JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1998, p. 281–283 0095-1137/98/$04.0010 Copyright © 1998, American Society for Microbiology
Vol. 36, No. 1
Long-Term Predictive Value of a Single Cytomegalovirus (CMV) DNA PCR Assay for CMV Disease in Human Immunodeficiency Virus-Infected Patients SHARON WALMSLEY,* TONY MAZZULLI,
AND
MEL KRAJDEN
Departments of Medicine and Clinical Biochemistry, The Toronto Hospital, Mount Sinai Hospital, and the University of Toronto, Toronto, Ontario, Canada Received 23 June 1997/Returned for modification 26 August 1997/Accepted 21 October 1997
Universal prophylaxis with oral ganciclovir is not cost-effective for the prevention of cytomegalovirus (CMV) disease in human immunodeficiency virus infection. For a preemptive strategy to be considered, patients at highest risk for CMV disease need to be easily and accurately identified. In this study, the sensitivity, specificity, positive predictive value, and negative predictive value of a single CMV DNA PCR assay for the subsequent development of CMV disease were 0.75, 0.89, 0.75, and 0.89, respectively. Specimen Preparation kit (Roche Diagnostic Systems, Inc., Branchberg, N.J.) according to the manufacturer’s directions. A total of 500 ml of whole blood was initially extracted, and the specimen was resuspended in 200 ml of the secondary extraction solution. Extracted nucleic acids were stored at 220°C, and aliquots of 25 ml were used for PCR amplification. CMV PCR was performed with the primer pair of Demmler (5), which targets a 435-bp sequence of CMV DNA within the fourth exon of the gene encoding the major immediate-early antigen of the CMV Towne strain. The CMV primer pair was supplied by Digene Diagnostics, Inc. (Silver Spring, Md.), with one of the primers labeled with biotin at the 59 end. CMV DNA was amplified according to the instructions supplied by the manufacturer. The Digene Sharp Signal System (DSSSA) (Digene Diagnostics, Inc.) was used to detect the PCR products (12). The PCR fragments were denatured and hybridized with a CMV-specific RNA probe. The RNA-DNA hybrids were captured onto the surface of streptavidin-coated microwells. Alkaline phosphatase-conjugated antibody specific for RNA-DNA hybrids, followed by a colorimetric substrate, was added. The optical densities at 405 nm at 1, 18, and 24 h were read. The color intensity was proportional to the amount of captured RNA-DNA hybrid in the well, with a maximum assay sensitivity at 24 h after substrate incubation. The DSSSA could detect approximately 50 input CMV copies. All PCR testing was performed in duplicate, and discrepant results were repeated. At baseline, 12 of 39 (31%) of patients had a positive DSSSA CMV assay. The proportion of patients with a positive PCR result increased as the CD4 count declined (Table 1). Of the patients with a positive assay, 9 of 12 (75%) developed CMV disease, including retinitis (n 5 6), retinitis with rectal ulcer (n 5 1), retinitis and myelopathy (n 5 1), and myelopathy (n 5 1), as outlined in Table 2. CMV disease was diagnosed at a mean of 8 months (range, 2 weeks to 23 months) after the positive DSSSA PCR result. The mean CD4 count at the time of diagnosis was 15/mm3. The other 3 of 12 (25%) patients had symptoms which may have been related to CMV but did not meet specific diagnostic criteria and therefore were classified as disease free in this study. Of these, one patient developed a rectal ulcer 9 months after the positive assay. Although a superficial culture was positive for CMV and the ulcer healed during therapy with intravenous ganciclovir, no histology or biopsy sample was obtained. A second patient developed an
Prior to the widespread use of highly active antiretroviral therapy, approximately 30 to 40% of patients with human immunodeficiency virus (HIV) infection and CD4 counts of ,50/mm3 developed cytomegalovirus (CMV) disease, most commonly retinitis (16). In a recent controlled trial, oral ganciclovir was shown to decrease the incidence of CMV disease relative to placebo (22). Although overall 50% of cases were prevented, 74% of patients receiving placebo did not develop disease. A recent analysis determined that the cost of prevention of one case would be $144,000 (19); therefore, the strategy of universal prophylaxis at a CD4 count of #50/mm3 is not cost-effective. An alternative approach would be to identify a patient subgroup at increased risk of disease and to target this group for prophylaxis or preemptive therapy. A number of different investigators are evaluating qualitative and quantitative CMV DNA PCR and antigenemia assays for their abilities to diagnose patients with CMV disease (1, 3, 4, 9, 13–15, 18, 23) and to predict