American Diabetes Association – 72nd Scientific Sessions

June 8-12, 2012; Philadelphia, PA; Report – CV Disease and Other Complications – Draft

Executive Highlights

ADA 2012 featured a large number of talks on complications, ranging from hypoglycemia to cardiovascular disease and retinopathy to sleep disorders. There were several presentations focused on how to better predict complications – some of the methods presented included hypoglycemia, hyperinsulinemia, type of therapy, advanced glycation end products, genetics, oxidative stress, and inflammation. These epidemiological studies should prove valuable in generating hypotheses. ADA brought relatively little discussion of new complications therapies in the pipeline, though Reata’s bardoxolone methyl (currently enrolling in phase 3; identifier: NCT01351675) was a notable exception – clearly there is great excitement about this therapy in development. Other talks focused on the risk of complications using sub-analyses from large trials such as ACCORD (hypoglycemia), DCCT/EDIC (inflammation and advanced glycation end products), and BARI-2D (use of insulin-providing or insulin-sensitizing therapies). We also noticed a particular focus on the negative implications of hypoglycemia and severe hypoglycemia on both a short-term (alterations in QTc interval, rises in inflammatory cytokines and markers of endothelial dysfunction, clot resistance to lysis, diminished vagal tone, and cardiac arrhythmias) and long-term (potentially cardiovascular disease and mortality) basis. With this in mind, we’re glad to see the emergence and expansion (and, indeed, establishment) of therapies for type 2 diabetes that don’t cause hypoglycemia (e.g., DPP-4s, GLP-1s, SGLT-2s). Of course, this is still a major challenge for type 1s, though we believe CGM has already helped many in this respect, and a bihormonal (insulin-glucagon) artificial pancreas might do so in the future.


Table of Contents 


Cardiovascular Disease and Other Complications

Symposium: Clinical Aspects of Hypoglycemia in Diabetes – Consequences and Prevention


Gabriella Gruden, MD (University of Turin, Turin, Italy)

Dr. Gruden presented new data from the EURODIAB IDDM Complications Study and the EURODIAB Prospective Complications Study on the relationship between severe hypoglycemia and QTc interval, CVD, and CVD markers. According to Dr. Gruden, the rationale for her group’s analysis came from: 1) studies suggesting that nocturnal hypoglycemia can induce alterations in QTc interval, 2) a lack of data on the relationship between hypoglycemia and CVD in type 1 diabetes (despite a link having been shown in type 2 diabetes), and 3) proposals that hypoglycemia induces rises in inflammatory cytokines and markers of endothelial dysfunction. The EURODIAB IDDM study recruited 3250 patients with type 1 diabetes aged 16-50 at 31 European centers. A variety of data, including severe hypoglycemia and QT interval data, was collected at baseline. Patients were then followed up with at six to eight years at which time morbidity and mortality and hypoglycemia data, along with blood sample data were collected. At baseline, 68% of patients did not report any hypoglycemic episodes over the past year, while 19% reported 1-2, and 13% reported ≥ 3. Severe hypoglycemia was found to be independently associated with QTc interval prolongation, independent of complications. Meanwhile, there was no association between severe hypoglycemia and CVD. Finally, severe hypoglycemia was not associated with differing levels of CVD markers and neither complications nor CVD were associated with nonsevere hypoglycemia in the year prior to follow-up.

  • The EURODIAB IDDM study recruited 3250 patients with type 1 diabetes of more than one-year duration were recruited at 31 centers in 16 European countries. Data on severe hypoglycemia was self-reported in questionnaires. Hypertension, retinopathy, microalbuminuria, macroalbuminuria, CVD, and QT intervals were also assessed at baseline. At baseline, 68% of patients did not report any hypoglycemic episodes over the past year, while 19% reported 1-2, and 13% reported ≥ 3.
  • Patients were then recalled for follow-up after six to eight years, at which time data on morbidity and mortality were collected and another hypoglycemia questionnaire (which now included questions about nonsevere hypoglycemia as well) was administered. A nested case- control study was also designed at follow-up; case subjects were chosen to have the highest complication burdens (while control patients had no complications) to enable subgroup analyses. Finally, blood samples used for analysis of CVD markers and cytokines were taken from both case (n=363) and control (n=168) patients taken at follow-up. VCAM-1, E-Selectin, IL6, TNF- α, HSP27, anti-HSP60 and HSP70 antibodies, and plasma CRP levels were measured.
  • In the EURODIAB IDDM study, severe hypoglycemia was independently associated with QTc interval prolongation. Logistic regression analysis showed that frequency of severe hypoglycemia was associated with longer diabetes duration, lower A1c, and prolonged QTc. The association between severe hypoglycemia and QTc interval prolongation was independent of diabetic microvascular/macrovascular complications. According to Dr. Gruden, these results support the hypothesis that severe hypoglycemia may represent not only a trigger, but also a risk factor, for QTc interval prolongation and cardiac arrhythmia.
  • Data from the EURODIAB Prospective Complications Study (PCS) (the follow up to the EURODIAB IDDM) do not support the hypothesis that severe hypoglycemia is a risk factor for cardiovascular disease (CVD). 2899 patients were included in this study (340 of whom had CVD at baseline), Morbidity/mortality and vital status data was missing for 335 and 383 patients, respectively, leaving 2181 patients in the analysis. In this cohort, patients with frequent hypoglycemia were older, had longer diabetes duration, and a greater prevalence of microvascular complications. The percent of patients with fatal or nonfatal CVD was similar among patients with 0, 1-2, or 3+ severe hypoglycemic episodes. The proportion of patients with both fatal and nonfatal CV events was comparable between patients who reported and didn’t report severe hypoglycemia, both at baseline and follow-up. Logistic regression analysis further showed that baseline severe hypoglycemia wasn’t associated with fatal and nonfatal CVD.
  • Finally, severe hypoglycemia was not associated with markers of CVD in the EURODIAB PCS. TNF-alpha levels were lower in samples from case subjects who did not report any severe hypoglycemic episodes versus those who reported one or more episodes, but the difference did not persist when adjusting for endothelial function. Per Dr. Gruden, these data show that cytokines previously postulated to link hypoglycemia to CVD do not differ in patients with or without a history of hypoglycemia. Further logistic regression analysis showed that neither complications nor CVD were associated with nonsevere hypoglycemia in the year prior to follow-up. Subjects with severe hypoglycemic episodes at follow-up had a 53% reduced risk of CVD, but the association wasn’t significant after adjustment for A1c.



