Impact of Extended Criteria Donor Transplants on Infections in Kidney and Liver Transplant Recipients and Marginal Donor Definition Criteria: A Systematic Review

Abstract

Background

Extended criteria donors (ECDs) have become an important source of organs, but their association with infection risk remains unclear. This systematic review assessed the impact of ECD versus standard criteria donors (SCDs) on infection rates in kidney and liver transplant recipients and analyzed the variability of ECD definitions.

Methods

Three different literature databases were searched up to June 2024 for randomized controlled trials or observational studies comparing infection rates between ECD and SCD recipients. Studies without quantitative data or comparator groups were excluded. Outcomes were measured at 30–180 days and 1, 3, and 5 years posttransplant, evaluating both bacterial and viral infections.

Results

A total of 11 119 articles were screened and 131 studies included. At 30 days, bacterial infections occurred at similar rates between ECD and SCD recipients (odds ratio [OR] 1.15, 95% confidence interval [CI] .86–1.53, P = .36). However, liver transplant recipients from ECDs had a higher infection risk (OR 1.73, 95% CI 1.12–2.68, P = .01). Definitions of ECD varied notably, especially in liver transplantation, while kidney transplantation criteria were more consistent.

Conclusions

Current evidence does not definitively establish a link between ECD and infection risk, highlighting the need for standardized donor definitions and routine reporting of infections as ECD outcomes.

Keypoints.

Extended criteria donor (ECD) use has been increasing in recent years. Few studies have investigated the impact of ECD on the infection rate after transplantation. Overall, no increased risk emerged except in the subgroup analysis in liver transplant recipients. In addition, a marked variability in ECD definitions was noticed underscoring the need for standardized criteria and infection reporting.

The widening gap between the growing demand for organs and their limited availability poses a central challenge in modern transplantation medicine. Alongside significant advances in surgical techniques, immunosuppressive therapies, and posttransplant care, the indications for organ transplantation have evolved from being a last-resort intervention to embracing earlier treatment strategies [1–4]. Early or preemptive transplantation, particularly in liver and kidney recipients, has been associated with improved survival and enhanced quality of life by alleviating the physical and psychological burdens of prolonged treatments such as dialysis [5–8].

Despite expanding indications and earlier referral, organ availability in many countries remains constrained by reliance on deceased donor. As a result, organ shortage continues to be a major limiting factor for access to transplantation. In 2023, the United States had more than 144 000 kidney and over 24 000 liver transplant candidates. Despite record numbers of transplants were performed (28 142 kidney and 10 659 liver), many patients never reached surgery. Indeed, 4002 kidney candidates died and 4855 were delisted as too ill, while among liver candidates, 6.4% died and 6.6% were removed for clinical deterioration. Similar trends were reported in Italy, with 2023 waitlist deaths of 2.1% (kidney) and 6.5% (liver) and a further 10%–25% delisted due to worsening condition [9, 10].

Within this context, the use of marginal donors, or extended criteria donors (ECDs), has emerged as a necessary response to organ shortages. In the transplantation literature, however, ECD status is defined in a heterogeneous manner. For kidney transplantation, ECD status commonly reflects combinations of advanced donor age (eg, ≥60–65 years or 50–59 years with comorbidities), history of hypertension or diabetes, elevated terminal serum creatinine, prolonged cold ischemia time (>5–12 hours), or donation after circulatory death (DCD). These factors have been also combined in a risk stratification score, namely the Kidney Donor Profile Index (KDPI). For liver transplantation, frequently used ECD criteria include older donor age, macrosteatosis, prolonged cold ischemia time, high-risk viral serologies, and DCD [3, 4].

Although ECDs carry a slightly increased risk of complications—such as delayed graft function [11–14]—clinical evidence suggests that their use can significantly reduce waiting list mortality [15, 16]. Indeed, the use of ECDs has become frequent in countries with well-organized organ procurement systems [17].

