Chapter  98:  Islet Transplantation in Patients with Type 1 Diabetes Mellitus

Patients and Procedures

Pancreas transplant candidates include: 1) type 1 diabetic patients with renal failure who may receive a cadaveric simultaneous pancreas/kidney transplant (SPK); 2) type 1 diabetic patients who may receive a cadaveric pancreas transplant after kidney (PAK) transplantation from either a cadaveric or a living-related donor; and 3) nonuremic type 1 diabetic patients with severely disabling and potentially life-threatening acute diabetic complications who may be offered a pancreas transplant alone (PTA) (American Diabetes Association, 2003; Steinman, Becker, Frost, et al., 2001).

In all cases, patients are usually excluded for evidence of prohibitive cardiovascular risk, active infection, recent malignancy, or other contraindications to major surgery. Evidence also suggests that graft loss is lower when patients are transplanted prior to extensive dialysis (Papalois, Troppmann, Gruessner, et al., 1996). In successful transplants, blood glucose normalizes immediately; glycosylated hemoglobin concentration (i.e., hemoglobin A_1c, HbA_1c) normalizes and remains normal while the graft is functional (Larsen and Stratta, 1996; Robertson, Sutherland, Kendall, et al., 1996).

Immunosuppressive Therapy

Rejection is the most common cause of graft loss, and lifelong immunosuppressive therapy is required to prevent graft loss. Current strategies attempt to prevent rejection while minimizing injury to the allograft and overall risk to the patient from immunosuppressive agents. Tacrolimus, favored over cyclosporine A since about 1994, is administered with prednisone and mycophenolate mofetil (MMF) for long-term maintenance immunosuppression. With this regimen, 1-year graft survival rates for all types of pancreas transplants are 82 to 86 percent (Gruessner and Sutherland, 2002). Some centers have successfully tapered or discontinued post-transplantation glucocorticoids over time to avoid exacerbating peripheral vascular disease and other organ damage (Jordan, Chakrabarti, Luke, et al., 2000).

Tacrolimus inhibits insulin secretion and can cause post-transplant diabetes mellitus. However, this complication is reversed in more than 80 percent of cases by decreasing tacrolimus dose (Jordan, Chakrabarti, Luke, et al., 2000). In a multicenter trial of tacrolimus primarily for SPK transplantation, 3 percent of patients had their immunosuppression changed from tacrolimus to cyclosporine A due to post-transplant diabetes (Gruessner, 1997). Another center reports at least 2-year outcomes without evidence of tacrolimus toxicity (Jordan, Shapiro, Gritsch, et al., 1999). Tacrolimus also affects kidney function in a dose-dependent manner (Wagner, Herget, and Heemann, 1996; Goral and Helderman, 1997).

Replacing azathioprine with MMF, combined with either cyclosporine A or tacrolimus, significantly lowered risks of acute rejection and graft loss (International Pancreas Transplant Registry, 2001). MMF inhibits the cellular and humoral immune response via a different mechanism and is associated with neither nephrotoxicity nor diabetes (Goral and Helderman, 1997). However, approximately 25 percent of renal transplant patients have discontinued MMF due to gastrointestinal upset, leukopenia, and infections (Jindal, Sidner, and Milgrom, 1997).

Indefinite immunosuppression may also be necessary to prevent recurrent autoimmune organ damage. When an identical twin receives a syngeneic pancreatic segmental organ graft without immunosuppression, selective autoimmune destruction of the beta cells in the transplanted organ occurs rapidly (Sutherland, Goetz, and Sibley, 1989). At least one publication has documented selective loss of beta cells in allogeneic pancreas transplants that were ultimately rejected (Tyden, Reinholt, Sundkvist, et al., 1996). If anti-islet autoimmunity persists long after diabetes onset, it could contribute to pancreas transplant rejection.

Outcomes of Whole-Organ Pancreas Transplantation

This overview of the outcomes of whole-organ pancreas transplant procedures addresses patient survival, graft survival and diabetic complications.

Patient Survival. Several studies comparing long-term survival after SPK versus kidney-alone transplants (KTA) report that pancreas transplantation confers a survival advantage (Smets, Westendorp, van der Pijl, et al., 1999; Tyden, Bolinder, and Solders, 1999; Becker, Brazy, and Becker, 2000; La Rocca, Fiorina, Astorri, et al., 2000; Fiorina, Folli, Maffi, et al., 2003). However, short-term, mortality and morbidity are substantially higher with SPK.

Recently three multivariate analyses of longitudinal registry data have attempted to assess the short- and long-term trade-offs of SPK versus cadaveric KTA and to quantify, if possible, the projected survival advantage.¹ Overall, these analyses show that survival after SPK is better than KTA in the long term, but during the early post-transplant period, survival is worse with SPK. Ojo, Meier-Kriesche, Hanson, and colleagues (2001) analyzed 13,467 uremic adults with type 1 diabetes who were wait-listed for transplant between 1988 and 1997. Operative and early infectious deaths were approximately twice as high for SPK compared to KTA. Time to equal mortality as wait-listed patients was 95 and 170 days after cadaveric KTA and SPK transplantation, respectively. By 5 years, however, the mortality risk relative to wait-listed patients was 0.40 for 4,718 SPK patients (95 percent confidence interval [CI] = 0.33–0.49) and 0.75 for 4,127 KTA patients (95 percent CI = 0.63–0.89).

Bunnapradist, Cho, Cecka, and co-workers (2003) analyzed survival of 3,642 SPK and 2,374 KTA patients with type 1 diabetes reported to UNOS during 1994-1997 and followed through 2000. After controlling for favorable donor and recipient factors in the SPK group, risk of death in the KTA group relative to the SPK group was 1.06 (95 percent CI: 0.88–1.28), suggesting SPK had neither a favorable nor adverse effect on patient survival at 3–6 years. Reddy, Stablein, Taranto, and colleagues (2003) analyzed 18,549 kidney recipients with type 1 diabetes transplanted 1987 to 1996. At 8 years, unadjusted survival was 72 percent for SPK (n = 4,602) and 55 percent for cadaveric donor KTA (n = 9,956).

Survival after PTA has been reported to be comparable to that after SPK (Sutherland, Gruessner, Dunn, et al., 2001). However, Venstrom, McBride, Rother, and colleagues (2003) found that from “1995-2000, survival for those with diabetes and preserved kidney function and receiving solitary pancreas transplant was significantly worse compared with the survival of waiting list patients receiving conventional therapy.” Of the 11,572 patients enrolled on the UNOS waiting list for pancreas transplants during this period, 5,379 received SPK, 838 received PAK, and 378 received PTA. The authors make the case for the comparability of transplant and wait-listed recipients on the grounds that solitary organ allocation is prioritized not by diabetes severity, but by time on the wait-list, for which the analysis was adjusted so that the groups were comparable. Compared to patients wait-listed for the same procedure, PTA and PAK recipients had a higher relative risk for overall mortality at followup of over 4 years. The relative risk for PTA was 1.57 (95 percent CI = 0.98–2.53; p = 0.06) and for PAK 1.42 (95 percent CI =1.03–1.94, p = 0.03). Survival of SPK recipients was far superior to wait-listed patients, but this analysis did not compare SPK to KTA.

Graft Survival. SPK cadaveric transplantation in patients with diabetic renal disease results in kidney graft survival that is at least equivalent to KTA. A followup study of SPK versus KTA observing patients over a 1- to 8-year period indicated that a pancreas transplant had no detrimental influence on long-term renal function (Hricik, Phinney, Weigel, et al., 1997). The longitudinal analysis by Bunnapradist, Cho, Cecka, and co-workers (2003) found no protective or detrimental effect on renal graft survival at approximately 5 years. Pancreas graft survival is slightly poorer than kidney graft survival (84.7 percent and 92 percent at 1 year, respectively (International Pancreas Transplant Registry, 2003). Aggregating all pancreas transplant procedures, at 3 years, pancreas graft survival is approximately 78 percent (International Pancreas Transplant Registry, 2003).

PAK transplants allow patients the benefits of a living-related donor kidney graft, if available, or a cadaveric kidney graft that is not associated with a simultaneously available pancreas graft. At 1 and 3 years after transplant, 78.5 and 63 percent of PAK transplant patients, respectively, have a functioning pancreas (International Pancreas Transplant Registry, 2003).

