Postprandial glucose regulation: New data andnew implications

Postprandial Glucose Regulation: New Data and New Implications Lawrence A. Leiter, MD1; Antonio Ceriello, MD2;Jaime A. Davidson, MD3; Markolf Hanefeld, MD4; Louis Monnier, MDS; David R. Owens, MD6; Naoko Tajima, MD7; and Jaakko Tuomilehto, MD8; for the International Prandial Glucose Regulation (PGR) Study Group* 1St. Michael's Hospital, Universityof Toronto, Toronto, Canada; 2University of Udine, Udine, Italy; 3Medical City Hospital, Dallas, Texas; 4Technical University, Dresden, Germany; SUniversity Institute off Clinical Research, Montpellie6 France; 6Llandough Hospital, Penarth, and Cardiff~UniversityMedical School, Cardiff, Wales; 7Jikei UniversitySchool ofiMedicine, Tokyo,Japan; and 8National Public Health Institute, Helsinki, Finland ABSTRACT

Background: Type 2 diabetes is characterized by a gradual decline in insulin secretion in response to nutrient loads; hence, it is primarily a disorder of postprandial glucose (PPG) regulation. However, physicians continue to rely on fasting plasma glucose (FPG) and glycosylated hemoglobin (HbAlc) to guide management. Objectives: The objectives of this article are to review current data on postprandial hyperglycemia and to assess whether, and how, management of type 2 diabetes should change to reflect new clinical findings. Methods: Articles were selected from MEDLINE searches (key words: postprandial glucose, postprandial hyperglycemia, and cardiovascular disease) and from our personal reference files, with emphasis on the contribution of postprandial hyperglycemia to overall glycemic load or cardiovascular (CV) risk. Results: About 33% of people diagnosed as having type 2 diabetes based on postprandial hyperglycemia have normal FPG. PPG contributes >_70% to the total glycemic load in patients who are fairly well controlled (HbAlc <7.3%). Furthermore, there is a linear relationship between the risk of CV death and the 2-hour oral glucose tolerance test (OGTT). Increased mortality is evident at OGTT levels of -90 mg/dL (5 retool/L), which is well below current definitions of type 2 diabetes. Biphasic insulin aspart was shown to be more effective at reducing HbAlc below currently recommended levels than basal insulin glargine (66% vs 40%; P < 0.001), and it reduced endothelial dysfunction more effectively than regular insulin (P < 0.01). Repaglinide achieved regression of carotid atheroscle-

rosis (intima-media thickness) in 52% of patients versus 18% for glyburide (P < 0.01) over I year, although levels of HbAlc and CV risk factors were similar for both treatment groups. Finally, acarbose reduced the relative risk of CV events by 49% over 3.3 years versus placebo in patients with impaired glucose tolerance (2.2% vs 4.7%; P = 0.03) and by 35% over >_1year in patients with type 2 diabetes (9.4% vs 6.1%; P = 0.006). Conclusions: All components of the glucose triad (ie, FPG, HbAlc , and PPG) should be considered in the management of type 2 diabetes. Therapy targeted at PPG has been shown to improve glucose control and to reduce the progression of atherosclerosis and CV events; therefore, physicians should consider monitoring and targeting PPG, as well as HbAlc and FPG, in patients with type 2 diabetes. (Clin Ther. 2005;27[Suppl B]:$42-$56) Copyright © 2005 Excerpta Medica, Inc. Key words: type 2 diabetes, postprandial glucose, postprandial hyperglycemia, cardiovascular disease.

INTRODUCTION Evidence is increasing in support of a more prominent role for postprandial glucose (PPG) regulation in the management of type 2 diabetes. A common feature of *Members of the International Prandial Glucose Regulation (PRG) Study Group appear in the Acknowledgments.

Accepted for publication July 6, 2005. doi:l 0.1016/j.dinthera.2005.11.020 0149-2918/05/$19.00 Printed in the USA. Reproduction in whole or part is nor permitted. Copyright © 2005 ExcerptaMedica, Inc.

type 2 diabetes is impaired PPG regulation; thus an approach that targets this deficiency, in addition to fasting plasma glucose (FPG) and glycosylated hemoglobin (HbAlc) , is logical and is finding increased clinical support. Data indicating that postprandial hyperglycemia contributes >70% to overall glycemic load in patients who are fairly well controlled (HbA~c <7.3%) 1 lend support to the concept. Casual PPG levels provide clinically relevant guidance to glycemic control, 2 and therapies that target postprandial hyperglycemia (eg, rapid-acting insulin analogues, which reach maximum concentration approximately twice as fast as regular human insulin 3) can lower HbAlc as effectively as or more effectively than therapies that target FPG (eg, long-acting basal insulins). 4,5 For example, in an open-label, parallel-group, 28-week study involving 233 patients with type 2 diabetes, 4 HbAlc was significantly lower (6.9% vs 7.4%; P < 0.01) in patients receiving twice-daily injections of biphasic insulin aspart 30 (BIAsp 30), an analogue premix containing 30% soluble biphasic insulin aspart and 70% protamine co-crystallized insulin aspart, than those receiving bedtime insulin glargine, a long-acting basal insulin analogue. Studies show that rapid-acting insulin analogues can reduce oxidative stress in the arteries, 6 improve endothelial dysfunction,7 and reduce levels of o~-dicarbonyls. 8 Therapies that lower PPG levels also have been shown to slow and even reverse the progression of atherosclerosis 9,1° and reduce the risk of cardiovascular (CV) events.11,12 The CV risks of postprandial hyperglycemia (eg, myocardial infarction [MI], stroke, and CV death) start at much lower glucose levels (90 mg/dL [5 mmol/L])13 than the risk for microvascular complications, such as retinopathy (200 mg/dL [11.1 mmol/L]),14,~5 on which current guidelines primarily are based.16,17 Physicians, therefore, may want to consider PPG targets in the management of their patients with type 2 diabetes, as well as HbAlc and FPG, and also include a therapeutic strategy aimed at reducing postprandial hyperglycemia. The objectives of this article are to review current data on postprandial hyperglycemia and to assess whether, and how, management of type 2 diabetes should change to reflect new clinical findings.

METHODS Articles were selected from MEDLINE searches (key words: postprandial glucose, postprandial hypeglycemia, and cardiovascular disease) and from our personal ref-

erence files. Emphasis was given to studies published since the release of the 2001 American Diabetes Association (ADA) Consensus Statement 18 that discuss the contribution of postprandial hyperglycemia to overall glycemic load or CV risk. Large pivotal studies published before the consensus statement (eg, the Framingham Study, the United Kingdom Prospective Diabetes Study [UKPDS], and the Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe [DECODE] study) were also included.

