Why Two Toxicology Dose Descriptors Matter
Animal toxicology studies describe effects at tested doses, while first-in-human (FIH) planning must use those findings to reason about a human starting dose. Two terms often appear in that reasoning: the no-observed-adverse-effect level (NOAEL) and the highest non-severely toxic dose (HNSTD). They both summarize a nonclinical study, but they answer different questions. A NOAEL concerns whether an effect was judged adverse; an HNSTD concerns whether toxicity reached a predefined severe level.
Neither term is a direct estimate of a safe human dose. Each is a dose level from a particular study, in a particular species, with a particular route, dosing schedule, duration, and set of assessments. The associated toxicokinetic (TK) data help describe the systemic exposure observed at that dose. Keeping the dose descriptor separate from its exposure measurements is essential: the descriptor is selected from toxicology findings, while TK characterizes exposure during the study.
How to Read the Dose Descriptors
A NOAEL is determined by evaluating the findings across dose groups and deciding which treatment-related changes are adverse. A finding is not automatically adverse merely because it differs from the control group. Its magnitude, duration, reversibility, functional consequences, relationship to other findings, and biological context matter. Conversely, a finding need not cause obvious illness to be adverse: for example, a persistent organ injury supported by clinical pathology and histology may be adverse even if animals appear outwardly well.
The NOAEL is the highest tested dose without observed adverse effects, not necessarily the dose at which there were no measurable effects. A treatment-related but non-adverse change may occur at the NOAEL. If adverse effects occur at every tested dose, the study has no NOAEL within its tested range. If no adverse effects occur at the highest tested dose, that dose is the study NOAEL, but the study has not established what dose above it would cause an adverse effect.
HNSTD is used particularly in nonclinical development of oncology drugs, where some toxicity may be expected or acceptable in the context of the intended treatment, but severe toxicity is a distinct concern. Its interpretation depends on the study’s criteria for severe toxicity, which should be defined before interpreting the results. Those criteria may include the nature, intensity, duration, reversibility, or clinical consequence of findings. HNSTD therefore does not mean “no toxicity” or “no adverse effects.” A dose can be below the severe-toxicity threshold while still producing adverse effects.
| Question | NOAEL | HNSTD |
|---|---|---|
| What determines the selected dose? | Whether any treatment-related adverse effect is observed | Whether toxicity meets the study’s definition of severe |
| Can findings occur at the selected dose? | Yes, if they are treatment-related but judged non-adverse | Yes, including adverse findings that do not meet the severe-toxicity definition |
| Typical context | General nonclinical toxicology and safety assessment | Often nonclinical oncology toxicology |
| What does the value represent? | A tested animal dose level | A tested animal dose level |
The two values may coincide, but they need not. If the first adverse finding is already severe, the NOAEL may be the highest dose below it and may also be the HNSTD. If adverse but non-severe findings arise at lower doses, the HNSTD can be higher than the NOAEL. If the study’s severe-toxicity criteria are not adequately specified, an HNSTD cannot be interpreted reliably simply from a summary table.
From Study Findings to a Dose and Its Exposure Record
A practical interpretation starts with the study design and findings, not with a single number copied from a report. Confirm the species, sex if relevant, route, dosing frequency, treatment duration, dose units, group sizes, and timing of assessments. Then identify treatment-related findings at each dose, judge adversity for NOAEL selection, and apply the prespecified severity criteria for HNSTD selection. These are separate evaluations, even when they use the same observations.
The selected dose should be reported with its dosing context, such as 12 mg/kg/day orally for 28 days in a specified species. A dose alone does not state how much drug reached systemic circulation. TK measurements can provide exposure summaries at the relevant dose, commonly including maximum concentration and area under the concentration-time curve, when measured and reported. These are exposure data, typically with units such as ng/mL or ng·h/mL; they are not alternative names for the NOAEL or HNSTD.
Exposure at the selected dose can vary with species, route, schedule, absorption, and clearance. It can also vary across animals in a group. When available, note whether the reported TK value is a group mean, median, or individual value, and whether it refers to parent drug or a metabolite. A missing exposure measurement does not change which dose was the NOAEL or HNSTD, but it limits the exposure information available for later FIH reasoning.