The use of these organs may be associated with an increased risk of posttransplant infections. However, this risk and its management have been well studied only for donors recognized as being at high risk of transmissible infections (such as people who inject drugs), those colonized with specific microorganisms (eg, donors colonized with carbapenem-resistant Enterobacterales), or donors with a documented infection at the time of organ retrieval (eg, bacteremia). There are no data on the indirect impact of ECDs—such as through increased risk of delayed graft function and other complications—on the occurrence of posttransplant infections, particularly in the perioperative period.

Finally, the lack of consensus on the definition and acceptance criteria for ECDs complicates direct comparison of outcomes across transplant centers and studies, underscoring the need for more standardized classification [18].

Given these challenges, this systematic review aims to evaluate the incidence of posttransplant infections in recipients of ECD organs compared with those receiving organs from standard criteria donors (SCDs) and to examine the variability in ECD definitions across studies.

METHODS

Study Design

We conducted a systematic review to investigate the impact of ECD utilization on the occurrence of posttransplant infections in liver and kidney transplant recipients, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [19]. The review was prospectively registered in the PROSPERO database (registration ID: CRD42024506773).

The study objectives were defined using the PECO (Population, Exposure, Comparator, Outcome) framework: the population consisted of adult patients undergoing liver and/or kidney transplantation; the exposure was defined as transplantation using donors with extended criteria as defined in the included studies; the comparator group included recipients of organs from SCDs; and the primary outcome was the occurrence of posttransplant infections as reported in the studies, diagnosed from transplantation to 30 days.

For the purpose of this review, we considered clinically relevant posttransplant infections as defined in the individual studies and grouped them into bacterial and viral categories for analysis. Where specified, bacterial infections most commonly included bloodstream infection, pneumonia, surgical-site or wound infection, intra-abdominal or biliary infection and urinary tract infection; asymptomatic bacteriuria was not included. Viral infections predominantly involved Cytomegalovirus infection or disease, but some studies also reported other herpesvirus infections (eg, herpes simplex virus [HSV], varicella-zoster virus [VZV], Epstein-Barr virus [EBV]).

A secondary aim of our study was to investigate the variability in extended criteria donor definitions used across studies.

Search Strategy

We performed an electronic search in the databases PubMed/MEDLINE, Scopus, and Embase to identify studies evaluating the use of marginal donors and their impact on infection rates and other relevant clinical outcomes in kidney and liver transplantation.

To ensure the inclusion of all relevant studies, the search incorporated both Medical Subject Headings (MeSH) terms and free-text keywords. Specifically, the following terms were included: “marginal donor,” “marginal kidney donor,” “marginal liver donor,” “marginal donor graft,” “marginal donor organ,” “suboptimal donor,” “suboptimal kidney donor,” “suboptimal liver donor,” “suboptimal donor graft,” “suboptimal donor organ,” “extended criteria donor,” “expanded criteria donor,” “expanded criteria donor organs,” “extended criteria donor organs,” “expanded criteria donor graft,” “extended criteria donor graft,” “marginality criteria donor,” “marginal DCD” “deceased marginal donor,” “donation after cardiac death,” “donor after circulatory death,” “expanding the donor pool,” “marginal organs,” and “marginal graft” AND “infection” [MeSH Terms], “risk factor for infection,” “infection risk,” “donor-derived infection,” and “risk stratification for infection.”

Given the limited number of studies specifically addressing infections in this context, the search was broadened to include as outcomes patient survival and graft survival. Thus, the following terms were also incorporated: “death” [MeSH Terms], “mortality” [MeSH Terms], “survival,” “death rate,” “graft loss,” “kidney graft loss,” “liver graft loss,” “primary liver graft non-function,” “primary kidney graft non-function,” “primary graft non-function,” “primary liver graft dysfunction,” “primary kidney graft dysfunction,” “loss of kidney graft function,” “loss of liver graft function,” “kidney graft failure,” “liver graft failure,” “graft failure,” “graft dysfunction,” “primary graft non-function,” “primary graft dysfunction,” and “loss of graft function.”

The search was restricted to studies published in English, and duplicate records were removed. Study selection was performed in 3 phases: initial screening by title, followed by abstract screening, and finally full-text review. The entire selection process was managed using Rayyan, a web-based platform designed to facilitate systematic reviews. Any disagreements during study selection process were resolved through discussion among the authors.

