- Split View
-
Views
-
Cite
Cite
Gregory A. Storch, Diagnostic Virology, Clinical Infectious Diseases, Volume 31, Issue 3, September 2000, Pages 739–751, https://doi.org/10.1086/314015
- Share Icon Share
Abstract
Diagnostic virology has now entered the mainstream of medical practice. Multiple methods are used for the laboratory diagnosis of viral infections, including viral culture, antigen detection, nucleic acid detection, and serology. The role of culture is diminishing as new immunologic and molecular tests are developed that provide more rapid results and are able to detect a larger number of viruses. This review provides specific recommendations for the diagnostic approach to clinically important viral infections.
Diagnostic virology is rapidly moving into the mainstream of clinical medicine as a result of the convergence of several independent developments. First, dramatic progress in antiviral therapeutics has increased the need for specific viral diagnoses. Second, technological developments, particularly in the area of nucleic acid chemistry, have provided important new tools for viral diagnosis. Third, the number of patients at risk for opportunistic viral infections has expanded greatly as a result of the HIV/AIDS epidemic. Finally, modern management of HIV infection and hepatitis C is providing a new paradigm for the integration of molecular techniques into management of chronic viral infections. These developments are not only increasing the use of diagnostic virology but are reshaping the field. The purpose of this article is to review the field of diagnostic virology at the beginning of the 21st century, to provide guidance about current use of the tools of diagnostic virology, and to provide a glimpse of important future developments.
Rationale for Specific Viral Diagnosis
Historically, diagnostic virology has had to justify its use. The reasons have been that traditional viral diagnostic techniques, especially culture, are slow, expensive, and often peripheral to clinical decision-making, particularly when no therapeutic agents are available. The availability of antiviral therapeutic agents such as acyclovir, ganciclovir, foscarnet, cidofovir, antiretroviral drugs, neuraminidase inhibitors, and IFN-α that are effective for specific viral infections but expensive (and in some cases potentially toxic) has created an obvious need for specific viral diagnosis.
In some cases, establishing a specific viral diagnosis limits other diagnostic procedures and may allow discontinuation of antibiotic therapy [1]. Likewise, in some cases, confirmation of a specific viral diagnosis helps in determining prognosis. A recent study of diagnostic testing for respiratory syncytial virus (RSV) documented that physicians usually believed that rapid test results influenced their management of cases [2]. Rapid RSV testing has also been used as the basis for patient placement to limit nosocomial transmission [3]. Finally, viral diagnosis may be important for public health purposes. For example, laboratory documentation of cases of rubella or rubeola can set in motion extensive vaccination campaigns.
Methods Used in Diagnostic Virology
Viral isolation and a number of methods for detection of viral antigens, nucleic acids, and antibodies (serology) are the core repertoire of techniques used for the laboratory diagnosis of viral infections, although some other techniques are also occasionally used (table 1). Viral isolation by means of cell culture is virtually always performed in designated virology laboratories. The other methods may be performed in those laboratories as well but may also be performed in diverse laboratory sections such as general microbiology, serology, blood bank, clinical chemistry, pathology, or molecular virology. The trend for viral diagnostic testing to be done outside of traditional virology laboratories is likely to accelerate as rapid diagnostic techniques based on immunologic and nucleic acid methodologies increasingly replace viral culture.
Cell culture. The modern era of diagnostic virology dates to the first descriptions of viral isolation in cell culture by Weller and Enders in 1948 [4] and Enders et al. in 1949 [5]. Indeed, the need for cell culture techniques is the raison d'être for virology laboratories as entities separate from other general clinical microbiology laboratories. While the relative importance of viral isolation as a diagnostic method is rapidly diminishing, it still remains necessary because it is the only technique capable of providing a viable isolate that can be used for further characterization, such as with phenotypic antiviral susceptibility testing. An additional advantage is that in contrast to most antigen and nucleic acid detection methods, viral culture allows detection of multiple viruses, not all of which may have been suspected at the time the culture was ordered.
Because no one cell culture type can support the growth of all medically relevant viruses, virology laboratories must maintain several different cell culture types. The minimum requirements are a primary monkey kidney cell line, used for the isolation of respiratory and enteroviruses, and a human fibroblast line, used for the isolation of cytomegalovirus (CMV), varicella-zoster virus (VZV), and rhinoviruses. A continuous human epithelial cell line such as HEp-2 is required for the isolation of RSV. Which cell lines are used for a specific specimen is determined by the information communicated from the ordering physician to the laboratory and by knowledge of the specimens usually isolated from a given specimen type. For example, CMV is the main virus isolated from urine specimens, and therefore any viral culture of urine must at least involve inoculation onto human fibroblast cells to allow CMV isolation.
Growth of viruses in cell culture is usually detected by visualizing morphological changes in the cells, known as cytopathic effect (CPE). The characteristics of the CPE are often sufficiently distinctive to allow the laboratory to be suspicious of which virus is responsible. When necessary, confirmation can be achieved by scraping the infected cells from the walls of the tube or vessel in which they are growing and preparing a fluorescent antibody stain with use of monoclonal antibodies specific for the virus or viruses thought to be responsible for the CPE. Cell cultures are typically viewed microscopically to detect CPE every 1–2 days for the first week of incubation. The time required to detect CPE varies from 1–2 days after inoculation for herpes simplex virus (HSV) to 1–3 weeks for CMV.