RESU LTS Glycemic Control and CV Complications The Framingham Study was one of the first to indicate that diabetes is associated with increased CV risk and that, after adjusting for associated risk factors, diabetes carries a 2- to 3-fold increased risk of CV disease (CVD). 19 Other studies have had similar findings,2°,21 confirming a relationship between hyperglycemia (elevated HbA~c level) and mortality. In a Finnish population sample of 229 patients aged 65 to 74 years with type 2 diabetes, Kuusisto et al 2° found that HbA~c >7.9% tripled the risk of coronary heart disease (CHD) over 3.5 years compared with HbAlc <6.0%, increasing the incidence of CHD from 8% to 22% (P < 0.05). The epidemiologic analysis of UKPDS 21 found HbAlc to be a statistically significant predictor of ischemic heart disease and other diabetic complications, with a 14% increment in the risk of MI for each 1% increment in HbAlc (P < 0.001). That being said, the interventional arm of UKPDS 22 did not achieve a statistically significant reduction in MI with intensive glycemic control (P < 0.05).

Postprandial Hyperglycemia and HbAlc With evidence clearly indicating the importance of optimal glycemic control, the goal of diabetes management is to reduce the long-term complications associated with diabetes, and this goal now is included in national and international treatment guidelines. 16,23-25 Until the mid-1990s, research and the available therapeutic armamentarium indicated that FPG was the day-to-day target for diabetes management. With the availability of therapies that target postprandial hyperglycemia, however, interest in understanding the relative contributions of both FPG and PPG levels to glycemic load has been stimulated. Yet data on whether or not targeting postprandial hyperglycemia would make a large enough contribution

to HbA k had been contradictory until the publication of a study by Monnier et al in 2003.1 The outpatient clinic study by Monnier et al 1 involved 290 consecutive type 2 diabetes patients admitted to the clinic following an overnight fast and subsequently given standardized meals. Results of the study indicated that the contribution of PPG to glycemic load varied according to the degree of glycemic control. In poorly controlled patients (HbAlc >10.2%), PPG contributed only 30% of the 24-hour AUC. In better controlled patients (HbAlc <7.3%), the contribution of PPG increased to ->70%. Between these levels of control (ie, HbAlc 7.3% to 10.2%), PPG and FPG made more or less equivalent contributions to glycemic load. Results of the study suggest that PPG could be a clinically relevant guide to glycemic control and that therapy should be tailored to manage both FPG and PPG according to the degree of glycemic control (Figure 1). A study conducted by E1-Kebbi et al, 2 involving 1827 patients with type 2 diabetes visiting the Grady Diabetes Clinic in Atlanta, Georgia, between 1991 and 1998, found that casual PPG reflected overall glycemic con-

trol. In contrast to the controlled conditions of the study by Monnier et al, 1 in which patients were served standard meals at regular times, in this study, HbAI~ was correlated with casual clinic PPG measurements taken 1 to 4 hours postmeal (Figure 2). Results of this study indicated that casual postmeal PPG had a strong linear correlation with HbA1c (r = 0.63; P < 0.001). A cutoff of 150 mg/dL (8 mmol/L) had a predictive value (calculated by receiver-operating characteristic analysis) of ~80% for poor control fie, HbA k >7.0%). The predictive value of casual measurements was strongest in insulintreated patients. The results of these 2 studies suggest that PPG monitoring can be used as a reliable guide in the management of individuals with type 2 diabetes. In fact, El-Kebbi et al suggest that casual PPG measurements, which are generally more easily available and more current, may help overcome the clinical inertia that may result from a delay in obtaining HbAlc results.

PPG Regulation and Glycemic Control Given the substantial contribution of postprandial hyperglycemia to overall glycemia, data are emerging suggesting that, in some patients, approaches that also • Postprandialglucose [] Fastinghyperglycemia

60cO

n

.23

40-

¢0

L..) 20-

0-

I (<7.3)

2 (7.3-8.4)

3 (8.5-9,2)

4 (9.3-10.2)

5 (>10.2)

HbA1c Quintile Figure 1. Contribution of postprandial glucose to glycemic load according to glycosylatecl hemoglobin (HbAlc) quintile (N = 290). Postprandial glucose contribution to 24-hour AUC increases with glycemic control, from 30% in poorly controlled patients (n = 58; HbAlc >10.2%) to _>70% in better controlled patients (n = 58; HbAlc <7.3%). Adapted with permission. 1

12.0

11.0

"% 10.0 <~ I C

9.0

8,0

7.0 50

I

100

I

150

I

200

I

250

I

300

I

350

I

400

Mean Casual PPG (mg/dL) Figure 2. Relationship between glycosylated hemoglobin (HbAlc) and casual postprandial glucose (PPG) levels. In a clinical setting, mean HbAlc had a strong linear correlation with mean casual PPG measurements taken 1 to 4 hours postmeal (N = 1827). Adapted with permission. 2

target PPG regulation may be more effective at achieving therapeutic goals. 4,s Of particular interest are the rapid-acting insulin analogues, which provide a shorter time to maximum concentration (within 40-50 minutes) and a more physiologically accurate insulin profile than conventional human insulin. Also of interest are the meglitinides and 0~-glucosidase inhibitors. A recent study by Raskin et al a found that BIAsp 30 injected BID was more effective at reducing HbA~c below currently recommended levels than bedtime insulin glargine, a long-acting basal insulin analogue. The open-label, parallel-group study involved 233 insulinnaive type 2 diabetes patients poorly controlled on metformin administered either alone or in combination. Patients were randomized to receive 10 to 12 units of insulin glargine OD at bedtime or 5 to 6 units of BIAsp 30 BID before breakfast and dinner. For both treatments, doses were titrated based on a treat-totarget regimen. After 28 weeks, HbAlc was lower in the BIAsp 30 group (6.9% vs 7.4%; P < 0.01). Furthermore, significantly more patients on BIAsp 30 met the ADA target goal of <7.0% for HbAlc (66% vs 40%; P < 0.001) and the American College of Endocrinology target goal of <6.5% (42% vs 28%; P < 0.05). The reduction in FPG was the same in both groups,

indicating that control was achieved by preferentially reducing PPG levels. Minor hypoglycemia (blood glucose <56 mg/dL [<3.1 mmol/L] without neurologic symptoms requiring assistance) was more frequent in the BIAsp 30 group than in the glargine group (3.4 vs 0.7 episodes/patient per year, respectively; P < 0.05). Only i major episode of hypoglycemia (an episode with neurologic symptoms requiring assistance) was reported in a subject in the glargine group. Similar results were achieved in a study comparing BIAsp 30 with neutral protamine Hagedorn (NPH) insulin, s In this 16-week, international, multicenter, double-blind trial in 403 people with type 2 diabetes not adequately controlled by oral agents and/or NPH insulin, patients were randomized to receive BIAsp 30 (n = 201) or NPH insulin (n = 202) immediately before breakfast and the evening meal. Oral agents were discontinued at the start of the study. Among the patients who switched from NPH monotherapy OD or BID to BIAsp 30 BID (n = 66), BIAsp 30 was significantly more effective than NPH insulin at lowering HbAlc (0.78% vs 0.58%, respectively; P = 0.03). This study supports intensifying therapy with an insulin analogue premix rather than adding a second dose of NPH insulin, which the authors note is a common

therapeutic strategy in patients poorly controlled on a single dose of NPH insulin.

