• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

AHCPR Health Technology Assessments. Rockville (MD): Agency for Health Care Policy and Research (US); 1990-1999.

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of AHCPR Health Technology Assessments

AHCPR Health Technology Assessments.

Show details

Simultaneous Pancreas-Kidney and Sequential Pancreas-After-Kidney Transplantation

Health Technology Assessment, Number 4

, M.D., FACP.

Created: .

Foreword

The Office of Health Technology Assessment (OHTA) evaluates the risk, benefits, and clinical effectiveness of new or unestablished medical technologies. In most instances, assessments address technologies that are being reviewed for purposes of coverage by federally funded health programs.

OHTA's assessment process includes a comprehensive review of the medical literature and emphasizes broad and open participation from within and outside the Federal Government. A range of expert advice is obtained by widely publicizing the plans for conducting the assessment through publication of an announcement in the Federal Register and solicitation of input from Federal agencies, medical specialty societies, insurers, and manufacturers. The involvement of these experts helps ensure inclusion of the experienced and varyinng viewpoints needed to round out the data derived from individual scientific studies in the medical literature.

OHTA analyzed and synthesized data and formation received from experts and the scientific literature. The results are reported in this assessment. Each assessment represents a detailed analysis of the risks, clinical effectiveness, and uses of new or unestablished medical technologies. If an assessment has been prepared to form the basis for a coverage decision by a federally financed health care program, it serves as the Public Health Service's recommendation to that program and is disseminated widely.

OHTA is one component of the Agency for Health Care Policy and Research (AHCPR), Public Health Service, Department of Health and Human Services.

  • Thomas V. Holohan, M.D., FACP
  • Director
  • Office of Health Technology Assessment
  • Clifton R. Gaus, Sc.D.
  • Administrator
  • Agency for Health Care Policy and Research
  • Questions regarding this assessment should be directed to:
  • Office of Health Technology Assessment
  • AHCPR
  • Willco Building, Suite 309
  • 6000 Executive Boulevard
  • Rockville, MD 20852
  • Telephone: (301) 595-4023

Introduction

Approximately 35 to 45 percent of insulin-dependent (type I) diabetics develop some degree of secondary kidney malfunction, and this condition may progress to frank end-stage renal disease (ESRD) (1) In the United States, kidney transplantation is a well established and generally effective treatment for ESRD that provides an attractive alternative to dialysis. Because a successful pancreas transplant can reduce or eliminate the need for insulin and diet therapy, the provision of simultaneous or sequential combined pancreas-kidney transplant has been advocated as a technique that may provide effective therapy for both ESRD and insulin deficiency. Although the necessity for administration of immunosuppressive drugs represents a significant disadvantage of pancreas transplant alone (PTA) for diabetes, the kidney transplant patient is already committed to immunosuppression to prevent rejection, and proponents of combined transplant believe this at least theoretically provides some additional rationale for those procedures. The evolution of surgical techniques has mitigated some of the problems associated with pancreatic exocrine drainage as well as vascular and infectious complications which are formidable adverse effects. Most combined procedures are performed as simultaneous pancreas-kidney transplants (SPK); sequential procedures are performed less frequently, and the large majority of those are pancreas-after-kidney transplants (PAK). Pancreatic transplantation poses additional risks over and above those of kidney transplantation; surgical complications are considerably more frequent, and at least within the first year after transplant there is a greater incidence of rejection episodes and a requirement for more intense immunosuppression than in kidney transplantation. The increased immunosuppression appears to engender a commensurately higher incidence of infection.

Advocates of combined pancreas-kidney transplant have asserted that the procedure will reduce, prevent, or mitigate secondary complications of diabetes, preclude complications due to elevations of blood glucose and, by freeing the patient from insulin administration and dietary restrictions, improve the quality of life (QOL) of transplant recipients. These potential benefits have understandable appeal to diabetics with ESRD and have resulted in increasing interest in the combined transplant procedures. Nevertheless, the appropriate recipient selection criteria and the comparative risk-benefit ratios of SPK transplant, pancreas-after-kidney (PAK) transplant, cadaver-kidney transplant (CAD/KT) alone, and living-related donor (LRD) kidney transplant have not been explicitly clarified. The overall clinical effectiveness of combined transplant and issues relating to its cost effectiveness, as well as determination of the most prudent transplant strategy for individual patients, has been controversial.

This assessment was prepared by the Office of Health Technology Assessment, Agency for Health Care Policy and Research (OHTA/AHCPR), for the Public Health Service at the request of the Health Care Financing Administration (HCFA). It sought an evaluation of the safety, efficacy, and effectiveness of SPK and PAK for patients with insulin-dependent diabetes and ESRD; PTA was not addressed in the request. Evaluation of cost effectiveness of health technologies is now required of OHTA assessments by Public Law 102-410 (Title IX of the Public Health Service Act) of October 1992.

Published articles in the English language addressing SPK and PAK transplantation were retrieved through a MEDLINE search of the National Library of Medicine and an additional literature search using the Paper Chase online information service. Medical subject headings used were "pancreas transplantation," "renal transplantation," and "human." Both searches were repeated coupling the additional headings "cost effectiveness," and "quality of life." A bibliography of 3,206 articles addressing cost effectiveness that were published between 1979 and 1990 was also reviewed.(2). The bibliographies of selected articles were examined so as to recover publications not otherwise captured. Additionally, many articles were forwarded or suggested by respondents to the notice of assessment published in the Federal Register. As part of the assessment process, letters were sent to all transplant centers identified by the United Network for Organ Sharing (UNOS) as performing, or having expressed an interest in, pancreas transplantation. These letters (see Appendix A) provided the centers an opportunity to participate in the assessment process and invited the submission of additional information. Data were also solicited from other agencies of the Public Health Service.

Material selected from this compilation of information and used in the prepartion of the assessment included reports of primary objective data addressing results of SPK and PAK transplants. For the most part, these consisted of case series or summaries of institutional experience with SPK and PAK, transplant registry reports, evaluations of secondary complications of diabetes and their relationship to SPK/PAK, and measures of QOL associated with these procedures. Review articles were not excluded from evaluation; although not sources of primary data, in a few instances they provided evidence of the appraisal of the utility of SPK and PAK by segments of the medical community and elucidated the strength of the evidence base and the relationship between the quality of evidence and resulting conclusions. Editorials, letters to editors, opinions, and case reports were not evaluated for the assessment. Other published material not considered included reports that attended to observations not directly related to the safety, effectiveness, cost effectiveness, and clinical utility of SPK and PAK; for example, data addressing parenteral nutrition in support of SPK/PAK, prophylaxis of posttransplant infection, metabolic and biochemical function after transplant, various methodologies for detecting rejection, the use of imaging techniques to evaluate graft function, histopathogic study of organ tissue samples after transplant, and detailed description of various surgical techniques, etc., were not considered. Case series of SPK and PAK published subsequent to 1989 were emphasized. This was based on the belief of the transplant community that results of SPK and PAK have improved over time and contemporary expectations of patient and graft survival should be based on recent data. In addition, the surgical procedure itself has been changing (e.g., pancreatic duct occlusion and enteric pancreatic drainage procedures have been replaced for the most part in the United States by bladder drainage), as have immunosuppressive regimens, with resultant alterations in morbidity and graft survival. A limited review of older publications indicated that such material was unlikely to alter conclusions derived from more contemporary data.

Draft copies of the assessment were reviewed by staff of OHTA/AHCPR, and Public Health Service officials selected by the National Institutes of Health (NIH) and the Food and Drug Administration (FDA), as well as nongovernmental consultants to OHTA.

Review of Available Information

Clinical Series

Wright et al(3). reported a series of 70 pancreas recipients (47 SPK and 23 PAK) who received transplants between 1984 and 1988. The stated purpose of the transplant was to reverse or slow the "progression of secondary complications of diabetes." The surgical procedure was changed during the observation period in that bladder drainage supplanted the enteric drainage technique. In addition, the immunosuppression regimen varied, as the authors reported that azathioprine, cyclosporine, steroids, and antilymphocyte globulin (ALG) were used "in most cases."

Although criteria for selecting patients for SPK or PAK were not stated, the authors indicated that the recipients' average pretransplant glycohemoglobin (HbA1C) level was 7.9 percent. One-half of the group suffered late complications and reoperations. Patient survival was reported as 82 percent and 75 percent at 1 and 4 years for SPK, and 91 percent at 1 and 4 years for PAK recipients. Pancreas graft survival for SPK at 1-4 years was reported as 59, 57, 50, and 45 percent, respectively. Corresponding figures for PAK were 41, 35, 35, and 35 percent. Although the authors stated that the studies had "yet to completely answer the question of what are the beneficial effects" of SPK and PAK, they concluded that the procedures should be offered to all "suitable" patients.(3)

Sutherland et al(4). described their institution's experience with several types of pancreas transplants. The report dealt with 276 transplants (in 222 recipients) dating from 1978. However, all transplant procedures were not equally represented across all periods of observation. The data were stratified into three time periods, viz: 1978-1984; 1984-1987; and 1987-1989 (Table 1). It is unclear when SPKs were begun, as the report included statements citing both 1985 and 1986. Surgical techniques for managing exocrine secretions evolved over time and included duct occlusion (44 cases), peritoneal drainage (44 cases), enteric drainage (89 cases), and bladder drainage (128 cases). Triple immunosuppressive therapy with azathioprine, cyclosporine, and prednisone (A-C-P) was initiated in 1983. This retrospective review demonstrated that odds ratios for success favored the bladder drainage technique, HLA-DR matching of donor and recipient, and that the specific procedures were themselves associated with differing likelihoods of success: specifically, success rate of SPK > PAK same donor > PAK different donor > pancreas transplant alone (PTA). One-year actuarial patient survival rates ranged from 87 percent (SPK) to 97 percent (PTA). Graft survival rates improved over time, as noted in Table 1. Of the 46 SPK recipients, 39 were primary cases (i.e., not retransplantation), and for those recipients, graft survival at 1 year was 74 percent (29/39). Graft survival in the seven retransplant cases was not given but could not have been greater than 28 percent (i. e., two of seven), because the overall 1-year graft survival for SPK patients was 31/46, or 67 percent. For PAK transplants, the authors reported differences in primary and retransplant graft survival rates in cases treated with the bladder drainage technique. One-year actuarial pancreas graft survival was reported as 56 percent for primary transplant (20 cases) and 48 percent for all patients (32 cases). From these data, the actuarial 1-year graft survival in retransplants may be calculated as 33 percent (12 cases). Within the entire group of 275 transplant procedures there were 37 second transplants, 15 third transplants, and two fourth transplants, for an overall retransplant proportion of 20 percent.

Table 1. Percent of 1-year graft survival [a].

Table

Table 1. Percent of 1-year graft survival [a].

No specific patient selection criteria were reported, nor were details provided regarding pretransplant diabetic treatment, indices of the effectiveness of glucose control, nor measures of the severity of the diabetes.

Sutherland et al(5). later described a retrospective multivariate analysis of factors associated with improved outcome in pancreatic transplants (SPK, PAK, and PTA) performed during the period from 1978-1991. They noted that no changes in transplant protocols or in immunosuppressive regimens were tested in randomized trials.

During the years 1985-1991, 250 procedures (194 primary and 56 retransplant) were performed using the bladder drainage technique. The authors reported the numbers of primary vs. retransplants by specific procedure: 104/12 for SPK (10.3 percent retransplant); 35/21 for PAK (37.5 percent retransplant); and 54/23 for PTA (29.8 percent retransplant). (This includes only 249 of the 250 cases.) Overall, 22.4 percent of all transplant procedures were retransplantations.

The multivate analysis revealed that for all transplants accomplished by the bladder drainage technique, the relative risk of pancreas graft failure was higher for retransplantation (vs. primary procedure), recipient age >45 years, organ preservation >30 hours, and HLA-A, B, DR mismatch of 2-3 vs. 0-1. HLA matching was also related to the recipients' need for exogenous insulin: insulin independence was noted for 80 percent of the 0-1 mismatch, 56 percent of the 2-3 mismatch, and 44 percent of the 4-6 mismatch groups. The authors noted that they believed the correction of uremia to be more important than the correction of hyperglycemia, and that end was best served by renal transplant from a LRD.

Their conclusions as to the benefit of pancreas transplant were complex, in that they stated "...there is little information on the effect of pancreas transplantation preventing secondary complications [of diabetes]," but went on to cite other references that claimed that with pancreas transplant "the recurrence of nephropathy is prevented" and that there was "stabilization or improvement in neurophysiological parameters." No primary data were presented relating to nephropathy or neuropathy. The authors pointed out that the main value of such a retrospective study was helping to plan prospective studies.

Olausson et al(6). reported results in a series of segmental pancreatic transplants comprising 50 SPK and 17 PAK performed using the bladder drainage technique between 1985 and 1990. No information was provided as to HLA-matching. Patient selection criteria, including severity of and treatment regimen prescribed for diabetes, were not described. Patient survival for all patients was 91 percent at 1 year and 89 percent at 2 years; for SPK alone, comparable figures were 89 percent and 87 percent. Actuarial 1-year graft survival for SPK patients was 76 percent for both pancreas and kidney and 83 percent for pancreas function alone.

Garvin et al(7). described a series of 100 patients who were evaluated for transplant, 44 of whom were provided SPK between 1985 and 1988. Thirty-eight patients had ESRD due to diabetes and were on dialysis at the time of transplant. All 44 patients had neuropthy or retinopathy, and five had "overt cardiac disease." Patient selection criteria were not described save for use of a pretransplant thallium scan as a screen for coronary artery disease. Six months into the second year of the series, this was altered to a stress thallium test with subsequent coronary angiography used for abnormal thallium tests. Preoperative histocompatibility matching included what the authors described as "minor emphasis" on HLA-A, B, and DR matching. The surgical procedure changed during the series, as did the prophylactic use of heparin. In the second year of the series, triple immunosuppression was used; acute rejection was treated with local radiation therapy and prednisone. In this series, 13.6 percent of the recipients died within 8 weeks of transplant (from acute myocardial infarction, sepsis, an bleeding).

Seventeen patients were insulin-independent from 4-50 months posttransplant. The last 25 cases of this series were reported to have had a 60 percent pancreas graft survival at 6 months. The authors stated that "the major purpose of SPK is to stabilize or reverse the secondary complications of diabetes." However, they noted that the "exact role [of pancreas transplant] remains to be determined.

Goldman et al(8). reported on six SPK and five PAK patients. No dates of accumulation of the series were given. Patient selection criteria were not specified, but those with angina and congestive heart failure, claudication, amputation, or ischemic foot ulcer were excluded. Pretransplant HbA1C levels were reported for 10 of the patients, with a mean value reported as 11. 6 percent, and the average pretransplant daily insulin dose was 42 units (range of 25-64 units). The posttransplant immunosuppression regimen varied during the course of the series. No meaningful graft survival rates could be provided, since 7 of the 11 patients had been followed for less than 1 year.

Stratta et al(9). retrospectively reviewed 38 SPK performed at the University of Nebraska between 1989 and 1991. Changes in transplant procedure during the period of study included the adoption of the duodenal segment with bladder drainage technique, use of triple immunosuppressive therapy with OKT3 induction, and the use of Minnesota antilymphocyte globulin (MALG) for steroid-resistant rejection. Details regarding patient selection criteria were not provided save for the comments that exclusionary criteria included "prohibitive cardiovascular risk, active infection, recent malignancy, history of noncompliance and identification of type II diabetes." Of the 38 cases, 15 (39 percent) were not on dialysis; the daily pretransplant insulin dose required ranged from 15-67 units daily, and the average HbA1C level was 10.9 percent. All had retinopathy, 29 percent were blind, and 84 percent had neuropathy. Pretransplant tissue matching included HLA-DR matching tests. With a mean followup of 15 months (4-to 29-month range), the authors reported a 95 percent graft survival; however, only eight of the 38 patients had been followed for 12 months or longer. The authors stated that SPK was "a logical way of protecting both the kidney and the patient from progressive diabetic complications," and that alternative treatments did "not provide sufficiently good metabolic control to prevent the progressive diabetic complications of retinopathy, neuropathy, nephropathy, and accelerated atherosclerosis." However, they also indicated their belief that long-term studies were necessary to determine the effect of pancreatic transplant upon the diabetic condition.

The same authors updated this experience in a report addressing the benefit and risk of SPK.(10) This paper extended the transplant interval by 8 months and the total number of cases to 61. Comparisons were made to nonrandomized, nonmatched groups: 31 diabetics and 31 nondiabetic patients provided renal transplant. The authors noted that the survival of transplanted kidneys was similar across all groups, and reported a 1-year actuarial pancreas graft survival of 93 percent. They again indicated that they believed pancreas transplantation was the only effective method for prevention of secondary diabetic complications, and that pancreas-kidney transplant is "the treatment of choice" for insulin-dependent diabetes and chronic renal failure. In a third paper, Stratta et al elaborated on their patient selection criteria and process, reporting on a total of 205 patients referred to their institution for consideration for pancreas transplantation. Immunologic matching procedures were as described in a previous paper (vide supra). Evaluation included the following diagnostic tests:

  • Ultrasound of gallbladder.
  • Mammography in females aged >35 years.
  • Electrocardiogram (EKG) or thallium stress test.
  • Echocardiography.
  • Coronary angiography "when indicated."
  • Vascular testing (Doppler arterial evaluation, transcutaneous, plethysmography, carotid Doppler).
  • Nerve conduction testing, gastric emptying, electromyography (EMG).

Two hundred and five patients were referred, and 151 were subject to initial screening; 137 of the latter group completed the process. Forty-five patients (33 percent) underwent coronary angiography; five underwent subsequent angioplasty or bypass. The authors did not consider a history of myocardial infarction, angioplasty, or bypass as a contraindication to pancreas transplant, but believed that smoking or obesity were relative contraindications. They stated that "hyperlabile diabetes with significant impairment of quality of life" should be considered as requirements only for PAK and SPK following a previous renal transplant. Ninety patients were accepted for SPK, 70 of whom underwent the procedure; in this group, the mean daily insulin dose was 40 units (range: 12-200 units), and the mean HbA1C level was 9.9 percent. Seven patients were accepted for PAK, with six subsequently transplanted; the mean daily insulin dose of these patients was 50 units (range: 30-75 units), and their mean HbA1C level was 10.4.(11)

These authors also compared 61 SPK recipients (both dialysis-dependent and predialysis) with "concurrent and historic" diabetic kidney transplant alone (KTA) recipients and "concurrent" nondiabetic KTA recipients.(12) All of the patients (n = 123) received cadaver organs. They reported no adverse effect of pancreatic transplant upon renal graft survival; they indicated that although the "morbidity rate" (rejection episodes, reoperations, readmissions, and rehospitalizations) of SPK recipients was higher, the difference was not statistically significant. The authors stated that "improved results" after SPK "may be attributed to selection bias."

Sollinger et al(13). described a series of 100 consecutive SPK transplant using the bladder drainage technique which were accomplished between December 1985 and December 1989. The first 17 patients were provided with a duodenal button, the remainder were treated with a duodenal segment bladder anastomosis. Ninety-two of these cases were tested for HLA-A, B, and DR matching, however, degree of match was not reported. Maintenance immunosuppression consisted of A-C-P. The 1-and 2-year actuarial graft survival was reported as 88 and 86.8 percent, respectively; actuarial patient survival was 94 and 92.8 percent. Pretransplant patient selection evaluation included a stress thallium test for all patients older than 30 years, with coronary angiography performed following positive thallium tests. The authors stated that patients with no or "minimal" coronary artery disease were accepted, but the term "minimal" was not further defined. Contraindications included prior amputation due to vascular disease, blindness, and age greater than 55 years. No data were provided regarding the pretransplant insulin regimen, blood sugar levels, HbA1C, or frequency of hypo-or hyperglycemic (H/HG) episodes. The authors stated that they only considered patients as potential recipients if the pancreas transplant could "stabilize or reverse already existing secondary complications," and that they believed SPK was "effective in influencing secondary diabetic complications. "

Sasaki et al(14). extended Sollinger's prior report(13). to 138 patients provided SPK between 1985 and 1990. The maximum age for SPK was given as less than 50 years (vs. 55 years as cited in Sollinger et al(13). ), although the authors stated that "age per se should not exclude patients." The minimum followup was 1-year, actuarial 5-year pancreas graft survival was reported as 76 percent.

Sollinger et al(15). updated their experience in a report published in 1993. The paper described a "subgroup" of 200 consecutive bladder-drained SPK accomplished between 1982 and 1993. The authors noted that all patients were required too undergo thallium stress testing, with cardiac catheterization performed when results were abnormal. Patients were excluded for "significant coronary artery disease." Other exclusion criteria included severe peripheral vascular disease and patients "who were expected to be unable to comply" with the postoperative regimen. No other patient selection criteria were specified. Actuarial pancreas graft survival and number of cases actually observed for 1-5 years posttransplant are seen in Table 2.

Table 2. Actuarial graft survival and number of cases 1-5 years posttransplant.

Table

Table 2. Actuarial graft survival and number of cases 1-5 years posttransplant.

The number of patients followed for longer than 2 years is still rather limited. For example, approximately 48 percent (96 cases), 27 percent (54), and 12 percent (24) of the 200 cases were actually evaluable at 3, 4, and 5 years, respectively. More than half of all cases appeared to have been followed for less than 3 years.

The authors stated that "others have shown an advantage of SPK transplantation in terms of stabilization or improvement of secondary complications of diabetes," but provided no date to that point. They claimed to have demonstrated that "SPK recipients have a better quality of life," and referenced unpublished data presented at the 1992 American Society of Transplant Surgeons' meeting.

Alexander et al(16). reviewed their experience with 29 SPK and 10 PAK procedures. Six different surgical procedures were employed over the reporting period. The 1-and 5-year pancreas graft survival figures for PAK were reported as 30 percent; corresponding figures for SPK were 52. 7 percent and 41 percent. However, for SPK performed between 1989 and 1991, 1- and 5-year graft survival was 58.2 percent. No details were provided regarding immunosuppression, tissue matching, complication rates, or patient selection criteria used.

Shafter et al(17). described a series of 26 SPK and one PAK performed between 1988 and 1991. The surgical procedure was altered after the first 20 transplants in order to decrease the frequency of wound complications. Maintenance immunotherapy consisted of A-C-P. Patient selection criteria were not specified save for the statement that "significant" CAD and major limb amputation were considered as contraindications. Pretransplant HbA1C levels were reported for only seven of the 27 patients, and averaged 10.8 percent. One-year actuarial pancreas graft survival rates were reported as 85 percent. One of the stated purposes of the transplant procedures was to eliminate the morbidity associated with fluctuations in glucose levels; mean posttransplant fasting blood sugar and HbA1C levels were reported for only 12 of the 27 patients, as 4.8 mmol/L and 6.6 percent, respectively. However, no data were provided that addressed the diminution of morbidity associated with reductions in glucose fluctuation. The authors stated that SPK and PAK could "stabilize or prevent secondary diabetic complications."

Bentley and Garrison(18). described results in eight PAK and 14 SPK performed between 1987 and 1991. No attempt was made to match for HLA, and the authors stated that there appeared to be no disadvantage for recipients of poorly matched grafts. The 1-year pancreas graft survival was 87 percent for SPK and 38 percent for PAK. No specific details were provided regarding patient selection criteria. The authors stated the ultimate goal of the procedures was to "prevent the development" of the secondary complications of diabetes.

Rosenloff et al(19). published a retrospective review of a series of 16 SPK, and one PAK, and three PTA and one "cluster" procedure accomplished during the years 1988-1991. Patient selection criteria were not specified save that all were required to undergo "cardiac evaluation" before transplant. The mean pretransplant fasting glucose for all patients was 209 +/-19 mg percent; posttransplant levels averaged 103 +/-8 mg percent. Preoperative HbA1C levels were reported as 11.2 percent; these decreased to an average of 5.1 percent after transplant. No data were provided regarding the pretransplant insulin regimens or diabetic complications of the recipients. Mean duration of followup was 526 +/-334 days. Pancreas graft survival at 1 year was reported as 86 percent. The authors found no relationship between HLA typing mismatch at any loci and graft survival.

Dawahra et al(20). summarized results of 226 SPK in 211 patients accomplished at the Edouard Herriot Hospital in Lyon, France. Segmental pancreas grafts with duct injection (occlusion) were performed in 166 cases. Of the whole pancreas grafts, enteric drainage was used in 14 cases, and bladder drainage in 46. Taken as a whole, this experience is unlike that reported in the United States, where segmental grafts are rarely employed and the majority of cases utilize bladder drainage. The authors evaluated patient and pancreas graft survival according to "time" (inferred to be various time periods from 1976 to 1992), type of immunosuppression, and surgical technique; the results were presented graphically. Unfortunately, the six graft survival curves were not labeled, and the reader can only determine that the 1-year pancreas graft survival ranged from approximately 10 to 75 percent for the various parameters considered. However, the report noted that the several surgical techniques did not differ as regards patient or graft survival, that segmental pancreas grafts had a lower rate of surgical complication, and that "overall" 1-year patient and graft survival rates were 90 and 70 percent, respectively. With respect to secondary complications of diabetes, the authors commented that it was not possible to determine whether "stabilization" of retinopathy was due to SPK or laser photocoagulation, or whether subjective improvement of neuropathy resulted from functioning pancreas grafts or reversal of uremia.

