The facility is a large (50–100 bed) level III NICU with 5 wards (A–E) that all have single-family rooms. There is no difference in the level of patient acuity across the different wards, and wards have shared staff. The catchment area of the facility includes a 30-mile radius, which includes 3 parishes. According to the facility’s visitation policy, parents are invited to stay over-night with their infant, and siblings > 2 years of age are encouraged to visit during daytime hours. All family and guests must wash hands before visitation. During the winter respiratory season (that covers RSV and influenza seasons), additional guidelines are in place that state visitors will have their temperature checked upon entrance to the unit and that any visitors with fever or reported respiratory symptoms will not be permitted into the unit.

Three days after being notified of the cluster, LDH epidemiologists visited the facility to review medical records of RSV-infected infants and infection control measures. The team manually abstracted data on patient demographics, clinical presentation, interventions and medications, preexisting conditions, and microbiology results. The team examined infection control policies and practices including use and availability of personal protective equipment, infection control signage, environmental cleaning, and visitation policies. This investigation was determined to be a public health response by LDH and therefore did not require institutional review board approval.

After the first patient was identified, the facility implemented enhanced surveillance for RSV. This included obtaining nasopharyngeal (NP) specimens from any NICU patient who had respiratory or other RSV-related symptoms, including cyanosis, apnea, or cough, as well as from asymptomatic infants on 2 (A and B) of 3 wards where the 7 patients resided. NP swabs were collected from all 7 patients within 2 days of symptom onset and immediately placed in commercial virus transport media and frozen at −70 °C.

Specimens from the 7 NICU patients were tested initially at the hospital laboratory using a commercial respiratory virus panel. These specimens and 23 RSV specimens collected during routine surveillance between October 2017 and January 2018 from patients located in 2 parishes of Louisiana—1 where the outbreak occurred and 1 in a parish neighboring the facility catchment area (we will refer to these routinely collected specimens as “community specimens”)—were then sent to the Centers for Disease Control and Prevention’s (CDC) Respiratory Viruses Branch laboratory for further characterization [14].

Hospital and all community specimens were tested using validated CDC molecular diagnostic assays to determine RSV subgroup [15]. Total nucleic acid was extracted from specimens using a magnetic silica-based platform (NucliSens easyMAG, bioMérieux, Durham, North Carolina). Nucleic acid extracts were then tested by an established pan-RSV real-time reverse-transcription polymerase chain reaction (rRT-PCR) assay [16] and a duplex rRT-PCR assay for RSV subgroup identification (A or B) [15]. WGS was performed on a subset using a 20-amplicon nested RT-PCR next-generation sequencing (NGS) method (details available upon request). In brief, RSV subgroup–specific primers were used to amplify the RSV genomes in overlapping amplicons. PCR products from each sample were pooled, purified, and used for library construction using Nextera XT DNA Sample Prep Kit (Illumina, San Diego, California). Paired-end sequencing was performed on the Ilumina MiSeq using 500-cycle MiSeq Reagent Nano Kit V2 (Illumina). A tailored NGS bioinformatics pipeline (vPipe) was used to perform read quality control and de novo assembly [17]. Sequences obtained from NICU patients and community specimens with the same RSV B genotype (including 2 specimens with RSV A co-detections) were aligned, and phylogenetic trees were computed using Bayesian analysis (Mr. Bayes version 3.2.6) and compared to 2 RSV B reference genomes (GenBank accession numbers KY924878 and KY249659). Additionally, signature nucleotides were identified by single-nucleotide polymorphism (SNP) analysis.

Characteristics of patients with RSV are shown in Table 1. Patient 1 was admitted to the NICU at 34 days of age for acute respiratory distress (day 0; Figure 1). This patient had first developed respiratory symptoms at home 2 days earlier (day −2). Patient 1 was born at the facility and had been discharged 30 days prior to readmission. RSV infection was confirmed by rRT-PCR at the facility’s laboratory upon admission on day 0. Between days 1 and 8, an additional 6 patients developed respiratory symptoms and were diagnosed with RSV at the facility laboratory (Table 1, Figure 2). On day 11, the facility notified LDH of the cluster.

No additional RSV infections were identified after the 7 patients originally reported. All the patients were born at the facility, and no patients were on respiratory support prior to onset of symptoms. Other than the first patient, no patients had ever left the NICU before onset of symptoms. Six patients (86%) were preterm with gestational ages at birth that ranged from 25 to 36 weeks (median gestational age, 34 weeks; Table 1). One patient had chronic lung disease. Median chronological age at symptom onset was 15 days (range, 7–147 days; Table 1). One patient was located on ward A, 5 on ward B, and 1 on ward C (Table 1, Figure 3). The 6 patients located on wards B and C were all located in adjacent single-family rooms (Figure 3).

