DRY PLANT EMP

Environmental Monitoring in Dry Food Plants: Salmonella, Cronobacter and Enterobacteriaceae

By Mthokozisi Nkosi, Food Safety Specialist & Lead Auditor, ASC Food Safety · 23 min read

A dry food plant needs an environmental monitoring programme built around Salmonella, with Cronobacter spp. added wherever infants eat the product and Enterobacteriaceae trended as the hygiene indicator. These pathogens cannot grow in a dry product, but they survive for months and even years in dust, residues and the product itself, so the programme must hunt for water, harbourage and residue. ASC designs and verifies these programmes, and our R1 450 (about USD 89 · EUR 77 · GBP 66 · AED 326) Automated EMP tool runs the schedule, results log and trend charts.

Why does a dry food plant need a different environmental monitoring programme?

Because the hazard behaves differently. In a wet ready to eat plant the worry is Listeria monocytogenes growing in drains and on wet surfaces. In a dry plant the worry is Salmonella that cannot grow in the product but survives for long periods in dust, residues and the food itself, and becomes more heat tolerant as water activity falls.

The Codex Code of Hygienic Practice for Low Moisture Foods (CXC 75-2015) covers foods with a water activity (a_w) of 0.85 or below, including cereal products, nut butters, dry protein products, chocolate, snacks, nuts, seeds and spices. Powdered formula has its own code, CXC 66-2008.

The FAO/WHO expert body JEMRA ranked cereals and grains, dried protein products, spices and dried herbs, and nuts and nut products as the four highest of seven low moisture food categories in its 2022 report, Microbiological Risk Assessment Series 26, stressing that bacteria persist in these foods even though they cannot grow.

A review by Beuchat and colleagues in the Journal of Food Protection (2013) states that some pathogens survive for months, even years, in low a_w foods and dry processing environments, often with increased heat tolerance. The authors conclude that preventing contamination through hygienic design, zoning and sanitation is more practical than removing it later. Finn and colleagues in Frontiers in Microbiology (2013) add that desiccation tolerance gives Salmonella cross protection against some sanitising agents and dry heat.

When the US FDA investigated the 2022 Salmonella Senftenberg outbreak linked to peanut butter from a Lexington, Kentucky plant, FDA’s analysis showed that an environmental sample collected at the facility in 2010 matched the outbreak strain. An old environmental positive may be a resident strain, not history.

Which organisms should a dry plant target: Salmonella, Cronobacter or Enterobacteriaceae?

Target Salmonella as the minimum in any dry plant where product can be contaminated from the environment. Add Cronobacter spp. where the consumers are infants, as in powdered formula and infant cereals. Count Enterobacteriaceae as a hygiene indicator alongside, never instead of, the pathogen tests. The risk assessment then decides whether anything else belongs.

Salmonella: the minimum

Annex II of CXC 75-2015 says that where end products may be contaminated from the environment, monitoring should as a minimum target Salmonella. The copy hosted by South Africa’s NICD includes the annexes.

Cronobacter spp.: when infants are the consumers

CXC 66-2008 names Salmonella enterica and Enterobacter sakazakii (Cronobacter spp.) as the organisms of greatest concern in powdered formula. It places neonates at greatest risk, particularly pre term, low birthweight and immunocompromised infants, along with infants under two months of age. Its Annex III covers monitoring for both, plus other Enterobacteriaceae, in the dry high hygiene areas.

A 2023 review in Microorganisms reports that Cronobacter survives up to two years desiccated, multiplies rapidly once rehydrated, and survives in foods of a_w 0.25 to 0.69 such as formula and infant rice cereals. If you blend infant cereals or pack milk powder for infant products, it belongs in your programme.

In the EU, Regulation (EC) No 2073/2005 Article 5(2) requires manufacturers of dried infant formulae and dried foods for special medical purposes for infants under six months, where these pose a Cronobacter risk, to monitor processing areas and equipment for Enterobacteriaceae. The US FDA’s March 2023 letter to powdered infant formula manufacturers encourages direct testing for Cronobacter spp. at some frequency within the processing environment.

Enterobacteriaceae: useful, but only as an indicator

CXC 75-2015 says high Enterobacteriaceae levels signal conditions that may support Salmonella, but testing for them alone is not sufficient, because low levels do not guarantee its absence. FDA’s January 2025 draft guidance on sanitation programmes for low moisture ready to eat foods notes that surface Enterobacteriaceae are not a reliable indicator of Cronobacter spp. and that data are insufficient to correlate them with Salmonella. Use the count for hygiene and the pathogen test for the pathogen.

