Evaluation of PCR-Based H5N1 Influenza Detection Methods in Milk from an Interlaboratory Comparison Study Demonstrating Method-Dependent Sensitivity Variability

• Evaluated the performance of PCR-based H5N1 influenza detection assays across 45 participants using blinded H5N1-spiked milk. • Detection rates ranged from ≥98% at high concentrations to only 24% at low levels. • Sensitivity varied based on the methods used. • This work highlights the critical need for interlaboratory comparison exercises to assess and improve laboratory testing in response to emerging threats. Since March 2024, H5N1 highly pathogenic avian influenza (HPAI) has been detected in U.S. dairy cattle. Infectious virus shed in unpasteurized milk, a matrix not previously evaluated for H5N1 detection, makes real-time reverse transcription quantitative polymerase chain reaction (rRT-qPCR) assays critical for animal diagnostics and food safety through screening and early detection of milk samples. This interlaboratory comparison exercise evaluated the performance of PCR-based H5N1 assays across 45 participants from government, academic, and commercial laboratories using blinded milk samples spiked with an inactivated H5N1 virus culture at concentrations ranging from 85 to 170,000 genome copies/100 μL. Participants tested samples using their routine protocols. The rate of detection (ROD), Ct values, and levels of detection where 50% of the samples yielded a positive result (LOD 50 ) were analyzed with methods grouped by extraction and PCR kit combinations for sensitivity comparison. For high-concentration samples (17,000 and 170,000 copies/100 μL), ROD was ≥99%; however, detection decreased significantly at lower concentrations, with only 34% of participants reliably detecting samples at 1,700 copies/100 μL and 24% at 85 copies/100 μL. Eight distinct method types were used by participants, with some methods consistently detecting low levels of virus, and others having a 0% detection rate at low copy levels. This study reveals substantial variability in H5N1 detection capabilities across laboratories, with method selection impacting sensitivity. This finding has important implications for human and animal health surveillance and emergency response, highlighting the need for evaluation of protocols and quality control measures for H5N1 testing in milk.