Outcomes of simultaneous pancreas–kidney transplantation from brain-dead and controlled circulatory death donors


  • M. S. Qureshi,

    1. Cambridge Transplant Unit, Addenbrooke's Hospital, National Institute for Health Research and Cambridge Biomedical Research Centre, Cambridge, UK
    Search for more papers by this author
  • C. J. Callaghan,

    Corresponding author
    1. Cambridge Transplant Unit, Addenbrooke's Hospital, National Institute for Health Research and Cambridge Biomedical Research Centre, Cambridge, UK
    • University of Cambridge Department of Surgery, Box 202, Level E9, Addenbrooke's Hospital, Cambridge CB2 0QQ, UK
    Search for more papers by this author
  • J. A. Bradley,

    1. Cambridge Transplant Unit, Addenbrooke's Hospital, National Institute for Health Research and Cambridge Biomedical Research Centre, Cambridge, UK
    Search for more papers by this author
  • C. J. E. Watson,

    1. Cambridge Transplant Unit, Addenbrooke's Hospital, National Institute for Health Research and Cambridge Biomedical Research Centre, Cambridge, UK
    Search for more papers by this author
  • G. J. Pettigrew

    1. Cambridge Transplant Unit, Addenbrooke's Hospital, National Institute for Health Research and Cambridge Biomedical Research Centre, Cambridge, UK
    Search for more papers by this author



Organ scarcity has prompted increased use of organs from donation after circulatory death (DCD) donors. An early single-centre experience of simultaneous pancreas–kidney (SPK) transplantation from controlled DCD donors is described here.


Outcomes of SPK transplants from DCD and donation after brain death (DBD) donors between August 2008 and January 2011 were reviewed retrospectively.


SPK transplants from 20 DCD and 40 DBD donors were carried out. Donor and recipient characteristics were similar for both groups, although pancreas cold ischaemia times were shorter in DCD recipients: median (range) 8·2 (5·9–10·5) versus 9·5 (3·8–12·5) h respectively (P = 0·004). Median time from treatment withdrawal to cold perfusion was 24 (range 16–110) min for DCD donors. There were no episodes of delayed pancreatic graft function in either group; the graft thrombosis rates were both 5 per cent. Similarly, there were no differences in haemoglobin A1c level at 12 months: median (range) 5·4 (4·9–7·7) per cent in DCD group versus 5·4 (4·1–6·2) per cent in DBD group (P = 0·910). Pancreas graft survival rates were not significantly different, with Kaplan–Meier 1-year survival estimates of 84 and 95 per cent respectively (P = 0·181).


DCD SPK grafts had comparable short-term outcomes to DBD grafts, even when procured from selected donors with a prolonged agonal phase. Copyright © 2012 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.


Simultaneous pancreas–kidney (SPK) transplantation is the preferred treatment for selected patients with diabetes and end-stage renal failure1, 2. The ongoing imbalance between the supply and demand of suitable donor organs has led to lengthening transplant waiting lists3 and a search for new sources of grafts. These sources include living donors4, paediatric deceased donors5, and older or overweight donation after brain death (DBD) donors6, 7. Pancreas procured from donation after circulatory death (DCD) donors may offer a further alternative. DCD donors are classified as uncontrolled when cardiorespiratory arrest has occurred unexpectedly, and controlled when potential donors, for whom prolongation of treatment is considered futile, undergo planned withdrawal of life-sustaining treatment in a hospital environment, usually intensive care8.

Unlike DBD donor organs, those from controlled DCD donors are subjected to additional warm ischaemia, not only during the time from cardiorespiratory arrest to commencement of cold perfusion, but also from hypotension and/or hypoxia in the period from withdrawal of treatment to cardiorespiratory arrest (the agonal phase). Although more than two-thirds of potential donors die within 4 h (younger age, high oxygen needs and more intensive ventilation requirements are associated with a shorter time to death9), the duration of the agonal phase is unpredictable. Acceptable cardiorespiratory parameters for pancreas donation during this phase have not yet been defined, although the time from withdrawal of treatment to perfusion has generally been less than 30 min10. These uncertainties limit the use of DCD pancreas, with just 32 grafts implanted in the USA in 200811. In addition, potential DCD donors may place a considerable strain on procurement services owing to the difficulty in predicting donor death and the need for prolonged waiting periods to increase organ recovery rates9, 12.

