Clozapine and Neutrophil Response in Patients of African Descent: A Six-Month, Multinational, Prospective, Open-Label Clinical Trial

Abstract

Introduction:

Clozapine is the most effective antipsychotic for treatment-resistant schizophrenia, but it is markedly underutilized, particularly in the US Black population, partly because of concern over clozapine-associated low absolute neutrophil count (ANC). People of African descent have a lower normative ANC range than the White population, which is associated with a specific “ACKR1-null” (“Duffy null”) CC genotype (SNP rs2814778) on the ACKR1 gene, termed benign ethnic neutropenia (BEN). The range of ANC variability and safety of clozapine have not been established in people with BEN or examined prospectively in people of African descent.

Methods:

We completed a multisite, 6-month, prospective, open-label clinical trial of clozapine treatment in people of African descent with schizophrenia spectrum disorders for whom clozapine was clinically indicated, with or without the ACKR1-null genotype. We examined clozapine safety and weekly ANC during clozapine treatment and evaluated ANC variability by ACKR1-null genotype, sex, study site, and clozapine dosing using repeated measures analysis of covariance. Genotype was assayed using TaqMan^® technology.

Results:

We enrolled 274 participants, of whom 227 (82.8%) completed 6 months of clozapine treatment. There was one case of severe neutropenia (<500 cells/mm³) (0.36%) over 1467.6 person-months of clozapine exposure. This participant recovered without sequelae after discontinuation of clozapine. Of the 249 participants with known genotypes,199 (79.9%) had the ACKR1-null genotype. Neutropenia (<1500 cells/mm³) occurred significantly more often in the ACKR1-null group (33% [65/199]) than in those with the T allele (6% (3/50); p<0.001). Fourteen (5%) patients discontinued due to adverse events. Rates of infection and fever were low and sialorrhea was the commonest side effect (N=187, 68%).

Conclusion:

To our knowledge, this is the largest prospective clozapine trial in people of African descent. Severe neutropenia was rare, despite the high prevalence (80%) of the ACKR1-null genotype. Our findings suggest that clozapine can be used safely in Black patients including those with BEN.

1. Introduction:

Schizophrenia is a chronic, disabling illness that often results in loss of productivity and poor quality of life (Correll and Schooler 2020). Approximately 30–50% of individuals with schizophrenia do not respond to currently available medications (Lally, Gaughran et al. 2016). In this subgroup of treatment-refractory patients, clozapine is the most effective antipsychotic medication, preventing hospitalization and enabling community integration (Lieberman, Saltz et al. 1991, Conley, Tamminga et al. 1999, Solanki, Singh et al. 2007, Volavka and Citrome 2011). Clozapine is the only medication that is US Food and Drug Administration (FDA)-approved to treat suicidal ideation (Meltzer, Li et al. 2003). Clozapine treatment is associated with substantial healthcare cost savings compared with treatment with other FDA-approved antipsychotic medications: $9,000/patient/year for vs. $23,000/patient/year, respectively, over 2 years (Velligan, Diamond et al. 2008).

Despite the substantial evidence of clozapine’s superior efficacy and its generic availability, clozapine is prescribed in the US at a rate considerably lower than the estimated prevalence of treatment-resistant schizophrenia (Lieberman 1998, Weissman 2002, Fayek, Flowers et al. 2003, Taylor, Young et al. 2003, Conley, Kelly et al. 2005, Morup, Kymes et al. 2020). Clozapine accounted for 11% of all US antipsychotic prescriptions in 1999, 9% in 2000, about 4% in 2008 (Meltzer 2012), and less than 5% in 2015 (Torrey, Knable et al. 2015). This substantial decline and continued low rate of use is in contrast to other countries that embrace higher use of clozapine in schizophrenia (i.e., 20% in Australia, China, England, Sweden, and Germany) (Conley, Kelly et al. 2005, Xiang, Wang et al. 2011, Bachmann, Aagaard et al. 2017).

One major barrier to clozapine use is concern over clozapine-associated neutropenia, usually defined as an absolute neutrophil count (ANC) of less than 1500 cells/mm³; severe neutropenia is defined as ANC < 500 cells/mm³. The risk of severe neutropenia is estimated to be less than 0.9% (Myles, Myles et al. 2019). Eight deaths from clozapine-associated agranulocytosis in Finland in the mid 1970’s (Crilly 2007) led to stringent FDA requirements for weekly ANC monitoring (Fayek, Flowers et al. 2003, O’Brien 2004, Conley, Kelly et al. 2005, Kelly, Dixon et al. 2006, Nielsen, Dahm et al. 2010) (Lieberman 1998, Kelly, Kreyenbuhl et al. 2007) and the Clozapine Risk Evaluation and Mitigation System (REMS) (Leung, Ehret et al. 2023), which poses additional barriers to clozapine use.

Concern over neutropenia is greater among people who self-identify as Black (Moeller, Chen et al. 1995, Kelly, Dixon et al. 2006, Pai and Vella 2012) and may contribute to their low rate of clozapine use and high rates of clozapine discontinuation compared with other populations. A retrospective cohort study of almost 400 chronic inpatients with schizophrenia treated with clozapine in Maryland, found that, compared with people who self-identified as White, those who self-identified as Black had greater improvement after six months of clozapine treatment, but were more likely to discontinue clozapine. After 2 years of treatment, their discontinuation rate was double that of their White counterparts (42% vs. 19%, p=0.041) (Kelly, Dixon et al. 2006). In a separate retrospective cohort of 1875 Maryland inpatients with schizophrenia, there were no racial differences in clozapine discontinuation because of lack of efficacy, medication nonadherence, death, non-white blood cell (WBC) hematologic adverse effects, or other side effects, while people who self-identified as Black had a significantly higher prevalence of neutropenia (5.3% vs. 2.4%, P = 0.001) (Kelly, Kreyenbuhl et al. 2007).

