Abstract
Background:
Cytomegalovirus (CMV) infection has broad implications for morbidity and mortality in renal transplant recipients (RTR). Routine surveillance for CMV replication with PCR-based quantitative nucleic acid testing (qNAT) assays is standard practice in most transplant centers, but the impact of assay sensitivity on antiviral decision-making and virologic outcomes has not been studied. We investigated the effects of an ultrasensitive CMV qNAT assay on multiple clinical outcomes, including time to detection and duration of CMV DNAemia.
Methods:
We conducted a single-center cohort study contrasting RTRs monitored with a qNAT with a higher lower limit of quantification (LLOQ> 300 IU/mL) with those monitored with a more sensitive qNAT (LLOQ> 35 IU/mL). Patients were stratified by donor (D)/recipient (R) CMV serostatus (D+/R-, high-risk; any R+, moderate-risk). CMV viral load monitoring was performed monthly post-transplantation, with the primary outcomes being time to CMV DNAemia and its duration.
Results:
1382 patients were analyzed from 2014–2016 and 2019–2021. Moderate-risk RTRs monitored with the more sensitive assay experienced a greater hazard for the development of a first episode of CMV DNAemia (aHR- 1.95 95% CI- 1.55 to 2.46) and an average of 24 ( 95% CI- 16.40 to 31.98) additional days of DNAemia. There was no difference in CMV end-organ disease or one-year all-cause mortality between moderate-risk RTRs.
Conclusions:
The more sensitive assay was associated with earlier detection and extended durations of CMV DNAemia in moderate-risk RTRs, without altering clinical outcomes. These findings inform optimal use of these assays and antiviral stewardship in RTRs.
Keywords: CMV, Renal transplant recipients, low level DNAemia
Key Summary:
The use of ultrasensitive CMV qNAT assays in moderate-risk CMV renal transplant recipients is associated with earlier detection and longer durations of CMV DNAemia without impacting CMV end-organ disease or one-year mortality.
Introduction
Cytomegalovirus (CMV) remains an important cause of morbidity and mortality in renal transplant recipients (RTR)1. The virus is responsible for a broad spectrum of conditions in this population, ranging from asymptomatic CMV DNAemia to tissue-invasive CMV disease2–6. Most transplant centers employ a strategy of routine CMV viral load monitoring using quantitative nucleic acid amplification testing (qNAT) assays to help determine the need for antiviral therapy7–10.
Despite the benefits of quantitative CMV viral load monitoring in the post-transplant period, universally accepted thresholds to guide antiviral management are lacking. Even with the international calibration standard proposed by the World Health Organization (WHO), substantial variation persists between different qNAT assays.11,12. As a result, transplant centers are recommended to establish institution-specific thresholds for antiviral management1.
The lack of universal CMV viral load thresholds for antiviral management is further complicated by the differing sensitivity of CMV testing platforms. The advent of more sensitive qNAT assays has enabled the detection of previously undetectable low-level CMV DNAemia, and the clinical implications of treating or not treating low-level DNAemia remain unexplored.
To address this knowledge gap, we analyzed the impact of a change in our institution’s qNAT testing platform on viral kinetics of previously undetectable levels of CMV DNAemia, and investigated the relationship between type of qNAT used and a variety of clinical outcomes in RTR, including: the time to the first episode of CMV DNAemia, the total duration of detectable CMV DNAemia, the duration of low-level CMV DNAemia, the duration of CMV DNAemia after the 1000 IU/mL threshold is reached in moderate-risk patients, peak CMV DNAemia, the magnitude of the initially detected CMV viral load, the odds of developing end-organ CMV disease, and one-year all-cause mortality.
Methods
Study Design and Participants
We conducted a cohort study of renal transplant recipients (RTR) at a single large tertiary-care hospital and transplant center. In April 2018, the CMV qNAT assay platform at this institution changed from one whose lower limit of quantitation (LLOQ) was 300 IU/mL (“high LLOQ qNAT”- Thermo Fisher 7500) to one whose LLOQ was 35 IU/mL (“lower LLOQ qNAT” – Roche COBAS). With the higher LLOQ qNAT assay, viral loads <100 IU/mL were reported as undetectable, those between 100–300 were reported as “detectable but not quantifiable”, and those ≥300 IU/mL were quantitated. In contrast, with the lower LLOQ qNAT assay, viral loads that were 1–35 IU/mL were reported as “detectable but not quantifiable”, and those ≥35 IU/mL were quantitated.
RTRs undergo routine post-transplant surveillance for CMV DNAemia with at least monthly CMV qNAT assays for the first 12 months post-transplantation. For this analysis, we included RTRs who had been transplanted over a three-year period before (January 1, 2014-December 31, 2016) and after (January 1, 2019-December 31, 2021) the assay change; we excluded those transplanted in 2017 and 2018 since these patients could have undergone monitoring using either assay. Patients needed to have received all transplant-related care (e.g., CMV monitoring) at the study institution and completed one year of post-transplantation follow-up with CMV viral load monitoring after the conclusion of antiviral prophylaxis.
