Is DCD for Liver Transplantation DNR?
* Corresponding author: John F. Renz, firstname.lastname@example.org
We need to improve our understanding of the pathophysiology of donation after cardiac death and develop new predictors of successful donation.
The U.S. Department of Health and Human Services' sponsored Organ Donation Breakthrough Collaborative (ODBC) has been a universally successful initiative to increase organ donation within the United States (1). To date, there has been a 23% increase in deceased donation (DD) since 2003 with over 8000 DD recorded in 2006 (2). The increases observed through the ODBC have principally resulted from the development and dissemination of ‘best practices’ with respect to donor identification, recruitment and management. An explicit goal of the ODBC is increased utilization of donors who do not meet neurologic brain death criteria—donation after brain death (DBD), but are recovered after meeting cardiopulmonary death criteria—donation after cardiac death (DCD) (3). The Institute of Medicine (4–6), United Network for Organ Sharing (UNOS) (2), and Joint Commission on Accreditation of Healthcare Organizations (7) has each endorsed DCD as an ethically proper technique for organ donation. Presently, DCD is the most rapidly increasing component of the DD pool: expanding from approximately 1% of DD in 1996 to greater than 8% of DD in 2006. This expansion has occurred concurrently with expansion of the DBD pool allaying early concerns that premature DCD would subvert DBD (2).
The true potential of DCD to supply future demand is unknown. Estimates have been proposed ranging from greater than 1000 DCD/year to a DCD supply that would exceed current demand; however, application of DCD has been erratic (1,2,5). While UNOS recommended all member transplant centers and organ procurement organizations (OPO) adopt a DCD protocol; DCD remains a practice that can account for as much as 20% of the donors in select OPOs with a significant minority (22%) of OPOs not performing a single procedure between June 2006 and May 2007 (2). The same practice occurs among transplant centers. In fact, preliminary region 9 as well as 2006 Scientific Registry of Transplant Recipient (SRTR) data suggest overall recovery and utilization of DCD for liver transplantation may be plateauing (2).
How can a technique with the potential to immediately impact the DD pool remain under-utilized? The short answer is DCD outcomes in liver transplantation are inferior. Unlike emerging data on kidney transplantation (8), multiple single-center series as well as SRTR data clearly demonstrate inferior allograft survival utilizing DCD (9–12). These data do not justify DCD utilization for many potential recipients that do not require immediate organ access. SRTR data examining transplanted DD liver allografts between 2000 and 2003 reveal an adjusted odds ratio of 1.85 for graft failure relative to DBD when utilizing a DCD allograft (p < 0.0001).1 Interestingly, graft failure was not associated with recipient disease severity at transplantation or center experience (13). A similar analysis by Abt et al. of SRTR data determined graft failure is typically not from an increased incidence of primary nonfunction, hepatic artery thrombosis or technical complications, but from the development of ischemic cholangiopathy (9). These data implicate inferior results are not a consequence of technical issues but an inherent limitation of the procedure.
Specific obstacles the OPO and transplant center must overcome to stimulate increased utilization of DCD allografts in liver transplantation include: recognition of the biologic differences between DBD and DCD, an inefficient allocation system, and unique DCD allograft failure patterns.
The principal impediment to expanded application of DCD is biology
Our lack of understanding of the physiologic events occurring during DCD and recognition that these processes are fundamentally different from the physiology occurring during DBD has resulted in inferior outcomes, confusing terminology, and variable methodologies associated with procurement. These limitations were recognized by a Consensus Conference on DCD convened in 2006 (14); however, inherent biologic limitations have precluded substantial progress in donor identification, methodology, and outcomes.
The physiologic changes occurring in DBD and DCD are fundamentally different and poorly understood. Progression to brain death in DBD results from cerebral insult and mass effect. The intracranial events may or may not include systemic ischemia. Thus, a DBD pattern is cerebral insult ± systemic ischemia followed by a period of stabilization prior to donation. The variable period of stabilization permits donor resuscitation and clinical assessment of DBD end organ function.
