INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Human cytomegalovirus (CMV) usually results in an asymptomatic infection in healthy individuals. In immunosuppressed patients, it is an important pathogen, with disease occurring as a result of acute primary infection or reactivation of a previously latent virus. In the human immunodeficiency virus (HIV)/AIDS population, CMV typically reactivates in those whose CD4 cell counts are below 100/mm³ (7). In transplant recipients, CMV may be acquired from the transplanted organ or tissue or from blood and blood product transfusions. CMV infection typically occurs within the first 3 months following transplantation and is a leading cause of organ rejection, morbidity, and mortality (18, 30).

The traditional "gold standard" method for detecting CMV infection is the isolation of virus in cell culture, which may take 28 or more days. In order to provide a more rapid result, it is now standard practice to perform rapid culture in shell vials with immunofluorescent staining, which can provide a result within 24 to 48 h (14, 15). However, studies have shown that cell culture and shell vial assays for CMV are relatively insensitive. Therefore, newer, more rapid, and more sensitive assays, such as the antigenemia assay, molecular amplification assays, and DNA hybridization assays, have been developed (2-5, 8, 20, 25-27, 34). These assays have focused on the detection of CMV in blood, which appears to be a good predictor of CMV organ disease in both HIV/AIDS and transplant patients (4, 5, 10, 11, 28, 29, 32).

The Hybrid Capture CMV DNA Assay version 2.0 is an improved version of the original Hybrid Capture CMV DNA Assay (2, 3, 16, 20, 25, 27, 34). It is a rapid, signal-amplified solution hybridization assay that utilizes unlabeled RNA probes, antibodies to RNA-DNA hybrids to amplify the signal, and a sensitive chemiluminescent detection system in a simple enzyme-linked immunosorbent assay-like format. A key difference between the original version and version 2.0 is that the probe diluent has been reformulated so that it is less viscous and easier to handle. In addition, several other reagents have been reformulated to improve the analytical sensitivity of the assay. No changes in the probes, antibodies, or procedure have been made. In this multicenter study, the performance of the version 2.0 Hybrid Capture CMV DNA Assay was compared to those of the pp65 antigenemia assay, traditional tube culture, and shell vial culture for the detection of CMV viremia in several different patient populations at three clinical sites.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Patients and specimens. The evaluation was conducted at three sites from January 1996 to September 1997. The protocol was approved by each site's ethics review committee. All patients gave written informed consent. Patients were eligible for entry into the study if they were HIV positive with suspected CMV disease, had received a solid-organ or bone marrow transplant within the past 6 months, or had some other underlying condition which put them at risk for CMV disease. A group of healthy, nonimmunocompromised volunteers were enrolled as controls. When blood specimens were collected for clinical reasons for CMV tube culture and shell vial, one additional EDTA tube of blood was collected at the same time for testing by the Hybrid Capture and antigenemia assays. Blood samples were not collected solely for the purposes of this study.