As noted, pancreas transplants alone are performed in highly selected patients. Graft survival data suggest that 78.2 and 62 percent of grafts are functioning at 1 and 3 years after transplant, respectively (International Pancreas Transplant Registry, 2003). Adverse outcomes and technical failure rates appear to be increased compared to SPK. Hospital admissions are higher at 73 percent versus 52 percent, respectively, for rejection; 53 percent versus 33 percent for infection, respectively; and 45 percent versus 13 percent for repeat laparotomy, respectively (Stratta, Weide, Sindhi, et al., 1997; Stratta, Taylor, Sindhi, et al., 1996).

The rate of technical failure (nonimmunologic graft loss) is higher for pancreas transplantation of any type than for other routine solid-organ transplants. However, the International Pancreas Transplant Registry reports improvement in technical failure rates comparing 1988-1989 cases to 2000-2001 cases: from 16 to 8 percent, respectively, for SPK; from 16 to 9 percent, respectively, for PAK; and from 19 to 13 percent for PTA, respectively (Gruessner and Sutherland, 2002). Immunologic failure rates have also improved significantly; those reported for SPK, PAK, and PTA transplants were 2 percent, 6 percent, and 9 percent, respectively, for 2000-2001 cases. Improvement in pancreas graft survival is largely due to improvements in immunosuppressive regimens.

Diabetic Complications and Quality of Life. Whole-organ transplantation has clear and positive effects on hypoglycemic and renal complications. Patients with hypoglycemia unawareness despite optimal medical management before transplant no longer have hypoglycemia following successful PTA (Kendall, Rooney, Smets, et al., 1997; Robertson, 1999). Pancreas grafts prevent nephropathy (Wilczek, Jaremko, Tyden, et al., 1995) and established renal lesions may be reversed in nonuremic patients after more than 5 years of normoglycemia (Fioretto, Steffes, Sutherland, et al., 1998). In contrast, histologic changes of diabetic nephropathy commonly recur in diabetic KTA patients within 2 years of transplantation, and progress to endstage disease after 10 years (Najarian, Kaufman, Fryd, et al., 1989).

Polyneuropathy is a common complication of diabetes; whether or not pancreas transplantation alleviates this complication is unclear. For example, Navarro, Sutherland, and Kennedy (1997) reported that progression was significantly delayed and motor and sensory nerve conduction improved in pancreas transplant patients with prior evidence of polyneuropathy compared to type 1 diabetic patients managed medically or with KTA. The effect was greatest 5 to 8 years post-transplantation. In another report, however, not all patients improved, nor did any patient characteristics predict response (Recasens, Ricart, Valls-Sole, et al., 2002).

Although available evidence is inconclusive, some studies suggest that retinopathy may stabilize or improve (Chow, Pai, Chapman, et al.; 1999; Koznarova, Saudek, Sosna, et al., 2000). Pancreas transplantation appears to have a beneficial effect on hypertension (Elliott, Kapoor, Parker, et al., 2001) and may improve cardiac function, but there is no discernable recovery from existing peripheral vascular disease in studies to date (Morrissey, Shaffer, Madras, et al., 1997; Knight, Schanzer, Guy, et al., 1998; Nakache, Merhav, and Klausner, 1999). Effects on progression of early asymptomatic vascular disease are uncertain.

Several studies assessed quality of life, primarily in patients successfully transplanted by SPK, comparing them to patients given SPK transplants who subsequently lost pancreas function, to patients receiving KTA, and to eligible patients not transplanted. Results for several measures generally support significantly improved quality of life after successful transplants (Adang, Engel, van Hooff, et al., 1996; Zehrer and Gross, 1994; Piehlmeier, Bullinger, Kirchberger, et al., 1994; Nakache, Tyden, and Groth, 1994; Kiebert, van Oosterhout, van Bronswijk, et al., 1994; Hathaway, Hartwig, Milstead, et al., 1994). In one study, PTA patients reported better quality of life with insulin independence and immunosuppression than with labile diabetes (Zehrer and Gross, 1991). However, it should be noted that available quality of life studies have serious shortcomings including: lack of comparable control groups; use of different quality of life instruments; use of instruments not validated in transplant patients; and potential selection bias (Holohan, 1995; Robertson, Holohan, and Genuth, 1998).

Table 1 arrays outcomes reported in the literature for whole-organ pancreas transplants contrasted with kidney transplant only or medical management in patients with type 1 diabetes. Where available, data from direct comparison studies were summarized; however, in some cases, summarized data represent indirect comparisons. In some cases, pancreas transplant results are from one type of transplant (e.g., PTA), in other cases from different types of pancreas transplants combined. Note that although registry data are from large numbers of patients, outcomes reported in individual papers typically include fewer patients and, thus, have greater uncertainty.

History of Islet Transplantation

In 1972, Ballinger and Lacy reported the first successful implant of purified rat islets into inbred (autograft) and non-inbred (allograft) diabetic rats (Ballinger and Lacy, 1972). All diabetic immunosuppressed controls died within a few weeks. Longer survival and normalization of blood glucose was observed in both the autografted and allografted rats, although the autografted animals had better results. Successful human islet autotransplantation was reported early (Najarian, Sutherland, Baumgartner, et al., 1980), but only in a small proportion of patients. Successful allotransplantation remained rare for several years.

An automated method for human islet isolation significantly improved yield (Ricordi, Lacy, Finke, et al., 1988) and allowed large scale isolation for clinical studies. Later, a standardized mixture of highly purified enzymes (Liberase) was developed and replaced the variable activities of collagenase lots for separating human islets, improving islet yield and integrity (Linetsky, Bottino, Lehmann, et al., 1997). These advances led to greater standardization of islet processing protocols, allowing clinical trials to proceed at multiple centers. Research continues in order to improve islet yield from autologous and cadaveric pancreata, investigate other islet sources, and develop better methods of immunosuppression and/or tolerance induction for long-term maintenance of transplanted islets.

The last summary of the International Islet Transplant Registry (ITR) reported on 240 islet autografts (140 well documented) between 1990 and 2000 performed at 15 institutions (International Islet Transplant Registry, 2001). These autografts were performed to preserve and restore islets to patients undergoing pancreatectomy, who would otherwise be left diabetic by the surgery. Of these cases, 47 percent were insulin independent at 1 year. However, among patients who received at least 300,000 islet equivalents (IE), 71 percent were insulin independent at 1 year and the rest had better diabetic control than patients undergoing total pancreatectomy without islet transplant (Wahoff, Papalois, Najarian, et al., 1995; Panaro, Testa, Bogetti, et al., 2003). Stable beta-cell function and normal levels of blood glucose after autotransplantation have been reported for up to 13 years (Robertson, Lanz, Sutherland, et al., 2001). Registry data show that increasing the yield of islets is an important success factor (Morrison, Wemyss-Holden, Dennison, et al., 2002).

After the introduction of the Ricordi isolation method, well-documented cases of insulin independence after human islet allotransplantation began to appear (Scharp, Lacy, Santiago, et al., 1990; Warnock, Kneteman, Ryan, et al., 1992; Ricordi, Tzakis, Carroll, et al., 1992; Gores, Najarian, Stephanian, et al., 1993; Bretzel, Brandhorst, Brandhorst, et al., 1999). In some cases, patients maintained insulin independence or graft function for several years (Alejandro, Lehmann, Ricordi, et al., 1997; Davalli, Maffi, Socci, et al., 2000; Cretin, Caulfield, Fournier, et al., 2001). Overall rates of insulin independence at and beyond 1 year, however, remained disappointingly low (11 percent overall, Table 2) until the advent of the Edmonton protocol.

Reports from the Islet Transplant Registry provide the largest dataset on outcomes of patients with type 1 diabetes who received islet allografts in the pre-Edmonton protocol era. Between 1990 and December, 2000, 355 such transplants had been reported to the Registry, of which 237 had at least 1 year of followup (Islet Transplant Registry, 2001). This includes nearly 90 percent of worldwide transplants completed in the same time period. A subsequent meeting presentation reported on a total of 466 well-documented patients transplanted between 1990 and August, 2002 (Brendel, Hering, Schultz, et al., 2002), with 1 year of followup for 270 of these cases. Table 2 summarizes outcomes reported in 2001 for all patients receiving islet transplants under pre-Edmonton protocols (up to 1999) followed for at least 1 year, and for various subgroups of these patients.

