Filter Integrity Testing for Cell and Gene Therapy

Most cell and gene therapy products cannot be terminally sterilized, so the sterilizing filter is the last line that stands between a contaminated batch and the patient. Filter integrity testing is the physical check that proves the filter did its job.

Quick answer

Filter integrity testing proves a sterilizing filter was not damaged during use. It does not look for microbes directly. Instead it checks the membrane, housing and seal for defects that would let organisms through. In cell and gene therapy this check is mandatory because most products cannot be terminally sterilized.

Why filter integrity testing matters in CGT

Cell and gene therapy products are built on living cells. Heat, radiation and most chemical sterilants kill those cells, so terminal sterilization is off the table for most formulations. That leaves sterilizing filtration as the core barrier of the aseptic process, against bacteria and fungi in the raw materials, media and buffers.

What makes that barrier believable is filter integrity testing. It never cultures the product to look for microbes. It checks the filter itself: the membrane, the housing and the seals. A tear, a pinhole or a failed seal lets organisms through while the product still looks fine, and the test catches the defect before the batch moves forward.

The risk hides in plain sight. Single-use, pre-assembled filter assemblies get handled during gamma irradiation, shipping, cleanroom set-up and fluid flow. Any of those steps can open a microbreak no one can see. The final sterility test will not reliably catch it either. A small CGT batch offers only a limited sample, and low-level contamination slips past often enough that regulators and the technical guidance for cell therapy products now treat integrity testing as a required control.

Dropping the post-use test is a serious error. The pre-use test catches install damage. The post-use test proves the membrane and seals survived the process fluid, pressure and temperature. Pre-use and post-use testing together form the evidence chain, and a missing post-use test counts as a major defect in inspection.

Filter integrity testing: the four methods at a glance

Four methods are in common use. Their physics, the membranes they fit, their limits and their reading rules differ, and they are not interchangeable.

MethodWhat it measuresBest forSensitivity to small defectsRead rule
Bubble point testPressure at which gas first breaks through a wetted membraneHydrophilic membranes, small areasModerateMeasured bubble point at or above the supplier value
Diffusion flow testGas flow across a wetted membrane below bubble pointBoth hydrophilic and hydrophobic membranes, large or high-risk areasHighMeasured flow at or below the supplier limit
Water intrusion testVolume of water pushed into a hydrophobic membraneHydrophobic vent filtersHigh for hydrophobic membranesMeasured intrusion at or below the supplier limit
Pressure hold testPressure drop across the closed systemWhole single-use assemblies and weldsLow, system-level onlyPressure decay at or below the set limit

The choice depends on whether the membrane is hydrophilic or hydrophobic, how the filter is built, what fluid runs through it and what the cleanroom can support. Picking the wrong method gives a false reading and puts the release decision at risk.

Bubble point test: principle and use

The bubble point test leads for hydrophilic sterilizing membranes such as PES, hydrophilic PVDF and mixed cellulose ester. The physics is a capillary model: liquid fills the pores and a surface-tension seal holds it there. Upstream gas pressure has to beat that tension before it can push the liquid out and form a steady stream of bubbles downstream. The pressure at the first continuous bubbles is the measured bubble point, and the filter passes if it meets or exceeds the supplier’s stated value.

Wetting is where the result is won or lost. PES and hydrophilic PVDF take purified water, and every pore has to fill. A dry spot lets gas sneak through and reports a bubble point that is too low, a false fail. Plenty of CGT teams rinse for a few seconds and go straight to pressure. The membrane is not fully wetted, the test fails, and release slips while the test is repeated. After wetting, clear the trapped air and raise pressure slowly. A fast ramp misreads the result and can even burst a good membrane.

