Low Torque Teno Virus Load Is Associated With Fibrosis in Protocol Biopsies Following Liver Transplantation: A Nordic Multicenter Cohort Study Authors: Delal Akdag, Jeanett Klubien, Andreas Arendtsen Rostved, Allan Rasmussen, Nikolai Kirkby, Gro Linno Willemoe, Deepthi Chiranth, Bo‐Göran Ericzon, Carl Jorns, William Bennet, Arno Nordin, Fredrik Åberg, Jens Georg Hillingsø, Susanne Dam Nielsen, Hans‐Christian Pommergaard Published: 2026 ABSTRACT ABSTRACTGraft inflammation and fibrosis are associated with progressive fibrosis and potential graft loss after liver transplantation, but diagnosis relies on invasive biopsies. Torque teno virus (TTV) may be a noninvasive marker of functional immunity, however its association with these changes remains unclear. In this Nordic multicenter cohort study, plasma samples collected with biopsies were analyzed for TTV load, and biopsies were scored for inflammation and fibrosis. Associations and performance of TTV versus tacrolimus were evaluated using logistic regression and ROC curves. TTV was categorized using the Youden Index. Among 287 liver transplant recipients, median age was 55, mean TTV load 5 log10 copies/mL and median time from transplantation to TTV measurement 12 months. Low TTV load was associated with higher odds of fibrosis (aOR 2.96 [95% CI 1.40–6.25], adjusted for time since transplantation). Compared with tacrolimus, TTV load remained associated with fibrosis (2.64 [1.19–5.85]), whereas tacrolimus did not (1.16 [0.95–1.42]). The model including both markers had the highest AUC (0.69 [0.60–0.78]). At the optimal threshold, TTV demonstrated 64% sensitivity and specificity, positive predictive value 20% and negative predictive value 93%. TTV load was not significantly associated with inflammation. In conclusion, TTV load may support risk‐stratified biopsy strategies, requiring external validation.Keywords: fibrosis, liver transplantation, Nordic multicenter cohort, protocol biopsy, torque teno virus Introduction As overall survival improves in liver transplant recipients, long‐term graft complications, such as inflammation and fibrosis, increasingly become an integral focus of post‐transplant care [1]. Fibrosis development is described due to continuous inflammation, which may be secondary to under immunosuppression [2, 3]. However, the most common cause in adults is recurrence of the primary liver disease [3]. Early detection is important to enable treatment of these histologic changes, to prevent progressive graft fibrosis and cirrhosis, as these conditions are associated with graft loss [4, 5]. A liver biopsy is the gold standard for diagnosing inflammation and fibrosis. However, the procedure is invasive, costly, and carries a risk of complications. Therefore, non‐invasive markers reflecting the net immune activity as a surrogate marker for inflammation and fibrosis risk are needed to select recipients who require a liver biopsy [6, 7]. Torque teno virus (TTV) has been proposed as a marker of functional immunity in solid organ transplant recipients, including liver transplant recipients [8, 9, 10]. With high TTV loads reflecting over immunosuppression and low under immunosuppression [10, 11]. However, few studies have examined TTV load as a marker of functional immunity in long‐term outcomes [12, 13], including fibrosis [14]. We aimed to (1) investigate whether TTV load in plasma, measured concurrently with protocol biopsies 1 year after transplantation, was associated with graft inflammation and fibrosis, and (2) compare the association between TTV load and tacrolimus trough levels with graft inflammation and fibrosis. We hypothesized that recipients with a low TTV load had an increased risk of inflammation and fibrosis, and that TTV load was a better marker of functional immunity than tacrolimus trough levels. Materials and Methods Study Design and Setting The Scandiatransplant organization provides the foundation for collaboration among the Nordic transplantation centers, maintains a shared waiting list for all liver transplant candidates, and serves as a platform for coordinated research efforts [15, 16]. Within this network, members of the present study group established the Nordic multicenter research project “Donor Specific HLA Alloantibodies in Liver Transplantation” (DSA study), an observational cohort designed to investigate the impact of donor‐specific HLA alloantibodies on post‐transplant outcomes (ClinicalTrials.gov ID: NCT02784080) [17]. As previously described by our study group [12], adult liver transplant recipients were enrolled between August 2015 and May 2021 from transplantation centers at Karolinska University Hospital (Stockholm, Sweden), Sahlgrenska University Hospital (Gothenburg, Sweden), Rigshospitalet (Copenhagen, Denmark), and Helsinki University Hospital (Helsinki, Finland). Follow‐up was conducted between