Fibrosis in One‐Year Protocol Biopsies After Liver Transplantation Is Associated With Graft Loss: A Multicenter Cohort Study

ABSTRACT

Subclinical inflammation and fibrosis are common after liver transplantation and are associated with progressive fibrosis and cirrhosis, which are causes of graft loss. However, the association between early inflammation and fibrosis with graft loss remains insufficiently studied. In this multicenter cohort study, protocol biopsies collected one year posttransplant were assessed for inflammation and fibrosis. Fibrosis risk was assessed in recipients with paired biopsies at 1 year and ≥ 2 years posttransplantation. Associations of inflammation and fibrosis with graft loss and mortality were assessed using Cox regression. Graft loss was defined as re‐transplantation or death due to liver complications. We included 446 liver transplant recipients. One year after transplantation, 54 recipients (18%) had inflammation, and 48 (12%) had fibrosis. In multivariable analysis, fibrosis was associated with graft loss, aHR 6.72 (95% CI: 1.36–33.29). In paired biopsies, recipients with fibrosis one year after transplantation had a fivefold higher risk of fibrosis at follow‐up than those without. Inflammation was not associated with graft loss or mortality, and fibrosis was not associated with mortality. To conclude, recipients with fibrosis one year after transplantation had an increased risk of graft loss. Protocol biopsies remain valuable for detecting subclinical changes, pending further research to determine the effect on graft loss prevention.

1. Introduction

The role of protocol biopsies in posttransplant care for liver transplant recipients remains under evaluation [1, 2, 3]. While the procedure is considered invasive, costly, and may lack therapeutic value when liver function tests are normal [1], subclinical histologic abnormalities, such as inflammation and fibrosis, are common [4, 5]. These findings are clinically relevant, as they are associated with an increased risk of late‐stage progressive fibrosis and cirrhosis [3, 4], which are known causes of graft loss [1, 6, 7]. Yet, the association between early graft inflammation and fibrosis during the first posttransplant years, when intervention remains possible, and subsequent graft loss warrants further investigation [4].

Graft inflammation is a well‐characterized driver of continuous wound healing, subsequent fibrosis, and cirrhosis of the liver [5, 8]. This mechanism and its consequences have been widely investigated in the pediatric population and adults transplanted due to cirrhosis secondary to HCV or HBV infection [5, 9, 10, 11]. Current immunosuppression regimens are highly effective, and antiviral therapies have reduced the burden of posttransplant cirrhosis in recipients with HCV and HBV [12]. However, fibrosis development is well documented in patients transplanted for other indications [7, 13, 14, 15], as graft fibrosis represents the common pathway of these diseases, including the recurrence of autoimmune liver disease, such as PSC and PBC [13, 14], and chronic rejection, often caused by underimmunosuppression, reflected in early graft inflammation and subsequent fibrosis [16, 17, 18].

A liver biopsy remains the gold standard for diagnosing fibrosis in liver transplant recipients [1, 6]. Graft fibrosis is a dynamic histological process that may regress in some recipients, while it progresses in others. Thus, identifying and characterizing recipients with fibrosis after transplantation is necessary to determine who may benefit from additional monitoring or treatment. Yet, data on the burden of fibrosis one year after transplantation, in an unselected cohort of adult liver transplant recipients, are sparse [9, 10, 15].

This study aimed to (1) determine the proportion of recipients with inflammation, fibrosis, and chronic rejection in protocol biopsies one year after liver transplantation; (2) assess if inflammation and fibrosis at one year were associated with fibrosis progression; and (3) investigate if inflammation, fibrosis, and relevant risk factors were associated with graft loss and mortality. We hypothesized that recipients with inflammation and fibrosis one year after liver transplantation were at increased risk of fibrosis progression, graft loss, and mortality.

2. Methods

2.1. Study Overview

The Nordic transplantation centers closely collaborate through the common organ exchange organization, Scandiatransplant [19]. The purpose of the organization is to maintain a waiting list for all Nordic liver transplant candidates and to form a network to support and enhance research. It is within this network that the research project “Donor Specific HLA Alloantibodies in Liver Transplantation” (DSA study) was established, including liver transplant recipients from Sweden, Denmark, and Finland (clinicaltrials.gov ID: NCT02784080) [20]. The DSA study is an observational cohort study, with prospectively collected data, on which the present study was based. All adult liver transplant recipients included in the DSA study with an available protocol liver biopsy, at any time point after transplantation, were included in the current study; indication biopsies were not included. (clinicaltrials.gov ID: NCT05492617).

