ABSTRACT
Donor derived cell‐free DNA (dd‐cfDNA) has become a useful biomarker for monitoring of kidney transplant rejection, however utilization after rejection is not well established. We conducted a retrospective cohort study of 64 kidney transplant recipients (KTRs) with biopsy‐proven rejection and concomitant dd‐cfDNA monitoring. The primary outcome was to examine the relationship between changes in dd‐cfDNA and allograft function after rejection. Secondary outcomes included the effect of a ≥61% dd‐cfDNA reduction on eGFR and proteinuria, and impact of persistently elevated dd‐cfDNA. The percent change in dd‐cfDNA from rejection to 6 months post‐rejection and 1‐year post‐rejection was negatively correlated with eGFR change at 1, 2, and 3 years. No correlation was seen with dd‐cfDNA values at 1 month. KTRs with at least 61% decrease in dd‐cfDNA from rejection to 1‐year post‐rejection showed improvement in eGFR at 1, 2, and 3 years. KTRs with persistently elevated dd‐cfDNA showed significantly worse change in eGFR to all time points compared to those without. Incidence of proteinuria at 1, 2, and 3 years trended lower in KTRs with ≥61% decrease in dd‐cfDNA. dd‐cfDNA was negatively correlated with eGFR, primarily at 6–12 months post‐rejection. In conclusion, persistently elevated dd‐cfDNA correlates with worse longitudinal changes in eGFR through 3 years post‐rejection.
Keywords: biomarker, biopsy, donor‐derived cell‐free DNA, graft rejection, kidney transplantation, rejection
1. Introduction
Kidney transplant rejection remains a major threat to long‐term graft and patient survival, despite advances in modern immunosuppression that have reduced its overall incidence [1, 2]. Early diagnosis is critical, however the recommended clinical markers, serum creatinine and urine protein, lack sensitivity and specificity and may lag behind earlier molecular or histological changes [3, 4, 5].
The advent of donor‐derived cell‐free DNA (dd‐cfDNA) has allowed for non‐invasive monitoring of rejection in solid organ transplant recipients. In the setting of allograft injury, unique donor DNA is released from allograft cells undergoing apoptosis and can be detected and quantified in the recipient bloodstream. Thresholds have been established as predictors of rejection with excellent sensitivity and specificity.
Prior studies have demonstrated that dd‐cfDNA values tend to decrease with treatment [6, 7, 8, 9, 10] although dd‐cfDNA response is heterogeneous [11, 12, 13, 14, 15]. Most existing studies are limited by short follow‐up durations of several months or up to one year following rejection. Consequently, there remains a gap in understanding the utility of long‐term longitudinal monitoring of dd‐cfDNA after rejection and its association with allograft outcomes. Therefore, this study aimed to characterize dd‐cfDNA trajectories following treatment of biopsy‐proven kidney transplant rejection and to assess their association with short‐ and long‐term allograft function for up to 3 years of follow‐up. In addition, we evaluated the impact of persistently elevated dd‐cfDNA levels on kidney function.
2. Materials and Methods
2.1. Patient Population
We conducted a single‐center, retrospective cohort study of kidney transplant recipients at the University of Kentucky Healthcare. Eligible participants were adults (≥18 years old) who underwent kidney transplantation between January 1, 2015, and May 31, 2023, with biopsy‐proven rejection. Patients must also have had concomitant dd‐cfDNA monitoring performed at time of rejection and at minimum one‐year post‐rejection. Exclusion criteria included borderline rejection, recipients of multi‐organ transplants, individuals with a history of non‐renal organ transplantation, and those with incomplete data. Participants were identified via our center's electronic medical record. This study was approved by the University of Kentucky Institutional Review Board (IRB# 97074).
2.2. Data Collection
Data were obtained through manual chart review. Collected data included recipient and donor demographics, rejection details, dd‐cfDNA measurements and specified clinical outcomes. The most recent dd‐cfDNA value obtained at the time of biopsy and prior to initiation of rejection treatment was designated as the rejection dd‐cfDNA value. Subsequent dd‐cfDNA levels were collected at approximately 30 days, 60 days, 90 days, 180 days, 1 year, 2 years, and 3 years following rejection. Due to the retrospective design, specific time windows were defined for each time point: day 30 (range: 14–44 days), day 60 (45–74 days), day 90 (75–135 days), day 180 (136–270 days), year 1 (271–639 days), year 2 (640–1004 days), and year 3 (1005–1365 days). If multiple dd‐cfDNA values were available within a given window, the value closest to the target time point was used.
Renal function was assessed using the estimated glomerular filtration rate (eGFR), calculated with the CKD‐EPI 2021 equation. The eGFR at the time of biopsy was considered the rejection value. Follow‐up eGFR values were collected at 1‐, 2‐, and 3‐years post‐rejection. Proteinuria was defined as the presence of ≥30 mg/dL of protein on a standard urinalysis. Data on donor specific antibodies (DSA) present on Luminex single antigen testing performed at the time of rejection were collected and relative intensity scores (RIS) were calculated based on mean fluorescence intensity (MFI) [16]. Repeat biopsies after rejection treatment were conducted on a patient‐specific basis and data were collected where available. Data were collected through 3 years post‐rejection, or until patient death or graft loss.
