American Journal of Transplantation 2011; 11: 1719–1726 Wiley Periodicals Inc.
C 2011 The Authors C 2011 The American Society of Journal compilation Transplantation and the American Society of Transplant Surgeons
doi: 10.1111/j.1600-6143.2011.03634.x
The Impact of Variation in Donation After Cardiac Death Policies Among Donor Hospitals: A Regional Analysis J. Y. Rheea , R. Ruthazera , K. O’Connorb , F. L. Delmonicoc , R. S. Luskinc and R. B. Freemand,∗ a
Transplantation, Tufts Medical Center, Boston, MA LifeCenter Northwest, Seattle, WA c New England Organ Bank, Waltham, MA d Dartmouth Hitchcock Medical Center, Lebanon, NH *Corresponding author: Richard B. Freeman,
[email protected] b
The Joint Commission requires all hospitals have a policy regarding donation after cardiac death. To this date however, a quantitative analysis of adult hospital donation after cardiac death (DCD) policies and its impact on transplantation outcomes has not been reported. Specific characteristics for DCD polices were identified from 90 of the 164 (54.9%) hospitals within the New England Organ Bank’s donor service area. Forty-five policies (50.0%) allow family members to be present during withdrawal of life-sustaining therapy (WLST) whereas eight (8.9%) prohibit this. Seventeen policies (18.9%) require WLST to occur in the operating room (OR); 20 (22.2%) specify a location outside of the OR. Fifty-six (62.2%) policies fail to state the method of determining death; however, some require arterial line (15 policies, 16.6%) and/or EKG (10 policies, 11.1%). These variables were not associated with organ recovery, utilization or donor ischemia time. Our regional analysis highlights the high degree of variability of hospital DCD policies, which may contribute to misunderstanding and confusion among providers and patients that may influence acceptance of this mode of donation. Key words: Asystole, DCD, ischemia, outcome, policy Abbreviations: CIT, cold ischemia time; ICU, intensive care unit; DBD, donation after brain death; ID, identification number; DCD, donation after cardiac death; IQR, interquartile range; DDT, donor death time; NEOB, New England Organ Bank; DKRI, donor kidney risk index; SBP, systolic blood pressure; DRI, donor risk index; UNOS, United Network for Organ Sharing; DSA, donor service area; WLST, withdrawal of life-sustaining therapy. Received 07 October 2010, revised 22 March 2011 and accepted for publication 11 April 2011
Introduction The limited number of organ donors, whether living or deceased, continues to be the dominant predicament faced by all areas of the world where solid organ transplantation is performed. This quandary has driven transplant centers to consider the use of donor organs with less than ideal characteristics for transplantation. Donation after cardiac death (DCD), one category of these so called expanded criteria donors, has experienced the greatest rate of growth during the past several years worldwide. From 2000 to 2008 (1), the United States experienced an almost sevenfold increase in the number of DCD donors donating at least one organ, comprising 9.1% of all deceased donors in 2008. In a recent 5-year period, the Netherlands experienced a 21% decrease in the number of brain dead donors (donation after brain death or DBD) donors and a 129% increase in the number of DCD donors (2). Researchers from the United Kingdom have also reported that in 2007, 23% of all deceased organ donors were DCD donors (3). This rapid increase in DCD has spurred the transplant community and other medical organizations (4–6), such as the Institute of Medicine and the Society of Critical Care Medicine, to develop recommendations and guidelines regarding DCD. However, much controversy and debate continue to resonate regarding the ethical and procedural aspects of DCD. Furthermore, the importance of DCD has compelled the Joint Commission to require all hospitals, to have a DCD policy (7) irrespective of whether they choose to perform DCD or not. Ethical concerns regarding declaration of death, premortem interventions and compromising palliative care for the purposes of organ donation continue to temper the pace of development of consensus regarding policy and practice (8,9). In a qualitative analysis of health care providers, all groups recognized the lack of standardization of DCD policies and called for ‘uniformity in the criteria used for the declaration of cardiopulmonary death’ (10). The lack of standardization in DCD practice was further highlighted in a recent survey of children’s hospitals’ DCD polices, where considerable variation existed in the type of procedures specified for the determination of death, premortem management and location of WLST (11). These results illustrate that with various recommendations from the Institute of Medicine (4) and other interested 1719
Rhee et al.
authorities, development of an individual hospital DCD policy can be extremely complex, balancing these recommendations with input from individual personnel, hospital philosophies and codes of ethics. To aid in this process, many hospitals often consult organ procurement organizations (OPO) during the development of their DCD policies. Whether the degree of variability in DCD practice observed within children’s hospitals also exists in adult hospitals, where DCD is practiced much more frequently, has not been studied. Because DCD in children’s hospitals represents less than 1% of all donors (1) and thus many more DCD donors come from adult hospitals, significant differences in policies and procedures among these institutions may result in widely disparate organ recovery results that may impact organ function. Moreover lack of consistency in these practices may continue to contribute to public confusion and hesitation in acceptance of DCD, as stated by previously published studies (8–10).
