Diagnosis of Cytomegalovirus Infections
Abstract
Cytomegalovirus (CMV) is recognized as the most common congenital viral infection in humans and an important cause of morbidity and mortality in immunocompromised hosts. This recognition of the clinical importance of invasive CMV disease in the setting of immunodeficiency and in children with congenital CMV infection has led to the development of new diagnostic procedures for the rapid identification of immunocompromised individuals with CMV disease, as well as fetuses and infants with congenital infection. Diagnosis of acute maternal CMV infection by the presence of IgM and low IgG avidity requires confirmation of fetal infection which is typically performed by CMV PCR of the amniotic fluid. Viral culture of the urine and saliva obtained within the first two weeks of life continue to be the gold standard for diagnosis of congenitally infected infants. PCR assays of dried blood spots from infants have not been shown to have sufficient sensitivity for the identification of most infants with congenital CMV infection. However, saliva PCR assays are currently being assessed as a useful screening method for congenital CMV infection. In the immunocompromised host, newer rapid diagnostic assays such as pp65 antigenemia and real-time CMV PCR of blood or plasma have allowed for preemptive treatment reducing morbidity and mortality. However, lack of standardized real-time PCR protocols hinders the comparison of the data across different centers and the development of uniform guidelines for the management of invasive CMV infections in immunocompromised individuals.
Cytomegalovirus (CMV) is the largest member of the virus family Herpesviridae and is a ubiquitous virus that infects almost all humans at some time in their lives. The virus was first isolated by three different groups of investigators; Rowe and colleagues, Weller and colleagues, and Smith simultaneously in 1956 [1]. It has since been recognized as the most common congenital viral infection in humans affecting between 20,000 and 40,000 infants each year in the United States [2]. Congenital CMV infection is a leading non-genetic cause of sensorineural hearing loss in children. In addition, CMV has been recognized as an important cause of morbidity and mortality in immunocompromised hosts such as patients with acquired immunodeficiency syndrome (AIDS), and recipients of solid organ and stem cell transplants [3–5]. This recognition of the clinical importance of invasive CMV disease in the setting of immunodeficiency and in children with congenital CMV infection has led to the development of diagnostic procedures for the rapid identification of immunocompromised individuals with CMV disease and infants with congenital infection.
DIAGNOSTIC METHODS FOR CMV
Serology
Serological tests are useful for determining whether a patient has had CMV infection in the past, determined by the presence or absence of CMV IgG. Many different assays have been described and evaluated for the detection of CMV IgG antibodies. Among these are complement fixation, enzyme-linked immunosorbent assay (ELISA), anticomplement immunofluorescence, radioimmunoassay, and indirect hemagglutination [6]. The detection of IgM antibodies has been used as an indicator of acute or recent infection. Many different assays are available but enzyme-linked immunosorbent assays (ELISAs) are the most widely used and are based on crude viral preparations. The IgM capture assays are widely employed and are based on selective binding of IgM antibody to the solid phase. Recombinant IgM assays using recombinant HCMV proteins and peptides have been developed in an attempt to standardize serological assays [7]. However, studies have shown poor correlation of results obtained with different commercial kits for IgM testing [8]. In addition, assays for IgM antibody lack specificity for primary infection because of false-positive results, because IgM can persist for months after primary infection, and because IgM can be positive in reactivated CMV infections [9–11].
Because of the limitations of the IgM assays, IgG avidity assays are utilized in some populations to help distinguish primary from non-primary CMV infection. These assays are based on the observation that IgG antibodies of low avidity are present during the first few months after the onset of infection and avidity increases over time reflecting maturation of the immune response. Thus, high anti-CMV IgG avidity represents longstanding infection in an individual. Avidity levels are reported as the avidity index which is the percentage of IgG bound to the antigen following treatment with denaturing agents [7].
Cell culture
The traditional method for detecting CMV is through conventional cell culture. This approach utilizes clinical specimens which are inoculated onto human fibroblast cells and incubated and observed for a period of time ranging from 2 to 21 days. In the standard tube cell culture technique, CMV exhibits a typical cytopathic effect (CPE) characterized by foci of flat, swollen cells where the CPE is directly related to a virus’s titer. However, this method is slow and requires 2–3 weeks until a result can be reported as negative.
Shell vial assay is a modified viral culture by a centrifugation-amplification technique designed to decrease the length of time needed for rapid virus detection. It utilizes fibroblast cell cultures propagated on cover slips contained in flat bottom plates. Centrifugation of specimen onto the cell monolayer greatly assists adsorption of virus, effectively increasing infectivity of the viral inoculum [12]. Viral antigens might be detected by monoclonal antibody directed at the CMV immediate-early (IE) viral antigen by indirect immuno-fluorescence after 16 hours of incubation [13]. This method was adapted to be performed in 96-well microtiter plates [14] allowing for the screening of larger numbers of samples.
