According to the pharmacopoeia specifications, parenteralia must be subject to a 100% visual inspection. In addition to damages to the primary container, the absence of particles is also checked. But do these medicinal products also have to be 100% free of particles?
Basically yes, but particle-freedom only applies to visible particles. Small, so-called sub-visible particles may (and will) be present according to the pharmacopoeia specifications. For example, according to the European Pharmacopoeia, up to 25 particles (<= 10 µm) per mL and up to 3 particles (<= 25 µm) may be found in containers up to 100 mL using the Light Obscuration counting method.
A distinction must also be made between intrinsic and extrinsic particles. Intrinsic particles, i.e. particles originating from the medicinal product or API itself are generally undesirable but may also be part of the medicinal product and described in the marketing authorisation. This is the case, for example, for suspensions.
Extrinsic particles on the other hand (i.e. foreign particles originating from outside such as fibres, glass and abrasion of machine parts) are undesirable under any circumstances. The following applies: the sterile, liquid medicinal product should be free of particles. Now, it is technically impossible to produce a medicinal product completely free of particles. In other words, it is not possible to manufacture a 100% particle-free batch. Units with particles must therefore be sorted out. This sorting takes place within the framework of the required 100% inspection during production.
But particle testing is also problematic. Because even if by definition (100% inspection) every container of a batch is visually inspected for visible particles, it is a probabilistic inspection. This means that the detection probability for a visible particle is not 100%. This means that no matter whether the visual inspection is performed manually (by human) or using a fully automatic optical inspection machine, units with particles could be not detected. The probability of detection depends on many factors, such as the size of the particle, its colour, its opalescence, its shape (needle-shaped or round), its position in the container, etc. Despite the probabilistic nature of particle detection, all detected units must be removed and eliminated.
In summary, it is not possible to produce a batch without particles, but at the same time, not all particles or not all units with particles can be detected. Now the question is how does this fit in with the requirement that sterile medicinal products for injection should be free of particles?
Pharmacopoeias have to be read in detail. The requirement laid down in the pharmacopoeias is not "free of particles" but "practically/essentially free from visible foreign particles". For a long time, this expression was difficult to understand and, above all, difficult to measure. This changed in 2014 with the introduction of USP Chapter <790> Visible Particulates in Injections. It describes an AQL test which - if passed with an AQL of 0.65 or better - defines a batch as essentially free from visible particulates. In practice, this means that first a validated 100% visual inspection of all units of a batch is performed where all units with detected particles are sorted out. In order to meet the probabilistic character of this test, a sample is taken from the good part of the tested batch in accordance with the AQL test plan and checked again. The number of units with particles permitted here is determined and depends, for example, on the batch size. This inspection is not mandatory in Europe, but can be understood as the state of the art and is carried out in most pharmaceutical companies as part of the batch release. Otherwise, it would also be difficult for the inspector during an inspection to answer the question of how to deal with the fact that the 100% inspection carried out don't find all units with particulate impurities. Another approach would be to perform a second 100% visual inspection.
Coming back to the initial question: do liquid, sterile medicinal products have to be free of particle? Yes, they do. Nevertheless, a batch with units containing particles is very likely to be released but a unit in which particles have been detected must be rejected and cannot be released.