2.3. Protocols
All patients received induction therapy with either anti‐thymocyte globulin or basiliximab, based on their immunologic risk. Maintenance immunosuppression consisted of triple immunosuppression with calcineurin inhibitor, mycophenolate, and prednisone. Rejection was categorized according to Banff criteria [17]. Specific details on treatment of rejection are available in supplemental material. In brief, T cell‐mediated rejection (TCMR) was treated with intravenous (IV) methylprednisolone, with the addition of rabbit anti‐thymocyte globulin for moderate‐severe TCMR. Antibody‐mediated rejection (AMR) was generally treated with IV methylprednisolone, plasmapheresis, and intravenous immunoglobulin (IVIG). In cases of inadequate response to initial therapy or markedly elevated DSA levels, bortezomib was added. Management of chronic rejection often mirrored treatment for acute rejection, with consideration given to adding belatacept to the maintenance immunosuppression regimen on a case‐by‐case basis.
Beginning in 2019, dd‐cfDNA monitoring was protocolized for kidney transplant recipients. In these patients, dd‐cfDNA was obtained every three months during the first two years post‐transplant. Additionally, dd‐cfDNA was collected on a for‐cause basis in the setting of allograft dysfunction, detection of new DSA, suspected or confirmed medication nonadherence, or other clinical concerns as determined by the treating provider. dd‐cfDNA testing was completed with AlloSure (CareDx).
2.4. Outcomes
The primary outcome was to evaluate the relationship between changes in dd‐cfDNA and eGFR following kidney allograft rejection. Percent changes in dd‐cfDNA from rejection to post‐rejection time points (30 days, 60 days, 90 days, 180 days, 1 year) were calculated. These values were then correlated with percent changes in eGFR from time of rejection to 1‐, 2‐, and 3‐years post‐rejection. Correlations were assessed for both the entire cohort, as well as a subgroup of patients with dd‐cfDNA ≥0.5% at time of rejection. Both analyses were conducted due to the anticipated impact of outliers and implication to currently known information regarding dd‐cfDNA. The correlation analysis included all rejection types. Exploratory subgroup analyses for acute TCMR, acute AMR and acute mixed rejection were performed.
Secondary outcomes included impact of persistently elevated dd‐cfDNA and assessing the effect of a threshold change in dd‐cfDNA on eGFR and proteinuria. Persistently elevated dd‐cfDNA was defined as rejection dd‐cfDNA ≥0.5% and dd‐cfDNA ≥1% at 1‐year post‐rejection. The impact of achieving a ≥61% decrease in dd‐cfDNA on eGFR was assessed at 1, 3, 6, and 12 months. The threshold of 61% was chosen based on prior data indicating an increase of 61% or more exceeds biological variability [18]. Analyses were conducted at discrete time points rather than as longitudinal trends. Given that dd‐cfDNA may rise after initial decline without recurrent rejection, we evaluated whether sustained reduction correlated with improved kidney function. Specifically, we compared percent change in eGFR between patients with a ≥61% decrease in dd‐cfDNA at 6 months that was maintained at 1 year, and those with an initial ≥61% decrease that was not sustained. Finally, we assessed characteristics that may be associated with sustained reduction in dd‐cfDNA.
2.5. Statistical Analysis
Descriptive statistics were presented as mean with standard deviation for normally distributed data, and as median with interquartile range (IQR) for non‐normally distributed data. The primary outcome, correlation between change in dd‐cfDNA and eGFR, was assessed using Spearman's rank correlation due to non‐normal distribution. Student's t‐test and Mann–Whitney U test were used for parametric and nonparametric continuous data analysis, respectively. Categorical data were evaluated using the chi‐square test or Fisher's exact test. A significant value of p‐value <0.05 was used for all statistical tests. IBM SPSS Statistics Version 28 was utilized for all analyses.
3. Results
3.1. Patient Characteristics
A total of 64 patients met inclusion criteria. Of these, 41 (64.1%) were Caucasian and the mean age at the time of transplant was 44 years. The most common cause of end‐stage renal disease was diabetes (28.1%). The majority of patients (95.3%) received induction with anti‐thymocyte globulin. Nine patients had experienced at least one prior rejection episode before the episode analyzed in this study. The median time to rejection was 422.5 days. Median eGFR at time of rejection was 48.89 mL/min/1.73m2. Additional baseline demographics are summarized in Table 1.
TABLE 1.
Demographic and rejection characteristics.
