Pancreas transplantation for type II diabetes

Pancreas transplantation aims to provide patients with type I diabetes with an alternative source of insulin. In 1998, Sasaki et al. reported a small number of patients with insulin-requiring type II diabetes who had received pancreas transplants with short-term success.² This and a few other single-centre retrospective reports focused on small numbers of patients thought to have type I diabetes at the time of transplant, which were subsequently classified as type II. Whilst short-term outcomes after pancreas transplantation for type II diabetics appeared similar to those with type I diabetes, concern existed about the effect of older age of onset, metabolic syndrome, insulin resistance, higher prevalence of obesity and cardiovascular comorbidity in such patients. It also became clear that no clinical or biochemical test exists that can reliably distinguish type I from type II diabetes mellitus. Neither fasting nor stimulated serum C peptide levels predict insulin production, especially in the presence of renal failure and impaired excretion.³ Type II diabetes can present at a young age with signs of autoimmunity whereas some with type I diabetes can present later in life with no evidence of an autoimmune disorder. There are in fact reasonable grounds to question the validity of the view that type I and type II diabetes are two distinct disorders with different aetiology and pathogenesis. The ‘accelerator hypothesis’ considers both types of diabetes to be the same disease of insulin resistance where a specific genetic inheritance and some environmental and behavioural factors determine the age of onset.⁴

The International Pancreas Transplant Registry (IPTR) began recording data on type II diabetes in the mid-1990s. Until 5 years ago, there had been a slow increase in the proportion of simultaneous pancreas–kidney (SPK) transplant recipients reported as having type II diabetes, reaching 7%. This proportion has remained stable over the last 5 years.^5,6 Most registry analyses confirmed that type II diabetics do as well as those with type I following pancreas transplantation.

Pancreas transplantation from living donors

The pancreas transplant database of all US transplants performed between 1967 and 2011 records 72 living-donor transplants out of approximately 25 000 pancreas transplants (0.3%).⁶ Only three centres in the USA have performed this procedure, more than 80% of reported cases coming from the University of Minnesota. A recent review article⁷ by the authors of the largest series quotes the worldwide experience as approximately 160 cases, with only two centres outside the USA contributing several cases each. One-year pancreas graft survival after living-donor pancreas transplantation in the Minneapolis series was inferior to graft survival after deceased donor transplantation. Long-term outcome was comparable. The experience regarding perioperative morbidity and the long-term risks for the donor is inadequate to allow comment. It is of concern that only 67 of the first 115 live pancreas donors were traced to complete a follow-up survey. The survey revealed that three of these donors needed to start insulin some time after donation and 10 had elevated HbA1c levels.

The fact that it remains a small fringe activity confined to very few centres presumably indicates that living-donor pancreas transplantation is not endorsed by the transplantation community in general. This is certainly the view that prevails in the UK and the rest of Europe. The discussion in the remainder of this chapter is confined to allogeneic pancreas transplantation from deceased organ donors.

Simultaneous pancreas–kidney transplantation (SPK)

There is little debate that diabetic patients with renal failure should be offered kidney transplantation and there is good evidence that their prognosis is poor on dialysis.⁸ Such patients will already be obligated to immunosuppression on account of kidney transplantation. Combined kidney and pancreas transplantation offers these patients the opportunity to become insulin independent as well as dialysis independent. The risks and potential benefits of SPK transplantation compared with kidney transplantation alone for diabetic patients are summarised in Box 9.1. The evidence for the risks and benefits alluded to in Box 9.1 is reviewed in the final part of the chapter.