Antigenemia
The antigenemia assay has been commonly used for more than a decade for CMV virus quantification in blood specimens. This assay depends on the use of monoclonal antibodies that detect the viral pp65 antigen, a structural late protein expressed in blood leukocytes during the early phase of the CMV replication cycle. Antigenemia is measured by the quantitation of positive leukocyte nuclei, in an immunoflurescence assay for the CMV matrix phosphoprotein pp65 in a cytospin preparation of 2×105 peripheral blood leukocytes ( PBL) [15–19]. This test is limited to detection of the virus in leukocytes; the demonstration of positive-staining signals in the nuclei of leukocytes indicates a positive result. The test not only gives a qualitative result but is also quantitative, correlating closely with viremia and clinical disease severity in immunosupressed populations [20–22].
The disadvantages of the antigenemia assay are that it is labor intensive with low throughput and not amenable to automation. It is also affected by subjective bias demanding skilled persons for accurate test performance and interpretation of results. The samples have to be immediately processed (within 6 hours) while delay will greatly reduce the assay’s sensitivity [23]. Particularly in neutropenic patients, false-negative results may occur, since the antigenemia test depends on the presence of a sufficient number of polymorphonuclear leukocytes [24].
Polymerase Chain Reaction Amplification
Polymerase chain reaction (PCR) is a widely available rapid and sensitive method of CMV detection based on amplification of nucleic acids. The techniques usually target major immediate early and late antigen genes in their well conserved regions [25–27], but a number of other genes have been used as targets for detection of CMV DNA. DNA can be extracted from whole blood, leucocytes, plasma, or any other tissue (tissue biopsy samples) or fluid (urine, CSF, BAL) [28–33]. Specimen deterioration with time after sample collection is not as problematic with PCR assays as other tests for CMV [34].
PCR for CMV DNA can be either qualitative or quantitative, in which the amount of viral DNA in the respective sample is measured. The threshold of the qualitative method needs to be carefully calibrated for preventing over-detection. The quantitative PCR (Real-Time PCR) allows for continuous monitoring of immunocompromised individuals to identify patients at risk for CMV disease for preemptive therapy and to determine response to treatment [35–37]. This method is generally more expensive compared to the antigenemia assay, but it is rapid and can be automated. Results are usually reported as number of copies/ml of blood or plasma.
Reverse transcriptase (RT-PCR) can be used to detect viral mRNA transcripts in peripheral blood leukocytes independent of the presence of DNA. The absence of circulating mRNA is associated with a lack of CMV-associated symptoms, irrespective of the presence or absence of CMV DNA, while its presence is detected only in the setting of disease [38]. The presence of CMV IE mRNA has been demonstrated in monocytes and polymorphonuclear leukocytes during active CMV infection [39, 40]. It appears to be less sensitive, however, than the pp65 antigen test and PCR in diagnosing CMV disease [41, 42].
Immunohistochemistry
Immunohistochemistry is performed primarily on tissue or body fluid samples. Slides are made from frozen sections of biopsy tissue samples (liver, lung) or by centrifuging cells onto a slide. Then monoclonal or polyclonal antibodies against early CMV antigens are applied and visualized by fluorescently labeled antibodies or enzyme labeled secondary antibodies which are visualized by the change of color of the substrate. The stained slides are then examined by fluorescent or light microscopy. This technique is more sensitive and very specific compared to plain histological microscopy but it is very labor intensive and requires experienced personnel to read the slides [43]. False negative results can also occur due to focal distribution of the virus [44].
Nucleic acid sequence-based amplification (NASBA)
The assay allows the specific nucleic assay sequence-based amplification of unspliced viral mRNAs (late pp67 mRNA expression) in a background of DNA using a specific isothermal technique of amplification. Studies suggest that NASBA may be more sensitive than the antigenemia assay for the detection of CMV infection in blood [45]. Whole blood samples can be stored prior to testing, and the test can be completed in a day. The method is standardized but the mRNA extraction procedure is time consuming. Published reports suggest the use of mRNA transcripts of late CMV genes for reliable prognosis [46]. Experience with NASBA in transplant, AIDS and SOT patients has been encouraging [47].
Hybrid capture assay
Hybrid capture assay uses RNA probes to detect and quantify viral DNA in an ELISA-type format where the resulting signal is measured. Because it detects DNA without amplification its sensitivity is questionable [48].
MATERNAL AND FETAL INFECTIONS
The natural history of CMV infection during pregnancy is complex and not fully understood. Primary maternal infections are more likely to be transmitted to the fetus and intrauterine transmission occurs in 30–40% of the cases of primary maternal CMV infection [2, 49–51]. However, unlike toxoplasmosis and rubella, preconceptional immunity to CMV is incomplete and intrauterine transmission and damaging fetal infection can occur in women who CMV seroimmune prior to pregnancy [52–56]. The majority (>90%) of CMV infections in pregnant women (primary and non-primary) are subclinical [57]. No reliable tests can define transmission of infection to the fetus. In most countries, pregnant women are not routinely screened for CMV infection and testing is usually performed when CMV infection is clinically suspected in the mother or fetus, or in women considered at high risk for acquiring CMV infection (http://www.cdc.gov/cmv/clinicians.htm) [58].