The shell vial culture method, first developed for CMV [6, 7], dramatically decreases the time required for detection of viruses in cell culture. The method involves centrifugation of the specimen onto the cell culture monolayer and incubation for 1–2 days, followed by fluorescent antibody staining of the cell culture, regardless of whether CPE is visible. In addition to detection of CMV, shell vial cultures have also been used to speed the detection of HSV, VZV, respiratory viruses, and the enteroviruses.
Another modification of traditional cell culture involves the use of genetically engineered cell lines. In these systems, genes are transfected into indicator cell lines to direct insertion of viral receptors on the surface of the cell and/or to direct expression of promoters that respond to a specific viral protein present in the specimen. Activation of the promoter triggers a reporter enzyme such as β-galactosidase that acts on a substrate to indicate the presence of the virus being sought. This approach has been most widely used for HSV [8] and HIV [9].
The virus for which culture remains most uniquely useful is HSV. Cell culture is also sometimes applied to the detection of CMV, VZV, adenovirus, RSV, influenza and parainfluenza viruses, rhinovirus, and the enteroviruses. It can also be used to detect measles, rubella, and mumps viruses, although those diseases are currently very unusual in the United States. For all of the viruses mentioned, the rapid tests described below are gradually replacing viral culture.
Antigen detection. Methods of antigen detection include fluorescent antibody (FA) staining, immunoperoxidase staining, and EIA. Of these, FA staining is the most widely used in diagnostic virology. Rapid viral diagnosis by means of FA staining was first described by Liu in 1956 [10] for detection of influenza and was pioneered for numerous viruses by Gardner and McQuillan [11]. The method was widely adopted by clinical laboratories during the 1980s, particularly for detection of respiratory viruses. The commercial availability of specific monoclonal antibodies was crucial. FA staining methods continue to be improved through the use of cytocentrifugation to prepare specimens [12, 13] and simultaneous staining with multiple different antibodies labeled with different fluorescent labels [14].
Antigen detection methods are particularly useful for viruses that grow slowly or are labile, making recovery in culture difficult. The most important targets have been RSV, influenza and parainfluenza viruses, and adenovirus in respiratory specimens; HSV and VZV in cutaneous specimens; rotavirus in stool specimens; and CMV and hepatitis B virus (surface antigen) in blood specimens. Viruses such as the enteroviruses and the rhinoviruses that have extensive antigenic heterogeneity and lack cross-reacting antigens are not suitable for antigen-detection techniques. The advantages of antigen-detection techniques are rapidity (results can be available within hours of receipt of the specimen into the laboratory) and lack of requirement for viral viability in the specimen, allowing greater flexibility in the handling and transport of specimens.
Nucleic acid detection. The development of PCR analysis in 1985 [15] made possible the diagnosis of viral infection through sensitive detection of specific viral nucleic acids. Any virus can potentially be detected in this way, and applications of PCR analysis and other nucleic acid amplification techniques continue to be developed. There is little doubt that during the next decade, applications of nucleic acid detection techniques will drastically reshape the field of diagnostic virology. By inclusion of a step employing the enzyme reverse transcriptase (RT), PCR analysis can be adapted to detect viral RNA. Viral load assays for HIV and hepatitis C virus (HCV) are examples of quantitative nucleic acid detection techniques. Multiplex techniques permit the simultaneous detection of more than one virus or even of a virus and a different class of pathogen. For example, a multiplex PCR analysis that detects the DNA of Epstein-Barr virus (EBV) and Toxoplasma gondii has been used for the diagnosis of mass lesions in the brains of patients with AIDS [16].
Nucleic acid amplification assays can be classified as target amplification assays or signal amplification assays. Examples of target amplification assays in addition to PCR include the ligase chain reaction, which has been used for detection of sexually transmitted disease agents [17], and the transcription-mediated amplification assay [18]. In signal amplification assays, the target itself is not amplified; rather, amplification is of a chemical signal used to detect hybridization of a probe with the target nucleic acid. Examples include the branched chain DNA (bDNA) assay [19] and the hybrid capture assay [20]. Signal amplification assays are less sensitive than target amplification assays but are also less prone to yield false-positive results due to contamination of reactions with laboratory-amplified nucleic acid.
Recent modifications of PCR analysis called “real-time” PCR are a dramatic development that may potentially greatly expand the applicability of PCR analysis for diagnosis of viral infections. In these assays, a fluorescent signal is generated as PCR takes place. The assays are run on highly specialized automated instruments that include optical systems to excite the fluorescent dyes and detect fluorescent emissions. Examples of real-time PCR instruments include the Light Cycler (Roche Diagnostics, Mannheim, Germany) and the Prism 7700, which utilizes “Taqman” technology (Perkin Elmer, Foster City, CA). The combination of amplification and signal detection markedly reduces the time required for nucleic acid detection and greatly reduces the likelihood of contamination of reactions with laboratory-amplified nucleic acid, since there is no need to open the tubes in which PCR has taken place. For example, PCR reactions run on the Light Cycler can be completed in 30 minutes.
