Prospective Study of Human Metapneumovirus Infection in Children Less Than 3 Years of Age

Abstract

Most lower respiratory tract infections (LRTIs) in children under the age of 3 years are due to respiratory syncytial virus (RSV). Epidemiological, host, and viral factors eventually account for the severity of LRTIs, but they do not completely explain it. Human metapneumovirus (hMPV) was recently identified in children with LRTIs. In a population-based prospective multicenter study (the PRI.DE study, conducted in Germany over 2 years), we tested 3,369 nasopharyngeal secretions from children younger than 3 years of age with LRTIs for RSV A and B, influenza viruses (IVs) A and B, and parainfluenza viruses (PIVs) 1 to 3. Of the children requiring intensive care (n = 85), 18% had hMPV infections, and 60% of these children were infected with hMPV in combination with RSV. We did not detect hMPV in a randomly selected subset of RSV-positive nasopharyngeal secretions (n = 120) from children not requiring intensive care support. hMPV was detected in <1% of virus-negative samples from patients without intensive care support (n = 620). Our data support the hypothesis that coinfections with RSV and hMPV are more severe than infections with either RSV or hMPV alone, at least in children younger than 3 years of age.

It is well accepted that the respiratory syncytial virus (RSV) is the most prevalent virus in patients up to 3 years of age with lower respiratory tract infections (LRTIs) (8, 18). Epidemiological, host, and viral factors account for the severity; but they do not completely explain it. Furthermore, precise data about the relative contributions of different pathogens to the extent of illness in children are not available. Problems encountered with the detection of respiratory viruses include the capture of a specimen for viral diagnostics and limitations caused by the transport of labile viruses.

In 2001, a new paramyxovirus associated with respiratory illness, the human metapneumovirus (hMPV), was discovered in The Netherlands (27). Data from that report suggest that hMPV is similar to RSV, in that hMPV infection usually occurs during the winter months, is common during childhood, and causes diseases with symptoms ranging from those that are mild to those of bronchiolitis and pneumonia. After its initial discovery in The Netherlands, hMPV has been identified around the globe (2, 6, 9, 10, 13, 19, 20, 23, 26).

Our intent was to test in a prospective German multicenter study (12) the extent to which hMPV infection contributes to the severity of illness in RSV-infected and RSV-non-infected children younger than age 3 years.

MATERIALS AND METHODS

Subjects.

The population-based prospective German multicenter study (the PRI.DE study) was performed in four different regions of Germany (north, Hamburg; west, Bochum; south, Freiburg; east, Dresden) covering inpatients (from university hospitals and community hospitals) and outpatients (from pediatric practices). The study ran from November 1999 to October 2001. Children up to 3 years of age with pneumonia, bronchitis, bronchiolitis, and croup and children up to 6 months of age with apnea were included (12). Demographic data, data on clinical signs and severity of disease, and the treatment regimen were recorded. Informed consent was obtained from the parents of the patients, and human experimentation guidelines for the conduct of clinical research were followed.

Materials.

Nasopharyngeal secretions (1 to 2 ml) were taken from the patients the first time that they were seen by a physician. According to a standard protocol, the samples were added to a standardized viral transport medium, immediately frozen in liquid nitrogen, and transported on dry ice to the Department of Virology (Ruhr-University Bochum, Bochum, Germany) for viral diagnostics.

Laboratory testing.

RSVs A and B, influenza viruses (IVs) A and B, and parainfluenza viruses (PIVs) 1 to 3 were detected by Hexaplex PCR, as described by the manufacturer (Prodesse, Milwaukee, Wis.). From November 1999 to October 2000, in parallel with the hexaplex PCR, the nasopharyngeal secretions were tested for RSVs, IVs, and PIVs by culture on three different cell lines (Vero, HEp-2, and LLC-MK2 cells). After 6 to 10 days of incubation the cells were studied by immunofluorescence. The culture results were in good accordance with the PCR results, thus allowing the use of PCR as a single test in the second year of the study (data not shown). Testing for hMPV was done retrospectively, after the enrollments had ended (spring 2003), with samples that were stored in liquid nitrogen and extracted immediately before analysis for hMPV. Specimens from both years of the study were tested for hMPV by reverse transcription-PCR (RT-PCR), as described below.

RNA extraction.

