Antimicrobial effectiveness testing: what "bacteriostatic" establishes, and what dilution does to it

The word bacteriostatic arrives on a diluent label as a bare adjective, and it tends to be read as a general property of whatever ends up in the vial — as though the container, and anything subsequently mixed into it, had acquired a standing resistance to microbial growth. It is in fact a compendial claim with a narrow evidentiary basis: a challenge test, run on a defined formulation, against a defined organism panel, scored against defined acceptance criteria across twenty-eight days. Everything the adjective means is contained in that test, and the boundaries of the test are where the meaning stops. This post sets out how antimicrobial effectiveness testing is constructed, what its acceptance criteria actually encode, why a preservative is not a stabiliser, and why the inference from a preserved diluent to a reconstituted peptide solution is not automatic.

What the label term refers to

Bacteriostatic Water for Injection is water for injection with a preservative added — most commonly benzyl alcohol at 0.9%, which is 9 mg/mL, with 1.1% appearing on some presentations — supplied in a multiple-dose container from which repeated withdrawals may be made. The compendial article sits near pH 5.7, with a specified range of 4.5 to 7.0. The comparison with unpreserved sterile water covers the diluent chemistry and where each is appropriate; what concerns us here is narrower, namely the evidentiary status of the adjective itself.

Two attributes are routinely collapsed in discussion of these containers, and they are not the same. Sterility is a property of the article as it leaves the manufacturer, established by the manufacturing process and its validation. Preservation addresses organisms introduced after the container is first entered. A preservative does nothing about the first question and a sterilising process does nothing about the second. “Bacteriostatic” speaks only to the second, and it speaks about inhibition of multiplication rather than killing.

The challenge test

The compendial method behind the claim is antimicrobial effectiveness testing, USP General Chapter <51>. Its design is direct: inoculate the product deliberately, then count what survives.

The organism panel is five strains, specified by ATCC number so that the challenge is reproducible between laboratories — Staphylococcus aureus ATCC 6538, Pseudomonas aeruginosa ATCC 9027, Escherichia coli ATCC 8739, Candida albicans ATCC 10231, and Aspergillus brasiliensis ATCC 16404. Three bacteria spanning Gram-positive, Gram-negative, and enteric representatives; one yeast; one mould. Each is inoculated separately into the product at a final concentration on the order of 10⁵ to 10⁶ colony-forming units per millilitre, and the inoculated containers are held and sampled at intervals across a twenty-eight-day period, with viable counts determined at each time point.

Acceptance criteria depend on which product category the article falls into. Category 1 covers injections and other parenterals including emulsions, otic products, sterile nasal products, and ophthalmic products made with aqueous bases or vehicles — the strictest of the four categories, and the relevant one here. For Category 1, bacteria must show not less than a 1.0 log reduction from the initial calculated count at 7 days, not less than a 3.0 log reduction at 14 days, and no increase from the 14-day count at 28 days. For yeast and moulds, the requirement is only that there be no increase from the initial calculated count at 7, 14, and 28 days. “No increase” is itself defined, as not more than 0.5 log unit above the previous measured value — a tolerance that accounts for the ordinary imprecision of plate counting rather than granting latitude on the result.

What the asymmetry in the criteria encodes

The most informative feature of those criteria is the difference between the bacterial and the fungal requirement, and it is the difference that the prefix -static is naming.

Against bacteria, the article must demonstrate a reduction: a thousand-fold decrease by day 14, which is a cidal performance. Against yeast and mould, it need only demonstrate stasis — the population must not grow. An article can pass Category 1 with the same mould count on day 28 that it had on day 1, and that is a pass, not a marginal result. The claim being licensed is that introduced organisms will not proliferate to a meaningful degree over the period tested, with a stronger showing required for the bacterial panel than the fungal one.

This has a straightforward consequence for interpretation. A preserved diluent is not a sterilant, and a passing antimicrobial effectiveness result does not establish that an organism introduced into the container is eliminated. It establishes bounded behaviour of a defined challenge under defined conditions. Contamination outside that challenge — a heavy inoculum, an organism unlike the panel — sits outside what the test was built to describe. Related reasoning applies to the bacterial endotoxin question, where the analyte is a heat-stable structural fragment that survives conditions lethal to the organism that produced it, and where preservation is irrelevant by construction.

The test is run on a formulation, not on a solution you make later

Here is where the inference most often breaks. Antimicrobial effectiveness testing qualifies a formulation as manufactured. The article tested is the diluent as supplied, at its label preservative concentration, at its own pH, with its own composition. A reconstituted peptide solution is a different article, and nothing about the diluent’s result transfers to it automatically.

