Clinical Utility of Cytomegalovirus (CMV) Serology Testing in High-risk CMV D+/R− Transplant Recipients


*Corresponding author: Atul Humar,


Late-onset cytomegalovirus (CMV) disease is a significant problem in D+/R− solid organ transplant (SOT) patients who receive antiviral prophylaxis. We assessed the clinical utility of CMV IgG and IgM serology testing for predicting late-onset CMV disease. We evaluated 352 D+/R− transplant recipients who participated in a trial comparing 100 days of ganciclovir versus valganciclovir prophylaxis. CMV serology was assessed on day 28, 56, 100, and 6 and 12 months post-transplant. IgG seroconversion occurred in 26.9% of patients by day 100, and in 63.4% and 75.3% by 6 and 12 months, respectively. IgM seroconversion occurred in 8.3%, 41.8% and 54.9% by day 100, month 6 and month 12, respectively. Seroconversion by day 100 (end of prophylaxis) was not predictive of subsequent CMV disease (CMV disease 13.3% if seropositive vs. 17.8% if seronegative; p = NS). However, at 6 months post-transplant, IgG serostatus was predictive of subsequent CMV disease between month 6 and 12 (CMV disease 1.3% if seropositive vs. 10.0% if seronegative; p = 0.002). In D+/R− patients, CMV serology testing is for the most part not clinically useful for predicting subsequent disease. However, seroconversion by 6 months may be useful for identifying patients at risk of late-onset CMV disease.


Cytomegalovirus (CMV) disease continues to be a common problem in solid organ transplant (SOT) recipients. CMV may result in viral syndromes or tissue invasive disease and may have indirect effects on graft function (1,2). One of the most important risk factors for CMV disease is the pre-transplant donor (D) and recipient (R) serostatus (3). Patients who are CMV D+/R− are at the highest risk for CMV disease post-transplant. Other recognized risk factors include the use of more potent immunosuppression, and specifically anti-lymphocyte antibody preparations (4). In the absence of prophylaxis, CMV disease may occur in up to 60% and 80% of D+/R− kidney and liver transplants recipients, respectively (1,2).

The high incidence of CMV disease in the D+/R− subgroup of patients has led to the widespread use of antiviral prophylaxis in this subpopulation. In randomized controlled trials, a 3-month course of intravenous or oral ganciclovir or valganciclovir has demonstrated efficacy for the prevention of CMV disease (5–7). Rates of CMV disease while on prophylaxis are very low. However, late-onset CMV disease (i.e., disease occurring beyond 3 months post-transplant) has become a common problem at many centers. In D+/R− SOT recipients who received antiviral prophylaxis for 100 days, CMV disease developed in approximately 13% and 18% of patients by 6 months and 1 year post-transplant, respectively (7). After 3 months post-transplant, many patients may not be followed as closely by their primary transplant centers, which compound the difficulty in diagnosis and management of patients who develop late-onset disease, and may result in unfavorable outcomes.

Simple methods to identify the risk of late-onset CMV disease in D+/R− patients receiving antiviral prophylaxis would be clinically useful. Seroconversion in a previously negative transplant recipient may imply the development of some degree of immunity to CMV and could be associated with a lower risk of subsequent CMV disease. It has been suggested that serological testing in such individuals may prove useful in identifying patients at risk of late-onset CMV disease. To answer this question, we assessed the clinical utility of regular CMV IgG and IgM testing in a large cohort of D+/R− SOT recipients receiving 3 months of antiviral prophylaxis.

Materials and Methods


Patients in this study were enrolled as part of a randomized, double-blind, double-dummy clinical trial comparing 3 months of valganciclovir with oral ganciclovir as primary prophylaxis for CMV in D+/R− SOT recipients. Inclusion and exclusion criteria are described in detail elsewhere (7). Informed consent was obtained from all patients for participation in the trial and for CMV serology and viral load testing at regular intervals. The study was approved by the Independent Ethics Committees/Institutional Review Boards of participating centers and was conducted in accordance with the Declaration of Helsinki.

