Нейтропения связана с терапией, истощающей клетки B, при множественных склерозах и нейромелите Оптика Спектрум.

Introduction

In the past 2 decades, the prevalence of B cell–depleting therapies (BCDTs), namely anti-CD20 and anti-CD19 monoclonal antibodies, for the treatment of neuroinflammatory disease has expanded remarkably. 1 These BCDTs have dramatically reshaped the outcomes for patients living with multiple sclerosis (MS) 2 and neuromyelitis optica spectrum disorder (NMOSD). 3 Evidence from both clinical trials and observational studies suggests that neutropenia remains a rare but known complication of BCDT, despite the limited expression of CD20 antigens on the surface of neutrophils. 4-7 It is hypothesized that the neutropenia is a result of a combination of neutrophil removal in the context of imbalance of granulopoiesis, possibly from concomitant infection or (monoclonal) antibody-dependent cellular cytotoxicity. 8

In clinical trials treating neuroinflammatory conditions with BCDT, proportions of patients experiencing neutropenia have been estimated to be as high as 4% (ocrelizumab 9 ), 1% (rituximab 10 ), and 11% (ublituxumab 11 ), respectively. Incidence rates are not always calculated for neutropenia specifically, but are often combined with lymphopenia. When incidence rates are reported for neutropenia, they are estimated to be 0.3 and 0.4 cases per 100 person-years for ofatumumab and ocrelizumab, respectively. 10-13 Neutropenia did not result in serious infections requiring termination of medication in any of the clinical trials.

Beyond incidence, little is known about the detailed natural history of patients experiencing neutropenia, particularly relapse rates and real-world treatment changes. Real-world monitoring provides a unique opportunity to reveal rare adverse events and confirm long-term inferences from clinical trials. This study estimates the real-world incidence of neutropenia among patients with neuroinflammatory conditions treated with BCDT and characterizes the clinical course and longitudinal outcomes.

Methods

This study included both a cohort study and a case series. The retrospective cohort study was completed by querying the electronic health record (EHR) for potential cases of incident neutropenia. Each case was then followed by detailed chart review, and these results are presented as a case series (eFigure 1). Patients were identified for the cohort study if they had at least one visit in the University of California San Francisco (UCSF) MS and Neuroinflammation Clinic; a clinical diagnosis of MS, NMOSD, or clinically isolated syndrome (CIS); at least one nonmissing demographic characteristic; and a valid start and stop date for any BCDT (defined as any of the following medications: ocrelizumab, rituximab, inebilizumab-cdon, ofatumumab, ublituximab-xiiy).

Case Definition

Cases of neutropenia were identified with the same criteria as given above with either a mention of “neutropenia” in the EHR or the additional criteria of an absolute neutrophil count (ANC) value below the reference range while on BCDT. Reference ranges for ANC varied between centers because of differences in measurement methods, population demographics, and calibration standards. Patients are considered neutropenic if their ANC falls below the reference range at the laboratory where the sample is processed. All ANC values were standardized to cells per microliter (cells/µL). In our population, the reference ranges at the processing laboratories were 1,500–7,800, 1800–6,800, or 2000–8,000 cells per microliter. When a reference range was not provided, a normal ANC was defined as 1800 cells/µL.

Clinical Outcomes

From the EHR, the following measures were manually extracted: diagnosis, disease onset, BCDT start and stop dates, ANC values and dates, body mass index ([BMI], kg/m 2 ), and when available, infections and course, hematologist note, bone marrow biopsy, and hospitalization summaries.

Statistical Analysis

Clinical and demographic characteristics were summarized using descriptive statistics. Incidence rates for neutropenia were calculated by dividing the number of first occurrences of neutropenia by total person-time on BCDT. Total person-time on BCDT was calculated by EHR query to pull all instances of medication records for inebilizumab-cdon, ocrelizumab, ofatumumab, rituximab, and ublituximab-xiiy with valid start dates. If the end date was missing, the current date was imputed. The records were then filtered and cleaned by participant to ensure clean temporal trajectories by aligning subsequent disease-modifying therapies (DMTs) and remove any overlap/overestimation. This cleaned dataset was used to calculate person-years on the specified DMTs. Confidence intervals were estimated with exact binomial calculations. Kaplan-Meier cumulative incidence curves estimated median time to recovery of neutropenia among patients experiencing neutropenia, who did not receive granulocyte-colony stimulating factor, filgrastim. Incidence rate ratios were calculated using negative binomial regression models, incorporating BCDT as a time-varying covariate offset by duration of time on each BCDT. Crude and adjusted models were estimated; adjusted models included covariates for age at start of BCDT, race, BMI, duration of disease at diagnosis, neurologic disease indication for BCDT, number of previous DMTs before BCDT, and cumulative time on BCDT. All statistics were performed using Stata/MP 18.0 for Mac.

