Abstract
Background.
The widespread use of normothermic machine perfusion (NMP) has enabled greater utilization of donation after circulatory death (DCD) liver grafts for transplantation. Use of NMP can cost an additional $40 000–$100 000 The aim of our study was to see whether the use of NMP would lower postoperative costs after DCD liver transplantation (LT).
Methods.
Retrospective data of all DCD LTs performed at our center between August 19, 2022, and May 31, 2024, were analyzed. The OrganOx metra device, back-to-base, was used for NMP at our center. United Network for Organ Sharing data were used to present national DCD LT volumes and waitlist outcomes.
Results.
Sixty-seven NMP and 44 static cold storage transplants were performed. In the NMP group, donors were older (50 versus 45 y, P = 0.0260), were at increased risk (US donor risk index 2.43 versus 2.12, P = 0.006), incidence of early allograft dysfunction (42% versus 75%, P = 0.0008) and postreperfusion syndrome (25% versus 48%, P = 0.0239) was lower, and recipients had better native kidney function at 3 mo (estimated glomerular filtration rate 73 versus 62 mL/min/1.73 m2, P = 0.0205). Use of NMP did not decrease postoperative direct costs. On multivariate analysis, an additional 56 min of cold ischemic time and the presence of postreperfusion syndrome increased postoperative direct costs by $14 700 and $23 100, respectively.
Conclusions.
Use of NMP does not decrease postoperative direct costs after DCD LT. However, with the use of NMP, a greater number of DCD liver grafts can be used, from a broader range of donors, wait time to transplant can be reduced, and waitlist survival can be improved while improving relevant clinical outcomes. The overall cost savings achieved by transplanting patients quickly at low Model for End-stage Liver Disease scores and improving waitlist morbidity should be further explored.
Liver transplantation (LT) from donation after circulatory death (DCD) donors has been on the rise during the past few years in the United States.1 The widespread use of normothermic perfusion techniques has enabled greater utilization of liver grafts from DCD donors.2,3 This surge can be attributed to improved outcomes of DCD LT with normothermic regional perfusion (NRP) and normothermic machine perfusion (NMP) techniques, compared with static cold storage (SCS) after super rapid recovery.4-6
Compared with SCS after super rapid recovery, normothermic perfusion techniques pose logistical challenges and incur additional organ acquisition costs. According to reports, factors that determine organ acquisition costs depend on the type of normothermic perfusion technique used (NRP and/or NMP) and the timing/logistics of NMP (back-to-base versus NMP at donor hospital). Purely from the point of pretransplant costs incurred with the use of normothermic perfusion technology, these modalities cost approximately an additional $9500 for NRP or $40 000–$100 000 for ex situ NMP, depending on the machine perfusion platform used.7-9 Do these additional organ acquisition costs offset the postoperative costs after LT? The aim of this study was to determine whether the use of NMP (back-to-base model) was associated with lower postoperative costs after DCD LT. This study also observed the trends of DCD LT, wait times, and waitlist mortality in relation to the implementation of normothermic perfusion for clinical use in the United States.
MATERIALS AND METHODS
We conducted a retrospective analysis of all DCD LTs performed at our center between August 20, 2022, and May 31, 2024, with follow-up through July 24, 2024. Follow-up for graft and patient survival and 12-mo readmission and need for kidney transplant were through March 21, 2025. Electronic medical records were reviewed to collect data for this study after Institutional Research Board approval. LTs from living donors, donation after brain death (DBD) donors, pediatric transplants at ages less than 18 y, and multiorgan transplants were excluded. DCD liver grafts that underwent NRP with or without NMP were excluded from primary analysis, but their outcomes are presented in the results. DBD LTs were excluded from our analysis as NMP technology, at our center, is usually reserved for high-risk DBD donors and to manage logistical issues. The main outcome of interest was direct costs incurred after transplantation. During a patient encounter, total costs incurred by the hospital are the sum of direct costs plus indirect costs. Resources used by a patient during a particular admission such as inpatient stay, operation charges, disposables, pharmaceuticals, healthcare provider professional charges, imaging and laboratory expenses, and so on fall under direct costs, whereas expenses such as electricity, water, equipment, medical records management, and so on contribute to indirect costs that are distributed across all patients. A summary of direct and indirect costs is provided in Table S1 (SDC, https://links.lww.com/TXD/A795). Organ acquisition costs, which include organ recovery, transportation charges, and the use of perfusion technology, have been included in our analysis and presented separately from direct costs. Our primary analysis centers on costs arising from the point of hospital admission for surgery. Expenses associated with preoperative waitlist period were not included. Direct costs were calculated as a micro-costing or bottom-up approach using patient-level medical records. Indirect costs and organ acquisition (procurement, machine perfusion, and travel) are calculated as a macro costing or top-down approach where overall costs are distributed among recipients.
The quality of the donor liver grafts has been defined using the donor risk index (DRI) for LT calculator.10 Agonal time at our center was defined as the duration of time during which donor systolic blood pressure is <80 mm Hg. Postreperfusion liver biopsies were used to stage fibrosis and steatosis of liver grafts, as donor liver biopsies were not routinely performed at donation.11,12 Anastomosis time was defined as the time between “liver out of ice” and reperfusion with portal blood. Cold ischemia time (CIT) was defined as the time from cross-clamp to either portal perfusion in recipients or the start of perfusion on the normothermic machine. IC was identified and classified as follows: grade A, normal; grade B, diffuse necrosis; grade C, multifocal progressive; grade D, confluence dominant, and grade E, minor form.13 Days alive and out of hospital has been considered a surrogate marker for quality of life in LT; it is measured as the number of days a patient is alive and spends outside a hospital setting in a given period of time.14
Medians with interquartile ranges or counts and percentages were used to summarize donor and recipient characteristics and recipient outcomes. Mann-Whitney U tests and Fisher exact tests were used to compare the SCS and NMP recipients. If all results were the same for Fisher exact tests, a P value of 1 was assumed. Pairwise deletion was used for bivariate analyses. Cumulative direct costs were summarized at index admission, 90 d, 180 d, and 365 d. Pairwise deletion was used for missing values. Kaplan-Meier curves were fitted to graft and patient survival. Survival times were censored at their last known follow-up. No death censoring was used as no deaths occurred without a graft failure. Log-rank tests were used to test for differences in time to failure or death between SCS and NMP.
A longitudinal multiple regression was fitted to cumulative direct costs across 4 time points (index admissions, 3 mo, 6 mo, and 12 mo post-LT) with a repeated effect for recipients. Clinically selected baseline fixed effects included donor body mass index (BMI), US DRI, agonal time, local recovery, CIT, recipient age, transplant Model for End-stage Liver Disease-sodium (MELD-Na), and postreperfusion syndrome (PRS). Additional baseline variables (distance between donor and recipient center and start time of surgery) were included as differences were significant between SCS and NMP groups. The SCS versus NMP effect, time frame for cumulative direct costs, and the interaction effect were also included. LT fiscal year was included to adjust for inflation. Three recipients (2.7%) were missing baseline agonal time data. Complete case deletion was used for the multivariate analysis.
Additional data were used to summarize deceased donor LT volumes and waitlist outcomes at a national and local level. The data reported here have been supplied by the United Network for Organ Sharing as the contractor for the Organ Procurement and Transplantation Network (OPTN). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the OPTN or the US Government. These data were current as of September 30, 2024.
