FTO predicts weight regain in the Look AHEAD clinical trial.

BACKGROUND: Genome-wide association studies have provided new insights into the genetic factors that contribute to the development of obesity. We hypothesized that these genetic markers would also predict magnitude of weight loss and weight regain after initial weight loss. METHODS: Established obesity risk alleles available on the Illumina CARe iSelect (IBC) chip were characterized in 3899 overweight or obese participants with type 2 diabetes from the Look AHEAD (Action for Health in Diabetes), a randomized trial to determine the effects of intensive lifestyle intervention (ILI) and diabetes support and education (DSE) on cardiovascular morbidity and mortality. Primary analyses examined the interaction between 13 obesity risk polymorphisms in eight genes and randomized treatment arm in predicting weight change at year 1, and weight regain at year 4 among individuals who lost 3% or more of their baseline weight by year 1. RESULTS: No single-nucleotide polymorphisms (SNPs) were significantly associated with magnitude of weight loss or interacted with treatment arm at year 1. However, fat mass and obesity associated gene (FTO) rs3751812 predicted weight regain within DSE (1.56 kg per risk allele, P=0.005), but not ILI (P=0.761), resulting in SNP × treatment arm interaction (P=0.009). In a partial replication of prior research, the obesity risk (G) allele at BDNF rs6265 was associated with greater weight regain across treatment arms (0.773 kg per risk allele), although results were of borderline statistical significance (P=0.051). CONCLUSIONS: Variations in the FTO and BDNF loci may contribute risk of weight regain after weight loss.

Effect of pioglitazone on body composition and bone density in subjects with prediabetes in the ACT NOW trial.

AIMS: This study examined the effects of pioglitazone on body weight and bone mineral density (BMD) prospectively in patients with impaired glucose tolerance as pioglitazone (TZD) increases body weight and body fat in diabetic patients and increases the risk of bone fractures. METHODS: A total of 71 men and 163 women aged 49.3 (10.7) years [mean (s.d.)]; body mass index (BMI), 34.5 (5.9) kg/m(2) were recruited at five sites for measurements of body composition by dual energy X-ray absorptiometry at baseline and at conversion to diabetes or study end, if they had not converted. RESULTS: Mean follow-up was 33.6 months in the pioglitazone group and 32.1 months in the placebo group. Body weight increased 4.63 ± 0.60 (m ± s.e.) kg in the pioglitazone group compared to 0.98 ± 0.62 kg in the PIO group (p < 0.0001). Body fat rose 4.89 ± 0.42 kg in the pioglitazone group compared to 1.41 ± 0.44 kg, (p < 0.0001) in placebo-treated subjects. The increase in fat was greater in legs and trunk than in the arms. BMD was higher in all regions in men and significantly so in most. PIO decreased BMD significantly in the pelvis in men and women, decreased BMD in the thoracic spine and ribs of women and the lumbar spine and legs of men. Bone mineral content also decreased significantly in arms, legs, trunk and in the total body. CONCLUSIONS: Pioglitazone increased peripheral fat more than truncal fat and decreased BMD in several regions of the body.

Determinants of glucose tolerance in impaired glucose tolerance at baseline in the Actos Now for Prevention of Diabetes (ACT NOW) study.

AIMS/HYPOTHESIS: The aim of the study was to examine the determinants of oral glucose tolerance in 602 persons with impaired glucose tolerance (IGT) who participated in the Actos Now for Prevention of Diabetes (ACT NOW) study. METHODS: In addition to the 602 IGT participants, 115 persons with normal glucose tolerance (NGT) and 50 with impaired fasting glucose (IFG) were identified during screening and included in this analysis. Insulin secretion and insulin sensitivity indices were derived from plasma glucose and insulin during an OGTT. The acute insulin response (AIR) (0-10 min) and insulin sensitivity (S(I)) were measured with the frequently sampled intravenous glucose tolerance test (FSIVGTT) in a subset of participants. RESULTS: At baseline, fasting plasma glucose, 2 h postprandial glucose (OGTT) and HbA(1c) were 5.8 +/- 0.02 mmol/l, 10.5 +/- 0.05 mmol/l and 5.5 +/- 0.04%, respectively, in participants with IGT. Participants with IGT were characterised by defects in early (DeltaI (0-30)/DeltaG (0-30) x Matsuda index, where DeltaI is change in insulin in the first 30 min and DeltaG is change in glucose in the first 30 min) and total (DeltaI(0-120)/DeltaG(0-120) x Matsuda index) insulin secretion and in insulin sensitivity (Matsuda index and S(I)). Participants with IGT in whom 2 h plasma glucose was 7.8-8.3 mmol/l had a 63% decrease in the insulin secretion/insulin resistance (disposition) index vs participants with NGT and this defect worsened progressively as 2 h plasma glucose rose to 8.9-9.94 mmol/l (by 73%) and 10.0-11.05 mmol/l (by 80%). The Matsuda insulin sensitivity index was reduced by 40% in IGT compared with NGT (p < 0.005). In multivariate analysis, beta cell function was the primary determinant of glucose AUC during OGTT, explaining 62% of the variance. CONCLUSION: Our results strongly suggest that progressive beta cell failure is the main determinant of progression of NGT to IGT.