Despite virtually unlimited potential applications, availability of diagnostic nucleic acid amplification assays remains limited because few such assays are currently licensed by the US Food and Drug Administration (FDA). The only assays currently licensed and available to laboratories in kit form are the Amplicor HIV-1 Monitor assay for HIV RNA (Roche Molecular Systems, Indianapolis), the Nuclisens assay for CMV pp65 RNA (Organon-Teknika, Durham, NC), and hybrid capture assays for CMV and human papillomavirus (HPV; Digene Corporation, Beltsville, MD).
Many laboratories have developed their own “home-brew” PCR assays. Implementation and use of these assays are regulated by the Clinical Laboratories Improvement Amendments of 1988 (CLIA '88). Availability of home-brew assays is largely limited to selected university laboratories whose personnel have the necessary skills and interest and to large commercial reference laboratories. The quality and performance of these assays vary widely. Interlaboratory certification programs to help standardize these assays are only beginning to become available.
Serology. The diagnosis of viral infections by detection of specific antiviral antibodies is a traditional method whose clinical utility is limited by the need for comparison of acute and convalescent antibody titers. However, detection of virus-specific IgM antibodies allows a diagnosis to be made from a single specimen. Viruses for which detection of virus-specific IgM antibodies are useful include EBV (IgM antibodies to the viral capsid antigen); CMV; hepatitis A virus; hepatitis B virus (IgM antibodies to the hepatitis B core antigen); parvovirus B19; measles, rubella, and mumps viruses; and the arboviruses such as St. Louis encephalitis virus.
Another area of utility of serological methods is for certain chronic infections such as with HIV or HCV, in which the presence of any antiviral antibodies is always (HIV) or usually (HCV) indicative of current infection. Finally, serology is uniquely useful for defining specific antiviral immunity. Viruses for which definition of immune status by serology is useful include VZV, CMV, EBV, HSV, measles and rubella viruses, parvovirus B19, hepatitis A (total antibodies), and hepatitis B (antibodies to the hepatitis B surface antigen).
Approach to the Diagnosis of Specific Viral Infections
Mucocutaneous infections. Methods used for the laboratory diagnosis of mucocutaneous infections caused by viruses are shown in table 2. The most common reason for viral diagnosis of mucocutaneous infections is the presence of vesicular or ulcerative lesions, which may be caused by HSV or VZV and occasionally by enteroviruses. When laboratory diagnosis is required, culture and FA staining are the procedures of choice. The sensitivity of culture for detecting HSV in genital lesions has been estimated to be 80% overall [21] and is higher for vesicular or pustular lesions than for crusted lesions [22]. Cultures for HSV usually become positive within 1–3 days. VZV is a more labile virus than HSV, and cultures are both less sensitive and slower, typically requiring 4–10 days for detection.
FA staining of a specimen obtained by scraping the base of a lesion with a scalpel blade or rubbing it vigorously with a polyester or rayon swab is more sensitive for detecting VZV than is culture [23] and should be considered the procedure of choice. FA staining for HSV can also be done effectively with use of the same specimen, especially if cytocentrifugation is used to process the specimen in the laboratory [12]. A recent study showed that PCR analysis was more sensitive than viral culture for detection of HSV in cutaneous lesions, suggesting the potential for future application of PCR analysis for routine detection of HSV [24].
Other common viral infections of the skin or mucous membranes are caused by HPVs and poxviruses (e.g., molluscum contagiosum and orf). These viruses are not cultured in clinical laboratories. The role of HPV testing of specimens from the female genital tract to detect high-risk HPV types that are associated with carcinoma of the cervix is not yet established, although it is likely that it will be used in the future in specific situations, such as for assessment of the cervical cancer risk of women with equivocal Papanicolaou smears [25].
When specific viral diagnosis is required, HPV infections can be diagnosed by DNA detection methods such as the hybrid capture assay [26] and PCR analysis [27] or by visualization of characteristic cytological or histologic changes. The hybrid capture assay, but not PCR analysis, is currently licensed by the FDA for HPV detection. Because of the different cervical cancer risks conferred by different papillomavirus types, useful tests will have to discriminate between high-risk and low-risk types. Both the hybrid capture assay and PCR analysis have this capability.
Respiratory infections. Methods available for the laboratory diagnosis of viral respiratory infections are shown in table 3. Specific diagnosis of viral respiratory infections is widely done for pediatric patients and is increasingly done for adult patients as well. The most widely employed methods are viral culture and antigen detection by FA staining or EIA. Suitable specimens include nasopharyngeal aspirates, washings, or swabs; bronchial washings; and bronchoalveolar lavage fluid. Antigen detection tests performed directly on the specimens have the greatest clinical utility because of the potential for rapid availability of results. Antigen detection procedures are available for RSV, the parainfluenza and influenza viruses, and adenoviruses, but not for rhinoviruses or coronaviruses. All of these viruses except the coronaviruses can be grown in cell culture.