those at highest risk (4, 6, 7, 17, 21). What remains unclear is the long-term predictive value of the assay results, the CD4 level at which testing should be initiated, and the optimal frequency of testing. Although a high-risk subset can be identified, it is unknown whether preemptive strategies would be effective and, if so, what threshold value of which assay should be chosen to initiate therapy. This will ultimately require standardized methodology (8). Therefore, the purpose of this study was to evaluate the long-term predictive value of a single CMV DNA PCR assay result for CMV disease in HIV. Thirty-nine patients attending the Toronto Hospital Immunodeficiency Clinic during the period from June to July 1994 consented to participate. The study was approved by the Toronto Hospital Institutional Review Board. At baseline, patients were examined for the presence of CMV disease through testing that included bedside fundoscopy. Blood samples were collected in anticoagulated EDTA tubes for testing by a CMV PCR assay. The patients were monitored prospectively for clinical signs and symptoms of CMV disease, which was diagnosed as previously described (22). CMV DNA testing was performed as follows (12). Nucleic acids were extracted within 4 h of blood collection by using the Amplicor Whole Blood * Corresponding author. Mailing address: The Toronto Hospital, Eaton Wing, Ground Floor, Room 219, 200 Elizabeth St., Toronto, Ontario, Canada M5G 2C4. Phone: (416) 340-3871. Fax: (416) 5955826. E-mail:
[email protected]. 281
282
NOTES
J. CLIN. MICROBIOL.
TABLE 1. Baseline CMV PCR results stratified by CD4 count No. of patients with CMV PCR-positive results/no. of patients tested (%)
CD4 count (cells/mm3)
.100................................................................... 1/12 (8%) 50–100 ................................................................ 2/10 (20%) ,50..................................................................... 9/17 (53%)
esophageal ulcer 12 months after the positive CMV assay. Because of concern about perforation, no biopsy was performed upon endoscopy. The patient died 1 month later without specific treatment for CMV. The third patient developed a severe wasting syndrome approximately 7 months after the positive assay. Despite enteral feeds, he died 1 month later. All patients with a positive baseline CMV DSSSA result died at a mean of 16 months following the positive test. One patient with a negative assay result at baseline was lost to further follow-up. Of the remaining 26 patients with negative assay results, 23 (89%) had not developed CMV disease within this study (Table 2). Eleven of these patients died of complications of HIV disease at a mean of 19 months (range, 4 to 33 months) after the assay. The other 12 patients remained alive and disease free after 39 months of follow-up. Of the surviving patients who remained disease free, 4 had CD4 counts of greater than 100/mm3, 4 had counts between 50 and 100/mm3, and 4 had CD4 counts of less than 50/mm3 at the last follow-up. A total of 3 of 26 (12%) patients with a negative PCR assay at baseline developed CMV retinitis at 10, 12, and 22 months, respectively, after the assay. Two of these patients died at a mean of 5 months after their CMV diagnosis; one remained alive at 15 months. Based upon these results, the sensitivity, specificity, positive predictive value, and negative predictive value of a single positive CMV DNA PCR for CMV disease were 0.75, 0.89, 0.75, and 0.89, respectively. This study demonstrates that a single CMV PCR assay has good positive (0.75) and negative (0.89) predictive values for the development of CMV disease in patients with moderately advanced HIV infection. These findings are consistent with those recently published by other investigators (4, 6, 21). The results of the studies cannot be directly compared, because of differences in the CMV assay used, differing populations, and the durations of follow-up. Shinkai et al. prospectively monitored 94 patients (median CD4 count of 54/mm3) enrolled in a natural-history study of HIV-related opportunistic infections (21). Plasma specimens were obtained every 3 months and assayed for CMV DNA by an in-house qualitative and quanTABLE 2. Clinical outcome of patients stratified by baseline CMV PCR resulta No. of patients Clinical situation
Baseline CMV PCR positive (n 5 12)
Baseline CMV PCR negative (n 5 26)
No CMV disease No. alive/no. deceased CMV disease with Retinitis Retinitis/rectal ulcer Retinitis/myelopathy Myelopathy No. alive/no. deceased
3 (25%) 0/3 9 (75%) 6 1 1 1 0/9
23 (89%) 12/11 3 (12%) 3 0 0 0 1/2
a One patient with a negative baseline PCR result had no clinical follow-up and was excluded.