Elkhammas et al(21). reported results in 150 consecutive SPK accomplished between 1988 and 1992. Patient selection criteria included stress multigated equilibrium radionuclide scan (MUGA) testing, coronary angiography was used for abnormal MUGA results, and bypass or angioplasty was provided to an unspecified number of patients. If "advanced secondary complications" of diabetes or inadequate cardiac reserve were present, patients were offered KTA or transplant was deferred. The proportion of patients so eliminated from consideration for SPK was not specified. The actual numbers of patients followed for each of the annual intervals was not reported. The authors reported actuarial pancreas graft survivals of 85 percent at 1 year and 76 percent at 4 years posttransplant.

In 1992, Van Oosterhout et al(22). reviewed the results of 24 SPK performed between 1986 and 1990 at the Leiden University Hospital. The SPK recipients were part of a larger prospective study to compare the results of SPK with KTA and insulin treatment. Patient selection was based on what was termed an "extensive pretransplantation analysis" completed during a 5-to 10-day hospitalization. Specific testing was not described save that all 24 patients underwent coronary angiography, and interventions for CAD were provided to two patients (one bypass, one angioplasty). The bladder drainage technique was used, and posttransplant immunosuppression was comprised of A-C-P. Np HLA-matching was performed. The authors reported that results in the first four patients, in whom the pancreas had been placed extraperitoneally, were poor. Intraperitoneal pancreas graft placement in the second group of patients provided superior outcomes, with 100 percent patient survival. Renal graft survival was reported as 62 percent both at 1 (n = 11) and 2 years (n = 10), and pancreas graft survival as 65 percent at 1 (n = 11) and 2 years (n = 10). The number of rejection episodes in SPK was "much higher" than had been observed in KTA.

Van Oosterhout et al(22). stated that the effect of normoglycemia following SPK in arresting secondary diabetic complications "has not yet been fully resolved." They believed that transplantation of the pancreas has been associated with no dramatic regressive effect on the course of the secondary complications of diabetes. They noted that retrospective studies of QOL had been performed with varying results, and that any benefit needed to be substantiated in term studies of larger samples prospectively compared with diabetic KTA recipients. Interestingly, their study was designed as a prospective, randomized trial of SPK vs. KTA and insulin therapy; it could not be completed because of the patients' unwillingness to accept randomization, even though the comparative risks and benefits were unknown. The patients preferred to "choose their own treatment modality."

Griffin et al(23). described 10 SPK and seven PAK performed over a 6-year period. Insulin-dependent diabetics with ESRD were not offered SPK/PAK if they had "good glucose control, stable complications, and a high quality of life." These criteria were not further defined. Selection for combined transplant was based on "severe visual impairment or some other disabling condition," "poor glucose control," or "a strong preference for the procedure. " Pretransplant mean daily insulin dose was 55 units (range: 20-114 units). Results were compared with a demographically similar group of diabetics provided only KTA during the same 6-year period. The actuarial pancreas graft survival rate at 1 year was 63 percent, and the patient survival rate was 92 percent. The number of cases followed for 1, 2, and 4 years posttransplant were eight, six, and four for SPK/PAK and 20, 18, and 11, respectively, for KTA. Renal graft survival in SPK was similar to that for KTA. Complications were more frequent in the SPK/PAK patients (vide infra). The authors concluded that the indications for pancreatic transplantation "remain unresolved," and that SPK or PAK transplantation should be restricted to patients "who are particularly unstable or exhibit progressive diabetic complications." However, it should be noted that this report included patients who required as few as 20 units of insulin daily, and patients selected for SPK/PAK at least in part on the basis of their preference for the procedure.

Brattstrom et al(24). described 65 20 (31 percent) functioned for 5 years or longer. Ten percent of cases had "slightly elevated" blood glucose levels; the HbA1C was normal in 73 percent. Five of 22 cases (20 SPK and 2 PTA) had symptomatic neuropathy at 5 years; 77 percent were working or attending school. However, serve cardiovascular disease was seen in one-fourth of pancreas graft recipients.

Many institutions' series of SPK and PAK have been multiply reported as cases accumulated.(25,26) Such additional reports add little to the results summarized above.

Sutherland et al(27). reported that in United States cases the "technical failure" rate of pancreas transplantation was 12 percent. Of the case series summarized above, none separated results based upon surgical success (i. e., technically successful procedures), and it was assumed that all combined transplant cases accomplished within the specified time periods of accumulation of each series were included for analyses. The definition of successful pancreas transplant, pancreas graft survival, or graft function varied slightly. This is important because it is unclear how much benefit, if any, accrues to a simple reduction in insulin dosage. Six reports (6,14,16,20,22,23). did not clearly define graft survival in terms of freedom from exogenous insulin. In Wright's series, (3). an unspecified number of pancreas recipients required oral hypolycemic agents. Many authors specifically stated that pancreas graft survival was associated with insulin independence, (4,7,9,10,15,17,18-20). whereas others were somewhat less precise, stating that "no insulin was required postsurgery" or that "no patient...required insulin in the immediate posttransplant period."(13).

Registry Data

The various case series of SPK and PAK are heterogeneous with respect to sample size, patient and graft survival, surgical procedures used, posttransplant immunosuppression, duration of followup, and patient selection. This circumstance makes it difficult to ascertain which of those factors may affect graft survival, operative morbidity, and ultimately the risks and benefits of the procedure. Such difficulties are compounded by the fact that 73 transplant centers have performed SPK or PAK, but few of these have published their results in the medical literature. Data from formal registries include case results not elsewhere reported and often provide a more reliable perspective of the clinical utility of medical or surgical procedures. Similar circumstances were noted in the OHTA assessment of lung transplantation, wherein the published data available at the time of preparation of the assessment led to the conclusion that 1-year survival of 65 percent should be regularly achievable. The UNOS Registry data which was published approximately 6 months after completion of the assessment indicated that the mean national 1-year survival was in fact only slightly over 50 percent.

Sutherland et al(28). reviewed information collected through the International Pancreas Transplant Registry (IPTR) which was subsequently submitted to UNOS. The data covered the period between October 1987 and November 1992. Pancreas transplant were performed by 73 institutions; the number of cases transplanted in each center during the 5-year period ranged from fewer than 10 to more than 200 (see Figure 1). SPK comprised 83 percent, and PAK 8.6 percent of these procedures; 4 percent of all cases were retransplants. The authors noted that pancreas transplant is much more common in the United States than in foreign countries; approximately 75 percent of all cases worldwide have been performed in the United States. Of the 2,103 pancreas transplants reported to to UNOS, several categories were excluded from this analysis. Pancreas-liver, pancreas-kidney-liver, pancreas-heart, and pancreas-heart-kidney transplants were not further analyzed. Also excluded were living-related donor (LRD) pancreas transplants, those that used enteric or ureter drainage of exocrine secretion, and transplants that used duct injection techniques. No followup data were available for 84 cases; the remaining 1,879 cases comprised 1,604 SPK, 166 PAK, and 109 PTA. Of these, 1,320 (70 percent) were listed as having functioning pancreatic grafts. However, only 690 cases had the potential for more than 1-year followup, and 124 of those did not have complete information available. Therefore, only 566, or 43 percent of all patients with functioning grafts, had actually been followed for more than 1 year. For the 1,604 SPK and 166 PAK patients, 1-, 2-, and 3-year actuarial patient and graft survival rates are shown in Table 3.

Figure 1. Pancreas transplants: UNOS data, 1987-1992.

Figure

Figure 1. Pancreas transplants: UNOS data, 1987-1992.

Table 3. 1993 IPTR report to UNOS [a].

Table

Table 3. 1993 IPTR report to UNOS [a].

Note: Updated unpublished information from the Registry indicated that the number of cases followed for 1-3 years were 1,048,682, and 357, respectively. Actuarial graft survival figures were similar to 1993 data (personal communication, Sutherland DE, November 23, 1993).

Only 12 of the SPK cases were retransplants; graft survival rates were 76 percent for primary cases and 62 percent for retransplants (P = not significant). For PAK, 27 percent were retransplants; graft survival was 51 percent in primary and 39 percent in retransplants (P =. 04).

Comparisons of 0-1 vs. 2-6 HLA-A, B, and DR loci mismatches revealed no effect for SPK (74 percent vs. 75 percent 1-year graft survival). For PAK, 1-year graft survival was 80 percent in the 0-1 mismatch cases and 44 percent in the cases with 2-6 mismatches (P = .03). However, it must be noted that only 16 SPK patients (1 percent), and 11 PAK patients (7 percent) had 0-1 antigen mismatches; even slight alterations in graft survival would have the potential to alter those statistics substantially.

Rate of graft loss was determined to be higher for patients older than 45 years. However, a multivariate analysis revealed that this was significant only for SPK.

Of note was the observation that of all diabetic kidney transplant recipients, 23 percent also received simultaneous pancreas grafts; this was true of more than half the cases transplanted in the high-volume pancreas transplant centers.

Although a total of 2,100 procedures were reported to UNOS, only 50 transplant cases during the 5-year period had graft function for as long as 4 years. Thus, long-term followup is still limited.

Transplant centers were separated into high-volume (20 or more cases) and low-volume (19 or fewer) categories. Both patient and graft survival were significantly higher for SPK in high-volume centers. Patient survival was 91 percent, and graft survival 78 and 67 percent, in the-and low-volume centers, respectively. Comparable analyses could not be done for PAK, because only one center qualified as high volume.

The relationship between transplant center census and outcome was reviewed by Hunsicker et al(29). based upon the UNOS data set available in 1993. Although the authors stated that much of the intercenter variability "remains unexplained" and may be due to "chance alone," they indicated that for all organ transplants other than kidney there was a clear center size effect. This was on the order of a factor of a 2-to 3.5-fold increase in risk of graft failure from the smallest centers to the "reference" fourth quintile.

Renal Graft Survival and Combined Transplants

The effect of pancreas transplantation on kidney allograft survival has been the subject of several reports. Sasaki et al(14). compared renal graft function in SPK with LRD kidney transplant, and reported superior, but not statistically significant, graft survival for LRD to 4 years. Schulak et al(26). retrospectively compared 32 SPK with 28 KTA. Three KTA were from LRD and four received six HLA antigen-matched organs. Selection for SPK was in part related to perceived cardiovascular After echocardiography, stress thallium scan, and coronary angiography in an unspecified proportion of cases, "high-risk" patients were not offered transplantation, and "moderate-risk" patients were offered only KTA. The authors stated that the kidney graft survival in SPK patients (86 and 77 percent at 1 and 4 years, respectively) was not significantly different from that in the cadaver KTA; however, all three LRD KTA had "excellent graft function." Rejection episodes were more frequent in the SPK than in the KTA group (75 percent vs. 54 percent); they also tended to be more severe; only 22 percent of the KTA rejection episodes were steroid-resistant, whereas that was true of 47 percent of episodes in SPK patients. The authors concluded that kidney survival in SPK patients is not inferior to that in KTA recipient, although no further comment was made regarding the superior graft survival associated with LRD.

Greussner et al(30). reviewed data on 39 SPK, 48 KTA, 10 PAK, and 31 PTA recipients. They noted that the 1-year renal allograft survival was 81 percent in SPK and 84 percent in KTA recipients (p < .01). If "technical failures" were excluded, renal allograft survival was similar (86 percent in SPK and 87 percent in KTA recipients). Tufveson et al(31). reviewed 1986 European data on renal allograft survival in 90 SPK recipients. Survival data were compared with those of group of 389 patients, comprising all diabetic patients listed in the European Dialysis and Transplant Association (EDTA) Registry who received KTA in Belgium, Germany, Denmark, Finland, France, Switzerland, and the United Kingdom. Patient survival was virtually equivalent in the two groups (89 percent for SPK and 90 percent for KTA recipients). Renal allograft survival at 1 year was 78 percent for SPK and 76 percent for KTA recipients. Pancreas graft survival at 1 year was recorded as 65 percent.

The United States Renal Data System (USRDS) 1992 Annual Date Report(32). described diabetic ESRD patients who received their first kidney transplant (SPK or KTA) from cadaveric donors between 1986 and 1989. A total of 3,158 patients were reported; 380 (12 percent) had received SPK. No significant differences in patient survival were noted, but kidney graft survival was significant higher for SPK than for KTA recipients for up to 28 months posttransplant, when the graft survival curves were observed to cross. The report stated that "none of the SPK kidney grafts remained beyond 29 months posttransplant."

Stratta et al(12). reviewed renal graft survival in a group of 61 SPK recipients who were compared with two groups of KTA recipients. The first was a nonrandomized "age-matched" group of 31 nondiabetic KTA patients. The authors noted that "age matching consisted of only selecting recipients between the ages of 19 and 55 years." The second comparison group consisted of 31 "concurrent and historical" diabetic patients who received KTA. Renal allograft survival was higher in the SPK group; however, the average duration of followup for SPK recipients was 16 months, whereas for diabetics provided KTA, followup averaged 24 months.

Cecka et al(33). reviewed data reported to the UNOS Renal Transplant Registry from 1987-1991. They noted that there were no significant differences in 1-year kidney graft survival rates based on the etiology of the ESRD. Of the 995 diabetics in the UNOS Registry who received a pancreas transplant in addition to their kidney, kidney graft survival was statistically superior at 1 year (81 vs. 77 percent). However, the data also indicated that a LRD kidney provided a very significant increase in probability of graft function over time. The "half-life" of 20,864 primary (i.e., not retransplant) cadever kidney transplants was 7.2 years. Comparative half-lives of LRD kidneys from parents (1,487 cases), one HLA-haplotype-matched siblings (1,206 cases), and HLA-identical siblings (970 cases) were 10.8, 12.2, and 26.9 years, respectively. The use of non-6-antigen-matched LRD kidneys therefore provided potential increases in expected transplant half-life ranging from 50-70 percent, and HLA-identical transplants were associated with renal half-life increases of more than 360 percent.

Petronis et at(34). compared results of SPK and KTA from information contained in the USRDS and the IPTR covering the years 1988-1989. A total of 1,441 kidney transplant patients, 340 (24 percent) of whom received an SPK, ere evaluated. Patient mortality after transplantation was very similar between the two transplant types. In the overall analysis, SPK transplants had a lower graft failure rate than did KTA transplants when no adjustments were made for another factors. However, when the data were adjusted for the effects of factors such as transplant center, time from ESRD to transplant, donor race, donor/recipient blood type, and use of pulsatile perfusion, the differences in kidney graft survival were not statistically significant. The authors concluded that SPK is not associated with a decrease or increase in kidney or patient survival in comparison with KTA. They noted that there is need for "examinations of the posttransplant endocrinologic and quality of life benefits of SFK transplantation. "

The data cited above indicate that overall kidney allograft survival in SFK does not differ significantly from that if cadaver-donor KTA (CAD/ KT). Notwithstanding, there is convincing evidence that potentially significant improvements in kidney allograft survival may result from LRD and HLA matching in KTA, and could provide benefits exceeding those from cadaver SPK.

Moreover, prospective HLA matching might improve pancreas graft function as well. Few of the case series of SPK and PAK address the issue of HLA matching. The information from those series and from reviews of collected series is retrospective. Bentley et al(18). believed that there was no disadvantage to recipients of poorly matched grafts in their small series of 14 SPK and eight PAK. However, Sutherland et al(5). reported that greater degrees of HLA match were associated with both lower relative risk of organ failure and a lower likehood of need for exogenous insulin (vide supra). They concluded that the correction of uremia was the most important intervention and that was most effectively accomplished by LRD renal transplant. Sutherland et al(35). in their review of the UNOS Pancreas Transplant Registry data base noted that HLA mismatch for PAK was associated with a significant risk of graft failure; for SPK the relative risk for 2-6 vs. 0-1 mismatches was 1.3 but did not reach statistical significance. Sasaki et al(14). reported that retrospective analyses of patient and graft survival revealed that both were significantly improved for HLA-DR 0-or 1-antigen mismatch vs. 2-antigen mismatch.

Although the proper role of pretransplant HLA matching in SPK/PAK has not been determined, its utility in improving renal graft survival has been established. Takemoto et al(36). reviewed data on allograft survival in 1,386 recipients of HLA-matched kidneys and performed two multivariate analyses; the first was intended to determine the effect of HLA match when other risk factors were considered (n = 1,340 patients) and the second to evaluate risk factors among the recipients (n = 1,223). Comparison groups were 22,188 primary transplants and 3,950 retransplants whose HLA antigens differed from those of the donor. Statistically significant improvements in 1-year allograft survival for HLA-matched cases were noted for the categories of: first transplants (88 vs. 79 percent), second transplant (78 vs. 72 percent), and "highly sensitized" recipients (86 vs. 74 percent). No advantage in graft survival was noted for HLA matching in third or subsequent transplants. In addition, paired kidneys from 470 donor were transplanted into HLA-matched and mismatched recipients. The 1-year renal graft survival rates of the matched and mismatched contralateral kidneys were 87 and 80 percent, respectively (P < .05). The estimated half-life of graft survival was 7.8 years for recipients of mismatched kidneys and 17.3 years for recipients of primary transplant matched kidneys.

Petronis et al(34). concluded that SPK is not significantly different from KTA with respect to a decrease or increase in kidney or patient survival up to a period of 2 years posttransplant. Of note were the data comparing total numbers of HLA mismatches for SPK and for KTA. For KTA, 0-1 mismatches comprised 12.4 percent of all transplants, whereas for SPK only 1.5 percent of transplants were 0-1 mismatches. Conversely, 68.2 percent of SPK, but only 52 percent of KTA, were 4-6 antigen mismatches. Data from Cecka et al(33). and Takemoto et al(36). suggest that equivalent kidney graft survival in SPK and KTA might not have been maintained if the period of followup had been extended to longer than 2 years.

Therefore, although it appears that SPK is not associated with renal allograft survival, which is inferior to the general population of KTA, this might not be the most appropriate comparison. If it is believed in a given patient that the greatest threat to health and life is ESRD, then the most efficacious strategy is that which will primarily increase the functional life of a transplanted kidney, irrespective of pancreatic function.

A retrospective review by Cheung et al(37). provided data to that point. The authors evaluated 1-, 2-, and 3-year patient and kidney survival in 435 transplant recipients. The group comprised 173 LRD kidney transplant (LRD/KTs), 148 cadaver kidney transplants (CAD. KTs), 97 SPK, and LRD/KTs followed by PAK (LRD/PAK). LRD/KTs were superior to SPK and to CAD/KTs for both patient and renal allograft survival; no differences were noted between LRD/KTs and LRD/PAK. Moreover, data from the UNOS Pancreas Transplant Registry(28). indicated that 3-year patient survival in 1,604 SPK was quite similar to that reported by Cheung et al(37). (81 vs. 79 percent). With regard to renal allograft survival at 3 years after SPK, the UNOS figures superior to those reported by Cheung et all(37). (73 vs. 63 percent), but were still inferior to LRD/KT (89 percent) and LRD/PAK (94 percent). Not surprisingly, they were quite close to the renal survival after CAD/ KT (69 percent).

SPK might provide renal allograft survival equivalent to cadaver KTA;i.e., a graft half-life slightly greater than 7 years. However, providing an LRD kidney or an HLA-matched kidney would be likely to increase renal allograft half-life significantly, perhaps to as much as 12-27 years. Therefore, foregoing an increase in expected kidney graft survival of 50 percent to as much as 360 percent in exchange for a projected (actuarial) .64 probability of being insulin-independent at 3 years postransplant could represent a poor strategy to improve health or QOL, particularly if it where also accompanied by a reduction in probability of patient survival. Figures 2 and 3 illustrate renal graft and patient survival associated with the various transplant procedures as reported by Cheung, (37). Sutherland, (28). and the UNOS Registry.(38)

Figure 2. Renal allograft survival .

Figure

Figure 2. Renal allograft survival .

Figure 3. Patient survival.

Figure

Figure 3. Patient survival.

Morbidity

Several authors have noted the adverse effects specifically associated with combined transplant procedures. Olausson et al(6). reported that complications were frequent in SPK recipients, including infections (in 57 percent), pancreatic leakage (28 percent), "cardiocerebrovascular" events (28 percent), and postoperative hemorrhage (10 percent). In the series of Goldman et al, (8). four of the 11 recipients experienced postoperative complications, and there were three deaths (at 3, 16, and 25 months, posttransplant). Stratta et al(9). reported at 31.6 percent rate of "major infections," and noted that 10 of the 38 cases in that series required a total of 15 reoperations.

Sollinger et al(13). in their description of 100 SPK transplant noted that morbidity was not trivial. Urine leak occurred in 13 percent, hematuria in 13 percent, urinary tract infection in 84 percent, and opportunistic infection in 25 percent. Both grafts were lost in 7 percent. The authors stated that the procedure was associated with "significant problems and occasional life-threatening complications."

In a subsequent update extended to 200 SPK procedures, Sollinger et al(15). reported a total of 54 "nonurologic" surgical complications. Infections complications were relatively frequent: 403 occurred within 180 days of transplant and 275 more than 180 days posttransplant. Rosenloff et al(19). categorized complications as surgical or infections; 10 surgical and 15 infections complications occurred in the entire group of 21 cases (one PAK, 16 SPK,, three PTA, and one "cluster" procedure). The former included instances such as perioperative myocardial infarction, bowel obstruction, abdominal abscess, and pancreatitis; the most frequent infections complication was cytomegalovirus (CMV) viremia (11 episodes).

Rosen et al(39). have described the excess morbidity associated with SPK as opposed to KTA at the Mayo Clinic. The authors evaluated 18 consecutive SPK and 18 consecutive KTA patients treated at their institution. They noted a statistically significant 2-fold higher rate of acute rejection episodes in the first year after SPK. (Refer to Table 5 on page 14 for additional information cited here and below on the complications of combined pancreas-kidney transplantation.)

Table 5. Complications of combined pancreas-kidney transplant.

Table

Table 5. Complications of combined pancreas-kidney transplant.

Elkhammas et al, (21). in their description of 150 consecutive SPK cases, reported posttransplant complications of graft rejection (in 85 patients), thromboses (in seven), metabolic complications (44), infection (102), and "major" urinary tract complication (45).

Van Oosterhout et al(22). noted a significantly higher incidence of acute graft rejection in SPK than in KTA cases. Their data are illustrated in Table 4.

Table 4. Acute organ rejection episodes, SPK/KTA.

Table

Table 4. Acute organ rejection episodes, SPK/KTA.

Combined transplant is associated with longer hospitalization than is KTA, and readmission to hospital is often required after SPK or PAK. Stratta et al(9). described an average of 3.2 readmissions per patient, Shaffer et al(17). reported a mean of 3.7 transplant-related readmissions for each case, and Bentley and Garrison(18). observed an average of 2. 8 readmissions per patient. Goldman et al(8). reported the mean length of hospital stay was 36 days. Sollinger(13). reported the average hospital stay as 29.4 days, with an average of 29. 6 readmission days and later updated that information, (15). noting that readmissions were required in 91 percent of patients (181/ 200), with the average number being slightly over three per case. Rosen et al(39). reported a more than 2.5-fold higher rate of hospital days for SPK than KTA: 1,271 for the 18 SPK and 483 for the 18 KTA recipients. This occurred despite the fact that the KTA recipients were initially more infirm than the SPK patients; pretransplant complications of diabetes (such as heart disease, stroke, and symptomatic peripheral vascular disease) "affected more kidney than pancreas/ kidney recipients," and 14 of the 18 KTA patients were on dialysis (for a mean of 13 months pretransplant), whereas dialysis was necessary for only six of the 18 SPK recipients (mean dialysis duration of 4 months).

Comparison of the risks and benefits between SPK/PAK and the alternative KTA with continued insulin therapy should be carefully assessed for each patient. The actual and potential additional risks of combined transplant vs. KTA comprise greater operative and perioperative morbidity, increased number of readmissions and a longer total inpatient stay, greater frequency of rejection episodes, increased intensity of immunosuppressive therapy, more frequent infections (in at least some cases directly related to intense immunosuppression), and a shorter expected half-life of the transplanted kidney if SPK is selected in lieu of LRD/KT or HLA-matched CAD/KT. In addition, the more intense immunosuppression may be associated with other long-term risks. Balanced against these adverse effects are the potential benefits of reduction or mitigation of the secondary complications of diabetes and an improved QOL incident to successful pancreas transplant. An evaluation in detail of the evidence for such benefits comprises the remainder of this assessment.