Antigen detection techniques are more sensitive than culture for RSV but less sensitive for the other respiratory viruses [25]. The sensitivity of FA staining for influenza and parainfluenza in comparison with culture varies widely. Many laboratories report sensitivities of ∼80%, although some laboratories report sensitivities approaching 100% [11]. The sensitivity of FA staining is enhanced when the laboratory uses cytocentrifugation for specimen preparation [13].
Several commercial assays are available for rapid detection of influenza virus. These assays are interesting because they are similar to widely used rapid tests for group A streptococci and can be performed by individuals without specialized expertise in virology, a circumstance which potentially makes rapid influenza diagnosis widely available outside of virology laboratories. The Directigen assay (Becton Dickinson, Cockeysville, MD) is an EIA that can be used to detect influenza A virus. An enhanced version, currently in development, can also detect influenza B virus. The Directigen assay has a sensitivity of ∼80% in relation to culture [28, 29].
FluOIA (Biostar, Boulder, Colorado) is an optical immunoassay that detects but does not distinguish between influenza A and B viruses. This assay had a sensitivity comparable to that of culture in one evaluation [30]. Zstatflu (ZymeTx, Oklahoma City, OK) is based on detection of influenza neuraminidase activity and also detects but does not distinguish between influenza A and B viruses. This assay had a sensitivity (in comparison with culture) of 76% for influenza A virus and 46% for influenza B virus in one evaluation, on the basis of analysis of nasal wash and nasal aspirate specimens [31]. It should be noted that this assay is currently licensed only for use with throat swab specimens. A rapid test called QuickVue influenza test (Quidel, San Diego) is licensed by the FDA for use on nasal swabs, aspirates, or washes but has not yet been evaluated in any peer-reviewed publications. The performance of all of these assays may be affected by the specimen used for testing. A recent study of adult volunteers with experimental influenza A infection showed that influenza virus titers were highest in nasal aspirate specimens, followed by nasopharyngeal swabs and throat swabs [32].
CNS infection. For diagnostic purposes, it is useful to divide viral infections of the CNS into three categories: acute meningitis in immunologically normal hosts, acute encephalitis occurring in immunologically normal hosts, and opportunistic infections occurring in immunocompromised individuals. In each category, nucleic acid amplification—based tests are rapidly becoming the most important means of laboratory diagnosis. Tests used for the laboratory diagnosis of acute meningitis and encephalitis are shown in table 4.
Currently, the most common viral causes of acute meningitis are the enteroviruses and HSV type 2 (HSV-2). A recent report suggests that VZV may also be an important cause, sometimes producing meningitis in the absence of cutaneous lesions [33]. For the detection of enteroviruses, RT-PCR performed on CSF is considerably more sensitive than viral culture [34, 35]. Primers that amplify a highly conserved segment of the 5′ nontranslated region of the enterovirus genome detect 60 of the 67 enterovirus serotypes [36], including those serotypes most commonly associated with meningitis [36, 37]. Rapid nucleic acid amplification techniques such as RT-PCR analysis to detect enteroviruses have the potential for decreasing necessary antibiotic usage, especially for patients who have been pretreated with antibiotics (for whom withdrawal of antibiotics on the basis of negative bacterial cultures of CSF and blood might otherwise be done reluctantly) and for patients with atypical clinical features. Even in cases with typical clinical features, rapid testing may be useful for shortening hospitalization and decreasing medical costs [1, 38, 39]. Unfortunately, no such enterovirus nucleic acid amplification assay has yet been licensed by the FDA, and accordingly these tests are not yet widely available.
In adult patients, HSV meningitis, usually caused by HSV-2, is a common cause of aseptic meningitis. This illness occurs with or without concomitant genital lesions and can have recurrent episodes. While primary infections may yield a positive culture of CSF, recurrent cases are always culture-negative. PCR analysis has been very useful in permitting rapid recognition of both primary and recurrent cases [40, 41]. Likewise, PCR analysis for VZV can be instrumental in identifying cases of acute meningitis caused by VZV, which can occur as a complication of either varicella or zoster and can occur with or without cutaneous lesions [33, 42].
HSV-1 is the most common cause of sporadic encephalitis in the United States. Early recognition of this entity is extremely important because early antiviral therapy can reduce morbidity and mortality, which can reach 70% among untreated cases [43]. The results of PCR analysis performed on CSF correlate very closely with the results of synthesis of HSV antibodies from brain biopsy and/or intrathecal specimens [44, 45]. In a recent meta-analysis that compared CSF PCR analysis to one of these reference procedures, the sensitivity of PCR analysis was 96% and the specificity was 99% [46]. For this reason, CSF PCR analysis is now widely accepted as a substitute for analysis of a brain biopsy specimen for diagnosis of HSV encephalitis. On the basis of a decision-analysis model, we recommend accepting a positive CSF PCR analysis result as diagnostic of HSV infection of the CNS. We also recommend that a negative result can be used as the basis for discontinuing therapy with acyclovir if the clinical suspicion for that diagnosis is low. However, when the clinical suspicion is high, acyclovir should be continued even when the HSV PCR analysis is negative, because the sensitivity of the test is <100%. In these cases, it may be useful to repeat the PCR analysis on a specimen obtained 1–2 days later and to consider other diagnoses that may mimic HSV encephalitis [47].