PPG and CV Risk Apart from its contribution to overall glycemic control, postprandial hyperglycemia also has become a focus of attention owing to the emergence of evidence that it is a CV risk factor in its own right. Suggestions for a link between CVD and postprandial hyperglycemia first emerged from epidemiologic studies, including the Framingham Offspring Study,26 the DECODE study, 13,27and the Funagata Diabetes Study.28 Framingham offspring data involving 3370 subjects suggest that, for every 40 mg/dL (2.1 retool/L) increase in 2-hour glucose levels at baseline, the risk of CHD, stroke, or peripheral vascular disease increased by 12% to 42%. 26 The DECODE analysis of data from 25,364 individuals reported that hazard ratios for death in individuals not previously known as diabetic and with normal FPG increased significantly compared with individuals with normal glucose tolerance (NGT) as 2-hour glucose levels increased (P < 0 . 0 0 1 ) . 27 Results from the Funagata Diabetes Study, 28 a Japanese cohort study involving 2651 individuals, were similar to those of the DECODE study. Over 7 years, the presence of impaired glucose tolerance (IGT) doubled the risk of CV death compared with NGT, from 0.17% to 0.52% (odds ratio, 2.303; P = 0.046), but fasting hyperglycemia had no effect on CV mortality (P = 0.908). One of the most important findings to emerge from the DECODE/DECODA (Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Asia) studies 27,29 was that -33 % of people classified as having type 2 diabetes based on postprandial hyperglycemia and >25% of people with IGT had normal FPG; thus, FPG measurements alone are an unreliable guide to identifying individuals at CV risk due to IGT or diabetes. Similar results were reported by Woerle et al. 3° The researchers systematically collected data from individuals in their clinic who underwent both 2-hour glucose and HbAlc testing between 1986 and 2002. They reported that 82% of the 132 patients found to have IGT after taking an oral glucose tolerance test (OGTT) had normal FPG levels. Pathophysiologic Mechanisms Several pathophysiologic mechanisms are emerging that could provide a causative link between postprandial hyperglycemia and CVD. As in the case of hyper-

tension, it is likely that both acute and chronic mechanisms are involved. As discussed earlier, chronically, PPG loads make a substantial contribution to overall glycemic exposure that increases as control improves. 1 A number of mechanisms may be responsible for acute toxicity of postprandial hyperglycemia to vascular tissues, including oxidative stress generation and, directly or indirectly, the activation of inflammatory mediators. Evidence is increasing that oxidative stress is central to the pathology of both type 2 diabetes and CVD. 31,32 During an overload of nutrients, the corresponding increase in citric acid cycle activity generates an excess of mitochondrial nicotinamide adenine dinucleotide and reactive oxygen species. 33 Cells that are dependent on insulin-regulated glucose transporters protect themselves by developing insulin resistance; however, endothelial cells are unable to protect themselves because glucose uptake occurs via facilitative diffusion. 34 Overexposed endothelial cells cannot downregulate the influx of nutrients and are particularly vulnerable to overfeeding and oxidative stress. 34 Oxidative stress contributes to endothelial dysfunction, an early manifestation of atherosclerosis, and it is known that endothelial dysfunction associated with oxidative stress increases the risk of CVD. 35,36 In a prospective, single-clinic study of 281 patients referred for assessment for coronary artery disease and followed for a mean of 4.5 years, the presence of a vitamin C effect on acetylcholine-induced vasodilation increased the relative risk (RR) of a CV event by 17% (P = 0.001). 36 Preclinical and clinical data indicate that acute hyperglycemia induces endothelial dysfunction37-39 and that this phenomenon is mediated by oxidative stress. 4°~2 For example, in vivo data from a rat heart perfusion model have indicated that coronary perfusion pressure is significantly increased by an acute glucose load (P < 0.001 vs control) and that this effect can be prevented by the antioxidant glutathione. 42 In a clinical setting, Ceriello et a143 reported that a high-carbohydrate meal that produced a greater degree of hyperglycemia than a low-carbohydrate meal (306 mg/dL [17 mmol/L] vs 234 mg/dL [13 mmol/L] at 120 minutes; P = 0.001) significantly reduced antioxidant capacity in 10 patients with type 2 diabetes. In this study, although plasma levels of total radical-trapping antioxidant decreased after both meals, the decrease was significantly greater after the high-carbohydrate meal (P =

0.001). Thus, oxidative stress leading to endothelial dysfunction appears to be a key pathophysiologic explanation for the observed relationship between postprandial hyperglycemia and CVD (Figure 3). 32 The proinflammatory response triggered by postprandial hyperglycemia and oxidative stress is wideranging. The presence of reactive oxygen species upregulates the production of advanced glycation end products (AGEs). These inflammatory mediators raise levels of C-reactive protein (CRP) and tumor necrosis factor-o~, and their increased presence is associated with the vascular complications of diabetes. 44,45 In a study of 48 patients with type 1 diabetes, skin AGE levels measured by enzyme-linked immunosorbent assay were progressively higher in patients with microalbuminuria and macroalbuminuria than in those with normal renal status (P < 0.001 across all groups). They were also higher in patients with proliferative retinopathy compared with those with less severe retinopathy (P < 0 . 0 0 4 ) . 44 The presence of reactive oxygen species also activates pathways regulated by the transcription factor nuclear factor-kappa[3 (NF-~c[3), which is known to have a central role in the pathogenesis of late diabetic complications. 46 NF-~[3 activation was reported to occur in response to acute, clamp-induced hyperglycemia in

a single-center glycemic clamp study involving 23 nondiabetic volunteers. 47 Acute hyperglycemia also increases production of highly reactive c~-dicarbonyls such as 3-deoxyglycosone (3-DG) and methylglyoxal (MG). 8 A 4-month crossover study of 21 patients with type 1 diabetes found that blood glucose levels correlated with levels of PPG (r = 0.55; P < 0.001) but not with HbAlc (r = 0.01; P = 0.95). 8 Another single-center clinical study involving patients with diabetes (n = 150) and normal, healthy controls (n = 21) found that glyoxalase I activity was significantly increased in diabetic patients overall (P < 0.001 vs healthy controls) and was further increased in those with complications (P < 0.05). 48

Postprandial Hyperglycemia and Atherosclerotic Markers The emergence of mechanisms linking postprandial hyperglycemia to vascular disease finds support in clinical data focusing on atherosclerotic markers. Early work from the Risk Factors in IGT for Atherosclerosis and Diabetes study 49 found that carotid intima-media thickness (IMT) in 582 people at risk for type 2 diabetes was more strongly correlated with 2-hour posfload glucose levels than with HbAlc or FPG (age/sex-adjusted correlation coefficients of 0.211,

I Hyperglycemia I Mitochondria Polyol pathway AGE Formation Hexosamineflux

t iN°SeN°s t NO

Adhesion molecules ,, Proinflammatory ~

ONOO-

""~Nitrotyrosine

/ Endothelialdysfunction . Diabeticcomplications Figure 3. Oxidative stress and pathogenesis of diabetic complications. Superoxide (0-2) overproduction produces an inducible nitric oxide synthase (iNOS) and endothelial nitric oxide synthase (eNOS) uncoupled state, further favoring the production ofO- 2 rather than nitric oxide (NO), even though this condition i eventually results in total NO overproduction. AGE = advanced glycation end product; ONOO- = i peroxynitrite. Adapted with permission. 31 I ..........................................................................................................................................................................................................................

