Microbial Content of Frozen Mulberries: What the Results Really Mean

    Frozen mulberries are not sterile, and a low total aerobic count does not by itself prove that a lot is safe. A useful microbiological assessment separates three questions: how many culturable organisms are present, whether hygiene or spoilage indicators are unusually high, and whether a defined pathogen or enteric virus is detected. Aerobic plate count, yeast and mold results mainly describe process hygiene and spoilage pressure. Salmonella, Listeria monocytogenes, pathogenic E. coli, hepatitis A virus, and norovirus answer different safety questions and require their own methods.

    There is no universal rule that every frozen mulberry lot must have total aerobic count below 10,000 CFU/g and yeast and mold below 100 CFU/g. Those numbers can be private specifications, but they are not a complete or globally applicable safety verdict. The correct limit depends on the destination law, whether the fruit will be eaten without a validated kill step, the customer's risk assessment, the laboratory method, and the sampling plan. Our practical recommendation is to approve the test panel and acceptance rules before production, then interpret the certificate of analysis together with process controls, traceability, packaging, and cold-chain records.

What "Microbial Content" Includes

    The phrase "microbial content" is often used as if one laboratory result describes everything on the fruit. It does not. A count is produced under a particular medium, incubation temperature, incubation time, dilution scheme, and reporting rule. Change the method and the recoverable population may change. That is why a specification should name the method or an accepted equivalent instead of stating only "TPC" or "microbiology passed."

    CFU/g means colony-forming units per gram. It is an estimate of organisms able to form visible colonies under the chosen conditions, not a direct census of every microbial cell. "Not detected in 25 g" is a presence-or-absence result for a defined analytical portion. It does not mean that every gram in a shipment has been tested. For that reason, the sample plan-how many units are taken, from which cartons, and how they are combined-belongs beside the limit.

    The FDA's current Bacteriological Analytical Manual describes APC as an indication of the level of microorganisms in a product and provides a procedure for frozen foods. Its separate yeast-and-mold chapter uses different media and incubation conditions. That separation is the practical lesson: results are comparable only when the method, sample preparation, reporting basis, and detection limit are known.

Why One "Acceptable Count" Does Not Fit Every Lot

    A number becomes meaningful only after its purpose is defined. You can use APC and yeast-and-mold limits to monitor whether field condition, washing, sorting, sanitation, or packing is drifting. If the fruit will be sold as a ready-to-eat retail product, add the hazard-specific criteria required by the market and risk assessment because the label and common use imply consumption without cooking. A jam factory with a validated heat process can control the final hazard through a different process. These are not interchangeable acceptance decisions.

    Destination rules also distinguish food categories. The consolidated EU microbiological-criteria regulation, for example, specifies Salmonella not detected in 25 g for pre-cut ready-to-eat fruit and vegetables and applies Listeria criteria according to the ready-to-eat category and the product's ability to support growth. The same regulation does not create one universal APC or yeast-and-mold limit for every frozen whole berry. A whole mulberry lot therefore cannot be declared globally compliant from two generic counts copied from another specification.

    Research values are useful context, not automatic purchase limits. A 2021 mulberry study reported microbial load below 1 × 10⁵ CFU/g under one calcium-hypochlorite pretreatment, while a 2015 study found that freeze-thawing reduced total aerobic bacteria by more than tenfold under its experimental conditions. The products, treatments, storage periods, and analytical designs were specific to those studies. Neither result proves that 100,000 CFU/g is a universal commercial limit or that freezing will reliably deliver the same reduction in another factory.

    Worked example: Suppose Laboratory A reports APC as 2.7 log₁₀ CFU/g and Laboratory B reports 3.7 log₁₀ CFU/g for two comparable lots tested by the same method. The first is about 10^2.7, or 501 CFU/g; the second is about 10^3.7, or 5,012 CFU/g. The difference is roughly tenfold, not "one unit." If your established process baseline is near the first result, the second result deserves a deviation review even when both numbers sit below a loose private ceiling of 10,000 CFU/g. Check raw-fruit condition, wash-water control, sanitation, hold time, and sample representativeness before release.

Where the Microorganisms Come From-and What Freezing Changes

    Mulberries are soft, high-moisture fruit. Damage and juice leakage create more opportunity for contamination to spread across sound fruit. Relevant sources begin before the freezing tunnel: soil contact, animal intrusion, agricultural water, wet harvesting conditions, unclean containers, worker hygiene, delayed cooling, contaminated wash or rinse water, equipment surfaces, and contact between raw and cleaned fruit. Codex berry-hygiene guidance specifically includes mulberries and highlights physical damage, clean water, worker hygiene, separation, and prevention of cross-contamination.