US laboratory neutropenia criteria are defined based on normative data from self-identified White populations (Wonodi, Oduguwa et al. 2017), whereas large population-based studies of US residents find that Black populations, compared with White populations, have lower mean WBC counts and neutrophil counts but similar lymphocyte counts. As a result, Black populations are reported to have a higher prevalence of neutropenia (4.5% vs. 0.79%) (Hsieh, Everhart et al. 2007). The “ACKR1-null” C/C genotype (SNP rs2814778) on the ACKR1 gene (previously called Duffy Antigen Receptor for Chemokines [DARC]) (McEvoy, Freudenreich et al. 1999, Procyshyn, Ihsan et al. 2001, Paz, Nails et al. 2011) is associated with a lower normative ANC range in Black populations than is seen in White populations (Nalls, Wilson et al. 2008) (Reich, Nalls et al. 2009). A lower normative range of ANC in non-White populations, not due to known medical causes of neutropenia and not associated with increased risk of infection (Merz, 2021), is termed benign ethnic neutropenia (BEN) and is not associated with repeated or severe infections or other health abnormalities associated with neutropenia (Haddy, Rana et al. 1999). BEN occurs in about 25–50% of Africans (Jeremiah, Umoh et al. 2011, Zeh, Amornkul et al. 2011, Dosoo, Kayan et al. 2012), in some ethnic groups in the Middle East, and up to 38% of certain Arab tribes (Denic, Showqi et al. 2009). In 2015 (after the beginning of this study), the Food and Drug Administration (FDA) issued new guidelines (i.e., lower thresholds) for clozapine monitoring in people with BEN. BEN was mentioned in the UK clozapine labeling earlier than in the US (Rajagopal 2005). There is little evidence that these lower ANC values are indeed “low” for this population or associated with increased risk of severe neutropenia. The pivotal trial that led to clozapine approval in the US included (Kane, Honigfeld et al. 1988) 23% Black participants, but no separate analysis by race was performed. A previous small study (N=26) from our group evaluated the safety of clozapine use in adults with neutropenia (mostly, but not all, Black). Clozapine use was associated with significantly higher mean ANC values but there were no cases of severe neutropenia (Richardson, Davis et al. 2016). However, the range of ANC variability and safety of clozapine have not been evaluated in large prospective studies in patients of African descent.

Here we present an evaluation of the safety of clozapine in a large, prospective, multinational open-label study in patients of African descent. We also examined the relationship between the ACKR1 genotypes, ANCs, and occurrence of clinically significant fluctuations in ANC during clozapine treatment. Prevalence of neutropenia and severe neutropenia were the primary endpoints.

2. Methods

2.1. Study Sites:

This study took place at 3 sites: Maryland (primarily at the Maryland Psychiatric Research Center (MPRC), University of Maryland School of Medicine, Baltimore, MD, USA and a referral network of clinics; in Nigeria (the Federal Neuropsychiatric Hospital Yaba, Lagos (FNPHY); and Howard University School of Medicine Department of Psychiatry, Washington, DC, USA.

2.2. Study Design:

This was a 6-month, open-label, single medication design with clozapine. Participants completed a baseline evaluation period and were seen weekly for 6 months.

2.3. Participants:

Inclusion criteria were a diagnosis of a major psychotic disorder eligible for clozapine treatment (schizophrenia, schizoaffective disorder, delusional disorder, or psychotic disorder NOS, by DSM-IV criteria (APA 2000) or a target symptom likely to benefit from clozapine treatment, as determined by the treating psychiatrist (e.g., suicidality, aggression, tardive dyskinesia), age 18–64 years, African ancestry (African, African-American or African-Caribbean) or other race with ANC < 2500 cells/mm³ in the previous 24 months consistent with BEN (DeFaria C and da Costa Dourado C. 2022), use of effective birth control (if a woman with reproductive potential), and ability to provide informed consent. Exclusion criteria were uncontrolled seizure disorder, current pregnancy or breastfeeding, history of myeloproliferative disorders, paralytic ileus, or a medical condition whose pathology or treatment would likely alter the presentation or treatment of schizophrenia or significantly increase the risks associated with protocol participation or affect the participant’s ability to mount an immune response. Participants with a current infection, fever (temperature > 38°C or 100.4°F), sickle cell anemia, current cancer, documented nutritional deficiencies or three successive ANC values below 500 cells//mm³ were also excluded. Participant enrollment occurred between June 2015 and June 2019, with final study participants completing the study in December 2019.

2.4. Ethical approval:

The study was conducted under an FDA-approved investigational new drug (IND) application and approved by the institutional review boards at the University of Maryland Baltimore, Howard University, FNPHY, and 6 other local research committees and hospital IRBs; monitored by a Data and Safety Monitoring Board; and registered in ClinicalTrials.gov (NCT02404155). All participants provided written informed consent and were required to score at least 10 out of 12 on the Evaluation to Sign Consent, (DeRenzo, Conley et al. 1998), which documented their capacity to provide informed consent.

2.5. Participant Recruitment and Screening:

Information about the study was disseminated to clinicians at the study sites. Participants were referred by their treatment team if eligible for clozapine treatment. All participants received at baseline a complete medical and psychiatric history, 12-lead electrocardiogram (EKG), clinical laboratory tests (including CBC and clinical chemistry panel), vital signs, and a physical examination. Some participants were included in an observation period of up to 6 weeks prior to baseline evaluations to establish natural patterns of ANC fluctuation without clozapine, if clozapine treatment did not need to be initiated immediately (data not analyzed or presented here). Participants’ diagnosis was based on the Structured Clinical Interview for DSM-IV Diagnosis (SCID-IV), administered by trained research staff) (First and Gibbon 2004) and information from the treating physician (including rationale for clozapine treatment and target symptoms). SCID results and clinical information on each subject were reviewed in a best estimate diagnostic framework. An experienced master’s degree level clinician, psychiatrist, or psychologist completed the SCIDs. Patients under court-ordered inpatient care were included in the protocol in accordance with the US 45CFR46 subpart C regulations for research with prisoners. The consent form included language regarding research participation not affecting parole or care, prisoner representatives on the reviewing IRBs, and approval of the research protocol by the US Office of Health Research Protections (OHRP) (Kearns A. 2023).

2.6. Clozapine Treatment/Dosing:

Prior to clozapine initiation, all US participants were registered in the clozapine REMS and assigned designees. Research staff assisted with the process to ensure all patients were enrolled. Clozapine in the US was supplied as a generic medication purchased for this study. In Nigeria, brand-name clozapine (Leponex^®) was provided by Novartis at a discount cost. Clozapine brand-name and generic products are bioequivalent (Yang, Yan et al. 2021). Dosing was at the discretion of the treating prescriber but guidance from the investigators suggested a slow titration and low dose increments. All clozapine prescriptions were dispensed by a research pharmacist at the study site. Clozapine adherence was monitored weekly by the study pharmacists who evaluated the pill containers and maintained medication dispensing logs. Clozapine non-adherence was defined as taking <75% of the prescribed dose during a study week or little to no clozapine detected in the plasma on more than one week. Study discontinuation occurred for participants nonadherent by either definition on at least two occasions.

To maximize the external validity of the study, participants were allowed to take adjunctive medications as clinically indicated by their treating prescriber.

