Bacteriostatic water vs sterile water

Last reviewed 26 August 2026

Two 10 ml vials of clear, colourless liquid. Both labelled water. The difference between them is one added substance, and almost everything else written about the pair follows from that single fact — including the number that causes more confusion in this category than anything else, the twenty-eight days.

Both names are titles of United States Pharmacopeia articles. What follows describes what those articles specify — composition, container designation, test regime — what the European Pharmacopoeia does and does not have to say about them, and what the published formulation literature reports about the preservative itself. Two terms recur and are worth fixing at the outset: single-dose and multiple-dose are container designations in the pharmacopoeial sense. They describe how many times a closure is expected to be entered. They are not quantities, and this page states none.

The difference in one line

Bacteriostatic water is water for injection with an antimicrobial preservative added, in a container designated for repeated entry. Sterile water is the same water with nothing added, in a container designated for a single operation. The container size limit, the labelling requirement, the 28-day convention and every compatibility question are all downstream of that.

What the two designations actually specify

The USP article Bacteriostatic Water for Injection is defined as water for injection containing one or more suitable antimicrobial agents. Note what that definition does not do: it does not name benzyl alcohol, and it does not fix a concentration. That is why the monograph carries a labelling requirement obliging the name and proportion of the added agent to appear on the container, and why both 0.9% w/v and 1.1% w/v benzyl alcohol presentations are marketed under the same article. The figure so often quoted as though it were part of the definition is in fact a property of a particular product, which is exactly the reason the label has to state it.

The packaging statement limits the container to not larger than 30 mL, and where glass is used it is preferably Type I or Type II. A 10 ml presentation sits comfortably inside that limit.

Sterile Water for Injection is the unadorned article: a sterile, non-pyrogenic, solute-free preparation, described on US labelling as containing no bacteriostat, no antimicrobial agent and no added buffer, packaged in single-dose containers. It is not a lower grade of water. It is the same water without the excipient.

Two adjacent USP articles are frequently conflated with it and should not be. Sterile Water for Irrigation and Sterile Water for Inhalation are separate articles with their own specifications and their own packaging statements. Sharing two words of a name does not make them interchangeable.

The European position, which is the one that matters in the UK

The European Pharmacopoeia, whose monographs the British Pharmacopoeia incorporates, handles this differently. Monograph 0169, Water for injections (Aqua ad iniectabile), covers water for injections in bulk and sterilised water for injections. The definition of the latter is short and decisive: water for injections in bulk that has been distributed into suitable containers, closed and sterilised by heat under conditions ensuring the product still complies with the test for bacterial endotoxins, and — in the monograph’s own words — Sterilised water for injections is free from any added substances.

The monograph sets numerical limits accordingly: bacterial endotoxins less than 0.25 IU/mL, total organic carbon a maximum of 0.5 mg/L, and conductivity limits that depend on container size — a maximum of 25 µS·cm⁻¹ for containers of nominal volume 10 mL or less, and 5 µS·cm⁻¹ above that. The looser limit on small containers is not a lower standard; it reflects the larger relative contribution of the container surface to a small volume.

What the European Pharmacopoeia does not contain is any article corresponding to bacteriostatic water. Because sterilised water for injections is defined as free from added substances, a preserved multiple-dose water cannot be that article, and no separate monograph creates one. “Bacteriostatic water” is a United States Pharmacopeia designation. A UK page describing it is describing an American pharmacopoeial article, and a UK label using the phrase is borrowing American terminology rather than citing a British Pharmacopoeia article. That is worth knowing before comparing two labels and concluding that one of them is deficient.

Benzyl alcohol: the substance doing the work

Name

Benzyl alcohol; IUPAC name phenylmethanol

Molecular formula

C₇H₈O

Relative molecular mass

108.14

CAS number

100-51-6 (PubChem CID 244)

Physical presentation

Clear, colourless liquid with a faint aromatic odour

Concentration in the preserved article

0.9% w/v = 9 mg/mL, which is approximately 83 mmol/L (9 g/L ÷ 108.14 g/mol)

Aqueous solubility

On the order of 40 g/L at ambient temperature, so 9 g/L sits well below saturation and no co-solvent is involved

The word bacteriostatic is doing precise work and is routinely over-read. A bacteriostatic concentration restricts the proliferation of organisms; it is not a sterilant, and it does not render a contaminated container clean. Aseptic technique is what keeps a container clean. The preservative constrains what happens to organisms introduced in spite of it, over a defined window, against a defined challenge. The distinction between bacteriostatic and bactericidal is the difference between restricting growth and killing, and the article is named for the former.

Benzyl alcohol and phenol are the two antimicrobial preservatives most commonly used in peptide and protein parenteral products, and the selection criteria set out in the review literature weigh antimicrobial performance against the preservative’s effect on the active ingredient, using methods including circular dichroism, fluorescence spectroscopy and differential scanning calorimetry (Meyer et al., 2007). That second criterion — the preservative’s effect on the material it is in contact with — is the subject of a later section here.

