How to Test for Allergens: ELISA, Lateral Flow, and PCR

Allergen testing — the analytical verification of allergen control — answers questions that observation alone cannot: is the allergen actually absent? Which test to use, when to use it, and how to interpret the results determines whether testing protects you or misleads you. This guide shows how to implement allergen testing that is scientifically sound and practically useful.

Step 1: Define the question before choosing the test

Clarify what you are trying to learn. Cleaning validation asks whether the changeover removes the allergen — that needs quantitative proof. Routine verification asks whether the program is working — periodic checks suffice. Supplier verification asks whether the ingredient contains undeclared allergen. Incident investigation asks where the allergen is. Label verification asks whether the product matches its declaration.

The purpose drives the method: whether you need quantitative or qualitative results, what sensitivity the decision requires, and how fast the answer is needed. The wrong test for the question gives a misleading answer with confidence. Define the question first.

Step 2: Understand ELISA — the quantitative standard

ELISA (enzyme-linked immunosorbent assay) uses antibodies to detect and quantify the specific allergen protein. Its strengths are real: quantitative, sensitive, specific, and widely accepted by auditors and regulators. Its limitations are equally real and must be respected. Food matrices interfere with the assay. Processing — especially heat — alters proteins and can cause under-detection. And it takes a laboratory, with the cost and turnaround that implies.

Use commercial kits validated for your allergen, interpret the mg/kg results against your thresholds (VITAL-informed), and deploy ELISA where the decision needs a definitive number: cleaning validation, supplier verification, confirmation of screening positives. ELISA is the workhorse — understand it, and respect its limitations.

Step 3: Understand lateral flow — the rapid on-site screen

Lateral flow devices (LFDs) are immunochromatographic strips giving yes/no or semi-quantitative results in minutes, on site, with simple technique. That speed makes them ideal for changeover verification, routine checks, and incident screening — decisions that cannot wait for the lab.

But know the limits: qualitative or semi-quantitative at best, detection limits higher than ELISA, susceptible to matrix interference, and not quantitative. They screen; they do not confirm. Confirm positives by ELISA before making significant decisions, and train the operators in the technique and the interpretation — a misread strip is worse than no strip. Screen rapidly; confirm quantitatively.

Step 4: Understand PCR — the DNA complement

PCR detects the allergen’s DNA rather than its protein. It is sensitive and specific, and useful in highly processed foods where DNA survives heat better than proteins do. But DNA is not protein, and allergenicity is a property of protein. The correlation between a DNA signal and the allergenic protein must be understood before the result drives decisions. Quantification is difficult, the cost is higher, and it needs a specialized laboratory.

Use PCR as the complement: the ELISA-negative but suspect case, the highly processed matrix. Interpret cautiously, with the protein correlation understood. PCR complements the protein methods; it does not replace them.

Step 5: Select the method fit for purpose

Decide on the question first — quantitative or rapid? Then check the allergen: is a kit available and validated for your matrix? Consider the matrix interference and run spike-and-recovery. Align the sensitivity: the method’s detection limit must sit below the threshold your decision needs. Weigh the speed against the decision timeline and the cost against a sustainable program budget.

The practical answer is often tiered: LFD screening on site for rapid decisions, ELISA confirmation in the lab for the definitive ones. Fit the method to the purpose; the tiered approach is both practical and rigorous.

Step 6: Validate the method on your matrix — spike-and-recovery

Before trusting any result, prove the method can detect the allergen in your product. Spike a known amount of the allergen into your representative product matrix, run the method, and measure the recovery. Test at levels around your decision threshold. Define adequate recovery in advance.

If the method cannot recover the spike, it is unsuitable for that matrix — find an alternative. Document the study. Spike-and-recovery is the essential check on the method’s eyes: a method that cannot find the allergen in your product proves nothing when it reports “not detected.” The matrix is the method’s challenge; prove it is met, then trust the results.

Step 7: Design the sampling

Plan the locations — worst-case and representative — and define the technique: swabs or product samples, consistent procedure, trained samplers. Take enough samples for the decision to rest on, statistically adequate. Time the sampling to the question: after changeover, at goods receipt, during the incident. Handle samples properly — preserved, transported, tested promptly — and document everything traceably.

Sampling is the result’s foundation. Unrepresentative sampling produces misleading results no method can fix. Sample thoughtfully.

Step 8: Interpret the results meaningfully

A positive means detected and quantified: compare against the threshold, assess the dose per serving against the reference values, and take the defined action. A negative means not detected — understand the detection limit, because “not detected” is not “absent” and not zero. For quantitative results, compute the mg/kg, the serving dose, the VITAL comparison. Trend the results over time and read the program’s health in the pattern. Acknowledge the method’s uncertainty.