|
All N = 64 |
Rejection dd‐cfDNA ≥0.5 c N = 51 |
|
|---|---|---|
| Age at transplant (years), mean ± SD | 44.2 ± 14.7 | 44.3 ± 14.1 |
| Male sex, n (%) | 36 (56.3) | 28 (54.9) |
| Race, n (%) | ||
| Caucasian | 41 (64.1) | 36 (70.6) |
| Black or African American | 18 (28.1) | 11 (21.6) |
| Other | 5 (7.8) | 4 (7.8) |
| Hispanic or latino/a, n (%) | 4 (6.3) | 2 (3.9) |
| Indication for transplant, n (%) | ||
| Diabetes mellitus | 18 (28.1) | 10 (19.6) |
| Hypertension | 7 (10.9) | 4 (7.8) |
| Glomerulonephritis | 14 (21.9) | 12 (23.5) |
| Polycystic kidney disease | 8 (12.5) | 8 (15.7) |
| Congenital | 7 (10.9) | 7 (13.7) |
| Previous renal allograft failure | 5 (7.8) | 5 (9.8) |
| Unknown/Other | 5 (7.8) | 5 (9.8) |
| Thymoglobulin induction, n (%) | 61 (95.3) | 48 (94.1) |
| cPRA, median [IQR] | 2.5 [0, 37.3] | 3.0 [0, 50] |
| Delayed graft function (DGF), a n (%) | 9 (14.1) | 6 (11.8) |
| Donor type, n (%) | ||
| Deceased donor | 43 (67.2) | 33 (64.7) |
| Living donor | 21 (32.8) | 18 (35.3) |
| Time to rejection (days), median [IQR] | 423 [199, 894] | 464 [215, 957] |
| Previous rejection on biopsy, n (%) | 9 (14.1) | 7 (13.7) |
| DSA at time of rejection, n (%) | 34 (53.1) | 29 (56.9) |
| eGFR at time of rejection (mL/min/1.73m2), median [IQR] | 42.01 [31.06, 57.28] | 48.89 [32.05, 58.84] |
| Banff pathology on biopsy, b n | ||
| Acute TCMR IA | 12 | 11 |
| Acute TCMR IB | 14 | 9 |
| Acute TCMR IIA | 9 | 9 |
| Acute TCMR IIB | 2 | 2 |
| Acute TCMR III | 1 | 0 |
| Active AMR | 21 | 20 |
| Chronic active TCMR | 9 | 5 |
| Chronic active AMR | 13 | 13 |
Abbreviations: AMR, antibody mediated rejection; cPRA, calculated panel reactive antibodies; DSA, donor specific antibody; IQR interquartile range; SD, standard deviation; TCMR, T cell‐mediated rejection.
DGF defined as requirement of dialysis within 7 days of kidney transplantation.
Patients may have more than one type of rejection on biopsy (e.g., mixed).
Subset of patients included in correlation analyses.
The types of rejections observed in the study are listed in Table 1. A total of 23 (35.9%) were observed to only have acute TCMR. The next most common finding in 16 (25.0%) of the rejections was mixed rejection: they had a component of acute or chronic TCMR and a component of acute or chronic AMR. Details regarding treatments administered for rejection are provided in the supplemental material.
3.2. Longitudinal Trend of dd‐cfDNA Values After Rejection
The median interval between donor‐derived cell‐free DNA (dd‐cfDNA) measurement and biopsy was 11.5 days [IQR 5, 18.5]. Descriptive statistics of average dd‐cfDNA values at time of rejection and all follow up points post‐rejection are presented in Table 2. The median dd‐cfDNA at time of rejection was 2.00%, though notable variation was observed across rejection subtypes. Patients with AMR, either in isolation or as part of a mixed rejection, demonstrated higher dd‐cfDNA levels at the time of rejection compared to those with TCMR alone. In contrast, patients with chronic active TCMR exhibited lower dd‐cfDNA levels at rejection than other groups.
TABLE 2.
Descriptive summary of dd‐cfDNA (%) over time post‐rejection.
| dd‐cfDNA at rejection | dd‐cfDNA +30 days | dd‐cfDNA +60 days | dd‐cfDNA +90 days | dd‐cfDNA +180 days | dd‐cfDNA +1 year | dd‐cfDNA +2 years | dd‐cfDNA +3 years | |
|---|---|---|---|---|---|---|---|---|
|
All n = 66 |
2.00 [0.78, 3.90] |
0.61 [0.31, 0.97] |
0.58 [0.21, 0.95] |
0.57 [0.17, 1.30] |
0.45 [0.14, 1.70] |
0.67 [0.18, 1.70] |
0.49 [0.21, 1.85] |
0.47 [0.19, 1.52] |
|
Only acute TCMR n = 23 |
1.10 [0.37, 2.10] |
0.37 [0.28, 0.84] |
0.28 [0.17, 0.74] |
0.20 [0.12, 0.69] |
0.13 [0.12, 0.42] |
0.21 [0.12, 0.69] |
0.17 [0.12, 0.60] |
0.37 [0.09, 2.24] |
|
Only active AMR n = 7 |
2.00 [1.20, 2.90] |
0.66 [0.55, 1.43] |
0.76 [0.61, 1.10] |
1.00 [0.70, 2.30] |
1.10 [0.64, 1.80] |
1.40 [0.71, 2.30] |
NA | NA |
|
Only chronic active TCMR n = 7 |
0.63 [0.12, 0.97] |
0.31 [0.14, 0.46] |
0.17 [0.12, 0.29] |
0.20 [0.12, 0.33] |
0.19 [0.12, 0.44] |
0.19 [0.12, 1.30] |
0.27 [0.25, 0.29] |
NA |
|
Only chronic active AMR n = 11 |
3.20 [1.70, 6.80] |
0.90 [0.59, 3.00] |
1.40 [0.75, 3.95] |
1.70 [1.07, 4.25] |
2.95 [0.43, 4.58] |
1.20 [0.69, 3.40] |
1.80 [0.46, 4.68] |
1.20 [0.59, 4.35] |
|
Mixed n = 16 |
3.10 [1.18, 7.24] |
0.76 [0.28, 1.08] |
0.46 [0.30, 0.91] |
0.52 [0.22, 1.10] |
0.52 [0.35, 3.90] |
0.79 [0.40, 1.93] |
0.71 [0.41, 3.85] |
0.22 [0.16, 0.39] |
Note: All values expressed as median [interquartile range].
