Characterizing Early Donor‐Derived cfDNA Kinetics in Stable Pediatric Living‐Donor Liver Transplant Recipients

ABSTRACT

Background

Early and precise assessment of graft health is vital for pediatric liver transplant patients; however, current monitoring methods depend largely on invasive biopsies and non‐specific biochemical tests. Donor‐derived cell‐free DNA (ddcfDNA) has emerged as a promising non‐invasive biomarker for detecting graft injury, but data on its early post‐transplant kinetics, baseline levels, and the influence of donor and recipient factors in stable pediatric populations remain limited.

Objective

To prospectively characterize the early postoperative patterns of ddcfDNA and explore baseline ranges in clinically stable pediatric recipients of living donor liver transplants (LDLT), while evaluating the influence of recipient and donor demographic and clinical variables on ddcfDNA dynamics for improved non‐invasive graft monitoring.

Methods

In a cohort of 22 stable pediatric LDLT recipients, ddcfDNA levels were measured longitudinally using the Trunome GrafAssure assay at postoperative days 1–2, 7, 10–14, and 30–31. Concurrent liver function tests and clinical data were also collected. Statistical analyses were performed to evaluate correlations and determine the significance of the observed relationships.

Results

ddcfDNA peaked on days 1–2 (~9.0%, 2.3 ng/mL), declined sharply by day 7 (~2.4%, 1.0 ng/mL), and stabilized by days 30–31 (~2.0%, 0.2 ng/mL). Liver enzymes decreased gradually. The absolute quantification values and the ddcfDNA percentage correlated with AST and ALT early post‐transplant, with associations diminishing over time. Recipient and donor characteristics did not significantly affect ddcfDNA levels.

Conclusion

dd‐cfDNA demonstrates rapid postoperative clearance and stable longitudinal trends in clinically stable pediatric liver transplant recipients. The baseline values and kinetic patterns characterized provide a preliminary reference framework for future comparative studies evaluating dd‐cfDNA behavior during graft dysfunction, rejection, or other adverse events in the early postoperative and surveillance periods.

Donor‐derived cell‐free DNA (ddcfDNA) in stable pediatric living donor liver transplant recipients peaks early after surgery and rapidly declines to a stable baseline by 1 month. These predictable kinetics, independent of donor or recipient characteristics, support ddcfDNA as a reliable non‐invasive biomarker for early graft monitoring and long‐term surveillance.

Abbreviations

ABOi

ABO incompatibility

ALT

alanine aminotransferase

AST

aspartate aminotransferase

BMI

body mass index

BSA

body surface area

ccfDNA

circulating cell‐free DNA

cfDNA

cell‐free DNA

CIT

cold ischemic time

ddcfDNA

donor‐derived cell‐free DNA

FDR

false discovery rate

gDNA

genomic DNA

GGT

gamma‐glutamyl transferase

GRWR

graft‐to‐recipient weight ratio

ICU

intensive care unit

IQR

interquartile range

IRI

ischemia–reperfusion injury

labPELD

laboratory pediatric end‐stage liver disease score

LDLT

living donor liver transplantation

LFT

liver function tests

LT

liver transplantation

POD

postoperative days

WIT

warm ischemic time

1. Introduction

Over the past few decades, advances in liver transplantation (LT) have substantially improved both patient survival and long‐term graft outcomes. However, the success of LT depends ultimately on the early recognition of graft dysfunction and immune‐mediated damage, especially rejection [1]. Liver function tests (LFT) are commonly employed as a screening tool to monitor graft health after LT, while invasive liver biopsy is still required for the definitive diagnosis of rejection. Although histopathological evaluation remains the gold standard for precise assessment of graft status [2, 3], it is not without limitations. The procedure carries risks such as infection, patient discomfort, and sampling error, and the cost and invasiveness of serial biopsies make them unsuitable for routine monitoring. Similarly, conventional biochemical markers of hepatocellular injury, such as aspartate aminotransferase (AST) and alanine aminotransferase (ALT), are non‐specific and often lack the sensitivity required to detect subclinical or early graft injury.

A pressing need in the field of transplantation is the development of non‐invasive approaches to detect and monitor donor organ injury as early and accurately as possible. Cell‐free DNA (cfDNA), a potential non‐invasive biomarker, has attracted significant research interest in recent years [4]. In the event of graft injury, the transplanted donor organ sheds donor‐derived cell‐free DNA (ddcfDNA) into the recipient's bloodstream, leading to elevated levels of ddcfDNA. The short half‐life of cfDNA makes it likely to provide a real‐time view of transplant health and allows early prediction of graft rejection [2], and the absolute quantification of ddcfDNA can provide a dynamic view of the transplanted organ [5]. In addition, personalized therapies can be optimized based on ddcfDNA levels to improve patient outcomes. While ddcfDNA has demonstrated utility in adult LT, its application in pediatric LT remains limited and underexplored. The longitudinal behavior of ddcfDNA, as well as its comparative performance against standard biochemical markers during the early post‐transplant period in pediatric LT, has not yet been comprehensively established [6].