Hybrid Capture CMV DNA Assay (version 2.0). The assay kits were provided by the manufacturer (Digene Corporation, Silver Spring, Md.), and the assay was performed according to the manufacturer's instructions for the qualitative detection of CMV DNA. Whole blood was collected in EDTA tubes and stored at 4°C or room temperature (RT) for up to 24 h or at 4°C for up to 48 h prior to testing. Just before processing, the blood was mixed thoroughly, and 3.5 ml was added to 10 ml of lysis buffer and incubated for 15 min at RT to allow for lysis of erythrocytes. Following centrifugation (10 min at 1,000 × g), the supernatant was removed, and the cell pellet was resuspended in 1.5 ml of cold lysis buffer, transferred to a 2-ml screw-cap tube, and incubated for 15 min at RT. The leukocytes were then pelleted by centrifugation (15 min at 1,000 × g), and the supernatant was removed. The cell pellet was either tested immediately or stored at 20°C until further testing. Cell pellets were denatured by the addition of 75 µl of sample diluent (buffered solution containing carrier DNA) and 50 µl of denaturation reagent (1.75 N NaOH). Positive (CMV plasmid DNA and carrier DNA in sample diluent) and negative (carrier DNA in sample diluent) calibrators were included in each test run. Specimens were denatured for 25 min at 70°C, transferred to a fresh 2-ml screw-cap tube and denatured for an additional 25 min at 70°C. The negative and positive calibrators were denatured for 50 min at 70°C. Following denaturation, 50 µl of probe mix was added to each specimen and calibrator and hybridized for 120 min at 70°C. The contents of each tube were then transferred to a capture tube (a polystyrene tube coated with antibodies specific for RNA-DNA hybrids) and incubated for 60 min at RT with shaking at 1,100 rpm. The solution was then decanted from the tubes, and 200 µl of Detection Reagent 1 (buffered solution containing alkaline phosphatase-conjugated antibodies specific for RNA-DNA hybrids) was added to each tube. After 30 min at RT, the tubes were washed 5 times, and 200 µl of Detection Reagent 2 (chemiluminescent substrate Lumiphos 530) was added to each tube. After 30 min, the light emission was measured in relative light units (RLU) by using a DCR-1 tube luminometer. A positive cutoff value, equivalent to 700 CMV DNA copies/ml of whole blood, was generated from the standards supplied with the assay.

Culture methods. Traditional tube and shell vial cultures were performed by the routine method used at each of the three study sites. At all three sites, CMV was cultured in the presence of minimal essential medium containing fetal calf serum. Differences in tube culture and shell vial methods between the sites are shown in Tables 1 and 2, respectively.

Antigenemia methods. The CMV pp65 antigenemia assay was performed according to each laboratory's routine procedure. Site 1 used the CMV-vue assay (Incstar Corporation, Stillwater, Minn.). Site 2 used a nonlicensed, improved fluorescent version of the CMV-vue assay (24), and Site 3 used an in-house fluorescent assay as previously described (27).

PCR for CMV DNA. PCR testing was performed on selected discrepant specimens by the Mayo Clinic Department of Microbiology. The leukocyte fraction of the specimens was separated from whole blood by using Histopaque-1119, and the resulting cell pellet was amplified by using primers specific for the immediate-early region of the CMV genome. The amplified product was transferred to a nylon membrane by the Southern blot procedure and was detected with oligonucleotide probes and a chemiluminescent label (9).

Consensus definition, sensitivity and specificity analysis, and statistical analysis. For this study, traditional tube culture and shell vial culture were assumed to be 100% specific (19, 22). Hybrid Capture- and antigenemia-positive results were considered true positives if the specimen was positive by at least one other method. Discrepant results were defined as those that were positive only by Hybrid Capture or antigenemia. Specimens that gave discrepant results were adjudicated by CMV PCR or by review of the patient's records for evidence of CMV infection within 2 weeks before or after the Hybrid Capture or antigenemia assay result (25). Specimens that were negative by Hybrid Capture or antigenemia but positive by tube culture or shell vial were considered false negatives by Hybrid Capture or antigenemia, respectively. Specimens that gave equivocal results in the Hybrid Capture assay were analyzed separately.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

A total of 561 specimens were collected from 402 patients. Results were not available for 34 shell vials (10 not done, 24 toxic), 31 tube cultures (7 not done, 23 toxic, 1 contaminated), and 30 antigenemias (13 not done, 7 invalid, 10 insufficient cells). An additional 27 specimens gave equivocal results with the Hybrid Capture assay and were not included in the sensitivity and specificity analysis, leaving a total of 462 specimen results for comparison. Patient and specimen characteristics are shown in Table 3. Table 4 shows the direct comparison of tube culture, shell vial, antigenemia, and Hybrid Capture for the detection of CMV in peripheral blood leukocytes. Overall agreement between the Hybrid Capture assay and the other assays ranged from 82 to 91%. When the consensus result was used (true positive defined as positive by shell vial, tube culture, or any two assays), the overall sensitivities of Hybrid Capture, antigenemia assay, shell vial, and tube culture were 95, 94, 43, and 46%, respectively (Table 5).

