Comparison of Cytomegalovirus Antigenemia and Shell Vial Culture in Allogeneic Marrow Transplantation Recipients Receiving Ganciclovir Prophylaxis

Comparison of cytomegalovirus antigenemia and shell vial culture in allogeneic marrow transplantation recipients receiving ganciclovir prophylaxis

VA Nicholson1, E Whimbey1, R Champlin2, D Abi-Said1, D Przepiorka2, J Tarrand3, K Chan4, GP Bodey1 and JM Goodrich5

1Section of Infectious Disease, 2Department of Hematology, and 3Division of Laboratory Medicine, 4Division of Pediatrics, The University of Texas MD Anderson Cancer Center, Houston, TX; and 5Division of Infectious Diseases, Southern Illinois University School of Medicine, Springﬁeld, IL, USA

Summary: be diagnosed with CMV disease coincident with or prior to having a positive CMV surveillance culture.1,7–9 Prophy- We prospectively monitored 61 allogeneic BMT patients lactic strategies result in administration of antiviral drugs for evidence of CMV infection and disease starting 7 to 30–40% of patients who would not have CMV infection days prior to transplant until day 110 after transplant. or disease.1,5 Patients receiving pre- and post-transplantation gan- To employ effectively an early treatment strategy for ciclovir prophylaxis were followed for the incidence of CMV infection requires a diagnostic test which is rapid and infection by the CMV antigenemia assay and shell vial risk stratifies patients for subsequent disease. Previously, cultures. The median age of all patients was 32 years centrifugation cultures have been used as a marker of CMV (range 5–54 years). Fourteen (25%) of 57 evaluable replication to initiate antiviral therapy in MT patients but patients became CMV antigenemia or culture positive. have lacked sensitivity.1,8 The CMV antigen assay appears The incidence of culture or antigenemia positivity in to be more sensitive than viral culture10–13 and provides CMV seropositive or seronegative patients with a sero- diagnostic results in a few hours as opposed to culture tech- positive donor was 29% (14 of 49 patients). The anti- niques which may take a few days to weeks.10,14,15 The genemia assay became positive a median of 29 days assay is based on the detection from the peripheral blood (range 12–89 days) after BMT as compared to 46 days of patients to a CMV late matrix tegument protein pp65.16,17 (range 26–98 days) by shell vial assay (P Ͻ 0.001). There The assay depends on the use of a pool of monoclonal anti- were no cases of CMV disease in the first 110 days after bodies directed against epitopes on this protein. In this transplant. This study demonstrates that despite the use study we investigated the use of the CMV antigen assay to of prophylactic ganciclovir, BMT patients developed monitor the development of CMV infection in allogeneic CMV infection but did not progress to disease in this BMT patients receiving prophylactic ganciclovir for the study, the CMV antigenemia assay may be used to prevention of CMV disease. The assay was compared with monitor for CMV infection during prophylaxis, and the viral culture and the centrifugation assay for sensitivity current regimens for CMV prophylaxis with ganciclovir and specificity. may require further evaluation to determine an optimal regimen to prevent CMV infection. Keywords: CMV; bone marrow transplant; antigenemia; Patients and methods diagnosis; viremia Patients The study group was comprised of 61 consecutive patients Cytomegalovirus (CMV) remains an important opportu- undergoing allogeneic MT between December 1992 and nistic infection following allogeneic BMT.1 Recently, September 1993 (Table 1). Patients were followed from day effective strategies for the prevention of CMV disease have −7 to day 100 after transplant. Blood was obtained weekly been developed in which ganciclovir has been given to for virus culture and the antigenemia assay. Urine cultures BMT patients either prophylactically (to all CMV seroposi- were collected weekly. Virus cultures were obtained from tive patients)1–6 or as early treatment (evidence of active other clinically relevant sites as needed (throat, tracheal CMV replication).1,7,8 Each of these strategies has limi- aspirate and bronchoalveolar lavage). tations. Twelve to 32% of patients given an antiviral based on an early treatment strategy (pre-emptive therapy based on a positive surveillance culture from blood or BAL) will CMV prophylaxis and treatment From March 1991 to April 1992, ganciclovir was given Correspondence: Dr J Goodrich, SIU School of Medicine, PO Box 19230, three times per week beginning 30 days post-transplant. In Springfield, Illinois 62794-1311, USA May 1992, patients who received T cell-depleted marrow Received 1 February 1996; accepted 18 July 1996 (TCD) were given ganciclovir prophylaxis five times per CMV antigenemia and allogeneic BMT VA Nicholson et al 38 Table 1 Characteristics of study population CMV antigenemia assay