Although most patients (73 percent; not shown) demonstrated evidence of insulin production 1 month or more after an islet allograft, only 41 percent of patients had functional islets at 1 year. Furthermore, only 11 percent of all patients remained insulin independent at 1 year. Thus, for the overwhelming majority (89 percent) of patients treated in the pre-Edmonton era, islet allotransplants did not achieve the intended outcome. Little, if any, mortality was associated with the procedure, since 96 percent of transplanted patients remained alive for at least 1 year. The Registry also reported that of 200 patients followed for at least 3 years after an islet allotransplant, 94 percent were alive and 19 percent retained some evidence of islet function (not shown; Islet Transplant Registry, 2001). However, the proportion remaining insulin independent at 3 years was not reported. In a later update (Brendel, Hering, Schultz, et al., 2002), the Registry reported functional graft survival in 24 percent, insulin independence in 4 percent, and overall survival in 95 percent of 235 patients followed for at least 3 years.

The overwhelming majority of patients given islet allotransplants in the pre-Edmonton era were treated either simultaneously with (simultaneous islet/kidney, SIK, 55 percent) or after (islet after kidney, IAK, 37 percent) a kidney transplant (Islet Transplant Registry, 2001). Only nine patients (4 percent) received an islet transplant alone (ITA). Immunosuppression regimens for most patients transplanted using pre-Edmonton protocols likely were primarily based on those used to manage kidney transplant recipients. The subsequent update (Brendel, Hering, Schultz, et al., 2002) included 138 SIK patients and 90 IAK patients. Among these, functional graft survival at 1 year was 51 percent and 40 percent, respectively, while insulin independence at 1 year was 9 percent and 15 percent, respectively.

Registry analyses identified several factors that influence outcomes of islet transplants (Islet Transplant Registry, 2001). These include the site of the transplant (data not tabulated; liver [n = 220] versus others [n = 17]); the number of islet equivalents transplanted per kilogram of body weight (≥6000 versus <6000); the duration of cold ischemia from cross-clamping to islet isolation (≤8 hours versus >8 hours); and the regimen used to induce immunosuppression (no anti-T cell antibody versus ATG, ALG, or antibody to IL-2R versus OKT3). A substantially greater proportion of transplants that were favorable on all four of these predictive criteria than of those that failed on one criterion or more remained functional (31 percent versus 9 percent, respectively) and maintained insulin independence (24 percent versus 5 percent, respectively) at 1 year after treatment (Table 2). Note also that success was more frequent among those patients transplanted in years 1998 and 1999 than among those transplanted earlier.

The Registry reports did not include any data on effects of insulin independence following islet allotransplantation on diabetic complications. Fiorina, Folli, Maffi, and co-workers (2003) followed 37 islet allotransplanted type 1 diabetic kidney transplant patients, of whom, 24 maintained islet function (C-peptide >0.5 ng/mL) for longer than 1 year and 13 lost or never achieved islet function during the first year, for an average of 63 months. Patients with successful islet transplants had significantly reduced: cardiovascular and all-cause mortality; microvascular-endothelial injury; atherothrombotic risk factors; and renal damage (as measured by urine albumin excretion) compared to those whose transplants were not successful (Table 3). Additionally, the cardiovascular death rate for successful islet transplant patients was similar to that of a control group of whole-organ pancreas transplant patients, and better than that of kidney-alone transplant patients. Patients with successful islet transplants had significantly lower exogenous insulin requirements than those with unsuccessful transplants at all timepoints.

An earlier small study (n = 8) reported that stable islet function after an allograft (n = 6) significantly reduced HbA_1c concentrations and insulin requirements (Alejandro, Lehmann, Ricordi, et al., 1997). Over 6 years of followup, these patients also remained free of the severe hypoglycemic episodes observed in the Diabetes Control and Complications Trial, even though they were not insulin independent. Other investigators also have reported that functional islets after allotransplants decrease hypoglycemia unawareness and improve hormonal counter-regulation in response to hypoglycemia (Meyer, Hering, Grosmann, et al., 1998).

Outcomes of Interest

Based on the Biological Response Modifiers Advisory Committee meeting (U.S. Food and Drug Administration, 2003), a consensus definition of success for islet transplantation is: Restoration of sustained euglycemia (i.e., absence of hyper- and hypoglycemia) with no or a reduced exogenous insulin requirement. Clinical outcome parameters that can be used together to measure success are insulin independence or percent of prior insulin use, hypoglycemic episodes, and quality of life.

• Insulin independence: Islet transplantation attempts to restore normal glucose metabolism by in vivo production of insulin regulated by changing glucose concentrations without the need for exogenous insulin supplementation. The percentage of patients who do not require exogenous insulin at yearly post-transplant intervals is a direct measure of success.

• Percent prior insulin use: failing absolute insulin independence, successfully transplanted patients may attain good control without glycemic excursions accompanied by a marked decrease in the need for exogenous insulin.

• Hypoglycemic episodes: Hypoglycemia unawareness and life-threatening hypoglycemic episodes can be a consequence of strict glucose control with exogenous insulin, and are indications for pancreas whole-organ or islet transplantation. Elimination of hypoglycemic episodes in conjunction with glycemic control is also a direct measure of transplant success.

• 

  Table 5. Measures for evaluation of transplantation (more...)

  Table 5. Measures for evaluation of transplantation and quality of life

  Survey	Description	Reference
  36-item Short Form health survey (SF-36)	Evaluates general quality of life. Survey addresses 8 areas of health status, summarized in 2 component scores:	Terada and Hyde, 2002; Ware and Sherbourne, 1992
  	Physical component: physical functioning; physical limitations; pain; general health perception
  	Mental component: vitality; social functioning; emotional limitations; mental health
  Health Utilities Index Mark 2 (HUI2)	Comprehensive description of health status in 8 core domains:	Furlong, Geeny, Torrance, et al., 2001
  	Vision; hearing; speech; ambulation; dexterity; emotion; cognition; pain.
  Hypoglycemia Fear Survey (HFS)	Addresses behaviors and worries related to potential hypoglycemic episodes.	Cox, Irvine, Gonder-Frederick, et al., 1987
  	Increased awareness of hypoglycemia correlates with decreased HFS scores in validation studies.
  Immunosuppressant QOL Survey (Memphis Survey)	Evaluates side effects of immunosuppressive therapy for organ transplantation.	Winsett, Stratta, Alloway, et al., 2001
  	Employs 4 subscales: emotional burden, life/role responsibilities, mobility, gastrointestinal distress.

  Quality of life: Islet transplantation can improve quality of life for patients by eliminating hypoglycemic and hyperglycemic episodes, the need for insulin injections, frequent self-monitoring of blood glucose levels, and dietary restrictions. General and specific standardized measures of quality of life may include the Health Utilities Index, SF-36, Immunosuppressant Quality of Life (QOL) Survey, and Hypoglycemia Fear Survey (Johnson, 2002). Characteristics of these instruments are briefly summarized in Table 5.

Although sustained euglycemia may be of highest clinical interest, in the absence of well-controlled studies, insulin independence may be the most persuasive measure available to establish the success of the procedure.

Improvement in long-term diabetic outcomes is the measure of ultimate success of islet transplantation in type 1 diabetes. The objective is to reduce or eliminate long-term diabetic outcomes such as retinopathy, neuropathy, nephropathy, and cardiovascular disease. Transplantation after complications have already become apparent may not be able to reverse or even stabilize the process. Transplantation prior to complications is more likely to delay or preclude their occurrence, but studies will require a minimum of 5–10 years in order to collect robust data on clinical outcomes.

Potential adverse events of islet transplant may be direct consequences of the procedure (e.g., hemorrhage or thrombosis from percutaneous access of the portal vein) or the continued immunosuppression needed to maintain viability and function of the transplanted islets. Adverse effects of immunosuppression may be near-term (e.g., mouth ulceration, diarrhea, anemia) or long-term (e.g., renal insufficiency, post-transplant lymphoproliferative disorders, other malignancies, cytomegalovirus or other infections).

