Despite having one of the world’s highest organ donation rates per million inhabitants, Spain’s proportion of liver grafts discarded at procurement in donors after brain death (DBD) has gradually increased in the last 10 years from 25% to 30% [1]. Increasing donor age, obesity epidemic, high incidence of metabolic syndrome, and previous or current alcohol consumption have undoubtedly worsened the quality of grafts offered for evaluation. The utilization of marginal donors, also referred to as “donors with extended criteria” (ECD), has augmented the donor pool with acceptable outcomes [2]. However, this strategy might be challenging for transplantation teams who must evaluate a larger number of potential donors with corresponding increases of travel costs, human resources, and operating room facilities, knowing that these efforts will eventually be unnecessary in almost one-third of cases. Unreimbursed travel and time costs are also burdensome for potential recipients and their families.

Although prerecovery percutaneous biopsy to identify fibrosis, steatosis, or necrosis in marginal donors has shown promising outcomes [3,4], it is rarely performed in Spain owing to logistical issues and concerns about the safety of the procedure.

Ultrasonography (US) as part of the routine donor workup has shown a low positive predictive value to identify steatosis [5]. It might be challenging to perform and interpret, especially in obese donors. More sensitive abdominal cross-sectional imaging is performed only in trauma DBD or when a solid undefined nodule is found on US.

Currently, there is no objective gold standard to accept or discard a liver graft. The decision is made by the transplant surgeon who simultaneously considers macroscopic appearance of the liver (before and after cold perfusion), past medical history of the donor, pre-procurement hemodynamic events with evolution of liver enzymes, and, if available, the result of biopsy. This decision is unavoidably influenced by the pressure of the local waitlist and the perceived risk of a specific donor-recipient match. Additionally, the increasing availability of dynamic ex-situ machine perfusion technologies is changing the current criteria for graft usability and final acceptance for transplantation. The sum of all these factors makes a surgeon’s decision somehow empirical and difficult to standardize among various transplantation centers. Indocyanine green (ICG) clearance is a dynamic test widely employed in liver surgery to estimate hepatic functional reserve in patients with cirrhosis [6]. In the field of liver transplantation, poor recipient ICG clearance measured shortly after implantation has been proven useful for predicting early graft malfunction and one-year graft loss [7]. Although the ICG clearance test performed on DBDs before the procurement operation has been correlated with malfunction in the recipient [8,9] and one-week graft survival [10], it has not been validated as a tool to discard a graft before procurement. A prospective pilot study on 29 DBDs published in 2020 by our group [11] showed that an ICG-plasma disappearance rate (ICG-PDR) below 15%/min was associated with graft rejection by the surgeon who was in charge of the retrieval. The present work aimed to internally validate our previous finding in a larger cohort of donors and set the best cut-off for ICG-PDR to avoid discarding any potential valid graft. Secondary objectives of this study were to externally validate the test on a cohort of donors evaluated by other Spanish centers, to correlate histological changes in discarded livers with PDR results, and to assess the effect of donor ICG clearance in graft malfunction on transplanted grafts.

Between March 2017 and August 2023, DBDs older than 18 years were prospectively included in this study at Gregorio Marañón General University Hospital, Madrid (HGUGM). To maintain consistency in the acceptance criteria, only donors evaluated by two surgeons (JMA, SC), both with more than 5 years of experience in liver procurement, were included.

The ICG-PDR was obtained in the operating room a few minutes before laparotomy, after intravenous injection of a weight dependent dose (0.25 mg/kg) of the dye dissolved in sterile water and flushed with 10 mL of saline solution. The ICG-PDR was measured using pulse digital densitometry (LiMon, Pulsion Medical System). A finger probe was used to detect fractional pulsatile changes in the optical absorption of the dye and the signal was transmitted to a portable display. The result of ICG-PDR, usually obtained in less than 10 minutes, was kept unknown to the surgeon in charge of the procurement. Grafts used for split technique or selected for ex-vivo machine perfusion were excluded from this study. Wedge biopsies of discarded grafts were taken to perform a deferred pathological study of formalin-fixed and paraffin-embedded specimens.

Donors from “HGUGM” were employed as a development cohort to assess the performance of ICG clearance to predict graft discard and choose the best cut-off to avoid false positive results. From March 2021 to May 2023, three other transplantation groups in Spain (“Río Hortega” University Hospital, Valladolid; “Virgen de las Nieves” University Hospital, Granada; and “Nuestra Señora de Candelaria” University Hospital, Santa Cruz de Tenerife) joined the protocol and recruited a cohort of donors for external validation.