David Brillon, MD (Weill Cornell Medical College, New York, NY)

After reviewing the literature published to date on the implications of hypoglycemia in the ACCORD trial, Dr. Brillon concluded: 1) patients with type 2 diabetes who experience symptomatic, severe hypoglycemia are at increased risk of death regardless of the level of intensity of glycemic control (Bonds et al., BMJ 2009); 2) the increased risk of death among participants in the intensive arm in ACCORD cannot be attributed to the increased of severe hypoglycemia in the intensive arm (Bonds et al., BMJ 2009); 3) certain baseline characteristics (female gender, African-American ethnicity, having less than high education, age, insulin use at baseline, peripheral neuropathy, lower BMI, nephropathy) may predict subsequent episodes of severe hypoglycemia (Miller et al., BMJ 2010); and 4) patients with poor cognitive function may be at higher risk (Punthakee et al., Diabetes Care 2012). After the presentation concluded, Dr. Philip Cryer (Washington University of St. Louis, St. Louis, MO), the moderator for the session, commented that the excess mortality observed in ACCORD could have been a result of chance or some non-glycemic factor; he stated that he doesn’t think we’ll ever know with certainty what caused the excess mortality in the intensive control arm of ACCORD.


Oral Sessions: Diabetic Foot Disease – What Contributes to Healing?


Benjamin Lipsky, MD (University of Washington, Seattle, WA)

Recipient of the Roger E. Pecoraro Award, Dr. Lipsky began his presentation on the history and advancement in the treatment of diabetes foot infections (DFIs) by briefly (and humorously) chronicling the study of podiatry since the 17th century. He then led the audience to the 20th century, explaining how modern study of DFIs began in the 1920s, steadily gaining more attention with the development of penicillin and safer amputation methods throughout the century; however, as Dr. Lipsky believes, “amputation is a good start, but we should be able to do better than that.” Thus, over the last 30 years, Dr. Lipsky and his team have worked to enhance understanding DFIs with the goal of bettering treatment, guidelines, and quality of life for patients suffering from such infections. During ADA, new treatment guidelines for DFIs developed by this group were published by the Infectious Disease Society of America. Thanks to attention to enhancement of treatment of DFIs, rate of amputations and length of hospitalization due to DFIs have decreased dramatically in the past two decades.

  • New guidelines for infection have brought treatment of DFIs to a more effective level: This update includes new recommendations for assessment, management, and therapy, as well as updated references and recommendations for future research questions and regulatory changes.



Agbor Ndip, MD (Manchester Royal Infirmary, Manchester, United Kingdom)

Diabetes is the single most common cause of end stage renal disease requiring renal replacement therapy – foot disease in diabetes further compounds the already poor survival of dialysis patients. In an effort to provide data to back up this assertion, Dr. Ndip presented results from a two-year prospective study that aimed to evaluate the impact of lower extremity complications on mortality in persons with diabetes receiving renal dialysis. At the end of this period, the overall mortality rate was 53.1%. Furthermore, patients who had a foot ulcer at the start of the study had a higher mortality rate than those who did not (59% vs. 41%). According to Dr. Ndip, these results strongly implicate foot disease as a predictor of mortality. He noted that his data demonstrate that an active ulcer increases one’s risk of dying by 120%, warranting future research to investigate if patient interventions will result in reduced ulcerations, amputations, and, ultimately, mortality among dialysis patients with diabetes.

  • All 192 diabetes patients underwent comprehensive assessment prior to and during the study but still showed slightly variable baseline statistics. For example, baseline A1c (7.8%), age (62±13 years), and gender (65% male, 35% female) may account any variability in data. During the study, all patients participated in routine, comprehensive foot examinations and follow-up reviews at three monthly intervals to assess risk factors, incident food ulcer/amputation, vascular procedures, hospitalizations foot related or not, and vital status.

Questions and Answers

Q: In this context, surely the foot ulcer is just the manifestation of a patient with more advanced diabetes complications; it’s sure that they are going to die. Are you suggesting that treating the foot ulcer would increase their chance of surviving?

A: I would argue that if we could prevent foot ulcers to start with, then we can at least attack the predictors of death. We can organize better mechanisms for care.

Q: Have you analyzed data about differences in death between those who had ulcers and those who had not? I would wonder if there was a higher rate of patients stopping dialysis or committing suicide, for example.

A: That is a very important point. That is ongoing as we speak.




Howard Foyt, MD, PhD (Cebix, La Jolla, CA)

Dr. Foyt was enthusiastic about Cebix’s new once-weekly C-peptide drug, Ersatta, for the disease- modifying treatment of diabetic peripheral neuropathy. Ersatta is a long-acting form of human C- peptide constructed through site-specific linkage to a single polyethylene glycol (PEG) moiety. Preclinical data show that Ersatta improves nerve conduction velocity and thus reverses the progression of peripheral neuropathy. Dr. Foyt described the recently completed randomized, blinded, placebo-controlled, multiple-ascending dose study in 30 people with type 1 diabetes. Participants were equally assigned to one of three groups: 0.3 mg, 1.0 mg (the estimated C-peptide “replacement” dose level), and 3.3 mg (each group included eight actively treated and two on placebo). The half-life of Ersatta was six to seven days (vs. one hour for naturally occurring C-peptide), supporting once weekly dosing. The drug was also safe and well tolerated with no reportable SAEs. These results have encouraged Cebix to undertake a phase 2 trial that includes 42 individuals. Encouragingly, Cebix’s Ersatta has also been granted fast track status by the FDA for Ersatta in the diabetic peripheral neuropathy indication. Eventually, the company hopes Ersatta could be used as a treatment for all microvascular complications associated with diabetes.

  • Dr. Foyt gave significant credit to Dr. Wahren, Cebix’s Chief Scientific Officer, for revitalizing the field of C-peptide study after years of researchers casting it off as a waste of resources. Because C-peptide is released along with insulin but has a much longerhalf-life (three minutes vs. 68 minutes), Dr. Wahren believes it must have some sort of physiological benefit.

Questions and Answers

Q: Do you plan to investigate C-peptide analogues that would be injected along with one’s normal insulin therapy, as to mimic physiological dosing?

A: Not at this time. We are focusing on a once weekly formula and have no current plan to mix this drug with insulin. Patients with diabetes already have a large number of required injections and we would like to avoid increasing that number as much as possible.