Maternal infection
The diagnosis of primary CMV infection is accomplished by documenting seroconversion through the de novo appearance of virus specific IgG antibodies in the serum of a pregnant woman known previously to be seronegative. The presence of IgG antibodies indicates past infection from 2 weeks to year’s duration. Women with primary CMV infection during pregnancy are at greatest risk for having a child with congenital CMV with intrauterine transmission of CMV occurring in approximately 40% of primary infections during pregnancy [51]. However, maternal reinfection with a different strain of CMV can occur and such reinfections have been associated with intrauterine transmission, damaging fetal infection, and long-term sequelae [52, 53, 59].
IgM assays have been assessed in pregnant women as an indicator of acute or recent infection. In addition to the methods for IgM detection listed previously, an IgM immunoblot utilizing structural and nonstructural viral proteins was been shown to have 100% sensitivity and 100% specificity at detecting mothers who transmitted CMV to their offspring when samples were obtained at 21–24 week gestation [60]. Recently, assays utilizing protein microarray technology have been developed to detect CMV antibodies in sera but are in the early stages of development and testing [61]. With most IgM assays, detection of IgM in the serum of a pregnant woman can indicate a primary infection. However, IgM can be produced in pregnant women with nonprimary CMV infections [62] and false positive results are common [63, 64] and can occur in patients with other viral infections. In addition, anti-CMV IgM can persist for 6–9 months in a primary CMV infection [7, 51, 65].
Because of the limitations of the IgM assays, IgG avidity assays are utilized to help distinguish primary from non-primary CMV infection. Studies from investigators in Italy examined the sensitivity of IgG avidity and IgM by immunoblot in serum samples obtained at 6–18 weeks gestation and at 20–23 weeks gestation in 124 pregnant women. They found that in early gestation, IgG avidity was able to detect all women who had an infected fetus or newborn. However, the sensitivity of IgG avidity for detecting women who transmitted the virus to their fetus was much lower at 20–23 weeks gestation (Sensitivity 63%). When IgM testing in addition to IgG avidity was performed at 20–23 weeks gestation, the sensitivity of detecting a mother who transmitted CMV to their offspring increased to 81% [66]. Other researchers have utilized microneutralization testing in combination to avidity testing for diagnosing recent primary CMV infection in the second trimester of pregnancy [67]. Based on these data, some investigators propose screening pregnant women with serum IgG and IgM. If IgM is positive, then serum IgG avidity should be performed to help determine recent or past infection. Using this algorithm, some argue the sensitivity is similar to documenting de novo seroconversion [65, 68, 69]. Lazzarotto et al examined a cohort of 2477 pregnant women referred to their center because of a positive screening CMV IgM. Of these women, 1110 were confirmed to be IgM positive by immunoblot and 514 had low/mod IgG avidity and thus were considered at risk of transmitting CMV to their fetus. Twenty-five percent (121/514) of infants were congenitally infected, similar to 53/183 (30%) of infants infected as a result of primary CMV infection during pregnancy documented by seroconversion. Among the 1110 women with confirmed IgM positivity, 336 had high avidity antic-CMV IgG and six (2.0%) had CMV infected infants [68].
Several studies have examined the utility of maternal virological tests in diagnosing recent primary infection and determining risk of transmission to offspring. These studies have shown that fewer than 50% of pregnant women have detectable CMV in their blood as assessed by either PCR or pp65 antigenemia at the time of serological diagnosis [64, 69–71]. Investigators from Italy tested sequential blood samples from a small group of pregnant women with primary CMV infection for CMV DNA, pp65 antigenemia and immediate-early (IE) mRNA. During the first month of infection, all three tests showed high sensitivity (80–100%), however, the ability to detect evidence of CMV infection in the blood dropped off rapidly after the first 30 days of infection [72].
Fetal infection
Detection of CMV in the amniotic fluid has been the standard for diagnosis of infection of the fetus. Viral isolation in tissue culture was first utilized however the sensitivity was found to be poor (70–80%) with a high rate of false negative results [7, 73–78]. With the advent of PCR, detection of CMV DNA in amniotic fluid has shown to improve prenatal diagnosis of congenital CMV infection [77, 79–82]. The highest sensitivity of this assay (90–100%) has been shown when amniotic fluid samples are obtained after the 21st week of gestation and at least 6 weeks after the first positive maternal serologic assay. This allows adequate time for maternal transmission of the virus to the fetus and diuresis by the fetal kidney which is the primary site of viral shedding [77–79, 82]. However, even when PCR on amniotic fluid is performed at the optimal time, false negative results may occur. In a recent study, investigators in Italy showed that among 194 women who underwent prenatal diagnosis of congenital CMV infection, 8 mothers with negative amniotic fluid culture results for CMV delivered infants who were confirmed to be CMV infected [70].