Worked Example: Selecting a NOAEL When Non-Adverse Findings Occur
In a hypothetical 28-day oral study in rats, the dose groups are 2, 6, and 18 mg/kg/day. At 2 mg/kg/day, no treatment-related findings are observed. At 6 mg/kg/day, a small, reversible change in a clinical chemistry measurement occurs without related tissue injury or functional consequences; the study team judges it non-adverse. At 18 mg/kg/day, persistent tissue injury is observed and judged adverse. Identify the NOAEL and state what the findings imply for HNSTD.
The NOAEL is the highest tested dose without an adverse effect. The 18 mg/kg/day group has an adverse effect, so it cannot be the NOAEL. The 6 mg/kg/day finding is treatment-related but judged non-adverse, so it does not disqualify that dose. Therefore:
Suppose the study’s severe-toxicity criteria also classify the tissue injury at 18 mg/kg/day as severe, while the lower-dose finding does not meet those criteria. The HNSTD would then be 6 mg/kg/day: it is the highest tested dose below the dose with severe toxicity. Here the NOAEL and HNSTD coincide, but for different reasons—absence of adverse effects at 6 mg/kg/day and absence of severe toxicity at that dose.
Worked Example: An HNSTD Above the NOAEL
In a hypothetical repeat-dose oncology study, animals receive 1, 3, 10, or 20 mg/kg per dose on the same schedule. Findings at 3 mg/kg per dose include an adverse but reversible blood-cell decrease. At 10 mg/kg per dose, the decrease is greater, but remains below the protocol’s prespecified severe-toxicity criteria and resolves during recovery. At 20 mg/kg per dose, severe, prolonged toxicity meets those criteria. Assume the findings at 1 mg/kg per dose are non-adverse. Determine the NOAEL and HNSTD.
The first adverse finding is at 3 mg/kg per dose. Thus the highest tested dose without an adverse finding is:
For HNSTD, apply the separate severe-toxicity criterion. The 10 mg/kg per dose group has adverse toxicity, but it does not meet the defined severe threshold. The 20 mg/kg per dose group does meet that threshold. Therefore:
This example shows why the labels are not interchangeable: 10 mg/kg per dose is not a NOAEL, because adverse toxicity occurred, but it can be the HNSTD because severe toxicity was not observed at that dose. Both descriptors remain specific to this study’s schedule and severity criteria.
What the TK Data Add
Suppose the second study also reports a mean AUC over the dosing interval of 8.4 mg·h/L at 1 mg/kg per dose and 63 mg·h/L at 10 mg/kg per dose. These values describe measured exposure in the animals at the NOAEL and HNSTD, respectively; they do not redefine either dose descriptor. Recording both dose and exposure preserves the distinction between the administered amount and the systemic concentration-time profile observed.
For example, if the table reports 63 mg·h/L at 10 mg/kg per dose, the appropriate statement is that the HNSTD was 10 mg/kg per dose and the associated mean AUC was 63 mg·h/L under the study conditions. It would be incorrect to call 63 mg·h/L the HNSTD: that is an exposure measurement, not a dose. These data can inform subsequent FIH assessments, but by themselves they do not determine a human starting dose.
Worked Example: Reporting Dose and Exposure Without Confusing Them
A hypothetical toxicology report gives a NOAEL of 4 mg/kg/day in dogs after oral dosing for 14 days. At that dose, TK data report a mean Cmax of 1.8 mg/L and a mean AUC over 24 hours of 12.6 mg·h/L. At 12 mg/kg/day, adverse liver findings occur. State what dose and exposure information the NOAEL provides.
The NOAEL is the dose-level finding from the toxicology assessment:
The associated exposure measurements are reported separately: mean Cmax = 1.8 mg/L and mean AUC over 24 hours = 12.6 mg·h/L at that dose. The 12 mg/kg/day group’s adverse liver findings explain why it is not the NOAEL, assuming the findings are treatment-related and the 4 mg/kg/day group has no adverse finding. The conclusion is limited to the specified dog study, oral route, and 14-day schedule; it does not establish a dose or exposure that is safe in humans.