Abbreviations: AMR, antibody mediated rejection; NA, not available; TCMR, T cell‐mediated rejection.
A decline in dd‐cfDNA was observed as early as 30 days following rejection, however, the pattern of change varied by rejection subtype. Patients with acute TCMR maintained reduced dd‐cfDNA levels throughout the study period. Conversely, individuals with active AMR, chronic active AMR, or mixed rejection demonstrated subsequent increases in dd‐cfDNA after the initial decline, with persistent elevations over time.
3.3. Correlation Between Changes in dd‐cfDNA and eGFR After Rejection
A correlation between percent change of dd‐cfDNA from time of rejection to 1‐, 3‐, 6‐, and 12‐months post‐rejection and percent change of eGFR from time of rejection to 1‐, 2‐ and 3‐years post‐rejection were assessed. Initial analyses included the full dataset; however, results were sensitive to outliers, particularly patients with dd‐cfDNA levels <0.5% at the time of rejection. Consequently, the correlation analyses reported here were limited to the 51 patients with dd‐cfDNA ≥0.5% at the time of rejection.
No significant correlation was found between the percent change in dd‐cfDNA at 30 days post‐rejection and eGFR change at 1, 2, or 3 years. At 3 months, percent change in dd‐cfDNA showed a significant negative correlation with eGFR change at 3 years (r = ‐0.524; p = 0.004), but not at 1 or 2 years. By 6 months, percent change in dd‐cfDNA was significantly correlated with eGFR change at all time points (1 year: r = −0.290; p = 0.043, 2 years: r = −0.462; p = 0.004, 3 years: r = −0.438; p = 0.020. The strongest correlations were observed with dd‐cfDNA measured at 12 months, which demonstrated significant negative correlations with eGFR change at 1 year (r = −0.328; p = 0.019), 2 years (r = −0.490; p = 0.0042) and 3 years (r = −0.651; p<0.001) (Table 3). Exploratory subgroup analyses stratified by rejection phenotype demonstrated generally similar trends between dd‐cfDNA reduction and eGFR outcomes in both acute TCMR and acute mixed rejection cohorts, although interpretation is limited by small sample sizes. Plots of percent change in dd‐cfDNA vs percent change in eGFR and rejection subgroup analyses are provided in the supplemental materials.
TABLE 3.
Correlation of percent change dd‐cfDNA and eGFR.
| % change eGFR from rejection to 1 year | % change eGFR from rejection to 2 years | % change eGFR from rejection to 3 years | ||||
|---|---|---|---|---|---|---|
| Correlation Coefficient a | p‐value | Correlation Coefficient a | p‐value | Correlation Coefficient a | p‐value | |
| % change dd‐cfDNA 1 month | 0.075 | 0.671 | 0.005 | 0.981 | −0.090 | 0.699 |
| % change dd‐cfDNA 3 months | −0.184 | 0.212 | −0.304 | 0.067 | −0.524 | 0.004 |
| % change dd‐cfDNA 6 months | −0.290 | 0.043 | −0.462 | 0.004 | −0.438 | 0.020 |
| % change dd‐cfDNA 12 months | −0.328 | 0.019 | −0.490 | 0.002 | −0.651 | <.001 |
Abbreviation: eGFR, estimated glomerular filtration rate.
Spearman's rank test.
3.4. Persistently Elevated dd‐cfDNA
Kidney transplant recipients with persistently elevated dd‐cfDNA, defined as a rejection dd‐cfDNA of ≥0.5% and levels remaining ≥1% one year after rejection, had worse allograft function, with steady declines in eGFR at 1, 2, and 3 years post‐rejection. In contrast, patients without persistently elevated dd‐cfDNA showed improvements in eGFR at all time points (1 year: −4.63% vs. +11.63%, p = 0.031; 2 years: −12.15% vs. +16.10%, p = 0.001; 3 years: −47.53% vs + 30.29%, p<0.001).
Eleven patients with persistently elevated dd‐cfDNA underwent subsequent biopsy after rejection treatment at a median of 128 days from first biopsy. All eleven patients were found to have rejection on biopsy (1 acute AMR, 3 chronic active AMR, 7 mixed). Of the fourteen patients without persistently elevated dd‐cfDNA and repeat biopsy (median 91 days from first biopsy), eleven patients had rejection on biopsy (4 acute TCMR, 2 acute AMR, 4 mixed).
3.5. Evaluation of a Threshold Change in dd‐cfDNA on Allograft Outcomes
A threshold of a 61% decrease in dd‐cfDNA was evaluated for a relationship with subsequent changes in eGFR (Table 4). A 61% reduction in dd‐cfDNA rejection at 1‐month post‐rejection was not associated with change in eGFR at 1, 2, or 3 years. At 3 months post‐rejection, a 61% dd‐cfDNA reduction appeared to be linked to improved eGFR, though this difference did not reach statistical significance. Patients who achieved a 61% decrease in dd‐cfDNA at 6 months post‐rejection showed significantly improved eGFR percent change at 1 year (+11.41% vs. +3.24%; p = 0.050) and 2 years (+19.06% vs. −3.86%; p = 0.034), but not at 3 years (+27.98% vs. −13.07%; p = 0.129). Finally, a 61% decrease in dd‐cfDNA at 1 year post‐rejection was associated with significantly improved eGFR at all follow‐up points: 1 year (+12.77% vs. +3.24%; p = 0.043), 2 years (+24.77% vs. −7.86%; p = 0.004), and 3 years (+30.76% vs. −46.49%; p<0.001).