Study design

Patients receiving a first heart, liver (including liver-kidney), kidney or kidney-pancreas allograft or second kidney allograft were randomly assigned in a 2:1 ratio to receive valganciclovir 900 mg once daily or oral ganciclovir 1000 mg three times daily. Treatment began within 10 days of transplantation and continued through day 100. For the purposes of the serology assessment, data from patients in both arms of the trial were analyzed together.


All patients were CMV seronegative pre-transplant and received an organ from a CMV seropositive donor. CMV serology (IgG and IgM) was assessed on days 28, 56, 100 and at months 6 and 12 post-transplant; serology was correlated with CMV disease and viremia. Plasma CMV load was assessed using the Cobas Amplicor CMV Monitor® Test (Roche Diagnostic Systems, Inc., Branchburg, NJ) at 2-week intervals until 5 months post-transplant and then at months 6, 8 and 12 post-transplant. All testing was performed by technologists blinded to the clinical status of the patient.

Clinical definitions

Symptomatic CMV disease cases were determined for the first 12 months post-transplant. An independent committee adjudicated all cases of CMV diseases using standard clinical criteria as outlined below (8).

CMV syndrome:  Fever ≥38°C (100.4°F) on two or more occasions (separated by 24 h within a 7-day period), and CMV detected in blood at a central laboratory, and one or more of malaise, atypical lymphocytosis of ≥5%, thrombocytopenia or an increase in hepatic transaminases to ≥ twice the upper limit of normal (non-liver transplant recipients).

Tissue-invasive CMV:  Symptoms or signs (including laboratory results) of organ dysfunction and evidence of localized CMV infection in a biopsy or other appropriate specimen (e.g., bronchoalveolar lavage, cerebrospinal fluid).

Laboratory methods

Serology testing for anti-CMV IgG and IgM was performed on all samples using the Abbott AxSYM™ enzyme immunoassay (Abbott Laboratories Ltd., Abbott Park, IL) as per manufacturer's instructions.


The clinical utility of IgG and IgM for prediction of subsequent CMV disease was calculated using 2 × 2 tables and comparisons done using corrected chi square tests. For the primary analysis, the first date of detectable antibody was defined as the timing of seroconversion regardless of the results of antibody testing on subsequent dates. In a secondary analysis, antibody results were analyzed for each date in isolation ignoring any previous results. All statistical analyses were performed using SAS version 8.0.


CMV disease

The randomized clinical trial PV16000 provided the patient pool for this substudy and included a total of 364 CMV D+/R− SOT recipients. At 12 months post-transplant, CMV disease occurred in 64 (17.6%) of these patients and the incidence of disease was similar in the two arms of the study. Results have been described in detail elsewhere (7). The majority of CMV disease was late-onset (60 cases), with only 4 (6.7%) cases occurring during prophylaxis. Forty-four cases of CMV disease (68.8%) occurred between day 100 and 6 months post-transplant. The remaining 16 cases (25%) occurred between 6- and 12 months post-transplant.


Of the 364 patients described above, samples from 352 (liver, n = 172 [including two liver-kidney]; kidney, n = 117; heart, n = 54; and kidney-pancreas, n = 9) were available for testing. All patients were seronegative pre-transplant. Cumulative rates of IgG and IgM detection for different time points post-transplant are shown in Figure 1. IgG seroconversion occurred in 26.9% of patients by day 100 (end of prophylaxis), and this reached 75.3% by 12 months post-transplant. The cumulative incidence of IgM detection was 8.3% and 54.9% by day 100 and 12 months, respectively. By the end of prophylaxis, CMV IgG seroconversion rates were 33.3% and 23.6% for ganciclovir- and valganciclovir-treated patients, respectively (p = 0.06). By 12 months, IgG seroconversion rates were similar in the two arms (ganciclovir 76.9%, valganciclovir 74.5%).

Figure 1.

Cumulative incidence of IgM and IgG seroconversion in cytomegalovirus donor positive/recipient negative solid organ transplant patients receiving 100 days of antiviral prophylaxis.

Different rates of seroconversion were observed in the various organ types (Figure 2). At the end of prophylaxis and at 6 and 12 months post-transplant, seroconversion rates were highest in liver recipients (40.5%, 86.2% and 91.8%, respectively) and lowest for kidney transplant recipients (9.0%, 35.6% and 57.7% respectively; Figure 2). These results also indicate that seroconversion was able to occur during the period of antiviral prophylaxis and under intense immunosuppression.