August 2023 and September 2024. The current study was based on a sub‐cohort of the DSA participants with an available TTV measurement ±2 months of their protocol biopsy. Protocol biopsies were performed annually after liver transplantation. The study population was comparable to the broader Nordic liver transplant population with respect to sex distribution, median age, and primary indications for transplantation [18, 19]. The present study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [20]. Ethics Declaration The DSA study and biobank were approved by the National Committee on Health Research Ethics (Protocol no. H‐15007823) and the Danish Data Protection Agency (RH‐2016‐203, I‐Suite no. 04779). Approval and exemption from collecting novel consents for the present study were obtained from the Danish National Committee on Health Research Ethics (Case no. 2110191, Doc. no. 1787215), as all liver transplant recipients with biobank material were informed and consented to participate in further research at the time of inclusion. The study protocol adheres to the ethical guidelines of the 2013 Declaration of Helsinki and complies with the 2018 Declaration of Istanbul. Outcome Measures To assess whether inflammation, fibrosis, and the severity of these changes were different between recipients with a low and high TTV load, we compared the proportion of inflammation, inflammation score, fibrosis, and METAVIR score in protocol biopsies across TTV categories. Fibrosis was evaluated according to the METAVIR classification (F0–F4). Clinically significant fibrosis was defined as a METAVIR score ≥ F2 [21]. Any inflammation was assessed using the interface (0–3), lobular (0–3), or portal inflammation (0–3). Inflammation was defined as a case when either the portal, interface, or lobular inflammation score was ≥ 2. The inflammation score (0–3) was determined as the highest score among the interface, lobular, or portal scores. For histological outcomes significantly associated with TTV load, we further compared the diagnostic value of TTV load and tacrolimus trough levels, as both may serve as surrogate markers of immune competence. To assess whether adding tacrolimus to the model performed as well as TTV load or improved discriminative ability for histological outcomes, the two markers were evaluated individually and in combination. TTV load was log10‐transformed to normalize data. Four patients (1.4%) had values below 2 log10 copies/mL; no values were undetectable or negative. These values were included as reported without imputation or censoring. Because no clinically established cutoff exists, TTV load was investigated as both a continuous variable (per log10 decrease) and a categorical variable (low versus high) in the multivariable models. TTV load was categorized based on the optimal thresholds, derived by using the Youden Index from receiver operating characteristic (ROC) curve analysis, selecting the point that maximized sensitivity and specificity for discriminating the presence of inflammation and fibrosis, respectively. The Youden index was chosen to evaluate the overall discriminatory performance of TTV rather than to establish a clinical decision threshold. This data‐driven approach identified cutoffs of 5.59 log10 copies/mL for inflammation and 4.94 log10 copies/mL for fibrosis. Data Sources Data were retrieved from the DSA study REDCap database. Data included sex, age, and ABO blood type of recipient and donor, transplantation center, reason for transplantation, bilirubin, International Normalized Ratio, creatinine, dialysis at the time of transplantation to calculate the Model for End‐Stage Liver Disease (MELD) score, cold ischemia time, history of re‐transplantation before inclusion in the current study, and the main immunosuppressive treatment at the time of the protocol biopsy (tacrolimus, ciclosporin, or everolimus), including trough levels. Liver Biopsies Protocol biopsies were defined as liver biopsies performed at a prespecified time point, without clinical suspicion of graft dysfunction or rejection. Protocol biopsies were performed and evaluated by trained pathologists at the respective transplant centers to assess inflammation and fibrosis in recipients. Data were entered into the REDCap database. Torque Teno Virus Plasma samples were stored in the DSA study's biobank at −80°C until analysis. One plasma sample per recipient was used. DNA was extracted from 200 μL thawed plasma samples using the eMAG/easyMAG Nucleic Acid Extraction System. The TTV R‐GENE kit by bioMérieux was used to perform PCR to quantify the TTV viremia load. Included in the R‐GENE kit were an extraction, inhibition internal control, and a range of four quantification standards. Detection and quantification of TTV‐DNA in all plasma samples were performed at the Department of Clinical