At IHU, we initially received 11 nasopharyngeal samples collected 3–10 days earlier from 11 of the LTCF patients. We first attempted virus isolation in a Biosafety Level 3 laboratory, as previously described (5). Within 3 hours of receipt, we tested samples by FilmArray Respiratory Panel 2 multiplex qPCR (BioFire Diagnostics, https://www.biofiredx.com); 4 samples tested HRV RNA–positive. All attempted cultures were negative.

The next day, we extracted DNA and RNA for next-generation sequencing (NGS) (Appendix). We performed qPCR on nucleic acid extracts by using the FTD Respiratory Pathogens 21 Assay (Fast Track Diagnostics/Siemens Healthineers, https://www.siemens-healthineers.com) and an in-house designed qPCR (Appendix). Results of qPCR targeting 46 pathogens were negative for all but 1 sample, which was Staphylococcus aureus DNA–positive.

We then performed NGS on DNA and RNA from the 4 HRV-positive, 2 additional HRV-negative samples, and a negative, sample-free control by using the Ligation Sequencing Kit for library preparation and sequencing on GridION (Oxford Nanopore Technologies, https://nanoporetech.com). Next, we performed viral nucleic acid enrichment by using the Respiratory Viral Research Panel (Twist Bioscience, https://www.twistbioscience.com) and sequenced on the MiSeq platform using MiSeq Reagent version 2 kit (Illumina Inc., https://www.illumina.com), according to manufacturer’s protocol (Appendix).

We conducted bioinformatic analyses of generated NGS reads by using various commercial and freely available tools, as well as in-house pipelines (Appendix). Six hours after starting NGS, 15 NGS reads generated from the samples of 2 patients (13 reads from 1 patient and 2 reads from the other), ranging from 468 to 2,024 nt in length, were found to be best matches with an HRV-B14 genome (GenBank accession no. NC_001490.1). We used those reads to obtain consensus sequences, one of 4,447 nt in 3 contigs for one sample, and the other of 1,019 nt for the other sample (Appendix). The negative control had no HRV reads.

We designed in-house PCR systems by using the Primer3Plus web application (https://www.primer3plus.com/index.html) to fill gaps in reconstructed HRV-B14 genomes (Appendix Table 1). Then, we used IQ-TREE (http://www.iqtree.org) to select the best model for each alignment and built maximum-likelihood phylogenetic trees. We visualized trees incorporating all complete HRV-B14 genomes by using iTOL version 7 (https://itol.embl.de).

Five days after we received the initial samples, we received additional samples from the LTCF outbreak; in total, multiplex qPCRs eventually detected HRV RNA in 13/19 nasopharyngeal swab samples from 10/13 patients. Unexpectedly, 9 samples tested positive with the BioFire assay and negative with the FTD assay, whereas we observed opposite for 2 other samples (Table; Appendix Table 2). An HRV-B14–specific qPCR we designed from partial genomes was positive for 5 patients (Appendix Table 3).

Figure 2

Figure 2. Phylogenetic tree of HRV-B14 from outbreak of severe respiratory illness and death among older adults, France, 2024. Tree includes HRV genomes downloaded from GenBank; accession numbers are indicated. Red font...

Pre-NGS viral nucleic acid enrichment enabled us to obtain 5 partial HRV-B14 genomes from 5 patients. Genomes were 913–6,644 nt in length and 13%–86% of the GenBank reference genome (accession no. NC_001490.1). In the phylogenetic tree, those 5 genomes clustered together with 100% bootstrap value near 5 other genomes available in public databases (Figure 2). However, 1 genome from patient P3 stood apart; its nucleotide similarity was 97.1%–98.1% with the 4 other genomes, whereas those 4 genomes had 99.9%–100.0% similarity to each other. Sequencing of the HRV virus capsid protein (VP) 1–encoding gene confirmed HRV-B14 for 4 of the initial 11 samples tested from the LTCF (Table).

Figure 3

Figure 3. Phylogenetic tree of HRV-B sequences circulating in Marseille during severe respiratory illness and death associated with outbreak of HRV-B14 among older adults, France, 2024. Genome alignment was performed by using...

We also tested 86 residue nasopharyngeal swab samples collected in southeastern France as part of routine diagnostic testing during the same timeframe as the cases from the LTC facility during late autumn 2024. VP1 virus genotyping for all 86 nasopharyngeal swab samples identified 9 additional HRV-B sequences, 3 of which were HRV-B14 on the basis of BLAST (https://blast.ncbi.nlm.nih.gov) similarity and phylogeny (Figure 3).

The outbreak among LTCF residents resulted in 18 severe illnesses, from which 8 (44%) patients died. Of note, metagenomic methods enabled us to identify HRV-B14 within ≈30 hours of receiving samples; without those methods, the cause of the outbreak would have largely remained elusive.

Viral nucleic acid enrichment successfully lengthened the HRV-B14 genomes from the LTCF outbreak, 1 of which greatly differed from the 4 others, indicating >2 closely related HRV-B14 strains were circulating during this outbreak. In addition, HRV-B14 was detected in 3% of HRV-positive samples received at our laboratory during the same period as the LTCF outbreak, indicating that HRV-B14 circulated in that geographic area in fall of 2024.

Previous metagenomics-based approaches have successfully identified various pathogens during outbreak investigations (6–10). For instance, in the United States, metagenomics used on 6 nasopharyngeal swab samples from 4 case-patients in a pulmonary ward and on 10 nasopharyngeal swab samples from 9 control outpatients enabled investigators to determine that case-patients’ HRV sequences were not genetically related, dispelling concerns of single-source nosocomial transmission <24 hours of receiving samples (7). That case demonstrates the benefits of metagenomics in routine settings.

During the LTCF outbreak, S. aureus DNA was the only other pathogen detected by qPCR, and only in 1 sample. Notwithstanding that finding, co-infections might have been missed because specific qPCR, serologic, or culture assays were not used for all possible infectious agents; because viral or microbial loads were too low for detection; or because our laboratory was not involved in first-line diagnosis and the time between respiratory sample collection and testing at our laboratory was up to 10 days.

Severe and life-threatening HRV infections previously have been reported among older persons (11,12). Illness severity as assessed through rates of hospital admission, intensive care unit admission, and death in those studies were similar to those previously reported among immunocompromised adults infected with the 2009 influenza A(H1N1) pandemic virus (11,12) or to rates among hematopoietic cell transplant recipients with RSV or influenza virus infection (13,14). We speculate that HRV could lead to severe disease and death in fragile older persons, as seen with RSV (1), but data are scarce. Moreover, those prior assessments did not address whether severity differed between viral genotypes, and no specific data on HRV-B14 epidemiology, infectiousness, or clinical severity are available.

In summary, our findings indicate >1 HRV-B14 strain was circulating and caused severe illness and death among a population of older adults in southeastern France. Rapid identification (≈30 hours) of the viral etiology in this case demonstrates that expanded use of real-time metagenomics for routine diagnostic testing in clinical virology laboratories, particularly targeting respiratory viruses other than vaccine-preventable influenza viruses and RSV, could enable early detection and management of such respiratory pathogens.