Symposium: Sleep Disorders and Diabetes – From Epidemiology to Intervention


Esra Tasali, MD (The University of Chicago, Chicago, IL)

This presentation examined whether sleep quality and deprivation contributed to the development of type 2 diabetes. In both scenarios, Dr. Tasali established that inadequate sleep quality and duration pose major risk factors. Sleep deprived individuals displayed increased insulin resistance, heightened caloric intake, and an inability to loose as much weight as their well-rested counterparts. Inadequate sleep quality reduced insulin sensitivity by 27%.

  • A meta-analysis was recently conducted, linking insufficient sleep duration to the incidence of type 2 diabetes. This analysis consisted of a 90,000-person testing group that drew data from studies with at least a three-year follow up. Limitations to these studies included self sleep reporting and the absence of control group for obstructive sleep apnea (OSA). Still, the meta analysis revealed that there was a significant risk for developing type 2 diabetes with inadequate sleep (defined as less than five to six hours of sleep).
  • Sleep deprivation is a major risk factor for developing type 2 diabetes. Three studies cited by Dr. Tasali pointed to reduced insulin resistance resulting form sleep deprivation. Participants in these studies were subjected to sleep deprivation (six nights with four hours of sleep, 14 nights with five hours of sleep, and six nights with five hours of sleep, respectively). Participants, even healthy, young individuals, displayed insulin resistance resulting from inadequate sleep. Caloric intake is also increased by sleep deprivation. Dr. Tasali cited a study that subjected participants to four versus nine hours of sleep. Those that slept only four hours consumed 300 more calories than their counterparts. Sleep deprivation also limits weight loss. Participants who slept 8.5 hours versus 5.5 hours lost 6.6 more pounds over a two-week period.
  • Sleep quality also contributed to acquiring type 2 diabetes. A study cited by Dr. Tasali found that reduced sleep quality, specifically resulting from sleep apnea, reduced insulin sensitivity by 27%.. In this study, 18-30 year-old participants were subjected to sleep fragmentation. This process uses acoustics or vibrations to reduce slow-wave sleep in order to mimic obstructive sleep apnea. Dr. Tasali’s own study confirmed these results, finding an increased diabetes risk and decreased insulin sensitivity resulting from OSA induced sleep fragmentation.

Questions and Answers

Q: Was there any difference between insulin secretion and sensitivity between an obese and non-obese person in sleep-deprived patients?

A: I don’t think there is a study that compared the two. However, one study that I am familiar with used overweight subjects. In separate studies, it has shown to reach the same outcome, but it has not been directly researched.

Q: When you look at incidences of hypoxia are they worse, better, or the same as subjects with non-hypoxic episodes.

A: The last study that I cited showed they have OSA that is characterized by arousals. We look at repertory events to see the occurrences of arousals. Sleep apnea in young people is categorized by arousals. In obesity, there could be differing physical characteristics in sleep apnea – for instance difficulty of contracting a diaphragm. What we should do is identify the population group who are most vulnerable to type 2 diabetes, especially, if they are to have a specific subtype of sleep apnea.


Naresh M. Punjabi, MD, PhD (Johns Hopkins, Baltimore, MD)

This presentation covered the negative impacts of sleep complications, specifically sleep apnea, hypoxia and, sleep fragmentation, on glucose metabolism. Fifteen or more occurrences of apneas in an hour results in a 36% increase in impaired glucose tolerance. Dr. Punjabi also established that insulin sensitivity is inversely related to the prevalence of apneas. When a person experienced sleep fragmentation or hypoxia, insulin sensitivity decreased by 30%. However, it remains inconclusive if treatment to mitigate these sleep complications has a positive effect on glucose metabolism.

  • Dr. Punjabi presented data establishing that a connection exists between sleep apnea and altered glucose metabolism. His Heart Health Study found that there is a close correlation between the prevalence of apneas (a cessation of airflow for at least 10 seconds during sleep) and glucose sensitivity. A person experiencing fifteen or more apneas an hour had a 36% higher impaired glucose tolerance. This reached 46% when adjusted for other outside factors. Participants in the study who experienced hypoxia (a decrease in airflow during sleep by at least 50% that is often the result of apneas) had the highest prevalence of impaired fasting glucose. Sleep apnea was also inversely related to insulin sensitivity. That is, the more prevalent the occurrence of apneas, the more sensitive to insulin a participant was.
  • Hypoxia and sleep fragmentation were found to decrease insulin sensitivity. In his study, Dr. Punjabi varied the participant’s oxygen concentrations over a period of five hours to mimic the effects of hypoxia. The change in oxygen concentrations persisted for one day followed by a day of normoxia. The results of his study were surprising: after only five hours of hypoxia, insulin sensitivity fell by an average of 30%. Dr. Punjabi also examined the effect of sleep fragmentation on insulin sensitivity, “torturing subjects” by embedding vibrators designed to wake participants every 30 seconds into a bed. Similar to the hypoxia results, the study showed that insulin sensitivity decreased by 25%-30% resulting from sleep fragmentation. Increased insulin compensation and decreased glucose effectiveness also resulted.
  • It is still inconclusive if CPAP treatment (treatment to mitigate sleep apnea and hypoxia) has a positive effect on glucose metabolism. According to Dr. Punjabi, two days of therapy would improve insulin sensitivity; however, because the study was not controlled, he “still does not know if treatment will make people better.” Another study tried to determine if reversing the effects of hypoxia in mice would result in better glucose metabolism. Treatment improved some aspects, but not all, of the metabolic symptoms in mice. Glucose tolerance and beta cell function did not fully recover.

Questions and Answers

Q: One caveat I found to your presentation is beta cell completion. I have done research that shows that actually, beta cell completion is not instantaneous. We determined it takes a couple of weeks to for beta cell function to return to some degree.

A: The time course for the study was about 48 hours. This is a very important issue you have raised. To fully answer your question, we need to be mindful of more studies, and to conduct more longitudinal work.


Oral Session: Frontiers of the Diagnosis and Treatment of Neural Dysfunction


Rodica Pop-Busui, MD, PhD (University of Michigan, Ann Arbor, MI)

BARI-2D evaluated efficacy of prompt revascularization versus medical intervention in people with coronary artery disease and type 2 diabetes. Within the medical intervention group, insulin-providing (IP) versus insulin-sensitizing (IS) treatments were compared. This sub-analysis evaluated the prevalence and incidence of diabetic peripheral neuropathy (DPN) in the BARI-2D cohort (N=2159). Using an intent-to-treat analysis, the authors found that that among people without DPN at baseline (N=1075), there was significantly lower progression to DPN in the IS (54%) compared to the IP (63%) group. There was no significant difference in the prevalence of DPN when including all patients (both those who had and did not have DPN at baseline).