Organ utilization and outcome data were provided by United Network for Organ Sharing (UNOS) data request dated August 2009. Donor and recipient data were linked by the UNOS ID. No recipient identifying characteristics were obtained. Organ donation consent rates for all potential DCD donors were not available. Transplanted recipients are listed in the UNOS registry as either being alive, dead or lost to follow-up or having been retransplanted. Liver graft failure was defined as death or retransplantation; kidney graft failure was defined as death, retransplantation or return to dialysis. Heart, lung and pancreas DCD organs outcomes were not analyzed due to extremely small numbers.
In an effort to ascertain how variable DCD practices are across a large donor service area, we characterize DCD policies for all hospitals served by the New England Organ Bank (NEOB) and examine whether differences in these policies are associated with organ recovery and utilization rates, graft outcome or patient survival.
The distribution of donors and organs are stratified by the DCD policy variables defined earlier. DDT, as well as the time from 50% drop from first systolic blood pressure (SBP) to asystole, time form SBP < 50 mmHg to asystole and time from last blood pressure to cessation of organized cardiac electrical activity, were all summarized for donors from whom a liver or kidney was recovered as medians and interquartile ranges (IQR) for each variable. Distributions of these times were compared between variables using the nonparametric Kruskal–Wallis test because of the skewed shape of these distributions. Power calculations were based upon the two conditions of a true hazard ratio of 1.1 or 1.5 and assumptions of overall 3-year event rate being similar to the sample and the two comparative groups having the same sample size.
Methods All hospitals served by the NEOB were included. Among these, only those that had a DCD policy on file and/or referred a potential DCD donor were selected for further analysis. Review of the hospital polices was conducted November 2009. Organ utilization rates were calculated based upon all DCD referrals from January 2000 to July 2009. Each hospital policy was reviewed by a single investigator for several qualitative and quantitative measures. Categories included indications/restrictions for application of a DCD protocol, tests required to determine death (e.g. mechanical asystole confirmed by arterial line or electrical asystole confirmed by EKG), whether the family is permitted to be present at withdrawal of life-sustaining therapy (WLST), location of WLST (e.g. ICU or OR), doctor responsible for declaring death, allowable premortem interventions and time from WLST to declaration of death after which organ could no longer be recovered. As per NEOB policy, all DCD donors had an observation period of 5 min in accordance with the Institute of Medicine recommendations (4). Categorization of these variables was based exclusively upon what was stated by each policy. Variables which were either not stated or left unclear, were categorized as such.
Organ utilization and outcome analysis All potential DCD donors referred to the NEOB from January 2000 to July 2009 were identified and retrospectively reviewed. All DCD donors were controlled, Maastricht category III (12) donors. We recorded demographic donor data, donor identification number (ID), cause of brain injury (COBI) and all variables corresponding to DCD polices outlined above. Donor death time (DDT) was calculated as minutes elapsed from WLST to declaration of death as defined by the individual donor hospital’s policy.
1720
The Donor Risk Index (DRI), as described by Feng et al. (13), was utilized in this study to adjust for donor characteristics when analyzing liver transplantation outcomes. Because the original DRI included DCD as an independent variable, we assigned a DRI-defined DCD coefficient, 0.411, to each donor’s DRI calculation. The DRI calculation was completed using the other six characteristics for each DCD donor from which a liver was transplanted. An equivalent adjustment using the Donor Kidney Risk Index (DKRI), as described by Rao et al. (14), was not possible as the DKRI includes recipient factors, such as number of B and DR mismatches that were not available at the time of this study.