The present study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [21].

2.2. Study Design and Setting

This was a Nordic multicenter cohort study. Liver transplant recipients were included between August 2015 and May 2021 from four transplantation centers in Scandinavia: (1) the Department of Transplantation Surgery, Karolinska University Hospital, Stockholm, Sweden; (2) Transplant Institute, Sahlgrenska University Hospital, Gothenburg, Sweden; (3) Department of Digestive Diseases, Transplantation and General Surgery, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark; and (4) Transplantation and Liver Surgery, Helsinki University Hospital and University of Helsinki, Helsinki, Finland. Follow‐up was conducted between August 2023 and September 2024. Data was collected prospectively and managed using Research Electronic Data Capture tools (REDCap) hosted at Rigshospitalet, Copenhagen University, Denmark.

2.3. 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). An approval and exemption from collecting novel consents for the present study was 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 conforms to the ethical guidelines of the 2013 Declaration of Helsinki and abides by the 2018 Declaration of Istanbul.

2.4. Outcome Measures

The study cohort comprised 446 liver transplant recipients. We determined the proportion of recipients with inflammation, fibrosis, and chronic rejection in recipients with one‐year protocol biopsies (n = 414). In recipients with available paired biopsy data (n = 96) from 1 year and ≥ 2 years after transplantation, we compared changes in fibrosis status, categorizing outcomes as either progression or regression. We additionally determined the association between one‐year inflammation and the presence of fibrosis in the subsequent biopsy (n = 92).

In a subset of 166 recipients, biopsies obtained at the time of transplantation were available. We compared fibrosis and METAVIR scores at the time of transplantation and one year later to determine the proportion of de novo versus persistent fibrosis.

For the Cox regression analysis, we assessed the first available protocol biopsy from each recipient for the presence of inflammation and fibrosis. Any subsequent inflammation or fibrosis events were not considered in the analysis. We investigated the association between inflammation and fibrosis and the outcomes: graft loss and death. Graft loss was defined as re‐transplantation or the recipient's death due to liver complications. Follow‐up time with graft loss as an outcome was calculated from the date of the first protocol biopsy to the occurrence of the outcome, censored at death due to causes other than graft loss or at the end of the study. Follow‐up time with death as an outcome was calculated from the date of the first protocol biopsy to the date of death, censored at the end of the study.

2.5. Inflammation, Fibrosis, Chronic Rejection Definitions

Fibrosis was assessed according to the METAVIR classification (F0–F4) and defined as a case when the METAVIR score was ≥ F2. Inflammation was assessed using the interface, lobular, and portal inflammation scores (each graded 0–3). Inflammation was defined as present when at least one of the scores was ≥ 2, and the overall inflammation category (0–3) was assigned according to the highest score among the three inflammation components [22]. Chronic rejection was defined as the recipient fulfilling all the following criteria: atrophy of bile ducts (0/1) = 1; ductopenia (0/1) = 1; sclerosing around bile duct (0/1) = 0; ductular reaction (0–3) ≤ 1; bile duct metaplasia (0–3) ≤ 1; edema (0/1) = 0; ballooning (0–3) ≥ 1; copper associated protein (0–3) ≤ 1. Only five cases of chronic rejection were observed in the one‐year protocol biopsies, with no associated deaths or graft losses during follow‐up. As a result, regression analyses using chronic rejection as an outcome were underpowered and affected by perfect data separation. Therefore, the impact on graft loss and death could not be estimated.

2.6. Data Sources

Data was retrieved from the DSA study REDCap database. Data included sex, age, and ABO blood type of recipient and donor, country of origin, reason for transplantation, MELD score, cold ischemia time, history of re‐transplantation before inclusion in the current study, and use of Tacrolimus, Ciclosporin, or Everolimus, including trough level, at the time of the first available protocol biopsy.