TABLE 4.
Impact of 61% threshold reduction in dd‐cfDNA on eGFR and proteinuria at 1‐, 2‐ and 3‐years after rejection .
| Achieved 61% decrease dd‐cfDNA by time point | Median % change eGFR 1 yr [IQR] | p‐value | Median % change eGFR 2 yrs [IQR] | p‐value | Median % change eGFR 3 yrs [IQR] | p‐value | Protein‐uria 1 yr, n (%) | p‐value | Protein‐uria 2 yrs, n (%) | p‐value | Protein‐uria 3 yrs, n (%) | p‐value | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 month | Yes |
8.07 [−5.86, 21.36] |
0.801 |
10.19 [−14.75, 47.89] |
0.860 |
11.49 [−37.42, 37.79] |
0.972 | 9 (45.0) | 0.172 | 6 (33.3) | 1.000 | 5 (38.5) | 0.387 |
| No |
7.73 [−20.29, 34.40] |
14.99 [−15.03, 34.01] |
9.25 [−27.69, 70.74] |
9 (69.2) | 3 (33.3) | 5 (62.5) | |||||||
| 3 months | Yes |
11.18 [−5.02, 30.14] |
0.327 |
13.03 [−16.23, 50.91] |
0.353 |
29.82 [−20.30, 80.10] |
0.188 | 12 (44.4) | 0.584 | 5 (20.8) | 0.124 | 7 (36.8) | 0.103 |
| No |
6.13 [−9.72, 19.57] |
1.70 [−18.37, 15.73] |
−13.07 [−42.98, 24.56] |
10 (52.6) | 6 (50.0) | 7 (77.8) | |||||||
| 6 months | Yes |
11.41 [−3.92, 40.60] |
0.050 |
19.06 [−11.18, 59.29] |
0.034 |
27.98 [−20.30, 80.10] |
0.129 | 8 (27.6) | <0.001 | 4 (17.4) | 0.056 | 7 (36.8) | 0.103 |
| No |
3.24 [−22.38, 11.19] |
−3.86 [−39.10, 13.67] |
−13.07 [−59.36, 26.15] |
16 (84.2) | 7 (53.8) | 7 (77.8) | |||||||
| 12 months | Yes |
12.77 [−0.62, 46.97] |
0.043 |
24.77 [−6.36, 68.49] |
0.004 |
30.76 [12.48, 93.13] |
<0.001 | 11 (60.9) | 0.195 | 4 (20.0) | 0.080 | 4 (23.5) | <0.001 |
| No |
3.24 [−9.01, 11.20] |
−7.86 [−40.51, 13.67] |
−46.49 [−77.26, ‐4.41] |
14 (42.3) | 8 (47.1) | 10 (90.9) | |||||||
Abbreviations: eGFR, estimated glomerular filtration rate; IQR, interquartile range.
Subsequent biopsy results were reviewed in the context of achieving a 61% dd‐cfDNA reduction. Among patients who did not achieve a 61% reduction and underwent repeat biopsy (median 145 days from first biopsy), all demonstrated rejection. In contrast, among patients who achieved a 61% reduction in dd‐cfDNA (median 97 days from first biopsy), no rejection was identified on repeat biopsy in 2 patients (13.3%) at 1 month; this increased to 3 patients (21.4%) at 3 months and remained stable through 6 months and 1 year.
Incidence of proteinuria at 1, 2 and 3 years trended lower in kidney transplant recipients with ≥61% decrease in dd‐cfDNA (Table 4). A significant reduction in proteinuria at 1 year was seen in those who achieved a 61% dd‐cfDNA decrease at 6 months (27.6% vs. 84.2%; p<0.001). A significant difference was also observed at 3 years of follow‐up for those who met threshold at 12 months (23.5% vs. 90.9%; p<0.001).
A sustained reduction in dd‐cfDNA (≥61% decrease in dd‐cfDNA at 6 months and maintained at 1 year) was associated with improved kidney function. Compared to those with ≥61% decrease in dd‐cfDNA at 6 months but not at 12 months, those with a sustained reduction showed improved median eGFR percent change at 1 year (+18.26% vs. −3.06%; p = 0.038), 2 years (+24.77% vs. −14.01%; p = 0.005) and 3 years following rejection (+30.76% vs. −54.40%; p<0.001). No patient or rejection characteristics were associated with sustained reduction in dd‐cfDNA. There were numerically fewer patients with sustained dd‐cfDNA who experience DGF after their transplant (76.7% no DGF vs. 25.0% DGF; p = 0.067) as well as fewer patients who had prior rejection episodes (95.8% no rejection vs. 4.2% rejection; p = 0.067). Neither presence of DSA at time of rejection, persistence of DSA or change in DSA was associated with sustained changes in dd‐cfDNA.
