Frequent Occult Infection with Cytomegalovirus in Cardiac Transplant Recipients despite Antiviral Prophylaxis

Patients.Seventy-five first heart transplant recipients (age, 50 ± 12 years; 54 males) surviving at least 6 months after transplant and at risk of CMV infection (11 R⁺/D⁻, 34 R⁺/D⁺, 21 R⁻/D⁺, and 9 R⁻/D⁻) were consecutively enrolled between 24 January 2002 and 19 August 2004 and monitored for up to 16 months posttransplantation. Immunosuppressive therapy has been described in detail previously (14) and involved induction with daclizumab (1) as well as maintenance with oral cyclosporine and prednisone together with either mycofenolate mofetil or sirolimus. All patients who were at risk of CMV infection (R⁻/D⁺, R⁺/D⁻, and R⁺/D⁺) received a standard course of intravenous ganciclovir at 5 mg/kg of body weight twice a day for the first 2 weeks after transplantation and 6 mg/kg daily for the following 2 weeks. Because of a higher risk for CMV disease, CMV-naïve patients receiving a heart from a CMV-seropositive donor (R⁻/D⁺) were subjected to more aggressive prophylaxis, including intravenous CMV immunoglobulin G (Cytogam; Medimmune, Inc., Gaithersburg, MD) for the first 4 months posttransplant (25), as well as an extended course of oral valganciclovir from 4 weeks posttransplant (900 mg daily) and continuing for an additional 10 weeks, on average, posttransplant (14). Starting at 2 weeks and continuing at months 1, 2, 3, 4, 6, 9, 12, and 16 posttransplant, PB and EMB specimens were collected and subjected to a sensitive two-stage CMV DNA PCR analysis. Blood samples collected in EDTA anticoagulant were processed within 24 h of collection by diluting with two volumes of phosphate-buffered saline (PBS) and separating into two fractions, PBMCs (interface) and PMNs (pellet), using Histopaque (Sigma-Aldrich Inc., St. Louis, MO). Following collection, the PBMC fraction was washed once in PBS, and the PMN fraction was mixed with two volumes of Hanks' balanced salts solution and two volumes of 3% dextran. Following a 20-min incubation of the PMNs, the supernatant was collected. The contaminating red blood cells in both leukocyte fractions were lysed with 1.6% NaCl. This method resulted in a >99% separation of granulocytes (including PMNs) from PBMCs based on direct histological evaluation of cytocentrifuged samples (data not shown). EMB samples were frozen at −140°C and batch processed into DNA for PCR analysis for viral and host cell DNA. A total of 670 PMN and 610 PBMC samples from 66 at-risk patients as well as 349 EMB samples from 55 at-risk patients were analyzed for CMV DNA by a two-stage analysis, starting with qualitative nested PCR (qualPCR). PMN, PBMC, or EMB samples positive by qualPCR for CMV DNA were subjected to a second-stage quantitative real-time (TaqMan) PCR. Our approach is based on many years of experience studying CMV latency (12, 20), where aliquots of 10⁵ bone marrow-derived mononuclear cells from healthy CMV-seropositive donors are consistently positive, whereas peripheral blood mononuclear cells from such individuals are consistently negative (data not shown). Screening of EMB specimens for adenovirus DNA was performed using real-time PCR.