There are relatively few reports on DCD pancreas transplantation13–17, but the results have been encouraging. The largest report is from an United Network of Organ Sharing registry analysis of 47 DCD SPK transplants that showed equivalent graft survival rates at 1, 3 and 5 years for organs from DCD and DBD donors14. Single-centre long-term follow-up has shown no difference in patient and graft survival at 8 years13. However, these findings may reflect more stringent DCD donor selection, as an adjusted survival analysis of the Scientific Registry of Transplant Recipients (SRTR) database has reported an increased, albeit not statistically significant, risk of failure associated with DCD organs10.

The extensive experience of the Cambridge Transplant Unit with kidney transplantation from controlled DCD donors12, 18 led to the use of pancreas from donors with agonal phases that were longer than those reported previously13. Here, the early results of SPK transplantation from DCD donors are reported, comparing these outcomes with those of transplants from DBD donors over the same interval.


All patients who received an SPK transplant between 1 August 2008 (implementation of the DCD SPK transplant programme) and 31 January 2011 were included in the study. Follow-up ceased on 30 April 2011. Only one pancreas after kidney transplant (DBD donor) and no pancreas transplants alone were performed during this interval. Donor and recipient information was obtained from a prospectively collected transplant database and retrospective case-note review. Information on organ offers was available from the National Health Service Blood and Transplant authority.

Donor selection

Absolute contraindications to pancreas donation were diabetes mellitus, acute necrotizing pancreatitis, chronic pancreatitis and pancreatic tumours. Although donor selection was on an individual basis, relative contraindications included pancreatic trauma, hyperamylasaemia, hyperglycaemia, high insulin requirement, body mass index (BMI) greater than 30 kg/m2 and age more than 55 years. Only controlled (Maastricht III) DCD donors were considered8.

Organ procurement

DBD pancreas recovery occurred after systemic heparinization (300 units/kg intravenously) followed by arterial perfusion via the distal abdominal aorta or right common iliac artery with 3–4 litres of University of Wisconsin (UW) solution (ViaSpan; Bristol-Myers Squibb Pharma, Garden City, New York, USA).

DCD donors had withdrawal of treatment in either the intensive care unit or the theatre anaesthetic room, in accordance with local hospital policy. Under UK law, pretreatment or predissection of potential DCD donors to aid organ recovery is prohibited and, specifically, administration of heparin is not permitted. Death was confirmed a minimum of 5 min after cardiorespiratory arrest, after which procurement could begin. DCD organ procurement employed a rapid recovery technique, with midline laparotomy followed by cannulation of the distal abdominal aorta or right common iliac artery, with outflow via the right atrium into the thorax. Perfusion fluids and volumes were the same as those for DBD donors. Heparinization was achieved by the addition of 25 000 units unfractionated heparin to each of the first 2 litres of UW perfusate. The liver and pancreas were preferentially removed en bloc to allow back-table identification of a potential accessory right hepatic artery. The duodenum was opened and flushed with UW solution before packing.

For both DCD and DBD donors, where the liver was also procured, dual perfusion was performed via the artery and portal vein, rather than the superior or inferior mesenteric veins, with the portal vein transected proximally to prevent pancreatic congestion. Machine perfusion was not used for either DCD or DBD kidneys.

There was no set time at which the procurement team ‘stood down’ if cardiorespiratory arrest did not occur after withdrawal of treatment, although in general this was a minimum of 4 h, as published previously for the Cambridge Transplant Unit DCD kidney programme12. There were no absolute agonal-phase cardiorespiratory parameters that were used to determine suitability for DCD pancreas donation. Instead, decisions were made on an individual basis, taking into account other factors such as the duration and haemodynamic instability of the agonal phase, warm ischaemia time, donor age and BMI.