    Freezing largely stops microbial growth while the product remains continuously frozen, but it is not a sterilization step. Some organisms may be injured or reduced, while survivors can remain. Viruses are a separate concern because they do not multiply in food, so APC cannot reveal them. FDA's fresh-and-frozen berry strategy states that hepatitis A and norovirus outbreaks have been linked to berries and emphasizes sanitary practices from the field through processing. In its FY2019–2023 surveillance assignment, FDA tested 1,558 frozen strawberry, raspberry, and blackberry samples, detecting hepatitis A virus in eight and norovirus in ten. Mulberries were not part of that assignment, so the data should be used as berry-category hazard evidence, not as a mulberry prevalence estimate.

    Blanching should not be inserted into every frozen-mulberry process description. It is common for many vegetables because it controls enzymes and can reduce microbes, but frozen berries intended for raw use may be processed without a microbiocidal step. Codex's berry guidance expressly covers frozen ready-to-eat berries produced without such a step. If a supplier uses blanching, pasteurization, irradiation, or another intervention, request the validated parameters and product effect; if not, build control around field hygiene, water, personnel, sanitation, testing, and intended-use instructions.

    Our IQF freezing guide explains the same operational boundary: individual quick freezing improves piece separation and handling, but it does not replace hygiene or hazard control. For a broader review of frozen-produce risks and test planning, see the frozen-food safety and testing guide.

Choose the Test Panel from the Intended Use

    Start with the product's real route to consumption. Will the berries go directly into a smoothie, retail pouch, thawed dessert, yogurt topping, or garnish without a validated kill step? Will they be cooked in a jam, bakery filling, or sauce under controlled time and temperature? Will the destination authority classify the product as ready-to-eat, an ingredient for further processing, or something else? The answer changes the hazard analysis and the release panel.

    For process trending, APC plus yeast and mold is a reasonable starting pair. Add an appropriate hygiene indicator when it has a defined role in your process plan. For food-safety release, select hazard-specific tests from the destination requirement and risk assessment rather than assuming that a low indicator count substitutes for them. A ready-to-eat berry program may need Salmonella and Listeria criteria and may need an enteric-virus strategy; a further-processed ingredient program should connect incoming-lot criteria to the validated downstream control.

    Method selection matters as much as the organism name. Specify a recognized reference method or a validated equivalent, the reporting unit, the analytical portion, and the limit of detection. ISO 4833-1 covers colony count at 30°C for aerobic organisms, while ISO 21527-1 covers viable yeast and mold enumeration in products with water activity above 0.95, including fruits. FDA BAM provides another recognized method family. Do not compare a result from one method with a historical limit built around another method unless your laboratory has established equivalence.

    Sampling is the most common hidden weakness. One composite taken from the top of one carton may miss variation across a production lot. Define the lot, number of sample units, carton locations, sampling time, sample mass, compositing rule, transport temperature, and laboratory receipt condition. If a pathogen criterion uses five sample units, a report based on one unit is not the same plan even when both say "not detected."

    Practical example: A hypothetical smoothie brand plans to use whole IQF mulberries without heating. Its proposed specification lists APC and yeast/mold only. Because the use is ready-to-eat, we would not treat those two counts as a complete safety panel. We would map the destination requirements, add the applicable pathogen criteria, assess enteric-virus prevention and verification, define the sample plan, and approve the label and handling instructions before accepting production. The action is to close the hazard gap before the first shipment, not after a low APC result creates false confidence.

How to Write a Defensible Microbiological Specification

    A useful specification is an agreement about decisions, not just a list of organism names. Put the exact product form at the top: whole IQF mulberry, species or commercial variety when relevant, unsweetened or formulated, intended use, and whether further heat treatment is expected. Then state each test, reference method, sampling plan, limit, unit, stage of testing, and action for a nonconforming result.

    For quantitative indicators, define whether the limit applies to each sample unit, the mean, or a multi-class plan with m, M, n, and c. For presence-or-absence criteria, state the analytical portion, such as "not detected in 25 g," and the number of units. Include the reporting rule for results below the method's quantification limit. Without that detail, "<10 CFU/g" may be misread as zero when it actually means the method could not quantify below ten under the reported conditions.