Yet another significant limitation is the paucity of data in the current literature on how donor and recipient variables affect ddcfDNA levels, particularly in stable transplant recipients. While clinical trials have validated the role of ddcfDNA as a non‐invasive marker in graft monitoring and rejection diagnosis [7, 8, 9, 10, 11], challenges persist in standardizing diagnostic thresholds and monitoring protocols, especially in pediatric transplantation. Establishing an accurate baseline is critical, as deviations from this baseline may signal graft injury, including acute rejection. However, these baseline levels are known to vary across organ types and are influenced by factors such as immune activation, surgical trauma, and individual recipient characteristics. The temporal pattern of ddcfDNA clearance and its return to baseline also differ among individuals and transplant types. Despite the increasing use of ddcfDNA in clinical practice, no universally accepted thresholds or monitoring frequencies have been defined, particularly during the high‐risk early post‐transplant period [12]. Pediatric patients present additional complexities, viz., physiological differences due to smaller body mass, graft size mismatch, and a higher rate of primary viral infections, all of which may significantly affect ddcfDNA dynamics [13]. These pediatric‐specific variables underscore the limitations of applying adult‐derived thresholds and interpretation frameworks to children.

To address these gaps, our study aimed to (1) establish a post‐transplant ddcfDNA baseline in stable pediatric liver recipients (2) assess intra‐patient variability over time (3) identify clinical or demographic factors influencing ddcfDNA release and (4) describe the longitudinal kinetics of ddcfDNA post‐transplant [14] using proprietary Trunome GrafAssure assay.

2. Materials and Methods

2.1. Study Design

This prospective observational study was conducted between August 2023 and November 2024 at Dr. Rela Institute and Medical Centre, Chennai in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Institutional Ethics Committee approval was obtained (ECR/1276/Inst/TN/2019/212, dated 31.07.2023). Written informed consent was provided by legal guardians of all pediatric participants.

Pediatric liver transplant recipients aged ≤ 16 years undergoing their first transplant either from living or cadaveric donors were screened for eligibility. Patients of any gender were included, provided that their legal guardians gave written informed consent and were willing to comply with study procedures and follow‐up assessments. Clinically stable post‐transplant children with a steady decline in transaminase levels were included in the study. Participants who developed rising or fluctuating transaminase levels suggestive of possible rejection, infection, or other graft‐related issues during the study period were excluded from the study. As we do not have paired liver biopsy samples, this rigorous selection process, focusing solely on stable patients with progressive decline in transaminases, helps with unbiased assessment of ddcfDNA dynamics. The other exclusion criteria were prior liver or combined organ transplantation, grafts from monozygotic twins, active malignancy, prior or active graft failure, ongoing infection, or any significant comorbidity likely to affect study outcomes.

2.2. Sample Processing and Donor Derived cfDNA Quantification

Peripheral blood (5 mL) was collected in EDTA or Paxgene circulating cell‐free DNA (ccfDNA) tubes at four defined time points in the postoperative days (POD) 1 or 2, 7, 10–14, and 30–31 as a part of their standard care. Sample omission at any single visit was not a basis for exclusion. Tubes were maintained at room temperature and processed within 2 h of collection to minimize leukocyte lysis and genomic DNA (gDNA) contamination of plasma. Plasma was isolated via a two‐step centrifugation protocol, an initial centrifugation at 1200 × rcf for 10 min at 4°C, followed by immediate transfer of the supernatant to fresh tubes and a second high‐speed clarification at 16000 × rcf for 10 min at 4°C to fully remove residual cellular debris. Plasma aliquots were stored at −80°C until extraction. To normalize extraction efficiency across samples, a known concentration of synthetic spike‐in control DNA (at a defined input copy number) was added to each plasma aliquot prior to extraction [15, 16, 17]. The cfDNA was extracted from a fixed input volume of plasma (1 mL) using the QIAamp Circulating Nucleic Acid Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The gDNA was isolated from the remaining cellular pellet using the QIAamp DNA Blood Mini Kit (Qiagen). Concentration and purity of extracted nucleic acids were assessed using the Quantus Fluorometer (Promega, USA) and DS‐11 Spectrophotometer (DeNovix, USA). Fragment analysis for cfDNA integrity was performed with the Agilent 2100 Bioanalyzer using a high‐sensitivity DNA chip for quality assessment and exclusion of gDNA contamination. The proprietary Trunome GrafAssure assay was used for determination of percentage ddcfDNA and direct absolute quantification of ddcfDNA concentration, reported as ng/mL of plasma, using a digital PCR–based methodology [18, 19].

2.3. Liver Function Tests

Serum levels of ALT, AST, Bilirubin (Total and Direct) and Gamma‐Glutamyl Transferase (GGT) were measured using standard hospital laboratory protocols at each sampling point (Normal reference ranges were ALT, < 45 U/L; AST, < 35 U/L; Total Bilirubin, 0.2–1.2 mg/dL; Direct Bilirubin, 0–0.2 mg/dL).