The arboviruses are also important causes of encephalitis in normal hosts. The most important viruses causing human infection in the United States in recent years are the California (LaCrosse), St. Louis, eastern equine, western equine, Venezuelan equine, and West Nile viruses. These infections are best diagnosed by serology, especially with tests that detect virus-specific IgM antibodies. Both serum and CSF should be tested whenever possible. The sensitivity of these tests approaches 100% by the 10th day of illness [48]. Culture and nucleic acid amplification techniques are less useful than serology for diagnosing these infections because of the transient and low levels of viremia that are characteristic. State public health laboratories are currently implementing diagnostic testing for West Nile virus. Acute-phase serum and CSF specimens should be submitted to the state public health laboratory when this diagnosis is considered. Use of an RT-PCR assay has been described but was not positive in all cases [49]. Colorado tick fever is also often considered along with the arboviral diseases, and is an important cause of CNS infection in parts of the western United States. The causative virus is a coltivirus, which causes infection of circulating erythrocytes and can be detected by culture, FA staining, or PCR analysis performed on a blood clot [50].
Testing for rabies, which is done in public health laboratories, should be considered for patients with a history of exposure or with encephalitis of unknown etiology. Early in the illness, the diagnosis is based on detection of rabies antigens or nucleic acids in saliva, CSF, or skin obtained by biopsy from the back of the neck. After the 8th day of illness, the diagnosis can also be made by detection of rabies antibodies in serum or CSF. Rabies antibodies can also be present in serum but not CSF as a result of preexposure immunization. Consultation with public health authorities is recommended whenever the diagnosis of rabies is considered.
The most important viruses causing infection of the CNS in immunocompromised patients are CMV, EBV, VZV, and JC polyomavirus. The neurological syndromes associated with each of these viruses in immunocompromised patients and the reported sensitivity and specificity of PCR analysis for diagnosis are shown in table 5 [51]. CMV is an important cause of encephalitis and radiculomyelitis in patients with advanced AIDS. The virus can only rarely be cultured from CSF in these cases but is detectable by PCR analysis [52]. Quantitation of the level of CMV DNA is useful to distinguish between clinically significant CMV CNS infection and cases with asymptomatic or presymptomatic infection [53]. EBV causes primary CNS lymphoma in patients with AIDS, and the detection of EBV DNA by PCR analysis performed on CSF from AIDS patients with contrast-enhancing mass lesions visualized by neuroimaging techniques correlates closely with the presence of CNS lymphoma [54]. PCR analysis for JC virus is useful for diagnosing progressive multifocal leukoencephalopathy, although the sensitivity in most laboratories is <100% [51]. The use of VZV PCR for diagnosis of VZV-associated myelitis and encephalitis is complicated by the observation that VZV DNA is sometimes detectable in CSF from patients with acute zoster who do not have neurological manifestations [55]. PCR analysis may be more useful for the diagnosis of VZV infections of the CNS occurring in the absence of cutaneous lesions (zoster sine herpete).
Gastrointestinal tract infection. Four viruses or groups of viruses are currently recognized as important causes of gastroenteritis: rotaviruses, enteric adenoviruses, caliciviruses, and astroviruses. None of these are cultured in clinical laboratories, and consequently, viral culture has no role in the laboratory evaluation of the patient with diarrhea. These viruses were discovered by electron microscopy of fecal material. Although negative-staining transmission electron microscopy can be used as a diagnostic test, newer, more convenient assays have replaced electron microscopy in all but a few laboratories. Tests used for the laboratory diagnosis of viral gastroenteritis are shown in table 6.
Rotavirus is the most important cause of gastroenteritis in young children. Because large amounts of rotavirus are shed in the stool during acute rotavirus infection, antigen detection tests performed on stool specimens are very sensitive for establishing the presence of this virus. A number of rotavirus antigen assays have been licensed by the FDA and are widely available. EIAs are generally more sensitive than agglutination-based assays [56].
Enteric adenoviruses are specific adenovirus serotypes (especially 40 and 41) that cause gastroenteritis, and unlike the serotypes that cause respiratory infections, do not grow in the cell cultures typically used in diagnostic virology laboratories. An FDA-licensed EIA called Adenoclone (Meridian Diagnostics, Cincinnati) is available for detection of these viruses in stool specimens. The caliciviruses such as Norwalk agent are the most common viral causes of outbreaks of gastroenteritis. These infections are mild and usually do not require hospitalization. Specific diagnosis is most important in the context of an outbreak. RT-PCR assays have been developed to detect caliciviruses in stool [57] and are becoming available in many state health department laboratories. Astroviruses can be detected in stool specimens by EIA and by RT-PCR assay [58]. Currently, these assays are not commercially available, and diagnostic specimens must be referred to research laboratories interested in this virus.