  • BARI-2D evaluated efficacy of prompt revascularization versus medical intervention in people with coronary artery disease and type 2 diabetes. Within the medical intervention group, insulin-providing (IP) versus insulin-sensitizing (IS) treatments were compared. The randomized study showed similar cardiovascular outcomes across these two groups. The formal study results were published in NEJM in 2009.
  • This sub-analysis evaluated the prevalence and incidence of diabetic peripheral neuropathy (DPN) in the BARI-2D cohort (N=2159). At baseline, mean age was 62 years, A1c was 7.7%, and duration of diabetes was 10 years with no difference between the IP and IS groups. DPN was defined as a score greater than two on the Michigan Neuropathy Screening Instrument (MNSI).
  • Using an intent-to-treat analysis, the authors found that that among people without DPN at baseline (N=1075), there was significantly lower DPN in the IS (54%) compared to the IP (63%) group. Correcting for A1c at baseline, this yields a hazard ratio of0.81 for IS versus IP. This effect was strongest in a further subgroup of people <65 years (HR=0.75), males (HR= 0.74), and had triglycerides ≥150 mg/dl (HR= 0.74).
  • There was no significant difference in the prevalence of DPN when patients starting with DPN were included. There was also no difference in the prevalence of DPN when only considering patients who had DPN at baseline (reversal rate).

Questions and Answers

Q: So more than 70% of your patients had DPN, how do you explain that?

A: Well that’s not quite right, that’s looking only at people who did not have it to begin with. If you’re looking at everybody, the proportion is closer to 50%.

Q: Was there any difference in weight?

A: I don’t know that off the top of my head, but I do know that there was no difference in blood pressure.


Saul Ganuth, MD (Case Western Reserve University, Cleveland, OH)

Advanced glycation end products (AGEs) are intracellular proteins conjugated with sugars, the formation of which has been linked to various pathologies and natural aging. This research group measured AGEs in skin biopsies from the DCCT. The original study measured six advanced glycation end products, and this new analysis adds four more: glucosepane (GSPNE), methylglyoxal (MG-H1) and glyoxal (G-H1), hydroimidazolones, and carboxyethyl-lysine (CEL). Including these new markers, GSPNE and fructose-lysine seem to correlate the most strongly with the incidence of retinopathy. Furosine is the strongest predictor of nephropathy, while MG-H1 and furosine are the strongest predictors of neuropathy. Importantly, these variables are significant despite adjustments for A1c during EDIC as well as other blood glucose control measures measured in the study.

  • Advanced glycosylation products (AGEs) are intracellular proteins conjugated with sugars, the formation of which has been linked to various pathologies and natural aging. Due to higher levels of blood sugars, people with diabetes have relatively higher concentrations of AGEs in their cells. AGEs have been associated with the so-called “metabolic memory.” Metabolic memory refers to the effect of differing outcomes within the intensive versus conventional treatment groups 10 years post-DCCT.
  • This research group measured AGEs in skin biopsies from DCCT. The original study measured six advanced glycation end products: fructose-lysine (furosine), fluorescence, pentosidine, CML, acid soluble collagen, and insoluble collagen. Furosine is the earliest glycosylation product in this panel. Levels of these AGEs correlated with increased incidence of retinopathy but not neuropathy.
  • Recently, samples were re-analyzed for four new AGEs: glucosepane (GSPNE), methylglyoxal (MG-H1) and glyoxal (G-H1), hydroimidazolones, and carboxyethyl- lysine (CEL). Glucosepane is the latest-developing AGE in this new panel.
  • Including these new markers, GSPNE and furosine seem to correlate the most strongly with the incidence of retinopathy. When correcting for these markers, the effect of A1c from EDIC loses significance, implying that these two AGEs may account for a proportion of this A1c-retinopathy association.
  • The new panel with all 10 AGEs also predicts microalbuminuria and neuropathy independently of A1c. Furosine is the strongest predictor of nephropathy, while MG-H1 and furosine are the strongest predictors of neuropathy.

Questions and Answers

Q: Do you think that we’ll need to look at tissue levels clinically if we use this as a marker?

A: Tissue levels are better, but they involve an invasive procedure as well as the measurement of ten products. So it’s probably not practical for clinical use, even though it’s invaluable from a research perspective. Auto-fluorescence, which is non-invasive, might be a practical way of making use of the biology that we’ve explained. We’ll let you know at the meeting next year if this is an important clinical marker that could be used instead of AGEs.


Georgia Pambianco, MS, MPH (University of Pittsburgh, Pittsburgh, PA)

Cardiovascular autonomic neuropathy (CAN) is thought to confer a high risk of mortality, coronary artery disease (CAD), heart failure, and tachycardia. Since there is no agreement on the optimal screening for autonomic neuropathy in type 1 diabetes, the long term predictive power of two simple tests, E/I ratio and 30:15 ratio, were assessed with regard to CAD incidence in the Pittsburgh Epidemiology of Diabetes Complications (EDC) study of childhood onset type 1 diabetes cohort. CAD was defined as clinic physician diagnosed angina, ischemic ECG changes, revascularization, confirmed MI, or CAD death. Risk factors examined were diabetes duration, age, A1c, hypertension, smoking, total cholesterol, HDL, triglycerides and LDL cholesterol, waist/hip ratio (WHR), gender, WBC, estimated glomerular filtration rate (GFR), and albumin excretion rate (AER). Forty-seven (19%) out of 248 participants developed CAD in the follow up period. Univariately, E/I ratio but not 30:15 ratio predicted CAD. In the fully adjusted model, E/I ratio remained independently predictive along with diabetes duration, WBC, and A1c. The study indicates that the E/I, but not the 30:15, is an independent predictor of CAD, and thus the preferred simple measure.

