Abstract
Background
Mechanical preservation (MP) of deceased donor kidney transplants showed a 30% to 50% reduction in delayed graft function (DGF) as defined by dialysis in the first week, when compared with cold storage. DGF is associated with longer hospital stays and increased costs. In this study, we sought to understand the impact of MP on rates of DGF and length of hospital stays in a contemporary cohort of deceased donor kidney transplants in the United States.
Material/Methods
All single deceased donor kidney transplants performed between January 1, 2010, and September 2, 2020, were identified in the Scientific Registry of Transplant Recipients database. Donor kidneys were considered pumped if the transplant center received the kidney on the pump.
Results
Multivariate logistic regression showed that MP had similar odds of reduction of DGF for all subsets of donors. The unadjusted rate of DGF for pumped brain-dead standard criteria donor (BDSCD) recipients was similar to that of donors stored on ice. The rate of DGF for expanded criteria donors (ECD) and donors after cardiac death (DCD) was lower in the recipients who received MP. The similar DGF rates in BDSCD donor recipients were due to longer cold ischemia times in MP kidneys. The lower DGF rates seen in ECD and DCD recipients of pumped kidneys did not translate into a shortened length of hospitalization after transplant.
Conclusions
As currently deployed, only DCD and ECD donor recipients of MP kidneys experienced a lower DGF rate. In all subsets of patients, MP did not appreciably shorten the hospital length of stay.
Keywords: Delayed Graft Function, Kidney Transplantation, Length of Stay, Organ Preservation
Introduction
Mechanical preservation (MP) of transplant organs promises to reduce injury to donor organs, potentially promote graft repair, and improve graft outcomes. Although some paired studies of hypothermic MP of deceased donor organs have suggested improved graft survival in some types of deceased donors, registry data examining hypothermic pulsatile MP use outside of research protocols have not shown a graft survival benefit associated with the current MP systems used in deceased donor kidney transplants in the United States [1,2]. Newer normothermic and oxygenated MP systems are being developed and have the prospect of improving preservation but are more costly and resource-intensive and will require validation of their efficacy in deceased donor kidney preservation.
Prospective and retrospective studies of MP with hypothermic pulsatile perfusion of deceased donor kidney transplants showed a 30% to 50% reduction in delayed graft function (DGF), as defined by the need for dialysis in the first week after trans plant [3–13]. DGF is associated with longer hospital stays after transplant, increased early readmission, and poorer long-term graft outcomes [14–16]. MP of deceased donor kidneys is significantly more expensive than cold storage of kidneys for transport. At our institution, MP adds $3000 to each procured kidney, and that does not include the upfront cost of the machines or replacement of disposables. One argument for a more expansive use of MP is the possibility that lowering the DGF rate would reduce hospital length of stays and offset the additional cost associated with this preservation method. In this study, we sought to understand the impact of cold pulsatile perfusion preservation on rates of DGF and length of hospital stays in a contemporary cohort of deceased donor kidney transplants in the United States.
Material and Methods
All single deceased donor kidney transplants performed between January 1, 2010, and September 2, 2020, were identified in the Scientific Registry of Transplant Recipients database standard analysis files. Multiorgan transplants, including kidney and dual or en bloc kidney transplants, were excluded since DGF rates are likely influenced by the early function of the non-kidney organ transplanted, while en bloc transplants have a higher rate of thrombosis, which would likely increase DGF but not necessarily be related to ischemia-reperfusion injury. Dual adult deceased donor transplants are rare, and implantation of 2 kidneys makes the analysis of cold ischemia time for the transplant problematic in analysis since each kidney procured would have a different cold ischemia time. Transplants missing data for preservation type (MP or ice) and/or DGF, as defined by dialysis in the first week after transplant, were excluded from the study. Donors were considered MP if the transplanting center received the donor kidney on the pump. Donor kidneys received on ice but then placed on MP were considered to be cold stored. The kidney donor profile index (KDPI; 2022 Mapping Table) of each donor was determined to stratify donor risk of DGF. Length of hospital stay was calculated from admission and discharge dates.