2.4. Parameters Assessed

The following parameters were collected from each patient to explore their potential associations with ddcfDNA: recipient characteristics (age, sex, body mass index (BMI), body surface area (BSA) and primary liver disease etiology), donor characteristics (age, sex, BMI, and ABO compatibility), and perioperative variables (graft size, warm and cold ischemia times). These clinical, demographic, and perioperative factors were evaluated in parallel with ddcfDNA to determine their impact on post‐transplant graft function and recovery.

2.5. Statistical Analysis

Descriptive statistics were used for summarizing demographic and clinical characters of the patients. Continuous variables were expressed as mean ± standard deviation or median with interquartile range (IQR: Q25‐Q75) to capture central tendency and dispersion. To normalize distribution skewness, ddcfDNA values were log‐transformed. Clinical and demographic categorical variables (e.g., recipient sex, age classification, donor characteristics) were encoded as categorical factors. ddcfDNA levels between two PODs in the same patients were compared using the Wilcoxon signed‐rank test. Sequential pairwise comparisons were conducted for POD 1–2, POD 7, POD 10–14, and POD 30–31 to evaluate postoperative decline and stabilization. Associations between ddcfDNA and biochemical parameters were assessed using Spearman rank correlation coefficients for each timepoint independently. Group comparisons of log‐transformed ddcfDNA levels utilized non‐parametric tests: Mann–Whitney U tests for dichotomous predictors and Kruskal‐Wallis tests for variables with more than two groups. Significant Kruskal‐Wallis results prompted post hoc pairwise comparisons via Mann–Whitney U tests adjusted for multiple comparisons using the Benjamini‐Hochberg false discovery rate method. Exploratory multivariate linear regression models were performed to identify predictors of ddcfDNA values, with appropriate categorical variable encoding and adjustment for confounders. All analyses were performed separately per timepoint and for overall averages, excluding missing data on a per‐analysis basis. Statistical significance was set at p < 0.05. Analytical procedures were conducted using GraphPad Prism 10.0.

3. Results

3.1. Demographic and Clinical Characteristics

Out of 35 pediatric liver transplant recipients initially recruited, 13 were excluded due to elevated or unstable transaminase levels. After applying the exclusion criteria, 22 children who underwent living donor liver transplantation (LDLT) with partial liver grafts were included in the final analysis. No deceased donor liver transplant (DDLT) recipients were included in the final analysis, as none of the DDLT patients met the study's stability criteria. The median age of the cohort was 12 months (IQR: 7.8–22.8 months) (Table 1), with 15 males (68.2%) and 7 females (31.8%). Donors were predominantly female (77.3%), with a median age of 31.5 years (IQR: 25.8–36). Recipients had a median BMI of 15.5 kg/m² (IQR: 14.3–17.2), BSA of 0.4 m² (IQR: 0.3–0.5); donors had a median BMI of 25.7 kg/m² (IQR: 22.6–28.2) and BSA of 1.6 m² (IQR: 1.5–1.7). The leading transplantation indications were biliary dysgenesis (n = 14) and metabolic liver disease (n = 6), with one case each of chronic liver disease and liver tumor. Pretransplant severity scores included a laboratory Pediatric End‐Stage Liver Disease score (labPELD) median of 23.3 (IQR: 12.2–32.3) and Child‐Pugh median of 8 (IQR: 7–10). ABO incompatibility (ABOi) was documented in three recipients.

TABLE 1.

Baseline demographic and clinical characteristics of liver transplant recipients and donors.

Category	Level	N (%)	Median (IQR)	p value
Recipient age	Infants (28 days–2 years)	17 (77.3)	12 months (7.8–22.8)	> 0.08
	Children (2–12 years)	5 (22.7)
Recipient sex	Male	15 (68.2)	—	> 0.4
	Female	7 (31.8)
Recipient BMI classification	Underweight	9 (40.9)	15.5 kg/m2 (14.3–17.2)	> 0.1
	Normal weight (18.5–22.9)	9 (40.9)
	Overweight (≥ 23)	4 (18.2)
Recipient BSA classification	Low (< 1.5)	7 (31.8)	0.4 m2 (0.3–0.5)	0.041
	Medium (1.5–1.7)	13 (59.1)
	High (> 1.7)	2 (9.1)
Disease etiology	Biliary dysgenesis	14 (63.6)	—	> 0.3
	Metabolic liver disease	6 (27.3)
	Chronic liver disease	1 (4.5)
	Tumor	1 (4.5)
ABO compatibility	Compatible	19 (86.4)	—	> 0.2
	Incompatible	3 (13.6)
Donor sex	Male	5 (22.7)	—	> 0.3
	Female	17 (77.3)
Donor BMI classification	Underweight	3 (13.6)	25.7 kg/m2 (22.6–28.2)	> 0.3
	Normal weight (18.5–22.9)	3 (13.6)
	Overweight (≥ 23)	16 (72.8)
GRWR	0.8 to < 4	20 (90.9)	2.93% (1.88–3.51)	—
	> 4	2 (9.1)
Warm ischemic time	≤ 30 min	5 (22.7)	33 min (30.8–37.3)	> 0.25
	31–40 min	13 (59.1)
	41–50 min	4 (18.2)
Cold ischemic time	0 to < 0.42 h	1 (4.5)	86 min (49–136)	> 0.4
	0.42 to < 1 h	7 (31.8)
	1 to 3.04 h	14 (63.6)

Note: Data are presented as number (percentage) or median (interquartile range). Bold indicates statistical significance (p < 0.05).

