Between the moment a tablet leaves the punch and the moment it is sealed in foil, it is checked at least six times, and almost every one of those checks is a number written into the batch record within minutes of being taken. Hardness, weight, thickness, friability and disintegration are the familiar four. The less familiar ones — the leak test, the metal check, the reconciliation — are the ones that decide whether a batch is releasable. This is what actually happens, and what is on paper afterwards.
Compression starts before the batch does
Nothing is compressed until the line clearance is signed. That is a physical check that the previous product, its documentation and its residues are gone, recorded against the previous batch number and countersigned. Then the press is set up and a set of trial tablets is taken and checked against the batch record limits before the batch is declared started. Machine speed matters here and is not a free variable: WHO Prequalification’s published deficiency list notes that “about 60% of the reviewed applications had protocols deficient,” including failure to validate a compression machine speed range. Running a validated product outside the validated speed range is a deviation, not an efficiency.
Once running, the press is sampled at intervals fixed in the batch record — commonly every 30 minutes, and additionally at every machine adjustment, every stoppage and every change of blend container. Each sample set produces individual weights, average weight, thickness and hardness, all logged with the time and the operator’s initials.
The four tests everyone asks about
Weight variation is the first control and the most frequent. Twenty tablets are weighed individually and against the average. The Indian and British pharmacopoeias set deviation limits of ±10% for tablets weighing 80 mg or less, ±7.5% between 80 and 250 mg, and ±5% at 250 mg and above. USP uses different bands on the same principle: ±10% up to 130 mg, ±7.5% from 130 to 324 mg, and ±5% above 324 mg. Which set applies depends on the pharmacopoeial standard the product is registered against, and it is worth confirming that before an audit asks.
Weight variation is not the same test as content uniformity. Weight variation is a proxy that only holds where the drug makes up a large enough share of the tablet. Content uniformity measures the active in individual units — the British Pharmacopoeia approach tests ten tablets and requires each individual content to fall between 85% and 115% of the average content, extending to twenty more if one fails, with not more than one of the thirty outside range.
Hardness, more properly breaking force, is measured in kilograms-force or newtons on a set of tablets at each interval. There is no universal pharmacopoeial limit; the range is product-specific and set during development. Its real function is as an early warning — hardness drifts before friability and disintegration fail.
Friability is a durability test, not a strength test. Tablets are tumbled in a friabilator at 25 rpm for four minutes, which is 100 revolutions, then reweighed. A weight loss below roughly 0.5% to 1% is generally considered acceptable for conventional compressed tablets. Friability is where you find out whether a tablet will survive the blister machine, the shipper and the road, and it is checked at defined points during compression rather than only at the end.
Disintegration is checked on six tablets. Pharmacopoeial limits vary by coating and by pharmacopoeia: the Indian Pharmacopoeia allows 15 minutes for uncoated tablets, 30 minutes for film-coated and 60 minutes for enteric-coated; the British Pharmacopoeia allows 15 minutes uncoated, 60 minutes film-coated and 5 minutes for effervescent tablets. If a tablet fails, the test is repeated on twelve more. Disintegration is not dissolution, and it does not replace it — dissolution is a finished-product test run by QC against the registered specification.
Between the press and the blister machine
Compressed tablets do not go straight to packing. They pass a deduster, and in most lines a metal detector, before going into labelled containers under a bulk quarantine status. Yield is reconciled at this point: theoretical against actual, with any difference outside the permitted limit raising a deviation before packing begins. Coated products go through the coating stage here, with their own in-process weight-gain and appearance checks.
The bulk then waits for QC. Nothing is blistered against an unreleased bulk unless the site operates a documented, approved arrangement for it — and if it does, an auditor will look at that arrangement first.
Blistering: the checks change shape
At the blister machine the questions are no longer about the tablet. They are about the pack.
Foil and film are verified against the batch record by grade and by printed lot. Overprinting — batch number, manufacturing date, expiry date — is checked at start-up and at intervals through the run, because a wrong expiry date on foil is a recall, not a deviation. Sealing integrity is checked by leak test: blisters are held under vacuum in a dye solution and inspected for ingress, typically at start-up and every 30 minutes thereafter. Pocket fill is checked visually and, on many lines, by an inline camera. Cartoning adds leaflet presence, carton coding and, where applicable, serialisation.
Reconciliation runs again at the end: foil issued against foil used and rejected, cartons issued against cartons packed, tablets in against packs out. The numbers have to close.
What gets checked, and how often
| Stage | Checked | Typical frequency | Recorded in |
| Before compression | Line clearance, machine set-up, trial tablets | Once per batch, before start | Line clearance record; BMR |
| Compression | Individual and average weight, thickness, hardness, appearance | ~Every 30 min, plus every adjustment or stoppage | In-process control sheet in the BMR |
| Compression | Friability | At defined points during the run | In-process control sheet |
| Compression | Disintegration | At defined points during the run | In-process control sheet |
| Transfer | Metal detection, deduster, yield reconciliation | Continuous; reconciliation once per batch | BMR reconciliation page |
| Bulk hold | QC release against specification | Once per batch | QC analytical report; release note |
| Blistering | Foil/film grade and lot, overprint accuracy | At start-up and at intervals | Batch packing record |
| Blistering | Leak/seal integrity test | Start-up, then ~every 30 min | Packing in-process sheet |
| Blistering | Pocket fill, appearance, carton coding | Continuous plus interval checks | Batch packing record |
| End of pack | Full material and unit reconciliation | Once per batch | Batch packing record |
Exact frequencies are fixed by the approved batch record for each product, not by a universal rule — the column above shows what is common, not what is mandated.
What an auditor actually looks for
Not the results. The gaps. An in-process sheet with no entries during a two-hour window, a hardness value recorded to a suspiciously consistent decimal, a reconciliation that closes exactly every time, an adjustment noted in the machine log with no corresponding in-process check — these are what get pulled. Revised Schedule M has sharpened the point by requiring data integrity controls in ALCOA+ terms, audit trails on critical data, quality risk management on ICH Q9 lines and an annual product quality review. The direction of travel is toward records that can be checked against each other, not just read.
How Salus approaches this
Salus Pharmaceuticals manufactures tablets at its WHO-GMP and ISO 9001:2015 certified facility in Baddi, Himachal Pradesh, and has done since 2005, with tablets and capsules supplied in blister packs and alu-alu. In-process controls, their frequencies and their acceptance limits are fixed in the approved batch manufacturing and packing records for each product, and the completed records are available for review during a customer or third-party audit.
Buyers and importers evaluating the site are welcome to ask for the in-process control specification for a specific product before an audit is scheduled.
Leave a Reply