Basic demographic variables of donors, along with other characteristics available before procurement (cause of death, past medical history, body mass index, latest laboratory tests, days of intensive care unit (ICU) stay before donation, need of vasoactive drugs, liver US findings) and ICG clearance parameters, were compared between accepted and rejected grafts. An US report was classified as pathological if any liver anomalous finding (higher echogenicity, abnormal contour, segmental or lobar hypertrophy/hypotrophy) was described. Uni- and multivariate analyses were performed to assess independent predictors of graft rejection. A cut-off was set for a predictor specificity of 100% by the analysis of receiver operating characteristic (ROC) curve.

In grafts that were accepted and transplanted, early allograft dysfunctions (EADs) in recipients were assessed according to Olthoff’s criteria [12] for the following conditions: 1) bilirubin ≥ 10 mg/dL on postoperative day (POD) 7, 2) international normalized ratio ≥ 1.6 on POD 7, and 3) alanine aminotransferase (ALT) or aspartate aminotransferase (AST) > 2,000 IU/L within the first 7 days. Donors’ characteristics were compared between patients who developed EAD and those with uncomplicated postoperative course.

The predictive model’s development followed the recommendations stated in the Transparent Reporting of a Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) Initiative [13]. This study was conducted in accordance with the ethical principles of the Declarations of Helsinki. It was approved by the Institutional Review Board of “HGUGM” (code “ICG-Donantes,” minutes ref. 26/2020). Written consent was obtained from all participating patients.

Statistical analysis

Quantitative variables are expressed as median and interquartile range (IQR). If data followed a normal distribution according to the Kolmogorov-Smirnov test, comparisons were made using Student’s t-test. Otherwise, non-parametric tests were used. Qualitative variables were analyzed using χ² and Fisher’s exact tests. Predictive analysis was performed using ROC curves. Statistical significance was set at p < 0.05.

Considering a ratio of accepted to rejected grafts of 3.4, according to the data offered by the National (Spanish) Transplant Organization [1], a sample number of at least 40 donors was needed for the development cohort to achieve an area under ROC curve (AUROC) ≥ 0.8 (good prediction) with a type α error of 0.05 and a β error of 0.20.

Validation of the predictor in the development cohort was performed by examining calibration and discrimination. The overall performance of the model was evaluated using the Brier score. In the validation set, the AUROC of the predictor was calculated and compared to that obtained in the training set. Statistical analysis was performed using Stata Statistical Software, Release 17 (StataCorp., LLC) and SPSS version 25 (IBM Corp.).

Seventy-three donors were evaluated by the “HGUGM” team. The ICG test could be performed in all but two donors who exhibited severe peripheral vasoconstriction and were unable to transmit a quality signal to the finger probe. The remaining 71 donors were included in the training cohort. Thirty (42.3%) grafts were discarded because of a poor quality at macroscopic evaluation, including those with parenchymal stiffness (16 cases), macroscopic steatosis (11 cases), irregular appearance after cold flush (2 cases), and severe atheromatous disease (1 case). Donors whose livers were rejected had a significantly higher body mass index than those with accepted grafts (median: 29.9 kg/m², IQR: 26.1–32.9 kg/m² vs. median: 26.0 kg/m², IQR: 23.3–29.3 kg/m², p = 0.002). They were more frequently classified as ECD by Eurotransplant criteria [14]. ICG-PDR was significantly lower in rejected donors (median: 14.3%/min, IQR: 11.0–18.5%/min vs. median: 24.0%/min, IQR: 19.1–27.3%/min, p < 0.001), while conventional liver laboratory parameters and US results did not show any differences between the two groups (Table 1).

