Incidence and risk for neutropenia/agranulocytosis among clozapine users: A retrospective cohort study

Abstract

Objective. To estimate the incidence and the risk of neutropenia or agranulocytosis (the outcome) associated with clozapine use (the exposure), and to identify risk factors. Methods. All data were derived from the computerized hospital database. Adult psychiatric patients were identified, and 95 incident clozapine users and 884 non-clozapine users were included. Cox proportional hazards regression was used to estimate the hazards ratio (HR) of developing the outcome after clozapine use adjusted for confounders. The interaction between clozapine and valproic acid was assessed a posteriori. Results. Throughout the 24-month follow-up, the incidence of neutropenia was 6.3% in the clozapine group and 5.8% in the non-clozapine group. One agranulocytosis was found in the non-clozapine group. The HR (95% CI) for neutropenia were: clozapine 1.33 (0.54–3.25) and age . 45 years 2.99 (1.63–5.48). Lithium, as an independent protective factor, reduced the risk for neutropenia by 85% compared with patients who did not receive lithium, HR 0.15 (95% CI 0.02–1.09). Valproic acid might potentiate the clozapine-associated neutropenia (HR 5.10, 95% CI 0.70–37.12). Conclusion. Clozapine might slightly increase the risk of neutropenia in psychiatric patients. Concerning clozapine-associated neutropenia, older patients are at increased risk and use of valproic acid concurrently with clozapine should be avoided.

Key Words::

• Clozapine
• neutropenia
• agranulocytosis
• cohort
• valproic acid
• psychosis

Acknowledgements

The authors thank Ms Pintip Wattanasukchai, the staff at the Division of Information Technology, Songklanagarind Hospital, for her excellent assistance in preparing the data for analysis. The help of Dr Alan Geater with proofing the manuscript is acknowledged. This study received financial support from the Research Funding for Graduate study, Prince of Songkla University.

Statement of interests: The present study had no conflict of interest. Decision about design and conduct of the statistical analysis, interpretation of the results, as well as preparation and submission of the manuscript was made by the authors and was not influenced by the funder.