Preservatives and the oral mucosa

Why preservatives exist

Microorganisms need water, and specifically they need water that is chemically available rather than bound up by solutes. The measure of this is water activity, written a_w, running from 0 to 1.0, where pure water is 1.0. Most bacteria stop growing below about 0.9, most yeasts and molds below about 0.7, and below roughly 0.6 microorganisms cannot proliferate.¹⁰

A conventional toothpaste is largely water plus humectants such as glycerin or sorbitol, giving it a water activity comfortably inside the range that supports growth. It is then stored in a bathroom, opened twice a day, and touched by a brush that has been in a mouth. Without a preservative system, that is a reliable route to a contaminated product.

The consequences are not hypothetical. Contaminated personal care products have caused infections, and the risk falls hardest on people who are immunocompromised or who have breaks in the oral mucosa. FDA guidance tells cosmetic manufacturers to verify the adequacy of preservation under reasonably foreseeable storage and consumer use. One way to do that is antimicrobial effectiveness testing, in which a product is deliberately inoculated with specified organisms and the surviving population is counted over time. Water activity testing answers a related but different question: whether the formulation contains enough available water for microorganisms to multiply.⁹

The systems in use

Several chemistries dominate oral care and cosmetics generally.

Benzoates and sorbates. Sodium benzoate and potassium sorbate are organic acid preservatives that work by crossing the microbial membrane in their undissociated form and acidifying the cytoplasm. They are pH dependent and only effective in acidic formulations, which constrains where they can be used.

Parabens. Esters of para-hydroxybenzoic acid, effective across a broad spectrum and a wide pH range. They became the most publicly contested preservative class following research on weak estrogenic activity. The honest summary is that parabens do show estrogenic activity in laboratory assays, that the potency is orders of magnitude below endogenous estradiol, and that assessments of systemic absorption and estrogenic potential at consumer-relevant exposures continue to be published and debated.¹ This is a live scientific question rather than a settled one in either direction.

Phenoxyethanol. A glycol ether, widely adopted as the primary paraben replacement, effective and generally well tolerated but not without its own sensitization reports.

Formaldehyde releasers and isothiazolinones. Both classes are effective and both carry substantial documented sensitization burdens. Systematic reviews and meta-analyses have quantified contact allergy prevalence to formaldehyde and its releasers² and to isothiazolinones,³ and these are among the better-characterized allergen classes in personal care.

Chlorhexidine is a different case and worth separating out. In oral care it is usually an active antiseptic rather than a preservative, and its adverse effect profile, including staining, taste disturbance and mucosal reactions, is documented in reviews of adverse events from home mouthrinse use.⁴

What the oral tolerance evidence shows

Toothpaste-induced contact stomatitis is a recognized clinical entity. A retrospective case series with an accompanying scoping review characterized it across 56 cases, describing presentation and the ingredients implicated.⁵ Reactions described in this literature include mucosal erythema, burning, desquamation, and perioral or lip involvement.

Two things need saying about how to read this. First, case series describe people who presented with a problem; they cannot tell you the rate in the general population. Set against the number of people who brush their teeth twice daily without incident, these reactions are uncommon. Second, they are real, reproducible on patch testing in individual patients, and worth knowing about if you are the person experiencing one.

The methodological difficulty running through the whole field is that patch testing is performed on skin. As covered in our article on surfactants and the oral mucosa, the non-keratinized oral mucosa differs from skin in barrier structure, permeability and immune surveillance. A positive skin patch test does not guarantee an oral reaction, and a negative one does not exclude it. Reviews of contact allergy in cosmetics continue to track the expanding allergen list,^6,7 but oral-specific data remains thinner than skin data.

In vitro work comparing the irritant potency of toothpaste ingredients on a common measure is one route toward better-grounded comparisons.⁸

Removing the requirement instead of meeting it

There is a third option beyond choosing a preservative or going without one, and it is a chemistry problem rather than an ingredient problem: formulate so that there is no free water for microorganisms to use.

An anhydrous formulation, one built on oils, fats or other non-aqueous carriers rather than water, can have a water activity far below the threshold that supports microbial growth. At sufficiently low water activity, organisms cannot multiply even though some may remain viable. A preservative system may then be unnecessary. This is the same principle that helps keep honey, dry spices and anhydrous ointments stable without added preservatives.¹¹

The trade-offs are real and worth naming. Anhydrous formulations behave differently in the mouth, have a different texture, cannot use water-soluble actives in the usual way, and are generally more expensive to produce per unit. They also still require verification rather than assumption: water activity should be measured, and microbiological testing performed on finished product, rather than the absence of water being asserted.

We make an anhydrous toothpaste, so we have an obvious interest in this argument. The chemistry stands independently of that, and readers should weigh the source accordingly. The relevant evidence is our finished-product water activity and microbiology report, published in full.¹²

What the finished-product testing shows

MCL Food Testing analyzed a finished toothpaste sample on June 17, 2026. The result was a water activity of 0.352, well below the roughly 0.60 growth floor described in the scientific literature and used in FDA guidance to distinguish products of greater concern for microbial proliferation.¹⁰ The same report found very low microbial counts.

These are complementary findings. The count results show that the tested sample had no measured category at or above 10 CFU/g. The water activity result explains why multiplication is strongly constrained in the formula even without an added preservative. Neither result means absolute zero organisms: a result of <10 is bounded by the method's reporting limit, and low water activity can stop proliferation without instantly killing every organism present.

What this page does not claim

• Preservatives are not presumptively harmful. They exist to prevent a documented safety problem, and for water-containing products they are the responsible choice.
• Preservative-free is not automatically safer. On a water-containing product it may be considerably less safe. The relevant question is water activity and verified microbiological stability, not the ingredient list.
• No claim is made here about disease. Nothing on this page says that any preservative causes illness, or that avoiding preservatives treats or prevents any condition.
• Contact reactions are uncommon. The case literature describes real phenomena, but case series cannot establish population rates, and most people use preserved oral care products without any reaction.