Formula plants should expect CXC 66-2008 to change. A JEMRA expert meeting in Rome on 15 to 19 June 2026 was held to inform its update, and Food Safety Magazine reported that the experts called for enhanced environmental monitoring with pathogen specific testing.

Why is water the biggest threat in a dry food plant?

Because water is what lets a surviving pathogen multiply and spread. Dry residues keep Salmonella and Cronobacter alive but static. A roof leak, condensation or a wet clean can turn a few surviving cells into a growing population that equipment and people then carry through the plant. Controlling water is controlling the hazard.

“Uncontrolled moisture (e.g. leaking roofs, leaking pipes, condensate, improper cleaning) is a major contributor to the presence of pathogens in low-moisture foods.”
Codex Alimentarius, CXC 75-2015, section 5.1

CXC 66-2008 says growth needs water, so the environment must be kept as dry as possible. It links wet cleaning to the spread of Salmonella and particularly Cronobacter, and calls for limited water systems, dry drains and no condensation in high hygiene areas.

In its 2023 letter, FDA reported water during production in areas meant to be dry, from leaks and condensation, and several firms with poor or no records showing that dry out after clean in place fully dried equipment surfaces. At a Sturgis, Michigan formula plant in 2022, FDA reported that five environmental subsamples were positive for Cronobacter sakazakii, five different strains. None matched the patients and FDA’s product samples were negative, but the organism was clearly in the plant.

In a US Department of Justice account of an earlier peanut butter outbreak with more than 700 cases, company officials hypothesised that moisture entered the process, naming a leaky roof among the factors.

Build a water register

Codex says a leak should be fixed and the area cleaned, disinfected and completely dried, and that monitoring should increase after maintenance, construction or wet cleaning. Link the two so every wet event generates samples.

• After heavy rain, log every drip, stain and puddle against EMP site codes.
• Map condensation risks: cold pipes, roof sheets over dryers, ducts and doors to wet areas.
• Record each wet clean and a dryness check before restart.
• Make each water event trigger Salmonella and Enterobacteriaceae samples around the spot.

What is a primary Salmonella control area, and how should a dry plant be zoned?

The primary Salmonella control area (PSCA) is the part of the plant where product or ingredients need the highest level of hygiene control, usually everything after the kill step. Its boundary, and the controls on people, equipment, air and water that cross it, decide where your environmental monitoring effort goes.

The term comes from Control of Salmonella in Low Moisture Foods, Grocery Manufacturers Association guidance from 2009 (see also Chen and colleagues in Food Protection Trends). Its seven control elements cover preventing ingress, stricter hygiene in the PSCA, hygienic design, preventing growth, raw material control, validating the kill step, and verification. Without a kill step, the whole process area may become the PSCA.

CXC 75-2015 section 5.2.4 asks for the strictest hygiene after a pathogen reduction step: controlled traffic, dedicated workers and tools, and air flowing from the strictest areas outwards. CXC 66-2008 calls these high hygiene areas, from drying to filling, with positive air pressure and personnel locks.

1. Find the kill stepRoasting, pasteurising, spray drying, extrusion or baking. Exposed product downstream sits in the PSCA.
2. Draw the boundary on the floor planEvery door, hatch, conveyor or duct crossing it is a route in and a candidate sample site.
3. Control what crossesFootwear, clothing, dedicated vacuums and tools, and maintenance toolboxes.
4. Check the airLog pressure differences and filters.
5. Weight the sampling planMost pathogen samples inside and at the edge of the PSCA, some outside to prove the barrier.

How should dry cleaning and controlled wet cleaning shape environmental monitoring?

Dry cleaning should be the default in the PSCA, controlled wet cleaning the exception, and the monitoring plan should be timed around both. Sample for Enterobacteriaceae before start up to check that cleaning worked, sample for Salmonella several hours into production to find harbourage, and add samples after every wet clean.

CXC 75-2015 defines dry cleaning as removing residues without water or detergents, and section 6.1.2 makes it the routine practice in the area that needs the most stringent hygiene control. It recommends vacuum cleaners dedicated to areas where possible, so their contents can be tested, and cleaned and disinfected in a designated area. Controlled wet cleaning uses a limited amount of water, controls where it spreads and ends with complete drying.

FDA’s 2025 draft agrees: dry clean where practical, limit wet cleaning, remove water fast, use a separate wet clean area, and build in sanitation breaks, such as one at the end of daily production. FDA also says it is not aware of any information showing that dry cleaning techniques, such as material flushes, can remediate a pathogen contamination event unless a sanitising treatment goes with them. The sanitisers used in these plants are often low moisture products, such as alcohol based ones. So write two modes: routine dry cleaning, and a remediation procedure with sanitising and drying steps.