Recently, CMV DNA quantification in amniotic fluid samples has been proposed as a means to evaluate the risk that a fetus can develop infection or disease. Several groups of investigators have shown that higher CMV DNA viral load in the amniotic fluid (≥ 105 GE/mL) was associated with symptomatic infection in the newborn or fetus [77, 83, 84]. However, other studies have failed to confirm a correlation between CMV DNA levels and the clinical status at birth [85, 86]. Rather, CMV viral load in the amniotic fluid correlated with the time during the pregnancy when the amniocentesis was performed, with higher CMV viral loads observed later in gestation [84, 85]. However, as with qualitative PCR on amniotic fluid, even when sampling was done at the appropriate time, very low or undetectable CMV DNA by quantitative PCR was found in infants infected with CMV [83, 85, 86].
In addition to CMV viral load, some investigators have examined the prognostic value of determining the CMV genotype in infected fetuses. Studies examining the two polymorphic CMV genes, gB and UL144, have failed to correlate a particular viral genotype with severity of fetal infection [87, 88]
Fetal blood sampling has been evaluated to determine the prognostic value of virologic assays in the diagnosis of congenital infection as well as nonspecific tests in determining severity of disease from CMV infection. The utility of CMV viremia, antigenemia, DNAemia and IgM antibody assays in fetal blood was examined for the diagnosis of congenital infection. Although these assays were highly specific, the sensitivity was shown to be poor (41.1%–84.8%) for identifying fetuses infected with CMV [7]. The same group found in 21 fetuses/infants with congenital CMV infection that all virologic parameters as well as IgM were higher in fetuses with ultrasound or clinical/laboratory abnormalities [89]. More recently, fetal thrombocytopenia has been shown to be associated with more severe disease in the fetus/newborn [78, 90]. However, investigators from Belgium documented fetal loss after funipuncture in an uninfected child. Thus, it is important to balance the value of cordocentesis against that known risk of miscarriage [78, 90].
Fetal imaging by ultrasound can identify structural and/or growth abnormalities and thus, can identify fetuses with congenital CMV infection that will be symptomatic at birth. The more common abnormalities on ultrasound include ascites, fetal growth retardation, microcephaly, and structural abnormalities of the brain [79]. However, the majority of infected fetuses will not have abnormalities on ultrasound examination [58]. In a recent retrospective study of 650 mothers with primary CMV infection by Guerra et al., among the 131 infected fetuses/neonates with normal sonographic findings in utero, 52% were symptomatic at birth. Furthermore, when the fetal infection status was unknown, ultrasound abnormalities predicted symptomatic congenital infection in only a third of infected infants [91].
Fetal MRI has been evaluated in a few small retrospective studies to assess its utility in detecting fetal abnormalities in utero. MRI appears to add to the diagnostic value of ultrasound with increased sensitivity and PPV of both exams versus ultrasound or MRI alone [92, 93]. However, more studies are needed to determine the true diagnostic and prognostic value of MRI in CMV infected fetuses.
CONGENITAL CMV INFECTION
Congenital CMV infection has been recognized as a leading cause of congenital infection and brain disease in children in the U.S. and Northern Europe. About 20,000 to 40,000 infants are born each year in the U.S. with congenital CMV infection, and of those, only about 10% of infants exhibit clinical findings suggestive of an intrauterine or congenital infection at birth or during neonatal period (symptomatic congenital CMV infection) [2, 94]. About 10% to 15% of the infants with subclinical or asymptomatic congenital CMV infection develop sensorineural hearing loss (SNHL) leading to the recognition that congenital CMV infection is a leading non-genetic cause of SNHL in the U.S. Many children with CMV-associated SNHL do have normal hearing at birth and the deficit can continue to deteriorate during early childhood [95–97]. Therefore, most infants with congenital CMV infection will not have detectable clinical abnormalities and the newborn hearing screening will not identify a significant proportion of children with CMV-associated SNHL. Since predictors of SNHL in children with congenital CMV infection have not been defined, in particular, among children with asymptomatic infection, it is currently not possible to identify infants at risk for CMV-related SNHL early in life. Early identification of congenitally infected infants at increased risk for SNHL is essential to provide appropriate monitoring and intervention measures during critical stages of speech and language development [98]. Therefore, detection of these at-risk infants early in life using rapid, reliable, and relatively inexpensive methods to screen newborns for congenital CMV infection was identified as a priority (NIDCD Workshop on Congenital Cytomegalovirus Infection and Hearing Loss: http://www.nidcd.nih.gov/funding/programs/hb/cmvwrkshop.htm).
The diagnosis of congenital CMV infection is typically made by the demonstration of the virus or viral antigens in newborn samples, urine or saliva. The detection of virus in urine and samples obtained from infants within the first two weeks of age is considered the gold standard method for the diagnosis of congenital CMV infection. In contrast to the symptomatic infants, most infants with asymptomatic congenital CMV infection are not identified because of the absence of clinical findings. Furthermore, identification of the virus or viral antigens in samples obtained from infants after the first two to three weeks of age may represent natal or postnatal acquisition of CMV and therefore, it is not possible to confirm congenital CMV infection in infants older than three weeks.
