Why Your Sterility Test Failed: A Root Cause Analysis

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-11-10

A positive sterility test is a data point, not a verdict. The single most likely reason a compendial sterility test turns positive is contamination introduced during the test itself, and the first thing to check is the session record for that specific test: isolator or cleanroom differential pressures, HEPA filter and glove integrity results, the hydrogen peroxide decontamination cycle printout, the environmental monitoring plates exposed during the run, and the analyst's gowning and transfer technique. The second most common reason is a media problem, usually a growth promotion or media sterility check that passed on paper but not in practice.

A sterility test failure root cause analysis starts with the test system, not the manufacturing suite. Most investigations run the other way: the positive result triggers a manufacturing review, three weeks pass, the batch record is clean, and everyone returns to the laboratory with less time and fewer samples than they started with. Reversing that order saves the batch and the investigation.

What follows is a working sequence for microbiology quality control and laboratory quality management teams: the mechanism behind each failure mode, ordered from most common to rarest, the evidence that confirms it, the fix, a troubleshooting table for the bench, a ten-step diagnostic sequence, and the preventive controls that stop the same failure from recurring.

What the Sterility Test Actually Measures

The compendial sterility test, described in United States Pharmacopeia Chapter <71> Sterility Tests and its national equivalents, is a presence or absence test. It does not count organisms and it does not prove that every unit in a batch is sterile. It samples a defined quantity of product, exposes that sample to conditions that will support growth if viable organisms are present, and reads the result after 14 days.

Two methods dominate. In membrane filtration, the product is drawn through a membrane with a nominal pore size of 0.45 micrometers, which retains bacteria and fungi. The membrane is rinsed with a validated volume of diluent to remove product residues that could inhibit growth, then placed onto two media. In direct inoculation, the product is added directly to the media, which is the standard approach for products that cannot be filtered, such as ointments, suspensions, and some devices.

The two media split the microbial world. Fluid Thioglycollate Medium (FTM) is incubated at 30 to 35 degrees Celsius and supports anaerobic and microaerophilic bacteria. Soybean Casein Digest Medium (SCDM) is incubated at 20 to 25 degrees Celsius and supports aerobic bacteria and fungi. Both are held for the full 14 days and examined at intervals for turbidity, colony formation, or both. FTM must be heated once before use to drive off dissolved oxygen and then cooled, and it must never be re-heated repeatedly, because each cycle pushes the medium toward oxidation and toward chemical changes that suppress growth.

Before any of that matters, three quality gates must pass.

Media sterility. A sample of every prepared lot is incubated to prove it is sterile.

Growth promotion. Each lot is challenged with fewer than 100 colony forming units of defined reference organisms and must show visible growth in the specified short window. For FTM that panel includes Staphylococcus aureus, Pseudomonas aeruginosa, Bacillus subtilis, and Clostridium sporogenes. For SCDM it includes Bacillus subtilis, Candida albicans, and Aspergillus brasiliensis. A lot that fails to support growth of any member of the panel is not fit for use, and a sterility test run on that lot is invalid.

Method suitability. Also called the bacteriostasis and fungistasis test, this gate proves that the product itself does not inhibit the challenge organisms under the exact conditions of the test. It is the gate most often treated as a formality and the one most often responsible for confusing results.

Two very different things get called a failed sterility test. An invalid test is one where the method, media, equipment, or handling did not meet requirements, so the result carries no information about the product. A positive test is one where growth appeared under valid conditions. The distinction determines whether the batch is retested, rejected, or held while a manufacturing investigation runs. Confusing the two is the most expensive mistake in the whole process.

A negative result, by contrast, has a statistical limit set by the number of units sampled and the volume filtered. Sterility testing is destructive, so the test cannot examine the entire batch. A passing result is evidence of a controlled process, not proof that each individual unit is sterile.

Root Cause Analysis for the Microbiology Laboratory

Root cause analysis systematically identifies the underlying faults behind a failure rather than the visible symptom [1]. That definition matters in a microbiology laboratory, where the visible symptom is always the same, growth in a tube, and the underlying faults range from a leaking glove to a raw material supplier.