titative competitive PCR assay. Oligonucleotide primers from the EcoRI fragment D region of human CMV strain AD169 were used for PCR. A total of 26 (28%) of 94 patients developed CMV disease at a mean follow-up of 12 months. The sensitivity, specificity, positive predictive value, and negative predictive value of one or more positive PCR results were 0.89, 0.75, 0.58, and 0.94, respectively. No data as to how many patients had a single positive assay versus repeated positive assays are presented. By using a quantitative PCR with a threshold of greater than 1,000 copies/ml as cutoff, the positive predictive value of the assay increased to 100%, but the sensitivity decreased to 35%. Dodt et al. (6) prospectively monitored 200 HIV-seropositive patients with CD4 counts of ,100/ mm3. They utilized an in-house nested PCR with primers from the fourth exon of the CMV IEAI gene of the Towne strain of CMV. The outer and inner primer pairs amplified 438- and 190-bp fragments of DNA, respectively. In addition, Dodt et al. used a pp65 antigenemia assay with mouse monoclonal antibodies (Dako) and an alkaline phosphatase staining technique. Both CMV PCR and antigenemia assays were performed every 2 months for 1 year. Thirty-eight patients (19%) developed CMV disease. The sensitivity, specificity, positive predictive value, and negative predictive value of any PCR-positive test were 0.95, 0.85, 0.57, and 0.99. For the antigenemia assay, which used a cutoff value of 10 positive cells per 105 peripheral blood mononuclear cells, 92% of patients with CMV disease had a positive test result compared to 12% without a diagnosis of CMV disease. Bowen et al. (4) prospectively monitored 97 HIV-infected patients with CD4 counts of less than 50/mm3 by using an in-house qualitative PCR assay. The primers amplified a 149-bp fragment of the CMV glycoprotein B gene. Patients with a positive baseline value were monitored monthly, whereas those with a negative baseline value were monitored every 3 months. Nineteen patients developed CMV disease over a 12-month period, including 16 of 27 (59%) who were PCR positive and 3 of 70 (4%) who were PCR negative at baseline. The three patients who developed disease despite a negative baseline PCR result were positive upon testing at the time of disease diagnosis. Results for the other 70 patients with repeated testing (235 samples in total) remained negative. Therefore, the sensitivity, specificity, and positive and negative predictive values of the baseline CMV PCR assays were 0.84, 0.86, 0.59, and 0.96, respectively. The methods for detecting CMV viremia vary among studies, making correlation between results uncertain. The two studies with repeated testing at 1- to 3-month intervals reported better sensitivities than the studies in which there was only a single evaluation (4, 6, 21). The former studies did not provide data on the number of patients who had only a single positive result despite repeated samples; thus, the optimal interval for testing remains undefined. Our data demonstrate a positive predictive value greater than that previously reported. This may be related to the longer duration of follow-up, appropriate classification of disease status, and the use of the DSSSA, which has been shown to be both sensitive and reproducible (2). Our data, together with those cited above, support the hypothesis that PCR-based or antigenemic assays for CMV viremia could be used to define a high-risk population in which trials of prophylactic or preemptive therapy for CMV could be designed (10). With improved assay standardization (8), specimen timing and frequency in the course of HIV infection could be assessed prospectively. Currently, the optimal assay for the detection of CMV viremia is unclear. Since CMV is a cell-associated virus, the use of whole blood as opposed to plasma or a defined number of white blood cells (as in the
VOL. 36, 1998
NOTES
