Summary Recommendations for Specific Scientific Programs

• Determine the factors responsible for pancreatic ß-cell growth and development.
• Define the molecular elements that make up the insulin-signaling pathway and the defects in this pathway that contribute to insulin resistance.
• Facilitate procedures to combine the results of all completed genome scans for type 2 diabetes.
• Convene a multidisciplinary planning group to develop a strategy for identifying diabetes and obesity susceptibility genes in various ethnic populations and characterizing their role in the pathophysiology of these conditions.
• Determine whether there are identifiable stages of type 2 diabetes or insulin resistance that can be determined and differentiated on the basis of altered expression of sets of genes or other biochemical or pathophysiologic criteria.

Therapy of Diabetes
Co-Chairs: David M. Nathan, M.D., and James Gavin, M.D., Ph.D.

The therapy of diabetes mellitus, although complex, has become more focused in the past decade. Insight into the pathophysiology of diabetes and its complications has led to the development of improved therapeutic strategies. Metabolic goals that result in improved long-term outcome, and the means of achieving those goals, have been established for type 1 diabetes on the basis of high-quality clinical research such as the Diabetes Control and Complications Trial (DCCT). Clinical data in type 2 diabetes, albeit not as definitive as for type 1 diabetes, support achieving and maintaining the metabolic goal of normoglycemia. (The specific means of achieving those goals, a quantitative estimate of the benefits, and the balance between benefits and adverse events remain to be established for type 2 diabetes.)

In addition to the therapies aimed at achieving glycemic goals that will prevent or delay the development of long-term complications, nonglycemic therapies have been developed and demonstrated to ameliorate long-term complications. Two prime examples of such therapy are treatment of hypertension and early stages of diabetic renal dysfunction with antihypertensive agents, and specifically with angiotensin-converting enzyme inhibitors, and laser photocoagulation of proliferative retinopathy and clinically significant macular edema.

Although cardiovascular disease causes the greatest mortality in diabetes, contributing to the tragically shortened lifespan for persons with both type 1 and type 2 diabetes, relatively little is known about the pathogenesis of macrovascular disease in the setting of diabetes mellitus. Moreover, the clinical studies that have advanced our understanding of the primary prevention of, and secondary intervention in, cardiovascular disease have almost uniformly excluded diabetic patients from participation. Therefore, many of the lessons regarding the beneficial effects of treating hypertension and dyslipidemia on cardiovascular disease have been extrapolated from studies in nondiabetic patients. Direct data are lacking on the benefits and risks in diabetes of specific therapies focused on cardiovascular disease. Nevertheless, when it comes to the treatment of cardiovascular disease risk factors, the consensus has been to treat patients with diabetes at least as aggressively as nondiabetic patients.

The research opportunities identified by the working group are described below.

Improve Adherence to Therapy-Behavioral Approaches
Despite well-publicized goals and "standards" of therapy, many of which are supported by high-quality research data, the implementation of the therapies and achievement of goals is limited at best. Metabolic control in type 1 and type 2 diabetes remains inadequate, the frequency of recommended examinations (feet, eyes, and measurements of glycemia, lipid status, and renal function) is inadequate, and therapy of diabetes and its complications is generally not aggressive. The treatment of diabetes, unique from the treatment of most other diseases, requires extensive patient involvement and affects most aspects of daily activity and lifestyle. Behavioral research has already identified key variables associated with adherence to treatment regimens, strategies for improving patient-care provider interactions, and self-management skills that can promote adherence. In addition, assessment tools that can identify barriers to adherence in individual patients and treatment strategies to lower these barriers have been developed.

The challenge is to improve application of currently available therapies that are known or suspected to be effective in improving the long-term outcome of persons with diabetes. Insight into the behavioral barriers that prevent application of effective therapies and the development of the means to lower those barriers will help implement current and future therapies.

Although the National Diabetes Education Program (NDEP) has launched a program to address these issues, it primarily deals with limitations of available resources, rather than the behavioral barriers that face both health care providers and patients in trying to implement currently available therapies of demonstrated efficacy. An improved appreciation of the behavioral barriers perceived by health care providers and patients and the development of innovative strategies to overcome those barriers is needed in order to maximize the benefit of available therapies. Specifically, two major areas that require increased research effort have been identified.

Achieving desirable outcomes: One area involves maximizing the ability of patients and providers to implement recommended therapy and achieve desirable outcomes.

• Apply behavioral theories and strategies to maximize diabetes self-management.
  • Develop and evaluate strategies that address social and cultural barriers to adherence.
    
  • Study interventions to decrease psychiatric and social comorbidities in individuals with disease (for example, depression, eating disorders, and family dysfunction, and their impact on glycemic control).
    
  • Develop methods to match patients to specific therapies and treatment goals.
    
  • Assess impact of therapies on quality of life and psychological status.
    
  • Develop a means of assessing quality of life in individuals with diabetes.
    
  • Analyze barriers between patients and health care providers and develop means of reducing them.
    
  • Develop means of improving communication between patients and health care providers.
    
  • Evaluate efficacy and cost-effectiveness of alternative models of health care delivery, including technological innovations such as computer-aided management.

Developing a healthy lifestyle: The second area for increased research effort involves maximizing adoption and maintenance of a healthy lifestyle. Obesity, smoking, and a sedentary lifestyle have been clearly identified as major risk factors for the development of diabetes and cardiovascular disease. Although the specific benefits of lifestyle modification on preventing or delaying the development of diabetes and cardiovascular disease are under active study in the diabetes prevention program, it is clear, in general, that adoption of a healthy lifestyle is desirable for a myriad of reasons but difficult to achieve. Previous research has identified effective strategies that improve initial weight loss and adherence to a healthy lifestyle. The following are specific areas identified for future research.

• Improve maintenance of weight loss.
  • Conduct basic studies on maintenance of behavior changes.
    