2.7. Liver Biopsies

Protocol biopsies were defined as a liver biopsy performed at a prespecified time point and without clinical suspicion of graft dysfunction or rejection. The protocol biopsies were performed and evaluated by trained pathologists at the respective transplantation center to determine the proportion of inflammation, fibrosis, and chronic rejection among recipients.

2.8. Statistical Methods

2.8.1. Descriptive and Comparative Statistics

Non‐normally distributed data was reported as median with interquartile range (IQR), categorical data as frequency counts and percentages of subjects within each category. Non‐normally distributed data was compared using the Chi‐squared/Fisher's Exact test.

In paired biopsy data, we estimated the absolute risk difference and risk ratio for subsequent fibrosis according to fibrosis status at year one. The Exact McNemar's test was used to evaluate the association between one‐year inflammation and subsequent fibrosis development, as well as to assess fibrosis progression from transplantation to 1 year after transplantation.

2.8.2. Cox Proportional Hazard Analysis

We computed univariable and multivariable Cox proportional hazard models to assess the association between inflammation, fibrosis, relevant risk factors with outcomes: graft loss and mortality (R‐package survival). All independent variables included in the models were assessed in univariable models and expressed as hazard ratios (HR) with 95% CI. Variables with a p‐value ≤ 0.10 were included in the multivariable model and expressed as adjusted hazard ratio (aHR) with 95% CI. The variables included inflammation, fibrosis, age at transplantation (per decade increase), donor age (per decade increase), recipient sex, recipient/donor sex‐mismatch, history of re‐transplantation, and reason for transplantation. The variable country was included in the Cox model using the strata() function in R to control for its effect without estimating a hazard ratio.

All models were tested for proportionality, linearity, and interaction assumptions. In the model with graft loss as the outcome, “history of re‐transplantation” violated the proportionality assumption and exhibited perfect data separation. Therefore, it was excluded from the analysis.

An interaction between inflammation and fibrosis was tested for both graft loss and mortality; however, the number of events was too low to allow an estimation in both analyses. Although the inflammation+/fibrosis+ group had a higher proportion of graft loss events (8%), two subgroups had no events at all. Therefore, we were not able to investigate the interaction between inflammation and fibrosis.

During the preparation of this work, the authors used OpenAI's text generation models, generative pre‐trained transformers (GPT) 5.0 and 5.2. The model was used solely to improve 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 published article.

Data was 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.

3. Results

3.1. Overview of Cohort

The study cohort comprised 446 liver transplant recipients who underwent 583 protocol biopsies during a median follow‐up of 5 years (IQR: 3.6–6.1) (Table 1). Among the entire cohort, 414 recipients (93%) had a one‐year protocol biopsy, while 130 recipients (29%) had a biopsy performed at a later time point (≥ 2 years). Paired biopsy data (at 1 year and ≥ 2 years) were available for 96 recipients.

TABLE 1.

Characteristics of liver transplant recipients.

