Every tablet is supposed to carry the same dose, and the quickest way to check that during manufacturing is to weigh the tablet on weighing balance. If the tablets of a batch are all close to the same weight, they carry close to the same amount of active ingredient. If the weights scatter, the doses scatter too. That is why weight variation, also called uniformity of mass, is one of the most frequent checks on a compression line, and why a failed result needs a proper investigation.
This article explains how the test works, what the common pharmacopoeial limits are, how to run through the calculation with a practical example, and what to check on the compression machine when the results start to vary.
Watch: Weight Variation
Weight Variation Calculator, which checks BP, Ph. Eur. and USP weight variation
Tablet Friability Calculator, which works out percentage weight loss over up to three runs.
What the test shows, and what it does not
- Weight variation test is an indirect check. It assumes that the active ingredient is spread evenly through the blend, so that a heavier tablet holds proportionally more of it.
- When that assumption holds, weight is an excellent stand-in for dose. When the blend itself is uneven, for example because the active segregated or was not mixed properly in granulation or blending area tablets can all weigh the same and still differ in strength.
- That is the reason the pharmacopoeias keep a separate content uniformity test, and require it for products that carry a low dose or a small share of active.
How the BP and Ph. Eur. uniformity of mass test works
- Weigh 20 tablets individually and record each weight.
- Calculate the average weight of the 20 tablets.
- Work out the limits from the average, using the percentage in the table below.
- Count the tablets outside those limits.
The tablets comply if no more than two are outside the limit, and none is outside twice the limit. The percentages below apply to tablets, both uncoated and film-coated.
| Average tablet weight | Allowed deviation from the average |
|---|---|
| 80 mg or less | ±10% |
| More than 80 mg and less than 250 mg | ±7.5% |
| 250 mg or more | ±5% |
The limits depend on the average weight, not on the target weight. A light tablet is allowed a wider percentage spread than a heavy one, because a small absolute difference is a larger share of its weight.
Worked example
Twenty tablets with a target weight of 500 mg are weighed. The individual weights, in mg, are 498, 503, 508, 495, 501, 512, 500, 504, 468, 507, 502, 497, 505, 499, 511, 496, 503, 493, 506 and 500.
- The total is 10,008 mg, so the average is 10,008 ÷ 20 = 500.4 mg.
- The average is above 250 mg, so the limit is ±5%: 500.4 × 0.95 = 475.4 mg lower side and 500.4 × 1.05 = 525.4 mg upper side.
- Twice the limit is ±10%, which is 450.4 mg to 550.4 mg.
- One tablet, at 468 mg, is outside the limit (it is 6.5% below the average). No tablet is outside 450.4–550.4 mg.
Only one tablet is outside the limit and none is outside twice the limit, so the sample complies and test is passed.
A second tablet below 475.4 mg would still be acceptable. A third one would fail the test, and a single tablet at 440 mg would fail it by itself. The Weight Variation Calculator does this arithmetic for you and marks each unit.
The USP approach: acceptance value
A note on terminology: the 10% / 7.5% / 5% table with cut-offs at 130 mg and 324 mg, which many textbooks call the “USP weight variation test”, is published in USP <2091> for dietary supplement tablets. For drug products, the USP route is <905>, described below.
The USP handles dose uniformity in the chapter on uniformity of dosage units (<905>). It allows weight variation to be used in place of a content uniformity test only for products that meet conditions set out in the chapter, such as tablets with a high proportion of active ingredient. The weights are converted to estimated contents using the assay of the batch, and the result is expressed as an acceptance value (AV).
AV = |M − X̄| + k × s, where X̄ is the mean of the estimated contents as a percentage of the label claim, s is their standard deviation, M is a reference value tied to the mean, and k is 2.4 for 10 units and 2.0 for 30 units. Stage 1 passes if the AV of 10 units is 15.0 or less. If not, 20 more units are tested, and Stage 2 passes if the AV of all 30 is 15.0 or less and no unit is outside the limits set by the chapter.
For example, take 10 tablets with an average weight of 400 mg and an assay of 99.0% of the label claim. If the individual weights are 398, 402, 401, 399, 403, 397, 400, 402, 398 and 400 mg, the mean estimated content is 99.0% and its standard deviation is about 0.50%. Because the mean lies inside 98.5–101.5%, M equals the mean, so AV = 0 + 2.4 × 0.50 = about 1.2, far below 15.0.
Common causes of weight variation
When weights vary on the press, the cause is nearly always one of these:
- Poor granule flow. Granules that flow unevenly fill the dies unevenly. Check the bulk and tapped density and the Carr’s Index of the blend. The Carr’s Index guide explains how to read them.
- Too many fines or a wide particle size range. Fine particles and large granules separate in the hopper and feed frame.
- Feeder and feed frame settings. Paddle speed, fill depth and feed frame condition all change die fill.
- Tooling. Worn or uneven punches and dies, and punches that stick, give different fill volumes.
- Press speed. The faster the turret, the shorter the die fill time, so a blend that runs well at a low speed may vary at a high one.
- Hopper level and vibration. A hopper that runs low, or a bridging blend, changes the head of powder over the feed frame.
In-process checks versus the pharmacopoeial test
The pharmacopoeial test is a release test, done on the finished batch. During compression, operators check weights much more often, usually at set intervals, with action limits that are tighter than the pharmacopoeial ones, so that drift is caught before a result comes close to failing. The right limits and frequency are those in your approved process and SOPs. Any trend towards one side, for example a slow rise in the average, deserves attention even when every tablet is still inside the limit.
Common mistakes
- Using the target weight instead of the measured average to set the limits.
- Applying the ±5%, 7.5% and 10% bands to the wrong weight range.
- Using weight variation where the monograph requires content uniformity.
- Mixing coated and uncoated tablet weights in the same test.
- Rounding each weight before averaging.
- Not recording the balance used and its calibration status.
Frequently asked questions
How many tablets are weighed?
The BP and Ph. Eur. test uses 20 tablets, weighed individually. The USP acceptance value test starts with 10 units.
How many tablets may be outside the limit?
No more than two may be outside the percentage limit, and none may be outside twice the limit.
Is weight variation the same as content uniformity?
No. Weight variation relies on the weight of each unit as an indicator of its dose. Content uniformity measures the amount of active in each unit directly.
What should I do if the test fails?
Stop, investigate the blend flow, the feeder and the tooling, and follow the deviation procedure. Check the compression reconciliation records for the same run as well.
Related: the Tablet Friability Calculator checks how well tablets resist abrasion and breakage.