Acellular human dermis promotes cultured keratinocyte engraftment.

In full-thickness skin injury, loss of dermis may result in compromised wound repair, including contracture, hypertrophic scarring, and wound breakdown. This report examines the effect of an acellular dermal matrix on in vivo skin repair. Human keratinocytes cultured onto a synthetic hydrophilic dressing were applied with (N = 9) and without (N = 11) an acellular dermal matrix to full-thickness skin defects on athymic mice. Host cells progressively repopulated the acellular dermal component of the grafts. All animals with dermal matrix revealed fully differentiated epidermis by postoperative day 21. Human keratinocytes persisted in all animals grafted with dermal matrix, compared to only 63.6% of those animals without a dermal component. Planimetric analysis revealed significantly reduced wound contraction (p = 0.016) in animals receiving the dermal matrix. Histologic, immunohistochemical, and electron microscopic analyses also were performed. These studies suggest that an acellular dermal matrix can effectively direct regeneration of normal skin morphology.

Burn injury results in decreased gastric acid production in the acute shock period.

Gastrointestinal complications including acute stress ulcerations occur after burn injury. The causes of acute gastric derangements are multiple, and tissue ischemia in the acute period of burn shock may promote breakdown of the gastric mucosal protective barrier. We compared gastric pH in mice after 25% total body surface area, full-thickness murine burn injury with that in unburned control animals. Animals were anesthetized with methoxyflurane and were resuscitated with 1 ml normal saline solution immediately after burn. Animals were killed at intervals up to 24 hours after burn injury, stomachs were removed and opened, and gastric mucosal pH was measured by use of a surface pH probe. In other animals mixed venous blood was obtained via direct inferior vena cava puncture, and blood gas analysis was performed at intervals up to 24 hours after burn injury. Unexpectedly, gastric mucosal pH increased in burned mice compared with that in controls. The peak difference, greater than one log pH unit, occurred at 3 hours after burn injury (pH 4.45 burn vs pH 3.34 control, p < 0.00001), and this difference was maintained through 12 hours (pH 4.88 burn vs pH 3.20 control, p < 0.005). In this model, shock was observed to begin as early as 1 hour after burn injury and reached its maximal period (base deficit, -27.8 mEq/L) at 12 hours after burn injury. In view of the higher gastric pH in burned mice with concomitant profound shock, gastric acid production appears to be impaired during this time, which suggests acute postburn gastrointestinal ulcerations may be primarily due to ischemia. Prevention of organ ischemia may play a key role in the prevention of acute gastric ulcerations after burn injury.

Effects of high-dose vitamin C administration on bacterial translocation and lung neutrophil sequestration in burned mice.

Severe burn injury produces shock and induces acute gastrointestinal derangements that may disrupt mucosal integrity and facilitate bacterial translocation (BT) to mesenteric lymph nodes, accompanied by endotoxemia. Antioxidant treatments may be beneficial after shock by acting as scavenger agents for highly reactive oxygen intermediates. We studied the effects of high dosages of vitamin C, a water-soluble antioxidant, on the incidence of BT and on levels of lung myeloperoxidase in burned mice. Myeloperoxidase is primarily found in neutrophils, and levels of myeloperoxidase in tissues reflect neutrophil sequestration. The doses of vitamin C used were equivalent on a weight basis to 1 gm/hr administered to humans over a 24-hour period, doses that have shown efficacy in improvement of resuscitation in other experimental burn models and currently are being used in clinical trials in patients with burns. Mice were anesthetized and received 32% total body surface area, full-thickness burn injury to the dorsum, followed by injection of 1 ml of Ringer's lactate (RL) for resuscitation. Mice were divided into three groups: (1) unburned, received anesthesia and RL injections; (2) burned, received vitamin C (14 mg/kg/hr) in 1 ml RL by intraperitoneal injection immediately after burn and via subcutaneous injection (0.5 ml) at 6 and 12 hours after burn; (3) burned, received identical injections of RL alone (control animals). Mesenteric lymph nodes were removed by use of sterile technique at 24 hours after burn and cultured; any growth was considered evidence of BT. The incidence of BT in burned mice was not altered by administration of vitamin C (normal, 10% BT; burn, 41.37% BT; burn+vitamin C, 45.83% BT). Similarly, burned animals that received vitamin C or RL alone did not differ in the levels of myeloperoxidase in the lungs (normal, 0.015 +/- 0.003 U/gm; burn, 0.2231 +/- 0.029 U/gm; burn+vitamin C, 0.281 +/- 0.041 U/gm).(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of small-volume bolus treatment with intravenous normal saline and 7.5 per cent saline in combination with 6 per cent dextran-40 on metabolic acidosis and survival in burned mice.