Characteristic Number CMV culture Number CMV culture Peripheral blood leukocytes (PBL) were isolated from or antigenemia or antigenemia 10 ml heparinized blood by the dextran sedimentation positive (%) negative (%) method. PBL were suspended in PBS at a concentration of 1.5 × 106 cells/ml and cytospin preparations were made. Male/Female 9/5 31/12 Slides were fixed in acetone, air-dried, and stored at 4°C Age median (range) 32 (9–48) 34 (5–54) if not stained within 24 h. Slides were stained with 35 ␮lof Underlying disease CML 4 (29) 11 (26) C10 and C11 monoclonal antibody directed against CMV AML 7 (50) 15 (35) phosphoprotein pp65 (Clonab, Biotest). Slides were coun- Myelodysplastic 0 (0) 3 (7) terstained with FITC-conjugated goat anti-mouse IG F(ab) syndrome (Biotest, Denville, NJ, USA) diluted in PBS 1% human CLL 1 (7) 4 (9) albumin and 0.0005% Evans blue. Positive cells showed a Lymphoma 1 (7) 3 (7) Aplastic anemia 0 (0) 1 (2) yellow–green nuclear staining. Positive controls consisted ALL 1 (7) 6 (14) of infected fibroblast or leukocytes from patients known to Donor be CMV antigenemia positive. Matched related 9 (64) 26 (60) Matched unrelated 2 (14) 12 (28) Mismatched 3 (2) 5 (12) CMV serology (D/R) Results Pos/Pos 11 (78) 18 (42) Pos/Neg 1 (7) 8 (19) Neg/Pos 2 (14) 9 (21) Sixty-one patients were enrolled in the study but four Neg/Neg 0 (0) 8 (19) patients were excluded from the analysis. Two patients Graft-versus-host excreted CMV prior to the start of monitoring, one patient disease died during conditioning therapy, and one patient was lost 0–1 6 (43) 23 (53) 2–4 8 (57) 20 (46) to follow-up prior to obtaining blood for culture and antigenemia assays. Fifty-seven patients were evaluated for the CML = chronic myelogenous leukemia; AML = acute myelogenous leuke- final analysis (Table 1). Eight patients who were CMV mia; CLL = chronic lymphocytic leukemia; ALL = acute lymphocytic leu- seronegative with seronegative donors were included. kemia. Forty-three (75%) of the patients remained culture and antigen negative and represented the control group. There were no significant differences between the culture and antigen negative and culture and antigen positive groups, except for positive donor recipients pairs (P = 0.02). week. Beginning in October 1992, most patients received Fourteen (25%) of 57 patients (29% of 49 CMV sero- the five times per week regimen unless the attending phys- positive donor/patient combinations) became culture or ician directed differently. At day −1, patients were switched antigenemia positive (Table 2). Eleven of the 14 patients to acyclovir 500 mg/m2 every 8 h until engraftment (ANC had a positive viral culture at some site (blood = 9, у1 × 109/l for 2 consecutive days). Three patients received urine = 1, sputum = 1). Thirteen of the fourteen patients ganciclovir 5 mg/kg three times per week (Table 2). All were positive for CMV antigenemia. Ten of the 13 patients patients received intravenous immunoglobulin 500 mg/kg who were antigenemia positive had corresponding positive weekly from admission to day 110. Patients who became virus cultures. All patients who were culture positive were CMV culture positive received ganciclovir treatment at a antigenemia positive. The exception was a patient who had dose of 5 mg/kg twice per day or foscarnet 60 mg/kg three a positive sputum culture on day 45 after transplant who times per day. The decision to use foscarnet was made by remained antigenemia negative. The median time to CMV the attending physician and was based on ganciclovir toxi- antigenemia or culture positivity was 29 days (range 12– city (neutropenia or thrombocytopenia). The treatment dur- 89 days) and 46 days (range 26–98 days), respectively (P ation was determined by the clinical judgement of the Ͻ 0.0001, paired log rank test) (Figure 1). All nine patients attending physician. Patients who were CMV seronegative who were blood culture positive were antigenemia positive. and received a marrow from a CMV seronegative donor A positive antigenemia test preceded all positive blood received CMV seronegative blood products. shell vial cultures by a median of 14 days. One patient who was antigenemia positive on day 89 after transplant had a positive urine culture on day 98 after transplant but Statistical evaluation remained blood culture negative. Three patients were antigenemia positive on days 12, 26 The statistical analysis was performed using the ␹2 test and and 29 after transplant without having a positive CMV cul- a two-tailed Fisher exact test as appropriate for examination ture from any site. Two of the three patients were on inad- of differences in proportions, and the Mann–Whitney test equate anti-CMV prophylaxis (acyclovir and ganciclovir for evaluation of differences between medians. Time-to- three times per week) when they became antigenemia event variables were estimated according to the method of positive. Both patients became antigenemia negative upon Kaplan and Meier18 and compared using paired and the institution of ganciclovir five times per week. One unpaired log rank test, as appropriate. patient did develop antigenemia on ganciclovir prophylaxis CMV antigenemia and allogeneic BMT VA Nicholson et al 39 Table 2 Cytomegalovirus antigenemia, virus culture, and antiviral therapy