For DCD donors, the warm ischaemia time was defined as the time from cardiorespiratory arrest to cold perfusion, and the cold ischaemia time as the interval from commencement of cold perfusion to reperfusion after implantation. Haemodynamic data from the agonal phase were collected; definitions of hypotension, oliguria and hypoxia were chosen as described previously12. Blood gas analysis and arterial lactate measurements were not performed on DCD donors, mostly because of lack of an arterial line or nearby blood gas analysers.

Recipient selection and implant techniques

Before December 2010, organ allocation in the UK was undertaken on a regional basis, and SPK recipients were selected by the on-call surgeon on the basis of blood group, HLA mismatch, presence of donor-specific HLA antibodies and donor age. In order to minimize the cold ischaemia time, recipients of potential DCD organs were admitted to the Cambridge Transplant Unit before withdrawal of treatment of the donor. In addition, the pretransplant cross-match was omitted in selected patients considered to be of low immunological risk19. When a preimplantation cross-match was deemed necessary, donor blood was obtained before withdrawal of treatment if possible and the cross-match performed using historical recipient serum.

From December 2010, DBD and DCD pancreas were allocated on a national basis according to a scoring system based on recipient sensitization, donor–recipient age difference, donor BMI, requirement for dialysis, waiting time, and distance from the donor hospital to the implanting centre. Recipient sensitization status was determined as HLA antibody reaction frequency (calculated reaction frequency), calculated as the percentage of a pool of 10 000 donors on the UK transplant database who were unacceptable owing to recipient antibody reactivity.

Donor pancreas was implanted intraperitoneally in the right iliac fossa, with systemic venous drainage to the inferior vena cava and enteric exocrine drainage via a Roux-en-SFY duodenojejunostomy. Extension venous grafts were avoided, and a feeding jejunostomy was placed routinely. The kidney was placed extraperitoneally in the left iliac fossa. Anastomosis time was defined as the time from organ out of ice to reperfusion.

Postoperative management

All recipients received thromboprophylaxis with 40 mg subcutaneous enoxaparin daily and were started on 75 mg aspirin at discharge, or if the platelet count was above 500 × 109/l. From 2009, patients also received an intravenous infusion of epoprostenol (prostacyclin) at 4 ng per kg per min at the time of pancreatic reperfusion, which was continued for 5 days after operation. Jejunostomy feeding began on the first day after surgery. Immunosuppression consisted of alemtuzumab induction therapy followed by tacrolimus and mycophenolate mofetil20.


Delayed graft function (DGF) of the pancreas was defined as the need for exogenous insulin to control hyperglycaemia within the first week after transplantation. Renal DGF was defined as the need for dialysis within the first week following transplantation, except when required for hyperkalaemia in the first 24 h after surgery. Primary non-function (PNF) of the kidney or pancreas was defined as failure of a graft to function ever, irrespective of cause. Kidney function was determined using the estimated glomerular filtration rate (GFR), which was calculated using the six-variable Modification of Diet in Renal Disease formula21.

Pancreas graft failure was taken as the time from transplantation to graft pancreatectomy, or return to insulin or oral hypoglycaemic medication. Kidney graft failure was defined as time to graft nephrectomy or return to renal replacement therapy. Cases were censored in the event of death with a functioning graft, and at the end of the follow-up interval.

Renal allograft rejection was generally biopsy-proven and defined by Banff criteria22, but when biopsy was contraindicated (hypertension or coagulopathy) antirejection treatment was commenced empirically. Pancreas rejection was usually made on clinical and biochemical grounds, with supportive evidence from a contemporaneous kidney biopsy. When pancreatic biopsy was performed, the diagnosis was based on the Banff criteria23.

Statistical analysis

Continuous data are presented as median (range) and were compared using the Mann–Whitney U test. Categorical variables were analysed by means of Fisher's exact test. Kaplan–Meier analysis was used for graft and patient survival, with comparison of groups by the log rank test. P < 0·050 was considered statistically significant. GraphPad Prism 5 was used for statistical analysis (GraphPad Software, La Jolla, California, USA).