Results

Incidence of Neutropenia

A total of 1,825 patients with MS, CIS, or NMOSD treated with BCDT were identified, reflecting 6,009 person-years of BCDT (ocrelizumab: 6,692 person-years, rituximab: 1,665 person-years, ofatumumab: 553 person-years, ublituximab: 37 person-years, inebilizumab: 61 person-years). Among these, 37 patients (2%) developed neutropenia, with an estimated incidence rate of 0.62 (95% CI 0.45–0.85) per 100 person-years. An additional 16 patients were flagged as experiencing neutropenia in the record extraction step but, on manual review, were found either to have neutropenia that preceded treatment with BCDT (n = 12) or to be a spurious diagnosis with records reflecting no evidence of neutropenia through chart review (n = 4). One further patient was identified through patient care of the authors and included. Most of the patients were female (69%); 625 identified as White, 9% as Black, and 8% as Asian. The median age at disease onset was 32 years and at start of DMT was 44 years. Most patients were diagnosed with MS (95%); however, 100 patients were diagnosed with NMOSD (5%). There was no statistically significant difference in age at onset, age at DMT start, duration of disease at onset, race, or ethnicity between patients who developed neutropenia and those who did not. There was no statistically significant difference in first choice of BCDT and last choice of BCDT among patients who developed neutropenia compared to those who did not. There were no differences in neurologic disease indication for BCDT, body mass index (BMI), or Expanded Disability Status Scale (EDSS) at diagnosis. Patients who developed neutropenia had a shorter duration of disease at diagnosis, compared with those who did not develop neutropenia ( p value 0.04), and a higher number of BCDTs ( p value < 0.001), longer total duration of BCDT (for the entire study period, p value <0.001), and longer duration of follow-up ( p value 0.01). Full summary of demographic and clinical characteristics of the cohort is given in Table 1 .

Table 1 Demographic and Clinical Characteristics Among Patients With and Without Neutropenia Demographic characteristics Without neutropenia (n = 1,788) Neutropenia (n = 37) p Value Total (n = 1,825) Age at disease onset 32.3 (25.2–41.3) 31.4 (28.1–46.5) 0.12 32.3 (25.2–41.3) Age at DMT start 43.8 (34.5–54.5) 38.0 (30.9–52.3) 0.43 43.7 (34.4–54.5) Sex N (%) 0.97  Female 1,236 (69.1) 25 (67.6) 1,261 (69.1)  Male 549 (30.7) 11 (29.7) 560 (30.7)  Nonbinary 3 (0.2) 0 (0.0) 3 (0.2) Race N (%) 0.9  White 1,104 (61.7) 24 (64.9) 1,128 (61.8)  Black or African American 167 (9.3) 3 (8.1) 170 (9.3)  Asian 132 (7.4) 4 (10.8) 136 (7.5)  Native American 32 (1.8) 0 (0.0) 32 (1.8)  Other 230 (12.9) 3 (8.1) 233 (12.8)  Decline 116 (6.5) 2 (5.4) 118 (6.5)  Unknown 7 (0.4) 0 (0.0) 7 (0.4) Ethnicity N (%) 0.49  Declined 124 (6.9) 1 (2.7) 125 (6.8)  Hispanic or Latino 298 (16.7) 3 (8.1) 301 (16.5)  Not Hispanic or Latino 1,355 (75.8) 32 (86.5) 1,387 (76.0)  Unknown 8 (0.4) 0 (0.0) 8 (0.4)  Unknown/declined 3 (0.2) 0 (0.0) 3 (0.2) Clinical characteristics  Neurologic disease N (%) 0.24  MS, subtype not specified 110 (6.2) 3 (8.1) 113 (6.2)  NMOSD 95 (5.3) 5 (13.5) 100 (5.5)  Primary progressive MS 173 (9.7) 2 (5.4) 175 (9.6)  Progressive relapsing MS 18 (1.0) 0 (0.0) 18 (1.0)  Relapsing remitting MS 1,359 (76.0) 26 (70.3) 1,385 (75.9)  Secondary progressive MS 33 (1.8) 0 (0.0) 33 (1.8) BMI 26.1 (22.5–31.2) 24.7 (22.2–27.6) 0.23 26.0 (22.5–31.2) Duration of disease 13.1 (7.7–20.5) 9.7 (6.8–14.6) 0.04 13.0 (7.7–20.4) Number of BCDTs 1.0 (1.0–1.0) 1.0 (1.0–2.0) <0.001 1.0 (1.0–1.0) Number of other DMTs before BCDT 1.0 (0.0–1.0) 0.0 (0.0–1.0) 0.62 0.0 (0.0–1.0) EDSS score at diagnosis 2.5 (1.5–4.0) 3.0 (2.0–3.5) 0.47 2.5 (1.5–4.0) First BCDT N (%) 0.22  Inebilizumab 20 (1.1) 1 (2.7) 21 (1.2)  Ocrelizumab 1,126 (63.0) 20 (54.1) 1,146 (62.8)  Ofatumumab 185 (10.3) 1 (2.7) 186 (10.2)  Rituximab 411 (23.0) 13 (35.1) 424 (23.2)  Ublituximab 46 (2.6) 1 (2.7) 47 (2.6) Last BCDT N (%) 0.48  Inebilizumab 35 (2.0) 2 (5.4) 37 (2.0)  Ocrelizumab 1,099 (61.5) 20 (54.1) 1,119 (61.3)  Ofatumumab 302 (16.9) 5 (13.5) 307 (16.8)  Rituximab 290 (16.2) 8 (21.6) 298 (16.3)  Ublituximab 62 (3.5) 1 (2.7) 63 (3.5) Duration of BCDT 827.5 (249.0–1799.5) 2,164.0 (987.0–2,643.0) <0.001 845.0 (251.0–1839.0) Duration of follow-up 4.4 (2.1–6.6) 6.4 (4.0–7.6) 0.01 4.4 (2.1–6.6) Abbreviations: BCDT = B cell–depleting therapy; BMI = body mass index; DMT = disease-modifying therapy; MS = multiple sclerosis; NMOSD = neuromyelitis optica spectrum disorder