2.7. Laboratory monitoring:

2.7.1. Absolute Neutrophil Count:

ANC was measured in peripheral venous blood collected weekly for six months at approximately the same time of day (9am-12pm) to avoid confounding of results by morning pseudo-neutropenia (Ahokas and Elonen 1999). All laboratories used automated cell counting technology. WBC counts at the Maryland and Howard University sites were performed by fully credentialed contract clinical laboratories (LabCorp^® and Quest Diagnostics^®) using Siemens Sysmex XN hematology analyzers (5-part differential analyzer). The Nigeria site used the Horiba Pentra ES60 (5-part automated differential). The Siemens Sysmex XP-300 (3-part differential hematology analyzer) was used for backup if reagents were unavailable in Nigeria. The coefficient of variance (COV) for Siemens Sysmex XN analyzer was <2.5% and the Horiba Pentra ES60 CV was < 3.0%. The mean variance in ANC among 60 participants whose aliquot of a single sample were analyzed by LabCorp and Quest and was within FDA standards for variance (300 cells/mm³; 9.09% difference (Kidd, Hoffman et al. 2016). Additionally, we conducted in-house replication testing on 4 samples and found a mean difference of 199.2 cells/mm³, 4.2% difference). Every site used the same protocol-specified criteria for neutropenia and response to changes in ANC.

Neutropenia was defined as an ANC < 1500 cells/mm³ and severe neutropenia as < 500 cells/mm³. An ANC below 1000 cells/mm³ triggered consultation with the study hematologist, twice weekly monitoring until two weeks after the count increased to > 1000 cells/mm³ and monitoring for signs of infection at all visits until > 1000 cells/mm³. These guidelines were modified criteria based on our FDA-approved IND.

2.7.2. Clozapine Plasma Concentrations:

Clozapine concentrations in venous plasma were assayed at least once during the clozapine titration phase (study Day 14) and when needed for clinical purposes) monthly during treatment, and at the conclusion of study participation. Blood samples were obtained prior to administration of the morning clozapine dose or 12 hours after the evening dose to assess trough concentrations. Blood samples from the Maryland and Howard University sites were assayed by fully credentialed clinical laboratories (LabCorp and Quest Diagnostics) using liquid chromatography, and tandem mass spectrometry (LC-MS/MS). The limits of detection (LOD) and quantification (LOQ) for Quest were both 4 ng/mL; the coefficient of variation (CV) was 20% For LabCorp, these were both 20 ng/mL; CV=8–9%. Blood samples in Nigeria were assayed by high performance liquid chromatography (HPLC) and calibrated using a clozapine standard (1 μg/mL in methanol). The LOD/LOQ were 30 and 50 ng/mL, respectively, CV= 21.5–25%. Future analyses and publications will examine details of plasma levels to symptoms and outcomes. They are reported here to include in the model for low ANC, however a full description of pharmacokinetic profile will be reported in future analyses and beyond the purpose of this manuscript.

2.7.3. Laboratory and Metabolic Monitoring:

A standard clinical chemistry panel (liver enzymes, BUN, creatinine, electrolytes), thyroid function test, and metabolic profile (lipids, fasting blood glucose) were done at baseline and repeated monthly. All values outside the laboratory’s reference range were evaluated by the medically accountable clinician.

2.7.4. TaqMan Genotyping:

We were characterizing the ACKR1 single nucleotide polymorphism (SNP, rs2814778). The ACKR1-null is defined as the Duffy null or homozygous FY−/− (CC) polymorphism. The comparison group were those of the, heterozygous FY+/− (CT), or homozygous FY+/+ (TT). Samples for genotyping were collected as whole blood at US sites and as blood spots (using FTADAN Elute cards [Whatman, plc, Maidstone, UK]) in Nigeria. Single-nucleotide polymorphism (SNP) genotyping was performed on genomic DNA isolated from blood using TaqMan polymerase chain reaction (PCR) technology. We used commercially available validated 5’->3’ exonuclease assay (Assay ID; C__15769614_10) for determining the genotype at the ACKR1 single nucleotide polymorphism (SNP, rs2814778) of ACKR1 gene (TaqMan SNP Genotyping Assays, Applied Biosystems Life Technologies Corporation, Carlsbad, California). All PCR was performed at the University of Maryland Center for Innovative Biomedical Resources (CIBR) on standard 96-well reaction plates, as previously described (Wonodi, Hong et al. 2009, Wonodi, Stine et al. 2011). For each PCR reaction, 2 ng of genomic DNA was used in a 5-μl reaction mixture (1.3 mM MgCl2, 200 μM dNTP mix, 0.25 μM of each primer, 5% DMSO, 1.5 U Taq polymerase). The PCR program includes denaturation of DNA (5 min at 94°C), 30 cycles of 30s at 94°C, 30s at 62°C, and 30s at 72°C. This was followed by a final extension of 10 min at 72°C. All PCR reactions were performed using the Biorad Multiplatform Thermocycler, which had 4 blocks of 384 wells (Bio-Rad Laboratories Inc., Hercules, CA, USA). PCR results were read by reporter dye fluorescence (Vic/Fam) and allelic discrimination analysis with 7900 Sequence Detection Systems (SDS) software v2.3 (Applied Biosystems) to assign genotype groups for ACKR1 polymorphism.

2.8. Fever and Infection Monitoring:

Because clozapine can raise basal body temperature by up to 1 degree F (Flanagan 2008), body temperature was measured at each visit by tympanic (infrared) or axillary thermometer. Fever was defined as a body temperature over 38.0 ° C (tympanic) or 37.5 ° C (axillary). At each visit, participants were asked about symptoms that might represent infection (sore throat, malaise, nasal congestion, rash, dysuria, hematuria, gum soreness, mouth ulcers, arthralgia, myalgia, chills, cough) and instructed to contact study staff promptly if these occurred between visits. Participants with possible infection were referred to their treating physician, with infectious disease consultation if appropriate. Evidence of infection triggered a clozapine blood concentration test, as clozapine blood concentrations can rise during infection (Clark, Warren et al. 2018).

2.9. Side Effect and Adverse Events Ratings:

Rating scales for movement disorders were performed monthly and the Side Effect Checklist was completed weekly.

1. Simpson-Angus Extrapyramidal Symptom Rating Scale (SAS) (Simpson and Angus 1970) The 11-item modified version of the SAS was used to assess extrapyramidal symptoms.

2. Abnormal Involuntary Movement Scale (AIMS): (Guy 1976) The 12-item AIMS total score was used to measure the presence of tardive dyskinesia (TD).

3. Barnes Akathisia Scale (BAS)(Barnes 1989): The 4-item BAS scale was used to assess objective and subjective components of akathisia.