United Network for Organ Sharing OPTN data were summarized at 6-mo increments. For each 6-mo time frame, DBD and DCD volumes were bi-annualized. Kaplan-Meier curves were fitted to wait time to transplant and to waitlist patient survival, allowing at least 1 y of follow-up time. Waitlist survival proportions were estimated for 1 y. Kaplan-Meier medians were estimated for time on waitlist to transplant. Transplants occurring between January 1, 2018, and June 30, 2024, were included in the volume counts for DBD and DCD. Patients waitlisted between January 1, 2018, and September 30, 2023, were included for Kaplan-Meier estimates, allowing for 1 y of follow-up. Pediatric recipients or waitlist patients, simultaneous and multiorgan transplants, and living donors were excluded from analyses.
SAS version 9.4 (9.4 TS1M8, Cary, NC) was used for all analyses on the 64-bit platform for Microsoft Windows. RStudio version 4.4.1 was used to create the forest plot. Comparisons were considered statistically significant at the 0.05 significance level. The study protocol was approved by The Ohio State University Institutional Review Board (No. 2023H0246).
RESULTS
A total of 120 DCD LTs were performed during this 21-mo study period. Of these, 67 transplants underwent NMP, 44 transplants were performed after SCS alone, and 9 DCD LTs underwent NRP (4 underwent NRP followed by NMP and 5 underwent NRP followed by SCS alone). A total of 83 DCD livers were assessed on NMP. Five DCD livers were declined as they did not meet our viability criteria: 3 livers were declined as they failed to clear lactate to <2 mmol/L within 6 h of perfusion/rising lactate; 1 liver was declined because of a rise in lactate to >2 mmol/L within 6 h, after initial lactate clearance to <2 mmol/L; 1 liver was declined because of poor arterial flows (<150 mL/min). Two DCD liver grafts were discarded because of unacceptable biopsy results only (fibrosis), which had resulted after these livers were placed on the pump. Once DCD liver graft was declined because of uncontrolled surgical bleeding on the pump from recovery injury. Four DCD liver grafts underwent NMP after NRP; the data of these transplants are presented but not included in the primary analysis. Another 4 DCD LTs were excluded from our analysis as they were multiorgan transplants (Figure 1).
FIGURE 1.
Flowchart of DCD liver grafts excluded from the analysis. DCD, donation after circulatory death; NMP, normothermic machine perfusion; NRP, normothermic regional perfusion; SCS, static cold storage; SLK, simultaneous liver and kidney transplantation.
Donor Characteristics
In the NMP group, donors were older (50 versus 45 y, P = 0.0260), liver grafts were from increased risk donors (USDRI 2.43 versus 2.12, P = 0.006), a greater percentage of livers were procured by local recovery surgeons (18% versus 0%, P = 0.003), and liver grafts were used from longer distances (141 versus 98 nautical miles, P = 0.0348), when compared with the SCS group. There were no differences in donor BMI, sex, hepatitis C virus status, predonation alanine transaminase, predonation aspartate transaminase (53 versus 40 U/L), predonation peak total bilirubin, cause of death, agonal time, withdrawal to cross-clamp time, hepatectomy time, CIT, use of open offer DCD liver grafts, macro steatosis, fibrosis stage, necrosis on reperfusion biopsy, and positive donor blood cultures between the NMP and SCS groups (Table 1).
TABLE 1.
Baseline donor characteristics of DCD liver transplants
| Variable | Total (N = 120) | SCS (N = 44) | NMP (N = 67) | P | NRP SCS (N = 5) | NRP NMP (N = 4) | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | ||
| Age, y | 120 | 47 (37–56) | 44 | 45 (35–54) | 67 | 50 (40–59) | 0.0260 | 5 | 46 (42–47) | 4 | 35 (29–51) |
| Sex | |||||||||||
| Male | 120 | 67 (56%) | 44 | 25 (57%) | 67 | 35 (52%) | 0.6991 | 5 | 4 (80%) | 4 | 3 (75%) |
| Female | 53 (44%) | 19 (43%) | 32 (48%) | 1 (20%) | 1 (25%) | ||||||
| BMI, kg/m2 | 120 | 28.3 (24.5–33.3) | 44 | 28.0 (25.5–33.4) | 67 | 28.2 (22.7–33.4) | 0.5092 | 5 | 31.7 (30.6–32.9) | 4 | 25.9 (25.3–32.6) |
| US DRI | 120 | 2.33 (1.95–2.64) | 44 | 2.12 (1.86–2.55) | 67 | 2.43 (2.08–2.74) | 0.0060 | 5 | 2.43 (1.83–2.60) | 4 | 1.81 (1.65–2.50) |
| Donor HCV status | |||||||||||
| NAT (–) | 120 | 115 (96%) | 44 | 41 (93%) | 67 | 65 (97%) | 0.3833 | 5 | 5 (100%) | 4 | 4 (100%) |
| NAT (+) | 5 (4%) | 3 (7%) | 2 (3%) | 0 (0%) | 0 (0%) | ||||||
| Last AST predonation, U/L | 120 | 45 (28–80) | 44 | 40 (26–58) | 67 | 53 (29–94) | 0.0908 | 5 | 40 (37–58) | 4 | 38 (25–81) |
| Last ALT predonation, U/L | 120 | 38 (21–72) | 44 | 36 (20–68) | 67 | 40 (23–79) | 0.4349 | 5 | 31 (24–38) | 4 | 53 (36–87) |
| Peak total bilirubin predonation, mg/dL | 120 | 0.5 (0.4–0.8) | 44 | 0.5 (0.4–0.7) | 67 | 0.5 (0.3–0.8) | 0.5375 | 5 | 0.5 (0.4–0.5) | 4 | 0.7 (0.4–0.9) |
| Cause of death | |||||||||||
| Anoxia | 120 | 78 (65%) | 44 | 29 (66%) | 67 | 44 (66%) | 0.9163 | 5 | 2 (40%) | 4 | 3 (75%) |
| CVA/stroke | 22 (18%) | 9 (20%) | 12 (18%) | 1 (20%) | 0 (0%) | ||||||
| Head/trauma | 13 (11%) | 3 (7%) | 7 (10%) | 2 (40%) | 1 (25%) | ||||||