Abbreviations: BMI, body mass index; BSA, body surface area; GRWR, graft‐to‐recipient weight ratio.

Postoperative recovery included a median Intensive care unit (ICU) stay of 5 days (IQR: 4–7.3) and a total hospital stay of 15 days (IQR: 11–19.3). The median graft weight was 240 g (IQR: 210–268) and the median graft‐to‐recipient weight ratio (GRWR) was 2.93% (IQR: 1.88–3.51). Ischemia times showed a median warm ischemic time (WIT) of 33 min (IQR: 30.8–37.3) and cold ischemic time (CIT) of 86 min (IQR: 49–136). Throughout the study period, all patients remained clinically stable without any episodes of biopsy‐proven or clinically suspected acute rejection.

3.2. Transplant Kinetics of ddcfDNA

Both the percentage and absolute concentration of ddcfDNA exhibited a characteristic decline following liver transplantation. On POD 1–2, ddcfDNA levels peaked, with a mean percentage of 9% ± 5.9% (IQR: 6.9–13) and an absolute concentration of 2.3 ± 5.9 ng/mL (IQR: 1.2–4.4). By POD 7, these values had fallen significantly to 2.4% ± 3.2% (IQR: 1.1–6.3; p < 0.001) and 1.0 ± 1.3 ng/mL (IQR: 0.3–1.9; p = 0.0076), representing a 3.75‐fold and 2.3‐fold decrease, respectively, compared with the initial peak (Figure 1; Table 2). No significant differences were observed between POD 7 and POD 10–14 for either ddcfDNA percentage (p = 0.35) or absolute concentration (p = 0.28), indicating that the major clearance of graft‐derived cfDNA occurred within the first postoperative week (Figures S1a,b and S2a,b).

FIGURE 1.

Line graph showing the donor‐derived cell‐free DNA and Liver Function Tests levels at various time points.

TABLE 2.

Temporal trends of liver function parameters and donor‐derived cell‐free DNA levels at serial postoperative time points in pediatric liver transplant recipients.

Details	T. Bilirubin (mg/dL)	D. Bilirubin (mg/dL)	AST (Units/L)	ALT (Units/L)	%ddcfDNA (%)	ddcfDNA Quantity (ng/mL)
S1	3 ± 1.6 (IQR: 1.9–4.4)	1.9 ± 1.2 (IQR: 0.95–2.7)	121 ± 230 (IQR: 62–276)	181 ± 166 (IQR: 97–242)	9 ± 5.9 (IQR: 6.9–13)	2.3 ± 5.9 (IQR: 1.2–4.4)
S2	1.3 ± 0.79 (IQR: 0.78–1.7)	0.77 ± 0.61 (IQR: 0.4–1.1)	52 ± 31 (IQR: 21–75)	84 ± 59 (IQR: 29–144)	2.4 ± 3.2 (IQR: 1.1–6.3)	1 ± 1.3 (IQR: 0.3–1.9)
S3	0.65 ± 0.48 (IQR: 0.49–1.1)	0.45 ± 0.38 (IQR: 0.2–0.9)	39 ± 16 (IQR: 33–49)	49 ± 30 (IQR: 27–74)	2.9 ± 1.7 (IQR: 1.2–4.5)	0.63 ± 0.63 (IQR: 0.39–0.76)
S4	0.38 ± 0.2 (IQR: 0.28–0.66)	0.21 ± 0.12 (IQR: 0.11–0.29)	34 ± 11 (IQR: 23–41)	32 ± 13 (IQR: 21–42)	2 ± 1.6 (IQR: 0.91–2.6)	0.16 ± 0.26 (IQR: 0.08–0.36)

Note: Data are presented as median ± standard deviation with interquartile range (IQR) in parentheses.

Abbreviations: %ddcfDNA, Percentage of Donor Derived Cell Free DNA in recipient plasma; ALT, Alanine Aminotransferase (U/L); AST, Aspartate Aminotransferase (U/L); D. Bilirubin, Direct Bilirubin (mg/dL); ddcfDNA quantity, Absolute concentration of Donor‐Derived Cell‐Free DNA (ng/mL); T. Bilirubin, Total Bilirubin (mg/dL).

This early decline and subsequent stabilization by POD 7 marked a critical kinetic transition point that can serve as a reference for post‐transplant monitoring. By POD 30–31, both metrics remained largely stable, with only a modest but statistically significant further decline in ddcfDNA levels (2.0% ± 1.6%, p = 0.027; 0.16 ± 0.26 ng/mL, p = 0.005 vs. POD 10–14), corresponding to approximately 4.5‐fold and 11.5‐fold reductions from the immediate postoperative peak. This minor late‐phase reduction likely reflects routine physiological clearance and normalization of cfDNA turnover rather than ongoing graft injury.