Measurement of C-peptide and HbA_1c (glycated hemoglobin) are biological outcomes that are indicators of graft function and glycemic control, respectively.

• C-peptide: C-peptide is an inactive cleavage product of insulin production. Because C-peptide is metabolized minimally by the liver, has a longer half-life than insulin, and measurement is not affected by the presence of exogenous insulin, serum C-peptide levels are better indicators of beta-cell function than the peripheral insulin concentration (Sacks, 1999), and thus, the preferred measure for monitoring post-transplant islet function.

• HbA_1c: Measurement of glycated hemoglobin is the standard method for assessing glycemic history over 2–3 months. HbA_1c is the standard assay for measurement of glycated hemoglobin and, thus, of post-transplant glycemic control.

A variety of metabolic measures are available to estimate pancreatic beta cell functional reserve. These include intravenous glucose tolerance tests (IVGTT), from which can be calculated the acute insulin response to glucose (AIR_g), glucose disposal (K_G), and areas under the curve for insulin and C-peptide (AUC_i and AUC_C-p, respectively); intravenous arginine stimulation, from which are derived the acute insulin response to arginine (AIR_arg) and the acute C-peptide response to arginine (ACPR); oral glucose tolerance testing (OGTT); and mixed meal stimulation. Several of these measures have been reported as near normal and stable over at least 5 years in pancreas organ transplant recipients (Robertson, Sutherland, Kendall, et al., 1996; Robertson, 2003). Islet transplant recipients have demonstrated results that are similar to those of segmental pancreas graft recipients, but lower than those of whole-organ transplant patients, despite exogenous insulin independence (Secchi, Taglietti, Socci, et al., 1999).

Various metabolic measures have been reported in conjunction with islet transplant outcomes by a few centers (Baidal, Froud, Ferreira, et al., 2003; Hering, Kandaswamy, Harmon, et al., 2004 [In press] Ryan, Lakey, Paty, et al., 2002). However, there does not yet appear to be consensus on the measures most predictive of continuing beta-cell function, on the clinical significance of impaired glucose tolerance in islet transplantation (Baidal, Froud, Ferreira, et al., 2003), nor is there consistent reporting among the majority of transplant centers.

Collaborative Islet Transplant Registry (CITR)

The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) initiated and funded the Collaborative Islet Transplant Registry (CITR) in September, 2001. As stated on the CITR website (http://spitfire.emmes.com/study/isl/index.html), the goals of the Registry are:

• “To develop and implement standards for reporting islet/beta cell transplants and their outcome.

• To collect and compile data on all islet/beta cell transplants in human recipients performed in the United States and Canada. [European transplant data will also be included under a funding arrangement with the Juvenile Diabetes Research Foundation]

• To increase the safety of islet/beta cell transplantation by disturbing electronically the pertinent information of submitted serious adverse event reports to all participating clinical centers in a timely fashion.

• To perform scientific analysis on islet/beta cell transplant data, with particular emphasis on:

  • - safety of islet/beta cell transplant product and procedure and protocol-regulated treatment products;

  • - number of islet/beta cell transplants and retransplants performed, categorized by transplant institution, donor tissue source and handling, recipient category, transplant technique and site, and recipient treatment protocols;

  • - efficacy of islet/beta cell transplants as defined by standardized outcome measures and as determined by donor factors, recipient demographics, donor-recipient matching, islet/beta cell processing and product characteristics, transplant technique and site, recipient treatment, and post-transplant events.

• To communicate comprehensive and current information on islet/beta cell transplantation to transplant institutions, the diabetes and general health care community, and the interested general public via the CITR website (http://www.citregistry.org), publications, and presentations.

• To stimulate prospective and retrospective studies on emerging issues of importance.”

The Registry is now completing collection of data from participating institutions. Unfortunately, data from the first CITR report were not yet available at the time this evidence report was being prepared.

Article Selection

All abstracts initially retrieved by the search strategy were reviewed by one researcher who also reviewed the fulltext articles to determine whether study selection criteria were met (MP). Selected papers were abstracted by a single reviewer (MP or JS) and evidence tables were fact-checked by a second reviewer (MP or JS).

Although a total of 2,052 abstracts were initially reviewed, very few articles were retrieved as almost all indexed clinical studies were completed prior to the adoption of the Edmonton protocol. Of the studies relevant to the Edmonton protocol, the vast majority were reviews, animal studies, or technical reports. Including articles published and retrieved during the preparation of this review, only 12 published studies (Owen, Ryan, O'Kelly, et al., 2003; Ryan, Lakey, Paty, et al., 2002; Paty, Ryan, Shapiro, et al., 2002; Johnson, Kotovych, Ryan, et al., [In press]; Markmann, Deng, Huang, et al., 2003; Goss, Schock, Brunicardi, et al., 2002; Kaufman, Baker, Chen, et al., 2002; Shapiro, Lakey, Ryan, et al., 2000; Ryan, Lakey, Rajotte, et al., 2001; Markmann, Deng, Desai, et al., 2003; Hering, Kandaswamy, Harmon, et al., 2004 [In press]; Hirshberg, Rother, Digon, et al., 2003) reported efficacy and adverse outcomes, and 2 additional (Casey, Lakey, Ryan, et al., 2002; Goss, Soltes, Goodpastor, et al., 2003) reported only adverse outcomes.

Additional sources of evidence. Due to the scarcity of published articles, additional sources of evidence were sought. Abstracts and presentations from scientific conferences were reviewed, and those meeting study selection criteria are summarized in this review as supplementary sources that provide preliminary results of studies anticipated to be fully reported in the next 2 years. The scientific conferences reviewed were:

• XIX International Congress of the Transplantation Society; 2002 August 25–30; Miami, FL (abstracts searched)

• City of Hope Rachmiel Levine Symposium; 2002 October 9–12; Anaheim, CA (attended)

• Islet Transplantation 2002 and Beyond: 2nd Annual Annenberg Symposium; 2002 December 5–7; Rancho Mirage, CA (attended)

• American Transplant Congress 2003: The Fourth Joint American Transplant Meeting, May 30, 2003 - June 4, 2003, Washington, DC (abstracts searched)

• 9th Congress of the International Pancreas and Islet Transplant Association. 8–11 July 2003, Dublin, Ireland (abstracts searched)

• 1st Islet Transplant Congress; 2003 November 13–16; Miami Beach, FL (attended)

In the future, the most comprehensive source of data will be the Collaborative Islet Transplant Registry (CITR), which is initiated and funded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). The registry is coordinated by the EMMES Corporation of Rockville, Maryland; Dr. Bernhard Hering of the University of Minnesota is the Medical Director.

CITR was initiated in September 2001. CITR is collecting extensive, retrospective data on all patients who have received islet transplants using Edmonton or subsequently developed transplant protocols, and will maintain an ongoing data collection process for new patients. The data elements are more comprehensive than those collected by the previous International Islet Transplant Registry, and require original data entry (i.e., data from the International Islet Transplant Registry has not been downloaded into the CITR database).

As of November 2003, CITR was still collecting data from the participating institutions. Thus, data and analyses from CITR were not available for this evidence report.

Technical Expert Panel and Peer Review

The development of this evidence report was subject to extensive expert review including ongoing guidance from a Technical Expert Panel (TEP) and document review by the TEP and by a panel of designated peer reviewers (Appendix B lists the members of the Technical Expert Panel and external peer reviewers).

TEP members provided ongoing guidance and review on all phases of this project including review of the draft report.

The draft report was also reviewed by a panel of external peer reviewers that included experts in endocrinology, pancreas transplantation, and islet transplantation, as well as a patient advocacy representative. Reviews were also solicited from the Immune Tolerance Network (ITN; currently overseeing multicenter studies of the Edmonton protocol), the Juvenile Diabetes Research Foundation (currently funding, along with the National Institutes of Health, ITN studies of the Edmonton protocol), and the American Diabetes Association. Comments were elicited from external peer reviewers using a structured comment form, compiled, and submitted with a description of comment disposition to the Agency for Healthcare Research and Quality.