Serological methods are unreliable for the diagnosis of congenital infection. Detection of CMV- IgG antibody is complicated by transplacental transfer of maternal antibodies. In addition, currently available tests for the detection of CMV-IgM antibody do not have the high level of sensitivity and specificity as viral isolation methods [99, 100].
Virologic Methods
Detection of CMV in the saliva and urine of infants is easily accomplished because newborns with congenital CMV infection shed large amounts of virus. Traditional tissue culture techniques and recent modification of the tube culture method, centrifugation-enhanced rapid culture methods (shell vial assay) using monoclonal antibodies to stain for immediate early protein, pp72 of CMV are considered the standard methods for the diagnosis of congenital CMV infection [6, 101–103]. The rapid culture methods have been shown to have comparable sensitivity and specificity to the standard cell culture assays and the results are available within 24 to 36 hours. A rapid method a 96-well microtiter plate and a monoclonal antibody to the CMV immediate early antigen and was shown to be 94.5% sensitive and 100% specific for detecting CMV in the urine of congenitally infected infants [14]. This microtiter plate assay has been adapted for use with saliva specimens with comparable sensitivity and specificity [104]. These rapid culture techniques are presently the standard for the diagnosis of congenital CMV infection.
The CMV antigenemia assay, which detects the pp65 protein in polymorphonuclear leucocytes, is used widely to diagnose CMV infections and monitor treatment in immunocompromised patients [105]. However, the utility of this assay in the diagnosis of congenital CMV infection has not been evaluated.
Nucleic Acid Amplification Methods
Polymerase chain reaction (PCR) amplification of viral DNA is a very sensitive method for the detection of CMV in a variety of clinical specimens. The PCR assay is used routinely for the diagnosis of CMV infection in allograft recipients at increased risk for invasive CMV disease and in other immunocompromised hosts. Quantitative PCR has also been proven to be useful in the monitoring of these patient groups for their response to antiviral therapy [106–108]. However, the usefulness of PCR or other nucleic acid amplification assays in the diagnosis of congenital CMV infection has not been defined. An early study by Demmler et al. found that PCR using primers targeting immediate early and late CMV antigen genes was 93% sensitive and 100% specific when testing urine samples from newborns with congenital CMV infection [26]. In a study by Warren et al., PCR was found to be 89.2% sensitive and 95.8% specific when compared with standard tissue culture and rapid culture techniques of saliva from CMV infected infants [109]. In another study, CMV was detected in the CSF of 60% (6/10) of infants with symptomatic congenital CMV infection [110]. Nelson and colleagues were able to detect CMV DNA in the serum samples of all 18 children with symptomatic congenital infection tested, in 1 of 2 children with asymptomatic infection and in 0 of 32 controls [111]. A possible disadvantage of PCR of the peripheral blood is that viremia may not be present in all infants with congenital CMV infection; thus, detection of CMV DNA in peripheral blood may not identify every infant with congenital CMV infection [111–114]. DNA hybridization assay has excellent sensitivity and specificity for the rapid diagnosis of CMV infections [115]. However, the requirement for virus concentration using high speed centrifugation and the need for hybridization using radio labeled probes renders this method cumbersome and impractical for the routine diagnosis of congenital CMV infection.
Since dried blood spots (DBS) are collected from all infants born in the U.S. for routine metabolic screening, there has been an increasing interest in utilizing PCR-based assays for the detection of CMV in newborn DBS samples. The advantages of DBS PCR for newborn CMV screening include: 1) the specimens are already routinely collected for metabolic screening from all newborns; 2) PCR can detect viral DNA in DBS samples from CMV-infected infants; 3) PCR does not requires tissue culture facilities; and 4) PCR is amenable to automation, so large numbers of specimens may be screened at relatively low cost. Most of the reports have studied selected infant populations and a prospective comparison of DBS PCR results to a standard (i.e., tissue culture) method for identifying CMV infection in an unselected newborn population has not been performed [116–121]. Early studies have examined the utility of DBS PCR on preserved blood spots that were obtained from infants in the nursery to diagnose congenital CMV infection retrospectively at the time of the detection of hearing loss. A study by Johansson et al. retrospectively tested DBS of 16 infants with proven congenital CMV infection and 14 were positive by a nested PCR assay [118]. A number of studies from the group of investigators in Italy examined DBS from newborns and reported the sensitivity of the DBS PCR assay approaching 100% with a specificity of 99% [116]. However, in a large multi-center study of more than 20,000 newborns, a DBS real-time PCR assay was compared with saliva rapid culture for identification of infants with congenital CMV infection and the results demonstrated that DBS PCR could only detect less than 40% of congenitally infected infants [122]. The sensitivity and specificity of the DBS PCR assay when compared with the saliva DEAFF were 30.4% (95%CI, 21.5 – 41.0) and 99.9% (95% CI, 99.9 – 100%), respectively. The positive predictive value of DBS PCR assay was 84.8% (95% CI, 67.3 – 94.3%) and the negative predictive value was 99.6% (95% CI, 99.5 – 99.7). These results have major public health implications because they indicate that such methods as currently performed will not be suitable for the mass screening of newborns for congenital CMV infection. The high specificity of the DBS PCR assay suggests that a positive DBS PCR result will identify infants with congenital CMV infection. However, the negative DBS PCR assay result does not rule out congenital CMV infection in newborns.