To determine the effect of MP on DGF, logistic regression was conducted. Subset analysis was conducted for brain-dead standard criteria donors (BDSCD), donors after cardiac death (DCD), and expanded criteria donors (ECD). Cold ischemia time was also analyzed to determine its impact on DGF in the setting of MP. Statistical analyses were performed with IBM SPSS Version 28.
Results
During the study period, 126 695 single deceased donor kidney transplants were identified in the dataset. For organ preservation and DGF, 1.3% and 1.4%, respectively, of transplants were missing data and were excluded from the analysis, leaving a study population of 124 903. Table 1 shows the characteristics of the donors based on whether they were received on ice or MP. The baseline characteristics of the donors were similar with regard to sex, race, and ABO type. The age of the donors was significantly higher in the MP donors, with a median age of 42 years, compared with 37 years for donors of organs received on ice. The MP donors were also more likely to have hypertension or diabetes mellitus, die of a cerebrovascular accident, have a higher creatinine level, and be donors after cardiac death. These donor differences translated into a higher median KDPI for the MP donors of 46%, compared with 37% for donors received on ice. The recipient characteristics were similar for sex, race, ethnicity, cause of end-stage renal disease, and ABO type. The recipients receiving MP kidneys were slightly older at 56 years, compared with 53 years for those with kidneys received on ice. The recipients of MP kidneys had a slightly longer median dialysis time. Recipients of kidneys received on ice were more likely to have panel reactive antibodies of greater than or equal to 80%. The median cold ischemia time for MP kidneys was significantly longer at 19 h, compared with 15 h for the kidneys received on ice. Pumped kidneys were more likely to be transplanted in the same organ procurement organization as they were procured. The DGF was 26.7% for kidneys received on ice and 27.6% for kidneys received on the pump.
Table 1.
Characteristic of study population based on preservation type.
| On ice (N=85,003) | On pump (N=39,900) | |
|---|---|---|
| Donor characteristics | ||
| Median age in years (25, 75%) | 37 (25, 49) | 42 (29, 53) |
| Gender Male | 61.3% | 61.8% |
| SRTR race | ||
| Asian | 2.6% | 1.9% |
| Black | 14.4% | 13.6% |
| Multi-Racial | 0.4% | 0.2% |
| Native American | 0.6% | 0.7% |
| Pacific Islander | 0.4% | 0.2% |
| White | 81.6% | 83.4% |
| Ethnicity Latino | 15.9% | 13.1% |
| ABO type | ||
| A | 36.6% | 37.7% |
| AB | 3.5% | 3.4% |
| B | 11.8% | 11.9% |
| O | 48.0% | 47.0% |
| Hypertension | 25.2% | 32.3% |
| Diabetes mellitus | 6.4% | 9.1% |
| Died of cerebrovascular accident | 26.5% | 28.1% |
| Donor after cardiac death | 14.7% | 31.9% |
| Median creatinine in mg/dL (25, 75%) | 0.9 (0.7, 1.3) | 0.98 (0.7, 1.43) |
| Median KDPI (25, 75%) | 37% (17%, 60%) | 46% (24%, 68%) |
| Recipient characteristics | ||
| Median age in years (25, 75%) | 53 (41, 63) | 56 (45, 64) |
| Gender Male | 59.3% | 61.7% |
| SRTR race | ||
| Asian | 7.3% | 6.1% |
| Black | 32.7% | 35.5% |
| Multi-Racial | 0.6% | 0.7% |
| Native American | 0.9% | 1.3% |
| Pacific Islander | 0.5% | 0.4% |
| White | 58.0% | 55.9% |
| Ethnicity Latino | 19.2% | 17.1% |
| On ice (N=85,003) | On pump (N=39,900) | |
| ABO type | ||
| A | 35.3% | 35.5% |
| AB | 5.3% | 5.2% |
| B | 13.5% | 13.4% |
| O | 45.9% | 45.8% |
| Percent preemptive transplant | 10.8% | 10.3% |
| Median time on dialysis (25, 75%) | 3.65 (1.49, 6.10) | 3.72 (1.64, 5.88) |
| Cause of ESRD | ||
| Diabetes mellitus | 26.2% | 30.2% |
| Primary glomerulonephritis | 17.4% | 16.3% |
| Hypertension | 23.8% | 25.2% |
| Other | 25.0% | 20.7% |
| Polycystic kidney disease | 7.6% | 7.6% |
| PRA greater than 79% | 12.4% | 8.9% |
| Transplant characteristics | ||
| Median cold ischemia time in hours (25, 75%) | 15.00 (10.00, 21.00) | 19.00 (14.18, 24.72) |
| Median HLA mismatch for A, B and DR | 4.0 | 4.0 |
| Transplant in same OPO as donation | 71.2% | 79.0% |
| Delayed graft function | 26.9% | 27.7% |
Figure 1 shows the rate of DGF based on the KDPI of the donor. The rate of DGF, as defined by the need for dialysis in the first week, was very similar across the KDPI spectrum, until the KDPI approached 80%, when the kidneys received on MP had a lower rate of DGF. Figure 2 shows the rate of DGF among DCD donors, and the rate of DGF is lower in kidneys pumped across the KDPI spectrum.
