To determine the efficiency of reverse-transcriptase polymerase chain reaction (RT-PCR) assays currently used in diagnosing enterovirus and rhinovirus infections, we compared results obtained with RT-PCR methods, which detect both enteroviruses and rhinoviruses simultaneously, with results obtained by conventional virus isolation. Both tests were performed on 591 specimens: 38 samples (6%) had positive results by both RT-PCR and isolation, 90 samples (15%) had positive results by RT-PCR only, and 7 samples (1%) had positive results only by virus isolation. In conclusion, RT-PCR was superior in rapidity and sensitivity to virus isolation and is recommended as the primary diagnostic tool for enterovirus and rhinovirus infections.
Human enteroviruses (HEVs; e.g., polioviruses, coxsackieviruses, and echoviruses) and human rhinoviruses (HRVs) are picornaviruses and comprise >150 serotypes that are important pathogens. Although most HEV infections are asymptomatic or cause only mild clinical symptoms, serious and even fatal infections can occur. The disease entities include paralysis, meningoenchephalitis, myocarditis, pleurodynia, and hand-foot-and-mouth disease, as well as other rashes, herpangina, and respiratory-tract infections . HRVs are considered the most frequent cause of upper respiratory-tract infections (in particular, the common cold) .
The diagnosis of HEV and HRV infections is conventionally based on virus isolation from clinical samples followed by a neutralization test (for HEVs) or an acid lability test (for HRVs). These methods are not only time-consuming, but also laborious, because a representative collection of susceptible cell types is needed. The utility of serological tests is limited because of the large number of HEV and HRV serotypes, and this has generally hampered the development of immunological tests, although some recent progress has been made in the production of broadly-reacting viral antigens.
Although the superior sensitivity, specificity, and rapidity of RT-PCR in direct detection of HEVs and HRVs from clinical samples has been well documented, less information is available about their efficiency in daily diagnostic use over longer time periods. To get a comprehensive picture of the impact of RT-PCR in the diagnosis of HEV and HRV infections, we evaluated findings for clinical samples analyzed by virus isolation and RT-PCR during a 5-year period (1996–2000).
Materials and methods. A total of 5970 clinical specimens, consisting of 1599 CSF samples, 812 stool samples, and 3559 respiratory-tract specimens (throat swab, bronchoalveolar lavage fluid, and nasopharyngeal aspirate specimens) were studied. Most of the samples were obtained from patients admitted to the Turku University Hospital (Turku, Finland) or to local hospitals in the southwestern part of Finland.
Prior to the nucleic acid extraction, nasopharyngeal aspirates and throat swab specimens were diluted in 1 mL PBS. Stool samples were suspended 1 : 10 in PBS and centrifuged to remove debris. Nucleic acids were isolated from 250 μL of a CSF specimen, respiratory-tract specimen dilutions, or stool sample supernatants by proteinase K-sodium docedyl sulphate treatment followed by phenol-chloroform extraction and ethanol precipitation. From April 1998 to the end of the study, a commercial kit (High Pure Viral Nucleic Acid Kit; Roche) was used according to the manufacturer's instructions.
Two previously described RT-PCR methods with primers from highly conserved regions of the picornavirus genome were used [3, 4]. These primers detect virtually all HEV and HRV serotypes. Primers amplifying a 120-bp fragment for both HEVs and HRVs were found to be most sensitive and were therefore used in the primary RT-PCR test . In 1996 and 1997, samples with positive results by the primary RT-PCR test were subjected to another RT-PCR test using a primer pair that gives rise to 530-bp amplification products for HRVs and 650-bp amplicons for HEVs . From 1998 to the end of the study, a liquid-phase hybridization assay, in which amplicons are identified by europium-labeled HEV probe and samarium-labeled HRV probe, was used . The sensitivity and specificity of the assays were previously evaluated with a representative collection of prototype enterovirus and rhinovirus strains as well as clinical isolates [3, 4].
Virus isolation was performed according to routine cell culture protocols; viruses were isolated from respiratory-tract specimens in cultures of LLC-Mk2 cells, A549 cells, Ohio strain of HeLa cells, and human foreskin fibroblasts, and from CSF and stool specimens in LLC-Mk2-cells, A549 cells, and human foreskin fibroblasts. The fibroblasts were replaced in 1998 with the CaCo cell line because of the latter's higher sensitivity to enteroviruses. HEVs were typed by neutralization with the World Health Organization antiserum pools, HRVs by an acid lability test, and other respiratory-tract viruses and herpes simplex viruses by immunoperoxidase staining [6, 7].