4. Side Effect Checklist (SEC) (Kelly, Sullivan et al. 2015): This 25-item scale inquired about commonly occurring antipsychotic side effects, including those specifically associated with clozapine and symptoms of infection.

Side effects were defined as SEC items newly endorsed after baseline or existing at baseline and the score increasing by at least 2 points in severity after baseline. Adverse events were those spontaneously reported to staff, clinically significant changes in lab values or noticed and reported by clinicians or research staff. Serious adverse events were defined in accordance with FDA regulations with the adverse event being life-threatening or resulting in death or hospitalization.

2.10. Statistical Analyses:

We used descriptive statistics to report participant sociodemographic and clinical characteristics overall and by study site and the frequency of ANC strata: <500, <700, <1000, and <1500 cells/mm³ by ACKR1 genotype and to characterize mean time with neutropenia. Mean and median daily clozapine dose was calculated based on the highest daily dose that was maintained for at least 2 consecutive weeks for each participant. The primary outcome measure was the status of severe neutropenia (<500 cells/mm³) and neutropenia which is a binary variable defined as 1 if ANC <1500 cells/mm³. To test the effect of clozapine on longitudinally measured ANC over the 24 weeks, we used the generalized estimating equation (GEE) while adjusting covariates including time, study site, age, clozapine plasma level, clozapine dose, and ACKR1 genotype. The genotype groups are defined as CT/TT vs. CC, where T is the minor allele. In addition, the mixed effect model was used for ANC as a continuous outcome. Further, we examined the ANC levels (i.e., a continuous measure) over the first two weeks vs the mean over the remaining 23 weeks (i.e., the acute vs. long-term effects) in both genotype groups using the mixed effect model. We also calculated person-months exposure to clozapine.

Secondary outcomes included frequency of side effects from the SEC, adverse events, and serious adverse events, which were evaluated by ACKR1 genotype group using the chi-square statistic. Other secondary outcomes involving psychiatric symptoms (psychosis, depression, global psychopathology) will be reported separately.

3. Results

3.1. Participants:

We enrolled 274 participants (124 at Maryland, 26 at Howard University, 124 in Lagos, Nigeria), including 178 (68%) inpatients, of whom 91 (51%) were court ordered for inpatient treatment. Two-hundred-twenty-seven participants (82.8%) completed the full 6 months of clozapine treatment. Four were discontinued before starting clozapine and 43 participants discontinued clozapine treatment: 14 (5% of all participants) due to adverse events, 13 (by participant choice, 6 due to treatment nonadherence, 2 due to lack of therapeutic response, 7 lost to follow-up or by prescriber choice, and one by administrative error (see CONSORT flow diagram, Figure 1). One participant was discontinued due to severe neutropenia and developed full agranulocytosis (complete loss of neutrophils). Nineteen per cent (24/124) of participants were discontinued at Maryland, 4% (5/124) in Nigeria, and 54% (14/26) at Howard University (see supplementary Figure 1 for each site’s CONSORT diagram). The study generated a cumulative 1467.6 person-months of exposure to clozapine. Tables 1 and 2 present participants’ baseline sociodemographic and clinical characteristics overall and by site, respectively.

Figure 1. CONSORT Flow Diagram (Total Study Sample).

This graph depicts the CONSORT flow diagram for the overall study. Site-specific CONSORT diagrams (Maryland, Howard and Nigeria) are included in supplemental figure 1.

Table 1. Participant Sociodemographic Characteristics.

This table presents baseline sociodemographic characteristics for the participants in the study. Three sites include Maryland (Maryland Psychiatric Research Center, University of Maryland School of Medicine and affiliated clinics and referrals), Howard University and Nigeria (Federal Neuropsychiatric Hospital Yaba, Lago, Nigeria). Sex is defined as sex at birth. Tribe is included for the Nigerian site. Hispanic/Non-Hispanic is listed only for US sites. First degree relative includes parent, child or sibling. Inpatient and outpatient status is where the participant entered the study. Race is selfidentified. Non-Black participants were permitted with a baseline ANC < 2500 cells/mm³ and no evidence of other causes for neutropenia

	STUDY SITE
	TOTAL (N=274)	MARYLAND (N=124)	HOWARD (N=26)	NIGERIA (N=124)
Age (years, SD)	40.7 (11.5)	37.5 (11.9)	51.9 (7.7)	41.6 (10.2)
Sex, n (%)
Male	168 (61.3)	96 (77.4)	14 (53.9)	60 (48.4)
Female	106 (38.7)	28 (22.6)	12 (46.2)	64 (51.6)
Race, n (%)
Black	255 (93.1)	107 (86.3)	25 (96.2)	12 3(99.2)
White	6 (2.2)	6 (4.8)	0 (0.0)	0 (0.0)
Native American	1 (0.4)	0 (0.0)	0 (0.0)	1 (0.8)
Asian	1 (0.4)	1 (0.8)	0 (0.0)	0 (0.0)
Mixed/Other	11 (4.0)	10 (8.1)	1 (3.9)	0 (0.0)
Marital Status, n (%)
Married	26 (9.5)	4 (3.2)	3 (11.5)	19 (15.3)
Widowed	8 (2.9)	3 (2.4)	2 (7.7)	3 (2.4)
Divorced/Separated	30 (11.0)	10 (8.1)	4 (15.4)	16 (12.9)
Never married	210 (76.6)	107 (86.3)	17 (65.4)	86 (69.4)
Ethnicity, n (%)
Not Hispanic or Latino	137/144 (95.1)	113/118 (95.8)	24/26 (92.3)
Hispanic or Latino	7/144 (4.9)	5/118 (4.2)	2/26 (4.2)
Nigerian Tribe, n (%)
Igbo				37 (29.8)
Yoruba				76 (61.3)
Other				11 (8.9)
First Degree Relative with Schizophrenia, n (%)
Yes	33/232 (14.2)	23/86 (26.7)	6/26 (25.0)	4/124 (3.3)
No	199/232 (85.8)	63/86 (73.3)	20/26 (75.0)	120/124 (96.7)
Patient Status, n (%)
Inpatient	178 (65.0)	107 (86.3)	0 (0.0)	71 (57.3)
Outpatient	96 (35.0)	17 (13.7)	26 (100.0)	53 (42.7)

Table 2. Participant Clinical Characteristics at Baseline.

This table presents participant clinical characteristics at baseline. The diagnosis was based on the diagnosis in medical records, clinician input and confirmation by a Structured Clinical Interview for DSM-IV (SCID). While no participants were taking clozapine immediately prior to entry into the study, any prior clozapine use was recorded. Psychotropic medications being taken at baseline were listed by class. Current (at baseline) and former smoker status was obtained by self-report results on two smoking questions. Current smokers were not recruited into the study in Nigeria.