Is adiposopathy (sick fat) an endocrine disease?

OBJECTIVE: To review current consensus and controversy regarding whether obesity is a 'disease', examine the pathogenic potential of adipose tissue to promote metabolic disease and explore the merits of 'adiposopathy' and 'sick fat' as scientifically and clinically useful terms in defining when excessive body fat may represent a 'disease'. METHODS: A group of clinicians and researchers, all with a background in endocrinology, assembled to evaluate the medical literature, as it pertains to the pathologic and pathogenic potential of adipose tissue, with an emphasis on metabolic diseases that are often promoted by excessive body weight. RESULTS: The data support pathogenic adipose tissue as a disease. Challenges exist to convince many clinicians, patients, healthcare entities and the public that excessive body fat is often no less a 'disease' than the pathophysiological consequences related to anatomical abnormalities of other body tissues. 'Adiposopathy' has the potential to scientifically define adipose tissue anatomic and physiologic abnormalities, and their adverse consequences to patient health. Adiposopathy acknowledges that when positive caloric balance leads to adipocyte hypertrophy and visceral adiposity, then this may lead to pathogenic adipose tissue metabolic and immune responses that promote metabolic disease. From a patient perspective, explaining how excessive caloric intake might cause fat to become 'sick' also helps provide a rationale for patients to avoid weight gain. Adiposopathy also better justifies recommendations of weight loss as an effective therapeutic modality to improve metabolic disease in overweight and obese patients. CONCLUSION: Adiposopathy (sick fat) is an endocrine disease.

Hypoglycemia.

Except in diabetic patients receiving insulin or sulfonylureas, hypoglycemia is a rare disorder. It is identified by modified Whipple's criteria consisting of neuroglycopenic symptoms, a blood glucose level equal to or less than 40 mg/dL, and relief of symptoms by glucose use. The sources of the body glucose are dietary intake, glycogenolysis, and [figure: see text] gluconeogenesis. The metabolism of glucose involves oxidation and storage as glycogen or fat. Causes of hypoglycemia include medications or toxins capable of decreasing blood glucose, disorders associated with fasting hypoglycemia, and postprandial hypoglycemic disorders. The most common type of hypoglycemia is insulin-induced hypoglycemia in diabetics. Insulinoma is rare; however, it is the most common hormone-secreting islet cell tumor. The diagnosis is made by the occurrence of hypoglycemia in the presence of symptoms of neuroglycopenia and inappropriately high levels of insulin and C-peptide. In hospitalized patients, the diagnosis is best made by prolonged fast. Most insulinomas are small and require invasive methods for precise localization. In surreptitious insulin use, hypoglycemia is associated with low plasma C-peptide. Postprandial hypoglycemia occurs in response to feeding and is generally caused by excessive insulin effect. It is seen in patients with postgastric surgery and rarely in early diabetes mellitus. Idiopathic postprandial hypoglycemia is rare and seems to be caused by subtle abnormalities of insulin response to food. Treatment of postprandial hypoglycemia consists of frequent small meals, with deletion of refined carbohydrate and increased protein intake. Primary treatment of insulinoma is surgical resection of the tumor.

Management of type 2 diabetes. Evolving strategies for treatment.

Type 2 diabetes is the most prevalent form of diabetes, accounting for approximately 90% of cases. This article examines the current classification, diagnostic criteria for diabetes, and screening recommendations and provides a therapeutic strategy for improving glycemic control in patients with type 2 diabetes.

Comparative efficacy and potency of long-term therapy with glipizide or glyburide in patients with type 2 diabetes mellitus.