Patients were categorized based on the type of PCR platform used for CMV testing (higher LLOQ vs. lower LLOQ) and further stratified by CMV donor/recipient serostatus (high-risk, donor CMV positive/recipient CMV negative, or D+/R-; moderate-risk, any recipient CMV positive, or R+); patients who were D-/R- were excluded from the analysis.
Data Collection
We collected all CMV viral load measurements in the first 12 months post-transplant for RTRs included in the cohort. CMV viral loads collected during the period of routine post-transplant antiviral (valganciclovir) prophylaxis (given for 6 months and 3 months after transplant in D+/R- and R+ patients respectively) were excluded from the analysis due to the low occurrence of breakthrough CMV DNAemia, and to minimize confounding related to immunosuppression management and antiviral underdosing. Patients were considered lost to follow-up if they lacked any viral load measurements between the 10th and 12th months post-transplantation (See Figure 1).
Figure 1:
Flow diagram illustrating the step-by-step patient selection process for the study, detailing the number of patients in each subgroup and those who died. The higher LLOQ assay quantifies viral loads greater than 300 IU/mL, reports viral loads between 100 – 300 IU/mL as detected but not quantified, and is unable to detect viral loads less than 100 IU/mL. The lower LLOQ assay quantifies viral loads greater than 35 IU/mL and reports viral loads between 1–35 IU/mL as detected but not quantified. Abbreviations: LLOQ- lower limit of quantitation. qNAT- quantitative nucleic acid testing.
Data were stored in the Emory renal transplant database, which directly retrieved the information from the electronic medical health record.
Variables and Definitions
Exposure
The primary exposure was the qNAT platform used to monitor patients for CMV DNAemia. We specifically compared outcomes among those patients monitored with the higher LLOQ qNAT assay with those monitored with the lower LLOQ qNAT assay and stratified our analysis by CMV risk status.
Outcomes
The primary outcomes of our study were the duration and the time to onset of CMV DNAemia. The time to CMV DNAemia was calculated as the time to the first episode of CMV DNAemia for each subgroup after the protocol-defined window of antiviral prophylaxis had ended. The duration of CMV DNAemia was measured by the number of days that a patient had detectable DNAemia. Each unique episode of CMV DNAemia was included. The duration of low-level CMV DNAemia was measured by the number of days a patient experienced CMV DNAemia where the first recorded viral load fell within the low-level DNAemia range. Low level CMV DNAemia was defined as the period of detectable but not quantifiable DNAemia, which differed between the two qNAT assays (100–300 IU/mL for the higher LLOQ assay and 1–35 IU/mL for the lower LLOQ assay). At our center, moderate-risk patients that are asymptomatic are typically only initiated on antiviral therapy after reaching a threshold of 1000 IU/mL, and so we also calculated the duration of CMV DNAemia in days after this threshold was reached. An episode of DNAemia was calculated from the first detected viral load, based on the definitions above, to the first of two consecutive undetectable viral load results, which were collected at least 5 days apart.
Secondary outcomes included peak CMV DNAemia, the magnitude of the first detected CMV viral load, the odds of developing CMV disease, and one-year all-cause mortality. Peak CMV DNAemia was quantified as the highest viral load detected during an episode of CMV DNAemia. Magnitude of the first detected CMV viral load was defined as the viral load at the first detection of DNAemia. Using ICD 9 and ICD 10 codes, patients who developed end organ CMV disease between the conclusion of antiviral prophylaxis and the first year after transplantation were identified. One-year all-cause mortality was defined as any death occurring during the first year after transplantation.
Covariables
Before the data analysis, we constructed a directed acyclic graph (DAG) to explore the relationship between the exposure and outcome. This graph helped identify potential confounding variables that might impact the results (see Supplementary Figures 1–2). Induction and maintenance immunosuppression were identified as potential confounders because local policy changes in immunosuppression protocols changed during the same period as the change in qNAT assay. These potential confounders were included as covariables in the Cox proportional hazards model, linear regression model, and logistic regression model. Notably, these variables consisted of the type of induction immunosuppression used at the time of transplantation, the maintenance immunosuppressive regimen used after transplantation (specifically whether the patient was treated with belatacept, tacrolimus, or other agent, in addition to an antimetabolite and corticosteroid).