For patient and liver graft survival analyses among the various policy alternatives, we compared Kaplan–Meier curves for unadjusted analyses and Cox proportional hazards models that included the DRI, as both a continuous and a binary variable (≥2 vs. 400 hospital beds 301–400 hospital beds 201–300 hospital beds 101–200 hospital beds 51–100 hospital beds ≤50 hospital beds
101 90 12 51 43 10 16 22 25 13 15
Table 2: Indications/restrictions for DCD Eligibility Mechanical Ventilation Ventilator dependent Not brain dead Severe Brain Injury GCS ≤5 Cardio/pulmonary disease (end-stage) Neuromuscular disease (end-stage) High cervical injury
22 33 66 50 15 12 7 6
Indications/restrictions The majority of hospitals (89/100, 89.0%) stated indications and restrictions for DCD donation in their policies (Table 2). Of these, 66 policies (66/89, 74.2%) specifically require failure to meet brain death criteria and 50 (50/89, 56.2%) articulate severe brain injury as an indication for DCD. Fifty-five (55/89, 61.8%) hospitals require that donors be mechanically ventilated, whereas 33 (33/55, 60%) of these further require that the potential donor be ventilator dependent. Twenty (20/89, 22.5%) hospitals stated age requirements, ranging from fetal demise at 20 weeks to 85 years old. Only one policy had a weight requirement, which was >350 g. Although brain injury was the predominant indication, 12 (12/89, 13.5%) policies expanded their indications to include end-stage cardiopulmonary disease whereas 7 (7.9%) policies included end-stage neuromuscular disease, such as amyotrophic lateral sclerosis (ALS). High cervical injury was noted in six policies (6.7%).
Premortem interventions
Number of hospitals
Duration of (procurement – deleted) recovery efforts ≤60 min 15 >60 min 33 NEOB discretion 24 Not stated 18 Determination of death Electrical asystole Mechanical asystole Both electrical and mechanical asystole Unclear Not stated
15 10 5 56 4
7 (7/11, 63.6%) of which required donor/family consent. Three (3/68, 4.4%) policies had a minimum mean arterial pressure requirement for the use of heparin. No policy completely prohibited the administration of heparin. Data on the use or prohibition of premortem vasodilators was not collected. Family in room Forty-five (45/90, 50%) hospitals offered family members the opportunity to be present at the time of WLST until death; however eight (8.9%) other hospitals specifically prohibited family members to be present (Table 3).
Premortem interventions Fifty-four (54/90, 60%) policies indicated that analgesics could be used to alleviate obvious discomfort (Table 3). Thirty-four (34/90, 37.3%) hospitals specifically approved use of intravascular monitoring catheters, whereas only one hospital had a stipulation that did not allow placement of premortem lines.
Location of WLST Polices specifying the location for where life sustaining therapy is withdrawn was specified as the operating room (OR) at 17 hospitals (18.9%), whereas 20 (22.2%) hospitals specified a location other than the OR, the intensive care unit (ICU) being the most common (Table 3). Twelve hospitals (13.3%) specified that WLST occur in either OR or ICU and 20 (22.2%) hospitals allowed family members to determine the location of WLST.
Of the 68 hospital policies that stated that heparin could be used after declaration of death, 42 (42/68, 61.8%) required consent by the family. Interestingly, only 11 (11/68, 16.2%) stated that pre-WLST heparin was permissible,
Allowable time prior to termination of organ recovery efforts Fifteen (16.7%) hospitals allow a maximum of 60 min from WLST for death to occur, after which all organ recovery
American Journal of Transplantation 2011; 11: 1719–1726
1721
Rhee et al. Table 4: Number of organs transplanted from DCD donors Organs transplanted donors with ≥1 transplanted organ Percent 1 kidney only Liver only 1 kidney + liver 2 kidneys Liver + 2 kidneys
33 5 9 176 66
11.4% 1.7% 3.1% 60.9% 22.8%
efforts are terminated (Table 3). Thirty-three (36.7%) hospitals specify a time greater than 60 min (90–180 min) and 24 (26.7%) hospitals left the decision of termination of organ recovery efforts at the discretion of the NEOB. Declaration of death All hospital policies did not allow any member of the transplant recovery team or any other physician involved in recovery efforts to determine when death has occurred (Table 3). Of the 90 (90/101, 89.1%) policies addressing declaration of death, 15 (15/90, 16.7%) required ‘absence of electrical activity on EKG’ and 10 (10/90, 11.1%) allowed absence of circulation with confirmation using arterial line; 5 (5/90, 5.6%) require both arterial line and EKG silence. However, the majority of policies (56/90, 62.2%) addressed declaration of death simply as ‘irreversible cessation of circulation and respiration,’ without elaboration on the method of confirmation. Four hospitals (4/90, 4.4%) did not directly address the declaration of death in their hospital policy at all. Organ utilization In total, 498 potential donors were identified (Tables 4 and 5). Three were excluded from further analysis due to missing data. Three hundred