]

0.123, and 0.10, respectively). This result, somewhat unexpected at the time, has found support since then in several studies. Sasso et aP ° reported that, in 234 male angiography patients with coronary artery disease (CAD) and NGT, the number of arteries with stenosis was independently correlated with postload glucose (r2 = 0.445; P < 0.001) but not with FPG. In this study, HbAlc was significantly correlated with degree of atherosclerosis (r2 = 0.630; P < 0.001), s° which may reflect the substantial contribution of PPG levels to HbAlc in people with fairly well-controlled glycemia (HbAlc <7.3%). 1 Similarly, Nakamura et al sl compared outcomes for patients with NGT (n = 24) versus IGT (n = 16) who received coronary stents at Seiyu Memorial Hospital in Japan and found a greater degree of stenosis at follow-up in patients with IGT (48.8% vs 30.6%; P = 0.01). Importantly, all of these patients had normal FPG. sl Finally, postload hyperglycemia, but not FPG, was significantly associated with arterial stiffness in 159 men with IGT or diabetes in stepwise regression analysis in a Japanese study published in 2004 (]3 = 0.005; P < 0.05) (Table). s2 These results, together with epidemiologic and pathophysiologic data, support the hypothesis that acute hyperglycemia has a direct, atherogenic effect across the full spectrum of glucose intolerance and

that FPG levels by themselves are a poor indicator of future CV risk.

PPG Regulation: Interventional Studies The first study to show that targeting PPG levels has the potential to improve clinical outcome was in an obstetric and not a CV setting, but the results are worthy of note. In a paper published in 1995, de Veciana et al s3 found that pregnant women with gestational diabetes requiring insulin therapy did better when PPG rather than FPG levels were used to guide management. The 66 women in the study were randomly assigned to either PPG (1 hour after meals) or FPG monitoring to achieve a PPG of <140 mg/dL (<7.8 retool/L) or a FPG of 60 to 105 mg/dL (3.35.9 retool/L). The study found that adjusting therapy according to PPG levels resulted in a greater reduction in HbAlc (3.0% vs 0.6%; P < 0.001), lower infant birth weight (3469 vs 3848 g; P = 0.01), lower rates of neonatal hypoglycemia (3% vs 21%; P = 0.05) and macrosomia (12% vs 42%; P = 0.01), and fewer cesarean section deliveries due to cephalopelvic disproportion (12% vs 36%; P = 0.04). In people with type 2 diabetes, there is increasing evidence that the acute vascular toxicity of postprandial hyperglycemia can be ameliorated by attention to

Table. Significance of relationship between arterial stiffness and clinical factors in people with normal glucose tolera n c e (n = 110), impaired fasting glucose (n = 10), impaired glucose tolerance (n = 30), or diabetes (n = 9). Unadjusted (P value)

Adjusted for Age and Height (P value)

Stepwise Regression Analysis (P value)

Age

<0.001

-

<0.001

Height FPG 2-hour PPG SBP DBP LDL-C HDL-C TGt

<0.001 NS <0.001 0.006 0.009 NS NS NS

-0.049 0.009 NS 0.02 NS NS NS

NS NS <0.05 NS NS NS NS NS

NS

NS

Factor ~

Fasting i n s u l i n ~

NS

NS = n o t significant; FPG = f a s t i n g p l a s m a glucose; PPG = p o s t p r a n d i a l glucose; SBP = systolic b l o o d pressure; DBP = diastolic

blood pressure; LDL-C = low-density lipoprotein cholesterol; HDL-C = high-density lipoprotein cholesterol; TG = triglycerides. ~Relationships determined by Pearson's correlation coefficients unless otherwise noted. t tRelationships determined by Spearman's correlation coefficients for variables with a skewed distribution. • ~N = 78. Adapted with permission, s2

PPG regulation, most notably with short-acting insulin analogues. Ceriello et al 6 reported that the shortacting analogue, insulin aspart, reduced postprandial oxidative stress more effectively than did regular insulin. The study involved 23 patients with type 2 diabetes and 15 normal matched controls. For the patients with type 2 diabetes, a standard meal was preceded with regular insulin (0.15 unit/kg body weight); for matched controls, a standard meal was preceded with insulin aspart (0.15 unit/kg body weight). Oxidative stress in the arteries was assessed by measuring nitrotyrosine at the start of the meal and 1, 2, 4, and 6 hours afterwards. The study found that nitrotyrosine levels were significantly elevated in people with diabetes versus control subjects (P < 0.001) and that they became further elevated postprandially. Insulin aspart was significantly more effective than regular insulin at reducing both PPG levels (P < 0.04) and nitrotyrosine levels after the meal (P < 0.03). Postprandial triglyceride levels were unchanged by insulin aspart versus regular insulin in the study, indicating that the reduction in oxidative stress was unlikely to be a result of lowering triglycerides, another CV risk factor. In a follow-up study with a similar design published 2 years later, 7 insulin aspart significantly improved endothelial function compared with regular insulin, thereby also providing proof-of-concept that oxidative stress leads to endothelial dysfunction. Twenty-three patients with type 2 diabetes and 10 normal controls again were given a standard test meal; however, in this study, the meal was preceded by 0.15 unit/kg body weight of regular insulin for controls and 0.15 unit/kg body weight insulin aspart for diabetes patients. Flowmediated vasodilation in the brachial artery was measured by blinded examiners using ultrasound at intervals after the meal. Meals had no effect on flowmediated vasodilation in the control subjects, while there was a significant decrease in the diabetes patients (P < 0.001). Insulin aspart significantly improved arterial function compared with regular insulin (P < 0.01). As in the previous study, triglyceride levels were similar in the 2 groups, implying that, by lowering postprandial hyperglycemia, insulin aspart had a beneficial effect on the vascular endothelium (Figure 4). Beisswenger et al 8 tested the hypothesis that reducing PPG levels would reduce the formation of reactive cz-dicarbonyls after a standard meal in patients with diabetes. The double-blind crossover study, involving 21 people with type i diabetes treated with insulin lispro,