Statistical Analysis
The baseline characteristics of each subgroup (higher LLOQ moderate/high CMV risk and lower LLOQ moderate/high CMV risk) were summarized by proportion and measures of central tendency. Baseline categorical data were presented as percentages. The incidence rate of CMV DNAemia was calculated for patients who were CMV moderate-risk (R+), with 9 months of follow-up time, and CMV high-risk (D+/R-), with 6 months of follow-up time. Follow-up time was defined as the period of monthly collection of CMV viral loads up to one year after transplantation.
Kaplan Meier curves were generated to visualize the time to the first episode of CMV DNAemia, contrasting the moderate-risk groups with each other and the high-risk groups with each other. We used Cox proportional hazards models for time to CMV DNAemia, linear regression for duration of CMV DNAemia, and logistic regression for the odds of developing CMV end-organ disease. The assumption of proportional hazards was checked both graphically and statistically. Mean peak CMV DNAemia and the mean first detected CMV viral load were tested for significant differences by T-test. After examination of the pattern of missing data suggested data were missing at random, missing values were imputed by multiple imputation. To assess the robustness of our findings and the potential influence of unmeasured confounders, we performed a sensitivity analysis to compute the E-value for each model13,14. Data analyses were performed using R and R studio version 4.2.2.
Results
Baseline Demographics
We identified 611 RTRs (509 CMV moderate-risk, 102 CMV high-risk) who were monitored with the higher LLOQ platform and 771 (641 CMV moderate-risk, 130 CMV high-risk) monitored with the lower LLOQ platform.
Baseline demographic characteristics were similar between the two groups (Table 1). A higher proportion of those tested with the higher LLOQ qNAT platform received thymoglobulin induction agent. A lower proportion of CMV high-risk patients received belatacept as part of their maintenance immunosuppression during the period when the lower LLOQ qNAT was in use. A total of 12993 CMV viral loads were measured corresponding to 428 discrete episodes of CMV DNAemia. The incidence rate of CMV DNAemia varied based on patient serostatus and qNAT testing platform. Notably, the absence of CMV viral loads from month 10 through month 12 can be attributed to laboratory collections occurring outside the study institution’s facilities or patient mortality.
Table 1:
Baseline patient characteristics of renal transplant recipients stratified by CMV risk status and quantitative nucleic acid amplification test performed (higher LLOQ vs. lower LLOQ). Demographic and clinical variables are summarized as means with standard deviations for continuous variables and as frequencies with percentages for categorical variables.
| Lower LLOQ high risk (n=130) | Lower LLOQ moderate risk (n=641) | Higher LLOQ high risk (n=102) | Higher LLOQ moderate risk (n=509) | Lower LLOQ combined (n=771) | Higher LLOQ combined (n=611) | |
|---|---|---|---|---|---|---|
|
| ||||||
| Age | ||||||
| Mean (SD) | 50.0 (13.9) | 52.4 (13.1) | 49.5 (12.4) | 50.9 (12.8) | 52.0 (13.2) | 50.7 (12.8) |
|
| ||||||
| Sex | ||||||
| Female | 43 (33.1%) | 302 (47.1%) | 38 (37.3%) | 229 (45.0%) | 344 (44.7%) | 267 (43.7%) |
| Male | 87 (66.9%) | 339 (52.9%) | 64 (62.7%) | 280 (55.0%) | 426 (55.3%) | 344 (56.3%) |
|
| ||||||
| Donor Type | ||||||
| Deceased | 87 (66.9%) | 490 (76.4%) | 58 (56.9%) | 331 (65.0%) | 576 (74.8%) | 389 (63.7%) |
| Living | 43 (33.1%) | 151 (23.6%) | 44 (43.1%) | 177 (34.8%) | 194 (25.2%) | 221 (36.2%) |
| Unreported | 0 (0%) | 0 (0%) | 0 (0%) | 1 (0.2%) | 0 (0%) | 1 (0.2%) |
|
| ||||||
| Induction Agent | ||||||
| Basiliximab | 128 (98.5%) | 624 (97.3%) | 88 (86.3%) | 457 (89.8%) | 751 (97.5%) | 545 (89.2%) |
