Impact of Valganciclovir Prophylaxis Duration on Cytomegalovirus Disease in High-Risk Donor Seropositive/Recipient Seronegative Heart Transplant Recipients

Abstract

Background

Few data support current guideline recommendations for 6 months over 3 months of antiviral prophylaxis for cytomegalovirus (CMV) disease prevention in donor seropositive/recipient seronegative (D+R-) heart transplant recipients (HTR).

Methods

We retrospectively assessed CMV disease and clinical outcomes in 310 adult HTR between 7/5/2005 – 12/30/2016 at a single center. Valganciclovir (VGCV) prophylaxis was given for 3 months in recipient seropositive (R+) and for 3 to 6 months in D+R- groups. A multivariable logistic model evaluated risk factors for CMV disease, with use of inverse probability weighting to correct for confounding. A nested matched cohort study within D+R- HTR was performed to assess CMV disease-associated morbidity. Chi-square and Mann-Whitney tests compared categorical and continuous variables, with p < 0.05 considered significant.

Results

The incidence of CMV disease among all, D+R-, and R+ groups was 8.7% (27/310), 26.5% (22/83), and 2.8% (5/180), respectively, and included syndrome in 22.2% (6/27) and end-organ involvement in 77.8% (21/27). In multivariable models, longer prophylaxis duration was not associated with reduced risk for CMV disease when assessed either as a continuous (p=0.28) or categorical (3 vs 6 months) variable (p=0.19). CMV disease in D+R- HTR was associated with higher rates of hospitalization (87.5% [14/16] vs 6.3% [1/16], p<0.001) and for a longer duration (median 4 days [IQR 3 – 7] vs median 0 days [IQR 0 – 0], p=0.001) than in matched D+R- controls.

Conclusions

CMV disease remains a major cause of morbidity in D+R- HTR. In contrast to documented benefit in D+R- lung and kidney recipients, VGCV duration longer than 3 months was not associated with a lower incidence of CMV disease in D+R- HTR, and should not routinely be used in this setting.

Introduction

Cytomegalovirus (CMV) disease is a major cause of morbidity and mortality in heart transplant recipients (HTR), and CMV donor seropositive [D+] recipient seronegative [R-] status is recognized as a major risk factor for CMV infection and disease. Among this D+R- high-risk group, antiviral prophylaxis with ganciclovir or valganciclovir (VGCV) is the most widely used prevention strategy in the United States [1–3], and has historically been given for approximately 3 months post-transplant. However, post-prophylaxis (“late-onset”) CMV disease is a major clinical problem in D+R- patients. In other high-risk D+R- organ transplant populations (kidney and lung transplant recipients), controlled clinical trials have demonstrated that durations of prophylaxis longer than 3 months are associated with a reduction in the incidence of CMV disease [4, 5]. However, few studies have assessed longer durations of antiviral prophylaxis specifically in D+R- HTR, who are at high risk for valganciclovir-associated toxicity and resistance.

Current CMV guidelines preferentially recommend 6 months over 3 months of VGCV prophylaxis in D+R- HTR, but these recommendations are based mostly on expert opinion and extrapolation from other D+R- organ transplant populations, rather than direct comparisons of 6 vs 3 months of prophylaxis in this setting [6, 7]. Important differences in the intensity of immunosuppression, risk for and consequences of rejection, higher risk for antiviral resistance, and other clinical factors in HTR make it difficult to directly extrapolate data on prophylaxis duration from other (non-heart) organ transplant populations. Only a single observational study has specifically assessed the impact of duration of antiviral prophylaxis on CMV disease rate in HTR [8]. In that study, longer duration of prophylaxis was not associated with a lower incidence of CMV disease. However, the durations of prophylaxis studied (≤ 6 vs ≥ 12 months) were different than those that are currently recommended either in international CMV consensus guidelines or studied in previous clinical trials. Other limitations of that study included a relatively small number of D+R- patients, use of a historical comparison group, and inclusion of a composite endpoint of CMV infection and disease rather than disease alone. Thus, direct comparisons of 3 vs 6 months of antiviral prophylaxis and CMV disease in high-risk D+R- HTR are needed.