  • Conduct long-term studies of lifestyle intervention, drug therapy, and combination therapies.
    
  • Develop means of matching patients to treatment regimens.
    
  • Develop means of improving long-term participation in physical activity.
    
• Develop and evaluate new strategies to maximize adoption and long-term maintenance of healthy lifestyles in the general population and in high-risk individuals.
  • Improve eating, exercise, and smoking behaviors in persons at risk for diabetes.
    
  • Study environmental interventions that may improve health, such as modifying school lunches.
    
  • Develop new means of assessing dietary intake and physical activity.

Innovative Therapy for Diabetes
Intensive treatment of type 1 diabetes, although effective in improving long-term outcome, is labor-intensive, is difficult to implement for many patients with type 1 diabetes, and does not achieve normoglycemia. When currently available intensive treatment methods are used, the level of glycemic control achieved, as measured by HbA1c, is at best 4 to 5 standard deviations above the mean value for the nondiabetic population. The implementation of intensive therapy is limited by the accompanying increased frequency of HbA1c severe hypoglycemia. In addition, weight gain that accompanies intensive therapy has resulted in an increasing prevalence of overweight type 1 diabetic patients. Finally, currently available methods are particularly problematic in children and adolescents, which is especially troublesome because our best clinical data support early intervention as being most effective.

The options for treatment of type 2 diabetes have expanded in the past three years. In addition to diet, exercise, sulfonylureas, and insulin, the recent availability of metformin, acarbose, and troglitazone and the potential for numerous combinations have provided new methods to achieve improved glycemia. Each of the medications has advantages and disadvantages, and the optimal medication regimen is unknown; however, the majority of patients with type 2 diabetes ultimately fail dietary therapy and oral agents and require insulin.

The limitations of therapy for type 2 diabetes parallel, in many respects, the problems with treatment of type 1 diabetes: inadequate methods to achieve normoglycemia, which are accompanied by side effects; requirement of considerable levels of adherence that are difficult to achieve in the older population; and nonaggressive application of treatment strategies.

The challenges are to refine currently available intensive therapies of type 1 and type 2 diabetes (directed at achieving normoglycemia), making them more accessible, user-friendly, and safe, and to develop new methods of therapy for both type 1 and type 2 diabetes.

Since many of the methods developed to treat type 1 diabetes could be applied to type 2 diabetes, and vice versa (monitoring methods and artificial pancreas, for example), we have avoided creating working groups within the committee defined by disease type (since we will not be dealing directly with autoimmunity, this should not pose a major problem). Rather, the working subgroups are defined based on a practical division of three fundamentally different approaches aimed at achieving normoglycemia. While we recognize that no artificial categorization of research will encompass all of the possibilities (for example, use of islets as the basis of a biosensor to monitor glucose, so-called "hybrid devices"), the proposed working structure will allow the process to move forward.

Biologic or cellular approaches to metabolic control: Although traditionally viewed as potential therapy for type 1 diabetes, insulin replacement via transplanted tissue could, under appropriate circumstances, be effective in the therapy of type 2 diabetes. Pancreatic transplantation has become increasingly successful in the last 10 to 15 years, with 70 to 90 percent of recipients normoglycemic and insulin-independent at one year post-transplant. It is the single most effective means of establishing normal glucose levels in type 1 diabetes. Islet transplantation has been successful in the setting of autotransplantation (chronic pancreatitis) and allografts in nondiabetics ("cluster" transplants in patients with metastatic cancer); however, fewer than 10 percent of islet transplants in type 1 diabetes have succeeded in eliminating insulin requirements and restoring normoglycemia.

Major research opportunities are listed below.

• Provide islets/ß-cells.
  • Study developmental biology of ß-cells.
    
  • Study use of islet stem cells for allografting.
    
  • Study ß-cell expansion.
    
  • Establish cell lines (human, neuroendocrine) for more refined engineering.
• Protect grafts through immunomodulation/tolerance induction, immuno-isolation of ß-cells and surrogate cells, and in vivo gene delivery (GLP-1, for example).
• Further develop and refine whole organ transplantation.
  • Evaluate the role of transplantation before complications begin.
    
  • Evaluate the cost-benefit ratio of early transplantation for patients with hypoglycemia unawareness or poor quality of life.
    
  • Develop improved protective measures for whole organs.
    
  • Use less toxic drugs.
    
  • Attempt immunomodulation.
    
  • Induce tolerance.

Mechanical approaches to metabolic control: Mechanical approaches to the treatment of diabetes hold great promise for improving metabolic control and quality of life for persons with diabetes. Arguably, the single greatest change in the management of both type 1 and type 2 diabetes in the past two decades has been the introduction and widespread implementation of reliable, accurate, and relatively user-friendly glucose self-monitoring devices. At present, state-of-the-art technology cleanly divides mechanical delivery devices and glucose-sensing technology, but the ultimate goal may be to develop a closed-loop delivery system by combining the two technologies.

• Expand use of mechanical delivery devices.

External insulin pumps have been available for almost 15 years. Although they are used clinically, their widespread application is limited by inconvenience, risk of interrupted delivery secondary to mechanical malfunction, kinking of catheters, and other factors such as tape allergies, skin infections, and the nonphysiologic route of subcutaneous insulin delivery. Totally implantable insulin delivery pumps (IIPs) are in a relatively advanced stage of development. Variable-rate, telemetrically controlled devices that require transcutaneous filling with insulin every one to three months have been extensively investigated. They appear to be extremely safe, delivering insulin intravascularly or intraperitoneally. The IIP therapy achieves metabolic control in type 1 and type 2 diabetes comparable to the levels achieved in the DCCT, with fewer episodes of severe hypoglycemia. Despite these initially promising outcomes, further research is necessary to complete their development.

