In drug development, the maximum tolerated dose (MTD) refers to the highest dose of an active substance that, over a defined duration of administration, still causes acceptable adverse effects and can therefore serve as an upper reference point for further clinical development.
The term is used primarily in early clinical phases to strike a balance between pharmacological activity and tolerability and to support rational dose selection for subsequent studies.
What the MTD is used for in early clinical studies
In classic Phase I studies (particularly in oncology, but also in other indications), the MTD often serves as a decision criterion for the dose to be further investigated in a Phase II study.
It is not identical to the “effective dose”; rather, it primarily describes the limit of tolerability under the conditions defined in the protocol (population, concomitant medication, dosing regimen).
Typical study designs for determining the MTD
Determination is often performed using dose-escalation designs. Traditionally, 3+3 designs are common; increasingly, model-based approaches are used, such as continual reassessment method (CRM) approaches or Bayesian optimal interval designs.
It is essential that the escalation rules are described prospectively in the clinical trial protocol and that safety data are assessed promptly (e.g., by a safety review team or a data safety monitoring board).
Dose-limiting toxicity (DLT) as the basis for the MTD
In many programs, the MTD is operationalized via dose-limiting toxicities (DLTs). DLTs are adverse events defined in the protocol (severity, laboratory values, duration) that make further dose escalation unacceptable for a proportion of participants.
For the MTD to be robust, DLT criteria must be clear, measurable, and plausible for the respective indication (e.g., CTCAE criteria in oncology) and linked to a defined observation window.
Distinction: MTD, recommended Phase II dose, and therapeutic window
In modern development programs, the MTD is increasingly supplemented or even replaced by the recommended Phase II dose (RP2D). In addition to acute toxicity, the RP2D also takes into account pharmacokinetics, pharmacodynamics, target engagement, and early signals of efficacy.
Especially with targeted therapies or biologics, a plateau effect may occur, meaning that a dose below the MTD is sufficient to achieve maximal biological activity. In this context, assessment of the therapeutic window and the exposure–response relationship is critical.
Common errors and practical pitfalls
A common error is overinterpreting the MTD as the “optimal” dose. An MTD derived from a small dose-escalation cohort can also be strongly influenced by population characteristics, concomitant medication, or supportive therapy.
Protocol deviations, delayed adverse effects, or incomplete safety assessment processes can also result in the MTD being set too high or too low. Therefore, rigorous safety management is crucial in practice, including timely SAE/SUSAR assessment and thorough documentation.
Regulatory classification and documentation:
Dose finding and safety monitoring in early clinical studies must be aligned with Good Clinical Practice principles as described in ICH E6 (R3). In Europe, requirements for the submission and conduct of clinical trials under EU Regulation 536/2014 (CTR) are also relevant.
The derivation of the MTD (including DLT definition, escalation scheme, safety data, and rationale for the subsequent dose) should be documented consistently in the clinical study report (CSR) and in regulatory dossiers.
FAQ
Is the maximum tolerated dose required in every Phase I study?
No. In many indications, determining an MTD is not necessarily the primary objective, for example if a meaningful dose can be derived early from pharmacodynamic and exposure data, or if safety limits can be clearly defined from non-clinical data.
How does the MTD differ from a “highest tested dose”?
The highest tested dose is simply the maximum dose administered during the course of the study. The MTD, by contrast, is a derived parameter linked to defined DLT criteria and an acceptable DLT rate.
What DLT rate is typically considered acceptable?
That depends on the context. In oncology programs, target ranges of around a 20–33% DLT rate are often discussed, whereas more conservative thresholds may be appropriate in other indications. What matters is prospective specification in the protocol and the medical rationale.
Regulatory references: ICH E6(R3) Good Clinical Practice; EU Regulation (EU) No. 536/2014 (Clinical Trials Regulation).
From a sponsor and CRO perspective, it is important that definitions and analysis rules are specified in advance in the protocol and statistical analysis plan. Consistent data collection, plausibility checks, and transparent reporting reduce room for interpretation vis-à-vis ethics committees, competent authorities, and HTA bodies.
From a sponsor and CRO perspective, it is important that definitions and analysis rules are specified in advance in the protocol and statistical analysis plan. Consistent data collection, plausibility checks, and transparent reporting reduce room for interpretation vis-à-vis ethics committees, competent authorities, and HTA bodies.
From a sponsor and CRO perspective, it is important that definitions and analysis rules are specified in advance in the protocol and statistical analysis plan. Consistent data collection, plausibility checks, and transparent reporting reduce room for interpretation vis-à-vis ethics committees, competent authorities, and HTA bodies.
From a sponsor and CRO perspective, it is important that definitions and analysis rules are specified in advance in the protocol and statistical analysis plan. Consistent data collection, plausibility checks, and transparent reporting reduce room for interpretation vis-à-vis ethics committees, competent authorities, and HTA bodies.
From a sponsor and CRO perspective, it is important that definitions and analysis rules are specified in advance in the protocol and statistical analysis plan. Consistent data collection, plausibility checks, and transparent reporting reduce room for interpretation vis-à-vis ethics committees, competent authorities, and HTA bodies.