Sixty SPK transplants were carried out between 1 August 2008 and 31 January 2011; 20 were from DCD donors and 40 from DBD donors. Median (range) follow-up was 14 (3–32) months. No patient was lost to follow-up.

Referrals for potential donation

Within the study interval, pancreas from 470 donors was offered to the Cambridge Transplant Unit, 193 from DCD and 277 from DBD donors. The proportion of offers accepted was almost identical: 57 (29·5 per cent) DCD versus 83 (30·0 per cent) DBD (P = 1·000). In the DCD group, six accepted potential donors (10·5 per cent) failed to progress to organ donation within the required interval after withdrawal of treatment. For a variety of reasons, 13 other potential DCD donors and 15 DBD donors either did not donate, had offers withdrawn, or the pancreas was not procured (for instance owing to damage or severe fatty infiltration). In total, 38 DCD and 68 DBD pancreas grafts were procured and assessed at the Cambridge Transplant Unit for SPK implantation. After assessment, 18 DCD and 28 DBD organs were deemed unsuitable for implantation (P = 0·547); the most common reason for non-implantation was concern over fatty infiltration of the organ (8 DCD versus 11 DBD; P = 0·767).

Donor characteristics

Donor characteristics were similar in the two groups (Table1). DCD donors were younger, but this did not reach statistical significance, and there was a higher incidence of death from cerebral anoxia in the DCD group (P = 0·014). For DCD donors, the median duration of the agonal phase and warm ischaemia time were 12 (6–89) and 13 (7–15) min respectively. The time from withdrawal of treatment to cold perfusion was 24 (16–110) min, and that from skin incision to organ perfusion was 3 (1–8) min. For six DCD donors the interval between withdrawal of treatment and organ perfusion was more than 30 min; these donors were aged 28 (15–53) years, with a BMI of 22·7 (17·2–24·2) kg/m2.

Table 1. Donor characteristics
 DCD (n = 20)DBD (n = 40)P
  • *

    Values are median (range).

  • DCD, donation after circulatory death; DBD, donation after brain death.

  • Fisher's exact test, except

  • Mann–Whitney U test.

Sex ratio (M:F)12:821:190·784
White ethnicity20391·000
Age (years)*31·5 (14–59)45·5 (14–58)0·076
Body mass index (kg/m2)*23·1 (17·2–31·8)24·3 (17·5–30·3)0·138
Height (cm)*177 (161–191)175 (152–200)0·197
Weight (kg)*70 (52–102)70 (48–100)0·541
Condition leading to   
 Traumatic brain injury440·422
 Anoxic brain damage840·014
 Cerebral infarct020·548
 Brain tumour020·548
Inotrope use at referral12320·127
Insulin use at referral6220·100
Serum glucose (mmol/l)*7·7 (5·7–11·6)7·0 (4·6–20·6)0·488
Serum amylase (units/l)*39 (6–157)49 (4–277)0·382
Serum creatinine (µmol/l)*81 (45–152)77 (27–240)0·783

Cardiovascular parameters during the agonal phase were generally favourable: just two of the 20 DCD donors had prolonged hypotension (systolic blood pressure less than 85 mmHg for more than 30 min), one had a urine output of less than 30 ml/h, and two had prolonged severe hypoxia (arterial oxygen saturation less than 70 per cent for more than 30 min, measured by peripheral pulse oximetry)12. In the group of six donors with withdrawal to perfusion times of more than 30 min, one donor had hypotension, oliguria and hypoxia, whereas another had hypotension and hypoxia only.

Recipient characteristics

Recipients of DCD and DBD organs were comparable (Table2). One patient who received a DCD graft was transplanted for type II diabetes mellitus. No patient had received a previous pancreas or kidney transplant. A higher proportion of recipients of DCD organs received grafts that had five or six HLA mismatches (P = 0·027). Only ten patients had a calculated reaction frequency of more than 0 per cent (none in the DCD group); among these, the median calculated reaction frequency was 48 (2–91) per cent. Twenty-three patients (38 per cent) had a preimplantation cross-match performed; this was less common in recipients of DCD grafts (2 versus 21 DBD recipients; P = 0·002).