Standard murine burn models include the administration of intraperitoneal (i.p.) saline solutions which are intended to resuscitate the animals during subsequent burn shock. Prehospital administration of small volumes of concentrated salt solutions has been recommended for the early treatment of haemorrhagic shock, and have also been utilized for burn shock. We studied the effects of bolus intravenous (i.v.) hypertonic saline (HS) or hypertonic saline/dextran-40 (HS + DEX) on animal survival and acid-base balance following 25 per cent total body surface area, full-thickness burn injury in mice. I.v. injections were administered via a tail vein immediately prior to burn injury. Some mice received 1 ml i.p. normal saline (NS) immediately after burn injury. Acid-base balance of vena caval blood was measured during the period of maximal metabolic acidosis following burn injury (12 h postburn). Preburn i.v. administration of 5 ml/kg of HS or HS+DEX, followed by 1 ml i.p. NS, only slightly decreased the degree of metabolic acidosis compared to animals receiving i.p. fluid alone, the standard resuscitation regimen for burned mice. Preburn i.v. administration of 0.2 ml volumes of HS or HS + DEX, without i.p. fluid administration, resulted in extremely high mortality. Immediate preburn i.v. administration of HS or HS + DEX did not eliminate metabolic acidosis in this murine burn model, and markedly increased the mortality when subsequent i.p. fluids were not administered. The degree of metabolic acidosis in the murine experimental burn model has not previously been clearly described. Furthermore, adequate fluid resuscitation of these animals may be difficult to achieve without indwelling vascular catheters which could deliver continuous i.v. fluids following burn injury.

Ultrastructural features of composite skin cultures grafted onto athymic mice.

Skin substitutes composed of cultured keratinocytes with or without a dermal substrate are now being used in the treatment of burns and other cutaneous wounds. Composite skin cultures (Graftskin, LSE), consisting of epidermal keratinocytes seeded on a fibroblast-containing collagen matrix and maintained at the air-liquid interface, develop a well differentiated epidermis in vitro with many of the morphological and biochemical features of intact skin. Basement membrane-associated antigens, developing hemidesmosomes and short segments of lamina densa are present at the dermal-epidermal junction in vitro, although the LSE lacks a continuous basement membrane. As epidermal differentiation proceeds, the culture develops a stratum corneum composed of electron-dense corneocytes surrounded by extracellular lipid. However, the intercorneocyte lipid lamellae do not exhibit the repeating pattern of broad and narrow electron lucent bands observed in electron micrographs of the stratum corneum of intact skin. In this study, LSE were grafted onto full thickness wounds in athymic mice. Animals were killed 6, 15, 30 and 60 d after surgery for examination by light and electron microscopy to identify any ultrastructural changes which occurred in the culture in response to the host environment. The grafted LSE integrated well into the host tissue and remained intact throughout the 60 d study period. At the dermal-epidermal junction, a continuous basement membrane with a well defined lamina densa was established by 15 d after surgery. An extensive network of anchoring fibrils was present by 30 d after surgery. Collagen fibrils within the dermal matrix condensed by 6 d after surgery and began organising into loosely packed bundles by 15 d after surgery.(ABSTRACT TRUNCATED AT 250 WORDS)

Current trends in the use of allograft skin for patients with burns and reflections on the future of skin banking in the United States.