3.3. Transplant Kinetics of LFTs

The temporal behavior of LFTs, particularly AST and ALT, also exhibited a gradual downward trend, on POD 1–2, AST was 121 ± 230 U/L (IQR: 62–276) and ALT was 181 ± 166 U/L (IQR: 97–242). By POD 7–8, AST had decreased to 52 ± 31 U/L (IQR: 21–75; p = 0.0017) and ALT to 84 ± 59 U/L (IQR: 29–144; p = 0.0006), yet both remained substantially above reference baseline. Further decline was noted by POD 10–14, with AST at 39 ± 16 U/L (IQR: 33–49; p = 0.63 vs. POD 7–8) and ALT at 49 ± 30 U/L (IQR: 27–74; p = 0.007), signifying stabilization only after this point. By Visit 4, AST reached 34 ± 11 U/L (IQR: 23–41; p = 0.03) and ALT 32 ± 13 U/L (IQR: 21–42; p = 0.003), consistent with full biochemical recovery and absence of further hepatocellular injury.

3.4. Correlation Analysis

Spearman correlation analysis between ddcfDNA (percentage and absolute concentration) and conventional biochemical parameters revealed time‐dependent associations. At the earliest postoperative timepoint (S1), ddcfDNA% showed significant positive correlations with total bilirubin (r = 0.54, p = 0.016), direct bilirubin (r = 0.56, p = 0.012), and AST (r = 0.56, p = 0.011), while ddcfDNA quantity correlated strongly with AST (r = 0.81, p < 0.0001) and moderately with ALT (r = 0.51, p = 0.021). By S2, these associations shifted; significant correlations were observed for ddcfDNA% with AST (r = 0.65, p = 0.0027) and GGT (r = 0.51, p = 0.026), and for ddcfDNA quantity with AST (r = 0.58, p = 0.0094) and GGT (r = 0.55, p = 0.0138), while bilirubin associations were no longer significant. At S3, the only significant association was between ddcfDNA quantity and AST (r = 0.56, p = 0.0105). By S4, correlations had further diminished, with significance observed only between ddcfDNA% and ALT (r = 0.53, p = 0.017) and between ddcfDNA quantity and AST (r = 0.49, p = 0.028).

3.5. Transplant Factors vs ddcfDNA

To evaluate factors potentially influencing ddcfDNA kinetics, various clinical, demographic, and perioperative variables were assessed, including recipient and donor age, sex, BMI, BSA, disease etiology, ischemia times, graft size, and ABO compatibility. No significant differences in ddcfDNA percentage or quantity across different recipient age groups (infants (28 days–2 years), children (2 years–16 years)) at any timepoint or averaged across the study (all p > 0.08). Likewise, no significant differences in ddcfDNA levels were found between male and female recipients at any timepoint or overall (all p > 0.4). Median ddcfDNA percentage (2.85% vs. 3.76%, p = 0.99) and absolute concentration (0.62 ng/mL vs. 0.71 ng/mL, p = 0.36) did not differ significantly, and decline trajectories were comparable (sex‐by‐time interaction p = 0.33).

Recipient BMI, which primarily reflected underweight status, did not significantly impact ddcfDNA trajectories (p > 0.1). BSA, however, demonstrated a statistically significant association with overall ddcfDNA quantity (Kruskal–Wallis p = 0.041). However, post hoc pairwise comparisons adjusted for multiple testing did not reveal significant differences between BSA subgroups (Medium vs. Low adjusted p = 0.1452; Medium vs. High adjusted p = 0.1143; Low vs. High adjusted p = 0.2222), indicating a trend rather than definitive group disparities.

Donor age and sex were not associated with significant variations in ddcfDNA levels at any timepoint or in averaged data (all p values > 0.2 for age and > 0.3 for sex). Similarly, donor BMI categories showed no significant relationships with ddcfDNA levels (all p > 0.3), indicating donor anthropometry does not materially affect ddcfDNA measurements.

Due to the limited number of recipients across the two GRWR groups (Group 1 (0.8‐ < 4) n = 20; Group 2 (> 4) n = 2), no formal statistical comparison between groups was performed. However, a descriptive trend was observed wherein both groups demonstrated elevated ddcfDNA levels at the initial timepoint, which declined progressively across the subsequent post‐transplant monitoring period. This pattern of early elevation followed by temporal attenuation was consistent across both ddcfDNA% and absolute ddcfDNA concentration (ng/mL), suggesting that the post‐transplant ddcfDNA trajectory may follow a similar kinetic course irrespective of GRWR classification in this cohort (Table S1).

At the earliest post‐transplant timepoint, ddcfDNA quantity was significantly higher in ABOi recipients than in compatible ones (median 2.62 vs. 1.18 ng/mL, p = 0.0117), but this difference resolved at subsequent timepoints (p > 0.2). A transient elevation in ddcfDNA quantity was observed in the small ABO‐incompatible subgroup (n = 3) at the earliest postoperative timepoint; however, this finding should be interpreted cautiously given the limited sample size. Exploratory analysis across primary liver disease categories, including biliary dysgenesis (n = 14), metabolic liver disease (n = 6), chronic liver disease (n = 1), and liver tumor (n = 1), did not demonstrate clear differences in ddcfDNA levels; however, interpretation is limited by the small number of patients in several subgroups (all Kruskal‐Wallis p > 0.3). Neither warm nor cold ischemia times correlated with ddcfDNA levels (p > 0.25 and p > 0.4, respectively).