Table 2. Recipient characteristics
 DCD (n = 20)DBD (n = 40)P
  • *

    Values are median (range).

  • DCD, donation after cardiac death; DBD, donation after brain death; DM, diabetes mellitus.

  • Fisher's exact test, except

  • Mann–Whitney U test.

Sex ratio (M:F)14:628:121·000
White ethnicity2040 
Age (years)*47 (24–56)42 (29–58)0·084
Body mass index (kg/m2)*25·3 (19·0–29·9)24·6 (17·8–34·4)0·981
Height (cm)*167 (148–188)171 (152–193)0·283
Weight (kg)*70 (53–105)70 (48–102)0·575
Type I DM19400·333
Duration of DM (years)*29 (12–50)26 (12–44)0·126
Dialysis type   
 Peritoneal dialysis880·127
Duration of dialysis (months)*14·5 (0–68)7·5 (0–84)0·158
Total HLA-A, -B, -DR   

Organs from DCD donors had shorter cold ischaemia times than those from DBD donors: median 8·2 (5·9–10·5) versus 9·5 (3·8–12·5) h respectively for pancreas (P = 0·004), and 11·2 (9·3–14·1) versus 12·5 (7·1–15·8) h for kidney (P = 0·025). As expected, there were no differences in anastomosis times between DCD and DBD groups: 39 (24–84) versus 39 (21–54) min respectively for pancreas (P = 0·378), and 41·5 (27–66) versus 42 (29–74) min for kidney (P = 0·580). The two recipients of DCD grafts who required a prospective cross-match had pancreas cold ischaemia times of 6·7 and 8·2 h.


The operative morbidity of DCD and DBD SPK transplantations was investigated by comparing the rates of reoperation within 30 days of transplant. Eleven recipients of DBD grafts had 15 early reoperations for: intra-abdominal bleeding (7), graft pancreatectomy for venous thrombosis (2), graft nephrectomy for venous thrombosis (1), abdominal compartment syndrome (1), drainage of an intra-abdominal collection (1), exploration of arterial anastomosis (1), wound closure (1) and planned relook laparotomy (1). The early reoperation rate was lower in the DCD group; four patients had five reoperations for: intra-abdominal bleeding (3), graft pancreatectomy for venous thrombosis (1) and small bowel resection for volvulus around the feeding jejunostomy (1). The difference in reoperation rates between groups was not significant (P = 0·753).

Pancreas graft survival was not significantly different in the two groups, with Kaplan–Meier 1-year survival estimates of 84 per cent for DCD transplants and 95 per cent for DBD organs (P = 0·181) (Fig.1). Three grafts failed in the DCD group owing to venous graft thrombosis, acute antibody-mediated rejection and chronic rejection. Two pancreas transplants from DBD donors failed, both as a result of venous thrombosis. The graft thrombosis rate was 5 per cent in both groups (DCD, 1 of 20; DBD, 2 of 40). There was no pancreatic DGF or PNF in either group.

Figure 1.

Kaplan–Meier survival analysis showing pancreatic graft survival after simultaneous pancreas–kidney transplantation from donation after circulatory death (DCD) and donation after brain death (DBD) donors. P = 0·181 (log rank test)

With respect to pancreatic endocrine function, there were no differences in haemoglobin (Hb) A1c level between recipients of DCD or DBD grafts at 3 or 12 months after transplantation: median 5·2 (4·4–6·5) versus 5·2 (4·6–6·8) per cent respectively at 3 months (P = 0·876), and 5·4 (4·9–7·7) versus 5·4 (4·1–6·2) per cent at 12 months (P = 0·910). In addition, serum amylase concentrations were similar between the DCD and DBD groups: median 66·5 (7–187) versus 60 (3–291) units/l respectively on day 2 after transplantation (P = 0·509), and 70 (41–126) versus 70 (18–205) units/l on day 30 (P = 0·423). Serum lipase values were also comparable in the two groups (data not shown).