The analytical burden is real. Rule-of-thumb triage collapses when the evidence is heterogeneous, when organism identifications, environmental monitoring trends, personnel records, equipment logs, and batch records all live in different systems and none of them line up on a single timeline [1]. Two practices fix most of that.

Binning. Group every positive result by organism genus and species, by media, by analyst, by shift, by test session, by isolator, by product, and by month. Real investigations rarely proceed one failure at a time. A cluster of positives caused by a single contaminated water bath will look unrelated until the results are sorted into a bin and the common element appears. Automated triage and binning of failures narrows the debugging scope to a small set of candidates instead of forcing a review of every event [2].

Data joining. Pull the environmental monitoring results, the decontamination cycle records, the maintenance log, the batch record, and the analyst's training file into one view for the relevant date range [3]. Modern failure analysis in manufacturing settings has moved toward exactly this: connecting separate databases so that a failure analyst can see the plant-level history and drill down to a single unit [3]. A sterility test investigation is the same problem at smaller scale.

Two further points from failure analysis practice apply directly. First, a single failure mode is often several failure modes wearing one coat. A structured review of equipment failures in high-voltage systems identified twelve distinct failure modes and fourteen corrective activities behind a class of failures that had been treated as one problem [4]. Expect a sterility investigation to generate more than one corrective action. Second, the method of investigation must fit the object being investigated. Advanced package failure analysis required purpose-built sample preparation and electrical characterization before root cause could be reached [5]. A sterility investigation that consists only of a batch record review is a method that does not fit the object.

A failure mechanism can sit latent for a long time and appear only when conditions align. Metallurgical failure analysis of a reactor orifice assembly found stress corrosion cracking driven by residual tensile stress created by geometric constraint during welding, with the crack propagating slowly and invisibly until the remaining ligament could no longer carry the load [6]. Contamination events behave the same way. A transfer technique that is marginally acceptable fails only on the day the isolator airflow is slightly off, the analyst is rushed, and the vial neck is touched.

The Causes, Most Common to Rarest

1. Contamination Introduced During the Test Itself

This is the leading cause by a wide margin, and it is the one investigators are most reluctant to accept because it implicates people and facilities rather than product.

The mechanism is direct. An organism enters the test system from the analyst's skin, gloves, or sleeve, from a non-sterile surface touched during transfer, from a glove port or sleeve leak on an isolator, from a compromised HEPA filter, or from a decontamination cycle that did not reach the required concentration or contact time. The organism then deposits on the membrane, in the medium, or on the container closure and grows during the 14-day incubation.

The evidence pattern is characteristic. The organism is a coagulase-negative Staphylococcus, Micrococcus, Cutibacterium, Bacillus, Aspergillus, or Penicillium species, all of which are normal skin or environmental flora. It appears in one medium only, because a single contamination event rarely lands in both. It appears in a single unit rather than across all units tested. It matches an organism recovered from the same session's environmental monitoring plates, from a glove print, or from a personnel isolate. And the bioburden history for the product is flat.

Confirmation requires more than a genus-level match. Identification to species by MALDI-TOF mass spectrometry or by 16S ribosomal RNA gene sequencing for bacteria, and by internal transcribed spacer sequencing for fungi, establishes what the organism is. Strain-level comparison by whole genome sequencing or pulsed-field gel electrophoresis establishes whether the product isolate and the environmental isolate are the same organism. A shared species is weak evidence. An indistinguishable genome is strong evidence.

The fix is procedural and physical. Retrain and re-qualify the analyst on gowning and transfer technique. Re-qualify the isolator with glove integrity testing, HEPA integrity testing, and a validated decontamination cycle. Increase the environmental monitoring density inside the isolator during testing so the next event has a data trail. Review the session's particle counts and differential pressures against the alarm history.

Safety note. Hydrogen peroxide used for isolator decontamination is corrosive to skin, eyes, and airways. Handle it with the specified gloves and eye protection, verify that residual levels are within the validated limit before entry, and never assume a completed cycle means a safe atmosphere without a residual check.

