Between June 1 and November 1, 2014, 40 infants at CMKC had PeV-positive CSF. Virus typing confirmed that 35 of these 40 infants were positive for PeV3, and 1 infant was positive for PeV4. Among the 4 remaining infants, CSF from 3 infants tested negative by VP1 RT-PCR, and CSF from 1 infant tested PeV-positive by RT-PCR but no residual specimen was available for sequencing.

Epidemiologic and clinical features of the 35 case-patients with PeV3 infection are listed in Table 1; all of these infants were hospitalized. Overall, 51% (18 of 35) were male, and the median age at illness onset was 24 days (range, 5–56 days). Common symptoms included fever (>38.0°C, 91% [32 of 35]), tachycardia (77% [27 of 35]), poor feeding (77% [27 of 35]), irritability (74% [26 of 35]), mottled skin (37% [13 of 35]), and rash (26% [9 of 34]); rash descriptions varied and were frequently nonspecific but included macules and papules both centrally and on extremities. In addition, 20% (7 of 35) of the infants had neurologic manifestations, including eye deviation and limb stiffening (hereafter referred to as seizures), that were consistent with partial or generalized seizures. Peripheral blood and CSF cell counts were typical of those previously described for patients with PeV infection [11, 12, 22]; peripheral white blood cell counts were not elevated, and a lack of CSF pleocytosis was found. The median length of stay was 3 days (range, 1–18 days). No deaths were reported.

The lone PeV4-positive infant experienced fever and difficulty breathing but did not require intensive care and was discharged home after 3 days. The infant was 6 days old at illness onset.

Among the 35 infants hospitalized with PeV3 infection, 8 required oxygen and 7 were admitted to the ICU. A comparison of infants admitted to the ICU with those who were not is shown in Table 1. All infants admitted to the ICU were aged <3 weeks (median, 11 days; range, 5–18 days) and were significantly younger than infants who did not require ICU admission (median, 27 days; range, 5–56 days; P = .0044). Tree of the 7 infants who required ICU admission experienced difficulty breathing, and 2 of 7 had a temperature of <35°C. It should be noted that 6 of the 7 infants who required intensive care experienced seizures (as defined earlier); all 6 of them were treated with acyclovir, and CSF samples from all of them subsequently tested negative for herpes simplex virus. Magnetic resonance imaging was performed for 3 of these 6 infants, and the results were abnormal for all of them, with white matter changes noted. Electroencephalography was performed for 2 of the 6 infants; for 1 infant, cephalopathic changes consistent with severe encephalopathy were found, and for the other infant, multifocal epileptiform discharges were seen. The median length of stay in the ICU was 4 days (range, 1–17 days). Two of the 7 infants who required ICU were discharged on anticonvulsant medication.

Complete PeV3 VP1 sequences were obtained from 35 patient CSF specimens. Phylogenetic analyses of these sequences with neighbor-joining and maximum-likelihood algorithms (Figure 1) gave similar results and showed 2 PeV3 lineages cocirculating during the investigation period. The largest PeV3 lineage (found in 30 patients) shared 97.9% to 100% nucleotide identity (NT ID) (99.5%–100% amino acid identity [AA ID]), and the smaller PeV3 lineage (5 patients) shared 99.8% to 100% NT ID (100% AA ID). Genetic distances between the 2 lineages, using the 35 VP1 sequences, ranged from 94% to 95% NT ID (98.5%–99% AA ID). Case-patients who required ICU admission were infected with either lineage, as were case-patients who experienced seizures. All PeV3 VP1 sequences acquired for this study were submitted to Gen Bank and assigned accession numbers KX356610 through KX356655.

Stool specimens (n = 18) were available from 17 PeV3 case-patients after discharge from the hospital, and 8 of these specimens tested positive for PeV3 (Figure 1); all CSF–stool pairs from individual patients exhibited ≥98.8% NT ID. Most infants provided a single stool specimen, and time between disease onset and the specimen-collection date varied among the infants. All specimens collected <25 days after onset (n = 7) tested positive for PeV3. In contrast, only 1 of 11 specimens collected beyond 25 days after disease onset (collected on day 43) tested positive for PeV3; the 10 PeV3-negative specimens were collected between 27 and 73 days after disease onset. One infant provided 2 specimens, 1 on day 24 after disease onset, which tested positive for PeV3, and 1 on day 44 after onset, which tested negative.

