Multiplex Respiratory Panels: How to Test for Coinfection Masking

To investigate whether high influenza A can mask low respiratory syncytial virus (RSV), select a clinically relevant pair, challenge RSV at its established LoD against a justified very-high influenza A level, and evaluate the reverse direction. Record each target’s actual call and the assay’s validity status for every replicate. Compare the low-target behavior with an appropriate baseline before attributing a missed detection to competition. A table of blood, mucin or medication challenges addresses a different interference question. FDA respiratory viral panel guidance, Section 6.B.iii(c).

This guide turns that recommendation into a study-planning matrix for multiplex respiratory nucleic acid amplification tests (NAATs). It explains what to choose, what to record and where the evidence stops. It does not prescribe a universal high concentration, sample size or pass rule. The central decision is whether your selected mixtures adequately challenge low-target detection on your assay, rather than whether another panel once produced a favorable result.

Read the guidance within its stated scope

The FDA document was issued October 9, 2009. It describes respiratory viral panel multiplex nucleic acid assays using 21 CFR 866.3980 and product code OCC, and identifies additional analyte-specific guidance for hMPV and influenza A subtype differentiation. Those are the document’s stated scope identifiers, not a determination that every contemporary respiratory panel belongs in that classification. Establish the classification and applicable controls for the actual intended use separately before applying a historical study design. FDA guidance, Sections 1–3.

Section 6.B.iii(c) recommends assessing clinically relevant co-infections with each pathogen probed by the assay. Pair selection should reflect combinations that occur or are widespread; influenza A with RSV is its example. The same section recommends a low target at LoD and another at a very high concentration, with evaluation placed in LoD, reproducibility or interference studies. The guidance permits alternative approaches that provide equivalent assurances of safety and effectiveness. Distinguish those design recommendations from applicable regulatory obligations.

The reciprocal challenge appears in Section 6.B.v, Precision (Repeatability/Reproducibility): evaluate low concentrations around the assay cut-off of one virus with high concentrations of the other, and vice versa. Keep this section distinct from Section 6.B.iv, which concerns cut-off determination. Likewise, the competitive-interference subsection does not specify a fixed replicate count or numerical acceptance threshold; broader precision recommendations address other design factors. FDA guidance, Section 6.B.v.

Why a strong target may hide a weak one

Preferential amplification is a documented concern in multiplex PCR. Elnifro and colleagues’ 2000 review describes PCR drift, involving stochastic effects particularly in early cycles, and PCR selection, involving properties that favor amplification of particular templates. It also discusses how more efficient amplification of other targets or nonspecific products can reduce amplification of the desired target. This provides a mechanistic reason to test mixtures; it does not predict a specific modern panel’s failure rate. Elnifro et al., multiplex PCR review.

Do not infer a universal cycle-threshold pattern or reagent-depletion mechanism from a missed call. Different panels have different amplification, detection and reporting designs. Near a detection limit, the low target may also be missed without a competitor. A useful investigation therefore distinguishes an observed result from its proposed explanation: first document the calls and conditions, then assess whether the high-target mixture changes low-target detection relative to its baseline.

FDA’s risk discussion includes the possibility that one virus out-competes another associated with a more serious infection, leaving only the less consequential virus reported. The practical implication for a validation team is to characterize that analytical risk. A positive call for one pathogen does not establish that a negative call for another excludes co-infection. The example does not support predicting a clinician’s treatment decision from the panel result alone. FDA guidance, Section 4.

Choose pairs using relevant co-detection evidence

Begin with the claimed targets and a documented rationale for the combinations selected. Use the guidance’s influenza A–RSV example as an example, not a universal benchmark that substitutes for considering the other pathogens your assay probes. As an editorial planning approach, map every claimed target to the combinations considered, the evidence supporting selection and the reason for any omission. Neither an unexplained small subset nor an automatic all-pairs rule answers the relevance question.

The 2019 NxTAG Respiratory Pathogen Panel publication reports 217 mixed detections among 1,211 NxTAG-positive prospective specimens, or 17.9%. Of those 217 mixed detections, 182 were dual detections, reported as 83.9%. Enterovirus/rhinovirus was involved in 47% of mixed detections and RSV in 31.3%. These are one study’s panel-based observations, not a population-wide co-infection rate or proof of a particular masking mechanism. The paper includes manufacturer-affiliated authors. Gonsalves et al., Section 3.1.3.

The 47% figure does not mean that enterovirus/rhinovirus plus RSV accounted for 47% of co-infections. Those percentages describe each target’s involvement across mixed detections, which can contain more than two pathogens. They do not establish a pair-specific frequency, the direction of a high/low concentration disparity, or that co-detection caused a clinical illness. An enterovirus/rhinovirus–RSV challenge may be a candidate to justify, but the marginal percentages alone cannot rank it as the most common pair.

