Two Food-Contact Listeria Positives in an RTE Hummus Plant

Do not release it, and do not release it on the strength of a finished-product test you run now.

Reasoning chain:

• Lm is an environmental contaminant, and it is in the plant on two food-contact surfaces that touch product during the production window. That gives the finished food a reasonable probability of contamination.
• Hummus is a growth-supporting RTE food (pH around 6.0, water activity around 0.97). For RTE foods that support Lm growth, the FDA position applied in practice is zero tolerance, so any Lm present is an adulterant.
• Finished-product testing cannot rescue the lot: Lm is distributed unevenly and at low level, so a handful of 25 g enrichments has poor sensitivity, and FDA does not treat product testing as a substitute for environmental control (see 21 CFR 117.165(a)(5)).

Actions:

1. Hold and segregate all product back to the last documented negative food-contact result - the 1,800 lb plus anything produced after Tuesday - with hold tags and a log.
2. Escalate to a documented hazard review with QA and plant management to dispose (destroy) or formally evaluate the lot.
3. Check what shipped from the same production window and start a recall evaluation, because the contaminating event clearly precedes Tuesday's sampling.
4. Release only after the root cause is corrected and verification sampling is negative.

Rank by proximity and by whether the vector can deliver Lm to product:

1. The refitted blade guard - highest priority. A guard removed for a bearing change is a classic maintenance-induced vector: hidden surfaces, lubricant residue, tools, and re-assembly without cleaning and sanitizing put Lm directly on a food-contact part. The timing matches the positives exactly.
2. Condensate drip tray 4 m away - condensate and drip pans are known Lm niches and can aerosolize or drip onto nearby equipment. Verify the tray is not above or draining toward product surfaces.
3. Floor drain with six positives - a reservoir feeding foot traffic, hoses and wheels; a transfer pathway, not usually a direct food-contact source.
4. The cleaning change - once-weekly wet foam instead of nightly increases soil and moisture residence time on equipment and in niches.

Sequence: swab to find the extent (vector swabbing of the guard, mounting bolts, gaskets, welds, the dripline under the tray) before cleaning, so you do not destroy the evidence; then disassemble, clean, and sanitize; then re-swab after cleaning. Fix niches (replace the drip tray design, seal or relocate, repair the drain) in the same window - intensifying sanitation on unrepaired niches just re-seeds the line.

Restart criteria (write them down and sign off):

• Root causes closed with objective evidence: guard cleaned/sanitized and its sanitary design reviewed, drip tray eliminated or redirected, drain repaired and its cleaning procedure changed, tools and maintenance procedure updated with a clean-before-reassembly step.
• Cleaning effectiveness verified: post-cleaning food-contact swabs at the repaired sites negative for Lm (and for Listeria spp. as an indicator), plus visual and ATP checks.
• At least three consecutive rounds of intensified food-contact and vector swabbing, taken across different production days with production running, all negative for Lm and with no upward trend in Listeria spp.
• Intensified sampling frequency (e.g., weekly food-contact sites for 8–12 weeks) documented in the EMP before returning to baseline.
• Corrective action record closed under the preventive controls written program (21 CFR 117.150) with the reason for the deviation.

Why one round fails: a negative sponge only shows the sampled square centimetre was clean that morning. It carries no information about whether contamination was re-introduced the next day, and one round gives no trend. Repeated sampling across days with production running is what distinguishes a fixed process from a cleaned surface, and it is the basis on which the hold can be lifted with a defensible record.