Cadaveric allograft skin can play a critical role in the care of patients with massive burns. It is difficult, however, to estimate current use and levels of enthusiasm for allograft skin in the United States. We report on a survey of 40 skin banks and 140 United States burn center medical directors as listed in the American Burn Association's Directory of Burn Care Resources for North America 1991-1992. Response rate was 45% for skin banks and 38% for burn directors. Overall, 12% of admitted patients were treated with allograft skin at the responding burn centers. Sixty-nine percent of burn center directors preferred to use fresh skin, although only 47% of skin banks were able to supply fresh cadaver skin. Tabulated survey results and a review and discussion of future directions in skin banking and replacement research are discussed in this paper and were presented to the Tissue Bank Special Interest group at the 1993 American Burn Association annual meeting.

Evaluation of Graftskin composite grafts on full-thickness wounds on athymic mice.

We used a living bilayered cultured skin replacement to close full-thickness wound defects on the dorsum of athymic mice. The skin replacement is composed of human fibroblasts that condense a bovine collagen lattice; the lattice is then seeded with cultured human keratinocytes. The collagen lattice with fibroblasts serves as a dermal template, and the overlying human keratinocytes form the epidermal component of this composite skin replacement. Twenty-four animals were grafted, and groups of six were killed and biopsied at 6, 15, 30, and 60 days after graft replacement. Twenty-four mice in the control group receiving grafts of fresh, split-thickness, human cadaver skin were biopsied at the same time points. "Take" of all grafts was excellent, with only one graft loss in the 48 mice (one Graftskin graft, at 15 days). Light microscopy revealed that vascular ingrowth into Graftskin occurred rapidly, and discrete dermal and epidermal layers were seen at all time points. Evidence of basement membrane formation was seen at 6 days after grafting by immunohistochemical staining for laminin and by electron microscopic visualization of lamina lucida and lamina densa zones at the dermal-epidermal junction. The results demonstrated that Graftskin formed a structurally complete skin replacement within 1 week of placement on full-thickness wounds on athymic mice, and effective skin coverage was provided for the 60-day observation period after grafting.

Development of a temporary living skin replacement composed of human neonatal fibroblasts cultured in Biobrane, a synthetic dressing material.

BACKGROUND: Preferred coverings for excised burn wounds when sufficient autograft skin is not available are fresh or cryopreserved cadaveric skin. Problems with supply, preservation, immune rejection, and potential infection transmission accompanying the use of allograft skin underscore the need for effective alternative temporary skin replacements. METHODS: We cultured human neonatal fibroblasts (HF) for 4 to 6 weeks in nylon mesh of Biobrane, a synthetic dressing consisting of a thin layer of silicone bonded to nylon mesh. Secreted matrix proteins were identified by immunostaining and quantitated, and growth factor-specific messenger RNAs were identified by reverse transcription-polymerase chain reaction. Living grafts (Biobrane/HF) were sutured to full-thickness, excised wounds on athymic mice; control animals received Biobrane alone. Wounds were observed and biopsy specimens were obtained at intervals during the subsequent 40 days. RESULTS: After 3 to 6 weeks of culture in Biobrane the HF proliferated and secreted matrix proteins including type I collagen, fibronectin, and decorin, as well as messenger RNA for several growth factors (acidic fibroblast growth factor, basic fibroblast growth factor, and keratinocyte growth factor). Biobrane/HF grafts were transferred to full-thickness wounds, resulting in rapid fibrovascular ingrowth from the wound and effective wound closure for up to 40 days with minimal inflammatory responses. Biobrane control grafts adhered initially to wounds, but within several days many grafts developed subgraft exudates; histologic sections revealed marked inflammatory responses in these wounds. By 20 days, most BB grafts were separating from the underlying wounds that were closing by epithelialization and contraction. CONCLUSIONS: The Biobrane/HF living skin replacement provides long-term biologic coverage of full-thickness wound defects in mice with rapid incorporation of a living tissue matrix into the wound bed. Because HF have been found to be relatively nonantigenic when transferred to allogeneic hosts, Biobrane/HF grafts could replace the use of cadaveric allograft skin for achieving temporary wound closure after burn wound excision. Biobrane/HF grafts may persist on human wounds for weeks or months, with long-term persistence perhaps primarily dependent on durability of the silicone rubber layer.