	STUDY SITE
	TOTAL (N=274)	MARYLAND (N=124)	HOWARD (N=26)	NIGERIA (N=124)
Psychiatric Diagnosis, n (%)
Schizophrenia	200 (73.0)	66 (53.0)	13 (50.0)	121 (98.0)
Schizoaffective Disorder	56 (20.4)	51 (41.0)	4 (15.0)	1 (1.0)
Bipolar Disorder	13 (4.7)	2 (2.0)	9 (35.0)	2 (2.0)
Other Diagnoses	5 (1.8)	5 (4.0)	0 (0.0)	0 (0.0)
Psychiatric Medications, n (%)
Prior Clozapine	45 (17.1)	33 (27.0)	1 (4.0)	11 (9.4)
Antipsychotic	252 (92.3)	123 (99.2)	16 (61.5)	113 (91.9)
Antidepressant	64 (23.4)	44 (35.5)	8 (30.8)	12 (9.8)
Mood Stabilizer	113 (41.4)	80 (64.5)	3 (11.5)	30 (24.4)
Anticholinergic Medications	177 (64.8)	105 (84.8)	8 (30.8)	64 (52.0)
Cigarette Smoking Status, n (%)
Current Smoker	70 (25.6)	50 (40.3)	20 (76.9)	0 (0.0)
Former Smoker	117 (42.7)	87 (70.2)	23 (88.5)	7 (5.7)
Never Smoker	87 (31.8)	13 (10.4)	3 (11.5)	117 (94)

3.2. Clozapine Dosing and Mean Plasma Concentration:

The mean (SD) highest maintained daily clozapine dose was 271.9 ± 156.2 mg (median 250mg range: 25–800 mg, N=274). This dose differed significantly by study site: Maryland 333.3 ± 168.8 mg, Howard University 230.4 ± 162.5 mg and Nigeria 218.8 ± 116.0 mg (F=19,96, df=2,854043, p<0.001). Mean plasma levels at the end of study for Maryland were 353.3 ± 205.7, 281.5 ± 575.5 at Nigeria and 229.2 ± 288.9 at Howard University.

3.3. ACKR1 Genotype and BEN Characterization:

ACKR1 genotype was available for 249 (91%) participants: 199 participants had the ACKR1-null genotype (79.9%), 39 (15.7%) CT and 11 (4.4%) TT. Thus, 50 (21.1%) participants had at least one T allele. There was a significant genotype difference by site: in Nigeria, 100% of participants were ACKR1-null and in the US 64.8% were ACKR1-null (Chi-Square=47.1, df=1, p<0.0001). (Maryland CC 82/122 67.2%, Howard ACKR1-null 10/20 50%).

We characterized the BEN phenotype by history of an ANC < 1500 cells/mm³ with no evidence of other causes (not by ACKR1-null genotyping). We obtained medical records with ANC data for 249 patients, of whom 57 (23%) had a known history of neutropenia. The majority of these 57 were ACKR-1 null (N=53, 93%). Thus, 53/199 (27%) of the ACKR-1 null group had known neutropenia as did 4/50 (8%) of the CT/TT allele group (Chi-Square=7.86, df=1, p=0.0051). The baseline ANC in the BEN phenotype was 1964.5 ± 1116.4 cells/mm³ vs 3353.4 ± 1428.2 cells/mm³ in the non-BEN phenotype (t=78.3, p<0.0001). All 7 people with low ANC who were self-identified as non-Black had T alleles.

3.4. ANC:

The mean baseline ANC differed significantly between the ACKR1-null (CC) and CT/TT groups: 2755.08 ± 1271.47 vs. 4360.58 ± 1779.13, respectively (t=35.26, df=1, p<0.0001). Figure 2 shows the weekly ANC by ACKR1 genotype. Based on the results of mixed model analysis, there were significant main effects of genotype (F=12.67; df=1, 242; p=0.0004) and time (F=2.20, df=23,4825, p=0.0008); but no significant time by genotype interaction (F=1.10; df=23, 4825; p=0.34). ANC increased after clozapine initiation in both groups, with a peak increase of 400–500 cells/mm³ occurring at 4 weeks after initiation. This significant bump in ANC was demonstrated by the significant time effect (F=2.20, df=23,4825, p=0.0008). We also examined the average expected changes in ANC after starting. The mixed model analysis revealed that the estimated combined mean from last 23 weeks was significantly higher than that of first two weeks, the difference was 328.4 ± 52.6 in CC genotype (t=6.248, df=5138.8, p<0.001) and 282.2 ± 104.9 in CT/TT genotype (t=2.69, df=5134.3, p=0.007), respectively.

Figure 2. Absolute Neutrophil Count (ANC) Over Time by Genotype.

This graph shows the mean ANC and standard error bars (standard deviation) by ACKR1 genotype over the 24 weeks. The C group is defined as the Duffy-Null, (aka ACKR1-null, CC genotype). The T group is defined here as anyone having a T allele present representing a combination of CT and TT genotype. Baseline is week zero and endpoint is week 24. There was a significant effect by genotype (F=12.67, df=1, 242, p=0.0004 and time effect (F=2.20, df=23,4825, p=0.0008) showing an increase in ANC with clozapine treatment. There was no time by genotype interaction (F=1.10, df=23, 4825, p=0.34).

Based on the GEE analysis, the ACKR1-null genotype was associated with 8-fold greater risk of neutropenia than the CT/TT genotypes (AOR 7.86, 95% CI 2.22–27.80, p=0.0014). There was no effect of clozapine dose on ANC (OR=0.999, 95% CI 0.997–1.001, p=0.405). Younger age was slightly but significantly associated with higher risk of neutropenia (OR=0.96, 95% CI 0.93–0.99, p=0.023).

At least one instance of neutropenia (ANC < 1500 cells/mm³) occurred in 75 (27.4%) participants, including 68 (27.3%) of those who were genotyped. There was significant genotype association with neutropenia prevalence: transient neutropenia occurred in 65/199 (33%) in the ACKR1-null genotype group vs. 3/50 (6%) in the CT/TT genotype group (Chi-Square=14.31, df=1, p=0.0002). Importantly, 96.5% (65/68) of the participants who developed neutropenia were ACKR-1 null. There was no significant difference in the mean (SD) frequency of neutropenia episodes between the ACKR1-null group (4.4 ± 4.4 times per participant and the CT/TT group (3.7 ± 1.5 times per participant) (t=0.27, df=66, p=0.78).

