Water / LOD Compensation

Water / Loss-on-Drying compensation is the arithmetic correction applied at dispense when the formula's target mass and the supplier CoA's assay value are quoted on different moisture bases. The two bases that matter in practice are 'as-is' (the actual mass of the material as it sits in the drum, including whatever water it contains) and 'anhydrous' (the mass of the material as if all water were removed). If the formula was developed against an anhydrous-basis assay (typical for most pharmacopeial APIs) but the operator weighs the actual material as-is, the difference between the two — the moisture content — must be added back via the LOD correction.

The math is one line: as-charged target = anhydrous-basis target ÷ (1 − LOD), where LOD is the loss-on-drying fraction expressed as a decimal (0.042 for 4.2% moisture). A 50.000 kg anhydrous-basis target with 4.2% moisture becomes 50.000 / (1 − 0.042) = 52.193 kg as-charged. The operator weighs 52.193 kg of the actual material; the active mass delivered after the 4.2% moisture is accounted for is exactly 50.000 kg.

Two pharmacopeial methods exist for measuring 'how much water is in this sample', and they are not interchangeable. Choosing the wrong one is the most common quiet failure mode in moisture-driven dispense errors.

USP General Notices 5.50.1 names three bases on which a pharmacopeial substance's strength can be expressed, and the formula, the CoA, and the dispense calculation must all be reconciled to one of them. The three are not the same and the differences compound across multiple-active formulations.

1. As-is basis — strength expressed against the actual material as it sits in the container, water and impurities included. The operator weighs against this directly; no correction is needed if the formula and CoA both use as-is. Common for excipients and for actives where moisture is naturally part of the dosed material.
2. Anhydrous basis — strength expressed against the active mass with all water removed (water by KF or moisture by LOD discounted). Common for APIs, especially those that exist as hydrates or that are hygroscopic. The standard recalculation: as-charged = anhydrous-basis × 1/(1 − moisture).
3. On-dry-basis — strength expressed against the active mass with all volatiles removed (water + residual solvents + any other LOD-detectable volatiles). Required for materials where residual solvent is part of the specification. The standard recalculation: as-charged = on-dry-basis × 1/(1 − LOD-total).

The general form when LOD compensation combines with potency factor and salt-to-base: as-charged target = label-claim-mass × SBF × PF × 1/(1 − LOD), with overage already layered into label-claim-mass at the formulation step. The order matters because rounding errors compound; the canonical order is SBF → LOD → PF → counter-balance.

Each as-charged target is what the operator actually weighs. The nominal and intermediate figures are shown for context only; the audit trail captures all of them so a regulator can reconstruct the math.

Once a container of API or excipient is opened, the material is exposed to ambient humidity and moisture content changes — usually upward for hygroscopic materials, sometimes downward for materials that started saturated. USP <1241> Water-Solid Interactions classifies materials by their water-uptake behaviour and that classification drives how long after first-open the original CoA LOD value remains valid before re-test is required.

1. Basis mismatch (anhydrous formula vs as-is CoA) — formula expects anhydrous-basis assay; CoA reports as-is. Operator weighs the nominal mass; product is silently over-strength by the moisture fraction. The mirror failure (as-is formula, anhydrous CoA) produces silent under-strength.
2. LOD substituted for KF on a hydrate — heating the sample drives off water of crystallisation along with surface moisture; reported 'moisture' is inflated; compensation factor is inflated; dispense over-charges by the difference. Common with iron sulphates, magnesium sulphates, certain calcium salts.
3. KF method drift unsuspected — Karl Fischer titration loses precision if the reagent degrades or if the cell contains organic interferences (aldehydes, ketones, mercaptans). Drift goes unnoticed until annual surveillance assay reveals a campaign-wide variance.
4. Hygroscopic material dispensed in non-controlled-humidity area — material absorbs moisture between scale-read and discharge; actual delivered active mass is below adjusted target. Repeat for every dispense over the open lot's life.
5. Re-test not triggered after first-open — original CoA LOD used for the entire shelf life of a moderately hygroscopic material; actual moisture drifts upward; dispense progressively over-charges. Caught only on the next lot or on annual stability surveillance.
6. Cumulative open exposure not logged — multiple opens of the same drum over a month, each adding moisture; no single open triggers a re-test threshold, but cumulative drift is significant. Failure of the policy implementation, not the policy itself.
7. Deliquescent material partially liquefied — dispense kiosk records the as-formulated mass; actual moisture is 5x the CoA; product is severely under-strength. Quality finds it on first finished-product assay; the entire lot of dispenses requires investigation.
8. Moisture spec exceeded but used anyway — lot LOD exceeds the formula's max-moisture spec but was dispensed against an 'extended use' rationale that wasn't change-controlled. Inspection finds the deviation chain incomplete; entire campaign requires investigation under §211.103.

