Clozapine Levels Are Clinical Clues—Not Treatment Targets

What is the “right” plasma clozapine concentration?

A new dose-response meta-analysis in Psychiatry Research offers several provocative numbers. But its most important contribution may be its warning against treating those numbers as universal thresholds.

Fornaro and colleagues examined the relationships between plasma clozapine and norclozapine concentrations, therapeutic response and adverse outcomes in people with schizophrenia-spectrum disorders. Their findings reinforce the value of therapeutic drug monitoring while challenging the temptation to practice by a single number.

What the researchers studied

The investigators searched PubMed/MEDLINE, Embase, Web of Science and Scopus through February 16, 2026. The review included cohort studies and clinical trials reporting clozapine exposure alongside efficacy or safety outcomes.

Thirty-four studies met inclusion criteria:

• Nine randomized clinical trials

• Twenty-five observational studies

• 134,111 participants in total

The researchers initially fitted linear models. They then used exploratory quadratic models to identify potential population-level concentration-response turning points.

A turning point is the vertex of a modeled curve—the point at which the predicted relationship changes direction. It is not necessarily the concentration at which a medication becomes effective or an adverse event begins.

The modeled clozapine turning points

The exploratory quadratic analyses produced the following estimates:

Outcome	Modeled clozapine turning point
Change in Brief Psychiatric Rating Scale score	535 ng/mL
Change in total PANSS score	898 ng/mL
Overall tolerability	707 ng/mL
Blood dyscrasia	357 ng/mL
EEG abnormalities or seizures	452 ng/mL
Salivary hypersecretion	225 ng/mL
Tachycardia	503 ng/mL

These values immediately attract attention. The PANSS estimate of 898 ng/mL, for example, could easily be misread as evidence for escalating clozapine until that concentration is reached.

That is not what the analysis demonstrates.

The most important sentence in the paper

The authors explicitly state that the modeled values should not be interpreted as universal therapeutic or toxic thresholds—or as targets for dose escalation.

These are exploratory vertices derived from heterogeneous studies. They describe patterns across populations; they do not identify the optimal concentration for an individual patient.

The blood-dyscrasia estimate of 357 ng/mL illustrates the danger of overinterpretation. It should not be treated as a hematologic toxicity cutoff, and it does not replace established absolute neutrophil count monitoring or clinical protocols.

Similarly, the EEG/seizure estimate of 452 ng/mL does not mean that concentrations below it are universally safe or that concentrations above it invariably produce seizures. Individual susceptibility, concomitant medications, medical conditions and the rate of change in exposure remain relevant.

Why dose and plasma concentration are not interchangeable

The same oral dose can produce very different plasma concentrations in different people. Clozapine exposure is influenced by metabolism, adherence, smoking status, interacting medications and other clinical variables.

A measured level must also be interpreted in relation to the timing and conditions of the blood draw. A concentration obtained under inconsistent sampling conditions cannot be treated as equivalent to a properly timed steady-state sample.

Therapeutic drug monitoring is therefore most useful when it answers a clinical question, such as:

• Is nonresponse occurring despite adequate exposure?

• Could nonadherence or inconsistent use explain the presentation?

• Has a change in smoking or metabolism altered clozapine exposure?

• Is an adverse effect emerging as the concentration increases?

• Has benefit plateaued despite progressively higher exposure?

The result should be interpreted alongside symptoms, functioning, tolerability, adherence and the patient’s goals.

Why higher is not automatically better

The findings support caution when considering escalation beyond a concentration that has already produced meaningful benefit.

If response has plateaued while adverse effects are increasing, pursuing a higher number may expose the patient to additional risk without predictable benefit. Conversely, a patient who responds well at a comparatively low concentration does not need escalation simply because the result falls below a preferred population average.

A plasma level is evidence. It is not a destination.

Important limitations

The apparent precision of the modeled numbers should not obscure the limitations of the evidence.

The analysis combined randomized trials with observational studies. The underlying populations, sampling methods, outcome definitions and monitoring practices were heterogeneous. Evidence at the highest and lowest concentrations may also have been sparse, making the shape of a quadratic curve less stable at its extremes.

Certainty was rated as moderate for 11 of the 21 evaluated outcomes and low for the other 10. Population-level associations also cannot establish that a particular concentration caused either benefit or harm in an individual patient.

The very large total sample size is impressive, but it does not eliminate heterogeneity or transform exploratory curve vertices into validated clinical thresholds.

The clinical takeaway

This meta-analysis does not provide a new concentration that every patient should reach. It provides a more useful lesson: clozapine therapeutic drug monitoring should remain individualized.

The goal is not to maximize the concentration. It is to identify an exposure that provides the best attainable balance between therapeutic benefit and adverse effects for a particular patient.

That requires combining the plasma result with clinical response, tolerability, adherence, sampling conditions, relevant metabolic factors and established safety monitoring.

The number informs the decision. It does not make the decision.

Reference

Fornaro M, Di Lorenzo C, Siskind D, de Bartolomeis A. Efficacy and Safety Outcomes of Different Plasma Clozapine and Norclozapine Levels in Schizophrenia Spectrum Disorder: A Dose-Response Meta-Analysis Exploring “Critical” Levels. Psychiatry Research. Published online July 26, 2026. Journal/DOI | PubMed