  • Since there is no agreement on the optimal screening for autonomic neuropathy in type 1 diabetes, the long-term predictive power of two simple tests, E/I ratio and 30:15 ratio, were assessed with regard to CAD. At study entry (1986-1988, n=658), the mean age was 28 years and mean diabetes duration was 19 years. Follow up (mean=10.3 years) was determined from the time of first assessment of the 30:15 ratio test (1994) in 248 participants. The E/I ratio was measured by means of heart rate response to deep breathing and dichotomized as ≥1.1 (normal) and <1.1 (abnormal). The 30:15 ratio was measured by the longest RR interval around the 30th heartbeat and the shortest around the 15th heart beat after standing up and dichotomized as >1.04 (normal) and ≤1.04 (abnormal).
  • Univariately, E/I ratio (HR 2.8, 95% CI 1.5, 5.1) but not 30:15 ratio (HR 1.3, 95% CI.7, 2.3) predicted CAD. In the fully adjusted model, E/I ratio remained independently predictive (HR 1.9, 95% CI 1.1, 3.5), along with diabetes duration (HR 1.1, CI 1.0, 1.1), WBC (HR1.2, CI 1.0, 1.4), and A1c (HR 1.3, CI 1.1, 1.7). The study indicates that the E/I, but not the 30:15, is an independent predictor of CAD, and thus the preferred simple measure. However, because the cohort was restricted to a mean diabetes duration of 27 years and free of CAD, no conclusions can be made about CAD incidence in shorter durations of diabetes. A lack of secondary sympathetic function tests in this analysis is also a limitation.


Symposium: Behavioral Interventions in Routine Clinical Care–What Works?


Paul Ciechanowski, MD, MPH (University of Washington, Seattle, WA)

Dr. Ciechanowski began with a discussion of how it feels to be stressed and explained how adding diabetes to the mix influences depression and anxiety. To address stress, he noted a problem-solving based treatment could be done with seven steps in a few minutes in a clinic visit and also mentioned a meta-analysis of problem-solving therapy (PST). The goal of PST is for patients to understand the link between symptoms and problems, be able to define their problems, show patients how they might resolve problems in a structured way, and to provide a positive experience with problem solving.

  • People with diabetes may often feel stressed as well as present with depression and anxiety. However, diabetes isn’t always the “big thing” or the primary problem itself. PST is one method to resolve problems (and may be more effective than other therapies) – Dr. Ciechanowski suggested if the problems resolve, the symptoms of depression and anxiety may often improve as well. As Dr. Ciechanowski described, PST helps patients understand the link between symptoms and their problems, as well as specify what the problem is. PST also provides an opportunity for the patient to have a positive experience with problem solving.
  • PST consists of seven steps. First, clarify the problem. When people are stressed, they may not be able to choose the right problem to solve. Define and write the problems down without judging. Then, set a goal. Next, brainstorm and generate possible solutions without shooting anything down (this is often where people get stuck). Determine advantages and disadvantages to each solution. Then, pick a solution. Describe the steps to the solution and implement them. Most importantly, come back and evaluate the solution. It is not important whether it was successful, but whether the patient was able to come up with solutions, act upon them, and take charge of the problem.
  • There is evidence to support PST as an effective treatment for depression compared to alternative psychosocial therapies and medications. Dr. Ciechanowski highlighted a meta-analysis of 22 PST studies that suggested significantly greater efficacy versus support therapy. Additionally, he noted PST has been shown to be effective in people with diabetes and other chronic diseases.

Questions and Answers

Dr. Elizabeth Walker (Albert Einstein College of Medicine, Bronx, NY) In DPP, we used a five-step problem solving to get to goal setting. Subsequently, it was used with visiting nurses as an intervention for diabetes and then with peers. I would like your experience with using it – I found that it was easier for community peers to learn and adopt problem solving for patients than it was for nurses. Thoughts?

A: We have noticed that people have to undo what they learned – not be psychoanalytical but stick to the seven steps. It’s hard for providers or therapists sometimes. Some people aren’t flexible enough to get it and stick to the structure – they get derailed at family of origin issues and make it too complicated.

Q: Is identifying the problem by the patient the most difficult part?

A: Yes. Sometimes patients have a hard time coming up with problems and don’t know where to begin. Once they get going, it’s fine. But there is a certain proportion of patients who have struggles starting.


Symposium: Women, Diabetes, and Cardiovascular Disease – The Perfect Storm


Kristin Newby, MD (Duke University, Durham, NC)

Dr. Newby opened by explaining diabetes itself is between a two and four-fold increase in the risk of cardiovascular events, an effect compounded when other factors (smoking, hypertension) are present. For women with diabetes, there is a doubling of risk factors compared to men, although conversely women are more responsive to exercise, alcohol rehabilitation, and diet interventions. Dr. Newby emphasized that hormone replacement therapy, as shown in the Women’s Health Initiative (WHI) and HERS trials, has no role in preventing cardiovascular events over four years, among women with or without diabetes. The WHI compared to HERS shows a similar relative increase in stroke, but no benefit on coronary heart disease events or death. She noted that aspirin is one area where guidelines are not gender neutral, largely because the Women’s Health Study (n=39,876), a randomized primary prevention trial, showed no significant difference in risk for cardiac events when women took aspirin versus placebo. Aspirin did, however, decrease the risk of ischemic stroke. Furthermore, she explained that up to 40% of diabetes patients are resistant to aspirin, and diabetes increases the prevalence of resistance to clopidogrel – the second line prevention therapy. Dr. Newby highlighted that a large randomized controlled trial of aspirin dosing in diabetes is currently underway, but for now recommendations based on the Women’s Health Study suggest aspirin as a class one medication only in high-risk women, and not as a routine myocardial infarction (MI) prevention therapy in younger women. To conclude, she asserted the need to design and analyze studies with gender in mind, to ensure women are representatively recruited to trials, and to raise women’s use of acute and chronic prevention therapies.

  • The Women’s Health Study (n=39,876), was a randomized, placebo controlled trial of low-dose aspirin (100 mg) in the primary prevention of cardiovascular disease. Women in the study were followed women for ten years. There was no significant risk reduction in overall cardiovascular events, with 477 cardiac events in women on aspirin compared to 522 on placebo (n=522), a risk ratio of 0.91 (p=0.12). Although there was a significant reduction in overall risk of strokes (0.83, p=0.04), there was an increase in the relative risk of hemorrhagic stroke (1.24, p=0.31), which raised concerns, according to Dr. Newby. Additionally, aspirin had no effect in preventing MI in women, with a relative risk of 1.02 (p=0.83). An age-based difference in aspirin efficacy surfaced, as women over 65’ use of aspirin, as compared to placebo use, led to 44 fewer MIs, strokes, or cardiovascular related deaths (p=0.008).



Elizabeth Barrett-Connor, MD (University of California San Diego, La Jolla, CA)

The highly regarded Dr. Barrett-Connor provided a packed room with an engaging overview of the burden and prevalence of cardiovascular disease in diabetes among different ethnic groups, maintaining a specific focus on women. Dr. Barrett-Connor began by emphasizing her view that the epidemiology of the diabetes burden has changed drastically in recent decades due to population growth, ethnic disparities, sex differences, changing life-styles, and growing stressors. Nevertheless, her conclusions were optimistic, stating that female death rates due to CVD are decreasing at the same rate as men. Presumably, that means that women have been changing their diet and responding to the interventions. Thanks to American surveillance of diabetes trends decades ago, the US has maintained a stronger level of attention to this epidemic than other nations. Dr. Barrett-Connor believes that we have achieved better treatment and prevention of type 2 diabetes and heart disease to some extent; however, access to affordable health care for women and minority groups still lacks in many regions.



