014; 130(1):52–60; doi: 10.1111/acps.12222

54.

Tse

, Barr

, Scarapicchia

, et al. Neuroleptic malignant syndrome: A review from a clinically oriented perspective. Curr Neuropharmacol, 2015; 13(3):395–406; doi: 10.2174/1570159x13999150424113345

55.

, Rothner

. Neuroleptic malignant syndrome in children and adolescents. J Child Neurol, 2001; 16(3):157–163; doi: 10.1177/088307380101600301

56.

Walther

, Stegmayer

, Wilson

, et al. Structure and neural mechanisms of catatonia. Lancet Psychiatry, 2019; 6(7):610–619; doi: 10.1016/s2215-0366(18)30474-7

57.

Wijdicks

EFM

. Neuroleptic malignant syndrome. In: UpToDate. ( Conner

. ed.) Wolters Kluwer: 2023.

58.

Wojcik

, Huebner

, Jorgensen

. Strategies for using the national death index and the social security administration for death ascertainment in large occupational cohort mortality studies. Am J Epidemiol, 2010; 172(4):469–477; doi: 10.1093/aje/kwq130

59.

Woods

. Chlorpromazine equivalent doses for the newer atypical antipsychotics. J Clin Psychiatry, 2003; 64(6):663–667.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.44 MB