What ISO 22716 does and does not tell you about microbial limits
QA teams reviewing ISO 22716 for the first time often go looking for a table of CFU limits or a list of organisms that must be absent from a finished product. They don't find one, and that gap causes confusion during audits and supplier reviews.
ISO 22716:2007, titled Cosmetics — Good Manufacturing Practices (GMP) — Guidelines on Good Manufacturing Practices, covers the production, control, storage, and shipment of cosmetic products. It does not set numeric acceptance criteria for bacteria, yeast, or mould counts, and it excludes research and development activity along with distribution of finished goods. Its scope is process control, not product specification.
That distinction has real consequences. In 2024, Suntegrity Skincare voluntarily recalled nine lots of its Impeccable Skin Sunscreen Foundation after inspectors found higher-than-acceptable levels of the mould Aspergillus sydowii in some tubes. Contamination like this is exactly what a GMP system is meant to catch before product reaches a shelf, but the specific pass/fail threshold that flags a batch as out of spec comes from elsewhere.
The numeric limits and pathogen-absence criteria that manufacturers test against are defined in ISO 17516 and regional cosmetic regulation, applied inside an ISO 22716-compliant manufacturing system. The sections that follow separate the two: how ISO 22716 governs process, and how ISO 17516 supplies the numbers that process is built to meet.
What ISO 22716 actually regulates
ISO 22716's main subject areas cover personnel training and hygiene, premises design, equipment maintenance, control of raw and packaging materials, production, storage and shipment of finished products, laboratory quality control, handling of nonconforming products, waste management, subcontracting, deviations, complaints and recalls, change control, internal audit, and documentation. That list is a process map, not a specification sheet.
Each of these areas reduces microbiological risk indirectly. Personnel hygiene procedures limit skin-flora transfer into product. Premises and equipment controls prevent cross-contamination between batches and biofilm buildup in poorly cleaned lines. Raw material controls catch contaminated inputs before they enter a formulation. Laboratory quality control sections require testing and record-keeping, but the standard leaves the choice of test methods and pass/fail thresholds to the manufacturer, informed by applicable microbiological standards, regulatory requirements, and product risk.
This is where manufacturers sometimes assume compliance with ISO 22716 alone satisfies microbiological release testing. It doesn't. A facility can run flawless GMP documentation, clean rooms, trained staff, controlled deviations, and still ship a batch that fails a microbial count if the acceptance criteria applied at QC were never defined against a recognized standard. ISO 22716 builds the conditions under which contamination is unlikely; it does not tell a QA manager what number on a lab report means the batch passes.
That gap is closed by pairing GMP process control with laboratory testing against defined criteria, which is the practical work covered in pharmaceutical, cosmetic and toiletries analysis. The next section covers where those numeric thresholds actually come from.
Where the CFU numbers actually come from
Microbiological Limits for finished cosmetics are set out in ISO 17516, not ISO 22716. ISO 17516:2014 splits products into two categories based on intended use and applies different total aerobic mesophilic count thresholds to each.
Category 1 covers products intended for children under 3 years of age, the eye area, or mucous membranes. These carry a tighter acceptance criterion, commonly cited as no more than 10² CFU per gram or millilitre. Category 2 covers everything else, general-purpose products like body lotions or shampoos, with a criterion of no more than 10³ CFU per gram or millilitre. Both counts capture aerobic mesophilic bacteria along with yeasts and moulds, not a single organism type.
The tighter limit for Category 1 reflects exposure risk rather than formulation difficulty. A contaminated eye cream reaches mucous membrane tissue with far less biological defense than skin does, so the acceptable bioburden is lower regardless of how clean the production run was.
None of this originates in ISO 22716 text. ISO 22716 tells a manufacturer how to build a facility and process that make contamination unlikely. ISO 17516, alongside applicable regional cosmetic regulation, tells the QA team what number on the certificate of analysis means the batch can ship. Malaysian manufacturers working under NPRA requirements need both read together, a point covered in more detail in this discussion of Malaysian cosmetic regulatory obligations.
A batch testing at 150 CFU/g might pass comfortably as Category 2 and fail outright as Category 1. Product classification, not just the lab result, decides which threshold applies, which makes accurate labelling and intended-use documentation part of the acceptance criteria conversation, not a separate administrative task.