Overview

Clinical and biological outcomes of islet transplantation from peer-reviewed, published studies constitute the main evidence for this systematic review. Only studies of protocols incorporating new preparation methods, sufficient islet mass, and glucocorticoid-sparing, reduced-dose immunosuppressive regimens were considered. The first of these protocols was developed at the University of Alberta at Edmonton and has since been called the Edmonton protocol. Subsequent protocols have varied different aspects of islet preparation, transplant procedure and/or immunosuppressive regimen seeking to improve outcomes. Reports on the outcomes of islet transplantation often combine results from patients treated using different protocols. This review makes no attempt to compare the outcomes of different protocols; however, the reader should note where indicated, that different protocols are being used.

Evidence Tables 1 and 2 summarize six patient series, reporting on 64 patients² in the 10 most recent publications from these series (Ryan, Lakey, Paty, et al., 2002; Owen, Ryan, O'Kelly, et al., 2003; Paty, Ryan, Shapiro, et al., 2002; Johnson, Kotovych, Ryan, et al., [In press]; Hering, Kandaswamy, Harmon, et al., 2004 [In press]; Goss, Schock, Brunicardi, et al., 2002; Goss, Soltes, Goodpastor, et al., 2003; Markmann, Deng, Huang, et al., 2003; Hirshberg, Rother, Digon, et al., 2003; Kaufman, Baker, Chen, et al., 2002) Older publications superceded by later updates of similar information are not summarized, but the corresponding citations are indicated in footnotes (Shapiro, Lakey, Ryan, et al., 2000; Ryan, Lakey, Rajotte, et al., 2001). Later publications detailing information already summarized in earlier publications are also indicated in footnotes (Casey, Lakey, Ryan, et al., 2002; Markmann, Deng, Desai, et al., 2003).

Although an attempt was made to summarize the most recent outcomes for each reporting center and pool results for an overall summary, this was not possible. First, some centers reported different outcomes on different numbers of patients in more than one publication, precluding an accurate synthesis. Second, different centers reported the same type of outcome in different ways. For example, HbA_1c was reported for either all patients, or for only those who remained insulin independent. Thus, a standardized data collection, such as that in progress by the Collaborative Islet Transplant Registry (CITR), will be needed for an accurate and complete data summary. For these reasons, data are generally presented here by center.

No attempt was made to compare islet transplant outcomes with those for whole-organ pancreas transplantation. Initial review of the islet transplant literature revealed the following limitations: small patient sample sizes from a small number of islet transplant centers; relatively short followup times; and, as noted, variably reported outcomes. In particular, success in islet transplantation is dependent on the use of protocols that have been only recently introduced, and for which there are minimal data beyond 1–2 years of follow-up. In contrast, outcomes of whole-organ pancreas transplantation are available for 5 to 8 years' post-transplant.

Evidence Table 1. Clinical islet transplantation: Patient and transplant characteristics reported in journal articles. All transplants are islet alone unless otherwise indicated.

Evidence table 2. Clinical islet transplantation: Outcomes reported in journal articles. All transplants are islet alone unless otherwise indicated.

Because islet transplant studies report pre- versus post-transplantation outcomes, the implicit comparison in these studies is to pretransplantation medical management. This is based on the assumption that, without intervention, patients would continue medical management and would require exogenous insulin, maintain higher HbA_1c, and experience hypoglycemic episodes at constant doses, levels, and numbers, respectively.

Patients

Sixty-four patients in all series combined had type 1 diabetes of 5 or more years' duration, stimulated serum C-peptide less than 0.5 ng/mL (0.16 nmol/L), and severe metabolic instability or recurrent hypoglycemia unawareness despite optimal medical management. Although a few patients with progressive complications of diabetes were initially enrolled in the University of Alberta series (Ryan, Lakey, Paty, et al., 2002), this indication was discontinued. Study patients meeting eligibility criteria may also have had retinopathy or neuropathy, but except for the patient reported by Kaufman, Baker, Chen, and co-workers (2002), patients neither had received nor needed a renal transplant.

Study protocols either called for 2 islet transplant procedures (generally the equivalent of islets prepared from two whole organs) or a second transplant within a few months if the first did not achieve insulin independence. Of the 64 enrolled patients, 52 had completed the islet transplant protocol at the time of the report; the rest were awaiting an additional transplant or, in the cases of two patients, had withdrawn from the protocol (Markmann, Deng, Huang, et al., 2003; Hirshberg, Rother, Digon, et al., 2003). In some cases, patients who had completed the protocol were given a supplemental islet transplant to achieve or maintain insulin independence. Patients were followed for up to 3 years in the University of Alberta, Edmonton series (Owen, Ryan, O'Kelly, et al., 2003: Ryan, Lakey, Paty, et al., 2002; Paty, Ryan, Shapiro, et al., 2002; Johnson, Kotovych, Ryan, et al., [In press]), up to 22 months at the National Institutes of Health (Hirshberg, Rother, Digon, et al., 2003), up to 1 year at the University of Minnesota (Hering, Kandaswamy, Harmon, et al., 2004 [In press]) and the University of Pennsylvania (Markmann, Deng, Huang, et al., 2003), and for no more than 4 months at Baylor (Goss, Schock, Brunicardi, et al., 2002; Goss, Soltes, Goodpastor, et al., 2003), and Northwestern University (Kaufman, Baker, Chen, et al., 2002).

Clinical Outcomes

Five centers published diabetic clinical outcomes on 47 patients ³ who completed an islet-alone transplant protocol (Owen, Ryan, O'Kelly, et al., 2003; Hering, Kandaswamy, Harmon, et al., 2004 [In press]; Markmann, Deng, Huang, et al., 2003; Goss, Schock, Brunicardi, et al., 2002; Hirshberg, Rother, Digon, et al., 2003). Of these, 44 patients (94 percent) achieved insulin independence over the 3-month, post-transplant period.

In the largest series, Owen, Ryan, O'Kelly, and co-workers (2003) reported all 26 patients who completed the Edmonton transplant protocol were initially insulin independent, and that 21 patients remained independent 1 year after their first transplant. In an earlier report on the same series, Ryan, Lakey, Paty, and co-workers (2002) reported 11 of 17 fully transplanted patients were insulin independent at a median of 20 months' followup. Because only 5,000 islet equivalents per kilogram (IEq/kg) on average were isolated from a single donor organ, and approximately 13,000 IEq/kg were required for insulin independence, a third, supplemental transplantation was required in seven cases (Owen, Ryan, O'Kelly, et al., 2003). During the course of this patient series, the original Edmonton protocol was modified with techniques such as islet culture and 2-layer perfluorocarbon organ storage to improve islet viability.

Hering, Kandaswamy, Harmon, and colleagues (2004 [In press]) reported on six patients, each transplanted with cultured islets prepared from a single donor after 2-layer perfluorocarbon organ storage. All patients received immunosuppression with humanized OKT3 anti-T cell monoclonal antibody (Ala-Ala), sirolimus, and reduced-dose tacrolimus. Four of six patients initially achieved insulin independence and remained independent at 12 months, while one patient required 60% of the pretransplant insulin requirement. The last patient was lost to followup.

Of seven patients receiving cultured islets and Edmonton immunosuppression (daclizumab; sirolimus; reduced-dose tacrolimus), Markmann, Deng, Huang, and co-workers (2003) reported that all patients were initially insulin independent. Five achieved independence after only one islet infusion, although in two of these cases the single infusion delivered pooled islets from two donor organs. In this series, six of seven patients followed for 13 months or less remained insulin independent, including one who received a supplemental transplant at 13 months.

Goss, Schock, Brunicardi, and co-workers (2002) reported on three patients, who remained insulin independent after 4 months or less followup. Hirshberg, Rother, Digon, and co-workers (2003) transplanted six patients, of whom, four initially achieved and two remained insulin independent at 17 to 22 months post-transplant. Kaufman, Baker, Chen, and co-workers (2002), reported on the only published transplant that was not islet alone. They transplanted islets into a patient with a functional renal allograft and achieved insulin independence after only one infusion of 4,100 IEq/kg. However, followup was only 4 months.

