Osmolality, osmolarity, and tonicity: three numbers for one reconstituted solution

A reconstituted peptide vial is usually described by one number: concentration, in milligrams per millilitre, derived from the label claim and the volume of diluent added. That number answers a mass question. It says nothing about how many dissolved particles are present, and for a solution that will meet a membrane — a filter, a dialysis cassette, a cell monolayer — the particle count is often the property that governs behaviour. Three terms describe that count, they are routinely used as synonyms, and they are not synonyms. Osmolality is measured. Osmolarity is converted. Tonicity is neither: it is a statement about a solution and a particular membrane, and it cannot be read off an instrument at all. This post separates the three, works through what each contributes in a typical reconstituted vial, and identifies where the distinctions stop being pedantic.

What the measurement counts

Osmotic concentration is a colligative property, which is a compact way of saying that it depends on how many solute particles are present and not at all on what they are. A mole of sucrose and a mole of benzyl alcohol depress the freezing point of water by the same amount. A mole of sodium chloride depresses it by roughly twice as much, because it dissociates into two particles. Molecular weight, charge, size and chemistry drop out of the accounting entirely once dissociation is taken into account; only the number of independent kinetic units survives.

Osmolality expresses that count per kilogram of solvent — milliosmoles per kilogram of water. Osmolarity expresses it per litre of solution — milliosmoles per litre. The difference looks trivial and is not, for two reasons. Volume varies with temperature while mass does not, so an osmolarity figure is implicitly tied to a temperature that is rarely stated. And in a solution with substantial dissolved or suspended material, a litre of solution contains measurably less than a kilogram of water, so the two numbers diverge in proportion to how much of the volume is occupied by things that are not solvent. For the dilute aqueous systems typical of peptide reconstitution the divergence is small, but it is a systematic bias rather than noise.

The practical consequence is that osmolarity is not measured. It is calculated, either from composition or by applying a conversion factor to a measured osmolality. USP general chapter <785> covers both quantities and takes freezing-point depression as the primary experimental route, with the molal cryoscopic constant for water taken as 1.860 degrees per osmole per kilogram. A companion chapter, <1785>, was developed to hold the practical and interpretive material that a test chapter cannot carry, and recent revisions to <785> include a cross-reference to it along with procedural updates reflecting instruments that use disposable sample cells.

There is a long-standing argument in the parenteral literature that osmolality is the more defensible quantity to record, precisely because it is the one the instrument produces. Whatever the convention in a given context, the analytical point stands: a recorded osmolality is a measurement, and a recorded osmolarity is a measurement plus an assumption.

Why the instrument choice matters more than usual here

Two osmometer designs dominate, and they do not measure the same population of solutes.

A freezing-point-depression osmometer supercools the sample, induces crystallisation, and reads the plateau temperature at which the heat of fusion holds the sample steady. Everything dissolved in the water contributes, volatile or not. A vapour-pressure osmometer instead measures the depression of solvent vapour pressure, usually via the dew-point temperature of the headspace above a small sample on a filter disc. Volatile solutes are effectively invisible to it, because they contribute to the vapour phase rather than depressing it. The clinical literature documents the consequence directly: freezing-point and vapour-pressure osmolality diverge in samples containing volatile alcohols, and vapour-pressure instruments are a recognised source of false negatives in exactly those cases.

This is not an abstract distinction for peptide work, because the most widely used reconstitution diluent contains a volatile solute by design. Bacteriostatic water is 0.9% benzyl alcohol. At a molecular weight of 108.14, 9 grams per litre is about 83 millimoles per litre, and benzyl alcohol does not dissociate, so it contributes on the order of 83 milliosmoles per kilogram on its own — a substantial fraction of physiological osmolality, contributed by the diluent before any peptide enters the vial. A freezing-point instrument will report it. A vapour-pressure instrument will largely not. The same vial, measured two ways, yields two numbers that differ by more than any plausible analytical imprecision, and neither instrument is malfunctioning.

The general rule that follows is the one that applies to any validated analytical method: an osmolality figure without its method is incomplete, and comparing figures generated by different techniques on preserved formulations is not a valid comparison.

The arithmetic of an actual vial

Working the numbers for a representative reconstituted vial is clarifying, because it inverts the intuition that the labelled substance dominates its own solution.

Take 5 mg of a peptide of molecular weight 4,000 reconstituted in 2 mL. The concentration arithmetic gives 2.5 mg/mL. Converted to molarity, that is 0.625 millimoles per litre — so the peptide itself contributes roughly 0.6 milliosmoles per kilogram. Under one part in four hundred of a physiological osmotic load.