Figure 1.
Delayed graft function (DGF) rate in all deceased donor recipients based on the kidney donor profile index (KDPI) of donor and preservation method. Created with Microsoft Excel, Version 2308 Build 16.0.16731.20542.
Figure 2.
Delayed graft function (DGF) rate in donor after cardiac death (DCD) recipients based on the kidney donor profile index (KDPI) of donor and preservation method. Created with Microsoft Excel, Version 2308 Build 16.0.16731.20542.
Table 2 shows the multivariate logistic regression analysis, specifically looking at the effect of donor preservation on rates of DGF. For all subtypes of donors, including BDSCD, ECD, and DCD, donors showed a similar risk reduction in rates of DGF. Figure 3 shows the distribution of cold ischemia time. On average, the recipients of DCD and ECD MP donors had a similar but slightly longer cold ischemia time than did recipients of donors preserved on ice. A larger difference was seen in the BDSCD donors, where the median difference was 4 h.
Table 2.
Risk reduction of DGF with mechanical preservation.
| Donor type | Odds ratio of DGF | 95% Confidence Interval | Significance | |
|---|---|---|---|---|
| Lower | Upper | |||
| Brain dead SCD donors | 0.617 | 0.593 | 0.642 | <0.001 |
| DCD donors | 0.682 | 0.645 | 0.721 | <0.001 |
| ECD | 0.601 | 0.54 | 0.67 | <0.001 |
Model included donor age, SRTR race, ethnicity, gender, donor creatinine, hypertension, diabetes mellitus, death by CVA, smoking, iontrope use, cocaine use, EtOH use, cold ischemia time, and time on dialysis prior to transplant; ECD analysis also included DCD status.
Figure 3.
Distribution of cold ischemia time. (A) Brain-dead standard criteria donor (BDSCD) recipients. (B) Expanded criteria donors (ECD) recipients. (C) Donor after cardiac death (DCD) recipients. Created with Microsoft Excel, Version 2308 Build 16.0.16731.20542.
Figure 4 shows the unadjusted DGF rate for BDSCD kidneys by cold ischemia time. The analysis shows that 3 to 6 h of additional cold ischemia time largely negated the reduction in DGF rate seen with MP kidneys.
Figure 4.
Delayed graft function (DGF) rate in brain-dead standard criteria donor (BDSCD) recipients based on cold ischemia time and preservation method. Created with Microsoft Excel, Version 2308 Build 16.0.16731.20542.
Figure 5 shows the distribution of hospital length of stay for recipients based on the presence or absence of DGF and preservation method, respectively. DGF was associated with longer hospitalizations, while preservation type did not affect length of stay. Subset analysis showed that hospital stay was almost identical for ECD and DCD donors regardless of preservation method, even though DGF rates were reduced in these subgroups (Figure 6A, 6B).
Figure 5.
Distribution of hospital length of stay. (A) Delayed graft function (DGF) or no DGF. (B) Preservation method. Created with Microsoft Excel, Version 2308 Build 16.0.16731.20542.