3.6. Graft Loss and Patient Death
Eight patients experienced death‐censored graft loss within 3 years following rejection, with all cases attributed to rejection. Of these, four were initially diagnosed with TCMR and four with mixed rejection. In addition, concomitant pyelonephritis was noted in two patients and recurrent IgA nephropathy in one patient. Most of these patients underwent subsequent biopsies that demonstrated further allograft injury. The median time from rejection to graft loss was 703 days (range: 135–1023 days). At the time of rejection, the median dd‐cfDNA level was 2.8% (range: 0.12%–13.1%) amongst those with graft loss. There were no notable trends with regards to dd‐cfDNA amongst those who experienced graft loss. Additionally, two patients died within 3 years of rejection. One death was attributed to infection, and the other to intracranial hemorrhage.
4. Discussion
In this study of kidney transplant recipients with biopsy‐proven rejection, a reduction in dd‐cfDNA was associated with improved eGFR through 3 years of follow‐up. The present study expands upon prior evidence suggesting that serial monitoring of dd‐cfDNA may be valuable not only for detecting rejection but also for assessing graft function following a rejection episode. Notably, persistently elevated dd‐cfDNA levels were linked to poorer allograft outcomes, while sustained decreases correlated with improved outcomes. Importantly, short‐term fluctuations in dd‐cfDNA should be interpreted cautiously within the context of overall clinical and laboratory assessment, and should not prompt premature changes in therapy.
Longitudinal trends in dd‐cfDNA after rejection have been previously described, but prior studies have only partially captured the patterns observed in our analysis. Most existing data are limited to 90 days after rejection [7, 9, 10, 11, 12, 13, 15, 19, 20], with only a few at this time extending to one year [14, 20]. Consistent with previous findings, we observed a rapid decline in dd‐cfDNA among patients with TCMR who received treatment, with levels remaining within the normal range thereafter [9, 11, 12, 13, 14]. In contrast, patients treated for AMR—whether isolated or mixed—initially showed a decline in dd‐cfDNA, followed by a rebound as early as 3 to 6 months after diagnosis. Although some studies have captured this rebound phenomenon [9, 13], most others have not [11, 12, 14]. CA‐TCMR also posed unique trends with dd‐cfDNA levels that were often low at the time of rejection, consistent with findings reported by Aubert et al. [6].
Our study identified a negative correlation between changes in dd‐cfDNA and eGFR following rejection. Although statistically significant correlations were observed, the proportion of variability in long‐term eGFR explained by dd‐cfDNA changes was modest, consistent with the multifactorial nature of allograft outcomes. The percent reduction in dd‐cfDNA at 6‐ and 12‐months post‐rejection correlated with improved eGFR at all time points, with the strongest correlation seen at 12 months. In contrast, reductions at 1‐ and 3‐ months showed no significant correlation. These findings suggest that short term elevations in dd‐cfDNA may not be significantly detrimental (compared to long term elevations) and sustained reduction in dd‐cfDNA may be a more reliable marker associated with long‐term renal function. Conversely, persistent elevations in dd‐cfDNA—potentially reflecting ongoing allograft injury—may be associated with worse allograft outcomes. Prior studies have similarly identified persistently elevated or rebound dd‐cfDNA trajectories to be linked to poorer outcomes, including death‐censored graft loss [9]. Exploratory subgroup analyses demonstrated generally similar directional trends among TCMR and mixed rejection cohorts, although interpretation was limited by small subgroup sample sizes, particularly for isolated AMR.
As noted with the correlation results, achieving and maintaining reduced dd‐cfDNA appears to be associated with better outcomes. However, some populations, such as AMR or mixed rejection, may initially exhibit a decline in dd‐cfDNA, followed by a rebound, with levels remaining elevated (e.g., ≥1%) long‐term. The clinical significance of this persistent elevation remains unclear, and it has been questioned whether it indicates ongoing allograft injury [9, 13, 14]. Currently, no standardized guidelines exist for modality of assessing dd‐cfDNA changes after rejection; however, several authors have proposed that percent change may be more appropriate than individual results [12, 13]. We assessed persistently elevated dd‐cfDNA using both absolute values (≥1%) and sustained reductions of at least 61% as secondary outcomes. In both analyses, sustained dd‐cfDNA reduction was associated with improved eGFR, suggesting that long‐term decreased in dd‐cfDNA are more clinically meaningful than transient reductions immediately following treatment. Although follow‐up biopsies were not routinely performed in this study, our findings are consistent with previous research that included such assessments. Studies that incorporated repeat biopsy assessment have demonstrated that elevated dd‐cfDNA levels following rejection treatment are associated with greater likelihood of persistent rejection [6, 7, 10]. More recent data suggest that dd‐cfDNA trends after rejection treatment may provide additive discrimination beyond eGFR trends alone for identifying persistent rejection [10].
The present study highlights the potential application for dd‐cfDNA monitoring after rejection, while also identifying areas for future research. First, this study further suggests a possible role for dd‐cfDNA monitoring following rejection as an adjunctive tool for assessing changes at a molecular level. Serial dd‐cfDNA measurements may help detect both early treatment response and sustained remission. Second, we identified scenarios in which dd‐cfDNA response may be misleading: early initial decline after AMR treatment and impending graft failure. In AMR, early dd‐cfDNA values alone may be an insufficient parameter to guide decisions about tailoring rejection treatment or adjusting immunosuppression. In cases of graft failure, dd‐cfDNA levels may appear “normal,” likely reflecting a reduction in viable tissue mass and, consequently, diminished DNA release. Finally, persistently elevated DNA levels are likely a strong indicator of ongoing tissue injury. We were unable to associate patient‐specific factors with persistently elevated levels, however, only 34 patients were eligible for inclusion in this sub‐analysis. Future studies should explore whether such elevations confer a high positive predictive value for persistent rejection or allograft injury. In contrast, a decline in DNA levels—sometimes observed even in cases of graft loss—may only offer a moderate negative predictive value.
