Nonclinical Development
Nonclinical development is where the safety case is built before a human, patient, or target animal is ever exposed. For medicines this is the toxicology and safety-pharmacology programme: dose-range-finding and pivotal repeat-dose studies, genotoxicity, safety pharmacology, and the toxicokinetics that connect animal exposure to the doses proposed for people. For devices it is biological evaluation under the ISO 10993 series — a risk-based assessment of the materials and their body contact, not a reflex battery of tests. For veterinary products it includes target-animal safety; for food-adjacent and cosmetic products, ingredient safety assessment carries the equivalent burden. The pivotal safety studies run under GLP — 21 CFR Part 58, the OECD principles, and their EU transposition — because their credibility must survive without the possibility of re-running them.
What this page does not claim
Phases are an organising device, not a regulatory mandate — real programmes overlap them, and a medical device, a food product and a small-molecule drug traverse them differently. Each page says where the differences bite.
What happens in this phase
Nonclinical development is where the safety case is built before a human, patient, or target animal is ever exposed. For medicines this is the toxicology and safety-pharmacology programme: dose-range-finding and pivotal repeat-dose studies, genotoxicity, safety pharmacology, and the toxicokinetics that connect animal exposure to the doses proposed for people. For devices it is biological evaluation under the ISO 10993 series — a risk-based assessment of the materials and their body contact, not a reflex battery of tests. For veterinary products it includes target-animal safety; for food-adjacent and cosmetic products, ingredient safety assessment carries the equivalent burden. The pivotal safety studies run under GLP — 21 CFR Part 58, the OECD principles, and their EU transposition — because their credibility must survive without the possibility of re-running them.
GLP is a study-integrity discipline, and its scope is exact: it governs nonclinical laboratory safety studies intended to support regulatory submissions. It does not govern clinical trials, and it does not govern routine quality-control testing — a distinction that matters because "GLP-like" is a phrase that hides decisions. Within its scope, the machinery is distinctive: a named study director holding single-point responsibility, an independent quality-assurance unit inspecting the study against its protocol, a master schedule, and archives that preserve raw data, specimens, and the final report for years. Bioanalytical work that generates the concentration data behind toxicokinetics is expected to follow validated methods per ICH M10.
The output of the phase is not a pile of studies but an argument: that the observed effects, the exposure margins, and the identified target organs justify the proposed first exposure — the first-in-human dose, the clinical investigation, the target-animal study. Quality's stake is that this argument is only as strong as the records beneath it. A pivotal study with unexplained data gaps, an audit trail that cannot attribute a result, or a test article whose identity and stability were never characterised does not merely weaken one report; it invites a regulator to ask what else in the package shares the defect.
- A nominated candidate with characterised identity, purity, and stability sufficient to dose reliably — the test-article characterisation GLP studies depend on.
- A nonclinical programme design mapped to the intended clinical plan or device classification, so studies answer the questions the next phase will ask.
- GLP-compliant facilities and a qualified study director for the pivotal safety studies, whether in-house or contracted.
- Validated or scientifically justified bioanalytical methods for the toxicokinetic and exposure work.
- Completed pivotal safety studies with final, QA-audited reports and archived raw data.
- An integrated safety assessment supporting the proposed first exposure — starting dose, margins, monitorable risks — or, for devices, a biological evaluation concluding acceptable biocompatibility for the intended contact.
- Toxicokinetic data linking observed effects to exposure, generated by validated bioanalytical methods.
- The nonclinical package assembled in submission-ready form, with every claim traceable to an archived study record.
SPEQ synthesis. Phase boundaries and gate criteria are an organising device for planning and review, not a regulatory mandate. Real programmes overlap phases and re-enter them; treat these as the questions worth answering, not a compliance checklist.
- Operate or oversee the GLP quality-assurance unit: protocol review, in-life and critical-phase inspections, and report audits against raw data.
- Qualify and oversee contract research organisations running outsourced studies, since accountability for the study does not transfer with the work.
- Guard the GLP scope boundary — which studies must run under GLP, which legitimately do not, and how each non-GLP study is labelled in the submission.
- Enforce test-article accountability: identity, stability, dose-formulation verification, and the chain from batch to animal.
- Own archive integrity — raw data, specimens, and reports retained, indexed, and retrievable for the life of the obligations.
- Approved study protocols and amendments, with study-director sign-off
- Final study reports with QA statements, supported by archived raw data
- Toxicokinetic and bioanalytical reports from validated methods
- Biological evaluation plan and report for device materials (ISO 10993-1 framework)
- Test-article characterisation and dose-formulation analysis records
- The integrated nonclinical summary that justifies first exposure
- Running a study "GLP-like" and discovering at submission that the claim is neither auditable nor accepted — the study must often be repeated, at the timeline's expense.
- Weak test-article characterisation, so observed effects cannot be firmly attributed to the intended substance at the intended dose.
- Treating device biocompatibility as a checklist of tests rather than a risk-based evaluation, generating cost without an argument.
- CRO oversight reduced to contract signature — no protocol review, no audit, no data ownership — until a finding at the CRO becomes the sponsor's finding.
- Archive decay: raw data and specimens that exist but cannot be retrieved or read when a regulator asks, years later.
Derived from the 6 standards SPEQ maps to this phase, across 6 regulatory bodies: FDA, OECD, EC, MHRA, ICH, ISO.
FREQUENTLY ASKED
What does GLP actually cover — and what does it not?
GLP governs nonclinical laboratory safety studies intended to support regulatory submissions — the toxicology, safety pharmacology, and similar work on which a first-exposure decision rests. It does not govern clinical trials, which run under GCP; it does not govern routine QC testing, which runs under GMP; and it does not govern discovery science at all. The confusion is expensive in both directions: running exploratory work under full GLP wastes resources, while running a pivotal safety study outside GLP can render it unusable for the submission it was meant to support, forcing a repeat.
Why is the study director such a central figure?
Because GLP deliberately concentrates responsibility in one named individual. The study director is the single point of study control: accountable for the protocol, the conduct, the interpretation, and the report, and for confirming GLP compliance in it. The quality-assurance unit inspects independently, but it advises and reports — it does not share the study director's accountability. This structure exists so that a regulator reading a study years later knows exactly who answered for it, and so that multi-site studies cannot diffuse responsibility until no one holds it.
Do medical devices go through nonclinical development too?
Yes, but it looks different. Instead of a toxicology programme sized to systemic drug exposure, a device undergoes biological evaluation under the ISO 10993 series: a risk assessment driven by the nature and duration of body contact, drawing on material characterisation, existing data, and targeted testing only where the assessment identifies a gap. Bench performance testing, software verification, and — where relevant — animal studies join it. The unifying principle is the same as for medicines: the evidence justifying human exposure must be generated under controls that make it credible without repetition.