The human parechoviruses (HPeV) are a recently recognized genus belonging to the family Picornaviridae. They were originally classified within the enterovirus (EV) genus with HPeV types 1 and 2 known as echovirus 22 and echovirus 23, respectively. The reclassification arose from sequence analysis in the early 1990s demonstrating distinct genetic differences between echovirus 22 and EV genus members.1 To date, 16 HPeV types have been characterized (www.picornaviridae.com).2
The clinical and epidemiologic patterns for EV have been well studied, as they are common human pathogens. In the neonatal/infant population, specifically, EV presentations are commonly that of a nonspecific febrile illness for which a sepsis evaluation is often performed in infants <3 months of age.3 Less common but more severe presentations include frank encephalitis, hepatitis, sepsis syndrome and/or myocarditis.4
Clinical presentations of HPeV mostly appear similar to EV. HPeV1 has been the most commonly reported type. Both HPeV1 and HPeV2 have been associated with mild gastrointestinal and respiratory symptoms.5,6 HPeV3, however, has been associated with more severe clinical manifestation in the form of sepsis-like and central nervous system (CNS) illnesses particularly in neonates and infants.7–9 HPeV3 was initially isolated from a patient in Japan during 1999 and has been detected with increasing frequency over the last decade in Asia, Europe, Canada, South America and more recently in the United States.7–16 However, much remains to be discovered regarding the characteristics of HPeV3 infection as well as those of HPeV types 4–16.
We previously reported clusters of HPeV3 CNS infection in a Midwestern United States pediatric population during 2006 to 2008.7 Renaud et al demonstrated a peak in HPeV CNS disease in Seattle, WA, in 2009.16
In the current study, we describe the clinical, laboratory and epidemiologic characteristics of predominantly HPeV3 CNS infection during 2009. Additionally, we present comparisons between patients with HPeV3 CNS infection to those with EV and those with similar presentations but without detectable HPeV or EV (controls). Our aim was to identify characteristics that could assist clinicians in differentiating between HPeV, EV and those with neither virus in the cerebrospinal fluid (CSF).
MATERIALS AND METHODS
We included 388 CSF specimens from children in this study (range = 1 day to 18 years; median age = 45 days; average age = 769 days). These CSF specimens were previously submitted for EV real-time reverse-transcription polymerase chain reaction (RT-PCR) assay during June through October 2009 as part of standard of care for patients treated at Children’s Mercy Hospitals and Clinics (CMH).
Laboratory Testing for HPeV
All previously tested CSF specimens were assayed for HPeV. Total nucleic acids were extracted from CSF by using the EasyMag automated extractor (bioMerieux, Durham, NC), and aliquots were stored frozen at –80°C since 2009. The extracts were initially tested at CMH by a 2-step real-time RT-PCR as described by Benschop et al (2008), with modifications as previously reported.7,17 Total nucleic acid extracts from HPeV-positive CSF specimens were later forwarded to the Centers for Disease Control and Prevention (CDC) and tested by parechovirus 1-step real-time RT-PCR18 to confirm HPeV positivity. The viral protein 1 (VP1) region was then sequenced by a nested PCR assay to determine HPeV type.19 All sequences were deposited in the GenBank database, accession numbers JF919622-JF919672.
Complete HPeV3 VP1 sequences (678 nucleotides) from the CMH strains, other VP1 clinical sequences from the CDC sequence database and strains from GenBank were aligned using Clustal W.20 Phylogenetic relationships were reconstructed by the neighbor-joining method,21 using evolutionary distances computed by the Kimura-2 parameter method.22 Pairwise, percent nucleotide identities were calculated from the Clustal W alignment using MegAlign (DNAStar, v.8.1.2; Madison, WI).23
Identification of Study Groups
Despite testing CSF for HPeV from all patients ≤18 years of age, HPeV was detected only in patients <6 months of age; therefore, the comparator groups included all EV patients <6 months of age and 1 each age-matched patient with neither EV nor HPeV detected (controls). Patients who had CSF enterovirus PCR testing performed were thus grouped into 3 groups: 1) HPeV-positive group (n = 66), 2) EV-positive group (n = 47) and 3) control group (n = 66).
