FDA's New Fresh-Cut Produce Guidance: The Layered Control System Processors Now Need to Prove
The FDA investigator walks in unannounced. You have thirty minutes before they need to be on the floor. Your food safety plan is current, your wash-water logs are filled out, and your environmental monitoring schedule is posted on the wall. Everything looks right, until the FDA investigator asks for 90 days of monitoring records and wants to see the trend.
That's where many fresh-cut and RTE produce processors are exposed: they can't show the controls are working consistently, which is what Part 117 requires and what FDA's new guidance spells out.
On August 11, 2026, FDA finalized its Guide to Minimize Biological Hazards in Ready-to-Eat Fresh-Cut Produce, a guidance document under the FSMA Preventive Controls for Human Food rule (21 CFR Part 117). It replaces FDA's 2008 fresh-cut guide, the reference this category has worked from for nearly two decades, and finalizes a draft issued in 2018. What it formalizes isn't new science; it's a documented, layered control system that many processors know they need but haven't built in a way that survives an inspection.
. This post breaks down what FDA expects under Part 117 as spelled out in the new guidance, where the compliance risk likely lives, and what changes when your documentation infrastructure catches up.
Who this guidance applies to (and who it doesn't)
One distinction matters before anything else: this guidance applies to fresh-cut processing processors, not farming.
For operations that grow and cut:
If you harvest, cool, and pack whole product and also run a fresh-cut room, you already work under two rules. What happens on your ranches and in your cooler (harvesting, field coring, vacuum or hydro-cooling, trimming, washing whole heads) is governed by the Produce Safety Rule (21 CFR Part 112). The moment product is shredded, chopped, diced, or torn, that activity is manufacturing/processing under the Preventive Controls for Human Food rule (21 CFR Part 117), and your operation is a farm mixed-type facility.
That's the operation this guidance was written for. FDA says it is intended for fresh-cut processing activities on farm mixed-type facilities, not for farms. Your field program doesn't change. What changes is the evidence FDA expects from the cut room, and from the handoff between the two.
The August 2026 guidance operates under 21 CFR Part 117, the Preventive Controls for Human Food rule, and it applies to manufacturing and processing facilities that transform whole produce into ready-to-eat, fresh-cut products. If your operation chops, dices, peels, shreds, slices, spiralizes, or tears fresh produce without a subsequent kill step, and the finished product has a water activity above 0.85, this guidance is written for you.
Covered products include chopped lettuce, sliced cantaloupe, diced onions, shredded carrots, coleslaw mixes, and fruit salads, among others.
That's the operation this guidance was written for. FDA says it is intended for fresh-cut processing activities on farm mixed-type facilities, not for farms. Your field program doesn't change. What changes is the evidence FDA expects from the cut room, and from the handoff between the two.
The four-layer control system FDA expects you to prove
Fresh-cut RTE produce has one defining characteristic that shapes the entire compliance picture: there's no kill step. Product goes from cutting room to consumer without heat treatment, acidification, or any lethal intervention. Where your hazard analysis identifies a hazard requiring a preventive control (and for RTE fresh-cut produce, biological hazards almost always do) FDA expects that control to be documented, monitored at adequate frequency, verified, and linked to corrective action procedures.
The August 2026 guidance organizes these into control measures, each aimed at a different source of contamination (Section XIII.F). Each preventive control is then subject to the same three management components under 21 CFR 117.140: monitoring, corrective actions, and verification, with associated records.
PCQIs will recognize this as the framework their food safety plan is built around.
Layer 1: Supply-chain controls
Under 21 CFR Part 117, Subpart G, when your hazard analysis identifies a hazard that's controlled before the produce reaches you, you need a supply-chain program to approve and verify the suppliers who control it. For fresh-cut produce, those suppliers are generally the farms that grow it, and pathogens on incoming produce are the hazard in question
This isn't a one-time supplier qualification at onboarding. For produce from farms covered by the Produce Safety Rule, FDA's examples use an onsite audit before first use and at least annually, by a qualified auditor, that considers the Produce Safety Rule. That can be your own audit or a third-party audit (LGMA, PrimusGFS, SQF, BRCGS, or GLOBALG.A.P.) provided the audit meets FDA's criteria for supplier verification. For farms not covered by the Produce Safety Rule, an annual written assurance can take the place of the audit. For a deeper look at what a functional supplier program looks like, see our guide to supplier quality management.