Experience has taught us the crucial importance of recipient selection for success in pancreas transplantation. Cardiovascular comorbidity is the most important factor leading to patient death in the early postoperative period after transplantation in diabetic patients. Patient selection needs to include a comprehensive medical evaluation, ideally performed by a multidisciplinary team within each transplant unit. Asymptomatic ischaemic heart disease is not uncommon in diabetics. The cardiac assessment should include as a minimum a 12-lead electrocardiogram (ECG), echocardiography, exercise tolerance test and a non-invasive test of myocardial perfusion (a radioisotope scan or dobutamine stress echocardiography). There is insufficient evidence to comment on the value of routine angiography.⁹ Any patient with an abnormality detected in non-invasive tests or those with symptoms of ischaemic heart disease should be investigated further, including angiography. Any correctable coronary artery disease should be dealt with prior to transplantation.¹⁰

Two per cent of SPK transplants between 2006 and 2010 were performed in recipients aged 60 or older. Strong evidence to differentiate between contraindications and relative contraindications does not exist (Box 9.2). Neither blindness nor previous amputation are regarded as contraindications. Hepatitis B, hepatitis C or human immunodeficiency virus (HIV) infection in the potential recipient are not contraindications to pancreas transplantation either. Box 9.2 is intended as a guide only. No attempt has been made to separate absolute and relative contraindications. The use of imprecise definitions such as ‘significant’, ‘severe’ or ‘recent’ is also intentional. Appropriate patient selection requires a balanced assessment by experienced clinicians of the risk factors versus potential benefits.

Pancreas after kidney transplantation (PAK)

Historically, the large majority of pancreas transplants performed have been SPKs. This was largely because of the relatively poor outcomes following solitary pancreas transplantation. Improved success of pancreas transplantation in the last 15 years^12,13 has encouraged many transplant units to offer pancreas transplantation for diabetic patients who have previously undergone successful kidney transplantation. Until the end of 2000, PAK transplants constituted 11% of all pancreas transplants performed in the USA and 5% of pancreas transplants performed outside the USA. Thereafter a rapid increase in PAK transplant rate was observed, reaching a peak of 400 (around a third of all pancreas transplants) in the USA in 2004.^6,11 There has since been a sharp decline in PAK numbers in the USA. More detailed analysis of activity and outcome in different forms of pancreas transplantation is given in the next section.

Contraindications to PAK transplantation are the same as those for SPK (Box 9.2). Within the confines of these contraindications, all diabetics who have previously undergone successful kidney transplantation are potentially suitable candidates for PAK. In practice the procedure is most useful for those diabetics who have a potential living donor for kidney transplantation. Since PAK outcomes are now approaching those of SPK, an elective and early living-donor kidney transplantation has obvious benefits to the potential recipient and has additional benefits to the overall pool of patients awaiting transplantation by releasing another deceased donor kidney. Clearly some time should elapse after kidney transplantation to allow for recovery from surgery and stabilisation of allograft function and immunosuppression. The optimal timing has not been determined. A study comparing early (within the first 4 months) with late (more than 4 months after kidney transplantation) PAK did not reveal any significant difference in the incidence of morbidity or outcome.¹⁴ More recent data suggest that longer intervals (greater than 3 years) between kidney transplantation and subsequent PAK transplantation can be associated with increased risk to the kidney graft.¹⁵

There are also unique immunological considerations for patients who are being considered for PAK. Previous kidney transplantation may have influenced sensitisation profiles for such patients. In the presence of good renal allograft function and no history of acute rejection following kidney transplantation, shared human leucocyte antigens (HLAs) between the pancreas allograft and the previous renal allograft may have a favourable impact on the outcome, an assertion that remains untested. Whether acute rejection following previous kidney transplantation predisposes PAK recipients to acute rejection of the pancreatic allograft is not known.

An important consideration, in particular for patients who are being assessed some considerable time after kidney transplantation, is the adequacy of kidney function. The level of kidney function at the time of PAK transplantation, independent of the time interval between the two transplants, is a factor associated with mortality risk.¹⁶ Criteria based on strong evidence do not exist but a creatinine clearance of greater than 40 mL/min is a useful guide as a minimum requirement.¹⁷ For recipients of kidney transplants who are not receiving calcineurin inhibitors, a higher creatinine clearance of not less than 55 mL/min is recommended. A renal allograft biopsy prior to PAK is useful for documentation of the baseline renal reserve and for the monitoring of the progression of allograft nephropathy (diabetic or otherwise).

