Cytomegalovirus DNA concentration in plasma predicts development of cytomegalovirus disease in kidney transplant recipients



The clinical significance of cytomegalovirus (CMV) DNA detection in post-kidney transplantation infection surveillance was examined by comparing the performance of three assays for detection of CMV in blood: the test for CMV-pp65-antigen in leukocytes, which is routinely employed in our laboratory, the quantitative plasma CMV-DNA-polymerase chain reaction (PCR; Cobas Amplicor CMV Monitor test®) and the qualitative plasma CMV-DNA-PCR (Amplicor CMV test®). Thirteen kidney transplant recipients were monitored with serial samples taken over a period of 3 months following transplantation. The quantitative CMV-PCR was the test with highest sensitivity, 95.9%, vs. 88.9% and 76.9% for the CMV-pp65 antigen assay and qualitative CMV-PCR, respectively. The virus load in the first positive specimens, assessed as DNA-copies/mL, was significantly associated with CMV disease because five of the six patients who developed disease, but only one of the seven who did not develop disease, had more than 3000 CMV-DNA-copies/mL. The number of CMV-pp65 antigen-positive cells in the first positive specimens did not have predictive value for development of CMV disease. Assessment of CMV in plasma by the quantitative CMV-PCR is especially useful since it has a high sensitivity and the amount of CMV DNA in plasma is a good predictor of CMV disease.

Cytomegalovirus (CMV) is still a major pathogen in organ transplant recipients despite the implementation of new diagnostic tools and improved strategies for anti-CMV prophylaxis and treatment. In a recent study on CMV infections in 477 kidney transplant recipients, we showed that among patients at risk for primary CMV infection, 68% acquired CMV infection, and of these, 81% developed CMV disease [1]. In patients at risk for CMV-reactivated infection, 69% had a reactivated CMV infection but only 26% developed CMV disease. This illustrates one of the major problems in the management of CMV infections in transplant recipients. Only a certain proportion of patients at risk will have an active CMV infection and even fewer will develop CMV disease.

Surveillance of CMV infections in transplant recipients is usually based on early detection of CMV in blood. This is based on the fact that generalized CMV endothelitis is a common feature of CMV infections in immunosuppressed patients and that CMV is either released into plasma or is picked up directly from infected endothelial cells by the granulocytes [2]. The CMV load in blood is assessed by applying quantitative methods and may be useful in the surveillance of CMV infections [3].

Recently, standardized, semiautomated CMV polymerase chain reaction (PCR) tests have been developed which have proved useful in monitoring CMV infection in a variety of organ transplantation settings [4,5].

The present study was undertaken in order to compare our routine test for quantitative assessment of CMV-pp65-antigen (Ag) in leukocytes with a quantitative CMV-DNA-PCR (Cobas Amplicor CMV Monitor®; Roche Diagnostics, Basle, Switzerland) and a qualitative CMV-DNA-PCR (Amplicor CMV®; Roche Diagnostics) for assessment of CMV DNA in plasma. We intended to evaluate the clinical utility of these assays in the surveillance of CMV infections in a cohort of kidney transplant recipients.

In our hospital, kidney transplant recipients are tested weekly for CMV-pp65-Ag in leukocytes during the first 3 months after transplantation. In the period from September 1998 to the end of March 1999 a parallel sample of ethylenediaminetetraacetic acid-plasma for CMV-PCR was obtained at the same time-points and immediately frozen and stored at −70 °C. The CMV-pp65-Ag test was positive at least once in 39 consecutive first kidney transplant recipients. None of these patients received ganciclovir prophylaxis or pre-emptive therapy.

A quantitative and qualitative CMV-PCR was performed on 120 plasma samples from 13 of the 39 patients who had CMV disease (n = 6) or asymptomatic CMV infection (n = 7) and were positive for CMV-pp65-Ag in at least three consecutive specimens. CMV-PCR was performed on plasma samples taken at the same time as the CMV-pp65-Ag positive samples. In addition, at least one sample that was taken before the first and after the last CMV-pp65-Ag-positive sample was examined. The median number of samples per patient was eight (range 5–18). The patients had given their informed consent to participate in the study and the Regional Ethics Committee had approved the study.

