Surgical Morbidity of Simultaneous Kidney and Pancreas Transplantation : A Single-Centre Experience in the Tacrolimus Era

Introduction. Simultaneous pancreas and kidney (SPK) transplantation is performed to restore normoglycaemia and renal function in patients with Type I diabetes mellitus and end-stage renal failure. We aimed to evaluate the impact of major postoperative complications to patient and gra survival outcomes. Method. Using a prospectively collected database over a 10-year period, major postoperative complications requiring return to operating theatre as well as patient and gra survival outcomeswere analysed retrospectively.Results. �etween January 2001 andMay 2010, 165 patients underwent �rst-time SPK transplantation.Median age of recipients was 39.8 years (range, 16.9–53.2). Enteric drainage was used in 149 patients, and bladder drainage was used in 16.Median follow-up timewas 5.2 years (range 1.1–10.3). Fiy-six patients (34%) returned to operating theatre at least once. Pancreatic allogra loss secondary to vascular thrombosis occurred in 12 patients (7%), and 2 patients (1.2%) required transplant pancreatectomy due to debilitating pancreatic enzyme leaks. At 1 and 5 years, patient survival was 98% and 94%; pancreas gra survival, 86% and 77%; kidney gra survival 96% and 89%, respectively. Conclusion. SPK is a safe and effective treatment for Type I diabetes mellitus and end-stage renal failure although surgical reintervention is required in approximately one-third of patients. Preventing vascular thrombosis remains a major challenge.


Introduction
e �rst successful simultaneous pancreas-kidney (SPK) transplantation was performed at the University of Minnesota Hospital in 1966 [1].Since then, SPK has become an effective treatment modality for patients with Type 1 diabetes mellitus (T1DM) and end-stage renal failure [2].
Despite numerous advances in surgical techniques, organ preservation, and immunosuppression over the past 25 years, SPK is still associated with signi�cant morbidity and rate of transplant gra loss remains signi�cant [3].Surgical complications in particular gra thrombosis and haemorrhage can lead to reintervention, gra loss, and increased length of hospital stay [4].
Our aim was to review the experience of SPK transplantation at the National Pancreas Transplant Unit (NPTU) in Australia over the last 10 years, focusing on the incidence of major postoperative complications and the reasons for gra loss.We also aimed to identify risk factors leading to gra loss within our cohort.

Patients and Methods
A retrospective study was conducted on 165 �rst-time SPK transplants performed at the NPTU between January 2001 and May 2010.is period was chosen because tacrolimusbased immunosuppression and enteric drainage of pancreas transplant gra became standard therapy in our unit in January 2001.Eligibility criteria for SPK transplantation included the following� T1DM, signi�cant renal impairment, and age less than 50 years.Exclusion criteria included a body mass index (BMI) greater than 30, cigarette smoking, and signi�cant untreated coronary artery disease.
Prior to being accepted onto the waiting list for transplantation, all patients were reviewed by a transplant physician and surgeon.Preoperative assessment included clinical evaluation, comprehensive biochemical studies, and cardiac stress tests.

Organ Procurement, Preservation, and Transplantation.
With the exception of one donation aer cardiac death (DCD), organs were recovered from heart-beating donors with certi�ed brain death.e donor pancreas procurement procedure has been previously described [5].In brief, the donor pancreas, duodenum, and spleen were recovered enbloc, and the root of the small bowel mesentery was stapled.Following 25,000 units of intravenous heparin and aortic �ush with several litres of cold (4 ∘ C) University of Wisconsin solution (UW), the organs were removed and placed in ice saline slush.e spleen was separated from the pancreas during back-table preparation.e donor duodenum was stapled proximally and distally and oversewn with 3-0 mono�lament absorbable suture.e donor common iliac artery, including its bifurcation, was anastomosed as a Y-gra to the donor splenic artery and superior mesenteric artery.e donor common iliac vein was used as an extension to the donor portal vein.
In the recipient, iliac artery exposure was performed through bilateral Rutherford-Morrison incisions or a midline laparotomy.e kidney gra was anastomosed to the recipient's le iliac vessels.An external ureteroneocystostomy was performed over a 7-French ureteric stent.e donor pancreas was positioned in the right iliac fossa in an intraperitoneal position.e pancreatic head was placed inferiorly, and the donor vessels were anastomosed to the recipient's right iliac vessels.
Since 2001, enteric drainage has been the preferred route of exocrine drainage in our unit.To achieve this, the donor duodenum segment and recipient jejunum were anastomosed in a side-to-side con�guration.In most cases, this was fashioned in two layers using a 3-0 mono�lament absorbable suture although a small number were performed with linear cutting stapler [6].For bladder-drained pancreases, the donor duodenum segment was anastomosed to the dome of the urinary bladder using 3-0 mono�lament absorbable suture.