thirty-four (334/495, 67.5%) donors had at least one organ recovered and 289 (289/495, 60.3%) of them had at least one organ transplanted [649 kidneys (123 discarded, 526 transplanted) and 131 livers (51 discarded, 80 transplanted)]. Of those donors that had at least one organ transplanted, the majority (176/289, 61.6%) had only kidneys transplanted (Table 4). Of the 161 potential donors in which no organs were recovered, two potential cases were aborted due to gross contamination and the remainder did not reach asystole within the allotted timeframe. In addition to kidney and liver DCD organs analyzed here, 1 heart, 3 double lungs and 11 pancreas DCD organs have been transplanted from this experience. For the nine donors referred from hospitals that did not allow family to be present during the WLST process, six became organ donors (nine kidneys, zero livers). One hundred forty two of 206 (68.9%) patients in hospitals where family members are permitted to be present during WLST had at least one organ transplanted (221 kidneys, 35 livers). 1722
WLST was performed in the OR for 158 potential donors; 108 (108/158, 68.4%) had at least one organ transplanted (178 kidneys, 24 livers). When WLST was required outside of the OR, 16 of the 23 patients (69.6%) patients became organ donors (27 kidneys, 1 liver). For hospitals allowing family members to determine location of WLST, 77 of 114 patients (67.5%) had at least one organ transplanted (116 kidneys, 13 livers). Hospitals allowing a maximum of 60 min for progression to death had 218 potential donors, 143 (143/218, 65.6%) had organs transplanted (222 kidneys, 34 livers). Hospitals allowing >60 min experienced a rate of 45 of 66 (68.1%) patients having transplanted organs (71 kidneys, 9 livers). For those hospitals requiring EKG silence for death, 162 potential donors were identified and 111 (111/162, 68.6%) of these became organ donors (178 kidneys and 28 livers transplanted). Forty-seven of 67 (70.1%) potential donors from hospitals using arterial line for declaration of death became organ donors (80 kidneys, 10 livers). These data suggest that individual hospital DCD policy variables were not significantly associated with organ recovery or transplantation rates of kidneys or livers. Donor Death Time Median DDT for potential donors from hospitals declaring death using arterial line was 23 min (n = 42; IQR 18–29 min) while the median DDT for hospitals using EKG was 24 min (n = 94; IQR 17–29 min; Table 6). Not surprisingly, hospitals allowing ≤60 min for progression to death recorded a shorter DDT of 20 min (35; 14–29 min) compared to 22 min (119; 17–29 min) for hospitals allowing >60 min. Hospitals prohibiting the presence of family members during WLST had a DDT of 19 min (5; 16–27 min) whereas those that offered the choice to family members had a DDT of 21 min (114; 16–30 min). In hospitals where WLST was not permitted in the OR, DDT was 21 min (31; 16–28 min) compared with 24 min (89; 16–30 min) in hospitals requiring that WLST was performed in the OR. These data suggest that there is no association between hospital DCD policy variable and duration of DDT. Power calculations For an expectant hazard ratio of 1.1 or 1.5, a sample size of over 140 and 2000 subjects, respectively, per equally sized group would be required to demonstrate a 10% difference with 80% power. Liver outcome/patient survival Of the 80 livers that were transplanted, 29 (36.3%) recipients reached the composite outcome of either retransplantation or death (Table 7). Of these29 patients, 19 (23.8%) died, and 10 (12.5%) recipients underwent subsequent retransplantation. American Journal of Transplantation 2011; 11: 1719–1726
Hospital DCD Policy Variability Table 5: Distribution of donors and organs based on hospital DCD policy variables 494 total potential donors
Non-donors
Actual donors
Kidneys transplanted
Livers transplanted
Family presence Choice Prohibited Unclear/not stated Total
64 3 95 162
142 6 185 333
221 9 296 526
35 0 45 80
Location of WLST OR Not OR Family choice Unclear/not stated
50 7 37 68
108 16 77 132
178 27 116 205
24 1 13 42
162
333
526
80
75 21 5 61 162
143 45 25 120 333
222 71 32 201 526
34 9 4 33 80
51 20 91 162
111 47 175 333
178 80 268 526
28 10 42 80
Total Duration of recovery efforts ≤60 min >60 min NEOB discretion Unclear/not stated Total Determination of death Electrical asystole Mechanical asystole Unclear/not stated Total
Table 6: Duration of warm ischemia time stratified by DCD policy variable Median
Interquartile range
5 114 91
20 21 24
16–27 16–30 19–32
0.21
89 31 67
24 21 23
16–30 16–28 18–28
0.88
92
21
17–32
20 22 22
14–29 17–29 17–31
0.41
23
18–29
0.40
24
17–29
21
16–29
N Family presence No family Family choice Family unclear Location of WLST OR Not OR Location choice Location unclear/not stated
Duration of recovery efforts ≤60 min 35 >60 min 119 Other 16 Determination of death Mechanical 42 asystole Electrical 94 asystole Unclear/both 132
p-Value
patients were missing several other components of the DRI and thus were not included in further analysis. With the mean DRI being 1.90 (±0.36), a DRI of ≥2 versus