CMV disease was defined as the detection of CMV in a blood specimen accompanied either by CMV syndrome with fever, muscle pain or leukopenia and/or thrombocytopenia (other causes excluded) or by organ involvement [1].

The assay for CMV-pp65-Ag in leukocytes was performed as a modification of a reported procedure [4]. The results were given as number of CMV-pp65-Ag-positive cells per 105 leukocytes.

The Amplicor CMV® and Cobas Amplicor CMV Monitor® tests are based on the amplification of the virus DNA-polymerase gene using biotinylated primers flanking a 365-base pair conserved sequence. Amplicons are then captured by specific hybridization on microwell plates (Amplicor CMV®) or magnetic beads (Cobas Amplicor CMV Monitor®) and detected by colorimetric reaction [5,6]. A 200-μL aliquot of plasma is processed to isolate CMV DNA after lysis of circulating viral particles.

In the qualitative Amplicor CMV® test, the presence or absence of CMV DNA is established by comparing the sample O.D. with the negative cut-off value, while in the quantitative Cobas Amplicor CMV Monitor® test, the amount of CMV DNA is automatically calculated and expressed in copies/mL over a 103−106 copies/mL dynamic range. A quantification standard, consisting of a synthetic DNA identical to the target except for a specific probe-binding sequence, is added in known amounts to the sample prior to DNA extraction.

Cobas Amplicor CMV Monitor® and Amplicor CMV® have analytical and clinical sensitivities of five copies/PCR reaction and 20 viral particles/PCR reaction, respectively. Both assays make use of Amperase® (Uracyl-N-glycosyilase) (Roche Diagnostics) to prevent carry-over contamination and use an internal control to ensure result integrity.

Non-parametric methods were generally used for statistical analysis. Distribution of variables within groups is generally presented as median values and ranges. The Mann–Whitney U-test calculated quantitative differences of specific variables between groups. Quantitative differences for time-dependent changes of the same variable were tested by the Wilcoxon matched pairs test. Kendall's τ (tau) was used as correlation coefficient. Two-by-two frequency tables were tested either by using χ2 or χ2 with Yates correction for lower observation numbers. All tests were considered two-tailed.

In immunocompromised patients clinical diagnosis of CMV disease is difficult and has to be supported by detection of CMV in specimens from the organs affected or, alternatively, by detection of CMV in blood specimens. Usually CMV can be detected in blood days or weeks before the onset of symptoms [1]. By monitoring serial blood samples for the presence of CMV, emerging infections may be treated before disease development, i.e. pre-emptive therapy, or at the early onset of disease, i.e. deferred therapy.

The sensitivity of the tests was compared in several ways. When CMV positivity was defined as a time-point where at least one of the tests were positive (99/120), 95 (95,9%) were positive with the quantitative CMV-PCR, 88 (88.9%) were positive with the CMV-pp65-Ag assay and 76 (76.8%) were positive with the qualitative CMV-PCR. Both the quantitative CMV-PCR and the CMV-pp65-Ag assay were more sensitive than the qualitative CMV-PCR (P < 0.05). The quantitative CMV-PCR tended to be more sensitive than the CMV-pp65-Ag assay; however, the difference was not statistically significant (0.05 < P < 0.1).

In this study the quantitative CMV-PCR detected 6.25 copies per PCR reaction (250 DNA-copies/mL) which is very close to the detection limit of five copies per PCR reaction given by the manufacturer. The CMV-pp65-Ag assay was positive for all blood samples containing more than 3480 CMV DNA-copies/mL, whereas among 43 samples harboring 249–2429 CMV DNA-copies/mL plasma, 33 (76,7%) blood samples were positive for CMV-pp65-Ag. Four specimens with CMV-pp65-Ag values ranging from 1 to 22 positive cells/105 leukocytes, were negative with the quantitative CMV-PCR. On the other hand, among 32 CMV-pp65-Ag-negative samples, 12 samples were positive with the quantitative CMV-PCR. However, there was a high degree of correlation between the quantitative tests for CMV-pp65-Ag and CMV-DNA when parallel specimens were tested and CMV was detected in at least one of the tests [τ = 0.55, P < 0.001 (n = 99)]. The good correlation between the quantitative CMV-PCR and the CMV-antigenemia test is in agreement with the findings of Caliendo et al. [7].