Immunosuppression and Medical erapy.
Maintenance immunosuppression typically consisted of a tacrolimus, mycophenolate mofetil and prednisolone.Induction therapy with monoclonal anti-CD25 antibody was used routinely aer 2006.
Recipients received broad-spectrum antibiotic prophylaxis for 3 days, antifungal prophylaxis for one week, and CMV prophylaxis for six months postoperatively.Subcutaneous heparin (5000 units three times daily) was administered during convalescence in hospital, and acetylsalicylic acid (100 mg daily) was commenced upon resumption of diet.

Statistical Analysis.
Pancreas gra failure was de�ned as the day upon which permanent insulin administration was commenced or patient death.Kidney gra failure was de�ned as the day upon which permanent dialysis was commenced or death.
Measured values are reported as percentages for categorical data, and medians +/− interquartile range (IQR; 25th to 75th quartile) for continuous data.Differences between groups were compared by the Chi-squared or Fisher's exact tests for categoric variables, Student's  test for continuous parametric variables, and the Wilcoxon test for nonparametric continuous variables.Patient and gra survival was analysed by the Kaplan-Meier method.Logistic regression and Cox regression were used to assess risk factors for organ loss.Statistical analysis was performed using NCSS (Kaysville, Utah, USA) [7].A value of    was considered signi�cant.

Demographics. Between January 2001 and May 2010, 165
patients underwent SPK transplantation.Patient demographics are presented in Table 1.Despite entry criteria onto the transplant program, 11 patients had a BMI greater than 30 and six patients were aged 50 years or older at the time of transplantation.Most patients had been dialysis-dependent for at least one year, while 7% of patients had not progressed to dialysis at the time of transplantation.Sixteen patients (10%) had peripheral vascular disease requiring amputations, 13 had minor (toe) amputations, and 3 had below-knee amputations.Nine of 18 patients with a history of coronary artery disease underwent revascularisation procedures prior to transplantation.
Organs were retrieved from around Australia, including 83 from New South Wales and the ACT, 23 from Queensland, 18 from Victoria, and 41 from Western Australia and South Australia combined.e longest pancreas cold ischaemia time (CIT) was 19.3 hours.

Surgical Complications and
Return to eatre.Eighty-�ve secondary operations were performed on 56 out of 165 recipients (34%; Table 2).Seven patients had two or more distinct surgical complications.Twenty patients underwent more than one reoperation.Ten of the 20 patients (50%) underwent drainage of intraabdominal collections (including those secondary to a pancreatic leak) on at least one occasion.Seven of the 20 (35%) resumed insulin administration within 6 months of transplantation.
irty-six patients (21.8%) returned to theatre within the �rst 30 days.Within this group, thrombosis was the most common reason for return to theatre (  ; 6.7%), followed by pancreatic leak (  , 4.2%; Table 2).In only one case of thrombosis was the pancreas gra able to be salvaged.One patient returned to theatre twice in this period-initially for postoperative haemorrhage and then for gra thrombosis.Aer the �rst 30 days, pancreatic leak was the commonest cause of return to theatre (  ; 4.2%), followed by small bowel obstruction (  ; 3.6%).
In total, 15/165 patients (9.1%) had con�rmed pancreatic leaks.One, which �stulated through the wound, was managed nonoperatively.Fourteen of the 15 patients with pancreatic leaks returned to theatre for a total of 26 reoperations (range 1-6); in eleven the exocrine drainage was revised, and in two (1.2%) the allogra was removed due to ongoing sepsis.
While 11/15 pancreatic leaks occurred within the �rst 30 days, in three cases the �rst manifestation of leak occurred more than six months postoperatively (at 7 months, 13 months, and 18 months).Despite extensive investigation at the time, no cause for the late leaks was established.
Leak occurred in 3/16 bladder-drained and 12/149 enterically drained pancreases (  ).Donor age (  ), degree of HLA-mismatch (  ), donor CMV status (  ), the degree of pancreatic enzyme rise in the immediate post-operative period (  ), and the requirement for  parenteral immunosuppressive therapy (beyond the induction period) within the �rst year of transplantation (  ) were not associated with pancreatic leak.

Pancreas and Patient
Survival.ere were �een deaths from 165 recipients during a median followup of 5.2 years.e most common con�rmed cause of death was opportunistic infection (  ), which included cryptococcal meningitis and fungal pneumonia.is was followed by malignancy (  ).At 1 and 5 years, patient survival was 98% and 94%, respectively.Kidney allogra survival at one and �ve years was 96% and 89%, while pancreas allogra survival was 86% and 77%, respectively (Figure 1).
Forty of 165 pancreas gras (24%) were lost over the 10year period.e most common cause of pancreas gra loss was organ failure, including failure due to rejection (  , 7.9%; Table 3).e most common cause of late gra loss was death with a functioning gra (  , 7.3%).In fact, recipient death accounted for 30% of all pancreas gras lost.Acute rejection coincided with gra thrombosis in three patients; two within the �rst 30 days, and one at 20 months.
Of the 40 pancreas gras lost, 23 (58%) were lost in the �rst year, including two due to recipient death (Table 4).Fourteen of the 40 (61%) were lost in the �rst month, the most common cause of which was thrombosis without evidence of rejection (  ).
Duration of dialysis greater than one year before SPK transplantation was associated with pancreas gra loss at one year (  ) and poorer long-term pancreas (RR = 2.3, 95% CI 1.1-4.8),but not kidney (RR = 2.5, 95% CI 0.9-6.7),gra survival.ere was no signi�cant difference in pancreas CIT between gras that failed and gras that were still functioning at one year (  ).e degree of serum lipase elevation within the �rst 24 hours of transplantation was also not associated with early gra loss (  ).