Specified organisms: why absence, not counts, is the criterion
Category 1 and Category 2 limits address total bioburden. A separate requirement under ISO 17516 asks for absence of specific organisms regardless of product category. These are qualitative pass/fail criteria, not permitted counts: a batch either has detectable Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus, or Candida albicans in a 1 g or 1 mL sample, or it doesn't.
Each organism maps to a distinct contamination pathway. P. aeruginosa thrives in water-based formulations and poorly maintained water systems, which is why ISO 22717 gives it dedicated detection guidance. S. aureus points to human skin or mucosal contact during handling. C. albicans signals fungal contamination, often from raw material or packaging sources. E. coli indicates faecal or hygiene-related contamination, a red flag for personnel or facility sanitation gaps that ISO 22716's premises and personnel clauses are meant to prevent in the first place.
Detection follows a staged process rather than a single plate reading. ISO 18415 describes an initial non-selective liquid enrichment step, which increases the recovery of stressed or low-level organisms and avoids inhibition from preservatives or surfactants in the formulation. Enrichment is followed by subculture onto selective or differential agar to isolate suspect colonies. From there, laboratories working from methods aligned with the FDA's Bacteriological Analytical Manual for cosmetics purify the isolate, run Gram staining, and confirm species identity using commercial identification systems such as API or Vitek strips. A tentative positive on selective media is not a confirmed result until this identification step closes it out.
A tentative positive is not a finding until it survives enrichment, isolation, and confirmatory identification.
Because contamination in a batch is rarely distributed evenly, testing a single sample unit can miss a pocket of organisms sitting in one corner of a tank or one filled unit. Plurality testing, drawing multiple independent units or portions rather than relying on one composite sample, raises the odds of catching uneven or low-level contamination that a single draw would pass by chance. It supplements a validated method; it does not replace one, and it does not substitute for the preservative system doing its job over the product's in-use life, a topic covered in more detail in this review of preservative efficacy testing in multi-use cosmetics.
Third-party testing as evidence, not just paperwork
An internal QC pass is useful, but auditors and regulators generally weigh independent verification differently. Third-party laboratory testing gives a manufacturer's ISO 22716 GMP system an external check against the acceptance criteria and specified-organism absence requirements drawn from ISO 17516 and regional cosmetic regulation.
That check is only as good as the laboratory's demonstrated competence. ISO/IEC 17025 accreditation assesses a laboratory's technical competence across method validity, staff competence, equipment maintenance, measurement traceability, and quality control, giving independent assurance that reported results are reliable. Accreditation is scope-specific: it names the exact methods, analytes, and matrices a laboratory is recognised as competent to test, not a blanket guarantee covering every service on a lab's menu. Manufacturers should confirm that a candidate laboratory's current accreditation scope actually covers the enumeration and specified-organism methods relevant to cosmetic products, rather than assuming coverage from a general accreditation claim.
KAS Lab, an ISO/IEC 17025-accredited laboratory based in Malaysia, offers pharmaceutical, cosmetic, and toiletries testing services covering microbial enumeration and specified-organism detection, as well as preservative efficacy and challenge testing. Manufacturers should verify the specific scope against their product's testing needs before relying on results for regulatory submissions or recall-defense documentation.
Making the three controls work as one system
ISO 22716 GMP practices, ISO 17516 acceptance criteria, and specified-pathogen screening function as one control system rather than three separate checkboxes. GMP prevents contamination at the source. Finished-product limits and pathogen-absence testing verify whether prevention worked. A result outside acceptance criteria, or the detection of a specified organism where it should be absent, should trigger an investigation, batch disposition, and a documented review of which other lots share the same raw material, equipment line, or production window, rather than routine release. That lot-scope assessment is what separates a contained deviation from a multi-lot recall, and it is the same logic behind the FDA's recall trigger of a reasonable probability of serious health consequences.
For QA teams, the practical implication is straightforward: build the deviation-to-disposition pathway before a failure happens, not after. Reviewing preservative efficacy alongside microbial limits is a related step worth checking, covered in evaluating preservative efficacy in multi-use cosmetic products.