| Other | 0 (0%) | 1 (0.2%) | 0 (0%) | 3 (0.6%) | 1 (0.1%) | 3 (0.5%) |
| Thymoglobulin | 0 (0%) | 6 (0.9%) | 12 (11.8%) | 39 (7.7%) | 6 (0.8%) | 51 (8.3%) |
| Missing | 2 (1.5%) | 10 (1.6%) | 2 (2.0%) | 10 (2.0%) | 12 (1.6%) | 12 (2.0%) |
|
| ||||||
| Maintenance Agent | ||||||
| Belatacept | 48 (36.9%) | 551 (86.0%) | 78 (76.5%) | 319 (61.0%) | 598 (77.7%) | 462 (75.6%) |
| Tacrolimus | 81 (62.3%) | 72 (11.2%) | 7 (6.9%) | 164 (31.4%) | 153 (19.9%) | 74 (12.1%) |
| Unknown | 0 (0%) | 1 (0.2%) | 0 (0%) | 0 (0%) | 1 (0.1%) | 0 (0%) |
| Other | 0 (0%) | 9 (1.4%) | 15 (14.7%) | 29 (5.5%) | 9 (1.2%) | 64 (10.5%) |
| Missing | 1 (0.8%) | 8 (1.2%) | 2 (2.0%) | 11 (2.1%) | 9 (1.2%) | 11 (1.8%) |
|
| ||||||
| Number of patients with missing CMV viral loads month 10 – month 12 | 13 | 73 | 9 | 100 | 86 | 109 |
|
| ||||||
| Number of discrete episodes of CMV DNAemia | 24 | 241 | 35 | 128 | 265 | 163 |
|
| ||||||
| Number of patients with more than one episode of CMV DNAemia | 4 | 31 | 7 | 27 | 35 | 34 |
|
| ||||||
| Number of recurrent CMV DNAemia episodes | 4 | 33 | 7 | 29 | 37 | 36 |
|
| ||||||
| Number of discrete episodes of low level CMV DNAemia | 4 | 156 | 8 | 38 | 160 | 46 |
|
| ||||||
| Number of discrete episodes of CMV DNAemia exceeding 1000 IU/mL threshold | -- | 64 | -- | 62 | -- | -- |
|
| ||||||
| Total number of CMV viral loads | 1106 | 6412 | 1055 | 4420 | 7518 | 5475 |
|
| ||||||
| Total number of low level CMV viral loads | 103 | 671 | 86 | 178 | 774 | 264 |
|
| ||||||
| Follow up time in person years | 65.00 | 482.25 | 51.00 | 381.75 | 547.25 | 432.75 |
|
| ||||||
| CMV DNAemia incidence rate per year | 0.97 | 0.76 | 1.24 | 0.41 | 0.78 | 0.51 |
|
| ||||||
| Number of patients with CMV end-organ disease | 0 | 2 | 4 | 2 | 2 | 6 |
Abbreviations: CMV- cytomegalovirus. LLOQ- lower limit of quantitation. SD- standard deviation
CMV Viral Kinetics
Tables 2, 3 and Figure 2 present the findings for the primary outcomes of time to CMV DNAemia, duration of CMV DNAemia, duration of low-level CMV DNAemia, and duration of CMV DNAemia after the 1000 IU/mL threshold is reached in moderate-risk patients.
Table 2: Moderate-risk CMV RTRs tested with the lower LLOQ assay have a greater hazard for developing a first episode of CMV DNAemia.
Univariable and multivariable Cox proportional hazards models for time to first episode of CMV DNAemia among renal transplant recipients stratified by CMV risk status and quantitative nucleic acid amplification test platform. Hazard ratios (HRs) and 95% confidence intervals are reported for the comparisons. Multivariable models are adjusted for induction and maintenance immunosuppression.
| Moderate Risk Comparison | High Risk Comparison | |||
|---|---|---|---|---|
| Variable | Unadjusted hazard ratio (95% CI) | Adjusted hazard ratio (95% CI)ψ | Unadjusted hazard ratio (95% CI) | Adjusted hazard ratio (95% CI) ψ |
| Lower LLOQ | 1.87 (1.51 to 2.31) | 1.95 (1.55 to 2.46) | 1.21 (0.82 to 1.77) | 1.35 (0.84 to 2.15) |
| Higher LLOQ | REF | REF | REF | REF |
- Adjusted for induction and maintenance regimen used
Abbreviations – CMV- cytomegalovirus. RTR- renal transplant recipient. LLOQ- lower limit of quantitation. CI- confidence interval
Table 3: Moderate risk CMV RTRs tested with the lower LLOQ assay have on average longer durations of CMV DNAemia, low-level CMV DNAemia, and CMV DNAemia after reaching a threshold of 1000 IU/mL.
Univariable and multivariable linear regression models for durations of CMV DNAemia, low-level CMV DNAemia, and CMV DNAemia duration after exceeding the 1000 IU/mL threshold among renal transplant recipients stratified by CMV risk status and quantitative nucleic acid amplification test platform. Point estimates and 95% confidence intervals are reported for the comparisons. Multivariable models are adjusted for induction and maintenance immunosuppression.