Clinical Characteristics of Infants Examined
We documented age and month of diagnosis, gender, clinical symptoms, laboratory results, chest radiograph results, use and duration of antimicrobials and length of hospitalization. Fever was defined as parent-reported (historical fever) or documented temperature (in emergency department or hospital records) of ≥38°C for patients ≤30 days of age and ≥38.3°C for those >30 days of age.7 Hypothermia was defined as a parent-reported or documented temperature <36.7°C.7 Length of hospitalization was assigned by the number of calendar days the patient spent in the hospital. Clinical symptoms were obtained either from caregiver report or hospital documentation by providers.
Meningitis was defined by age-specific white blood cell (WBC) count parameters. CSF pleocytosis was considered present with >28 WBCs in the first 30 days of life or >8 WBCs after 30 days of life.7 CSF glucose was considered abnormal if it was below threshold detection by assay (<20mg/dL) or <50% of concurrent serum glucose. CSF protein was considered elevated at >150mg/dL in neonates <10 days and at >58mg/dL in patients older than 10 days.7 Normal peripheral WBC count was defined as 5000 to 15,000/μL. Neutropenia was defined as an absolute neutrophil count (ANC) <500 neutrophils/μL and lymphopenia as lymphocyte counts <1500/μL. Normal ANC = 1500 to 8500/μL; normal absolute lymphocyte count (ALC) = 4000 to 10,500/μL. Normal platelet count was considered between 100,000 and 450,000 cells/μL with thrombocytopenia and thrombocytosis defined as less and greater than these values, respectively. The study was approved by the Institutional Review Board at CMH.
GraphPad InStat version 3.06 (GraphPad Software, Inc., La Jolla, CA) was used for statistical analysis. Mean and standard deviations were calculated for continuous variables. An unpaired t test was used for comparing 2-group continuous variables. This form of comparison was also utilized when comparing the 3 groups individually with one another. One-way analysis of variance was used when collectively comparing the 3 groups of continuous variables. The Tukey–Kramer multiple comparison was subsequently utilized for posttest analysis. A 2-sided Fisher exact test was used to compare categorical variables. SigmaStat (Systat Software, Inc., Chicago, IL) was used for multivariate stepwise analysis. A P value <0.05 was considered significant.
HPeV Prevalence and Genotype Analysis
HPeV was detected in 66 of the 388 (17%) CSF specimens tested by the 2-step HPeV real-time RT-PCR whereas EV was detected in 54 of the 388 (14%) from June through October 2009. The majority of HPeV-infected children were male (40/66). Overall 82% presented at <60 days of age and 100% <6 months of age. HPeV was detected from June to October with 52% of cases occurring during the month of August (Fig., Supplemental Digital Content 1, http://links.lww.com/INF/B373).
Total nucleic acids from the 66 HPeV-positive CSF specimens were later sent to the CDC and retested by a 1-step real-time RT-PCR, which detected parechovirus in 56 of the 66 (85%) specimens. Parechovirus genotyping was not attempted on the 10 total nucleic acid extracts that tested negative with the CDC 1-step RT-PCR test. Specimens positive by 2-step RT-PCR assay had lower Ct values (median 33.6; range 22.2–40) versus specimens positive by 2-step PCR only and negative by 1-step RT-PCR (median 38.6; range 35.6–40). Fifty-one specimens (91%; 51/56) positive by 1-step RT-PCR assay were successfully typed by sequencing the VP1 region and all of these were identified as HPeV3. Three of the CDC real-time RT-PCR positive specimens lacked sufficient volume to attempt genotyping. Two specimens failed sequencing reaction and had high real-time CT values (39.3 and 38.3), indicating low viral genome copy number. Overall, the HPeV3 VP1 sequences shared ≥93% nucleotide identity (NT ID). All of the HPeV3 viruses in our study were closely related to other viruses detected recently in the United States, with the exception of 1 virus that was clearly of Asian origin (Fig., Supplemental Digital Content 2, http://links.lww.com/INF/B374). CMH 2009 HPeV3 viruses and the US HPeV3 viruses from 2005 shared 95.6%–100% NT ID. With the exception of the CMH HPeV3 virus of Asian origin, the CMH HPeV3 viruses were 98.2%–100% identical to one another. The strain of Asian origin (USA-MO-09-11620) shared 99% NT ID with 2 2005 HPeV3 viruses from Thailand and only 94%–95.6% NT ID to the other CMH HPeV3 viruses. The father of this patient had recently traveled in Asia.