Patient Antigenemia Drug (days on antiviral Culture (site day Drug (days on antiviral therapy (day after transplant) therapy before positive after transplant before positive culture) test)

1 yes (33) GCV (3) yes (blood, 47) GCV (14) 2 yes (28) GCV (10) yes (blood, 38) GCV (25) 3 yes (56) None. Off all antiviral therapy yes (blood, 59) None 7 days prior to +Ag 4 yes (33) GCV (25) yes (blood, 62) None Discontinued 2 days prior to + culture. 5 yes (34) GCV (12) yes (blood, 48) GCV (26) 6 yes (89) FOS (5) yes (urine, 98) None, FOS discontinued. 7 yes (29) GCV (1) no GCV 8 yes (26) GCV (2) no GCV 9 yes (81) GCV (7) yes (blood, 105) None off all antivirals. 10 no ACV (47) yes (Tracheal aspirate, 45) ACV. Patient did not engraft. 11 yes (21) ACV (5) Persistently Ag + yes (blood, 34) FOS (5) On FOS QOD due to until started on GCV. increased creatinine. Started on GCV and became Ag and culture negative. 12 yes (12) ACV (21) no GCV. Became Ag negative. 13 yes (16) ACV (16) Persistently Ag + until yes (blood, 30) None. Taken off GCV due to started on GCV. seizures, restarted after + blood culture. 14 yes (14) ACV (15) Persistently + on yes (blood, 34) GCV (3) Remained + until ACV given FOS and became − for Ag and culture.

GCV = ganciclovir; ACV = acyclovir; FOS = foscarnet; Ag = antigenemia; +=positive; −=negative. Patients 4, 5 and 8 received ganciclovir prophylaxis three times per week.

0.4 specificity of antigenemia could not be determined for CMV disease. However, the sensitivity and specificity could be determined for a positive CMV culture. The sensitivity of antigenemia for a positive viral culture at any site 0.3 was 90% with a specificity of 93%. The positive and negative predictive values were 77% and 98% respectively. Since viremia may be predictive for subsequent CMV dis- 0.2 ease, an analysis was done with only patients who were viremic. The sensitivity and specificity of CMV antigenemia for predicting subsequent viremia was 100% and 91%

Proportion positive respectively. The positive and negative predictive values 0.1 were 69% and 100% respectively.