The test is cheap, fast and easy inside a Grade A or B cleanroom, which suits the small-batch, fragmented CGT workflow. Media and buffer filters usually use it. The downsides are real, though. The bubble point test misses the tiniest pinholes that diffusion flow would catch. Its reading moves with the wetting fluid, the temperature and the membrane lot, because temperature shifts surface tension, so ambient temperature must be recorded and the result corrected when needed. It only works on hydrophilic membranes. Water will not wet a hydrophobic PTFE or PVDF vent filter, so a bubble point test there returns useless numbers, and a hydrophobic filter needs the water intrusion test instead.

Diffusion flow test: the most sensitive option

The diffusion flow test, also called the forward flow test, is the most sensitive and the most flexible of the four. It works on hydrophilic and hydrophobic membranes alike. Below the bubble point, gas dissolves into the pore liquid and diffuses across the membrane along a concentration gradient, releasing a steady flow downstream. An intact membrane keeps that flow within the supplier’s limit. A pinhole or tear opens a path for bulk gas flow, and the reading shoots past the limit.

Sensitivity is the reason to pick it. The bubble point test catches only large breaks. Diffusion flow finds micron-scale pinholes, which matters for high-risk steps, large filter areas and clinical batches where a miss is costly. Viral vector material filtration and allogeneic CAR-T batches are common cases where the diffusion flow test is the recommended choice. Because it runs below bubble point, it does not stress the membrane, so the test itself cannot damage the filter.

The cost is time and equipment. A dedicated high-precision integrity tester is needed, with annual calibration and higher upkeep. A single test takes longer than a bubble point test, and running several filters in parallel stretches the time the cleanroom stays occupied. Wetting still has to be done properly. A dry area lets bulk gas through and pushes the reading artificially high, a false pass risk.

Method validation matters. Many teams take the supplier’s diffusion flow limit as given and never challenge it with their own process fluid. Protein in media changes the membrane over time and shifts the baseline, so the limit has to be set under process conditions, not under pure water. The reading also trends. Several batches drifting slowly upward while still inside the limit is a warning sign. The cause could be membrane aging, poor wetting or a slow line leak, and it deserves investigation before a batch fails.

Water intrusion test: for hydrophobic vent filters

The water intrusion test exists for hydrophobic membranes such as PTFE and hydrophobic PVDF. Water cannot wet these pores on its own, so under set pressure only a small volume enters the surface and none crosses the membrane. The volume that enters during the test window is the water intrusion. At or below the supplier limit, the filter passes. Above it, water crosses the membrane and the microbial barrier is gone.

In CGT cell culture, hydrophobic membranes almost always sit in the vent line of a culture bag or bioreactor, as 0.2 micrometer vent filters that keep bacteria and fungi in the room air from reaching the culture. These vents fail in a way of their own. Condensate and aerosols collect in the housing during culture. Once the hydrophobic membrane gets fully wetted, it loses its repellency, the culture loses its protection, and the gas path clogs. The water intrusion test is therefore mandatory for vent filters at the end of a culture run, to prove the membrane survived long ventilation and condensate exposure. Inspectors flag this gap often, because labs assume the vent filter left the factory sterile and skip the post-culture test.

The boundary is strict. The water intrusion test cannot be used on hydrophilic membranes, because water crosses them directly. Pressure has to stay within the supplier’s range. Too much pressure forces water into the pores and can wet the membrane permanently, destroying the hydrophobic property. After the test the filter has to be dried so no water stays in the pores.

The practical traps are clear. Wrong pressure, too short a test window, or treating a vent filter’s result as proof of liquid filtration all cause trouble. A hydrophobic vent filter only sterilizes gas and must never be used to filter liquid media. A single-use assembly that cannot be taken apart has to be tested as a whole, and a line leak has to be told apart from a membrane break.

Pressure hold test: checking the whole closed system

The pressure hold test, also called the pressure decay test, is different in kind from the other three. Bubble point, diffusion flow and water intrusion test the membrane itself. The pressure hold test checks the entire closed system: tubing, welds, seals and thermal seals. Gas is charged into the closed assembly, the source is closed, and the pressure drop over a fixed time is measured. At or below the set decay limit, the system is sealed.