Now add what else is in the vial. A lyophilised cake is rarely pure peptide; bulking agents and lyoprotectants are present precisely because a few milligrams of peptide cannot form a cake on its own. Forty milligrams of mannitol (MW 182.17) in the same 2 mL is about 110 milliosmoles per litre. Forty milligrams of sucrose (MW 342.3) is about 58. The excipient outweighs the peptide’s osmotic contribution by two orders of magnitude, which is the arithmetic consequence of it being present at ten times the mass and one twentieth the molecular weight.

Counterions add a smaller but non-zero term. A peptide carrying four trifluoroacetate counterions per molecule contributes four times its own molarity in additional particles — about 2.5 milliosmoles per litre at the concentration above. Buffer salts, where present, dissociate and count double or triple. And the diluent itself, if it is bacteriostatic water, contributes the 83 milliosmoles computed earlier.

The ranking that emerges is stable across most formulations: diluent and excipients first, counterions and buffer salts second, peptide last by a wide margin. A vial whose osmolality reads 150 mOsm/kg is not telling you about the peptide. This is the same category of observation as the gap between peptide content and chromatographic purity — a number attached to a vial that is arithmetically dominated by material other than the named substance.

Tonicity is not a property of the solution

Osmolality is measurable because it is a property of the liquid. Tonicity is not, because it is a property of a pairing: this solution, that membrane.

The reason is that membranes are selective. A solute that crosses freely equilibrates across the barrier and ends up at equal concentration on both sides, which means it generates no sustained osmotic gradient and exerts no sustained effect on cell volume. Only solutes the membrane excludes — effective osmoles, in the standard terminology — contribute to tonicity. The textbook demonstration is 5% dextrose, which is very nearly iso-osmotic with cell contents at around 300 mOsm and yet functionally hypotonic, because dextrose enters cells readily and the osmotic gradient it appears to create does not persist.

Benzyl alcohol is the same kind of solute. It is small, appreciably lipophilic, and crosses biological membranes without difficulty — which is inseparable from how it acts as a preservative in the first place. The 83 milliosmoles it contributes are therefore real osmoles and largely ineffective ones. A vial reconstituted in bacteriostatic water and an otherwise identical vial reconstituted in sterile water will differ by a clearly measurable margin in osmolality, and by far less than that margin in tonicity against any given membrane. Two instruments and two biological endpoints can all be correct while appearing to disagree.

Dissociation supplies a second layer of the same lesson. The textbook osmolarity of 0.9% sodium chloride, computed with complete dissociation assumed, is 308 mOsm/L. Measured values come in near 286 mOsm/kg, because sodium chloride is not fully dissociated in solution — the osmotic coefficient sits around 0.93 rather than 1. The calculated figure and the measured figure differ by about 7%, in the direction that matters: the calculation overstates. For reference, this is the reason a solution that arithmetic predicts should be mildly hypertonic behaves as isotonic in practice.

For peptide work the downstream relevance is concentrated in assay systems. A cell-based experiment is sensitive to effective osmolality; a filtration step is sensitive to the physical properties of the bulk solution; a dialysis or buffer-exchange step is sensitive to both, in different terms. None of those is answered by one number.

What the figure establishes, and what it does not

Osmolality has a limited and genuine evidentiary role, and it is worth stating both halves.

What it establishes is formulation consistency. Because the measurement is dominated by excipients, counterions and diluent, it is sensitive to exactly the errors that are hardest to see by other means: a bulking agent weighed at the wrong level, a reconstitution volume off by a material fraction, a buffer prepared at the wrong strength, a diluent substituted for another. It is a fast, cheap, non-specific check on whether a solution was assembled as intended, and its insensitivity to the peptide is what makes it useful as an orthogonal check rather than a redundant one.

What it does not establish is anything about the peptide. It does not distinguish intact material from degradation products, since fragmentation of a species present at 0.6 mOsm produces changes below the resolution of the method. It does not report solubility; a solution at its saturation limit and one well below it can return the same figure. It says nothing about particulate matter, which by definition is not in solution and therefore not counted. And it has no bearing on identity or purity, which is why a certificate of analysis carries it, when it carries it at all, alongside those tests rather than in place of any of them.

The habit worth forming is to read an osmotic figure as three separate claims with three separate warrants. The osmolality is an instrument reading, and the instrument type is part of the reading. The osmolarity is that reading plus a conversion, and the conversion is formulation-specific. The tonicity is neither, and cannot be inferred from either without knowing which of the counted particles the membrane in question actually stops.

What the measurement counts

Why the instrument choice matters more than usual here

The arithmetic of an actual vial

Tonicity is not a property of the solution

What the figure establishes, and what it does not

Further reading