	Total n = 446 n (%)	Fibrosis n = 50 a n (%)	Without fibrosis n = 363 a n (%)	p
Sex, female	180 (41)	22 (44)	142 (39)	0.51
Sex‐mismatch	143 (32)	18 (36)	115 (32)	0.54
Recipient age at transplantation, years, median (IQR)	55 (45–63)	56 (45–62)	55 (45–63)	0.92
< 40	75 (17)	9 (18)	60 (17)	0.91
40–49	78 (18)	8 (16)	66 (18)
50–59	131 (29)	14 (28)	110 (30)
60–69	141 (32)	16 (32)	113 (31)
> 70	21 (5)	3 (6)	14 (4)
Donor age, years, median (IQR)	58 (44–68)	61 (50–68)	58 (43–68)	0.49
Country				0.01
Denmark	93 (21)	10 (20)	83 (23)
Finland	95 (21)	4 (8)	91 (25)
Sweden	258 (58)	36 (72)	189 (52)
MELD score, median (IQR)	13 (9–18)	13 (9–17)	12 (9–18)	0.80
< 20	290 (84)	41 (89)	223 (84)	0.33
> 20	56 (16)	5 (11)	44 (16)
Missing	100 (22)	4 (8)	96 (26)
ABO‐blood type match				0.25
Identical	410 (93)	46 (94)	334 (93)
Compatible	28 (6)	2 (4)	24 (7)
Mismatch	3 (1)	1 (2)	1 (0)
Cold ischemia time, hours, median (IQR)	6 (5–8)	7 (6–8)	6 (5–8)	< 0.01
Missing	26 (6)	1 (2)	16 (4)
History of re‐transplantation	29 (7)	3 (6)	22 (6)
Reason for transplantation				0.93
Autoimmune liver disease (any cause)	179 (40)	19 (38)	145 (40)
Primary sclerosing cholangitis	125 (28)	16 (32)	98 (27)
Primary biliary cirrhosis	33 (7)	1 (2)	28 (8)
Cirrhosis (any cause)	209 (47)	23 (47)	169 (47)
Cirrhosis (alcohol)	106 (24)	14 (29)	84 (23)
Cirrhosis (viral)	27 (6)	2 (4)	22 (6)
Cirrhosis (autoimmune)	40 (9)	4 (8)	34 (9)
Hepatocellular carcinoma	89 (20)	7 (14)	75 (21)
Acute liver failure	6 (1)	1 (2)	5 (1)
Hepatitis C	23 (5)	2 (4)	18 (5)
Metabolic liver disease	46 (10)	5 (10)	40 (11)
Other	39 (9)	6 (12)	33 (9)
Immunosuppression at the time of protocol biopsy				0.34
Tacrolimus	336 (75)	40 (89)	265 (83)
Ciclosporin	53 (12)	4 (9)	49 (15)
Everolimus	7 (3)	1 (2)	4 (1)
Missing	50 (11)	5 (10)	45 (12)
Through level, microgram/L, median (IQR)
Tacrolimus	6 (5–7)	6 (5–8)	6 (5–7)	0.20
Ciclosporin	143 (122–176)	142 (121–176)	149 (137–215)	0.39
Everolimus	6 (5–6)	—	—

Note: The first column includes all liver transplant recipients, the two latter columns include recipients with available fibrosis data. More than one reason for transplantation can apply per recipient. Percentages were calculated on available data for each variable. Missing variables were indicated when missing in > 5% of total cases.

^a Fibrosis was assessed in the first available protocol biopsy for each recipient. Fibrosis data were missing for 33 recipients, resulting in 50 recipients with fibrosis and 363 without.

Six recipients (1.3%) experienced graft loss during follow‐up, and 25 recipients (6%) died. Graft loss occurred in 3 of 50 recipients with fibrosis and 3 of 363 recipients without fibrosis at the one‐year protocol biopsy. Of the six recipients with graft loss, four were due to recurrence of the primary liver disease (two with primary sclerosing cholangitis, one with HCV, subsequent cirrhosis and HCC in the explant, and one with recurrence of cholestatic disease in the form of strictures), one with a hepatic arterial thrombosis, and another with liver complications in relation to a bacterial infection. The recipients with HCC and bacterial infection both subsequently died.

3.2. Recipient Characteristics

Characteristics of the entire cohort and recipients with and without fibrosis are presented in Table 1. Among the entire cohort (n = 446), most were male (60%), from Sweden (60%), with a median age of 55 and a median donor age of 58. Five percent of recipients had a history of re‐transplantation. The most common reasons for transplantation were cirrhosis of any cause (47%), autoimmune liver disease (40%), and hepatocellular carcinoma (20%).

3.3. Proportion and Severity of Inflammation, Fibrosis, and Chronic Rejection One Year After Transplantation

In one‐year protocol biopsies, 18% of recipients had inflammation, 12% had fibrosis, 29% had either inflammation or fibrosis, and 1% had chronic rejection. METAVIR scores and inflammation categories in the one‐year biopsies are presented in Figure 1. Most recipients (82%) had none or mild inflammation, 15% had moderate inflammation, and 3% had severe inflammation. Similarly, the majority (88%) of recipients had none or mild fibrosis (F1), 11% moderate fibrosis (F2), and 1% severe fibrosis (F3) one year after transplantation. No recipients had cirrhosis (F4) in the one‐year protocol biopsies. Additionally, moderate–severe inflammation was significantly associated with a higher proportion of advanced METAVIR scores (p < 0.01) (Figure 2).