antigenemia assays) may alter the amount of detectable CMV DNA in the sample. Patients with clinical CMV or advanced HIV may have significant leukopenia. Using serial titrations of CMV DNA PCR-positive specimens, we have demonstrated that HIV-positive patients with CMV infection have 6 to 9 log10 genomes of CMV DNA/ml in whole blood (data not shown). Bowen et al. (3) also found a median pretherapy CMV load in blood of 4.95 log10 genomes/ml (range, 3.6 to 7.05). Thus, a decrease in leukocytes by 1 to 2 log10 would be expected to have minimal impact on the results of qualitative CMV DNA assays. Zipeto et al. (24) have also demonstrated that the amount of CMV DNA in plasma correlates with the amount of CMV DNA in leukocytes; thus, whole-blood titration likely provides a relative measure of the amount of CMV DNA. For clinical use, a single or yearly evaluation as used in this study would be more convenient and cost-effective than monthly evaluations. In our study, a greater percentage of patients with positive assay results had CD4 counts of less than 100/mm3, which may mark the appropriate level for initiation of screening. However, in our cohort, one patient with a CD4 count of 180/mm3 at baseline had a positive assay result and developed CMV retinitis 2 years later. Further, with the introduction of highly active antiretrovirus therapy, the epidemiology of CMV disease appears to be changing, and disease might occur at even higher CD4 counts (11). Therefore, screening at infrequent intervals even earlier during disease may need to be considered to increase the window of opportunity during which preemptive strategies could be initiated. Although the sensitivity of the assay was only 75% in this study, previous studies of patients with CMV disease have shown that 5 to 44% do not test positive for viremia at the time of diagnosis (1, 3, 6, 9, 18, 20, 21). It is possible that hematogenous seeding of the eye occurs earlier, possibly at a time of increased HIV replication, and is subsequently cleared from the plasma. Reactivation of CMV in the eye at a later time may not be associated with viremia. Testing earlier in the course of HIV infection might increase the ability to detect this group. This will be important to clarify, since preemptive strategies will not be effective if a significant proportion of patients never develop positive markers for infection before disease. Although our findings need to be confirmed in a larger cohort, our data suggest that a single assay for CMV has a good predictive value for the subsequent development of CMV disease and can identify a high-risk subset and, more importantly, a low-risk subset for the development of CMV disease. Since the incidence of CMV disease has decreased with the use of highly active combination antiretrovirus therapies, the identification of patients who might or might not benefit from prophylaxis or preemptive therapy is essential for their cost-effective utilization. This work was supported by a grant from Allelix Biopharmaceuticals. We thank Prema Shankaran for laboratory expertise in performing the PCR assays and Jennifer Clark-Diprata for collecting patient specimens. REFERENCES 1. Bek, B., M. Boeckh, J. Lepenies, B. Bieniek, K. Arasteh, W. Heise, K. Deppermann, G. Bornhoft, M. Stoffler-Meilicke, I. Schuller, and G. Hoffken. 1996. High-level of quantitative pp65 cytomegalovirus (CMV) antigenemia assay for diagnosis of CMV disease in AIDS patients and follow-up. J. Clin. Microbiol. 34:457–459. 2. Boivin, G., J. Handfield, G. Murray, E. Toma, R. Lalonde, J. G. Lazar, and M. G. Bergeron. 1997. Quantitation of cytomegalovirus (CMV) DNA in leukocytes of human immunodeficiency virus-infected subjects with and without CMV disease by using PCR and the SHARP signal detection system. J. Clin. Microbiol. 35:525–526. 3. Bowen, E. R., P. Wilson, A. Cope, C. Sabin, P. Griffiths, C. Davey, M.
4.
5. 6.
7.
8.
9. 10. 11. 12.
13. 14.
15.
16.
17.
18.
19. 20.
21.
22.
23.
24.
283
Johnson, and V. Emery. 1996. Cytomegalovirus retinitis in AIDS patients: influence of cytomegaloviral load on response to ganciclovir, time to recurrence and survival. AIDS 10:1515–1520. Bowen, E. F., C. A. Sabin, P. Wilson, P. D. Griffiths, C. C. Davey, M. A. Johnson, and V. C. Emery. 1996. Cytomegalovirus (CMV) viraemia detected by polymerase chain reaction identifies a group of HIV-positive patients at high risk of CMV disease. AIDS 11:889–893. Demmler, G. J., G. J. Buffone, C. M. Schimbor, and R. A. May. 1988. Detection of cytomegalovirus in urine from newborns by using polymerase chain reaction DNA amplification. J. Infect. Dis. 158:1177–1184. Dodt, K. K., P. H. Jacobsen, B. Hofmann, C. Meyer, H. J. Kolmos, P. Skinhoj, B. Norrild, and L. Mathiesen. 1997. Development of cytomegalovirus (CMV) disease may be predicted in HIV-infected patients by CMV polymerase chain reaction and the antigenemia test. AIDS 11:F21–F28. Drouet, E., A. Boibieux, S. Michelson, R. Ecochard, F. Biron, D. Peyramond, R. Colimon, and G. Denoyel. 