Microbiology, Rigshospitalet, Denmark. Statistical Methods Descriptive and Comparative Analyses After log‐transformation, TTV load was normally distributed and therefore reported as mean and standard deviation (SD). The remaining continuous variables were non‐normally distributed and reported as medians with interquartile ranges (IQRs). Categorical data were presented as frequency counts and percentages of subjects within each category and TTV groups. Proportions of inflammation and fibrosis were compared across TTV categories using Fisher's exact test, while inflammation score and METAVIR scores were compared using the Mann–Whitney U test. Regression and Predictive Modeling Univariable and multivariable logistic regression analyses were performed to assess the association between TTV load and the outcomes inflammation and fibrosis, expressed as odds ratios (OR) and adjusted odds ratios (aOR) with 95% confidence intervals [95% CI]. Based on previous research and a research project currently under submission by our study group (July 2026), we know that the proportion of fibrosis among liver transplant recipients is low 1 year after transplantation [22]. To prevent overfitting, we included one variable for every 10–15 cases in a prioritized order including the most plausible confounders from previous studies. Accordingly, in the model with inflammation as the outcome, we included TTV load (per log10 decrease or low versus high), time since transplantation (per month increase), and age at transplantation (per decade increase). In the model with fibrosis as the outcome, we included TTV load and time since transplantation. To assess whether the underlying reasons for transplantation influenced the association between TTV load and fibrosis, post hoc sensitivity analyses were performed in which the fibrosis model was additionally adjusted for clinically relevant transplantation indications, including autoimmune liver disease, cirrhosis, and hepatocellular carcinoma. These variables were evaluated individually due to the limited number of fibrosis cases to avoid model overfitting. Model assumptions, including linearity, collinearity, residuals, and outliers, were tested. Linearity of TTV load was evaluated by comparing models with TTV included as a linear term and as a restricted cubic spline, and the spline model did not improve fit. None of the models violated the above assumptions. To compare the predictive ability of TTV load and tacrolimus trough levels for histological outcomes, each variable was first investigated separately and then jointly in a multivariable logistic regression model. We then compared the area under the ROC curves for the three multivariable logistic regression analyses, using DeLong's method. In addition, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall accuracy were compared for the models with and without tacrolimus to investigate whether it improved diagnostic performance. During the preparation of this work, the authors used OpenAI's text generation models, generative pre‐trained transformers 5.2. The model was solely used to improve and clarify readability and language throughout the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. Data were analyzed using RStudio 2022.07.1+554. Statistical significance was set at a p‐value < 0.05. p‐values were not adjusted for multiple tests. The number of missing values was reported where applicable. No imputation was used. Results Recipient Characteristics The study cohort included 287 liver transplant recipients. The median time from transplantation to protocol biopsy was 12 months (IQR 11–13). Using the Youden Index cutoff for fibrosis, 154 recipients were categorized with a high TTV load and 133 with a low TTV load. Figure 1 shows the distribution of TTV load for the study cohort and according to fibrosis status. Distribution of log10‐transformed torque teno virus (TTV) load. (A) Distribution of TTV load for the whole cohort (n = 287). (B) Distribution of TTV load stratified by fibrosis status. p‐value < 0.05 was considered significant and was calculated using the Mann–Whitney U test. Characteristics of the entire cohort, and by TTV group, are presented in Table 1. The proportion of females was comparable between groups. However, recipients with a low TTV load were younger (median 53 vs. 57 years) and were more often transplanted due to autoimmune liver diseases (49% vs. 37%), and less frequently due to cirrhosis (40% vs. 53%) or hepatocellular carcinoma (17% vs. 25%). Characteristics of recipients stratified according to fibrosis status are presented in Table S1. TABLE 1Characteristics of liver transplant recipients.TotalHigh TTVLow TTV p n = 287 n = 154 n = 133 n (%) n (%) n (%)Sex, female110 (38)57 (37)53 (40)0.62Sex‐mismatch93 (32)51 (33)42 (32)0.78Recipient age at TX, years, median (IQR)55 (47–62)57 (50–63)53 (44–61)< 0.01*< 4045 (16)18 (12)27 (20)40–4949 (17)18 (12)31 (23)50–5990 (31)53 (34)37 (28)60–6994 (33)59 (38)35 (26)> 709 (3)6 (4)3 (2)Donor age, years, median (IQR)58 (45–68)58 (47–68)59 (43–69)0.77CountryDenmark70 (24)23 (15)47 (35)< 0.01*Finland56 (20)52 (34)4 (3)Sweden161 (56)79 (51)82 (62)TTV load, log10 copies/mL, mean (SD)5.0 (1.3)6.0 (0.9)3.9 (0.8)< 0.01*Time from transplantation to biopsy, months, median (IQR)12 (11–13)12 (11–12)12 (11–14)< 0.01*> 1 year79 (28)26 (17)53 (40)MELD score, median (IQR)12 (8–17)12 (9–18)11 (8–15)0.05*< 20198 (87)83 (82)115 (91)> 2030 (13)18 (18)12 (9)Missing59 (21)21 (11)38 (40)ABO‐blood type matchIdentical265 (94)137 (91)128 (98)0.01*Compatible15 (5)13 (8)2 (1)Mismatch2 (1)1 (1)1 (1)Cold ischemia time, h, median (IQR)6.3 (5.0–8.0)6.0 (4.5–8.1)6.5 (5.4–8.2)0.01*Prior re‐transplantation10 (4)6 (4)4 (3)0.76Reason for transplantationAutoimmune liver disease (any cause)122 (43)57 (37)65 (49)0.04*Primary sclerosing cholangitis88 (31)37 (24)51 (38)< 0.01*Primary biliary cirrhosis22 (8)14 (9)8 (6)0.33Cirrhosis (any cause)134 (47)81 (53)53 (40)0.03*Cirrhosis (alcohol)68 (24)44 (29)24 (18)0.04*Cirrhosis (viral)17 (6)11 (7)6 (5)0.35Cirrhosis (autoimmune)28 (10)13 (8)15 (11)0.42Hepatocellular carcinoma60 (21)38 (25)22 (17)0.09Acute liver failure1 (0)0 (0)1 (1)0.46Hepatitis C12 (4)7 (5)5 (4)0.74Metabolic liver disease31 (11)10 (7)21 (16)0.01*Other23 (8)15 (10)8 (6)0.25Immunosuppression at the time of biopsyTacrolimus238 (87)121 (91)117 (79)< 0.01*Ciclosporine32 (12)28 (7)4 (21)Everolimus4 (2)1 (2)3 (0)Missing13 (5)4 (3)9 (7)Trough levels, μg/L, median (IQR)Tacrolimus5.8 (4.9–7.4)5.9 (4.9–8.0)5.7 (4.9–6.9)0.21Ciclosporine128.0 (117.5–151.8)142.0 (120.3–159.8)114.0 (98.0–125.5)0.10Everolimus5.5 (4.1–6.2)——— Note: The first column includes all liver transplant recipients; the latter two columns group the cohort according to low or high TTV. TTV was categorized by determining the optimal cutoff point for fibrosis, using Youden's Index. Percentages were calculated based on existing data. Variables with more than 5% missing data are presented where relevant. *p‐value < 0.05 were considered significant.Abbreviations: MELD, Model for End‐stage Liver Disease; TTV, torque teno virus; TX, transplantation. Characteristics of liver transplant recipients. Note: The first column includes all liver transplant recipients; the latter two columns group the cohort according to low or high TTV. TTV was categorized by determining the optimal cutoff point for fibrosis, using Youden's Index. Percentages were calculated based on existing data. Variables with more than 5% missing data are presented where relevant. *p‐value < 0.05 were considered significant. Abbreviations: MELD, Model for End‐stage Liver Disease; TTV, torque teno virus; TX, transplantation. Proportion of Inflammation, Fibrosis, and Severity by TTV Groups Figure 2 shows the distribution of inflammation and fibrosis, along with corresponding inflammation and METAVIR scores, in protocol biopsies stratified by TTV group. Proportion of recipients with inflammation and fibrosis in protocol biopsy by TTV category. None of the recipients had a METAVIR score of F4 in either group. p‐value < 0.05 was considered significant and calculated using Fisher's Exact test or the Mann–Whitney U test. Fibrosis was defined as a METAVIR score ≥ F2. Inflammation was defined as an inflammation score ≥ 2. The proportion of recipients with any inflammation did not differ significantly between the low and high TTV groups (16/106 [15%] vs. 24/135 [18%], p = 0.61). Although significantly fewer recipients in the low TTV group had no inflammation (17/106 [16%] vs. 39/135 [29%], p = 0.03) and a higher proportion had mild (73/106 [69%] vs. 72/135 [53%], p = 0.02), the overall inflammation score did not significantly differ between groups (p = 0.22). The proportion of recipients with fibrosis was significantly higher in the low TTV group than in the high TTV group (24/133 [18%] vs. 12/154 [8%], p = 0.01). This difference was primarily driven by a lower proportion of recipients with no fibrosis (51/133 [39%] vs. 83/154 [54%], p = 0.01) and a higher proportion with moderate fibrosis in the low TTV group (22/133 [17%] vs. 9/154 [6%], p < 0.01) resulting in a significantly different overall METAVIR score between the groups (p < 0.01). None of the recipients had a METAVIR score of F4 in either group. Association of TTV Load and Recipient Characteristics With Inflammation Univariable and multivariable logistic regression analyses of risk factors for inflammation are presented in Table 2. TABLE 2Logistic regression analyses to assess risk factors for inflammation in liver transplant recipients.InflammationUnivariableMultivariableOR [95% CI] p aOR [95% CI] p TTV, pr. log10 decrease1.01 [0.79–1.29]0.961.10 [0.85–1.42]0.47TTV, low vs. high0.82 [0.41–1.64]0.581.03 [0.50–2.10]0.94Time since