Diabetes Mellitus Part 2

Retinopathy

Given a long enough duration, retinopathy occurs in almost all patients with type 1 diabetes mellitus and in most patients with type 2 diabetes mellitus who are on conventional treatment that does not come close to normalizing glycemic levels [see Table 2].21 The most common form of retinopathy is nonproliferative retinopathy (also termed background retinopathy). It begins with loss of capillary pericytes, the supporting cells of the retinal vasculature, a loss leading to capillary dilatations that are seen on direct fundoscopy as microaneurysms [see Figure 6a]. Microaneurysms measure 50 to100 |im in diameter and can occur anywhere in the retina. However, they tend to cluster near the macula, the area responsible for central vision and visual acuity. Small dot hemorrhages form when microaneurysms leak blood. Hard lipid exudates form on leakage of serum [see Figure 6a]. These lesions are usually benign unless they occur quite close to the macula and in sufficient number to cause clinically significant macular edema. The latter is a feared complication that can decrease central vision and acuity. Capillary closure, which actually begins in the phase of background retinopathy, increases; and in the phase of preproliferative retinopathy, enough capillaries become obstructed to cause ischemia of the retina. Infarctions of the retinal nerve layer appear as soft (cotton wool) exu-dates. The retina responds to further ischemia with proliferation of new blood vessels from its surface [see Figure 6b]. In this phase of proliferative retinopathy, ischemic retina releases vascular en-dothelial growth factor (VEGF), which stimulates new vessel formation. These new vessels grow forward into the vitreous. They are extremely fragile and can bleed into the vitreous, causing temporary loss of vision until the blood is reabsorbed. If no reab-sorption occurs, blindness can result unless successful vitrecto-my is carried out. Proliferative vessels that cover more than one fourth of the disk diameter and that occur within 1 disk diameter of the disk [see Figure 6b] are especially likely to bleed. Even after reabsorption of the vitreous blood, fibrous scars form that can cause traction on the retina and can lead to retinal detachment, another cause of profound and often permanent loss of vision.

Nephropathy

Diabetic nephropathy [see Figure 7] is the complication associated with the highest mortality. Between 35% and 45% of patients with type 1 diabetes mellitus and a somewhat smaller percentage of patients with type 2 diabetes mellitus experience significant nephropathy.22-24 Histologically, the earliest change is thickening of the capillary basement membrane. Subsequently, mesangial material accumulates diffusely throughout the glo-merulus [see Figure #]. Ultimately, there is loss of podocytes and development of peritubular fibrosis. Excretion of low but abnormal levels of albumin in the urine is a marker of the incipient phase of nephropathy.25 As glomeruli become increasingly filled with mesangial matrix products, albuminuria increases and eventually gross proteinuria appears. Microalbuminuria is defined as excretion of 30 to 300 mg of albumin a day or an albu-min-creatinine ratio between 30 and 300 in a random urine specimen. Clinical proteinuria is defined as excretion of more than 0.5 g of total protein a day. This level of excretion can be detected by a positive dipstick urine test for protein. The nephrotic syndrome may also eventually occur.

Figure 4 The cellular actions of insulin begin with binding to its plasma membrane receptor. As a result, certain tyrosine molecules in the intracellular portion of the transmembrane receptor are autophosphorylated, creating tyrosine kinase activity in the receptor. Several intracellular insulin receptor substrates (IRS) are then tyrosine phosphorylated by the receptor. Phosphorylated IRS docks and either activates or inactivates numerous enzymes (e.g., phosphatidylinositol-3-kinase [PI-3 kinase]) and other mediating molecules. Among the chief effects of these insulin-stimulated cascades are translocation of glucose (Glut-4) transporters to the plasma membrane, where they facilitate glucose diffusion into the cell; shifting of intracellular glucose metabolism toward storage as glycogen by activating glycogen synthase; stimulation of cellular uptake of amino acids, phosphate, potassium, and magnesium; stimulation of protein synthesis and inhibition of proteolysis; and regulation of gene expression via insulin regulatory elements (IRE) in target DNA molecules. Numerous intermediates in these various pathways, along with the molecules mentioned above, are products of candidate genes whose mutation could produce the state of insulin resistance characteristic of type 2 diabetes mellitus. Red connectors between insulin chains A and B and among insulin receptor subunits a and 3 indicate S-S bonds. The A chain also has an intramolecular S-S bond.