The results of further study of the sensitivity and specificity of the four assays in each of the different patient populations are shown in Table 6. The sensitivities and specificities of Hybrid Capture and antigenemia remained high in all patient populations with no statistically significant difference among the groups of patients. However, the sensitivities of shell vial and tube culture showed substantial variability among the different patient groups and were consistently lower than Hybrid Capture and antigenemia in all patient groups.

Seventeen specimens were positive by Hybrid Capture but negative by shell vial, tube culture, and antigenemia. Eleven were available for testing by PCR, of which seven were positive. The medical records of the four PCR-negative patients revealed no evidence of CMV infection or disease. Review of the medical records of the six patients for whom a specimen was not available for PCR testing identified four patients with evidence of CMV infection within ±2 weeks of the positive Hybrid Capture result (two antigenemia positive, one shell vial positive, and one urine culture positive). Therefore, including these data in the calculations, the specificity of Hybrid Capture was 98% (347 of 353 specimens).

The specificity of Hybrid Capture was further evaluated in 131 healthy nonimmunosuppressed individuals. Only one subject initially tested positive for CMV DNA by Hybrid Capture. A second specimen from this subject was negative. This individual was CMV IgG positive. Three other specimens from three different subjects gave equivocal results by Hybrid Capture.

Twenty-seven (5.5%) of the 489 specimens analyzed gave equivocal results (as defined in the Methods section) by Hybrid Capture. Sixteen specimens were from solid-organ transplant recipients, ten were from patients with HIV/AIDS, and one was from a bone marrow transplant recipient. Evidence of CMV infection by shell vial, tube culture, antigenemia, or PCR was present in 13 (48%) of these specimens. Review of the medical records of the remaining 14 patients failed to provide evidence of clinical CMV infection.

When all 27 Hybrid Capture equivocal results were treated as negatives and included in the overall analysis, the sensitivities of Hybrid Capture, antigenemia, shell vial, and tube culture were 91, 93, 43, and 46%, respectively, and the specificities of Hybrid Capture and antigenemia were 96 and 94%, respectively. When the Hybrid Capture equivocal results were treated as positives, the sensitivity of Hybrid Capture increased to 95% and the specificity reduced to 90% while the sensitivities of the other methods were almost unchanged (93, 42, and 45%, respectively).

Twenty specimens were positive by the pp65 antigenemia assay but negative by all other methods. Seventeen (85%) of these specimens showed only one pp65-positive nuclei per 10⁵ cells. Three specimens showed two, three, and four pp65-positive nuclei respectively. Review of the medical records of these patients identified only one (5%) with evidence of CMV infection within ±2 weeks of the positive antigenemia result. Therefore, including these data in the calculations, the specificity of the antigenemia assay was 95% (334 of 353 specimens). The mean number of pp65-positive nuclei per 10⁵ cells in the consensus positive specimens was 56.4 ± 112.3 (median, 13; range, 1 to 500) while the mean number of pp65-positive nuclei per 10⁵ cells in the consensus negative specimens was 1.2 ± 0.9 (median, 1.0; range, 0.5 to 4) (P < 0.05, Mann-Whitney U Test).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Immunocompromised patients remain at significant risk of developing CMV infection and disease. The recent introduction of highly active antiretroviral therapy (HAART) in patients with HIV/AIDS has resulted in a dramatic decrease in CMV retinitis in this group of patients and has changed the population at risk for CMV disease (21). However, there continues to be a need for highly sensitive assays for the detection of CMV in these patients. In transplant recipients, the use of prophylactic and preemptive strategies for the prevention of CMV has continued to place emphasis on the need for highly sensitive and specific laboratory tests which can identify those patients most likely to benefit from anti-CMV therapy. We and others have previously evaluated several laboratory assays for detecting CMV viremia, including blood cultures, antigenemia assays, molecular amplification assays, as well as version 1 Hybrid Capture assay (2-5, 8, 20, 25-27, 34). Although these assays have demonstrated some utility in confirming the diagnosis of CMV disease and predicting those patients likely to develop disease, there continues to be a need for improved assays. The commercially available Hybrid Capture CMV DNA Assay version 2.0 appears to offer improved sensitivity in a standardized assay which can be used for comparison between different testing centers.