2. Media, Reagent, and Growth Promotion Failures

A media failure produces either a false positive, from a lot that was not sterile, or a false negative, from a lot that cannot support growth. The false negative is the more dangerous of the two because it silently releases product.

Mechanisms include overheating during preparation, which caramelizes sugars and generates Maillard reaction products that suppress growth. Overheating also oxidizes the upper layer of FTM, producing a pink band that signals an anaerobic medium no longer fit for anaerobes. Expired media, incorrect weighing, incomplete dissolution, pH drift outside the specified range, storage at the wrong temperature or under the wrong atmosphere, and an autoclave cycle that ran at the wrong pressure or hold time all produce the same outcome.

Process parameter drift as a root cause is well documented in engineering failure analysis. A welded joint in a steam line failed because unsuitable welding parameters produced large grain structures and low strength at the weld root, with residual tensile stress concentrated there [7]. The parallel is exact. A media preparation step that drifts outside its validated parameters produces a medium that looks acceptable and behaves poorly.

Confirmation is straightforward and should be routine. Review the lot's certificate of analysis and in-house growth promotion records. Re-incubate a retained sample of the same lot. Repeat growth promotion with the reference panel on the suspect lot. Pull the autoclave chart and verify the cycle parameters against the validated load. Measure the pH of the prepared medium. Inspect the FTM for oxidation and for particulate matter.

The fix starts with lot release discipline. No medium enters the laboratory without in-house growth promotion and sterility results, regardless of what the supplier's certificate says. Preparation records must capture weight, volume, heating time, and sterilization cycle. Prepared medium gets a defined shelf life and defined storage conditions. FTM is heated once, cooled, and used, never boiled and reheated.

Safety note. Autoclave loads operate at roughly 121 degrees Celsius under pressure. Allow the chamber and the load to reach atmospheric pressure and a safe handling temperature before opening, and use the specified thermal protection when removing hot carboys or bottles.

3. Method Suitability Gaps and Product Inhibition

This is the failure mode that produces results nobody can explain, because the product kills the challenge organism before it can grow, and the investigation keeps looking for contamination that never existed.

The mechanism is residual antimicrobial activity. Many products contain preservatives, antibiotics, or biocides. Others bind to the membrane and are not removed by the rinse volume used. Some have a pH or osmolarity outside the range that supports growth. When the neutralization is incomplete, the challenge organism in the suitability test is inhibited and the test fails, or, more subtly, an organism that enters the test dies before day 14 and the test reads negative when it should read positive.

The evidence pattern is a growth promotion test that passes on un-inoculated medium, a method suitability test that is marginal or that passes only with a rinse volume never used in routine testing, or a positive result in a product test that cannot be subcultured onto fresh medium. A positive that will not grow on subculture is a warning, not a confirmation.

Confirmation is a repeat suitability test with a defined challenge and a deliberate comparison of neutralizer conditions. Standard neutralization approaches include beta-lactamase for beta-lactam antibiotics, lecithin and polysorbate 80 for many preservatives and surfactants, sodium thiosulfate for oxidizing agents, and EDTA or specific enzymes for other product classes. The neutralizer must itself be shown to be non-toxic to the challenge organisms, which is a step that gets skipped.

Design weaknesses concentrate risk in exactly this way. Failure analysis of a vehicle drive shaft found fatigue cracks initiating at the root fillet of the spline gear, where small geometric dimensions produced a high stress concentration, and poor manufacturing tolerance at that same fillet made it worse [8]. In a sterility method, the equivalent stress concentration is the least robust step: a rinse volume that barely works, a neutralizer that barely neutralizes, a filter that loads heavily. Robust methods have margin at every step.

Safety note. Beta-lactamase powders can cause sensitization in exposed workers. Weigh and handle them in a contained area with appropriate respiratory and skin protection, and follow the facility's occupational health program.

4. Sample Handling, Transport, and Chain of Custody

Delay, temperature abuse, and physical damage between sampling and testing introduce organisms or invalidate the sample.

The mechanism is either contamination from the outside in, through a damaged container or a compromised closure, or loss of viability from excessive heat, freezing, or prolonged hold. Sampling from a non-homogeneous product, such as a suspension that settled during transport, produces a sample that does not represent the batch. A mislabeled container produces a result on the wrong lot, which is a data integrity event rather than a microbiological one.