We conducted a retrospective cohort study to determine whether a 6 month duration of VGCV prophylaxis was associated with a decreased incidence of CMV disease compared to 3 months of prophylaxis in a large cohort of high-risk D+R- HTR.

Methods

Study design

All patients ≥18 years old who received a first heart transplant at University of Washington Medical Center from 7/5/2005 – 12/30/2016 (no overlap from a prior cohort study [9]) and who had a minimum 1 year of follow up (shorter follow-up if death occurred earlier than 1 year), were included in the cohort. All D+R- HTR who survived and had followup for ≥3 months after completing VGCV prophylaxis were included in our primary analysis of risk factors for CMV disease. We also performed a nested matched cohort analysis within D+R- HTR to evaluate morbidity associated with CMV disease, as described in the section below. Prespecified clinical and laboratory variables were abstracted from electronic transplant and clinical databases by trained personnel using standardized data collection forms. This study was approved by the Institutional Review Board/Human Subjects Division of the University of Washington Medical Center (HSD #00003994).

CMV prophylaxis and monitoring

All CMV seropositive (R+) or seronegative (R-) patients who received an organ from a seropositive donor (D+) received CMV prophylaxis, initially with IV ganciclovir (5mg/kg/day), then oral VGCV (900mg once daily), once oral medications were tolerated. Both medications were adjusted for renal function according to manufacturer recommendations. The duration of antiviral prophylaxis was generally either 3 or 6 months, at the discretion of the primary clinical team. After discontinuation of VGCV prophylaxis, patients underwent CMV testing for clinical signs or symptoms of CMV disease. Routine surveillance virologic monitoring in otherwise asymptomatic patients was not performed, consistent with CMV international consensus guidelines [7].

CMV disease definition and risk factors

All potential cases of CMV disease were identified from a prospectively maintained database and confirmed by chart review by two separate reviewers. CMV disease cases were classified as proven, probable, or possible according to published definitions [10], and patients were censored at time of second organ transplant. Any discordance in adjudication of CMV disease was resolved through consensus review by a third senior reviewer (author APL). Risk factors for CMV disease, including the duration of antiviral prophylaxis, were evaluated in D+R- recipients following first transplant.

Morbidity associated with CMV disease

Following our primary analysis, we assessed the morbidity of CMV disease with a nested matched cohort study. Cases were D+R- HTR with a diagnosis of CMV disease and controls were D+R- HTR matched by duration of VGCV prophylaxis and transplant date. The outcome measures (hospitalization and GCSF use by 30 days; invasive bacterial or fungal infection, biopsy-proven acute rejection, and mortality by 90 days) were assessed starting at the same time post-transplant of CMV disease diagnosis of the corresponding matched case patient.

Immunosuppression therapy, rejection, and other standard practices

Immunosuppression regimens were stable during the study period. Induction immunosuppression included either anti-thymocyte globulin (ATG) or an interleukin-2 receptor antagonist (for the few patients unable to tolerate ATG). Maintenance immunosuppression consisted of a combination of tacrolimus, mycophenolic acid, and prednisone. Patients generally underwent cardiac biopsy weekly for the first two weeks post-transplant, every other week until 3 months post-transplant, every month until 6 months post-transplant, then every other month until 1 year post-transplant; after 1 year they underwent yearly trans-thoracic echocardiogram. All cardiac biopsies were reviewed for the diagnosis of rejection. Acute cellular rejection was defined as allograft biopsy findings consistent with ISHLT-2004 guidelines grade ≥2R, which is equivalent to previous ISHLT-1990 grade ≥3A. Trimethoprim/sulfamethoxazole was given for pneumocystis prophylaxis for an indefinite duration.