Figure 2.

Cumulative rate of IgG seroconversion by transplant organ type in cytomegalovirus donor positive/recipient negative solid organ transplant patients receiving 100 days of antiviral prophylaxis.

Serology and CMV disease

The timing of seroconversion in relation to CMV disease is shown in Table 1. Results were similar for both the ganciclovir and valganciclovir arm of the study. For all patients combined, the majority of patients who seroconverted did so either in the absence of CMV disease (51.1%) or after the onset of CMV disease (11.4%). A smaller percentage of patients seroconverted before the onset of CMV disease and a few patients had CMV disease and either did not seroconvert, or no subsequent serology data was available (Table 1). These data suggest that serology would not be useful for the diagnosis of acute CMV disease in the clinical setting.

Table 1.  Summary of onset of IgM or IgG seroconversion versus Endpoint Committee-defined cytomegalovirus disease

CMV IgM or IgG seroconversion*
Ganciclovir %
(n = 120)
Valganciclovir %
(n = 232)
Total %
(n = 352)
  1. *Seroconversion indicates IgM or IgG positive. The first date where IgM or IgG is positive is used as the date of seroconversion.

  2. **Or no serology data available subsequent to cytomegalovirus (CMV) disease event.

Seroconverted before CMV disease onset2 (1.7)13 (5.6)15 (4.3)
Seroconverted at/after CMV disease onset17 (14.2)23 (9.9)40 (11.4)
Seroconverted and no CMV disease62 (51.7)118 (50.9)180 (51.1)
No seroconversion** and had CMV disease4 (3.3)5 (2.2)9 (2.6)
No seroconversion and no CMV disease35 (29.2)73 (31.5)108 (30.7)

Prediction of CMV disease

The value of IgG and IgM seroconversion for predicting CMV disease was assessed at each time point. Any patient with CMV disease prior to each time point assessed is excluded from the analysis.

IgG seroconversion by day 28, 56 and by the end of prophylaxis (day 100) were not predictive of subsequent CMV disease. For example, in patients who seroconverted by day 100, the rate of subsequent CMV disease was 13.2% in seropositive patients versus 17.8% in seronegative patients, p = NS; Table 2). If only the day 100 IgG result was analyzed in isolation (ignoring any previous results), the rate of CMV disease in seropositive patients was 9.6% versus 17.4% in seronegative patients (p = 0.2).

Table 2.  Value of IgG seroconversion for predicting cytomegalovirus (CMV) disease
TimeSubsequent disease n (%)p-value
  1. *Excludes patients who have already had CMV disease prior to the time point being assessed for prediction.

Day 100*
 Seroconversion11 (13.2)72 (86.8)0.34
 No seroconversion41 (17.8)189 (82.2) 
6 months*
 Seroconversion2 (1.3)151 (98.7)0.002
 No seroconversion11 (10.0)99 (90.0) 

IgG seroconversion by 6 months post-transplant was predictive of subsequent CMV disease between months 6–12 (CMV disease rate 1.3% in IgG seropositive patients versus 10.0% in IgG seronegative patients [p = 0.002]). Therefore, seroconversion by 6 months post-transplant had a protective effect in terms of subsequent CMV disease. A similar finding was seen when only the 6-month IgG measurement was analyzed in isolation ignoring any previous serology results (CMV disease rate 1.6% in seropositive patients versus 8.7% in seronegative patients; p = 0.02).

In contrast to IgG, IgM seroconversion was not predictive of subsequent CMV disease (Table 3)

Table 3.  Value of IgM seroconversion for predicting cytomegalovirus (CMV) disease
TimeSubsequent disease n (%)p-value
  1. *Excludes patients who have already had CMV disease prior to the time point being assessed for prediction.