In this study, we compared the version 2 Hybrid Capture assay with the pp65 antigenemia assay, conventional tube culture, and shell vial assays for the detection of CMV in blood at three different study sites. The incidence of CMV viremia at these sites varied considerably depending on the assay used to detect it. The difference in these incidence rates may be accounted for by differences in laboratory techniques as well as by differences in the patient populations studied at each site. More than 70% of the patients enrolled at site 3 were HIV/AIDS patients whereas >70% of patients at sites 1 and 2 were transplant recipients. Prior to HAART, patients with HIV/AIDS tended to have a higher prevalence of CMV infection than transplant recipients (7, 18). Using the Hybrid Capture assay, the incidence of CMV viremia detected at all three sites was similar to that detected by antigenemia and was consistently higher than those detected by shell vial and tube culture. The increased positivity rate of Hybrid Capture and antigenemia was seen even when the results were analyzed by patient group (Table 4). These results likely reflect a true increase in sensitivity of Hybrid Capture and antigenemia compared to culture methods and not simply differences in methodologies used for shell vial and tube cultures at each site. This increased sensitivity was maintained even when consensus definitions for true positives and negatives were used. The clinical significance of this increased sensitivity is not known, as these assays may be detecting patients with subclinical CMV viremia who may or may not go on to develop active CMV disease.

The specificity of Hybrid Capture was further evaluated by testing 131 healthy nonimmunocompromised individuals for the presence of CMV viremia. Only one specimen from a CMV IgG-positive volunteer tested positive by Hybrid Capture. This individual had no clinical symptoms to suggest CMV infection, and testing of a new sample from this same patient was negative. It is not clear what caused the initial false positive by Hybrid Capture. It may have been due to nonspecific binding of the RNA probes or anti-RNA-DNA antibodies, or to laboratory error in performing the assay. Of note, three of the healthy volunteers had equivocal results by Hybrid Capture. Repeat samples for testing were not available for these individuals.

A total of 27 (5.5%) of the 489 specimens analyzed gave equivocal results by Hybrid Capture. Thirteen (48%) of these equivocal results showed some evidence for the presence of CMV infection either by shell vial, tube culture, antigenemia assay, PCR, or clinical history, while the remaining 14 patients had no laboratory or clinical evidence of CMV infection. This makes the interpretation of equivocal results difficult. We recommend that when no patient information with which to interpret the equivocal results is available, then equivocal results should be interpreted as negative with the awareness that approximately 50% of these may represent low-level CMV infections. This group of patients would benefit from repeat testing on a new specimen as recommended in the Hybrid Capture assay package insert or from serial monitoring for evidence of clinical CMV.

Using the same consensus definitions for true positives and negatives that were used to evaluate the Hybrid Capture assay, we evaluated the antigenemia assay. The results of this study are consistent with previous studies evaluating the antigenemia assay in which we and others have shown that patients with active CMV disease tend to have higher antigenemia levels than those without disease and that an increasing antigenemia level correlates with an increased risk of developing active CMV disease (5, 10, 13, 26, 33). Therefore, as with the equivocal results using the Hybrid Capture assay, which may represent low levels of CMV DNA, patients with a low-level positive antigenemia may benefit from repeat testing of a new specimen or serial monitoring for CMV.

Overall, Hybrid Capture appears to be more sensitive than standard tube culture and shell vial assays for the detection of CMV viremia and demonstrates sensitivity and specificity equivalent to that of the pp65 antigenemia assay. However, unlike the pp65 antigenemia assay, Hybrid Capture can utilize specimens that have been stored for up to 48 h (32) and is a standardized method. Moreover, Hybrid Capture detects CMV DNA in leukocytes, a marker that may correlate more closely with clinical symptoms and response to antiviral therapy (12, 13). Further study will be required to determine the clinical utility of the Hybrid Capture version 2 assay in predicting and detecting active CMV disease and monitoring patients' response to anti-CMV therapy.