Confirmation relies on records, not on cultures. Pull the sampling time, the receipt time, the courier's temperature data, the container inspection record, and the chain-of-custody signatures. Look for gaps over a few hours, missing temperature logs, and any transfer of custody without documentation.

The fix is a validated transport container with a qualified hold time, a sample receipt inspection step that documents container condition, and a defined maximum interval between sampling and incubation. Where a sample fails receipt inspection, invalidate the test rather than reading the result.

5. Equipment and Facility Failures

Equipment causes both false positives and invalid tests. An incubator that drifts above or below its range stresses organisms and can suppress growth. A filtration unit with a cracked manifold or a failed seal lets unfiltered product bypass the membrane. An isolator with a glove puncture or a sleeve failure admits room air at the moment of transfer. A decontamination cycle that aborted partway leaves viable spores on surfaces.

Detection of hidden physical defects depends entirely on the detection limit of the method chosen. Extremely high-speed X-ray radiography at sub-15 micrometer resolution can now reveal hidden dynamics inside electronic packages that conventional imaging misses [9]. The lesson transfers directly: a vacuum decay test with an inadequate sensitivity, or a visual inspection of a stopper, will not find a leak that a dye ingress test or a helium leak test finds easily.

Confirmation means pulling the records. Review incubator calibration, temperature mapping, and continuous monitoring charts against the test period. Review the filtration unit's integrity test, if one is performed. Review glove integrity, HEPA integrity, particle counts, differential pressure, and the decontamination cycle printout for the exact session.

The fix is a preventive maintenance schedule with documented calibration, continuous monitoring with alarms that are reviewed rather than silenced, and a rule that an excursion produces an automatic invalidation and a documented assessment, not an informal judgment at the bench.

6. Container and Closure Integrity

A container that leaks is not sterile, whatever the test says. Vials with cracks at the neck, stoppers with incomplete seating, and seals that failed during capping admit organisms during incubation or during shipping.

The complicating factor is that ingress can happen after the product was sterile. A positive result from a leaking vial is a genuine positive test result and a genuine container defect, but it is not evidence that the fill process was contaminated. Both conclusions get drawn, and only one is correct.

Confirmation uses container closure integrity testing: dye ingress with methylene blue under vacuum, vacuum decay, helium leak detection, or high-voltage leak detection, depending on the container type. Visual inspection under magnification, with attention to the neck and the sealing surface, catches gross defects.

The fix is a validated container closure system, a capping process under control, and a change control gate that prevents stopper or vial supplier changes without a re-qualification.

7. True Product Contamination

This is the cause everyone fears and the one that appears in the smallest share of investigations. It is also the most consequential, because a genuine positive on a lot indicates that the manufacturing process is not in control.

The mechanism is contamination introduced anywhere upstream of the fill: raw materials carrying bioburden, a water system that has developed a biofilm, component sterilization that was not validated for the load, a hold time exceeded, an HVAC or cleanroom classification failure, a personnel practice that transferred organisms into the critical zone, or an aseptic process that is not capable.

Confirmation is a manufacturing investigation, and it is built on trend data rather than a single point. Bioburden results on the bulk, the components, and the water system. Environmental monitoring trends by location and grade. Aseptic process simulation results. Media fill acceptance rates. Deviation history for the same line. The organism's identity compared with the resident flora of the facility. If the product isolate matches an organism that has been recovered repeatedly from the filling line environment, the case for product contamination is strong.

The fix is a contamination control strategy rather than a discrete corrective action. Hold time validation, sanitization frequency, filtration validation, supplier qualification, gowning qualification, and a formal review of the aseptic process simulation program.

8. Rarer Causes: Data Integrity, Identification Error, and Mixed Events

A meaningful share of apparent failures turn out to be paperwork. A result transcribed onto the wrong line. A plate read upside down. An audit trail that shows a result edited after the fact. A sample labeled with the wrong lot. An incubator log with a gap. An identification run against a contaminated reference library.