Table 1 . Donors’ characteristics of accepted versus rejected grafts

Characteristic	Accepted (n = 41)	Rejected (n = 30)	p-value
Age (yr)	65 (50–78)	68 (59–77)	0.150
Sex (male)	21 (51.2)	13 (43.3)	0.649
Cause of death			0.807
Stroke	34 (82.9)	23 (76.7)
Trauma	4 (9.8)	4 (13.3)
Anoxia	3 (7.3)	3 (10.0)
Cardiac resuscitation	6 (16.7)	3 (11.1)	0.403
BMI (kg/m2)	26.0 (23.3–29.3)	29.9 (26.1–32.9)	0.002*
ICU stay (day)	2 (1–4)	2 (1–4)	0.540
Pressors	23 (62.2)	20 (80.0)	0.135
Hypertension	22 (53.7)	22 (73.3)	0.092
Diabetes mellitus	10 (25.0)	8 (26.7)	0.875
Dyslipidemia	11 (27.5)	11 (36.7)	0.414
Ultrasonography (pathological)	6 (14.6)	5 (18.5)	0.670
Marginal donor (Eurotransplant criteria)	27 (65.9)	29 (96.7)	0.002*
ICG-PDR (%/min)	24.0 (19.1–27.3)	14.3 (11.0–18.5)	< 0.001*
ICG-R15 (%)	2.7 (1.5–5.7)	11.7 (6.2–18.0)	< 0.001*
GGT (IU/L)	30 (18–65)	40 (20–71)	0.343
AST (IU/L)	25 (19–52)	27 (20–52)	0.876
ALT (IU/L)	18 (12–44)	19 (15–33)	0.946
Bilirubin (mg/dL)	0.5 (0.3–0.8)	0.6 (0.3–0.8)	0.628
Creatinine (mg/dL)	0.7 (0.6–0.9)	0.8 (0.6–1.2)	0.215
Platelets (103/mm3)	196 (158–264)	212 (113–267)	0.961
Sodium (mEq/L)	144 (140–150)	144 (141–148)	0.955

Quantitative variables are expressed as median (interquartile range). Categorical variables are expressed as total counts (%).

ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; GGT, gamma-glutamyl transferase; ICU, intensive care unit; ICG-PDR, indocyanine green plasma disappearance rate; ICG-R15, indocyanine green retention rate at 15 minutes.

*Statistically significant.

In multivariate regression analysis, ICG-PDR was the only independent predictor of graft rejection (odds ratio = 0.788, 95% confidence interval [CI]: 0.697–0.892, p < 0.001) (Table 2). ROC analysis showed that ICG-PDR had a good accuracy in predicting graft discard or acceptance, with an AUROC of 0.857 (95% CI: 0.768–0.947) (Fig. 1). Setting a specificity of 100%, the corresponding cut-off was 13.5%/min. Below this value, none of the donors had suitable livers for transplantation. With such a cut-off, sensitivity, positive predictive value, and negative predictive value were 43%, 100%, and 71%, respectively. Thirteen out of 30 discarded grafts (43.3%) could have been correctly identified in this cohort of donors.

Table 2 . Univariate and multivariate analyses of donors’ variables for predicting graft discard in the training cohort (n = 71)

Donor variable	Univariate analysis (p)	Odds ratio	95% CI	p-value
Age	0.113	-	-	-
ICG-PDR	< 0.001	0.788	0.697–0.892	< 0.001
Arterial hypertension	0.095	-	-	-
BMI	0.005	1.114	0.995–1.247	n.s.
Extended criteria for donation	0.011	2.311	0.218–24.459	n.s.

Variables with a p < 0.05 at univariate analysis were entered in a multivariate regression model.

ICG-PDR, indocyanine green plasma disappearance rate; BMI, body mass index; CI, confidence interval; n.s., not statistically significant.

Fig 1. Receiver operating characteristic curves of ICG-PDR in liver donors in the prediction of graft discard by procurement surgeon both in the training cohort (red line) and the validation cohort (blue line). An area under curve of 1.0 is characteristic of an ideal test, whereas an area of 0.5 or less indicates a test of non-diagnostic value. ICG-PDR, indocyanine green plasma disappearance rate.

Internal validation

For internal validation of PDR, the Hosmer-Lemeshow test was computed by partitioning the study population into five groups based on the predicted probability of graft suitability obtained from the risk prediction model. The chi-square statistic used to compare the observed and expected events had p = 0.455, showing a good calibration. The calibration belt (Fig. 2A) showed a statistic value of 1.55 with a p-value of 0.214, indicating that the hypothesis of good calibration could not be rejected. The calibration plot for internal validation is shown in Fig. 2B. The global accuracy of the test was good, with a Brier score of 0.150 (95% CI: 0.107–0.194).

Fig 2. Calibration belt (A) showing deviations from the bisector (45% line of perfect fit) at 80% (inner belt: light grey area) and 95% (outer belt: dark grey area) confidence levels, and calibration plot (B) indicating that predicted and empirical probabilities are similar and that the built prediction model fits the data well. PDR, plasma disappearance rate; CI, confidence interval.

External validation

The external validation cohort consisted of 17 donors. Characteristics of training and validation cohorts of donors are shown in Table 3. The two populations did not show any significant differences in demographics, donor characteristics, or liver-specific test results. Rate of discard was 42.3% in the training set and 52.9% in the validation set, showing no significant (p = 0.426) difference.