The reasons for the low sensitivity of DBS PCR in identifying infants with congenital CMV infection are not entirely clear. PCR testing of peripheral blood has been widely used as a standard diagnostic method to detect invasive CMV infections in immunocompromised individuals including allograft recipients and patients with AIDS [106, 108]. However, it is likely that the pathogenesis of congenital CMV infection is different from that in immunocompromised hosts since such patients usually experience acute CMV infection or symptomatic reactivation shortly before testing, whereas congenitally infected infants may have acquired CMV infection months before birth and thus are no longer viremic when tested as newborns. Therefore, based on the results of our large multi-center newborn CMV screening study, it appears that DBS are probably not appropriate samples for newborn CMV screening. These findings underscore the need for further evaluation of high throughput methods performed on saliva or other samples that can be adapted to large-scale newborn CMV screening.
Several previous studies that included smaller number of subjects examined the utility of testing saliva samples with PCR-based methods demonstrated the feasibility and high sensitivity of these methods [54, 109, 123]. However, none of these studies have included screening of unselected newborns as well as a direct comparison of saliva PCR assay with the standard rapid culture method of saliva or urine. Although a more recent study from Brazil in which more than 8,000 newborns were screened for congenital CMV infection demonstrated the utility of saliva PCR assay to screen newborns for CMV, the PCR assay was not directly compared to the standard culture based assay [55]. As part of an ongoing multicenter newborn screening study, the utility of real-time PCR of saliva samples in identifying infants with congenital CMV infection is being evaluated. Thus far, the results are promising and showed that real-time PCR of saliva samples has excellent sensitivity (100%; 95% CI, 95.8 – 100%) and specificity (99.9; 95% CI, 99.8 – 99.8). The PPV and NPV of saliva PCR were 91.4% and 99.6%, respectively [124]. A major advantage of the saliva real-PCR assay used in that study was that there was no need for processing of saliva samples for DNA extraction. This elimination of the DNA extraction step will make it easier to adapt this assay for screening large number of newborns in a high throughput fashion. These findings demonstrate that saliva PCR could become a useful approach to screen newborns for congenital CMV infection.
Future Directions
Continual advances are being made in our understanding of the natural history and pathogenesis of congenital CMV infection and the role of antiviral therapy for congenitally infected children. It is hoped that the ongoing work in developing and standardizing molecular diagnostic methods will result in the availability of reliable, rapid, and simple methods for routine clinical use in the future. In addition, there is also growing interest in examining the feasibility of a newborn CMV screening program in conjunction with universal newborn hearing screening. It is somewhat disappointing that DBS PCR assays have not been shown to have sufficient sensitivity for the identification of most infants with congenital CMV infection. Nevertheless, the development of saliva PCR assays could have the potential to adapt these methods in a high throughput fashion to screen large number of newborns for congenital CMV infection. In addition, the ability to measure virus burden in saliva specimens from infants with asymptomatic congenital CMV infection using saliva PCR assays could provide the means to identify at-risk infants early in life thus, ensuring judicious use of resources by targeting at-risk children for follow-up and monitoring.
PERINATAL CMV INFECTIONS
Perinatal infections can be acquired by three routes, 1) contact with virus in maternal genital tract secretions during delivery, 2) ingestion of breast milk containing virus or 3) through transfusions of CMV seropositive blood. Transmission via breast milk and through blood transfusion can result in severe symptoms in premature, very low birth weight (VLBW) infants [125, 126]. For definitive diagnosis of perinatal CMV infection, it is important to demonstrate an absence of viral shedding for first two weeks of life. There is no standard method for diagnosis of perinatal CMV infections. Viral culture and CMV DNA detection by PCR using urine and/or saliva are the preferred diagnostic method [127, 128], although CMV excretion does not begin until 3 to 12 weeks after exposure [129, 130]. Recently, investigators have utilized quantitative plasma PCR assays in infants that acquire CMV perinatally. However, similar to blood PCR assays to diagnose congenital infection, not all infants who shed virus in their urine or saliva as a result of perinatal infection have detectable CMV DNA in their blood [127]. In perinatal CMV infection, serological assays have the same limitations described above for infants with congenital CMV infection.
CMV INFECTIONS IN IMMUNOCOMPROMISED HOSTS
CMV is one of the most common and difficult opportunistic pathogens which complicates care of immunocompromised patients. Advances in diagnostic and therapeutic modalities have reduced the frequency of life-threatening CMV complications and improved overall survival. CMV disease in immunocompromised hosts can result from primary infection, reinfection with a new virus strain or reactivation of the latent virus. The severity of CMV disease varies depending on the population, type of transplantation as well as on the level of immunosupression and can range from a self-limiting febrile illness to multi-system disease. CMV also has a number of indirect effects that contribute to increased morbidity and poorer outcome after transplantation. Correct and timely diagnosis of CMV infections is critical in the appropriate clinical management and the diagnosis should be based on appropriate clinical findings together with detection of the virus, viral antigens or DNA in blood, plasma or affected tissue [131].