Qualitative PCR.We developed methods based on earlier work (see the reference list in reference 12) to amplify viral DNA from clinical samples with a low background and a sensitivity of 3 to 5 CMV genomes/10⁵ cell equivalents. Aliquots of 5 × 10⁶ freshly separated PB cells were lysed overnight at 60°C in 250 μl buffer containing 50 mM KCl, 10 mM Tris-HCl (pH 8.0), 2 mM MgCl₂, 0.45% NP-40, 0.45% Tween 20 (Sigma-Aldrich Inc., St. Louis, MO), and 100 μg/ml proteinase K (Invitrogen Corp., Carlsbad, CA). Following lysis, samples were boiled for 5 min, cooled on ice, and stored at −80°C until use. DNA was isolated from EMB samples using the QIAGEN DNeasy Tissue kit (QIAGEN Inc., Valencia, CA) according to the manufacturer's instructions. The initial amplification of DNA was with primers IEP2AII (^5′ATGGAGTCCTCTGCCAAGAGAAAGATGGAC^3′) and IEP4BII (^5′CAATACACTTCATCTCCTCGAAAGG^3′), followed by IEP3B (^5′TCTGCCAGGACATCTTTCTC^3′ and IEP3A (^5′GTGACCAAGGCCACGACGTT^3′). For the initial qualPCR, 5 μl (1 μg; 10⁵ cell equivalents) of extracted PBMC and PMN DNA or 50 ng EMB DNA was added to 45-μl reaction mixtures containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂ (Roche Diagnostics Corporation, Indianapolis, IN), 200 μM of each deoxynucleotide triphosphate (Invitrogen Corporation), 1 U Taq polymerase (Roche Diagnostics Corporation), and 1 μM of each primer. For the second qualPCR, 5 μl out of the first reaction mixture was added to a 45-μl reaction mixture as used for the initial qualPCR. qualPCR mixtures were layered with 50 μl of mineral oil and subjected to one cycle of 94°C for 3 min, followed by 30 cycles of 1 min of denaturation at 94°C, 1 min of annealing at 62°C, and 2 min of extension at 72°C, and a final extension step of 7 min at 72°C. Positive and negative controls were included in each run, and, after qualPCR, each sample was electrophoresed through a 2% agarose gel containing 0.2 μg/ml ethidium bromide with appropriate DNA size markers. Amplicons were visualized on a UV transilluminator and photographed.

Quantitative PCR.We developed methods to quantify viral load in the qualPCR-positive samples using SYBR green detection of real-time PCR products able to detect a range of 3 to 100,000 CMV genome copies/10⁵ cells. Real-time PCR was performed in a GeneAmp 5700 Sequence Detection System (Applied Biosystems, Foster City, CA). Five microliters of sample was added to 45 μl of reaction mix containing 25 μl SYBR green PCR Master Mix (Applied Biosystems) and 10 pmol of each primer. The primers used for the real-time PCR were also located in the gene encoding the IE1 protein, exon 2 (IE1-2F, ^5′TCGTTGGAATCCTCGGTCA^5′; and IE1-2R, ^5′GGCCGAAGAATCCCTCAAAA^5′). Real-time PCR detection of adenovirus DNA (51 serotypes) employed a 25-μl reaction volume containing 45 μM of each hexon-specific primer (^5′GCCACGGTGGGGTTTCTAAACTT^3′ and ^5′GCCCCAGTGGTCTTACATGCACATC^3′) and 5 μM probe (6-carboxyfluorescein^5′TGCACCAGACCCGGGCTCAGGTACTCCGA^3′6-carboxytetramethylrhodamine) (9). Conditions for the real-time PCR were established empirically and included incubation at 50°C for 2 min to enable uracil-DNA-glycosylase (in Master Mix) to act on samples and 94°C for 10 min to activate AmpliTaq Gold polymerase, followed by 40 cycles (for CMV) or 42 cycles (for adenovirus) of 15 s of denaturation at 94°C and 1 min of annealing and extension at 60°C. A standard curve of serial dilutions of known amounts of plasmid DNA was used in each run as an internal control and to determine the copy number in the samples. In addition, a natural CMV DNA-positive sample was run to assess consistency. The detection limit of this test was 5 CMV DNA copies/10⁵ cells. Each run included positive and negative samples, and specificity of the amplified products was assessed on each run by dissociation curve analysis to ensure that all products had an expected, uniform melting temperature. In addition, a selection of CMV-positive and CMV-negative samples was subjected to quantitative PCR analysis for CMV, Epstein-Barr virus, and human herpesvirus type 6 in the laboratory of Lawrence Cory and Meei-Li Huang at the University of Washington, Seattle, confirming the CMV DNA levels in positive samples but failing to detect any correlation between CMV DNA detection and the less frequent detection of Epstein-Barr virus or human herpesvirus type 6 (data not shown).