Patients most commonly presented with congestion (n = 7 [100.0%]), cough (n = 4 [57.1%]), tachypnea (n = 4 [57.1%]), and poor feeding (n = 4 [57.1%]) (Table 2). Three patients had abnormal chest radiographs, including 2 with acute signs of lower respiratory disease: 1 with left lower lobe infiltrates, and 1 with bilateral infiltrates. Three patients required additional respiratory support after onset of symptoms (Table 1): 1 required bilevel positive airway pressure (patient 2), 1 required high-flow nasal cannula (patient 4), and 1 required continuous positive airway pressure and was intubated 5 days after symptom onset (patient 3). Four patients required intravenous fluids, and 1 received antibiotics. No patients received palivizumab either before or after becoming ill. Median length of hospitalization from symptom onset to hospital discharge was 9 days (range, 3–125 days; Table 1). All patients recovered and were discharged from the facility.

At the hospital laboratory, all patient specimens were identified as RSV B and were negative for influenza, coronavirus, parain-fluenza, and human metapneumovirus by rRT-PCR. Patient 2 had a co-detection of RSV with rhinovirus/enterovirus by rRT-PCR at the hospital laboratory (Table 1). CDC retested the 7 NICU patients and 22 community controls using RSV pan and duplex rRT-PCR assays. All 7 specimens from NICU patients were identified as RSV B. Of the 22 community specimens, 14 were identified as RSV B, 6 were identified as RSV A, and 2 were identified as A and B co-detections.

Except for 1 community specimen with very low viral load (cycle threshold value, 37.2), complete genome sequences were obtained from 28 specimens. Average coverage was 99 times to 171 times and genome coverage exceeded 99.1% for all samples tested. Nucleotide alignment of the obtained G gene sequences revealed that all RSV B strains were genotype BA and all RSV A strains were genotype ON1.

Phylogenetic analysis of the G gene alone segregated patient 1 and patients 2–7 into 2 clusters (Figure 4A), with patients 2–7 clustering with some community patients. Phylogenetic analysis of the RSV WGS further segregated NICU patients 2–7 and community patients into unique clusters (Figure 4B). NICU patients 2–6 were 100% identical, and patient 7 had 99.99% nucleotide identity by pairwise uncorrected comparison to patients 2–6. By WGS, patient 1 shared 99.5% nucleotide identity with patients 2–7, and the community specimens had 99.16%–100.00% nucleotide identity. In contrast, NICU and community specimens shared 97.60%–98.23% (KY924878) and 96.03%–96.97% (KY249658) nucleotide identity when compared to reference strains of RSV B.

SNP analysis identified 3 unique signature nucleotides in patients 2–7 that were absent in patient 1 and the community specimens; 2 unique SNPs were found in the M2–1 gene and 1 in the polymerase (L) gene. The unique SNPs and nucleotide agreement among viruses found in patients 2–7 suggest a single introduction of RSV into the NICU that resulted in infections among those patients, and distinguish that introduction from that of patient 1, which did not result in additional infections.

After identification of the first patient, the NICU implemented contact precautions, in addition to standard precautions, for all symptomatic infants. All NICU healthcare workers were asked to self-report any respiratory symptoms at the start of each shift. Additionally, all parents of NICU patients were sent a letter outlining guidelines issued during the winter respiratory season that included routine screening of visitors for fever and disallowing visitation of anyone with respiratory symptoms. LDH, in consultation with CDC, recommended additional precautions to limit transmission, including restriction of visitation from children < 12 years of age, use of surgical face masks during visitation by nonstaff, increased hand hygiene stations, enhanced environmental cleaning, and cohorting of staff. In the context of this cluster, droplet precautions were also recommended and implemented by the facility. Notices were placed outside the rooms of patients, as were signs describing appropriate contact and droplet precautions for visitors and staff.

The staff reported that the mother and siblings of the first patient (patient 1) had respiratory symptoms at the time of their NICU readmission. In addition, NICU staff observed that 2 patients were visited by family members with respiratory illnesses prior to the implementation of enhanced infection control measurements, but the specific timing of those visits relative to the patients’ onset of symptoms was not known. No NICU healthcare workers reported any symptoms and none were tested for RSV.