Demographic and Clinical Characteristics Among Patients With and Without Neutropenia

Abbreviations: BCDT = B cell–depleting therapy; BMI = body mass index; DMT = disease-modifying therapy; MS = multiple sclerosis; NMOSD = neuromyelitis optica spectrum disorder

At the onset of neutropenia, 22 patients (60%) were being treated with ocrelizumab, 11 (30%) with rituximab, 2 (5%) with inebilizumab, 1 (3%) with ublituximab, and 1 (3%) with ofatumumab. Drug-specific incidence rates were 0.60 (95% CI 0.37–0.87), 0.66 (0.32–1.12), 3.28 (0.05–9.63), 2.70 (2.11–11.24), and 0.18 (0.14, 0.78) for ocrelizumab, rituximab, inebilizumab, ublituximab, and ofatumumab, respectively (eFigure 2). Weighted by time on specific BCDT, the pooled incidence rate was estimated at 0.25 (0.10–0.45) per 100 person-years. The median time to neutropenia from starting any BCDT was 1.3 years (IQR 0.3–2.0), and the median (IQR) time from last infusion of current BCDT was 4 months (1–6). There was no statistically significant difference in seasonal incidence of neutropenia (eFigure 3).

Negative binomial regressions estimated incidence rate ratios with rituximab as the reference group. In crude analyses, inebilizumab showed a statistically significantly higher incidence rate compared with rituximab, but the statistical significance was lost in adjusted models. In the adjusted model, as disease duration at diagnosis increased, there was a decreased incidence rate compared with the first quantile of disease duration. Otherwise, there were no statistically significant differences across BCDTs, and no other statistically significant associations were identified including for age at start of BCDT, race, BMI, neurologic disease indication for BCDT, number of previous DMTs before BCDT, and cumulative time on BCDT ( Table 2 and eTable 1 and eFigure 4).