• Complete clinical trials.
• Move to next generation of pumps with further miniaturization, development of improved catheters, super-concentrated insulin formulations that are compatible with the devices and catheters, do not aggregate, and allow further miniaturization of the pumps.
• Accelerate use of glucose-sensing devices.

Despite the enormous success of glucose self-monitoring, the technical challenges of developing methods for continuous monitoring of blood glucose and several highly publicized industry failures have overshadowed recent progress in the field of glucose sensing. Several approaches to continuous glucose measurement, which are minimally invasive, are close to clinical applicability.

The intended application of a specific sensor must be clearly defined, since it determines the required technology. Potential, important functions include (1) informing patients of high and low blood glucose levels (alarm system); (2) providing glucose data that assist in selection of appropriate doses; (3) closing the loop, or driving an insulin delivery device. Below are specific areas of research in sensors.

• Complete development of sensors.
• Investigate the relationship between blood glucose levels and dynamics in different tissues, including relative concentrations, characteristics of lag times, determinants of partitioning.
• Use innovative, speculative approaches, such as employing isolated islet electrical activity as sensor.
• Develop optical approaches. Two methods are most promising: Fourier-transformed near-infrared spectral analysis using transmitted or reflected light, and measurement of changes in index of reflection. Proof-of-concept research is necessary, focusing on the reliability of such approaches. A major focus of the research should be examining the inter-individual variability conferred by different skin types.
• Conduct independent clinical assessment of any sensing devices developed by industry.
• Integrate sensor and delivery systems and create a true artificial pancreas.
  • Examine servo-control parameters for control of glycemia with insulin delivered peripherally (subcutaneously or intravascularly) or intraperitoneally.
    
  • Analyze safety for clinical use, minimizing risk of over- and under-treatment.
    
  • Develop algorithms.
    
  • Develop interfaces such as hard wiring or telemetry.

Pharmacologic approaches to metabolic control: In general, basic science investigation has advanced understanding of the pathophysiology of diabetes and its abnormal metabolism, providing the substrate for the development of new therapies. In addition to improving the application of established therapies, including the investigation of whether specific therapies and combinations are more effective for specific subgroups of type 2 diabetes, general emphasis needs to be placed on continued basic research. Specifically, the following areas of research should be addressed with the aim of identifying new drug targets:

• ß-cells:
  • Cell biology: apoptosis, proliferation, factors that affect ß-cell mass/growth.
    
  • Insulin gene expression.
    
  • Development of imaging methods to assess mass.
    
  • Investigation of secretagogues, e.g., GLP-1, GIP.
• Insulin resistance:
  • Definition of insulin-signaling cascades and new targets, such as phosphatase inhibitors.
    
  • Compartmentalization.
    
  • Definition of exercise-induced signaling pathways, hypoxia, and other alternative pathways.
    
  • Use of insulin sensitizers such as PPAR agonists, identifying nuclear targets, and cellular mechanism of action.
• Genes involved in diabetes:
  • Characterization of the genetically heterogeneous nature of diabetes, especially as it affects response to specific drugs.
    
  • Pharmacologic manipulation.
    
  • Gene therapy such as IL4, cytokines, and GLUT 4.
• Obesity:
  • Pathogenesis.
    
  • Drug targets such as uncoupling proteins (UCP) and new fl-adrenergic receptors.
    
• Prevention of hypoglycemia in intensive therapy.
• Diabetes management in pregnancy.

Microvascular Complications
Co-Chairs: Michael Brownlee, M.D., and Daniel Porte, Jr., M.D.

In the 1990s, the central therapeutic problem in diabetes mellitus is not management of its acute metabolic derangements but prevention and treatment of its chronic complications. In the United States, diabetes is the leading cause of new blindness in people aged 20 to 74, and it accounts for 35 percent of all new cases of end-stage renal disease (ESRD). Diabetics are the fastest-growing group of renal dialysis and transplant recipients. More than 60 percent of diabetics are affected by neuropathy, which includes distal symmetrical polyneuropathy, mononeuropathies, and a variety of autonomic neuropathies causing erectile dysfunction, urinary incontinence, gastroparesis, and nocturnal diarrhea. Approximately 60 percent of type 2 diabetics have hypertension. Accelerated lower extremity arterial disease in conjunction with neuropathy makes diabetes account for 50 percent of all nontraumatic amputations in the United States. Diabetics have a death rate from coronary heart disease that is 2.5 times that of nondiabetics. Heart disease in diabetics appears earlier in life and is more often fatal.

The Microvascular Breakout Session focused exclusively on the three complications that occur only in hyperglycemic individuals-diabetic retinopathy, diabetic nephropathy, and diabetic neuropathy. Presentations and discussion about the pathogenesis and the genetics of microvascular complications were followed by consideration of each of the three specific microvascular complications. In each area, major accomplishments were identified, and important unanswered research questions were articulated. Suggestions about new or innovative programs that might help address these questions were elicited and are listed at the end of each section of this summary.

Pathogenesis of Microvascular Complications
The Diabetes Control and Complications Trial (DCCT) demonstrated that mean glycated hemoglobin (HbA1c) is the dominant predictor of retinopathy, nephropathy, and neuropathy, and that intensive therapy reduced appearance and progression of these complications by 45 to 78 percent. Several mechanisms by which hyperglycemia may cause microvascular and neurologic damage have been identified. These include increased polyol pathway activity and associated changes in intracellular redox state, increased diacylglycerol synthesis with consequent activation of specific protein kinase C (PKC) isoforms, increased nonenzymatic glycation of both intracellular and extracellular proteins, and increased formation of reactive oxygen species. Tissue injury then results from acute changes in protein function and chronic changes in gene expression. It is likely that the pathologic effectors induced in different tissues by these mechanisms are specific for each particular complication (for example, VEGF in diabetic retinopathy and TGF-ß in diabetic nephropathy). Evidence that abnormal PKC activation, abnormal protein glycation, and increased flux through the polyol pathway play prominent roles in the pathogenesis of microvascular complications is particularly strong in animal models, because a specific PKC isoform inhibitor prevents diabetes-induced changes in retinal blood flow and glomerular filtration rate, while an inhibitor of advanced glycation product formation prevents retinal acellular capillary formation and increased mesangial volume. Both prevent diabetes-induced albuminuria. Similarly, treatment with an aldose reductase inhibitor prevents slowing of nerve conduction velocity. Antioxidants such as vitamin E and lipoic acid also partially prevent several early vascular and neurological abnormalities in diabetic animals.