Early in type 1 diabetes mellitus, kidney size and glomerular filtration rate (GFR) may actually be greater than normal. However, in both types of diabetes, GFR begins to decline, and after clinical proteinuria develops, GFR almost inexorably falls to the level of ESRD [see Figure 7]. Unlike the risk of retinopathy, the risk of nephropathy does not continue to rise with increasing duration. The incidence of nephropathy peaks at approximately 15 to 17 years and declines somewhat thereafter.26 The prevalence of nephropathy remains approximately constant after that time. If the dipstick test has not revealed proteinuria by 25 to 30 years of diabetes duration, the risk of ESRD decreases. Coincident with or shortly after the development of microalbuminuria, hypertension often appears. Hypertension in turn further aggravates diabetic nephropathy and is an important component in the progression to renal failure.

Figure 5 The diagram represents the gold standard for measuring the sensitivity of glucose metabolism to insulin, utilizing a glucose insulin clamp. When steady state is reached, glucose metabolized/unit time = glucose infused/unit time. Assuming endogenous glucose production is suppressed to zero, insulin sensitivity = (glucose metabolized/unit time) – plasma insulin. For each dose of insulin, the more exogenous glucose required to sustain plasma glucose at its basal levels, the greater the insulin sensitivity. Conversely, individuals who require lesser amounts of glucose than usual to maintain the basal plasma glucose level are insulin resistant. The latter is usually the case in type 2 diabetes mellitus.

Neuropathy

Neuropathy has protean manifestations in diabetes. The most common presentation is peripheral symmetrical sensorimotor neuropathy, which causes numbness or tingling in the toes and feet.27 At this point, symptoms are only mildly disturbing and require no specific treatment. These symptoms may even abate over time as neuropathy becomes more severe and hypoesthesia or anesthesia takes the place of paresthesias and dysesthesias. Ultimately, insensate feet become very vulnerable to trauma, and neuropathic foot ulcers are frequent causes of hospitaliza-tion and even amputation. Testing sensation with a nylon monofilament providing a calibrated 10 g point pressure is an effective way to screen for high risk of foot ulcers. Patients who cannot detect the pressure of the nylon filament have a 30- to 40fold increased risk of foot ulcer.28 In some instances, neuropathy is manifested by severe pain that can interfere with sleep and normal daily activities. The distribution of pain can suggest mononeuropathy and radiculopathy. Abrupt onset of cranial neuropathies that most commonly give rise to extraocular muscle weakness and diplopia has been attributed to microinfarcts caused by thrombosis of nutrient blood vessels. Carpal tunnel syndrome and other entrapment syndromes are more frequent in diabetic patients than in nondiabetic patients.

Involvement of the autonomic nervous system is also common and can become debilitating. Manifestations include male impotence and female anorgasmia, difficulty voiding and urinary retention, impaired gastric emptying with early satiety and emesis, diarrhea, orthostatic hypotension, and decreased sweating and vasomotor tone in the lower extremities. The combination of decreased sympathetic tone and loss of vagal control of the heart rate can produce persistent resting sinus tachycardia; sudden death can result.

A form of diabetic neuropathy called amyotrophy occurs most commonly in elderly men with diabetes. It is manifested by severe, unremitting pain and weakness in the thigh muscles. Severe depression, cachexia, and weight loss may mark the 1- to 2-year course of this form of neuropathy. Sometimes confused with painful neuropathy are rare muscle infarcts, usually occurring in the thigh muscles. These infarcts are marked by abrupt onset of severe pain lasting several months. Magnetic resonance imaging of the affected area can demonstrate the presence of necrosis.

Diabetic neuropathy may be another microvascular complication, but the pathogenesis is still not completely understood.29 De-myelinization of nerves is manifested by decreases in motor and sensory nerve conduction velocities. Axonal degeneration is reflected in decreased amplitudes of action potentials. Histological-ly, swelling is seen at the axonal nodes. An inflammatory component to diabetic neuropathy has also been suggested.30