The discussion above relating to SPK and PAK highlights the difficult choice facing patients. The experience of the local pancreas transplant centre, long waiting times for SPK transplantation and individual risk factors (age, obesity, race, etc.) that may adversely affect pancreas transplant outcome may favour early live-donor kidney transplantation followed by PAK. On the other hand, most fit patients in experienced centres without unduly long waiting times are likely to fare better with SPK.¹⁵

Pancreas transplantation alone (PTA)

Universally agreed criteria that constitute indications for PTA are more difficult to find compared with those for SPK or PAK. The risk of transplantation for non-uraemic diabetics needs to take account of not only perioperative morbidity but also long-term risks associated with immunosuppression. At present the majority of patients developing type I diabetes will be better served by insulin injections at the onset of disease. Clear indications for PTA are life-threatening complications of diabetes in patients managed with insulin, namely intractable hypoglycaemic unawareness and cardiac autonomic neuropathy. In these two subpopulations of diabetic patients pancreas transplantation is truly life saving. Other indications for PTA are more controversial. The presence of two or more diabetic complications has been advocated as an indication.¹⁸ In the absence of conclusive data on some of the diabetic complications, as reviewed in the final section of this chapter, it is difficult to defend the logic behind the assertion that two or more complications constitute an indication for PTA. Disabling and intractable symptoms of diabetic neuropathy or early nephropathy with preserved renal function may be considered as indications. In diabetics with overt nephropathy and impaired renal function, the nephrotoxicity of the immunosuppressive drugs will need to be considered. Available evidence does not permit precise guidelines but patients with creatinine clearance less than 40 mL/min may be better served by SPK transplantation.

Criteria for eligibility for pancreas donors

Contraindications to organ donation are clinical HIV infection, diagnosis of Creutzfeldt–Jakob disease (CJD), history of malignancy (except non-melanoma skin cancer and certain primary central nervous system tumours) and active systemic sepsis. Additional specific contraindications to pancreas donation are the presence of diabetes mellitus or gross pancreatic disease in the donor (including trauma, acute or chronic pancreatitis, excessive fatty infiltration). Neither hyperglycaemia nor hyperamylasaemia should preclude pancreas donation providing the pancreas appears normal on inspection. Increasing donor age, cerebrovascular/cardiovascular cause of death and donor obesity are associated with poorer pancreas transplant outcomes.22–24 A precise upper age limit for pancreas donors is difficult to define because of many confounding variables, the retrospective nature of the studies and small sample sizes of individual reports. IPTR analyses use 45 as the age cut-off distinguishing ‘young’ and ‘old’ donors. Historically no more than 3% of US pancreas donors were older than 49 and this has remained largely unchanged.^6,11 In practice an upper age limit of around 55 is used by most transplant units worldwide. The general consensus is that the ideal pancreas donor will be younger than 45.

It is similarly difficult to quote an acceptable weight limit for donors. Body mass index (BMI) > 30 should be regarded as a relative contraindication. Pancreas grafts from paediatric donors (age > 4 years) and selected non-heart-beating donors can be used with excellent results.^25,26 The influence of donor related factors on pancreas transplant outcome is discussed on p. 168.

Pancreas retrieval operation

The pancreas is a close neighbour of the liver and shares important vascular structures with it. During multiorgan retrieval procedures priority clearly needs to be given to the liver. The key to successful retrieval of both the liver and the pancreas is good cooperation between the two retrieval teams. The optimum scenario is for both organs to be retrieved and transplanted by the same team. Specific anomalies in the arterial blood supply to the liver that preclude successful liver and successful pancreas transplantation are very rare.

It is not intended to give a detailed description of the surgical procedure for pancreas retrieval in this section, but several pertinent points about pancreatic retrieval from multiorgan donors are highlighted below.