Although much progress has been made, the possible interrelationships between these various hyperglycemia-accelerated pathways have not been systematically studied. Different mechanisms may play a dominant role at different stages or in different cell types of target tissues. Also, development of interventions for modifying known pathways is limited, and many molecular elements of relevant signaling pathways are unknown.

Several specific areas for research were identified:

• Conduct clinical trials of agents that prevent, stabilize, or reverse complications in animals. These trials are urgently needed.
• Conduct studies to determine the sequence of events in the pathogenesis of hyperglycemia-induced injury.
• Develop new transgenic and gene deletion animal models to elucidate the elements involved in hyperglycemia-induced activation/inhibition of gene expression.
• Conduct studies of genetic modulators of the susceptibility of tissues to hyperglycemia-induced injury. These will require central resources for familial characterization and banking of material for genetic studies.

Genetics of Microvascular Complications
At present, there is considerable evidence for a genetic basis for susceptibility to diabetic nephropathy. At least three studies have shown that the incidence of persistent proteinuria rises during the early years of type 1 diabetes and then declines. In one study, 82 percent of diabetic siblings of probands with nephropathy had either end-stage renal disease or persistent albuminuria, while only 17 percent of siblings of probands free of diabetic nephropathy had significant clinical disease. Similarly, the cumulative risk of nephropathy after 25 years was that 71.5 percent in the proband had significant proteinuria and 25.4 percent in the proband did not. Similar concordance rates are observed in type 2 diabetes, although limited studies of the natural history of nephropathy in these patients may make identification of susceptibility genes more difficult.

In a nondiabetic genetically hypertensive rat model that develops early and severe renal disease, genotyping studies have identified a major 50cM locus named Rf-1 that has a very high LOD score for the renal failure phenotype, but no correlation with blood pressure. Using such animal studies as a paradigm, important genetic loci may be identified in future human studies.

Despite this progress, many challenges still remain to be overcome. For example, data to support a genetic predisposition for retinopathy are very limited. For neuropathy, very few family studies have been conducted. It is difficult to collect sibling pairs concordant for complications for genome scanning approaches and to collect parents of affected offspring for transmission disequilibrium approaches. The possible polygenic nature of complications susceptibility could make genome scanning studies of families unrealistic.

The following specific areas for research were identified:

• Develop centralized resources for the collection of additional sibling pairs concordant for nephropathy or the collection of both parents and an affected offspring.
• Conduct systematic studies of familial/ethnic aggregation of proliferative diabetic retinopathy and diabetic neuropathy using these resources.
• Develop better surrogate endpoints (intermediate phenotypes) to identify subjects for genetic analysis earlier.

Diabetic Retinopathy
The earliest retinal abnormalities induced by diabetes are impaired autoregulation of retinal blood flow and decreased retinal blood flow, associated with loss of endothelial-supporting pericytes from retinal capillaries. These changes are followed by endothelial cell permeability changes and retinal capillary occlusion (perhaps by leukostasis of activated monocytes or accumulation of extracellular matrix), which result in the classic signs of nonproliferative retinopathy, dot and blot retinal hemorrhages, hard exudates, cotton-wool spots (nerve fiber layer infarctions), and retinal microaneurysms. When the degree of retinal nonperfusion due to capillary occlusion becomes extensive, vasoproliferative factors (predominantly VEGF and to a lesser extent insulin-like growth factor [IGF-1]) are produced that stimulate new vessel formation on the retina, into the vitreous and, in severe cases, on the iris and the anterior chamber angle. This proliferative phase of diabetic retinopathy includes a glial component that can contract and lead to retinal detachment. Macular edema, which can develop at either stage of diabetic retinopathy, results from a severe leakage of fluid from retinal capillaries mediated by increased VEGF production and endothelial cell dysfunction. VEGF production can be stimulated by hyperglycemia, advanced glycation end products, reactive oxygen species, and IGF-1 in vitro. VEGF effects on endothelial cells appear to be mediated by activation of PKC pathways.

The DCCT trial demonstrated that reducing hyperglycemia will reduce the incidence and progression of diabetic retinopathy in type 1 diabetics, and smaller studies suggest the same is true for type 2 diabetics. Photocoagulation is highly effective in reducing the risk of blindness in proliferative disease if treatment occurs early enough. Hypertension has been shown to be an independent risk factor for diabetic retinopathy by the Wisconsin study, and the EUCLID study suggests that antihypertensives, particularly angiotensin- converting enzyme inhibitors, may have a protective effect. Hypercholesterolemia may be an additional risk factor as well, and excessive growth hormone action also has been postulated to play a role.

While considerable progress has been made, there are significant challenges to progress in this area. The initiating mechanism for background retinopathy and the major stimulus for VEGF production in vivo in proliferative retinopathy are not clear. The essential role of growth hormone and other growth factors in the pathogenesis of diabetic retinopathy remains to be fully evaluated. Promising drug therapies for the treatment and prevention of proliferative retinopathy are limited by the lack of efficient means of selective drug delivery to the eye. Fifty percent of diabetic patients do not receive appropriate ophthalmologic care, resulting in unnecessary vision loss. Mechanisms for interactions among government, academia, and industry are not well developed.