| Moderate Risk Comparison | High Risk Comparison | |||
|---|---|---|---|---|
| Variable | Unadjusted Linear Regression Estimate (95% CI) | Adjusted Linear Regression Estimate (95% CI)ψ | Unadjusted Linear Regression Estimate (95% CI) | Adjusted Linear Regression Estimate (95% CI) ψ |
| CMV DNAemia duration | ||||
| Lower LLOQ | 24.14 (17.24 to 31.04) | 24.79 (17.73 to 31.85) | 10.31 (−8.93 to 29.55) | 20.75 (−6.43 to 47.93) |
| Higher LLOQ | REF | REF | REF | REF |
| Low-level CMV DNAemia duration | ||||
| Lower LLOQ | 22.94 (14.85 to 31.03) | 23.71 (15.59 to 31.84) | -10.92 (−32.28 to 10.43) | -0.69 (−28.26 to 26.89) |
| Higher LLOQ | REF | REF | REF | REF |
| CMV DNAemia duration after exceeding 1000 IU/mL threshold | ||||
| Lower LLOQ | 22.87 (15.38 to 30.36) | 24.19 (16.40 to 31.98) | ||
| Higher LLOQ | REF | REF | ||
- Adjusted for induction and maintenance regimen used
Abbreviations- CMV- cytomegalovirus. RTR- renal transplant recipient. LLOQ- lower limit of quantitation. CI- confidence interval
Figure 2:
A- Kaplan Meier curve analysis comparing time to first episode of CMV DNAemia for higher LLOQ CMV moderate risk patients (n=509) with lower LLOQ CMV moderate risk patients (n=641) over a one year follow up period after transplantation. Time to CMV DNAemia was significantly shorter for patients monitored with the LLOQ platform (log-rank test p-value <0.0001). The median time to CMV DNAemia for the lower LLOQ group was 333 days and could not be calculated for the higher LLOQ group given that the survival probability was greater than 50% at the end of the study period. The horizontal blue line represents the time at which antiviral prophylaxis was discontinued. B- Kaplan Meier curve analysis comparing time to first episode of CMV DNAemia for higher LLOQ CMV high risk patients (n=102) with lower LLOQ CMV high risk patients (n=130) over a one year follow up period after transplantation. Time to CMV DNAemia was not significantly different by (log-rank test p-value = 0.78). The median time to CMV DNAemia for the higher LLOQ group and lower LLOQ group was 363 days and 360 days, respectively. The horizontal blue line represents the time at which antiviral prophylaxis was discontinued. Abbreviations- CMV- cytomegalovirus. LLOQ- lower limit of quantitation. qNAT- quantitative nucleic acid testing.
Among patients tested with the lower LLOQ platform, the hazard rate for the first episode of CMV DNAemia was 1.95 times higher (95% CI: 1.55 to 2.46) compared to those tested with the higher LLOQ platform, adjusting for the types of induction and maintenance immunosuppression regimens administered. For the high-risk group, a numerical trend suggested a greater hazard rate in patients tested with the lower LLOQ platform than those tested with the higher LLOQ platform; however, this trend did not achieve statistical significance when accounting for the induction and maintenance immunosuppression regimens used. Non-significant p-values in the Schoenfeld residuals test and graphical assessment of scaled Schoenfeld residuals versus time indicated that the proportional hazards assumption was valid.
Figure 2 displays the Kaplan-Meier curves illustrating the time to the first episode of CMV DNAemia for both moderate and high-risk CMV patients. In the moderate-risk group, patients tested with the lower LLOQ platform had earlier onset of CMV DNAemia compared to those tested with the higher LLOQ platform. However, for the high-risk group, there was no difference in the time to the initial episode of CMV DNAemia between the lower and higher LLOQ testing platforms.
Linear regression modeling allowed precise estimates of CMV DNAemia duration produced by platforms with differing LLOQ. Regarding the duration of CMV DNAemia, on average, moderate-risk patients tested with the lower LLOQ platform experienced CMV DNAemia 24.79 days longer than those tested with the higher LLOQ platform, adjusting for induction and maintenance immunosuppression regimens. In the high-risk group, a numerical trend was noted toward longer durations of CMV DNAemia in patients tested with the lower LLOQ platform, although this difference was not statistically significant. For moderate-risk patients who developed low level CMV DNAemia initially, on average, those tested with the lower LLOQ platform experienced CMV DNAemia 22.94 days longer than those tested with the higher LLOQ platform, adjusting for induction and maintenance immunosuppression regimens. There was no significant difference noted for high-risk patients who developed low level CMV DNAemia. Finally, for CMV DNAemia duration after exceeding the 1000 IU/mL threshold, moderate-risk patients tested with the lower LLOQ platform had on average 24.19 days of additional CMV DNAemia compared to those tested with the higher LLOQ platform, adjusting for induction and maintenance regimens used (See Table 3).
Regardless of CMV risk status, patients tested with the lower LLOQ platform had a lower peak CMV viral load and a lower first detected CMV viral load compared to those tested with the higher LLOQ platform (See Table 4). Moderate risk CMV RTRs tested with the lower LLOQ platform had a longer time to peak CMV and 1000 IU/ml viral loads compared to those tested with the higher LLOQ platform.