• LOD/KF-corrected dispense coverage % — fraction of moisture-sensitive dispenses where the recalculation engine ran with the correct basis (target 100%).
• Method-mismatch rate — fraction of dispense events where the lot's measured-by method (LOD vs KF) does not match the specification's required method (target 0%; non-zero is an immediate corrective action).
• Re-test trigger compliance % — fraction of open events on hygroscopic-class lots where the policy-required re-test was performed before next dispense (target 100%).
• Moisture-out-of-spec dispense rate — fraction of dispense events where lot LOD exceeded formula max-moisture spec (target 0%; every one is a deviation).
• CoA-vs-actual moisture variance — for surveillance re-tests, average absolute delta between CoA LOD and re-tested LOD (drift >2% suggests storage/packaging issue with the supplier).
• Hygroscopicity classification coverage — % of moisture-sensitive items with a USP <1241> class assigned (target 100% for items with KF or LOD specs).
• Open-event logging compliance — % of opens on class 3+ items where ambient humidity and re-close timestamp were captured (target 100%).

1. Item master holds: assay basis (anhydrous / as-is / on-dry), moisture method required (LOD / KF / either), max-moisture specification, USP <1241> hygroscopicity class, re-test policy by class (first-open, every-open, time-based), counter-balance partner.
2. Lot master holds: CoA moisture value, method actually used (with auto-flag if method ≠ item-required), measurement date, open-events log (timestamp + ambient RH + re-close timestamp per open), re-test history.
3. CoA import (electronic CoA from supplier) auto-populates assay, basis, moisture, method, and flags any mismatch with the item-required method as a deviation before the lot can be released.
4. Dispense screen shows: anhydrous-basis target, LOD/KF value applied, as-charged target, the recipe ('50.000 / (1 − 0.042) = 52.193 kg'), and the lot's hygroscopicity class with re-test status.
5. If re-test policy threshold is hit (e.g. moderately hygroscopic + 24h since open + open-to-dispense >24h), the kiosk blocks dispense until re-test result is entered, then re-computes the target with the new LOD.
6. Out-of-spec lots (LOD > max-moisture) cannot be dispensed without deviation + QA disposition + change-controlled extended-use rationale; the audit trail captures all four artefacts.
7. Counter-balance excipient is re-computed in the same step — if the active goes up due to moisture compensation, the diluent comes down to hold total batch mass.
8. Two-person e-signature gates the next dispense step; the verifier sees the same recipe the operator weighed to (Part 11 §11.50 + §11.70).
9. BMR/BPR/DHR renders the basis-corrected target, the LOD/KF value applied, the method used, the CoA reference, the open-events log entries that affected the result, and the operator + verifier signatures — contemporaneously, no post-batch transcription.
10. Periodic moisture surveillance (annual or per stability protocol) feeds back into the CoA-vs-actual variance KPI; persistent drift triggers a supplier scorecard alert and potential change control on the moisture specification.

See the FAQs below for short answers on common operational questions — when LOD compensation is mandatory, how to handle a CoA without an LOD result, what happens when LOD and KF disagree, and how moisture compensation interacts with overage and PF.