Clozapine outperformed other antipsychotics across six diagnoses — but not in BPD

Clozapine appears in treatment guidelines across the world as an option for several psychiatric disorders, but the evidence supporting most of those listings is thin [s1]. A cohort study published in The Lancet Psychiatry on 2 November tested its effectiveness across seven diagnoses using national registers from two countries [s1].

The design does most of the work

The study used nationwide register-based data from Finland and Sweden covering individuals aged 16 or older diagnosed with schizophrenia, schizoaffective disorder, delusional disorder, bipolar disorder, psychotic depression, major depressive disorder, or borderline personality disorder [s1].

Clozapine was compared with other oral antipsychotics as a group, and for bipolar disorder additionally against mood stabilisers [s1]. The primary outcome was all-cause psychiatric hospitalisation; secondary outcomes were a composite of all-cause hospitalisation or mortality, all-cause discontinuation (for which clozapine was compared with olanzapine), and disorder-specific hospitalisations [s1].

The methodological centrepiece is a within-individual design, in which each patient serves as their own control [s1]. That structure removes confounding by anything stable about a person — genetics, childhood adversity, baseline illness severity — because the comparison is between periods when the same individual was and was not taking the drug. It does not remove confounding by things that change over time, including illness trajectory, and it is a design that can be distorted when treatment is started in response to worsening.

Data were analysed separately in each country and combined using meta-analytical methods [s1].

Who was in it

The study population was 505,474 individuals, of whom 283,809 (56.1%) were women and 221,665 (43.9%) were men, with a mean age of 41.6 years (SD 4.4) [s1]. Data on ethnicity were unavailable [s1]. In total 19,910 individuals used clozapine [s1].

The distribution of use is itself the finding that most needs stating. Parsed by disorder, clozapine was used by 12.2–18.7% of people with schizophrenia (n=5258 in Sweden, 10,115 in Finland), 8.8–20.7% with schizoaffective disorder (n=1131, n=1591), 1.8% with delusional disorder (n=341, Sweden only), 0.5% with major depressive disorder (n=118, n=331), 0.5–0.6% with bipolar disorder (n=490, n=371), 0.3–0.6% with psychotic depression (n=39, n=111), and 0.3% with borderline personality disorder (n=36, Sweden only) [s1].

Outside the schizophrenia spectrum, in other words, the numbers are small — 39 people with psychotic depression in one country, 36 with borderline personality disorder. Estimates built on counts like those carry wide uncertainty regardless of how large the total cohort is.

The results

Clozapine use was associated with reduced psychiatric hospitalisation risk versus other oral antipsychotics in every disorder except borderline personality disorder [s1].

The largest reductions were in schizophrenia (meta-analysed adjusted hazard ratio 0.70, 95% CI 0.67–0.72) and schizoaffective disorder (0.71, 0.67–0.74), followed by delusional disorder (adjusted HR 0.73, 0.60–0.89), major depressive disorder (0.74, 0.66–0.84), psychotic depression (0.76, 0.61–0.96), and bipolar disorder (0.77, 0.69–0.87) [s1].

There was no evidence of increased all-cause hospitalisation or mortality risk associated with clozapine use for any disorder examined [s1]. For every disorder except borderline personality disorder, all-cause discontinuation rates were lower on clozapine than on olanzapine [s1]. For bipolar disorder, clozapine outperformed mood stabilisers on all-cause psychiatric hospitalisation and other antipsychotics on several disorder-specific outcomes [s1].

A comment published alongside it in The Lancet Psychiatry is titled "Clozapine: superior yet underused" [s2].

What this cannot settle

Clozapine carries monitoring requirements that exist for a reason — agranulocytosis, myocarditis, and other serious adverse effects — and this study's mortality analysis, which found no excess signal [s1], is not a substitute for the safety literature that produced those requirements. A register study measures what happened to people who were prescribed the drug under a monitoring regime, not what would happen without one.

The within-individual design also cannot address why clozapine was started when it was. In most health systems it is used after other treatments fail, which means clozapine periods follow periods of established treatment failure — a pattern that can cut either way in a within-person comparison.

And the confidence intervals for the rarer indications rest on very few exposed individuals [s1]. The schizophrenia and schizoaffective results are the ones supported by substantial numbers.

Why it matters

The finding that clozapine reduced psychiatric hospitalisation across six of seven disorders, with no mortality signal, in two national cohorts, is the largest transdiagnostic evidence base the drug has [s1]. The authors argue it could inform several treatment guidelines and clinical decision making [s1]. The borderline personality disorder exception is equally worth carrying forward — it is the one diagnosis where the pattern did not hold [s1].

This article is informational and does not constitute medical advice. Decisions about antipsychotic treatment involve monitoring requirements and individual risk factors that no population study can adjudicate.

Sources

• [s1] Transdiagnostic effectiveness and safety of clozapine in individuals with psychotic, affective, and personality disorders: nationwide and meta-analytic comparisons with other antipsychotics. The Lancet Psychiatry, 2 November 2025. https://doi.org/10.1016/S2215-0366(25)00297-4
• [s2] Clozapine: superior yet underused. The Lancet Psychiatry, 2 November 2025. https://doi.org/10.1016/S2215-0366(25)00326-8