Table 4: Regardless of CMV serostatus, RTRs tested with the lower LLOQ assay had lower mean peak CMV viral loads and lower first detected CMV viral loads. Moderate risk CMV RTRs had a longer time to peak CMV viral load when tested with the lower LLOQ platform.
Comparisons between higher LLOQ and lower LLOQ moderate and high-risk groups including peak CMV DNAemia and first CMV viral load detected. Means, standard deviations, and t-tests are reported where applicable.
| Moderate Risk (CMV R+) Comparison | T-test p-value (95% CI) | High Risk (CMV D+/R-) Comparison | T-test p-value (95% CI) | |||
|---|---|---|---|---|---|---|
| Higher LLOQ (n=509) | Lower LLOQ (n= 641) | Higher LLOQ (n= 102) | Lower LLOQ (n= 130) | |||
| Peak CMV DNAemia log10 mean (std dev) IU/mL | 3.06 (0.77) | 1.94(0.90) | <0.001 (0.94 to 1.29) | 4.94 (1.01) | 3.84 (1.29) | <0.001 (0.66 to 1.54) |
| Time to Peak CMV viral load (std dev) IU/mL | 42.43 (60.80) | 66.45(57.46) | 0.03 (−46.76 to −1.28) | 47.00 (47.43) | 42.66 (38.38) | 0.68 (−16.28 to 24.97) |
| Time to 1000 IU/mL threshold | 9.45 (26.20) | 42.17 (52.73) | <0.001 (−47.38 to −18.06) | -- | -- | -- |
| First CMV viral load detected log10 mean (std dev) IU/mL | 2.81 (0.74) | 1.54 (0.58) | <0.001 (1.12 to 1.42) | 3.90 (1.25) | 2.87 (1.25) | <0.001 (0.54 to 1.51) |
Abbreviations: CMV- cytomegalovirus. RTR- renal transplant recipient. LLOQ- lower limit of quantitation. CI- confidence interval
Clinical Outcomes
There was no difference in the odds of developing CMV disease based on qNAT platform used for the moderate-risk CMV patients (See Supplementary Table 1). When considering one-year mortality, moderate-risk CMV patients tested with the lower LLOQ platform exhibited a risk that was 0.91 times (95% CI: 0.45–1.84) that of patients tested with the higher LLOQ platform. Among high-risk CMV patients, no deaths occurred in the group tested with the lower LLOQ platform, whereas 3 patients tested with the higher LLOQ platform died within one year (See Supplementary Table 2 for more information regarding cause of death).
Sensitivity Analysis
We conducted an E-value sensitivity analysis for the moderate-risk comparison for the primary outcomes. For the adjusted hazard ratio, the E-value was determined to be 2.57, with a lower confidence limit E-value of 2.09. For the linear regression coefficient for CMV DNAemia, the E-value was 3.32, with a lower confidence limit E-value of 2.62. For the linear regression coefficient for low level CMV DNAemia, the E-value was 3.46, with a lower confidence limit E-value of 2.53. For the linear regression coefficient for CMV DNAemia duration after the 1000 IU/mL threshold was reached, the E-value was 4.61, with a lower confidence limit E-value of 3.01.
Discussion
Impact of Ultrasensitive CMV PCR Platform
Our study, drawing on a robust renal transplant database at Emory, examined CMV viral kinetics and clinical outcomes in this patient population. We found that use of a more sensitive qNAT platform was associated with earlier CMV DNAemia detection and longer durations of detectable DNAemia, regardless of initially detected viral load, in CMV moderate-risk patients. Among all patients, use of the more sensitive assay was also associated with lower peak and initial CMV viral loads. The E-value sensitivity analysis added further strength to our study by indicating that any unmeasured confounding would need to be considerable to negate our observed associations.
Low-level CMV DNAemia
Prior research offers insight into the kinetics of low-level CMV DNAemia that can guide antiviral stewardship. A study by Natori et al. identified clinical predictors of clearance for low-level CMV DNAemia (viral loads between 137 and 999 IU/mL) in solid organ transplant recipients. Remarkably, spontaneous clearance of DNAemia was common even among high-risk CMV serostatus patients, especially if they had experienced a prior CMV DNAemia episode15. This suggests that careful monitoring without immediate antiviral treatment could be feasible for patients with low-level CMV DNAemia. Furthermore, antiviral stewardship programs have advocated for preemptive therapy in CMV prevention and the use of CMV cell-mediated immune assays to guide duration of surveillance monitoring16,17. Preemptive therapy in liver transplant recipients was shown to enable immune system priming, leading to the development of neutralizing antibodies and reducing the risk of delayed-onset post-prophylaxis CMV disease, compared to those on antiviral prophylaxis18. These findings, coupled with our own, suggest that controlled exposure to low-level CMV DNAemia could potentially enhance CMV-specific immunity and antiviral stewardship, without compromising patient outcomes.