Table 3 . Comparison of donor’s characteristics between development and validation cohorts

Characteristic	Development cohort (n = 71)	Validation cohort (n = 17)	p-value
Age (yr)	68 (55–79)	66 (47–74)	0.234
Sex (male)	38 (53.5)	11 (64.7)	0.404
Cause of death			0.755
Cerebrovascular accident	57 (80.3)	14 (82.4)
Trauma	8 (11.3)	1 (5.9)
Anoxia	6 (8.5)	2 (11.8)
Cardiac resuscitation	9 (14.3)	4 (23.5)	0.359
BMI (kg/m2)	27.7 (24.5–31.1)	25.2 (22.7–30.0)	0.252
ICU stay (day)	2 (1–4)	3 (1–4)	0.865
Pressors	49 (72.1)	13 (76.5)	0.487
Hypertension	44 (62.0)	11 (64.7)	0.834
Diabetes	18 (25.7)	2 (11.8)	0.338
Dyslipidemia	22 (31.4)	2 (11.8)	0.136
Extended criteria donor (%)	56 (78.9)	11 (64.7)	0.218
ICG-PDR (%/min)	19.5 (14.7–25.5)	19.0 (12.5–22.9)	0.495
ICG-R15 (%)	5.4 (2.2–10.9)	5.7 (3.0–10.5)	0.669
GGT (U/L)	31 (18–66)	40 (30–65)	0.264
AST (U/L)	26 (19–52)	31 (22–54)	0.369
ALT (U/L)	19 (12–43)	25 (20–45)	0.145
Bilirubin (mg/dL)	0.6 (0.3–0.8)	1.0 (0.4–1.3)	0.081
Creatinine (mg/dL)	0.8 (0.6–1.0)	0.9 (0.7–1.1)	0.229
Platelets (103/mm3)	210 (151–265)	220 (164–265)	0.670
Sodium (mEq/L)	144 (140–148)	143 (140–148)	0.965
US findings (pathological)	11 (16.2)	0 (0)	0.083
Rate of discard (%)	42.3	52.9	0.426

Quantitative variables are expressed as median (interquartile range). Categorical variables are expressed as total counts (%).

ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; GGT, gamma-glutamyl transferase; ICU, intensive care unit; ICG-PDR, indocyanine green plasma disappearance rate; ICG-R15, indocyanine green retention rate at 15 minutes; US, ultrasonography.

ICG-PDR in the validation cohort showed a good discrimination power, with an AUROC of 0.806 (95% CI: 0.582–1.000), not statistically different from that of the training set (p = 0.673) (Fig. 1). Out of the nine donors whose livers were discarded in the external cohort, five (55.6%) donors had an ICG-PDR of less than 13.5%/min. Therefore, they could have been recognized as unsuitable. Concurrently, no donor with an ICG-PDR < 13.5%/min had a valid graft.

Pathological findings

Considering the two cohorts as a whole, 39 discarded grafts were analyzed. Nine biopsies were missing. Therefore, only thirty specimens were available for pathological assessment. Among 16 biopsies coming from donors with a PDR ≤ 13.5%/min, 12 with bridging fibrosis or cirrhosis, 2 with severe steatosis, 1 with moderate steatosis, and 1 with non-specific features were found. Among the 14 specimens from donors with PDR > 13.5%/min, 7 with periportal fibrosis, 3 with severe steatosis or steatohepatitis, 2 with ischemic changes, and two with mild steatosis or considered non-pathological were found (Fig. 3).

Fig 3. Histologic findings for biopsy specimens of rejected grafts according to the cut-off of indocyanine green plasma disappearance rate (PDR) selected for discard.

Donor plasma disappearance rate and transplant outcomes

Of 49 accepted grafts in pooled cohorts, three were assigned to external centers and lost to follow-up and one was not transplanted after finding progression of hepatocellular carcinoma in the recipient. The remaining 45 transplanted patients were established as the study cohort. Fifteen (33.3%) patients with EAD were observed. Donors’ median PDR among patients with EAD versus patients with uncomplicated postoperative course was not statistically significantly different (24.0%/min, IQR: 17.2–27.1%/min vs. 22.7%/min, IQR: 19.1–27.1%/min, p = 0.782). Donors’ variables that showed differences between EAD and non-EAD patients were donor age (median: 75 years, IQR: 58–80 years vs. median: 60 years, IQR: 41–73 years, p = 0.037) and the use of vasoactive drugs prior to donation (86.7% vs. 56.7%, p = 0.043) (Table 4).

