Antipsychotic Response Does Not Run on One Clock

When should an antipsychotic be judged effective—or ineffective?

The usual answer is expressed in weeks. But a new study in The Lancet Psychiatry suggests that the more clinically useful question may be: Which symptoms are we waiting to improve?

Leucht and colleagues analyzed individual-participant data from six antipsychotic trials involving 2,079 people with acute schizophrenia. Their findings show that the major symptom domains of schizophrenia improve on meaningfully different timelines.

Five symptom domains, five response trajectories

The investigators evaluated five domains derived from the Positive and Negative Syndrome Scale:

• Hostility and excitement

• Positive symptoms

• Anxiety and depression

• Negative symptoms

• Cognitive and disorganization symptoms

Response was defined as a reduction of at least 50% from baseline within a particular domain. After adjustment for baseline severity, age and sex, the median times to response were:

Symptom domain	Median time to response
Hostility and excitement	3 weeks
Positive symptoms	4 weeks
Anxiety and depression	4 weeks
Negative symptoms	8 weeks
Cognitive and disorganization symptoms	8 weeks

All five domains showed substantial improvement during the first weeks of treatment. Improvement then continued more gradually, beginning to plateau at approximately week 15.

The critical finding was not that some symptoms improved while others did not. It was that clinically important improvement occurred at different speeds.

Why the four-week difference matters

In clinical practice, a patient may show early improvement in hallucinations, delusions, hostility or excitement while continuing to experience reduced motivation, diminished emotional expression, social withdrawal or disorganized thinking.

If these residual symptoms are treated as evidence that the medication has completely failed, an otherwise useful treatment may be changed prematurely.

The study suggests that negative and cognitive-disorganization symptoms may require approximately four to five additional weeks to reach the same 50% response threshold as positive, affective or excitement-related symptoms.

This does not establish an eight-week rule for every patient. It supports a more precise approach: assess response by symptom domain rather than relying exclusively on a global impression of whether the patient is “better.”

The study also challenges a familiar assumption

Negative symptoms are often viewed as largely unresponsive to antipsychotic medication. In this analysis, they improved more slowly and less completely—but they did improve.

At six months, the average reductions from baseline were:

• 76% for hostility and excitement

• 64% for positive symptoms

• 61% for cognitive and disorganization symptoms

• 57% for anxiety and depression

• 50% for negative symptoms

Exploratory path analyses found that only a small proportion of negative-symptom improvement was statistically mediated by improvement in other domains. This raises the possibility that some negative symptoms may improve independently rather than solely as a secondary consequence of reduced psychosis, anxiety or depression.

That interpretation remains hypothesis-generating. Statistical mediation cannot establish a biological mechanism, and the study did not incorporate biomarkers or formal neurocognitive testing.

What clinicians should—and should not—take from the findings

The study does not support passive waiting when treatment is unsafe, poorly tolerated or clearly ineffective. Persistent severe positive symptoms, worsening behavior, adverse effects, nonadherence, inadequate dosing and substance use may all require earlier intervention.

It also does not mean that every residual negative symptom will eventually respond to the current medication. Primary negative symptoms must still be distinguished from secondary contributors such as depression, sedation, extrapyramidal effects, persistent psychosis and environmental deprivation.

The practical lesson is narrower and more useful: the persistence of one symptom domain should be interpreted in the context of how the other domains are changing.

A patient whose positive symptoms are improving at week four but whose motivation and organization remain impaired presents a different clinical picture from someone whose hallucinations, delusions and behavioral disturbance remain unchanged.

Important limitations

This was an individual-participant data analysis of available trials, not a systematic review of all eligible studies. The authors had access to six of 20 potentially eligible long-term trials.

The analyzed medications were primarily olanzapine and amisulpride, with additional data involving risperidone, quetiapine and ziprasidone. The dataset did not include clozapine, partial dopamine agonists or xanomeline–trospium, so the identified timelines should not automatically be generalized to those treatments.

There was no long-term placebo group. Some trials permitted benzodiazepines, which could have influenced anxiety, excitement or sedation. Most participants were White, nearly two-thirds were men, and people with lived experience were not involved in the study.

The cognitive-disorganization domain was also based on PANSS items such as conceptual disorganization, difficulty with abstract thinking and poor attention. It should not be equated with neurocognitive performance measured through formal testing.

A better question for follow-up visits

Antipsychotic response is not a single event occurring at a predetermined week. It is a multidimensional process.

Instead of asking only, “Has this medication worked?” clinicians may obtain more useful information by asking:

• Which symptoms changed first?

• Which symptoms remain functionally important?

• Is the trajectory still moving in the right direction?

• Are residual symptoms expected to respond to medication?

• Is the current trial sufficiently effective and tolerable to continue?

The larger lesson is straightforward: do not judge every symptom by the same clock.

Reference

Leucht S, Harrer M, Illing S, et al. Temporal course of symptom domain improvement in schizophrenia: individual participant data analysis of six antipsychotic drug trials. Lancet Psychiatry. 2026;13(8):688–698. Journal/DOI | Open-access article | PubMed