Secondary Complications of Diabetes After Combined Transplants

Many of the published case series cited above have predicated that combined or sequential pancreas-kidney transplant retards, ameliorate, or prevents the secondary complications of diabetes.(5,7,9,10,15,17,18,) This argument has been used as a rational for the use of SPK and PAK as providing benefits over and above insulin independence (vide supra).

Retinopathy

Konigsgrainer et al(40,41). published two case series that compared ophthalmologic evaluations in a total of 25 SPK patients with functioning pancreas grafts to 14 SPK patients whose graft function had failed. The authors concluded that pancreatic function had a beneficial effect upon retinopathy. However, they also stated that "no firm conclusion can be drawn from this study" and that a "large prospectively randomized trial is needed. "(40).

Ulbig et al(42). followed 38 patients with successful SPK for 38 months; no control group was studied. They concluded that there was a beneficial effect of pancreas graft function in that retinopathy "tended to stabilize" after 3 years, although they cautioned that "we cannot be sure that this beneficial effect...is caused by combined renal and pancreatic transplantation. "

However, Bandello et al (43,44). published two reports of diabetic retinopathy in a total of 23 SPK recipients with functioning pancreas grafts compared to 13 patients who had undergone KTA. No differences between the groups in visual acuity or retinal vascular disease as assessed by angiography were noted.

The same investigators(45). provided additional data for a total patient sample of 32 SPK and 18 KTA, and again found no differences in visual acuity, retinal photography, or angiography between the groups. Patient selection criteria were specified. Thirty of 64 eyes in the SPK and 14 of 35 in the KTA patients were excluded because of neovascular glaucoma, secondary retinal detachment, laser treatment, cataract surgery, optic neuropathy, or CMV retinitis.

Scheider et al(46). evaluated visual acuity and retinopathy in a group of 30 recipients of successful SPK transplants and 15 SPK patients whose pancreas grafts had failed, and found no differences between the groups. A subset of 27 successful SPK and 13 pancreas failures were followed for a total of 52 months with no apparent differences in eye disease due to diabetes.

Petersen et al(47). compared visual acuity, neovasculariztion, and retinopathy in four patients with successful SPK with a group of seven SPK recipient whose pancreas grafts had failed. No differences between groups were noted for any of the parameters evaluated.

Other investigators also found little benefit of pancreas transplant as regards diabetic retinopathy. Zech et al(48,49). published two series describing 21 SPK recipients with functioning pancreas grafts. They concluded that a functioning pancreas graft did not seem to reverse or halt diabetic retinopathy. Scheider et al(50). published an abstract report that compared 19 SPK recipients with a group of six SPK recipients whose pancreas grafts were rejected. Duration of followup was 39 and 30 months for the patient and comparison groups, respectively. Sequential funduscopic photography and visual acuity scores were recorded. No significant progression was observed in retinopathy nor visual acuity for either the study or comparison groups. The authors concluded that normoglycemia resulting from successful pancreas transplant was unable to ameliorate retinopathy. Detailed descriptions of the study parameters were lacking, as the document was published in abstract form.

In perhaps the most comprehensively reported study of pancreas transplant and retinopathy, Ramsay et al(51). compared ophthalmoloogic examinations in 22 successful pancreas transplants and 16 demographically comparable patients whose pancreas grafts had failed. The duration of observation was 2 years. The successful transplants' average HbA1C was reported as 7 percent, whereas the mean level of those with graft failure was 12 percent, confirming the expected superior blood glucose control in the study group. Each ophthalmologic examination included visual acuity measurements, seven-field stereophotography, and macular (fluorescein) angiography. A single reviewer evaluated and rated the (masked) retinal stereophotographs. In the study patients, visual acuity remained stable in 47 percent, improved in 18 percent, and decreased in 35 percent of eyes examined. In the comparison group, 36 percent remained stable, 11 percent improved, and 52 percent decreased in acuity. The mean pretransplant retinopathy score in the study group was 6.09 (zero = minimal, 13 = end-stage); the comparison group's pretransplant score was similar (6.7). The proportions exhibiting stable, progressive, or improved retinopathy in the study and comparison groups, respectively, were: 56, 44, and 0 percent; and 46, 50, and 4 percent. Differences between study and control patients in visual acuity and retinopathy were not significantly different, nor was the rage of progression of retinopathy. A series of regression models evaluated the potential influence of age, sex, duration of diabetes, type of organ donor, previous renal transplant, baseline retinopathy, blood pressure, HbA1C level, and outcome of pancreatic transplant; no significant relation or interaction between these factors and retinopathy was found. The authors calculated that the power of their study to detect a 30, 35, or 40 percent lower rate of progression of retinopathy in the study group was .72, .82, and. 89, respectively. Thus, subtle differences between the study and control groups may have gone undetected. It is of note that the studies of Konigsgrainer, (41). Bandello, (44). and Scheider(46). all indicated that the mean HbA1C levels of patients with failed pancreas grafts were not much higher than those of patients with functioning grafts. The former demonstrated a mean level of 7.7 percent (in a total of 41 patients) and the latter a mean of 6.6 percent (75 patients). This degree of control of blood glucose after KTA has not been reported in much of the literature published by United States investigators.

The studies CiTed above have significant methodologic flaws, which variably include small sample sizes, lack of detail regarding patient selection criteria, absent or suspect comparison groups, and advanced disease of many of the patients at time of transplant which may make extrapolation of findings perilous. For example, using SPK recipients with failed grafts as controls for patients with functioning pancreas grafts presupposes that the patients are otherwise directly comparable, and that the cause of the pancreas graft failures has no concomitant direct or indirect effect upon retinopathy. Likewise, comparison of SPK and KTA patients assumes that the patient groups are clinically comparable, when this is very unlikely to be the case. Nevertheless, it appears that the available evidence does not permit a conclusion that successful SPK with resultant normoglycemia is likely to ameliorate or improve diabetic retinopathy.

Nephropathy

Bohman et al(52). reported results of rental biopsies performed from 24-36 months after transplant in six diabetics provided KTA and two patients provided SPK transplants. They reported light microscopy changes compatible with diabetic nephropathy and electron microscopic (EM) evidence of glomerular basement membrane (GMB) thickening in the kidney-only recipients, but not in the SPK patients. No data were provided regarding patient selection criteria or renal function parameters for the cases reported. Of note were the reported HbA1C values for the KTA patients, viz, a mean of 12.6 percent and range of 9. 4-16.4 percent. These values are high and bespeak poor control of hyperglycemia. Other data cited in the section on retinopathy indicate that much better glucose control has been attained in diabetic KTA patients.(41,44,46)

Bilous et al(53). compared renal biopsy results in insulin-dependent diabetics before and after successful pancreas transplantation (PAK) with biopsies obtained from diabetics who were give KTA. They concluded that normoglycemia can prevent the progressive of diabetic glomerulopathy. However, the data do not permit the conclusion that functioning pancreas grafts will retard, prevent, or ameliorate diabetic nehropathy. Only 12 patients were studied; the "controls" consisted of 13 patients matched for age, sex, duration of diabetes, and length of survival of the renal allograft. No details were provided as to how the 12 study subjects were selected from the larger number of PAK transplants accomplished at that institution. Functioning renal allografts were present a variable period of time (range: 1.0-7.2 years) before accomplishment of the pancreatic transplant. The postpancreas-transplant renal biopsies were obtained as little as 23 months and as long as 10 years after transplant. Before the pancreas transplants, the immunosuppressive regimen consisted of azathioprine and prednisone; five of the 12 study group patients had cyclosporine added after pancreas transplant. None of the 13 controls was treated with cyclosporine. Blood sugar control in KTA group appeared to be less than optimal, with HbA1C values reported as 9.0-15.3 percent. The authors stated that no significant changes in glomerular morphometric indexes were noted after pancreas transplant. In comparing light microscopy examination of biopsy specimens, the KTA patients were reported to have exhibited "minimal to marked" mesangial expansion, whereas the PAK patients exhibited "minimal" expansion. Pancreas-after-kidney transplant patients had smaller mean glomerular volumes as compared with KTA recipients. However, not all PAK recipients could be matched for duration of disease to a KTA recipient. No differences were found between creatinine clearance rates in KTA and non-cyclosporine-treated PAK patients. In contrast to Bohman et al, (52). no difference in GMB thickness was noted between groups.

The renal transplants in the combined transplant group were present an average of 4.2 years before, and 4.4 years after pancreatic transplant; the duration of the renal transplant was "at least 5 years" in the comparison (KTA) group. Given the fact that the expected half-life for transplanted cadaver kidneys is approximately 7.2 years (vide supra) and that the patients in the study and comparison groups were followed for 5.0-8.6 years with no documented difference in renal function, the clinical significance of the greater mesangial expansion in KTA recipients is unclear.

Morel et al(54). evaluated renal function in type I diabetic patients after renal, pancreas, or SPK transplant procedures. Of 62 patients who received a pancreatic graft that functioned for at least 1 year, 53 were analyzed at 1 month, 42 at 1 year, but only 16 at 2 years posttransplant. Results were compared with "matched" KTA patients (27 paired with SPK and 12 paired with PAK recipients). Details of the matching procedure were not provided, but demographic data seemed to be comparable. No differences in renal function were found between pancreatic, SPK, or kidney-only recipients at 1-month, 1-year, or 2-year observation points.

Diabetics receiving KTA will not inevitably develop diabetic nephropathy in the transplanted kidney. Mauer et al(55). reported significant variability in the rates of development of diabetic renal lesions after transplant in 18 patients in whom tissue samples were obtained from 6 to 13 years subsequent to transplant. Five patients exhibited little or no histologic abnormalities associated with diabetic nephropathy, six had "definite but not advance changes," and in seven cases such changes were moderate to marked. The relationship of nephropathy to mean blood sugar levels was complex and not direct; for example, a rank order of severity of hyperglycemia had no statistically significant relationship to the rank order histologically documented nephropathy, but mean blood sugar and histologic abnormalities were associated (correlation coefficient = .61). Although statistically significant, the correlation is rather modest (i.e., it would imply that one can ascribe only 36 percent of the variability in renal lesions to hyperglycemia). They concluded that even in a population with significant diabetic nephropathy before transplant (and thus a demonstrated risk for that complication), "organ sensitivity to the diabetic state" may determine the risk of progression, and that such factors appeared to be intrinsic to the kidney allograft itself.

Wilczek et al(56). compared renal biopsy results from insulin-dependent diabetics provided KTA with a nonrandomized comparison group of SPK patients. The transplanted kidneys of 34 KTA recipients (14 LRD and cadaver-donor kidneys) were biopsied between 1-16 years posttransplant. A comparison group of 20 patients provided SPK was biopsied between 1.0 and 6.5 years post-SPK. Forty-five total biopsies were performed in the KTA and 36 in the SPK recipients. The posttransplant immunosuppression regimen after KTA included azathioprine and steroids in 23 patients, and cyclosporine with steroids "in some cases combined with azathioprine" in 11. Recipients of SPK were immunosuppressed with A-C-P. The authors noted that renal lesions varied considerably among the KTA patients. However, the average "nephropathy score" was significantly higher in KTA than SPK recipients. Electron microscopy (EM) was performed on a proportion of both groups (34 of the 45 biopsy specimens in 26 of 34 PAK patients and 28 of 36 specimens in 15 of 20 SPK patients). Nineteen of the 34 KTA specimens, but only two of the 28 SPK specimens had glomerular basement membrane (GBM) thickening. The authors concluded that functioning pancreas grafts reduce or prevent "the various signs of diabetic nephropathy." Difficulties in extrapolating the results of this study to the general population of type I diabetics with ESRD result from deficiencies in several of the study parameters, including lack of detail provided regarding patient selection criteria for the SPK and KTA groups; lack of information as to the size of SPK and KTA recipient pool from which the groups were selected; variable time of biopsy posttransplant (1-16 years for KTA and 1.0-6.5 years for SPK); and unspecified selection criteria for the subgroups subjected to EM. Moreover, the finding that an unstated proportion of KTA recipients had "absent or minor" lesions on light microscopy was not further explored. The authors' data indicated that approximately 56 percent of the KTA specimens examined by EM did not exhibit significant GBM thickening. This phenomenon has been observed by other investigators, (55). and it appears that at least a subset of ESRD patients will evidence no progression of diabetic pephropathy after kidney transplant.

Moreover, euglycemia resulting from successful pancreas transplant appears to have little effect on even mild established lesions of diabetic nephropathy. Fioretta el al(57). compared diabetic renal lesions in 13 pancreas transplant recipients with those in a comparison group of 10 diabetic KTA recipients. The groups appeared to be demographically similar at study entry. Kidney biopsies were obtained at baseline and 5 years after pancreas transplant in the study group. Baseline biopsies were performed in the comparison group at the time of evaluation for pancreas transplant and 5 years later. Seven of the comparison patients never received a pancreas graft; three were transplanted with resultant graft failure. Of note, two of the pancreas transplant recipients, but none of the comparison group, subsequently underwent renal transplantation. Creatinine clearance fell in the pancreas recipients (average 102 mL/ min to 68 mL/min) but was unchanged in the comparison group (102 vs. 91 mL/min.). The mean 24-hour urinary albumin excretion was 40 mg in the pancreas transplant group and 19 mg in the comparison group (normal = <22 mg); 7/ 13 pancreas recipients and 4/10 comparison patients had abnormal proteinuria at 5 years posttransplant. There was no evidence of a benefit ef euglycemia (successful pancreatic transplant) on glomerular structures; GBM width and mesangial fractional volume did not significantly differ between groups. In fact, one could argue based upon the total urinary albumin excretion and the necessity for subsequent renal transplantation that the natural history of diabetic nephropathy in native kidneys was less severe than that after PTA. The authors concluded that pancreas transplant has no beneficial effects on "established" diabetic nephropathy.

These data provide little direct evidence for prevention or amelioration of diabetic nephopathy secondary to SPK or PAK vs. PAK vs. KTA. Indeed, as previously discussed, the evidence indicates that the duration of renal graft survival and renal function parameters after SPK are comparable, but not clearly superior, to those after cadaver KTA.

Neuropathy

Several investigators have addressed the effect of pancreatic transplant on diabetic neuropathy. Solders et al(58). compared motor nerve conduction velocities (NCV) and distal latency, sensory NCV and amplitude, and R-R variations of the EKG (as a measure of autonomic neuropathy) in 13 SPK recipients, and 15 diabetic KTA and 15 nondiabetic KTA recipients. The authors found that after 2 years of normoglyecemia following SPK, only a mirror improvement in nerve conduction occurred; this was not significantly different from the comparison group of diabetic KTA. No improvement was noted in autonomic neuropathy. The authors concluded that the slight improvement that occurred posttransplant was due to the correction of uremia, and not to the normoglycemia present in all SPK recipients. They subsequently published an additional report with a posttransplant observation period extended to 4 months.(59) In the latter study, the authors compared electorneurographic evaluations and R-R variations in 13 SPK recipients to 18 patients who had either lost pancreatic graft function (14) or were provided only KTA. They measured sensory and motor NCV, distal latency and amplitude, and R-R variations. Both groups improved in objective measures of neuropathy, and in the 24-to 48-month interval the SPK, but not the comparison group, exhibited continuous improvement. However, the differences were not statistically significant, and even after 48 months, all patients still had measurable evidence of neuropathy. The R-R interval variations were abnormal in both groups and no significant differences were noted.

Naouri et al(60). evaluated autonomic and peripheral nerve function in 21 of 100 patients provided SPK between 1984 and 1989 (51 of those 100 had expired or had nonfunctioning grafts). Autonomic function was assessed in 13 patients awaiting SPK, six patients at 12 months or less postransplant, seven patients at 13-24 months after SPK, and 11 patients transplanted a mean of 44 months previously. Tests used were anorectal manometry, R-R interval variation, and blood pressure response to standing. The cardiovascular testing revealed significant autonomic dysfunction in all cases, with no changes related to time from transplant. The 11 patients more than 2 years posttransplant had significantly improved sphincter compliance. Peripheral nerve function was studied in three groups, viz: awaiting transplant, 24 months posttransplant, and 49 or more months posttransplant. Tests included NCV and action The authors stated that peripheral nerve function "tended to improve" after successful SPK. However, the actual published data indicated significant differences were found only for median nerve (not tibial o peroneal) motor and sensory function.

Murat et al(61). measured gastric emptying in 32 diabetics with chronic renal failure (23 of whom were on dialysis) and compared the measurements with those obtained in 18 successful SPK. No significant relationship was noted between gastric emptying and nerve conduction, "parasympathetic cardiac scores," and HbA1C levels. Gastric emptying of solid foods was not greater in successful SPK recipients, whereas rates of gastric liquid emptying were higher in SPK than comparison cases.

Vial et al(62). measured motor and sensory NCV and amplitudes in 20 patients after SPK. Subject available for evaluation at 1-4 years posttransplant numbered 18, 16, 10, and 5, respectively. The data were mixed: motor NCV improved in the five patients tested at 4 years and the 16 at 2 years. Improvement in sensory NCV was significant only for the 1-and 2-year observations, and varied between specific nerves tested. Sensory amplitudes were improved only in the 2-year posttransplant interval, and motor amplitudes showed no significant improvement. The authors had published virtually the same data in a separate earlier report.(63)

Gaber et al(64). studied autonomic function in 12 SPK and four KTA patients. They recorded blood pressure and heart rate, skin temperatures, capillary blood flow, skin temperature recovery time, EKG R-R interval variation, heart rate change with Valsalva, and transcutaneous recordings of gastric rhythm. The authors observed improvements in vasomotor function after SPK and, contrary to Naouri and Solders, concluded that overall autonomic function was significantly superior for SPK vs. KTA patients. Gaber et al(65). reported similar data on a smaller sample of eight patients, at least some of whom appeared to be a subset of the 16 reported above.

Kennedy et al(66). studied neuropathic changes in 61 patients who received pancreatic grafts that functioned for more than a year. Results were compared with a group of diabetics on insulin therapy alone. The investigation was complicated in that the study group of 61 included seven SPK, 19 PAK, and 35 PTA recipients; 34 received cadaveric and 27 LRD pancreas grafts. The comparison group consisted of 48 patients who were awaiting pancreas transplant ( n = 33) or who had received a prior graft that failed (n = 15). Twenty-one of this group had undergone prior transplant. The numbers of study and comparison group patients seen at various time intervals after entry in the study varied; in the study group, 61 were evaluated at entry and 12 months after transplant, 27 after 24 months, and 11 after 42 months. In the comparison group, all 48 were evaluated at entry and after 12 months, 21 at 24 months, and 12 again at 42 months. Neurologic evaluation included clinical examination, and sensory and motor NCV. Cardiorespiratory reflex tests ("Valsalva ratio" as a measure of autonomic function) were performed in a subset of patients (51, 26, and 8 patients at 12, 24, and 42 months followup, respectively). The effects of pancreas transplant on neuropathy were variable. Comparisons between the two groups indicated that a higher percentage of patients with functioning pancreas grafts tended to improve and a lower percentage to worsen over time. The authors attributed the differences to pancreas transplants, and although they stated that peripheral nerve function had a "tendency to improve," they noted that the neuropathy was "only slightly improved 42 months after transplantation," and speculated that the "small" degree of improvement was most likely due to preexisting structural damage to the nervous system. Interestingly, they observed that the increase in sensation and in endurance which most pancreas transplant recipients reported "did not correspond to any change in objective and electrophysiologic results." These authors had previously published similar data on a smaller group of SPK patients(67). seen at the same institution with similar results.

Comi et al(68). compared NCV, action potentials, and symptoms of autonomic and peripheral neuropathies in 16 SPK patients with successful grafts to nine diabetic patients with KTA. One-year posttransplant followup was accomplished in all patients; at 2 years, 12 SPK and six KTA were Symptoms of autonomic neuropathy but not peripheral neuropathy had improved in the SPK patients at 1 year. NCV improved in both groups at 1 year, with the improvement in SPK patients being statistically significant; action potential changes were minor in both groups. By year 2, the remaining patients in the SPK group exhibited further increases in NCV. It is of note that the KTA patient's mean NCV were greater than those of the SPK patients in three of five nerves tested at year 1, and in two of five nerves at year 2; the SPK group had uniformly lower NCV in all nerves at baseline. Although the authors concluded that nerve function improvement occurred solely as a result of persistent normoglycemia, they added a cautionary note that "only long-term followup studies will clarify the exact role of pancreas transplantation in the treatment of diabetic neuropathy. "

Secchi et al(69). studied 22 patients, of whom 14 received SPK and eight KTA. Selection criteria for the differing transplant procedures were not specified. Motor and sensory NCV, action potential amplitudes, and sensations of cold and warmth were measured for up to 2 years after transplant. NCV improved in both groups during the first year; by year 2, 94 percent of the tested nerves in the SPK group evidenced improvement, and in the KTA group, improvement was present in 78 percent of tested nerves. Action potential amplitudes were similar for both groups, as were cold and warmth thresholds. Of interest, the mean HbA1C levels were similar at 2 years for both groups, viz, 5.9 percent for the SPK and 6.8 percent for the KTA patients.

Nusser et al(70). evaluated 26 SPK recipients with functioning grafts and 13 SPK recipients whose pancreas grafts were lost. They compared EKG R-R variations associated with respiration, lying/standing, and Valsalva maneuvers, and used questionnaires to elicit autonomic signs and symptoms. Patients were followed for 36 months posttransplant. The authors reported that only the respiratory R-R variations were significantly improved in successful SPK cases. Autonomic symptoms did not differ between groups. The authors concluded that the improvements in neuropathy which successful SPK produces were "marginal" and that neuropathy was resistant to amelioration.

Muller-Felber et al, (71). investigators at the same institution as Nusser et al, (70). published data in the same year that addressed 53 SPK recipients followed for a mean of 40.3 months posttransplant. During the study, 14 patients had loss of pancreas function over an average followup duration of 21 months. Nine patients lost renal graft function at an average and followup of 30 months posttransplant. The investigators recorded signs and symptoms of neuropathy and measured NCV, distal latencies, and action potential amplitudes. Although symptoms of polyneuropathy improved, neurologic signs were unchanged. NCV increased but sensory action potential amplitudes remained unchanged. The authors concluded that polyneuropathy could be stabilized or "partly reversed" with sucessful pancreas transplant. It was not clear whether some of the patients in the series described by Nusser et al and Muller-Felber et al were multiply reported.

Boucek et al(72). evaluated autonomic neuropathy as measured by cardiovascular reflex testing (R-R respiratory variation, heart rate, and blood pressure response to orthostasis) in nine SPK recipients and 10 KTA recipients. Patient selection criteria were not detailed. Followup extended to 24 months in all patients and to 4 years in seven SPK recipients. All test results were abnormal in both groups at baseline, and none improved over 24 months. No significant changes were noted in the SPK patients followed for 4 years. The authors believed that the lack of improvement in autonomic function was due to extensive nerve damage.

Hathaway et al(73). evaluated autonomic function in 15 SPK and 13 KTA recipients. Six of the latter group received LRD kidneys. Parameters measured were: capillary pulse amplitude, both baseline and in response to hand elevation; reflex vasoconstriction in the hand; cardiac R-R interval; Valsalva ratio; skin temperature recovery time; subjective gastrointestinal symptom scores; cutaneous electrogastrography; and gastric emptying time. Tests were performed before, and 1 year following transplant. The authors reported that, although KTA patients improved on every parameter, the only statistically significant difference pre-vs. posttransplant was the gastrointestinal symptom score. SPK recipients exhibited improvement in vasomotor and cardiovascular function, gastric motility, and symptom scores. The authors concluded that euglycemia plays an important role in reversal of autonomic neuropathies. However, of the three tests of cardiac function, SPK recipients demonstrated significantly greater improvements in only one, viz, Valsalva ratios. Likewise, in the three tests of gastric function, superiority of the SPK group was seen only in the subjective symptom scores. Moreover, the authors did not comment as to whether such statistically significant differences as they described were in fact clinically important. For example, the pretransplant Valsalva ratios for KTA and SPK recipients were 1.2 +/-0.1 and 1.1 +/-0.1, respectively. Posttransplant ratios were 1.3 +/- 0.1 for KTA and 1.3 +/-0.1 for SPK recipients. In view of the lack of detailed information as to patient selection criteria for SPK and KTA, this study provides suggestive but hardly conclusive evidence that some autonomic functions may be improved by a functioning pancreas graft, or perhaps alternatively by KTA and assiduous insulin therapy.