Relation of microvascular complications to glycemia

The appearance of microvascular complications in the 1930s generated a 50-year debate about whether diabetic retinopathy, nephropathy, and neuropathy were the direct result of the metabolic abnormalities, most notably hyperglycemia, or whether they were a parallel independent consequence of diabetes that had formerly been usually preempted by death from extreme metabolic disequilibrium (i.e., diabetic coma). This debate ultimately came to encompass type 2 diabetes mellitus as well. The debate was not merely academic, because it was reflected in quite different approaches to treatment. A belief in the metabolic hyper-glycemic cause of retinopathy, nephropathy, and neuropathy impelled the physician to work with inadequate means to help the patient achieve as close to normal blood glucose levels as possible. Conversely, a belief in the metabolically independent nature of these complications encouraged a somewhat more laissez-faire approach, which attempted primarily to eliminate the immediate symptoms, such as polyuria, that were produced by plasma glucose levels exceeding the renal threshold (> 180 mg/dl). Furthermore, the risks associated with the more aggressive approach to hyperglycemia reinforced the arguments of the conservative practitioners. A large body of evidence was eventually built up that supported but did not prove the so-called glucose hypothe-sis.31 The Diabetes Control and Complications Trial (DCCT) and United Kingdom Prospective Diabetes Study (UKPDS) ended this debate for type 1 and type 2 diabetes mellitus, respectively.

Table 2 Diabetic Retinopathy

Stage*	Pathologic Process	Manifestations
Background	Loss of capillary integrity	Microaneurysms
	Leakage, exudation, diapedesis	Dot hemorrhages
		Hard exudates
	Early capillary closure	Macular edema
Preproliferative		Blot hemorrhages
	Capillary closure	Soft exudates
	Microinfarcts	Intraretinal microvas-cular abnormalities
	Ischemia
		Venous beading
		Macular edema
Proliferative	Forward growth of new	Preretinal hemorrhage
	large vessels Fibrosis	Vitreous hemorrhage
	Traction on retina or vitreous	Retinal detachment
		Macular edema

*Loss of visual acuity may occur from macular edema at any stage. Blindness may occur from severe macular edema, vitreous hemorrhage, or retinal detachment.

Figure 6 (a) This fundus photograph reveals nonproliferative (or background) retinopathy in a diabetic patient. Microaneurysms (arrows) occur at end capillaries. Punctate (or dot-and-blot) hemorrhages (H) and hard exudates (C) can also be seen. The hard exudates form three distinct circles (termed circinate retinopathy), which indicate leakage of plasma proteins from abnormal vessels located in the centers of the three circles. Lesions in the area of the macula (M) are potentially more dangerous, as they may lead to macular edema requiring laser therapy. (b) In prolif-erative retinopathy, new vessels grow from the retina into the vitreous. This fundus photograph reveals fine, tangled, new vessels originating from several areas of the disk (arrows). The vessels often form arcades and characteristically have thin walls and are fragile. They tend to bleed into the vitreous; the scars that form can cause retinal detachment and loss of vision. Proliferation within one disk diameter of the disk (termed neovascularization of the disk) is particularly dangerous, as these vessels are especially prone to bleed and form traction scars.

The DCCT32 was a randomized clinical trial that enrolled 1,441 nonobese patients, aged 13 to 39 years, with type 1 diabetes mellitus. Half of the patients with diabetes of 1 to 5 years’ duration participated in a primary prevention trial that excluded all patients with retinopathy or microalbuminuria, and half of the patients with diabetes of 1 to 15 years’ duration participated in a secondary intervention trial that included only patients who already had mild to moderate nonproliferative diabetic retinopathy but less than 200 mg/day of urinary albumin excretion. In both of these DCCT trials, patients were randomly assigned either to receive conventional treatment (no more than two insulin injections a day) or to receive intensive treatment (three to four insulin injections a day or use of a continuous subcutaneous insulin infusion [CSII] pump; self-monitoring of blood glucose at least four times a day; premeal target blood glucose levels of 70 to 120 mg/dl; glycated hemoglobin [HbA1c] goal of less than 6.05%; and very frequent contacts between patient and treatment team). An HbA1c difference of 1.8% (8.9% versus 7.1%) was maintained between the two treatment groups for up to 9 years.

Over a mean follow-up of 6.5 years, intensive treatment produced substantial benefits. The risks of de novo development (primary prevention trial) or of progression (secondary intervention trial) of retinopathy were reduced by 27% to 76%; the development of microalbuminuria was reduced by 35%; macroalbu-minuria (i.e., proteinuria) was reduced by 56%; and development of clinical neuropathy, confirmed by abnormal nerve conduction velocities or autonomic nervous system function tests, was reduced by 60%.32 Patients in the primary prevention cohort, with a mean diabetes duration of 2.5 years, had a greater response to intensive treatment than did patients in the secondary prevention cohort, with a mean diabetes duration of 8.5 years.