| Other | 7 (6%) | 3 (7%) | 4 (6%) | 0 (0%) | 0 (0%) | ||||||
| Agonal time, min | 116 | 16 (13–22) | 42 | 15 (13–19) | 66 | 16 (12–21) | 0.7046 | 5 | 78 (72–81)a | 3 | 91 (84–160)a |
| Withdrawal to cross-clamp time, min | 116 | 25 (21–31) | 42 | 24 (22–30) | 66 | 24 (20–28) | 0.4174 | 5 | 87 (86–88) | 3 | 112 (91–160) |
| Hepatectomy time, min | 100 | 29 (26–35) | 37 | 29 (26–33) | 56 | 30 (26–36) | 0.4411 | 3 | 28 (26–33) | 4 | 25 (20–38) |
| Local recovery | |||||||||||
| No | 120 | 104 (87%) | 44 | 44 (100%) | 67 | 55 (82%) | 0.0030 | 5 | 2 (40%) | 4 | 3 (75%) |
| Yes | 16 (13%) | 0 (0%) | 12 (18%) | 3 (60%) | 1 (25%) | ||||||
| Open offer | |||||||||||
| No | 120 | 104 (87%) | 44 | 39 (89%) | 67 | 58 (87%) | 1.0000 | 5 | 4 (80%) | 4 | 3 (75%) |
| Yes | 16 (13%) | 5 (11%) | 9 (13%) | 1 (20%) | 1 (25%) | ||||||
| Distance between donor and recipient center, nautical miles | 120 | 119 (56–169) | 44 | 98 (8–137) | 67 | 141 (61–205) | 0.0348 | 5 | 207 (165–247) | 4 | 186 (28–390) |
| Cold ischemia time, h | 120 | 4.5 (3.9–5.5) | 44 | 4.2 (3.8–4.9) | 67 | 4.6 (4.1–5.5) | 0.0830 | 5 | 5.2 (4.9–5.7)b | 4 | 5.7 (4.4–6.1)b |
| NMP duration, h | 71 | 11 (8–13) | 0 | NA | 67 | 11 (8–13) | NA | 0 | NA | 4 | 8 (6–12) |
| Total preservation time, h | 120 | 12.4 (4.5–16.7) | 44 | 4.2 (3.8–4.9) | 67 | 15.5 (13.5–18.0) | NA | 5 | 5.2 (4.9–5.7) | 4 | 14.8 (12.4–17.1) |
| Macro steatosis, % | |||||||||||
| Minimal (<5%) | 78 | 54 (69%) | 32 | 23 (72%) | 42 | 28 (67%) | 0.8072 | 3 | 2 (67%) | 1 | 1 (100%) |
| Mild (5%–33%) | 22 (28%) | 8 (25%) | 13 (31%) | 1 (33%) | 0 (0%) | ||||||
| Moderate (33%–66%) | 2 (3%) | 1 (3%) | 1 (2%) | 0 (0%) | 0 (0%) | ||||||
| Severe (>66%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | ||||||
| Fibrosis stage | |||||||||||
| F0 (no fibrosis) | 108 | 95 (88%) | 41 | 38 (93%) | 60 | 50 (83%) | 0.4253 | 4 | 4 (100%) | 3 | 3 (100%) |
| F1 (portal fibrosis) | 12 (11%) | 3 (7%) | 9 (15%) | 0 (0%) | 0 (0%) | ||||||
| F2 (periportal fibrosis) | 1 (1%) | 0 (0%) | 1 (2%) | 0 (0%) | 0 (0%) | ||||||
| F3 (bridging fibrosis) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | ||||||
| Necrosis on reperfusion biopsy | |||||||||||
| Minimal (<5% necrosis) | 105 | 23 (22%) | 41 | 11 (27%) | 58 | 10 (17%) | 0.6611 | 4 | 2 (50%) | 2 | 0 (0%) |
| Mild (5%–33% necrosis) | 60 (57%) | 21 (51%) | 35 (60%) | 2 (50%) | 2 (100%) | ||||||
| Moderate (34%–67% necrosis) | 16 (15%) | 6 (15%) | 10 (17%) | 0 (0%) | 0 (0%) | ||||||
| Severe (>67% necrosis) | 6 (6%) | 3 (7%) | 3 (5%) | 0 (0%) | 0 (0%) | ||||||
| Positive donor blood cultures | |||||||||||
| No | 118 | 110 (93%) | 44 | 42 (95%) | 65 | 59 (91%) | 0.4701 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 8 (7%) | 2 (5%) | 6 (9%) | 0 (0%) | 0 (0%) | ||||||
aAgonal to cross-clamp time was presented here for the NRP group, as the NRP start time was not available for them.
bCold ischemic time in NRP group also includes NRP run time, as the NRP stop time was not available for them.
ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass index; CVA, cerebrovascular accident; DCD, donation after circulatory death; DRI, donor risk index; HCV, hepatitis C virus; IQR, interquartile range; NAT, nucleic acid testing; NMP, normothermic machine perfusion, NRP, normothermic regional perfusion; SCS, static cold storage.
Recipient Characteristics and Intraoperative Variables
There were no differences in recipient characteristics such as age, BMI, sex, MELD score, MELD-Na score, cause of liver disease, presence of hepatocellular carcinoma, redo LT, baseline creatinine, baseline estimated glomerular filtration rate, need for preoperative dialysis, and wait time to transplant between the NMP and SCS groups. The start time of surgeries was timed between 4 am and 12 pm to a greater extent (79% versus 14%, P < 0.0001), and there was a significantly lower incidence of PRS (25% versus 48%, P = 0.0239) in the NMP group. There was no difference in the operation time, anastomosis time, incidence of postreperfusion cardiac arrest, and the need for intraoperative blood product transfusion, such as red blood cells, platelets, fresh frozen plasma, and cryoprecipitate (Table 2).
TABLE 2.
Baseline recipient characteristics and intraoperative variables
| Variable | Total (N = 120) |
SCS (N = 44) |
NMP (N = 67) |
P | NRP SCS (N = 5) |
NRP NMP (N = 4) |
|||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | ||
| Age, y | 120 | 59 (50–64) | 44 | 59 (51–63) | 67 | 58 (49–64) | 0.5996 | 5 | 64 (61–64) | 4 | 57 (42–69) |
| Sex | |||||||||||
| Male | 120 | 74 (62%) | 44 | 27 (61%) | 67 | 41 (61%) | 1.0000 | 5 | 4 (80%) | 4 | 2 (50%) |
| Female | 46 (38%) | 17 (39%) | 26 (39%) | 1 (20%) | 2 (50%) | ||||||
| BMI, kg/m2 | 119 | 28.9 (25.4–33.3) | 44 | 30.5 (26.1–34.1) | 67 | 27.4 (25.0–31.8) | 0.0533 | 5 | 35.4 (34.5–37.1) | 3 | 24.3 (23.4–37.0) |
| MELD at transplant | 116 | 16 (13–19) | 43 | 16 (13–19) | 64 | 16 (14–19) | 0.9238 | 5 | 17 (17–18) | 4 | 18 (15–21) |
| MELD-Na at transplant | 120 | 19 (15–22) | 44 | 19 (15–22) | 67 | 19 (15–22) | 0.9138 | 5 | 17 (15–20) | 4 | 21 (16–23) |