The timing comes from CXC 75-2015 Annex II: Salmonella samples after several hours of production detect organisms shed from harbourage, and Enterobacteriaceae samples just before start up show whether cleaning worked.

What should you sample in a dry plant: dust, vacuum contents, scrapings, filters or sponges?

Sample the material where Salmonella actually accumulates. In a dry plant that means residue and dust as much as surfaces: scrapings, dedicated vacuum cleaner contents, sifter tailings and dust collector fines, plus large area sponges that are only slightly moistened. Swabs of clean stainless steel alone miss most of it.

CXC 75-2015 section 6.5 calls sampling of dust and product residue a critical verification activity, and its Annex II names scrapings and vacuum residues as useful samples. CXC 66-2008 Annex III agrees, and separates environmental samples (equipment exteriors, floors, platforms) from line samples taken inside equipment after the dryer.

Dry or pre moistened sponges?

CXC 75-2015 Annex II says that in areas needing stringent hygiene control, wipes and sponges should be slightly moistened, not wet or dripping, and the area left completely dry afterwards. Use a small, measured volume of neutralising broth matched to that area’s sanitiser, then dry the spot. A dripping sponge is a wet event of your own making.

How much material?

No law fixes a residue sample size, so agree the test portion with your laboratory. EU criteria for dried infant formula test 25 g for Salmonella and 10 g for Cronobacter; collecting at least that much residue allows a comparable test. Codex allows pooling if sensitivity is not reduced, but a positive pool needs follow up to find the site. Follow a recognised sampling method such as ISO 18593:2018. Ask the laboratory for ISO 6579-1 for Salmonella, ISO 22964 for Cronobacter spp. and ISO 21528-2 for Enterobacteriaceae counts, all of which cover environmental samples.

Ingredient systems belong on the map. In the 2022 chocolate outbreak, WHO reported 151 cases in 11 countries, 89% in children under ten, with matching strains found in buttermilk tanks at the Belgian plant.

How should a dry plant set zones, sites and sampling frequencies?

Use the four zone model, define each zone around residue and the PSCA boundary, and let a scored risk assessment set each site’s frequency. Codex fixes no sample number; frequency should reflect product, area and volume, and follow the findings. General zone principles are in our EMP design guide.

In ASC’s RA19 structure each site is scored on a five by five matrix and the residual rating sets frequency: Critical daily, High weekly, Medium monthly, Low to Medium quarterly, Low twice a year. A four week matrix rotates sites across shifts and days. After a wet clean, leak or building work, affected sites step up until results settle.

CXC 66-2008 expects line samples in formula plants. Before testing any food contact surface for a pathogen, write down which product you will hold and how you will release it.

How should a dry plant trend Enterobacteriaceae results?

Trend Enterobacteriaceae as counts, by site and by zone, against alert and action limits you set from your own baseline. Expect to find them, because Codex says they are part of the normal flora in dry processing areas. What matters is a rising count, a site that keeps coming back, or a jump after a wet event.

CXC 75-2015 Annex II tells manufacturers to expect Enterobacteriaceae and plan for exceedances. A plant reporting “not detected” everywhere probably has the wrong method or sites, and FSSC 22000 clause 2.5.7 (d) makes a long run without positive results a review trigger.

Link the environment to the product

The CXC 66-2008 product criterion for Enterobacteriaceae is n=10, c=2, m=0/10 g: one or two positives signal a trend towards loss of control, three or more signal loss of control. The EU criterion is n=10, c=0, not detected in 10 g, with parallel Cronobacter testing unless a plant level correlation exists. When product results move, check the environmental trend for that line.

1. Set a baselineUse a few months of counts per site before fixing alert and action limits.
2. Chart by zone each monthSamples above alert and Salmonella positive rate per zone.
3. List repeat sitesRepeat exceedances get an investigation as well as a re clean.
4. Overlay eventsMark wet cleans, leaks and building work on the chart.
5. Review and recordClause 2.5.7 (c) expects regular trend analysis. Minute it.

What does a positive result mean in a dry plant, and what happens to the product?

It depends on the organism, the zone and the history, which is why Codex says the decision criteria must be written before the programme starts. An Enterobacteriaceae exceedance in zone 3 calls for cleaning and a look at the trend. A Salmonella or Cronobacter result on a food contact surface is a contamination event that puts product on hold.

CXC 75-2015 lists responses from no action, through intensified cleaning, source tracing and holding and testing product, up to product disposition, with strain subtyping where contamination persists. FDA’s 2025 draft adds whole genome sequencing, sampling before and after cleaning, equipment disassembly, and restoring surfaces to a clean, dry, sanitary condition. The draft also states that finished product testing alone cannot ensure the absence of a hazard in the food.