Although it would undoubtedly be difficult to conduct a prospective, randomized trial of the effect of SPK or PAK on diabetic neuropathy, all of these case series had significant methodologic flaws. These included: lack of detailed descriptions of patient selection criteria; comparisons of groups likely to have been clinically, if not demographically, dissimilar; combining results in SPK and PAK recipients and cadaveric and LRD (segmental) pancreas donors; incomplete reporting of data regarding glucose control in the study and comparison groups, and so Notwithstanding, taken as a whole, the data do not constitute compelling evidence for improvement of peripheral or autonomic neuropathy after successful pancreas transplantation.

Despite not directly addressing the issue of diabetic neuropathy, a recent study of microangiopathy is worthy of comment. Cheung et al(74). described an "intravital" microscopy system couple to a computer-assisted imaging system using a videotimer with high-resolution videocamera and display. Intravenous injections of sodium fluorescein provided imaging of nailbed capillary morphology and transcapillary leakage. Eleven SPK patients were studied from 14 days to 30 months posttransplant; three were studied twice and one three times. Groups selected as comparisons ("control groups") included 10 healthy volunteers, five type I diabetics, and 11 pre-SPK patients (five of whom were also included in the posttransplant group). No description of the selection criteria for any of these groups was provided. Capillary morphometric measurements did not differ between groups, nor did flow velocities. However, capillary leakage times increased 12 months after successful SPK in some cases, approximating those of the healthy volunteers. The authors concluded that successful SPK can reverse diabetic microangiopathy. This study provided interesting information regarding small vessel damage secondary to diabetes and the possibility of improvement after SPK. However, the data provided permitted the reader to calculate that the mean time to capillary leakage of the normal volunteers was 187 seconds; of the 16 SPK subjects, six had leakage times not significantly different from the pre-SPK mean of 30 seconds, and only seven were at or above 165 seconds. Taken together, with the fact that the measures were taken at widely varying posttransplant times ranging from 11 days to 2.0 and 2.5 years, and considering the small sample size and absent selection criteria for inclusion into the various groups, the reliability of the conclusions that may be drawn is limited. This report provides interesting observations of the potential effects of pancreas transplantation (or of normoglycemia) on microvascular diabetic disease; unfortunately, the flaws in the experimental design do not permit a conclusion that these benefits will accrue from SPK.

Cheung et al(75). provided additional data on eight SPK and 48 "controls," the latter comprising groups of eight patients each of uremic and nonuremic diabetics and nondiabetics, and diabetics and nondiabetics KTA recipients. No patient selection criteria were specified. Results were, in general, similar to their prior report. The authors concluded that it was "evident" that improvement in early capillary leakage in the SPK recipients was due to pancreas transplantation. However, based upon the very limited data provided, that conclusion appeared to be quite tenuous.

The available data concerning the effects of successful pancreas transplantation on the diabetic complications of retinopathy, nephropathy, and neuropathy do not permit the conclusion that amelioration or improvement of complications constitutes a rationale for transplant. The American Diabetes Association (ADA) has also concluded that pancreas transplantation has not been shown to prevent or ameliorate the secondary complications of diabetes.(76) The National Institutes of Health (NIH) is supporting an ongoing registry of the effect of pancreas transplantation at selected medical centers upon the secondary complications of diabetes (personal communication, NIH, November 2, 1994). When available, these data may provide reliable evidence as to whether any improvement or mitigation in such complications results from combined transplantation.

Sutherland(77). has noted that influencing secondary complications of diabetes is not the primary goal of SPK/PAK today. He stated that the "...effect of the pancreas transplant on the course of established secondary diabetic complications are (sic) variable" and that "...it is the overall effect on quality of life that is most important." Nevertheless, a number of recent reports have alleged that SPK or PAK is an effective therapy for diabetic retinopathy, nephropathy, and neuropathy.(3,5,7,9,15,17,18)

Quality of Life

Zehrer and Gross(78,79). reported a series of evaluations of quality of life (QOL) in 131 pancreas transplant patients. The 1991 study retrospectively reviewed outcomes in 131 pancreas recipients including SPK, PAK, kidney-after-pancreas transplant, and PTA recipients. Twenty-six of the patients (20 percent) had one or more pancreas retransplants. The authors presented several indices from patient self-reports and compared recipients with failed grafts with those whose graft function remained intact. The questionnaire that was used had been composed specifically for this study. The authors calculated an "index of well-being" derived from a single seven-point question combined with the score on Campbell's Index of General Affect. Assessment of activities of daily living was accomplished by a modified self-report version of the Karnofsky Index. Questions about health care burden and satisfaction were developed by the authors. Patients also rated their own general health on a four-point scale ("excellent," "good," "fair," "poor"), and reported comparative changes resulting from pancreatic transplant ("more healthy," "less healthy," "about as healthy").

The authors reported that health and ability to perform daily activities were related to pancreas graft function; indices of well-being and general affect, overall life satisfaction, and specific satisfaction were described as significantly higher for patients with functioning grafts.

A subsequent report the following year evaluated QOL in the same patients but separated them according to the categories of PTA (n = 62), SPK (n = 28) or sequential pancreas-kidney transplant ( n = 41). An unspecified number of the sequential procedures were kidney-following-pancreas transplants, a procedure not under consideration in this assessment. This paper provided some additional detail regarding the QOL measurement techniques used. The authors noted that the four-point global health rating was based on the General Social Survey, and the questionnaire included the Medical Outcome Study (MOS) short form. Hospitalizations, emergency room visits, and days sick in bed were also recorded. The authors concluded that patients with functioning grafts had significantly more positive health perceptions, reported greater ability to function socially, stated they were nearer to normal with regard to activities of daily living, and were more likely to view themselves as healthy since transplant than those patients whose pancreas grafts had failed. Gross et al(79). noted that the "observational nature" of the study design "limits our ability to infer that the observed differences are benefits of successful pancreas transplantation." They postulated that a "prospective study...can best estimate the type and size of both benefits and adverse effects," a view that has also been clearly expressed by Piehlmeier.(80,81)

Zehrer et al(82). described QOL measures in a group of 14 SPK recipients. This appeared to be a followup of previously reported series. Fourteen patients with functioning SPK, 16 with functioning KTA (cadaver), and 23 "control group" patients completed QOL questionnaires at baseline and 1-year posttransplant. The questionnaire included the Diabetes Control and Complications Trial (DCCT) Research Group instrument (vide infra) and the SF-36 form. Both transplant groups reported improvements in three of the SF-36 scales, and the SPK group improved in one additional scale. The SPK group had higher overall SF-36 scores. Although the term "control group" was used, no details were provided as to selection process for the 23 type I diabetics, nor were selection criteria specified for the smaller SPK and KTA groups. No data were provided as to how patients were selected for SPK or for KTA. The authors (again) stated that larger numbers of patients and longer followup times are required for "more complete evaluation."

There are, however, significant limitations of these studies which call into question whether they in fact support an unambiguous conclusion that pancreatic transplantation has been demonstrated to provide significant improvements in QOL. The investigations were retrospective, and included patients subjected to very different transplant procedures (SPK, PAK, KAP, and PTA). Both primary and retransplantation cases were included. Moreover, it is not clear that the groups with failed vs. functioning grafts differed only on the basis of pancreatic function. It is of note that the patients with functioning pancreas grafts were surveyed on average 33 months posttransplant, whereas those with nonfunctioning grafts were questioned 53 months posttransplant. A difference of nearly 2 years in time from transplant to survey between the two groups is disquieting in patients with chronic disease subject to long-term immunosuppression, and makes it difficult to maintain that the groups were truly comparable. Other factors not directly related to graft function could also have affected perceptions of QOL. For example, the failure of a transplanted organ is always an undesirable outcome and could engender disappointment and depression resulting in perceptions of dissatisfaction with health status. Moreover, the underlying physiologic factors responsible for graft failure could themselves have resulted in symptoms and reduced the perceived QOL.

The clinical significance of the magnitude of the QOL differences reported between groups was uncertain. In the initial study, the lowest ratings of activities of daily living ("requires considerable assistance and medical care" and "disabled") were chosen by 6 percent of those with functioning and 14 percent of nonfunctioning grafts. Analyses of the reported decrements of QOL scores from patients with functioning vs. nonfunctioning grafts for categories of "index of well-being," "general affect," and "overall life satisfaction" revealed that the average reductions of those scores in patients with loss of pancreatic graft function were 13, 12, and 14 percent, respectively.

In the second study, (79). comparisons of patients with functioning vs. nonfunctioning grafts in the combined kidney-pancreas graft patients revealed no differences in proportions hospitalized or seen in an emergency room within the previous year, nor were differences noted in the number of days in the prior month during which health problems inhibited normal activity. Days sick in bed in the previous month were not significantly different between those with functioning or nonfunctioning grafts after adjustments were made for case-mix, nor did Karnofsky ratings differ. Although only 27 percent of successful graft recipients were working full time and 3 percent part time, the corresponding figures for those with failed pancreas grafts were 41 percent and 7 percent, respectively.

The authors published an additional study(83). which compared 37 patients (21 SPK, 14 PAK, and 2 KAP) with functioning kidney and pancreas transplants to a nonrandomly selected group of 50 diabetic recipients of cadaver kidneys (KTA). The 37 were selected from the larger group of 131 reported above. It is unclear whether the comparison group possessed certain characteristics that could have been responsible for both a lesser satisfaction with life quality and the fact that they were not selected for, or deferred, pancreas Additional questions selected from the DCCT were posed to patients. The authors reported no differences in overall life satisfaction (in contrast to the studies previously noted), higher levels in two of the six MOS domains, and slightly greater satisfaction on the DCCT items (scores of 70 vs. 63) for recipients with functioning pancreas grafts.

Nakache et al(84). compared two groups of diabetic patients in Sweden: 14 who received combined pancreas-kidney transplant (group I), and 16 "chosen to match" them (group II). The latter comprised eight SPK recipients who experienced "immediate" pancreas graft failure, four LRD renal transplants, and four cadaver KTA. The authors reported that group I patients were less likely to be paid a sickness pension, had fewer lost workdays and nights spent in hospital, were more likely to be employed fulltime, and exhibited fewer complaints of deteriorated sexual function than did group II patients. However, there were no significant differences between groups in terms of proportions employed, in numbers engaged in regular sports activities, in self-reports of physical well-being, in mean perception of health, or in ratings of "rehabilitation to normal living." Moreover, it was not made clear that such differences as were observed between the groups could be ascribed to the presence of a functioning pancreas. Eight of the 16 group II patients had in fact previously received a pancreas graft with resultant "immediate failure" of the graft. The question as to whether physiologic factors that might have been related to rapid graft failure could have also been responsible for deleterious effects upon QOL was not addressed in the report. In addition, the two groups were not well matched for age: group I patient ages ranged from 25-40 years, while group II patients were older, with an age range of 31-51 years, and the difference in mean ages was greater than 4 years. In chronically ill patients, greater age could itself be associated with poorer QOL.

Nakache et al(85). updated this report and described QOL measures in eight SPK and 10 KTA patients whose grafts functioned for 7 years. Differences noted in the earlier study were comparable to those observed between the various groups at 7 years.

Nathan et al(86). reported metabolic and QOL evaluations in a group of 33 SPK patients and a comparison group of 18 KTA and one SPK patient who had lost pancreatic graft function. The latter group was specifically chosen from a larger population in an attempt to provide similarities to the SPK patients in age, duration of disease, and baseline complications. The authors noted that because of the lack of randomization, their report was "largely descriptive." The questionnaires used included the QOL survey developed for the DCCT (the "diabetes quality-of-life survey" or DQOL) and "a more general quality-of-life instrument" from Campbell. Baseline QOL assessments were done for 28 of 33 SPK, but only for five of 19 in the comparison group. Only the first QOL assessment completed 6 months or more after transplant was evaluated. Twenty-five of the 27 SPK, and 14 of the 19 renal recipients with 6 months or longer followup completed at least one general and DCCT survey instrument at 18 +/-10 months posttransplant. The mean scores of successful SPK recipients on all four subscales of the DCCT questionnaire were significantly higher than those for the kidney transplant patients. No differences were noted between groups in the general QOL assessment scores. The authors concluded that SPK improves QOL in diabetic patients with ESRD but noted that they believed it was arguable as to whether improved QOL justifies the added expense, increased adverse effects, and longer duration and greater frequency of hospitalization in SPK compared with KTA.

This study again indicates the methodologic difficulty in selecting an appropriate comparison group. The authors noted that their selection of a "control group" was "far from ideal." The SPK patients were generally selected before beginning renal dialysis (36 percent dialyzed vs. 78 percent in the renal transplant group), and there was a higher incidence of autonomic neuropathy in the renal transplant group. The authors indicated that patients in that group were not provided pancreas transplant because of personal preference, LRD KTA, or abnormal urodynamics. It is unclear whether such intergroup differences could have affected posttransplant subjective estimates of QOL.

Piehlmeier et al(80,81). reported results of a cross-sectional survey directed at determining QOL in 157 diabetic patients who had received pancreas and/or kidney transplant. Five patient categories were studied, viz:

  • Group A: 29 pretransplant patients not on dialysis.
  • Group B: 44 pretransplant patients on dialysis.
  • Group C: 31 patients post-SPK with both grafts functioning.
  • Group D: 29 posttransplant patients with intact renal function but on insulin therapy.
  • Group E: 15 posttransplant patients on dialysis and insulin.
  • Group F: 9 patients post-PTA on insulin.

Patients were asked to respond to 276 questions, 59 of which were specifically directed at diabetic somatic complaints. The questionnaire was intended to encompass six components of QOL (vide infra). Statistical analyses consisted of calculations of mean component scores, intergroup differences were tested by Student's T-Test, Spearman correlation coefficients were calculated between component scores and overall QOL, and multiple regression analyses were used to determine variables contributing most to the overall QOL.

The authors reported that for "leisure time activities and social contact" and for "physical and mental status," the scores of both groups C and D were superior to those of group B; for "overall QOL," groups C and D were more satisfied than other groups; for "partnership, sexual and family life," scores of groups C and D did not differ; and for "emotional satisfaction" scores of groups C and D did not differ.

The lowest QOL scores were reported in patients on dialysis, but in no component were QOL scores of groups C (post-SPK, both grafts functioning) and D (post-KTA on insulin) significantly different. Ratings of "satisfied" or "very satisfied" were recorded by 74 percent in group C and 69 percent in group D; 26 percent in group C and 34 percent in group D were employed. Therefore, this research indicated that successful kidney and SPK transplant both tended to improve QOL, but there was little if any evidence that a functioning pancreas graft provided QOL improvement over and above that resulting from a functioning kidney graft. The lowest QOL scores were associated with ESRD requiring dialysis.

The authors stated that there was some disagreement as to how to measure They remarked that "while various quality-of-life instruments have been carefully developed and used in different patient populations, their benefit in transplantation research and specifically in diabetic patients is rare. " However, they believed that their approach was more appropriate at least insofar as it followed the disease-specific approach of the DCCT Research Group. They recommended that future work should be based upon prospective designs which "in spite of randomization problems will provide a more appropriate account of changes in QOL due to transplantation."

In an apparent followup to these reports, Piehimeier et al(8). mailed questionnaires to all patients who had been provided SPK or who were on the SPK waiting list. Thirty-eight patients who responded in both 1989 and 1993 were included in the analysis. The questionnaire response rate was not specified. All patients replied to two questionnaires within a time span of 3.0-3.5 years and were divided into four groups: viz, (1) eight pretransplant later given SPK; (2) 15 with both grafts functioning, evaluated after transplant; (3) nine patients with pancreatic graft loss and renal graft function evaluated after transplantation; and (4) six patients on the waiting list. The questionnaire was identical to that previously described. Group A QOL scores increased after transplant, but differences were not statistically significant. The scores of patients in group B remained essentially unchanged. In group C, the scores showed "no marked amelioration in overall QOL." Scores for grou D patients increased over time. The authors concluded that QOL improves after successful pancreas-kidney transplant, although it is unclear how the data permitted such a conclusion, inasmuch as none of the reported differences were statistically significant.

Bentdal et al(89). reported their assessment of QOL as estimated by rehabilitation and ability to work in 27 SPK patients (11 had both grafts functioning and 16 had lost pancreatic function). No difference was found between groups in ability to work. The authors stated that it was "probable' that assessment of emotional and psychological well-being would have demonstrated "important differences" between groups but provided no evidence to that point.

Johnson et al(90). used karnofsky scale ratings to evaluate QOL in terms of physical ability in 15 SPK and 2 PAK recipients. Assessments were performed by a single social worker. The authors concluded that improvement followed SPK/PAK. However, the data were confusing ratings were reported for followup periods described as "posttransplant up to 1 month," "2 months posttransplant," and "longer than 3 months postoperatively," whereas the Materials and Methods section of the paper stated that QOL "...was assessed at pretransplant and at 3, 6 and 9 months and 1-year posttransplant."

Secchi et al(91). evaluated QOL in 31 SPK, nine KTA, and 44 uremic pretransplant patients on dialysis. Multiple-choice questions were posed which purported to assess physical well-being, social well-being, and working capacity. Details were not provided as to the methodology used to construct or to validate the questionnaire. The authors stated that SPK recipients enjoyed a better QOL than did KTA recipients. Intergroup differences were observed in the physical well-being and social life components; however, physical activity and working capacity were similar for SPK and KTA recipients.

Milde et al(92). evaluated QOL in 44 patients who received pancreas and kidney transplant (either SPK or "sequential" procedures), and compared the responses of 31 with functioning pancreas grafts with 13 whose grafts had failed. The measurement instrument for QOL was apparently constructed for this study and consisted of the following components: Karnofsky Index; Pearson and Byar's Subjective Fatigue Checklist; Myer's Resources and Social Support Questionnaire; Parfrey's Physical Symptom Scale; Gill's Adjective Checklist; Health Locus of Control Scale; the Beck Depression Inventory; Multiple Affect Adjective Checklist; Simmon's Self-Esteem Scale; Campbell's Index of Psychological Affect and Index of Life Satisfaction; and Cantril's Self-Anchoring Scale.

The reference that the authors cited for the Pearson and Byar checklist was a 1956 publication of the United States Air Force School of Aviation Medicine. The Myer's questionnaire war referenced to at text entitled Handbook of Tests and Measures for Black Populations; Parfrey's scale was referenced to an abstract in Transplant Proceedings; references cited for the Beck Depression Inventory, the Multiple Affect Adjective Checklist, and the Cantril Scale were published between 1965 and 1967. The authors did not indicate that these psychometric tests developed for specific populations (e.g., fatigue measurements by the Air Force Flight Surgeons' School, or social function questionnaires developed for black patients or constructed as long as 28 years ago have been validated in current populations of transplant patients.

The report stated that there were no statistically significant differences between the two groups in terms of major demographic variables; however, the percentages of patients in the groups of successful vs. failed pancreas grafts appeared to differ substantially in certain respects, as detailed in Table 6.

Table 6. Successful versus failed grafts[a].

Table

Table 6. Successful versus failed grafts[a].

Thus, although these differences may not have been statistically significant, the groups hardly appear to have been homogenous with respect to these characteristics.

Notwithstanding, the authors reported that the groups did not differ with respect to levels of fatigue and activity or Karnofsky scores, physical symptoms, level of happiness, energy, locus of control, depression, anxiety, hostility, self-esteem, sense of well-being, or present QOL and health. The group with pancreatic failure was significantly more satisfied with the support they received. eight percent of the successful transplant and 46 percent of the failed pancreas transplant group were employed. However, the authors concluded that recipients of successful pancreas transplants perceived their improvement in health and QOL to be significantly superior to that of those with failed grafts on the basis of a significant time by group interaction when repeated-measure analyses of variance were performed; that is, the group with functioning pancreas grafts perceived their prior state of health to have been much lower than did those patients with failed grafts, and expected their future health to be superior. The authors stated, however, that "the multiple comparisons, of course, may have resulted in significance occurring by chance," and pointed out that "prospective, long-term followup studies" were needed to identify the impact of pancreas transplantation on QOL.

Stratta el al(10,12). published two reports of 61 consecutive SPK procedures performed at their institution. The authors stated that QOL was assessed "by patient interview and written questionnaire" (not further specified). They reported that QOL was improved after transplant. In one study, (10). the 61 SPK recipients were compared with a nonrandomized age-matched group of KTA recipients; QOL was compared for seven categories of normal social activity, full-time vocational rehabilitation, part-time vocational rehabilitation, full-time school, disability, or working (nonmedical), and delayed rehabilitation. The only statistically significant differences between SPK and KTA patients was a higher proportion of disability among KTA patients.

Hathaway et al(93). described QOL assessments in 15 SPK and 15 KTA recipients. The SPK recipients were disproportionately white, female, high school graduates, younger, and with a shorter duration of renal disease than KTA recipients; these differences were statistically significant. Sickness Impact Profile (SIP), QOL Index (QLI), Adult Self-Image Scales (ASIS), and a one-item General QOL question were used to evaluate QOL. All SIP scores improved from baseline to 6 months posttransplant, with slightly higher scores reported for SPK recipients. Scores for QLI exhibited similar patterns. Scores for ASIS were higher for SPK recipients both at baseline and 6 months. The Global QOL measure differed little between groups and improved for each group at 6 months. The authors noted that "several scales" showed significant improvement for SPK recipients "that was not seen to the same extent for KTA recipients" and concluded that a combined transplant may afford greater improvement in QOL. However, the published data do not permit even such a tentative conclusion. Selection criteria for SPK vs. KTA recipients were not specified; there was significant differences between groups in terms of race, sex, age, educational level, and duration of ESRD; followup was limited to 6 months; and the reported intergroup differences were marginal.

Kiebert et(94). retrospectively evaluated QOL in 17 SPK recipients and 23 KTA recipients with functioning grafts, as well as a third group of 11 SPK recipients with failure of one or both grafts. Patients selection criteria, clinical and demographic data pertaining to the various patient groups, and the QOL measurement instrument were not described in detail. The authors noted that a "comparable improvement and satisfaction" followed either a successful SPK or KTA, and that SPK provided further improvement in "mobility to perform daily activities."

Schareck et al(95). retrospectively evaluated QOL in 54 SPK recipients. Patients were interviewed by telephone. Although details regarding the questionnaire were not provided, the authors stated that patients' reports of "well-being" rose postoperatively.

Several investigators.(83,86) have included the DQOL in their evaluations of QOL after pancreas transplant. This instrument was developed by the DCCT Research Group.(96) It is instructive to review the methodology used by the DCCT group in the construction of their questionnaire, and how reliability, validity, and intertest reliability were assessed. Patient criteria for entry into the DCCT and the DQOL study were the same and included restriction to the 13-to 40-year age range, duration of insulin-dependent diabetes of 1-15 years, less than 130 percent ideal body weight, and generally good health without advanced complications of diabetes. The study population consisted of 136 adults and 56 adolescents, with a mean duration of diabetes of 8 years. The questionnaire was composed of 46 questions which addressed four primary scales, viz, "satisfaction," "impact," "diabetes worry," and "social/vocational worry." The authors stated that items used in the DQOL were derived from a review of the diabetic literature, clinical experience of knowledgeable health professionals, and patients with type I diabetes. In distinction to other QOL measures, "worry scales" were included because of the observation that concerns or worries may importantly affect the patient or family. Drafts of the DQOL were circulated for critical review to patients, experienced diabetologists, and nurses who were members of the DCCT group. Validity of the DQOL was tested against several instruments: the Symptom Checklist-90-R (SCL); the Bradburn Affect Balance Scale (ABS), a measure of psychologic well-being used in various large population studies; and the Psychosocial Adjustment of Illness Scale (PAIS), which assesses the personal aspect of illness on different aspects of an individual's life. Analysis of the DQOL for internal consistency revealed highly consistent results in general, with the diabetes worry scale ranking lowest in that parameter. Test-retest reliability was quite good. The satisfaction and impact scales were highly correlated with the SCL, ABS, and PAIS scales; the diabetes worry and social/vocational scales were significantly correlated only with the SCL and the psychological distress scale of the PAIS.

In the patient population in which this test was developed, the DCCT Research Group reported that most were "generally satisfied" and "not worried" and indicated that diabetes made only a modest impact on their lives. Therefore, these diabetic patients apparently did not consider their QOL was significantly degraded by their disease.

The authors concluded that the DQOL questionnaire exhibited validity in terms of other accepted psychometric tests, but noted that the was then no "gold standard" for assessing validity of any QOL measure. The development of this measurement tool was relatively precise and statistically rigorous and conformed with accepted psychometric test development practices in evaluation of parameters such as reliability and validity. However, the patient population tested was fairly homogeneous with respect to demographic and clinical status (matching the DCCT requirements), and the authors suggested further research to validate the DQOL in patients with "divergent health characteristics." Therefore, the homogeneity of the test population upon which the DQOL questionnaire was based raises questions as to the conclusions that may be drawn from its application for other than research purposes in QOL assessments of diabetic pancreas transplant patients.