BACKGROUND: Long-term studies on the comparative efficacy and relative potency of glipizide and glyburide are sparse and controversial. METHODS: In a randomized prospective trial, we compared the effectiveness and relative potency of glipizide and glyburide over a 15-month period in 18 patients with type 2 diabetes mellitus (DM2) (9 on glyburide and 9 on glipizide) who were unresponsive to diet therapy. Glycemic control was assessed using 4 methods: 1) quarterly fasting plasma glucose (FPG), and 2-hour postprandial plasma glucose after a standard breakfast; 2) insulin and glucose response to Sustacal (test meal) challenge every 3 to 6 months; 3) quarterly hemoglobin A1c; and 4) intravenous glucose tolerance testing every 6 months to measure first and second phase insulin secretion. Patient characteristics were similar in each treatment group. RESULTS: Similar doses of glipizide (11 mg/day) or glyburide (10 mg/day) resulted in comparable reduction of FPG and hemoglobin A1c and increase in first phase insulin response to intravenous glucose tolerance testing. There was greater reduction in FPG and 2-hour postprandial plasma glucose with glipizide than with glyburide in 6 months. Contrary to the Physicians' Desk Reference, but consistent with another short-term study, our long-term study demonstrated that glipizide and glyburide are equipotent at similar doses in controlling hyperglycemia in DM2. CONCLUSIONS: Glipizide and glyburide are effective in controlling hyperglycemia with similar doses in DM2. Glipizide exhibits greater reduction in FPG and 2PPG at 6 months. Additional studies are needed to validate equipotency of these drugs.

Adrenal androgen and glucocorticoid dissociation in premenopausal rheumatoid arthritis: a significant correlate or precursor to onset?

Controlled studies of adrenal steroids in premenopausal women with rheumatoid arthritis (RA) have revealed subtle and inconsistent decreases in glucocorticosteroid (GCS) function, but prominent deficiencies of adrenal androgens (AA). Such findings have suggested that hypoandrogenicity may predispose to RA in younger women. However, recent prospective studies of serum cortisol and dehydroepiandrosterone sulfate (DHEAS) levels before (x = 12 yrs) the onset of the disease (pre-RA) offer an alternative perspective. Significant dissociation of serum cortisol and DHEAS levels was found only in the subgroup of premenopausal women who developed RA before age 50. This subgroup alone had significant deficiency in serum DHEAS levels. Aggregate data imply that the documented deficits of DHEAS (and other AA) in such young females are a correlate of relative adrenal insufficiency, and that subtle GCS dysfunction may either contribute to development of RA in such young women as well as pubertal girls or may predispose to earlier onset of disease.

Immunogenetic analysis suggests different pathogenesis for obese and lean African-Americans with diabetic ketoacidosis.

OBJECTIVE: When presenting with diabetic ketoacidosis (DKA), lean and obese patients differ in their subsequent clinical course. Although lean patients tend to remain insulin dependent, most obese patients recover endogenous insulin secretion and discontinue insulin therapy. The aim of this study was to determine whether obese African-American patients with DKA could be determined to have type 1 or type 2 diabetes based on insulin secretion or the presence of immunological and genetic markers. RESEARCH DESIGN AND METHODS: This was a prospective study that analyzed the clinical characteristics, insulin secretion indices, immunological markers (islet cell, GAD, ICA512, and insulin autoantibodies), and HLA susceptibility genes (DR/DQ) in 131 patients with DKA (77 obese and 54 lean), 51 obese patients with hyperglycemia but no DKA, and 25 nondiabetic subjects. All subjects were African-American. Beta-cell function was evaluated by the C-peptide response to glucagon (1 mg i.v.) within 48 h of resolution of DKA or hyperglycemia. RESULTS: The acute C-peptide response was lower in obese DKA patients (1.0+/-0.1 ng/ml) than in obese patients with hyperglycemia (1.7+/-0.2 ng/ml, P < 0.01), but was higher than that in lean DKA patients (0.2+/-0.1 ng/ml, both P < 0.01). The overall prevalence of autoantibodies in obese subjects with DKA (17%) and obese subjects with hyperglycemia (16%) was lower than that in lean subjects with DKA (65%, P < 0.01). Obese patients with hyperglycemia and positive autoantibodies had lower rates of insulin secretion than those without antibodies. Regardless of body weight, all DKA patients with GAD autoantibodies carried the DQB1*0201 allele. However, there were no significant differences in HLA distribution between the three patient groups. CONCLUSIONS: Our results indicate that most obese African-American patients with DKA have type 2 diabetes characterized by higher insulin secretion, the absence of autoimmune markers, and a lack of HLA genetic association. In contrast, most lean African-American patients with DKA have metabolic and immunological features of type 1 diabetes. At presentation, assessment of beta-cell function and determination of autoimmune markers allow for correct classification of diabetes in African-Americans with hyperglycemic crises.

