Baseline neutrophil count, benign ethnic neutropenia risk, and clozapine haematological monitoring: A multi-ethnic retrospective cohort study from the United Arab Emirates

Abstract

Background Concerns about haematological toxicity and absolute neutrophil count (ANC) monitoring may limit clozapine use, particularly in multi-ethnic populations where benign ethnic neutropenia (BEN) and lower baseline ANC are common. We characterised baseline ANC distribution, nationality-derived BEN-risk, post-initiation ANC outcomes, and haematological monitoring patterns in a multi-ethnic United Arab Emirates clozapine cohort. Methods We conducted a retrospective cohort study at a tertiary psychiatric hospital in the United Arab Emirates. Adults with documented clozapine administration and at least one post-initiation ANC between 1 January 2018 and 31 July 2024 were included. Nationality-derived BEN-risk, baseline ANC, post-initiation ANC outcomes within a five-band hierarchy, monitoring intensity, and pre-specified sensitivity and exploratory analyses were examined. Results Among 239 patients, 179 formed the final study cohort, contributing 284.7 on-clozapine person-years. Overall, 158 patients (88.3%) had no post-initiation ANC <2.0 × 10 9 /L, 16 (8.9%) entered the 1.5–2.0 × 10 9 /L low-ANC monitoring range, 2 (1.1%) reached conventional-threshold neutropenia, 3 (1.7%) reached severe neutropenia, and none reached agranulocytosis range. Five patients (2.8%) met the clinically meaningful ANC <1.5 × 10 9 /L threshold; all were BEN-risk, all had concomitant valproate exposure, and four had concomitant lithium exposure. Baseline ANC was lower in BEN-risk patients and statistically accounted for much of the BEN-risk-to-outcome association in an exploratory analysis. Conclusions Low ANC during clozapine treatment was uncommon and strongly shaped by baseline ANC and BEN-risk context. Baseline ANC-informed and BEN-aware interpretation may reduce unnecessary treatment interruption while preserving haematological safety. BEN-risk was inferred from nationality, not confirmed genetically; all five low-ANC cases received concomitant valproate; associations are observational.