However, the final report of the DCCT did address issues of QOL as affected by intensive insulin theraphy vs. standard therapy. In that study, intensive therapy included such elements as the administration of insulin three or four times daily (or constant administration by use of an external infusion pump); self-monitoring of blood glucose at least four times daily; weekly 3:00 a.m. blood glucose determinations; monthly HbA1C measurement; adjustment of insulin dosage according to glucose measurements, dietary intake, and anticipated exercise; and monthly medical visits. Conventional therapy required one or two daily insulin injections, daily blood or urine glucose monitoring, dietary and exercise education, and medical examination every 3 months. Despite the arduous regimen in the intensively treated group and the three-fold increase of severe hypoglycemic reactions resulting from that treatment, there were "no significant differences in the mean total scores on the trial's quality-of-life questionnaire."

In sum, the reports cited above provide very limited evidence of improved QOL resulting from transplantation of the pancreas. Many of the studies had significant methodologic flaws which variably included: the retrospective nature of the reviews; a lack of demonstrated comparability of patient groups; differing transplant procedures; inclusion of both primary and retransplant cases; different posttransplant intervals at interview for the groups compared; different ages of patients with functioning vs. nonfunctioning grafts; lack of demonstrated validity and reliability of the test instruments for the patient population under study; and the potential for selection bias. Moreover, because a wide variety of questionnaires and survey instruments were used and few of the investigators utilized identical measures of QOL, any comparisons across the individual reports are quite tenuous.

It is evident that insulin-dependent diabetes mellitus may diminish the QOL. However, the magnitude of such effect varies greatly between patients, and such variations need to be considered when evaluating the most appropriate application of a procedure such as pancreas transplantation. Some patients may be required to administer insulin and measure their blood glucose level several times daily, suffer the results of frequent hypo-and/or hyperglycemic (H/ HG) episodes, require emergency treatment and hospitalization for blood glucose fluctuations, and experience excess mortality as well as morbidity from poor glucose control. Others, such as the patients evaluated for the development of the DCCT/ DQOL questionnaire, experience much less morbidity from their disease. No published case series of pancreas transplant has provided information as to the pretransplant diabetic treatment regimen, the severity and morbidity of the diabetes and its treatment regimen, and the resultant impairment of QOL in patients selected for SPK or PAK. No prospective studies of the effects of SPK or PAK on QOL have been published. Indeed, several transplant centers have informed OHTA that they do not consider the severity of diabetes to be of significance in evaluating patients for pancreas transplant (personal communication, University of Rochester, March 1, 1993; personal communication, University of Minnesota, February 25, 1993). In the absence of comprehensive data regarding the severity of diabetes and the arduousness of the treatment regimen before pancreatic transplantation, it is not possible to determine accurately the benefits expected in terms of improvement in QOL. This issue is critical given that even the most optimistic reports described above provide only modest evidence of improvements in QOL.

Cost-Effectiveness Analyses

A formal cost-effectiveness analysis (CEA) of SPK or PAK could not be completed. Reliable cost data were not available for either procedure; only isolated information on charges and payments were accessible. Moreover, it was not possible to determine clearly the differences in effectiveness between SPK/PAK and the alternative treatment of KTA with continued insulin and diet theraphy. As previously discussed, there was no compelling evidence that the secondary complications of diabetes were affected by SPK or PAK. Arguments that the QOL was improved were not well supported by objective data from prospective studies using adequate, reliable, and valid measurement techniques; the methodologic problems of the majority of studies of QOL following SPK/PAK have been elaborated on previously in some detail. Therefore, a model for CEA was constructed to provide comparisons of the relative cost effectiveness of SPK/PAK and KTA over a range of QOL estimates.

In this model, a number of costs based on the best available charge and payment data were used. Estimates of effectiveness were based on assumed differences in QOL that were assigned to each of the procedures. This paradigm then permitted a series of inferences regarding the cost effectiveness of the alternative treatments based on variations in cost and QOL. It also indicated the quantitative differences in QOL which must be substantiated to support the cost effectiveness associated with the levels of resource requirements for all of the alternative treatments. As will be discussed, the model was constructed so as to give the benefit of any doubt to SPK.

Cost-effectiveness analyses evaluate the cost and consequences associated with health care interventions for the purpose of making informed choices among competing technologies to provide the most efficient use of available resources. Results are expressed as the net costs to produce a unit of output quantified in terms of health, such as diseases detected, lives or years of life saved, or quality-adjusted life years (QALYs). These units provide a common basis for direct comparisons of the results of applying different technologies.

Although the addition of years of life is a direct and manifest basis upon which to compare technologies, it is uncommon that technologies can be so obviously distinguished. More often, differences in patient outcomes due to the application of alternative interventions are far more subtle; this circumstance led, in part, to the development of the QALY concept. There are disagreements both as to the most valid measures of QALY and the appropriateness of comparing QALYs. Underlying assumptions include the premises that 1 year of life in perfect health is equivalent to 2 years at half that quality, and that the addition of one QALY at age 75 is viewed essentially the same as the addition of one QALY at age 25. Although a comprehensive discussion of these issues is beyond this assessment, the QALY concept has gained general recognition as an acceptable tool for evaluation of the comparative clinical utility of two or more technologies.

One common approach to calculation of QALYs is based on utility theory principles which were originally developed for economic decisionmaking under conditions of uncertainty (von Neumann-Morgenstern, or vN-M utilities) (98) Such utilities are derived from a series of "standard gambles" or preferential choices wherein an individual selects either one or another of two alternatives or is indifferent to the choice. (Alternative methods used to calculate utilities have included the "time trade-off" technique.) An iterative process of such discrete choices permits derivation of utilities of various health states. These utilities can then be used as the adjustment weights for QALY. Some believe that the calculations of these weights are best based on preferences expressed by patients, not clinicians or the general public, following the observation that patients are more likely to assign higher QOL to their clinical condition. Others have argued that weights assigned by the general public ("consumers") should be emphasized.(98) Examples of such preference weights are seen in Table 7 (page 29).

Table 7. Sample utilities for selected health states [a].

Table

Table 7. Sample utilities for selected health states [a].

In this assessment, estimated preference weights for QALYs were used in a CEA model (vide infra), because available data do not permit a direct comparison of the effectiveness of SPK or PAK with the alternative of kidney transplant alone and the associated treatment of H/HG.

It has been previously noted that there are only limited data regarding the expected duration of pancreas graft functional survival; i.e., a relatively small number of patients with functioning pancreas grafts have been followed for 1 year or longer, and only about 50 for 4 years posttransplant. Actuarial graft survival from registry data has been calculated only to 3 years, which is less than the expected survival of diabetic renal transplant recipients. Moreover, as previously described, there are few data that address the additional benefit accruing from SPK/PAK relative to KTA. It is clear that successful SPK or PAK will render patients euglycemic and obviate or reduce the need for insulin administration and other treatment necessitated by H/HG; actuarial predictions indicate that this may be the case in approximately 35 percent (PAK) to 64 percent (SPK) of patients at 3 years posttransplant.

Other benefits are much less certain. There is no evidence that adding pancreas to kidney transplant increases life expectancy, nor are there compelling data that secondary complications of diabetes are affected. Moreover, the QQL after SPK/PAK vs. KTA has not been clearly elucidated (vide supra). Many of the reports of QQL cited previously have commented that prospective studies are needed and that carefully constructed, reliable, and validated measurement instruments are lacking. The QQl questionnaire developed for the DCCT represents a potentially useful instrument, but it was not developed for, nor has it been validated in, populations of transplant recipients. No published studies of patient utility preferences address SPK or PAK, and an extensive bibliography of cost-effectiveness studies did not include a single reference addressing panreatic transplant.(2)

Little appears to have changed since Evans stated in 1991 that "data on the cost of transplantation are sorely lacking."(99). Reliable, well-documented costs of SPK and PAK have not been reported. Charges and limited payment data are accessible and in some circumstances have been used aas surrogates by experts in the field.

The OHTA published a notice in the Federal Register announcing this assessment and soliciting information on the subject. Reimbursement data for SPK and KTA were also solicited from private and Federal third-party payers. Payments for the SPK transplant hospitalization made by private insurers in 1993 to transplant centers located in those geographic regions in which the majority of pancreas and combined kidney-pancreas transplants are performed in the United States averaged $152,700; payments made for KTA in the same geographic regions averaged $77,300 (personal communication, Health Insurance Association of America [HIAA], October 26, 1993. Medicare does not currently cover pancreas transplant (SPK, PAK, or PTA). In 1989 the average Medicare reimbursement for KTA (transplant hospitalization alone) averaged $43,110; payments for the renal transplant portion of SPK averaged $47,223 (personal communication, Dr. Paul Egger, October 1, 1993. Office of Research, HCFA). In 1990, "total 1-year" Medicare payments for renal transplant recipients for the entire calendar year after the day of admission averaged $79,716. This payment includes coverage of immunosuppressive drugs as well as payments made for all care, whether related to the transplant or not, in the year following renal transplantation.

Medicare payment levels for KTA (transplant alone) increased by an average of 3. 5 percent per annum from 1985-1989.(99) Extrapolating the payment level at the same rate of increase from the reported $43,110 in 1989, the Medicare payment in 1993 would be $49,469. The total 1-year costs for ESRD patients who received renal transplants increased at an annual rate of 7.3 percent from the 1985-1989 era (most of the increase was related to outpatient costs). Again, extrapolating at the same rate of increase, total 1-year payments in 1993 would be approximately $98,000.

In addition, OHTA wrote to medical/surgical specialty societies and to every United States medical center that performed pancreas transplants to provide an opportunity for their participation in the assessment process; in that letter (see Appendix A), information on costs of SPK and PAK were sought. In response, the University of Rochester stated that they receive from private insurers a total payment for SPK of approximately $73,000; this complies $33,000 for the renal transplant, an "add-on" payment of $32,000 for pancreas transplant, and "professional [surgeon] fees" of approximately $8,000. The University of Minnesota stated that the "overall average charge" for SPK "from time of approval...to 1-year posttransplant" was $188,722, and that the comparable charge for KTA was $143,384. These charge data were based on 43 patients provided SPK and KTA from 1990-1991. It was not specified what specific services or procedures were covered in addition to the transplant itself. The University of Nebraska stated that the "hospital charges for the [SPK] transplantation average $110,000," and that the pretransplant evaluation charges average $10,000; these figures do not include physician or surgeon fees. The American College of Surgeons forwarded information from the Ohio Solid Organ Transplantation Consortium; that organization stated that for 207 SPK and PTA performed between 1990 and 1993 the average charges for transplant "from admission to discharge" were $76,032, with average first year readmission charges for 116 patients of $37,169 (physician/surgeon fees excluded). The Mayo Clinic stated that pancreas transplant "costs" were approximately $100,000 for the transplant and first hospital stay and averaged $150,000 for "all charges in the first year;" the latter figure includes all physician and surgeon fees. No other medical centers provided cost or charge data.

Sollinger et al(15). stated that the charges for SPK (including professional fees) were analyzed for "a subset of SPK patients" transplanted in "fiscal year 1991 to 1992." The "average first admission charges" for SPK were $67,694 and were $41,791 for KTA. It should be noted, however, that the actual total charges were likely to have been considerably higher; 91 percent of these patients (181/200) required an average of 3.3 readmissions each after the SPK transplant procedure. This information is summarized in Table 8.

Table 8. Charge and payment data.

Table

Table 8. Charge and payment data.

In sum, reported charges for the SPK transplant hospitalization alone ranged from $68,000 -$110,000 (not including professional fees). Total charge for SPK transplant and 1 years of followup care range from approximately $97,000 - $189,000.

Therefore, data available to OHTA indicated that the range of payment level for KTA were reported as $41,000 (University of Rochester) to $77,000(HIAA). The most reasonable estimate of 1993 Medicare payment levels for KTA is $50,000, and total 1-year payment for all care (including immunosuppressive drugs) in the year after transplant was approximately $80,000 in 1990 and would be expected to have been $98,000 in 1993. Charges for KTA were reported by only one institution and averaged $143,384 in 1990-1991 (University of Minnesota).

However, the use of charges may be seriously flawed. Saywell analyzed hospital costs for heart transplant at Methodist Hospital, Indiana. From 1984-1987, there was a progressive decrease in hospital costs from $57,638 to $28,526 per case; however, in 1989, the average charges for heart transplant were still $57,000.(100).

The information obtained was thus quite heterogeneous with respect to charges and payments made for SPK/PAK and KTA. This was not unexpected; Evans noted that "there is extreme variability in transplantation procedure charges, both within and across transplantation centers."(99). As described above, SPK and PAK procedures are associated with significantly greater morbidity than are KTA, and rehospitalizations are more frequent after SPK/PAK than subsequent to KTA. Data cited in this assessment indicated an average of about three readmissions after SPK/PAK; these generally lasted between 7 and 11 days. Although these add to the overall cost, it is clear that not all charges, costs, or payments cited above include such rehospitalization. Moreover, summary Medicare data for rehospitalization due to complications of the KTA are not separately recoverable; for that reason, the Medicare program calculates total 1-year payments which in sum may significantly exceed the true cost of KTA.

For purposes of the CEA model used in this assessment, a range of "costs" of SPK were selected as follows:

  1. The most reliable data for average payments made by private insurers in 1993 to transplant centers was obtained from HIAA: average, $153,000.
  2. The Ohio Solid Organ Transplantation Consortium data represents the lower range of 1-year total charges: $97,000.
  3. Payment levels reported by the University of Rochester represent the lowest payment for SPK transplant hospitalization alone: $73,000.
  4. The highest total 1-year charges were reported by the University of Minnesota: $189,000 in 1990-1991.
  5. The total 1-year charges reported by the Mayo Clinic were chosen to represent a middle ground between the Ohio and Minnesota data: $150,000.

Comparative data regarding the cost of KTA were selected as follows:

  1. The average payments made by private insurers in 1993 as reported by HIAA were selected because 15-20 percent of all KTA are covered by private payers and not Medicare: $77,000.
  2. Medicare payments for kidney transplant hospitalization alone: $43,000 in 1989, and $50,000 in 1993.
  3. Medicare total 1-year payments: $80,000 in 1990, and $98,000 in 1993.

Because there are few data addressing the relative effectiveness of SPK/PAK vs. KTA, a paradign was constructed to provide a basis for CEA. An hypothetic comparison of 100 patients provided SPK and 100 provided KTA and maintained on insulin treatment was devised. The model used this comparison so as to provide the most favorable parameters for combined transplant; SPK has a higher pancreas graft survival rate than does PAK, and the latter has the additional economic disadvantage of the cost of an additional transplant procedure. Moreover, SPK represent over 80 percent of all pancreas transplants, and PAK less than 9 percent.(28) The following conditions, based upon the data reviewed in detail above, were assumed:

  1. There are no differences between SPK and KTA in patient life expectancy or changes in secondary complications of diabetes and their sequelae.
  2. Actuarial estimates of pancreas graft survival for 1, 2, and 3 years after SPK as reported by Sutherland et al (e.i., 75, 71, and 64 percent) were assumed to be accurate.
  3. The expected rate of graft failure was calculated at 6-month intervals and assumed to have been present for that entire period (e.g., 12.5 percent graft failure at 0-6 months, 25 percent at 7-12 months, 27 percent at 13-18 months, 29 percent at 19-24 months, 32.5 percent at 25-30 months, and 36 percent at 31-36 months.
  4. Pancreatic graft failures incur the additional cost of treatment of hypo-or hyperglycemia (H/HG) when graft failure occurs.
  5. No retransplants were performed when pancreas grafts failed.
  6. No deaths or renal graft failures occur during the hypothetical 3-year observation period posttransplant.
  7. Quality-of-life utility preferences for SPK and KTA were estimated on the following bases:
    1. A weight of .84 has been reported for KTA (vide supra); however, diabetic renal transplant cases report QOL as somewhat lower than that for nondiabetic recipients.
    2. The maximum preference weight that recipients of successful SPK could conceivably achieve would be .95, i.e., virtually perfect health.
    3. Assumed QALY preference weights for SPK of .95 and .90 provide moderately optimistic estimates of improved QOL due to SPK.
    4. Assumed preference weights of .7, .75, and .8 for KTA are moderately pessimistic (i.e., lowest weight is 17 percent lower than the average patient rating for KTA reported by Torrance and Feeny(98).

Charge ad payment levels for SPK and KTA were chosen as described previously. Kidney transplant alone requires expenditures for continued insulin therapy (i.e., the cost of insulin, syringes, alcohol skin wipes, glucometer and glucometer test strips, etc.) and the more significant additional costs necessitated by any episodes of H/HG. In any CEA model, costs of this treatment of H/ HG in KTA patients must be compared with costs of SPK/PAK over the projected duration of pancreas graft survival. The American Diabetes Association (ADA) informed OHTA that there are no reliable estimates of the average annual cost for treatment of H/HG in insulin-dependent diabetics. The ADA has, however, estimated the total direct and indirect costs of diabetes (Direct and Indirect Costs of Diabetes in the United States in 1992, American Diabetes Association, Alexandria, VA). Indirect costs comprise the present value of future productivity losses caused by the morbidity and morality of the disease. These costs are not relevant to the CEA model, inasmuch as we are limited to 3-year followup data for pancreas graft survival. The ADA estimate of direct costs includes those for physicians and hospital outpatient services, equipment, emergency room care, hospitalizations, pharmaceuticals, and laboratory tests. These costs were reported for all (7.2 million) diabetics; type I cases were not separately reported. The mean annual direct costs per patient were $6,280. Even if the costs for type I diabetics significantly differed from the overall mean, it is very unlikely that they would be outside the range of the annual costs of treating HG that were used in this model.

Moreover, the monograph indicated that the most significant component of costs of diabetes is that due to complications of the disease; for example, 47 percent of the 1987 United States total hospital costs for diabetic complications resulted from heart disease. It has been reported that both type I and type II diabetics exhibit similar complications and that the overall costs of treatment may be comparable.(101). Therefore, the mean annual direct costs for type I diabetics may not substantively exceed the ADA estimate of $6,000-$7,000. Estimated costs of the insulin therapy components were based on published data that reported the range of insulin dosage of pancreas transplant recipients(11). as 12-200 units per day. These costs are outlined in Table 9. The cost for insulin therapy alone then ranges from several hundred to several thousand dollars per year; added to this are the costs for treating H/HG. In fact, the latter expenses are far more significant to CEA. Depending upon the severity of the diabetes and effectiveness of insulin therapy, these would include costs of additional physician visits, emergency room admissions, or hospitalizations as necessitated by the individual patient's clinical condition. These would be much greater than the direct costs of insulin therapy. For purposes of this model, annual costs of treating H/HG (i.e., insulin therapy plus additional care as described previously) were assumed to range from $5,000-$40,000 annually, as described below.

Table 9. Costs of insulin therapy (excluding costs of laboratory testing).

Table

Table 9. Costs of insulin therapy (excluding costs of laboratory testing).

By adding the costs of SPK and the costs of treating H/HG for those proportions of SPK recipients whose grafts fail at the various time intervals, one can estimate a 3-year total cost for SPK. Likewise, the cost of KTA plus the same estimated costs of treatment of H/HG over the hypothetical 3-year period provide a comparable expected cost for KTA. An example of such calculations is seen in Table 10.

Table 10. Total costs for 100 transplant recipients[a].

Table

Table 10. Total costs for 100 transplant recipients[a].

This process was repeated for estimated annual costs or treating H/HG ranging from $5,000-40,000/year in increments of $5,000.

Likewise, taking estimates of QALY preference weight for SPK assumed above (.95,. 9) and multiplying them by the proportion of patients with surviving grafts at the appropriate time intervals, and adding that number to the proportion with failed grafts times the utilities for KTA (.7, .75, .8) will provide a summary of QALY at those preference weights, as seen in Table 11.

Table 11. Total QALY for 100 SPK recipients.

Table

Table 11. Total QALY for 100 SPK recipients.

This process was repeated for all preference weights assigned to SPK and to KTA. The assumption that QALY preference weights after failure of a pancreas graft would be no lower than the weights of KTA tends to favor SPK. The calculations were performed on the Excel Spreadsheet program.

This model is not without flaws. Pancreas retransplant following SPK is not uncommon in some centers; epidemiologic data indicated that the assumptions of no patient deaths or renal graft failures during the hypothetic 3-year observation period are likely to be wrong. However, these assumptions are either neutral in the CEA model or favor SPK. In addition, QALY preference weights and annual costs of diabetes treatment are unlikely to be independent variables; i. e., patients whose diabetes is so severe as to occasion large annual costs for treatment of blood glucose fluctuations are also likely to have a relatively lower QOL. However, this fact would not alter the results of the model. Although there are few objective data to support the specific preference weights assigned to SPK, they were selected so as to favor SPK over KTA. No consideration was given to expenses or benefits beyond the 3 years after transplant; for PAK after LRD kidney transplant, the increase in renal graft survival over that expected after cadaver transplant could conceivably improve its cost effectiveness beyond the temporal limits of this model. However, assumptions beyond 3 years restricted by the lack of data addressing SPK and PAK outcomes beyond that time period. The calculations of cost and QALY for each 6-months period were based on an assumption that all graft failures were present for the entire interval; this is not likely to be the case. To assess the effect of this assumption, several sets of calculations were performed assigning all failures to the midpoint of each time interval; the resulting differences in outcome (i.e., cost/ QALY) of the model were negligible and would no have sufficed to alter the findings.

Results

Some components of this model used costs, charges, or payments limited to the admission for the transplant procedure alone, whereas others included transplant and variable-related care for the year after SPK or KTA. Data previously cited regarding readmission rates indicated that the total cost of SPK may considerably exceed transplant costs; e.g., Medicare total 1-year payments for KTA are nearly twice the payments for transplant alone, and SPK routinely generates greater readmissions than KTA. This distinction must be kept in mind in considering results of these analyses.

Figures 4-6 are graphic representations of the relative cost effectiveness of SPK vs. for transplant hospitalization alone at the payment levels for SPK and KTA reported by private insures (HIAA), and for the 1993 Medicare payment level for KTA, under various QALY preference weights. Each figure includes the calculated cost per QALY of SPK or KTA across differing levels of reimbursement and two QALY preference weights. It is apparent that as the annual cost of treating H/HG increases, SPK becomes more attractive on the basis of cost effectiveness. Increasing the disparity in preference weights between SPK and KTA also operates to the advantage of SPK.

Figure 4. Cost-effectiveness analysis model: SPK/KTA transplant only.

Figure

Figure 4. Cost-effectiveness analysis model: SPK/KTA transplant only.

Figure 5. Cost-effectiveness analysis model: SPK/KTA transplant only.

Figure

Figure 5. Cost-effectiveness analysis model: SPK/KTA transplant only.

Figure 6. Cost-effectiveness analysis model: SPK/KTA transplant only.

Figure

Figure 6. Cost-effectiveness analysis model: SPK/KTA transplant only.

Figure 4 (above) is based on relative preference weights of .95 for SPK and .7 for KTA. It is doubtful that SPK would result in that favorable a weight, as it would imply virtually perfect health, a state unlikely to be attained as a result of that procedure. Likewise, a weight as low as .7 would uncommonly be associated with KTA, because it is 17 percent lower than the mean preference weight chosen by renal transplant recipients. Nevertheless, at payment levels of $77,000 and $50,000 for KTA, SPK becomes equally cost effective only when the annual expenses of treating H/HG are $17,000 and $29,000, respectively.

Reference to Figure 5 demonstrate that for preference weights of .98 (SPK) and .75 (KTA), cost-effective equivalence of SPK and KTA for the two KTA payment rates occurs at annual H/HG expenses of $24,000 and $36,000. This scenario also seems improbable; although SPK may conceivably provide that high a QOL, it is less likely that many diabetes with successful kidney transplant would rate their state of health as 11 percent lower than the mean preference weight selected by renal transplant recipients as cited above.

The most plausible of the assumptions may be the preference weights of .9 for SPK and .8 for KTA; i.e., SPK patients perceive their posttransplant state to represent a 7 percent improvement in QOL over what would have been achieved with KTA, and diabetic kidney recipients rate their health state as 5 percent lower than the preference weights chosen by KTA recipients and only 25 percent superior to dialysis. In this scenario, SPK is as cost effective as KTA only when annual costs of treating H/HG are $28,000 (for KTA payment of $77,000) and $40,000 (for KTA payment of $50,000).

Figures 7-9 use data supplied by the Mayo Clinic, which were stated to represent total costs for transplant and continued care necessitated by the transplant during the subsequent year. Comparison was made with Medicare projected 1993 total 1-year payment levels because, although the Mayo Clinic did not specify the time frame of their cost data, it was assumed to have been contemporary information. For the three QALY preference weight ratios used in this model, cost-effective equivalence of SPK occurred at annual H/HG costs of approximately $9,000, $14,000, and $18,000.