Several limitations should be considered when interpreting the findings of this study. First, it was a retrospective analysis with a small sample size, limiting the ability to infer causality and reducing the capabilities of statistical analysis. Second, our center does not always perform biopsies to confirm treatment efficacy; therefore, we cannot definitively attribute changes in dd‐cfDNA to histological resolution of rejection. The observed dd‐cfDNA associations in this study reflect longitudinal changes in renal function (eGFR) rather than direct assessment of histological resolution. However, our findings align with prior studies that include biopsy confirmation, lending credibility to our findings. Third, our center does not follow a protocolized dd‐cfDNA monitoring after rejection, so outcome assessment at fixed time points may overlook data outside those windows. Finally, our cohort includes heterogeneity in allograft injury (e.g., some biopsies with BK nephropathy, pyelonephritis in addition to rejection) as well as patients with dd‐cfDNA <0.5% at time of rejection. While this variability complicates interpretation, we believe it also reflects the complexity and nuance of real‐world clinical practice. Despite these limitations, we believe our data expands upon previous knowledge, with the addition of knowledge through 3 years post rejection which has not been previously studied. We believe the data on establishing cutoffs or evaluation of persistently elevated dd‐cfDNA can generate further hypothesis and research. The ability of dd‐cfDNA to monitor efficacy of AMR treatment is a significant opportunity for further research, which the present study supports further evaluation.
5. Conclusion
In conclusion, a decrease in dd‐cfDNA after rejection was correlated with improved eGFR; however, this benefit appears to depend on a sustained decline in dd‐cfDNA over 6 to 12 months. These findings suggest that serial dd‐cfDNA monitoring after rejection, in conjunction with standard clinical monitoring, may provide additional insight into the allograft response to rejection treatment. Elevated dd‐cfDNA levels likely reflect continued allograft damage. Future studies should further investigate how different types of rejection may present with different dd‐cfDNA trends, as well as further characterize outcomes in those with persistently elevated dd‐cfDNA.
Author Contributions
Lauren Schumacher, Olivia Philippart, Jesica Yau, and Hasan Fattah contributed to the conception and design of the paper. Lauren Schumacher, Olivia Philippart, Fawad Shuaib, Sravanthi Paluri, Ana Lia Castellanos, and Hasan Fattah contributed to data collection. Lauren Schumacher and Aric Schadler analyzed and interpreted the data. Lauren Schumacher wrote the original manuscript draft. Olivia Philippart, Jesica Yau, Hasan Fattah, and Lauren Schumacher contributed to manuscript review and editing. Hasan Fattah provided supervision. All authors have reviewed and agreed to the final version of the manuscript.
Funding
The project described was supported by the NIH National Center for Advancing Translational Sciences through grant number UL1TR001998. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Conflicts of Interest
All other authors have no conflicts of interest to disclose.
Schumacher L., Philippart O., Yau J., et al. “Reduction in Donor‐Derived Cell‐Free DNA After Rejection Treatment is Associated With Improved Long Term Allograft Function.” Clinical Transplantation 40, no. 6 (2026): e70603. 10.1111/ctr.70603
Reduction in donor‐derived cell free DNA is associated with long‐term allograft function after kidney transplant rejection
References
- 1. HU M.‐K., Schold J. D., Srinivas T. R., and Kaplan B., “Lack of Improvement in Renal Allograft Survival Despite a Marked Decrease in Acute Rejection Rates Over the Most Recent Era,” American Journal of Transplantation 4, no. 3 (2004): 378–383, 10.1111/j.1600-6143.2004.00332.x. [DOI] [PubMed] [Google Scholar]
- 2. HU M.‐K., Schold J. D., and Kaplan B., “Long‐Term Renal Allograft Survival: Have we Made Significant Progress or is it Time to Rethink Our Analytic and Therapeutic Strategies?,” American Journal of Transplantation 4, no. 8 (2004): 1289–1295, 10.1111/j.1600-6143.2004.00515.x. [DOI] [PubMed] [Google Scholar]
- 3. Bloom R. D. and Augustine J. J., “Beyond the Biopsy: Monitoring Immune Status in Kidney Recipients,” Clinical Journal of the American Society of Nephrology 16, no. 9 (2021): 1413–1422, 10.2215/cjn.14840920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Kidney Disease: Improving Global Outcomes (KDIGO) Transplant Work Group , “KDIGO Clinical Practice Guideline for the Care of Kidney Transplant Recipients,” American Journal of Transplantation 9, no. S3 (2009): S1–155, 10.1111/j.1600-6143.2009.02834.x. [DOI] [PubMed] [Google Scholar]