CMV. Viral culture and serological testing were the traditional methods used to diagnose CMV infection, but both are being increasingly replaced by a confusing array of alternative methods. The reason for seeking replacements is that neither culture nor serology had adequate sensitivity and specificity for defining clinically significant infection. Newer direct detection methods include the pp65 antigenemia assay performed on peripheral blood leukocytes [59] and a variety of nucleic acid detection methods, including PCR analysis, the hybrid capture assay [60], and nucleic acid sequence—based amplification (NASBA) [61]. The latter is unique in that it detects CMV RNA, possibly a more specific indicator of active infection than CMV DNA. Eventually, quantitative nucleic acid amplification assays are likely to become the definitive tests for CMV because of a growing body of information that supports a relationship between the level of CMV in blood and the likelihood of present or future symptomatic disease [62–66].
The choice of test begins with the purpose of the testing. Important considerations include the clinical syndrome under investigation and whether the patient is immunologically normal or immunocompromised. Test methods for different diagnostic needs related to CMV are shown in table 7. In general, active infection is best diagnosed by direct tests for the presence of the virus, although an exception is the diagnosis of CMV mononucleosis in otherwise normal hosts, in which testing for CMV-specific IgM antibodies is useful [67]. A caveat in using this test is that for some patients with acute EBV infection, tests are false-positive for CMV-specific IgM antibodies [67], suggesting that testing for acute EBV infection should be performed simultaneously. Congenital infection can be conveniently diagnosed by urine culture, which must be performed within the first week of life to distinguish congenital infection acquired in utero from neonatal infection occurred during or shortly after birth.
For the diagnosis of systemic CMV infection in immunocompromised patients, the pp65 antigenemia assay [59] probably has the best combination of sensitivity and specificity among widely available test methods. An additional advantage is that it provides a quantitative estimate of the level of CMV viremia. Qualitative PCR analysis performed on peripheral blood leukocytes is too sensitive for the diagnosis of clinically significant systemic infection, having a low positive predictive value [68]. PCR analysis performed on plasma may be more specific [69], but further clinical evaluation is required. Localized CMV infection such as pneumonitis or gastrointestinal tract disease is best diagnosed by histologic examination of biopsy tissue, supplemented when necessary with immunohistochemistry.
Preemptive therapy is now a widely used approach for the control of CMV after solid organ or bone marrow transplantation. This approach is based upon the use of a laboratory marker that signals the onset of CMV infection before significant clinical manifestations have occurred. For this purpose, leukocyte or plasma PCR analysis, the pp65 antigenemia assay, and the hybrid capture assay are all appropriate. PCR analysis is preferred for bone marrow transplantation patients because of its greater sensitivity [70, 71]. In AIDS patients who are asymptomatic, detection of CMV by any of a variety of methods signifies an increased risk of future symptomatic disease [72], but such a finding is not currently used as an indication for starting therapy with anti-CMV drugs.
Epstein-Barr virus. EBV is the cause of most cases of infectious mononucleosis. It is also associated with primary CNS lymphoma in patients with AIDS and posttransplantation lymphoproliferative syndrome in recipients of solid organ and bone marrow transplants. The diagnostic approach differs for each of these clinical syndromes. Because cultures for EBV are too slow and cumbersome to be carried out in clinical laboratories, other diagnostic methods must be used. Serological methods are used to diagnose infectious mononucleosis and other manifestations in immunologically normal individuals, while molecular methods are used in immunocompromised individuals.
The diagnosis of infectious mononucleosis can be made in most cases on the basis of characteristic clinical findings, the presence of atypical lymphocytes, and the demonstration of heterophile antibodies, usually done with a rapid slide test. EBV-specific serological tests are reserved for young children, especially those <4 years old, who may fail to develop heterophile antibodies with acute EBV infection [73], and for adults suspected of having “heterophile-negative” EBV infection. Most EBV antibody panels include tests for IgG and IgM antibodies to the viral capsid antigen (VCA) and antibodies to early antigen (EA) and nuclear antigen (EBNA) [74]. Some laboratories also distinguish between diffuse and restricted components of EA because the presence of antibodies to the diffuse component is more specific for recent infection.
Testing for IgM anti-VCA is very useful in the diagnosis of acute EBV infection because these antibodies are typically detectable for only a few months after infection. Testing for IgG VCA is useful in defining EBV immune status, since these antibodies are detectable throughout the life of the individual. Antibodies to EA arise within a few weeks of infection, but they are not detected in all patients with acute infectious mononucleosis. Antibodies to the restricted component of an early antigen may persist for several years after infection [75]. Antinuclear antibodies arise later after infection than antibodies to EA, and they persist for life.
Thus, measurement of antibodies to EA (especially the diffuse component) and antinuclear antibodies is occasionally useful for determining the timing of infection. Table 8 summarizes the expected pattern of these antibodies in various EBV-related disease states. EBV serological testing should not be used for the evaluation of chronic fatigue because the relationship between EBV and this syndrome has not been validated.
