Long-term treatment with clozapine and other antipsychotic drugs and the risk of haematological malignancies in people with schizophrenia: a nationwide case-control and cohort study in Finland

Lancet Psychiatry. 2022 May;9(5):353-362. doi: 10.1016/S2215-0366(22)00044-X. Epub 2022 Mar 22.


Background: Clozapine is the most efficacious treatment for schizophrenia and is associated with lower overall mortality than are other antipsychotic drugs, despite the risk of agranulocytosis. Preliminary reports over the past 10 years suggest a possible risk of haematological malignancies, but the issue has remained unsettled. We aimed to study the risk of haematological malignancies associated with use of clozapine and other antipsychotics.

Methods: We did a nationwide case-control (and cohort) study of people with schizophrenia, using prospectively gathered data from Finnish national registers. A nested case-control study was constructed by individually matching cases of lymphoid and haematopoietic tissue malignancy with up to ten controls without cancer by age, sex, and time since first schizophrenia diagnosis. For the case-control study, we restricted inclusion criteria to malignancies diagnosed on a histological basis, and excluded individuals outside of the age range 18-85 years, and any patients that had a previous malignancy. Analyses were done using conditional logistic regression adjusting for comorbid conditions.

Findings: For the case-control study 516 patients with a first-time diagnosis of lymphoid and haematopoietic tissue malignancy during years 2000-17 and diagnosed after their first diagnosis of schizophrenia were identified. 102 patients were excluded due to diagnosis that was without a histological basis, five patients were excluded because of their age, and 34 were excluded for a previous malignancy, resulting in 375 patients being matched to controls. We selected up to ten controls without cancer (3734 in total) for each case from the base cohort of people with schizophrenia. For the cohort study, data for 55 949 people were included for analysis. Cumulative incidence of haematological malignancies during the mean follow-up of 12·3 years (SD 6·5) was 102 (0·7%) cases among 13 712 patients who had used clozapine (corresponding to event rate of 61 cases per 100 000 person-years), and during mean follow-up of 12·9 years (SD 7·2) was 235 (0·5%) malignancies among 44 171 patients having used other antipsychotic medication than clozapine (corresponding to 41 cases per 100 000 person-years). Of the 375 individuals with haematological malignancies (305 lymphomas, 42 leukaemia, 22 myelomas, 6 unspecified) observed from 2000-17, 208 (55%) were males and 167 (45%) were female. Ethnicity data were not available. Compared with non-use of clozapine (most had used other antipsychotics and a few had used no antipsychotics), clozapine use was associated with increased odds of haematological malignancies in a dose-response manner (adjusted odds ratio 3·35, [95% CI 2·22-5·05] for ≥5000 defined daily dose cumulative exposure, p<0·0001). Exposure to other antipsychotic drugs was not associated with increased odds. A complementary analysis showed that the clozapine-related risk increase was specific for haematological malignancies, because no such finding was observed for other malignancies. Over 17 years of follow-up of the base cohort, 37 deaths occurred due to haematological malignancy among patients exposed to clozapine (26 with ongoing use at time of haematological malignancy diagnosis, and 11 in patients who did not use clozapine at the exact time of their cancer diagnosis), whereas only three deaths occurred due to agranulocytosis.

Interpretation: Unlike other antipsychotics, long-term clozapine use is associated with increased odds of haematological malignancies. Long-term clozapine use has a higher effect on mortality due to lymphoma and leukaemia than due to agranulocytosis. However, acknowledging that the absolute risk is small compared with the previously observed absolute risk reduction in all-cause mortality is important. Our results suggest that patients and caregivers should be informed about warning signs of haematological malignancies, and mental health clinicians should be vigilant for signs and symptoms of haematological malignancy in patients treated with clozapine.

Funding: The Finnish Ministry of Social Affairs and Health and Academy of Finland.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Adult
  • Aged
  • Aged, 80 and over
  • Agranulocytosis* / chemically induced
  • Agranulocytosis* / drug therapy
  • Antipsychotic Agents* / adverse effects
  • Case-Control Studies
  • Clozapine* / adverse effects
  • Cohort Studies
  • Female
  • Finland / epidemiology
  • Hematologic Neoplasms* / chemically induced
  • Hematologic Neoplasms* / drug therapy
  • Hematologic Neoplasms* / epidemiology
  • Humans
  • Leukemia* / chemically induced
  • Leukemia* / drug therapy
  • Male
  • Middle Aged
  • Schizophrenia* / drug therapy
  • Schizophrenia* / epidemiology
  • Young Adult


  • Antipsychotic Agents
  • Clozapine