Statistical analysis

The incidence and characteristics of CMV disease were stratified by donor/recipient serostatus. The duration of antiviral prophylaxis was assessed for all D+R- patients who survived at least 3 months after discontinuation of prophylaxis. To assess antiviral prophylaxis duration and other risk factors for CMV disease, we included univariate and multivariable models for D+R- patients who received 3 ± 1 months or 6 ± 1 months of prophylaxis; putative risk factors included baseline factors, time-dependent factors [acute rejection], and duration of antiviral prophylaxis. The multivariable logistic model that assessed risk factors for CMV disease included covariates with p-value <0.2 in univariate analysis as well as the duration of antiviral prophylaxis, which was forced into the model.

Inverse probability weighting was assessed to correct for confounding by indication, with consideration of the following as potential risk factors for choosing a specific duration of antiviral prophylaxis: recipient age, recipient gender, donor age, need for granulocyte-colony stimulating factor within 3 months after transplant, acute cellular rejection within 3 months after transplant, dose of anti-thymocyte globulin used for induction immunosuppression, reason for transplant, and pre-transplant diabetes mellitus. Chi-square and Fisher’s exact (categorical variables) and Mann-Whitney (continuous variables) tests were used to compare groups. A p-value <0.05 was considered significant. Cumulative incidence curves and Wilcoxon log-rank tests were used to examine time-dependent incidence of CMV disease. The Kaplan-Meier method was used to estimate survival.

Matched data were compared using McNemar’s test for proportions and Wilcoxon signed rank test for continuous variables.

Analyses were done in RStudio software version 1.0.153 (RStudio, Inc., Boston, MA, USA) or Stata software version 13.1 (Stata Corporation, College Station, TX, USA).

Literature review

A literature search, current through December 19, 2018, was performed using PubMed with the search terms cytomegalovirus, heart, cardiac. Studies were included if the duration of antiviral prophylaxis (among studies with 3 or 6 months duration) and incidence of CMV disease in D+R- heart transplant recipients could be accurately determined based on the data provided.

Results

Study population

Between July 5, 2005 and December 31, 2016, 332 patients underwent a first heart transplant. Twenty-two patients were excluded from the cohort: 15 received a transplant at another center (clinical details unavailable), 4 had received a prior heart transplant at our center, and 3 patients did not have sufficient information available; 310 patients were included in the analysis and their characteristics are shown in Table 1. Overall survival of the cohort at last follow-up was 87.1% [270/310] at a mean follow-up of 4.6 years, median of 3.7 years [IQR 1.8 – 7.3] and 0.6% [2/310] underwent re-transplantation.

Table 1.

Demographics of the study cohort.

Characteristic	Value (n = 310)
Recipient age
Mean	52.2
Median	54.9
IQR	44.8–62.3
Male sex (%)	226 (72.9)
Reason for transplant (%)
Ischemic cardiomyopathy	136 (43.9)
Idiopathic cardiomyopathy	89 (28.7)
Other	85 (27.4)
Donor/Recipient serostatus (%)
D-/R-	47 (15.2)
D+/R-	83 (26.8)
R+	180 (58.1)
ATG for induction (%)	305 (98.4)
Maintenance Immunosuppression (%)
Tacrolimus + MMF + Prednisone	269 (86.8)
Othera	41 (13.2)
Follow up time (months)
Mean	54.9
Median	44.1
IQR	21.8–87.3
Rejection ≥2R by last follow up (%)	23 (7.4)
Survival
1-year survival (%)	285 (91.9)

^aOther includes use of: Sirolimus, Azathioprine

Incidence and characteristics of CMV Disease

The incidence of proven CMV disease in the entire cohort (all donor/recipients serostatus combinations) was 8.4% [26/310], and was significantly higher in the D+R- group (p<0.001, data not shown). The clinical characteristics of CMV disease (syndrome, specific type of end-organ) are also shown in Table 2. 88.5% [23/26] of those with CMV disease were deemed ill enough to require hospitalization at time of diagnosis (90.5% of D+R- [19/21] and 80% of R+ [4/5] cases). An additional 17 cases of possible gastrointestinal CMV disease (12 in D+/R-, 4 in R+, and 1 in D-/R-) were not included because they did not meet the strict criteria for “proven” CMV disease. There were no cases of breakthrough CMV disease during receipt of antiviral prophylaxis, and no deaths within 90 days of CMV disease diagnosis.