found a highly significant correlation between PPG excursions and postprandial excursions of the ct-dicarbonyls 3-DG and MG (r = 0.55; P < 0.001). Levels of these compounds were not correlated with HbAlc (r = 0.01; P = 0.95), supporting the notion that 0t-dicarbonyl formation is an unwelcome and direct consequence of acute hyperglycemia, not chronic hyperglycemia. Results of a double-blind, randomized, placebocontrolled study from the University of Heidelberg found that regulating PPG levels with acarbose reduced NF-~c[3 activation after 8 weeks of treatment in 20 patients with type 2 diabetes (P = 0.045). s4 The authors commented that "the importance of the study lies in the observation that every food intake provides an activation of proinflammatory response in patients with type 2 diabetes that are already characterized by elevated basal inflammatory levels." NF-K]3 mediates the expression of a host of atherosclerotic mediators, including cytokines, adhesion molecules, clotting factors, endothelin-1, and the receptor for AGEs. These studies suggest that PPG regulation benefits the vascular environment by improving oxidative stress, the inflammatory profile, and endothelial function and set the stage for a trial by Esposito et al. 9 These researchers found that regulating PPG results in clinically significant regression of atherosclerosis. In this randomized, single-center study, carotid IMT was measured in 175 people with type 2 diabetes after 1 year of treatment with either repaglinide (n = 88) or glyburide (n = 87). Over the study period, repaglinide produced a significant regression of mean carotid IMT compared with glyburide (P = 0.02) (Figure 5). Carotid IMT regression was observed in 52% of patients in the repaglinide group compared with 18 % in the glyburide group (P < 0.01). Interestingly, these resuits appeared to be related to a reduction in PPG excursions, since HbAlc was similar for both treatments. Repaglinide, however, was significantly more effective than glyburide at lowering PPG levels (P < 0.01). Furthermore, in the patients who had the greatest reduction in postprandial hyperglycemia, there was the greatest regression of carotid IMT (r = 0.21; P = 0.01). Finally, the results could not be explained by differences in classical CV risk factors such as body mass index, lipids, and blood pressure, which were not significantly different between the groups. Esposito et al also noted that repaglinide significantly reduced levels of the cytokine interleukin-6 (IL-6) as well as CRP compared with glyburide (P = 0.02),

-0- Regular insulin -O- Insulin aspart Controls

B

A

Triglycerides

Arterial Function

1.8-

16.0 14.0 12.0 o~ 10.0 1.1_

1.6--

"~ 1.4-

8.0 6.0 4.0 2.0 0

E

E 1.2 v

1.0 0

I

I

I

I

2

4

0.8

I

0

6

I

1

I

2

I

4

I

6

Time (h)

Time (h) C Glycemia

17.0 15.0 .-. 13.0 J

-~ 11.0 O E 9.0 E "-" 7.0 5.0

/

3.0 0

I

I

L

2

I

4

I

6

I

Time (h) Figure 4. Effect o f insulin aspart on endothelial function. (A) Insulin aspart significantly improved postmeal, flowmediated vasodilation (FMD) in the brachial artery versus regular insulin (N = 33; P < 0.01). (B) Triglycerides were unchanged, suggesting that the beneficial effect o f insulin aspart on the artery was due to (C) an acute reduction in hyperglycemia. Adapted with permission. 7

and changes in IL-6 and CRP were significantly correlated to reductions in PPG (r = 0.35, P < 0.001; and r = 0.26, P = 0.01, respectively). In this study, baseline serum levels of proinflammatory cytokines and CRP generally were higher in the people with type 2 diabetes versus controls (eg, IL-6, 2.4 vs 2.0, P = 0.02; CRP, 3.5 vs 1.4, P = 0.02). 9 Elevated levels of inflammatory cytokines have been shown to be related to future CV events in apparently healthy men and in those with known CAD. ss,s6 In a study of 124 hypertensive outpatients, 57 CRP levels at baseline were significantly correlated with progression of carotid atherosclerosis, as indicated by plaque number and plaque score, (r = 0.404, P = 0.002 and r = 0.436, P < 0.001 per annum, respectively). Esposito et al 9 speculate that postpran-

dial hyperglycemia exacerbates the inflammatory process in people with type 2 diabetes, a process interrupted in this trial by regulating PPG levels. The Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM), an international, double-blind, placebo-controlled, randomized trial involving 1429 patients with IGT, found a 49% RR reduction for any CV event in patients treated with acarbose over a mean follow-up of 3.3 years (2.2% vs 4.7%; P = 0.03). 11 A single-center, ultrasound substudy over the same period involving 132 of the 1429 STOP-NIDDM participants found that acarbose 100 mg TID significantly reduced the progression of carotid IMT relative to placebo (P < 0.05). l° Patients on acarbose had an annual mean IMT progression rate simi-

Repaglinide (n = 88)

Glyburide (n = 87)

I

o¢ -

-0.02 E m

t~ c-

L) ¢-

T

tl

-0.04

-0.06

Figure 5. Regression of atherosclerosis by targeting postprandial hyperglycemia. Repaglinide produced a significant regression of mean carotid intima-media thickness (IMT) over a 12-month period versus glyburide (N = 175). *p = 0.02.9

lar to those previously reported for comparable healthy subjects, which was half the rate of patients taking placebo. Over a mean 3.3-year follow-up, IMT increased by 0.05 mm in the placebo group versus only 0.02 mm in the acarbose group (P = 0.027). 1° Considered as a whole, this body of data suggests that PPG regulation ameliorates the pathologic vascular environment of patients with IGT and type 2 diabetes and can slow, or possibly reverse, atherosclerosis. However, the question is: Does PPG regulation translate into better outcomes? There is increasing evidence suggesting that it does. Since STOP-NIDDM, other work has shown that the cardioprotective effect of PPG regulation also may extend to type 2 diabetes. In 2004, Hanefeld et a112 published a meta-analysis of the 7 randomized, doubleblind, placebo-controlled, long-term studies of acarbose in type 2 diabetes. Data from patients treated with either acarbose (n = 1248) or placebo (n = 932) for at least 1 year were pooled to obtain the primary outcome of time to first CV event. Most patients also were taking a concomitant medication, either a sulfonylurea (31% in the acarbose group vs 38% in the

placebo group), metformin (4% vs 5%), or insulin (11% vs 12%). The study found that patients taking acarbose remained event-flee significantly longer than patients on placebo (P = 0.006). During the course of the 7 studies, 9.4% of the placebo group experienced a CV event versus 6.1% of the acarbose group, a significant 35% RR reduction (P = 0.006). There was also a 64% reduction in RR of MI (0.72% vs 2.04%; P = 0.012; Figure 6). The majority of patients (56.5%) were on state-of-the-art concomitant CV medications, so the cardioprotective effect of PPG regulation appears to be additive to that of other cardioprotective therapies.

