Enterovirus serotypes in patients with central nervous system and respiratory infections in Viet Nam 1997–2010

Data and samples from 5 prospective observational clinical studies on CNS (n = 3) and respiratory (n = 2) infections, conducted in Viet Nam between 1997 and 2010 were used [14‐16, 18, 19]. Patients were enrolled according to study-specific case definitions for CNS infection, acute respiratory infection and lower respiratory tract infection. The three CNS studies were conducted to determine the etiology of CNS infections in children in Children’s Hospital 1 in Ho Chi Minh City [HCMC] (2004), and in adults in the Hospital for Tropical Diseases (1997–2008) and in a network of 12 provincial hospitals in the southern and central part of Viet Nam (2007–2010). In these three studies, CSF was obtained from all participants, while throat and rectal swabs were only taken from children. The two respiratory infections studies were conducted to study lower respiratory tract infection in hospitalized children under 2 years of age in two main paediatric hospitals in HCMC (2009–2010) and the antibiotic use in out-patients with acute respiratory infections in Children’s Hospital 1 in HCMC (2009–2010). Respiratory swabs were taken from all enrolled children. Patients positive for generic EV RT-PCRs in CSF or swabs from these studies and with diagnostic specimens available were included in this study. The geographical distribution of included patients is shown in Additional file 1: Figure S1.

Viral RNA was extracted from clinical samples using either the MagNA Pure 96 platform (Roche Applied Science, Darmstadt, Germany) or the QIAamp Viral RNA Mini kit (QIAgen GmbH, Hilden, Germany) following the manufacturer’s instructions.

EV-A71 detection and typing of samples from patients with CNS infection was performed using real-time RT-PCR as described previously [12]. In brief, 2 μl of viral RNA were subjected to one-step real-time RT-PCR reaction using the SuperScript III One-Step qRT-PCR system with Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA, USA). The cycling conditions included one cycle of 60 °C for 3 min, followed by 15 min at 53 °C and 2 min at 95 °C, and 45 cycles of 15 s at 95 °C, 1 min at 53 °C (including data acquisition) and 15 s at 72 °C.

Nested RT-PCR for viral protein 1 (VP1) and complete viral protein 4 and partial viral protein 2 (VP4/VP2) was carried out using previously described assays [20‐22]. VP1 RT-PCR of samples from patients with CNS infection was primarily performed using the assay described by Leitch et al., [22] and alternatively, the VP1 assay described by Oberste et al., [20]. The VP4/VP2 assay described by Mirand et al., [21] was applied to all samples from patients with respiratory infection and to samples from patients with CNS infection with negative VP1 RT-PCRs. There was no major modification of the original assays except the use of Superscript III reverse transcriptase with Platinum Taq (Invitrogen) in the first round RT-PCR and Hotstar Taq DNA polymerase (QIAgen) in the second round PCR and the adaptation of thermal cycling conditions as per supplier’s instructions. In the three assays, 5 μl of viral RNA were used in the first round one-step RT-PCR and then 0.5 μl of the first round PCR product was transferred to the second round PCR reaction.

PCR products of the 2nd round PCR were analysed in agarose gel, purified with cold ethanol and were then subjected to DNA sequencing using the Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems Inc., Foster City, CA, USA).

Nucleotide sequences of the VP1 and VP4/VP2 were assembled using Vector NTI ® Express software v7.1 (Thermo Fisher Scientific, Waltham, USA). All the sequences have been submitted to GenBank (MH021887-MH021955). Multiple sequence alignments were performed using BioEdit software v7.0.9 (Ibis Therapeutics, CA, USA) and with the inclusion of reference prototype sequences obtained from GenBank. Neighbour-joining trees were constructed in MEGA software v7.0.26 (www.​megasoftware.​net) using the maximum composite likelihood nucleotide substitution model with 1000 bootstrap replicates.

Serotype assignments for VP1 and VP4/VP2 sequences were performed using an automated phylogenetic-based enterovirus typing tool developed by Kroneman et al., [23] available at http://​www.​rivm.​nl/​mpf/​enterovirus/​typingtool/​. Because the automated typing tool can determine VP4/VP2 sequences at species level only [23], further serotype determination for VP4/VP2 sequences was based on highest nucleotide identity score (HNIS) and highest amino acid sequence similarity (HAASS) with reference prototypes [24, 25]. In case of discrepancy, final assignment was based on HAASS and further confirmed using BLAST [21, 26].