One participant (ACKR1-null genotype) had severe neutropenia (<500 cells/mm³) after 6 weeks of clozapine treatment and recovered without sequelae after discontinuation of clozapine. This participant had experienced severe neutropenia during previous treatment with clozapine. During 1467.6 person-months of cumulative clozapine exposure, severe neutropenia occurred in only the one participant. An additional twelve (4%) participants had ANCs below 1000 cells/mm³, eleven with the ACKR1-null genotype. The lowest ANC recorded in the ACKR1-null group besides the one case with severe neutropenia was 720 cells/mm³.

Four participants were treated with lithium: 3 starting at weeks 15, 16 and 21 and one being treated at baseline. For none of these participants was lithium being used to raise ANC.

3.5. Fever and Infection:

No deaths occurred during the study. Six participants (2.2%) had fever during clozapine treatment-- one for five consecutive weeks who was rapidly titrated to 400 mg in 5 weeks. The other fevers did not appear to be titration related. Twenty (7.3%) participants had an infection, with mean onset at 12.4 ± 7.8 weeks. These infections included 1 chickenpox, 1 tonsilitis, 5 urinary tract infections, 1 ear infection, 1 pneumonia, 1 cold, 1 diverticulitis, 1 face abscess, 1 bug bite, 2 skin infections, 1 respiratory virus/pleural effusion, 1 scalp wound, 3 unknown infections.

The infections resulting in clozapine discontinuation were diverticulitis, 2 unknown and respiratory infection/pleural effusion. The mean ANC at onset of infection was 4388.4 ± 2975.4 cells/mm³. There was no significant difference in infection prevalence between those with the ACKR1-null genotype (15/199, 7.54%) and those with the CT/TT genotype (4/50, 8.0%) (Chi-Square=0.012, df=1, p=0.91).

3.6. Vital Signs and Laboratory Measures:

There were significant increases in diastolic blood pressure (3.72 ± 11.1 mm Hg; F=1.87, df=24,53, p=0.006) and heart rate (9.99 ± 15.6 bpm; F=4.99, df=24,53, p<0.0001) from baseline to end of clozapine treatment, with no significant site differences. There were no changes over time in systolic blood pressure, temperature or respiratory rate. Seven people were started on beta blockers during the study. There was a significant increase in body weight over time (10.2 +/− 18.2 lbs., F=8.16, df=24,4126, p<0.0001), but no significant site difference (F=0.97, df=45, 4126, p=0.523). The mean increase in BMI (weight) from baseline to endpoint at Maryland was 1.67 ± 3.27 kg/m²(11.03 ± 20.71 pound), 0.23 ± 3.07 kg/m² (1.60 ± 22.25 pounds) for Howard and 1.62 ± 2.43 kg/m² (10.18 ± 15.40 pounds) for Nigeria. Baseline BMI in Nigeria was significantly lower than the US sites with baseline values of 24.80 ± 5.10 kg/m² at Nigeria, 30.72 ± 6.51 kg/m² at Maryland and 32.38 ± 7.94 kg/m² at Howard (t=57.89, df=2, p<0.0001).

There was no time or time by site effect on triglycerides, fasting blood glucose, or high-density lipoprotein (HDL). There was however a time effect (F=3.2, df=6,770, p=0.0041) and time by site effect (F=4.57, df=10,764, p<0.0001) for total cholesterol and low-density lipoprotein (LDL; time 3.16, df=6,781, p=0.0046, time by site F=4.32, df=10,776, p<0.0001). These effects were driven by an increase at the Howard site and a decrease from baseline at the Nigeria site.

3.7. Side Effects and Adverse Events:

Table 3 shows the prevalence of the side effects listed on the Side Effect Checklist. The commonest side effect was sialorrhea/salivation (68%). Most side effects had higher prevalence at the Maryland site, except for dry mouth, sedation and weight loss, which were more commonly seen at the Howard University site. Additional adverse events not rated on the SEC are listed in Table 4. There were 2 myocarditis cases (0.7%), 1 seizure (0.4%), 2 stroke or transient ischemic attack (0.7%), and 3 new onset diabetes (1.1%). The myocarditis cases occurred at weeks 3 and 4 at a clozapine dose of 100 mg. One participant was increased from 25 to 100 mg: the other from 50 to 100 mg. Neither participant had a fever). One stroke occurred during the screening process, so was not considered study related. I other occurred at week 14 and didn’t follow a clozapine dose increase. The seizure occurred at week 3 just after a decreased in valproate dose, which may have been the cause. The participant remained in the study. There were 25 serious adverse events, none resulting in death. These are noted in Table 4 with an asterisk. All resulted in hospitalization or were considered the potential for being life threatening.

Table 3. Side Effects Based on the 25-item Side Effects Checklist.

This table presents the findings from the Side Effect Checklist, an a priori list of 25 potential side effects associated with anti-psychotic medication. Side effects whose prevalence did not significantly differ by site are indicated with with * (p>0.05)

	STUDY SITE
	TOTAL n (%)	MARYLAND (N=124) n (%)	HOWARD (N=26) n (%)	NIGERIA (N=124) n (%)
Abdominal Pain	91/274 (33.0)	63 (51.0)	4 (15.0)	24 (19.0)
Anorexia	51/274 (19.0)	33 (27.0)	5 (19.0)	13 (10.0)
Bruising	12/274 (4.0)	11 (9.0)	0 (0.0)	1 (1.0)
Constipation	88/274 (32.0)	70 (56.0)	6 (23.0)	12 (10.0)
Diarrhea	64/274 (23.0)	43 (35.0)	6 (23.0)	15 (12.0)
Dizziness	112/274 (41.0)	65 (52.0)	11 (42.0)	36 (29.0)
Dry Mouth	65/274 (24.0)	32 (26.0)	13 (50.0)	20 (16.0)
Enuresis	70/274 (26.0)	47 (38.0)	8(31.0)	15 (12.0)
Fever*	35/274 (13.0)	20 (16.0)	0 (0.0)	15 (12.0)
Headache	86/274 (31.0)	55 (44.0)	2 (8.0)	29 (23.0)
Insomnia	53/274 (19.0)	30 (24.0)	1 (4.0)	22 (18.0)
Malaise	144/274 (53.0)	79 (64.0)	9 (35.0)	56 (45.0)
Mucosal issues*	22/274 (8.0)	11(9.0)	0 (0.0)	11 (9.0)
Nausea	77/274 (28.0)	60 (48.0)	3 (12.0)	14 (12.0)
Rash	36/274 (13.0)	23 (19.0)	1 (4.0)	12 (10.0)
Restlessness	79/274 (29.0)	46 (37.0)	7 (27.0)	26 (21.0)
Salivation	187/274 (68.0)	93 (75.0)	12 (46.0)	82 (66.0)
Sedation*	145/274 (53.0)	66 (53.0)	17 (66.0)	62 (50.0)
Sore Throat	49/274 (18.0)	42 (34.0)	1 (4.0)	6 (5.0)
Stiffness	66/274 (24.0)	44 (35.0)	6 (23.0)	16 (13.0)
Tinnitus	37/274 (14.0)	31 (25.0)	1 (4.0)	5 (4.0)
Tremors*	73/274 (27.0)	36 (29.0)	3 (12.0)	34 (27.0)
Urticaria	59/274 (22.0)	37 (30.0)	7 (27.0)	15 (12.0)
Vomiting	61/274 (22.0)	47 (38.0)	3 (12.0)	11 (9.0)
Weight Loss	29/274 (11.0)	14 (11.0)	7 (27.0)	8 (6.0)