Outcome of antigenemia positive patients 0.0 –10 10 30 50 70 90 110 CMV antigenemia was detected in four patients at 12, 14, Days post-marrow transplant 16 and 21 days post-transplant while still on acyclovir. One had resolution of antigenemia following institution of gan- Figure 1 Kaplan–Meier product limit estimate of developing a positive ciclovir prophylaxis. The other three patients developed CMV antigenemia (–––) vs CMV centrifugation culture (–––) in allogeneic BMT patients. positive cultures for CMV 12–14 days after detection of antigenemia and were treated with therapeutic doses of ganciclovir (5 mg/kg twice per day). five times per week and subsequently became negative after CMV antigenemia was detected in eight patients on gan- ganciclovir was discontinued for seizures. None of the eight ciclovir or foscarnet prophylaxis. Two patients remained patients who were CMV seronegative with seronegative culture negative. One of these became antigenemia negative donors became CMV culture positive or antigenemia after 14 days even though the ganciclovir was discontinued positive. when the patient had a seizure. The other became antigenemia negative when the ganciclovir prophylaxis was increased from three times per week to five times per week; Sensitivity and specificity interestingly, antigenemia recurred 1 month later in this None of the patients in this study were diagnosed with patient and resolved without any further changes in therapy. CMV disease despite being viremic. The sensitivity and Six patients who were CMV antigenemia-positive sub- CMV antigenemia and allogeneic BMT VA Nicholson et al 40 sequently developed positive cultures for CMV and were while receiving acyclovir, was switched to ganciclovir and treated. The CMV antigen positive patients were on became antigenemia negative. Other studies in BMT, liver, prophylactic ganciclovir a median of 5 days (range 1–12 kidney, and heart transplants have mentioned a dissociation days) before becoming antigen positive. between cultures and antigenemia similar to what was Two antigenemia-positive and 17 antigenemia-negative found in our study.11,12,23,25 Whether this dissociation patients died in the study prior to day 110 (log rank test, reflects an increased sensitivity of antigenemia for CMV P = 0.18). The primary causes of death were regimen replication or false positivity remains to be determined. related toxicity (n = 1), relapse of underlying disease Finally, none of the patients in this study infected with (n = 6), disseminated aspergillus (n = 4), Candida krusei CMV progressed to CMV disease. This raises the question sepsis (n = 3), Torulopis glabrata sepsis (n = 2), RSV pneu- about the need to monitor patients during antiviral prophy- monia (n = 2), and liver failure (n = 1). No deaths were laxis. In this study, we did not monitor for emergence of attributable to CMV disease. ganciclovir resistance. Ganciclovir resistance has been described in immunocompromised hosts receiving ganciclovir.26–28 Persistent CMV infection due to antiviral Discussion resistance may be missed without a system of monitoring. In addition, our study is limited owing to the small number In this study, we have detailed our initial experience with of patients who became infected. In a larger study of 216 the CMV antigenemia assay in allogeneic BMT patients allogeneic marrow transplant patients, some patients who receiving prophylactic antivirals. Rapid methods for the excreted CMV while on ganciclovir prophylaxis did pro- diagnosis of CMV have become clinically important since gress to CMV disease. These patients had risk factors sug- the advent of effective antivirals and the discovery and con- gesting a greater depth of immunosuppression as shown firmation of the predictive value for CMV disease by the by their progression to CMV disease despite prophylactic recovery of CMV from the blood or bronchoalveolar fluid therapy (J Goodrich and E Whimsey; personal in allogeneic BMT patients.7,8,19,20 The current clinical stan- communication). These data suggest a need at least to