Three centers reported that 28 of 37 patients (76 percent of those completing a transplant protocol) maintained insulin independence for 1 year (Owen, Ryan, O'Kelly, et al., 2003; Hering, Kandaswamy, Harmon, et al., 2004 [In press]; Hirshberg, Rother, Digon, et al., 2003). Only one published study (from the Edmonton group) reported on patients with 2 years of followup: four of six patients remained insulin independent (Ryan, Lakey, Paty, et al., 2002). Patients who did not achieve or who lost insulin independence tended to use far less insulin than before transplantation. Ryan, Lakey, Paty, and co-workers (2002) reported daily insulin doses were 57 percent of pretransplant doses for six patients who resumed insulin use at a median of 10 months; three of these patients continued to produce C-peptide. Markmann, Deng, Desai, and co-workers (2003) described one patient who lost insulin independence but continued to produce detectable C-peptide and required only 50 percent of the pretransplant insulin dose.

All series reported abatement of hypoglycemic episodes in insulin-independent transplant patients. In three series reporting on 26 patients completing the transplant protocol, hypoglycemic episodes were also abated in nine patients with continuing C-peptide secretion who required some exogenous insulin at 1 year (Ryan, Lakey, Paty, et al., 2002; Hering, Kandaswamy, Harmon, et al., 2004 [In press];Hirshberg, Rother, Digon, et al., 2003). All series described severe, recurrent hypoglycemic episodes (usually requiring the assistance of others and despite intensive medical management; see Evidence Table 1) as the most common indication for islet transplantation. However, only one report quantified the number of such episodes in the year pre- (5–100 episodes in each of six patients) and post-transplant (0 episodes, all patients; Hering, Kandaswamy, Harmon, et al., 2004 [In press]).

Although not mentioned in published reports, the Edmonton group has provided greater detail on patient indications for their center in subsequent meeting presentations: reduced hypoglycemia awareness as defined by the absence of adequate autonomic symptoms at plasma glucose levels less than 54 mg/dL; metabolic lability/instability characterized by two or more episodes of severe hypoglycemia associated with blood glucose less than 54 mg/dL and requiring assistance of another person within 12 months; or two or more hospital visits for diabetic ketoacidosis over the previous 12 months (Shapiro, 2003). Standard, quantifiable eligibility criteria for islet transplant alone are being developed and evaluated (Ryan, Shandro, Vantyghem, et al., 2003).

While insulin independence is the gold standard for clinical outcomes, analogous to a functioning whole pancreas transplant, the consensus recommendation of the U.S. Food and Drug Administration (FDA) Biological Response Modifiers Advisory Committee was that restoration of sustained euglycemia (i.e., absence of hyper- and hypoglycemia) with no or a reduced exogenous insulin requirement should be the primary definition of success for islet transplantation (U.S. Food and Drug Administration, 2003). Clear information on this outcome was available from two publications with at least 1 year of followup (Ryan, Lakey, Paty, et al., 2002; Hering, Kandaswamy, Harmon, et al., 2004 [In press]). Of 23 patients completing a transplant protocol, 19 (83 percent) were euglycemic.

Biological Outcomes

In each series and for all insulin-independent patients, mean HbA_1c decreased to levels recommended to avoid or delay progression of diabetic complications (i.e., <7 percent at a minimum, American Diabetes Association, 2003; recommended <6.5 percent, U.S. Food and Drug Administration, 2003); mean pretransplant baseline levels were all greater than 7 percent except for one case report (6.9 percent; Kaufman, Baker, Chen, et al., 2002). Patients with measurable graft function, but requiring some exogenous insulin, were seldom reported separately. However, in one study, two patients who did not achieve complete insulin independence but produced measurable C-peptide for several months were able to achieve HbA_1c levels near 7 percent or less on reduced doses of exogenous insulin (Hirshberg, Rother, Digon, et al., 2003). In another study, HbA_1c remained elevated in one patient despite measurable C-peptide and reduced insulin requirements (Hering, Kandaswamy, Harmon, et al., 2004 [In press]).

In the largest series, post-transplant fasting and stimulated C-peptide levels increased to near those of nondiabetic controls, but post-transplant stimulated glucose levels remained higher than controls, although lower than pretransplant (Ryan, Lakey, Paty, et al., 2002; data not shown). Although an attempt was made to abstract metabolic outcomes data across centers, results were not reported consistently for one or a few specific tests to allow comparison (Evidence Table 3). Metabolic outcomes await standardization and routine data collection.

Interestingly, Paty, Ryan, Shapiro, and co-workers (2002), reporting on a subset of seven patients with prolonged insulin independence and absence of hypoglycemic symptoms from the University of Alberta series, found that glucagon responses to hypoglycemia were similar to those of nontransplanted type 1 diabetic patients and that epinephrine response and hypoglycemic symptom recognition were not restored. The glucagon results are similar to those reported for pre-Edmonton islet transplants (Meyer, Hering, Grossmann, et al., 1998), although in this study, glycemic thresholds and/or peak incremental responses of epinephrine, norepinephrine, and cortisol, as well as symptom recognition, improved in islet transplanted patients. In contrast, glucagon responses to hypoglycemia and symptom recognition reportedly were fully restored and maintained long-term in whole pancreas transplant patients (Kendall, Rooney, Smets, et al., 1997; Paty, Lanz, Kendall, et al., 2001).

Evidence Table 3. Types of metabolic testing reported by transplant centers.

Quality of Life

No general measures of pre- versus post-islet transplant quality of life have been published; only one publication reports limited data on specific measures. Johnson, Kotovych, Ryan, and co-workers (In press) summarized results of patients in the University of Alberta series who completed the Hypoglycemia Fear Survey (HFS) and other quality of life indices at baseline (pretransplant), between the first and second islet infusions, at 1, 3, 6, and 12 months, and annually thereafter. Forty-six patients completed baseline surveys and 35 transplanted patients completed additional surveys. Results of the last survey for each transplant patient were compared to baseline results. Transplant patients had slightly lower fear of hypoglycemia by HFS at 1 month than pretransplant patients (Evidence Table 4) but a much larger reduction by month 3. Overall, HFS in pretransplant patients was significantly higher than for post-transplant patients at a median followup of 11.9 months (p < 0.0001).

The authors note that these data are limited by the small sample size and their cross-sectional nature. They represent the beginning of longitudinal data collection on a growing population of patients. As presented at the Annenberg Second Annual Symposium (Johnson, 2002), Johnson and colleagues are collecting data from both general and disease-specific quality of life instruments for eventual analysis of quality of life trends over time.

Evidence table 4. Results of Hypoglycemia Fear Survey post-islet transplant compared to baseline.

Long-Term Diabetic Complications

Only Ryan, Lakey, Paty, and colleagues (2002) reported on long-term diabetic complications. Of 17 subjects who had completed the transplant protocol at the time of this publication and were followed for 3 years or less, three patients had progression of their retinopathy requiring laser photocoagulation. Nine patients either started or increased antihypertensive therapy. Cholesterol rose in 15 patients and in 11 required statin therapy. There were no major changes in neuropathy. Serum creatinine and urine protein did not change significantly except for two patients with pre-existing renal impairment who suffered serious deterioration in renal function; both patients' renal dysfunction was stabilized by withdrawing potentially nephrotoxic tacrolimus.

Adverse Events

Portal vein hypertension, thrombosis and hemorrhage. One center reported branched portal vein thrombosis (main portal veins remained patent in all patients) in two of 34 transplanted patients (Ryan, Lakey, Paty, et al., 2002; Casey, Lakey, Ryan, et al., 2002; Owen, Ryan, O'Kelly, et al., 2003). One of these patients also developed a hepatic hematoma requiring surgery, but without other long-term consequences. Four additional patients transplanted early in this series experienced hemorrhage from percutaneous portal vein access (Ryan, Lakey, Paty, et al., 2002). Of six patients described in the initial experience at the U. S. National Institutes of Health (NIH) between December, 2000 and June, 2001, symptomatic portal vein thrombosis occurred in one and intra-abdominal hemorrhage requiring transfusion occurred in another (Hirshberg, Rother, Digon, et al., 2003). Various measures have been adopted to minimize risks of these complications (Goss, Soltes, Goodpastor, et al., 2003; Owen, Ryan, O'Kelly, et al., 2003; Baidal, Froud, Ferreira, et al., 2003; Froud, Yrizarry, Alejandro, et al., [In press]). Four other groups did not encounter portal vein thrombosis or hemorrhage among 24 transplanted patients (Evidence Table 2; Hering, Kandaswamy, Harmon, et al., 2004 [In press]; Markmann, Deng, Huang, et al., 2003; Goss, Soltes, Goodpastor, et al., 2003; Kaufman, Baker, Chen, et al., 2002).