| Cause of liver disease | |||||||||||
| Alcohol | 119 | 46 (39%) | 44 | 12 (27%) | 66 | 29 (44%) | 0.3437 | 5 | 3 (60%) | 4 | 2 (50%) |
| MASH | 40 (34%) | 18 (41%) | 20 (30%) | 1 (20%) | 1 (25%) | ||||||
| PSC/PBC/autoimmune | 12 (10%) | 5 (11%) | 7 (11%) | 0 (0%) | 0 (0%) | ||||||
| Other | 21 (18%) | 9 (20%) | 10 (15%) | 1 (20%) | 1 (25%) | ||||||
| Hepatocellular carcinoma | |||||||||||
| No | 120 | 81 (68%) | 44 | 30 (68%) | 67 | 44 (66%) | 0.8389 | 5 | 5 (100%) | 4 | 2 (50%) |
| Yes | 39 (33%) | 14 (32%) | 23 (34%) | 0 (0%) | 2 (50%) | ||||||
| Redo liver transplant | |||||||||||
| No | 120 | 119 (99%) | 44 | 44 (100%) | 67 | 66 (99%) | 1.0000 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 1 (1%) | 0 (0%) | 1 (1%) | 0 (0%) | 0 (0%) | ||||||
| Baseline creatinine, mg/dL | 120 | 0.95 (0.74–1.32) | 44 | 0.96 (0.74–1.50) | 67 | 0.93 (0.70–1.26) | 0.5812 | 5 | 1.03 (1.00–1.38) | 4 | 0.88 (0.56–1.16) |
| Baseline eGFR, mL/min/1.73 m2 | 120 | 82 (57–90) | 44 | 80 (54–90) | 67 | 84 (57–90) | 0.4244 | 5 | 81 (43–90) | 4 | 86 (73–90) |
| Preoperative dialysis | |||||||||||
| No | 120 | 120 (100%) | 44 | 44 (100%) | 67 | 67 (100%) | 1.0000 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) | ||||||
| Wait time to transplant, d | 120 | 65 (18–211) | 44 | 87 (32–214) | 67 | 44 (14–179) | 0.1404 | 5 | 218 (62–296) | 4 | 54 (28–107) |
| Operation room time, min | 120 | 312 (257–364) | 44 | 312 (243–360) | 67 | 318 (262–374) | 0.4421 | 5 | 288 (273–315) | 4 | 282 (243–330) |
| Start time of surgery | |||||||||||
| 8 pm–3:59 am | 120 | 9 (8%) | 44 | 4 (9%) | 67 | 4 (6%) | <0.0001 | 5 | 1 (20%) | 4 | 0 (0%) |
| 4 am–11:59 pm | 63 (53%) | 6 (14%) | 53 (79%) | 1 (20%) | 3 (75%) | ||||||
| 12 pm–7:59 pm | 48 (40%) | 34 (77%) | 10 (15%) | 3 (60%) | 1 (25%) | ||||||
| Anastomosis time, min | 119 | 27 (25–33) | 44 | 27 (25–33) | 66 | 27 (23–32) | 0.4916 | 5 | 38 (34–43) | 4 | 26 (26–30) |
| Intraoperative RBC transfusion, units | 108 | 6 (4–9) | 39 | 6 (4–8) | 61 | 6 (4–9) | 0.8815 | 4 | 5 (3–5) | 4 | 7 (5–11) |
| Intraoperative platelet transfusion, units | 101 | 3 (2–4) | 39 | 3 (2–4) | 54 | 3 (2–4) | 0.7459 | 5 | 2 (1–2) | 3 | 4 (4–6) |
| Intraoperative plasma transfusion, units | 112 | 6 (4–10) | 40 | 6 (5–11) | 64 | 6 (3–10) | 0.5462 | 4 | 4 (2–7) | 4 | 10 (3–15) |
| Intraoperative cryoprecipitate transfusion, units | 100 | 4 (2–6) | 36 | 4 (2–6) | 56 | 4 (3–6) | 0.3572 | 5 | 2 (2–2) | 3 | 6 (4–6) |
| Postreperfusion syndrome | |||||||||||
| No | 120 | 80 (67%) | 44 | 23 (52%) | 67 | 50 (75%) | 0.0239 | 5 | 4 (80%) | 4 | 3 (75%) |
| Yes | 40 (33%) | 21 (48%) | 17 (25%) | 1 (20%) | 1 (25%) | ||||||
| Postreperfusion cardiac arrest | |||||||||||
| No | 120 | 119 (99%) | 44 | 44 (100%) | 67 | 66 (99%) | 1.0000 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 1 (1%) | 0 (0%) | 1 (1%) | 0 (0%) | 0 (0%) | ||||||
| Fiscal year | |||||||||||
| 2022/2023 | 120 | 44 (37%) | 44 | 24 (55%) | 67 | 18 (27%) | 0.0049 | 5 | 1 (20%) | 4 | 1 (25%) |
| 2023/2024 | 76 (63%) | 20 (45%) | 49 (73%) | 4 (80%) | 3 (75%) | ||||||
BMI, body mass index; eGFR, estimated glomerular filtration rate; IQR, interquartile range; MASH, metabolic dysfunction-associated steatohepatitis; MELD-Na, model for end-stage liver disease-sodium; NMP, normothermic machine perfusion; NRP, normothermic regional perfusion; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; RBC, red blood cell; SCS, static cold storage.
Clinical Outcomes
In the NMP group, peak alanine transaminase within 7 d was lower (585 versus 1321 U/L, P < 0.0001), peak aspartate transaminase within 7 d was lower (1579 versus 3359 U/L, P < 0.0001), international normalized ratio on postoperative day 7 was lower (1.1 [interquartile range (IQR), 1.0–1.1] versus 1.1 [IQR, 1.0–1.2], P = 0.0056), incidence of early allograft dysfunction (EAD) was lower (42% versus 75%, P = 0.0008), there was lesser need for hemodialysis within 14 d post-LT (1% versus 11%, P = 0.0353), and recipients had higher estimated glomerular filtration rate at 3 mo (73 versus 62 mL/min/1.73 m2, P = 0.0205), compared with the SCS group. There was no difference in total bilirubin on postoperative day 7, length of stay in intensive care unit, total length of stay in the hospital, peak creatinine within 14 d, incidence of acute kidney injury within 14 d, new need for hemodialysis within 3 mo, incidence of bacterial and fungal infections within 3 mo, incidence of clinically treated rejection, biopsy proven rejection, grade of rejection (rejection activity index), and multiple episodes of acute cellular rejection within 6 mo between the NMP and SCS groups. At a median follow-up of 304 d, there was no difference in biliary complications between the NMP and SCS groups, the percentage of patients who underwent endoscopic retrograde cholangiopancreatography, the number of endoscopic retrograde cholangiopancreatographies completes on these patients, percentage of patients who had percutaneous transhepatic catheter, rate of anastomotic stricture, grade of anastomotic stricture, bile leak and IC, and the grade of IC were similar (Table 3).
TABLE 3.