Multivariate linear regression models incorporating these categorical and continuous variables, with appropriate encoding and confounder adjustment, confirmed the lack of independent predictors influencing ddcfDNA levels across all assessed postoperative time points and overall averages.

4. Discussion

This study represents one of the few prospective efforts to systematically characterize the early kinetics and establish baseline thresholds of ddcfDNA in pediatric liver transplant recipients. While ddcfDNA is increasingly recognized as a sensitive, non‐invasive marker of graft injury and rejection in adults, its application in children faces challenges due to physiological differences and limited pediatric data. To address these gaps, we aimed to establish post‐transplant ddcfDNA baselines in stable pediatric recipients and describe the longitudinal trend and intra‐patient variability over time. Given the limited literature in pediatric LT, we contextualized our findings using available evidence from other solid organ transplants. While absolute ddcfDNA levels and thresholds may differ by organ type, the temporal trend of ddcfDNA decline and stabilization of levels post‐transplant appear consistent across organs [20], supporting broader applicability of our findings to pediatric transplant monitoring.

4.1. Impact of Transplant Related Factors

A comprehensive longitudinal evaluation demonstrated no statistically significant or sustained impact of any recipient or donor characteristic on ddcfDNA peak values, baseline concentrations, or their pattern of decline post‐transplant. The ddcfDNA levels were consistent across recipient age groups from infants to children indicating minimal age‐related variability in ddcfDNA pattern under stable conditions similar to [21] other study. While infancy might transiently influence early ddcfDNA due to metabolic or immune factors, overall, age is not a major confounder of baseline ddcfDNA measurements [22, 23].

Similarly, recipient sex also did not significantly affect ddcfDNA percentages or absolute quantities at any postoperative timepoint; this aligns with existing solid organ transplant literature demonstrating minimal sex‐related differences under stable conditions [24, 25]. Although one lung transplant study reported higher ddcfDNA in females, this likely reflects a small sample size and is considered an isolated observation [26].

No significant differences were observed across recipient BMI categories (p = 0.167) as well, indicating that body composition does not substantially affect biomarker release or clearance, which aligns with kidney transplant studies showing minimal BMI influence on ddcfDNA after adjusting for confounders [27], although one reported a negative correlation suggesting possible need for body size normalization in overweight patients, a finding limited by small sample sizes [22]. Recipient BSA showed an initial statistically significant association with ddcfDNA quantity; however, post hoc analyses adjusting for multiple comparisons did not confirm meaningful subgroup differences, consistent with Dandamudi et al. [13]. This suggests that any subtle effect of physiological size on circulating ddcfDNA is minimal and clinically insignificant in stable pediatric liver transplant recipients. Overall, ddcfDNA levels primarily reflect graft‐specific physiology rather than recipient anthropometry within the studied BMI and BSA ranges.

Donor demographics, including age group, sex, and BMI, did not significantly influence recipient ddcfDNA patterns, indicating that donor factors do not confound its use as a biomarker for graft injury [2, 10, 24, 28]. While Mirza et al. [25] reported higher ddcfDNA levels in recipients of male donor kidneys, this finding remains isolated and lacks broader validation, supporting ddcfDNA's robustness independent of donor demographic variables.

Neither WIT nor CIT significantly influenced ddcfDNA levels beyond the immediate post‐reperfusion peak. Specifically, patients exhibited a range of WIT durations (≤ 30 min: 22.7%; 31–40 min: 59.1%; 41–50 min: 18.2%) and CIT durations (0 to < 0.42 h: 4.5%; 0.42 to < 1 h: 31.8%; 1 to 3.04 h: 63.6%). Despite this variation, statistical analysis showed no significant differences in ddcfDNA between these ischemia time categories at any measured timepoint. This aligns with evidence that ischemia–reperfusion injury (IRI), primarily drives early ddcfDNA elevation, while the length of ischemia does not cause sustained ddcfDNA alterations during stable graft function [10, 25, 28]. Although Goh et al. and Foley et al. reported similar median WIT but differing CIT durations across liver and kidney transplants, longer ischemic times such as WIT > 30 min or CIT > 10 h are associated with poorer graft outcomes including increased graft loss risk [29]. Mechanistically, cold ischemia induces hypoperfusion and hypoxia, and warm ischemia causes cellular and inflammatory damage; however, these effects appear transient and do not translate into prolonged ddcfDNA elevation beyond the immediate post‐transplant phase [12, 30, 31].