Characteristics of HPeV-positive, EV-positive and Control (Neg-HPeV/EV) Patients
There was no difference in gender predominance when the HPeV group was compared with EV and Neg-HPeV/EV groups (Table 1). However, the age of presentation for the EV group was younger than that of the HPeV group (P = 0.048). The frequency of the presenting symptoms of irritability and fever was not different between HPeV and the other groups. However, fever was of longer duration and the maximum in-hospital temperature was also higher in HPeV patients compared with both EV and Neg-HPeV/EV groups (P < 0.0001). On day of discharge, more HPeV patients (12/63; 19%) remained febrile than did the EV patients (5/47, 11%). One patient each in the HPeV and EV groups experienced seizure activity.
All EV-positive and Neg-HPeV/EV patients were hospitalized whereas 63 of the 66 HPeV-positive patients were hospitalized. While 8 HPeV-positive patients (12%) required care in the pediatric intensive care unit (PICU), only 1 EV-positive patient and no Neg-HPeV/EV patients required PICU care. Six HPeV-positive patients admitted to the PICU exhibited clinical signs of septic shock with decreased perfusion and hypotension requiring volume resuscitation. Of the remaining 2 HPeV PICU patients, 1 was admitted for severe neutropenia and the other for a generalized seizure.
CSF pleocytosis was rare (2%) in HPeV patients compared with EV (41%) and Neg-HPeV/EV patients (13%) (P < 0.001; Table 1). Additionally, the average CSF WBC count of HPeV patients was lower than that of EV patients (P < 0.001). There was no difference in CSF WBC counts between HPeV and Neg-HPeV/EV patients (P = 0.075). The CSF glucose was lower in EV compared with HPeV patients (P < 0.001) whereas the protein in EV was higher (P < 0.001). This significantly higher protein difference persisted even when accounting for age-specific values of normal. Herpes simplex virus PCR testing on CSF was ordered in 16 EV-positive patients, 18 HPeV-positive patients and 15 control patients. All were negative for herpes simplex virus.
Complete blood count results between the 3 groups were also different (Table 1). The average peripheral WBC count for HPeV patients was lower than that of both EV and Neg-HPeV/EV patients (P < 0.001). The ANC and ALC were correspondingly lower in HPeV than both other groups (Table 1). The lower mean absolute monocyte count of the HPeV group was not statistically significant (P = 0.40). There was no significant difference in C-reactive protein values among those tested.
There was no difference between antibiotic use or length of antimicrobial therapy between HPeV and the other 2 groups. Average duration of hospitalization was longer in HPeV patients compared with the EV and Neg-HPeV/EV groups (P = 0.005).
This study adds to mounting evidence that HPeV CNS infection is clinically relevant to neonatal and infant populations. We report an HPeV infection rate of 17% (66/388) in CSF originally submitted for EV testing from June through October 2009 in the Midwest United States. HPeV prevalence was as higher than EV has been for most recent years. All HPeV were detected in summer through early autumn.7 All typeable HPeV strains from CSF (n = 51) were HPeV3.
The world-wide distribution of HPeV3 is illustrated by the phylogenetic tree, using CMH and other selected complete VP1 gene sequences (Fig., Supplemental Digital Content 2, http://links.lww.com/INF/B374). All but 1 2009 CMH isolate clustered with recent US HPeV3 viruses (98.2%–100% identical), whereas European and Asian viruses formed separate clusters. We speculate that the CMH HPeV3 strain that was similar to Asian isolates was transmitted from father to child, given the father’s recent Asian travel and the virus’ genetic relationship (99% NT ID) to Thailand and Japan strains.
Our seasonality data confirm previous observations from across the globe including Europe8,11 and recently also from Seattle, WA.16 Despite indications of summer through fall seasonality, year-round testing is needed to ensure that HPeV does not also occur outside this interval. We observed little activity in 2006 (2% prevalence; 4/218 CSF) and 2008 (0% prevalence; 0/242 CSF) in contrast to 20077 and 2009. Similar intermittent HPeV activity was noted, but in even-numbered years, in the United Kingdom and the Netherlands.8,9 Thus, HPeV3 seems to cause periodic CNS infection outbreaks. Interestingly, our 17% CNS HPeV infection rates are identical in 20077 and 2009.
