Topic · Stability
“The accelerated arm is back, and the project team wants a shelf life.”
Shelf life and stability studies
Which rules from ICH Q1A(R2) and Q1E apply when stability data come in, and how CovaSyn makes the evaluation reproducible: check significant change, compute shelf life from long-term data, see trends early.
Regulation
Obligations and deadlines
What the guidelines set out, as stated in our Insights articles. The table does not replace the guideline text; check your case against the decision trees.
| Framework | What applies | Threshold or deadline |
|---|---|---|
| ICH Q1A(R2), forced degradation | On one batch of drug substance: temperature in 10 °C steps above accelerated conditions, humidity where relevant, oxidation, hydrolysis across a pH range and photolysis (Q1B). The results should identify likely degradation products. | No target degradation level is set; 5 to 20 % per condition is common practice |
| ICH Q1A(R2), significant change for a drug product | Assay change from the initial value, any degradation product exceeding its acceptance criterion, or failure of appearance, physical attributes, functionality, pH or dissolution for 12 dosage units. | 5 % from the initial value, regardless of the specification limit |
| ICH Q1A(R2), significant change for a drug substance | Significant change means failure to meet the specification. | No 5 % assay rule |
| ICH Q1A(R2), intermediate condition | Significant change in the accelerated arm with long-term storage at 25 °C / 60 % RH: test at 30 °C / 65 % RH. If the long-term condition is already 30 °C / 65 % RH, there is no intermediate condition. | New filing: at least 6 months of data from a 12-month study, at least three primary batches, time points 0, 6, 9, 12 |
| ICH Q1A(R2), refrigerated products | Significant change in the accelerated arm (25 °C) between months 3 and 6: shelf life from long-term data. Within the first 3 months: a discussion of short excursions from the label storage condition, supported by additional testing on one batch. | Month 3 and month 6 |
| ICH Q1E, shelf life and extrapolation | Shelf life comes from regression of long-term data against the one-sided 95 % confidence bound. Batch poolability is tested by ANCOVA at the 0.25 significance level. | Without significant change up to 2 times the long-term period, at most +12 months; after significant change at accelerated conditions up to 1.5 times, at most +6 months |
Sources: our articles on significant change, on ICH Q1E shelf-life calculation and on forced degradation study design, linked below.
CovaSyn
How CovaSyn does it
The computations are deterministic and version-pinned: same data, same result, weeks later too. The assessment and what goes into the filing stay with you.
Check significant change
CovaSyn flags the 5 % threshold per storage condition and reports value and time point. Which decision tree applies is your call.
Shelf life under ICH Q1E
Regression of long-term data against the one-sided 95 % confidence bound, pooled across batches where poolability allows. Arrhenius serves as supporting kinetics.
Plan early instead of catching up
The Arrhenius fit shows before the accelerated data arrive whether 40 °C is likely to trigger. You then place the intermediate arm on time.
Detect OOS and OOT
Out-of-specification values are flagged, and trends in the series show up before a point breaks the limit.
Design forced degradation
The fitted degradation rate per condition gives stress times that land each arm in the 5 to 20 % window. Mass balance comes with it.
Traceable for inspection
Every computation is logged with timestamp, tool and version, and outputs carry a SHA-256 checksum.
What the computation does not replace
- An Arrhenius prediction is not a shelf life. It supports decisions; it is not a filing number.
- Kinetic modelling does not replace the intermediate-condition data that ICH Q1A requires.
- Arrhenius assumes one dominant degradation mechanism across the temperature range. If a new pathway switches on at 40 °C, the extrapolation is wrong without R² showing it.
Insights
Read on and try it yourself
Significant Change Under ICH Q1A: What Triggers It
Significant change ICH Q1A explained: the exact assay, degradant, dissolution and physical triggers, and what intermediate testing you owe once it fires.
Read articleICH Q1E Shelf Life Calculation from Accelerated Data
ICH Q1E shelf life calculation from accelerated stability data: the method, a worked example with real Arrhenius output, and when extrapolation is not OK.
Read articleForced Degradation Study Design (ICH Q1A R2)
Forced degradation study design under ICH Q1A(R2): stressor choice, target degradation, mass balance and stability-indicating methods, with real tool output.
Read articleOOT vs OOS in Stability Data: Catching Drift Early
Out of trend stability data explained: how OOT differs from OOS, when regression, CUSUM and Shewhart methods fire, and a worked CovaStab example.
Read articleMass Balance in Forced Degradation, and Why It Fails
Why mass balance in forced degradation rarely closes: non-chromophoric and volatile degradants, adsorption and response factors, plus how to close the gap.
Read articleShelf-life prediction with Arrhenius: from weeks of data to a shelf-life estimate in minutes
From a few weeks of accelerated stability data to a shelf-life estimate in minutes: on an amoxicillin example. Why the limiting parameter is not assay. And why the tool itself tells you when not to rely on extrapolation alone.
Read article
Questions
Common questions
Is 5 % a change from the initial value or from the specification?
From the batch's initial value. A batch that starts at 100.0 % and reaches 94.0 % has a significant change even if the lower specification limit is 90.0 %.
Can a good Arrhenius model replace the intermediate study?
No. ICH Q1A requires generated data at the intermediate condition. Kinetic modelling helps you place the intermediate arm on time, not argue it away.
Next step
Run your own stability series.
Create an account, get your API key, use the tools in Claude, ChatGPT or Cursor.