Troubleshooting QC Material Failures in Microbiology

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

The most common cause of a failed microbiology QC result is not a defective lot of material. It is the test system wrapped around that material. The control organism has drifted through too many passages, the inoculum is the wrong density, the incubator is running cold, or the medium lost water and its pH during preparation. The first thing to check is the pattern of the failure, not the lot number. A failure that hits every organism on one lot of plates points at the plates. A failure that hits one drug for one organism points at the drug, the dilution, or the way the zone was read. A failure that follows the same organism across several different methods points at the organism itself.

That single observation, made in the first few minutes, eliminates most of the wrong turns. This guide works from the most common cause to the rarest, gives the mechanism behind each failure mode, and provides a diagnostic sequence that isolates one variable at a time.

What Counts as a QC Material Failure

A QC material in microbiology is any preparation with a known, expected result that is run alongside patient specimens to prove the system is working. That includes reference strains used for antimicrobial susceptibility testing, growth promotion organisms for culture media, identification panel controls, biochemical reagent controls, antisera and latex agglutination reagents, stain controls, and the positive and negative controls inside molecular and antigen assays.

The expected range is the interval of results that a properly functioning system should produce. For disk diffusion it is a zone diameter range in millimeters. For broth microdilution it is a range of two-fold dilutions, usually spanning three to five wells. For qualitative tests it is simply the correct answer, positive or negative, with a defined appearance.

A control result outside that range can come from three different places, and it pays to name them before touching anything:

1. The material is genuinely defective. It arrived nonviable, it is contaminated, it is the wrong organism, or it has degraded in storage.
2. The method failed. The inoculum, the atmosphere, the temperature, the incubation time, the medium chemistry, or the reading step was wrong.
3. The result is a statistical outlier in a system that is working. Even a perfectly controlled process produces an out-of-range result at a low but non-zero frequency, and that frequency compounds when a single panel carries a dozen drug and organism combinations.

A fourth category exists and is easy to miss. The test was performed correctly and the result was recorded incorrectly. Transcription errors, a zone entered from the wrong row, a result attached to the wrong organism, or a QC worksheet signed before the plate was read all mimic real failures.

First Response: Read the Pattern Before You Touch the Material

Work through these questions in order. Each one narrows the search space.

1. Which control failed? Write down the organism, the lot, the method, the drug or reaction, the value obtained, and the expected range.
2. Is it one control or all controls? All controls failing on one medium lot is a material or preparation problem. One control failing across several media is an organism problem.
3. Is it one method or every method? A failure confined to disk diffusion but absent from the broth MIC for the same organism suggests an inoculum, agar depth, or reading issue.
4. Is the deviation directional or random? Values consistently 2 to 4 mm low for every drug suggest a systematic error such as heavy inoculum. A single value wildly off in one direction suggests a reading or transcription error.
5. What changed since the last acceptable run? A new lot, a new technologist, a new incubator, a maintenance visit, a water bath cleaning, a change in disinfectant, a new shipment of plates.
6. When was the working culture last subcultured, and from what? If nobody can answer, that is the answer.
7. Was the material stored and handled as written? Check the temperature log, the freezer alarm history, and the opened date on the container.

Only after those questions are answered should the material be suspected. Many laboratories quarantine a lot before checking the McFarland standard, and the lot turns out to be fine.

The Failure Pattern Table

Cause 1: The Control Organism Itself

This is the most frequent root cause in laboratories that have not tightened their culture management, and it is the first thing to rule out.

Passage level

Every reference strain has a history. The original vial from a culture collection such as ATCC is the reference. Everything derived from it by subculture is a passage. Standard practice limits the number of passages from the reference culture before a fresh vial is brought up, and the widely used limit in susceptibility testing is five passages. Beyond that, the organism can drift in its expression of resistance mechanisms, lose a plasmid, or change its growth rate enough to shift zone diameters.

The practical control is a two-tier system. Keep a reference stock at or below -70 °C in a cryoprotectant, ideally in multiple aliquots so a vial is never refrozen. Keep a working stock derived from one aliquot, and date it. Retire the working stock on a fixed schedule, not when it starts to look tired.