Data analysis.Continuous variables are expressed as means ± standard deviations of the means or as medians with 25th to 75th percentiles given in parentheses, as appropriate, if not otherwise specified. Student's t test was used to compare continuous variables that showed a normal distribution. Chi-square or Fisher's exact analyses were applied to test differences in categorical variables. In order to compare patient's survival free from CMV detection in PMN versus PBMC serologically dictated subgroups, we applied a nonparametric test for paired data (Wilcoxon signed-rank test). The estimated incidence of CMV detection in PBMCs, PMNs, and EMB specimens was displayed using Kaplan-Meier plots. Differences in incidence of the first CMV detection between serological subgroups of patients were tested with the log-rank test. Differences in the time from transplant to peak CMV titer were analyzed by using a Mann-Whitney test.

In order to distinguish differences in CMV load among positive samples, we assigned each patient sample to a quartile and considered samples above the 75th percentile of this distribution to represent high-grade infection. This study was performed with approval of the Human Subjects Institutional Review Board at Stanford University.

CMV DNA detection in the study population.Although collected and assayed at the same time as for patients at risk of CMV infection, CMV DNA was not detected in any samples collected from the R⁻/D⁻ subgroup of nine patients, reinforcing the importance of serology in predicting risk of CMV infection. Conversely, most of the 66 at-risk patients receiving antiviral prophylaxis had an active CMV infection based on detection of viral DNA in PB leukocytes over the first 16 months after transplant (Fig. 1). In the PMN fraction, which becomes positive for CMV DNA only during active infection, 58 patients presented at least one CMV-positive sample, an estimated cumulative incidence of 91% ± 4% of patients with active CMV infection. CMV was detected in the PBMC fraction in a pattern similar to that of PMNs, although viral DNA in PBMCs may represent latency or productive infection (8, 12). Sixty patients became positive, accounting for an estimated cumulative incidence of 92% ± 4%. Although detection in PMNs was considered an unambiguous indicator of active CMV replication (5), both fractions became positive for viral DNA in most at-risk patients (54 patients; 82%). CMV DNA was detected in about 200 samples regardless of leukocyte fraction, with four (6%) patients positive only in PMNs and six (9%) positive only in PBMCs. This extends earlier observations on high-risk heart transplant recipients (15), from whom comparable numbers of PBMC and total PB leukocytes had been shown to be CMV DNA positive. Acute CMV disease occurred in three patients, one R⁻/D⁺, one R⁺/D⁺, and one R⁺/D⁻, during the study period, a frequency (5%) that was in line with previous reports (11). CMV detection was less frequent in the R⁻/D⁺ subgroup than in the subgroup receiving standard antiviral prophylaxis, as previously reported (14). Thus, despite achieving the expected level of suppression of acute CMV disease, neither standard anti-CMV prophylaxis in the R⁺ subgroup nor aggressive anti-CMV prophylaxis in the R⁻/D⁺ subgroup eliminated viral reactivation and replication. Prophylaxis achieved the goal of reducing CMV replication levels but had little apparent impact on the incidence of reactivation.

• Open in new tab
• Download powerpoint

FIG. 1.

Kaplan-Meier analysis of the incidence of the first CMV DNA detection in PB cells and cardiac EMB samples from heart transplant recipients receiving anti-CMV prophylaxis. Results of qualPCR detection in PMNs (gray line), PBMCs (black line), and EMB specimens (dashed line) are shown.

Time course of CMV in PMNs and PBMCs.When the time to first CMV detection was calculated for the patient subgroup that became CMV positive in both leukocyte fractions, the median time (25th to 75th percentile) to CMV DNA detection was 54 (21 to 79) days in PBMCs and 95 (33 to 166) days in PMNs (P = 0.02 by Wilcoxon signed-rank test). Overall, CMV DNA was detected earlier in PBMCs than in PMNs. More specifically, paired data analysis revealed that the earlier appearance in PBMCs was significant only in the 40 R⁺ (55 [21 to 79] versus 60 [30 to 124] days; P = 0.048) patients and not in the 21 R⁻/D⁺ (53 [20 to 138] versus 87 [18 to 143] days; P = 0.9) patients. Within the two R⁺ patient subgroups, the time to first CMV detection was significantly shorter in PBMCs than in PMNs only in the 27 R⁺/D⁺ (48 [20 to 79] versus 70 [30 to 105] days; P = 0.02) patients (Fig. 2). This relationship was not observed in the 13 R⁺/D⁻ (58 [21 to 96] versus 53 [30 to 141] days; P = 0.5) patients (Fig. 2). However, the small size of the this subgroup remains a confounding factor.