FIGURE 1.

Proportion of inflammation and fibrosis score among liver transplant recipients one year after transplantation. (A) Proportion of recipients within each inflammation category, ranging from 0 to 3. (B) Proportion of METAVIR score within each category, ranging from METAVIR score F0–F4.

FIGURE 2.

Proportion of METAVIR Score by Inflammation category in protocol biopsies one year after transplantation. The proportion of recipients with higher METAVIR scores significantly increased with inflammation severity, p < 0.01. The number of recipients within each METAVIR score is shown in the bars. p‐value was calculated using Fisher's Exact test.

3.4. Fibrosis Progression

Out of 446 recipients, 96 had paired biopsy data from one year and more than 2 years after transplantation. Among the 96 recipients with paired biopsy data, the subsequent biopsy was performed at year 2 after transplantation in 26 recipients, year 3 in 52 recipients, and year 4 in 18 recipients. Overall, the recipients were comparable to the remaining cohort in terms of age and sex. While there were more recipients transplanted due to PSC (39% vs. 25%) and fewer due to cirrhosis (36% vs. 50%) in the paired data.

Of all recipients with paired data, 89 had no fibrosis, and eight had fibrosis at year one. Among the 89 recipients without fibrosis at year one, 11 developed fibrosis (10 with METAVIR F2 and one with F4) in the subsequent biopsy, while 78 stayed fibrosis‐free. Of the eight recipients with fibrosis at year one, three showed regression to no fibrosis, and five continued to have fibrosis, all remaining with METAVIR F2. The risk of subsequent fibrosis was 62.5% (5 of 8) among recipients with fibrosis at year 1 compared with 12.4% (11 of 89) among those without, corresponding to an absolute risk difference of 50.1% (95% CI 17.5–77.9) and a risk ratio of 5.1 (2.3–10.9).

In the same subset of recipients with paired biopsy data, we furthermore analyzed whether inflammation at year one was associated with subsequent fibrosis development. Among all recipients, 74 recipients had no inflammation, and 18 had inflammation at year one. Of the 74 recipients without inflammation at year one, 13 (18%) developed fibrosis in the subsequent biopsy, and 61 (82%) remained fibrosis‐free. Of the 18 recipients with inflammation at year one, none developed fibrosis in the subsequent biopsy. The association between inflammation at year one and subsequent fibrosis development was not statistically significant (p = 0.473).

3.5. Fibrosis at the Time of Transplantation Versus One Year After Transplantation

In a subset of 166 recipients with available biopsies from the time of transplantation, we assessed fibrosis progression between the time of transplantation and one year after. The subset of recipients was comparable to the overall cohort with respect to sex (p = 0.16), age (p = 0.12), and the most common reasons for transplantation (p > 0.05). However, donor age was lower in the subset compared to the overall cohort (p < 0.01).

At one year after transplantation, 12 recipients (7%) had fibrosis and 154 (93%) had no fibrosis. Among those with fibrosis at one year (all METAVIR F2), 11 (92%) had no fibrosis at transplantation, and 1 (8%) had fibrosis. In this case, the fibrosis was portal, periportal, and pericellular (METAVIR F1–F2) and was described as progressed compared with the time‐zero biopsy. Of the 154 recipients without fibrosis at 1 year, none (0%) had fibrosis at transplantation, while 11 (7%) had mild fibrosis (METAVIR F1) that regressed or remained stable by year one. Overall, more recipients developed fibrosis between transplantation and 1 year than regressed (p < 0.01).

3.6. Inflammation and Fibrosis as Risk Factors of Graft Loss

Univariable and multivariable Cox regression analyses of risk factors for graft loss are presented in Table 2.

TABLE 2.

Cox regression analysis to assess risk factors of graft loss in liver transplant recipients.