A common gap: supplier documents collected at onboarding, filed somewhere, and not looked at again until someone asks for them. FDA's examples show what ongoing verification looks like instead. Each year, the PCQI reviews the supplier's latest audit results along with its compliance history: FDA warning letters, import alerts, recalls, and state enforcement actions. A supplier network that was compliant 18 months ago may have material gaps today.
Verification also happens at the dock. In each of FDA's examples, receiving staff check every incoming load against the approved-supplier list and document that check, which is the written receiving procedure 21 CFR 117.420 requires.
Cyclospora is where the supply chain matters most. The guidance names Cyclospora cayetanensis as a potential hazard in salad mixes, including those containing lettuce and carrots, and identifies fecally contaminated agricultural water and infected field workers as reasonable routes of contamination. FDA also states that it knows of no antimicrobial suitable for fresh-cut produce that would significantly minimize Cyclospora cross-contamination in wash water. That puts the control where contamination happens: on the farm, verified through your supply-chain program. The recent outbreak linked to iceberg lettuce grown in central Mexico shows what's at stake.
FDA's own examples make this concrete. In one, a fresh-cut processor sources lettuce from a farm that uses a contract harvesting crew. The processor audits the farm annually and also reviews the harvester's records: Cyclospora training for hand harvesters, water quality certificates, and sanitation of harvest equipment. In another, cut melon arrives through a U.S. distributor that imports from foreign farms. FDA treats those farms, not the distributor, as the suppliers. The processor can use the distributor's audit documentation, but it still reviews that documentation every year. Buying through a broker or distributor doesn't take verification off your plate.
Layer 2: Antimicrobial wash-water validation
Wash water is a cross-contamination vector. When antimicrobial efficacy degrades as organic load accumulates, pathogens in one batch can spread to the next. That's the job the antimicrobial does: it limits cross-contamination. It isn't a kill step. FDA says antimicrobial wash water doesn't significantly reduce pathogens already attached to or internalized in produce, and it knows of no antimicrobial suitable for fresh-cut produce that would significantly minimize Cyclospora cross-contamination. The August 2026 guidance frames wash-water treatment as a process preventive control under 21 CFR 117.135 and includes two worked examples: sodium hypochlorite for fresh-cut leafy greens (Appendix 5) and peracetic acid for fresh-cut onions (Appendix 6).,
As a process preventive control, wash-water treatment requires defined parameters (antimicrobial concentration and, for chlorine systems, pH, validated under worst-case organic load) , monitoring at adequate frequency per 21 CFR 117.145, records documenting each monitoring activity, corrective action procedures for deviations, and verification that monitoring is occurring and effective.
The corrective action side is just as specific. When a reading falls outside its critical limit, FDA's examples segregate all product processed since the last reading that was within limits, and rewashing that product is not an acceptable fix. That's where timestamped actual readings earn their keep: without them, you can't show how far back the hold needs to go.
"Adequate frequency" is deliberate and significant. FDA doesn't define it prescriptively. In the guidance's examples, the processor sets it from its own validation data, based on how quickly antimicrobial levels fluctuated under worst-case conditions, and then demonstrates that monitoring happens at that frequency without gaps. Chlorine checked once per shift doesn't demonstrate it. In FDA's examples, a processor using manual titration checks every 30 minutes at the point in the flume where concentration runs lowest. A processor using continuous sensors verifies them against a titration kit twice per shift. Either way, the records have to show every check happened at the validated frequency and location.
Validation is where the guidance adds the most. Before relying on a wash-water control, FDA's examples validate it under worst-case conditions: the product and processing conditions that create the highest organic load, with samples taken at the point in the flume where antimicrobial concentration is consistently lowest. The processor then sets an operating limit above the critical limit, so normal fluctuation triggers an adjustment before it becomes a deviation. Validation is required for process controls under 21 CFR 117.160.
Layer 3: Environmental monitoring and sanitation controls
For RTE produce with no kill step, the processing environment is a primary hazard source. Listeria monocytogenes and Salmonella can establish harborage points in cutting, washing, and packaging areas and cross-contaminate product before it ever leaves the facility.