Racial differences in the correlation between gonadal androgens and serum insulin levels.

OBJECTIVE: We previously demonstrated a direct correlation between serum insulin levels and gonadal androgens (testosterone and androstenedione) in a group of obese hyperandrogenic predominantly black women. Subsequent work by others in predominantly white women showed conflicting results. To examine these potentially important racial differences further, 14 premenopausal females from each ethnic group, of similar age, BMI, and waist-to-hip ratio, were studied. RESEARCH DESIGN AND METHODS: We measured baseline gonadal androgens, dehydroepiandrosterone (DHEA), dehydroepiandrosterone sulfate (DHEAS), and leutinizing hormone (LH)/follicle-stimulating hormone ratio. Serum glucose, insulin, and C-peptide were measured at baseline and during a 2-h oral glucose tolerance test (area under the curve [AUC]). Insulin sensitivity was measured by glucose decrement during the first 15 min of an intravenous insulin tolerance test. RESULTS: Simple correlation analysis revealed a significant direct correlation in blacks (but not whites) between gonadal androgens and AUC for glucose, insulin, and C-peptide. Race-by-covariate interaction models reinforced the simple correlation finding. Cholesterol level was also correlated to all androgens in blacks, but not in whites. We also found that whites had higher serum triglycerides and greater AUC glucose than blacks. CONCLUSIONS: We conclude that there is a significant direct correlation between gonadal androgens and stimulated glucose, insulin, and C-peptide in blacks but not in whites. Thus, the previously reported direct correlation between gonadal hyperandrogenism and hyperinsulinemia may be a race-dependent phenomenon, hitherto an unreported observation. The implications of these findings are discussed.

Management of diabetic ketoacidosis.

Diabetic ketoacidosis is an emergency medical condition that can be life-threatening if not treated properly. The incidence of this condition may be increasing, and a 1 to 2 percent mortality rate has stubbornly persisted since the 1970s. Diabetic ketoacidosis occurs most often in patients with type 1 diabetes (formerly called insulin-dependent diabetes mellitus); however, its occurrence in patients with type 2 diabetes (formerly called non-insulin-dependent diabetes mellitus), particularly obese black patients, is not as rare as was once thought. The management of patients with diabetic ketoacidosis includes obtaining a thorough but rapid history and performing a physical examination in an attempt to identify possible precipitating factors. The major treatment of this condition is initial rehydration (using isotonic saline) with subsequent potassium replacement and low-dose insulin therapy. The use of bicarbonate is not recommended in most patients. Cerebral edema, one of the most dire complications of diabetic ketoacidosis, occurs more commonly in children and adolescents than in adults. Continuous follow-up of patients using treatment algorithms and flow sheets can help to minimize adverse outcomes. Preventive measures include patient education and instructions for the patient to contact the physician early during an illness.

Does syndrome X exist in hypertensive elderly persons with impaired glycemic control?

BACKGROUND: This report focuses on the glycemic state in relation to insulin and lipid levels of a cohort of elderly hypertensive persons to estimate the prevalence of syndrome X. METHODS: A cross-sectional study was performed at the University of Tennessee, Memphis, and the General Clinical Research Center (GCRC) on 95 participants in the Trial of Nonpharmacologic Interventions in the Elderly (TONE) study who agreed to participate in an ancillary study. A standard oral glucose tolerance test (OGTT) with insulin and C-peptide levels and a fasting lipid profile were obtained. RESULTS: In this sample of healthy elderly participants with hypertension who were taking an antihypertensive medication, 43 (45.3%) had normal glucose tolerance (NGT), 41 (43.2%) had impaired glucose tolerance (IGT), and 11 (11.6%) had undiagnosed non-insulin-dependent diabetes mellitus (NIDDM). Fasting hyperinsulinemia occurred in only one participant, who was in the IGT group. Hypertriglyceridemia and low high density lipoprotein (HDL) occurred in four persons, none of whom had hyperinsulinemia. Persons in the NIDDM and IGT groups had decreased beta cell function compared to persons in the NGT group, but did not have increased peripheral insulin resistance as estimated from the OGTT data. CONCLUSIONS: Our data demonstrated that in this cohort of elderly hypertensive participants with a high prevalence of central obesity, impaired glycemic control was common, but was not associated with fasting hyperinsulinemia or peripheral insulin resistance. Furthermore, we conclude that syndrome X essentially did not occur in these participants and postulate that the primary etiology for their impaired glycemic control is beta cell dysfunction. Further research is needed to elucidate these relationships.