Data Extraction

Data were extracted independently by 2 pairs of authors (M. C. C., M. C. and C. G., L. C.) using a standardized extraction form; any discrepancies were resolved by consensus or, if needed, after discussion with a fifth reviewer (C. B.). For each study, we collected information on study characteristics (first author, year of publication, country, study design, sample size, and main population characteristics), the definition of marginal or extended criteria donors (terminology used and specific criteria applied), graft characteristics (including donor age, relevant comorbidities such as hypertension or diabetes, cause of death, and DCD status), and clinical outcomes (posttransplant infections, graft survival, and patient survival). In cases of unclear or incomplete information, we contacted the corresponding authors for clarification or additional data.

Risk of Bias Assessment

The quality of the studies was independently assessed by 4 reviewers using the Cochrane Risk of Bias tool (RoB 2.0) for randomized controlled trials [20] and Risk of Bias in Nonrandomized Studies of Interventions tool for observational studies [21] (Supplementary Table 1).

Studies were categorized as having low, moderate, or high risk of bias. Any discrepancy was resolved through discussion with a fifth reviewer (C. B.).

Data Synthesis

Quantitative analysis for each outcome was performed by pooling effect estimates for ECDs against SCDs.

Treatment effects were calculated as odds ratios (ORs), with 95% confidence interval (CI) for dichotomous data, using a random-effects model. Statistical significance was assessed using the Z-test for dichotomous data, with a 2-sided P-value <.05.

Heterogeneity among studies was assessed using the χ² test, with a P-value <.10 indicating substantial heterogeneity, and the I² statistic, where values >50% suggested substantial heterogeneity. Subgroup analyses were performed according to the type of solid organ transplantation for primary and secondary outcomes. Sensitivity analyses were also conducted according to risk of bias, the use of DCD as an ECD criterion, and the publication period of the study (we divided studies published before or after 2010), in order to assess the robustness of the outcomes. Publication bias was evaluated through visual inspection of funnel plots and Egger's test. All statistical analyses were conducted using MedCalc for Windows (version 19.6.1, MedCalc Software Ltd, Ostend, Belgium).

In addition to the meta-analysis, we performed a descriptive analysis to summarize the variability in marginal donor/ECD definitions across the included studies and to provide context for their clinical implications.

RESULTS

The search yielded a total of 11 119 records. After removing 3626 duplicate records, 7493 unique records were screened. In the title screening phase, 6041 articles were excluded, followed by the exclusion of 1089 articles after abstract screening. Finally, 225 full-text articles were excluded because they did not meet the inclusion criteria. After this selection process, a total of 131 studies were included in this systematic review. The study selection process is summarized in the PRISMA flow diagram (Figure 1).

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.

A total of 131 observational studies were included in this systematic review (Supplementary Table 2). Among these, 14 studies (10.7%) were prospective and 117 (89.3%) were retrospective. The studies were conducted across different geographical regions, with 39 studies (29.8%) originating from the United States and Canada, 61 studies (46.6%) from Europe, 24 studies (18.3%) from Asia, 5 studies (3.8%) from South America, and 2 studies (1.5%) from Australia. The publication period spanned from 2001 to 2024, reflecting advancements in the field of transplantation over 2 decades. Regarding the types of transplants analyzed, the majority focused on kidney (70 studies, 53.4%), followed by liver transplants (56 studies, 42.7%) and combined kidney–liver transplants (5 studies, 3.9%).

Overall, infection outcomes were reported in 16 studies for 30-day bacterial infections (1238 ECD vs 1869 SCD recipients) and in 10 studies for 30-day viral infections (1219 vs 5116 recipients). At 1 year, bacterial and viral infections were reported in 3 and 6 studies, respectively (205 vs 998 and 355 vs 1099 recipients), while only 2 studies reported 3-year viral infections (82 vs 148 recipients). In contrast, survival outcomes were much more frequently reported, with 84 studies contributing to 1-year patient survival and 81 to 1-year graft survival (19 590 vs 90 849 and 11 344 vs 104 945 recipients, respectively; see Supplementary Table 3).