The guidance positions environmental monitoring as the primary verification tool for sanitation controls in RTE fresh-cut facilities. Your environmental monitoring program (EMP) should identify sanitation control effectiveness, detect harborage sites before they become contamination events, and generate corrective action triggers when results indicate problems.
The guidance doesn't lay out EMP design itself. For that, it points readers to FDA's guidance on controlling Listeria monocytogenes in ready-to-eat foods, which is still a draft, issued in January 2017.
The discipline that matters most is trending. In Zones 2 through 4, a single positive is a signal; a pattern of positives in the same zone over multiple sampling periods is a harborage problem. Zone 1 is different. A positive on a food-contact surface in a plant with no kill step calls for immediate corrective action and can put product on hold by itself.
Paper logs and emailed lab reports make that pattern hard to see. When trending means manually compiling results across weeks, it tends to happen late, or not at all.
Under 21 CFR 117.165(a)(3), 117.150(a)(1), and 117.190, processors must document not just that sampling occurred, but that results were reviewed, positives triggered defined corrective actions, and those actions were carried through. Paper can meet that requirement. It just makes the loop slow to close and hard to prove later.
For operations that cool and pack whole product next door to the cut room, the guidance gets specific about that boundary. FDA recommends keeping floor flumes in the cooling and packing operation separate from fresh-cut flumes where practical, since floor flumes can spread aerosols into fresh-cut room air and onto nearby equipment. It also recommends restricting movement of lift trucks, bins, totes, tools, and people from receiving and storage into processing and packaging, giving special attention to containers that can't be sanitized, such as wooden bins, and separating intact produce from finished product through linear product flow with reverse air flow, or a physical partition, with positive air pressure in the fresh-cut area.
Layer 4: Time-temperature controls
For fresh-cut RTE produce, temperature control is a growth-inhibition control, not a kill step. FDA recommends that fresh-cut produce be held under refrigeration. The FDA Food Code, the model code for retail, sets 41°F (5°C) or below for foods that need time/temperature control for safety, a category that includes cut leafy greens, cut melons, and cut tomatoes. Confirm the parameter validated for your own product and process, as some operations have validated parameters that differ. Refrigeration slows pathogen growth; it doesn't stop it. FDA notes that Listeria monocytogenes can grow slowly at refrigeration temperatures, which is why time matters as much as temperature.
Dock staging delays, fluctuating cold-room setpoints, and temperature excursions during loading all expose product to temperature abuse, and pathogen growth adds up with every hour it lasts. Whether time/temperature is a preventive control in your plan depends on your hazard analysis; the guidance leaves that determination to you. Either way, the same trending logic applies: a processor who can show cold-room temperatures stayed within range consistently, with documented corrective actions on the occasions they drifted, has a defensible control.A clipboard log that records'OK' or '<41°F' instead of the actual reading, or shows a drift with no corrective action beside it, does not Statistical process control is the methodology that makes this kind of trend visible before it becomes a deviation.
CGMP isn't enough for high-risk RTE operations
This is where processors most frequently underestimate their exposure.
21 CFR Part 117 contains two distinct categories of requirements. CGMP requirements (Subparts A and B, including 21 CFR 117.35, 117.37, 117.40, and 117.80) establish baseline facility, equipment, and hygiene standards. Preventive controls requirements (Subpart C, 21 CFR 117.126 through 117.190) establish a risk-based, documented control system with monitoring, verification, and corrective action obligations.
Many processors have well-documented CGMP programs and assume that satisfies FDA's expectations for high-risk RTE operations. It doesn't. The guidance shows where the line moves. A CGMP measure becomes a preventive control when you apply the management components to it: monitoring, corrective actions, and verification. FDA's example is handwashing. It's a routine CGMP in most plants, but a facility with a history of hygiene lapses may need to manage it as a sanitation preventive control.
When a hazard analysis under 21 CFR 117.130 identifies a hazard requiring a preventive control, and for RTE fresh-cut produce, biological hazards almost always qualify, the associated controls escalate to full Subpart C requirements. That means documented monitoring procedures and records, documented corrective action procedures and records when deviations occur, documented verification activities confirming controls are working, and records in a format that allows FDA review.