How the excipient itself is controlled

Benzyl alcohol has its own European Pharmacopoeia monograph (0256). It specifies a content of 98.0 to 100.5 per cent, and names benzaldehyde, benzoic acid and ethyl benzoate as specified impurities, with benzaldehyde limited to 0.05 per cent. It also sets a peroxide value of not more than 5.

Those three impurities are not arbitrary. They are the oxidation sequence: benzyl alcohol oxidises on standing to benzaldehyde, and benzaldehyde onward to benzoic acid. The peroxide value limit is the one worth noticing, because peroxide species carried in an excipient are a recognised oxidation risk to whatever is dissolved alongside them. A preserved diluent is a formulated liquid with a specification and a degradation pathway of its own. Solute-free water is not.

Where the 28-day figure actually comes from

Two documents produce it, and neither is about peptides.

The first is USP General Chapter <51>, Antimicrobial Effectiveness Testing. A preserved article is challenged with specified organisms and survivors are counted at intervals across a 28-day window. For Category 1 products — aqueous-based injectable and comparable preparations — the acceptance criteria for bacteria are not less than a 1.0 log reduction from the initial count at 7 days and not less than a 3.0 log reduction at 14 days, with no increase thereafter; for yeasts and moulds, no increase from the initial count at 7, 14 and 28 days. Twenty-eight days is the end of the observation window.

The second is USP General Chapter <797>, which assigns a beyond-use date of 28 days to an opened or entered multiple-dose container of a preserved sterile preparation, unless the manufacturer specifies otherwise, and cross-references <51> when it does so.

So the number states two things: how long a preservative system was demonstrated to hold against a defined microbial challenge, and how long a re-entered multiple-dose closure is conventionally regarded as usable. Both are properties of the vehicle and its container.

Neither is a statement about anything dissolved in the water.

The formulation that circulates in this category — that reconstituted material “lasts 28 days because bacteriostatic water preserves it” — welds together two unrelated questions. Microbial proliferation inside a re-entered container is one of them. Chemical degradation of the dissolved solute is the other, and it runs on its own kinetics: hydrolysis of the backbone, deamidation at asparagine and glutamine residues, oxidation at methionine and cysteine where those residues are present, and physical aggregation. Those routes are the subject of a substantial solid-state and solution-state literature (Li et al., 2005; DeHart and Anderson, 2012). None of them is inhibited by an antimicrobial preservative. Taking a lyophilised solid into aqueous solution raises degradation rates by orders of magnitude relative to the solid state whichever water is used, and benzyl alcohol does not alter that.

The corollary for the unpreserved article is cleaner than most pages admit. Sterile water carries no preservative, so the 28-day convention has no application to it at all — not a shorter figure, no figure. There is no preservative whose demonstrated effectiveness window could be quoted, and a single-dose designation does not contemplate a second entry in the first place.

What the formulation literature reports about the preservative

Benzyl alcohol has been studied as a variable in its own right, largely in the recombinant-protein formulation literature, because multiple-dose presentations of biologics require a preservative and the preservative is in contact with the material it preserves. The studies below are reported for what they examined and what their authors reported.

Recombinant proteins in aqueous solution

• Lam, Patapoff and Nguyen (1997) examined the interaction of benzyl alcohol with recombinant human interferon-γ in liquid formulations. They reported conformational change and aggregation, and reported that the dynamics of the interaction varied with buffer species and concentration, ionic strength, and the concentrations of both the protein and benzyl alcohol.
• Tobler and colleagues (2004) followed the same protein by hydrogen–deuterium isotope exchange with mass spectrometry. They reported that the protein’s unfolding rate was markedly sensitive to benzyl alcohol concentration, particularly in the presence of salt; that the tertiary-structure changes previously reported by circular dichroism appeared to involve only a limited portion of the molecule while the hydrophobic core remained stable; and that dynamic light scattering detected small aggregates forming over days.
• Zhang and colleagues (2004) examined recombinant human interleukin-1 receptor antagonist in aqueous solution. They reported that benzyl alcohol produced only minor structural disruption but shifted the molecular population toward partially unfolded, aggregation-prone states, and characterised the interaction as relatively weak and hydrophobically driven, with sucrose partially counteracting the effect.

Lyophilised material reconstituted with a preserved diluent

• Roy and colleagues (2005) is the study closest to the case at hand, because it starts from a freeze-dried solid rather than a liquid. They examined lyophilised formulations of recombinant human interleukin-1 receptor antagonist reconstituted with a bacteriostatic concentration (0.9% w/v) of benzyl alcohol in water, and tested the hypothesis that the extent of aggregation would correlate with the degree of structural perturbation introduced during freeze-drying. They reported greater aggregation on reconstitution with benzyl alcohol than with water alone; that higher protein concentration and sucrose reduced structural perturbation during freeze-drying while sodium chloride destabilised the protein; and that reduced storage temperature after reconstitution lowered the extent of preservative-associated aggregation they observed.