Progression to donation in DCD involves an initial clinical insult that can be much more variable. There are DCD donors whose brain has survived a stroke or trauma while others have brain injury entirely from anoxia. Analysis of 914 DCD donors reported to the SRTR between January 2002 and May 2007 indicate the cause of donor death as anoxia 29%, cerebrovascular accident 23%, trauma 42% and other 6% (2). It is the anoxia population who may experience a different mechanism of injury as cerebral ischemia is the consequence of systemic ischemia. In the DCD donor there is also a variable period of stabilization that permits assessment of end-organ function but the patient's neurologic status may limit resuscitation to optimize organ recovery. Furthermore, end-organ function will be compounded by a new episode of ischemia secondary to withdrawal of support. The resulting DCD clinical sequence is insult-limited stabilization-ischemic insult-donation versus the DBD clinical sequence of insult-stabilization-donation.
The scientific processes underlying DBD and DCD are just beginning to be appreciated. Thus, the remaining paragraph is largely speculative. DCD is associated with the liberation of inflammatory mediators, activation of the coagulation cascade, immune surveillance, apoptosis and necrosis (15). Unrecognized neural and/or humoral ‘effector molecules’ released by solid organs during periods of ischemia could impact both the sympathetic outflow from the viable remnant brain as well as local organ function in a paracrine effect. The liver during a DCD recovery is therefore exposed to solid organ ischemia/acidosis-stimulated neurologic response, portal vein-mediated local visceral ischemic response from the splanchnic circulation, as well as internal hepatic ischemia. The timing and process are fundamentally different than the cranial-dominated outflow of ‘effector molecules’ in DBD associated with the dying brain from insult/mass effect.
Limited understanding of the biology interrupts each phase of the organ placement process resulting in inefficient allocation
There are no accurate predictors of DCD progression within the time frame generally recognized for liver transplantation. The absence of ‘predictors’ of successful progression to liver donation from a DCD candidate is a substantial impediment for OPOs seeking to increase donation utilizing limited financial and manpower reserves. While a DCD evaluation tool has been proposed by University of Wisconsin (16), this instrument has not been reliably validated in other OPOs and the predictive time frame of one hour is too long. Thus, OPOs are forced early in the DCD donation process to resource allocate in a manner where exponential efforts are necessary to achieve linear gains.
Without a defined starting point, it is not surprising that language rapidly diverts. Time from extubation? Time from hemodynamic parameters? Time from oxygenation parameters? Unclear language yields unclear methodologies resulting in an allograft that is significantly less ‘portable’ than a DBD allograft and an allocation process that is operationally different. For example, from an OPO resource management perspective, a healthy 75-year-old male DBD candidate, blood type O, with normal renal function and hemodynamically stable has a reasonable chance of proceeding to successful liver donation. Elderly DBD donors typically have several centers interested and the allograft can still be successfully placed even if the recovering center declines for recipient issues.
Progression to donation and allograft placement are less predictable in a healthy 45-year-old male DCD candidate with a similar clinical scenario. In practice, the efficiency of placing a DCD liver allograft after the procurement surgeon has declined the allograft, for whatever reason, is less than DBD organs with significant additional risk factors. Although difficult to prove, SRTR data presented at the National Conference of Donation after Cardiac Death suggests this by evaluation of the adjusted odds of discard for liver allografts by donor type. The adjusted odds of discard for DCD allografts procured between 01 January 2000 and 31 December 2004 was greater than three-times that of DBD. Furthermore, these odds increased to eight times that of DBD if the DCD allograft included renal allograft extended-donor criteria (13). Continuing to apply the current allocation scheme to DCD requires exponential effort for linear gains which is inefficient and unsustainable.