For a combined enterovirus/rhinovirus assay result, retain that combined identity in your evidence map. Do not silently relabel it as rhinovirus alone. Similarly, use the RSV types or subtypes that the assay actually reports. These distinctions make the pair rationale traceable to the claimed result rather than to a broader organism name. Keep additional pairing proposals separate from combinations explicitly named by FDA.

Define high and low before filling the matrix

The NxTAG paper’s competitive-interference subsection reports one pathogen at 1.00 × 10^5 TCID50/mL and another at 3×LoD, with no interference observed under the evaluated conditions. This is a published design precedent and counterevidence to any assumption that competition necessarily produces masking. It is not a universal maximum clinical titer, an acceptance threshold or a requirement for a different assay. The narrative does not provide the pair-by-pair competitive-interference replicate results needed to recreate a complete experimental ledger. Gonsalves et al., Section 3.3.2.2.

Keep that precedent separate from the FDA recommendation to use a low target at LoD. A 3×LoD challenge is not automatically a C95 challenge. FDA describes C95 as a low-positive concentration producing positive results approximately 95% of the time, C5 as a high-negative concentration producing positives approximately 5% of the time, and a separate moderate-positive level. The NxTAG repeatability section calls 1×LoD low positive and 3×LoD moderate positive. Use your assay’s established concentration-response evidence to describe what a selected low level represents. FDA guidance, Section 6.B.v; NxTAG study, Section 3.3.3.

For each planned mixture, write down the low target’s actual concentration, units, relevant LoD or cut-off study and relationship to that benchmark. For the high target, document the concentration, units and rationale for calling the challenge very high. Do not substitute a concentration ratio for the underlying measurements. Keep TCID50/mL and nucleic-acid copy units explicit rather than assuming they are interchangeable. These are planning records, not additional numerical limits supplied by the guidance.

A reciprocal high/low planning matrix

The table is an original editorial template, not an FDA form or completed experiment. The first two rows operationalize the guidance’s named pair. The remaining two illustrate an additional candidate requiring an assay-specific selection rationale. “Record after testing” means no result is asserted. Set the planned replicate count and acceptance criteria in the protocol before generating results; no number is supplied here as a regulatory minimum.

For interpretability, the proposed protocol also includes a matched low-target-alone condition, high-target-alone condition and appropriate negative/control conditions. These are editorial design suggestions to help separate baseline low-level dropout, unexpected calls and invalid results from an effect associated with the mixture. Keep the relevant specimen matrix and processing conditions comparable between the baseline and challenged samples, and explain any differences. A loss in one mixed replicate is a finding to investigate, not by itself proof of competition.

A reusable replicate record identifies the pair, direction, material identity, matrix, concentration and units for each target, the planned condition, run and replicate identifier, each target’s initial call, assay validity and any available underlying signal. Retain site, operator, lot and instrument identifiers when those factors are part of the study. Link retests to the original record instead of overwriting it. This makes the summary auditable without inventing experimental observations.

Size the experiment to the question and planned analysis. A small set of perfect detections does not establish a population detection probability with certainty. Conversely, combining unlike targets or directions can hide the condition in which losses occurred. Keep replicate allocation and the handling of invalid or equivocal results explicit before testing. Justify the planned sample size and acceptance rule for the intended analytical claim; Section 6.B.iii(c) does not supply a universal numerical recipe.

Choose the study placement that preserves the question

FDA names three possible placements. Their common feature is the high/low target challenge; placement alone does not establish adequate coverage. Use the comparison below to choose where the challenge fits while retaining the necessary condition-specific evidence. FDA guidance, Section 6.B.iii(c).

For reproducibility, Section 6.B.v discusses sources of variation and recommends near-cut-off levels as well as a moderate positive. It also calls for site-specific and combined summaries of values above and below the cut-off, including invalid and, where applicable, equivocal percentages. A moderate-positive mixture should not silently stand in for all near-cut-off behavior. A combined study needs to preserve both questions in its design and analysis.

CLSI lists MM17, second edition, published May 31, 2018, as a guideline for analytical validation and verification of qualitative multiplex nucleic acid assays. Its public catalog describes reference/control materials, analysis, reporting and an error-based approach, and labels it FDA-recognized. The catalog explicitly says that specific validation/verification protocols are not included. Use it as a framework rather than as a source of a fixed competitive-interference recipe. The 2009 FDA guidance cites the earlier MM17-A, published in 2008. CLSI MM17 catalog.

For the wider analytical-validation context, see IVD analytical performance validation. Keep the paired-target evidence identifiable when integrating it with that broader program.