Our cohort exclusively comprised pediatric recipients of partial liver grafts from living donors (left lateral segments) and the graft size, evaluated using GRWR, could not be formally compared between groups due to the inadequate sample size in Group 2 (n = 2). Nevertheless, trend analysis revealed a consistent pattern of initial ddcfDNA elevation at the early post‐transplant timepoint in both groups, followed by progressive decline across the monitoring period. This trajectory is consistent with the typically observed kinetics of ddcfDNA release following liver transplantation, wherein peak levels in the immediate post‐operative period reflect IRI and early graft handling, before declining progressively as the graft stabilizes, suggesting a shared post‐transplant ddcfDNA kinetic irrespective of GRWR classification in this cohort. Importantly, both groups demonstrated GRWR values well above the commonly accepted at‐risk threshold of < 0.8%, indicating adequate graft sizing in all recipients and negligible risk for small‐for‐size syndrome. It is therefore plausible that dd‐cfDNA levels stabilized early in the postoperative period. Observations in kidney and lung transplantation further support this, indicating that graft size or donor‐recipient size mismatch may cause temporary ddcfDNA fluctuations but do not contribute to sustained elevations in stable grafts [6, 28, 32].

A noteworthy exception was observed in recipients undergoing ABO‐incompatible transplantation, who exhibited elevated ddcfDNA quantities at the earliest postoperative timepoint. This transient rise likely represents heightened immunologic and ischemic stress unique to ABO mismatch scenarios, but the levels normalized promptly under standard immunosuppressive therapy, consistent with reports in heart and kidney transplant cohorts by Oren et al. [33] and Hirai et al. [34] respectively.

Multivariate analysis showed no significant association between demographic or clinical variables and ddcfDNA levels in this stable pediatric LDLT cohort. This simplifies clinical interpretation and supports the use of common pediatric reference baselines across recipients. Our results further suggest that ddcfDNA primarily reflects graft‐specific injury rather than being substantially influenced by stable recipient, donor, or perioperative factors. These findings are consistent with reports from several solid organ transplantation studies indicating that ddcfDNA levels are largely independent of donor, recipient, and perioperative factors [14, 22, 24]. However, this relationship has not been consistently observed across all solid organ transplant populations [25]. In our pediatric LDLT cohort, these findings support the utility of ddcfDNA as a stable biomarker that is minimally influenced by transplant‐related factors, reinforcing its potential as a real‐time tool for graft surveillance.

4.2. ddcfDNA Kinetics

Having established that ddcfDNA is a robust and largely independent variable, we next characterized its temporal pattern following transplantation. The hallmark finding of our study is the distinct kinetic profile of ddcfDNA, characterized by a peak on POD 1–2 followed by a sharp decline by POD 7 and subsequent stabilization. This early peak likely reflects IRI, surgical trauma, and systemic inflammation rather than immunologic graft rejection—consistent with findings from previous studies in both kidney and liver transplants [2, 9, 35, 36]. Statistical analyses confirmed a significant reduction in ddcfDNA levels (p < 0.001 for ddcfDNA% and p = 0.0076 for ddcfDNA qty) within the first postoperative week reflecting clearance of graft‐derived cfDNA after IRI, with levels stabilizing by POD 7. This early plateau, consistent with prior liver transplant studies, marks a clinically relevant threshold for defining graft stability, where subsequent rises could indicate rejection or other injury.

Prior studies by Cucchiari et al. [36], Shen et al. [35] (Kidney), and Zhao et al. [9](Liver), have also demonstrated that IRI particularly in deceased donor grafts is associated with elevated early ddcfDNA levels, which reflect the extent of initial graft injury and the pace of recovery. The liver, being especially vulnerable to IRI, may exhibit ddcfDNA levels peaking as high as 80%–90% immediately post‐reperfusion, returning to baseline (< 10%) by POD 10 in patients without complications [2, 9, 16, 37, 38]. This characteristic ‘L‐shaped’ decline is consistent with previous observations in adult and pediatric transplants [2, 25, 35, 39]. Our findings concur with studies by Florman et al. and Kanamori et al. who report early ddcfDNA peaks (~80%) with rapid decline and stabilization by week 2 (~5%) in adult liver recipients [2, 20]. In our other longitudinal study in adult liver cohort, we likewise observed stabilization by POD 7, with the 10% rejection threshold [16]. Similarly, our pediatric cohort also demonstrated early normalization, with ddcfDNA stabilizing by POD 7 and remaining low through POD 30–31, suggesting effective graft recovery and minimal ongoing injury.

Importantly, ddcfDNA levels correlated moderately with liver transaminases before POD 7, consistent with findings by Sorbini et al. around POD 5 [28]. While dd‐cfDNA declined rapidly in the early postoperative period, conventional LFTs also approached functional stabilization by POD 10–14. Unlike AST and ALT, which are not liver‐specific and may be elevated due to infections, medications, or hemodynamic fluctuations, ddcfDNA reflects direct turnover of donor‐derived hepatocytes, offering greater organ specificity. This distinction is especially meaningful in pediatric recipients, where variable enzyme levels complicate interpretation and invasive biopsies pose greater risk. Additionally, prior research shows ddcfDNA offers superior specificity and earlier detection of graft injury compared to LFTs; for example, ddcfDNA rises sharply (~90%) during rejection or hepatic hematoma, whereas LFT changes may be more subtle or nonspecific [35, 38].