The Missing Link in Diabetes Achievement of glycemic control remains elusive for most patients with type 2 diabetes and, as discussed, CVD continues to be a leading cause of death in these individuals. A review of the data on PPG and CV risk suggests that postprandial hyperglycemia may be the missing link. The UKPDS achieved relatively good control of FPG with intensive treatment using insulin or sulfonylureas: from year 6 to year 10 there was little rise (7.8-8.5 retool/L) in mean FPG levels. 22 By marked contrast, HbAlc levels continued to rise. The most logical explanation for this finding is the lack of focus on postprandial hyperglycemia, which was neither monitored nor treated in the UKPDS. One might speculate that the results of the UKPDS would have been different if PPG levels had been monitored and controlled with short-acting insulin analogues. Another key finding from UKPDS was that control of FPG significantly decreased the risk of microvascular complications, but the results for macrovascular disease were more equivocal. 22 Again, one can speculate that better cardioprotection might have been achieved with more emphasis on PPG control. Although speculative, such considerations have important implications for management. The current recommendations for glycemic control are based on the sharply increased risk of microvascular complications (in particular, retinopathy) above an HbAlc cut point of approximately 6%. $8 By contrast, there is no cut point for the risk of CVD with postprandial hyperglycemia. The DECODE study indicated that there is a linear relationship between the risk of CV death and OGTT 2-hour glucose levels, with an increased mortality risk evident at levels of -90 mg/dL

100

--

Acarbose

--

Placebo

--7--7~ I

I

v

i

98

*

> UA

> u

96

0

94

c-

c~

92 90 I

I

0

100

I

200

I

300

400

1

500

I

600

I

700

I

800

Time After Randomization (d) Figure 6. Kaplan-Meier survival curve for time to experience a cardiovascular (CV) event (myocardial infarction [MI]) during administration of either acarbose or placebo. A meta-analysis of 7 randomized, doubleblind, placebo-controlled, long-term studies (N = 2180) found a significant reduction in risk o f MI in patients taking acarbose (n = 1248) versus those taking placebo (n = 932).*Log rank test, P < 0.001; Cox proportional model, P = 0.012. Adapted with permission. 12

(5 mmol/L), 13 which is well below the ADA definition of type 2 diabetes at a casual glucose level of 200 mg/dL (11.1 mmol/L). 16 Coutinho et al s9 found a progressive relationship between 2-hour glucose levels and CV risk that extended well below the diabetic threshold in a meta-regression analysis of 20 studies involving 95,783 people followed for >12 years (13 = 0.531, P = 0.002). If the goal is optimal glycemic control to reduce the risk of diabetes complications in our patients, then it is no longer acceptable to do a partial job. The weight of evidence suggests that postprandial hyperglycemia may be the "first strike" against the body, setting the stage for the deadly onset of atherosclerosis, long before HbAI~ levels rise high enough to trigger microvascular complications.

Management Considerations: Timing and Targets for PPG Measurement Increasing evidence on the contributions of postprandial hyperglycemia to both HbAlc and CV risk is influencing treatment guidelines, with PPG targets becoming standard worldwide. Two recent additions are worthy of note. First, in its 2003 guidelines, s8 the ADA added a PPG target of <180 mg/dL (<10.0 mmol/L) and recommended targeting PPG for therapy if HbAI~

goals are not met despite reaching FPG goals. Second, postprandial hyperglycemia is acknowledged to be a possible CV risk factor in the joint European CV prevention guidelines, z3 which set targets in 2004 for self-monitored PPG levels much lower than those of the ADA, at 70 to 135 mg/dL (4.0-7.5 mmol/L). There is currently no clear consensus on the frequency of monitoring PPG levels to meet postprandial treatment goals. The ADA recommends self-monitoring 1 to 2 hours after the start of a meal. 16 The European CV guidelines do not suggest a time scale but provide the target range of 70 to 135 mg/dL (4.0-7.5 mmol/L). 23 The data from EI-Kebbi et al 2 suggest that casual measurements up to 4 hours after a meal reflect glycemic control, with a PPG >150 mg/dL (8.3 mmol/L) having an 80% predictive value of poor control (ie, HbAlc >7.0%). Monnier et al 6° found that, in controlled conditions, the probabilities for predicting HbAlc <7% (calculated by receiver-operating characteristic analysis) were significantly higher after meals, namely at 11 AM, 2 PM, and 5 PM, than in a fasting state at 8 AM (0.93, P < 0.001; 0.92, P = 0.002; 0.92, P = 0.005, respectively, vs the 8 AM value of 0.87). However, the actual cut point values of PPG that predicted poor control varied from 126 to 198 mg/dL (7-11 mmol/L), depending on the time of the meal, somewhat con-

firming the results of E1-Kebbi et al. In the Monnier study,6° the following levels were predictive of poor control (HbAlc >7.0%): >198 mg/dL (11 retool/L) after breakfast (11 AM);>162 mg/dL (9 mmol/L) at 2 PM~ after lunch; and >126 mg/dL (7 mmol/L) at the extended postlunch period (5 PM). Taken together, these results suggest that, for practical purposes, a target range for casual PPG measurements would best serve the needs of patients and their physicians. The question remains: Which target range is most appropriate? If increased mortality risk starts at a 2-hour PPG level -90 mg/dL (5 retool/L), according to DECODE, 13 this would be the ideal lower limit; however, the European CV guidelines recommend starting lower (70-135 mg/dL; 4.07.5 mmol/L).23 An appropriate level of >150 mg/dL (8.3 retool/L) would be suggested by the results of E1-Kebbi et al~2 yet the results of Monnier et a160 suggest levels from 126 to 198 mg/dL (7-11 mmol/L), depending on the time of day. Whichever targets future guidelines committees recommend, it is clear that PPG monitoring can no longer be considered optional. CONCLUSIONS

There is increasingly suggestive evidence that postprandial hyperglycemia may be the missing link that explains the connection between type 2 diabetes and CVD. The body of data now clearly shows that therapeutic goals (HbAlc <6.5% or <7.0%) in our patients with type 2 diabetes frequently cannot be met without considering all components of the glucose triad: FPG, HbAlc , and PPG. Glycemic control can be greatly improved by utilizing intensive therapy that includes treatments targeted at postprandial hyperglycemia, and, for some patients, may be impossible to attain without PPG regulation. Therapy targeted at PPG ameliorates the proatherogenic vascular environment in people with type 2 diabetes and reduces the progression of atherosclerosis. There is increasing evidence that this may translate into a reduction in CV events. Therefore, monitoring not only for FPG and HbAlc but also for PPG should be considered and therapies that target this important component of diabetes risk should be used. ACKNOWLEDGM ENTS

Financial support for the preparation of this manuscript was provided by Novo Nordisk.