    Link the release decision to a documented response. An indicator excursion should trigger a hold, trend review, investigation, and defined retest or disposition procedure. A pathogen detection is not handled as an ordinary quality deviation; follow the applicable food-safety, withdrawal, recall, notification, and corrective-action requirements. The current EU regulation, for example, requires corrective action and cause investigation after unsatisfactory results and specifies withdrawal or recall for food-safety-criterion failures in product already placed on the market.

    A certificate of analysis should identify the product, batch or lot, production date, sampling date, laboratory, methods, results, units, detection limits where relevant, and authorization. Review the laboratory's competence and scope; a polished PDF is not proof that the named method or matrix falls within accreditation. XMSD's current certification and document page shows the types of commercial evidence that can support a review, but the approved order specification remains the controlling document.

Process Evidence Matters More Than a Passing Snapshot

    Finished-product testing samples a very small fraction of a commercial lot. It is evidence, not a substitute for prevention. A stronger program connects the result to field and facility controls: agricultural-water risk management, exclusion of damaged or ground-contact fruit, worker hygiene, prompt cooling, hygienic containers, controlled wash water, separation of raw and cleaned zones, sanitation verification, preventive maintenance, environmental control, and traceability.

    Trend data are especially valuable. Review APC and yeast/mold results by farm, harvest date, line, shift, sanitation cycle, and season. A single passing lot can sit inside a worsening trend; several stable lots can show that controls are behaving consistently. Set an internal alert level below the contractual rejection limit when enough process data exist. The alert is not a legal limit-it is an early-warning tool that gives the processor time to correct drift before a finished lot fails.

    Packaging protects the already-controlled product from handling damage, moisture loss, odor transfer, and exposure after packing. It does not sanitize the fruit. Check seal integrity, liner condition, carton cleanliness, coding, and whether the pack survives the actual frozen route. XMSD's frozen-food packaging overview can help you frame questions about inner bags, cartons, labels, and pallet protection without confusing packaging performance with microbial validation.

    Practical example: A hypothetical arrival lot has a passing COA but several cartons contain large ice masses, wet outer board, and a damaged inner liner. The laboratory result describes the submitted sample at the time of testing; it does not erase the arrival evidence. Place the affected lot on hold, preserve photographs and temperature records, map the damaged cartons, sample under an agreed investigation plan, and decide disposition from the combined evidence. The action protects both safety and quality while the root cause-loading exposure, seal failure, or temperature fluctuation-is investigated.

A Practical Release and Arrival Decision

    Before production, define the product, use, market, hazards, methods, sample plan, limits, and corrective actions. Before shipment, match the COA to the lot and review process deviations, sanitation evidence, traceability, packaging, and temperature control. At arrival, verify seal condition, coding, frozen state, evidence of thawing or leakage, temperature records, and the agreed inspection sample. Keep safety and quality findings separate: freezer burn or soft texture may be a quality failure, while a pathogen result or uncontrolled contamination event is a safety matter.

    Do not average away a serious signal. If one sample unit is positive under a presence-or-absence criterion, four negative units do not cancel it. If one carton has lost lot identity or its seal is open, isolate it even when neighboring cartons look normal. Your written plan should say who places the lot on hold, which records are preserved, how additional sampling is authorized, and who can release or reject the stock. This turns laboratory data into a controlled decision instead of an informal debate after arrival.

    If you need a quick decision rule, use this one: never approve a frozen mulberry lot because two indicator counts look low. Approve it because the intended use is defined, applicable hazards are covered, the methods and sampling plan match the specification, preventive controls are credible, the results comply, and the shipment arrives with intact traceability and cold-chain evidence.

References

• U.S. FDA: Bacteriological Analytical Manual
• U.S. FDA: FY2019–2023 Microbiological Surveillance Sampling of Frozen Berries
• U.S. FDA: Strategy to Prevent Norovirus and Hepatitis A Outbreaks Associated with Berries
• EUR-Lex: Consolidated Regulation (EC) No 2073/2005 on Microbiological Criteria for Foodstuffs
• Codex Alimentarius: Draft Annex on Hygiene for Fresh and Frozen Berries, Including Mulberries
• ISO 4833-1: Colony Count at 30°C
• ISO 21527-1: Enumeration of Yeasts and Molds in High-Water-Activity Foods
• Kim et al. (2015): Effects of Freezing Temperature on Quality of Mulberry
• Powpongchan (2021): Pretreatments and Freezing Methods for Frozen Mulberry Quality