Day 100*
 Seroconversion4 (16.7)20 (83.3)0.99
 No seroconversion48 (16.8)238 (83.2) 
6 months*
 Seroconversion3 (3.2)91 (96.8)0.38
 No seroconversion10 (6.3)150 (93.7) 

Serology and CMV viremia

The timing of seroconversion in relation to CMV viremia (regardless of whether symptoms occurred) is shown in Table 4. Neither IgG nor IgM seroconversion at day 100 were predictive of subsequent detectable viremia (>400 copies/mL, data not shown).

Table 4.  Summary of onset of IgM or IgG seroconversion versus cytomegalovirus (CMV) viremia

CMV IgM or IgG seroconversion*
Ganciclovir %
(n = 120)
Valganciclovir %
(n = 232)
Total %
(n = 352)
  1. *Seroconversion indicates IgM or IgG positive. The first date where IgM or IgG is positive is used as the date of seroconversion.

  2. **Or no serology data available subsequent to viremia event.

Seroconverted before onset of viremia12 (10.0)25 (10.8)37 (10.5)
Seroconverted at/after onset of viremia44 (36.7)81 (34.9)125 (35.5)
Seroconverted and no viremia25 (20.8)48 (20.7)73 (20.7)
No seroconversion** and had viremia5 (4.2)10 (4.3)15 (4.3)
No seroconversion and no viremia34 (28.3)68 (29.3)102 (29.0)

Quantitative CMV IgG results

Quantitative CMV IgG results were analyzed at day 100 post-transplant to see if predictive value could be improved. In patients with positive serology, the median IgG level was 34.8 AU/mL (range 15 to >250 AU/mL). Using a cut-off value of ≥25 AU/mL for defining seropositivity, the rate of subsequent disease in seropositive patients was 11.1% versus 16.9% in seronegative patients (p = 0.48). Using a cut-off value of ≥50 AU/mL for defining seropositivity, the rate of subsequent disease in seropositive patients was 9.5% versus 16.7% in seronegative patients (p = 0.55). Using a cut-off value of ≥75 AU/mL, the rate of subsequent disease in seropositive patients was 9.1% versus 16.5% in seronegative patients (p = NS). Therefore, no improvement in predictive value was observed by analyzing quantitative IgG results.


While antiviral prophylaxis is effective in suppressing CMV viral replication and disease in D+/R− transplant patients during prophylaxis, there remains a high incidence of late-onset CMV disease (∼20%) between months 3–12 post-transplant (7). If D+/R− patients at risk of developing late-onset CMV disease could be readily identified, this subpopulation could then be targeted for more aggressive monitoring or antiviral therapy. Identification could be based on clinical or laboratory assessments. For example, patients with acute rejection and more potent immunosuppression are recognized to be at risk of late-onset CMV disease (9). Assessment of the development of host immunity to CMV may be another approach to identifying such patients. T-cell studies may prove useful for this but are not readily available in most clinical settings (10,11). Serology has been proposed as a useful surrogate marker (12) and could help identify which patients may be at risk of late-onset CMV disease.

We assessed the value of serology at various time points post-transplant for predicting late-onset CMV disease. IgG or IgM seroconversion by day 100 and IgM seroconversion by 6 months did not appear to be protective in terms of CMV disease since the incidence of subsequent disease was similar in those that seroconverted compared with those that did not. While there was a trend to lower subsequent disease rates in patients seropositive at day 100 versus those who were seronegative, this was not significant. Also, based on our data, serological monitoring while on prophylaxis would not be indicated. Thus, overall, serologic assessment was only of limited clinical utility. However, in patients who IgG seroconverted by 6 months, the rate of subsequent CMV disease was very low at 1% compared with 10% in those that were seronegative. In clinical practice, this may have some modest utility, since a single IgG measurement at month 6 could be used to identify D+/R− patients at risk of late-onset CMV disease. The results also indicate that further prophylaxis in seropositive patients would neither be necessary or cost-effective. Additionally, in any individual patient presenting with suggestive symptoms, CMV disease must still be included in the differential diagnosis, regardless of the serostatus. The other main findings of this paper were that seroconversion is delayed in transplant recipients but that it does still occur while patients are on antiviral prophylaxis.