• Open in new tab
• Download powerpoint

FIG. 2.

Differences in length of time to the first detection of CMV DNA based on cell type and CMV serostatus at transplant. (A) Time to first CMV detection in PBMCs versus PMNs in R⁺/D⁺ patients (P = 0.016 by Wilcoxon signed-rank paired analysis). (B) Time to first CMV detection in PBMCs versus PMNs in R⁻/D⁺ patients (P = 0.9 by Wilcoxon signed-rank paired analysis). (C) Time to first CMV detection in PBMCs versus PMNs in R⁺/D⁻ patients (P = 0.5 by Wilcoxon signed-rank paired analysis). The thick line indicates the median, the lower and upper sides of the rectangles indicate 25th and 75th percentiles, respectively, and the bars indicate the 90th and 10th percentile limits.

Subsequent quantification with real-time PCR revealed higher peak levels in PMNs than in PBMCs on a background that varied from a minimum of 5 (the cutoff) to a maximum of >4,000 DNA copies/10⁵ leukocytes. The majority of positive samples had low CMV DNA levels, such that overall median (25th to 75th percentile) levels were 34 (8 to 110) CMV genome copies/10⁵ leukocytes in PMNs and 17 (5 to 57) copies/10⁵ leukocytes in PBMCs. Overall, CMV load was significantly higher in PMNs than PBMCs (P < 0.01), consistent with active viral replication in these patients. These levels were substantially lower than has been observed in untreated immunocompromised patients (5, 6, 15) and suggests that prophylaxis reduces infection levels without altering the reactivation dynamics. Time course analysis revealed that CMV DNA was detected less frequently and remained at lower median levels during prophylaxis than afterwards (11), and this was evident in both the PMN and PBMC fractions. In particular, during and after prophylaxis the incidence of CMV-positive PMN samples was 19% and 36% (P < 0.01), and the median genome copy numbers were 23 and 67 per 10⁵ PMNs (P = 0.05), respectively. Similarly, after prophylaxis the incidence of CMV-positive samples in PBMCs rose to 38% (P < 0.01) and the median DNA copy number rose to 34 genomes/10⁵ PBMCs (P < 0.01). We also compared the time course of CMV detection in CMV-naïve patients (R⁻/D⁺) and CMV-seropositive patients (R⁺). First, we analyzed differences in cumulative incidence of CMV DNA detection. We then compared the times to peak CMV DNA levels in PMNs and PBMCs. CMV DNA was detected significantly earlier and more often in PMN or PBMC fractions from the R⁺ subgroup than in those of the R⁻/D⁺ subgroup (Fig. 2 and 3A and B), although the time to peak CMV DNA level and the frequency of CMV DNA-positive patients were the same in the two patient subgroups regardless of the PB leukocyte fraction that was assessed (data not shown). The incidence of CMV detection began to diverge at about 1 month posttransplantation, at the time when prophylaxis of the R⁺ subgroup was terminated.

• Open in new tab
• Download powerpoint

FIG. 3.

CMV DNA detection based on donor/recipient CMV serostatus at transplant. The Kaplan-Meier estimated incidence of CMV detection in PBMCs (A) or in PMNs (B) is shown. Determinations of the statistical significance of R⁺ and R⁻/D⁺ differences employed the log-rank test.

Time course and prevalence of high CMV DNA levels.Aiming to analyze the effect of preexisting CMV infection (based on serostatus) and the influence of prophylaxis regimen on the kinetics of high CMV DNA levels, we studied the time course of detection of CMV DNA above the 75th percentile of the entire cohort of samples, levels greater than 110 CMV DNA copies/10⁵ PMNs. As shown in Fig. 4A, we found that R⁺ patients had a higher frequency of detection of CMV DNA at or above this level than R⁻/D⁺ patients. This difference was particularly evident at month 4 after transplant, when prophylaxis was ongoing in the R⁻/D⁺ subgroup but was 2 to 3 months after prophylaxis had been discontinued in the R⁺ subgroup. Consistent with the frequency of high CMV load being detected following termination of prophylaxis, the frequency was highest in R⁻/D⁺ patients during the 5- to 7-month period following discontinuation of prophylaxis after the fourth month. Thus, in either the R⁺ or R⁻/D⁺ subgroups, high-grade infection frequency peaked 2 to 3 months after prophylaxis was discontinued.