Purity

A working culture that has picked up a contaminant will produce a result that looks like a resistance problem. A rapid-growing contaminant can overgrow the control organism in the subculture broth and change the inoculum density. A Gram stain of the working culture before every QC setup is cheap and catches this within minutes. Two morphologies where one is expected means the stock is contaminated.

Identity

Labeling errors happen. A vial labeled as one collection number can contain another, particularly when several strains are brought up on the same bench on the same day. Confirm the identity of a new working stock by a method independent of the one being controlled. A MALDI-TOF identification, a short biochemical panel, or a set of spot tests is enough. Record the confirmation with the date and the operator.

Viability

A control that has been in a freezer for years, thawed repeatedly, or stored in a frost-free freezer that cycles above -20 °C will lose viability unevenly. The organisms that survive are not a random sample. They are the subpopulation that tolerated the conditions, and that subpopulation can have different growth characteristics from the original.

Cause 2: Inoculum Density and Preparation

The 0.5 McFarland standard is the anchor for susceptibility testing and it is 1.5 × 10^8 CFU/mL for a typical organism. Three things routinely go wrong with it.

The standard itself has drifted. A barium chloride and sulfuric acid turbidity standard is stable for months when sealed and stored in the dark, but a latex particle standard can settle, and either type can be prepared incorrectly. Compare the in-house standard against a commercial sealed standard, or use a calibrated spectrophotometer to verify the suspension density directly. A turbidity standard is a reference material and deserves the same control as any other.

The suspension is made from the wrong material. Direct colony suspension should use well-isolated colonies from an overnight plate, not growth from a broth that may still be in log phase, and not colonies scraped from a selective medium that has stressed the organism. For some organisms, especially those that grow in clumps such as staphylococci, vigorous vortexing and a few minutes of standing are needed so the densitometry reads the suspension rather than the clumps.

The suspension sits. Inoculum prepared 30 minutes before use can settle, and the actual density delivered to the plate falls. Prepare suspensions immediately before use, and vortex again before swabbing.

Safety note

McFarland standards made with barium chloride and sulfuric acid are corrosive and toxic if ingested or inhaled. Prepare them in a fume hood with gloves and eye protection, label them clearly with the hazard, and store them away from acids and oxidizers.

Cause 3: Incubation Conditions

The standard incubation for disk diffusion and broth microdilution is 35 ± 2 °C in ambient air for 16 to 18 hours. Each part of that specification is a potential failure point.

Temperature. An incubator with a faulty thermostat or an overworked door gasket can hold 32 °C while the display reads 35 °C. A calibrated thermometer or an independent thermocouple placed on the shelf where plates actually sit is the only reliable check. Log it daily, including weekends, because failures do not take weekends off.

Atmosphere. Adding CO2 changes the pH of unbuffered media and shifts zone diameters for several drug classes. A CO2 incubator that is used for both CO2-requiring organisms and routine susceptibility panels is a common source of slowly drifting, unexplained failures. Keep separate incubators, and label them.

Time. A plate left in for 20 hours instead of 18 can show larger zones for some organisms and smaller zones for others. Timers on the incubator or a documented loading time solves this.

Stacking and airflow. Plates stacked too high, or placed on the bottom shelf directly over the heating element, incubate at different temperatures. Spread them, and leave space between stacks.

Cause 4: Storage, Cold Chain, and Viability

Plates stored at 2 to 8 °C in sealed bags keep their moisture for the labeled in-use period. Plates stored in the door of a refrigerator that is opened constantly, or in a fridge without a temperature alarm, will dry, crack, or accumulate condensation on the lid. Condensation dripping onto the agar surface produces a confluent lawn instead of isolated colonies, and the resulting control failure looks like an organism problem.

Freeze-thaw cycles are the second common storage failure. A working stock that is taken out of a -70 °C freezer, thawed, used, and refrozen loses viability with every cycle. Aliquot the stock so that each vial is used once.