The following specific areas for research were identified:

• Expand basic studies on mechanisms of initial injury in experimental retinopathy.
• Establish a national database of consenting patients, updated with regard to diabetes and retinal status, and of clinical trials of the most promising pharmacologic agents.
• Develop technologies for delivering drugs selectively to the eye, particularly antiproliferative agents. These include methods to achieve timed-release, prolonged action, increased solubility/permeability, and improved implantation materials.
• Develop innovative approaches such as telemedicine technologies for more effective screening of diabetic eyes by primary care physicians. These technologies include image acquisition technologies, storage capacities, and comparative scanning software.
• Develop new mechanisms to facilitate and expand industry/academic/NIH/Food and Drug Administration collaboration in clinical trials.

Diabetic Nephropathy
The earliest renal abnormality induced by diabetes in type 1 diabetes patients is glomerular hyperfiltration that is partially reversible by correction of hyperglycemia. Thickening of the capillary basement membrane and expansion of the glomerular mesangial matrix occur after three to five years of diabetes, in part as a result of TGF-ß overproduction. In 25 to 40 percent of type 1 patients, fixed microalbuminuria develops after 5 to 15 years of diabetes.

Microalbuminuria is a critical marker and perhaps also a determinant of future decline in renal function, as well as a marker for increased risk of other complications. Microalbuminuria evolves into macroalbuminuria, perhaps due to glycation-induced changes in extracellular matrix molecules, which is accompanied by a progressively decreasing glomerular filtration rate due to capillary occlusion by expanding mesangial matrix. The clinical course in type 2 diabetics appears to be similar, and there is an equal risk of developing nephropathy in the two types. Intraglomerular hypertension has been postulated to play a role in the pathogenesis of nephropathy, and treatment with antihypertensives, particularly angiotensin-converting enzyme inhibitors, reduces both albuminuria and the rate of decline of renal function. The DCCT demonstrated that reducing hyperglycemia reduces the incidence and progression of diabetic nephropathy in type 1 patients, and experimental pancreas transplantation can reverse extracellular matrix accumulation. Familial and ethnic group studies suggest that genetic influences play a critical role in determining which patients progress to ESRD. In animal models, various manifestations of diabetic nephropathy are prevented by isoform-specific PKC inhibitors, glycation product inhibitors, antihypertensives, and antioxidants. The rate of survival of diabetics treated for end-stage renal disease with dialysis is lower than that of nondiabetics, while survival of diabetics treated by transplantation is similar to the rate for nondiabetics.

Several challenges to progress in this area were identified. Current surrogate endpoints of developing glomerular pathology are either not specific enough (microalbuminuria) or not sensitive enough (glomerular filtration rate) for either genetic studies or intervention studies. Molecular elements of relevant signaling pathways are largely unknown. The strong genetic factors underlying diabetic nephropathy are currently unknown. The reason microalbuminuria is a predictor of other vascular disease is unknown. The reasons for reduced survival in diabetics with ESRD treated with dialysis are not known.

Specific areas for research are listed below.

• Identify sensitive and specific indicators of developing glomerular pathology, and then validate them in clinical trials.
• Pursue investigations of the fundamental cellular mechanisms involved in hyperglycemia-induced glomerular dysfunction. A promising approach is high-throughput molecular biology technologies for quantifying changes in gene expression and identifying the genes involved.
• Develop collaborative multicenter studies of the genetic basis of nephropathy in both type 1 and type 2 diabetes.

Diabetic Neuropathy
Diabetes affects almost every part of the somatic peripheral and autonomic nervous system. Diabetes may damage small fibers, large fibers, or both. Small nerve fiber dysfunction typically occurs first in the lower extremities, where it is manifested by loss of thermal sensitivity and pin prick. Pain and paresthesias are often present. Large fiber damage is characterized by reduced vibration perception, light touch and position sense, weakness, and depressed tendon reflexes. Autonomic neuropathy causes disturbances in cardiovascular, gastrointestinal, genitourinary, and pseudomotor function. In diabetic peripheral sensory neuropathy, both types of fibers are typically involved, with axonal degeneration in a spatial distribution that suggests multifocal ischemia. Capillary closure is frequently observed in the vasa nervorum, and endoneurial blood flow and oxygen tension are reduced.

The DCCT established unequivocally that the effects of inadequate insulin action (as monitored by the level of hyperglycemia) are associated with the incidence of diabetic neuropathy. The intensively treated groups had an overall incidence reduction of 64 percent. The polyol pathway is one pathogenic factor in experimental diabetic neuropathy, and treatment with aldose reductase inhibitors prevents long-term neuropathic changes in diabetic dogs. Aminoguanidine, a glycation-product inhibitor that also inhibits NOS, prevents diabetes-induced changes in neuronal blood flow and concomitant electrophysiologic abnormalities. Antioxidants (alpha-lipoic acid, taurine), essential fatty acid supplements (gamma-linoleic acid), prostenoid analogues, cAMP analogues, osmolytes (myo-inositol), and neurotrophic agents have also shown promise in animal models of diabetic peripheral neuropathy.

Axonal regrowth and myelination after injury is markedly impaired in diabetic animals, and reduced gene expression or axonal transport is observed for neurotropins such as nerve growth factor and neurotrophin-3 (NT-3). There is reduced gene expression of IGF-1 as well.

Defects in the neural response to hypoglycemia are a major side effect of intensive insulin treatment and can be reproduced in part by one or two episodes of hypoglycemia in normal man. In addition, the effects of repeated severe hypoglycemic episodes on cognitive function have been fairly well established. It is not yet clear whether the hypoglycemia associated with intensive insulin therapy is deleterious, but its presence is the major limitation to the application of intensive insulin therapy.