Figure 7. Cost-effectiveness analysis model: SPK/KTA total 1-year coasts.

Figure

Figure 7. Cost-effectiveness analysis model: SPK/KTA total 1-year coasts.

Figure 8. Cost-effectiveness analysis model: SPK/KTA total 1-year coasts.

Figure

Figure 8. Cost-effectiveness analysis model: SPK/KTA total 1-year coasts.

Figure 9. Cost-effectiveness analysis model: SPK/KTA total 1-year coasts.

Figure

Figure 9. Cost-effectiveness analysis model: SPK/KTA total 1-year coasts.

Figure 10 illustrates the results of application of this model for SPK/KTA preference weights of .9/.8, using charge data supplied by the University of Minnesota. Because these data were based on transplants performed from 1990-1991, the comparison data used for KTA were Medicare payment levels reported for 1990. Both charges and payment levels are for transplant plus 1 year of followup care. The Medicare data include cost of immunosuppressive drugs and all care provided during the year, even if unrelated to transplant; this comparison is therefore likely to provide a somewhat more optimistic view of the relative cost of effectiveness of SPK than is actually the case. Comparison of SPK charges was not made to the KTA charges reported by Minnesota (i.e., approximately $143,000). There were several reasons for this: Medicare pays for great majority of KTA performed in the United States, and the 1990 payment data are accurate and reliable; moreover, because this assessment was initiated at the request of HCFA, CEAs are more appropriately directed at actual Medicare reimbursement levels and not at charge data, which exceed Medicare payments by approximately 79 percent. Although it may be argued that Medicare payment levels undervalue KTA, that issue is beyond the purview of this document. At a QALY preference ratio (SPK/KTA) of .9/.8, cost-effective equivalence of SPK and KTA occur at annual costs of treatment of H/HG of approximately $41,000.

Figure 10. Cost-effectiveness analysis model: SPK/KTA total 1-year costs.

Figure

Figure 10. Cost-effectiveness analysis model: SPK/KTA total 1-year costs.

Table 12 contains a summary of the results of the application of this model to CEA, and presents the cost-effectiveness equivalence points of SPK and KTA based upon the sensitivity analyses described previously. It is clear that for the most probable QALY weights and the cost/charge/ payment levels available for SPK and cited above, SPK is unlikely to be as cost effective as KTA unless the cost of treating H/HG is above $15,000 annually. The Ohio Consortium charge data might argue for cost effectiveness of SPK, since they reported that charges for transplant plus readmissions in the 1990 to 1993 time interval were $97,000. However, both charges and payments provided by other sources were considerably higher. The payment data fom the University of Rochester covered only the transplant itself, but total annual costs are unlikely to exceed those specified by the Ohio Consortium.

Table 12. Summary of CEA model.

Table

Table 12. Summary of CEA model.

For purposes of comparison, the model was also applied to PAK, but was limited to evaluation of cost effectiveness for QALY preference weights of .9 for PAK and .75 for KTA. It was feelt that little additional information would be gained by running the model through all of the preference weights used to compare SPK and KTA. Because no data were acquired regarding the cost of pancreas transplant alone, the total cost of PAK was assumed to be equivalent to that of the simultaneous procedure, and the HIAA SPK payment level of $153,000 was chosen. It was further assumed that the payment effectively occurred at the same moment in time, viz, at initiation of the model. That is, the time interval between kidney transplant and PAK was assumed to have been short enough that discounting of costs would have negligible effects upon the outcome. These assumptions greatly favor the cost-effectiveness of PAK, since it is unlikely that a sequential procedure involving two separate transplants and hospitals admissions could be achieved at a cost equivalent to a combined procedure accomplished in one admission. Nevertheless, this analysis revealed that PAK was not as cost effective as KTA even at annual costs of $40,000 for treatment of H/ HG (Figure 11). The results indicated that at least over a 3-year period after pancreas transplant, and under the most optimistic cost assumptions, PAK is at a significant economic disadvantage when compared with STA and SPK.

Figure 11. Cost-effectiveness analysis model: SPK/KTA total 1-year costs.

Figure

Figure 11. Cost-effectiveness analysis model: SPK/KTA total 1-year costs.

These sensitivity analyses, which systematically varied estimates or assumptions of costs and measures of effectiveness over a range of values, were used to test the robustness of conclusions of cost effectiveness with respect to those assumptions and thus the subsequent confidence that may be placed in the results of the analyses. Variation of the costs of KTA from $50,000 to $98,000 and the assignment of preference weights of .95 and SPK and .7, .75, and .8 for KTA, did not significantly alter the inferences that could be drawn from the application of this model. In addition, the model made no attempt to consider the actual or potential economic effect of out-of-pocket payments for SPK, the opportunity costs associated with higher readmission rates, lost work or school time, travel costs, or financial impact upon family members, which may further diminish the cost effectiveness of SPK. Although for some of the scenarios evaluated by this model, the differences in cost effectiveness (favoring KTA) were not vast, the magnitude of the observed differences were dependent on the assumptions that QALY weights approximating near perfect health result from combined transplant, and that weights after KTA are inferior to those reported in the general population of kidney recipients. To date, objective data have not verified those assumptions.

It appears that for the majority of the cost/charge data analyzed, SPK may be an economically defensible procedure if it provided improvements in QOL comparable to thee estimated preference weights when the procedure was applied in severe diabetics; i.e., those with the highest costs for treating H/ HG. However, as was discussed previously, the published case series of SPK and PAK provide little evidence that is the patient population in whom the procedure has been used. Indeed, as previously noted, several transplant centers informed OHTA that the severity of the diabetes and/or the intensity of the pretransplant treatment regimen were not considered as important selection criteria for o[ancreatic transplant (vide infra). This is problematic, since this model was constructed to provide a rather optimistic evaluation of SPK which itself is highly likely to exhibit an economic advantage over PAK, in light of SPK's superior pancreas graft survival and PAK's inherent handicap in requiring two separate transplant procedures. In addition, pancreas retransplantation would place both SPK and PAK at an addition economic disadvantage. Retransplantation rates may be fairly high in some centers; on review indicated that 10 percent of SPK (12/104) and 38 percent of PAK (21/35) were transplants. Moreover, as noted above, the IPTR reported 1-year graft survival rates for primary vs. retransplantation cases of 76 and 62 percent for SPK, and 51 and 39 percent for PAK,

There may be additional costs not addressed in our model. Patients have been required in some instances to pay additional fees over and above any third party reimbursements (personal communication, NIH, February 10, 1994). These fees may be $25,000 or more, and such additional out-of-pocket expenses could affect the cost effectiveness of combined pancreas-kidney transplantation. Moreover, such payments raise disturbing ethical issues; all transplant programs benefit from significant governmental funding based on the premise that the organ procurement, distribution, and transplant systems are a public good. Although organ donation is promoted and generally viewed as a contribution to society at large, the altruistic motivation may be adversely affected by the perception of potential donors that members of middle and lower socioeconomic strata are less likely to be selected as recipients based on their inability to bear high-of-pocket expenses.

The results of this CEA model must be considered in light of the fact that the assumption underlying the model significantly favor SPK. For example, data from the IPTR (vide supra) indicated that there was a 12 percent "technical failure rate" for all United States pancreas transplants (85 percent of which were SPK). The model assumed that all transplants were successful. Moreover, the Medicare total 1-year payments for KTA include those for all care (for any cause) which were provided in the year after transplant, as well as payments for all immunosuppressive drugs. This appears unlikely to be true for the charge data for SPK. In addition, no differences in renal graft survival were assumed to exist between SPK and KTA; reference to Figure 2 indicates that renal graft survival may well be greater for KTA-LRD and HLA-matched CAD/KT. Finally, the QOL preference weights were clearly optimistic for SPK and relatively pessimistic for KTA, even though no objective data had demonstrated those magnitudes of benefit.

In sum, within the limits of the available 3-year actuarial pancreas graft survival data, SPK and PAK may be comparable to KTA on the basis of cost effectiveness when provided to complicated diabetics who have incurred large annual expenses for care secondary to H/HG, and when a significant improvement in QOL consequent to pancreas transplant has been clearly demonstrated. SPK or PAK in different circumstances could conceivably be justifiable on other grounds but would not be cost effective.

Information Provided by Other Sources

As part of this process, notice of the assessment was published in the Federal Register. Information was also solicited from other agencies of the Public Health Service. In addition, letters inviting participation in the assessment process and soliciting information on SPK and PAK were sent to every transplant center performing SPK or PAK (see Appendix A).

The NIH (personal communication, March 26, 1993) provided an extensive bibliography of reports related to SPK and PAK. They informed OHTA of their opinion that both SPK and PAK have been successfully used. They further stated that the most systematic data on costs were contained in the National Cooperative Transplantation Study, and cited data indicating the average cost for combined transplant was $85,000-$90,000, with other followup costs of $10,000-$15,000 in the first year; comparable first year costs for KTA were about $54,000. NIH believed that the costs of the combined surgery would be exceeded by 3 or 4 years of costs for dialysis. The annual cost to the Nation of the kidney disease of diabetes mellitus was estimated as $2 billion, and successful use of combined pancreas-kidney transplant could prove highly cost effective if it reduced that annual expense. The most common cause of ESRD is kidney disease due to diabetes.

NIH-sponsored research gives special emphasis to health problems of minority patients, and is also directed at new methods for improved survival and reduced immunogenicity of pancreas transplants and of pancreatic islet cell tissue. NIH is currently supporting studies at the Universities of Minnesota, Iowa, Chicago, Texas, Kentucky, and Temple University.

NIH provided the following information on patient and graft survival: 4-year patient survival for SPK was given as 65 percent; PAK was 77 percent. Four-year pancreas graft survivals were reported as 40 percent for SPK and 35 percent for PAK. The NIH opinion was that QOL and metabolic parameters were better over the long term after combined transplant than following KTA. It was further stated that both SPK and PAK "are now considered safe and effective treatments for diabetic nephropathy," and that "in comparison to the costs of regular dialysis, this surgical treatment could prove to be very cost effective."

FDA stated (personal communication, Shaffer D, February 17, 1993) that they regulate some adjuncts to transplantation such as perfusion/cold storage solutions and immunosuppressive agents. However, FDA does not regulate the actual transplantation of organs and consequently had no pertinent information or statistics pertinent to the assessment.

The Division of Organ Transplantation, Health Resources and Services Administration (HRSA), provided no institutional review of SPK or PAK, but provided data from the registry of the United Network for Organ Sharing regarding SPK/PAK outcomes (this material was presented in the body of the assessment document).

The American College of Surgeons (personal communication, Stombler R, May 27, 1993) did not provide an evaluation of SPK or PAK, but forwarded material from the Ohio Solid Organ Transplantation Consortium. That information has been cited elsewhere in this assessment.

The American Society of Transplant Surgeons had corresponded with OHTA before the initiation of this assessment (May 6, 1991). They stated that they believed Medicare should reexamine its policy with respect to pancreas transplantation. The Society provided an undated "White Paper" entitled "Medicare Coverage of Pancreas Transplantation." This document stated that pancreas transplantation is most frequently used in patients with diabetic nephropathy necessitating a kidney transplant "since such patients are already obligated to immunosuppresion by lieu of the kidney transplant...." They stated that successful pancreas transplant in diabetics who also receive a kidney transplant has been shown to: prevent recurrence of diabetic nephropathy in the transplanted kidney....slow or stabilize other secondary complications of diabetes....provide insulin independence for the patient, obviating diabetic control problems, and improving quality of life." The Society believed that recipients will be afforded "provision of a higher quality of life and protection against secondary complications." They recommended that the Medicare coverage policy be amended to include SPK or sequential kidney-pancreas transplantation; they indicated that the "significant" additional hospital costs were "$10,000-$25,000 on average."

The University of Cincinatti (personal communication, Alexander J, Munda R, April 6, 1993) stated that the duration of hospitalization after SPK or PAK averaged 21 days, and that the number of readmissions in the first year was three, with hospital stay averaging 11 days for such readmissions. Patients were followed for at least 1-year posttransplant. Measures of severity of diabetes evaluated pretransplant were glycohemoglobin levels, documented episodes of H/ HG, and/or hospitalizations due to such occurrences. Other criteria include freedom from or minimal neuropathy, peripheral vascular disease, and retinopathy. Contraindications include alcoholism; chemical dependency; active infection; active ulcer disease; uncontrolled malignancy;or lack of compliance. Their pretransplant selection protocol includes EKG; noninvasive vascular assessment of lower extremities and carotid vessels; MUGA stress test of coronary angiogram, if indicated. The most common preoperative clinical condition that needs correction is cholelithiasis. The University of Cincinatti performs an average of 15 kidney-pancreas transplant per year. They reported patient survival of 95 percent, pancreas graft survival of 90 percent, and kidney graft survival of 90 percent but did not specify the duration of followup associated those figures.

The Washington Hospital Center (personal communication, Sasaki T, January 28, 1993) stated that they happen been performing SPK since 1989. Forty transplants were performed, 22 of which were accomplished in 1992. They reported patient survival of 91.5 percent, kidney survival of 85.1 percent, and pancreas survival of 85.1 percent; the duration of followup associated with these survival rates was not specified. They stated that "the quality life is also the best for simultaneous pancreas-kidney transplant when compared to kidney alone in type I diabetics." The Center also stated that although they have not analyzed the data regarding secondary complications, it is their "strong impressions that these patients do better in terms of neuropathy and limb ischemia."

New England Deaconess Hospital reported (personal communication, Shaffer D, December 29, 1992) that they have performed 35 SPK and one PAK since 1988. Patient and graft survival rates were comparable to those of cadaver kidney transplant alone. They believe that patients generally feel their overall QQL was improved. They had no data regarding the effect of functioning pancreas grafts on neuropathy or neuropathy. They recommended PAK in patients with a related kindney donor are between the ages of 18 and 55 years, have established diabetic nephropathy, and are free of "significant cardiovascular disease," substance abuse, "major psychiatric illness," or a history of active infection or cancer (except nonmelanoma skin cancer) unless free of disease for more than 5 years.

The Massachusetts General Hospital (personal communication, February 17, 1993) provided two published articles from their institution which they believed appropriate to the questions posed in OHTA's correspondence.

The University of Rochester (personal communications, Reed A, March 1, 1993, and November 19, 1993) reported they had performed two SPK. They also stated that the hospital was reimbursed $3,9000 for the kidney portion of the transplant, and an "incremental payment from som insurance companies of approximately $32,000 to cover additional costs to the hospital for the pancreas." They also provided data submitted to Blue Cross/Blue Shield indicating their current rates: kidney transplant = $33,510, pancreas add-on = $31,516, professional (surgeon) fees = $7,830, totalling $72,856. The correspondence stated that the age range preferred is 20-55 years, but patients "in exceptionally good health" over 55 may be considered. Pretransplant testing "includes EKG, stress thallium, and coronary angiography," all usually performed at referring institutions. They prefer at least one HLA-DR antigen match with the donor, but that is not a prerequisite. The University of Rochester stated that with respect to patient selection "severity of the patient's diabetes or the type of insulin regimen employed is of no particular concern."

The University of Nebraska (personal communication, Stratta R, February 25, 1993) stated that they had completed 105 pancreas transplants in 102 82 were SPK, eight were PAK, 20 were solitary pancreas transplants, and three were combined liver-pancreas transplants. The mean duration of pretransplant diabetes was 23.7 years. The mean daily insulin dose was 42 units (range: 12-200), and the mean pretransplant glycohemoglobin level was 9.6 percent. At the time of SPK transplant, 44 patients were not on dialysis and 38 were dialysis-dependent. The mean HLA-A, B, C match was .76, and the mean HLA-DR match was .46. At the time of correspondence, 80 patients were alive; 76 of those had functioning pancreas grafts. Mean 1-year pancreas graft survival was given as 92 percent ("58 of 63" cases). The mean initial hospital stay was 20 days; 120 readmissions occurred in 61 patients. Twenty-one patients had no readmissions, 29 had one readmission, 15 had two readmissions, and 17 had three or more readmissions in the first 3 months. Causes for readmission included "rejection (45); infection (25); pancreas-specific morbidity such as dehydration, hematuria, or pancreatitis (43); and miscellaneous causes (7)." The mean length of readmission was 7 days, and the mean total hospitalization in the first 3 months posttransplant was 30 days. The University reported that hospital charges "for the transplant hospitalization average $110,000," and "total charges for the evaluation average $10,000."

The University of Minnesota (personal communication, Sutherland D, February 25, 1993) provided a short bibliography. In addition, they stated that the mean durations of hospitalization for SPK and KTA were 26 and 12.7 days, respectively; the mean number of readmissions "over the entire period of followup" were 3.0 for SPK and 2.6 for KTA, and "during the first year" were 2.6 (SPK) and 1.4 (KTA). They stated that they consider "all uremic diabetic patients who are candidates for a kidney transplant as candidates for a pancreas transplant," and that exclusion criteria such as severe "peripheral vascular disease" are used rather than inclusion criteria. Inclusion criteria are used "only for pancreas transplants alone." Because KTA recipients will be immunosuppressed "unless there is a reason not to make them insulin independent as well as dialysis free, we will do so with the addition of a pancreas." They stated that they encourage all patients to undergo LRD kidney transplant if possible, with PAK being subsequently used; CAD/KT is not, however, a contraindication to PAK. Measures of severity of pretransplant diabetes are considered an indication only for isolated pancreas transplantation, but "lability is not a necessary criterion for adding a pancreas to patients who are obligated to immunosuppression" because "such patients have already demonstrated they have inadequate control to prevent development of secondary complications. " The University of Minnesota believed "the addition of a pancreas will prevent recurrence of diabetic nephropathy" and that "their quality of life is improved by making them insulin independent...." They did not believe that glycohemoglobin was an adequate measure of diabetic lability, but that elevated glycohemoglobin "coupled with frequent hypoglycemic episodes" indicate lability is almost certainly present, although "this criteria is mostly for pancreas transplant alone patients. " The pretransplant diabetes treatment regimen does not influence their selection or exclusion of patients; age limitations are less important than specific problems such as advanced vascular disease; comorbid conditions influence selection, although correctable coronary artery disease would be treated and the team "proceed with the pancreas transplant." The "main focus" of the pretransplant protocol workup is cardiac evaluation. Symptomatic patients undergo coronary angiography; asymptomatic patients undergo stress thallium testing and angiography for abnormal results. Some asymptomatic patients with a normal stress thallium test will undergo angiography "at the discretion of the cardiologist." All candidates undergo an ultrasound evaluation of the gallbladder. The University stated that they do not use tissue matching as a specific criteria for SPK transplants. The "total charges" (including "evaluation, pretransplant maintenance care, transplantation, organ acquisition charges, and posttransplant care from time of approval. ..to 1-year posttransplant" reflecting "professional fees as well") were $188,722 for SPK and $143,384 for KTA. These charge data were based on 21 KTA and 22 SPK performed from July 1990 to June 1991.

The Mayo Clinic (personal communication, Munn S, April 26, 1993) informed OHTA that for SPK, PAK, and isolated pancreas transplant, the mean duration of hospitalization was 22 days, number of readmissions was four, and mean total inpatient stay for readmissions was 40 days. The mean glycohemoglobin level pretransplant was 10.5 percent, and documented episodes of hypoglycemia have occurred in nearly every transplant patient. The mean daily insulin dose in transplant recipients was 36 units. Simultaneous pancreas-kidney and PAK are restricted to patients under the age of 46 years, due to the higher mortality rate associated with age. Virtually all SPK/PAK recipients have advanced retinopathy and neuropathy. No efforts are made to match tissue types for SPK "because of the limited number of combined organs available," although they "accept no more than two mismatches at the HLA loci for pancreas transplantation either done alone or subsequent to a previous kidney transplant." The Mayo Clinic stated that the "additional cost for a pancreas when combined with a kidney transplant is approximately $3,000; when the pancreas transplant is performed on its own, the cost is somewhere between $75,000 and $100,000 for the surgery and first hospital stay, and between $125,000 and $175,000 for all charges in the first year." The Mayo Clinic also provided a detailed pancreas transplant protocol book. The protocol provides some criteria for patient selection criteria. The Clinic believes "there is a subgroup of diabetic patients in whom the benefits of intervention may outweigh the risks." They further state that "there is hope, partially supported by the literature, that successful pancreas transplantation may prevent the secondary complications. ..and improve the recipient's quality of life." Age limitations were given as 1-50 (<45 years for males); type I diabetes; blood pressure <140.85; no clinical evidence of peripheral vascular disease; adequate peripheral arterial supply as determined by Doppler examination and plethysmography; not requiring narcotics or large amounts of analgesic medication; no evidence of recent retinal hemorrhage; no history of myocardial infarction; no active neuropathic ulcer; insulin requirements less than 2 units/kg/day; no previous renal transplant recipients with panel reactivity of >80 percent or early (<1 mo) graft loss; no other contraindications including previous multiple operations; ability to understand risks and benefits, cooperation, and compliance with the medical program. Required "laboratory test" results include: satisfactory cardiac function (ejection fraction > 50 percent), normal exercise thallium ventriculogram (angiography reserved for remaining questions of coronary disease); if preuremic or had prior renal transplant, GFR (iothalamate) of 60 mL/min/1.7 m(2) or 50 mL/min/m(2) if on cyclosporine; normal amylase and lipase; calcium less than 10.2 mg/ dL; if prior renal transplant, Hb=8-10g; no hyperchylomicronemia. As regards financial consideration, the protocol manual stated that about 50 percent of Blue Cross/Blue Shield plans provide benefits for the procedures, and it was estimated that 80 percent of other group insurance plans so cover. Patients are expected to guarantee hospital and medical charges not covered by third-party payers. If necessary, the Clinic will assist potential recipients in fund raising activities if they do not have funds "to guarantee the personal responsibility." A limited number of patients who are area residents and are without required funding may be selected by the Pancreas Transplant Selection Committee. For Medicare patients, additional payment may be requested, because no coverage is available for pancreas transplant.

The American Diabetes Association (ADA) provided a prepublican Technical Review and Position Statement, considered an "endorsed" statement of the ADA and approved by the Association's Committee on Professional Practice, and by the Executive Committee. The document included the following statements regarding pancreas transplantation:

Life expectancy: "It is too early...to estimate its ultimate effect on life expectancy, and little relevant data exist.... The critical question...remains unanswered."

Graft functional survival: "UNOS data for 1,021 cases from 1987-1990 show 1-year graft survival rates of 77 percent for simultaneous pancreas-kidney transplant, 52 percent for pancreas transplants subsequent to a kidney transplant...." "Three-year graft survival rates of approximately 50 percent were reported by one center."

Metabolism: "Successful pancreas transplantation has been shown to eliminate the need for exogenous insulin.... However, in pancreas transplant recipients, insulin and glucagon responses to a glucose challenge and certain other biochemical indices may differ from those of nondiabetic patients. The long-term clinical significance of these disturbances is unclear."

Quality of Life: "Successful pancreas transplant...eliminate the need for insulin injections and frequent glucose monitoring, and enable the patient to resume a more normal lifestyle. Several studies have detailed significant improvements in recipients' quality of life."

Effect on Chronic Complications: "For several reasons, only preliminary and incomplete data exist with regard to the effect of pancreas transplantation on these complications.... Presently, there are no reports of long-term, prospective, randomized controlled studies of the effect of pancreas transplantation on the complications of diabetes. Hence, prevention or attenuation of long-term complications is not a reason for considering transplantation at the current time."

Transplantation Safety: "With current selection procedures, 1-year survival of combined pancreas-kidney transplant recipients is similar to that of kidney-only transplants. The mortality risk of pancreas transplant procedures...is...small, but not negligible."

Transplant Costs: "All organ transplants are expensive and cost is an increasingly important consideration in establishing the appropriateness of medical care procedures. Only limited, incomplete data exist on the cost of pancreas transplants. One report indicated that the average hospital charge for pancreas-only transplants in 1988 was $70,000. Another study reported the median cost in 1988 to be $67,000...."

The ADA also recommended that: "Pancreas transplants should be considered an acceptable therapeutic alternative to continued insulin therapy in patients with end-stage renal disease." However, the ADA also commented that "These patients should also have significant clinical problems with exogenous insulin therapy. ..."

"The current data on the long-term effects of transplantation on life expectancy and diabetic complications are insufficient and, per se, do not justify pancreas transplantation."

"Institutions that perform islet or pancreas transplantations should be tertiary care centers that have an active kidney transplant program...."

Pretransplant Patient Screening

Many transplant centers have established relatively detailed and complex screening procedures or protocols for patients considered for SPK or PAK. Elements of such pretransplant evaluations have variably included:

  • EKG and thallium stress tests.
  • Coronary angiography.
  • Coronary bypass grafting.
  • Ultrasound abdominal examination.
  • Prophylactic cholecystectomy.
  • Mammography for females over 35 years of age.
  • Prostate-specific antigen screening.
  • Upper gastrointestinal series or endoscopy.
  • Barium enema or endoscopy, if over 35.
  • EMG and/or nerve conduction testing.