Kidney graft survival was also similar, with estimated 1-year survival rates of 89 and 98 per cent for the DCD and DBD groups respectively (P = 0·206). Two kidneys failed in the DCD group (acute antibody-mediated rejection and chronic rejection) and one in the DBD group (venous thrombosis). Rates of DGF were higher among kidneys from DCD donors (7 of 20, 35 per cent) than DBD donors (6 of 40, 15 per cent), but the difference was not statistically significant (P = 0·101). There were no episodes of renal PNF among recipients of organs from DCD donors, but one occurred in the DBD group owing to venous thrombosis (P = 1·000).

There were no statistically significant differences in estimated GFR between DCD and DBD transplants: median 55·7 (26·2–118·3) versus 50·4 (17·5–102·1) ml per min per 1·73 m2 respectively at 1 month (P = 0·396), 64·3 (33·9–113) versus 59·9 (32·4–111·9) ml per min per 1·73 m2 at 3 months (P = 0·441), and 48·5 (30·1–81·4) versus 53·5 (32·6–100·2) ml per min per 1·73 m2 at 12 months (P = 0·583) after transplantation. Serum creatinine levels were also equivalent in recipients of DCD and DBD grafts (data not shown).

In the DCD group, six recipients had seven episodes of acute rejection of either the pancreas or kidney. The seven episodes included six instances of biopsy-proven rejection of the kidney. Nine recipients of DBD grafts had ten episodes of acute rejection (P = 0·542); seven episodes were biopsy-proven and two patients had successful pancreatic biopsies showing acute rejection. Only one patient did not have a biopsy attempted.

The postoperative length of hospital stay was similar, with a median of 17 (10–54) days for recipients of DCD organs and 17 (9–36) days for recipients of DBD organs (P = 0·535). One recipient of a DCD graft died from oesophageal cancer 14 months after transplantation with a functioning pancreatic and renal allograft. There were no deaths in the DBD group; this difference was not significant (P = 0·140). Of the six recipients of organs from DCD donors with withdrawal of treatment to perfusion times of more than 30 min, five have functioning pancreas and renal grafts, whereas one died with a functioning graft, as described above. The median duration of follow-up for this group was 16 (7–34) months.


This single-centre study demonstrated comparable short-term outcomes for SPK transplantation from DCD and DBD donors, with no difference in pancreatic allograft survival, endocrine function or graft thrombosis rates. Renal allograft survival and function in the two groups were also similar. DGF rates in DCD kidneys were higher, but the difference was not statistically significant. In addition, there were no significant differences in postoperative length of hospital stay, need for early reoperation, or rates of acute rejection. Analysis of referrals for potential donation showed that the Cambridge Transplant Unit appeared to be similarly selective with both DCD and DBD donors, and just as likely to discard either graft once procured.

These findings mirror those of Fernandez and colleagues13 from the University of Wisconsin. The Wisconsin group reported 37 DCD SPK transplants over 11 years, with a 1-year pancreas survival rate of 83 per cent, compared with 84 per cent in the present series. Importantly, however, the Cambridge results were achieved despite accepting DCD donors with longer times from withdrawal of treatment to perfusion and without prior heparinization of the donor. Fernandez and co-workers13 defined this as the warm ischaemia time, with a mean duration of 17·5 (range 6–48) min; 20 of 37 donors had a warm ischaemia time of 15 min or less in their series. By contrast, in the Cambridge cohort, the time from withdrawal of treatment to cold perfusion ranged from 16 to 110 min. Nonetheless, DBD and DCD outcomes were relatively similar, perhaps emphasizing that brain death also has deleterious effects on the pancreas24–26.

