PharmaCalc

Stability testing evaluates the degradation of an active pharmaceutical ingredient (API) and the overall product quality over time under defined storage conditions. ICH Q1E guidance provides the statistical methodology for analyzing stability data and establishing the shelf life (expiration dating period) that ensures the product meets specification throughout its labeled storage period.

Stability Testing Conditions (ICH Q1A/Q1B):

• Long-Term: 25°C ± 2°C / 60% RH ± 5% RH. Typically 0, 3, 6, 9, 12, 18, 24 months (up to 36 months for 3-year shelf life).
• Intermediate: 30°C ± 2°C / 75% RH ± 5% RH. Optional; performed in certain regions.
• Accelerated: 40°C ± 2°C / 75% RH ± 5% RH. Typical duration: 0, 3, 6 months. Used to detect potential degradation pathways.

Regulatory Requirements:

• Stability data is mandatory for all NDA, ANDA, and MAA (Marketing Authorization Application) submissions.
• The lower 95% confidence limit at the proposed expiration date must not exceed the lower acceptance limit (e.g., if lower limit is 90%, assay should not drop below 90%).
• Long-term data through the proposed shelf life is required in the initial submission.
• Ongoing stability monitoring continues after approval (post-approval stability commitment).

ICH Q1E recommends linear regression of assay results (% label claim) versus time (months). The slope indicates degradation rate. The shelf life is the time at which the lower 95% confidence limit of the regression line intersects the acceptance limit.

Linear Regression Equation:

Where ŷ = predicted assay (% LC), a = intercept (assay at t=0), b = slope (%/month), t = time (months).

Lower 95% Confidence Interval:

Where:

• t(α,n−2): Critical t-value (2-tailed, α=0.05, df = n−2). For n=7: t ≈ 2.571.
• s: Standard error of the regression (residual SD).
• t̄: Mean of time points.

Shelf Life Determination:

The calculated shelf life is typically rounded down to a standard interval (12, 24, 36, 48, 60 months) for commercial labeling.

Scenario: A pharmaceutical company is establishing the shelf life for a tablet product containing aspirin. Long-term stability data has been collected at 25°C/60% RH at 7 time points over 24 months. The acceptance limit is 90% LC. Linear regression and 95% confidence interval analysis are used to determine the expiration date.

1. Calculate Regression Statistics:
   Using linear regression (least squares method):
   n = 7 time points
   t̄ = (0+3+6+9+12+18+24)/7 = 72/7 = 10.29 months
   ȳ = (100.2+99.8+99.1+98.6+98.0+97.1+96.3)/7 = 689.1/7 = 98.44%
2. Calculate Slope (b) and Intercept (a):
   Using standard regression formulas (detailed algebra omitted for brevity):
   b ≈ −0.174 %/month (negative slope = degradation)
   a ≈ 100.4 % (intercept at t=0)
3. Calculate Fit Quality (R²):
   R² ≈ 0.9985 (excellent fit; nearly perfect linearity)
4. Calculate Standard Error and Residual SD:
   Residual sum of squares (RSS) ≈ 0.85
   s (residual SD) = √[RSS/(n−2)] = √[0.85/5] ≈ 0.41%
5. Determine Shelf Life (Lower 95% CI = 90%):
   At t = shelf life, the predicted assay is:
   ŷ = 100.4 + (−0.174) × t
   
   The lower 95% CI at that time point must equal 90%:
   90% = ŷ − 2.571 × 0.41 × √[1/7 + (t − 10.29)²/Σ(tᵢ − t̄)²]
   
   Solving iteratively (or using regression software):
   t (shelf life) ≈ 59.8 months (raw calculation)
6. Apply Commercial Rounding:
   Raw shelf life: 59.8 months
   Rounded down to standard interval: 48 months (4 years)

Interpretation: The stability data shows a linear, predictable decline in assay at approximately 0.174% per month at 25°C/60% RH. With this degradation rate, the product's lower 95% confidence limit reaches the 90% acceptance limit at approximately 60 months. For commercial labeling and a conservative margin, a 48-month (4-year) shelf life is assigned. The product will meet the 90% LC specification throughout this period when stored at room temperature.