no evidence of islet autoimmunity. However, this approach has significant limitations driven in part by the nature of the tests used and by the heterogeneity of type 2 diabetes [1]. Of course, the ultimate demonstration of an immune‐mediated cause of diabetes would require documenting an immune infiltrate within islets [2]. This is extremely invasive, has a significant risk of serious complications and for these reasons is almost never undertaken (nor should it).
Several authors have argued persuasively that if the ultimate goal of diabetes management is to achieve glycemic control safely and effectively, then the underlying diagnosis may be less important [1]. In the absence of therapeutic choices this may be less important but given the proliferation of therapeutic classes in diabetes pharmacotherapy a greater understanding of the pathophysiologic abnormalities resulting in hyperglycemia may help optimize therapy. This is perhaps best illustrated by the demonstration that several forms of monogenic diabetes respond to sulfonylureas and do not necessarily need insulin therapy [3]. Early detection and treatment of hemochromatosis may prevent loss of pancreatic islet function and diabetes associated with generalized lipodystrophy responding to leptin therapy serve as other examples.
How does hyperglycemia arise? In the postprandial state glucose concentrations are the net result of stimulation of insulin secretion, suppression of glucagon secretion, the ability of insulin (insulin action) and glucose (glucose effectiveness) to suppress endogenous glucose production and stimulate glucose uptake by the tissues, and the rate of gastric emptying. In type 1 diabetes, all defects are ultimately secondary to insulin deficiency whereas in type 2 diabetes the relative contribution of these parameters is more variable [4, 5].
Defective and/or delayed insulin secretion is the hallmark of all forms of diabetes and the presence of hyperglycemia implies that the degree of insulin secretion is inadequate for the prevailing insulin action [6]. Increasing secretory demands may overwhelm the ability of β‐cells to assemble insulin leading to protein misfolding and a cascade of mechanisms, including the unfolded protein response, that result in dysfunction and if unchecked the death of the β‐cell [7, 8]. Common genetic variation associated with type 2 diabetes has helped to identify multiple pathways associated with the synthesis and secretion of insulin. It is important to note, however, that the genetic predisposition to fasting hyperglycemia differs from that predisposing to glucose intolerance [9, 10]. This is concordant with the observation that impaired fasting glucose can occur independently of impaired glucose tolerance and vice‐versa in prediabetes [11].
Abnormalities of glucagon suppression are observed in both type 1 and type 2 diabetes and were initially attributed to insulin deficiency within the islet [12]. However, insulin restraint of α‐cell secretion may not be as important as previously thought [13, 14]. Certainly there are other paracrine regulators of glucagon secretion [15]. Abnormal glucagon secretion arises early in prediabetes and occurs independently of defects in insulin secretion [14]. Underlining its importance in the pathogenesis of diabetes is the observation that people with diabetes‐associated genetic variation exhibit defects of α‐cell function [16, 17].
The factors altering the ability of insulin and of glucose itself to suppress endogenous glucose production (and release into the circulation) as well as stimulate uptake are less well understood although weight, adiposity and physical activity all influence these parameters. Genetic predisposition to defects in insulin action, for example, is less well characterized. Certain syndromes such as polycystic ovarian syndrome are associated with diabetes through effects on insulin resistance [18, 19].
The final variable affecting glycemic control is upper gastrointestinal function. The stomach and proximal small bowel function in unison to dampen fluctuations in the rate of appearance of ingested calories into the duodenum and jejunum after meals. Multiple neural and hormonal inputs regulate gastric volume and wall tension (allowing accommodation of ingested food), pyloric tone and the rate of gastric emptying [20]. Although this integrated system regulates satiety and, to a lesser extent, caloric intake, there has been no clear association with predisposition to diabetes. For example, most of the benefits of bariatric surgery occur through changes in caloric intake and weight loss [21] and are likely independent of the rate of gastric emptying – outcomes after sleeve gastrectomy and Roux‐en‐Y‐Gastric Bypass (RYGB) are broadly comparable – despite a far higher rate of gastric (pouch) emptying after RYGB [22].
Genetic and environmental influences on disease presentation
Having somewhat downplayed the notion that there are two discrete forms of diabetes, one must acknowledge that the extremes of the disease spectrum are fairly easy to recognize and seem to differ in their genetic predisposition. Both type 1 and type 2 exhibit genetic predisposition to the disease but the environmental contribution to type 2 diabetes is much more prominent [23].
Common genetic variation in more than 200 loci has been associated with type 2 diabetes. The risk conferred by most of these variants is small – indeed knowledge of genetic variation at the 18 loci with greatest effect on disease risk did not appreciably alter the performance of a prediction model utilizing anthropometric information and family history [24]. Although later models incorporating genetic information from additional loci improved their predictive performance, especially in younger adults, it remains apparent that genetic information is unhelpful in predicting type 2 diabetes risk at an individual level [25].
In contrast, genetic variation in the human leukocyte antigen (HLA) confers more than half of the genetic risk of type 1 diabetes. HLA binds and processes antigen‐presentation to the immune system. A handful of other loci involved in immune response pathways confer significant additional risk. Other variants (~50) also contribute smaller effects. Most of the loci associated with type 1 diabetes alter immune regulation [26–28].
The environmental events, if any, that trigger the cascade of immune‐mediated destruction in type 1 diabetes are not well defined. Often hyperglycemia is detected for the first time in the context of a febrile illness, but it is well accepted that significant islet destruction has occurred by the time hyperglycemia develops. In contrast, excess caloric intake and physical inactivity are associated with type 2 diabetes. However, this is not uniform, so that of all patients coming to bariatric surgery in the United States (BMI > 35 kg/m2) only a third have diabetes. The reasons for this heterogeneity are unclear at present [29]. Another interesting observation is the association of shift work and circadian misalignment with type 2 diabetes – this may be mediated through effects on insulin synthesis and secretion. Genetic variation in the melatonin receptor (MTNR1B) – a key part of circadian signaling – is associated with type 2 diabetes [30, 31].
Measuring insulin secretion
Insulin is secreted by the β‐cell in response to various stimuli. However, peripheral insulin concentrations may be an imperfect measure of insulin secretion because they represent the net sum of two opposing processes. The first is actual insulin secretion into the portal vein while the second is hepatic extraction of insulin that occurs across the liver as insulin reaches the systemic circulation. It is uncertain whether hepatic extraction is an active or a passive process, but it is likely proportional to the magnitude of insulin secretion [32] and declines as β‐cell function declines [33].
C‐peptide arises from the post‐translational processing of insulin as preproinsulin which folds upon itself to form specific disulfide bonds, resulting in a dimeric structure after cleavage of the connecting (C‐)peptide. This peptide is secreted in a 1:1 ratio with insulin but does not undergo hepatic extraction. Therefore, in theory, C‐peptide concentrations serve as a better measure of insulin secretion than do insulin concentrations themselves. However, C‐peptide which is cleared by the kidney (and therefore cannot be used reliably in renal dysfunction or failure) has a half‐life of ~30 minutes and therefore accumulates in the circulation compared to insulin (half‐life of < 5minutes). Deconvolution of insulin secretion rates from C‐peptide concentrations requires knowledge of the kinetics underling C‐peptide