Serology is much less useful in the diagnosis of certain specific manifestations of EBV infection that occur in immunocompromised individuals. In patients with AIDS, EBV is highly associated with primary CNS lymphoma, and EBV PCR analysis performed on CSF is useful for establishing this diagnosis [54]. EBV is also associated with lymphoproliferative disease after solid organ or bone marrow transplantation. Quantitative PCR testing of peripheral blood leukocytes may be useful for the early recognition of posttransplantation lymphoproliferative disorder, particularly in high-risk patients such as pediatric transplant recipients, and can also be useful for monitoring response to therapy [76]. Quantitative testing is required because many transplant recipients have EBV detectable in peripheral blood mononuclear cells after transplantation, and measurement of the level of DNA is required to distinguish those with posttransplantation lymphoproliferative disorder.
Viral hepatitis. Five organisms, hepatitis viruses A–E, are currently known to cause hepatitis. All are either difficult or impossible to culture, and hence all current laboratory testing involves detection of specific antigens, antibodies, or nucleic acids. Although no molecular tests are currently licensed by the FDA, molecular testing for HCV is widely used by clinicians caring for patients with HCV, and molecular testing for hepatitis B virus is coming into wider use as well. Blood banks are currently using molecular tests to screen blood units for HCV on an experimental basis. Table 9 shows diagnostic tests for viral hepatitis and their specific uses.
HIV and other human retroviruses. Several tests for HIV have become part of the mainstream of modern medical practice. The indications for use of these tests are summarized in table 10. In most circumstances, the laboratory diagnosis of HIV is based on a positive EIA performed on serum or plasma. HIV EIAs are widely available and have sensitivity and specificity >99.9% [77]. However, false-positives still make up a substantial percentage of all positives when these tests are used on low-risk populations, such as blood donors. For this reason, all first-time-positive HIV EIAs should be confirmed with an additional test, usually a western blot. Since it takes ∼22 days after infection for HIV antibodies to become detectable [78], the HIV EIA may be falsely negative during the first few weeks after infection. HIV RNA is detectable in plasma before the HIV EIA becomes positive and can be an indicator of acute HIV infection. It is important to keep in mind that false-positive HIV RNA tests have occurred [79], and the diagnosis of HIV should not be based solely on a viral load assay.
Rapid tests for HIV antibodies are now available and make it possible to perform HIV antibody testing within minutes. One such test, called SUDS (Murex Diagnostics, Norcross, GA), has been licensed by the FDA. Rapid testing is recommended by the US Public Health Service for situations in which individuals in need of testing are at high risk of not returning for a separate visit to receive test results [80]. Tests to measure HIV antibodies in oral transudate [81] and in urine [82] perform well and have also been licensed by the FDA.
Antibody assays cannot be used to diagnose HIV infection in infants born to HIV-infected mothers because of transplacental passage of antibodies, regardless of the infection status of the baby. HIV DNA PCR testing is now widely used for this purpose and allows determination of HIV infection status within the first few months of life [83]. Assays for HIV RNA (see below) can also be used [84].
Assays for plasma HIV RNA, often referred to as viral load assays, are now an essential component of the management of care for patients infected with HIV [85]. The level of HIV RNA predicts the rate of progression and can be used to monitor the response to therapy. The only assay currently licensed by the FDA is the Roche Amplicor HIV-1 Monitor (Roche Diagnostics, Indianapolis), which is an RT-PCR assay. An “ultrasensitive” modification of the assay has a detection limit of 40 copies per mL of plasma. Other test methods, including branched-chain DNA and NASBA, are also under development. It should be noted that the Roche assay has lower sensitivity for detection of some subtypes of HIV-1 that occur predominantly outside of the United States and does not detect HIV-2.
HIV drug resistance is currently evaluated by either genotypic or phenotypic methods. Most genotypic testing is performed by amplification of the RT and protease genes by RT-PCR, followed by nucleotide sequencing. Codons that have been associated with drug resistance are analyzed [86]. A disadvantage of genotypic testing is that minority viral populations, which constitute <20%–30% of the total population, are not detected. In addition, the potential interactions among multiple different mutations are not well understood.
Phenotypic testing is performed by amplifying the RT and protease genes and transferring them into a viable retroviral vector, which is then grown in the presence of varying concentrations of the drugs to be tested [87, 88]. The advantage of phenotypic testing is that it provides an “in vivo” view of the behavior of the patient's virus. This is particularly helpful when multiple mutations are present. Several studies have shown that genotypic susceptibility testing improves the management of HIV [89, 90]. Studies to evaluate the impact of phenotypic testing are under way.
HIV tests vary in their ability to detect HIV-2 in addition to HIV-1. EIAs used by blood banks all detect HIV-2, but some other commercial assays detect only about two-thirds of individuals with HIV-2 infection [91]. Testing for human T-cell leukemia virus types I and II is analogous to that for HIV. Commercial EIAs are widely available, since testing of donated blood is required. Confirmatory assays are available through reference laboratories.
Miscellaneous viral infections. The laboratory approach to a number of other viral infections is summarized in table 11. None of these infections is diagnosed primarily by culture. Virus-specific IgM antibody assays are the preferred method for rapid diagnosis of measles, rubella, mumps, and parvovirus B19. Some laboratories are also skilled in using immunofluorescence of nasopharyngeal secretions to provide a rapid diagnosis of measles [92]. Virus-specific IgG antibody assays are used to determine past infection and specific immune status. PCR analysis performed on serum is used to diagnose aplastic crisis caused by parvovirus B19. Because this manifestation occurs early after infection, parvovirus IgM titers may not be positive at the time of illness, although parvovirus IgM and IgG antibodies will become detectable within a few days. Serum PCR analysis is also used to diagnose chronic parvovirus B19 infection, which may cause chronic RBC aplasia in immunocompromised patients [93]. Many patients with chronic parvovirus infection do not have a diagnostic serological response.