Outcomes
| Total (N = 120) | SCS (N = 44) | NMP (N = 67) | P | NRP SCS (N = 5) | NRP NMP (N = 4) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | N | Median (IQR) or n (%) | |
| Peak ALT within 7 d, U/L | 120 | 898 (355–1568) | 44 | 1321 (766–2439) | 67 | 585 (246–1040) | <0.0001 | 5 | 1195 (321–2400) | 4 | 836 (538–1969) |
| Peak AST within 7 d, U/L | 120 | 2100 (1047–3840) | 44 | 3359 (1847–5245) | 67 | 1579 (659–2839) | <0.0001 | 5 | 1468 (1328–4473) | 4 | 2133 (1539–4321) |
| Bilirubin on POD 7, mg/dL | 118 | 1.4 (0.9–3.2) | 44 | 1.3 (0.9–2.2) | 65 | 1.8 (1.0–3.8) | 0.2540 | 5 | 1.1 (0.9–1.6) | 4 | 2.5 (1.0–5.7) |
| INR on POD 7 | 115 | 1.1 (1.0–1.2) | 44 | 1.1 (1.0–1.1) | 62 | 1.1 (1.0–1.2) | 0.0056 | 5 | 1.0 (1.0–1.1) | 4 | 1.0 (1.0–1.2) |
| PNF | |||||||||||
| No | 120 | 119 (99%) | 44 | 44 (100%) | 67 | 66 (99%) | 1.0000 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 1 (1%) | 0 (0%) | 1 (1%) | 0 (0%) | 0 (0%) | ||||||
| EAD | |||||||||||
| No | 115 | 52 (45%) | 44 | 11 (25%) | 62 | 36 (58%) | 0.0008 | 5 | 3 (60%) | 4 | 2 (50%) |
| Yes | 63 (55%) | 33 (75%) | 26 (42%) | 2 (40%) | 2 (50%) | ||||||
| Length of ICU stay (initial ICU stay), d | 120 | 2 (1–3) | 44 | 2 (1–3) | 67 | 2 (1–3) | 0.9648 | 5 | 1 (1–2) | 4 | 3 (2–7) |
| Total length of hospital stay, d | 120 | 8 (7–11) | 44 | 8 (7–11) | 67 | 8 (7–10) | 0.3258 | 5 | 8 (7–12) | 4 | 7 (5–14) |
| Renal function | |||||||||||
| Peak creatinine within 14 d, mg/dL | 120 | 1.60 (0.92–2.54) | 44 | 1.71 (1.36–2.60) | 67 | 1.57 (0.84–2.50) | 0.1055 | 5 | 2.84 (1.54–5.52) | 4 | 1.68 (0.84–2.46) |
| AKI | |||||||||||
| None | 120 | 96 (80%) | 44 | 33 (75%) | 67 | 55 (82%) | 0.4530 | 5 | 4 (80%) | 4 | 4 (100%) |
| Stage 1 | 15 (13%) | 6 (14%) | 9 (13%) | 0 (0%) | 0 (0%) | ||||||
| Stage 2 | 8 (7%) | 4 (9%) | 3 (4%) | 1 (20%) | 0 (0%) | ||||||
| Stage 3 | 1 (1%) | 1 (2%) | 0 (0%) | 0 (0%) | 0 (0%) | ||||||
| New need for HD/CRRT within 14 d | |||||||||||
| No | 120 | 113 (94%) | 44 | 39 (89%) | 67 | 66 (99%) | 0.0353 | 5 | 4 (80%) | 4 | 4 (100%) |
| Yes | 7 (6%) | 5 (11%) | 1 (1%) | 1 (20%) | 0 (0%) | ||||||
| 3-mo outcomes | |||||||||||
| eGFR, mL/min/1.73 m2 | 112 | 66 (47–90) | 43 | 62 (28–79) | 61 | 73 (51–90) | 0.0205 | 5 | 57 (44–70) | 3 | 90 (66–90) |
| New need for HD (3 mo) | |||||||||||
| No | 113 | 102 (90%) | 43 | 36 (84%) | 62 | 59 (95%) | 0.0870 | 5 | 4 (80%) | 3 | 3 (100%) |
| Yes | 11 (10%) | 7 (16%) | 3 (5%) | 1 (20%) | 0 (0%) | ||||||
| Bacterial infections- treated | |||||||||||
| No | 112 | 106 (95%) | 43 | 41 (95%) | 61 | 57 (93%) | 1.0000 | 5 | 5 (100%) | 3 | 3 (100%) |
| Yes | 6 (5%) | 2 (5%) | 4 (7%) | 0 (0%) | 0 (0%) | ||||||
| Fungal infections | |||||||||||
| No | 112 | 111 (99%) | 43 | 42 (98%) | 61 | 61 (100%) | 0.4135 | 5 | 5 (100%) | 3 | 3 (100%) |
| Yes | 1 (1%) | 1 (2%) | 0 (0%) | 0 (0%) | 0 (0%) | ||||||
| 6-mo outcomes | |||||||||||
| ACR (clinically treated) | |||||||||||
| No | 87 | 65 (75%) | 37 | 28 (76%) | 43 | 30 (70%) | 0.6212 | 4 | 4 (100%) | 3 | 3 (100%) |
| Yes | 22 (25%) | 9 (24%) | 13 (30%) | 0 (0%) | 0 (0%) | ||||||
| Biopsy-treated ACR rejection | |||||||||||
| No | 87 | 74 (85%) | 37 | 32 (86%) | 43 | 35 (81%) | 0.7623 | 4 | 4 (100%) | 3 | 3 (100%) |
| Yes | 13 (15%) | 5 (14%) | 8 (19%) | 0 (0%) | 0 (0%) | ||||||
| Biopsy RAI | 13 | 4 (4–5) | 5 | 4 (4–5) | 8 | 5 (4–6) | 0.7041 | 0 | NA | 0 | NA |
| Multiple episodes | |||||||||||
| No | 87 | 82 (94%) | 37 | 36 (97%) | 43 | 39 (91%) | 0.3662 | 4 | 4 (100%) | 3 | 3 (100%) |
| Yes | 5 (6%) | 1 (3%) | 4 (9%) | 0 (0%) | 0 (0%) | ||||||
| 12-mo outcomes | |||||||||||
| Readmission | |||||||||||
| No | 106 | 35 (33%) | 40 | 13 (33%) | 58 | 21 (36%) | 0.8296 | 5 | 0 (0%) | 3 | 1 (33%) |
| Yes | 71 (67%) | 27 (68%) | 37 (64%) | 5 (100%) | 2 (67%) | ||||||
| Need for kidney transplant | |||||||||||
| No | 106 | 104 (98%) | 40 | 38 (95%) | 58 | 58 (100%) | 0.1641 | 5 | 5 (100%) | 3 | 3 (100%) |
| Yes | 2 (2%) | 2 (5%) | 0 (0%) | 0 (0%) | 0 (0%) | ||||||
| Biliary complications | |||||||||||
| Median time of follow-up for biliary complications, d | 120 | 304 (178–459) | 44 | 410 (253–492) | 67 | 262 (151–432) | 0.0066 | 5 | 330 (288–332) | 4 | 207 (120–455) |
| ERCP | |||||||||||
| No | 118 | 75 (64%) | 44 | 25 (57%) | 65 | 45 (69%) | 0.2235 | 5 | 3 (60%) | 4 | 2 (50%) |
| Yes | 43 (36%) | 19 (43%) | 20 (31%) | 2 (40%) | 2 (50%) | ||||||
| No. of ERCP (if yes) | 43 | 2 (2–4) | 19 | 2 (2–3) | 20 | 3 (2–4) | 0.3004 | 2 | 2 (1–2) | 2 | 3 (1–4) |
| PTC | |||||||||||
| No | 118 | 116 (98%) | 44 | 44 (100%) | 65 | 63 (97%) | 0.5141 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 2 (2%) | 0 (0%) | 2 (3%) | 0 (0%) | 0 (0%) | ||||||
| Anastomotic stricture | |||||||||||
| No | 118 | 86 (73%) | 44 | 31 (70%) | 65 | 49 (75%) | 0.6599 | 5 | 3 (60%) | 4 | 3 (75%) |
| Yes | 32 (27%) | 13 (30%) | 16 (25%) | 2 (40%) | 1 (25%) | ||||||
| Anastomotic stricture, grade | |||||||||||
| No | 118 | 86 (73%) | 44 | 31 (70%) | 65 | 49 (75%) | 0.2130 | 5 | 3 (60%) | 4 | 3 (75%) |
| Mild | 7 (6%) | 1 (2%) | 6 (9%) | 0 (0%) | 0 (0%) | ||||||
| Moderate | 8 (7%) | 3 (7%) | 4 (6%) | 0 (0%) | 1 (25%) | ||||||
| Severe | 17 (14%) | 9 (20%) | 6 (9%) | 2 (40%) | 0 (0%) | ||||||
| Bile leak | |||||||||||
| No | 118 | 108 (92%) | 44 | 40 (91%) | 65 | 60 (92%) | 1.0000 | 5 | 5 (100%) | 4 | 3 (75%) |
| Yes | 10 (8%) | 4 (9%) | 5 (8%) | 0 (0%) | 1 (25%) | ||||||
| Ischemic cholangiopathy | |||||||||||
| No | 118 | 108 (92%) | 44 | 40 (91%) | 65 | 59 (91%) | 1.0000 | 5 | 5 (100%) | 4 | 4 (100%) |
| Yes | 10 (8%) | 4 (9%) | 6 (9%) | 0 (0%) | 0 (0%) | ||||||
| Ischemic cholangiopathy, grade | |||||||||||
| B | 10 | 2 (20%) | 4 | 1 (2%) | 6 | 1 (1%) | N/A | 0 | 0 (0%) | 0 | 0 (0%) |
| C | 1 (10%) | 0 (0%) | 1 (1%) | 0 (0%) | 0 (0%) | ||||||
| D | 6 (60%) | 3 (7%) | 3 (5%) | 0 (0%) | 0 (0%) | ||||||
| E | 1 (10%) | 0 (0%) | 1 (1%) | 0 (0%) | 0 (0%) | ||||||
| Days out of hospital, mo | |||||||||||
| 3 | 113 | 80 (73–83) | 43 | 80 (70–84) | 62 | 81 (74–83) | 0.9792 | 5 | 80 (79–82) | 3 | 82 (50–86) |
| 6 | 88 | 172 (161–175) | 37 | 170 (157–175) | 44 | 173 (164–175) | 0.2891 | 4 | 173 (170–174) | 3 | 174 (142–178) |