Males predominated among the combined 124 HPeV patients from 2007 and 2009, but males also predominated in our EV control group. Male gender does not differentiate HPeV from EV. The median age of the 2009 EV group (31.3 days) are at least 10 days younger when compared with our 2009 (40.8) and 2007 (46.2 days) HPeV groups. The clinical importance of this difference is unclear. Although CSF from all children ≤18 years of age was tested, HPeV-positive patients were mostly <60 days of age, and all <5 months old.
Abnormal CSF was rare in both our 2007 and 2009 HPeV groups, with pleocytosis in <10%. Low peripheral WBC count, ALC and ANC were noted in both years of HPeV groups. The normal CSF findings and low peripheral WBC and ALC distinguish the HPeV groups from the EV group. Our HPeV patient findings are similar to previous findings from studies conducted in United States and Netherlands.11,16 Our data do not provide a reason for this more intense fever with leukopenia.
HPeV patients had longer duration of fever with a higher maximum temperature during hospitalization. In both the EV and HPeV groups, there were likely some postdischarge febrile days that we did not capture. However, proportionally more HPeV patients (19%) were still febrile on discharge than EV patients (11%). Thus we likely would have missed more febrile days in the HPeV group than the EV group, resulting in underestimation of the duration of fever in HPeV-infected patients and supporting the idea that HPeV caused a longer duration of fever than EV. The 1 extra mean hospital day compared with the EV group may have been due to longer and higher maximum temperatures in HPeV patients. Clinicians likely were uncomfortable discharging highly febrile young infants, who unbeknown to them had HPeV infection (because results were unavailable during the hospitalization). In contrast, a laboratory-confirmed diagnosis of EV and the shorter fever duration in EV patients likely gave clinicians’ confidence to discharge EV patients sooner. Alternatively, HPeV patients may have been more ill-appearing than EV patients so hospitalizations were longer regardless of fever. This is supported by our observation that more HPeV children (8/63) appeared severely ill to warrant PICU admissions whereas only 1 EV-infected child (1/47) was admitted to PICU. Like the EV group and when compared with the HPeV group, the Neg HPeV/EV group had significantly shorter duration of fever and hospitalization, lower maximum temperature in the hospital and higher peripheral leukocyte values, including total WBC counts, ANCs and ALCs. The shorter hospital stay may again be due to the clinician greater level of confidence in discharging infants who by 48 hours of hospitalization were no longer febrile and who had negative sepsis workups.
Non-US studies show that HPeV CNS infection can be similar to EV infections.7,8,11 But in 1 Netherlands study,11 8 of the 11 neonates with HPeV infection from 1994 to 2006 showed mild-to-severe CNS white matter abnormalities in magnetic resonance imaging studies. Three of 8 had neurodevelopmental delay at follow-up. In our study, only 6 of the 66 of our HPeV patients had neuroimaging studies and all were normal. Normal findings in neurologic examinations and no CSF pleocytosis likely made CMH providers feel that further CNS imaging was unwarranted. We did not perform developmental follow-up. This may be of low yield given none of our patients had abnormal discharge neurologic examinations.
We detected no dual CNS infections (HPeV plus other viral or bacterial agents). No HPeV patient had a bacterial bloodstream infection. However, 2 HPeV-positive children had concomitant urinary tract infections (1 Staphylococcus aureus and 1 Escherichia coli infection), but these were evident within 24 hours of admission.
Limitations to our study include its retrospective nature. Chart documentation may not be as complete as data from prospective studies. Additionally, testing CSF originally sent for testing in EV season potentially precludes us from identifying HPeV with other clinical presentations or outside the typical EV season. Although fever and irritability was the most common presenting symptom noted in the current and prior HPeV patient groups, it may not be a surprising finding given that these symptoms without a known focus of infection generally provoke a sepsis workup, including CSF testing. In addition, because EV types were not identified in this study, variable clinical presentations could be expected in other years due to variability in EV types circulating from year-to-year.
In conclusion, HPeV CNS infection can be the sole cause of sepsis-like syndrome in young infants. It should be considered particularly when initial laboratory data show a normal CSF without pleocytosis but low peripheral WBC, ALC and/or ANC counts. Hospitalized HPeV patients appear to have an extra febrile day compared with EV patients or those with no proven CSF viral pathogen. Multiple days with in-hospital temperatures >39°C were characteristic of HPeV but not EV infections.