Nanette Wenger, MD (Emory University, Atlanta, GA)

Half a century ago, heart disease was thought of as a “man’s” disease. Today, it is known that cardiovascular disease is the number one killer of women in the United States. Dr. Wenger is a pioneer in the movement that changed this paradigm, and in her talk today she questioned why cardiovascular disease disparately affects women with diabetes compared to men with diabetes (relative risk for fatal CHD is 50% higher in women with diabetes than men with diabetes). She approached the discussion by comparing disparities arising from a biological origin versus those occurring due to gender-based biases. According to Dr. Wenger, women with diabetes have higher CV risk profiles than diabetic men due to higher rates of hypertension, atherogenic dyslipidemia, worse LDL control, and clustering of comorbidities. Furthermore, Dr. Wenger presented a plethora of evidence for gender-based bias that prevent adequate CV treatment for women with diabetes.

  • Dr. Wenger’s summary clearly demonstrated gender-based bias in access to CV treatment for women with diabetes. For example, women receive fewer and lower doses of lipid lowering medications. Studies have shown that women have less access to aspirin, primary and secondary preventive therapies, and invasive intervention when necessary. Women on Medicare plans were also 19% less likely to receive appropriate care and LDL control compared to those on private plans, suggesting that gender-based biased is enhanced among minority groups.


Oral Session: Can We Rein in the Costs of Diabetes with Better Diabetes Care?


Suma Vupputuri, PhD (Emory University, Atlanta, GA)

Dr. Vupputuri reviewed a retrospective cohort study she helped conduct on the costs of chronic kidney disease (CKD). The study analyzed over 25,000 patients using the standard zero to four-stage index that categorizes patients based on estimated glomerular filtration rate (eGFR). Results suggested healthcare costs increased with higher stage CKD. Furthermore, patients who progressed from baseline (stages zero, one, or two) to higher stages of CKD had significantly higher healthcare costs compared to those who did not progress. A large percent (often >40%) of these costs stemmed from incident CKD-related disorders as opposed to basic CKD care.

Symposium: Current Status of Acute Complications of Pediatric Diabetes


Nicole Glaser, MD (University of California Davis, Davis, CA)

Dr. Glaser provided an informative overview of pediatric DKA and HHS. She prefaced her talk by describing the challenge she encountered in squeezing the abundance of noteworthy information concerning these conditions into a short 30-minute presentation. Dr. Glaser explained that much of the available data simply guide future studies and do not provide conclusive results that drastically change treatment approaches. Nevertheless, the dangers of pediatric DKA and HHS are well understood, and high-risk groups have been identified. Thus, Dr. Glaser emphasized the availability of opportunities for intervention and improvement of pediatric DKA and HHS treatment.

  • A number of high-risk groups have been identified in relation to DKA. Children under the age of two, ethnic minorities, and those with no access to health insurance are disproportionately affected by DKA. In addition, delayed diagnosis/diagnostic error contributes to increasing rates of pediatric DKA. Interestingly, there is a decreased frequency of DKA observed in children with at least one first-degree relative with type 1 diabetes.
  • Cerebral edema is common in children with DKA and is associated with detectable neurological injury. In 2006, Glaser et al. demonstrated that ventricle diameter was significantly smaller in children who had experienced DKA. Furthermore, they found that the apparent diffusion coefficient of water in brains of these children was lower than in those who had not experienced DKA. According to the study, this suggests that DKA causes vasogenic edema. Of interest, 55% of children with ventricular narrowing had some abnormality in mental status compared to just 21% of children who showed no ventricular narrowing. The subtle neurological dysfunction in children with DKA is concerning. For example, studies find that children who have had at least one episode of DKA demonstrate a significant deficiency in contextual memory compared to (two groups were nearly identical in every other aspect).
  • HHS frequency and mortality rates are rising: The extent and severity of this condition in children are often underestimated. Although adequate fluid resuscitation and electrolyte replacement is hard to optimize, Dr. Glaser indicated that it is imperative that we mobilize to improve our treatment options and provide prospective data to help augment fluid treatment of children with HHS. A 13-center prospective randomized trial at the Pediatric Emergency Care Applied Research Network will evaluate the rate of fluid infusion and rehydration optimization.

Questions and Answers

Q: Could you also explain acute memory deficits with reverse causation?

A: I think what speaks to that being less likely is that the groups otherwise were extremely similar, especially A1c and glucose levels. It seemed that they at least were controlling their diabetes similarly, so affects from this treatment did not seem to be affecting memory especially.


Barbara Anderson, MD, PhD (Baylor College of Medicine, Houston, TX)

In this talk, Dr. Anderson sought to present contemporary empirical evidence associating depression, anxiety, and psychological distress to pediatric diabetes and discuss psychosocial screening barriers. Her goal was to answer the question, “Should we be screening all patients/families with diabetes for depression?” Ultimately, Dr. Anderson’s answer was yes, citing the prevalence of depressive symptoms among individuals with diabetes and the short and long-term consequences of untreated depression in these individuals.

  • Most studies report a two to three times increased risk of depression in both youth and adults with type 1 diabetes compared to medically well individuals. Of 237 children with type 2 diabetes, 19.4% had neuropsychiatric disease as diagnosed by a psychiatrist (Katz LE et al., Ped Diab 2005). According to Dr. Anderson, the complex interrelationships between ADHD, depressive symptoms, parenting stress due to diabetes-specific family contact, and deteriorating glycemic control get caught in a vicious cycle, ultimately resulting in more ER visits and hospitalizations, which return to further complicate these contributing factors.
  • Commonly reported barriers to psychosocial screening include questions about who should conduct it, the stigma surrounding it, and the question of reimbursement. To quell some of these concerns, Dr. Anderson believes that psychology trainees at the internship and fellowship level are competent for the job of screening children with diabetes for depressive symptoms, which she argues would come at no extra cost to hospitals.
  • Dr. Anderson strongly asserts that all patients diagnosed with diabetes should be screened for psychosocial distress. Considering the strong link between depression and short and long-term complications, as well as current ADA guidelines for individuals with type 1 diabetes, Dr. Anderson strongly believes that we must overcome the barrier of reimbursement now and document the cost-effectiveness of screening in future research.