transplantation, per month increase0.91 [0.78–0.99]0.090.89 [0.78–1.01]0.08Age at TX, per decade increase0.70 [0.54–0.92]0.01*0.69 [0.52–0.89]< 0.01* Note: We investigated TTV load both as a continuous and a categorical variable. TTV was categorized by determining the optimal cutoff point for inflammation, using Youden's Index. *p‐value < 0.05 was considered significant.Abbreviation: TTV, torque teno virus. Logistic regression analyses to assess risk factors for inflammation in liver transplant recipients. Note: We investigated TTV load both as a continuous and a categorical variable. TTV was categorized by determining the optimal cutoff point for inflammation, using Youden's Index. *p‐value < 0.05 was considered significant. Abbreviation: TTV, torque teno virus. A low TTV was not associated with increased odds of inflammation in either the univariable or the multivariable model adjusted for time since transplantation and recipient age (OR 1.01 [95% CI 0.79–1.29] and aOR 1.03 [0.79–1.34], respectively). Similarly, TTV load per log10 decrease was not associated with inflammation in any model (OR 1.01 [0.79–1.29]; aOR 1.03 [0.79–1.34]). Recipients' age per decade increase was significantly associated with lower odds of inflammation in both the univariable and multivariable model adjusted for TTV load and time since transplantation, aOR 0.68 [0.52–0.89]. Time since transplantation showed no significant association in any model (multivariable model, p = 0.08). Association of TTV Load and Recipient Characteristics With Fibrosis Univariable and multivariable logistic regression analyses of risk factors for fibrosis are presented in Table 3. TABLE 3Logistic regression analyses to assess risk factors for fibrosis in liver transplant recipients.FibrosisUnivariableMultivariableOR [95% CI] p aOR [95% CI] p TTV, pr. log10 decrease1.34 [1.01–1.76]0.04*1.43 [1.07–1.92]0.02*TTV, low vs. high2.61 [1.25–5.44]0.01*2.96 [1.40–6.25]< 0.01*Time since transplantation, per month increase0.96 [0.89–1.04]0.320.94 [0.87–1.02]0.15 Note: We investigated TTV load both as a continuous and a categorical variable. TTV was categorized by determining the optimal cutoff point for fibrosis, using Youden's Index. *p‐value < 0.05 was considered significant.Abbreviation: TTV, torque teno virus. Logistic regression analyses to assess risk factors for fibrosis in liver transplant recipients. Note: We investigated TTV load both as a continuous and a categorical variable. TTV was categorized by determining the optimal cutoff point for fibrosis, using Youden's Index. *p‐value < 0.05 was considered significant. Abbreviation: TTV, torque teno virus. Low TTV load was associated with increased odds of fibrosis in both the univariable and multivariable models adjusted for time since transplantation (OR 2.61 [1.25–5.44]; aOR 2.96 [1.40–6.25], respectively). When including TTV load as a continuous variable, each log10 decrease in TTV load remained significantly associated with increased odds of fibrosis in both models (OR 1.34 [1.01–1.76]; aOR 1.43 [1.07–1.92]). Time since transplantation was not significantly associated with fibrosis in any model (multivariable model, p = 0.15). In post hoc sensitivity analyses, models additionally adjusted for autoimmune liver disease, cirrhosis, or hepatocellular carcinoma showed consistent associations between TTV load and fibrosis, with each log10 decrease in TTV load remaining significantly associated with increased odds of fibrosis, aOR 1.48 [1.10–1.99], aOR 1.42 [1.06–1.91], and aOR 1.43 [1.07–1.92], respectively. Comparison of Models Including TTV Load and Tacrolimus Trough Levels for Fibrosis The results of multivariable logistic regression analyses and AUCs for fibrosis are presented in Table 4. TABLE 4Comparison of TTV and Tacrolimus trough levels association with fibrosis.TTVTacrolimusAUC [95% CI] p aOR [95% CI] p aOR [95% CI] p TTV + time since transplantation2.96 [1.40–6.25]0.01*——0.64 [0.54–0.74]—Tacrolimus + time since transplantation——1.18 [0.97–1.43]0.100.63 [0.53–0.73]0.83 (vs. TTV)TTV + tacrolimus + time since transplantation2.64 [1.19–5.85]0.02*1.16 [0.95–1.42]0.140.69 [0.60–0.78]0.07 (vs. TTV) Note: The first model included TTV (low vs. high) and time since transplantation (per‐month increase), with corresponding adjusted odds ratios (aORs) and AUC. The second model included tacrolimus (per μg/L decrease) and time since transplantation (per month increase). The third model included TTV (low vs. high), tacrolimus (per μg/L decrease) and time since transplantation (per month increase). aOR not shown for “time since transplantation.” The p‐values are calculated by comparing the second and third models to the first using DeLongs method. *p‐values < 0.05 was considered significant.Abbreviations: aOR, adjusted odds ratio; TTV, torque teno virus. Comparison of TTV and Tacrolimus trough levels