Table 2.

Incidence and characteristics of CMV disease.

Proven CMV disease in each serostatus group	D+/R-n=21/83 (25.3%)	D-/R-n=0/47 (0%)	R+n=5/180 (2.8%)
Ganciclovir Resistant CMV	3/21 (13.6)	0	0
Time of diagnosis post- transplant, median months (IQR)	9.8 (6.5 – 14.4)	NA	16.7 (6.4 – 20.5)
CMV Disease type, no. of patients
CMV Syndrome	4/21 (19.0)	0	1/5 (20.0)
Tissue Invasive Disease (proven)	17/21 (80.9)	0	4/5 (80.0)
GI disease	15/21 (71.4)	0	3/5a (60.0)
Pneumonia	1/21 (4.8)	0	2/5a (40.0)
Retinitis	1/21 (4.8)	0	0

^aOne R+ patient with both GI and Pneumonia

Risk factors for CMV disease among CMV D+R- patients

Since the majority of CMV disease occurred among D+R- patients, we assessed risk factors specifically within this high-risk group. Among the 74/83 (89.1%) of D+R- patients who survived at least 3 months after completing prophylaxis (and thus were assessable for post-prophylaxis CMV disease), 58/74 (78.4%) received either 3 +/− 1 month [40/74, 54.1%] or 6 +/− 1 months [18/74, 24.3%] of antiviral prophylaxis, and were included in the analyses comparing 3 vs 6 months of prophylaxis. The remainder of patients (16/74 [22%]) received other durations of prophylaxis. Among patients who received either 3 or 6 months of prophylaxis, there were no baseline or time-dependent risk factors associated with CMV disease in univariate or multivariable models (Table 3). The dose of ATG did not differ between groups (mean 4.1, median 4.5, IQR 4.0–4.5 mg/kg in patients with CMV vs mean 4.3, median 4.5, IQR 4.5–4.5 mg/kg in patients without CMV). Four (9.5%) of the patients without CMV disease were diagnosed with acute cellular rejection ≥ 2R at a median of 13.7 months post-transplant (IQR 10.8 – 32.7); no patients with CMV disease developed subsequent rejection.

Table 3.

Risk factors for Proven CMV disease among D+R- OHT recipients^a

Comparison	CMV Disease n = 16	No CMV Disease n = 42	Unadjusted ORb	p-value	Adjusted ORc	p-value
Recipient Age
Mean	54.5	50.5
Median	60.5	53.8	1.02 (0.98, 1.07)	0.31
IQR	45.5–63.2	39.9–61.4
Male sex (%)	14 (87.5)	30 (71.4)	2.80 (0.55, 14.23)	0.21
Diabetes mellitus pre-tx (%)	7 (43.8)	14 (33.3)	1.56 (0.48, 5.05)	0.46
Ischemic cardiomyopathy as reason for transplant (%)	7 (43.8)	9 (21.4)	2.85 (0.83, 9.78)	0.10	2.95 (0.84, 10.37)	0.09
ATG for induction (%)	16 (100)	42 (100)		NA
Prophylaxis length
(6+/− 1 month)	7 (47.3)	11 (26.2)	2.19 (0.66, 7.30)	0.20	2.28c (0.66, 7.91)	0.19
(3 +/− 1 month)	9 (56.3)	31 (73.8)

^aIncluding only:

1. D+/R- recipients

2. Patients whose prophylaxis was 3 or 6 months (+/− 1 month)

3. OHT recipients who survived for at least 3 months after discontinuation of prophylaxis.

^bMultivariable analysis adjusted for duration of prophylaxis and ischemic cardiomyopathy

^cElevated odds ratio for CMV disease was associated with 6 months (+/− 1 month) of prophylaxis

To specifically assess durations of antiviral prophylaxis recommended in CMV guidelines, we directly compared the incidence of CMV disease among D+R- patients who received either 3 or 6 months of prophylaxis. We assessed the potential for confounding by indication (i.e. duration of prophylaxis) by using inverse probability weighting, and did not identify any factors associated with prophylaxis duration [data not shown]. Thus, inverse probability weighting was not deemed necessary for subsequent analyses. The incidence of proven CMV disease was not statistically different for 3 vs 6 months of prophylaxis, even after adjustment for other factors in a multivariable model (p=0.19, Table 3). The cumulative incidence of proven CMV disease is shown in Figure 1 and was not different for HTR who received 3 vs 6 months of prophylaxis.