Symposium: The Holy Grail of Advancing Therapeutics for Diabetic Kidney Disease


Sharon Adler, MD (Harbor-UCLA Medical Center, Los Angeles, CA)

Dr. Adler reviewed some of the newer pharmacologic agents used to treat diabetic nephropathy (DN). Combining existing medication classes is currently a contentious issue in nephrology. OnTARGET was a randomized controlled trial comparing three arms: telmisartan (an angiotensin II inhibitor), ramipril (an ACE inhibitor), and a combination of the two. The patient cohort had diagnoses of diabetes, coronary artery disease, and additional risk factors. Surprisingly, the trial failed to show any differences between the groups and even showed additional renal risk with the combined therapy. Triple therapy is also a potential choice (adding spironolactone), but it likely carries more risk than benefit. Dr. Adler concluded that combination therapy is not well studied enough to either rule the practice out or to use it routinely. She shared a summary of an unpublished study by Zhang et al. (Diabetes 2012) in which early outgrowth cells (a type of bone-derived stem cell) reduced oxidative injury to the kidneys in a mouse model. Results also showed a great improvement in albumin-creatinine ratio over a short period of time, seemingly due to the amelioration of oxidative effects and reversal of fibrosis. Moving forward, Dr. Adler proposed the hot field in diabetic nephropathy drug development is inflammation and fibrosis. Reata’s bardoloxone methyl represents a new class of anti-inflammatory, anti-oxidant medications that modulates the Nrf2-KEAP pathway. Pirfenidone is a small molecule antifibrotic that shows promise for diabetic nephropathy. It has shown in experimental models to reverse and inhibit fibrosis. Like bardoloxone, it improved eGFR but not ACR in diabetic nephropathy.

  • Dr. Adler reviewed some of the newer pharmacologic agents used to treat diabetic nephropathy (DN). She noted that glycemic control can only go so far in treating and preventing diabetic kidney disease, since the upstroke on the “J” curve for mortality and A1c suggests a floor for the amount of glycemic control that physicians should expect from their patients.
  • One failed agent for diabetic nephropathy was alagebrium, a drug that acted to reverse glycosylation of AGEs. It also modulated some of the oxidant pathways of DN by attenuating superoxide activity.
  • Combining existing medication classes is currently a contentious issue in nephrology. OnTARGET was a randomized controlled trial comparing three arms: telmisartan (an angiotensin II inhibitor), ramipril (an ACE inhibitor), and a combination of the two. The patient cohort had diagnoses of diabetes, coronary artery disease, and additional risk factors. Surprisingly, the trial failed to show any differences between the groups. The only variant outcome it did show was that serum creatinine suffered when the two agents were combined versus using them individually. Many nephrologists took issue with the interpretations of this study due to specifics of the study design. A year after OnTARGET, a meta-analysis showed that proteinuria was improved in dual therapy.
  • Triple therapy is also a potential therapy choice, but it likely carries more risk than benefit. Combining spironolactone (a potassium-sparing diuretic), and ACE inhibitor, and an angiotensin II blocker lowers proteinuria and SBG but raises serum creatinine and potassium in diabetic patients. Unfortunately, the hyperkalemia was more clinically significant than the changes in proteinuria and blood pressure.
  • Dr. Adler concluded that combination therapy is not well studied enough to either rule the practice out or to use it routinely. Taken together, she believes that the data surrounding combination ACE inhibitors and angiotensin II blockers is not convincing and recommended against its use.
  • As an aside, Dr. Adler noted that several studies have suggested avoiding a systolic blood pressure below 120 mmHg in patients with diabetic nephropathy. These included ROADMAP, ACCORD, IDNT, ONTARGET, and INVEST.
  • Finding new targets has been largely driven by unbiased “-omics” studies. In particular, looking at the transcriptomes in diabetic nephropathy has revealed some targets for future medications. The results of these studies have pointed to proteins mediating an expected culprit, oxidant species, as potential mediators of the disease.
  • Dr. Adler shared a summary of an unpublished study by Zhang et al. (Diabetes 2012) in which early outgrowth cells (a type of bone-derived stem cell) reduced oxidative injury in a mouse model. The research group extracted the stem cells from bone, then cultured and reintroduced them into circulation. They found a great improvement in albumin-creatinine ratio over a short period of time, seemingly due to the amelioration of oxidative effects and reversal of fibrosis. Mysteriously, however, the early outgrowth cells were not found in the kidney. Nearly every other organ showed plentiful early outgrowth cells growing within them after treatment. Further investigating this effect, Zhang et al. showed that early outgrowth cell-conditioned medium (the extracellular matrix exposed to the stem cells) was able to induce similar effects in cultured kidney cells in vitro. This suggests that there is a hormonal effect rather than a direct interaction leading to this improvement in kidney function.
  • Dr. Adler thinks the hot field in diabetic nephropathy drug development is inflammation and fibrosis. Persistent inflammation causes fibrosis in the kidney (and most other tissues). Since there is chronic inflammation in the kidney in diabetics and fibrosis is an important of diabetic nephropathy, this suggests that addressing inflammation is a valuable target. Bardoloxone methyl represents a new class of anti-inflammatory anti-oxidant medication that modulates the Nrf2-KEAP pathway. Early results have shown rapid eGFR improvements (+10.1 ml/min/1.73m2) at four weeks. The reasons for this are not yet well understood, however, and larger trials are underway.
  • Pirfenidone is a small molecule antifibrotic that shows promise for diabetic nephropathy. It has shown in experimental models to reverse and inhibit fibrosis. Like bardoloxone, it improved eGFR but not ACR in diabetic nephropathy. Gastrointestinal symptoms limited tolerability (and study completion) in the high dose group in clinical trials, however.
  • The thrombin-thrombomodulin Activated Protein C (APC) pathway (a pathway usually involved in blood clotting) confers cytoprotection in diabetic nephropathy. APC enhances the survival of the endothelium, the activity of the podocytes, and overall seems to have an anti-apoptotic effect in kidneys.