- 5. Josephson M. A., “Monitoring and Managing Graft Health in the Kidney Transplant Recipient,” Clinical Journal of the American Society of Nephrology 6, no. 7 (2011): 1774–1780, 10.2215/cjn.01230211. [DOI] [PubMed] [Google Scholar]
- 6. Aubert O., Ursule‐Dufait C., Brousse R., et al., “Cell‐Free DNA for the Detection of Kidney Allograft Rejection,” Nature Medicine 30, no. 8 (2024): 2320–2327, 10.1038/s41591-024-03087-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Parajuli S., Garg N., Dodin B., et al., “Changes in Donor‐Derived Cell‐Free DNA Before and After Rejection and De Novo DSA Detection in Primary and Repeat Kidney Transplant Recipients,” Clinical Transplantation 38, no. 11 (2024): e70019, 10.1111/ctr.70019. [DOI] [PubMed] [Google Scholar]
- 8. Hinojosa R. J., Chaffin K., Gillespie M., and Villarreal V. H., “Donor‐Derived Cell–Free DNA May Confirm Real‐Time Response to Treatment of Acute Rejection in Renal Transplant Recipients,” Transplantation 103, no. 4 (2019): e61, 10.1097/TP.0000000000002579. [DOI] [PubMed] [Google Scholar]
- 9. Bunnapradist S., Leca N., Zaky Z. S., et al., “Associations Between Donor‐Derived Cell‐Free DNA Dynamics and Clinical Outcomes After Kidney Allograft Rejection: A Prospective, Multicenter Study,” American Journal of Transplantation 25, no. 12 (2025): 2543–2553, 10.1016/j.ajt.2025.07.2470. [DOI] [PubMed] [Google Scholar]
- 10. Ku E., Winnicki E., Legaspi S., Lin F., McCulloch C. E., and Adey D., “Donor‐Derived Cell‐Free DNA Levels and Prediction of Resolution of Rejection in Kidney Transplant Recipients,” Clinical Transplantation 40, no. 1 (2026): e70443, 10.1111/ctr.70443. [DOI] [PubMed] [Google Scholar]
- 11. Benning L., Morath C., Fink A., et al., “Donor‐Derived Cell‐Free DNA (dd‐cfDNA) in Kidney Transplant Recipients With Indication Biopsy—Results of a Prospective Single‐Center Trial,” Transplant International 36 (2023): 11899, 10.3389/ti.2023.11899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wolf‐Doty T. K., Mannon R. B., Poggio E. D., et al., “Dynamic Response of Donor‐Derived Cell‐Free DNA Following Treatment of Acute Rejection in Kidney Allografts,” Kidney360 2, no. 4 (2021): 729–736, 10.34067/kid.0000042021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Shen J., Guo L., Yan P., et al., “Prognostic Value of the Donor‐Derived Cell‐Free DNA Assay in Acute Renal Rejection Therapy: A Prospective Cohort Study,” Clinical Transplantation 34, no. 10 (2020): e14053, 10.1111/ctr.14053. [DOI] [PubMed] [Google Scholar]
- 14. Steggerda J. A., Pizzo H., Garrison J., et al., “Use of a Donor‐Derived Cell‐Free DNA Assay to Monitor Treatment Response in Pediatric Renal Transplant Recipients With Allograft Rejection,” Pediatric Transplantation 26, no. 4 (2022): e14258, 10.1111/petr.14258. [DOI] [PubMed] [Google Scholar]
- 15. Ranch D., Fei M., Kincade E., Piburn K., Hitchman K., and Klein K., “Utilization of Donor‐Derived Cell‐Free DNA in Pediatric Kidney Transplant Recipients: A Single Center Study,” Pediatric Transplantation 28, no. 1 (2024): e14582, 10.1111/petr.14582. [DOI] [PubMed] [Google Scholar]
- 16. Vo A. A., Sinha A., Haas M., et al., “Factors Predicting Risk for Antibody‐Mediated Rejection and Graft Loss in Highly Human Leukocyte Antigen Sensitized Patients Transplanted After Desensitization,” Transplantation 99, no. 7 (2015): 1423–1430, 10.1097/tp.0000000000000525. [DOI] [PubMed] [Google Scholar]
- 17. Loupy A., Haas M., Roufosse C., et al., “The Banff 2019 Kidney Meeting Report (I): Updates on and Clarification of Criteria for T Cell– and Antibody‐Mediated Rejection,” American Journal of Transplantation 20, no. 9 (2020): 2318–2331, 10.1111/ajt.15898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Bromberg J. S., Brennan D. C., Poggio E., et al., “Biological Variation of Donor‐Derived Cell‐Free DNA in Renal Transplant Recipients: Clinical Implications,” Journal of Applied Laboratory Medicine 2, no. 3 (2017): 309–321, 10.1373/jalm.2016.022731. [DOI] [PubMed] [Google Scholar]
- 19. Gupta G., Moinuddin I., Kamal L., et al., “Correlation of Donor‐Derived Cell‐Dree DNA With Histology and Molecular Diagnoses of Kidney Transplant Biopsies,” Transplantation 106, no. 5 (2022): 1061–1070, 10.1097/tp.0000000000003838. [DOI] [PubMed] [Google Scholar]
- 20. Akifova A., Budde K., Amann K., et al., “Donor‐Derived Cell‐Free DNA Monitoring for Early Diagnosis of Antibody‐Mediated Rejection After Kidney Transplantation: A Randomized Trial,” Nephrology Dialysis Transplantation 40, no. 7 (2024): 1384–1395, 10.1093/ndt/gfae282. [DOI] [PMC free article] [PubMed] [Google Scholar]