Respiratory coronavirus infections in children younger than two years of age.

Abstract

The human coronaviruses (HCoV) comprise many antigenic variants, with the two serogroups represented by strains OC43 and 229E being the most common ones. HCoV infections have been associated with common colds in children and adults and with wheezing in young children with recurrent respiratory infections.1, 2 Reinfections with HCoV are common, and the virus may be carried for extended periods in the upper respiratory tract.2 Pitkäranta et al3 recently found HCoV, using the reverse transcriptase polymerase chain reaction (RT-PCR) technique, in 17% of the middle ear fluid (MEF)/nasopharyngeal aspirates (NPA) in children with acute otitis media (AOM). In Finland 40% of children were reported to have antibodies against the strain OC43 by the age of 2 years and practically all of them had antibodies by the age of 6 years.4, 5 However, no extensive studies to define better the role of HCoV in respiratory infections have been undertaken. Although HCoV infections are common, routine diagnostic panels for respiratory infections do not usually include a test for HCoV. Specific diagnosis of coronavirus infections has been very cumbersome. Virus isolation in cell culture is difficult, insensitive and time-consuming. Serologic methods (complement fixation, enzyme-linked immunosorbent assay) do not allow rapid virus identification. Virus detection in NPA with an indirect immunofluorescence assay using monoclonal antibodies is considered to be sensitive,6 but it is not commonly used to detect viruses in clinical specimens.7 The recently developed RT-PCR enables the detection of HCoV RNA directly from clinical samples with a high level of sensitivity.3, 6–8 We have adopted a previously published duplex RT-PCR technique for HCoV strains OC43 and 229E and combined it with microplate hybridization. The purpose of this study was to evaluate the applicability of this technique to the analysis of large numbers of specimens and to determine the presence of HCoV in a set of NPA and MEF specimens collected from infants in association with upper respiratory infection or AOM. Patients and methods. The specimens were derived from the collections of the Finnish Otitis Media (FinOM) Cohort Study carried out in the Hervanta area, in Tampere, Finland, from May, 1994, to June, 1997. After obtaining informed consent from the parents, we enrolled 329 children at the age of 2 months, and the clinical follow-up continued up to 24 months of age. The families were encouraged to take the child to the study clinic whenever he or she was suffering from acute respiratory infection, and especially if AOM was suspected. During these visits NPA was obtained routinely with a Pediatric Mucus Extractor (Orion Diagnostica, Espoo, Finland). Criteria for the diagnosis of AOM were a visually abnormal tympanic membrane (with regard to color, position and/or mobility) suggesting middle ear effusion, together with at least one of the following symptoms or signs of acute infection: fever, earache, irritability, diarrhea, vomiting, simultaneous respiratory infection or acute otorrhea not caused by otitis externa. In the case of AOM with effusion, MEF specimens were drawn from the inflamed ear(s). The NPA and MEF samples were frozen immediately after collection and stored at -70°C for 1 to 4 years before the analysis. The time-resolved fluoroimmunoassay method9 was used to detect antigens of adenoviruses, respiratory syncytial virus, parainfluenza virus types 1, 2 and 3 and influenza viruses A and B in the NPA and MEF samples. The samples had been analyzed for human rhinovirus by a previously published, combined isolation-RT-PCR method.10 A systematic report on occurrence of these viruses will be prepared separately. HCoV RT-PCR and hybridization. Extraction of viral RNA from NPA and MEF samples was performed with a commercial RNA isolation procedure (RNeasy, Qiagen GmbH, Hilden, Germany). RT-PCR for HCoV RNA was carried out by previously published methods with minor modifications.8, 10 Pairs of primers complementary to the HCoV nucleocapsid protein gene for both HCoV OC43 and 229E, 50 pmol of each, were used in the same reaction. At the reverse transcription step the virus-specific oligonucleotide primers were 5′-GCAAGAATGGGGAACTGTGG (OC43) and 5′-GACTATCAAACAGCATAGCAGC (229E). In the PCR the 5′-biotinylated forward primers were 5′-AGGAAGGTCTGCTCCTAATTC (OC43) and 5′-GGTACTCCTAAGCCTTCTCG (229E). Reverse transcription was carried out in a final volume of 40 μl including 5 μl of RNA. The mixtures were incubated for 60 min at 37°C and then heated for 10 min at 65°C. PCR was performed in 96-well plates in a final volume of 100 μl with 5 μl of complementary DNA. Forty cycles of PCR were run using published parameters.8 The microplate hybridization assay was carried out as previously published for human rhinovirus detection10 with some minor modifications. Five microliters of the PCR product were allowed to bind to streptavidin-coated microwells (Labsystems, Helsinki, Finland) and, after alkaline denaturation, exposed for 60 min at 36°C to the hybridization probes: 5′-TATTGGGGCTCCTCTTCTG for the HCoV OC43 and 5′-ACAACACCTGCACTTCCAAA for the 229E. Dinitrophenyl coupled to the 5′-end of the probes was detected by the immunoperoxidase reaction.10 Interpretation of