In fully closed, automated CGT processes and the single-use systems that carry them, this test confirms that sterile connect welds and thermal seals do not leak. The weld joint is the weak point of the closed system, and a cold weld or a crushed tube is hard to see. The pressure hold test finds it. It cannot stand in for a membrane test, though. A decay only shows that the system leaks somewhere. It cannot tell whether the leak is at a joint or in the membrane, so it does not replace bubble point, diffusion flow or water intrusion.

The setup has to match the part. Larger tubing holds pressure longer for the same leak, so the pressure, hold time and decay limit have to be set per assembly size, not copied across the line. Flexible plastic stretches under pressure and the reading drifts slowly; that is the material, not a leak, and validation has to separate the two. In CGT the test confirms weld seals after sterile connection, thermal seals after sampling, and the whole single-use assembly after build. The gas used must be sterile, from a 0.2 micrometer filter, so the test does not contaminate the sterile surfaces. Pressure is released after the test to avoid fatigue cracks.

How to choose a filter integrity test

Filter integrity testing starts with the membrane, and integrity test method selection then weighs the risk. Hydrophilic PES, PVDF and MCE liquid filters take the bubble point test first. High-risk batches and large areas move up to the diffusion flow test. PTFE and hydrophobic PVDF vent filters take the water intrusion test by default. Whole single-use lines and sterile welds take the pressure hold test. The methods are not free to swap.

By sensitivity the order is diffusion flow first, then bubble point, then water intrusion for hydrophobic membranes only, then pressure hold last. Pressure hold only finds large system leaks and cannot see small membrane breaks. By effort the bubble point test is fastest and simplest. Diffusion flow is slower and needs more gear. Water intrusion needs pressure control. Pressure hold is simple to run but heavy on method validation per line size.

Risk level sets the plan. Media and buffers are low-risk, small-area liquid filters, so the bubble point test is enough. Viral vector filtration and high-potency cell batches are high-risk even on hydrophilic membranes, so diffusion flow is preferred. Bioreactor and culture bag vents are fixed to the water intrusion test. Single-use line welds and thermal seals take the pressure hold test.

ScenarioMembraneRecommended test
Media or buffer, small areaHydrophilic PES, PVDF, MCEBubble point test
Viral vector or high-potency cell batchHydrophilic, any areaDiffusion flow test
Large filter area, high-risk releaseAnyDiffusion flow test
Culture bag or bioreactor ventHydrophobic PTFE, PVDFWater intrusion test
Single-use line, sterile weld, thermal sealWhole systemPressure hold test

Changing methods needs change control. Moving from bubble point to diffusion flow, or changing the membrane material, is a major change. A risk assessment and a comparison study have to prove the two methods are equivalent, the SOP has to be updated, staff retrained, and the process revalidated if needed. Swapping a method without assessment is not allowed.

Validation and batch records

Filter integrity testing stands on a chain of evidence: bacterial retention validation, method validation, instrument calibration, staff qualification and record control.

Bacterial retention validation is the base layer. Under pharmacopoeial method, Brevundimonas diminuta is the challenge organism at a concentration of 10 to the power of 7 CFU per square centimeter of membrane area, tested under the process temperature, pressure, fluid and pH. It proves the filter retains organisms under process conditions. Integrity testing proves the actual filter used in each batch was not broken. The two are separate and both are required.

Method validation covers precision, repeatability, detection limit and interference. It checks whether the process fluid, temperature, pH and protein change the result, and whether wetting time, ramp rate and hold time move it. Many CGT teams copy the supplier’s pure-water parameters and ignore how media protein coats the membrane and shifts the baseline. That is a common inspection finding. Challenged membranes with artificial defects have to be made to prove the method can find a set defect size.

The instrument needs outside calibration every year, with the certificate archived, and a daily performance check against a standard leak. Operators need training on each method’s principle, wetting, pressure control and reading, and they have to pass before they touch a filter. Training also covers the failure path: when integrity testing fails, the batch is held, the deviation process starts, the root cause is found, CAPA is applied, the product risk is judged, and the batch is scrapped if needed.