| 12 | 49 | 352 (339–357) | 26 | 350 (338–357) | 21 | 352 (341–357) | 0.6524 | 1 | 356 (356–356) | 1 | 324 (324–324) |
| Acquisition costs, $, fiscal year | |||||||||||
| 2022/2023 | 44 | 79 213 | 24 | 79 213 | 18 | $79 213 | NA | 1 | $79 213 | 1 | $79 213 |
| 2023/2024 | 76 | $98 036 | 20 | $98 036 | 49 | $98 036 | NA | 4 | $98 036 | 3 | $98 036 |
| Direct costs, $ | |||||||||||
| Index admission | 120 | $45 849 ($36 972–$56 307) | 44 | $45 175 ($36 800–$55 091) | 67 | $46 131 ($38 123–$57 289) | 0.6885 | 5 | $40 091 ($35 596–$47 983) | 4 | $45 705 ($34 551–$83 411) |
| 3 mo | 113 | $53 777 ($42 616–$75 262) | 43 | $51 585 ($42 616–$74 595) | 62 | $56 755 ($43 215–$79 697) | 0.6886 | 5 | $46 857 ($41 147–$50 673) | 3 | $53 292 ($42 581–$155 792) |
| 6 mo | 88 | $59 252 ($44 376–$81 415) | 37 | $57 398 ($48 340 – $81 328) | 44 | $60 383 ($44 079–$84 011) | 0.6940 | 4 | $44 506 ($38 386–$56 305) | 3 | $54 180 ($43 253–$161 525) |
| 12 mo | 49 | $62 610 ($43 982–$96 334) | 26 | $71 010 ($52 929 – $96 334) | 21 | $62 461 ($43 982–$89 954) | 0.7240 | 1 | $37 755 (NA) | 1 | $162 559 (NA) |
ACR, acute cellular rejection; AKI, acute kidney injury; ALT, alanine transaminase; AST, aspartate transaminase; CRRT, continuous renal replacement therapy; EAD, early allograft dysfunction; eGFR, estimated glomerular filtration rate; ERCP, endoscopic retrograde cholangiopancreatography; HD, hemodialysis; ICU, intensive care unit; INR, international normalized ratio; IQR, interquartile range; NMP, normothermic machine perfusion; NRP, normothermic regional perfusion; PNF, primary nonfunction; POD, postoperative day; PTC, percutaneous transhepatic catheter; RAI, rejection activity index; SCS, static cold storage.
The readmission rate within 12 mo; need for kidney transplantation within 12 mo; days alive; and spent out of the hospital at 3, 6, and 12 mo were similar between the NMP and SCS groups (Table 3). Time to failure and the proportion of recipients with graft or patient survival at 6 and 12 mo were also similar (Figure 2).
FIGURE 2.
Kaplan-Meier graft and patient survival. Kaplan-Meier analyses were used to estimate graft and patient survival proportions. Proportions of graft and patient survival were similar between the NMP and SCS groups. CI, confidence interval; NMP, normothermic machine perfusion; SCS, static cold storage.
Cost Analysis
In the bivariate analysis, direct costs associated with DCD LT were similar between NMP and SCS groups, at index admission ($46 131 versus $45 175), 3 mo ($56 755 versus $51 585), 6 mo ($60 383 versus $57 398), and 12 mo ($62 461 versus $71 010; Table 3). A longitudinal multiple regression was fitted to estimate the estimated cumulative direct costs at index admission ($59 528 versus $67 120), 3 mo ($73 753 versus $79 041), 6 mo ($82 138 versus $96 956), and 12 mo ($96 382 versus $101 768), and there was no difference between the groups (Figure 3). Based on the multivariate analysis, the baseline variables CIT and PRS were the only significant variables that increased the direct costs posttransplant. An increase in CIT by 1 SD (0.93 h or 56 min) increased the direct cost by $14 700 (95% confidence interval, $4020-$25 300, P = 0.0074), and the presence of PRS increased the direct cost by $23 100 (95% confidence interval, $3250-$43 000, P = 0.0231). Other variables, such as donor age, donor BMI, US DRI, distance between the donor and recipient center, recipient age, MELD-Na, use of local recovery, agonal time, use of NMP, surgery start time, and the direct costs between NMP and SCS at different time points, did not significantly impact posttransplant direct costs (Figure 4).
FIGURE 3.
Estimated cumulative direct costs adjusted for baseline covariates by group. A longitudinal multiple regression was fitted to calculate cumulative direct costs across 4 time points (index admissions, 3 mo, 6 mo, and 12 mo). There was no difference in cumulative direct costs between the NMP and SCS groups across 4 time points (index admissions, 3 mo, 6 mo, and 12 mo). NMP, normothermic machine perfusion; SCS, static cold storage.
FIGURE 4.
Forest plot of multivariate analysis depicting mean shift in cumulative postoperative direct costs. One standard deviation change of a continuous variable or the presence of a categorical variable is associated with a shift in cumulative postoperative direct costs by a given estimate. BMI, body mass index; MELD-Na, Model for End-stage Liver Disease-sodium; NMP, normothermic machine perfusion; SCS, static cold storage; USDRI, US donor risk index.
Trends in LT, Nationally and at Our Center
Nationally, since the latter half of the year 2021, when both the major NMP platforms received Food and Drug Administration approval for clinical use in the United States, and use of NRP for clinical use was on the rise, the following trends were observed: steady rise in the DCD LT volumes, decrease in median wait time to transplant, and increase in the waitlist survival probability. Our center started using NMP technology in August 2022. We observed an increase in the DCD transplant volume and a decrease in the median wait time to transplant, whereas the waitlist survival probability remained stable (Figure 5).
FIGURE 5.
Temporal trends in deceased donor liver transplant volumes, wait time to transplant, and waitlist survival probability. National data: Transmedics OCS and OrganOx metra got FDA approval for clinical use in the United States in September 2021 and December 2021, respectively. Temporal trends depict an increase in the DCD liver transplant volume, a decrease in the wait time to liver transplant, and an increase in waitlist survival probability. Local data: OrganOx metra was used by our center from August 19, 2022. Temporal trends depict an increase in the DCD liver transplant volume and a decrease in the wait time to transplant, whereas the waitlist survival probability remained the same. Data are based on Organ Procurement and Transplantation Network (OPTN) data with follow-up through September 30, 2024. DCD, donation after circulatory death; FDA, Food and Drug Administration.