Effect of pH and temperature on the observed aggregation

• Thirumangalathu and colleagues (2006) examined recombinant human granulocyte colony stimulating factor. They reported that benzyl alcohol at 0.9% w/v accelerated aggregation at pH 7.0; that the effect was much greater at 37 °C than at 25 °C; that 1.0 M sucrose partially counteracted it; and that at pH 3.5 benzyl alcohol did not induce aggregation of the protein at all.

The scope of these reports, stated plainly

Each of these papers concerns a specific recombinant protein, in a defined buffer system, examined by a specific analytical method. They are studies of folded proteins of tens to hundreds of residues with tertiary structure and defined conformational states. They are not studies of the short synthetic sequences commonly held as lyophilised solids in a laboratory — a seven-residue or fifteen-residue peptide such as those described in the entries for TB-500 and BPC-157 does not present the conformational landscape these methods were built to measure, and results obtained on interferon-γ do not transfer to it.

What this body of work establishes is narrower, and more useful, than a generalisation: an antimicrobial preservative is a formulation variable that can be measured, and in each of the systems above the authors measured it and reported an effect. It does not establish what occurs with any given peptide. Where no study exists for a particular sequence, that is an absence of data, not a finding of no effect — and it is worth saying so rather than filling the gap in either direction.

Choosing between them is a compatibility question, not a quality question

Neither article is a higher grade of the other. They are the same water with a different specification for added substance and a different container designation. Four questions decide which is appropriate to a given piece of work.

1. Will the closure be entered once, or repeatedly? The multiple-dose designation exists for the second case, and the 28-day convention exists only because of it. The single-dose designation does not anticipate repeated entry, and the resealing behaviour of a closure across multiple entries is a property of containers specified for that purpose.
2. Does the analytical work read in the near-ultraviolet? Benzyl alcohol carries an aromatic chromophore. It contributes background to absorbance-based concentration determination and resolves as its own peak in reversed-phase HPLC with UV detection — which is precisely how it is routinely quantified. Solute-free water contributes neither. Where the material under study is itself chromophoric the interaction has to be thought about rather than assumed away: a copper(II) peptide complex such as GHK-Cu has visible-region absorbance of its own, and the vehicle’s contribution sits in a different part of the spectrum.
3. Is there a documented incompatibility, or simply no reason to introduce an excipient? The preceding section is the honest answer for most short synthetic peptides: no study, therefore no basis for asserting either an effect or its absence. Where a protocol has no need for a preserved vehicle, the unpreserved article removes the variable entirely.
4. Does the method require a solute-free vehicle by definition? Some measurements are confounded by any added substance, and the USP article for sterile water is worded as solute-free for exactly that reason.

The arithmetic does not change

Reconstitution arithmetic is indifferent to which diluent is used. Concentration is the mass of lyophilised solid divided by the volume of diluent added: 10 mg of solid taken into 2 ml gives 5 mg/ml; the same 10 mg into 1 ml gives 10 mg/ml. Choosing a preserved or unpreserved vehicle changes the composition of the liquid, not that relation.

Two points of precision are worth stating because they are usually skipped. First, the dissolved solid contributes its own volume, so the final volume slightly exceeds the volume of diluent added. At milligram quantities in millilitre volumes the displacement is small relative to the tolerance of the measuring device, but it is real and it scales with solid loading — which is why a concentration computed from diluent volume is properly described as nominal unless the final volume is measured. Second, in a preserved vehicle the benzyl alcohol concentration is a property of the diluent, so it is unchanged by how much solid is dissolved into it; 9 mg/mL remains 9 mg/mL.

NovoVita’s reconstitution calculator resolves the relationship between vial strength, diluent volume and the resulting concentration, and its vial presets fill in strength only. The boundary there is deliberate and it is the same boundary this page observes: the volume is arithmetic and is ours to calculate; a target amount is not, and the calculator does not produce one.

