Digital PCR sharpens wastewater influenza surveillance when H5 and seasonal viruses overlap
Wastewater surveillance has become a valuable population-level layer of influenza monitoring, but a broad influenza A signal can be hard to interpret when seasonal influenza and avian influenza are both circulating. A new dual-duplex reverse transcription digital PCR strategy addresses that problem by pairing the CDC Flu SC2 assay components with the European Commission Joint Research Centre’s JRC-HA and JRC-MP assays to separate HPAI H5N1 clade 2.3.4.4b from seasonal human influenza A in mixed wastewater samples.
The practical challenge is straightforward: matrix-gene assays are excellent at confirming influenza A is present, but they do not tell clinicians, veterinarians, or environmental health teams whether the signal comes from human seasonal strains, avian strains, or both. In a setting where clade 2.3.4.4b continues to circulate widely in wild and domestic birds and has caused spillover into mammals and occasional human infections, a more specific environmental assay has obvious surveillance value.
Why matrix-gene screening alone leaves too much ambiguity
The influenza A matrix gene remains the standard screening target because it is conserved across strains. That strength becomes a weakness in wastewater, where a single amplicon can reflect multiple sources and multiple hosts. Mixed sewer inputs, changing shedding patterns, and matrix effects all complicate interpretation, and that problem deepens when seasonal influenza activity rises at the same time as avian influenza circulation.
The JRC approach was designed to add resolution without abandoning broad detection. The JRC-HA assay targets a conserved hemagglutinin region in HPAI H5N1 clade 2.3.4.4b, while the JRC-MP assay retains broader influenza A detection through the matrix protein target. In effect, one assay asks whether influenza A is present and the other asks whether the signal fits the clade-specific avian signature most relevant to current European surveillance concerns.
In silico analysis favored clade-level discrimination
The first step was a comparative in silico assessment of publicly available primer and probe sequences from three H5 assays alongside the JRC-HA and JRC-MP designs. Predicted detection of HPAI H5 clade 2.3.4.4b was broadly similar across assays, but the JRC-HA design offered better discrimination against non-2.3.4.4b HPAI clades. Both JRC assays also showed the lowest predicted cross-reactivity with low pathogenicity avian influenza H5 sequences.
That matters because environmental surveillance does not benefit from vague positivity. A signal that separates clade 2.3.4.4b from other H5 lineages can focus confirmatory testing and help public health teams decide whether the wastewater result suggests a broader avian event or a more specific high-concern lineage.
Viral isolates from northern Italy confirmed analytical performance
The assay was then tested against recent HPAI H5N1 clade 2.3.4.4b isolates from wild birds and poultry collected in northern Italy during October through December 2025. Sequence alignment showed high conservation across primer and probe binding regions, indicating that recent viral evolution had not materially compromised assay inclusivity. Two substitutions appeared in the JRC-HA probe-binding region in three samples, but both quantitative reverse transcription PCR and RT-dPCR still detected influenza A in all 10 samples, including low-concentration material.
A small technical adjustment improved the assay’s behavior in the most difficult samples. Lowering the JRC-HA RT-dPCR annealing temperature from 60°C to 56.4°C improved partition separation for three samples without undermining specificity, and in silico analysis suggested that stable primer-dimer formation remained unlikely. The limit of detection was 5 copies per reaction, consistent with earlier work.
One sentence captures the operational value here: the assay still detected avian influenza at the edge of analytical sensitivity.
The CDC influenza B assay was negative in all 10 samples, while the CDC influenza A assay detected influenza A in all 10 with performance comparable to the JRC assays. Broad detection worked exactly as intended, but it could not separate avian from seasonal human influenza A in a mixed specimen.
Spiked wastewater showed how mixed circulation changes interpretation
Because confirmed HPAI-positive wastewater samples were unavailable, the investigators created three simulated co-circulation scenarios by spiking defined amounts of avian and seasonal influenza A RNA into influenza-negative wastewater extracts. That design does not replace field testing, but it does mirror the interpretive problem that environmental labs face when both viruses are present in the same catchment.
When only HPAI H5N1 clade 2.3.4.4b RNA was added, the JRC-HA, JRC-MP, and CDC influenza A assays produced comparable copy numbers, suggesting detection near the analytical limits of the JRC methods. In the second scenario, low concentrations of HPAI H5N1 RNA were combined with seasonal influenza A RNA at 35 copies/µL. Here, the JRC-HA and JRC-MP assays selectively detected the avian RNA, while the CDC influenza A assay detected both sources without distinction.
The third scenario created the most realistic stress test: low concentrations of HPAI H5N1 RNA were added to a high seasonal influenza A background of 3,500 copies/µL. Under those conditions, the JRC-HA assay still detected HPAI H5N1 RNA without cross-reactivity, while the JRC-MP assay detected both avian and seasonal influenza A but yielded lower copy number estimates than the CDC influenza A assay. The avian-specific signal remained visible even when seasonal influenza dominated the sample.
Early-warning use depends on field validation, not just bench performance
The authors propose using CDC Flu SC2 together with JRC-HA and JRC-MP in a dual-duplex workflow. In that framework, CDC Flu SC2 preserves routine community surveillance for influenza A and B, and the JRC assays follow as a second step to identify HPAI H5N1 clade 2.3.4.4b when an influenza A signal appears. The result is not a different public health response in every case, but a sharper one: less ambiguity, better prioritization of confirmatory testing, and cleaner integration with veterinary, wildlife, domestic animal, and human surveillance data.
Field conditions still need to be tested. The evaluation relied on animal-derived viral material and spiked wastewater extracts rather than naturally contaminated wastewater samples, so matrix effects in real sewer networks remain unresolved. Influenza viruses also evolve quickly, which means sequence-based reassessment and assay updating will be necessary as new H5 clades emerge. For now, the main advance is more modest and more useful: wastewater influenza surveillance can move from broad detection toward clade-aware interpretation without giving up the sensitivity needed for early warning.