Viral RNA was extracted from 200 μl of nasopharyngeal aspirate specimens with a QIAamp viral RNA Mini kit (Qiagen, Hilden, Germany) or a viral RNA robot kit (Qiagen).

hMPV RT-PCR.

Complementary cDNA was synthesized by using 10 μl of eluted RNA and a One-Step RT-PCR kit (Qiagen). The PCR assay was carried out according to the instructions of the manufacturer. New PCR primers were designed for amplification of a part of the hMPV nucleoprotein (N) gene (Primer 3 output; http://www-genome.wi.mit.edu). The sequence of forward primer hMPV-N1f was 5′-TCTACAGGCAGCAAAGCAGA-3′, and that of reverse primer HMPV-N1r was 5-TTTGGGCTTTGCCTTAAATG-3′. The primer positions are 742 to 761 and 946 to 965, respectively (GenBank accession number AF371337). The primers amplified a 224-bp region of the N gene. Amplification conditions consisted of 30 min at 50°C, 5 min at 94°C, 55 cycles of PCR for 15 s at 94°C and 1 min at 65°C, and a final extension step at 72°C for 10 min. Each RNA sample was run with a housekeeping gene to verify RNA integrity. To exclude the possibility of contamination, negative and positive samples were always run in parallel. Each PCR assay could detect at least 50 copies of the viral target, as assessed by using plasmid clones with the respective hMPV amplicon cloned into the pCR2.1 TOPO vector (TopoTA cloning kit; Invitrogen, Carlsbad, Calif.).

hMPV isolation.

Nasopharyngeal samples positive for hMPV by RT-PCR were inoculated on rhesus monkey kidney (LLC-MK2) cells in 24-well plates and were incubated for 3 weeks. The medium was renewed 3 days, and the culture supernatants were removed and tested for hMPV by RT-PCR.

RESULTS

A total of 2,386 outpatients were eligible during both years of the multicenter study of acute LRTIs; of these, informed consent was obtained for 1,487 (62.3%) outpatients, who took part in the study. The respective values for inpatients were 2,924 and 2,068 (70.7%). The main diagnoses for the inpatients were bronchiolitis (n = 1,232; 60.3%) and pneumonia (n = 726; 35.1%), whereas outpatients consulted a pediatrician mainly due to bronchiolitis (n = 777; 52.3%), bronchitis (n = 573; 38.5%), and croup (n = 201; 13.5%). Among the 3,555 pediatric outpatients and inpatients for whom informed consent was obtained, 3,369 specimens were sufficient for complete analysis. RSV A was found in 916 (27.2%) of the specimens and RSV-B was found in 206 (6.1%). PIVs 1 to 3 were detected in 360 (10.7%) of the patient samples, and IVs were detected in 133 (3.9%) of the patient samples. Fifty-six (1.7%) of the samples were found to be positive for more than one virus. Thus, after the first diagnostic approach, 53.6% (n = 1,807) of the children with LRTIs remained negative for the viruses under study.

First reports of hMPV infection in children indicated that the clinical symptoms were largely similar to those of the respiratory tract illness caused by human RSV (27). Therefore, we primarily focused on nasopharyngeal aspirate samples negative for the common respiratory viruses. We analyzed the RSV-, IV-, and PIV-negative nasopharyngeal secretions obtained from October 2000 to April 2001 of the second study year. A total of 620 samples were analyzed for the presence of hMPV by RT-PCR. Only two samples, one obtained on 27 November 2000 from an outpatient in Hamburg and one obtained on 18 December 2000 from an inpatient in Hamburg, were positive for hMPV. Viral growth was subsequently confirmed in cultures of both samples. Both children had only mild respiratory disease.

It became apparent quite recently that hMPV and RSV coinfections occur (14). To actually test the hypothesis that infections with RSV and hMPV together might be more severe than infection with either RSV or hMPV alone, we focused on children who needed intensive care support. In our prospective study, a total of 85 children from both study years were admitted to the intensive care units of the respective centers (Table 1). We found RSV A in 24 patients, RSV B in 7 patients, PIV 3 in 3 patients, PIV 1 in 1 patient, IV in 2 patients, and hMPV in 15 patients. The remainder of the samples were negative for all viruses. Children proved to be positive for hMPV, with a total prevalence of 18%. Among the 15 samples of nasopharyngeal secretions positive for hMPV, 9 were also positive for RSV (60%), 1 was positive for PIV 1, 1 was positive for PIV 3, and 4 were negative for the viruses tested for in this study (Table 2). The nine hMPV-positive samples were distributed between both study years (first year, n = 3; second year, n = 6).