^*Not significantly different by site (p> 0.05)

Table 4. Adverse Events Based on Spontaneous Self-Report or Staff Observation.

This table presents adverse events based on spontaneous reports by participants, clinically significant laboratory values, or by clinician and research staff report. Those designated as meeting criteria for a serious adverse event (considered life threatening or resulting in hospitalization) are marked with an asterisk.

	STUDY SITE
	TOTAL (N=274) n (%)	MARYLAND (N=124) n (%)	HOWARD (N=26) n (%)	NIGERIA (N=124) n (%)
Abnormal Ejaculation/Erectile Dysfunction	2 (0.7)	2	0	0
Chest Pain	6 (2.2)	5	1	0
Diabetes/Hyperglycemia	3 (1.1)	2	0	1
Disorientation/Confusion	2 (0.7)	2	0	0
Edema	2 (0.7)	1	0	1
Pleural Effusion	1 (0.4)*	1	0	0
Myocarditis	2 (0.4)*	2	0	0
Dehydration	3 (1.1)*	3	0	0
Unresponsive	1 (0.4)*	1	0	0
Pulmonary Embolism	1(0.4)*	0	1	0
Seizure	2 (0.7)*	1	1	0
Stroke	2 (0.7)*	1	1	0
Transient Ischemic Attack	1 (0.4)*	0	0	1
Clinical Worsening	3 (1.1)*	2	1	0
Prolapsed Rectum	1 (0.4)*	1	0	0
Diverticulitis	1 (0.4)*	0	1	0
Severe UTI	1 (0.4)*	1	0	0
Drug Fever	1 (0.4)*	1	0	0
Skin Infection/Cellulitis	2 (0.7)*	2	0	0
Pneumonia	1 (0.4)*	1	0	0
Severe Constipation	1 (0.4)*	1	0	0
Diffuse Ileus	1 (0.4)*	1	0	0

3.8. Extrapyramidal Symptoms:

Extrapyramidal side effects (EPS) assessed by the Simpson Angus Scale (SAS) decreased significantly over time (F=2.98, df=6,1117, p=0.0068) with prevalence differing by site (F=27.59, df=1,262, p<0.0001). There was no significant time by site interaction. EPS assessed by the Abnormal Involuntary Movement Scale (AIMS) also improved significantly over time (F=16.7, df-2,270, p<0.0001), but there was no site or site by time effect. Akathisia, assessed with the Barnes Akathisia Scale (BAR) significantly improved over time (F=7.6, df=2,270, p=0.00062) with a significant site effect (F=6.23, df=2,353, p=0.0022) but no significant site by time interaction.

4. Discussion:

This is the largest clinical trial of clozapine treatment in individuals of African descent. A recent meta-analysis which included 21 randomized clinical trials with clozapine found that less than 15% of the enrolled participants were Black (Siskind, McCartney et al. 2016). This is consistent with the under-representation of Black participants in clinical trials in general (Mosenifar 2007), and the underutilization of clozapine by Black patients (Moeller, Chen et al. 1995, Kelly, Dixon et al. 2006, Kelly, Kreyenbuhl et al. 2007, Pai and Vella 2012, Green, Trivedi et al. 2022). Given the high prevalence of BEN in those of African ancestry and the association of a genetic polymorphism of the ACKR-1 gene with low ANC, testing the safety of clozapine in this population is essential to improving their clinical care.

The data generated by this large, prospective safety study in participants of African descent (many with BEN) supports the safety of clozapine in this population. We observed only one case (0.36% prevalence) of severe neutropenia over 1467.6 person-months of clozapine exposure. Participants with BEN or the ACKR1-null genotype were not at higher risk for severe neutropenia. These findings should reassure prescribers about the safety of clozapine treatment in this population.

Our findings suggest that people in Sub-Saharan Africa have a high prevalence of ACKR1-null genotype and may be at high risk of BEN. We found that 64.8% of the Black patients enrolled at the US study sites had the ACKR1-null genotype, similar to prior findings (i.e., 68.4%) (Denic, Showqi et al. 2009). This group also has a significantly lower starting ANC than those without the ACKR1-null genotype. Among participants with BEN based on ANC history, 93% are ACKR1-null genotype.

Our data have several other implications for clinical care: First, clozapine initiation is associated with a significant increase in mean ANC values up to 700–800 cells/mm³ over the first month of treatment, remaining slightly elevated thereafter, consistent with previous studies in the US (Richardson, Davis et al. 2016) and in Nigeria (Wonodi I. 2017). Second, about one-quarter of Black patients will have an ANC below 1500 cells/mm³ during the first 6 months of clozapine treatment, largely among those with the ACKR1-null genotype. Neutropenia is transient and occurs randomly with a frequency of about 4 times within a 24-week period of time, often with no consistent pattern. Third, the ACKR1-null genotype is the most significant predictor of neutropenia, conferring 8 times the risk of neutropenia as the other alleles. Fourth, the risk of severe neutropenia is quite low, even in individuals with BEN. We had only one case during 1467.6 person-months of clozapine exposure. Fifth, transient ANC fluctuations below 1500 cells/mm³ occur in about one-third of Black persons and rarely require discontinuation of clozapine. Sixth, clozapine can be safely used in patients with BEN. We observed only 2 cases of myocarditis, 2 of stroke or transient ischemic attack, and one seizure. Only 5% of patients were discontinued for an adverse event.