Our study had several limitations. First, this study was performed in the context of a larger study comparing valganciclovir versus ganciclovir for prophylaxis of CMV disease. Therefore assessing the clinical utility of CMV serology testing was not the primary purpose of the original study. Also, the addition of cell-mediated immunity data to the serology data would have been useful to interpret the development of immunity against CMV post-transplant. Finally, further refinement of serological testing, including the use of avidity testing would have likely been useful. Overall, however, since the primary purpose of this study was to assess the clinical utility of a simple, readily available diagnostic test, we believe the current study fulfills this objective.

Previous studies have assessed the utility of serology for diagnosis of CMV disease (13–15). These studies have demonstrated that serology is generally of limited clinical utility for the acute diagnosis of CMV disease, when compared with molecular assays. For example, in a study of 30 liver transplant recipients, positive PCR and antigenemia test results generally preceded the onset of symptoms; however, IgM detection occurred at a median of 2.5 weeks after onset of disease (13). In a study of 42 liver and kidney transplant recipients, IgG or IgM serological responses did not differ between symptomatic and asymptomatic patients, and usually occurred only after the onset of symptoms (14). The use of a more sensitive recombinant antigen ELISA IgM assay (similar to the one used in our study), was found to have some clinical utility in predicting CMV disease in one study of 40 liver transplant recipients, although again DNAemia generally preceded detection of antibody (15). Based on these data, in the past several years, molecular detection methods have generally replaced serology and culture based tests for the diagnosis of active CMV disease (16–18). There are, however, no previous studies assessing the utility of CMV serology testing in a large cohort of D+/R− patients receiving antiviral prophylaxis, for the purposes of identifying patients at risk of late-onset CMV disease.

Several other interesting observations were noted in this study. Firstly, seroconversion was observed during antiviral prophylaxis, thereby allowing the patient to develop humoral immunity in the absence of CMV disease (IgG seroconversion rates were 26.9% by day 100, but CMV disease was observed in just 1% of patients while on prophylaxis). Secondly, kidney transplant recipients appeared to have the lowest rates of seroconversion compared with other organs. Specifically, end of prophylaxis seroconversion rates were very low in kidney recipients (9.0%) versus 40.5% in liver transplant recipients (p < 0.001). This may be indicative of greater viral suppression in this organ or alternatively may be an effect of differences in immunosuppressive medications between organs. For example, the percentage of patients receiving mycophenolate mofetil at day 100 was 52% for kidney recipients versus 30% for liver transplant recipients (p < 0.001). Additionally, there was also a trend towards a lower rate of seroconversion at day 100 in patients receiving valganciclovir compared to those receiving oral ganciclovir (p = 0.06). This is likely reflective of the increased bioavailability of valganciclovir compared with oral ganciclovir and the resultant increased suppression of viral replication. Finally, most seroconversion occurred in the absence of clinical disease, which supports viral load data suggesting that asymptomatic viral replication is much more common than symptomatic disease (7). A point of interest, however, is that seroconversion was more common than viremia during the period of antiviral prophylaxis. This discrepancy may be due to the use of a plasma-based test or due to infrequent monitoring which may have missed low level or transient viremia,

In summary, this study evaluates the clinical utility of serology testing for predicting CMV disease in D+/R− organ transplant recipients. For the most part, IgG or IgM testing at regular intervals was not clinically useful. However an IgG measurement at 6 months post-transplant could be clinically useful in identifying patients at risk of late-onset CMV disease. Closer clinical follow-up or monitoring may be indicated in patients who have not seroconverted by 6 months post-transplant although this strategy would likely only pick up a small minority of late-onset CMV disease.


There follows a complete list of the members of the Valganciclovir Solid Organ Transplant Study Group in Alphabetical order by country:

In Australia: Josie Eris, Royal Prince Alfred Hospital, Camperdown, NSW; Anne Keogh, St Vincent's Hospital, Darlinghurst, NSW; Tim Mathew, Queen Elizabeth Hospital, Woodville, SA; Geoff McCaughan, Royal Prince Alfred Hospital, Camperdown, NSW; Kathy Nicholls, Royal Melbourne Hospital, Parkville, VIC and Simone Strasser, Royal Prince Alfred Hospital, Camperdown, NSW.