Results. Both RT-PCR and virus isolation were performed on 591 samples (table 1). The RT-PCR assay was the most sensitive method for detection of HEVs and HRVs. Overall, 90 (15%) of the specimens had positive results by the RT-PCR only, and 38 (6%) of the specimens had positive results by both tests. HEVs grew in cultures of 7 specimens (1%) that had negative results by the RT-PCR test. The superior efficiency of RT-PCR was most evident in detection of HEVs in CSF and respiratory-tract specimens, as well as detection of HRVs in respiratory-tract specimens. HEV RNA was found in 56 (18%) of 301 CSF samples tested by RT-PCR and isolation, whereas HEVs could be isolated from only 18 (6%) of the CSF specimens. Of those 18 CSF specimens, 15 also had positive results by RT-PCR. HEV RNA was detected in 31 (14%) and HRV RNA in 30 (13%) of the 229 respiratory-tract specimens. HEV and HRV isolation findings were positive for 8 (4%) and 6 (3%) of the respiratory-tract specimens, respectively.
A total of 5970 clinical specimens were tested by at least 1 of the 2 assays during the 5-year study period. Virus isolation was performed on 4857 samples, and RT-PCR was performed on 1704 samples (figure 1). Overall, HEV and HRV isolation had positive findings for 158 (3%) and 64 (1%) of the samples, respectively, whereas HEV RNA was found in 189 (11%) of samples and HRV RNA in 214 (13%) of samples tested by RT-PCR.
Discussion. In this study, we analyzed the efficiency of RT-PCR testing and conventional virus isolation in clinical laboratory use during a 5-year period. Virus isolation in cell culture has been the most widely used method for the diagnosis of HEV infections, and it also makes serotyping of the isolated enterovirus possible. However, typing by the neutralization assay includes an additional passage in cell culture that makes the method laborious and time-consuming. In spite of the recently introduced accelerated method of detection of HEVs by immunostaining of the inoculated cells , the results are only rarely available during the acute phase of illness. Due to problems with HRV isolation, which suffers from low sensitivity and the need for subsequent typing by acid treatment and cell culture passage, the test is used mainly in laboratories that are particularly interested in the etiological diagnosis of respiratory-tract infections.
We have developed and applied RT-PCR technology for human picornaviruses testing during the past 15 years and here we report results obtained during a 5-year period (1996–2000) in which the number of samples tested by both the most recent molecular methods and conventional virus isolation has been sufficient for clinical comparison. Earlier studies have shown the advantage of RT-PCR in the diagnosis of HEV infections of the CNS [9–11]. In agreement with these reports, our study shows that the sensitivity of RT-PCR is superior to that of virus cultivation in the detection of HEVs in CSF samples. For instance, only one-third of all the HEV meningitis cases reported in our study could be diagnosed by virus isolation. The low sensitivity of cell culture may be due to a low concentration of infectious viruses in CSF or the presence of HEV serotypes, which grow poorly in standard cell cultures.
HRVs have been shown to be the most common cause of upper respiratory-tract infections both in adults [2, 12] and in children . They have been associated with more-serious respiratory diseases in hospitalized patients as well [14, 15]. Recently, HEVs have also been reported to be common causes of respiratory-tract infections in children . We found that the RT-PCR test was the most sensitive method for detecting HRVs, as well as HEVs, in respiratory-tract specimens. Our results indicate that HRVs may indeed be important pathogens among hospitalized patients.
On the basis of our experience, RT-PCR is clearly the most suitable method for specific detection of HEVs and HRVs and should be used as a method of choice for specific laboratory diagnosis. Although it is sufficient in most clinical cases to know that the infection is caused by HEVs or HRVs, further identification of the virus strains, usually determined by conventional serotyping, may sometimes be needed for epidemiological purposes or for selection of proper treatment (e.g., identification of polioviruses or determination of which antiviral drugs to use in future treatment). Typing of HEV can also be carried out using molecular methods that take advantage of amplification and analysis of the VP1 capsid protein gene sequences, a rapid and precise method for accomplishing this purpose . Rapid diagnosis of HEV meningitis and other infections caused by these viruses decreases both the unnecessary use of antibiotics and the costs in clinical practice . The RT-PCR assay is also able to detect HEV and HRV strains that do not grow in standard cell culture conditions. Simultaneous identification of both HEVs and HRVs from a single sample makes analysis of large numbers of samples and various sample-types possible. Because picornavirus drugs are under development and in clinical evaluation, the need for rapid specific diagnostic procedures may significantly increase in the near future.
The database management by Katja Rannikko is acknowledged.