Reduction of bacterial translocation and intestinal structural alterations by heparin in a murine burn injury model.

Burn injury produces acute gastrointestinal (GI) derangements that may predispose the burn victim to bacterial translocation (BT). We studied the effects of heparin on gastrointestinal (GI) anatomic alterations and BT after 25% and 32% total body surface area (TBSA), full-thickness murine burn injuries. Heparin (100 U/kg) was administered with 1 mL of normal saline (NS) resuscitation solution immediately postburn and 4 hours and 18 hours postburn in volumes of 0.5 mL NS. Mice with 25% TBSA burns treated with heparin maintained small intestine weight, measured 24 hours postburn, and ileal mucosal height was preserved, whereas burned, untreated mice lost organ weight and mucosal height. Bacterial translocation was decreased in mice with 25% TBSA burn injuries treated with heparin (35.0% vs. 10.7%, p < 0.025). After 32% TBSA burn injuries, BT was also decreased in heparin-treated animals (64.3% vs. 31.6%; p < 0.025). Analysis of mixed venous blood gases showed that heparin did not affect the severe metabolic acidosis that follows burn injury in this animal model, indicating that general tissue perfusion was not improved. Heparin administered in the acute postburn period ameliorates GI structural and functional damage in this murine burn model and decreases BT.

Composite grafts of human keratinocytes grown on a polyglactin mesh-cultured fibroblast dermal substitute function as a bilayer skin replacement in full-thickness wounds on athymic mice.

We have developed and tested in athymic mice a new, cultured, dermal-epidermal graft composed of two human cell types coupled with a biodegradable dermal scaffold. Cultured, proliferating human keratinocytes (HK) were applied to the surface of a living dermal tissue replacement that is composed of human fibroblasts cultured on a polyglactin mesh. After 4 to 6 days of coculture, proliferating HKs achieved confluency on the surface of the living dermal tissue replacement. Grafts were then transferred to full-thickness wounds on the dorsum of athymic mice. Sixteen animals were grafted, and the mean percentage of graft take (original wound area covered) on day 20 after grafting was 51.25%. Staining with antibody specific for human involucrin confirmed the presence of HKs on closed wounds, and staining with antibody specific for human laminin revealed a continuous layer of laminin at the dermal-epidermal junction on day 20. Animals closed with living dermal tissue replacement alone markedly contracted, whereas application of living dermal tissue replacement-HK grafts appeared to retard contraction. Because polyglactin mesh fibers are absorbed by hydrolysis rather than by enzymatic degradation, this living composite graft may be more resistant to destruction when placed on excised human wounds than are composite grafts, which are composed of a collagen matrix. The inclusion of the living dermal substitute may ultimately provide better skin quality than is achieved from the use of cultured keratinocytes alone. Fragility of the epidermal layer is probably due to the short-term culture of HKs on the living dermal tissue replacement, and further efforts to develop a thicker epithelial layer may improve graft durability.

Effects of fluid resuscitation, burn eschar excision, and blockade of afferent pain responses on bacterial translocation and acid-base balance after murine burn injury.