Lyophilized organisms require the correct rehydration volume. A vial rehydrated with half the stated volume is not a more concentrated stock, it is an osmotic shock. Use a calibrated pipette, not a graduated transfer pipette.

Shipping damage. Media shipped in summer or winter can arrive frozen or overheated. Receiving should record the temperature of the shipping container, inspect the plates for condensation, cracking, or ice crystals, and quarantine the lot until growth promotion is verified in the laboratory, not just at the manufacturer.

Cause 5: Medium and Reagent Chemistry

Preparation errors

Agar that is boiled for too long loses water and concentrates the salts. Agar that is autoclaved at 121 °C for longer than 15 minutes at 15 psi darkens, drops in pH, and can inactivate labile supplements. Supplements such as blood, antibiotics, and chromogenic substrates must be added after the base has cooled to the specified temperature, and a base that is too hot will destroy them. A supplement that is added when the base is too cool will not mix and will pool at the bottom of the plate.

Weighing errors are more common than laboratories like to admit. A balance that has not been checked with certified weights, or a recipe that has been scaled up in the head rather than on paper, produces a medium that grows organisms but does not inhibit them correctly.

pH

Mueller-Hinton agar is specified at pH 7.2 to 7.4 at room temperature. A shift of a few tenths changes the ionization of several antimicrobial agents and moves zone diameters in a consistent direction. Measure the pH of the finished, cooled medium with a calibrated meter, not with pH paper, and record it.

Water quality

Deionized water with residual chlorine or a high bioburden will inhibit control organisms and grow contaminants. The water system needs its own monitoring, and a change in the building water supply can show up first as a series of unexplained QC failures in media prepared in-house.

Glassware and residue

Detergent residue on glassware inhibits growth. So does residual disinfectant on a bench, on a rack, or in a water bath that has just been shock-treated. A control that fails only on plates poured on a particular day is a hint that the glassware from that day is the problem.

Safety note

Boiling and autoclaving agar generates superheated liquid that can boil over and cause serious scalds. Loosen caps, use heat-resistant gloves and a face shield, and never move a flask straight from the autoclave to a cold bench.

Cause 6: Instrument, Reader, and Measurement Error

An optical density reader that has drifted, a disk diffusion zone reader that needs calibration, a colorimeter used for biochemical strips, or a spectrophotometer used to standardize inoculum can all produce a consistent bias that looks exactly like a reagent problem.

Zone measurement itself is a skill. The correct practice is to measure the fully inhibited zone with a ruler placed on the underside of the plate, in good light, without removing the lid, and to record to the nearest whole millimeter. Reading with the plate held up to a ceiling light, tilting the plate, or including faint hazes of growth all bias the result.

Agar depth is another measurement issue. A 100 mm round plate filled with 25 mL of agar gives a depth of roughly 4 mm, which is what the zone diameter tables assume. A plate poured thin gives larger zones, and a plate poured thick gives smaller zones, for every drug at once.

Cause 7: The Statistics of Expected Ranges

CLSI control ranges are derived from multi-laboratory studies and are set to capture the central portion of the distribution of results from a properly performing system. They are not the full biological range. By design, a small proportion of results from a correct system will fall outside the range, and the probability of at least one such result on any given day increases with the number of drug and organism combinations tested.

This has two implications for troubleshooting QC material failures in microbiology. First, a single out-of-range result in an otherwise well-controlled system should be repeated before a full investigation is launched. If the repeat is in range and no other indicator is abnormal, the laboratory documents the repeat and continues. Second, repeated single-drug outliers across several days are not statistical noise. That pattern is a real signal and deserves the full sequence.

The error to avoid is repeat-until-acceptable. Repeating a control five times and recording the one acceptable value is not QC, it is fabrication, and it defeats the entire purpose of laboratory quality management.

Cause 8: Genuinely Defective Material

This is the rarest category and the one most often blamed first. A material is truly defective when it arrived nonviable, is contaminated at the point of manufacture, carries the wrong organism, or has a physical defect such as a cracked lyophilized vial or a leaking ampoule.