Nearly all of 26 patients evaluated for pressure changes during the procedure experienced transient moderate increases of portal venous pressure that correlated with the volume and number of islets transfused (Casey, Lakey, Ryan, et al., 2002). However, only one of the two patients with thrombosis had a detectable post-infusion change in main portal venous pressure. Investigators attributed the hepatic hematoma in a third patient to a higher dose of anticoagulant therapy used to moderate increases in portal venous pressure. No patients showed evidence of sustained portal vein hypertension over 17 months of followup.

Other acute procedural complications. The Edmonton group reported transient acute bradycardia in two of 30 transplanted patients (Ryan, Lakey, Paty, et al., 2002). A second group reported moderate abdominal pain in 71 percent and nausea in 59 percent of eight transplanted patients (Goss, Soltes, Goodpastor, et al., 2003). Of six patients transplanted at the University of Minnesota, one experienced a severe rash attributed to antifungal therapy and one had a transient but severe elevation of liver transaminase activity in the serum (Hering, Kandaswamy, Harmon, et al., 2004 [In press]). The remaining three centers reported no other serious procedural complications among 16 transplanted patients (Markmann, Deng, Huang, et al., 2003; Kaufman, Baker, Chen, et al., 2002; Hirshberg, Rother, Digon, et al., 2003).

Adverse effects of continued immunosuppression. Published studies reported no instances of post-transplant lymphoproliferative disorder (PTLD), other malignancy, reactivation or transfer of cytomegalovirus (CMV) infection, opportunistic infections, or other long-term consequences of immunosuppression (Evidence Table 2). The absence of CMV infection is noteworthy since at least 16 CMV-negative patients were transplanted with islets from CMV-positive donors in two of the series (Ryan, Lakey, Paty, et al., 2002; Hirshberg, Rother, Digon, et al., 2003). In contrast, nearly 85% of similarly mismatched solid organ transplants (CMV-positive donor to CMV-negative recipient) reportedly developed CMV infection (Preiksaitis, Lakey, LeBlanc, et al., 2002). The absence of PTLD is encouraging but may reflect the small sample size thus far; incidences in the first year after kidney or heart transplant are 0.2% and 1.2%, respectively (Riddell, 2001).

Three studies reported effects of immunosuppressive therapy. In the largest series, mouth ulcers (n = 15), diarrhea (n = 10) and anemia (n = 8) were most common among 17 patients who completed the protocol (Ryan, Lakey, Paty, et al., 2002). Other adverse events included nausea and vomiting (n = 2), arthralgias, and rheumatoid arthritis (n = 1, each).

The first two doses of anti-CD3 humanized monoclonal antibody (Ala-Ala) used for immunosuppression at the University of Minnesota were associated with mild-to-moderate infusion reactions (Hering, Kandaswamy, Harmon, et al., 2004 [In press]). They also reported transient neutropenia in three of six patients, and mild-to-moderate mouth ulcers and weight loss in an unspecified number.

Recurrent oral ulcers, episodic diarrhea, and normocytic anemia occurred in all six patients in the initial NIH experience (Hirshberg, Rother, Digon, et al., 2003). Leg edema and generalized fatigue each occurred in five of six patients; transient, mild thrombocytopenia in four; and two patients experienced severe neutropenia that recovered after myeloid growth factor treatment or when a myelotoxic antibiotic regimen was discontinued. Infections of the skin and urinary tract each occurred in one patient. One patient developed interstitial pneumonitis from sirolimus toxicity, and one experienced severe diarrhea and fatigue.

Patient tolerance of immunosuppressive regimens has been a concern for some, but not all groups. Markmann, Deng, Huang, et al. (2003) discontinued immunosuppression without completing the transplant protocol for one patient due to concern that sirolimus might have been preventing healing of a traumatic foot wound. Immunosuppression was discontinued for one NIH patient due to sirolimus-related pneumonitis, resulting in loss of most graft function. A second NIH patient with graft function but requiring some exogenous insulin elected to discontinue immunosuppression due to intolerable diarrhea, fatigue, weight loss, and renal dysfunction, but also retained minimal islet function (Hirshberg, Rother, Digon, et al., 2003). Similarly, two patients transplanted at the University of Minnesota discontinued immunosuppression while having partial islet function and subsequently rejected their grafts (Personal communication, Hering B, December 2003). No patients who completed the transplant protocol at the Universities of Miami or Alberta discontinued all immunosuppression (Personal communication, Alejandro R, December 2003; Personal communication, Shapiro J, December 2003).

Autoantibodies and sensitization. Only 2 centers reported on the development of islet autoantibodies. Of six patients followed by Hering, Kandaswamy, Harmon, and colleagues (2004 [In press]), three were positive for glutamic acid decarboxylase (GAD) and/or islet cell antigen (ICA) autoantibodies to pancreatic islet cells pre- and post-transplant and two additional islet recipients became GAD-positive post-transplant (only one of four islet cell autoantibody-positive patients did not achieve insulin independence). Insulin autoantibodies were positive in six patients pretransplant, and in five patients post-transplant.

Of 17 Edmonton patients who completed the transplant protocol at the time of the publication, three lost C-peptide secretion and two of these had developed GAD and ICA autoantibodies (Ryan, Lakey, Paty, et al., 2002). One patient with continued C-peptide secretion but requiring exogenous insulin had a rise in GAD and ICA antibodies.

Only one center reported on allosensitization to donor leukocyte antigens (Markmann, Deng, Huang, et al., 2003). All nine patients were negative for panel reactive antibodies (PRA), a test to measure what percentage of a patient serum sample reacts to a panel of known human leukocyte antigens, both pre- and post-transplant. One patient became PRA-positive after terminating immunosuppression. While these are insufficient data for conclusions, the lack of PRA is encouraging. In kidney and kidney-pancreas transplants, as much as 20% PRA has been detected (Pelletier, Hennessy, Adams, et al., 2002) and patients who lose their first transplant and develop broadly reactive antibodies (>50% PRA) have poorer graft survival upon retransplant than those who do not develop broad reactivity (Cecka, 1996).

Hepatic tissue changes. One study reported that periportal steatosis was detected by magnetic resonance imaging (MRI) after islet transplant in two of four patients (Markmann, Rosen, Siegelman, et al., 2003). Steatosis correlated with post-transplant functional islet grafts and may be a consequence of locally elevated insulin concentration in the portal tracts where functioning islets are located. Long-term consequences, if any, of the tissue changes are uncertain although the Edmonton group noted normal liver function tests in patients whose MRI scans showed evidence of steatosis (U.S. Food and Drug Administration, 2003). Additionally, similar hepatic changes observed in patients transplanted with autologous islets reportedly had no clinical consequences through 18 years of followup (U.S. Food and Drug Administration 2003).

Overview

This section of the Results chapter supplements the published reports with evidence from recent meetings and discusses more robust data on key outcomes such as insulin independence at 1 and 2 or more years of followup, glycemic control, and adverse events. Where abstracts from the same patient series were presented at different meetings, only the abstract with the most recent and/or detailed information was selected. Where different outcomes for the same group of patients were reported in different abstracts, both were selected and summarized. Abstracts that duplicated information available from published papers were excluded. Selected abstracts from the four most recent meetings are summarized in Evidence Tables 5 and 6. These meetings are:

• American Transplant Congress, 2003: The Fourth Joint American Transplant Meeting, May 30, 2003 - June 4, 2003, Washington, DC

• 63rd Scientific Sessions of the American Diabetes Association, June 13–17, 2003, New Orleans, LA

• 9th Congress of the International Pancreas and Islet Transplant Association, July 8–11, 2003, Dublin, Ireland

• 1st Islet Transplant Congress, November 13–16, 2003, Miami, FL

Because summary data from the Collaborative Islet Transplant Registry is not yet available, a summary of results from transplant groups attending the 2002 Annenberg Second Annual Symposium, in Rancho Mirage, CA represents the only available effort to date to collate islet transplant data from active centers (Evidence Table 7).