Table 2 Incidence Rate Ratios From Crude and Adjusted Negative Binomial Models Crude (n = 1,825) Adjusted (n = 1,825) B cell–depleting therapy (BCDT)  Rituximab Reference Reference  Inebilizumab 4.96 a (1.16, 21.10) 3.10 
(0.57, 17.02)  Ocrelizumab 0.92 
(0.45, 1.88) 0.91 
(0.42, 1.95)  Ofatumumab 0.27 
(0.04, 2.10) 0.23 
(0.03, 1.77)  Ublituximab 4.36 
(0.58, 32.64) 3.23
 (0.37, 28.20) Age at BCDT start b 
 0.84 
(0.61, 1.17) Race 
 0.87 
(0.69, 1.09) BMI 
 0.97 
(0.91, 1.04) Duration of disease at diagnosis b 
 0.64 a 
 (0.45, 0.91) Neurologic disease 
 0.87
 (0.67, 1.12) Number of other DMT therapies before BCDT 
 1.11
 (0.72, 1.70) Cumulative time (d) of BCDT therapy b 
 0.84
 (0.52, 1.36) Abbreviations: BMI = body mass index; DMT = disease-modifying therapy. 95% CIs in brackets. a p < 0.05. b Modeled as quantiles.

Incidence Rate Ratios From Crude and Adjusted Negative Binomial Models

Abbreviations: BMI = body mass index; DMT = disease-modifying therapy.

95% CIs in brackets.

p < 0.05.

Modeled as quantiles.

Clinical Course of Neutropenia

At neutropenia ascertainment, 12 patients (32%) were asymptomatic (ascertained by routine laboratory monitoring), 14 (37%) developed neutropenic fever, and 6 (16%) reported a preceding infectious prodrome. Within the month of neutropenia diagnosis, 12 patients (71%) had a confirmed infection: 6 were viral infections, 4 were bacterial, and 1 was fungal. The median nadir ANC was 390 cells/µL (IQR: 40–960); 13 patients (35%) sustained a nadir ANC below 100 cells/µL, among whom 6 had a nadir ANC of zero cells/µL. All patients experiencing nadir ANC less than or equal to 100 cells/µL were receiving either ocrelizumab or rituximab. The full summary of clinical characteristics of neutropenia is given in Table 3 , and comparisons of demographic and clinical characteristics between BCDTs can be found in eTable 2.

Table 3 Neutropenia Course (n = 37) Median (IQR) or N (%) BCDT at time of neutropenia  Inebilizumab 2 (5)  Ocrelizumab 22 (60)  Ofatumumab 1 (3)  Rituximab 11 (30)  Ublituximab 1 (3) Age at nadir neutropenia (y) 40 (34–58) Fever 14 (37) Asymptomatic 12 (32) Preceding infectious prodrome N (%) 6 (16) Coinciding infectious symptoms/syndrome 17 (46) Confirmed simultaneous infection N (%) 12 (71) a  Cellulitis 1 (6)  COVID-19 PNA 1 (6)  COVID-19 URI 1 (6)  CMV viremia and E. coli UTI 1 (6)  CMV colitis 1 (6)   E. coli UTI 1 (6)  Fungal pharyngitis/esophagitis 2 (12)  Gram-positive PNA 1 (6)  HSV mucositis; Gram-negative meningitis 1 (6)  Rhinovirus URI 1 (6)  Rhinovirus PNA 1 (6) Emergency department visit N (%) 21 (57) Hospitalization N (%) 20 (54) Bone marrow biopsy N (%) 5 (13) Filgrastim administered N (%) 5 (13) Contributing concomitant medications N (%)  None 34 (92)  IVIG 1 (6)  Mycophenolate mofetil 1 (6)  Oxcarbazepine 1 (6) Time from neurologic disease onset to nadir (mo) 87 (52–138) Time from most recent infusion to nadir (mo) 4 (1–6)  Inebilizumab 4 (2–7)  Ocrelizumab 5 (4–7)  Ofatumumab 0.3 (0.3–0.3)  Rituximab 0.6 (0–2)  Ublituximab Time from last normal WBC to nadir (mo) 4 (2–7) ANC (cells x 1,00^6/L) nadir median (IQR) 390 (40–960)  ≤100 c 13 (35)  101–500 5 (13)  501–1,000 10 (27)  1,001–2,500 7 (19) Adverse event grade d  Grade 1 14 (38)  Grade 2 2 (5)  Grade 3 21 (57) Time from nadir to ANC >500 (d) 9 (4–29) Time from nadir to ANC >2,500 (d) 34 (7–201) Restarted on B cell–depleting therapy 30 (81) Restarted the same therapy 28 (93) Restarted a different therapy 2 (7) Ocrelizumab to rituximab  Time from nadir to restarting B cell–depleting therapy (mo) b 5 (1–11)  Recurrence of neutropenia at last follow-up visit 13 (35)  Time from nadir to last follow-up visit (mo) 21 (3–70) Abbreviations: ANC = absolute neutrophil count; BCDT = B cell–depleting therapy, PNA = Pneumonia, URI = Upper Respiratory Infection, CMV = cytomegalovirus, UTI= urinary tract infection, HSV = Herpes Simplex Virus, IVIG Intravenous Imunoglobulin, WBC = white blood cell a Among patients with coinciding infectious symptoms/syndromes. b Three patients restarted B cell–depleting therapy before nadir and after the first diagnosis of neutropenia. c Six patients with nadir ANC of 0. d Based on definitions given in Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. 14