association with fibrosis. Note: The first model included TTV (low vs. high) and time since transplantation (per‐month increase), with corresponding adjusted odds ratios (aORs) and AUC. The second model included tacrolimus (per μg/L decrease) and time since transplantation (per month increase). The third model included TTV (low vs. high), tacrolimus (per μg/L decrease) and time since transplantation (per month increase). aOR not shown for “time since transplantation.” The p‐values are calculated by comparing the second and third models to the first using DeLongs method. *p‐values < 0.05 was considered significant. Abbreviations: aOR, adjusted odds ratio; TTV, torque teno virus. In the multivariable model adjusted for TTV load and time since transplantation, low TTV load was significantly associated with increased odds of fibrosis (aOR 2.96 [1.40–6.25]). In contrast, the model including tacrolimus trough level and time since transplantation, lower tacrolimus trough levels were not associated with fibrosis (aOR 1.18 [0.97–1.43]). In the combined model, adjusted for TTV load, tacrolimus trough level, and time since transplantation, low TTV load remained significantly associated, whereas tacrolimus trough level did not (TTV load: aOR 2.64 [1.19–5.85]; tacrolimus: aOR 1.16 [0.95–1.42]). Predictive and Diagnostic Value of TTV Load as a Marker of Fibrosis ROC curves and corresponding AUCs for models including TTV load, tacrolimus trough levels, and their combination for prediction of fibrosis are presented in Figure 3. Receiver operating characteristic (ROC) curves and areas under the curve (AUCs) comparing the performance of three multivariable models for predicting fibrosis: TTV load and time since transplantation (solid line), tacrolimus trough level and time since transplantation (dashed line), and TTV load, tacrolimus trough level, and time since transplantation (dotted line). When we compared the discriminatory ability of the three models, the combined model, including both TTV and tacrolimus, showed the greatest AUC (0.69 [95% CI 0.60–0.78]), followed by the model with TTV load (0.64 [0.54–0.74]) and the model with tacrolimus (0.63 [0.53–0.73]). However, the AUC of the combined model did not significantly differ from the model with TTV load (p = 0.07) (Table 4). When comparing the predictive value of the models, the TTV‐only model performed better as a rule‐out tool than the combined TTV and tacrolimus model. For the TTV model, sensitivity and specificity were both 64%, PPV was 20%, and NPV 93%. Whereas in the combined model, sensitivity was 58%, specificity 64%, PPV 21%, and NPV 90%. Additionally, to identify a TTV threshold capable of ruling out fibrosis, we determined the highest TTV value observed among recipients with fibrosis. A TTV load > 7.45 log10 copies/mL (negative test) was not observed in any fibrosis patient (NPV of 100%). Thus, a cut‐off of TTV load < 7.45 log10 copies/mL (positive test) correctly identified all fibrosis cases (sensitivity = 36/36 = 100%). However, only 12 of 251 non‐fibrotic patients exceeded this threshold (TTV load > 7.45 log10 copies/mL) corresponding to a specificity of 12/251 = 5%, indicating that while high TTV strongly suggests the absence of fibrosis, most patients without fibrosis still had TTV values below this threshold. Discussion In this multicenter cohort study, TTV load, measured concurrently with protocol biopsies, was associated with graft fibrosis. Liver transplant recipients with a low TTV load had a higher proportion of fibrosis and threefold the odds of fibrosis compared to those with a high TTV load, independent of time since transplantation. When comparing TTV load association with fibrosis to that of tacrolimus trough level, only TTV load was significantly associated with fibrosis, and adding tacrolimus did not improve model performance significantly. At the Youden index cutoff, a high TTV load identified recipients without fibrosis with a NPV of 93%. TTV load was not associated with inflammation. The recipient's age was significantly associated with lower odds of inflammation, independent of TTV load and time since transplantation. TTV load measured in blood inversely correlates with immune function in solid‐organ transplant recipients, including liver transplant recipients [10, 11]. This corresponds with the findings of the present study, in which recipients with a low TTV load had a higher proportion of fibrosis than those with a high TTV load (18% vs. 8%, p = 0.01), supporting the immune‐mediated role in the development of fibrosis after transplantation. This relationship was confirmed in the regression analysis, where recipients with a low TTV load had threefold higher odds of fibrosis, independent of time since transplantation. Few studies have investigated the role of TTV load and its utility as a screening marker for long‐term