Record outcomes before naming the cause

Use separate fields for the observable calls and the interpretation. “Strong target only” is an observed combination; “weak target lost” describes that same low-target absence and is not a second mutually exclusive outcome category. The following ledger vocabulary is an editorial proposal. Apply the assay’s actual reporting and validity rules, including target-specific validity if relevant, rather than assuming an internal-control signal alone determines every result.

The primary analytical summary should identify each pair, direction and concentration, count the relevant observed outcomes and state denominators. Include the baseline comparison and any planned analysis of underlying signals. A weak-target detection fraction in contrived positive mixtures is not automatically clinical sensitivity or negative percent agreement. A set containing only positive mixtures has no negative-comparator observations from which to estimate an NPA.

For diagnostic comparative studies, FDA’s March 13, 2007 statistical guidance recommends tabular comparisons and appropriate accuracy or agreement measures with two-sided 95% confidence intervals, expressed as fractions and percentages. Sensitivity and specificity describe comparison against a suitable reference standard; PPA and NPA describe comparison against a non-reference standard. Report relevant site and subgroup results. Choose terms that describe the actual study instead of applying clinical reporting labels to every analytical matrix. FDA statistical guidance, Sections 4–6.

Do not discard ambiguous results or apply an unexplained rule that every invalid result must be counted as a negative in every agreement denominator. Describe the result categories, accounting and prespecified analysis. The statistical guidance warns that ignoring equivocal outcomes can bias estimates and discusses context-dependent handling. The respiratory guidance separately recommends clinical retesting according to the assay instructions, use of the resulting final results in analysis, and reporting initial retest and final invalid/equivocal rates. FDA statistical guidance, Section 6, recommendation 2; FDA respiratory guidance, Section 6.D.iv.

That planned retest process differs from changing a valid disagreement after a resolver test. Preserve the original comparative table when investigating discordance; outcomes revised through discrepant resolution should not be used to estimate accuracy or agreement. For clinical co-infections, Section 6.D.iv of the respiratory guidance recommends showing candidate-test results for reference-detected co-infections and reference results for candidate-detected co-infections. These two reporting perspectives are distinct from reversing concentrations in an analytical experiment. FDA respiratory guidance, Section 6.D.iv; FDA statistical guidance, Section 6, recommendation 3.

The separate guide to discordant IVD results explains how to retain the original comparison while reporting follow-up investigations.

Keep the conclusion within the tested conditions

A useful report states which target materials, directions, concentrations, matrices and assay conditions were evaluated, what happened to each target, and how the data met the prespecified criteria. If no interference was observed, qualify that finding by those conditions. Do not extend one panel’s result to every chemistry or extrapolate two-target contrived mixtures to all possible clinical co-infections. Label the planning template and the published NxTAG findings separately so they cannot be mistaken for data generated by your team.

The labeling section of the 2009 respiratory guidance recommends explaining that a positive result does not exclude bacterial infection or another viral co-infection, that detection does not necessarily establish the cause of illness, and that a negative result does not exclude respiratory infection or stand alone as a basis for management. Those recommendations remain relevant to interpreting the evidence boundary: favorable bench results do not establish that clinical masking is impossible. FDA guidance, Section 7.

The guidance cautions, in its interfering-substances discussion, that spiking may not accurately model the in-vivo situation. Applying that caution to contrived target mixtures is an editorial inference: controlled mixtures help isolate a challenge but do not reproduce every specimen context. State the materials and processing actually studied rather than claiming that a numerical concentration is a universal clinical worst case. FDA guidance, Section 6.B.iii(a).

Keep competitive interference separate from inclusivity and cross-reactivity. Cross-reactivity asks about an unintended target call; competitive interference asks whether a co-target alters intended-target detection. Substances and microorganisms outside the panel remain separate interference considerations. The useful deliverable is a traceable reciprocal matrix with actual observations and bounded conclusions, not an assurance of clearance or perfect diagnosis.

Sources

• Respiratory Viral Panel Multiplex Nucleic Acid Assay - Class II Special Controls Guidance for Industry and FDA Staff

• CLSI MM17 - Validation and Verification of Multiplex Nucleic Acid Assays (Second Edition)

• Multiplexed detection and identification of respiratory pathogens using the NxTAG respiratory pathogen panel (Gonsalves et al., Methods 158:61-68, 2019)

• Statistical Guidance on Reporting Results from Studies Evaluating Diagnostic Tests - Guidance for Industry and FDA Staff (March 2007)

• Multiplex PCR: Optimization and Application in Diagnostic Virology (Elnifro et al., Clin Microbiol Rev 13(4):559-570, 2000)