dard has been the centrifugation assay for rapid detection monitor certain patients at high risk for progression to of CMV from a variety of sites.21,22 CMV disease. We have found the antigenemia assay to be a sensitive In conclusion, the antigenemia assay was sensitive, with and specific assay for the detection of CMV in BMT a relatively high positive and negative predictive value. All patients. The sensitivity of antigenemia for a positive cul- patients who were viremic had a positive antigenemia test ture at any site was 90% with a specificity of 93%. The a median of 14 days before a positive culture suggesting one patient who did not have a positive antigenemia test the test would be adequate for monitoring patients. The had a positive culture from an endotracheal aspirate but did question of false positive results remains to be resolved not die of CMV disease. Since all patients in this study definitively. received prophylactic antivirals, we were unable to answer the question of the sensitivity and specificity of antigenemia for CMV disease. Viremia has been shown to be a marker Acknowledgements for CMV disease in BMT patients.8,20 The positive predictive value of antigenemia for CMV viremia was 69% which We wish to acknowledge the technical help of Sandhya Misra. is comparable to another study.23 Additionally, all patients with viremia were positive by antigenemia an average of 14 days prior to the first positive centrifugation culture pro- References viding the opportunity to institute antiviral therapy. In cases of CMV antigenemia while being given ganciclovir, most 1 Goodrich JM, Boeckh M, Bowden R. Strategies for the pre- patients had been on ganciclovir or foscarnet for less than vention of cytomegalovirus disease after marrow transplan- 7 days on a five times per week regimen. A previous study tation. Clin Infect Dis 1994; 19: 287–298. has suggested that it may take up to 2 weeks before a 2 Atkinson K, Downs K, Golenia M et al. Prophylactic use of ganciclovir in allogeneic bone marrow transplantation: patient started on ganciclovir seven times per week will 8 absence of clinical cytomegalovirus infection. Br J Haematol become CMV culture negative. In the cases where patients 1991; 79: 57–62. remained persistently antigenemia positive and become cul- 3 Yau JC, Dimopoulos MA, Huan SD et al. Prophylaxis of cyto- ture positive, an increase in both ganciclovir frequency and megalovirus infection with ganciclovir in allogeneic marrow dose terminated viremia. This suggests a threshold for gan- transplantation. Eur J Haematol 1994; 47: 371–376. ciclovir antiviral effect which may vary due to the level of 4 von Bueltzingsloewen A, Bordigoni P, Witz F et al. Prophy- immunosuppression or other factors.24 Two of the patients lactic use of ganciclovir for allogeneic bone marrow transplant who did not clear their antigenemia were pediatric patients recipients. Bone Marrow Transplant 1993; 12: 197–202. ages 9 and 10 while in our experience it is less common 5 Goodrich JM, Bowden R, Kellar C, Meyers JD. Prevention of for adults to be persistently infected. cytomegalovirus disease after allogeneic marrow transplant by ganciclovir prophylaxis. Ann Intern Med 1993; 117: 173–178. Three of the patients had a positive antigenemia test 6 Winston DJ, Ho WG, Bartoni K et al. Ganciclovir prophylaxis without having a positive culture. The reason for this is not of cytomegalovirus infection and disease in allogeneic marrow clear. Two of the three patients were antigenemia positive transplants. Ann Intern Med 1993; 117: 179–184. on the first and second day of ganciclovir prophylaxis, 7 Schmidt GM, Horak DA, Niland JC et al. A randomized, con- respectively. The third patient became antigenemia positive trolled trial of prophylactic ganciclovir for cytomegalovirus CMV antigenemia and allogeneic BMT VA Nicholson et al 41 pulmonary infection in recipients of allogeneic bone marrow 19 Meyers JD, Flournoy N, Thomas ED. Risk factors for cyto- transplants. New Engl J Med 1991, 324: 1005–1011. megalovirus infection after human marrow transplantation. 