outcomes [12, 13, 14]. Engel et al. evaluated TTV load in relation to fibrosis, using the liver allograft fibrosis (LAF) score and the Ishak fibrosis score, in 80 liver transplant recipients [14]. When restricting the analysis to recipients with TTV measured 1 year after transplantation, the authors observed a moderate inverse correlation between TTV load and fibrosis severity. When including the entire cohort, the correlation weakened but remained statistically significant. In multivariable linear regression, lower TTV load remained independently associated with higher fibrosis scores after adjusting for age, sex, and immunosuppression intensity. Despite methodological differences between our study and Engel et al. (e.g., biopsy indication, sample size, and statistical methods), the findings are consistent. Specifically, both studies suggest that lower TTV load, as a surrogate marker of under‐immunosuppression, was associated with fibrosis. Although we did not find a significant difference in the proportion of any inflammation (inflammation score > 2) between the TTV groups, significantly fewer recipients with a low TTV load had no inflammation, and more had mild inflammation compared to the high TTV group. The study by Engel et al. did not find TTV load to be correlated with inflammation score, as determined by the rejection activity index, when assessed within or after the first year after transplantation. The most common cause of graft inflammation and fibrosis in adult transplant recipients is recurrence of the original liver disease, and idiopathic post‐transplant hepatitis is often a cause of graft inflammation [3]. In our cohort, the distribution of transplantation indications was comparable between recipients with and without fibrosis, and additional adjustment for clinically relevant transplantation indications did not alter the association between TTV load and fibrosis. These findings suggest that differences in the underlying indication for transplantation are unlikely to fully explain the observed relationship between TTV load and fibrosis. However, disease‐specific mechanisms may contribute differently to graft injury, and larger studies are needed to evaluate whether the association between TTV load and fibrosis varies according to the underlying liver disease etiology. TTV load is higher in recipients receiving triple immunosuppression treatment, compared to those on monotherapy [23]. Reflecting that, with heavy immunosuppression, a corresponding increase in TTV load is observed. However, tacrolimus trough levels alone do not appear to correlate with TTV load [14, 23]. These findings possibly explain why the current study and the study by Engel did not find an association between tacrolimus trough level and fibrosis. While TTV load remained significantly associated with fibrosis, even after adjusting for tacrolimus trough level, possibly suggesting that TTV load may be a better marker of net immune function than tacrolimus trough level. We further investigated the diagnostic and predictive value of TTV load as a marker of fibrosis. TTV load demonstrated modest overall discriminatory ability (AUC: 0.64–0.69), which is consistent with previous research [23]. The fact that TTV only has a modest overall ability to classify fibrosis correctly likely reflects that it may be a marker of net immune function, whereas fibrosis is a multifactorial and not solely immune‐mediated process. Using the Youden Index threshold of 4.94 log10 copies/mL, TTV could potentially serve as a practical rule‐out tool for graft fibrosis. In our cohort, a TTV load above the threshold identified a low‐risk group with an NPV of 93% (161/174), corresponding to a 7.5% miss rate, where 13 of 174 patients classified as low‐risk had fibrosis. If validated in a randomized study, these findings suggest that recipients with TTV ≥ 4.94 and no additional indicators of graft injury may be considered low risk and managed with routine follow‐up without immediate biopsy. In contrast, recipients with TTV < 4.94 should be considered at increased risk and undergo stepwise evaluation. This could include reassessment of immunosuppressive exposure, graft age, and liver biochemistry, followed by non‐invasive fibrosis testing (e.g., elastography), with biopsy considered when multiple clinical indicators align with potential fibrosis. To further evaluate TTV as a fibrosis rule‐out marker, we identified the highest TTV value observed among recipients with fibrosis. No fibrosis case had a TTV > 7.45, however, only 12 of 251 non‐fibrotic recipients exceeded this threshold (specificity 5%), indicating that while the threshold rules out fibrosis, the high false‐positive rate limits its usefulness for reducing protocol biopsies. We want to address the strengths and limitations of our study. First, the multicenter design of our study