	Graft loss
	Univariable		Multivariable
	HR (95% CI)	p	aHR (95% CI)	p
Inflammation	1.44 (0.15–13.87)	0.75
Fibrosis	7.20 (1.45–35.67)	0.02*	6.72 (1.36–33.29)	0.02*
Recipient age, per decade	1.24 (0.62–2.49)	0.55
Donor age, per decade	2.28 (0.98–5.29)	0.06	2.34 (0.96–5.69)	0.06
Sex, male versus female	0.68 (0.14–3.34)	0.63
Sex‐mismatch	1.03 (0.19–5.64)	0.97
Cirrhosis (any cause)	1.16 (0.23–5.75)	0.86
Autoimmune liver disease	0.76 (0.14–4.14)	0.75
Hepatocellular carcinoma	2.04 (0.37–11.14)	0.41
Hepatitis C	3.50 (0.41–29.93)	0.25

Note: Variables with a p‐value < 0.1 were included in the multivariable analyses. History of re‐transplantation was excluded from the analyses due to violation of the proportionality assumption and perfect data separation. When adjusting for the transplantation center in the multivariable model with the strata function(), fibrosis and donor age remains significantly associated with graft loss: fibrosis: 12.58 (1.53–103.08), p = 0.02, donor age: 3.33 (1.12–9.92), p = 0.03. *p‐values < 0.05 were considered significant.

In both the univariable and multivariable analyses, recipients with fibrosis had a significantly increased risk of graft loss, with an HR of 7.20 (95% CI: 1.45–35.67) and aHR of 6.72 (1.36–33.29), respectively. In the univariable analysis, increasing donor age was associated with an increased risk of graft loss, with an HR of 2.34 (0.96–5.69). However, this was not significant in the multivariable model, p = 0.06. Inflammation was not statistically associated with graft loss in the univariable analysis, HR of 1.44 (0.15–13.87).

Age, sex, recipient/donor sex‐mismatch, cirrhosis, autoimmune liver disease, and HCC as a reason for transplantation were not significantly associated with an increased risk of graft loss.

3.7. Inflammation and Fibrosis as Risk Factors of Mortality

Univariable and multivariable Cox regression analyses of risk factors for mortality are presented in Table 3.

TABLE 3.

Cox regression analysis to assess risk factors of all‐cause mortality in liver transplant recipients.

	All‐cause mortality
	Univariable		Multivariable
	HR (95% CI)	p	aHR (95% CI)	p
Inflammation	0.56 (0.13–2.43)	0.44
Fibrosis	0.95 (0.29–3.19)	0.93
Recipient age, per decade	1.84 (1.22–2.79)	< 0.01*	1.86 (1.15–3.03)	0.01*
Donor age, per decade	1.10 (0.86–1.40)	0.46
Sex, male versus female	2.27 (0.91–5.65)	0.08	1.79 (0.70–4.55)	0.22
Sex‐mismatch	1.36 (0.62–2.99)	0.45
History of re‐transplantation	2.32 (0.70–7.75)	0.17
Cirrhosis (any cause)	4.03 (1.61–10.03)	< 0.01*	2.89 (1.13–7.36)	0.03*
Autoimmune liver disease	0.66 (0.29–1.53)	0.34
Hepatocellular carcinoma	1.89 (0.82–4.36)	0.13
Hepatitis C	0.74 (0.10–5.45)	0.77

Note: Variables with a p‐value < 0.1 were included in the multivariable analyses. *p‐values < 0.05 were considered significant.

In the univariable analysis, inflammation and fibrosis were not associated with an increased mortality risk, with HRs of 0.56 (95% CI: 0.13–2.43) and 0.95 (0.29–3.19), respectively. However, the recipient's age and cirrhosis (any etiology) were associated with increased mortality risk in both the univariable and multivariable analyses, aHR 1.78 (1.13–2.82) and 3.31 (1.32–8.32), respectively.

The donors' age, recipients' sex, recipient/donor sex mismatch, autoimmune liver disease, and HCC as a reason for transplantation were not significantly associated with an increased mortality risk.

4. Discussion

In this multicenter cohort study, liver transplant recipients with fibrosis one year after transplantation had a sixfold higher risk of graft loss compared to recipients without fibrosis. Most recipients with graft loss were due to recurrence of the primary liver disease. Moreover, fibrosis was common in our cohort, as 12% had significant graft fibrosis (METAVIR score ≥ F2) in their protocol biopsy one year after transplantation. In paired biopsy data, recipients with fibrosis at one year had a fivefold higher risk of fibrosis on the subsequent biopsy. In the paired analysis, we did not find a significant association between one‐year inflammation and subsequent fibrosis. Additionally, inflammation and fibrosis were not statistically associated with an increased mortality risk in our cohort.