Discussion
Whole-organ pancreas transplantation is an effective treatment for patients with Type I diabetes mellitus.It leads to improved quality of life, independence from exogenous insulin administration, normalisation of glucose homeostasis, and can ameliorate diabetes-mediated organ dysfunction [8][9][10][11][12].Despite continuing improvements in patient and allogra survival aer SPK transplantation, surgical complications leading to organ loss or reoperation are common [13][14][15].is series examines the outcomes of Australia's largest whole-organ pancreas transplant service over the last decade.
A variety of techniques to manage pancreatic exocrine secretions have been described.ese include enteric drainage, bladder drainage, duct occlusion, and open drainage into the peritoneal cavity.e practice in our unit is to use enteric drainage (ED) in preference to bladder drainage (BD).While BD avoids exposing the operative �eld to enteric �ora and permits monitoring of amylase secretion, ED has been shown to be at least equivalent to bladder drainage with regard to patient and gra survival [16,17].Moreover, ED is associated with a lower incidence of urinary tract infections and urological complications, avoids the metabolic derangement associated with bicarbonate loss, and has a lower incidence of anastomotic revision [18][19][20][21].
We also advocate the orientation of the head of the pancreas in a caudal direction towards the pelvis.is con-�guration facilitates ease of conversion to bladder drainage in the event of a pancreatic leak, rather than resorting to gra pancreatectomy.
Pancreatic leak, while infrequent, was the most common cause of return to theatre.Salvage operations were effective at preserving organ function� in only two of �een patients was the pancreas lost due to the consequences of a leak.Our rate of allogra loss due to pancreatic leak (1.2%) is comparable to those of large international centres.A 1.3% rate of pancreas loss secondary to leak in ED gras was reported by UNOS [16].Meanwhile, in a large single-centre experience, Sollinger et al. documented a leak rate of 5.7% in 610 enteric-drained pancreases, with the pancreas being lost in almost half of these [18].While pancreatic leak is oen amenable to re-intervention, pancreatic thrombosis remains the major challenge faced by transplant surgeons [4,22].In this series, primary thrombosis accounted for 12 of 28 (50%) death-censored gra failures and was responsible for 11/14 pancreatic gras (78.6%) lost within a month of transplantation.Large series have documented thrombosis to be responsible for the majority of technical failures and early gra loss [4,23].Factors that have been shown to predispose to gra thrombosis include recipient obesity, preservation time, and nontraumatic causes of donor death [4,24].
Both medical and surgical strategies are used to minimise the risk of gra thrombosis.While we routinely use a short venous extension to the donor portal vein, it has been postulated that this may increase the risk of thrombosis [18].Medical adjuncts to pancreas transplantation include the use of heparin or low-molecular-weight heparin [25], and commencing aspirin when oral intake begins.ere is, however, little available evidence to support this in reducing gra thrombosis.
Our relaparotomy rates of 21.8% at one month and 34% overall are similar to those reported elsewhere [14,15,18].Congruent with the recent �ndings of Manrique et al. [14], reoperation was associated with poorer gra, but not patient, survival.e �ve-year patient, pancreas, and kidney survival aer SPK transplantation in this series are 94%, 77%, and 89%, respectively.is compares favourably with United States data in which �ve-year patient and pancreas-gra survival rates were 87% and 71-73%, respectively, for patients who received a SPK transplant [26,27].
ere is evidence that SPK transplantation before the commencement of dialysis results in improved renal allo-gra outcomes [28].Although we observed a trend towards improved kidney gra survival with preemptive transplantation, this did not reach statistical signi�cance (log rank   ).Prior dialysis for more than one year, however, was associated with poorer one-year (  ) and longterm pancreas-gra survival (log-rank   ).

Conclusion
SPK transplantation has established bene�ts in improving quality of life and survival in Type 1 diabetic patients.Nevertheless it is an operation with well-de�ned morbidity, including that of pancreatic leak and thrombosis.In a country as large as Australia, our service faces the added logistical issues posed by recipients and donors being in distant locations.Despite this, we have demonstrated that SPK transplantation can be performed safely and successfully with functional and survival outcomes comparable to those achieved by large international centres.

Con�ic� of �n�eres�s
e authors declared that there is no Con�ict of interests.

F 1 :
Patient and gra survival.
T 1: Recipient and donor demographics.
a In the setting of CMV pancreatitis.b Following transplant renal biopsy.
T 3: Causes of pancreas loss.
T 4: Risk factors for pancreas gra loss at one year.