The authors disclose the following. L.A. Leiter has received research grants, honoraria, and consultancy fees from Eli Lilly and Company, GlaxoSmithKline, Novartis Pharmaceuticals, Novo Nordisk, and sanofiaventis. A. Ceriello has received honoraria from Bayer, Eli Lilly and Company, GlaxoSmithKline, Novo Nordisk, sanofi-aventis, and Takeda Pharmaceuticals North America, Inc. J.A. Davidson has received research grants from Amylin Pharmaceuticals, Eli Lilly and Company, GlaxoSmithKline, Myogen, Novartis Pharmaceuticals, Novo Nordisk, and sanofi-aventis, and consultancy fees and/or honoraria from Bristol-Myers Squibb, Eli Lilly and Company, GlaxoSmithKline, Hoffmann-La Roche, Novartis Pharmaceuticals, Novo Nordisk, Pfizer Inc, sanofi-aventis, and Takeda Pharmaceuticals North America, Inc. M. Hanefeld has received honoraria from Bayer, GlaxoSmithKline, Novo Nordisk, sanofi-aventis, and Takeda Pharmaceuticals North America, Inc. L. Monnier has received honoraria and consultancy fees from Novo Nordisk. D.R. Owens has received honoraria and consultancy fees from Novo Nordisk and sanofi-aventis. N. Tajima has received research grants from Novo Nordisk and consultancy fees from Eli Lilly and Company, Novo Nordisk, and sanofi-aventis. J. Tuomilehto has received research grants from AstraZeneca, Novartis Pharmaceuticals, Novo Nordisk, and Unilever, and honoraria from AstraZeneca, Bayer, Novartis Pharmaceuticals, Novo Nordisk, and sanofiaventis. The authors wish to acknowledge the excellent assistance of Helen Byrt, PhD, in the preparation of this manuscript. The authors founded the International Prandial Glucose Regulation (PGR) Study Group in April 2002 to encourage credible scientific debate about the importance of prandial glucose regulation and to increase awareness of prandial glucose regulation as an important requirement to enhance diabetes management. REFERENCES 1. Monnier L, Lapinski H, Colette C. Contributions of fasting and postprandial plasma glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: Variations with increasing levels of HbA(1 c). Diabetes Care. 2003;26:881-885. 2. EI-Kebbi IM, Ziemer DC, Cook CB, et al. Utility of casual postprandial glucose levels in type 2 diabetes management. Diabetes Care. 2004;27:335-339.

3. Siebenhofer A, PlankJ, Berghold A, et al. Short acting insulin analogues versus regular human insulin in patients with diabetes mellitus. Cochrane Database Syst Rev. 2004;(4):CD003287. 4. Raskin P, Allen E, Hollander P, et al, for the INITIATE Study Group. Initiating insulin therapy in type 2 diabetes: A comparison of biphasic and basal insulin analogues. Diabetes Care. 2005;28:260-265.

5. Christiansen JS, Vaz JA, Metelko Z, et al. Twice daily biphasic insulin aspart improves postprandial glycaemic control more effectively than twice daily NPH insulin, with low risk of hypoglycaemia, in patients with type 2 diabetes. Diabetes Obes Metab. 2003;5:446-454. 6. Ceriello A, Quagliaro L, Catone B, et al. Role of'hyperglycemia in nitrotyrosine postprandial generation. Diabetes Care. 2002;25:1439-1443. 7. Ceriello A, Cavarape A, Martinelli L, et al. The post-prandial state in type 2 diabetes and endothelial dysfunction: Effects of insulin aspart. Diabet Med. 2004;21:171-175. 8. Beisswenger PJ, Howell SK, O'Dell RM, et al. Alpha-dicarbonyls increase in the postprandial period and reflect the degree of hyperglycemia. Diabetes Care. 2001 ;24:726-732. 9. Esposita K, Giugliano D, Nappo F, Marfella R, for the Campanian Postprandial Hyperglycemia Study Group. Regression of carotid atherosclerosis by control of postprandial hyperglycemia in type 2 diabetes mellitus. Circulation. 2004;110:214-219. 10. Hanefeld M, ChiassonJL, KoehlerC, et al. Acarbose slows progression of intima-media thickness of the carotid arteries in subjects with impaired glucose tolerance. Stroke. 2004;35: 1073-1078. 11. ChiassonJL, Josse RG, Gomis R, etal, for the STOP-NIDDM Trial Research Group. Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: The STOP-NIDDMtrial. JAMA. 2003;290:486-494.

12. Hanefeld M, Cagatay M, Petrowitsch T, et al. Acarbose reduces the risk for myocardial infarction in type 2 diabetic patients: Meta-analysis of seven long-term studies. Eur HeartJ. 2004;25:10-16. 13. The Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe (DECODE) Study Group and the European Diabetes Epidemiolagy Group. Is the current definition for diabetes relevant to mortality risk from all causes and cardiovascular and noncardiovascular diseases? Diabetes Care. 2003;26: 688-696. 14. Harris MI. Undiagnosed NIDDM: Clinical and public health issues. Diabetes Care. 1993;16:642-652. 15. Krolewski AS, Laffel LM, Krolewski M, et al. Glycosylated hemoglobin and the risk of microalbuminuria in patients with insulin-dependent diabetes mellitus. N EnglJ Med. 1995; 332:1251-1255.

16. American Diabetes Association. Standards of medical care. Diabetes Care. 2005;28(Suppl 1):$4-$36. 17. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus: Report of'the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care. 1997;20:1183-1197. 18. American Diabetes Association. Postprandial blood glucose. Diabetes Care. 2001 ;24:775-778. 19. KannelWB, McGee DL. Diabetes and cardiovascular disease. The Framingham Study. JAMA. 1979;241:20352038. 20. Kuusisto J, Mykkanen L, Pyorala K, Laakso M. NIDDM and its metabolic control predict coronary heart disease in elderly subjects. Diabetes. 1994; 43:960-967. 21. Stratton IM, Adler AI, Nell HA, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): Prospective observational study. BMJ. 2000;321:405412.

22. UK Prospective Diabetes Study (UKPDS) Group. Intensive bloodglucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33) [published correction appears in Lancet. 1999;354:602]. Lancet. 1998;352:837-853. 23. de Backer G, Ambrosioni E, BorchJohnsen K, et al. European guidelines on cardiovascular disease prevention in the clinical practice. Third work group of the European societies and other societies on cardiovascular disease prevention in the clinical practice [in Spanish]. NeuroIogia. 2004;19:440-450. 24. Canadian Diabetes Association. Clinical practice guidelines for the prevention and management of'diabetes in Canada: 2003. Available at www.diabetes.ca/cpg2003. Accessed August 26, 2005. 25. American College of Endocrinology. Consensus statement on guidelines for glycemic control. Endocr Pratt. 2002;8(Suppt 1):5-11. 26. MeigsJB, Nathan DM, D'Agostino RB Sr, Wilson PW, for the Framingham Offspring Study. Fasting and postchallenge glycemia and cardiovascular disease risk: The Framingham Offspring Study. DiabetesCare. 2002; 25:1845-1850. 27. The Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe (DECODE) Study Group and the European Diabetes Epidemiology Group. Glucose tolerance and mortality: Comparison of WHO and American Diabetes Association diagnostic criteria. Lancet. 1999;354: 617-621. 28. Tominaga M, Eguchi H, Manaka H, et al. The Funagata Diabetes Study. Impaired glucose tolerance is a risk factor for cardiovascular disease, but not impaired fasting glucose. Diabetes Care. 1999;22:920-924. 29. Qiao Q, Nakagami T, Tuomilehto J, et al, for the International Diabetes Epidemiology Group and the