Figure 6.
Distribution of hospital length of stay. (A) Expanded criteria donors (ECD) recipients and preservation method. (B) Donor after cardiac death (DCD) recipients and preservation method. Created with Microsoft Excel, Version 2308 Build 16.0.16731.20542.
Discussion
MP in a contemporary cohort of deceased donor kidney transplants as currently deployed did not have a large effect on DGF and had minimal effect on the length of hospitalization after transplant. Our analysis indicates that the primary reason for this was significantly longer cold ischemia times in MP donors, especially in BDSCD donors. Among ECD and DCD donors, the cold ischemia times were similar, indicating that these donors were more urgently implanted than the BDSCD donors. The effect of pumping improved DGF rates in those subgroups. The effect of pumping on DGF was negated in recipients of BDSCD donors by longer cold ischemia times. Transporting kidneys from one organ procurement organization to another typically increases cold ischemia time due to longer transport distances. The percentage of kidneys procured and transplanted in the same organ procurement organization was higher in the MP kidneys, suggesting that the additional cold ischemia time is not likely explained by increased travel times for the kidneys.
The strength of this study is the large number of patients in a contemporary era of transplantation. However, a number of limitations exist in the study. Although it appears that prolonged cold ischemia time of pumped BDSCD kidneys explained the reason that DGF rates were not better in the pumped cohort, placing kidneys received on ice by the center on the pump could also be a factor influencing this outcome. The dataset used did not have data on this particular scenario, and if a large number of donor kidneys received on ice were later placed on the pump, this could have influenced the results. Second, DGF rates are influenced by center and physician practices regarding the aggressiveness of dialysis use after transplant [17]. Some centers and physicians may have a lower threshold for implementing dialysis in the post-transplant setting, and DGF rates are based on dialysis use. These data have no means of adjusting for practice differences. Finally, DGF in this study was defined by an operational definition of dialysis in the first week after transplant. Although ischemia-reperfusion injury of the graft explains most of DGF based on this definition, other causes of renal dysfunction, such as hyperacute rejection, graft thrombosis, or primary non-function, could bias the results if not equally distributed among pumped and cold-stored organs.
The major benefit of MP is that it reduces the incidence of DGF after implantation. DGF is associated with longer and costlier hospitalizations after transplant [14]. On the other hand, MP increases the cost of organ procurement substantially over that of cold storage. The economic argument for MP depends on the fact that overall rates of DGF would be reduced. As MP is currently implemented in the United States, our study shows that it did not improve DGF rates in most cases and did not shorten hospital stays. The effect of MP on DGF reduction is largely negated by as little as 4 to 6 h of additional cold ischemia time, and pumping was associated with longer cold ischemia times. Many logistical considerations, such as availability of operating room space, staff, desire for daytime surgery, or delays in crossmatching, affect cold ischemia time, but to realize the full benefits of pumping, cold ischemia time still needs to be minimized, and the urgency of implantation needs to be like that of cold-stored donor organs. Also, the benefits of MP regarding reduction in DGF may be overestimated by transplant teams, giving them a reduced sense of urgency for implantation. On the other hand, MP does allow for some delay in implantation without increasing DGF rates, to address logistical implantation delays and allow for more flexible scheduling of surgeries when staff is well rested.
In the subset analysis of length of hospital stay, a minimal difference in length of stay was seen in ECD and DCD recipients even though both these groups did have a significant difference in DGF rate between the 2 preservation methods. We can only speculate on the reason for the lack of effect. The operational definition of DGF as the need for dialysis in the first week after transplant does not capture the full spectrum of ischemia-reperfusion injury seen in kidney transplants. It is possible that MP only retards ischemia-reperfusion injury in donors with mild injuries and has less effect on those with more profound injuries. Receiving a dialysis treatment because the recipient is hyperkalemic but non-oliguric after transplant is much different than a recipient who remains oligoanuric for days after transplant. Longer duration of dialysis with DGF is associated with poorer graft outcomes [18]. MP appears to improve early perfusion of the graft, but this effect may have little impact on established acute tubular necrosis [13]. The recipients of kidneys with more profound injury would likely be the patients needing longer hospitalizations and more dialysis treatments after transplant, and MP may have little effect on these transplants. Also, since DGF is an operational definition of ischemia-reperfusion injury, it does capture other causes of early graft dysfunction, such as early rejections or surgical mishaps, that would likely require longer hospitalizations but would not be influenced by preservation potentially biasing the results.