results. Preparations of RD cell cultures infected with HCoV 229E (American Type Culture Collection, Rockville, MD) and OC43 (provided by Kathryn V. Holmes, University of Colorado, Denver, CO) were used as positive controls. At each step several negative controls were included. The results were calculated from the optical density values. The cutoff value of positive samples was defined as the mean of all the negative samples in each plate (altogether 20 samples) plus 5 times the standard deviation of the mean. If the optical density value was less than the mean of the negative controls plus 3 times standard deviation of the mean, the sample was considered to be negative. The samples yielding values between these two thresholds were reassayed by a confirmatory test for the hybridization step.10 Eighteen samples were reassayed, and one of them was confirmed to be positive. This was also documented by gel electrophoresis. According to comparative tests using serial dilutions of the prototype strains, the sensitivity of the HCoV RT-PCR was of the same range as that of virus isolation in RD cells. Results. Of the 2005 NPA and 1133 MEF specimens collected in the FinOM Cohort study, 1475 (74%) NPA and 391 (35%) MEF specimens were available for the analysis of HCoV RNA. They were derived from 279 children. Thirty-five tested NPA samples (2.4%) were positive for HCoV, 17 for the type OC43 and 19 for the type 229E (Fig. 1). One sample was positive for both OC43 and 229E. Thirteen NPA-positive children had acute otitis media as the main diagnosis at sample collection. A coinciding MEF(s) specimen was available in 4 cases; all 4 were negative for HCoV. HCoV RNA was also detected in 10 MEF specimens (3%); 7 samples were positive for OC43 and 3 for 229E. In all cases a coinciding NPA was analyzed, but they were negative for HCoV. Other viruses were found in 9 of the 35 HCoV-positive NPA specimens; 7 samples were positive for rhinovirus, 1 for respiratory syncytial virus and 1 for both rhinovirus and influenza B virus. One HCoV-positive MEF sample was positive for rhinovirus. Rhinovirus was the most commonly detected virus in the analysis of the entire material (S Vesa, M Kleemola, S Blomqvist, A Takala, T Kilpi and T Hovi, manuscript in preparation).Fig. 1: Absorbances of 1866 nasopharyngeal aspirates and middle ear fluid samples analysed for human coronavirus OC43 and 229E RNA. Original absorbance values have been transformed to relative values. −, negative; i, intermediate; i(+), intermediate, confirmed to be positive; +, positive sample.Clinical diagnoses recorded in association with the HCoV-positive NPA are shown in Table 1. The main diagnoses at the time of the HCoV infection were AOM (13) and upper respiratory infection (13). The detected HCoV infections occurred from August to March during the study years. Specimens available did not allow systematic analysis of the length of HCoV infection. Eight of the children had a HCoV-negative MEF before the HCoV-positive MEF specimen (time range, 9 days to 3 months) and 3 of these children also had a negative MEF ˜1 month after the detection of HCoV in the MEF samples. In 18 children with HCoV-positive NPA sample, the next NPA was taken; <1 month afterwards (range, 2 to 34 days). Only 1 child had 2 HCoV-positive NPA samples, collected 5 days apart.TABLE 1: Clinical diagnosis associated with human coronavirus-positive NPADiscussion. Little is known about HCoV infections in small children, partly because of the difficulties of conventional diagnostic methods. Now RT-PCR has been introduced for epidemiologic studies and also for clinical diagnostic purposes. In the present study we found that HCoV were detectable by RT-PCR in 2.4% of NPA samples of children with acute respiratory infection and/or AOM and in 3% of MEF samples of children with AOM. The 2.4% coronavirus infection rate is lower than in some previous studies with children,2, 3 possibly because our children were very young. Indeed several serologic studies suggest that the frequency of HCoV infections in children <2 years old is relatively low (antibody prevalence, 3 to 8%), but with older children and during the epidemics, which may occur every 2 to 3 years, the incidence can rise to 19%.5, 11–13 Hence the apparent paucity of the observed infections may also reflect a specific epidemiologic situation in the population.4, 5 Variable incidences have also been reported in adults.1, 14 According to serologic studies almost one-half of the HCoV infections are asymptomatic11 and therefore could have remained unrecognized in the current study. In conclusion we have adopted an RT-PCR-hybridization assay for large scale testing of HCoV OC43 and 229E in respiratory specimens. A pilot analysis of stored NPA and MEF specimens revealed a HCoV infection in 2.4% of the specimens, about one-half of them collected in association with AOM. This result justifies the addition of this test to the diagnostic panel of comprehensive respiratory virus detection studies. Acknowledgments. This work was partly supported by Wyeth-Lederle Vaccines and Pediatrics, Merck & Co., Inc. and Pasteur Mérieux Sérums et Vaccins and Helsinki University Central Hospital Research Fund. We thank Virva Jäntti for her advice concerning database searches and Kristiina Aitkoski, Annamari Harberg and Leena Palmunen for their excellent technical assistance.