Spearman's correlation analysis in our cohort revealed moderate associations between ddcfDNA and hepatic enzymes (AST r = 0.48, ALT r = 0.52), indicating partial overlap in the information these markers provide. Inconsistent correlations with GGT in our study highlight that not all hepatic enzymes reflect graft recovery or injury as accurately as ddcfDNA. This aligns with Kanamori et al., who reported a strong correlation between ddcfDNA and transaminases, but a weaker association with GGT, which may normalize later or be more indicative of biliary rather than hepatocellular injury [2]. This temporal advantage supports the utility of ddcfDNA as a more sensitive and timelier marker of graft stabilization.

Based on the evidence and observations discussed above, we describe exploratory postoperative ddcfDNA ranges in clinically stable pediatric liver transplant recipients that may reflect distinct phases of graft recovery and stabilization. The early range, observed around POD 7, ranges from approximately 2%–2.9% for ddcfDNA percentage and 0.6–1.0 ng/mL for absolute quantity, marking the critical initial graft recovery phase. A second range observed after POD 30 ranges from approximately 2.0%–2.4% and 0.16–0.6 ng/mL, potentially reflecting physiological steady state graft recovery in clinically stable recipients. These observations provide preliminary longitudinal reference patterns for stable pediatric LDLT recipients and may help guide future monitoring strategies. However, larger multicenter cohorts including unstable grafts are required to establish clinically validated thresholds for graft injury or rejection.

These proposed values concur with previous findings of Beck et al. who reported a mean ddcfDNA fraction of 3.5% in stable adult liver transplant recipients by POD 10 [39]. Similarly, Zhang et al. reported baseline ddcfDNA levels ranging from 3.3% to 5% [12], while Levitsky et al. proposed a higher threshold of 5.3% to differentiate normal from acute rejection [7]. The consistency of these values across diverse cohorts and transplant centers underscores the reproducibility of ddcfDNA as a biomarker, despite differences in demographics and transplant management.

While percentage dd‐cfDNA remains the primary clinically validated metric for monitoring graft health, absolute dd‐cfDNA concentrations provide complementary kinetic and quantitative information. Our analyses demonstrate that interpreting both metrics together enhances sensitivity and specificity for detecting graft‐related changes. Early studies focused solely on %dd‐cfDNA, but subsequent experience with our platform supports the utility of reporting both measures.

Importantly, the continued stability of ddcfDNA levels beyond the early postoperative phase in clinically stable patients supports the clinical utility of these dual baselines as essential comparators. They enable sensitive detection of subclinical rejection or graft dysfunction in future studies and routine monitoring, providing a nuanced framework tailored to the dynamic post‐transplant timeline.

Despite these promising findings, our study has certain limitations. The relatively small sample size and single‐center design may limit the generalizability of our results and preclude robust multivariate analysis. Several subgroup analyses were limited by small sample sizes and should therefore be interpreted as exploratory observations requiring validation in larger multicenter cohorts. Additionally, the absence of adolescent recipients restricts applicability across the full pediatric age spectrum. When LFT trends toward normalization in the postoperative period, we infer the absence of rejection; however, obtaining liver biopsies at these time points could provide deeper insight. On the other hand, invasive tissue biopsies themselves may elevate ddcfDNA levels, potentially confounding interpretation. Importantly, this study included only clinically stable pediatric liver transplant recipients; therefore, ddcfDNA patterns and thresholds in unstable grafts, rejection, or graft dysfunction remain to be established in this population. Future multicenter studies involving both stable and unstable grafts, coupled with immune marker profiling, are essential to further validate ddcfDNA as a robust and independent biomarker.

5. Conclusion

In summary, our study characterizes the early dynamics and baseline ranges of ddcfDNA in stable pediatric living donor liver transplant recipients, a niche and understudied population. Percent ddcfDNA, as the primary clinically validated metric, and absolute ddcfDNA, providing complementary kinetic information, together offer a preliminary framework for understanding ddcfDNA behavior in clinically stable pediatric liver transplant recipients. These baseline values were consistent across donor and recipient demographic variables, suggesting broader applicability within this population. Collectively, these reference ranges represent an important foundational step in the development of ddcfDNA as a biomarker in pediatric liver transplantation and provide the comparator dataset against which future investigations can evaluate ddcfDNA dynamics in the setting of graft dysfunction, rejection, or other adverse events during the early postoperative and surveillance periods.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: (a) Distribution of %ddcfDNA across serial postoperative time points in stable pediatric LDLT recipients.

Figure S1: (b) Individual patient trajectories of %ddcfDNA across four postoperative time points in stable pediatric LDLT recipients. Each colored line represents an individual patient.

Figure S2: (a)Distribution of absolute ddcfDNA concentration across serial postoperative time points in stable pediatric LDLT recipients.

Figure S2: (b) Individual patient trajectories of absolute ddcfDNA concentration across four postoperative time points in stable pediatric LDLT recipients. Each colored line represents an individual patient.

Table S1: Recipient and donor characteristics, graft metrics, and %dd‐cfDNA levels across four postoperative time points in stable pediatric LDLT recipients.

Data Availability Statement

Research data are not shared.