DISCUSSION
The use of back-to-base NMP in DCD LT was associated with decreased incidence of EAD, PRS, and renal replacement therapy in the early postoperative period, and better native kidney function at 3 mo. Although the use of NMP did not decrease postoperative direct costs, longer CIT and the presence of PRS significantly increased postoperative direct costs. At our center, the use of back-to-base NMP has enabled greater utilization of DCD liver grafts, use of high-risk DCD liver grafts, use of DCD liver grafts procured by local recovery teams, and use of DCD liver grafts from longer distances without compromising patient and graft outcomes while decreasing the wait time to LT at our center. Nationally, observed trends that correlate with normothermic perfusion techniques include a rise in DCD LT volumes, a decrease in median wait time to transplant, and an increase in waitlist survival.
Wehrle et al3 published their data about pre- and posttransplant costs associated with back-to-base NMP in both DBD and DCD LT. They identified no difference in pre- and post-LT costs with the use of NMP and also found that recipients with greater severity of complications incurred higher costs. These results encourage increased utilization of NMP, especially for DCD LTs. With respect to DCD LTs reported in this study, the only limitation we see is the comparison of SCS liver grafts transplanted during the first 3 y and 8 mo of their study period to NMP LTs performed during the past 1 y, with very few DCD SCS LTs performed since the start of NMP. Therefore, comparing 74 DCD SCS LTs performed between January 2019 and September 2022 to 37 DCD NMP LTs performed in 1 y (October 2022–September 2023).
The biggest strength of our study is that we present clinical outcomes of DCD LT at a granular level after NMP, from a single center that performs high-volume DCD LTs in the United States, using the back-to-base NMP model. We also performed a significant number of DCD LTs with SCS alone, in the era of NMP. This enabled us to report clinical outcomes and postoperative costs, comparing the NMP and SCS DCD LTs performed within the same time frame. Our study also analyzes variables that influence postoperative costs after NMP in DCD LT, which has not been explored previously. Assessing liver graft viability on NMP has allowed us to implement local recovery for DCD procurements. This approach allowed us to accept organs from donors who were unlikely to pass within predetermined time criteria. As a result, we significantly reduced the occurrence of “dry runs,” a major cost burden in DCD LT. However, our study did not evaluate the financial impact of this reduction in “dry runs.” A few of the limitations of this study are the retrospective nature of this study from a single center, small sample size, and lack of viability assessment data of NMP liver grafts. Viability assessment and DCD liver graft selection were not the focus of our study and will be reported in our future studies. Another limitation of our study was the lack of data on the overall cost savings achieved with NMP. The economic implications of increasing the transplant volume, reducing wait times for transplant, transplanting patients at low MELD scores, thereby reducing waitlist morbidity, hospital admissions events, and the cost savings associated with it have not been explored.
Results from our study should be interpreted with caution. We were able to achieve excellent results from DCD LT before the NMP era, with 1-y graft survival anywhere between 90% and 92% while keeping ischemic cholangiopathy (IC) rates of <3%.15 We attribute these results to careful donor selection and keeping our CIT short. We continue to follow those rigid criteria especially with DCD SCS LTs. Our prior experience and good results with DCD LT with SCS and the use of higher-risk DCD liver grafts in the NMP group were probably the reason we did not see any significant improvement in clinical and cost outcomes with NMP. With the use of NMP, we have expanded our DCD donor criteria significantly, where our average DCD donor age increased from about 38 y old in the pre-NMP era to 47 y in this study, and the US DRI increased from 1.82 in the pre-NMP era to 2.33 in this study.16 Expanding our donor criteria may also have attributed to the increased incidence of IC we see in our study, which was 8%. The presence of IC can be associated with increased postoperative costs.17 Utility of sequential NRP followed by NMP, hypothermic machine perfusion (HMP) during transit, and NMP at the donor hospital should be explored in the reduction of IC rates and the costs associated with it.
Should all DCD liver grafts undergo NMP? This is a question we keep asking ourselves periodically. In a recent survey, about 1 in 4 transplant centers in the United States, which performed DCD LTs, did not use NMP technology. 53% of the centers that performed DCD LTs routinely used NMP for all DCD LTs, whereas 47% of the centers used NMP only when indicated.18 At our center, we have slowly moved toward using NMP for the majority of our DCD LTs. The decreased incidence of EAD, PRS, use of DCD liver grafts procured by local recovery teams, high-risk recipients with either high MELD score or previous abdominal surgery, ability to manage transplant logistics, and performing transplants during the daytime are the main factors that drive our NMP use. However, we perform a notable number of DCD LTs using SCS alone, whereas keeping the CIT short, if it is from a young standard criteria donor, can be performed during the daytime; there are multiple transplants to perform on the same day and if the recipients are low risk and not complex. Just the adoption of normothermic perfusion technology by centers, if not for all their DCD LTs, has enabled an increase in DCD LT volumes, reduced waiting time to LT, and increased waitlist survival probability, both nationally and at our center.
Our study findings suggest that CIT was associated with increased postoperative costs, where approximately an additional hour of CIT would lead to an increase of postoperative direct costs by about $14 700. It is well known that longer CIT is associated with worse short-term allograft and patient outcomes, especially in DCD LT.19,20 In our study, CIT in the NMP group ranged from 2.4 to 6.8 h. Emerging data suggest that a prolonged time from cross-clamp to start of NMP (>2 h 45 min) was associated with inferior short-term outcomes.21 Although common knowledge, it has recently been shown that increased postoperative complications are associated with higher postoperative costs, especially in DCD LT.3 So efforts should be made in keeping the CITs short after DCD LT, irrespective of the use of back-to-base NMP, to improve short-term allograft and patient outcomes and potentially reduce postoperative costs. Although the notion is speculative, it is reasonable to consider that the cost savings with “NMP alone at the recipient center” may be counterbalanced by the complications and associated expenses with prolonged CIT. It remains to be explored whether sequential NRP followed by NMP, use of HMP in transit, or NMP at the donor hospital offers a more optimal solution. Our findings also suggest that the presence of PRS was associated with an increase in postoperative direct cost by $23 100. This may partly be explained by the additional pharmaceuticals used in the management of PRS, such as angiotensin II, vitamin B and C infusion, prothrombin complex concentrate, and fibrinogen concentrate, with the costs of angiotensin II, vitamin B, and vitamin C being very expensive. Our study did not explore the individual pharmaceutical costs and the quantity of pharmaceuticals used.
Resuscitation of liver grafts with the use of NRP, along with the start of NMP at the donor center, with minimal interruption to resuscitation, may look like the ideal scenario for the use of DCD liver grafts. Wide variation in experience and practice regarding NRP use,22 along with challenging logistics and increased costs associated with NMP use at the donor center, preclude centers from universally adopting this ideal scenario. Standardization of NRP use and its techniques may encourage widespread adoption among Organ Procurement Organizations. We are not entirely sure how logistics and costs can simultaneously be controlled using NMP at donor centers. Intermediary solutions that address the SCS before back-to-base NMP, such as keeping the CIT short or the use of HMP during organ transport, should be considered. Although excellent clinical outcomes can be achieved with the use of HMP in comparison with SCS alone,23 the clinical utility of HMP either in conjunction with NMP or as an alternative to NMP needs to be explored in future studies.