At 30 days posttransplant, bacterial infections occurred at similar rates between ECD and SCD recipients, with an OR of 1.15 (16 studies; 1238 vs 1869 recipients, OR 1.15, 95% CI .86–1.53, P = .36) and moderate heterogeneity among studies (I² = 36.4%, P = .07). At 1 year, bacterial infections appeared more frequent in the ECD group (3 studies; OR 1.75, 95% CI .72–4.25; P = .22), although the difference was not statistically significant, and the analysis showed high heterogeneity (I² = 82.7%). No data were available for bacterial infections at 180 days, 3 years, or 5 years.

For viral infections, no significant difference was observed at 30 days (10 studies; OR 1.44, 95% CI .77–2.68, P = .25), though heterogeneity was substantial (I² = 77.4%). At 1 year, there was a trend toward increased viral infections in ECD recipients (6 studies; OR 1.58; 95% CI .93–2.67, P = .09; I² = 55.6). Limited data were available at 3 years (2 studies; OR 1.09, 95% CI .56–2.13, P = .81, I² = 0.0%), and no data were reported for viral infections at 180 days or 5 years. Results are summarized in Table 1 and Figure 2.

Table 1.

Bacterial and Viral Infections at Different Time Points

Outcome	Studies	No. of Patients (ECD vs SCD)	Odds Ratio (95% CI)	Heterogeneity (I2; P-Value)	Publication Bias (P-Value Egger's Test)
30 –d
Overall—bacterial infection	16	1238 versus 1869	1.15 (.86–1.53) P = .36	36.4% P = .07	.44
Liver—bacterial infection	5	238 versus 545	1.73 (1.12–2.68) P = .01	0.0% P = .47	.73
Kidney—bacterial Infection	11	1000 versus 1324	1.00 (.71–1.41) P = .99	39.8% P = .08	.29
Combined kidney–liver—bacterial Infection	0	NA	NA	NA	NA
Overall—viral infection	10	1219 versus 5116	1.44 (.77–2.68) P = .25	77.4% P < .001	.71
Liver—viral infection	3	722 versus 4576	1.85 (.87–3.92) P = .11	79.3% P < .01	.59
Kidney—viral infection	7	497 versus 540	1.25 (.44–3.56) P = .68	72.6% P < .01	.16
Combined kidney–liver—viral infection	0	NA	NA	NA	NA
1 y
Overall—bacterial infection	3	205 versus 998	1.75 (.72–4.25) P = .22	82.7% P = .003	.35
Liver—bacterial Infection	0	NA	NA	NA	NA
Kidney—bacterial Infection	3	205 versus 998	1.75 (.72–4.25) P = .22	82.7% P = .003	.35
Combined kidney–liver—bacterial infection	0	NA	NA	NA	NA
Overall—viral infection	6	355 versus 1099	1.58 (.93–2.67) P = .09	55.6% P = .05	.73
Liver—viral infection	4	202 versus 260	1.28 (.77–2.13) P = .35	0.0% P = .47	.87
Kidney—viral infection	2	153 versus 839	2.21 (.60–8.17) P = .23	88.1% P = .004	NA
Combined kidney–liver—viral infection	0	NA	NA	NA	NA

Abbreviations: ECD, extended criteria donor; NA, not available; SCD, standard criteria donor.

Figure 2.

Forest plot for bacterial infection at 30 d posttransplant.

In the subgroup of liver transplant recipients, ECD showed a significantly higher risk of bacterial infections at 30 days (5 studies; OR 1.73, 95% CI 1.12–2.68, P = .01, I² = 0.0%), whereas no significant difference in viral infections were observed at any time point. In contrast, no difference between ECD and SCD was found for either bacterial or viral infections at any of the evaluated time points in kidney transplant recipients. No data were available for either bacterial or viral infections in combined kidney–liver transplant recipients.

The impact of ECD versus SCD on patient and graft survival is reported in Supplementary Table 3.

After excluding studies with serious or critical risk of bias (13 of 16 studies reporting 30-day bacterial infections and 8 of 10 studies reporting 30-day viral infections), the direction and magnitude of the effect estimates remained essentially unchanged, and no significant differences between ECD and SCD were observed for the primary outcomes (Supplementary Table 4).