Before a manufacturer complaint is justified, the laboratory needs evidence: the lot number, the expiration date, the receipt and storage records, the results of a parallel test against a different lot, and a documented subculture showing the organism's identity and purity. Manufacturers cannot investigate without that package, and laboratories that file complaints without it lose credibility with their suppliers.

A Diagnostic Sequence That Isolates One Variable at a Time

The discipline here is changing one thing at a time. A technologist who changes the inoculum, the medium, and the organism in the same repeat learns nothing.

1. Verify the organism. Gram stain the working culture, check purity on a nonselective plate, and confirm identity by an independent method.
2. Verify the inoculum. Compare the suspension against a sealed standard or a calibrated reader. Check that the suspension was made from an overnight plate and used promptly.
3. Repeat with a freshly prepared inoculum from the same working culture. If the result corrects, the problem was in the preparation.
4. Repeat with a fresh subculture from the reference stock. If the result corrects, the working culture had drifted.
5. Repeat with a different lot of medium or reagent. If the result corrects, the lot is defective or was damaged.
6. Repeat on a second method. Disk diffusion against broth microdilution, or a different identification platform. If the second method agrees with the expected result, the first method or its reading is at fault.
7. Check the environment. Incubator temperature with a calibrated device, atmosphere, timer settings, and the plate fill volume.
8. Escalate. If all of the above are correct and the result is still out of range, the material is defective and the manufacturer should be contacted with the full evidence package.

Keep the original failed result. Discarding it and starting over destroys the only evidence of what went wrong.

Scenario Playbook

One Antimicrobial Agent Fails While the Rest of the Panel Passes

The failure is confined to a single drug, so the organism, the inoculum, and the incubation are probably fine. Check the drug's stability in the panel format. Some agents degrade faster than others, and a panel near the end of its shelf life can fail on the least stable agent first. Repeat the drug by a second method. If disk diffusion gives a value in range and broth microdilution does not, the panel is suspect. If both methods fail the same way, look at whether the control organism is the right one for that drug, and whether the reading endpoint for that drug was interpreted correctly.

All Zone Diameters Are Too Small

This is the classic heavy-inoculum picture. Compare the suspension to a sealed 0.5 McFarland standard side by side, and check the reader if one is used. If the inoculum is correct, check the incubator temperature with an independent thermometer. If the temperature is correct, check the agar depth. A plate poured thick reduces every zone.

All Zone Diameters Are Too Large

This is the light-inoculum or thin-plate picture. It also appears when a plate has been incubated too long or when the inoculum was made from a broth culture that had not reached the density assumed. Check the fill volume and the incubation time first.

The Growth Control Fails to Grow

A general-purpose medium that does not support the growth of a control organism has a chemistry problem, not a selectivity problem. Consider residual disinfectant on glassware or the bench, a pH that is far off, an autoclave cycle that caramelized the sugars in the medium, or a control organism that is no longer viable. Test the same lot with a second control organism. If both fail, the medium is at fault. If one grows and one does not, the organism is at fault.

A Selective Medium Grows the Organism It Should Inhibit

Selective agents lose potency with heat and with storage. A lot that was overheated during preparation, or that sat in a warm loading dock, can grow organisms it should suppress. Test the lot with a known susceptible strain alongside the target organism, check the preparation record for temperature and time, and check the expiry. If the lot is within date and was prepared correctly, the preparation method itself may need review.

A Chromogenic Medium Gives the Wrong Colony Color

Chromogenic substrates are heat-labile. A base that was poured too hot, or a medium that was exposed to light for long periods, will give weak or absent color development. Check the preparation temperature against the manufacturer's specification, protect plates from light, and test reference strains that should give each of the expected colors.

An Identification Panel Returns the Same Wrong Answer for Several Organisms

When unrelated isolates all identify as the same organism, suspect contamination of a shared reagent. The inoculum water, the saline, the suspension tube, or the panel lot itself. Culture the water, check the panel lot, and repeat with fresh reagents. A single wrong identification is a different problem and usually points at the organism or the panel well.