Selection Logic and Important Limitations
The selection can be expressed as two distinct rules. Let \(D\) denote the set of tested dose levels, and let \(A(d)\) indicate whether an adverse treatment-related effect is observed at dose \(d\). The study NOAEL is the greatest tested dose for which \(A(d)\) is false, if at least one such dose exists. Let \(S(d)\) indicate whether the prespecified severe-toxicity criterion is met. The HNSTD is the greatest tested dose for which \(S(d)\) is false, if at least one such dose exists.
These rules explain why a dose-level label cannot be inferred from a number alone. For NOAEL, the adversity judgments must be supported by the study’s findings. For HNSTD, the severe-toxicity criterion must be stated and applied consistently. If every tested dose has an adverse effect, the set used for the NOAEL is empty; no NOAEL has been demonstrated in the tested range. If every dose meets the severe-toxicity criterion, no HNSTD has been demonstrated in that range.
The rules also assume that the study has interpretable dose groups and adequate observations to make the relevant judgments. Small groups, missing assessments, inconsistent dose administration, or insufficient recovery information can weaken the conclusion. Findings may differ by sex or individual animal, and a summary value can conceal those patterns. Review the underlying data and study interpretation rather than treating a reported descriptor as self-validating.
For FIH decisions, a NOAEL-based approach and an HNSTD-based approach provide different kinds of nonclinical anchor. A NOAEL identifies a dose without observed adverse effects under the study conditions. An HNSTD identifies a dose below the study’s severe-toxicity threshold and may include lesser adverse toxicity. Which descriptor is relevant depends on the drug, intended clinical context, study design, and applicable guidance; neither is universally appropriate for every program. The human starting-dose decision also requires other evidence and a justified translation from animal findings. In particular, a numerically high HNSTD should not be treated as evidence that a comparable human dose is safe.
Common Mistakes and Diagnostics
- Calling a NOAEL a dose with no effects. The correction is that it permits treatment-related findings judged non-adverse; it excludes observed adverse effects.
- Treating the HNSTD as the NOAEL. Check whether adverse findings occurred below the severe-toxicity threshold. If so, the HNSTD may exceed the NOAEL.
- Reporting exposure as the dose descriptor. Keep administered dose units, such as mg/kg/day, separate from concentration or AUC units.
- Assuming the highest tested dose must be the HNSTD or NOAEL. It qualifies only if it meets the relevant criterion; a study may have no qualifying level within the tested range.
- Ignoring study context. A value without species, route, schedule, duration, and severity criteria is difficult to interpret and cannot be generalized automatically to another study or to humans.
A useful diagnostic is to ask whether the report gives enough information to reproduce the label: for the NOAEL, are adverse findings and their dose groups described? For the HNSTD, are severe-toxicity criteria explicit? Are TK measurements tied to the corresponding dose and schedule? If any answer is no, the value may still be reported, but its interpretation should be qualified.
Check Your Understanding
Answer each question, then compare with the solutions below.
- At 2 mg/kg/day, a treatment-related change is judged non-adverse; at 8 mg/kg/day, treatment-related adverse effects occur. What is the NOAEL, assuming these are adjacent tested doses and no higher dose without adverse effects was tested?
- A study has adverse but non-severe findings at 5 mg/kg per dose, and severe toxicity at 15 mg/kg per dose. If 10 mg/kg per dose was the highest tested dose without severe toxicity, can 10 mg/kg per dose be the HNSTD? Can it be the NOAEL if adverse findings were also present there?
- A report states “HNSTD exposure = 42 mg·h/L.” What clarification is needed?
- If adverse effects occur at every tested dose, what can be concluded about the NOAEL?
Solutions. (1) The NOAEL is 2 mg/kg/day: the change there is explicitly judged non-adverse, while adverse effects occur at 8 mg/kg/day. (2) Yes, 10 mg/kg per dose can be the HNSTD if it is the highest tested dose that does not meet the prespecified severe-toxicity criterion. It cannot be the NOAEL if adverse findings occurred at that dose. (3) The report should identify the administered dose that is the HNSTD and say what exposure metric 42 mg·h/L represents, including the relevant sampling interval and summary statistic. Exposure is not itself a dose descriptor. (4) No NOAEL has been established within the tested dose range; the study does not show that a lower, untested dose would or would not be a NOAEL.