Although such interventions are not properly part of the transplant technology per se, they are worthy of brief mention in this assessment since they increase the overall cost of SPK/PAK, engender some additional, if slight, morbidity or mortality, and provide a potential database which could address patient selection criteria.

There is little question that vascular disease, and particularly coronary artery disease, is a significant cause of mortality and morbidity in diabetic patients with ESRD. Several investigators have noted that asymptomatic myocardial ischemia is common in that patient population, and that aggressive diagnostic evaluation and intervention may be associated with reductions in cardiac mortality. Several case series of SPK/PAK addressed the alteration of pretransplant assessment to include stress testing, angiography, and prophylactic bypass grafting in an attempt to improve recipient survival. Although the hypothesis is not unreasonable, there are no comprehensive data available that clearly delineate the improvement in outcomes that have resulted from such approaches.

Although the prevalence of cholelithiasis may be increased in diabetics with ESRD, prophylactic cholecystectomy is a less tenable procedure, and there are not sufficient published data to demonstrate that the risk/benefit ratio is such that the procedure is prudent and appropriate for the pretransplant population. Lowell et all(104). retrospectively reviewed 61 SPK and 12 PTA for evidence of the development of cholelithiasis; a comparison group was selected that included 87 KTA recipients, 35 of whom had type I diabetes. In the KTA groups, the incidence of cholelithiasis in diabetics was reported as 27 percent vs. 12 percent in the nondiabetic patients. Fifty-six of the 73 pancreas recipients were "eligible for followup. " Seventeen had ultrasonographically detected gallstones, and 14 of these underwent cholecystectomy. However, only three of the 14 were symptomatic. This retrospective analysis did not demonstrate that routine pretransplant screening for cholelithiasis reduced morbidity or improved outcome of SPK. However, the performance of cholecystectomy in patients of whom nearly 80 percent were without symptoms raises serious questions as to the proper role of a procedure that will clearly increase cost and morbidity (and potential mortality) in the absence of a clearly demonstrated benefit.

Careful evaluation of available evidence by the United States Preventive Services Task Force led that body to recommend that mammography screening not be used in women under the age of 50 save for specific patients deemed to be at high risk, not including transplant recipients. Although immunosuppression subsequent to renal transplantation is associated with a higher incidence of cancers, breast cancer occurring before age 50 has not been demonstrated disproportionately to present in that population, in contrast to malignancies such as lymphoma. Moreover, PSA screening is of indeterminate efficacy in reducing morality from prostate cancer, and its relationship to posttransplant immunosuppression is uncertain. The rationale for its inclusion in screening potential SPK/ PAK recipients is not evidence-based.

It would appear that many of the screening tests used before selection for SPK/PAK are based on what is believed, at a given institution, to be the most prudent clinical judgment. Nonetheless, the quality of evidence supporting much of the pretransplant testing and resulting interventions is quite weak, even though these procedures may occasion adverse effects and be very costly in the aggregate.

More importantly, it is likely that the majority of patients selected for SPK or PAK have had considerable pretransplant evaluation. It is therefore of concern that in contrast to the quantity of data amassed, few details regarding specific patient selection criteria have appeared in the various case series of SPK or PAK.

In this regard, Ratner et al(105). have reported a retrospective analysis of 53 combined kidney-pancreas transplants performed at their institution which revealed seven cases (13 percent) had pretransplant C-peptide levels consistent with a diagnosis of non-insulin-dependent diabetes (NIDDM or type II). The various case series of SPK and PAK (vide supra) did not report preoperative C-peptide levels, raising the possibility that an unknown proportion of SPK or PAK recipients may in fact have been type II, not type I diabetics.

Discussion

The DCCT(97). indicated that intensive insulin therapy slows the onset or reduces the progression of some secondary complications of diabetes. Intensive therapy reduced the average risk of retinopathy by 76 percent in the primary prevention and 54 percent in the secondary prevention cohorts, respectively. Diabetic neuropathy appeared in 10 percent of the conventional therapy group and 3 percent of the intensively treated group at 5 years. Although the risk of albuminuria was reduced with intensive therapy, severe nephropaty was uncommon: a creatinine clearance below 70 mL/1.73 m(2) body surface area developed in only two patients of the intensive treatment group and five of the conventionally treated patients. There were no significant differences between the groups' mean scores on the trial's QOL questionnaire despite the "added demands" of intensive therapy.

The DCCT results may have relevance to the comparative clinical effectiveness and cost effectiveness of PTA vs. insulin therapy in the earlier stages of insulin-dependent diabetes. Indeed, comparison of PTA to conventional insulin therapy alone might now be considered inappropriate. However, these findings were not considered significant to this assessment of the clinical utility and cost effectiveness of SPK or PAK inasmuch as the patients eligible for the DCCT trial were dissimilar to the patients reported in the various case series of SPK-PAK. Eligibility for the DCCT trial was restricted to patients aged 13 to 39 years who had no hypertension, hypercholesterolemia, or "severe diabetic complications or medical conditions. "In the DCCT primary prevention chohort (n =726), eligibility further required a duration of diabetes of 5 years or less, no retinopathy, and albuminuria of less than 40 mg/24 hours; in the secondary prevention cohort (n =715), duration of diabetes of 1-15 years was permitted, as was "very mild to moderate retinopathy," and albuminuria of 200 mg/24 hours or less. Creatinine clearances in the primary prevention group were reported as 127-138 mL/minute, and in the secondary prevention group as 125-130 mL/minute.

The case series of SPK and PAK cited in this document comprised patients with multiple, severe complications of diabetes; these variably included severe retinopathy and/or blindness; serious and sometimes debilitating neuropathy; macrovascular disease; amputation; and in every case, ESRD requiring transplant. Moreover, a number of case series reported the pretransplant duration of diabetes, (11,,17,26,41). which ranged from 20 to 24 years (median 23 years). Such patients are not comparable to those enrolled in the DCCT, and the trial results were therefore not considered to be generalizable to the SPK/PAK transplant population.

A rational evaluation of the clinical usefulness of any procedure must begin with an unambiguous postulate of the benefit that may be expected from its application. An explicit statement of all anticipated benefits of a procedure is always critical, but its relevance necessarily increases in circumstances wherein the adverse effects are significant, the costs are high, the clinical condition in which the procedure is applied is not immediately life-threatening, and alternative treatments exist. The great majority of published reports of pancreas-kidneys transplantation have alleged that the benefits accruing from that technology are the prevention, amelioration, or reduction in the secondary complications resulting from diabetes mellitus, particularly retinopathy, neuropathy, and nephropathy. These complications may be associated with significant morbidity and some mortality and greatly increase the burden of suffering of diabetics. Although it appears that some patients may have shown limited improvement or at least stabilization of some secondary diabetic complications subsequent to SPK or PAK, careful analyses of reports which have addressed the measurements of the severity of such complications after transplant do not support the premise that pancreas transplant has been demonstrated to be often effective in that regard, nor that it is possible to predict which patients are more likely to benefit. Nearly all studies of secondary complications have serious methodologic flaws, and for each category of complication the available data conflict as to the degree or even the existence of beneficial effects consequent to pancreas transplant. The ADA(76). has stated that kidney-pancreas transplant cannot be used on the basis of expected mitigation of these secondary complications.

Alternative postulated benefits have addressed improvements in the QOL of recipients of successful pancreas-kidney transplantation. However, the data are far from clear that QOL has been generally improved. Freedom from insulin injection and frequent blood glucose monitoring is an undeniable convenience, the magnitude of which varies in proportion to the severity of the diabetic state and the effectiveness of the clinical regimen used. However, the precise relationship between these exercises required of insulin-dependent diabetics an measures of QOL has not been clearly elaborated; nor can it be concluded that patients reported in the various case series of SPK and PAK were selected on the bases of the severity of diabetes, the difficulty in managing blood sugar levels, the intensity or inconvenience of the treatment regimen, or demonstrably poorer QOL that those offered only KTA. It has not been demonstrated that objective criteria which are generally believed to be closely related to life quality, such as rates of employment or school attendance, physical activity, social interaction, or frequency of use of medical care are improved in recipients of successful SPK or PAK. Although there may be a higher level of subjective satisfaction in recipients of successful combined transplant, the methodologic problems associated with the various studies make any such conclusion highly tenuous. Consequently, the risk-benefit ratio for this class of patient has not been clearly established. If, in fact, the major benefit which may be expected from SPK/PAK is the elimination of insulin injection and blood glucose monitoring, more data of demonstrated reliability and validity may be necessary to justify a procedure of "significant morbidity and some mortality."(13). associated with charges of up to $189,000 (and potentially higher for PAK) and resulting in insulin independence in no more that approximately 35 percent to 65 percent of recipients at 3 years.

Moreover, the data available make it difficult to develop appropriate treatment strategies that relate any anticipated benefits in QOL to both the morbidity and mortality of the transplant procedure and the probability of pancreas graft survival over time. This is problematic, particularly since there is no information which relates subjective improvements in QOL resulting from SPK to the reduction in QOL (or duration) of life that may result from decreased renal graft survival if SPK is chosen in preference to LRD/KT or HLA-matched CAD/ KT. Increased rigor in the application of the appropriate psychometric and survey techniques(106). may permit future studies which provide data of increased validity and reliability.

The proportion of SPK or PAK recipients who may remain insulin independent over time poses additional problems in determining the proper role for those procedures. Given that the best long-term data are limited to actuarial 3-year pancreas graft survival rates for SPK and PAK of approximately 65 and 35 percent, respectively, and that a relatively small number of patients with functioning pancreas grafts have been followed for more than 1 year, the issue of the appropriateness of retransplantation must be addressed. Retransplant will greatly increase the individual case cost and occasion additional morbidity. The effects on QOL are indeterminate, inasmuch as no studies have delineated QOL changes in relation to retransplantation. Although the published literature provides little explicit guidance for patient selection in primary transplants, there are no data addressing those issues for retransplantation. Moreover, retransplantation has been considered a "cost-ineffective use of scarce donor organs,"(107). and such concerns need be carefully considered in evaluating the ultimate clinical utility of pancreas-kidney transplantation.

The ADA position paper(76). states that pancreas transplant should be considered as a viable treatment option "in selected cases." Unfortunately, the bases for such selection are elaborated neither in the ADA's document nor in the many published series of SPK OR PAK. Simultaneous pancreas-kidney transplant and PAK are therapeutic options which may ultimately prove to be of predictable benefit to insulin-dependent diabetics with specific clinical characteristics. However, the morbidity, expense, and potential deleterious effects on renal graft survival if SPK is chosen over LRD/KT or HLA-matched CAD/ KT currently mitigate against widespread use of these transplant procedures in the general population of insulin-dependent diabetics with ESRD. Therefore, SPK/PAK has not been clearly demonstrated to be an appropriate therapeutic option save perhaps for patients with so-called "labile" or "brittle" diabetes who have demonstrated resistance to control, and who are transplanted not only with a therapeutic intent but also with the design of prospectively acquiring data intended to clarify the actual risk-benefit ratios of SPK or PAK. The NIH has stated that the combined procedure may be safe and efficacious when used for "appropriate and specifically defined indications" such as "labile insulin-dependent patients not responding well to exogenous insulin therapy." They have also indicated that the procedure is best carried out by "institutions meeting a preset criterion of experience by multidisciplinary surgical teams" (personal communication, NIH, May 26, 1994).

Notwithstanding, further clinical studies of SPK and PAK as well as retransplant procedures are required to provide this necessary information. Although many of the published case series cited(5,22,40,80,92). in this assessment have called for prospective studies to clearly establish the benefits of pancreas transplant, none have yet been accomplished. Such studies should be encouraged in an environment of prospective trials with formal patient selection criteria accompanied by careful analyses of objective evidence of improvement in physiologic or QOL parameters. Simultaneous pancreas-kidney/PAK procedures may be best evaluated by those centers that have both a high volume of SPK/PAK and demonstrated superior results in terms of patient morbidity and mortality and pancreas graft survival.(28)

Remuzzi et al(108). recently noted that after nearly 30 years of pancreas and kidney/pancreas transplantation and "enormous expenditures," the risk-benefit profiles of the procedures have not been established by any controlled studies, and the effect of SPK upon secondary complications of diabetes is indeterminate. Although the procedure is technically feasible and "in selected patients" results in improved QOL, the risks are severe. Remuzzi et al(108). noted that there are presently "insufficient data to allow an objective judgment" as to the relative clinical utility of SPK vs. cadaver KTA. They concluded that these questions can be answered only by comparative trials of SPK and KTA.

Summary and Conclusions

Simultaneous or sequential combined pancreas-kidney transplantation may be accomplished with a relatively low mortality, but with morbidity which significantly exceeds that associated with KTA. The surgical approaches have evolved over time, and in the United States most combined transplants use cadaver organs and provide bladder drainage of exocrine pancreatic secretions with systemic (vascular) absorption of insulin secreted by the transplanted pancreas. Other techniques such as segmental pancreas transplant, pancreatic duct occlusion, pancreatic exocrine drainage, etc., are now uncommonly used. Simultaneous pancreas-kidney transplants represent the great majority of combined transplants, with PAK comprising somewhat less than 9 percent of all combined transplants in the United States. It is nearly always accomplished by transplant of single cadaver donor pancreas and kidney; the procedure provides significantly superior pancreas graft survival while avoiding the morbidity and expense of two separate transplant procedures. Pancreatic graft survival rates have slowly improved as surgical techniques and immunosuppression regimens have evolved, and they are now in the range of 35 and 65 percent at 3 years for PAK and SPK, respectively.

Virtually all studies report significant morbidity after pancreas transplant. Readmissions are frequently required for treatment of complications such as graft rejection, metabolic acidosis, hematuria, infection, and problems with wound healing. Approximately three posttransplant readmissions are required on average. Thus, compared with KTA, combined transplant has been reported to have greater adverse effects and a longer duration and greater frequency of hospitalizations.(83) Although it appears that most complications occur within the first year posttransplant, it has not yet been determined whether the more intensive immunosuppression required in combined transplants vs. that required in KTA will be associated with additional deleterious long-term effects.

The case series were heterogenous with respect to reports of pancreas graft survival. The IPTR registry survival figures were, in the aggregate, somewhat lower than than those reported in many of the various case series. This was not surprising, as OHTA has noted the same phenomenon in lung and heart-lung transplant data.(109)

Many reports stated that SPK or PAK prevented or ameliorated secondary complications of diabetes (specifically: nephropathy, retinopathy, and neuropathy). A careful review of the published data did not provide unambiguous support for those contentions. It cannot be concluded with reasonable certainty that prevention, improvement, amelioration, or mitigation of secondary complications are likely to result from SPK or PAK or provide a valid rationale for the procedures. The ADA review of pancreas transplant has arrived at the same conclusion. Although some degree of stabilization or amelioration apparently my follow SPK or PAK in individual cases, the weight of evidence is insufficient to permit a conclusion that such can be considered an expected benefit.

A number of evaluations of QOL after SPK (and a much smaller number after PAK) have recently been published. Unfortunately, virtually all of the studies suffer from significant methodologic deficiencies (vide supra). Although many of the researchers in this area stated that prospective studies were needed, none have thus far been published. There is no consistent evidence that objective measures of QOL, such as return to full-or part-time employment or school, reduction in medical care requirements, days spent sick or in hospital, or intensity of physical or social activities, etc., are improved after SPK or PAK. Improvements in subjective impressions of QOL have been more often recorded. However, not all reports have found improvement even in subjective QOL estimates.

This is not unique to combined panceas-kidney transplant. Formal attempts to evaluate QOL are relatively recent; measurement tools of acceptable reliability and validity and methodologically sound research documenting levels of QOL resulting from specific medical interventions are rare. Taking all of the available information in the aggregate, and considering the quality of the evidence available, there are perhaps suggestive but hardly unequivocal data indicating improved QOL in combined transplant recipients.

Because the vast majority of SPK use single-donor cadaver organs, the procedure cannot provide the increased renal graft survival that has been associated with LRD or an HLA-matched kidney transplant. This may not be a trivial concern: an organ recipient who chooses SPK over LRD/KT or HLA-matched CAD/KT must accept a reduction in expected kidney half-life of 40-70 percent to as much as 360 percent in return for an approximate probability of .65 of being insulin-free at 3 years. Pancreas-after-kidney transplant provides the opportunity of receiving an LRD renal transplant and the resultant superior renal graft survival. However, the recipient must undergo two surgical procedures and accept a significantly lower likelihood of long-term pancreas graft survival, approximating 35 percent at 3 years. Consequently, nearly two-thirds of PAK recipients may not remain insulin-free 3 years after their second transplant procedure. Although some institutions claim to have realized higher graft survival with PAK, that has not been the usual experience.

Without valid and reliable data as to the benefits which have been demonstrated for SPK and PAK, it was not possible to quantitate confidently the differences in clinical effectiveness between SPK/PAK and KTA. In addition, reliable nationwide data regarding the costs of combined transplant were not accessible. Therefore, a CEA model was constructed which used the available but limited charge/ payment data and assumed QALY preference weights for SPK and KTA. Sensitivity analyses were used to test the robustness of the assumptions underlying the model. Under this model, assuming that the postulated improvement in QALY preference weights which were selected to favor SPK could be validated, SPK was generally likely to be equal in cost effectiveness to KTA when the annual cost of treating H/HG in the KTA recipients was high, i.e., in the range of $15,000 per year and up. Although the model does not provide definitive information regarding cost effectiveness, the results were consistent with concerns expressed in this assessment as to the appropriateness of combined transplant in patients with relatively mild or uncomplicated diabetes. Notwithstanding, although the model generally did not indicate that SPK was cost effective compared to KTA with continued insulin therapy, the absolute differences were not large.

However, the limitations and the underlying assumptions of the CEA model used in this assessment may have resulted in an immoderately optimistic evaluation. Evans et al(110). reviewed data from the National Cooperative Transplantation Study and noted that significant markup (as high as 200 percent) of organ procurement charges had been included in billings to third-party payers. His investigation revealed that inflation-adjusted billed charges for hearts and livers rose by 64.1 and 61.8 percent, respectively, between 1983 and 1991. The report commented that the relative cost effectiveness of transplants may be threatened. He believed that "the transplant community would do well to become more fiscally responsible than it is today.

The most disturbing characteristic of the many published series of SPK and PAK is the lack of detailed description of patient selection criteria. Aside from frequent (but not consistent) reference to serious contraindications to the surgery, it was not possible to determine what clinical characteristics of recipients placed them above the threshold for selection for SPK or PAK. In particular, it was unclear as to whether or not the severity of the diabetes, the intensity of the required insulin regimen, and the complications attendant to attempts to control H/HG played a significant role in patient selection. In fact, several transplant centers informed OHTA that these were not considered important in choosing recipients. Selection criteria were not adequately clarified in detailed formal protocols for patient pretransplant evaluation, several of which were forwarded to OHTA. Moreover, it was not apparent what clinical utility could be ascribed to some of the pretransplant testing such as ultrasound studies of the gallbladder and prophylactic cholecystectomy in asymptomatic patients; routine mammography in females; PSA screening in males; etc. Although such pretransplant investigations have not been demonstrated to be useful in patient selection or to result in improved graft survival or reduction of morbidity and mortality, this assessment considered that these studies might be viewed as variations in clinical judgment in the face of uncertainty, wherein the transplant team chose what they believed to be the most prudent course.

Additional data are needed in order to permit a logical and scientifically defensible conclusion as to the ultimate clinical effectiveness of combined kidney-pancreas transplantation. These include clear and unambiguous documentation of specific patient selection criteria, with explicit justification based on objective data for applying the procedure in patients who do not suffer from significant complications of H/HG or the management thereof. The NIH has recommended combined transplant for "labile insulin-dependent diabetics not responding well to exogenous insulin therapy" when used by institutions meeting "a present criterion of experience" (vide supra). In addition, prospective and longer followup in a larger number of patients is required to permit a reliable assessment of long-term pancreas graft survival; as mentioned, current survival data are based on a relatively small numbers of cases. The ultimate utility of some of the diagnostic testing and therapeutic interventions used in pretransplant remains unclear. Because such interventions in the aggregate add to the expense and potentially to the morbidity of SPK/PAK, the transplant community should seriously consider review and publication of such data, perhaps in association with the issue of appropriate selection criteria.

Prospective studies of alterations in QOL secondary to combined transplant are needed, and these investigations should use uniform, reliable, and validated measurement instruments. This might best be accomplished by the formation of a cooperative national group or groups that could most efficiently use the available expertise in this area.

Reliable data regarding the costs of SPK and PAK are not currently accessible. To determine the true cost effectiveness of combined transplant as compared with alternative treatments, transplant centers and private payers should cooperate in making charge and payment data available to the public. Although OHTA formally solicited cost data from every transplant center performing SPK/PAK, only five institutions or organizations responded. Although those centers are to be commended for their cooperation and willingness to participate in this public process, the overall response was inadequate to address this significant public concern. It should be noted that CEAs are now required of the assessment process by Federal statute (Public Law 102-410, Title IX of the Public Health Service Act), indicating the increasing interest of Congress in this aspect of technology assessment.

Finally, retransplantation is a significant issue from a medical, economic, and ethical viewpoint. Although on a national level this does not yet appear to be very frequent, in those centers performing the largest numbers of SPK and PAK, the retransplant rate is not insignificant. Increases in rates of retransplantation could require careful reexamination of SPK/PAK, particularly as regards issues of cost effectiveness, expected medical benefit of repeated transplant, and social and ethical concerns regarding the most appropriate use of scarce donor organs.

Appendix A. Letter to transplant centers

Chief, Transplantation Program
[Transplant Center...]

Dear Doctor,

The Office of Health Technology Assessment (OHTA) is conducting an assessment of simultaneous pancreas-kidney (SPK) and pancreas-after-kidney (PAK) transplantation. To that we are soliciting information from all UNOS member transplant centers that perform such procedures. As you may aware, we have already published a notice of assessment in the Federal Register requesting information on this technology, but wish in addition to invite each transplant center to participate in the process.

If you choose to respond to this request, you should be aware that OHTA will discuss in the assessment only those data and analyses for which a source can be cited. If you have no objection to such citation, you should enclose a written statement of consent. Otherwise, in accordance with 42 U.S.C. 199a-1(c), no information may be disclosed that would permit identification of an individual or an entity supplying information.

We are specifically interested in any information you may care to submit regarding the following factor in SPK/PAK transplantation:

  1. Data available concerning the duration of hospitalization at time of transplant, number of readmission and total (or average) inpatient stay resulting from such readmissions.
  2. Graft survival rates at 2 or more years posttransplant.
  3. Duration of posttransplant patient followup (range, mean, and median).
  4. Data on SPK/PAK recipient selection criteria; for example:
    1. Measures of severity of pretransplant diabetes (e.g., glycohemoglobin and blood glucose levels, documented episodes of hypoglycemia or ketoacidosis, hospitalization and/or emergency department admissions due to diabetes, secondary complications, etc.).
    2. Diabetes treatment regimen used before selection for transplant.
    3. Age and rationale for age limitations, if any.
    4. Comorbid conditions.
    5. Formal patient selection protocol, if any (e.g., some institutions routinely perform pretransplant testing including such items as electrocardiogram, stress test, coronary angiography, ultrasound evaluation of the gallbladder, etc.).
    6. Use of pretransplant tissue matching.
  5. General operative approach(es) used at your institution, e.g., bladder, enteric drainage, etc.
  6. Posttransplant immunosuppressive regimen.
  7. Costs of the entire transplant procedure at your institution, including any preoperative testing and/or treatment to correct or ameliorate preexisting clinical conditions (as e.g., No. 4 above), or hospitalization for both the transplant and readmissions for disorders incident to the transplant, etc. We request this information since there are now statutory requirements (PL 102-410) that OHTA also evaluate the cost-effectiveness aspects of assessed
  8. Retransplant rates, including graft survival rates following retransplant.

Thank you for your cooperation in this endeavor. Your participation in the assessment process is appreciated.