Figure 1.

Cumulative incidence of CMV disease^a according to duration of prophylaxis

^aIncluding only:

1. D+/R- recipients

2. Patients whose prophylaxis was 3 or 6 months (+/− 1 month)

3. OHT recipients who survived for at least 3 months after discontinuation of prophylaxis.

We performed a separate analysis of all possible/probable/proven cases of CMV disease in a multivariable model, which also showed no association between CMV disease and 3 vs 6 months of valganciclovir prophylaxis (p=0.27, data not shown). Lastly, we analyzed valganciclovir prophylaxis duration as a continuous variable among D+R- recipients who received any duration of prophylaxis (n=74) and again found no association between CMV disease and duration of valganciclovir (p=0.28). Figure 2A shows the duration of prophylaxis between groups of patients who did or did not develop CMV disease in a box and whiskers plot. Figure 2B shows the CMV disease status (red = CMV disease, blue = no CMV disease) for each patient according to the duration of valganciclovir prophylaxis that each patient received. As shown, the proportion with CMV disease was similar among those who received longer or shorter durations of antiviral prophylaxis.

Figure 2.

Duration of prophylaxis and CMV disease among D+R- recipientsa^a

A. Median, IQR, and range of vanganciclovir prophylaxis duration among groups of patients who did and did not develop CMV disease.

B. The CMV disease status (red= CMV disease, blue=no CMV disease) for each patient is shown according to the duration of prophylaxis for that individual patient.

*This patient was misclassified as D-R- and thus did not receive any antiviral prophylaxis, and was not included in any of the analyses.

^aIncluding only:

1. D+/R- recipients who received any length of VGCV prophylaxis

2. OHT recipients who survived for at least 3 months after discontinuation of prophylaxis.

Measures of morbidity in patients with CMV disease

We matched 16 D+R- CMV disease case patients 1:1 with D+R- control patients who did not have CMV disease; cases and controls were matched on duration of prophylaxis (3 vs 6 months) and transplant date. Demographics of the cases and controls are shown in Supplemental Table 1. There were no clinically significant fungal or bacterial infections in either cases or controls in the 90 days following CMV disease diagnosis. A significantly higher proportion of CMV disease patients were hospitalized as compared to controls (87.5% [14/16] vs 6.3% [1/16], p<0.001) and they also had a higher number of hospital days within the next 30 days (median 4 [IQR 3 – 7] vs median 0 [IQR 0 – 0], p=0.001). There were few doses of G-CSF given in either group (Table 4). There was no difference in 90 day mortality or acute allograft rejection between groups (0 deaths; 0 episodes of rejection).

Table 4.

Morbidity associated with CMV disease

	Cases n=16 (%)	Controls n=16 (%)	p-valuea
Patients hospitalized (%)	14 (87.5)	1 (6.3)	< 0.001
Days hospitalized
Mean	4.9	0.125	0.001
Median	4	0
IQR	3–7	0–0
Patients who received GCSF (%)	3 (18.7)	1 (6.3)	0.62
Doses GCSF
Mean	0.31	0.06	0.34
Median	0	0
IQR	0–0	0–0

^aProportions compared using McNemar’s test, continuous variables compared using Wilcoxon signed rank test

Analysis of prior cohort studies

Ten studies included adequate data to assess the incidence of CMV disease after antiviral prophylaxis for 3 or 6 months in D+R- HTR; details of those studies are in Supplemental Table 2. The pooled 6-month CMV disease incidence was 10.5% [10/95] and the pooled 1-year CMV disease incidence was 21.2% [41/193]; although these estimates come from a heterogeneous group of studies, they were similar to 6-month and 1-year estimates in our cohort of D+R- recipients (6-month CMV disease incidence 5.2% [3/58] and 1-year CMV disease incidence 17.2% [10/58]).