Figure 7 Relation of the developing histopathologic changes in the kidney to the development of renal functional abnormalities. Note that GFR is actually elevated early, corresponding to early renal hypertrophy. The appearance of microalbuminuria (albumin excretion > 30 mg/day) indicates that the patient is at considerable risk for overt nephropathy and end-stage renal disease (ESRD), but not all such individuals suffer this fate. Blood pressure begins to rise at about the time that microalbuminuria appears, and hypertension further damages the kidney.

Figure 8 (a) The normal glomerulus with a large filtration surface has a lacy appearance. (b) There is diffuse deposition of extramesangial material throughout, as well as thickening of capillary basement membranes in a diabetic glomerulus. The GFR through such a glomerulus is reduced.

The main adverse effect of intensive treatment was a threefold increase in the risk of severe hypoglycemic episodes characterized by coma, convulsions, or the required assistance of others to treat and reverse the episode.32,34 At least one such event per year was experienced by 25% of intensively treated patients, and 50% had experienced more than one such episode by the end of the study34;14% experienced 10 or more episodes. The overall rate of severe hypoglycemia was 62 events per 100 patient-years for intensive treatment, compared with 19 events per 100 patient-years for conventional treatment. In addition, intensive treatment caused greater weight gain; one third of the patients exceeded 120% of ideal body weight (approximate BMI, 27) by the end of the study.32 Intensive treatment was also more expensive than conventional treatment.35 However, the cost was partly offset by projected decreased costs of a lower rate of complications,36 and the estimated cost per year of quality life gained was $28,661, a figure thought to represent a good value.

The UKPDS37’38 enrolled 5,102 patients with newly diagnosed type 2 diabetes mellitus, a mean age of 53 years, and a mean BMI of 28. After a 3-month dietary run-in, 1,138 patients were randomly assigned to a continuation of diet treatment only as long as their FPG remained below 270 mg/dl and they had no hyperglycemic symptoms. In the study, 2,729 patients were randomly assigned to intensive treatment, 1,573 to receive one of three sulfonylurea (SU) drugs, and 1,156 to receive insulin. In two thirds of the clinical sites, 342 patients were also randomized to intensive treatment with metformin. The goal of intensive treatment was an FPG of less than 108 mg/dl. Of the conventional-treatment patients, 80% ultimately required drugs to maintain their treatment goals of an FPG of less than 270 mg/dl and freedom from symptoms, although nearly 60% of their total treatment time was spent on diet therapy alone. Likewise, in the intensive-treatment groups, metformin therapy had to be added to the SU therapy, and insulin had to be substituted for or added to oral-drug therapy to maintain the stringent treatment goal.

Despite these drug crossovers, after 10 years of follow-up, patients who received intensive treatment showed a 25% decrease in the risk of serious microvascular complications (vitreous hemorrhage, need for laser treatment, and renal failure), compared with patients given conventional treatment.37 This important benefit was associated with an HbAlc difference of 0.9% (7.9% for conventional therapy; 7.0% for intensive therapy). Serious hypoglycemia occurred in 3% of insulin-treated patients each year and in 1% to 2% of SU-treated patients. These rates were much lower than that experienced with intensive treatment in patients with type 1 diabetes mellitus in the DCCT.

These two trials provided experimental proof that microvascular and neuropathic complications could be prevented or at least substantially delayed by maintaining blood glucose levels as near to normal as treatment techniques would safely allow. Although these two experimental trials did not prove that hyperglycemia caused microvascular complications, both trials provided additional strong evidence supporting that hypothesis. In the DCCT, the risk of retinopathy was directly related to the preceding mean HbA1c difference in a similar exponential fashion in each of the two treatment groups.39 The risk of retinopathy was decreased by about 44% for each proportional 10% decrease in HbA1c (e.g., a decrease in HbA1c from 10% to 9.0%). Microalbuminuria and neuropathy showed similar risk relations with glycemia. In the UKPDS, the risk of microvascular complications was also directly related to the mean HbA1c in an exponential fashion.40 The risk of these complications was decreased by about 37% for every absolute decrease of 1% in HbA1c. These similarities suggest that similar biologic processes are at work. Neither the UKPDS nor the DCCT analyses indicated any glycemic threshold in the diabetic range of HbA1c, below which there was no further risk of microvascular complications.40,41 This observation sets normo-glycemia as the ultimate goal of treating type 1 and type 2 diabetes mellitus. Furthermore, the benefits of previous intensive treatment (or the adverse effects of previous conventional treatment) are still demonstrable 7 years after the DCCT was completed, during which time interval the mean HbA1c concentrations in both groups were nearly identical (approximately 8.0%).42 Thus, sustained periods of glycemic exposure are associated with prolonged consequences. An unacceptable level of hyperglycemia continues to have adverse effects even after some improvement in metabolic control, and a marked reduction in hyper-glycemia with intensive treatment continues to have beneficial effects even after some worsening in metabolic control.