We tested effects of fluid resuscitation, early burn excision/grafting, and blockade of afferent stimuli from the burn wound on bacterial translocation and acid-base balance after murine burn injury. Burn excisions were performed with patients either 15 minutes or 2 hours after burn injury under anesthesia, and excised wounds were immediately closed with murine allograft skin. Twenty-four hours after 25% total body surface area (TBSA) burn injury and 48 hours after 32% TBSA injury, mesenteric lymph nodes were cultured. Incidences of bacterial translocation in 25% and 32% TBSA burns were 31.6% and 68.4% of animals, respectively. Burned animals were in severe shock, and metabolic acidosis reached a nadir 12 hours after burn injury, with base deficit -27.8 +/- 0.6 mEq/L; 5% to 10% of animals died acutely after burn injury. After excision/grafting of burned mice at 2 hours after burn injury, the incidence of bacterial translocation was unchanged (35.7% with 25% TBSA burn, 73.3% with 32% TBSA burn), and mortality did not change. When 32% TBSA excisions were performed exactly 10 minutes after burn injury, four of the 13 mice died within several hours, and five (55.5%) of the nine survivors translocated. Rates of bacterial translocation in mice receiving anesthesia or excision/grafting without burn injury were 15.0% and 20%, respectively (p = NS compared with normal mice). Subcutaneous implantation of normal or burned skin into normal animals neither elicited shock nor increased the incidence of bacterial translocation. Increasing amounts of fluid resuscitation in the 25% TBSA burn model provided only delayed improvement of acid-base balance; increased amounts of fluid did not decrease bacterial translocation.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of corticotropin releasing factor on acid-base alterations and bacterial translocation in a murine model of thermal injury.

Corticotropin-releasing factor (CRF) is a 41 amino acid polypeptide produced by the hypothalamus which has been shown to decrease inflammation and tissue oedema when administered following burns, cold and acid injuries in some animal models, and to increase mesenteric blood flow. We determined whether systemic administration of CRF to burned mice would decrease metabolic acidosis and protect the gastrointestinal (GI) tract from ischaemic injury leading to bacterial translocation (BT). Synthetic CRF was administered by intraperitoneal injection in doses of 20 and 200 micrograms/kg to mice immediately following 25 and 32 per cent TBSA burn injuries; the doses were repeated at 8 and 16 h postburn. Severe metabolic acidosis, measured 12 h after burn injury, was not improved in mice which received CRF treatment. Bacterial translocation, measured by quantifying bacteria in mesenteric lymph nodes harvested from animals 48 h postburn, was also not decreased with CRF treatment. CRF does not improve general tissue perfusion nor decrease GI derangements leading to bacterial translocation in this animal model of burn injury.

Development of a device for measuring adherence of skin grafts to the wound surface.

Adherence of a biological graft to the wound surface is the most important factor influencing the ultimate success of graft viability. A machine has been developed to test the adherence of biological graft materials to a substrate such as a wound surface. The peeling mode, which yields reproducible quantitative measurements of adherence, is a standard method for testing adhesives. The device is designed to continuously measure the force required to peel the graft from the substrate at a constant rate. This force is a function of the energy of adhesion per unit area of adhered surface. This device has been used to measure the peeling force of (2 x 2 cm) skin grafts which are applied to full-thickness wounds on mice. Results of tests on adherence of autografts on mice show that the peeling force increases significantly with time over the first 9 days of healing. Thus, this device is useful in quantitative comparison of various skin grafting techniques and artificial grafts.

Evaluation of a biodegradable matrix containing cultured human fibroblasts as a dermal replacement beneath meshed skin grafts on athymic mice.