TABLE 1.

Characteristics of infants with a positive hMPV results requiring intensive care

Center	Date of sampling (mo/day/yr)	Age (days)	Multiplex PCR result	Nosocomial infection
Freiburg	12/03/99	288	Negative	No
Hamburg	12/20/99	1,014	PIV-1	No
Bochum	01/05/00	901	Negative	No
Hamburg	01/11/00	57	Negative	No
Hamburg	01/11/00	63	RSV-A	No
Freiburg	01/18/00	28	Negative	No
Hamburg	01/24/00	353	RSV-A	No
Hamburg	02/21/00	31	RSV-A	No
Freiburg	08/14/00	45	PIV-3	Yes
Hamburg	12/08/00	14	RSV-A	No
Hamburg	01/12/01	42	RSV-A	No
Hamburg	02/12/01	18	RSV-A	No
Hamburg	03/08/01	39	RSV-A	No
Hamburg	03/20/01	135	RSV-A	No
Hamburg	03/21/01	31	RSV-A	No

TABLE 2.

Prevalence of hMPV in nasopharyngeal secretions grouped by study center

Study center	Total no. of specimens examined	No. of ICU patientsa	No. of hMPV-infected patients
			Total positive	RV positiveb	RSV positivec
Bochum	505	5	1	0	0
Dresden	471	25	0	0	0
Freiburg university hospital	624	17	0	0	0
Freiburg community hospital	511	15	3	1	0
Hamburg community hospital	1,051	25	11	2	9

^aPatients who required support in an intensive care unit (ICU).

^bRV, respiratory viruses, including IVs A and B and PIVs 1 to 3.

^cRSV A and B.

Our further intent was to support the hypothesis that RSV and hMPV are a coinfecting pair of viruses that cause severe respiratory distress in children younger than 3 years of age. For this purpose 120 RSV-positive samples from children not requiring intensive care support from the second study year were randomly picked from all centers and were tested for the presence of hMPV. We could not detect hMPV in any sample (data not shown).

DISCUSSION

The present article presents data on the prevalence of hMPV in a prospective multicenter study covering children younger than 3 years of age. The national prospective PRI.DE study has clearly shown with a population-based sample of children <3 years of age that a single hMPV infection contributes to the etiology of severe LRTIs in a smaller proportion of patients than human RSV does and that double infection, like that with RSV and hMPV, must be considered a cause of severe lower respiratory tract disease.

Recent studies have suggested that hMPV should be added to the list of human respiratory viral pathogens affecting children (4, 15, 20, 21, 22, 24, 27). Our results, which cover data from 1 study year, indicate an hMPV prevalence of less than 1% in children <3 years of age who were negative for the main respiratory viruses. The prevalence of hMPV as a cause of acute LRTIs differs markedly in the literature, ranging from 1.5 to 25% (4, 16, 19, 29). Thus, the prevalence of hMPV in this study seems to be lower than that in studies from other European countries. One reason may be that, because of the study design, we detected only community-acquired hMPV infections. The difference in incidence is also likely related to the difference in the patient populations studied. Two published studies (11, 19) suggested that the incidence of hMPV may vary substantially from year to year. Our data for RSV-, IV-, and PIV-negative samples reflect data from only 1 year of the study. However, recent data from an ongoing prospective study on the prevalence of hMPV in Germany (unpublished data) in children <3 years show that the prevalences of the three types of viruses are identical (data not shown).