Mixed events are also common and rarely recognized. A single positive may have two contributing causes: a marginal method suitability that allowed a small number of organisms to survive a partially neutralizing rinse, plus a technician whose transfer technique was slightly off that day. Chasing one cause and closing the investigation leaves the other in place.

Detection requires audit trail review on the laboratory information management system, second-person verification of critical readings, and a habit of comparing the written record against the physical evidence. The principle that drove yield analysis in semiconductor manufacturing applies: pull the historic data across systems, build a visualization that shows the whole picture, then drill down to the unit [3]. Classifier-based fault segregation has been used to separate fault types with high accuracy in transformer diagnostics, and the same logic applies to sorting laboratory positives by organism and session [10].

Troubleshooting Table

Symptom	Likely cause	How to confirm	Fix
Growth in one medium only, organism is a skin commensal	Contamination during transfer	Identify to species, compare with environmental monitoring and glove print isolates, strain type by whole genome sequencing	Retrain, re-qualify gowning, review isolator decontamination cycle, increase in-session monitoring
Growth in both media, same organism, multiple units of one lot	Product or process contamination	Bioburden trend, water system data, aseptic process simulation results, deviation history	Hold lot, open manufacturing investigation, contamination control strategy review
Media turbid in the negative control	Media lot not sterile	Re-incubate retained medium from the same lot	Reject lot, supplier investigation, strengthen receipt testing
Growth promotion fails on one reference organism	Media preparation or sterilization fault	Repeat growth promotion on the suspect lot with fresh reference cultures, pull autoclave chart	Reject lot, review preparation records, revalidate the autoclave load
Suitability test marginal, product test positive	Incomplete neutralization or inadequate rinse	Repeat bacteriostasis and fungistasis test with increased rinse volume and alternate neutralizer	Revise method, revalidate rinse volume, document neutralizer toxicity check
Positive result that will not subculture onto fresh medium	Organisms killed by residual antimicrobial agent or disinfectant	Subculture onto non-selective media, test for residual hydrogen peroxide or chlorine in the filtrate	Increase rinse, validate residual limit, review disinfectant carryover
Single unit positive, clean environmental monitoring, clean aseptic process simulation	Sporadic laboratory event or low-level product contamination	Strain typing, session record review, binning against prior positives	Tiered decision, targeted monitoring, do not assume laboratory cause without evidence
Incubator temperature excursion during incubation	Equipment failure, invalid test	Review continuous monitoring chart and calibration record	Invalidate affected tests, investigate excursion, restore and verify calibration
Glove or HEPA integrity test failure	Facility contamination pathway	Review integrity test records, particle counts, differential pressure logs	Decontaminate, recertify, invalidate tests performed since last passing test
Fungal growth first seen at day 10 to 14	Slow-growing environmental mold	Identify by internal transcribed spacer sequencing, review fungal environmental monitoring counts	Review HVAC and humidity control, extend monitoring, evaluate incubator placement
Sample received outside the validated hold time with no temperature log	Handling or transport failure	Review courier records, receipt inspection form, chain of custody	Invalidate test, qualify transport container, set a defined maximum interval

A Ten-Step Diagnostic Sequence

Work through these in order. Each step either removes a cause or produces evidence for the next.