To understand the characteristics of infants with PeV3 infection and their families, we performed birth chart reviews and parent interviews (Table 2). One of the 35 case-patients with PeV3 infection was born and resided in Texas and was excluded. Among the remaining 34 case-patients, 26 met our criteria for birth chart review, including 22 case-patients whose infection status was detected before August 19, 2014, and 4 case-patients whose infection status was detected after this date and were aged <2 weeks at onset. Twenty-two (85%) of these 26 birth chart reviews were completed. Parents of 23 (68%) of the 34 case-patients agreed to be interviewed; the median time between illness onset and the interview date was 24 days (range, 1–74 days). Both birth chart reviews and interviews were available for 16 cases.

Birth chart reviews to explore mother and child health at delivery were unrevealing; for example, most births were uncomplicated, case-patients were not born prematurely, and the mothers were not reported as being febrile during delivery or in the week before. Parent interviews revealed that almost half of the case-patients (11 of 23) had a household contact who was ill during the week before disease onset. Ill contacts (n = 20) typically experienced symptoms representative of mild respiratory (n = 15) or gastrointestinal (n = 5) illness or had a rash (n = 2); the specific illness of 1 contact was not described. Most (21 of 23) of the infants had a household contact aged <11 years; the median age of these contacts was 3 years (range, 13 months to 10 years). Ten case-patients had a household contact aged <11 years who was ill during the week before onset. Stool specimens were available from 7 contacts of 6 case-patients. Siblings of 2 case-patients had PeV3-positive stool, and both case-patient (CSF)–sibling (stool) pairs exhibited ≥99.8% NT ID (Figure 1); 1 sibling was reported to have developed a rash after the case-patient was discharged from CMKC, and the sibling of the other case-patient reported no symptoms. For 4 of the case-patients with PeV3 infection, a family history of neurological issues was reported, including history of seizures, attention-deficit hyperactivity disorder, and multiple sclerosis; 1 of these 4 patients experienced seizures during the hospitalization for PeV3 infection.

When we compared PeV3-positive case-patients aged ≤7 days (n = 5; listed in Table 3) with older case-patients, no apparent differences were observed in the data from the birth chart reviews or parent interviews, apart from the infant's birth hospital. Four (80%) of the 5 infants whose illness onset occurred in their first week of life were born at the same hospital (hospital A) within a period of 34 days. In contrast, only 2 (10%) of 20 case-patients who were >7 days old at illness onset were delivered at hospital A. Viruses from the 4 infants delivered at hospital A were part of the larger lineage and exhibited ≥99.7% NT ID.

We provide here a clinical and epidemiologic description of PeV3 infection in young infants and, by typing all available specimens, have provided an in-depth type-specific disease characterization. We show that PeV3 can cause severe neurologic illness in neonates and, in some cases, associated white matter changes. All infants who required intensive care became ill during their first 3 weeks of life and were significantly younger than those who were not admitted to the ICU.

Our clinical findings are consistent with data from previous PeV investigations. We found that common symptoms during PeV3 infection include fever [7, 9, 11, 12], tachycardia [11], poor feeding [7], irritability [7, 9, 11, 12], mottled skin [23], and rash [9]. Severe PeV3 infections that required intensive care management were more common in younger infants (aged <3 weeks) in our investigation, as previously reported for untyped PeV infections [7]. We also identified seizures [3, 11, 24] associated with white matter changes as an important presentation, which was observed in 7 of 34 infants (6 of 7 in the ICU). As expected, all infants in our investigation who presented with seizures were evaluated for HSV infection and empirically treated with intravenous acyclovir. At CMKC, acyclovir treatment was stopped once the infant tested negative for HSV and positive for PeV. By comparison, CMKC detected only 1 infant (aged <30 days) with central nervous system HSV infection during the investigation period. A wider availability of diagnostics for PeV could lead to more judicious use of acyclovir in this very young and vulnerable population. Furthermore, infection outcomes vary substantially among patients with HSV and those with PeV. Unlike the potentially fatal neurologic complications of HSV infection in infants [25], most infants in our investigation who were infected with PeV3 were discharged from the hospital in healthy condition after several days. Although longer-term neurologic sequelae after PeV infection have been observed [7, 26], we did not assess long-term outcomes in this investigation. In addition, all case-patients in our investigation survived, although PeV3 infection has been associated with rare fatalities [27, 28].