Abnormal histological findings are common in liver transplant recipients, even when liver function tests are normal [3], and the prevalence increases over time after transplantation [23]. More importantly, 10%–20% of cases with histological changes are silent and categorized as moderate‐to‐severe [4, 24]. These findings support the use of protocol biopsies as they identify recipients with significant histological changes, thereby guiding and justifying changes in treatment.

Fibrosis is a well‐known risk factor for recurrent cirrhosis, subsequent graft loss, and mortality in recipients transplanted due to hepatitis C virus‐related cirrhosis [10, 25]. However, in the postdirect‐acting antiviral (DAA) era, investigations have shown that the burden of HCV‐related graft fibrosis is decreasing, and survival rates have increased [12]. In the current study, liver transplant recipients with fibrosis had a higher risk of graft loss, despite a low HCV prevalence (< 6%). Among the eight recipients with fibrosis at the one‐year protocol biopsy, one recipient had a pretransplant diagnosis of chronic hepatitis C. Recurrent hepatitis C may therefore have contributed to fibrosis development in this recipient; however, the low number of cases did not allow assessment of the association between recurrent hepatitis C and early graft fibrosis. Furthermore, fibrosis was prevalent in 12% of recipients one year after transplantation. Few studies have investigated the burden and long‐term outcomes of early fibrosis in an adult transplant population with a low prevalence of HCV infection [2, 3, 26]. However, Liukkonen et al. conducted a study including 875 HCV‐negative recipients who underwent liver transplantation between 2000 and 2019 at the Helsinki University Hospital, Finland. The study had a comparable follow‐up time to the present study of 5 years [4]. Liukkonen et al. found that fibrosis, defined as a METAVIR score ≥ F2 compared to F0 in the first available protocol biopsy, was not associated with increased risk of graft loss, with an HR of 1.91 (95% CI: 0.67–5.40). The inability to reach statistical significance may be attributed to the low proportion of fibrosis in their cohort compared to the present study (3% vs. 12%, respectively), as well as the categorization of fibrosis into stages (F0, F1, F2–F4), which may have further diminished statistical power.

In our study, we also found that donor age was a risk factor for graft loss. This association is well recognized from previous extensive research [7, 27]. However, in a systematic review and meta‐analysis, the authors describe how the added risk of increasing donor age can be balanced by reducing other risk factors such as ischemia time and MELD‐score [28].

A strength in the study by Liukkonen et al. was the number of follow‐up protocol biopsies 5 and 10 years after transplantation. The authors demonstrated that graft inflammation and mild fibrosis were predictors of subsequent fibrosis progression. Our findings align with this, as we observed that recipients with fibrosis at year one had a 50% higher absolute probability and a 5‐fold risk of fibrosis in the subsequent biopsy compared to those without fibrosis. However, we did not find that inflammation at 1 year was associated with later graft fibrosis. Although the paired cohort was comparable in age and sex to both our overall cohort and that of Liukkonen et al., there were differences in the reason for transplantation. Specifically, more recipients were transplanted due to PSC and fewer due to cirrhosis (including alcoholic cirrhosis). This imbalance and small sample size may partly explain our inability to reproduce the findings of Liukkonen et al.

In a subset of 166 recipients with available time‐zero biopsies, we assessed whether fibrosis observed at one year after transplantation was already present at the time of transplantation. The subgroup was comparable to the overall cohort, except for donor age, which was lower in the subset. In 11 of 12 fibrosis cases (92%) detected at one year, the fibrosis was de novo. In the remaining case (8%), fibrosis was already present at transplantation but had progressed by year one. Among the 154 recipients without fibrosis at 1 year, 11 (7%) had mild fibrosis at transplantation, all of which had regressed or remained stable by the time of the one‐year protocol biopsy. Overall, we found that significantly more recipients developed fibrosis between transplantation and one year after transplantation, than regressed. Whether the lower donor age in this subset may have contributed to the lower degree of fibrosis observed at transplantation remains unclear, as the influence of donor age on fibrosis mainly has been studied in HCV‐infected liver transplant recipients and in the posttransplant setting rather than at the time of transplantation [29, 30].