Quarantine and preserve. Hold the batch. Do not discard the positive unit, the medium, or the membrane. Photograph the container and the medium as found, with the label visible. Record the day growth was first observed.
Subculture before anything else. Transfer from the positive medium onto Tryptic Soy Agar with 5 percent sheep blood and onto Sabouraud dextrose agar. Incubate at 30 to 35 degrees Celsius and at 20 to 25 degrees Celsius. Store the isolate in duplicate at ultralow temperature for later strain typing.
Perform a Gram stain and a direct microscopic examination. Record morphology, arrangement, and Gram reaction. A pure Gram-positive coccus in clusters points in a different direction than a Gram-negative rod.
Identify to species. Use MALDI-TOF mass spectrometry or 16S ribosomal RNA gene sequencing for bacteria, and internal transcribed spacer sequencing for fungi. Genus-level identification is not enough to support a root cause conclusion.
Pull the session record. Differential pressure, particle counts, glove integrity, HEPA integrity, decontamination cycle printout, environmental monitoring plates exposed during that session, and the analyst's training and qualification file for the test date.
Pull the media and equipment records. Growth promotion and sterility results for the exact lots used, autoclave charts, incubator calibration and continuous monitoring for the full 14 days, and the filtration unit's integrity test.
Repeat the quality gates. Run growth promotion on the retained media lot. Run the bacteriostasis and fungistasis test with the same product lot. A gate that now fails is a strong candidate for the invalid test category.
Bin the result. Compare the organism, the media, the analyst, the shift, the isolator, and the month against every positive result from the previous 12 months. A pattern changes the investigation from a single event to a system problem.
Strain type and compare. Compare the product isolate against environmental monitoring isolates from the same session and against the facility's resident flora. Indistinguishable genomes support a source attribution. Different genomes exclude it.
Classify and act. Decide whether the test is invalid or the product is positive, document the basis, and assign corrective and preventive actions with owners and due dates. Where the evidence does not discriminate, say so and describe the risk-based disposition.

How to Fix It: Corrective and Preventive Actions by Failure Tier

Corrective actions address the specific event. Preventive actions address the system that allowed it. Investigations that produce only corrective actions tend to repeat.

For laboratory contamination, the corrective action is retraining and re-qualification of the analyst, and the preventive action is a change to the monitoring program: more settle plates inside the isolator, routine glove prints, a defined response when a session's particle count exceeds its action level, and a periodic review of transfer technique by direct observation rather than by reading a training file.

For media failures, the corrective action is rejection of the lot and a supplier notification. The preventive action is in-house release testing for every lot, a documented preparation record, a defined shelf life, and a periodic review of autoclave load validation.

For method suitability gaps, the corrective action is a method revision with a validated rinse volume and a qualified neutralizer. The preventive action is a review of every product method at defined intervals to confirm that the neutralization still works and that no formulation change has invalidated it.

For equipment failures, the corrective action is repair and recalibration. The preventive action is a preventive maintenance schedule, alarm review as a documented task rather than an informal habit, and a rule that any excursion automatically invalidates the affected tests.

For genuine product contamination, the corrective action is a batch disposition decision and a manufacturing investigation. The preventive action is a contamination control strategy review with attention to hold times, water system sanitization, component sterilization validation, and aseptic process simulation performance.

For data integrity events, the corrective action is correction of the record with a documented justification. The preventive action is audit trail review as a routine task, second-person verification of critical readings, and access controls that prevent retrospective edits.

Prevention: Microbiology Quality Control and Laboratory Quality Management

Prevention is cheaper than investigation and far cheaper than a rejected batch.

Environmental monitoring with trends, not just results. A single plate within limits means little. A trend line showing an increase in a specific location, a specific shift, or a specific organism is the early warning that lets you act before a sterility test fails.

Aseptic process simulation. Media fills are the closest available measure of whether the aseptic process is capable. They test the process, the personnel, and the facility together, and they generate the organism data that tells you what your resident flora looks like.

Gowning qualification and requalification. Gowning is a skill that degrades without practice and observation. Qualification at hire and at defined intervals, with glove prints and sleeve sampling, catches the drift before it reaches a product test.

Media quality control on receipt and after preparation. Every lot, every time. Supplier certificates are a starting point, not a release criterion.

Equipment calibration and monitoring. Incubators, autoclaves, isolators, and filtration units all need calibration, mapping, and continuous monitoring where the failure mode is silent.

Training and proficiency testing. Analysts should periodically read seeded media and identify unknown organisms. Proficiency testing finds the gap between what a person knows and what a person does.

Deviation and change control. A change to a stopper supplier, a filtration membrane, a disinfectant, or a media supplier is a potential root cause of the next failure. Change control exists to make that connection before the failure does.

Periodic review of positive results. An annual review of every positive sterility test result, binned by organism and by source, is the single most useful input to the next year's contamination control plan. Failure analysis practice has moved toward exactly this kind of historic data review, using stored records to isolate failure modes faster than a unit-by-unit investigation can [3]. Machine-learning classification of fault types has been used to segregate failure modes in other industries with high accuracy, and the same triage logic applies to sorting laboratory positives [10].

