How Is Sample Size Calculated for CCIT?

Sample size for Container Closure Integrity Testing (CCIT) release testing is calculated using AQL-based attribute sampling, aligned with ANSI/ASQ Z1.4 at General Inspection Level II. Each unit is classified as pass or fail against a validated, product-specific threshold. The plan specifies a sample size n and an acceptance number c.

What is the acceptance number, and why is c = 0 recommended?

The acceptance number is the maximum number of failures permitted before the lot is held. For sterile injectable CCI, a zero-acceptance-number (c = 0) plan is recommended. A single confirmed failure triggers the defined control strategy. This provides the strongest statistical protection for a given sample size, consistent with the patient risk of a compromised sterile closure.

The core relationship is: P(accept) = (1 – p)n — where p is the true defect fraction and n is sample size. At n = 315, a 1% defective lot is accepted only 4.4% of the time. At n = 800, that same lot is accepted less than 0.03% of the time.

What sample sizes apply across typical batch sizes?

Why is 100% testing recommended for lots of 50 or fewer?

For very small lots — clinical batches, early commercial runs, stability samples — standard AQL table structures provide poor statistical protection at small n. Since deterministic CCI methods are non-destructive, 100% inspection is operationally feasible, eliminates sampling uncertainty entirely, and provides maximum regulatory defensibility.

Does sample size guarantee batch quality?

No. It determines the probability of detecting a defective batch. Sampling cannot create quality — it provides evidence that a validated, well-controlled process is performing as expected. The foundation is a validated method and a capable process.