Meshed, expanded split-thickness skin grafts (MSTSG) frequently achieve poor results when used to cover full-thickness wounds. Poor cosmetic and functional results occur in part because the epithelium that grows across the skin graft interstices lacks a dermis. We used a living dermal replacement composed of either polyglycolic acid (PGA) or polyglactin-910 (PGL) mesh containing confluent, cultured human fibroblasts. These grafts were applied to full-thickness wounds on athymic mice; widely expanded, 3:1 ratio human MSTSG was then placed over the dermal graft. Histologic examination of wounds during a 99-day period after graft placement showed that PGA/PGL-fibroblast grafts vascularized to the wound, and the MSTSG simultaneously vascularized to the PGA/PGL-fibroblast graft. Epithelialization from the MSTSG bridges proceeded rapidly across the surface of the PGA/PGL-fibroblast grafts, resulting in an epithelialized layer that covered a densely cellular substratum that resembled dermis. Basement membrane formation at the dermal-epidermal junction of the epithelialized interstices was confirmed by immunohistochemical microscopy. Minimal inflammatory reaction to the PGA/PGL-fibroblast grafts was seen. Grafts composed of PGA or PGL biodegradable meshes combined with cultured fibroblasts vascularize in full-thickness wounds, resulting in formation of organized tissue beneath the epithelialized surface that resembles dermis.

Microbial contamination in allografted wound beds in patients with burns.

Microbial wound contamination has been recognized as a cause of autograft skin graft failure. Human cadaver allograft is believed to decrease or control microbial wound contamination, and by its adherence to the wound bed, to indicate a sufficiently low microbial count to allow successful application of autograft. Suboptimal "take" of more fragile cultured skin grafts after removal of adherent allograft led us to reexamine the microbial population of these wound beds. Immediately before autografts were placed, wound beds beneath adherent allograft were biopsied for quantitative microbiologic analysis. Eighty tissue biopsy specimens from 21 patients were examined. No patients had signs of infection or sepsis at the time of biopsy. Fifty-seven percent of all patients had positive cultures beneath adherent allograft skin. Twenty-one percent of all cultures revealed "infected" wound beds (greater than 100,000 colonies per gram), and 30% of all wounds were "colonized" (less than 100,000 colonies per gram). These data suggest that take and vascularization of allograft does not guarantee that a wound bed is free of microbial contamination.

Cyclosporine A for prolonging allograft survival in patients with massive burns.

Cyclosporine A (CsA) immunosuppression was used in three patients with massive burns to prolong skin allograft survival. Cyclosporine A kinetic studies in patients with burns revealed markedly accelerated blood clearance and high variability in drug absorption when compared with studies in renal transplantation patients. Doses required to maintain therapeutic levels varied widely. While patients were receiving adequate maintenance therapy with CsA immunosuppression the allograft was tightly adherent without gross or microscopic rejection and was indistinguishable from autograft. Ultimately, patients' wounds were permanently covered with sequential autografts by recropping limited donor sites. There were not unusual septic complications, although prophylaxis for opportunistic infections was used. The disadvantage of allograft use is its early rejection and obligatory replacement until permanent coverage with autograft can be accomplished. Cyclosporine A can prolong allograft survival and allow autograft coverage from limited donor sites in a sequential fashion. This may lead to increased survival in patients with massive burns.

Reversal of diabetic somatic neuropathy by whole-pancreas transplantation.

To answer the crucial question regarding reversibility of diabetic somatic neuropathy by whole-pancreas transplantation, metabolic studies and electron microscopic morphometry of the sciatic and testicular nerves were performed monthly for 2 years in three groups of highly inbred rats: (1) NC, 47 nondiabetic controls; (2) DC, 90 untreated alloxan-induced diabetic controls; and (3) DT, 230 diabetic rats given syngeneic pancreaticoduodenal transplants 6, 9, 12, 15, 18, and 21 months after induction of diabetes mellitus (DM). Six diabetic nerve lesions were quantitated by a "blind" protocol: (1) loss of myelinated axons, (2) intraaxonal glycogen deposits, (3) axons with glycogen deposits, (4) demyelinated axons, (5) degenerating axons, and (6) loss of intact axoglial junctions in paranodal terminal myelin loops. In the DT group, testicular nerve specimens were obtained just before transplantation and at death so that each animal served as its own control. As we have observed previously in untreated diabetic controls, all six nerve lesions progressed relentlessly for 2 years, in contrast to nondiabetic controls (p less than 0.01). Whole-pancreas transplants produced complete metabolic control of DM for life and reversed all six lesions in both sciatic and testicular nerves, even when done late in the course of DM. There was complete reversal of the nerve lesions when pancreatic transplantation was done within 15 months of the onset of DM. These results provide the first demonstration of reversal of diabetic somatic neuropathy by any form of DM therapy and extend our previous work in which whole-pancreas transplants were found to prevent both diabetic neuropathy and nephropathy and reverse mesangial enlargement in the kidney.

