Depyrogenation oven qualification and the FD value

Killing an organism and destroying what it left behind are different problems. Bacterial endotoxin — lipopolysaccharide from the outer membrane of Gram-negative bacteria — survives moist heat sterilisation comfortably. A vial can come out of an autoclave sterile and still be pyrogenic, which is why glassware destined for parenteral products goes through a dry heat depyrogenation step as well.

The same model, different constants

The arithmetic mirrors F0 exactly. Lethality per interval is 10 raised to (T − T_ref)/z, summed across the cycle. What changes is the reference temperature and the z-value, because the mechanism being modelled is the thermal destruction of a large molecule rather than the killing of a spore.

The much larger z-value has a practical consequence worth noticing: depyrogenation lethality is far less sensitive to small temperature errors than sterilisation lethality is. A 1 °C error changes FD by about 5 %, against 26 % for F0. That is a genuine relief, and it is more than offset by the difficulty of measuring 250 °C accurately in the first place.

The classic cycle and where it comes from

250 °C for 30 minutes is the cycle everyone quotes, and it is not arbitrary: at the reference temperature each minute contributes exactly one unit of lethality, so 30 minutes at 250 °C delivers FD = 30, which is the figure associated with the required 1000-fold endotoxin reduction. Cycles at other temperatures are designed to accumulate the same total.

As with F0, the accumulation happens throughout the cycle, not only at the hold. A tunnel or an oven with long ramps contributes meaningfully on the way up and down, and a cycle designed around the hold alone is being scored conservatively.

Endotoxin challenge is the evidence

Thermal data says what the cycle delivered against a model. The acceptance evidence that the model is right is an endotoxin challenge: vials spiked with a known quantity of endotoxin — typically enough to demonstrate the 3-log reduction with margin — placed at the worst-case positions, run through the cycle, and then recovered and assayed.

• Challenge vials go where the thermal study said the cold spots are, not where they are convenient to place.
• A positive control that did not go through the cycle demonstrates that the assay would have found the endotoxin had it survived.
• Recovery efficiency matters: endotoxin that cannot be washed back out of the vial looks like endotoxin that was destroyed.
• The assay itself — LAL or a recombinant equivalent — has its own validation and interference considerations.

Tunnels are not ovens

A batch oven and a continuous depyrogenation tunnel pose different qualification problems. The oven has a load pattern and a cycle; the tunnel has a belt speed, a temperature profile along its length, and items that experience the profile as they travel. The lethality calculation integrates along the path rather than over time at a point, which means belt speed is a critical parameter in exactly the way hold time is for an oven — and a tunnel qualified at one belt speed is not qualified at another.

Tunnels also sit between rooms of different classification, so the airflow balance across them is part of the qualification for contamination control reasons that have nothing to do with heat.

Measuring at 250 °C

Thermocouples used at depyrogenation temperatures live a harder life than sterilisation probes. Insulation degrades, junctions drift, and repeated thermal cycling to 250 °C and above ages a thermocouple noticeably. Pre- and post-study calibration is not a formality here — it is the check that catches a probe which shifted during the campaign, and shifts are common enough that finding none is the pleasant outcome rather than the expected one.