DCD allograft failure patterns are fundamentally different from DBD
Perhaps the most significant disincentive on behalf of the transplant center for the utilization of DCD originate from recognition that DCD allograft failure patterns are fundamentally different from DBD. The vulnerability of biliary epithelia to ischemia-reperfusion injury results in an increased incidence of biliary cast syndrome and ischemic cholangioapthy among DCD allografts (17). These biliary complications have been reported in 15–37% of DCD recipients (11,18,19) and typically present within the first few months after transplantation. The spectre of ischemic cholangiopathy dramatically limits the applicability of these allografts with respect to recipient selection. Successful algorithms to resolve biliary cast syndrome/ischemic cholangiopathy through percutaneous biliary drainage procedures, repetitive endoscopic procedures or surgical clearence have not been developed with sparse existing literature on this topic. A large percentage of these recipients progress to re-transplantation or death from complications of repetitive cholangitis. The development of ischemic cholangiopathy requiring re-transplantation is a significant clinical dilemma because it generates a patient with urgent need for re-transplantation with MELD scores that are typically less than 20. These patients have been repetitively hospitalized for sepsis and are typically malnourished. Thus, they need immediate graft function from a premium allograft but are not in a position to compete within their region. UNOS has not addressed this issue as a function of policy leaving prioritization to the discretion of regional review boards which can be variable.
Increasing interest in DCD brings us full circle in the history of organ transplantation as all early donors were uniformly DCD prior to the recognition of brain death criteria (3). However, past performance does not guarantee future results. The current donor population is dramatically different than the early donor pool. Examination of age, cause of death, length of hospitalization and level of care clearly indicate a donor population that is continually increasing in complexity as a result of medical advances coupled with relaxation of traditional criteria to expand the donor pool (2). Unfortunately, this has occurred with few concomitant technologic advances with respect to clinical organ preservation. Funding basic science initiatives in ischemia-reperfusion injury and technological advances in machine perfusion must be a priority.
Fundamental change is required to re-ignite interest in utilization of DCD for liver transplantation. Until scientific or technologic advances are available, our current strategies, techniques and policies are insufficient to maximize the donor-pool potential. How can we maximize DCD utilization in this limited environment? I believe the problem cannot be solved at the level of OPO or transplant center. Rather, a prospective, donor-specific database initiative at the multi-regional OPO level involving organizations, such as the American Society of Transplant Surgeons, American Society of Transplantation, National Association of Transplant Coordinators or Association of Organ Procurement Organizations is necessary with the following mandate: develop an efficient predictive tool for donation and standardize data-driven methodology. An efficient tool to predict candidate progression to donation will enhance OPO resource utilization and relax the current need for exponential efforts to achieve linear gains. Clarifying methodology will crystallize the language of donation. This will streamline allocation, minimize cold ischemia, and form the foundation for studies designed to identify powerful predictors of posttranplant allograft function.
The transplant community must match this effort by optimizing organ utilization within the context of our present allocation scheme. Data-driven recovery techniques should be a priority. The development of regional DCD ‘recovery teams’ utilizing highly experienced procurement surgeons employing protocol-driven methodologies and collecting follow-up graft survival data would significantly advance this effort. Early data supports experienced surgical analysis and technique in the field increases outcomes and the confidence of the accepting transplant center thereby improving allocation efficiency and allograft ‘portability’ (20). Furthermore, we must design policy that recognizes DCD allograft failure patterns and provides an avenue for re-transplantation in the event of ischemic cholangiopathy.
Fundamental changes in DCD application will be dependent upon scientific/technologic advances. Until that is achieved, we must make an organized effort to improve our understanding of the physiology of DCD and develop novel predictors for successful donation. These efforts will clarify the terminology of DCD, standardize procurement, identify predictors of graft failure, improve outcomes and ultimately permit expansion of this very important donor population.
This manuscript is dedicated to the memories of Carla Ciner, Rafael Reinoso and Henry Holliday.
1Relative risk adjusted for donor age, gender, race, indication and recipient age, creatinine, body mass index, MELD, cold ischemia and warm ischemia.