The Problem With “Close Enough”
Picture a 60 mL cough syrup bottle that measures out to 57 mL when a patient actually pours it. On the shelf, nothing looks wrong — the label says 60 mL, the carton is sealed, the fill line looks right. But over a ten-day course, that patient is quietly short-changed roughly 5% of every dose. Multiply that across a production lot of tens of thousands of bottles and a fill-volume problem that seemed cosmetic becomes a genuine efficacy and compliance issue.
USP General Chapter <698>, Deliverable Volume, exists precisely to catch this before it reaches a patient. It applies to oral liquids and other liquid dosage forms labeled to contain 250 mL or less, and it asks a simple question with a surprisingly detailed answer: when someone empties this container the way it's meant to be used, does it actually deliver what the label promises?
Two Container Types, Two Different Bars
USP <698> doesn't treat every liquid container the same way, because the risk profile is different depending on whether a patient gets one shot at the dose or several.
- Multiple-unit containers — the standard multi-dose bottle a pharmacist or caregiver draws repeated doses from over days or weeks. Because there's room to average out a slightly light pour here or there, USP <698> only sets a floor: no single container may fall below 95% of its labeled volume, and the batch average must be at or above 100%. There's deliberately no upper ceiling — a bottle that delivers a bit more than labeled isn't a patient-safety concern the way a bottle that delivers less is.
- Single-unit containers — a one-time-use ampoule, vial, or unit-dose cup. Here there's no second chance to make up a shortfall, so the tolerance tightens on both sides: every container must land within 95%–110% of labeled volume. Too little under-doses the patient; too much can matter just as much depending on the drug's therapeutic index.
It's tempting to assume the single-unit window (95%–110%) is simply the “stricter” version of the multiple-unit rule and use it everywhere for safety. In practice this over-corrects: a multi-dose bottle that delivers 112% of label on one container hasn't actually failed anything under USP <698>, and flagging it as an OOS wastes an investigation on a non-issue.
The Formula
% of labeled volume = (Measured volume ÷ Labeled volume) × 100Walking Through a Real Batch
Suppose a QC analyst is releasing a batch of 60 mL multi-dose antihistamine syrup. Ten bottles are selected at random and their contents measured:
58.9, 61.2, 60.5, 57.3, 62.0, 59.8, 60.1, 61.5, 58.2, 60.9 mLConverting each to a percentage of the 60 mL label and averaging gives a batch mean of 100.1% — comfortably at or above the 100% requirement. Scanning the individual values, the lowest reading is 57.3 mL, which works out to 95.5% of label. That clears the 95% floor by a narrow margin, so every container passes and the batch is released on Stage 1 with no further testing.
Now change one number: suppose that same bottle had measured 56.4 mL instead — 94.0% of label. The average barely moves, but that one bottle now sits in the 90%–95% borderline band. Under USP <698>, this doesn't fail the batch outright. Because the average still clears 100% and only one container is borderline (not below 90%, and not more than one container affected), the standard permits testing 20 additional bottles and pooling all 30 results together under Stage 2's slightly more forgiving tolerance — up to one borderline unit is still allowed at n=30.
Where a result does fail outright, with no retest available, is a container below 90% of label (multiple-unit) or two or more containers landing in the borderline band. At that point the standard doesn't offer a second attempt — the result goes to an out-of-specification investigation instead.
Common Mistakes to Avoid
- Assuming a borderline result is an automatic failure. USP <698> explicitly builds in a retest path for isolated borderline containers — treating every borderline unit as a hard fail wastes material and manufacturing time on batches that would have passed at Stage 2.
- Evaluating the 20 additional Stage 2 containers as their own separate group instead of pooling all 30 results together. The standard requires one combined average and one combined pass/fail call across all 30 units, not two independent judgments.
- Using the single-unit tolerance window on a multi-dose container “to be safe.” This produces false failures on bottles that are perfectly compliant under the rule that actually applies to them.
- Treating a container below 90% (multiple-unit) or outside 90%–115% (single-unit) as retestable. These results are excluded from the Stage 2 pathway entirely and should route straight to an OOS investigation.
- Rounding percentages before comparing against the 95% or 90% cutoffs. A reading of 94.96% rounds to 95% on a report but is still, arithmetically, below the floor — comparisons should be made on the unrounded value.
Frequently Asked Questions
Does USP <698> apply to injectable products?
No. Chapter <698> is scoped to oral and other non-parenteral liquid dosage forms labeled at 250 mL or less. Injectable and parenteral volume requirements are covered under separate chapters with their own fill-volume and withdrawable-volume criteria.
How many containers actually need to be tested for a routine release?
Ten containers for Stage 1. Testing stops there if the batch passes outright. Only batches that land in the defined borderline zone move to Stage 2, which adds 20 more containers for a pooled total of 30 — it is not automatically a 30-unit test for every batch.
Can a batch be released if only the average passes but a few individual containers don't?
It depends on how many containers miss the mark and by how much. One borderline container with an otherwise compliant average can still pass through Stage 2 retesting. Two or more borderline containers, or any single container below the extended retest floor, cannot be released on the strength of the average alone.
Is deliverable volume the same test as fill volume?
No, and the distinction matters. Fill volume is what was put into the container at manufacture; deliverable volume is what actually comes out when the container is used as intended, accounting for residual film left behind by viscous formulations, dead volume in the container geometry, and pour technique. A batch can have a correct fill volume and still fail deliverable volume if the formulation clings to the container walls.
What should happen if a batch fails Stage 2?
A Stage 2 failure is not retestable under USP <698> — there is no Stage 3. It triggers a formal out-of-specification investigation to determine root cause (fill-pump calibration, container geometry, viscosity drift, closure integrity) before any disposition decision is made on the batch.
Need to run these numbers on an actual batch? The Deliverable Volume Calculator on PharmainformCalc applies this exact Stage 1 / Stage 2 logic — enter your container type, labeled volume, and measured results, and it tells you whether the batch passes, needs Stage 2 retesting, or fails outright.