FDA criticises Non-Application of Statistical Process Control (SPC) in Validation

With the introduction of the updated FDA guidance on process validation, a process validation life cycle was introduced in 2011. One of the stages in the cycle is the so-called Continued Process Verification stage 3, which shows whether the process remains permanently in a validated state. Many companies use statistical process control (SPC) to show the 'state of control' in this stage 3.  

In a recent Warning Letter, the FDA criticised a drug manufacturer's statistical process control. What was criticised?

No control charts created

The methods used in the 'Continued Process Verification' to show the 'state of control' are not specified. One possibility mentioned in the Process Validation Guidance is statistical process control. This is what the company inspected by the FDA had planned.

Was it intended that way? Yes, that is how it was intended. Citing section 21 CFR 210/211, the FDA criticises the lack of activities that show that the process is permanently 'under control'. They are referring to the manufacturing and engineering departments, which have not created control charts to monitor and control the manufacturing process. This is exactly what the company had specified in its internal procedures. This was rectified during the inspection, with the result that four previously undetected trends were discovered. The usefulness of an SPC was thus demonstrated directly during the inspection.

The inspected company also recognised this and responded to the inspection report by stating that it was committed to complying with its internal procedures on SPC in future. This was not enough for the FDA. It also expects an investigation into how it was possible that the control charts were not created in the first place.

Scientific approaches in the context of process validation

One of the innovations in the update of the FDA Process Validation Guidance was the introduction of a more scientific approach to process validation. This is also clearly reflected in this Warning Letter. The FDA criticises an unscientific approach to a validation run. The mixing uniformity was criticised as insufficient in terms of sample locations and sampling (number, sample size, sample preparation).

Conclusion: Statistical process control is not a mandatory requirement in the context of process validation. However, if it is required by internal company specifications, then it must also be implemented. And in this case, the advantage of an SPC was demonstrated very impressively, which was then followed up. Four previously undiscovered trends were identified. Furthermore, a scientific approach to process validation is required.

You can find the entire Warning Letter on the FDA website.

The ECA will conduct a Live Online SPC Training on 08/09 October 2024.