In our study, graft fibrosis was not associated with all‐cause mortality. This finding aligns with data previously published from the Nordic Liver Transplant registry (1985–2009); in a cohort of 3299 liver transplant recipients, the authors reported that death due to liver‐related causes accounted for only 9% of all mortality cases. Whereas malignancies and cardiovascular cases were the primary causes [31]. We, however, did find that the recipient's age and cirrhosis (any etiology) as a reason for transplantation were significantly associated with increased mortality risk. This aligns with existing evidence, as cirrhosis is a recognized marker of a high comorbidity burden and is associated with all‐cause mortality risk.

Graft inflammation was not associated with graft loss or mortality in the current study. Inflammation of the graft is an inherent part of the pathogenesis behind many conditions and diseases that occur after transplantation, thereby serving as a driver of long‐term graft fibrosis, cirrhosis, and graft loss [5, 32, 33]. Pathophysiologically, fibrosis is the result of continuous wound healing due to repeated liver injury. This probably explains why inflammation itself is not directly associated with graft loss, but instead serves as a mediator of fibrosis; however, the limited follow‐up duration in this study may also have limited our ability to capture this association.

In this study, we found that fibrosis was associated with graft loss, and recipients with paired data showed that fibrosis progressed rather than regressed in the subsequent biopsy. Protocol biopsies are valuable in detecting subclinical inflammation and early fibrosis. While less invasive alternatives exist, they have limited sensitivity for early fibrosis and the ability to distinguish between the underlying causes of inflammation and fibrosis [1]. The findings of the current study support a potential role for protocol biopsies to monitor recipients at increased risk and to identify those who may benefit from immunosuppressive adjustment, as inflammation and early‐stage fibrosis may be reversible. However, given the observational design of this study, our findings demonstrate an association between one‐year fibrosis and subsequent graft loss but do not establish causality. Additionally, the frequency and timing of protocol biopsies to identify fibrosis at a reversible stage remain important research topics.

A strength of the current study is its large, multicenter design, where more than 90% of the cohort had an available protocol biopsy within the first year after transplantation. Additionally, we were able to investigate time‐zero biopsies; although this was in a subset of recipients, the data support the significance of the findings of the one‐year protocol biopsies. We would also like to address the limitations of the study. First, protocol biopsies were not assessed by the same pathologist across transplantation centers. This may have introduced interobserver variability, potentially affecting the study's generalizability. Second, we had very few cases of graft loss, limiting our ability to adjust for potential confounders. As graft loss is a multifactorial outcome, residual confounding due to factors not accounted for in the analysis cannot be excluded. Third, only 96 recipients had paired biopsy data, which limited the analysis of long‐term histological progression to a descriptive analysis. Fourth, acute rejection is a known risk factor for fibrosis; data on acute rejection were missing in more than 50% of patients, and we were therefore unable to adjust for its potential confounding effect. Finally, adults who underwent liver transplantation during infancy or childhood were not included in the present study. Therefore, our findings cannot be generalized to this patient population.

To conclude, we found that graft fibrosis one year after liver transplantation was significantly associated with graft loss. Protocol biopsies remain valuable for detecting inflammation and early fibrosis, guiding clinicians in the management of their patients. However, further studies are needed to determine whether early interventions targeting fibrosis can improve long‐term graft outcomes and to clarify whether the observed association between fibrosis and graft loss reflects a causal relationship.

Funding

This work was supported by the Danish Cancer Society (grant number R306‐A18169).

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 the Novo Nordic Foundation, Independent Research Foundation Denmark, Rigshospitalet Research Foundation, EU Horizont, 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 the Danish Cancer Society, Independent Research Fund Denmark, Svend Andersen Fonden, ctDNA Center Denmark, Research Council Copenhagen University Hospital Rigshospitalet, and the advisory board for AstraZeneca. The other authors declare no conflicts of interest.

Data Availability Statement

Due to their sensitive nature, the data generated and analyzed during the current study are not publicly available.