Mechanistic Insights
Our study demonstrates that the more sensitive assay detects a greater number of viral loads that fall within the low-level CMV DNAemia range, regardless of CMV risk status. For moderate-risk CMV RTRs this amounts to earlier detection before reaching the 1000 IU/mL threshold at which antiviral treatment is initiated. At the study institution, patients are treated with antivirals until they achieve two consecutive negative viral loads. The more sensitive assay detects these tail viral loads for a longer period for the moderate-risk CMV patients leading on average to an additional 24 days of CMV DNAemia. This bears significant implications for antiviral stewardship in renal transplant recipients. Given that we treat patients with a renally adjusted dose of valganciclovir, an extra 24 days of CMV DNAemia could lead to additional costs ranging from $103 to $821 per CMV DNAemia episode (based on current valganciclovir prices), underlining the economic impact detecting prolonged CMV DNAemia. For time to CMV DNAemia, the frequency of viral load monitoring could mask differences in the high-risk group, and a biweekly screening interval might provide more detailed insight into viral kinetics. The lack of a difference in CMV DNAemia duration in the high-risk patients may have been due to a smaller sample size and a shorter follow up period. We performed a power analysis and calculated an effect size of 0.372, which corresponds to an 80% chance of detecting a difference of 13.5 days or more.
Limitations and Future Directions
We acknowledge some limitations, such as a smaller sample size and shorter follow up period for high-risk CMV patients, the potential lack of generalizability beyond the renal transplant population, and lack of direct antiviral use and immunosuppression reduction data analysis. Future investigations should extend to liver, lung, and heart transplant recipients to determine the implications of low-level CMV DNAemia detection in these patients. Direct analysis of antiviral use patterns was not possible with our database due to missing parameters such as creatinine clearance, which is essential to differentiate between an antiviral treatment dose vs secondary prophylaxis dose. However, the differences in CMV DNAemia duration serve as a surrogate for differences in antiviral use duration given that the duration of antiviral use is dependent directly on the duration of CMV DNAemia. Future studies examining the impact that ultrasensitive assays have on the time to reduction of immunosuppression maintenance regimens are needed.
Conclusion
In summary, our study underscores the clinical implications of employing an ultra-sensitive PCR platform in monitoring CMV DNAemia in RTRs. We observed earlier detection of CMV DNAemia, longer durations of DNAemia, and lower peak and initial viral loads, particularly among moderate-risk patients. Despite these variations in viral kinetics, we did not observe significant differences in the odds of developing end-organ CMV disease or the risk for one-year mortality. Furthermore, our findings highlight the potential economic impact of prolonged DNAemia duration and point to the importance of optimized antiviral stewardship. Given these complex implications, our findings encourage a careful consideration of using ultrasensitive CMV qNAT assays when designing institutional protocols for the treatment of CMV DNAemia. Future research in diverse organ transplant populations will be crucial in further refining the role of these platforms in CMV surveillance.
Supplementary Material
Acknowledgements:
This work was supported by:
[TL1TR002382, UL1TR002378] to VB, [K23AI144036] to MW
“Supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Numbers TL1TR002382 and UL1TR002378, and the National Institute of Allergy And Infectious Diseases under Award Number K23AI144036. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.”
Footnotes
Author statements:
All the authors have seen and approved the submission of this manuscript in its entirety. All the authors contributed significantly to the preparation of this manuscript. This manuscript is not being considered for publication elsewhere.
Conflict of Interest:
The authors do not have associations that might pose a conflict of interest.
De-identified data available upon request.