Specific challenges to research in this area were identified. There is no defined area within NIH having diabetic neuropathy as its primary focus. The potential interrelationships among mechanisms implicated in hyperglycemia-induced nerve damage are not understood. The molecular elements of relevant signaling pathways are largely unknown. Standardized measures of small fiber sensory perception are unsatisfactory. The clinical potential of pharmacological approaches showing promise in experimental animals is unknown. Mechanisms by which chronic, repeated hyperglycemia and hypoglycemia may damage the central nervous system have not been systematically studied.

The following specific areas for research were identified.

• Establish a mechanism to focus on diabetic nerve disease at NIH to evaluate, promote, and be responsible for multidisciplinary approaches to diabetic neuropathy, combining the neurobiology of peripheral nerve cells with the pathobiochemistry and pathophysiology of diabetic nerve disease.
• Pursue investigations of the fundamental cellular mechanisms involved in hyperglycemia-induced peripheral nerve dysfunction. A promising approach is high-throughput molecular biology technologies for quantifying changes in gene expression.
• Develop standardized surrogate endpoints for sensory function in man.
• Conduct short-term studies in diabetic patients, followed by a long-term, multicenter primary prevention/secondary intervention study with clinically significant endpoints to establish the therapeutic potential of promising agents.
• Develop neuron-specific vectors for targeted neurotropin therapy.
• Study systematically the mechanisms by which chronic, repeated hyperglycemia and hypoglycemia damage the central nervous system.

Macrovascular Complications
Co-Chairs: Willa A. Hsueh, M.D., and Antonio M. Gotto, Jr., M.D.

The association between clinically recognized diabetes mellitus and an increased risk for cardiovascular disease (CVD) has been documented in diabetic populations throughout the world: Among diabetic patients, morbidity and mortality rates from myocardial infarction are increased, as are the risks for recurrent infarction, congestive heart failure, cerebrovascular disease, and peripheral vascular disease. It is well established that heart and blood vessel diseases are the most common cause of death in the diabetic population. The heavy burden of atherosclerotic disease in diabetic patients necessitates that CVD prevention become an important focus for research and public health policy in diabetes care.

Many strides have been made in elucidating the relation between the metabolic derangements of diabetes and the increased risk for the development of macrovascular disease (i.e., stroke, heart attack, and peripheral arterial disease). The same risk factors that predict large vessel disease in the general population also affect the diabetic. In general, about 50 percent of excess heart disease in diabetics can be attributed to associated abnormalities in other known CVD risk factors. However, diabetic status appears to confer risk such that even after correction of other risk factors, the diabetic remains at high risk for macrovascular disease. Although absolute rates of CVD vary widely, the presence of diabetes increases CVD risk two- to fourfold compared with rates in the corresponding nondiabetic population. Indeed, the excess risk is so high that being diabetic is considered by some groups to place a patient in the same risk category as having established coronary and other atherosclerotic disease. Furthermore, the increased risk for CVD associated with diabetes is relatively greater in women than in men. Also, the increased prevalence of diabetes in several U.S. minority populations indicates these individuals may also be at high risk for macrovascular disease. The degree of risk is, at least in part, influenced by the degree of diabetic control. Hyperglycemia and its associated metabolic abnormalities appear to increase the risk further. Thus, diabetes may increase CVD risk by both direct and indirect effects.

Type 1 diabetes is a leading cause of premature cardiovascular disease, and type 2 diabetes is a significant risk factor for cardiovascular disease in middle and older ages. As the US population ages, the importance of glucose intolerance as a risk factor for CVD will increase since its prevalence increases markedly with age, affecting more than 50 percent of those older than age 65. Preliminary data indicate that diabetes is one of the most important CVD risk factors in the elderly. Moreover, as methods are developed to reduce chronic microvascular complications, cardiovascular complications will become an increasingly important source of morbidity and mortality among those with diabetes. Thus, it is critical to determine the efficacy and appropriateness of treatment protocols in all diabetics and in subgroups of diabetic patients. For example, the most appropriate treatment to prevent macrovascular complications in a young, insulin-dependent diabetic with sustained, poorly controlled hyperglycemia may not be the best for an elderly noninsulin- dependent diabetic with mild hyperglycemia. More information is needed in all of these areas to design better and more specific treatment and prevention regimens.

In contrast with extensive studies on other major CVD risk factors such as elevated blood pressure and cholesterol, evaluations of the role of glucose intolerance and its related abnormalities as causes of CVD have been limited. Available data on hyperglycemia as a risk factor for CVD have numerous limitations. The group agreed that it is time to develop new programs to improve our understanding of how diabetes increases the risk for CVD and to test treatments to reduce this risk.

The group discussed opportunities for two broad categories of studies: (1) laboratory investigations to improve understanding of the factors that contribute to the excess macrovascular disease seen in diabetics and (2) clinical studies to test the ability of currently available interventions to reduce macrovascular complications.

Laboratory Investigations
Improved molecular and cellular biological techniques and the ability to manipulate genes in animals have tremendously advanced our knowledge of the factors and processes involved in the development of atherosclerosis. Steps involved in atherosclerosis include (1) endothelial cell damage with loss of the vascular protective effects of nitric oxide (NO), (2) platelet adherence to damaged areas with clot formation and release of growth factors and cytokines, (3) inflammatory reaction with increased production of adhesion molecules and interaction of circulating cells with the vascular wall, (4) incorporation of lipid into the vascular wall, and (5) vascular smooth muscle cell (VSMC) migration to the intima, proliferation, and extracellular matrix production. Many long-recognized risk factors for coronary artery disease (CAD) such as hypertension, low HDL cholesterol, and high triglycerides, are prevalent in type 2 diabetics. In recent years, other factors have been identified that appear to contribute to the increased risk for CAD.