After withdrawal of treatment, DCD donors may show prolonged cardiovascular and respiratory stability before dying relatively rapidly. Using the time from withdrawal to perfusion as a surrogate for warm ischaemic damage may therefore be inadvisable. Recent analyses of kidneys transplanted from DCD donors did not show a statistically significant association between agonal-phase characteristics and long-term graft function12, 27. In contrast, DCD livers appeared more sensitive to ischaemic injury, and a systolic blood pressure of less than 50 mmHg for more than 15 min was associated with poorer graft outcome27. The tolerance of the pancreas to warm ischaemia is likely to be between these extremes, as animal models have shown decreased islet viability after warm ischaemia of more than 30 min28–30. This finding was reinforced in a study of human islets from 12 DCD donors, in which all isolates were biologically functional when the duration of systolic blood pressure below 50 mmHg was less than 25 min26.

Patient numbers in the present study were too small to analyse outcomes on the basis of agonal-phase cardiorespiratory parameters, although more than a quarter of the DCD donors had withdrawal of treatment to cold perfusion times longer than 30 min, with good early graft outcomes. Postwithdrawal cardiorespiratory parameters are therefore likely to be of more use than the extubation to perfusion time13, although further analysis of how agonal-phase cardiorespiratory characteristics affect pancreatic transplant function is essential as experience accrues. Ultimately, decisions on the use of organs from DCD donors must be made individually, taking into account other donor and recipient risk factors10 and graft appearance.

Donor and recipient characteristics were generally similar for DCD and DBD grafts; donors were younger in the DCD group, although this failed to reach statistical significance. Importantly, every effort was made to minimize the cold ischaemia time when using DCD organs. This was achieved by selective omission of the preimplantation cross-match19, or by using peripheral blood to perform the cross-match before withdrawal of treatment in the donor. In addition, planned recipients of both DCD and DBD organs were admitted before the donor went to theatre (DBD) or had treatment withdrawn (DCD). In this study, cold ischaemia times in the DBD group were thus substantially lower than those reported elsewhere14, although this policy did, on several occasions, result in the recipient not receiving a transplant because the organs were unsuitable on inspection or, for DCD donors, asystole did not occur within 4 h after treatment withdrawal.

Although there were no statistically significant differences between DCD and DBD pancreas or kidney graft survival, the number of patients in each group was relatively small and the duration of follow-up short. The survival curves suggest slightly poorer outcomes with DCD donor organs, which may become significant with longer follow-up or larger numbers. Of note, two DCD pancreas grafts failed within the first year owing to rejection; this may have been triggered by inflammation related to the warm ischaemic insult from DCD organ recovery. This appears unlikely, however, because the early clinical course and all biochemical markers of pancreatic inflammation were indistinguishable in the DCD and DBD groups. Notably, HbA1c levels were similar in both groups at 1 year. However, a recent risk-adjusted SRTR database analysis of pancreas transplant outcomes demonstrated a marginally increased risk associated with use of DCD grafts (hazard ratio 1·39), although this was not statistically significant (P = 0·10)10. Even so, transplantation from this donor source is still likely to carry a survival advantage over remaining on the waiting list or receiving a deceased donor kidney alone1, 31.

This single-centre analysis has demonstrated that SPK transplants from controlled DCD donors have acceptable short-term outcomes, which are similar to those of organs from DBD donors. This was achieved despite using donors with a longer agonal phase than usually accepted. This most likely reflects the short cold ischaemia times that were achieved, but further analysis of registry data and collection of postwithdrawal variables are needed to define what constitutes acceptable and unacceptable agonal-phase characteristics for DCD pancreas transplantation. Although there has been a continued rise in DCD donation in the UK over the past 5 years, this has led predominantly to increases in DCD kidney, rather than pancreas, transplantation. The Cambridge Transplant Unit, however, now performs approximately equivalent numbers of DCD and DBD SPK transplants. DCD donors thus represent a significant source of additional pancreas for transplantation and, given the acceptable outcomes achieved, their prudent use is to be encouraged.


The authors acknowledge the assistance of Olivera Gjorgjimajkoska with interpretation of tissue typing data, and thank the donor coordinators who collected agonal phase measurements. Finally, the authors thank Julia Ertner and Stephanie Smith, pancreas transplant coordinators, who maintained a prospective database of SPK transplant outcomes. There were no external funding sources.

Disclosure: The authors declare no conflict of interest.