To get an early indication of CMV disease, the assays for detection of CMV in blood have to be positive, before, or at least when the symptoms appear. In our patient population the first positive CMV test appeared approximately 45 (24–54) days after transplantation for primary CMV infection and approximately 30 (13–44) days for reactivated infection (Table 1). However, there was a statistically significant difference only for the qualitative CMV-PCR (P < 0.05). Among the six patients with CMV disease, CMV was found in blood by all three methods before or at the time of disease manifestation. There was no statistically significant difference between the tests with regard to time to appearance of first positive test after transplantation. This was true both for primary infection, reactivated infection, asymptomatic infection and symptomatic infection (Table 1).

Table 1.  Characteristics of first CMV-positive blood specimen from 13 kidney transplant recipients. CMV-pp65-Ag in leukocytes (CMV-pp65-Ag), quantitative CMV-PCR (CMV-Qn-PCR) and qualitative CMV-PCR (CMV-Ql-PCR) were used for assessment of CMV. The results are given as median and range
Category of patientsNo. of patientsFirst positive test (days after transplant)Viral load of first positive test
  • a CMV-pp65-Ag-positive cells/105 leukocytes.

  • b

    CMV-DNA-copies/mL plasma.

  • c

    P = 0.023 vs. asymptomatic CMV infection.

All patients1329 (13–54)31 (13–54)33 (13–54)8 (1–612)1850 (249–18 400)
Primary CMV infection541(24–54)48 (27–54)48 (27–54)22 (2–650)4210 (305–18 400)
Reactivated CMV infection827 (13–41)30 (13–37)29 (13–44)8 (1–21)607 (249–11 400)
Asymptomatic CMV infection733 (20–41)29 (19–48)33 (19–48)8 (2–22)344 (249–4210)
CMV disease628 (13–54)34 (13–54)32 (13–54)9 (1–650)5430 (309–18 400)c

We further tested whether the viral load of the first positive specimen or the maximum viral load was predictive of later development of symptomatic infection. As shown in Table 1, the load of CMV DNA in plasma in the first positive specimen of consecutive post-transplant specimens was considerably higher in patients who developed CMV disease (P = 0.022), whereas this was not the case for the first positive CMV-pp65-Ag test. Five out of six patients who developed CMV disease but only one out of seven who did not develop CMV disease, had more than 3000 CMV DNA-copies/mL in the first positive specimen.

Maximum viral load and the time of its appearance after transplantation is often used to characterize a CMV infection [8]. In our study, however, no differences were seen in maximal virus loads between those developing CMV disease and those who were asymptomatic when employing any of the two quantitative CMV tests (results not shown). In this and previous studies we have found that the systemic virus load shows a high variability also among patients who develop CMV disease so that no virus load can be used as a cut-off to predict the development of CMV disease. Several factors other than virus load are of importance for the development of CMV disease. Nordøy et al. showed that the cytokine profile after transplantation is highly predictive of CMV disease [9]. High levels of interleukin-8 and low levels of macrophage inflammatory protein 1 were recorded among those who developed disease, even before the onset of CMV infection. A more reliable prediction of CMV disease may thus be achieved by combining tests for viral load and for a relevant cytokine.

In conclusion, both the quantitative (Cobas Amplicor CMV Monitor test®) and the qualitative (Amplicor CMV test ®) CMV-DNA-PCR, as well as the quantitative CMV-pp65-Ag assay have the sensitivity required to detect CMV in blood before the onset of CMV disease in kidney transplant recipients. A high level of CMV DNA in plasma, but not the number of CMV-pp65-positive leukocytes, in the first positive specimen of sequential specimens, was highly predictive of later development of CMV disease.