Table 4 . Comparison of donor variables and cold ischemic time between patients who developed early allograft dysfunction (EAD) and those with a normal postoperative course

Variable	EAD (n = 15)	Non-EAD (n = 30)	p-value
Age (yr)	75 (58–80)	60 (41–73)	0.037*
Sex (male)	7 (46.7)	17 (56.7)	0.526
BMI (kg/m2)	27.5 (24.5–29.6)	24.8 (22.3–27.8)	0.170
Cause of death (stroke vs. others)	12 (80.0)	25 (83.3)	0.542
Pressors	13 (86.7)	17 (56.7)	0.043*
Cardiac resuscitation	4 (26.7)	3 (10.3)	0.166
Arterial hypertension	11 (73.3)	13 (43.3)	0.055
Diabetes mellitus	4 (26.7)	4 (13.3)	0.241
Dyslipidemia	5 (33.3)	6 (20.0)	0.266
Ultrasonography (pathological)	3 (20.0)	3 (10.0)	0.311
ICU stay (day)	2.5 (1.0–3.3)	2.0 (1.0–4.0)	0.900
GGT (IU/L)	36.0 (24.0–65.3)	26.5 (14.8–46.8)	0.359
AST (IU/L)	37 (23–56)	28 (18–49)	0.408
ALT (IU/L)	25 (15–51)	19 (12–41)	0.352
Bilirubin (mg/dL)	0.8 (0.5–1.1)	0.5 (0.3–0.8)	0.135
Platelets (103/mm3)	204 (137–258)	214 (153–264)	0.731
ICG-R15 (%)	2.7 (2.3–2.6)	3.1 (1.6–5.7)	0.588
ICG-PDR (%/min)	24.0 (17.2–27.1)	22.7 (19.1–27.1)	0.782
Marginal donor	11 (73.3)	16 (53.3)	0.167
Cold ischemic time (min)	335 (293–420)	303 (238–363)	0.112

Quantitative variables are expressed as median (interquartile ranges). Categorical variables are expressed as total counts (%).

ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; GGT, gamma-glutamyl transferase; ICU, intensive care unit; ICG-PDR, indocyanine green plasma disappearance rate; ICG-R15, indocyanine green retention rate at 15 minutes.

*Statistically significant.

This study internally validated ICG-PDR as a useful test to discard liver grafts in a prospective study of 71 DBD evaluated by a transplantation team in Spain. A cut-off for PDR of 13.5%/min was found. Below this value, all grafts were deemed unsuitable due to marginal quality. This result was externally validated in a prospective cohort of 17 donors evaluated by three other Spanish transplantation centers, showing well-concordant outcomes. A microscopic confirmation of parenchymal structural damage, namely bridging fibrosis, cirrhosis, or severe steatosis, was found in the great majority of specimens from donors with PDR ≤ 13.5%/min.

Some authors in the early ’90 showed a possible application of the ICG clearance test in liver donors. Wesslau et al. [15] have studied ICG clearance using a thermodilution method in 41 donors. They found that out of 21 grafts accepted for transplantation, only two had a PDR < 15%/min. Interestingly, these two cases had a graft failure.

In more recent years, relying on modern non-invasive trans-cutaneous digital probes, Zarrinpar et al. [10] from UCLA University, USA found that ICG-PDR in 53 donors was the only donor variable associated with non-survival of the graft with an ideal cut-off of 19.3%/min (meaning by “non-survival” both discard before transplantation and 7-days graft failure after transplantation). In the present study, we could not find any correlation between donor ICG clearance and EAD in the 45 patients who underwent transplantation. This could be explained by the inherent design of the present study, which aimed primarily to validate a method to discard grafts before procurement, and not to predict the risk of graft dysfunction. However, it would have been questionable to set both surgeons’ turndown of organ and graft failure after transplantation as the end-point of the study. In the latter situation, the recipient’s conditions, ischemic time, and technical variables could also intervene, making the ICG-PDR cut-off difficult to interpret and apply in clinical use.