There are three distinct groups of immunocompromised patients potentially affected by CMV infection: patient with solid organ transplantation, hematopoetic stem cell transplant recipients and patient with HIV/AIDS. The latter two groups can exhibit the most severe CMV disease due to severely impaired cellular immunity.
Solid organ transplantation (SOT)
CMV infection in SOT remains one of the major causes of extended hospitalization resulting in a significant part of the overall cost of care provided to these patients. It should be expected that 70–90% of patients undergoing SOT will be infected with CMV without antiviral therapy [132]. The source of the virus can be environmental exposure to the virus, the transplant recipient’s latent viral infection, virus contained in the transplanted organ or virus in a blood product. Clinical manifestations of CMV infection in SOT recipients can be expressed as an acute systemic febrile illness with symptoms such as fever, malaise, arthralgia and rash. Alternatively, CMV can affect specific organs in 10–30% of patients with CMV disease. Due to the immunomodulatory properties of CMV, infection can have indirect effects resulting in opportunistic infections with fungi or bacteria, or graft rejection [131, 132].
Serologic methods are of limited usefulness for identification of CMV disease in immunocompromised individuals [133]. However, these assays are used for pretransplant assessment of the solid organ transplant donor and recipient. CMV-seronegative transplant patients (no pre-existing CMV-specific immunity) that receive an organ from a CMV-seropositive donor are at highest risk for CMV disease [134, 135]. Serological testing is also useful to screen the donors of blood products to minimize the risk of CMV infection for seronegative recipients [136]. In addition, pre-transplant serostatus in the recipient is determined to assess the potential for reactivation of latent virus in seropositive recipients when undergoing immunosupression.
Detection of the virus in blood or CMV tissue after SOT indicates CMV infection onset with all related clinical signs and symptoms [137]. The expected time of onset is dependent on the recipient’s pre-transplant immunostatus, level of the immunosuppression and on the presence or absence of antiviral prophylaxis. It is important to monitor the levels of CMV viremia by either the pp65 antigenemia assay or by whole blood or plasma PCR in order to begin preemptive antiviral therapy promptly. The level of pp65 antigenemia used to start preemptive therapy varies depending on the center and is usually around 100 positive cells/200,000 leukocytes for seropositive recipients [138–140]. The use of plasma viral load measurement in risk assessment has limited value but may be helpful in specific cases [141] The viral load level that triggers antiviral therapy varies from institution to institution [142] and thresholds utilized in the assays may also vary according to the type of transplants and the individual PCR assays [143, 144].
Although the detection of the virus in tissues might often be difficult, the demonstration of the virus by immunohistopathology or in-situ techniques in addition to histopathological changes compatible with CMV infection is required for the diagnosis of CMV end organ disease [145–147].
Hematopoietic stem cell transplantation (HSCT)
CMV infection continues to cause significant morbidity in HSCT despite the use of CMV prophylaxis and preemptive therapy following allogenic HSCT. CMV disease usually presents as pneumonitis with interstitial pattern on radiographs with respiratory distress and hypoxemia or as gastrointestinal disease with mucosal inflammation or erosion anywhere in GI tract. The affected are mostly patients belonging to higher risk group: seronegative recipients (R-) of bone marrow from seropositive donors (D+) or seropositive recipients (R+) with non-ideal histocompatibility match [148]. The indirect effects of CMV can be presented via its immunomodulatory properties potentially enhancing graft versus host disease (GVHD) and opportunistic infections. The breakthrough and late onset CMV infections in high risk patients impact negatively the outcome in HSCT patients. It is important to screen blood products using serological techniques as well as nucleic acid amplification techniques before they are administered to HSCT patients. This allows reduction of exposure of seronegative recipients to CMV containing products.
Rapid detection techniques (both antigen and nucleic acid detection) are commonly and effectively used to monitor individuals at risk for CMV disease and for a timely start of preemptive antiviral therapy. HSCT recipients are usually monitored on a weekly basis. The pp65 antigenemia assay has been used to monitor for active CMV infection and a positive antigenemia is defined as >10 infected cells per 200,000 leukocytes. Typically, pre-emptive antiviral therapy is initiated at this threshold [149] but this can vary widely [150]. However, pp65 antigenemia is prone to false negative results in this population due to neutropenia commonly observed following HSCT.
More recently, blood quantitative or real-time PCR assays are utilized for detection of viral DNA to screen for CMV infection. However, there are no uniform viral load threshold values because of the absence of a standardized PCR protocol across the different centers and laboratories [151]. In general, the PCR threshold for the initiation of antiviral therapy is around 10,000 viral copies/ml of blood [152, 153]. When compared to antigenemia and PCR, the shell vial assay is considered to be insufficiently sensitive to be routinely used after allogenic HSCT to monitor for CMV infections [154, 155].