Molecular and Antigen Controls Fail

Work through the control hierarchy in order. A failed extraction control means the problem is upstream of amplification. A failed internal control in a sample that also failed the target suggests inhibition. A failed positive control with a valid extraction control points at the amplification reagents or the instrument. Contamination of the amplification area with amplicon shows up as a positive result in the negative control, and it requires a decontamination and workflow review, not a reagent change.

Stain, Reagent, and Antisera Controls Fail

For a Gram stain, control organisms with known reactions are run with each new lot and each new batch of working reagents. A decolorizer that is too strong turns Gram-positive organisms pink, and one that is too weak leaves Gram-negative organisms purple. Both are timing and reagent-freshness problems, not organism problems. For antisera and latex reagents, check the storage temperature, the expiry, the rotation step, and whether the reagent was allowed to come to room temperature before use.

Corrective Action, Documentation, and Review

A QC failure is documented with the same care as a patient result. The record should contain the date, the control organism, the lot, the method, the value obtained, the expected range, the immediate action taken, the investigation, the root cause, and the resolution.

Structured feedback to staff works. In a quality improvement project in an emergency department, monthly report cards comparing individual prescribers against their peers improved adherence to facility guidelines for urinary tract infection management [1]. The same principle applies to laboratory quality management. Sharing QC failure patterns by bench, by shift, and by technologist turns a series of individual corrections into a system-level improvement.

When a failure follows a change in how a reusable item is cleaned, disinfected, or sterilized, verify the item as well as the process. A study of 3D-printed anatomical models used in-house quality control with heat map deformation analysis after steam sterilization and disinfection, and then quantified structural change with external microscopy, because the process itself can damage the material it is meant to prepare [2]. The lesson transfers directly. A new disinfectant or a longer sterilization cycle can degrade a rack, a carrier, or a reusable component, and the resulting QC failure will look like a reagent problem for weeks.

Review QC data in aggregate at least monthly. A technologist who sees only today's result misses a slow drift across a lot. A Levey-Jennings style plot of zone diameters or MIC values over time, with the mean and control limits, shows drift before it produces a failure.

Prevention: Building a System That Fails Less Often

Write the procedure down to the level of the step that can go wrong. Not "prepare Mueller-Hinton plates," but the grams per liter, the pH target, the sterilization cycle, the cooling temperature before supplement addition, the fill volume, and the storage conditions.

Qualify every lot on receipt. Test a new lot in parallel with the current lot using the full set of control organisms before it enters service. Record the result and the date.

Control the reference stock. Two tiers, dated, aliquoted, with a passage record. Retire working stocks on schedule, not on suspicion.

Calibrate and log the environment. Incubator temperatures daily with an independent device, refrigerator and freezer temperatures continuously with alarms, timers checked, balances checked with certified weights, pH meters calibrated before each use.

Train and assess competency. Direct observation of inoculation, reading, and recording, plus blind unknown samples. A technologist who has never been observed reading a zone diameter may be reading every zone correctly, or may have been taught a shortcut years ago.

Keep the failed result. Investigation depends on evidence, and evidence that has been discarded cannot be reviewed.

Trend, do not just pass or fail. The control result that is at the edge of the range every day for two weeks is a warning. The system is moving.

Review the corrective action log for repeats. The same failure in the same bench in the same month is not bad luck. It is an unresolved root cause.

Common Mistakes and Limitations

Blaming the lot first. Most failures originate in the inoculum, the working culture, or the incubator. Quarantining a lot before checking these wastes time and strains the supplier relationship.

Repeat until acceptable. Running a control repeatedly and recording only the passing result defeats microbiology quality control entirely and is a reportable quality failure.

Replacing the control organism without investigating. Swapping in a fresh vial makes the failure disappear without anyone learning why it happened, and it will happen again.

Adjusting the expected range. The published range belongs to the method and the organism. A laboratory that widens it to fit its results is measuring its own drift, not the material.