Several mechanisms drive the divergence, and all of them are ordinary formulation chemistry.

Concentration is the first and most obvious. Preservative activity is steeply concentration-dependent, and any addition to the container dilutes the preservative below its qualified level. The arithmetic of reconstitution applies to the excipient as much as to the peptide: the number on the diluent label describes the diluent, not the mixture.

Partitioning is the second. Benzyl alcohol is a small amphiphile, and its mode of action depends on reaching and disordering microbial membranes. Anything in the container that sequesters it reduces the concentration available to do that. Sorption of preservatives into elastomeric closures is a well-characterised effect and one of the reasons extractables and leachables work on stoppers is not purely about what migrates outward; migration runs in both directions. Surfactants, lipids, and hydrophobic peptide surfaces offer further partitioning compartments.

Solution conditions are the third. Preservative efficacy is pH-sensitive, and a peptide, its counterion, and any buffering species will move the pH of the reconstituted solution away from the diluent’s own value. That shift interacts with peptide charge state and solubility at the same time, which is why reconstitution conditions are a formulation variable rather than a procedural detail.

None of this means a reconstituted solution is unpreserved. It means the qualified statement attaches to the diluent, and the mixture’s behaviour is an open question unless someone has tested the mixture.

A preservative is not a stabiliser

The second common conflation is between microbiological and chemical stability. They are independent, and preservation addresses only the first.

Every chemical degradation pathway that operates in aqueous peptide solution continues to operate in a preserved one. Oxidation at methionine and tryptophan, deamidation at asparagine, disulfide scrambling, adsorptive loss to container surfaces, and photodegradation of aromatic residues are all indifferent to whether benzyl alcohol is present. This is the central reason the distinction between lyophilised and reconstituted storage matters more than the choice of diluent: the solid state suppresses mobility, and no additive to the aqueous phase reproduces that.

Benzyl alcohol is also not chemically inert. It oxidises to benzaldehyde and onward to benzoic acid, and benzaldehyde is an electrophile that can condense with primary amines — the N-terminus and lysine side chains — to form Schiff bases. Separately, studies of preserved protein formulations have characterised benzyl alcohol as capable of perturbing conformation and promoting aggregation in susceptible molecules, an effect sufficiently reproducible that preservative compatibility is a routine screen in parenteral formulation development. Whether this matters for a given sequence is an empirical question about that sequence, but the general point holds: a preservative adds a reactive small molecule to the system, which is a stability variable rather than a neutral one, and a reason forced degradation work is run on the intended formulation rather than on the peptide in water.

The twenty-eight-day convention, read correctly

The compendial position on multiple-dose containers is that antimicrobial agents in bacteriostatic or fungistatic concentration are to be present in preparations held in such containers, in adequate concentration at the time of use, to prevent multiplication of organisms inadvertently introduced during withdrawal of a portion of the contents. Paired with this is the familiar convention that a multiple-dose container is not used beyond 28 days after initial entry unless the manufacturer specifies otherwise.

Notice what that figure is and is not. It is a microbiological convention, aligned to the 28-day window of the antimicrobial effectiveness test, and it governs container entry. It is not a chemical stability statement about the contents. A peptide in solution may degrade measurably well inside 28 days, or be chemically unremarkable well beyond it; the number carries no information either way, because it was never derived from a stability study of that peptide. Retest and shelf-life figures come from a different exercise entirely — ICH Q1A-style stability programmes run on the specific material, under specified conditions, with stability-indicating methods. Two numbers with different provenance should not be read as one.

Where the evidence sits

The chain is short and worth stating in order. A preservative claim rests on a challenge test. The challenge test is run on the diluent as supplied. Its criteria require reduction of a bacterial panel and mere stasis of a fungal one, over 28 days, against a standardised inoculum. Reconstitution changes the concentration, the partitioning environment, and the pH of the system that was qualified. And none of it speaks to the chemical fate of the peptide, which is governed by mechanisms the preservative does not touch and, in one or two respects, may participate in.

Read that way, the adjective on the label is useful and honest: it describes a real, tested capability of a real article. What it cannot do is extend to a mixture nobody tested, or stand in for a stability figure it was never derived from. As with a certificate of analysis, the discipline is the same — identify which article the claim attaches to, identify the method behind it, and decline to carry the conclusion further than the method reaches.