Hyperinsulinemia and increased resistance to insulin action appear to contribute to CVD risk, especially among type 2 diabetics. The role of insulin in atherogenesis is complex: Insulin stimulates VSMC migration, growth, and matrix production and enhanced clot formation, while insulin resistance is associated with impaired NO production. Even in the absence of diabetes, patients with the insulin-resistance syndrome, a constellation of metabolic abnormalities associated with increased insulin resistance, are at significantly increased risk for CAD. Insulin resistance is associated with hypertriglyceridemia, which, in turn, is associated with a preponderance of small, dense LDL. This LDL phenotype is perhaps more susceptible to oxidative modification, a process that promotes atherosclerosis by impairing endothelial cell function, stimulating inflammation and adhesion, and promoting VSMC changes. Diabetes is also associated with hemostatic abnormalities, increases in several clotting factors, increased platelet adhesion, enhanced activity of lipoxygenase pathway, and decreased lysis of fibrin clots.

Hyperglycemia itself appears to affect multiple mechanisms that increase atherosclerosis. Hyperglycemia enhances oxidation, thrombosis, inflammation, matrix production, and the formation of advanced glycosylated end products and other metabolites that potentially can damage the vasculature. Paradoxically, intensive treatment for hyperglycemia can also have potentially adverse effects upon cardiovascular risk factor levels. Data from the DCCT demonstrate that the increased weight associated with tight control was also associated with increased intra-abdominal fat, triglycerides, total and LDL cholesterol, apolipoprotein B, and blood pressure and with decreased HDL cholesterol and apolipoprotein A1. Such effects may explain why it has been difficult to prove that normalizing glucose levels will eliminate the excess risk for macrovascular disease seen in diabetic patients.

A major impediment to improved understanding of how diabetes increases CVD risk is the lack of adequate animal models to explore the link between diabetes and atherosclerosis. The committee strongly recommended additional resources for the development and utilization of such models, including animals with specific genetic defects and the exploration of how nutritional variation alters the CVD risk associated with diabetes.

Given the toll of macrovascular disease in diabetic patients, a key consideration for future planning is the integration of the work on atherosclerosis with that on diabetes. One promising avenue is the number of animal models that have been developed in recent years that appear to mimic one or more components of the atherosclerotic lesion in man. These include the apolipoprotein E knockout mouse, the fat-fed LDL receptor knockout mouse, and the miniature swine. Eight varieties of genetic alterations in the mouse lead to some form of vascular injury. Preliminary data in which animals have been made diabetic (by administration of streptozotocin) or insulin resistant (by administration of a high- fructose diet) suggest that these alterations advance the atherosclerotic process. There are also animal models of diabetes. In addition to the NOD model of type 1 diabetes, there are a number of natural and manipulated gene models of insulin resistance and type 2 diabetes. With the possible exception of the miniature swine, however, there are no good animal models to study the accelerated atherosclerosis of diabetes. Additional models are needed to facilitate study of diabetic cardiovascular complications. It is critical and timely to proceed with crossing the atherosclerotic models with models of type 1 and type 2 diabetes to understand how the atherosclerotic process is accelerated in the presence of diabetes and insulin resistance.

Important Issues to Address in Diabetes-CVD Animal Models
How do hyperglycemia and glycation alter endothelial and VSMC behavior in response to vascular injury?

How does insulin affect vascular cells at the molecular and physiologic levels?

How do specific alterations in circulating lipids, particularly remnant particles, affect the development and expansion of the atherosclerotic lesion and the interaction of lipids with the arterial wall?

How is vascular inflammation altered or enhanced in the patient with diabetes?

How does increased oxidation in the presence of insulin resistance and diabetes affect atherosclerosis?

How do hyperglycemia and associated metabolic abnormalities alter hemostasis and interactions at the surface of unstable atherosclerotic plaques?

How do antidiabetic drugs affect the vasculature and coagulation system, and what impact do they have on atherosclerosis development and progression?

Answering these questions will require (1) extensive molecular, cellular, histologic, and physiologic characterization and comparison with human lesions, (2) multidisciplinary collaborative efforts, (3) core animal facilities for consistent production and ease of distribution of tissues, as well as facilities for accurate measurement of lipids, lipoproteins, and clotting factors, and (4) development of consistent reproducible approaches to assess atherosclerosis extent and progression in small and large animal models, both invasively and noninvasively. Research in this area will provide a critical understanding of the atherosclerotic process in diabetes and a means by which to test potential interventions by diet, exercise, or pharmacologic or genetic manipulation.

Clinical Trials and Observational Studies
Observational data collected over the past 40 years in the U.S. population have indicated that risk factor reduction has produced a decrease in morbidity and mortality from CAD and stroke. Evidence from subgroup analyses in these populations suggests that the diabetic patient also has the same relative decrease in risk. Since this group is at a higher risk of heart disease, therapeutic interventions that benefit nondiabetics can be antici-pated to have even greater absolute benefit for the diabetic patient.

Limited data from clinical trials confirm that diabetic patients appear to benefit from reduction of conventional CAD risk factors. Diabetics accounted for 10 to 14 percent of two large clinical trials involving the class of cholesterol-lowering drugs called HMGCoA reductase inhibitors, or statins. In the Scandinavian Simvastatin Survival Study (4S), simvastatin was given to men and women who had coronary disease and high levels of low-density lipoprotein (LDL) cholesterol, while in the Cholesterol and Recurrent Events (CARE) trial, pravastatin was given to men and women who had suffered a myocardial infarction, but who had only average levels of cholesterol (mean total cholesterol of 209 mg/dL). In both studies the diabetics experienced similar decreases in CAD events and LDL cholesterol concentrations compared with the overall group, suggesting statin therapy as a beneficial strategy in the treatment of diabetics with established CAD. However, these represent observations drawn from subgroup analyses; large trials are currently underway investigating the effect of statins, fibric acid derivatives (or fibrates, another class of lipid-modifying agents), or a combination of the two on CAD risk in diabetic populations. Similar benefits have been observed with blood pressure control. In the Systolic Hypertension in the Elderly Program (SHEP), blood pressure lowering with thiazides showed a greater absolute benefit in stroke and cardiovascular disease morbidity and mortality in diabetic compared with nondiabetic participants.