Records have to be complete. Each test logs the filter model, filter lot, method, wetting fluid, ambient temperature, pressure, test time, measured value, supplier limit, pass or fail call, operator and time. The raw pressure and flow data go with the batch release package alongside the media fill, environmental monitoring and final sterility data. Only logging pass or fail, or back-filling records after the fact, is not allowed.

Common gaps found in inspections

Filter integrity testing shows up often in CGT aseptic findings. The gaps fall into four groups: wrong method, poor execution, missing validation and weak system management.

Wrong method selection is the most serious. Using a pressure hold test in place of a membrane test, assuming a sealed line means an intact membrane, running a bubble point test on a hydrophobic vent filter, or picking a hydrophobic PTFE filter for liquid and then a water intrusion test, all break the barrier proof. These are design-level errors, and inspectors rate them high risk.

Poor execution is common at the bench. Wetting is rushed with no set time. The ramp rate is uncontrolled. The result is not corrected for process temperature. Ambient temperature is not logged. An integrated single-use filter gets only the pre-use test, and the post-use test is dropped. A vent filter after a long culture run gets no water intrusion test. Some staff assume gamma-irradiated single-use parts leave the factory intact and cannot break in use, forgetting that fluid shock, pressure cycles and condensate can damage the membrane during the run. The post-use test is the forced proof that the barrier held.

Missing validation covers the absence of a bacterial retention report, no method suitability study under the process fluid, no comparison after a filter or supplier change, and no challenged-membrane data to prove the method finds defects. Many teams keep only the supplier’s filter sheet and never validate under their own media and buffers.

Weak system management shows up when a failure investigation stops at operator error without finding why the membrane broke, when CAPA is never confirmed effective, when training lacks a hands-on check, when the instrument is not calibrated on time, and when trend data is never reviewed. A slow drift toward the limit is a warning the team never acted on.

The fix is layered. Short term: review past batches for risk, add microbial testing if needed, and retrain staff with a hands-on check. Medium to long term: complete the validation data set, rewrite the SOP with a clear method matrix that separates membrane tests from pressure hold, build a monthly trend review, and trigger a risk assessment on any process, material or staff change.

Where integrity testing sits in the aseptic system

Filter integrity testing is a core part of aseptic assurance, not a standalone task. It is one of five defenses: consumable sterilization control, media fill, environmental monitoring, personnel monitoring and final sterility testing. Raw materials rely on sterilizing filtration, and integrity testing proves that filtration worked. If it fails, the filtered material is treated as contaminated no matter what the environment or final test shows.

The media fill proves the process, staff and environment can run sterile. Integrity testing proves the filtration unit operation is reliable. A closed process can slim down its media fill, but it cannot drop integrity testing. Environmental and personnel monitoring control the risk people and the room bring. Integrity testing controls the risk the material brings. The two risks differ and back each other up. The final sterility test is the last gate, but its small sample only catches high-level contamination. Integrity testing catches a broken barrier right after filtration and stops a bad batch before it costs more.

The data also feeds continuous improvement. Several batches drifting upward on bubble point, diffusion flow or water intrusion, even while still inside the limit, should trigger a review of consumable lots, operator habits, cleanroom swings and membrane aging. In deviations, changes and annual product reviews, integrity testing data is a core input that keeps the aseptic system under control.

Limitations

Integrity testing infers barrier status. It does not culture or count microbes, so it must pair with the final sterility test rather than replace it. A pass reading only proves the filter if the right method was chosen and validated for the actual fluid and temperature. Copied supplier limits are a known gap. Small CGT batches limit the sample size of the final sterility test, which makes the pre-use and post-use checks matter more, not less. A pressure hold test cannot see small membrane breaks, and treating it as a membrane test is a major defect.