Why “test a few packs” does not work

Contamination in dry product is low and uneven, which is why the Codex formula criteria need 60 samples of 25 g for Salmonella and 30 of 10 g for Cronobacter, with no positives. A few negative packs prove little. A doubtful lot needs a statistically meaningful release plan, diversion to a validated kill step, or destruction. Decide which in advance.

How do pests, raw materials and toxins fit into a dry plant programme?

Pests and raw materials are the two main routes by which Salmonella reaches a dry plant, so the monitoring plan should sample where they meet the building. Toxins such as aflatoxin are a separate hazard, managed through supplier control and product testing rather than environmental swabs.

The 2009 GMA guidance states that birds, rodents and insects carry and spread Salmonella into facilities. Finn and colleagues cite a Japanese oil meal plant where 46% of 41 captured rodents carried Salmonella, mainly those caught in heavily contaminated areas. Stored product insects breed in residue and move dust between stock and clean areas, so treat insect activity in the PSCA as a hygiene failure and sample around it. R638 regulation 10(3) already requires effective measures against insects, rodents and birds.

• Add zone 4 samples at bait station lines, dock doors and the raw material store.
• Sample residue around insect traps showing activity near the PSCA.
• Treat raw nuts, grain, cocoa beans and spices as possible carriers of Salmonella until a validated kill step has been applied, and keep their dust out of the PSCA.

Keep toxins separate. Aflatoxins from Aspergillus flavus and Aspergillus parasiticus are controlled through suppliers, storage and lot testing; see our article on the revised Codex peanut aflatoxin code.

What does this mean for South African dry food manufacturers?

South African law sets the hygiene baseline through R638, but it does not prescribe a pathogen environmental monitoring programme. The detailed expectation comes from certification schemes, especially FSSC 22000 clause 2.5.7, and from customers and export markets that apply Codex, EU or US requirements.

R638 of 22 June 2018, under the Foodstuffs, Cosmetics and Disinfectants Act 54 of 1972 and enforced by municipal environmental health practitioners, requires a certificate of acceptability (regulation 3(1)), pest control (regulation 10(3)) and processing records (regulation 10(16)). Its limit of 100 viable micro-organisms per cm² on food contact surfaces, in regulation 6(4)(b), is a hygiene count, not a Salmonella programme. Codex codes are not law here on their own, but auditors, customers and importing countries apply them.

FSSC 22000 Version 7 applies clause 2.5.7 to food chain categories BIII, C, I and K. Many dry food plants sit in category CIV, processing of ambient stable products, and some grain and nut handlers in BIII, pre process handling of plant products. The clause requires a risk based programme, a documented procedure, trend analysis, and review at least annually and on triggers such as repeat pathogen detection. Category D (processing of feed and animal food, which takes in pet food) is not listed, though customers often expect a programme anyway. BRCGS Food Safety Issue 9 clause 4.11.8 and SQF Edition 9 clause 2.4.8 carry similar expectations; SQFI expects Edition 10 audits to start on 2 January 2027. See our FSSC 22000 clause 2.5.7 guide.

Exporters face their market’s rules: Regulation 2073/2005 for infant products into the EU, and for US customers 21 CFR 117.165(a)(3), which calls for environmental monitoring where environmental contamination of a ready to eat food needs a preventive control. Pet food is scrutinised too: in 2012 the CDC reported 49 human Salmonella Infantis infections linked to dry dog food from a single plant.

What should a dry plant manager change on Monday?

Start with water and residue, not the swab count. Map water sources, draw the PSCA, move pathogen sampling onto dust, scrapings and vacuum contents, write the product hold rule, and trend Enterobacteriaceae. These changes close the most common gaps.

1. Walk the plant for waterStart the water register this week.
2. Mark the PSCADraw it and check controls at every crossing.
3. Re point your sitesSwap some clean surface swabs for residue and vacuum samples.
4. Fix the organism listSalmonella throughout; Cronobacter spp. for infant products; Enterobacteriaceae for hygiene.
5. Check technique and laboratorySlightly moistened sponges, the right neutraliser and an agreed test portion.
6. Write the decision treeActions per organism and zone, plus hold and release rules.
7. Schedule the trend reviewMonthly, with clause 2.5.7 (d) triggers logged.

Planning help: choosing an EMP consultant, EMP costs and common EMP audit nonconformities.

Frequently asked questions

Published by ASC Food Safety, South African food safety and quality consultants. This article is general guidance and not a substitute for site specific advice.