ICG clearance is known to be influenced both by hepatic arterial blood flow and liver-specific extraction rate [16]. A matter of concern is the reliability of ICG-PDR in the context of brain-dead donation, when hemodynamic stability is often dependent on vasoactive drugs. ICG clearance test might not be trustful during high volume fluid resuscitation and in presence of severe hemodynamic shock [17]. In this respect, Tang et al. [9] have analyzed 3 months graft survival in 90 patients. DBDs had one or multiple measurements of ICG clearance within 6 h of procurement. A poor ICG retention rate at 15 minutes (ICG-R15) was a reason for attempting to improve maintenance measures in donors. An ICG-R15 below 11% was considered the maintenance target before procurement. Notably, the mean donor age in this Chinese cohort was 31 years, which was far from the Western reality of a significantly older donor population. In their scenario, ICG clearance could reflect a state of physiological instability related to brain death and/or sepsis rather than a reliable indicator of liver quality.

In 2020, our group at “HGUGM” published results of a pilot study of 29 DBDs [11], showing a good correlation between donor’s ICG-PDR and decision of the surgeon (blinded to PDR results) to accept or discard the graft. Thereafter, we recruited more donors and started a multicentric study to validate our results. We set a cut-off with 100% specificity to avoid any false positive results. Our purpose was to ideally adopt the ICG clearance test, which requires non-invasive bedside equipment already available in many ICUs, as a screening before a donation is pursued by on-site procurement coordinators.

The overall rate of graft discard in the present dataset was 39 out of 88 (44.3%). This rate is somewhat higher than expected according to national statistics [1] possibly because the cohort did not include all consecutive donors available for logistic reasons. Applying ICG-PDR cut-off as a filter before donation, 18 (46.2%) donors with unsuitable grafts could have been identified correctly, thus avoiding unnecessary travels and waste of financial and human resources.

Surprisingly, none of the donors’ laboratory tests or ultrasound findings could suggest graft turndown, although the majority of discarded grafts showed advanced pathological injury. A possible explanation for this was that the study cohort was composed of donors already selected for liver evaluation, having preemptively discarded those with gross abnormalities in the workup.

This study has some limitations that should be pointed out. First, only 17 donors could be recruited for external validation. Therefore, AUROC comparison between training and validation cohorts might lack a sufficient power. Despite this, both datasets were comparable for all donor characteristics. Most importantly, no valid graft would have been missed in the validation cohort using the established cut-off.

Second, approximately 20% of specimens for pathological study were eventually missing for analysis. Additionally, the spectrum of variability and combination of histological findings made it difficult to establish a straightforward correlation between ICG elimination and microscopic liver anatomy. While a PDR below 13.5%/min was almost invariably associated with advanced histological damage, findings in discarded grafts from donors with a PDR above 13.5%/min were quite variable, ranging from periportal fibrosis or severe steatosis to apparently healthy liver. Although this subgroup of grafts was unviable for transplantation according to surgeon’s criteria, they still showed reasonably preserved ICG elimination. They might be an interesting target for future studies with machine perfusion technologies, including hypothermic oxygenated machine perfusion (HOPE), normothermic machine perfusion (NMP), and controlled oxygenated rewarming. It is significant that most of the clinical trials thus far [18] comparing dynamic perfusion techniques with cold static storage have included donors with ECD grafts. The definition of ECD [14] (any donor with at least one of the following criteria: age > 65 years, ICU stay with ventilation > 7 days, body mass index > 30 kg/m², steatotic liver > 40%, serum sodium > 165 mmol/L, ALT > 105 U/L, AST > 90 U/L, serum bilirubin > 3 mg/dL) is quite broad. Thus, grafts of very different qualities might have been grouped in the same category. As a matter of fact, in our training cohort, almost 80% of donors fulfilled the ECD criteria. A more objective stratification inside the ECD group by the donor’s ICG clearance determination might help select which grafts would benefit more from these perfusion techniques. However, in the current study, we intentionally excluded grafts selected for NMP or HOPE because the surgeon’s criteria of acceptance might have been different in such cases. In addition, these technologies were not available at our center during the first part of the study period.

Despite the above-mentioned limitations, this is the first study to validate pre-retrieval ICG clearance test in DBD as a tool to predict graft discard in a Spanish population. External validation in a larger international cohort is warranted to explore the applicability of this strategy in other countries.

The authors would like to thank José María Bellón of the “Instituto de Investigación Sanitaria Gregorio Marañón” (IiSGM) for his support in statistical analysis.

None.

No potential conflict of interest relevant to this article was reported.

Conceptualization: SC, JMA, JALB, JMPP, AMM. Data curation: SC, JMA, KP, AJPA, MSH, DP, NZC, MABG, AGMT, MFM, SHK. Methodology: SC, JMA. Writing - original draft: SC. Writing - review & editing: SC, JMA, JALB.