References
- 1.Razonable RR, Humar A. Cytomegalovirus in solid organ transplant recipients-Guidelines of the American Society of Transplantation Infectious Diseases Community of Practice. Clin Transplant 2019;33(9):e13512. DOI: 10.1111/ctr.13512. [DOI] [PubMed] [Google Scholar]
- 2.Meesing A, Razonable RR. New Developments in the Management of Cytomegalovirus Infection After Transplantation. Drugs 2018;78(11):1085–1103. DOI: 10.1007/s40265-018-0943-1. [DOI] [PubMed] [Google Scholar]
- 3.George MJ, Snydman DR, Werner BG, et al. The independent role of cytomegalovirus as a risk factor for invasive fungal disease in orthotopic liver transplant recipients. Boston Center for Liver Transplantation CMVIG-Study Group. Cytogam, MedImmune, Inc. Gaithersburg, Maryland. Am J Med 1997;103(2):106–13. DOI: 10.1016/s0002-9343(97)80021-6. [DOI] [PubMed] [Google Scholar]
- 4.Munoz-Price LS, Slifkin M, Ruthazer R, et al. The clinical impact of ganciclovir prophylaxis on the occurrence of bacteremia in orthotopic liver transplant recipients. Clin Infect Dis 2004;39(9):1293–9. DOI: 10.1086/425002. [DOI] [PubMed] [Google Scholar]
- 5.Snydman DR. The case for cytomegalovirus prophylaxis in solid organ transplantation. Rev Med Virol 2006;16(5):289–95. DOI: 10.1002/rmv.514. [DOI] [PubMed] [Google Scholar]
- 6.Walker RC, Marshall WF, Strickler JG, et al. Pretransplantation assessment of the risk of lymphoproliferative disorder. Clin Infect Dis 1995;20(5):1346–53. DOI: 10.1093/clinids/20.5.1346. [DOI] [PubMed] [Google Scholar]
- 7.Dioverti MV, Lahr BD, Germer JJ, Yao JD, Gartner ML, Razonable RR. Comparison of Standardized Cytomegalovirus (CMV) Viral Load Thresholds in Whole Blood and Plasma of Solid Organ and Hematopoietic Stem Cell Transplant Recipients with CMV Infection and Disease. Open Forum Infect Dis 2017;4(3):ofx143. DOI: 10.1093/ofid/ofx143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Humar A, Gregson D, Caliendo AM, et al. Clinical utility of quantitative cytomegalovirus viral load determination for predicting cytomegalovirus disease in liver transplant recipients. Transplantation 1999;68(9):1305–11. DOI: 10.1097/00007890-199911150-00015. [DOI] [PubMed] [Google Scholar]
- 9.Natori Y, Alghamdi A, Tazari M, et al. Use of Viral Load as a Surrogate Marker in Clinical Studies of Cytomegalovirus in Solid Organ Transplantation: A Systematic Review and Meta-analysis. Clin Infect Dis 2018;66(4):617–631. DOI: 10.1093/cid/cix793. [DOI] [PubMed] [Google Scholar]
- 10.Razonable RR, Asberg A, Rollag H, et al. Virologic suppression measured by a cytomegalovirus (CMV) DNA test calibrated to the World Health Organization international standard is predictive of CMV disease resolution in transplant recipients. Clin Infect Dis 2013;56(11):1546–53. DOI: 10.1093/cid/cit096. [DOI] [PubMed] [Google Scholar]
- 11.Fryer JF, Heath AB, Minor PD, Collaborative Study G. A collaborative study to establish the 1st WHO International Standard for human cytomegalovirus for nucleic acid amplification technology. Biologicals 2016;44(4):242–251. DOI: 10.1016/j.biologicals.2016.04.005. [DOI] [PubMed] [Google Scholar]
- 12.Preiksaitis JK, Hayden RT, Tong Y, et al. Are We There Yet? Impact of the First International Standard for Cytomegalovirus DNA on the Harmonization of Results Reported on Plasma Samples. Clin Infect Dis 2016;63(5):583–9. DOI: 10.1093/cid/ciw370. [DOI] [PubMed] [Google Scholar]
- 13.Mathur MB, Ding P, Riddell CA, VanderWeele TJ. Web Site and R Package for Computing E-values. Epidemiology 2018;29(5):e45–e47. DOI: 10.1097/EDE.0000000000000864. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.VanderWeele TJ, Ding P. Sensitivity Analysis in Observational Research: Introducing the E-Value. Ann Intern Med 2017;167(4):268–274. DOI: 10.7326/M16-2607. [DOI] [PubMed] [Google Scholar]
- 15.Natori Y, Alghamdi A, Husain S, et al. Clinical predictors of progression and clearance of low-level CMV DNAemia in solid organ transplant recipients. Transpl Infect Dis 2020;22(1):e13207. DOI: 10.1111/tid.13207. [DOI] [PubMed] [Google Scholar]
- 16.Jorgenson MR, Descourouez JL, Kleiboeker H, et al. Cytomegalovirus antiviral stewardship in solid organ transplant recipients: A new gold standard. Transpl Infect Dis 2022;24(5):e13864. DOI: 10.1111/tid.13864. [DOI] [PubMed] [Google Scholar]
- 17.Jorgenson MR, Descourouez JL, Schulz LT, Saddler CM, Smith JA. CMV antiviral stewardship: navigating obstacles to facilitate target attainment. Curr Opin Organ Transplant 2023;28(1):8–14. DOI: 10.1097/MOT.0000000000001032. [DOI] [PubMed] [Google Scholar]
- 18.Singh N, Winston DJ, Razonable RR, et al. Effect of Preemptive Therapy vs Antiviral Prophylaxis on Cytomegalovirus Disease in Seronegative Liver Transplant Recipients With Seropositive Donors: A Randomized Clinical Trial. JAMA 2020;323(14):1378–1387. DOI: 10.1001/jama.2020.3138. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.