Effect of pancreas transplantation on diabetic somatic neuropathy.

To determine whether pancreas transplantation is capable of preventing diabetic somatic neuropathy, metabolic studies and electron microscopic morphometry of the sciatic nerve were performed monthly for 2 years in four groups of highly inbred rats: (1) NC-28 nondiabetic controls; (2) DC-82 untreated alloxan-diabetic controls; (3) WPT-122 diabetic rats that received a syngeneic whole-pancreas transplant; and (4) IT-90 diabetic rats that received intraportal injections of 1500 to 2000 syngeneic pancreatic islets. Five diabetic nerve lesions were quantitated by a "blind" protocol: intra-axonal glycogen deposits, axons with glycogen deposits, demyelinated axons, intact axoglial junctions in paranodal terminal myelin loops, and basal lamina thickness of vasa nervorum. Untreated diabetic control animals had significant and progressive increases in all five nerve lesions compared to nondiabetic controls (p less than 0.01). Whole pancreas transplants produced precise metabolic control of diabetes and prevented development and progression of all five diabetic nerve lesions throughout the 2-year study period. Pancreatic islet transplantation produced strict metabolic control and prevented diabetic neuropathy for the first 6 months, but then diabetes recurred and nerve lesions that were similar in severity to those in untreated diabetic rats developed. The finding that whole pancreas transplantation prevents diabetic somatic neuropathy adds to and extends our previous studies showing that whole-pancreas transplants prevent diabetic nephropathy.

Comparison of the metabolic control of diabetes achieved by whole pancreas transplantation and pancreatic islet transplantation in rats.

To compare the long-term effectiveness of whole pancreas transplantation and pancreatic islet transplantation in controlling the metabolic disorders of alloxan diabetes, metabolic studies were performed monthly for 2 years in 4 groups of highly inbred rats: (1) NC-116 nondiabetic controls; (2) DC-273 untreated alloxan-diabetic controls; (3) PDT-182 rats that received syngeneic pancreaticoduodenal transplants shortly after induction of diabetes with alloxan; and (4) IT-92 rats that received an intraportal injection of at least 1500, but usually 2000, syngeneic pancreatic islets shortly after induction of diabetes with alloxan. Whole pancreas transplantation maintained strict metabolic control throughout the 2 years of study. In group PDT, hyperglycemia was abolished; plasma glucose concentration was maintained tightly within the normal range; markedly depressed plasma insulin levels were raised to above normal; glucose tolerance tests had insulin levels above normal and glucose levels that increased less and declined more rapidly than normal; and body weight gain and growth approached normal. In contrast, pancreatic islet transplantation failed to maintain precise metabolic control. In group IT, plasma glucose concentration initially fell to normal but then was elevated significantly above normal beginning with the 3rd posttransplant month; plasma insulin level declined progressively after the 6th posttransplant month; glucose tolerance tests had a diabetic glucose tolerance curve as a result of a markedly deficient plasma insulin response; and body weight gain and growth were significantly less than in group PDT. The results of these long-term metabolic studies may explain the effectiveness of whole pancreas transplantation and the ineffectiveness of pancreatic islet transplantation in preventing diabetic nephropathy.