More adverse effects were reported at US sites than at the Nigeria site. This could be due to the lower rates of adverse events (AE) reporting in Africa due to cultural differences or higher prevalence of concomitant medications in the US, among other factors. In particular, constipation was significantly more prevalent in the US than in Nigeria, although discontinuation rates due to constipation were similar. This difference may be due to dietary differences; the typical US diet contains low levels of fiber which may be less than the typical Nigerian diet (Nutrition 2021). Those who are at highest risk of developing neutropenia are younger. A one-year difference in age was associated with a 4% reduced risk of developing neutropenia. A prior study found that ANC in children and adolescents with BEN was lower than the ANC typically found in adults with BEN; ANC increased as they got older. (Ortiz, Meier et al. 2016).The risk of severe clozapine-associated neutropenia and of death appears to be greater at older ages (Ibáñez, Vidal et al. 2005, Ingimarsson, MacCabe et al. 2016).

This study has several strengths. This is the first prospective clozapine study involving Black patients. It has a relatively large sample size for a study of this type, multinational study sites, high study completion rate and low rate of missing genotype data (9% missing). We used centralized training procedures and coordination to ensure data reliability across sites, especially for the EPS ratings and AE evaluations. The study teams met weekly throughout the entire study period by videocalls, telephone and instant messaging (Slack^™) to troubleshoot, report and discuss. Weekly collection of ANC and side effect data ensured that significant or transient fluctuations were unlikely to be missed.

This study has several imitations. First, it was impossible to ensure or verify complete consistency of laboratory procedures, data collection and reporting, across all 3 sites. Logistical barriers made it impossible to cross-validate blood testing procedures due to changes in time and temperature in transatlantic transport. However, all sites used validated laboratory methods, there was extensive cross-training of staff across sites, and all sites followed the same study protocol. We may have over- or under-estimated nonadherence, as we used weekly pill counts and medication dispensing logs, with no automated technology to verify participant handling of medication and we had plasma concentrations. Additionally, the primary outcome of severe neutropenia was reliably measured using validated laboratory technology with low variability at all sites. Finally, the open-label study design may have allowed bias by both research staff and participants.

The data from this study support the 2015 change made by the FDA to lower the WBC/ANC thresholds for discontinuing clozapine treatment in patients with BEN. This change should eliminate the practice of using lithium to maintain ANC within “normal” ranges (Nykiel, Henderson et al. 2010). Experts warn against this practice due to higher rates of clozapine discontinuation and under-prescribing of clozapine with lithium treatment (Verdoux, Quiles et al. 2023). Without this 2015 change for lower ANC thresholds in patients who may have BEN, over 25% of our participants would have had clozapine treatment interrupted and unnecessary daily ANC monitoring instituted. We believe that further changes are appropriate, such as making the clozapine REMS an educational program or eliminating the safe use requirements of clozapine REMS, which too often serve as a barrier to patients receiving optimal care. During the COVID-19 pandemic, many patients received clozapine treatment outside the clozapine REMS guidelines (Siskind, Honer et al. 2020, Wadoo, Karim et al. 2021) without evidence of higher mortality or increased prevalence of severe neutropenia and despite COVID-19 infection being associated with decreased ANC (Bonaccorso, Ricciardi et al. 2021). Additionally, countries with less stringent monitoring do not have higher rates of severe neutropenia and adverse ANC outcomes (Ingimarsson, MacCabe et al. 2016).Other medications with significant risks of neutropenia, such as cancer chemotherapy agents, do not require a REMS with ANC monitoring (Leung, Ehret et al. 2023).

Our finding of a significant association between clozapine-associated neutropenia and the CC allele of the ACKR1 gene (ACKR1-null ) is consistent with prior studies (Nalls, Wilson et al. 2008, Reich, Nalls et al. 2009) and contributes to the discussion over genotyping patients considered for clozapine treatment, as has been recommended by others (Legge, Pardinas et al. 2019, Meyer 2020). We find that clozapine can be used safely in the Black population and this population does not have a higher rate of severe neutropenia. We expect about 27% of those with the ACKR1-null genotype to develop transient neutropenia (by current ranges) but not progress to severe neutropenia or agranulocytosis (loss of all neutrophils). We also expect those with the CT/TT alleles to rarely experience neutropenia.

A rigid interpretation or routine requirement for genotyping could reduce the use of the lower ANC criteria for BEN patients and subsequently reduce clozapine utilization. Rather than employing routine genotyping, a more reasonable approach is to generally accept that a Black patient with an episode of neutropenia is likely one with the ACKR1-null and potentially a BEN patient (95.5% of all neutropenia cases in our study were ACKR-1-null genotype and 93.0% of those with historical information on low ANC were ACKR-1-null genotype). There remains confusion among clinicians on how to designate their patient with BEN for the clozapine REMS, likely leading to underuse of the BEN algorithm. We suggest that Black race and an ANC < 1500 cells/mm³ is most likely an ACKR-1 null patient and can be designated BEN in the clozapine REMS system. The more challenging designation is initiating a Black patient on clozapine and not having full ANC history to make the determination of having BEN in advance of clozapine treatment. We found that 27% of those with ACKR-1-null genotype developed neutropenia during 6 months of prospective clozapine treatment and also about 27% had a historical data documenting past neutropenia. If no ANC history is available which may suggest neutropenia, and a patient is genotyped and found to be ACKR-1 null the rate of neutropenia incidence is 33%. TT allele patients are unlikely to have any episodes of neutropenia. However, if neutropenia develops at any point in a Black patient once starting clozapine, with no other known causes, clinicians should immediately change their designation of the patient in the clozapine REMS to BEN to allow for the less stringent monitoring guidelines.

Also, the terminology of BEN perpetuates a misnomer that lower ANC is a “condition” or abnormality (Merz and Achebe 2021). A better terminology may be benign neutropenia or familial or hereditary neutropenia (Andreou, Jayaram et al. 2023). We chose to use BEN here as it remains the prevailing terminology during the design and conduct of the study and in FDA guidelines. With regard to clozapine REMS and strict ANC monitoring, a broader discussion of the utility of the restrictive monitoring system is in order as rates of myocarditis, constipation and other side effects as well as mortality may be more common than severe neutropenia (de Leon, Arrojo-Romero et al. 2023).The risk of suicidality from disruption of clozapine treatment may be more harmful than the risk of dying from agranulocytosis (loss of neutrophils and resulting infection). The clozapine REMS may adversely impact access to treatment (Leung, Ehret et al. 2023) and at the very least it should be adapted to be less restrictive for ANC monitoring. Our data provide evidence for the safety of clozapine and the rare occurrence of severe neutropenia and agranulocytosis.

In summary, clozapine-associated severe neutropenia is rare in patients theoretically considered at high risk because of BEN or of African descent. Every effort should be made to reduce barriers to using clozapine, particularly in Black patients.

Supplementary Material

1

Supplementary Figure 1. This graphic shows the participant CONSORT by all 3 sites, Maryland, Nigeria and Howard University.