In the subgroup analysis on DCD defining ECD as unique criteria, the pooled effect estimates for posttransplant infections in the 2 groups were very similar to those of the main analysis (data shown in Supplementary Table 5). In the subgroups analysis according with the time period (<2010 and ≥2010), the pooled effect estimates for infection incidence in ECD versus SCD recipients were also similar (data shown in Supplementary Tables 6–7).

The quality of the studies is reported in Supplementary Table 1. A substantial proportion of studies showed a serious overall risk of bias, particularly in the domains of confounding bias and selection of reported results, whereas the classification of exposure was generally at low risk. Moderate risk of bias was observed mainly in the domains of missing data and outcome measurement, underscoring potential limitations in study design and data reporting.

The criteria used to define ECD varied substantially across studies. For liver transplantation (Figure 3), the following ECD parameters were considered: DCD (40/56, 71.4%), older donor (>60 years) (18/56, 32.1%), prolonged ischemia time (10/56, 17.9%), donor comorbidities (9/56, 16.1%), donor body mass index (BMI) (> 25 kg/m²) (3/56, 5.4%), hepatic steatosis (10/56, 17.9%), hepatitis C virus antibody positivity (7/56, 12.5%), hepatitis B virus core antibody positivity (4/56, 7.1%) and split liver grafts (4/56, 7.1%).

Figure 3.

Liver transplantation extended criteria donor (ECD) definitions.

For kidney transplantation (Figure 4), ECD parameters included advanced donor age (>60 years) (36/70, 51.4%), DCD (28/70, 40%), donor comorbidities (31/70, 44.3%), age stratified criteria (20/70, 28.6%), prolonged cold ischemia (4/70, 5.7%), donor BMI (> 25 kg/m²) (1/70, 1.4%), dual kidney transplantation (1/70, 1.4%), elevated serum creatinine levels (>1.5 mg/dL) (41/70, 58.6%), a history of hypertension (38/70, 54.3%), diabetes mellitus (5/70, 7.1%), high KDPI (>85%) (3/70, 4.3%), and ABO (A, B, and O blood group) incompatibility (1/70, 1.4%).

Figure 4.

Kidney transplantation extended criteria donor (ECD) definitions.

For combined kidney–liver transplants, DCD was the only ECD parameter considered in all the 5 studies included (100%).

DISCUSSION

This systematic review is, to our knowledge, the first study investigating the impact of ECD use on the posttransplant infection risk in kidney, liver, and combined kidney–liver recipients. Overall, similar rate of early bacterial infections in the pooled population of kidney and liver transplant recipients was observed. However, a significantly increased OR for early posttransplant bacterial infections was observed in the subgroup analysis of liver transplant recipients. In addition, we found marked heterogeneity in the criteria used to classify donors as ECD, particularly in liver transplantation. Our findings highlight the need to systematically assess infections in studies evaluating the use of ECD, in order to plan appropriate monitoring and management strategies, and the need of harmonized ECD criteria to ensure reproducibility across studies.

The higher risk of posttransplant infections observed in ECD recipients, particularly among liver transplant recipients, is consistent with previous reports, which highlight factors such as prolonged ischemia times, older donor age, and preexisting comorbidities as key contributors to increased infection susceptibility [1, 2, 22, 23]. While bacterial infections were not significantly different at 30 days posttransplant, their incidence tended to increase over time, and viral infections showed a similar, nonsignificant trend at 1 year. These findings align with prior studies demonstrating that ECD recipients may have a higher incidence of delayed graft function and compromised organ quality, often require intensified management, and are therefore more prone to both nosocomial and opportunistic infections [3–6]. The role of DCD in exacerbating ischemia-reperfusion injury and contributing to infectious risks is also noteworthy, as previously reported in both kidney and liver transplant populations [8, 11].

A key challenge identified in this review was the substantial heterogeneity in the definition of ECD across the included studies, particularly in liver transplantation. While kidney ECD criteria are relatively standardized—often defined by donor age >60 years, history of hypertension, or elevated serum creatinine—liver donor criteria remain inconsistent. Factors such as macrosteatosis, donor age, and cold ischemia time were variably reported, complicating cross-study comparisons and the synthesis of evidence [24].