DECODA Study Group. Comparison of the fasting and the 2-h glucose criteria for diabetes in different Asian cohorts. Diabetologia. 2000;43: 1470-1475. 30. Woerle HJ, Pimenta WP, Meyer C, et al. Diagnostic and therapeutic implications of relationships between fasting, 2-hour postchallenge plasma glucose and hemoglobin alc values. Arch Intern Med. 2004;164: 1627-1632. 31. Ceriello A. New insights on oxidative stress and diabetic complications may lead to a "causal" antioxidant therapy. Diabetes Care. 2003;26:

1589-1596. 32. Ceriello A, Hanefeld M, Leiter L, et al. Postprandial glucose regulation and diabetic complications. Arch Intern Med. 2004;164:2090-2095. 33. Maddux BA, See W, Lawrence JC Jr, et al. Protection against oxidative stress-induced insulin resistance in rat L6 muscle cells by micromolar concentrations of alpha-lipoic acid. Diabetes. 2001 ;50:404-410. 34. Ceriello A, Motz E. Is oxidative stress the pathogenic mechanism underlying insulin resistance, diabetes, and cardiovascular disease? The common soil hypothesis revisited. Arteriosder Thromb Vast Biol. 2004;24:816-823. 35. Perticone F, Ceravolo R, Pu]ia A, eta]. Prognostic significance of endothelial dysfunction in hypertensive patients. Circulation. 2001 ;I 04:191 -I 96. 36. HeitzerT, SchlinzigT, Krohn K, et al. Endothelial dysfunction, oxidative stress, and risk of cardiovascular events in patients with coronary artery disease [published correction appears in Circulation. 2003;I 08:500]. Circulation. 2001 ;I 04:2673-2678. 37. Kawano H, MotoyamaT, Hirashima O, et al. Hyperglycemia rapidly suppresses flow-mediated endotheliumdependent vasodilation of brachial artery. J Am Coil Cardiol. 1999;34: 146-154. 38. Giugliano D, Marfella R, Coppola L, et al. Vascular effects of acute hyperglycemia in humans are reversed

39.

40.

41.

42.

43.

44.

45.

46.

47.

by L-arginine. Evidence for reduced availability of nitric oxide during hyperglycemia. Circulation. 1997;95: 1783-1790. Ceriello A, Taboga C, Tonutti L, et a[. Evidence for an independent and cumulative effect of postprandial hypertriglyceridemia and hyperglycemia on endothelial dysfunction and oxidative stress generation: Effects of short- and long-term simvastatin treatment. Circulation. 2002; 106:1211-1218. Marfella R, Verrazzo G, Acampora R, eta]. Glutathione reverses systemic hemodynamic changes induced by acute hyperglycemia in healthy subjects. AmJ Physiol. 1995;268:EI 167El 173. Ting HH,Timimi FK, Boles KS, etal. Vitamin C improves endotheliumdependent vasodilation in patients with non-insulin-dependent diabetes mellitus.J Clin Invest. 1996;97:22-28. Ceriello A, Quagliaro L, D'Amico M, et al. Acute hyperglycemia induces nitrotyrosine formation and apoptosis in perfused heart from rat. Diabetes. 2002;51 :I 076-I 082. Ceriello A, Bortolotti N, Motz E, et al. Meal-induced oxidative stress and low-density lipoprotein oxidation in diabetes: The possible role of hyperglycemia. Metabolism. 1999;48: 1503-I 508. Beisswenger PJ, Makita Z, Curphey TJ, et al. Formation of immunochemical advanced glycosylation end products precedes and correlates with early manifestations of renal and retinal disease in diabetes. Diabetes. 1995;44:824-829. Beisswenger PJ, Szwergold BS, Yeo KT. Glycated proteins in diabetes. Clin Lab Med. 2001 ;21:53-78. Brownlee M. Biochemistry and molecular cell biology of diabetic complications. Nature. 2001 ;414:813-820. Schiekofer S, Andrassy M, ChenJ, et a]. Acute hyperglycemia causes intracellular formation of CML and activation ofras, p42/44 MAPK, and nuclear factor kappa[3 in PBMCs

[published correction appears in Diabetes. 2003;52:1307]. Diabetes.2003;

52:621-633. 48. McLe[lan AC, Thorna[ley PJ, Benn J, Sonksen PH. G[yoxalase system in clinical diabetes mellitus and correlation with diabetic complications. Clin Sd (Lond). 1994;87:21-29. 49. Temelkova-Kurktschiev TS, Koehler C, Henkel E, et al. Postchallenge plasma glucose and glycemic spikes are more strongly associated with atherosclerosis than fasting glucose or HbAlc level. Diabetes Care. 2000; 23:1830-1834. 50. Sasso FC, Carbonara O, Nasti R, et al. Glucose metabolism and coronary heart disease in patients with normal glucose tolerance.lAMA. 2004; 291 :I 857-I 863. 51. Nakamura N, Ueno Y, Tsuchiyama Y, eta[. Isolated post-challenge hyperglycemia in patients with normal fasting glucose concentration exaggerates neointima[ hyperplasia after coronary stent implantation. CircJ. 2003;67:61-67. 52. Ohshita K, Yamane K, Ishida K, et al. Post-challenge hyperglycaemia is an independent risk factor for arterial stiffness in Japanese men. Diabet Med. 2004;21:636-639. 53. de Veciana M, Major CA, Morgan MA, et al. Postprandial versus preprandial blood glucose monitoring in women with gestational diabetes mellitus requiring insulin therapy. N EnglJ Med. 1995;333:1237-I 241. 54. RudofskyJr G, Reismann P, Schiekofer S, et al. Reduction of postprandial hyperglycemia in patients with type 2 diabetes reduces NF-kappa[3 activation in PBMCs. Horm Metab Res. 2004;36:630-638. 55. Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation. 2000;I 01 :I 767-I 772. 56. Blankenberg S, Tiret L, Bickel C, et al, for the AtheroGene Investigators. Interleukin-18 is a strong predictor

of cardiovascular death in stable and unstable angina. Circulation. 2002;I 06: 24-30. 57. Hashimoto H, Kitagawa K, Hougaku H, et al. Relationship between Creactive protein and progression of early carotid atherosclerosis in hypertensive subjects. Stroke. 2004;35: 1625-I 630. 58. American Diabetes Association. Clinical practice recommendations 2003. Introduction. Diabetes Care. 2003; 26(Suppl I ):$I-$2. 59. Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783 individuals followed for

12.4 years. Diabetes Care. 1999;22: 233-240. 60. Monnier L, Colette C, Lapinski H. Global assessment for quality and safety of control in type 2 diabetic patients. Eur J Clin Invest. 2004;34: 37-42.

Address correspondence to: Lawrence A. Leiter, MD, Division of Endocrinology and Metabolism, St. Michael's Hospital, 61 Queen Street East, #6121Q, Toronto, Ontario, Canada M5C 2T2. E-maih leiter@ stub.toronto.on.ca