enhances generalizability across different transplant populations and clinical routines. Second, TTV and biopsy assessments were performed within a narrow two‐month time window, strengthening the temporal association between viral load and histological findings. Third, all TTV measurements were quantified in the same laboratory, minimizing analytical and inter‐laboratory variability and ensuring consistency in viral quantification across the cohort. The limitations of our study include, first, although TTV load demonstrated possible rule‐out capability, it cannot be used as a stand‐alone diagnostic marker for fibrosis. The PPV was low, meaning that a low TTV load does not confirm fibrosis. Second, TTV load was assessed using a single measurement per recipient in our cohort. Because TTV is known to fluctuate, including intra‐individually, serial measurements may provide a more accurate reflection of individual immune function and associated fibrosis risk. Third, protocol biopsies were evaluated locally at each transplantation center, potentially introducing inter‐observer variability in the histopathological assessment of inflammation and fibrosis. This may influence the precision of fibrosis classification and should be considered when interpreting the findings. Furthermore, fibrosis was assessed using the METAVIR classification rather than a transplant‐specific fibrosis scoring system. Although METAVIR ≥ F2 is an established threshold for clinically significant fibrosis after liver transplantation, transplant‐specific scores may provide a more detailed assessment of fibrosis patterns specific to liver allografts. Fourth, for 22 recipients (7.7%), TTV load was measured after the protocol biopsy. Because data on changes in immunosuppressive therapy between the biopsy and TTV measurement were unavailable, we cannot exclude that post‐biopsy adjustments in immunosuppression influenced TTV load in this small subgroup, potentially affecting the observed associations. Fifth, because inflammation subtype data were missing in some cases, inflammation was categorized as “any inflammation” based on the highest available interface, lobular, or portal score. As these patterns differ in clinical relevance, collapsing them into a single category may have introduced misclassification, reducing our ability to detect subtype‐specific associations and potentially underestimating the effect. Sixth, leukocyte count data were not available in our cohort, preventing adjustment. As TTV load has been reported to be associated with leukocyte counts, residual confounding cannot be excluded. Finally, although we adjusted for tacrolimus exposure, residual confounding by immunosuppressive regimens cannot be excluded. TTV load may serve as an accessible, minimally invasive marker of functional immune status. In this cohort, low TTV load was associated with a threefold higher odds of fibrosis and could be used to guide a risk‐stratified biopsy strategy. Patients with high TTV levels appear to be at low risk and may not require routine protocol biopsy, whereas low TTV levels could help identify recipients who warrant closer follow‐up or further diagnostic evaluation. In summary, incorporating TTV into post‐transplant surveillance may support more targeted monitoring and reduce unnecessary biopsies, although external validation is required before routine clinical implementation. Funding The Danish Cancer Society, grant number R306‐A18169, supported this research. Conflicts of Interest F.Å. declares consulting fees from AstraZeneca, Takeda, Ipsen, Boehringer‐Ingelheim, and lecture fees from AbbVie, Sandoz, Astellas, Siemens Healthineers, and Norgine. A.N. declares lecture fees from Norgine. S.D.N. declares unrestricted research grants from Novo Nordic Foundation, Independent Research Foundation Denmark, Rigshospitalet Research Foundation, E.U. Horizon, Svend Andersen Fonden, and Kirsten og Freddy Johansen Fonden. Lecture fees and advisory board from Gilead, Takeda, and GSK. H.C.P. declares unrestricted research grants from Danish Cancer Society, Independent Research Fund Denmark, Svend Andersen Fonden, ctDNA Center Denmark, Research Council Copenhagen University Hospital Rigshospitalet, and advisory board for AstraZeneca. D.A., A.A.R., A.R., N.K., G.L.W., D.C., B.‐G.E., C.J., W.B., and J.G.H. declare no conflicts of interest. Supporting information Table S1: Characteristics of liver transplant recipients by fibrosis status. Data Availability Statement Data sharing not applicable to this article as no datasets were generated or analysed during the current study. References Associated Data Supplementary Materials Table S1: Characteristics of liver transplant recipients by fibrosis status. Data Availability Statement Data sharing not applicable to this article as no datasets were generated or analysed during the current study.