University of Wisconsin (UW) solution was first developed as a pancreatic preservation solution²⁷ and remains the benchmark for pancreas preservation. Other preservation solutions have been used but experience with these is more limited.²⁸ Recent experience with histidine–tryptophan–ketoglutarate (HTK) solution and Celsior solution suggests that both these solutions perform equally well as UW solution with short (< 12 h) ischaemia times. UW solution may perform better in longer cold ischaemia times exceeding 12 hours (for further details see Chapter 6).²⁹

The cold ischaemia tolerance of the pancreas is somewhere between that of the liver and the kidney. In pancreas allografts perfused with UW solution, 20 hours was thought to be the limit for successful preservation, beyond which a time-dependent deterioration in outcome occurred.³⁰ Earlier data failed to demonstrate a clear benefit from a preservation time cut-off below 20 hours. Most surgeons intuitively aim for shorter preservation times. The mean (± SE) preservation time for the 2000–2004 cohort of pancreas transplants in the USA was 13.3 (± 5.9) hours for SPK, 15.3 (± 5.5) hours for PAK and 16.2 (± 5.8) hours for the PTA group.³¹

More recent data suggest that ischaemia time is of greater importance in recipients of suboptimal grafts and in such cases ischaemia times over 12 hours are likely to be associated with poorer outcomes. Registry data reveal that pancreas transplant surgeons have taken this message on board, resulting in a trend towards shorter preservation times. The median cold ischaemia time for all US pancreas transplants has been under 12 hours since 2006.^32,33

The pressure gradient between mean arterial pressure and portal venous pressure that maintains blood flow through the pancreas can be significantly diminished during the perfusion of the abdominal organs in retrieval operations. Particular attention is required to maintain an adequate gradient if a cannula for perfusion is placed in the portal venous system as well as the aorta. Many transplant units perfuse abdominal organs with an aortic cannula only. Some evidence supports the view that additional portal perfusion is unnecessary.³⁴ For the interests of the pancreatic allograft, aortic perfusion alone is the most ‘physiological’ state that allows satisfactory perfusion and adequate drainage of the effluent.

It is common practice to flush the donor duodenum through a nasogastric tube with an antiseptic or antibiotic solution during the retrieval operation. No evidence exists to demonstrate the superiority of any solution used for duodenal decontamination.

Povidone–iodine during cold storage may be toxic to duodenal mucosa.³⁵ If povidone–iodine is used for flushing the allograft duodenal segment during organ retrieval, further flushing of the duodenal segment with preservation solution on the back table should be considered.

Donor duodenal contents should be submitted for bacterial and fungal culture. The results may be important in guiding the management of infection in pancreas transplant recipients.³⁶

General considerations

In SPK transplantation, pancreatic implantation is usually performed first because of the lower ischaemia tolerance of the pancreas. It is easier to implant the pancreatic graft on the right side. The renal allograft can also be placed intra-abdominally with anastomoses to the left iliac vessels. Alternatively, and perhaps more easily, an extraperitoneal renal transplant on the left side can be performed using the same incision or through a separate left iliac fossa incision. A further and equally satisfactory alternative is to implant the renal graft ipsilaterally, using more caudal segments of the recipient right iliac vessels.

In PAK transplantation, even in the presence of a right-sided renal allograft, the preferred site for the pancreas would be on the right-hand side cranial to the renal allograft. The arterial inflow to the pancreatic allograft usually comes from the right common iliac artery. Unless the lower aorta is clamped the renal graft does not suffer any ischaemia during pancreatic implantation.

Severely atherosclerotic and calcified vessels in some diabetic recipients can be a challenge during pancreatic implantation. Iliac Y grafts used for reconstruction offer greater flexibility in choosing a suitable arterial anastomotic site in the recipient vessels.