1993. Polymerase chain reaction detection of cytomegalovirus DNA in peripheral blood leukocytes as a predictor of cytomegalovirus disease in HIV-infected patients. AIDS 7:665–668. Grundy, J. E., A. Ehrnst, H. Einsele, V. C. Emery, H. Hebart, H. G. Prentice, and P. Ljungman. 1996. A three-center European external quality control study of PCR for detection of cytomegalovirus DNA in blood. J. Clin. Microbiol. 34:1166–1170. Hansen, K. K., A. Ricksten, B. Hofmann, B. Norrild, S. Olofsson, and L. Mathiesen. 1994. Detection of cytomegalovirus DNA in serum correlates with clinical cytomegalovirus retinitis in AIDS. J. Infect. Dis. 170:1271–1274. Hardy, W. D. 1996. Lessons learned from HIV pathogenesis and therapy: implications for better management of cytomegalovirus disease. AIDS 10(Suppl. 1):S31–S35. Jacobson, M. A., M. Zegans, P. R. Pavan, J. J. O’Donnell, F. Sattler, N. Rao, S. Owens, and R. Pollard. 1997. Cytomegalovirus retinitis after initiation of highly active antiretroviral therapy. 1997. Lancet 349:1443–1445. Krajden, M., P. Shankaren, C. Bourke, and W. Lau. 1996. Detection of cytomegalovirus in blood donors by PCR using the Digene SHARP signal system assay: effects of sample preparation and detection methodology. J. Clin. Microbiol. 34:29–33. Mazzulli, T. 1993. Improved diagnosis of cytomegalovirus infection by detection of antigenemia or use of PCR methods. Clin. Microbiol. News. 15:97–100. Mazzulli, T., R. H. Rubin, M. J. Ferraro, R. T. D’Aquila, S. A. Doveikis, B. R. Smith, T. H. The, and M. S. Hirsch. 1993. Cytomegalovirus antigenemia: clinical correlations in transplant recipients and in persons with AIDS. J. Clin. Microbiol. 31:2824–2827. Mazzulli, T., S. Wood, R. Chua, and S. Walmsley. 1996. Evaluation of the Digene hybrid capture system for detection and quantitation of human cytomegalovirus viremia in human immunodeficiency virus-infected patients. J. Clin. Microbiol. 34:2959–2962. Pertel, P., R. Hirschtich, J. Phair, J. Chmiel, L. Poggensee, and R. Murphy. 1992. Risk of developing cytomegalovirus retinitis in persons infected with the human immunodeficiency virus. J. Acquired Immune Defic. Syndr. 5: 1069–1074. Ranger-Rogez, S., J. P. Couty, T. Nicot, F. Denis, J. P. Rogez, A. Riche, and P. Weinbreck. 1996. Interest of quantitative pp65 cytomegalovirus antigenemia assay for human immunodeficiency virus-positive patients. J. Clin. Microbiol. 34:3255–3256. Rasmussen, L., S. Morris, D. Zipeto, J. Fessel, R. Wolitz, A. Dowling, and T. C. Merigan. 1995. Quantitation of human cytomegalovirus DNA from peripheral blood cells of human immunodeficiency virus-infected patients could predict cytomegalovirus retinitis. J. Infect. Dis. 171:177–182. Rose, D. N., and H. S. Sacks. 1997. Cost-effectiveness of cytomegalovirus (CMV) disease prevention in patients with AIDS: oral ganciclovir and CMV polymerase chain reaction testing. AIDS 11:883–887. Salzberger, B., C. Franzen, G. Fatkenheuer, O. Cornley, A. Schwenk, H. Rasokat, V. Diehl, and M. Schrappe. 1996. CMV-antigenemia in peripheral blood for the diagnosis of CMV disease in HIV-infected patients. J. Acquired Immune Defic. Syndr. 11:365–369. Shinkai, M., S. A. Bozzettee, W. Powderly, P. Frame, and S. A. Spector. 1997. Utility of urine and leukocyte cultures and plasma DNA polymerase chain reaction for identification of AIDS patients at risk for developing human cytomegalovirus disease. J. Infect. Dis. 175:302–308. Spector, S. A., G. F. McKinley, J. P. Lalezari, T. Samo, R. Andruczk, S. Follansbee, P. D. Sparti, D. V. Havlir, G. Simpson, W. Buhles, R. Wong, and M. J. Stempien. 1996. Oral ganciclovir for the prevention of cytomegalovirus disease in persons with AIDS. N. Engl. J. Med. 334:1491–1497. Wetherill, P. E., M. L. Landry, P. Alcabes, and G. Friedland. 1996. Use of a quantitative cytomegalovirus (CMV) antigenemia test in evaluating HIV1 patients with and without CMV disease. J. Acquired Immune Defic. Syndr. 12:33–37. Zipeto, D., M. G. Revello, E. Silini, M. Parea, E. Percivalle, M. Zavattoni, G. Milanesi, and G. Gerna. 1992. Development and clinical significance of a diagnostic assay based on the polymerase chain reaction for detection of human cytomegalovirus DNA in blood samples from immunocompromised patients. J. Clin. Microbiol. 30:527–530.