Neutropenia Course (n = 37)

Abbreviations: ANC = absolute neutrophil count; BCDT = B cell–depleting therapy, PNA = Pneumonia, URI = Upper Respiratory Infection, CMV = cytomegalovirus, UTI= urinary tract infection, HSV = Herpes Simplex Virus, IVIG Intravenous Imunoglobulin, WBC = white blood cell

Among patients with coinciding infectious symptoms/syndromes.

Three patients restarted B cell–depleting therapy before nadir and after the first diagnosis of neutropenia.

Six patients with nadir ANC of 0.

d Based on definitions given in Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. 14

Twenty patients (54%) required hospitalization, among whom 4 required admission to an intensive care unit. Patients who were hospitalized had a lower median ANC compared with those who were not (100 vs 700 cells/µL, p value not significant.) There was no difference in type of BCDT, sex, age at nadir, duration of neurologic disease at nadir, or BMI between those patients who required hospitalization and those who did not (eTable 3). Using the Common Terminology Criteria for Adverse Events Version 5.0, 21 patients (57%) experienced Grade 3 neutropenia, 3 (12.5%) patients experienced Grade 2 neutropenia, and 12 (37.5%) experienced Grade 1 neutropenia. Grade 3 adverse event is defined as “severe or medically significant but not immediately life-threatening; hospitalization or prolongation of hospitalization indicated; disabling; limiting self-care ADL.” Grade 2 is defined as “moderate; minimal, local, or noninvasive intervention indicated; limiting ag-appropriate instrumental ADL.” Grade 1 is defined as “mild; asymptomatic or mild symptoms; clinical or diagnostic observations only; intervention not indicated.” 14 There was no difference in Grade 3 severity (neutropenia requiring hospitalization) by calendar month of diagnosis (eFigure 5).

Time to Recovery of ANC

All patients recovered to normal ANCs. Among the 26 patients who did not receive filgrastim and for whom quantitative follow-up ANC values were available, Kaplan-Meier cumulative incidence plots estimated the median overall time to ANC recovery to normal to be of 11 days (95% CI 7–26). Comparison between rituximab and ocrelizumab, the 2 therapies with more than 1 patient included, revealed shorter median time to recovery among patients receiving rituximab (11 days) than those with ocrelizumab (28 days) (log-rank p value 0.836) ( Figure and eFigure 6).

Days to Absolute Neutrophil Count (ANC) Recovery Among Patients Not Receiving Filgrastim

Normal neutrophil count defined as greater than 1,800 cells/dL.

Subsequent Treatment Decisions

After recovery, 30 patients (81%) continued BCDT: 30 (93%) restarted their previous BCDT while 2 (7%) changed within the anti-CD20 class, and at the time of submission, 5 subsequent treatment decisions have not been finalized. The median (IQR) time to redosing BCDT was 5 months. Of note, one patient with asymptomatic, mild neutropenia (ANC <1,000 cells/µL) was redosed on ocrelizumab after the initial neutropenia was detected but before nadir; their course remained asymptomatic and did not require hospitalization.

Clinical Course of Recurrent Neutropenia

Among the 37 patients, 13 experienced recurrence (10 female, 10 White). The incidence of recurrent neutropenia was estimated at 0.22 (95% CI 0.13–0.40) per 100 person-years among all patients treated with BCDT.

The mean time to recurrence from initial neutropenia was 251 (SD: 355) days. The mean nadir ANC for recurrence was 1,030 cells/µL; 2 patients demonstrated nadirs less than 500 cells/µL, and zero patients experienced a nadir of 0 cells/µL. Between initial and recurrent neutropenia episodes, all the patients remained on the same BCDT; however, 3 patients did not receive additional doses of BCDT between initial and recurrent episodes of neutropenia. Among patients with recurrence, 7 patients were receiving ocrelizumab, 5 were receiving rituximab, and 1 was receiving ublituximab-xiiy.