As currently deployed in the United States, MP did reduce DGF rates in DCD and ECD recipients, mainly because cold ischemia times in these cases were similar to cold-stored kidneys. In BDSCD recipients, because of longer cold ischemia times in pumped kidneys, the effect on DGF was negated. To realize the full benefits of MP, kidneys need to be transplanted with the same urgency as cold-stored kidneys. The lower DGF rates in ECD and DCD recipients of pumped kidneys did not translate into a clinically significant shortening of the length of hospital stay, and therefore the cost of MP was not offset by cost reductions due to shorter hospital stays.
Conclusions
The study shows that if cold ischemia times of deceased donor kidneys are similar between MP and cold-stored organs, the actual rate of DGF was reduced, but the addition of as little as 6 to 12 h of additional cold ischemia time on mechanically perfused kidneys negated the beneficial effect. To appreciate the full benefit in terms of reduced rates of DGF, mechanically perfused kidneys should be transplanted with the same urgency as cold-stored kidneys. Although mechanically perfused ECD and DCD donors experienced lower DGF rates due to more urgent implantation, the length of hospitalization was not improved appreciably, and we hypothesize that the lack of effect is likely due to the lack of impact of MP on kidneys with more severe ischemia-reperfusion injury. Further studies with better data regarding DGF, including such variables as the length time on dialysis or number of dialysis treatments after transplant, urine output, and velocity of creatinine improvement, could provide a fuller picture of the benefits of MP.
Abbreviations
- BDSCD
brain-dead standard criteria donor
- DCD
donor after cardiac death
- DGF
delayed graft function
- ECD
expanded criteria donor
- KDPI
kidney donor profile index
- MP
mechanical preservation
Footnotes
Conflict of interest: None declared
Publisher’s note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher
Declaration of Figures’ Authenticity: All figures submitted have been created by the authors, who confirm that the images are original with no duplication and have not been previously published in whole or in part.
Financial support: The purchase of data from the Scientific Registry of Transplant Recipients was received from Sacred Heart Hospital in Pensacola, Florida
Statement: Funding for the purchase of data from the Scientific Registry of Transplant Recipients was received from Sacred Heart Hospital in Pensacola, FL, USA.
References
- 1.Moers C, Smits JM, Maathuis MH, et al. Machine perfusion or cold storage in deceased-donor kidney transplantation. N Engl J Med. 2009;360(1):7–19. doi: 10.1056/NEJMoa0802289. [DOI] [PubMed] [Google Scholar]
- 2.Sandal S, Luo X, Massie AB, et al. Machine perfusion and long-term kidney transplant recipient outcomes across allograft risk strata. Nephrol Dial Transplant. 2018;33(7):1251–59. doi: 10.1093/ndt/gfy010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bathini V, McGregor T, McAlister VC, et al. Renal perfusion pump vs cold storage for donation after cardiac death kidneys: A systematic review. J Urol. 2013;189(6):2214–20. doi: 10.1016/j.juro.2012.11.173. [DOI] [PubMed] [Google Scholar]
- 4.Cannon RM, Brock GN, Garrison RN, et al. To pump or not to pump: A comparison of machine perfusion vs cold storage for deceased donor kidney transplantation. J Am Coll Surg. 2013;216(4):625–33. doi: 10.1016/j.jamcollsurg.2012.12.025. discussion 633–34. [DOI] [PubMed] [Google Scholar]