When FDA investigators examine an RTE fresh-cut facility, they're asking whether your procedures executed consistently, whether deviations were caught and corrected, and whether the controls were verified as effective. The enforcement consequence when the answer is unsatisfying: 483 observations and, in cases with repeated documentation failures, warning letters.FDA publishes warning letters on its website, and 483s, while not routinely posted, can be obtained through FOIA requests. Both affect your customer relationships and retailer standing.
Where paper-based systems create blind spots
Every quality manager reading this knows how manual compliance tracking works in practice. A technician records a check on paper. The form goes to a supervisor at end of shift. The stack goes into a binder. The binder goes into a filing cabinet.
When an FDA investigator asks for 90 days of EM records for Zone 3 on Line 2, someone goes through binders. When they ask whether positive results triggered documented corrective actions, someone searches email threads. That process takes hours. Part 117 requires records to be made promptly available on request, and every hour spent searching is an hour the investigator spends wondering what else is hard to find.
More consequentially, paper-based systems create blind spots that prevent proactive response. A trending positive EM rate, two hits in six weeks, then three in the next six, then four, doesn't surface until someone manually compiles the data. By the time the pattern is visible, product may already have been compromised.
Under 21 CFR 117.145(b), you must 'monitor the preventive controls with adequate frequency to provide assurance that they are consistently performed.' Assurance requires evidence. Evidence requires records that capture actual readings at the frequency you've set, and make them easy to find.
FDA is specific about what counts. Records must contain the actual values and observations from monitoring, recorded at the time the activity is performed (21 CFR 117.305). The guidance's example: record "22 ppm" free chlorine, not "> 20 ppm," "OK," or "Pass." A check mark tells an investigator someone looked. An actual reading tells them what the control was doing.
Not sure how your current documentation would hold up? Use FDA's finalized guidance document to walk your four control layers against the requirements before your next inspection.
What changes when your documentation infrastructure catches up
Here's what that looks like in practice for fresh-cut operations:
Supplier documentation stops living in a spreadsheet. Expiring certifications surface automatically before they lapse, instead of being discovered when someone asks about a grower you qualified 18 months ago. You can see the status of your entire supplier network without pulling files. For operations managing dozens of grower relationships, especially across a seasonal transition, it's the difference between knowing where your supplier program stands and hoping it's current.
In-process parameters generate a trend. When wash-water concentration or cold-room temperature starts to drift, SPC alerts flag it before it becomes a deviation. When a reading does fall out of spec, the supervisor can be notified immediately, so the corrective action can be recorded against that reading while the run is still going. Understanding statistical process control is the underlying methodology here; the practical result is that your monitoring data shows a story, not just a list of check marks.
Environmental monitoring closes the loop. Positive EM and the corrective actions they trigger live in the same system, with root cause, corrective measures, and follow-up documented together.. That's the closed-loop documentation trail 21 CFR 117.150(a)(1), 117.165(a)(3), and 117.190 call for. Historical EM data is searchable and reportable, so trend analysis that would otherwise take hours of manual compilation becomes a quick report you run, not a project.
Audit prep stops being an event. When an FDA investigator arrives unannounced, your documentation is already there, organized, searchable, and accessible in minutes. Fresh-cut operations have the same time pressure as other food and beverage manufacturers, but with less margin: no kill step means every hour of undetected deviation matters.
If you want to see how your current audit readiness stacks up before the next inspection, our audit readiness checklist will help you.
The next step depends on where your gaps are
FDA's finalized fresh-cut produce guidance doesn't change what controls you need to run. Processors with sound food safety programs have been running supply-chain controls, wash-water monitoring, environmental monitoring, and temperature controls for years.
What it spells out is the evidence standard Part 117 already sets: documented, monitored, verified, and linked to corrective action. If your current system requires manual record retrieval, spreadsheet compilation, or memory of what happened to a form filled out six weeks ago, the gap between what you're doing and what you can demonstrate is a compliance risk.
Check your supplier verification records.. Pull your last 90 days of EM data. See whether your corrective action records close every loop. If any of those feels uncertain, that's the conversation worth having before an investigator asks the same question.
Talk to a SafetyChain specialist about how digital food safety and quality programs apply to your fresh-cut operation.