Ignoring the second morphology on the plate. A mixed working culture often announces itself with a single odd colony. It is easy to pick around it and hard to explain later.

Using a thermometer that has never been calibrated. Two uncalibrated devices that agree with each other prove nothing.

Recording QC after the fact. A worksheet completed at the end of the shift from memory does not detect anything.

Extending in-use dating. A plate that is still inside its printed expiry but has been open in a refrigerator for two weeks is not the same material.

Limitations. Some failure modes, particularly lot-specific defects and low-frequency contamination, cannot be resolved in the laboratory and require the manufacturer's investigation with full lot traceability. Judgments about whether to continue reporting patient results, whether to accept a repeat, and when to bring a new lot into service belong to the laboratory director, who weighs the failure pattern against the clinical impact of the affected test. Individual laboratory situations need that director's review, not a generic rule.

Safety Notes

Reference strains are handled at biosafety level 2 in most clinical laboratories. Work in a biological safety cabinet when there is a risk of aerosol generation, use gloves and a gown, decontaminate the bench before and after, and dispose of cultures in biohazard waste.

Iodine and decolorizer used in Gram staining are irritants and stain skin and clothing. Use gloves and eye protection and work in a ventilated area.

Many biochemical reagents contain sodium azide, which is toxic and reacts with copper and lead plumbing to form explosive metal azides. Never pour azide waste down a sink. Collect it for hazardous waste disposal.

Autoclave and boiling agar carry scald risk, and pressurized vessels require trained operators and documented cycle records.

Frequently Asked Questions

What is the first thing to check when a QC organism fails?

Check the pattern before the material. Determine whether the failure affects one organism or all organisms, one method or every method, and whether the deviation is directional or random. That answer eliminates most of the possible causes in minutes.

How many times can I repeat a failed QC before I have to investigate?

One repeat is appropriate for a single out-of-range result in an otherwise well-controlled system. If the repeat is also out of range, stop repeating and start the diagnostic sequence. Repeating until the result passes is not acceptable.

Should I stop reporting patient results when a QC fails?

Patient results from the affected run are generally held until the failure is resolved, because the control result is the evidence that the run was valid. The laboratory director decides the scope of the hold and whether previously reported results need review.

Can I use a patient isolate as a QC organism?

No. A QC organism must have a known, stable, characterized result and a documented passage history. A patient isolate has neither, and its identity and resistance profile can change.

Why does a control fail on only one drug in the panel?

A single-drug failure points at the drug itself, its dilution in that panel format, its stability at that point in the shelf life, or a reading error at that well. It rarely points at the organism.

How often should QC organisms be subcultured?

Follow a written schedule tied to passage number rather than calendar time alone, and retire the working stock after a defined number of passages from the reference culture. date every subculture so the passage history is visible on the bench.

What should I do if a new lot of plates fails QC on first use?

Do not put the lot into service. Repeat the QC once to rule out a preparation error on your side, then quarantine the lot, notify the manufacturer with the lot number and the results, and continue with the current in-date lot.

Do I need to investigate a single out-of-range result that corrects on repeat?

Document it and continue. A single outlier that corrects, in a system with verified inoculum, temperature, and medium, is expected statistical variation. Two outliers in the same week are a trend and warrant investigation.

Further Reading

• Development of edible films using chickpea protein-derived amyloid fibrils and their application in chilled fresh beef preservation.
• Using Nobel Prize case-based learning in Medical Immunology to cultivate critical thinking dispositions for medical undergraduates.
• Bacterial threats to blueberry (Vaccinium spp.) production: A synthesis of pathogen diversity, epidemiology and control strategies.
• Development of postbiotic supplemented polylactic acid films for improving quality and shelf life of rainbow trout.

Sources

1. Peer Comparison Improves Guideline Adherence for Urinary Tract Infection in a Military Emergency Department: A Quality Improvement Project.
2. Accuracy and Sterilizability of In-House Printed Patient-Specific Aortic Model for Surgeon-Modified Stent Grafts-A Workflow Description for Emergency Aortic Endovascular Procedures.