When the Indicator Strip Says Pass, Ask the Culture Dish
A Routine Batch With an Unruly Result
The first sign was not dramatic. A batch of prepared media, golden and clear coming out of the autoclave, began to cloud after a few days at room temperature. The indicator strip from the run had turned the expected color. The cycle log showed a completed program. By every visible signal, the load had passed.
The cultures disagreed.
Contamination in a single batch is not, by itself, a lab crisis. Media prep is a high-volume operation, and occasional losses happen. What made this different was the mismatch: the sterilization record said success, while the biological outcome said otherwise. At McKnight Lab, where reproducibility is treated as a daily practice rather than a retrospective virtue, that mismatch demanded a closer look. If a green light could be wrong, then every protocol built on that green light needed examination.
What an Indicator Strip Can and Cannot Prove
Chemical indicator strips are useful, but they are often asked to carry more meaning than they can bear. Many strips respond to heat, steam, or time within a specified range. They offer evidence that a load was exposed to sterilizing conditions. They do not, on their own, prove that every surface of every vessel reached the required temperature and pressure for the required duration.
That distinction matters most in loads that are not uniform. A tightly packed run of large flasks, sealed bottles, and dense agar volumes can create cold spots or impede steam penetration. The indicator strip may sit in a representative location and respond appropriately while an interior volume lags behind. For years, our media-prep workflow had treated the strip as a summary judgment. The audit forced us to ask a more precise question: a summary of what, exactly?
Reconstructing a Failed Load
The investigation began with the mundane work of reconstruction. We pulled autoclave cycle records, maintenance logs, and media-prep worksheets. We reviewed how bottles had been arranged, how much liquid each contained, and where indicators had been placed. Then we ran controlled load-pattern tests using biological indicators, including spore preparations placed in the positions most likely to resist steam penetration.
The results were not catastrophic, which made them more instructive. The autoclave was not failing outright. Instead, the audit identified a calibration drift in a temperature sensor that had gone unquestioned because the displayed values remained plausible. The drift was small enough to escape casual notice and large enough to matter at the margins of a dense load. Combined with a load configuration that had gradually changed as our media volumes grew, it produced a condition in which the chemical indicator could pass while some vessels did not receive the sterilization margin we had assumed.
No single person had done anything reckless. The failure was systemic: a trusted instrument, a familiar workflow, and a green light that had quietly become a substitute for verification.
The Human Factors of a Green Light
Labs are good at writing protocols for actions and less good at writing protocols for doubt. A green light is a cognitive shortcut. It reduces the need to re-examine a routine decision, and that is precisely why it is dangerous when its meaning is overstated.
During the audit, several lab members described the same internal calculation: the strip passed, the log looked normal, and the media were needed for Monday. None of those observations was unreasonable. Together, however, they formed a chain of assumptions rather than a chain of evidence. The sterilization audit did not reveal a lab that ignored procedures. It revealed a lab that had confused a convenient signal with a sufficient one.
That is a reproducibility problem as much as an equipment problem. Published methods often describe sterilization in a single sentence, as if autoclave performance were a fixed background condition. In practice, equipment drifts, loads change, and personnel turn over. A method is only as reproducible as the infrastructure assumptions beneath it.
Rewriting the SOP Around Verification
The audit produced a revised media-prep standard operating procedure with several deliberate redundancies. Loads are now mapped by configuration, not just by total volume. Biological indicators are used on a defined schedule and after significant changes in load pattern. Chemical indicators remain in use, but they are no longer treated as standalone proof of sterility. Media are quarantined until release criteria are met, and cycle records are reviewed against calibration status before a batch enters general use.
The changes add time. They also remove a category of silent failure. Perhaps most importantly, the new SOP includes a short section titled 'What This Test Does Not Tell You,' which is now part of onboarding for every lab member who touches media prep. The section is not an accusation of carelessness. It is an acknowledgment that trust in equipment should be calibrated as carefully as the equipment itself.
Trust Is a Protocol, Not a Feeling
Open science depends on more than shared data and published methods. It depends on the honesty of the verification behind those methods. When we report that cells were grown in sterilized media, a reader should be able to assume that sterilization meant something defensible, not merely that an indicator changed color.
The autoclave is working correctly again. The sensor was recalibrated, the load patterns were adjusted, and the revised SOP has held through repeated audits. But the lesson we keep is less about one machine than about the culture of evidence around it. A green light is an invitation to ask a question, not a substitute for having asked it. In a lab, the most dangerous instrument is sometimes the one everyone has stopped questioning.