- 5.Ciancio G, Gaynor JJ, Sageshima J, et al. Machine perfusion following static cold storage preservation in kidney transplantation: Donor-matched pair analysis of the prognostic impact of longer pump time. Transpl Int. 2012;25(1):34–40. doi: 10.1111/j.1432-2277.2011.01364.x. [DOI] [PubMed] [Google Scholar]
- 6.Gallinat A, Moers C, Treckmann J, et al. Machine perfusion versus cold storage for the preservation of kidneys from donors ≥65 years allocated in the Eurotransplant Senior Programme. Nephrol Dial Transplant. 2012;27(12):4458–63. doi: 10.1093/ndt/gfs321. [DOI] [PubMed] [Google Scholar]
- 7.Gill J, Dong J, Eng M, et al. Pulsatile perfusion reduces the risk of delayed graft function in deceased donor kidney transplants, irrespective of donor type and cold ischemic time. Transplantation. 2014;97(6):668–74. doi: 10.1097/01.TP.0000438637.29214.10. [DOI] [PubMed] [Google Scholar]
- 8.Jochmans I, Moers C, Smits JM, et al. Machine perfusion versus cold storage for the preservation of kidneys donated after cardiac death: A multicenter, randomized, controlled trial. Ann Surg. 2010;252(5):756–64. doi: 10.1097/SLA.0b013e3181ffc256. [DOI] [PubMed] [Google Scholar]
- 9.O’Callaghan JM, Morgan RD, Knight SR, Morris PJ. Systematic review and meta-analysis of hypothermic machine perfusion versus static cold storage of kidney allografts on transplant outcomes. Br J Surg. 2013;100(8):991–1001. doi: 10.1002/bjs.9169. [DOI] [PubMed] [Google Scholar]
- 10.Tedesco-Silva HJ, Mello Offerni JC, Ayres Carneiro V, et al. Randomized trial of machine perfusion versus cold storage in recipients of deceased donor kidney transplants with high incidence of delayed graft function. Transplant Direct. 2017;3(5):e155. doi: 10.1097/TXD.0000000000000672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tingle SJ, Figueiredo RS, Moir JA, et al. Machine perfusion preservation versus static cold storage for deceased donor kidney transplantation. Cochrane Database Syst Rev. 2019;3(3):CD011671. doi: 10.1002/14651858.CD011671.pub2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wang W, Xie D, Hu X, et al. Effect of hypothermic machine perfusion on the preservation of kidneys donated after cardiac death: A single-center, randomized, controlled trial. Artif Organs. 2017;41(8):753–58. doi: 10.1111/aor.12836. [DOI] [PubMed] [Google Scholar]
- 13.Zhong Z, Lan J, Ye S, et al. Outcome improvement for hypothermic machine perfusion versus cold storage for kidneys from cardiac death donors. Artif Organs. 2017;41(7):647–53. doi: 10.1111/aor.12828. [DOI] [PubMed] [Google Scholar]
- 14.Kim DW, Tsapepas D, King KL, et al. Financial impact of delayed graft function in kidney transplantation. Clin Transplant. 2020;34(10):e14022. doi: 10.1111/ctr.14022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Iqbal K, Hasanain M, Rathore SS, et al. Incidence, predictors, and outcomes of early hospital readmissions after kidney transplantation: Systemic review and meta-analysis. Front Med (Lausanne) 2022;9:1038315. doi: 10.3389/fmed.2022.1038315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shoskes DA, Cecka JM. Deleterious effects of delayed graft function in cadaveric renal transplant recipients independent of acute rejection. Transplantation. 1998;66(12):1697–701. doi: 10.1097/00007890-199812270-00022. [DOI] [PubMed] [Google Scholar]
- 17.Louvar DW, Li N, Snyder J, et al. “Nature versus nurture” study of deceased-donor pairs in kidney transplantation. J Am Soc Nephrol”. 2009;20(6):1351–58. doi: 10.1681/ASN.2008070715. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Budhiraja P, Reddy KS, Butterfield RJ, et al. Duration of delayed graft function and its impact on graft outcomes in deceased donor kidney transplantation. BMC Nephrol. 2022;23(1):154. doi: 10.1186/s12882-022-02777-9. [DOI] [PMC free article] [PubMed] [Google Scholar]