In the last 10 years the most common technique used for pancreas transplantation has been intra-abdominal implantation of the whole pancreas together with a donor duodenal segment. A list of previously employed techniques is given below (these are all associated with poorer outcomes and are rarely performed nowadays):

Management of exocrine secretions

Drainage of the exocrine secretions of pancreatic grafts into the recipient’s bladder was the most common technique, accounting for 90% of US pancreas transplants during the 1980s and early 1990s. The popularity of this technique was due to its perceived safety, primarily less serious consequences of anastomotic leaks (compared with enteric drainage) in the days of higher corticosteroids and unrefined immunosuppression. The ability to monitor amylase levels in the urine has been considered an additional advantage of bladder drainage. The latter point is contentious. The evidence with respect to the usefulness of urinary amylase monitoring is reviewed later in this chapter. The unphysiological diversion of pancreatic exocrine secretions into the urinary bladder causes frequent complications, often leading to chronic and disabling symptoms. As a consequence, conversion of the urinary diversion to enteric drainage is required in many patients. Complications of pancreatic transplantation, including specific problems associated with bladder drainage, are discussed in detail later in the chapter.

As discussed earlier, enteric drainage has replaced bladder drainage as the method of choice for the management of exocrine secretions in the USA. Data regarding the prevalence of enteric versus bladder drainage for pancreas transplantation in Europe are not easy to find. Historically, enteric drainage has been the preferred method in Europe and it is likely that the majority of European pancreas transplant units employ this technique.

Any part of the recipient’s small bowel can be used for anastomosis with the allograft duodenum. No data exist to demonstrate the superiority of one particular site over others. Roux-en-Y loops, which were commonly used, are becoming rare and a simple side-to-side entero-enterostomy is preferred.³¹

Delayed endocrine function from the transplanted graft is uncommon and insulin infusion should be discontinued at the time of reperfusion. Achieving insulin independence for the first time in many years in the recipient is a gratifying part of the operation for the surgeon. Patients can become hypoglycaemic at this stage. Blood sugar levels should be checked frequently and a low rate of dextrose infusion is often required.

Management of the venous drainage

Drainage of the venous outflow from pancreas grafts into the portal circulation was first described by Calne in 1984.³⁷ This complex surgical technique using a segmental graft and gastric exocrine diversion in a paratopic position has never gained popularity. Drainage of the venous outflow into the systemic circulation at the level of the lower inferior vena cava has been the norm in pancreatic transplantation. Since the mid-1990s there has been a resurgence of interest in portal venous (PV) drainage. The technique described by Shokouh-Amiri et al.³⁸ involves placement of the graft slightly more cranially in the abdominal cavity, with the utilisation of the superior mesenteric vein (SMV) for the venous drainage. Several studies, including prospective randomised comparisons, have shown that this offers at least equivalent outcome to that of systemic venous (SV) drainage, with no compromise in safety.^39–41 The impetus for PV drainage was to achieve a more physiological insulin delivery. A theoretical benefit was considered to be avoidance of hyperinsulinaemia, which has been linked with atherogenesis.⁴² Systemic drainage does not always cause hyperinsulinaemia. Nor may it be the only factor, since hyperinsulinaemia occurs in non-diabetic recipients of kidney transplants receiving steroids.^43,44 None of the studies of metabolic function after PV drainage have shown a clear benefit in terms of glucose metabolism, lipid profiles or atherogenesis. There is some evidence suggestive of an immunological advantage to PV drainage in the form of a reduction in the incidence of acute rejection. The mode of antigen delivery is known to modulate the immune response, which has been proposed as the mechanism responsible for the observed reduction in acute rejection rates in portal venous drained grafts.^39–41

Technically, PV drainage may be attractive for re-transplants or for patients who have had previous lower abdominal surgery. It requires the use of a long donor iliac Y graft to reach a suitable arterial anastomotic site on the recipient vessels. Suturing the graft portal vein to the recipient SMV or its main feeding tributary requires delicate handling of these fragile vessels. In obese patients and in those with thickened or foreshortened bowel mesentery, the SMV may not be easily accessible and PV drainage may be difficult to perform.