• Open in new tab
• Download powerpoint

FIG. 4.

Frequency of high-level CMV DNA in heart transplant recipients receiving anti-CMV prophylaxis. (A) Rate of samples showing >110 DNA copies/10⁵ PMNs (high-grade infection) in the R⁻/D⁺ (open circles) and R⁺ (filled circles) patient groups over 16 months posttransplant. (Time periods were as follows: month 1, 14 to 45 days; month 2, 46 to 76 days; month 3, 77 to 105 days; month 4, 106 to 161 days; months 5 to 7, 162 to 208 days; months 8 to 10, 209 to 304 days; months 11 to 13, 305 to 426 days; months 14 to 16, 427 to 486 days). The standard prophylaxis period used for R⁺ patients is indicated by dark gray shading, and the aggressive prophylaxis period used for R⁻/D⁺ patients is indicated by light gray shading. (B) Distribution of patients developing low-grade (≤110 copies/10⁵ cells) and high-grade (>110 copies/10⁵ cells) infection according to CMV serological status.

Table 1 shows the demographic and clinical characteristics of the 20 patients who reached levels above 110 CMV genome DNA copies/10⁵ PMNs and the 46 patients who remained at or below 110 CMV genome copies. The frequency of high-level samples in R⁺ patients was twice that in the R⁻/D⁺ subgroup. Over a third (39%) of the R⁺ subgroup experienced high-level infection, whereas only 15% (P = 0.04) of the R⁻/D⁺ subgroup had CMV DNA levels in this range. Thus, the overwhelming majority (85%) of patients who experienced high-grade infection were CMV seropositive at the time of transplant. Together with our recent report examining the impact of CMV infection on clinical endpoints in an overlapping cohort of these patients (14), this study demonstrates that viral levels are suppressed by the aggressive antiviral regimen in use on R⁻/D⁺ patients (Fig. 4B) and that this prophylaxis regimen effectively reversed the traditional higher risk for CMV infection than that of R⁺ patients (Fig. 2C). Note that demographic and baseline characteristics of patients with versus those without high-grade CMV infection were similar (Fig. 4), limiting the possibility that factors other than serology and prophylaxis regimen influenced high-grade CMV occurrence.

CMV DNA detection in EMB samples.Over 80% of patients consented to provide EMB samples for viral DNA analysis to compare infection in the transplanted heart to infection of PB cell fractions. We tested 349 EMB samples from 55 of the patients distributed across the high- and low-grade infection subgroups. CMV DNA was detected in only two allograft samples, each from a different patient. EMB samples from high-risk patients (R⁻/D⁺) receiving aggressive prophylaxis were uniformly negative, consistent with the lower CMV DNA levels in leukocytes from this subgroup (Table 1; Fig. 4). The only patients with positive EMB samples were CMV seropositive at the time of transplant (one R⁺/D⁺ and one R⁺/D⁻), and these infections were coincident with high-grade systemic infection in PB cells and followed termination of anti-CMV prophylaxis. CMV DNA levels were at the highest levels recorded for these patients, whether evaluated in PBMCs (346 and 2,152 CMV genome copies/10⁵ PBMCs, respectively) or PMNs (524 and 936 CMV genome copies/10⁵ PMNs, respectively) on the day EMB samples tested positive. Although neither patient had signs of clinically evident CMV disease or cellular acute rejection at the time the positive samples were collected (2), one had a high incidence of other infections (bacterial as well as herpes zoster) and an acute rejection episode requiring increased immunosuppressive therapy during the previous month, and the other experienced a rejection episode requiring additional immunosuppressive treatment 1 month afterwards. In addition to CMV, adenovirus has been implicated in chronic vascular disease in pediatric heart transplant patients (24), suggesting that this virus may contribute to disease independently or in combination with CMV. We detected adenovirus DNA in two patients, three times in one patient (16, 66, and 124 days posttransplant) and twice in another patient (78 and 427 days posttransplant), suggesting that infection with this virus was also infrequent within the transplanted heart. The detection of adenovirus did not coincide with CMV detection, acute rejection, adverse events, or other infections.