Bottom line

Filter integrity testing is the physical proof that a sterilizing filter held its barrier in cell and gene therapy. The four methods cover different membranes and risks, and no single test fits all. Method choice, execution, validation and deviation handling each have to hold, or the barrier cannot be shown to work. Quality and process staff must keep the four methods and their roles straight, build a testing plan that fits their process, and give inspectors verifiable evidence.

FAQ

What is filter integrity testing

Filter integrity testing is a physical check that a sterilizing filter was not damaged in use. It does not culture for microbes but checks the membrane, housing and seals for tears, pinholes or failed seals that would let organisms through. A passing result, read before and after filtration, shows the filter can still retain microbes under process conditions.

Why is filter integrity testing required for CGT products

Most CGT products use living cells and cannot be terminally sterilized, so filtration is the main barrier. A broken filter lets contaminants into the batch with no visible sign. Integrity testing proves the barrier held, and the technical guidance for cell therapy products, together with the regulations for investigator-initiated trials, treats it as a required control rather than an optional check.

What is the difference between bubble point and diffusion flow tests

A bubble point test finds the pressure where gas first breaks through a wetted membrane and suits hydrophilic filters with small areas. A diffusion flow test measures gas crossing the membrane below that pressure and finds micron-scale pinholes the bubble point test misses. Diffusion flow is the more sensitive choice for high-risk or large-area filters.

Can a pressure hold test replace a filter integrity test

No. A pressure hold test checks the whole closed system for leaks but cannot tell whether the leak is at a weld or in the membrane. It only finds large system leaks and misses small membrane breaks that would still let organisms through. It must be used with, not instead of, bubble point, diffusion flow or water intrusion tests, and only after sterile welds or for whole-assembly checks.

How do you choose the right filter integrity test

Pick by membrane and risk. Hydrophilic liquid filters take the bubble point test first and move to diffusion flow for high-risk or large areas, while hydrophobic vent filters take the water intrusion test and whole single-use lines take the pressure hold test. The methods are not free to swap without a risk assessment and a study proving they are equivalent.

What does wetting do in a bubble point test

Wetting fills every membrane pore with liquid so surface tension seals it. A dry spot lets gas slip through and reports a bubble point that is too low, a false fail that wastes a release cycle. Purified water wets PES and hydrophilic PVDF, but the soak must be complete, trapped air cleared from the lines, and pressure raised slowly.

What is a water intrusion test used for

A water intrusion test is for hydrophobic membranes such as PTFE vent filters. Water cannot wet these pores, so only a small volume enters the surface under set pressure. Above the supplier limit, water crosses the membrane and the barrier fails, which is why the test is mandatory for culture vent filters after a long run, when condensate can wet the membrane.

How is a bacterial retention test different from integrity testing

A bacterial retention test challenges the filter with Brevundimonas diminuta at 10 to the power of 7 CFU per square centimeter under process conditions to prove it retains organisms. Integrity testing proves the actual filter used in a batch was not broken. The two are separate steps and both are required for a valid barrier claim.

What counts as passing an integrity test

A bubble point result at or above the supplier value passes. A diffusion flow or water intrusion result at or below the supplier limit passes, and a pressure hold result at or below the set decay limit passes. The limits must be set under the actual process fluid and temperature, not copied from pure-water supplier data.

What happens when filter integrity testing fails

The batch is held and the deviation process starts at once. Staff isolate the material, find the root cause, separate a factory defect from install damage or poor wetting, apply CAPA and judge the product risk, and the batch may be scrapped. Trend data from past batches is reviewed to catch a slow drift toward the limit before the next failure.

Does filter integrity testing detect microorganisms

No. Integrity testing infers microbial security from the physical state of the filter. It finds defects that would let organisms through but does not count or culture them. The final sterility test looks for microbes, and its small sample only catches high-level contamination, which is why integrity testing is done right after filtration.

How often should filter integrity testing be done

It is done per batch, before and after filtration of each protected material. Pre-use testing catches install damage and post-use testing proves the membrane survived the run, while vent filters get a water intrusion test at the end of the culture. Method limits are reviewed each month and on any process, material or staff change.

Comments are closed