In conclusion, normothermic perfusion technology has enabled increased DCD LT volumes with increased use of high-risk DCD liver grafts, reduced wait time to LT, and increased waitlist survival probability without adversely affecting short-term DCD LT outcomes. Although the use of back-to-base NMP did not decrease postoperative costs, longer CIT was associated with higher postoperative costs. Strategies such as sequential NRP followed by NMP, keeping the CIT short, and the use of HMP may be helpful in decreasing postoperative costs, along with improving clinical outcomes. Standardization of normothermic perfusion technologies can improve widespread adoption of these technologies by the Organ Procurement Organizations and the transplant centers.
ACKNOWLEDGMENTS
The authors thank all donors, recipients, and their families, without whom this work would not be possible.
Supplementary Material
Footnotes
The authors declare no funding or conflicts of interest.
S.R.P., W.K.W., A.D.S., and M.A. participated in conception or design of work. S.R.P., A.J.L., M.I., L.V.S., L.G., A.N., W.C., A.O., A.L., N.S., and S.B. participated in acquisition, analysis, or interpretation of data for the work. S.R.P., A.J.L., A.D.S., and M.A. drafted the work. M.I., L.V.S., L.G., A.N., W.C., A.O., A.L., N.S., S.B., and W.K.W. reviewed the draft critically for important intellectual content. All authors approve of the final version to be published and agree to be accountable for all aspects of work.
Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantationdirect.com).
Contributor Information
April J. Logan, Email: april.logan@osumc.edu.
Manoj Iyer, Email: manoj.iyer@osumc.edu.
Lauren Von Stein, Email: lauren.vergara@osumc.edu.
Leonid Gorelik, Email: leonid.gorelik@osumc.edu.
Annelise Nolan, Email: annelise.nolan@osumc.edu.
Wei Chen, Email: wei.chen2@osumc.edu.
Ayato Obana, Email: ayato.obana@osumc.edu.
Ashley Limkemann, Email: ashley.limkemann@osumc.edu.
Navdeep Singh, Email: navdeep.singh@osumc.edu.
Sylvester Black, Email: sylvester.black@osumc.edu.
William K. Washburn, Email: ken.washburn@osumc.edu.
Austin D. Schenk, Email: austin.schenk@osumc.edu.
Musab Alebrahim, Email: musab.alebrahim@osumc.edu.
REFERENCES
- 1.Organ Procurement and Transplantation Network. National data. Available at https://optn.transplant.hrsa.gov/data/view-data-reports/build-advanced. Accessed April 11, 2024.
- 2.Oniscu GC, Mehew J, Butler AJ, et al. Improved organ utilization and better transplant outcomes with in situ normothermic regional perfusion in controlled donation after circulatory death. Transplantation. 2023;107:438–448. [DOI] [PubMed] [Google Scholar]
- 3.Wehrle CJ, Zhang M, Khalil M, et al. Impact of back-to-base normothermic machine perfusion on complications and costs: a multicenter, real-world risk-matched analysis. Ann Surg. 2024;280:300–310. [DOI] [PubMed] [Google Scholar]
- 4.Brubaker AL, Sellers MT, Abt PL, et al. US liver transplant outcomes after normothermic regional perfusion vs standard super rapid recovery. JAMA Surg. 2024;159:677–685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gaurav R, Butler AJ, Kosmoliaptsis V, et al. Liver transplantation outcomes from controlled circulatory death donors: SCS vs in situ NRP vs ex situ NMP. Ann Surg. 2022;275:1156–1164. [DOI] [PubMed] [Google Scholar]
- 6.Hefler J, Leon-Izquierdo D, Marfil-Garza BA, et al. Long-term outcomes after normothermic machine perfusion in liver transplantation—experience at a single North American center. Am J Transplant. 2023;23:976–986. [DOI] [PubMed] [Google Scholar]
- 7.Bakhtiyar SS, Maksimuk TE, Gutowski J, et al. Association of procurement technique with organ yield and cost following donation after circulatory death. Am J Transplant. 2024;24:1803–1815. [DOI] [PubMed] [Google Scholar]
- 8.Webb AN, Izquierdo DL, Eurich DT, et al. The actual operative costs of liver transplantation and normothermic machine perfusion in a Canadian Setting. PharmacoEcon Open. 2021;5:311–318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Saharia A. The cost of perfusing organs—is it worth it? Available at https://cdn.ymaws.com/transplanttexas.org/resource/resmgr/events/2021/doc_saharia__CostofPerfusing.pdf. Accessed November 11, 2024. [Google Scholar]
- 10.Feng S, Goodrich NP, Bragg-Gresham JL, et al. Characteristics associated with liver graft failure: the concept of a donor risk index. Am J Transplant. 2006;6:783–790. [DOI] [PubMed] [Google Scholar]
- 11.Batts KP, Ludwig J. Chronic hepatitis. An update on terminology and reporting. Am J Surg Pathol. 1995;19:1409–1417. [DOI] [PubMed] [Google Scholar]
- 12.Neil DAH, Minervini M, Smith ML, et al. Banff consensus recommendations for steatosis assessment in donor livers. Hepatology. 2022;75:1014–1025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Croome KP, Mathur AK, Aqel B, et al. Classification of distinct patterns of ischemic cholangiopathy following DCD liver transplantation: distinct clinical courses and long-term outcomes from a multicenter cohort. Transplantation. 2022;106:1206–1214. [DOI] [PubMed] [Google Scholar]
- 14.Frasco PE, Mathur AK, Chang YH, et al. Days alive and out of hospital after liver transplant: comparing a patient-centered outcome between recipients of grafts from donation after circulatory and brain deaths. Am J Transplant. 2023;23:55–63. [DOI] [PubMed] [Google Scholar]
- 15.Limkemann AJ, Singh N, Helfrich K, et al. Safely expanding the liver donor pool by utilization of organs from donation after circulatory death with comparable results to donation after brain death, a large single-center experience. J Gastrointest Surg. 2022;26:1453–1461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Punjala SR, Logan AJ, Subramanian J, et al. Outcomes of liver transplantation from hepatitis C virus–positive DCD donors and its utilization among centers in the United States. Transplantation. 2025;109:186–195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Jay CL, Lyuksemburg V, Kang R, et al. The increased costs of donation after cardiac death liver transplantation: caveat emptor. Ann Surg. 2010;251:743–748. [DOI] [PubMed] [Google Scholar]
- 18.Punjala SR, Logan A, Han J, et al. Variation in DCD liver transplant protocols among transplant centers in the United States. Transplant Direct. 2024;10:e1650. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Sibulesky L, Li M, Hansen RN, et al. Impact of cold ischemia time on outcomes of liver transplantation: a single center experience. Ann Transplant. 2016;21:145–151. [DOI] [PubMed] [Google Scholar]
- 20.Paterno F, Guarrera JV, Wima K, et al. Clinical implications of donor warm and cold ischemia time in donor after circulatory death liver transplantation. Liver Transpl. 2019;25:1342–1352. [DOI] [PubMed] [Google Scholar]
- 21.Lee C, Mathur AK, Mao S, et al. Prolonged time from cross-clamp until normothermic machine perfusion start is associated with an increased risk of early allograft dysfunction following DCD liver transplant. Liver Transpl. 2025;31:616–622. [DOI] [PubMed] [Google Scholar]
- 22.Sellers MT, Philip JL, Brubaker AL, et al. Normothermic regional perfusion experience of organ procurement organizations in the US. JAMA Netw Open. 2024;7:e2440130–e2440130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rijn R, Schurink IJ, Vries Y, et al. Hypothermic machine perfusion in liver transplantation—a randomized trial. N Engl J Med. 2021;384:1391–1401. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.