Prevalence of undiagnosed non-insulin-dependent diabetes mellitus and impaired glucose tolerance in a cohort of older persons with hypertension.

OBJECTIVE: To determine the prevalence of undiagnosed non-insulin-dependent diabetes mellitus (NIDDM) and impaired glucose tolerance (IGT) in a cohort of older persons with hypertension. To examine the usefulness of screening for NIDDM in this study population. DESIGN: Cross-sectional study. SETTING: University of Tennessee, Memphis and the General Clinical Research Center (GCRC). PATIENTS: Ninety-five participants in the Trial of Nonpharmacologic Interventions in the Elderly (TONE) study who agreed to participate in an ancillary study. MEASUREMENTS: A standard oral glucose tolerance test (OGTT) with insulin and C-peptide levels was performed before the beginning of the TONE intervention. RESULTS: In this cohort, 43 participants (45.3%) had normal glucose tolerance (NGT), 41 (43.2%) had IGT, and 11 (11.6%) had undiagnosed NIDDM. The positive predictive value for NIDDM of a fasting glucose > or = 115 mg/dL in our participants was 57%. Hyperinsulinemia occurred in only one participant, a subject in the IGT group. CONCLUSIONS: Our data demonstrate that undiagnosed NIDDM is common in our cohort of older persons who are being treated for essential hypertension and that impaired glucose tolerance may be more common than in the general population of the same age. Further, our data show that the vast majority of this older, obese, hypertensive cohort did not have fasting hyperinsulinemia. We also infer that a fasting glucose alone has a low positive predictive value for screening of NIDDM in our older cohort. As the prevalence of NIDDM and impaired glucose tolerance in older hypertensive patients in the general population is unknown, we believe that further investigation is needed to characterize the relationship of hypertension, glycemic status, and hyperinsulinemia in the general population.

Hyperglycemic crises in urban blacks.

BACKGROUND: The hospital admission and mortality rates of patients with diabetic emergencies, such as diabetic ketoacidosis (DKA) and hyperglycemic hyperosmolar nonketotic syndrome (HHNS), are higher in black patients than in white patients with diabetes. However, there is limited data describing the precipitating events and response to treatment in black patients. Analysis of their clinical characteristics and response to medical therapy is needed to evaluate the impact of programs designed to reduce the development of these acute metabolic complications. METHODS: A prospective evaluation was conducted of 144 consecutive patients with DKA and 23 patients with HHNS admitted to a large inner-city hospital between July 1993 and October 1994. RESULTS: In patients previously diagnosed as having diabetes, poor compliance with insulin therapy was the major precipitating cause for DKA (49%) and HHNS (42%). Alcohol or cocaine abuse was a contributing factor for noncompliance and was present in 35% and 13% of patients with DKA and in 44% and 9% of patients with HHNS, respectively. Newly diagnosed diabetes accounted for 17% of patients with DKA and HHNS. Obesity (body mass index > 28 kg/m2 [the weight in kilograms divided by the square of the height in meters]) was present in 29% of patients with DKA and in 17% with HHNS and was most common in patients with DKA who were newly diagnosed as having diabetes (56%). Patients were treated by residents, who used a low-dose insulin protocol with an algorithm for insulin adjustment in 88 of 144 patients with DKA and 14 of 23 patients with HHNS. Although there was no difference in mortality rates or time needed to correct hyperglycemia or ketoacidosis, the use of the protocol significantly reduced the risk of hypoglycemia (5%) compared with patients treated without a protocol (23%) (P < .01). CONCLUSIONS: In urban black patients, poor compliance with insulin therapy was the main precipitating cause of acute metabolic decompensation, and substance abuse was a significant contributing factor for noncompliance. Obesity is common in black patients with DKA; it was present in more than half of those with newly diagnosed diabetes. Improved patient education and better access to medical care might reduce the development of these hyperglycemic emergencies.

NIDDM: new aspects of management.