Meeting Abstracts

Islet transplantation alone. The abstracts in Evidence Tables 5 and 6 reported on 134 patients entering an islet transplant-alone protocol, including five reported by Berney, Bucher, Mathe, and co-workers (2003a; 2003b) and five reported by Cagliero (2003), but not including the 32 patients from the Immune Tolerance Network (ITN) trial reported by Shapiro, Hering, Ricordi, and colleagues (2003) or the 75 patients combined from 3 centers and reported by Shapiro (2003b) that are most likely included in some individual centers' reports. Not all abstracts reported the number of patients that completed the transplant protocol. Patients were selected primarily for severe hypoglycemia unawareness.

Of four centers that have followed 104 patients for 12 months or more, insulin independence at the time of the report and with variable followup times ranges from 50 to 90 percent (Shapiro, 2003; Shapiro, Hering, Ricordi, et al., 2003; Alejandro, Ferreira, Froud, et al., 2003; Alejandro, 2003; Hering, Kandaswamy, Parkey, et al., 2003; Maffi, Bertuzzi, De Taddeo, et al., 2003). Results beyond 1 year have been reported by the Edmonton transplant center: Shapiro, Lakey, Paty, and co-workers (2003) report Kaplan-Meier projections of 84 percent insulin independence at 1 year (30 patients with 1 or more years of followup), and 64 percent at 2 years (15 patients with 2 or more years of follow-up) in patients who completed the transplant protocol. Of particular note, Hering, Kandaswamy, Parkey, and colleagues (submitted for ATC 2004) reported that of 18 single-donor transplant recipients, 16 were initially insulin independent and 11 remained so at more than 1 year of follow-up. Other centers report on small numbers of patients with less than 1 year of follow-up. In general, for patients who did not achieve or retain insulin independence, centers reported decreases in pretransplant insulin doses of 25 to 75 percent.

Evidence Table 5. Clinical islet transplantation: Patient and transplant characteristics reported in meeting abstracts from the American Transplant Congress (ATC), June, 2003; the 63rd Scientific Sessions of the American Diabetes Association (ADA), June, 2003; and the 9th Congress of the International Pancreas and Islet Transplant Association (IPITA), July, 2003 and the 1st Islet Transplant Congress, November, 2003.

Evidence table 6. Clinical islet transplantation: Outcomes reported in meeting abstracts from the American Transplant Congress (ATC), June, 2003; the 63rd Scientific Sessions of the American Diabetes Association (ADA), June, 2003; and the 9th Congress of the International Pancreas and Islet Transplant Association (IPITA), July, 2003 and the 1st Islet Transplant Congress, November, 2003.

Other notable results include 79 percent insulin independence at 2 years for patients receiving cultured islets at the Edmonton center, and 2 of 3 patients remaining insulin independent at greater than 4 years post-transplant (Shapiro, 2003). Ninety-nine percent of 75 patients pooled from 3 transplant centers (Universities of Alberta [Edmonton], Minnesota, and Miami) demonstrated primary islet function, 96 percent were C-peptide positive at 1 year, and 85 percent insulin independent at 1 year (Shapiro, 2003). For 32 patients entered into the ITN trial, Shapiro, Ricordi, Hering, and colleagues (2003) noted that results varied by center, with 90 percent insulin independent at three centers with long-standing experience, 67 percent at a fourth center, but a much lower average across remaining centers.

Where reported, insulin-independent patients experienced no hypoglycemic reactions (Alejandro, Ferreira, Froud, et al., 2003; Goss, Brunicardi, Feliciano, et al., 2003; Zavala, Hanaway, Peddi, et al., 2003). HbA_1c was reported primarily for insulin-independent patients, in whom mean levels decreased to well under 7 percent and, in most cases, less than 6.5 percent. In two series (Shapiro, 2003; Hering, Kandaswamy, Parkey, et al., submitted for ATC 2004), HbA_1c was maintained at or below 6.5 percent for up to 3 years post-transplant (total n = 13).

Adverse events (islet transplantation alone). Evidence Table 6 summarizes data on adverse events reported at recent meetings by three multicenter groups (the ITN trial and the Edmonton/University of Minnesota/University of Miami and Baylor/University of Miami collaborations) and six single institutions (Universities of Miami, Minnesota, Cincinnati, Alberta, and Emory, and the San Raffaele Institute). These abstracts report on more than 124 patients.⁴ None reported CMV infection or PTLD in any patients given islet-alone transplants. A recent summary presented to the FDA's Biologic Response Modifiers Advisory Committee confirmed that neither adverse event has been reported after islet transplant (U.S. Food and Drug Administration, 2003).

Serious adverse events ranged from none (Goss, Brunicardi, Feliciano, et al., 2003, IPITA #011) to 15 of 32 patients (Shapiro, Hering, Ricordi, et al., 2003; ATC #3, Multicenter ITN trial). Frequent complications included hypercholesterolemia or other dyslipidemia, hemorrhage, and neutropenia and/or leukopenia (n = 121). Occasional complications included portal thrombosis, mouth ulcerations, mild deterioration of renal function, and acneiform rash. Hemobilia, severe tubulopathy, acute pyelonephritis, and interstitial pneumonitis each occurred in one patient.

Islet and kidney transplantation. A few transplant centers report on a total of 30 kidney transplant patients who received an islet transplant. Lehmann, Weber, Zuellig, and co-workers (2003) performed eight islet/kidney transplants; six patients have completed the islet transplant protocol and five achieved and remain insulin independent after a mean of 15 months. Berney, Bucher, Mathe, and colleagues (2000a; 2003b) reported on 2 simultaneous islet/kidney, and 3 islet after kidney transplant patients, but did not report results separately from islet-alone transplants. C-peptide positivity was achieved in all patients, and insulin independence in some over a short followup time.

The GRAGIL 1B trial used a different glucocorticoid-free immunosuppressive regimen and shipped islet preparations to different transplant centers, but was less successful. In this trial, Berney, Bucher, Kessler, and co-workers (2003) report achieving insulin independence initially in six of seven patients with prior kidney transplants tapered off glucocorticoids, but after 3–12 months' followup only one remained insulin independent. Froud, Ferreira, Hafiz, and co-workers (2003) report three of three islet after kidney transplants achieved and remained insulin independent, one after a single islet infusion. Cagliero (2003) reported on five islet-after-kidney transplants, followed for 3 to 23 months; three are insulin independent and two require 25 to 30 percent of their pretransplant insulin doses while awaiting a third transplant.

No centers reported on hypoglycemic reactions. Where reported, HbA_1c levels were normal for most patients, even if some insulin was needed to maintain good glycemic control.

Two deaths were reported, both in patients given simultaneous islet/kidney transplants. One died from a reaction to the OKT3 antibody used for immunosuppression (Berney, Bucher, Mathe, et al., 2003a, 2003b; ATC #401/IPITA #014). The other died from severe pneumonopathy; it is uncertain whether an investigational drug included in this patient's immunosuppression regimen contributed to the adverse outcome (Berney, Bucher, Kessler, et al., 2003; IPITA #013, the GRAGIL 1B trial). Acute rejection episodes occurred in three patients given simultaneous islet/kidney transplant, but rejection of the renal allograft reversed when immunosuppression was modified.

Annenberg 2002 Data Summary

At the “Islet Transplantation 2002 and Beyond: 2nd Annual Symposium,” December 5–7, 2002, at the Annenberg Center for Health Sciences, Rancho Mirage, CA, a number of islet transplant centers pooled their data for a brief summary, shown in Evidence Table 7. At that time, 14 centers reporting had performed 263 islet infusions on 155 patients; 113 of these patients had completed the centers' transplant protocols. At 3 months, after one infusion, 32 patients were insulin independent and after two infusions 65 were insulin independent. This summary does not supply sufficient information to allow the calculation of insulin independence or euglycemia percentages, but reflects the experience of several transplant centers that are using a variety of protocols.

At the 1st Islet Transplant Congress, November 13–16, 2003, in Miami, FL, it was reported that over 75 centers worldwide have initiated transplant programs, and that over 300 patients have received islet transplants since 1999. However, no overall summary of outcomes for all these patients has as yet been reported.

Evidence table 7. 2nd Annenberg Islet Symposium: Transplant Center Results as of December, 2002.