← Drug Development Suite Technical Operations · Vitalis Therapeutics Inc.

Stability Program

Download Word

Vitalis Therapeutics Inc. — The stability program for VitaFlow (VTX-401) under ICH Q1A(R2): study design and conditions, position at 18 months, the 24 months refrigerated plus 28 days in use at room temperature claim sought, the timing decision that made it achievable, and post-approval obligations.

18 mo
Data available
24 mo
Required for claim
3
Primary batches
0
Recovery levers
Contents
  1. What Stability Establishes
  2. Study Design
  3. Bracketing and Matrixing
  4. Position at 18 Months
  5. When a Result Goes Out of Trend
  6. The Shelf-Life Claim
  7. Stability and Distribution
  8. The Timing Decision
  9. Post-Approval Obligations
  10. Governance Position

1. What Stability Establishes

A stability program answers one question: how long does this product remain within specification, stored how?

The answer becomes the shelf life on the label, and it is not an estimate or a projection. It is the period actually supported by real-time data collected under ICH Q1A(R2) on batches made by the commercial process at the commercial site.

This is the constraint most resistant to management, and it is worth understanding why. Every other problem in this program has a lever. Enrolment behind curve? Activate more sites, spend more money. Method transfer failing? Embed a team on site. PPQ campaign tight? Sequence it earlier.

Stability has no lever at all. 24 months of real-time data takes 24 months. No amount of money, escalation, parallel working or executive attention produces month nineteen early. The only decision available is when to start — and that decision has to be made long before anyone knows whether the program will succeed.

2. Study Design

StudyConditionTimepointsPurpose
Long-term5°C ± 3°C0, 3, 6, 9, 12, 18, 24, 36 monthsEstablishes the shelf life. The claim can only be as long as the data.
Accelerated25°C ± 2°C / 60% RH ± 5%0, 3, 6 monthsDetects degradation pathways quickly and supports excursion assessment. Does not substitute for long-term data on a refrigerated product.
In-use25°C, pen in use after first dose0, 7, 14, 21, 28 daysSupports the 28-day room-temperature in-use claim in the target product profile.
PhotostabilityICH Q1B option 2single studyConfirms whether light protection is required in the packaging.
Thermal cycling5°C ↔ 25°C, 3 cyclessingle studySimulates distribution excursions.
Accelerated data does not substitute for long-term data on a refrigerated product. This is a common misconception, and it matters because accelerated studies finish in six months while long-term studies take years.

Accelerated conditions reveal degradation pathways quickly — what breaks down, into what, under stress. That is genuinely useful for formulation work and for assessing excursions during shipping. But a shelf-life claim for a product stored at 5°C ±3°C requires data at 5°C ±3°C. Stress testing tells you what could go wrong; only real-time data tells you whether it does.

Three primary batches are on study. Three primary stability batches, all at commercial scale from the registration campaign. Two are on stability; the third follows the registration batch blocked by GC-01.

Two studies exist to support specific label claims rather than the shelf life itself. In-use stability underwrites the 28-day room-temperature claim in the Target Product Profile — the difference between a pen a patient can carry and one that must return to a refrigerator between doses. Thermal cycling simulates distribution excursions, so that a shipment briefly out of refrigeration does not automatically become a rejected consignment.

3. Bracketing and Matrixing

A full stability design tests every combination of batch, strength, container size and timepoint. That is expensive, and two reduced designs are permitted where justified.

DesignWhat it reducesWhen it is defensible
BracketingTests only the extremes of a variable — smallest and largest container, highest and lowest strength — and infers the middle.Where the extremes genuinely bound the behavior. Requires that intermediate configurations are compositionally identical.
MatrixingTests a subset of configurations at each timepoint, rotating across the schedule.Where variability between configurations is expected to be low, and enough timepoints remain to establish a trend.
This program uses neither, and the reason is worth stating. VitaFlow launches as a single strength in a single presentation — one prefilled pen, one concentration. There is nothing to bracket and nothing to matrix.

That is a hidden benefit of the narrow launch scope set in the Target Product Profile. A program launching three strengths in two container types would face a stability design several times larger, and would likely have had to justify a reduced design to the agency — another argument to have, in a review where argument costs time.

3. Position at 18 Months

AttributeSpecificationPosition at 18 months
AppearanceClear, colorless to slightly yellowconforms at all timepoints
Assay95.0–105.0% of label99.1% at 18 months — no meaningful trend
Related substances, total≤3.0%1.4% at 18 months, rising ~0.05%/quarter
Aggregates (SEC)≤1.0%0.3% at 18 months — the attribute to watch
pH6.8–7.47.1, stable
SterilitySterileconforms
Deliverable volumeWithin device specificationconforms
The attribute to watch is aggregation, and the reason is clinical rather than chemical. Peptide aggregates can affect immunogenicity — the likelihood that a patient's immune system reacts to the product. At 0.3% against a ≤1.0% limit there is comfortable margin, and the trend is flat.

Related substances are rising at roughly 0.05% per quarter from 1.4% against a ≤3.0% limit. Extrapolated naively, that reaches the limit somewhere past five years — comfortably beyond the 24-month claim. Degradation is rarely linear, which is why the claim rests on measured data rather than on the extrapolation.

5. When a Result Goes Out of Trend

Two different failures, frequently confused, with different consequences.