Discussion

In a large cohort of HTR who received standardized clinical care and had excellent overall survival, we confirmed that the highest incidence of CMV disease occurred in D+R- recipients and was associated with significant morbidity. We also demonstrated that a 6 month duration of prophylaxis was not associated with a reduced incidence of CMV disease compared to 3 months. These findings extend the results of a prior study in HTR [8], and suggest that routine use of 6 months of valganciclovir prophylaxis in HTR may not be warranted, especially given the significant cost, toxicities, and risks for resistance with longer durations of ganciclovir/valganciclovir.

The apparent lack of benefit on CMV disease incidence with 6 months of antiviral prophylaxis in the present study is in contrast to randomized, controlled studies in kidney and lung transplantation demonstrating benefits of longer duration of antiviral prophylaxis [4, 5]. However, in those populations, even prior to the conduct of controlled trials, observational studies demonstrated reduced CMV disease incidence with longer durations of prophylaxis [11–13], and served as the rationale to conduct controlled trials that demonstrated benefits of longer prophylaxis durations. There are multiple potential explanations for the lack of observed benefit of longer prophylaxis duration on CMV disease incidence in our study. The generally high level of immunosuppression in HTR might counterbalance any potential benefit of longer duration of prophylaxis. It is possible that there was bias in selecting longer duration of prophylaxis for patients anticipated to be at higher risk for CMV disease, but we used inverse probability weighting method to correct for potential confounding by indication. However, since this study is not a randomized controlled trial, the possibility of unmeasured differences between the groups remains possible. Also, the sample size in this study might not have been large enough to detect a difference, although at least a trend towards lower CMV disease incidence would have been expected, as shown previously in kidney and lung transplant recipients.

Given the continued high rates of CMV disease and associated morbidity in D+R- HTR, novel strategies such as vaccination, immune-guided prophylaxis duration, or adoptive immunotherapy should be studied in this population. Newer antiviral agents that appear to have lower hematologic toxicity might be more tolerable than current options, but would need to be specifically studied [14, 15]. Approaches such as CMV monitoring and pre-emptive therapy after completion of antiviral prophylaxis have not specifically been reported in HTR, but have not been associated with reduced CMV disease incidence in other D+R- settings [16–19]. A recent randomized trial of PET vs prophylaxis in D+R- liver transplant recipients demonstrated a significant reduction in CMV disease with PET, and is also an approach that merits evaluation in the setting of D+R- HTR [20].

Strengths of this study included the use of well-validated definitions of CMV disease and prophylaxis durations reflective of those recommended in current international consensus CMV guidelines [6, 7]. Additionally, the presence of confounding by indication was assessed by inverse propensity weighting analysis. The observed overall incidence of CMV disease in our D+R- HTR cohort was similar to previously published cohorts and provides some reassurance of the generalizability of results to other heart transplant populations. We acknowledge that the size of our cohort limited the difference in CMV disease that could be detected between the 3 month vs 6 month groups; however, the number of patients in this study was still relatively large in comparison to many studies focused on D+R- HTR, and we did not observe even a trend towards benefit of a longer duration of prophylaxis. Other potential limitations included the retrospective single center design and lack of randomization to different prophylaxis duration.

In summary, the results of the present study suggest that longer durations of VGCV prophylaxis did not reduce CMV disease incidence compared to standard 3 month duration in high-risk D+R- HTR. Thus, given the toxicities, cost, and potential for resisistance, VGCV prophylaxis durations longer than 3 months can not routinely be recommended in high-risk D+R- HTR, and the duration of CMV prophylaxis should be specifically studied in randomized controlled trials in HTR, rather than extrapolated from other transplant populations. Novel strategies to prevent CMV disease and associated morbidities in D+R- HTR are warranted.

Key Points:

Valganciclovir prophylaxis longer than 3 months for prevention of CMV disease in high-risk donor seropositive/recipient serongegative heart transplant recipients was not associated with reduced incidence of CMV disease compared to 3 months, and cannot be routinely recommended.