Samples from a total of 203 patients were included in this study, including from 79 patients with CNS infection and 124 with respiratory infection. When analyzing for the monthly distribution of cases, there was no clear peak among CNS cases, whereas two peaks of respiratory cases were found in April and November (Fig. 1).

Among patients with CNS infection (n = 79), 51 were children (median age 2 years, interquartile range [IRQ]: 1–5) and 28 were adults (20 years, IRQ: 17–29.5). All adults had EV detected in CSF and 93% (26/28) had meningitis as discharge diagnosis. Of the 51 children, 10 had EV detected in CSF, and the remaining (41/51) had EV detected in respiratory and/or rectal swabs. Encephalitis was the most common discharge diagnosis (30/41). Fifteen deaths were noted, and all were children with EV detected in swabs only (Table 1).

Among patients with respiratory infections, the median age was 12.7 months (IQR: 5.7–24.3). Inpatients (n = 65) were younger than outpatients (n = 59), due to study enrolment criteria (Table 2). Bronchiolitis (as assessed at the discretion of treating physicians) was the most common clinical diagnosis (57% in outpatients and 61% in inpatients). Thirty-five percent of inpatients had a clinical diagnosis of pneumonia (Table 2).

In patients with CNS infection, co-infection was more common and was found in 27% (14/51) children versus 11% (3/28) adults (Table 1). Most co-infected children (13/14) had EV detected in swabs only. Of the 15 documented fatal cases, 7 were serologically confirmed to have co-infection with dengue virus (DENV; n = 3) and Japanese encephalitis virus (JEV; n = 4) (Additional file 4: Table S1). Among the 3 co-infected adults, 2 had cytomegalovirus (CMV) DNA detected in CSF and 1 had positive CSF culture with Mycobacterium tuberculosis (Table 1).

In both patient groups of respiratory infection, there were high rates of co-infection with other respiratory viruses. Co-infection with rhinovirus (RHV; 69%) and respiratory syncytial virus (RSV; 27%) was more common than other viruses, including adenovirus (ADV; 14%), parainfluenza viruses (PIV; 9%), human metapneumovirus (hMPV; 9%), human bocavirus (BoV; 7%), human coronavirus (HCoV; 4%), influenza virus (Flu; 2%), and parechovirus (PEV; 1%). The rate of co-infection with RHV and RSV was significantly higher in the inpatient group as compared to the outpatient group (P < 0.001) (Table 2).

Enterovirus serotype determination was successful in 38% (77/203) of patients: 19/38 (50%) CSFs (6 children, 13 adults) and 27/41 (66%) swabs from patients with CNS infections, and 31/124 (25%) of swabs from patients with respiratory infections. A total of 26 different serotypes belonging to four enterovirus species (A-D) were identified.

Twenty-one different serotypes belonging to 3 enterovirus species: EV A (n = 12, 3 serotypes), B (n = 30, 16 serotypes), and C (n = 4, 2 serotypes) were identified in patients with CNS infection. The identification of these serotypes was achieved by EV-A71 specific RT-PCR (n = 8; Additional file 4: Table S2), VP1 (n = 22; Additional file 2: Figure S2) and VP4/VP2 (n = 16; Additional file 3: Figure S3) sequencing. EV B was the only species detected in adults and the most common species in both swabs and CSF from children (Table 3). EV-A71 was the most commonly detected single serotype, in 9 swabs and 1 CSF of 10 children. Three of these children had a clinical diagnosis of HFMD while the other 7 children had a clinical diagnosis of encephalitis (Additional file 4: Table S2). There was no specific serotype associated with fatal outcome (Additional file 4: Table S1).

Among patients with respiratory disease 13 different serotypes of four enterovirus species were identified by VP4/VP2 sequence analysis including A (n = 2, 2 serotypes), B (n = 11, 7 serotypes), C (n = 10, 3 serotypes), and D (n = 8, 1 serotype). CV-A24 (n = 8) and EV-D68 (n = 8) were the two most common serotypes detected (Table 3) and these viruses had a close genetic relationship in phylogenetic tree analyses (Additional file 3: Figure S3).