Consistent with previous reports that suggested that older siblings might pose a risk for PeV exposure [29], results of our parent interviews indicated that almost half of all case-patients had ill household contacts aged <11 years. However, data from households without recognized PeV3 infection were not available for comparison. Results of previous retrospective studies have suggested that PeVs are a common source of illness in the community; PeV has been detected in 3% to 17% of CSF specimens from hospitalized infants [12, 22, 30, 31], 13% of blood specimens from infants who required intensive care for late-onset sepsis [32], 3.7% of plasma specimens from children who presented to an emergency department [33], and 16.3% of stool specimens from children who presented to a medical center [10]. Despite these high estimates, PeVs often go undetected because of mild or asymptomatic presentations [2, 34] combined with limited application of available diagnostic tests. Although data on PeV3 circulation are limited, our cases occurred mostly within the proposed PeV3 season (late summer and fall) [9, 18].

A subset of case-patients who became ill in their first week of life were of particular concern, because infection in this age group had been reported only rarely before this investigation. To explore the possibility of perinatal infection, maternal demographics and health at delivery were reviewed closely using labor and delivery charts, but this review was unrevealing. Maternal age has been suggested to be associated with severity of illness [35], but we did not detect such an association.

We noted that 4 of 5 infants aged ≤7 days at illness onset were delivered at hospital A within a similar time period. Although it was not possible to distinguish between infections acquired in the healthcare setting versus those acquired in the community, this finding supports the recently reported possibility of healthcare-associated acquisition of PeV3 infection [13]. By comparison, it has been known for some years that PeV1 (previously named echovirus 22) can be transmitted within healthcare settings and can cause respiratory [36] or gastrointestinal [37] illness.

PeVs have been detected in a variety of specimens, including those from respiratory secretions [38] and stool [2, 10, 39]. We found that the stool of PeV3 case-patients could test positive 43 days after illness onset. By comparison, in infants followed longitudinally in previous investigations, the median duration of shedding was 51 [2] to 58 [39] days. Prolonged shedding after symptom resolution highlights the need for constant vigilance in infection-prevention and -control practices.

Phylogenetic clustering of the PeV3 VP1 sequences indicated cocirculation of 2 PeV3 strains, which differed by ≥5% NT ID, among patients treated at CMKC. Cocirculation of different PeV3 lineages had not been observed in PeV3 outbreaks previously, but it is common in large outbreaks of enterovirus disease [40]. This observation in a local PeV3 outbreak might be attributable to the large catchment area of CMKC. Severe illness was observed in both PeV3 clusters.

Our investigation was subject to several limitations. First, PeV testing is specialized, and its availability is limited. By directing testing toward CSF specimens from hospitalized infants aged <6 months, CMKC focuses on young infants and the more severe end of the clinical spectrum; therefore, the generalizability of these results is unknown. Second, parent interviews might have been subject to recall issues, especially for the families of infants who became ill earlier in the season. Last, we were not able to sequence 5 of the 40 original CSF specimens collected; the 5 infants from whom the specimens were not collected were aged ≥18 days, but no other differences were noted.

PeV3 should be considered alongside HSV, enteroviruses, and other pathogens as a cause of sepsis-like illness and meningitis in young infants. In addition to informing clinical management, expanded testing for PeVs might allow for a better understanding of virus circulation and disease burden.