DefinitionConsequence
Out of specification (OOS)A result outside the acceptance criterion.Formal investigation. If confirmed on the primary batches, the shelf-life claim is affected and marketed batches may be implicated.
Out of trend (OOT)A result within specification but inconsistent with the established trend.Investigated as a signal. No immediate product impact, but it may predict an OOS at a later timepoint.
OOT is the one that carries information, and the one most often ignored. A result inside specification feels like a pass. But stability is a trend study — the question is not whether month 18 conforms, it is whether the trajectory from month 0 to month 18 supports a claim at month 24.

A single in-specification result that breaks trend can indicate an analytical problem, a batch difference, or the beginning of a degradation pathway that was not previously visible. Waiting for it to become an OOS wastes the warning.

An OOS on a stability batch after approval triggers a different chain entirely: investigation, assessment of marketed inventory, and potentially a field alert or recall. That is why the ongoing commitment at §7 is a regulatory obligation rather than a monitoring courtesy.

4. The Shelf-Life Claim

The program seeks 24 months refrigerated plus 28 days in use at room temperature. Today's data supports 18 months refrigerated.

PointLong-term dataSupports
Data available today18 monthssupports an 18-month shelf life
Data at Gate 5 submission~24 monthssupports the 24-month claim sought
Data at the action date~30 monthsmargin against the claim
Extrapolation under ICH Q1ELimitedExtrapolation beyond available long-term data is permitted only where accelerated data show no significant change and the long-term trend is minimal. This program does not rely on extrapolation for its 24-month claim — the data will be there.
Extrapolation is possible and this program is not relying on it. Under ICH Q1E a sponsor may, in defined circumstances, claim a shelf life somewhat beyond the available long-term data — where accelerated data show no significant change and the long-term trend is minimal.

It is a legitimate tool and it carries a cost: a claim resting on extrapolation invites review questions, and a review question is time the program does not control. With 24 months of data expected by the September 2028 submission, the program is choosing to file on measured data. That choice is only available because stability started early enough.

The in-use claim is the one genuinely at risk. Room-temperature in-use data generating; 28-day claim not yet supported. If the 28-day claim is not supported, the fallback is a shorter in-use period — a real commercial loss, since it constrains how patients can carry and use the pen, but not a filing obstacle.

7. Stability and Distribution

Shelf life assumes storage conditions are maintained. Distribution is where that assumption meets reality.

StudyProtects
Thermal cyclingRepeated excursions between refrigerated and ambient during transit and handling.
In-use stabilityThe 28-day room-temperature period after a patient starts a pen.
Short-term excursion dataAssessment of a specific shipment that went out of range — without it, an excursion means automatic rejection of the consignment.
Excursion data is a commercial asset, not just a scientific one. A cold-chain product without it must reject any consignment that leaves its range, however briefly. With it, a sponsor can assess a specific excursion against evidence and release stock that is genuinely unaffected.

The difference is measured in destroyed inventory and stock-outs. This is a case where a modest study, run years before launch, prevents an operational problem that would otherwise recur for the life of the product — and it only exists if someone thought about distribution while designing a stability program.

5. The Timing Decision

Stability on registration batches began before Phase 3 read-out. That was a governance decision rather than a technical one, and it is worth examining because it is the kind of choice a program manager actually influences.

OptionEffect on the filing dateEffect on cost
Start early (what this program did)Protects the Sep 2028 submission. 24 months of data available without extrapolation.Spends manufacturing and testing cost on a program with a 50% conditional probability of approval.
Start at Phase 3 read-outData would be roughly 12 months at submission. Claim would rest on extrapolation, or the shelf life would be shorter.Saves that spend on the ~50% of outcomes where the program does not proceed.
Start after toplineSubmission delayed by 12–18 months, or filed with a materially shorter shelf life.Minimal spend at risk.
There is no technically correct answer, and that is the point. This is not a question Technical Operations can resolve, because it is not a question about chemistry. It is a question about how much a filing date is worth against capital spent on an asset that may fail — which is precisely the kind of trade a gate exists to put in front of people with the authority to make it.

The decision was taken at Gate 4 as part of the Stage 4 authorization. It cost money the program would not have needed if Phase 3 fails. It is also the reason the program can file on measured data rather than argue for extrapolation.

6. Post-Approval Obligations

Stability does not end at approval.

The pattern repeats across CMC. Process validation continues for the life of the product (Stage 3 in the Process Validation Plan). Stability continues for the life of the product. Neither is a project that closes at launch.

For a program manager this matters at closure: the Post-Launch Review and the program closure report have to hand these obligations to a permanent owner, because the program that created them will not exist to carry them.

7. Governance Position

QuestionAnswer
Is the shelf-life claim at risk?No — data will reach 24 months before submission on the current schedule.
Is the in-use claim at risk?Yes — 28-day room-temperature data still generating; commercial consequence only, not a filing obstacle.
Does the third registration batch affect the shelf life?Not the claim, but the package — three primary batches are expected in the dossier. Two would invite a review question.
Can anything recover a stability slip?No — this is the one workstream with no recovery lever. Anything that restarts the clock costs months no budget can buy back.

The third row connects back to GC-01. The blocked registration batch does not threaten the 24-month claim, which rests on batches already on study. It threatens the completeness of the stability package — and a dossier presenting two primary batches where three are expected is not a rejection, but it is a question, and questions consume review time the program does not control.