Reference and working standards

Almost every quantitative result in a QC lab is a comparison. An HPLC assay compares the sample peak with a standard peak. A titration depends on a volumetric solution whose strength was set against a primary standard. If the standard is wrong, every result calculated from it is wrong — and the error is silent, because system suitability will still pass.

That is why regulators expect reference standards to be controlled as tightly as the samples themselves. 21 CFR 211.194 asks for complete records of the testing and standardisation of reference standards, reagents and standard solutions, and WHO TRS 957 Annex 1 has a dedicated section on reference substances and reference materials.

A standard is only as good as its paperwork. If you cannot show where it came from, how it was stored and what potency you used, an inspector will treat the results that depend on it as unsupported.

A primary reference standard is one accepted as having the appropriate quality without comparison to another chemical substance. In practice, most labs buy pharmacopoeial standards: USP Reference Standards, Ph. Eur. Chemical Reference Substances (CRS), IP Reference Substances or WHO International Chemical Reference Substances. When a pharmacopoeial standard is used for its official purpose, its suitability is established by the issuing body and you do not need to re-test it.

A secondary or working standard is an in-house material — often a well-characterised batch of the API — that has been qualified against the primary standard. Labs use working standards because primary standards are expensive and supplied in small quantities. The working standard carries a potency assigned by your own lab, so your lab is fully responsible for it.

Qualification is a documented protocol, approved before testing begins. Choose a batch of high purity — usually a production batch that meets the specification with a good margin — and homogenise it before sub-dividing. Then test it against the current lot of the primary standard.

1. 1Identity: confirm by a specific technique, typically IR comparison with the primary standard spectrum, plus HPLC retention time.
2. 2Water or loss on drying: determine by Karl Fischer titration or LOD so potency can be expressed on the correct basis.
3. 3Purity tests: related substances, residual solvents and residue on ignition, as relevant to the substance.
4. 4Assay against the primary standard: typically replicate preparations (for example six) of both standard and candidate, often by two analysts or on two days, using the compendial or a validated method.
5. 5Assign potency: calculate the mean, check precision against the protocol limit (for example %RSD), and state the basis — as is, dried or anhydrous.
6. 6Approve and release: QA reviews the raw data and issues a certificate showing the working-standard lot, the primary lot used, the assigned potency, storage conditions and the retest date.

Always record which primary-standard lot was used. If that lot is later withdrawn by the pharmacopoeia, you need to know which working standards — and which results — are affected.

A working standard is given a retest date, not an indefinite life. Many labs re-qualify every 12 months, but the interval should reflect the stability of the substance and is defined in your SOP. Hygroscopic or unstable substances may need shorter intervals or water determination at each use.

Re-qualification usually repeats identity, water/LOD and assay against the current primary standard. If the new assay differs from the assigned value by more than a set amount, investigate before changing the value. A drifting potency is a stability signal, not just a number to update.

• Pharmacopoeial standards normally carry no expiry date: check the official catalogue to confirm the lot is still current before each use.
• Do not use a working standard past its retest date, even 'just for this run'.
• Keep a register linking each working-standard lot to its primary lot, certificate and retest date.

On receipt, record the standard in the reference-standard register: name, supplier, lot, quantity, date received, storage condition and certificate. Store it exactly as the label says — often 2–8 °C or below −15 °C, protected from light and moisture — in a dedicated, access-controlled and temperature-monitored location.

Let refrigerated vials reach room temperature in a desiccator before opening, so moisture does not condense on the powder. Follow the label on drying: some USP standards must be dried before use, others are used as is with a correction for water.

• Label every container with name, lot, assigned potency and basis, storage condition, date opened and retest or use-by date.
• Keep a usage log for each lot: date, analyst, quantity taken, test and sample reference, balance used.
• Weigh standards above the balance minimum weight; small weighings of expensive standards are a common source of error.
• Prepared standard solutions get their own label: identity, concentration, solvent, preparation date, expiry, analyst and the standard lot used.
• Return to storage promptly and never put unused material back into the original container.

Impurity reference standards are used to identify impurity peaks, check resolution in system suitability and quantify specified impurities. They are often available only in milligram quantities and may have lower or less certain purity than the API standard.

When an impurity standard is used for quantitation, its purity must be known and taken into account. If an impurity is quantified against the API standard instead, the method may apply a relative response factor. Peak-identification mixtures supplied by the pharmacopoeia are for identification and system suitability only — do not use them for quantitation unless the monograph allows it.

In-house impurity standards (for example isolated degradants) need characterisation proportionate to their use: structure confirmation, purity and a documented assigned value if they are used quantitatively.

A volumetric solution (VS) such as 0.1 M sodium hydroxide or 0.1 M perchloric acid is a reagent whose exact strength must be known. You prepare it to the nominal concentration, then standardise it against a primary standard — for example potassium hydrogen phthalate for sodium hydroxide — to find its factor (actual molarity ÷ nominal molarity).

Standardise in replicate (often three determinations), check the agreement against your SOP limit and record the factor on the label with the date and next standardisation due date. Re-standardise at the interval set in your SOP and whenever the solution may have changed, for example after a temperature change for non-aqueous titrants. Karl Fischer reagent is treated the same way: its water equivalence factor is determined regularly.

• Dry primary standards as the pharmacopoeia directs before weighing.
• Record the burette or titrator used — a burette accuracy problem looks exactly like a wrong factor.
• Discard and re-prepare if the solution is past its use-by date or shows turbidity or precipitation.

The same observations appear in inspection reports year after year. Most are documentation gaps rather than science problems, which means they are easy to prevent with a good register and disciplined labelling.

• Working standard used past its retest date, or primary lot used after it was superseded.
• No qualification protocol, or potency assigned from a single determination.
• Assigned potency basis (as is vs dried) not stated, so the wrong correction was applied in calculations.
• Usage log quantities do not reconcile with the remaining stock.
• Standard solutions without preparation or expiry dates, or used beyond their established solution stability.
• Storage temperature not monitored, or excursions not assessed.
• Volumetric solution factor on the label differs from the factor in the standardisation record.

A quick self-check: pick one reported assay result and trace it back to the working-standard certificate, the primary lot and the balance calibration. If any link in that chain is missing, fix the system, not just the record.