Four of the 13 patients were asymptomatic at both the initial neutropenia diagnosis and recurrence, including lack of an infectious prodrome in preceding weeks. Two patients required hospitalization with recurrence, and 3 received filgrastim as guided by the consultant hematologist. Five underwent bone marrow biopsy; none of these biopsies revealed an additional hematologic or oncologic process to explain the neutropenia. Three patients had more than 1 recurrence, and all had fluctuating low levels of neutropenia both with and without symptoms. A complete description of recurrence clinical courses and comparisons with the initial course can be found in eTables 3 and 4, and full reports of bone marrow biopsy can be found in eTable 5.

Discussion

This comprehensive analysis from a large clinical cohort of over 6,000 person-years on BDCTs estimated the incidence rate for neutropenia of 0.62 per 100 person-years. Although still rare, this is higher than what is reported in clinical trials for these agents, 10,11,13,15 and in many patients, the course was recurrent and not benign.

The severity of neutropenia in this cohort differed from that in clinical trials, where most of the neutropenia cases were categorized as Grade 1 (i.e., mild; asymptomatic or mild symptoms; clinical or diagnostic observations only; intervention not indicated). 14 Indeed, while many in this cohort had mild courses (35% were asymptomatic), most required hospitalization and severe neutropenic fevers requiring intensive care admission were observed. The ANC was <100 cells/µL at nadir in one-third of patients and 0 cells/µL in 16% of patients. All patients recovered to normal ANCs, with the median time to recovery of 11 days. While a previous analysis suggested that neutropenia associated with ocrelizumab correlated with younger age, male sex, and lower BMI, our study did not identify any clinical or demographic risk factors of neutropenia in multivariable negative binomial regression models. 16

While BCDTs may differ in structure, in our population, there were no statistical differences in incidence rates across the BCDTs evaluated. This is in contrast to MS clinical trials, where ofatumumab was reported to confer less risk of neutropenia. 15 With its fully human amino acid sequence and subcutaneous route of administration, ofatumumab could be theorized to pose a lower risk of neutropenia. This was not the case in this study. Ofatumumab given intravenously at higher doses for other indications has shown comparable rates of neutropenia with other BCDTs administered intravenously. 17 Thus, it will be of interest to compare neutropenia occurrence in patients who receive the recently approved subcutaneous formulation of ocrelizumab vs those who received IV administration in MS.

Most of the patients studied restarted BCDT after resolution of neutropenia, including in most cases their initial agent, without recurrence. However, in this case series substantially larger than previous reports, recurrence was identified in over one-third of patients, with an incidence rate of recurrence of 7.5 per 100 person-years among those with 1 episode of neutropenia. Half of these recurrences were asymptomatic in both initial and recurrent neutropenia episodes, but several symptomatic recurrences required hospitalization. Because none of the bone marrow biopsies revealed additional hematologic or oncologic contributors to the neutropenia, reserving this invasive procedure for patients whose ANCs do not improve promptly or in whom other underlying processes are suspected seems reasonable.

This study, describing the granular clinical course for all episodes of neutropenia identified in a cohort reflecting 6,000 person-years of follow-up, had several limitations inherent to single-center, retrospective studies using EHR data. These include incomplete ascertainment because not all neutropenia events occurred within the health system queried (e.g., patients who live outside catchment and present to local hospital systems), as well as the potential for systematic bias from monitoring practices, which limit the ability to detect asymptomatic neutropenia. Another limitation is the lack of granular detail on dosing frequencies (i.e., extended interval dosing). Finally, our sample size limits the power to detect variation by indication or between different medications within the B cell–depleting therapies.

Overall, our study of a large, diverse cohort of individuals with MS and NMOSD with 6,009 person-years on BCDT does continue to suggest that neutropenia is a rare complication of BCDT for the management neuroinflammatory disease. In many cases neutropenia associated with BCDT can be mild and self-limited. While lower ANC values were found among hospitalized patients, the clinical chart review highlighted the personalized clinical decisions that had resulted in these hospitalizations. Factors considered when deciding on hospitalization included comorbidities, symptomatology, and degree of neutropenia. However, because more severe cases were identified (16% with ANC of 0), and there was an unexpectedly high rate of recurrence, often severe, these findings underlie the importance of more detailed discussion of risks with patients and consideration of possible risk mitigation strategies with a subcutaneous route of admission or extended interval dosing. 8