This variability complicates comparisons across studies and highlights the need for an international consensus on ECD classification to facilitate more precise risk stratification and clinical decision-making. Emerging predictive models such as the KDPI [25] for kidney and the Donor Risk Index [26] for liver transplantation have been proposed to improve donor selection and optimize organ allocation strategies, and their broader implementation may contribute to better infection risk assessment in the context of marginal graft utilization. Future research should focus on refining these predictive models, optimizing perioperative management strategies, and exploring novel immunosuppressive protocols tailored to the specific risks associated with ECD transplantation. Although the use of marginal donors inevitably presents clinical challenges, ongoing advances in transplantation medicine, combined with more precise risk stratification and improved preservation techniques, may enhance long-term outcomes and expand the availability of life-saving organ transplantation a broader patient population [27, 28].

Given the higher rate of early infections after transplantation of ECD livers, risk mitigation may include enhanced early monitoring (with prompt work-up for fever, graft dysfunction, or organ failure) in recipients of high-risk grafts (eg, older donors, DCD, prolonged cold ischemia, macrosteatosis), perioperative optimization to limit ischemia and maintain hemodynamic stability, and—in selected cases—microbiological surveillance tailored to donor/graft features. At this stage, we deem that providing indication about enhanced antimicrobial use in patients receiving ECD organs is not supported by current evidence and could be dangerous in terms of ecological exposure. Thus, further data are needed to optimize the management of infectious risk in ECD organ recipients.

A major strength of this study lies in its systematic approach, adhering to PRISMA guidelines, which ensures methodological rigor and minimizes selection bias. In addition, the inclusion of a large number of studies from different geographical regions enhances the generalizability of the findings. By evaluating both bacterial and viral infections at multiple time points, this study provides a nuanced overview of the infectious risks associated with ECD organs, thereby helping clinicians make more informed decisions.

However, several limitations must be acknowledged. First, in this meta-analysis recipient characteristics were not considered. Given that recipient factors—such as age, comorbidities, and immune status—can significantly influence infection risk and overall transplant outcomes, future studies should aim to assess the interaction between donor and recipient characteristics to provide a more comprehensive framework for matched risk stratification. Second, the heterogeneity in ECD definition across studies complicates direct comparisons and may affect the robustness of pooled estimates. Third, residual confounding cannot be ruled out, as adjusted analyses were not performed in the vast majority of included studies. Among potential sources of bias that were not addressed, it is worth noting the use of advanced organ preservation techniques, such as hypothermic machine perfusion, which may mitigate early complications and improve long-term graft outcomes. Moreover, reporting of perfusion techniques is very limited and heterogeneous; consequently, we were unable to analyze perfusion modality as a potential modifier of infection risk and could only acknowledge this factor qualitatively.

Finally, although statistical methods were used to explore publication bias, the possibility of underreporting negative results remains a concern. Future research should focus on multicenter prospective studies with harmonized ECD definitions and a standardized approach to infection surveillance.

This systematic review provides a comprehensive analysis of the impact of ECD transplantation on posttransplant infections and explores the heterogeneity of extended criteria donor definition. The available data suggest that the risk of early posttransplant bacterial infections may be increased in ECD liver recipients, but not in ECD kidney recipients, compared with recipients of SCD grafts, although the variability in findings highlights the need for improved reporting and standardization in future studies. The marked heterogeneity in donor classification criteria across studies underscores the need for consensus in defining marginal donors. Establishing standardized donor classification systems would not only improve the comparability of future studies but also facilitate the development of tailored strategies to optimize outcomes.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial support. The work reported in this publication was funded by the Italian Ministry of Health, RC-2025-2791682.

Patient consent. This study is a review of the existing literature and did not involve the use of individual human data. Therefore, the requirement for patient consent is not applicable.

Data availability. This systematic review and meta-analysis is based exclusively on data extracted from previously published studies. All data underlying this article are available within the article and its Supplementary material.