Multiple mechanisms by which increased glucose concentrations may cause damage to the retina, kidney, and nerves have been discovered [see Figure 9]. (1) Glucose itself can react nonenzymatically with free amino groups in N-terminal amino acids and lysine residues of proteins. HbA1cis one such molecule. This reaction sets into motion cross-linking of proteins that ultimately generate harmful advanced glycation end products (AGEs).43,44 Such products include carboxymethyllysine and pentosidine. Concentrations of long-lived AGEs were higher in tissues of conventionally treated patients in the DCCT than in tissues of intensively treated patients in the DCCT.45 AGEs correlated with HbA1c and, independent of HbA1c, with the presence of retinopathy, nephropathy, and neuropathy.45 (2) Three-carbon dicarbonyl products of glucose and lipid metabolism, glyoxal and methylglyoxal, also react readily with amino groups in proteins and produce other AGEs, one of which is argpyrimidine. AGEs react with specific cellular receptors and can stimulate numerous potentially dangerous processes.43,44 (3) Hyperglycemia can also secondarily produce oxidative stress in tissues, with depletion of glu-tathione and formation of reactive oxygen species and damaging free radicals.46 (4) When glucose is insufficiently metabolized by insulin-stimulated routes [see Figure 1], it can overflow into the sorbitol (polyol) pathway via the enzymes aldose re-ductase and sorbitol dehydrogenase.47 Accumulation of sor-bitol and fructose in vulnerable tissues such as nerves produces osmotic damage, loss of myoinositol essential to nerve membrane integrity, and reduction of Na+, K+-ATPase activity.47 (5) Elevated glucose levels increase protein kinase C, an enzyme whose activity influences numerous cellular processes with damaging potential,48 such as stimulating neovascularization and epithelial cell proliferation, increasing collagen synthesis, increasing vascular permeability, increasing apoptosis (programmed cell death), increasing oxidative stress, and mediating the actions of VEGF and transforming growth factor-3. (6) Elevated glucose levels also increase the production of VEGF, a molecule that stimulates angiogenesis. VEGF is present in high concentrations in human diabetic ocular tissues and in kidneys of animals with experimentally produced diabetes. It is a logical candidate to mediate development of proliferative retinopa-thy. (7) Hyperglycemia stimulates nitric oxide synthase to produce nitric oxide, a molecule that itself generates damaging free radicals.46 (8) Excess blood glucose also overflows into the hexosamine pathway, resulting in deleterious products.49 A single mitochondrial defect that leads to overproduction of reactive oxygen species can result in at least three of the above pathways and has been proposed as the primary culprit.50 A number of these pathways are also mutually reinforcing, setting up vicious circles that can accelerate tissue damage.

The therapeutic importance of elucidating the mechanistic links between hyperglycemia and microvascular/neuropathic complications lies in our current inability to normalize blood glucose consistently. Therefore, drug therapies that intercept pathogenetic processes downstream from glucose hold promise for preventing these complications, even in the presence of hy-perglycemia. An inhibitor of AGE formation, aminoguanidine, has been successful in animal experiments, but human trials have revealed unacceptable toxicity. Several inhibitors of aldose reductase, catalyzing the first step in the polyol pathway, have been studied in clinical trials, but none have shown sufficient clinical benefit or an acceptable adverse-effect profile to warrant approval in the United States. Nonetheless, such drugs have been effective in animal models. Current clinical trials are testing the effects of antioxidants such as vitamin E and a relatively non-toxic oral inhibitor of protein kinase C. Antagonists to VEGF and other growth factors to be administered by systemic or local injection are also in development.

Figure 9 Multiple pathways have been described that may link high blood glucose levels to the microvascular and neuropathic complications of diabetes (see text). There are good reasons to believe that genetic factors, possibly operating through such pathways, may explain the observation that some individuals with consistently high blood glucose levels do not experience complications, whereas other individuals with near-normal blood glucose levels do experience complications.