The diagnosis of CMV pneumonia, the most common manifestation of CMV disease in HSCT, is usually arrived at by the detection of CMV bronchoalveolar lavage (BAL) or lung biopsy specimens in the presence of clinical findings [151, 156]
CMV infection in HIV and AIDS patients
The diagnosis of CMV in HIV and AIDS patients provides additional rationale for optimization of antiretroviral therapy and consideration for preemptive anti-CMV therapy [157]. The occurrence of CMV disease in patients infected with HIV is closely related to the CD4 T-cell counts. About 10% of patients with CD4 counts less than 250 cells/mm3 and more that 20% of patients with counts less than 100 will develop CMV disease [158–161]. CMV disease is typically seen when HIV viral load is >100,000 copies/ml of plasma [160] or when p24 antigen is increased along with a low CD4 count [162]. CMV disease in HIV patients most often manifests as retinitis, esophagitis and enteritis. Other manifestations include peripheral neuropathy, polyradiculoneuritis, pneumonitis, gastritis or hepatitis and colitis. In HIV infected children, the presence of CMV infection was associated with more rapid progression into AIDS and death. However, with current highly active antiretroviral therapy (HAART), reconstitution of the immune function results in reduced occurrence of CMV disease in HIV-infected individuals [163, 164].
In CMV seropositive patients with AIDS, detection of CMV viremia was predictive of death and provides prognostic information in addition to CD4 counts and HIV viral load [165]. Quantitative PCR of CMV has been useful with high CMV viral load correlating with clinical symptoms and the presence of disease [166]. Also, pp65 antigenemia assay level of >100nuclei/200,000 cells has been associated with the occurrence CMV disease in 63% of patients, compared to 3% in antigenemia negative patients [160]. Additionally, the pp65 antigenemia from the CSF can be successfully used in confirmation of CMV polyradiculomyelopathy seen almost exclusively in AIDS patients [167, 168].
Another promising technique for detection of CMV in AIDS patients is pp67 mRNA NASBA, which has shown to be comparable to pp65 antigenemia with high specificity and positive predictive value. An additional benefit of this novel assay is that it can be performed on frozen or lysed samples [47].
Table 1
Summary of diagnostic tests for CMV
| Method | Specimen | Comments |
|---|---|---|
| Serology | Blood | IgG only valuable for establishing past infection, IgM has poor sensitivity and specificity to detect recent infection |
| Cell Culture | Blood, Urine, Saliva | Conventional culture time consuming Rapid culture methods (Shell vial assay) |
| Antigenemia | Blood | Rapid, semi-quantitative results, labor intensive |
| Polymerase Chain Reaction (PCR) | Blood, Urine, Saliva, tissue, | Very sensitive, detects viral DNA and/or RNA, allows quantitation of viral load, not standardized |
| Immunohistochemistry | Tissue | High specificity but low sensitivity due to tissue distribution of virus, labor intensive |
| NASBA* | Blood, tissue | Can store samples prior to testing, limited data in immunocompromised patients |
| Hybrid Capture Assay | Blood, tissue | Poor sensitivity based on few published reports |
Table 2
Laboratory diagnosis of cytomegalovirus infection by patient population
| Maternal Infection |
| IgG seroconversion with two consecutive maternal blood samples most accurate IgM in addition to low IgG avidity |
| Fetal Infection |
| CMV PCR of amniotic fluid obtained after 21wga and 6 weeks past positive maternal serology |
| Congenital Infection |
| Detection of virus or viral antigens in saliva or urine using standard culture method and rapid culture methods. |
| CMV PCR of blood is highly specific but insufficiently sensitive. PCR assays of saliva or urine are promising. |
| Perinatal Infection |
| Viral culture or PCR of urine; Proof of absence of CMV shedding in the first two weeks of life |
| Solid Organ Tranplant Recipients |
| Risk assessment based on CMV serostatus of the donor and the recipient. Diagnosis based on pp65 antigenemia assay of whole blood, PCR –based techniques of blood, plasma or PBMCs. |
| End organ disease diagnosed by tissue immunohistopathology or in-situ techniques. |
| HCST Recipients |
| Risk assessment based on CMV serostatus of the donor and the recipient. Diagnosis based on pp65 antigenemia assay of whole blood (false positive results at low PBMC counts), PCR –based techniques of blood, plasma or PBMCs. PCR or immunohistochemistry detection of the virus in BAL or lung biopsy. |
| HIV/AIDS Patients |
| Pp65 antigenemia of whole blood or CSF; Quantitative PCR of blood, plasma or PBMCs. |
Acknowledgments
Financial support: National Institute on Deafness and other Communication Disorders (N01 DC50008, S.B., K23 DC008539, S.R.), and National Institute of Child Health and Human Development (1 R03 HD061090, S.B)
Footnotes
Potential conflict of interest: None