Based on this and our prior 2007 study, we suggest potential benefits to routine HPeV RT-PCR on CSF of <6-month old infants undergoing a sepsis workup during summer/fall in the United States. Clinicians could decide on earlier discharge for patients with confirmed HPeV. CMH now offers routine HPeV RT-PCR. We plan to prospectively measure the impact of this HPeV testing on patient management when there are sufficient numbers of HPeV cases for analysis.
1. Hyypiä T, Horsnell C, Maaronen M, et al. A distinct picornavirus group identified by sequence analysis. Proc Natl Acad Sci USA. 1992;89:8847–8851
2. Knowles NJ, Hovi T, Hyypiä T, et al.King AMQ, Adams MJ, Carstens EB, Lefkowitz EJ Picornaviridae. In: Virus Taxonomy: Classification and Nomenclature of Viruses: Ninth Report of the International Committee on Taxonomy of Viruses. 2011 San Diego Elsevier:855–880
3. Baraff LJ. Management of fever without source in infants and children. Ann Emerg Med. 2000;36:602–614
4. Abzug MJ. Presentation, diagnosis, and management of enterovirus infections in neonates. Paediatr Drugs. 2004;6:1–10
5. Benschop K, Thomas X, Serpenti C, et al. High prevalence of human Parechovirus (HPeV) genotypes in the Amsterdam region and identification of specific HPeV variants by direct genotyping of stool samples. J Clin Microbiol. 2008;46:3965–3970
6. Harvala H, Robertson I, McWilliam Leitch EC, et al. Epidemiology and clinical associations of human parechovirus respiratory infections. J Clin Microbiol. 2008;46:3446–3453
7. Selvarangan R, Nzabi M, Selvaraju SB, et al. Human parechovirus 3 causing sepsis-like illness in children from midwestern United States. Pediatr Infect Dis J. 2011;30:238–242
8. Harvala H, Robertson I, Chieochansin T, et al. Specific association of human parechovirus type 3 with sepsis and fever in young infants, as identified by direct typing of cerebrospinal fluid
samples. J Infect Dis. 2009;199:1753–1760
9. Benschop KS, Schinkel J, Minnaar RP, et al. Human parechovirus infections in Dutch children and the association between serotype and disease severity. Clin Infect Dis. 2006;42:204–210
10. Boivin G, Abed Y, Boucher FD. Human parechovirus 3 and neonatal infections. Emerging Infect Dis. 2005;11:103–105
11. Verboon-Maciolek MA, Krediet TG, Gerards LJ, et al. Severe neonatal parechovirus infection and similarity with enterovirus infection. Pediatr Infect Dis J. 2008;27:241–245
12. Yamamoto M, Abe K, Kuniyori K, et al. Epidemic of human parechovirus type 3 in Hiroshima city, Japan in 2008. Jpn J Infect Dis. 2009;62:244–245
13. Sedmak G, Nix WA, Jentzen J, et al. Infant deaths associated with human parechovirus infection in Wisconsin. Clin Infect Dis. 2010;50:357–361
14. Drexler JF, Grywna K, Stöcker A, et al. Novel human parechovirus from Brazil. Emerging Infect Dis. 2009;15:310–313
15. Ito M, Yamashita T, Tsuzuki H, et al. Isolation and identification of a novel human parechovirus. J Gen Virol. 2004;85(Pt 2):391–398
16. Renaud C, Kuypers J, Ficken E, et al. Introduction of a novel parechovirus RT-PCR clinical test in a regional medical center. J Clin Virol. 2011;51:50–53
17. Benschop K, Thomas X, Serpenti C, et al. High prevalence of human Parechovirus (HPeV) genotypes in the Amsterdam region and identification of specific HPeV variants by direct genotyping of stool samples. J Clin Microbiol. 2008;46:3965–3970
18. Nix WA, Maher K, Johansson ES, et al. Detection of all known parechoviruses by real-time PCR. J Clin Microbiol. 2008;46:2519–2524
19. Nix WA, Maher K, Pallansch MA, et al. Parechovirus typing in clinical specimens by nested or semi-nested PCR coupled with sequencing. J Clin Virol. 2010;48:202–207
20. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994;22:4673–4680
21. Saitou N, Nei M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol. 1987;4:406–425
22. Kimura M. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol. 1980;16:111–120
23. Tamura K, Dudley J, Nei M, et al. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol. 2007;24:1596–1599