What has not been adequately examined in a clinical trial is whether the control of hyperglycemia and insulin resistance will affect CAD incidence and which glucose-lowering drugs may maximize such a benefit. While there is observational evidence that lower levels of glycosylated hemoglobin are associated with a lower incidence of CAD, manipulation of glycosylated hemoglobin concentrations to reduce CAD risk has never been tested or established in a clinical trial. In the University Group Diabetes Program (UGDP), there was an increase in CAD events with sulfonylurea treatment. In the DCCT, a beneficial trend for total CVD was noted though insufficient numbers were available to detect a significant effect on CAD events. Results from a recent Veterans Administration pilot study of intensive treatment of hyperglycemia in type 2 diabetics failed to suggest any short-term benefit of improved glucose control. The UK Prospective Diabetes Study (UKPDS) is scheduled to report its findings in 1998. Given the suggested benefits of lipid-lowering and antihypertensive therapy in diabetic patients, the interventions to control glucose and insulin resistance likely will be in addition to statins, antihyper-tensives, or other background therapy.

Addressing this question is particularly timely in view of the multiple choices of drugs to control glycemia that are now available for treatment of type 2 diabetes. In addition to insulin, four oral agents are available for lowering blood sugar levels. Three of these compounds were introduced within the last two years. Insulin has long been a treatment for type 2 diabetes; however, patients with type 2 diabetes are generally insulin resistant and hyperinsulinemic, particularly at the early stages of their disease. A major unresolved question is whether increasing insulin levels with exogenous insulin alters atherosclerosis progression. Current data suggest that in middle-aged white males, hyperinsulinemia is a risk factor for CAD; this issue is unclear in women, African Americans, and Hispanics and other ethnic groups. A related question is whether lowering levels of insulin resistance will affect progression of atherosclerosis.

Several classes of drugs with different mechanisms of action can be used to lower glucose in the type 2 diabetic patient. Metformin is a biguanide that suppresses hepatic glucose production and has a modest effect to improve insulin action in skeletal muscle. Metformin decreases circulating LDL cholesterol and triglyceride levels and increases HDL cholesterol slightly. Its use is associated with weight loss and improvement in insulin resistance. Sulfonylureas, which stimulate pancreatic insulin secretion, have been available for many decades for the treatment of type 2 diabetes. They increase circulating insulin levels and are associated with weight gain. In addition, they bind to vascular potassium channels that induce vasoconstriction, an action that may have adverse effects on the coronary arteries of patients with atherosclerotic disease. Thiazolidinediones are insulin-sensitizing agents that improve insulin-mediated glucose uptake into skeletal muscle and improve many components of the insulin resistance syndrome, such as hypertriglyceridemia, low HDL cholesterol, hypertension, and high circulating free fatty acids. In cultured cell systems and in animals, these agents impair VSMC growth and migration by inhibiting the mitogen activated protein kinase (MAPK) pathway through activation of the novel steroid receptor, peroxisomal proliferating activating receptor (PPARg), which has recently been identified in VSMC. Thus, this class of agents could have vascular protective effects. Since we now can individualize therapy targeted at glucose control, it is critical to investigate the optimal therapeutic strategies for prediabetic and diabetic patients who are at marked increase risk for CAD.

1. The panel enthusiastically recommended a large-scale clinical trial to determine whether improving glucose control and reducing levels of insulin and insulin resistance will decrease the incidence of CAD in a diabetic population.
2. It was the subcommittee's recommendation that such a trial be jointly supported by the NIDDK, by the NHLBI, and by industry.
3. In order to foster the development of such trials, the subcommittee felt very strongly that it was important for clinical trials to be supported by the NIH and not by industry alone. It was further recommended that an incentive be given to pharmaceutical companies for participating with NIH in the cofunding of trials. The most apparent way to accomplish this would be through an extension of the five-year exclusivity agreement for drugs, with limited patent protection. A demonstration project might be carried out to test the efficacy of such an approach.
4. The subcommittee recommended that the NIH support a trial to test the optimal forms of revascularization in the diabetic heart disease patient.

   The risk factors for a clinical CVD event and the most appropriate preventive treatments are not necessarily the same for all diabetic patients. Recent clinical trial data suggest that surgical treatment of CAD warrants careful consideration in the diabetic patient. In a subgroup of the Bypass Angioplasty Revascularization Investigation (BARI) trial, overall survival following percutaneous transluminal coronary angioplasty (PTCA) was compared with survival following coronary artery bypass grafting (CABG), and a very high mortality rate was observed with the diabetic participants who had undergone PTCA. The sub- group difference was very striking and contrasted with that in nondiabetic participants: The diabetic participants who had PTCA had ten times the mortality of the diabetic participants who had a left internal mammary artery grafted to a coronary artery. The clinical implications of this finding are important: Hundreds of thousands of coronary bypass surgeries and angioplasties are performed on CAD patients each year in the United States, and many of those patients are diabetic.

5. The subcommittee recommended that these studies incorporate measurements to determine the association of genes with risk for coronary heart disease, cerebrovascular disease, and peripheral arterial disease.

   In addition to clinical trials, attempts are underway to recruit diabetic patients and multiple other family members for studies to discover the genetics of diabetes. It is important to recognize that genetic factors may help to explain the wide variation in CVD risk among diabetic patients from different populations around the world. An effort must be made to determine which genes are involved in increasing the risk for macrovascular disease in the diabetic population, whether these are the same as the genes affecting risk in the nondiabetic population, and whether some are the same genes that increase the risk for diabetes per se.