BK virus is a polyomavirus that can cause hemorrhagic cystitis in bone marrow transplant recipients. The virus is difficult to grow in the laboratory, but PCR analysis performed on urine readily allows detection of BK virus. A recent study reported a correlation between the presence of BK virus in plasma of kidney transplant recipients and clinically significant nephropathy [94].
The diagnosis of human herpesvirus (HHV) types 6–8 remains difficult for clinical laboratories. None of these viruses can be grown in typical, hospital-based, diagnostic virology laboratories. The presence of virus-specific IgM antibodies can be used to diagnose acute HHV-6 or HHV-7 infection in young children with roseola. An alternative is to demonstrate by PCR analysis the presence of HHV-6 or HHV-7 DNA in peripheral blood leukocytes from a patient who does not have virus-specific antibodies [95]. The diagnosis of HHV-6 and HHV-7 infection in immunocompromised patients is currently problematic because serological testing may be nonspecific in this setting, and PCR analysis of peripheral blood leukocytes does not distinguish those patients who have clinically significant illness attributable to the virus from those who have asymptomatic reactivation. The clinical utility of laboratory diagnosis of HHV-8 infections is undetermined.
The diagnosis of unusual viral infections, especially those acquired overseas, requires consultation with public health laboratories, usually at the level of the state health department. Laboratory diagnosis of hantavirus infection can be achieved by detection of hantavirus-specific IgM antibodies [96] or by detection of hantavirus RNA by means of RT-PCR analysis performed on peripheral blood leukocytes [97]. Some state public health laboratories perform hantavirus tests, and others forward specimens to the Centers for Disease Control and Prevention. Dengue can also be diagnosed either by detection of dengue-specific IgM antibodies, which typically become detectable within a few days following defervescence, or by detection of dengue RNA in peripheral blood leukocytes by RT-PCR analysis [98].
Other exotic viral infections, including Lassa, Marburg, and Ebola virus infections, yellow fever, Congo-Crimean hemorrhagic fever, and Rift Valley fever are also diagnosed by means of IgM serology, PCR analysis (usually performed on blood), and culture. Serum and whole blood samples are usually required. Throat swabs, urine specimens, and CSF specimens may also be useful, depending upon the infection. Public health laboratories should be consulted regarding the choice of specimen, specimen collection, and transport.
Antiviral Susceptibility Testing
With the increasing use of antiviral drugs, antiviral drug resistance has become a clinical problem. However, the field of antiviral susceptibility testing is relatively undeveloped, testing is not widely available, and testing methods are not standardized. The viruses other than HIV for which antiviral susceptibility testing has greatest clinical relevance are HSV and CMV. The methodological considerations for detecting resistance are different for these 2 viruses.
For HSV, phenotypic testing with a culture-based method is preferred. The reasons for this are that HSV grows rapidly in the laboratory, making culture-based testing practical; in addition, whereas most resistance to acyclovir is related to changes in viral thymidine kinase, the genetic basis for the changes is diverse, complicating the application of genotypic methods [99]. The most widely used culture-based method is the plaque-reduction assay. The results of this assay are usually expressed as the 50% inhibitory concentration, or the concentration of drug required to produce a 50% decrease in the number of plaques that appear in culture, in comparison with the number that appear in the absence of drug. This test requires 1–2 weeks to perform and can provide data for a range of drugs, including acyclovir, foscarnet, and cidofovir. Other culture-based test methods include the Hybriwix assay (Diagnostic Hybrids, Athens, OH), which measures the effects of drugs on synthesis of HSV DNA [100], and the ELVIS assay (Diagnostic Hybrids, Athens, OH), which uses a genetically engineered cell line to compare viral growth in the presence and absence of drugs [101].
For CMV, culture-based phenotypic susceptibility testing is impractical because of the slow and fastidious growth of the virus. Approximately 80% of CMV resistance to ganciclovir is the result of a limited number of mutations in a specific gene designated UL97, and these mutations can be detected by a PCR-based assay [102–104]. This assay can be carried out either on an isolate recovered from a patient or directly in a patient specimen if the level of virus is sufficiently high to allow effective PCR amplification.
The other source of ganciclovir resistance is mutations in the DNA polymerase gene [104]. The only method currently available to detect these mutations is nucleotide sequencing. Sequencing can be carried out either on a viral isolate or directly from a clinical specimen with a sufficiently high level of virus. CMV resistance detection is performed only in a limited number of university and commercial reference laboratories, which have special interests in CMV.
Acknowledgments
I am grateful to Max Arens, Richard Buller, William Mulford, and the technologists who work in the Virology Laboratory at St. Louis Children's Hospital for their ongoing collaboration with me in the day-to-day work of diagnostic virology.
Comments