Common Mistakes and Limitations

Treating every positive as product contamination. The most common error. It triggers a manufacturing investigation before the laboratory has been cleared, and it wastes the sample and the timeline.

Discarding the positive unit before subculture. Without a viable isolate, strain typing is impossible and the investigation loses its strongest discriminator.

Accepting a genus-level identification as attribution. Staphylococcus epidermidis is on every analyst's skin and in every cleanroom. Genus and species tell you almost nothing about source. Whole genome comparison tells you a great deal.

Reading only at day 14, or reading carelessly at intervals. Slow-growing fungi and stressed organisms can appear late. Interval readings exist to catch the day growth first appears, which is itself diagnostic information.

Assuming a negative retest resolves the failure. A single negative retest may reflect sampling statistics rather than a clean batch. The basis for any retest decision must be documented before the retest is run.

Blaming the analyst by default. Operator contamination is the leading cause, but the conclusion requires evidence. An investigation that starts with the operator and stops there will miss a media lot problem or a facility issue that is still active.

Ignoring residual disinfectant. Hydrogen peroxide, chlorine, and quaternary ammonium residues carried into the filtrate or the medium can suppress growth and produce a false negative, or kill a contaminant and produce a positive that will not subculture.

Not binning. A single positive looks random. Twelve positives sorted by organism, analyst, and month frequently do not.

Closing the investigation with one corrective action. A structured failure review can surface a dozen distinct failure modes and more corrective activities than the team expected [4]. Single-action closures tend to recur.

The limitation to state plainly: this is a framework, not a disposition. Batch release, retest decisions, and deviation classification belong to the qualified microbiologist and the quality unit, and in animal health manufacturing those decisions also involve the veterinarian responsible for the product. Every investigation has to be documented against its own data, and no template substitutes for that record.

Safety Notes for the Bench

Handle broken vials and ampoules with tongs and puncture-resistant gloves. Glass from a failed container is often contaminated on the outside as well as the inside.

Perform all subculturing from positive media inside a biological safety cabinet, and decontaminate the surface before and after.

Autoclave and dispose of positive media, membranes, and filtration hardware as biohazardous waste according to the facility's waste procedure. Do not accumulate positive cultures on open benches.

Follow the facility's chemical hygiene plan for hydrogen peroxide, sodium thiosulfate, chlorine-based disinfectants, and beta-lactamase powders.

Frequently Asked Questions

What is the most common reason a sterility test fails?

Contamination introduced during the test itself. The organism is usually a skin or environmental commensal, appears in a single medium or a single unit, and matches an isolate from the same session's environmental monitoring.

How quickly should an investigation start?

Immediately. Subculture the positive medium and preserve the isolate on the day growth is observed. Identification, session records, and media records can be gathered in parallel while the batch is on hold.

Can a sterility test be repeated after a positive result?

A retest may be justified when the investigation shows a laboratory or materials fault. The decision, the justification, and the number of retests must be defined in the procedure before the retest is performed, and the compendial requirements apply.

What does growth in only one of the two media mean?

It usually points toward a contamination event rather than a product problem, because a single event rarely reaches both media. It can also indicate an organism that grows only under one set of conditions, so identification matters more than the pattern alone.

Is a positive result always a batch failure?

No. A positive result from a leaking container is a real positive and a real defect, and it can occur after a sterile fill. The investigation has to separate the container defect from a process contamination event.

How do I tell laboratory contamination from product contamination?

Strain-level comparison is the strongest tool. If the product isolate is indistinguishable from an environmental or personnel isolate by whole genome sequencing, the laboratory or facility source is supported. If it matches the resident filling line flora, product contamination is supported.

Why did the media pass growth promotion and still fail?

Growth promotion is performed on a sample of the lot under controlled conditions. A lot can pass that test and still be damaged by later overheating, oxidation, or improper storage before use.

What records should I review first?

The session record for the test date, the media lot release records, the incubator continuous monitoring chart for the full incubation, and the analyst's training and qualification file. Those four cover most of the common causes.