Do Particulate-Inspection Methods Need Scrutiny?

In recent months, there has been an increase in product recalls because of particulate contamination. Throughout the production process, pharmaceutical manufacturers take various measures to keep foreign particles out of their drug products. In case any of these measures fails, drugmakers perform inspections intended to detect foreign particles and prevent contaminated products from being released. The recent recalls, including that of American Regent’s caffeine and sodium benzoate injection on May 5, 2011, highlight the importance of particulate inspection, and they might lead observers to ask whether current inspection methods are sufficiently effective.

“A firm should evaluate the effectiveness of its 100% final-inspection processes to identify any areas for improvement,” says a spokesperson for the Office of Compliance within FDA’s Center for Drug Evaluation and Research. “Improvement would seem to be indicated if quality-assurance samples of finished products showed an adverse pattern of visible-particulate detection even after 100% final inspection had been done.”

The reference inspection method described in all of the pharmacopeias is human visual inspection. These documents outline the recommended inspection conditions in various levels of detail. “In many respects, visual inspection is the very backbone of product release,” says Ronald Iacocca, senior research advisor for physical and structural characterization at Eli Lilly and Company. “Some individuals can see particles as small as 50 μm. The human eye is indeed the perfect camera,” he adds.

In addition, FDA permits drugmakers to use alternative methods, including automated inspection systems, as long as they are validated appropriately.

From companies’ perspective, “a large volume of a few product types is desirable to justify the significant capital investment and validation expense of an automated inspection system,” says John Shabushnig, senior manager at Pfizer and leader of the Parenteral Drug Association’s (PDA’s) Visual Inspection Interest Group and member of the US Pharmacopeia’s Visual Inspection Expert Panel. According to a 2008 PDA survey, 33% of firms used manual methods for particulate-matter inspection, 24% used semiautomated methods (e.g., a human inspector and a machine for container transport), and 43% used automated systems, says Shabushnig.

Human and machine inspection techniques are effective, but neither is perfect. Each method’s sensitivity depends on the size and location of the particle, and on the amount of contrast. Manufacturers must control variables such as lighting type, lighting intensity, and inspection time to achieve reliable and reproducible inspection results.

An analysis of several studies involving various groups of inspectors revealed that the probability that a human will detect a single 50-µm particle in clear solution in a 10-mL vial with diffuse illumination between 2000 and 3000 lux is slightly greater than 0%, says Shabushnig. The probability increases to approximately 40% for a 100-µm particle and becomes greater than 95% for particles equal to or greater than 200 µm. The results vary with the formulation and container type.

In general, the sensitivity of automated inspection systems is similar to that of human visual inspection, but automation can provide greater consistency than humans can. Both methods provide good sensitivity when they have been qualified or validated appropriately. “Current inspection methods are adequate when properly qualified and operated in a controlled manner, but there are still opportunities for improvement,” says Shabushnig.

Fortunately, equipment manufacturers are working to design better inspection devices. For example, companies are continuing to develop technologies that see inside powders and lyophilized cakes, as well as through opaque suspensions. “A practical, commercial system may be available in the near future to fill this gap in our current inspection capabilities,” says Shabushnig.

Improved inspection algorithms could reduce the time needed to train an inspection system to examine a new product or container. This training currently is time consuming and expensive. In addition, advances in the ability to inspect opaque or translucent containers and products would help drugmakers. “We would also benefit from advances in image processing to reduce the number of false rejects associated with automated inspection systems,” says Shabushnig.

But pharmaceutical professionals say that current techniques and methodologies for particulate inspection are generally effective, even if they have not achieved perfection. “Industry has demonstrated good performance with manual and automated inspection methods. The important point is that any inspection method, manual or automated, must be well characterized and controlled,” says Shabushnig.