Six things this category consistently gets wrong

1. “Bacteriostatic means it sterilises.” It does not. Bacteriostatic describes restriction of growth, not killing, and no preservative concentration makes a contaminated container clean. The article is named accurately and is read inaccurately.
2. “The 28 days is how long the reconstituted material lasts.” The 28 days is the observation window of USP <51> and the beyond-use convention of USP <797> for a re-entered preserved container. It is a statement about the vehicle and the closure. The chemical stability of a dissolved solute is a separate question with separate kinetics.
3. “0.9% is part of the definition.” The USP article specifies one or more suitable antimicrobial agents and obliges the label to declare which and how much. Both 0.9% w/v and 1.1% w/v benzyl alcohol presentations are marketed under it. Read the label rather than the category name.
4. “0.9% benzyl alcohol is the same as 0.9% sodium chloride.” A coincidence of number, nothing more. Bacteriostatic Sodium Chloride Injection is a different USP article again — a 0.9% sodium chloride solution with benzyl alcohol added — and it is neither of the two articles discussed here.
5. “Bacteriostatic water is a British Pharmacopoeia article.” It is not. Ph. Eur. 0169 defines sterilised water for injections as free from any added substances, and the European Pharmacopoeia carries no preserved-water monograph for the phrase to refer to.
6. “Sterile water is the cheaper, lower grade.” Both are the same water. One has an excipient added and a container designation that anticipates repeated entry; the other has neither. Which is appropriate depends on the work, not on a ranking.

The two presentations NovoVita stocks

Both are supplied as 10 ml vials — bacteriostatic water also in a 3 ml size — and both are diluents rather than research compounds.

• Bacteriostatic Water 10ml (BA10) — sterile water containing 0.9% benzyl alcohol as a bacteriostatic preservative, in a sealed multi-dose vial. Also stocked as Bacteriostatic Water 3ml (BA3).
• Sterile Water 10ml (WS10) — sterile water containing no preservative and no solute, in a sealed single-use vial.

Both are specified for storage at 15–25 °C, not to be frozen, and not to be used if the solution is discoloured or the seal is broken. Neither carries a purity figure, and that omission is deliberate: a peptide purity specification has no meaning applied to a diluent.

References

1. United States Pharmacopeia. Bacteriostatic Water for Injection (monograph). USP–NF. Definition, Packaging and storage, and Labeling requirements.
2. United States Pharmacopeia. Sterile Water for Injection (monograph). USP–NF. pH and composition figures quoted above are from representative US manufacturer labelling for products meeting the article.
3. United States Pharmacopeia. General Chapter <51>, Antimicrobial Effectiveness Testing.
4. United States Pharmacopeia. General Chapter <797>, Pharmaceutical Compounding — Sterile Preparations.
5. European Pharmacopoeia. Water for injections (Aqua ad iniectabile), monograph 0169.
6. European Pharmacopoeia. Benzyl alcohol, monograph 0256.
7. Meyer BK, Ni A, Hu B, Shi L. Antimicrobial preservative use in parenteral products: past and present. J Pharm Sci. 2007;96(12):3155–67. PMID 17722087. DOI 10.1002/jps.20976.
8. Lam XM, Patapoff TW, Nguyen TH. The effect of benzyl alcohol on recombinant human interferon-γ. Pharm Res. 1997;14(6):725–9. DOI 10.1023/A:1012190120061.
9. Tobler SA, Holmes BW, Cromwell MEM, Fernandez EJ. Benzyl alcohol-induced destabilization of interferon-gamma: a study by hydrogen–deuterium isotope exchange. J Pharm Sci. 2004;93(6):1605–17. PMID 15124217. DOI 10.1002/jps.10589.
10. Zhang Y, Roy S, Jones LS, Krishnan S, Kerwin BA, Chang BS, Manning MC, Randolph TW, Carpenter JF. Mechanism for benzyl alcohol-induced aggregation of recombinant human interleukin-1 receptor antagonist in aqueous solution. J Pharm Sci. 2004;93(12):3076–89. PMID 15514986. DOI 10.1002/jps.20219.
11. Roy S, Jung R, Kerwin BA, Randolph TW, Carpenter JF. Effects of benzyl alcohol on aggregation of recombinant human interleukin-1-receptor antagonist in reconstituted lyophilized formulations. J Pharm Sci. 2005;94(2):382–96. PMID 15614819. DOI 10.1002/jps.20258.
12. Thirumangalathu R, Krishnan S, Brems DN, Randolph TW, Carpenter JF. Effects of pH, temperature, and sucrose on benzyl alcohol-induced aggregation of recombinant human granulocyte colony stimulating factor. J Pharm Sci. 2006;95(7):1480–97. PMID 16729274. DOI 10.1002/jps.20619.
13. Li B, O’Meara MH, Lubach JW, Schowen RL, Topp EM, Munson EJ, Borchardt RT. Effects of sucrose and mannitol on asparagine deamidation rates of model peptides in solution and in the solid state. J Pharm Sci. 2005;94(8):1723–35. PMID 15986465. DOI 10.1002/jps.20372.
14. DeHart MP, Anderson BD. Effects of water and polymer content on covalent amide-linked adduct formation in peptide-containing amorphous lyophiles. J Pharm Sci. 2012;101(9):3142–56. PMID 22437444. DOI 10.1002/jps.23092.
15. National Center for Biotechnology Information. PubChem Compound Summary for CID 244, Benzyl alcohol.