In Canada: Atul Humar, Toronto General Hospital, Toronto, ON; Richard Lalonde, Montreal Chest Institute, Montreal, QC; Paul Marotta, London Health Sciences Center University Campus, London, ON; Jutta Preiksaitis, University of Alberta Hospital, Edmonton, AB and Eric Yoshida, Vancouver Hospital and Health Sciences Center, Vancouver, BC.

In France: Iradj Gandjbakch, Pitie-Salpetriere Hospital, Paris; Yvon Lebranchu, Bretonneau Hospital, Tours; Christophe Legendre, Saint-Louis Hospital, Paris and Faouzi Saliba, Hopital Paul Brousse, Villejuif.

In Ireland: Oscar Traynor, St Vincents University Public Hospital, Dublin.

In Italy: Paolo Angeli, Azienda Ospedaliera Di Padova, Padova and Francesco Menichetti, Ospedale Cisanello, Pisa.

In New Zealand: Ed Gane, Auckland Hospital, Auckland.

In the United Kingdom: Ali Bakran, Royal Liverpool University Hospital, Liverpool; John Forsythe, Edinburgh Royal Infirmary, Edinburgh; Nigel Heaton, Kings College Hospital, London; Peter Lodge, St James Hospital, Leeds; Derek Manas, Freeman Hospital, Newcastle Upon Tyne; Peter Morris, Churchill Hospital, Oxford; Jayan Parameshwar, Papworth Hospital, Papworth Everard and Nizar Yonan, Wythenshawe Hospital, Manchester.

In the United States: Barbara Alexander, Duke University Medical Center, Durham, NC; Emily Blumberg, Hospital of the University of Pennsylvania, Philadelphia, PA; Daniel C Brennan, Barnes Jewish Hospital, St Louis, MO; Robert Brown, Columbia Presbyterian Medical Center, New York, NY; Ronald W Busuttil, UCLA School of Medicine, Los Angeles, CA; Ken Chavin, Medical University South Carolina, Charleston, SC; David Conti, Albany Medical Center, Albany, NY; Angelo DeMattos, Oregon Health Sciences University, Portland, OR; Ed Dominguez, University of Nebraska Medical Center, Omaha, NE; Howard J Eisen, Temple University School of Medicine, Philadelphia, PA; Dan Fishbein, University of Washington, Seattle, WA; Thomas Fishbein, Mt Sinai Medical Center, New York, NY; Robert Fisher, Medical College of Virginia Hospital, Richmond, VA; Richard Freeman, New England Medical Center, Boston, MA; Chris Freise, University of California San Francisco, San Francisco, CA; Marquis Hart, UC San Diego Medical Center, San Diego, CA; Thomas Heffron, Emory University, Atlanta, GA; Ray E Hershberger, Oregon Health Sciences University, Portland, OR; Richard J Howard, University of Florida, Gainesville, FL; Sandra A Kemmerly, Alton Ochsner Medical Institution, New Orleans, LA; Richard Knight, Mt Sinai Medical Center, New York, NY; Bernard Kubak, UCLA School of Medicine, Los Angeles, CA; Shimon Kusne, University of Pittsburgh, Pittsburgh, PA; Steven Mawhorter, Cleveland Clinical Foundation, Cleveland, OH; Martin Mullen, Loyola University Medical Center, Maywood, IL; Carlos Paya, Mayo Clinic, Rochester, MN; Mark Pescovitz, Indiana University Medical Center, Indianapolis, IN; John Pirsch, University of Wisconsin Medical School, Madison, WI; Timothy L Pruett, University of Virginia Health Systems, Charlottesville, VA; Jeffrey Punch, University of Michigan Medical Center, Ann Arbor, MI; John Rabkin, Oregon Health Sciences University, Portland, OR; Robert Rubin, Massachusetts General Hospital, Boston, MA; John Scandling, Stanford University Medical Center, Palo Alto, CA; Michael Shapiro, Hackensack University Medical Center, Hackensack, NJ; Randi Silibovsky, Albert Einstein Medical Center, Philadelphia, PA; Kenneth Washburn, University of Texas Health Service Center, San Antonio, TX and Sam Weinstein, LifeLink Transplant Institute, Tampa, FL.

This study was funded by F. Hoffmann-La Roche, Basel, Switzerland.