Sincerely,

  • Thomas V. Holohan, M.D., FACP
  • Director, Office of Health Technology Assessment

Appendix B. Abbreviations

  • A-C-P = azathioprine, cyclosporine, prednisone
  • ADA = American Diabetes Association
  • ALG = antilymphocyte globulin
  • ATG = antithymocyte globulin
  • CAD = coronary artery disease
  • CAD/KT = cadaver kidney transplant
  • CEA = cost-effectiveness analysis
  • CMV = cytomegalovirus
  • DCCT = Diabetes Control and Complications Trial
  • DQOL = Diabetes Quality-of-Life Survey
  • EDTA = European Dialysis and Transplant Association
  • EKG = Electrocardiogram
  • EM = electron microscopy
  • ESRD = end-stage renal disease
  • FDA = Food and Drug Administration
  • GBM = glomerular basement membrane
  • H/HG = hyper/hypoglycemia
  • HbA1C = glycohemoglobin
  • HCFA = Health Care Financing Administration
  • HIAA = Health Insurance Association of American
  • HLA = histocompatibility locus antigen
  • IPTR = International Pancreas Transplant Registry
  • KAP = Kidney-after-pancreas
  • KTA = Kidney transplant alone
  • LRD/KT = living-related donor kidney transplant
  • LRD = living-related donor
  • MALG = Minnesota antilymphocyte globulin
  • MESH = medical subject headings
  • NCV = nerve conduction velocities
  • NIH = National Institutes of Health
  • OHTA = Office of Health Technology Assessment
  • PAK = pancreas-after-kidney transplantation
  • PTA = pancreas transplant alone
  • QALY = quality-adjusted life years
  • QOL = quality of life
  • SPK = simultaneous pancreas-kidney transplantation
  • UNOS = United Network for Organ Sharing
  • USRDS = United States Renal Data System

References

1.
Nathan DM. Long-term complications of diabetes mellitus N Engl J Med 1993. 328: 1676–1685. [PubMed: 8487827]
2.
Elixhauster A, Luce BR, Taylor WR, et al. Cost-benefit and cost-effectiveness analysis from 1979 to 1990: A bibliography Med Care 1993. 31(Suppl): JS1–JS150.
3.
Wright FH, Smith JL, Ames SA, et al. Function of pancreas allografts more than 1 year following transplantation Arch Surg 1989. 124:–. [PubMed: 2662939]
4.
Sutherland DE, Dunn DL, Goetz FC. A 10-year experience with 290 pancreas transplants at a single institution Ann Surg 1989. 210: 274–288. [PMC free article: PMC1357985] [PubMed: 2673082]
5.
Sutherland DE, Gruessner R, Gillingham E, et al. A single institution's experience with solitary pancreas transplantation: A multivariate analysis of factors leaing to improved outcome. Clin Transpl . 1991: 141–152. [PubMed: 1820112]
6.
Olausson M, Nyberg G, Norden G, et al. Outcome of pancreas transplantations in Goteborg, Sweden, 1985-1990 Diabetologica 1991. 34 (Suppl 1) S1–S3. [PubMed: 1936668]
7.
Garvin PJ, Carney KM, Aridge D. Evolution of synchronous renal and pancreatic transplantation Am J Surg 1989. 258: 265–629. [PubMed: 2589601]
8.
Goldman J, Oh HK, Webb MG, et al. Allogeneic whole pancreas transplantation in insulin-dependent diabetes mellitus H Ford Hosp Med J 1990. 38: 246–251. [PubMed: 2086553]
9.
Stratta RJ, Taylor RJ, Zorn BH, et al. Combined pancreas-kidney transplantation in Nebraska Nebraska Med J 1991. 76: 385–391. [PubMed: 1784320]
10.
Stratta RJ, Taylor RJ, Wahl TO. Recipient selection and evaluation for vascularized pancreas transplantation Transplantation 1993. 55: 1090–1096. [PubMed: 8497888]
11.
Stratta RJ, Taylor RJ, Ozaki CF, et al. A comparative analysis of results and morbidity in type I diabetics undergoing preemptive versus postdialysis combined pancreas-kidney transplantation Transplantation 1993; . 55: 1097–1103. [PubMed: 8388585]
12.
Stratta RJ, Taylor RJ, Ozaki CF, et al. The analysis of benefit and risk of combined pancreatic and renal transplantation versus renal transplantation alone Surg Gynecol Obstet 1992; 1993. 17: 177(2): 163–171. [PubMed: 8342097]
13.
Sollinger HW, Knechtle SJ, Reed A, et al. Experience with 100 consecutive simultaneous kidney-pancreas transplants with bladder drainage Ann Surg 1991. 214: 703–711. [PMC free article: PMC1358495] [PubMed: 1741650]
14.
Sasaki T, Pirsch JD, D'Allesandro AM, et al. Simultaneous pancreas-kidney transplantation at University of Wisconsin-Madison Hospital. Clin Transpl . 1991: 159–178.
15.
Sollinger HW, Ploeg RJ, Eckhoff DE, et al. Two hundred consecutive simultaneous pancreas-kidney transplants with bladder drainage Surgery 1993; . 114: 736–744. [PubMed: 8211688]
16.
Alexander JW, First MR, Hariharan S, et al. Recent contributions to transplantation at the University of Cincinnati. Clin Transpl . 1991: 159–178. [PubMed: 1820114]
17.
Shaffer D, Madras PN, Sayhoun AI, et al. Combined kidney and pancreas transplantation: A 3-year experience Arch Surg 1992. 127: 574–578. [PubMed: 1575627]
18.
Bently Bentley FR, Garrison RN. Superior results with combined kidney-pancreas transplants Am Surgeon 1992. 58: 136–140. [PubMed: 1550306]
19.
Rosenloff Rosenlof LK, Earnhardt RC, Pruett TL. Pancreas transplantation: An initial experience with systematic systemic and portal drainage of pancreatic allografts Ann Surg 1992. 215: 586–595. [PMC free article: PMC1242509] [PubMed: 1632680]
20.
Dawahra M, Cloix P, Martin X, et al. Simultaneous transplantation of kidney and pancreas in diabetes patients Transpl Proc 1993. 25: 2227–2229. [PubMed: 8516880]
21.
Elkhammas EA, Henry ML, Tes RJ, et al. Combined kidney-pacreaspancreas transplantation at the Ohio State University hospitals. Clin Transpl . 1992: 191–197. [PubMed: 1306697]
22.
Van Oosterhout EC, van der Woude FJ, Lemkes HH, et al. Simultaneous pancreas and kidney transplantation. A feasible procedure in selected patients Netherlands J Med 1992. 41: 194–207. [PubMed: 1494398]
23.
Griffin PJ, Owens D, Krishman H, et al. Pancreatic transplantation in a small program programme Br J Surg 1994. 81: 98–101. [PubMed: 8313137]
24.
Brattstorm Brattstrom C, Tibell A, Tyden G, et al. Outcome in 22 patients with pancreas transplants functioning beyond 5 years Transpl Proc 1994. 26: 414–415. [PubMed: 8171479]
25.
Cheung AH, Sutherland DE, Gillingham KJ. Simultaneous pancreas-kidney transplant versus kidney transplant alone in diabetic patients Kidney Intl 1992. 41: 924–929. [PubMed: 1513114]
26.
Schulak JA, Mayes JT, Hricik DE. Kidney transplantation in diabetic patients undergoing combined kidney-pancreas or kidney-only transplantation Transplantation 1992; . 53: 685–687. [PubMed: 1549868]
27.
Sutherland DE, Gruessner A, Moudry-Munns K. International pancreas transplant registry report Transpl Proc 1994. 26: 407–411. [PubMed: 8171477]
28.
Sutherland DE, Gruessner A, Moudry-Munns K, et al. Analysis of United Network for Oregon Sharing (UNOS) USA Pancreas Transplant Registry data according to multiple variables. Final report to UNOS and the Department of Health and Human Services. 1993.
29.
Hunsicker LG, Edwards EB, Breen TJ, et al. Effect of center size and patient l-mix covariates on transplant specific patient and graft survival in the United States Transpl Proc 1993. 25: 1318–1320. [PubMed: 8442127]
30.
Gruessner RW, Dunn DL, Tzardis PJ, et al. Simultaneous pancreas and kidney transplants versus single kidney transplants and previous kidney transplants in uremic patients and single pancreas transplants in nonuremic diabetic patients: Comparison of rejection, morbidity, and long-term outcome Transpl Proc 1990; . 22: 622–623. [PubMed: 2327003]
31.
Tufveson G, Brynger H, Dimeny E, et al. Renal transplantation in diabetic patients with or without simultaneous pancreatic transplantation, 1986: Data from the EDTA Registry Nephrol Dial Transpl 1991. 191: 1–4. [PubMed: 2057109]
32.
[no authors listed] United States Renal Data System (USRDS). Annual Data Report. . 1992
33.
Cecka JM, Terasaki PI. The UNOS Scientific Renal Transplant Registry-119. Clin Transpl . 1991: 1–11. [PubMed: 1820108]
34.
Petronis KR, Sutherland DE, Moudry-Munns KC, et al. A comparison of simultaneous kidney-pacreas transplantation and kidney transplantation alone among diabetic end-stage renal disease patients in the United States The Urban Institute Health Policy Center 1993;(in press)
35.
Sutherland DE, Gruessner A, Moudry-Munns K. Analysis of UNOS United States of America Pancreas Transplant Registry data according to multiple variables. Clin Transpl . 1992: 45–59. [PubMed: 1306718]
36.
Takemoto S, Terasaki PI, Cecka JM, et al. Survival of nationally shared HLA-matched kidney transplants from cadaveric donors N Engl J Med 1992. 327: 834–839. [PubMed: 1508242]
37.
Cheung AH, Matas AJ, Gruessner RG, et al. Should uremic diabetic patients who want a pancreas transplant receive a simultaneous cadavere kidney? Transpl Proc 1993. 25: 1184–1185. [PubMed: 8442081]
38.
[no authors listed] UNOS Scientific Registry Report to the Division of Organ Transplantation, Health Resources and Services Administration November 28, 1993.
39.
Rosen CB, Frohnert PP, Velosa JA, et al. Morbidity of pancreas transplantation during cadaveric renal transplantation Transplantation 1991. 51: 123–127. [PubMed: 1846252]
40.
Konigsgrainer Konigsrainer A, Miller K, Kieselbach D, et al. Course of diabetic retinopathy after pancreas transplantation Transpl Proc 1990. 22: 689–90. [PubMed: 2327019]
41.
Konigsgrainer Konigsrainer A, Miller K, Steurer W, et al. Does pancreas transplantation influence the course of diabetic retinopathy? Diabetologica 1991. 34: S86–S88. [PubMed: 1936704]
42.
Ulbig M, Kampik A, Thurau S, et al. Long-term followup of diabetic retinopathy for up of 71 months after combined renal and pancreas transplantation Graefe's Arch Clin Exp Ophthalmol 1991. 229: 242–245. [PubMed: 1869059]
43.
Bandello F, Vigano C, Secchi A, et al. The influence of combined kidney-pancreas transplantation on advanced diabetic retinopathy Transpl Procl 1992. 24:–. [PubMed: 1604640]
44.
Bandello F, Vigano C, Secchi A. Effect of pancreas transplantation on diabetic retinopathy. A 20-case report Diabetologica 1991. 34 (Suppl 1) S92–S94. [PubMed: 1936706]
45.
Caldara R, Bandello F, Vigano C, et al. Influence of successful pancreaticorenal transplantation on diabetic retinopathy Transpl Proc 1994; . 26:–. [PubMed: 8171518]
46.
Scheider A, Meyer-Schwickerath E, Nusser J, et al. Diabetic retinopathy and pancreas transplantation. A 3-year followup Diabetologica 1991. 34: (Suppl 1) S95–S99. [PubMed: 1936707]
47.
Petersen MR, Vine AK. Progression of diabetic retinopathy after pancreas transplantation Opthalmology 1990. 97: 496–502. [PubMed: 2326028]
48.
Zech JC, Trepsat D, Gain-Guegnon M, et al. OpthalmologicalOphthalmological followup of type I(insulin-dependent) diabetic patients after kidney and pancreas transplantation Diabetologica 1991. 34: (Suppl 1) S89–S91. [PubMed: 1936705]
49.
Zech JC, Trepsat D, Gain-Guegnon M, et al. OpthalmologicOphthalmologic followup follow-up of type-I diabetic patients after kidney and pancreas transplantation Transpl Proc 1992. 24:–. [PubMed: 1604650]
50.
Scheider A, Thurau S, Ulbig M, et al. Diabetic retinopathy remains stable after combined kidney and pancreas transplantation for type-I (insulin-dependent) diabetes mellitus Diabetologica 1989; . 32:–.
51.
Ramsey Ramsay RC, Goetz FC, Sutherland DE, et al. Progression of diabetic retinopathy after pancreas transplantation for insulin-dependent diabetes mellitus N Engl J med Med 1988. 318: 208–214. [PubMed: 3275895]
52.
Bohman SO, Tyden G, Wilczek, et al. Prevention of kidney graft diabetic nephropathy by pancreas transplantation in man Diabetes 1985. 34: 306–308. [PubMed: 3918902]
53.
Bilous RW, Mauer SM, Sutherland DE, et al. The effects of pancreas transplantation on the glomerular structure of renal allografts in patients with insulin-dependent diabetes N Eng J Med 1989. 321: 80–85. [PubMed: 2659996]
54.
Morel P, Sutherland DE, Almond S, et al. Assessment of renal function in type-I diabetic patients after kidney, pancreas, or combined kidney-pancreas transplantation Transplantation 1991. 51: 1184–1189. [PubMed: 2048195]
55.
Mauer SM, Goetz FC, McHugh LE, et al. Long-term study of normal kidneys transplanted into patients with type-I diabetes Diabetes 1989. 38: 516–523. [PubMed: 2647558]
56.
Wilczek HE, Jaremko G, Groth CG. Pancreatic graft protects a simultaneously transplanted kidney from developing diabetic nephropathy: A-1-to 6-year followup study Transpl Proc 1993. 25: 1314–1315. [PubMed: 8442125]
57.
Fioretta Fioretto P, Mauer SM, Bilous RW, et al. Effects of pancreas transplantation on glomerular structure in insulin-dependent diabetic patients with their own kidneys Lancet 1993. 342: 1193–1196. [PubMed: 7901527]
58.
Solders G, Wilczek H, Gunnarsson R, et al. Effects of combined pancreatic and renal transplantation on diabetic neuropathy: A 2-year followupfollow-up study Lancet 1987. 2: 1232–1235. [PubMed: 2890854]
59.
Solders G, Tyden G, Persson A, et al. Improvement of nerve conduction in diabetic neuropathy: A followup follow-up study 4 years after combined pancreatic and renal transplantation Diabetes 1992. 41: 946–951. [PubMed: 1628768]
60.
Naouri A, Martin X, Dubernard JM. Evolution of diabetic neuropathy after kidney-pancreas transplant Transpl Proc 1992. 24: 875–876. [PubMed: 1604651]
61.
Murat A, Pouliquen B, Cantarovich D, et al. Gastric emptying improvement after simultaneous segmental pancreas and kidney transplantation Transpl Proc 1992. 24:–. [PubMed: 1604641]
62.
Vial C, Martin X, Lefrancois N, et al. Electrophysiologic evolution of diabetic polyneuropathy after combined pancreas and renal transplantation Transpl Proc 1992. 24: 869–870. [PubMed: 1604648]
63.
Vial C, Martin X, Lefrancois N, et al. Sequential electrodiagnositc electrodiagnostic evaluation of diabetic neuropathy after combined pancreatic and renal transplantation Diabetologica 1991. 34(Suppl 1): S100–S102. [PubMed: 1936669]
64.
Gaber AO, Cardoso S, Pearson S, et al. Improvement in autonomic function following combined pancreas-kidney transplantation Tranpl Proc 1991. 23: 1660–1662. [PubMed: 1989324]
65.
Gaber AO, Oxley D, Karas J. Changes in gastric emptying in recipients of successful combined pancreas-kidney transplants Dig Dis 1991. 9: 437–443. [PubMed: 1804583]
66.
Kennedy WR, Navarro X, Ramsay RC, Goetz FC, Sutherland DE, et al. Progression of diabetic retinopathy after pancreas transplantation for insulin-dependent diabetes mellitus N Engl J Med 1988. 318: 208–214. [PubMed: 3275895]
67.
Van Der Vliet JA, Navarro X, Kennedy WR, et al. The effect of pancreas transplantation on diabetic neuropathy Transplantation 1988. 45: 368–370. [PubMed: 3125636]
68.
Comi G, Galardi G, Amadio S, et al. Neurophysiological study of the effect of combined kidney and pancreas evaluation Diabetologica 1991. 34: (Suppl 1) S103–S107. [PubMed: 1936670]
69.
Secchi A, Martenenghi Martinenghi S, Galardi G, et al. Effects of pancreatic transplantation on diabetic polyneuropathy Transpl Proc 1991. 23: 1658–1659. [PubMed: 1989323]
70.
Nusser J, Scheuer R, Abendroth D, et al. Effect of pancreatic and/or renal transplantation on diabetic autonomic neuropathy Diabetologica 1991. 34: (Suppl 1) S118–S120. [PubMed: 1936674]
71.
Muller-Felber W, Landgraf R, Wagner S, et al. FollowupFollow-up study of sensory-motor polyneuropthy polyneuropathy in type-I diabetic subjects after simultaneous pancreas and kidney transplantation and after graft rejection Diabetologica 1991. 34: (Suppl 1) S113–S117. [PubMed: 1936673]
72.
Boucek P, Bartos V, Vanek I, et al. Diabetic autonomic neuropathy after pancreas and kidney transplant Diabetologica 1991. 34: (Suppl 1) S121–S124. [PubMed: 1936675]
73.
Hathaway D, Abell T, Cardoso S, et al. Improvement in autonomic function following pancreas-kidney versus kidney-alone transplantation Transpl Proc 1993. 25: 1306–1308. [PubMed: 8442122]
74.
Cheung TW, Cox KL, Ahlfors CE, et al. Reversal of microangiopathy in long-term diabetic patients after successful simultaneous pancreas-kidney transplants Transpl Proc 1993. 25: 1310–1313. [PubMed: 8442124]
75.
Cheung AT, Perex Perez RV, Basadonna GP, et al. Microangiopathy reversal in successful simultaneous pancreas-kidney transplantation Transpl Proc 1994. 26: 493–495. [PubMed: 8171520]
76.
American Diabetes Association American Diabetes Association Technical Review and Position Statement Diabetes Care 1992. 15: 1668–1672. [PubMed: 1468299]
77.
Sutherland DE. State of the art in pancreas transplantation Trapl Proc 1994. 26: 316–320. [PubMed: 8108998]
78.
Zehrer CL, Gross CR. Quality of life of pancreas transplant recients recipients Diabetologica 1991. 34: (Suppl 1) S145–S149. [PubMed: 1936683]
79.
Gross CR, Zehrer CL. Health-related quality-of-life outcomes of pancreas transplant recipients Clin Transpl 1992. 6: 165–171. [PubMed: 10147929]
80.
Piehlmeier W, Bullinger M, Nusser J, et al. Quality of life in diabetic patients prior to or after pancreas transplantation in relation to organ function Transpl Proc 1992. 24: 871–873. [PubMed: 1604649]
81.
Piehlmeier W, Bullinger M, Nusser J, et al. Quality of life in type-I diabetic patients prior to and after pancreas and kidney transplantation in relation to organ function Diabetologica 1991. 34: (Suppl 1) S150–S157. [PubMed: 1936684]
82.
Zehrer CL, Gross CR. Comparison of quality of life between pancreas/kidney and kidney transplant recipients: 1-year followup Transpl Proc 1994. 26: 508–509. [PubMed: 8171528]
83.
Gross CR, Zehrer CL. Impact of the addition of a pancreas to quality of life in uremic diabetic recipients of kidney transplants Transpl Proc 1993; . 25: 1293–1295. [PubMed: 8442117]
84.
Nakache R, Tyden G, Groth C. Quality of life in diabetic patients after combined pancreas-kidney or kidney transplantation Diabetes 1989. 38: (Suppl 1) S40–S42. [PubMed: 2642856]
85.
Nakache R, Tyden G, Groth C. Long-term quality of life in diabetic patients after combined pancreas-kidney transplantation or kidney transplantation Transpl Proc 1994; . 26: 510–511. [PubMed: 8171529]
86.
Nathan DM, Fogel H, Norman D, et al. Long-term metabolic and quality-of-life results with pancreatic/renal transplantation in insulin-dependent diabetes mellitus Transplantation 1991. 52: 85–91. [PubMed: 1858159]
87.
Campbell A, Converse P, Rogers W. The Quality of American Life: Perceptions, Education, and Satisfaction, New York: Russell Sage Foundation, 1976.
88.
Piehlmeier W, Bullinger M, Kirchberger I, et al. Prospective study of the quality of life in type-I diabetic patients before and after organ transplantation Transpl Proc 1994. 26: 522–523. [PubMed: 8171535]
89.
Bentdal OH, Fauchald P, Brekke IB, et al. Rehabilitation and quality of life in diabetic patients after successful pancreas-kidney transplantation Diabetologica 1991. 34: (Suppl 1) S158–S159. [PubMed: 1936685]
90.
Johnson JL, Schellberg J, Munn SR, et al. Does pancreas transplantation really improve the patient's quality of life? Transpl Proc 1990. 22: 575–576. [PubMed: 2326987]
91.
Secchi A, Di Carlo V, Martenenghi Martinenghi S, et al. Effect of pancreas transplantation of life expectancy, kidney function, and quality of life in uremic type-I diabetic patients Diabetologica 1991. 34: (Suppl 1) S141–S144. [PubMed: 1936682]
92.
Milde FK, Hart L, Zehr P. Quality of life of pancreatic transplant recipients Diabetes Care 1992. 15: 1459–1463. [PubMed: 1468271]
93.
Hathaway DK, Hartwig MS, Milstead J, et al. Improvement in quality of life reported by diabetic recipients of kidney-only and pancreas-kidney allografts Transpl Proc 1994. 26: 512–514. [PubMed: 8171530]
94.
Kiebert GM, van Oosterhout EC, Lemkes H, et al. Quality of life after combined renal-pancreatic transplantation Transpl Proc 1994. 26:–.
95.
Schareck WD, Hopt UT, Geisler F, et al. Quality of life after combined pancreas-kidney transplantation Transpl Proc 1994. 26:–. [PubMed: 8171533]
96.
Diabetes Control and Complications Trial (DCCT) Research Group. Reliability and validity of a diabetes quality-of-life measure for the diabetes control and complications trial Diabetes Care 1988. 11: 725–732. [PubMed: 3066604]
97.
The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus N Engl J Med 1993; . 329: 976–986. [PubMed: 8366922]
98.
Torrance GW, Feeny D. Utilities and quality-adjusted life years Intl J Technol Assess Health Care 1989. 5: 559–575. [PubMed: 2634630]
99.
Evans RW, Mannimen DL, Dong FB. The National Cooperative Transplantation Study Final Report (Pub. no. BHARC-100-91-020, Vol III, Ch 12), Seattle, WA: Battlelle-Seattle Research Center, 1991.
100.
Saywell RM, Woods JR, Halbrook HG, et al. Cost analysis of heart transplantation from the day of operation to the day of discharge J Heart Transpl 1989. 8: 244–252. [PubMed: 2661775]
101.
Lesse Leese B. The costs of diabetes and its complications Soc Sci Med 1992; . 35: 1303–1310. [PubMed: 1439913]
102.
Trochu JN, Cantarovich D, Renaudeau J, et al. Assessment of coronary artery disease by thallium scan in type-1 diabetic uremic patients awaiting combined pancreas and renal transplantation Angiology 1991; . 42: 302–307. [PubMed: 2014921]
103.
Manske CL, Wang Y, Rector T, et al. Coronary revascularization revascularisation in insulin-dependent diabetic patients with chronic renal failure Lancet 1992. 340: 998–1002. [PubMed: 1357450]
104.
Lowell JA, Stratta RJ, Taylor RJ. Cholelithiasis in pancreas and kidney recipients with diabetes Surgery 1993. 114: 858–863. [PubMed: 8211705]
105.
Ratner R, Delone P, Saski T, et al. Combined kidney-pancreas transplantation in patients with unrecognized NIDDM Proc 54th Ann Sci Session, Am Diabetes Assoc., , New Orleans, LA 1994. 156:–.
106.
Guyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life Ann Int Med 1993. 118: 622–629. [PubMed: 8452328]
107.
Evan Evans RW, Manninen DL, Dong FB. Is retransplantation cost-effective? Tranpl Proc 1993. 25: 1694–1696. [PubMed: 8442240]
108.
Remuzzi G, Ruggenenti P, Mauer SM. Pancreas and kidney/pancreas transplant transplants: Experimental medicine or real improvement? Lancet 1994. 343: 27–31. [PubMed: 7905046]
109.
Holohan TV, Handlesman H. Quality of evidence as it relates to organ transplantation J Transpl Coordination 1993. 3: 162–165.
110.
Evans RW. Organ procurement expenditures and the role of financial incentives JAMA 1993. 269; 3113–3118. [PubMed: 8505813]

AHCPR Pub. No. 95-0065

PubReader format: click here to try

Views

  • PubReader
  • Print View
  • Cite this Page