Patients over age 40 should be made aware of the triad of risk factors for the disease and taught to recognize common signs and symptoms. Those with high-risk profiles should be tested regularly and counseled regarding preventive and therapeutic strategies. For obese patients whose weight cannot be brought under control with diet and exercise alone, a trial of anorectic agents should be considered.

Differing effects of pancreas-kidney transplantation with systemic versus portal venous drainage on cholesteryl ester transfer in IDDM subjects.

OBJECTIVE: Cholesteryl ester transfer (CET) is accelerated in patients with IDDM treated with conventional (subcutaneous) insulin therapy (CIT) and a number of other disorders associated with premature cardiovascular disease. We have shown that in IDDM this disturbance is closely linked to iatrogenic hyperinsulinemia (HI), because it was reversed when insulin was administered by the intraportal (i.p.) route. In this study, we sought to determine whether HI after successful pancreas-kidney transplantation (PKT) has the same adverse effect on CET. RESEARCH DESIGN AND METHODS: CET was measured by both mass and isotopic assays and compared in two groups of euglycemic non-insulin-requiring IDDM PKT patients with either systemically draining allografts and persistent HI or grafts with portal vein anastomoses that were normoinsulinemic (PK-P). A third group of eight nondiabetic kidney transplant (KT) patients receiving the same immunosuppressive drugs served as control subjects. RESULTS: CET in pancreas-kidney transplantation subjects with systemic venous drainage (PK-S) was increased (P < 0.001) to the same level we have reported previously in IDDM patients receiving CIT and was significantly higher (P < 0.001) than in those subjects with PK-P. CET in the PK-P group did not differ from that of the KT control patients. CONCLUSIONS: CET is affected by variations in systemic insulin levels in pancreas transplant patients with allografts that have differing venous drainage. Because high systemic insulin levels are linked to the activation of (ET, euglycemic HI IDDM pancreas allograft recipients may continue to be at high risk for macrovascular complications.

Results of pancreas transplantation with portal venous and enteric drainage.

PURPOSE: The standard method for pancreatic transplantation involves drainage of exocrine secretions into the urinary bladder with venous outflow into the systemic circulation. Despite the high success rate associated with this approach, it often leads to complications, including chemical cystitis, reflux pancreatitis, metabolic acidosis, and hyperinsulinemia. The authors developed a new technique of pancreatic transplantation with portal drainage of endocrine secretions and enteric drainage of exocrine secretions (PE), which theoretically should be more physiologic. PROCEDURES: All patients were insulin-dependent diabetics with end-stage renal disease who underwent combined kidney-pancreas transplantation. Between 1990 and 1994, 19 patients have been transplanted using intraperitoneal placement of the pancreas allograft with exocrine drainage into a Roux-en Y loop and venous drainage into the portal circulations (PE). A comparison group of all patients undergoing standard systemic-bladder (SB) transplantation between April 1989 and March 1993 (n = 28) also was studied. Patient follow-up ranges from 6 months to 5 years for the SB patients (mean = 2.5 years) and 6 months to 4 years for the PE patients (mean = 1.6 years). Routine follow-up includes documentation of the clinical course and detailed endocrine studies. FINDINGS: Patient and graft actuarial survival at 1 and 3 years is no different for SB and PE patients. Urinary tract infections occurred in 89.3% of the SB patients (2.8/patient) versus 26.3% of the PE patients (0.25/patient, p < or = 0.0001). None of the PE patients experienced hematuria compared with 53.6% of the SB patients (p < or = 0.0001); however, two PE patients had melanotic episodes. The incidence of urinary retention and reflux pancreatitis was 32.1% versus 5.3% (p < or = 0.028) for SB and PE groups, respectively. Patients in the SB group required sodium bicarbonate therapy (mean = 55 mEq/day) although no PE patient required routine therapy; despite this, SB patients experienced more episodes of acidosis (44 vs. 5). Endocrine studies indicate no difference in glycosylated hemoglobin or fasting and stimulated glucose values throughout the follow-up period. In contrast, hyperinsulinemia was evident in both fasting and stimulated tests for the SB patients, with values consistently two- to fivefold higher than those of the PE group. CONCLUSIONS: These results indicate that PE and SB pancreas transplantation are equivalent in terms of patient and graft survival and suggest that the PE approach is associated with a decreased incidence of metabolic and bladder-related complications. In addition, the PE approach eliminates the state of peripheral hyperinsulinemia that characterizes the SB procedure. Continued follow-up will be necessary to determine if long-term outcomes will differ for patients with PE and SB grafts.