8 Goodrich J, Mori M, Gleaves C et al. Early treatment with J Infect Dis 1986; 153: 478–488. ganciclovir to prevent cytomegalovirus disease after allo- 20 Meyers JD, Ljungman P, Fisher LD. Cytomegalovirus geneic marrow transplant. New Engl J Med 1991; 325: excretion as a predictor of cytomegalovirus disease after mar- 1601–1607. row transplantation: importance of cytomegalovirus viremia. 9 Eisele H, Ehninger G, Hebart H et al. Polymerase chain reac- J Infect Dis 1990; 162: 373–380. tion monitoring reduces the incidence of cytomegalovirus dis- 21 Gleaves CA, Smith TF, Shuster EA, Pearson GR. Rapid detec- ease and the duration and side effects of antiviral therapy. tion of cytomegalovirus in MRC-5 cells inoculated with urine Blood 1995; 86: 2815–2820. specimens by using low-speed centrifugation and monoclonal 10 van der Bij W, Schirm J, Torensma R et al. Comparison antibodies to an early antigen. J Clin Microbiol 1984; 19: between viremia and antigenemia for the detection of cyto- 917–919. megalovirus in blood. J Clin Microbiol 1988; 26: 2531–2535. 22 Gleaves CA, Reed EC, Hackman RC, Meyers JD. Rapid diag- 11 Boland GJ, de Gast GC, Hene RJ et al. Early detection of nosis of invasive cytomegalovirus infection by examination of active cytomegalovirus (CMV) infection after heart and kid- tissue specimens in centrifugation culture. Am J Clin Pathol ney transplantation by testing for early antigenemia and influ- 1987; 88: 354–358. ence of cellular immunity on the occurrence of CMV infec- 23 Boeckh M, Bowden RA, Goodrich JM et al. Cytomegalovirus tion. J Clin Microbiol 1990; 28: 2069–2075. antigenemia detection in peripheral blood leukocytes after 12 van der Berg AP, Klompmaker IJ, Haagsma EB et al. Anti- allogeneic marrow transplantation. Blood 1992; 80: 1358– genemia in the diagnosis and monitoring of active cytomegalovirus infection after liver transplantation. J Infect Dis 1991; 1364. 164: 265–270. 24 Atkinson K, Arthur C, Bradstock K et al. Prophylactic gan- 13 Gerna G, Zipet D, Para M et al. Monitoring of human cyto- ciclovir is more effective in HLA-identical family member megalovirus infections and ganciclovir treatment in heart marrow transplant recipients than in more heavily immune- transplant recipients by determination of viremia, antigenemia, suppressed HLA-identical unrelated donor marrow transplant and DNAemia. J Infect Dis 1991; 164: 488–498. recipients. Bone Marrow Transplant 1995; 16: 401–405. 14 van der Bij W, Torensma R, VanSon WJ et al. Rapid immuno- 25 Gerna G, Zipeto D, Parea M et al. Monitoring of human cyto- diagnosis of active cytomegalovirus infection by monoclonal megalovirus infection and ganciclovir treatment in heart trans- antibody staining of blood leukocytes. J Med Virol 1988; 25: plant recipients by determination of viremia, antigenemia, and 179–188. DNAemia. J Infect Dis 1991; 164: 488–498. 15 The TH, van der Ploeg M, van den Berg AP et al. Direct 26 Erice A, Chou S, Biron KK et al. Progressive resistance due to detection of cytomegalovirus in peripheral blood leukocytes. ganciclovir-resistant cytomegalovirus in immunocompromised A review of the antigenemia assay and polymerase chain reac- patients. New Engl J Med 1989; 320: 289–293. tion. Transplantation 1992; 54: 193–198. 27 Knox KK, Drobyski WR, Carrigan DR. Cytomegalovirus iso- 16 Grefte JMM, van der Gun BTF, Schmolke S et al. Cytomega- late resistant to ganciclovir and foscarnet from a marrow trans- lovirus antigenemia assay: identification of the viral antigen plant patient. Lancet 1991, 337: 1292–1293. as the lower matrix protein. J Infect Dis 1992; 166: 683–684. 28 Boivin G, Chou S, Quirk MR et al. Detection of ganciclovir 17 van der Giessen M, The TH, van Son WJ. Cytomegalovirus resistance mutations and quantitation of cytomegalovirus antigenemia assay. J Clin Microbiol 1991; 12: 2909–2910. (CMV) DNA in leukocytes of patients with fatal disseminated 18 Kaplan EL, Meier P. Nonparametric estimation from incom- CMV disease. J Infect Dis 1996; 173: 523–528. plete observations. J Am Stat Assoc 1958; 53: 457–481.