Little information on the clinical presentation and the impact of hMPV infection exists. In the first description of hMPV by van den Hoogen et al. (27), the virus was isolated from 28 epidemiologically unrelated children; 13 were infants and 14 were younger than 5 years of age. These children had a range of diseases, from mild respiratory problems to bronchiolitis and pneumonia. In our study the two children positive for hMPV and negative for the other viruses tested had only mild symptoms. Serological studies in The Netherlands indicate that by age 2 years the proportion of children with evidence of previous hMPV infection is approximately 55% and that by age 5 to 10 years the proportion is 100%. Recently, Greensill et al. (14) detected hMPV coinfection in 70% of infants with RSV bronchiolitis receiving ventilatory support in a pediatric intensive care unit. However, from their report it is not clear whether the patients became infected with RSV and/or hMPV during their stay in the hospital (nosocomial infections) due to viral endemicity. In our prospective study the nasopharyngeal secretions were obtained at the time that the patients saw a physician. Thus, in our prospective study we detected only community-acquired hMPV infections. We detected hMPV coinfections in circa 30% of RSV-infected infants admitted to an intensive care unit and coinfection with RSV in 60% of hMPV-infected infants admitted to an intensive care unit. Overall, our data show that the percentage of children in an intensive care unit infected with hMPV was 18%, whereas <1% of the virus-negative samples from the second year were infected with hMPV. In our study the various centers had high or low incidences of hMPV-positive samples among the patients in the intensive care units of the hospitals (Table 2). Unfortunately, we have no insight how it happened that all nine of the RSV-positive patients were admitted to a hospital in Hamburg. However, we must keep in mind that differences in the management of RSV-infected children may lead to differences in the prevalence of RSV and hMPV coinfection in intensive care units. Another explanation might be a particularly severe RSV season in Hamburg. In a separate project we analyzed about 100 RSV isolates from all study centers during the first study year (1999-2000) with regard to their genetic relatedness by sequencing the first hypervariable region of the RSV G protein (25). The circulating RSV strains could be divided into three clusters. In contrast to the dominant clusters in Dresden, Bochum, and Freiburg, cluster 2 was dominant in Hamburg. Moreover, there was a significant correlation between RSV cluster 2 and disease severity (25). In summary, our data parallel those from Greensill et al. (14). Our data clearly show that RSV and hMPV coinfections are more severe than either hRSV infection or hMPV infection alone. From the literature it also became apparent that RSV and hMPV coinfections are not unique among RSV-infected patients. hMPV has been reported to coinfect patients with severe acute respiratory syndrome (5) as well as IV-infected patients. In our study we detected children who were coinfected with hMPV and PIVs.

Some general problems may be encountered when the results of the different studies in the literature that contain data on the prevalence of hMPV in children infected and not infected with common viruses that cause LRTIs are compared. Problems associated with sample retrieval and transport may influence virus stability and may thus have led to underestimates of the rates of RSV, PIV, and IV detection in hMPV-positive samples in previous studies and, as a consequence, a failure to detect coinfections. Therefore, in our prospective study the nasopharyngeal aspirates were immediately frozen in liquid nitrogen at the time of sampling. We must also keep in mind that the rates of detection of other respiratory viruses, especially RSV, vary between the different studies. No information that can be use to compare the sensitivities of the methods used to detect RSV, IV, and PIV as well as those of the PCRs used for hMPV detection is available. The test systems used to detect RSV, PIV, and IV by various investigators include antigen detection by enzyme-linked immunosorbent assay or direct immunofluorescence assay. In our study the nasopharyngeal secretions from the first study year were tested for the respiratory pathogens RSV, IV, and PIV in parallel by RT-PCR and by culture on three different cell lines for detection of the respective viruses. The results of culture and those of PCR were in good accordance. We did not isolate any virus under study from PCR-negative samples. We detected slightly more virus-positive samples by RT-PCR than by culture (data not shown). The use of PCR is particularly advantageous for hMPV detection because this virus is fastidious and difficult to grow in most cell lines; in addition, rapid antigen detection tests are not available. However, some problems with regard to the detection of hMPV by RT-PCR exist. Cote et al. (7) performed a systematic evaluation of the target regions (the N, M, L, F, and P genes) used for the detection of hMPV by RT-PCR assays. Their results indicated that PCR assays designed to amplify the N gene had superior specificities and sensitivities. In this regard, the N gene harbors more conserved regions than the other viral genes (1, 3, 7, 17, 28). Nonetheless, all primer sequences used so far, as described in the literature, are not conserved among all hMPV lineages described to date. Therefore, it is possible that in all published studies specific genetic lineages of hMPV have been missed. Furthermore, in some studies hMPV was detected directly from the respective specimens; in others it was performed after amplification by culture.

Our hMPV RT-PCR assay showed no cross-reactivities with viral cultures and/or clinical samples positive for human RSV, PIVs 1 to 3, or IVs A and B (data not shown).

This article is the first to present data on the prevalence of hMPV from a prospective multicenter study in Germany. The national prospective PRI.DE study has shown with a population-based sample of children <3 years of age that a single hMPV infection does not add much to the etiology of LRTIs. It does show, however, that double infection, like that with RSV and hMPV, must be considered a cause of severe lower respiratory tract disease.