European manufacturing rules are pushing drug makers to verify sterilizing grade filter integrity before product flows, and engineering is responding with single use assemblies and automated test platforms that turn the procedure into a designed in step.
Biopharma manufacturers are being asked to prove their sterilizing-grade filters work before any drug touches them, and the practical answer is showing up not in regulatory filings but in the hardware. A tightening of European manufacturing rules has turned a long-debated quality check called PUPSIT, pre-use post-sterilization integrity testing, into a near-default expectation inside new production lines, and the engineering community is responding with pre-assembled single-use flow paths and automated test platforms designed to make the procedure less painful than the regulation implies.
The mechanical idea is straightforward. A sterilizing-grade filter is the last barrier before a sterile drug product enters its final container. Conventional testing happens after the filter has been used: a bubble point or forward-flow measurement runs on a membrane that has already processed product. PUPSIT runs the same integrity test, but earlier, on a filter that has been autoclaved or otherwise sterilized but not yet exposed to the batch. The point is to catch damage introduced by sterilization, transport, or installation before a compromised membrane processes a single dose.
PUPSIT has existed in pharmacopeial guidance for years, but the 2022 revision of EU GMP Annex 1 tightened the language around sterile manufacturing, and Annex 1 now generally expects pre-use post-sterilization verification unless a manufacturer can produce a documented, risk-based rationale for skipping it. PIC/S, the international cooperation scheme that aligns inspectorates across dozens of non-EU regulators, has been propagating a similar expectation, so the rule is migrating outward from Europe even where national pharmacopeias have not formally changed.
The case for the test is that post-use integrity testing has a known blind spot. A filter that is mechanically damaged before use can still process an entire batch and pass an after-the-fact test, because the test measures what the filter can hold back under challenge, not the state of every square centimeter of membrane. The cost of a missed failure is not theoretical: a batch is scrapped, a line is shut down, and a regulatory conversation follows. The argument against has been that PUPSIT adds wet, vent, and test steps that themselves create contamination risk on the sterile side of the system, that the procedure is operator-dependent in manual setups, and that the evidence base for the test preventing real-world failures is thinner than the regulatory language suggests. That debate is not closed, and a piece of industry guidance written by users of the technology continues to push back on making the test universal.
Inside the equipment vendors, the response to the new expectation has been to redesign the flow path so the operator is no longer the integration point. Martin Glanz, a senior principal scientist at Cytiva and lead author on a 2026 study in Applied Microbiology and Biotechnology, describes the shift in a Genetic Engineering & Biotechnology News interview as a move away from testing a single cartridge in isolation toward testing the entire assembly: tubing, connectors, vent filters, and the sterilizing-grade filter as a single, preassembled, single-use unit. The integrity test, in that design, becomes a property of the system rather than a manual procedure on one component. Glanz, who works for a major supplier of the filters and integrity testing equipment being adopted, frames this as a quality upgrade rather than a regulatory burden, and the underlying engineering is real even if the framing comes from a vendor with commercial interest in the outcome.
The second-order change is in the test instrument itself. Manual wetting, manual venting, and manual bubble-point reading are giving way to benchtop or at-line platforms that walk an operator through the steps, capture the data electronically, and produce an audit-ready record. The same trend is showing up in autoclave loading, sterile connection, and pre-use flush procedures: anything that used to depend on a trained hand now ships as a fixture, a script, or a closed disposable. The result, in factories that have made the change, is that PUPSIT compliance looks less like a procedure operators perform and more like a property of the bag-and-instrument set they pick up at the start of a batch.
The unresolved question is whether the new engineering is keeping pace with the new expectation, or papering over a procedure whose clinical value is still contested. Annex 1 is now in force, inspectors are writing observations about pre-use verification, and PIC/S member authorities are aligning their expectations, so the regulatory pressure is not going to soften. What remains to be seen is whether the next generation of single-use assemblies and automated test platforms will close the manual-intervention gap that critics of PUPSIT have pointed to for years, and whether the evidence base for the test itself will catch up to the confidence the rules now place in it.