Interpret scientifically so the decisions are sound. The meaning of a result lives in its context: the limit, the threshold, the trend.

Step 9: Manage the limitations honestly

Acknowledge the caveats openly. Matrix effects cause interference. Processing alters proteins and detection. Sampling is always a limitation — you test a fraction. Every method has blind spots. Make sure the decision-makers understand these limitations, so a “negative” is never read as a guarantee. Combine complementary methods for better coverage.

Honest testing means the limitations are acknowledged. Overconfidence in a test result is the dangerous outcome — worse than no testing, because it stops further inquiry.

Step 10: Build the testing into a system

Schedule the routine testing on a defined, risk-based frequency. Define the triggers: validations, incidents, supplier events. Trend the results and use them to improve the program. Review the methods periodically and update as better ones appear. Budget the program so it is sustained, and document everything retrievably.

Testing belongs to the program — integrated, not ad hoc. Systematize it.

Working notes

Spike before you trust. Spike-and-recovery is the first step: the method proven on your matrix before any result is believed.

Screen fast, confirm carefully. LFD for the rapid decisions, ELISA for the definitive ones. The tiered approach is practical and rigorous.

Trend the negatives too. “Not detected” results, trended at a consistent method limit, reveal the program’s health in their pattern.

Train the samplers. Consistent swabbing technique from trained, assessed samplers — the sampling is the result’s foundation.

Budget the program. Planned, sustained testing costs less than ad hoc testing with its gaps and its rush premiums.

How it goes wrong in practice

The blind ELISA is the false confidence: a validation passes on ELISA negatives, but spike-and-recovery was never performed. When someone finally checks, the matrix is suppressing detection — the method was blind to the allergen in this food. Re-validation with a suitable method gives the true results, and the procedure gets strengthened. The blind method is the dangerous one. Spike first, always.

The LFD positive is the tiered system working: a changeover check comes back positive on the strip, the line is held, and ELISA confirmation quantifies a low level — below the threshold by the VITAL assessment. The decision is informed rather than panicked: screen rapid, confirm quantitative. Confirm the positives; the quantitative context keeps decisions sound.

The PCR confusion is the meaning problem: PCR positive — DNA detected — triggers the alarm, but the protein ELISA is negative. In the highly processed product, the DNA survived while the protein denatured. The decision has to be nuanced, and that nuance has to be communicated: the methods mean different things. DNA does not equal allergen.

The untrended data is the missed signal: testing performed and filed, the gradual increase across months unnoticed because nobody trended. Then trending gets instituted, the program’s health becomes visible, and the data starts paying for itself. Review the data; extract the value.

Common mistakes

Trusting a method never validated on your matrix. The kit works in the manufacturer’s validation foods, so it is assumed to work in yours. Matrix interference then blinds it. Run spike-and-recovery on your actual product before any result is trusted.

Treating LFD as quantitative. A strong line on the strip gets read as “high level” and a faint line as “low,” driving decisions the device was never designed to support. LFD screens; ELISA quantifies. Confirm before deciding.

Reading “not detected” as “absent.” The negative gets reported as zero allergen, and the decision-makers treat it as a guarantee. Communicate the detection limit with every negative: not detected at this limit, not proven absent.

Using PCR as a standalone allergen test. The DNA result drives protein decisions without understanding the correlation. Use PCR as the complement to protein methods, interpreted cautiously, in the situations where DNA detection adds value.

Sampling conveniently instead of representatively. Swabs from the easy surfaces, product from the top of the pallet. The allergen is where it is worst, not where sampling is easy. Design the plan around worst-case and representative locations.

Filing results without trending. Each result is read alone and filed, so the gradual drift — the program slowly deteriorating — goes unnoticed. Trend everything; the pattern is where the early warning lives.

Checklist — allergen testing methods

• [ ] Purpose defined — the question clarified first; method selected for fit; documented
• [ ] ELISA understood — the quantitative standard; strengths and limitations respected; results interpreted against thresholds
• [ ] LFD deployed properly — rapid on-site screening; limitations known; positives confirmed by ELISA; operators trained
• [ ] PCR used as complement — DNA vs. protein understood; applied where it adds value; interpreted cautiously
• [ ] Method fit for purpose — allergen, matrix, sensitivity, speed, and cost considered; tiered approach where practical
• [ ] Matrix validated — spike-and-recovery performed; recovery adequate; documented; unsuitable methods replaced
• [ ] Sampling designed — representative and worst-case; technique consistent and trained; handling proper; documented
• [ ] Results interpreted — positives and negatives read against limits; dose assessed; trends analyzed; actions defined
• [ ] Limitations managed — honestly acknowledged; communicated to decision-makers; methods complemented
• [ ] System integrated — routine scheduled; triggers defined; results trended; methods reviewed; budgeted; documented