Aliza Thompson, MD (US Food and Drug Administration, Silver Spring, MD)

Dr. Thompson began by defining surrogate endpoints as biomarkers used in therapeutic trials as a substitute for a clinically meaningful endpoint, and highlighted that they are expected to predict treatment effects on outcomes of clinical interest – mortality and end-stage renal disease (ESRD) in the case of diabetic nephropathy studies. In terms of standards for accepting surrogate endpoints, there must be 1) a pharmacologic or physiologic basis for thinking the biomarker is on the causal pathway; 2) observational data showing a relationship between the biomarker and clinical outcomes of interest; and 3) data from interventional trials showing that a treatment’s effect on outcomes is predicted by its effect on the biomarker. After positing blood pressure as a successful surrogate endpoint, she moved to diabetic nephropathy, commenting on the complexity of deciding if creatinine levels can serve as a surrogate endpoint. Dr. Thompson heavily emphasized that risk markers do not automatically become surrogate endpoints, utilizing the examples of low hemoglobin and albuminuria. According to Dr. Thompson, potential surrogate endpoints fail because the assumed relationship between the marker and clinical outcome is wrong or the drug had unexpected adverse effects that minimize its benefit or resulted in harm. Currently, using a surrogate endpoint in a trial may allow shorter trials and smaller numbers of patients, but there is a post-marketing commitment to complete studies verifying the drugs’ clinical benefit. She additionally emphasized that using surrogate endpoints should mean that trials gather even more safety data, because there is a lower tolerance for risk due to residual uncertainties on the validity of the surrogate endpoints predictive ability on primary endpoints.

  • Dr. Thompson presented blood pressure as a successful example of a surrogate endpoint, referring to trials that demonstrated treatment effects on blood pressure predict treatment effects on stroke incidence. The multiplicity of drugs and mechanisms involved in these trials helped establish blood pressure as a surrogate endpoint by raising the likelihood of it being on cardiovascular events’ causal pathway.
  • Although doubling of creatinine levels are linked to ESRD and death, it is unclear whether lesser changes in creatinine and glomerular filtration rate (GFR) decline can act as surrogate endpoints in diabetic nephropathy studies. Dr. Thompson noted that the National Kidney Foundation and FDA are co-sponsoring a workshop in December 2012 to examine the data supporting various levels of GFR decline as endpoints in clinical trials.
  • Risk markers do not equal surrogate endpoints, as shown by the failure of ameliorating low hemoglobin to improve cardiovascular mortality in the NHS, CHOIR, and TREAT trials despite being an established risk factor for cardiovascular events. Dr. Thompson drew an analogy to albuminuria, a clear marker of ESRD risk, which has little data from intervention trials showing that treatment effects on albuminuria predict treatment effects on renal outcome in people with diabetic nephropathy. Since the causal pathway of ESRD is unclear, it will be difficult to establish that albuminuria is actually mediating the progression of renal disease in nephropathy.



Peter Chuang, MD (Mount Sinai School of Medicine, New York, NY)

Dr. Chuang reviewed the types of natural oxidants and antioxidants, emphasizing that oxidant stress occurs when there is an imbalance of these molecules with an excess of reactive oxidants. After explaining the variety of markers for oxidative stress (oxidative modification products, direct measurements, loss of endogenous antioxidants), he demonstrated that in people with type 1 diabetes, products of oxidative modification – malondialdehyde and protein carbonyl groups – were elevated compared to control, while the antioxidants were reduced. Additionally, he noted that in a Joslin study (n=410) where participants with type 1 and type 2 diabetes were followed for eight to ten years, significantly elevated levels of TNRF predicted the development of end-stage renal disease. Utilizing Brownlee’s Unified Mechanism of Hyperglycemia Tissue Damage, he demonstrated that reactive oxidative species activate pro-inflammatory pathways, and furthermore that inflammation is complicit in the pathogenesis of diabetes and diabetic kidney disease through a variety of pathways. In people with type 2 diabetes, activation of the NFκB receptor due to reactive oxygen species and accumulation of glycation end products leads to the activation of inflammatory pathways, making the pathway involving this receptor an important target for diabetic kidney disease treatment. Other potential targets include growth factors and transcriptional pathways involved in the inflammatory process underlying diabetic kidney disease.

Oral Sessions: Diabetic Dyslipidemia


Juraj Koska, MD, PhD (Phoenix VA Healthcare System, Phoenix, AZ)

Dr. Koska described the results of the Risk Factors, Atherosclerosis, and Clinical Events in Diabetes (RACED) sub-study (n=262) of VADT regarding the effect of intensive glucose-lowering therapy (INT) on standard plasma lipids and cholesterol content. The Vertical Auto Profile (VAP) test was used to measure lipoprotein subclasses in plasma at baseline and nine months following randomization. Despite increasing BMI, intensive glucose-lowering therapy reduced triglycerides and increased total HDL2 cholesterol. While total LDL cholesterol did not change, LDL4 was substantially reduced from baseline, while LDL2 was significantly increased. This effect of intensive glucose-lowering therapy on LDL2 and LDL4 was independent of diabetes medications used (rosiglitazone or insulin). Thus, although intensive glucose-lowering therapy lacked benefits on CV risk as reported in VADT, this therapy has been shown to have favorable lipid profile effects.

Meet the Expert Sessions


Henry N. Ginsberg, MD (Irving Institute for Clinical and Translational Research, New York, NY)

Responding to a variety of questions about lipid management in type 1 and type 2 diabetes, Dr. Ginsberg endorsed the wide use of statins among people with type 2 diabetes, saying that small-to- moderate increases in average blood glucose were probably counteracted by statins’ well-demonstrated CV benefits. He indicated that LDL-lowering – whether by statins, diet, or bile acid sequestrants – seems to be the clearest way to cause CV benefit. The evidence of benefit is murkier for lowering triglyceride levels and/or raising HDL (especially given the recent negative results for Roche’s HDL- raising drug dalcetrapib – Dr. Ginsberg’s opinion is that low HDL is a marker of CV risk, but not a particularly useful therapeutic target).

  • In the midst of discussing the ambiguities and imperfections of various CV trials, Dr. Ginsberg shared insight into the design of the ACCORD study (he was a researcher in ACCORD Lipid). He said that the 6.0% A1c target for intensive control was not chosen because the researchers thought it should be the glycemic target in the real world, but because they calculated that to show benefit, they would need a 1.5% A1c difference from the standard control group (for whom the A1c target was 7.5%, because the researchers thought that a higher target in the standard control group would be unethical). We found this a fascinating insight into ACCORD and the compromises involved in all study design, and we can’t help wondering how the results might have been different if ACCORD’s enrollment had been large enough for the glycemic targets to be more realistic.

Questions and Answers

Dr. Ginsberg: Here is a short preamble to start us thinking – probably just about every patient with type 2 diabetes ought to be on a statin. Maybe we can start with that.

Q: Can you talk about the potential risk of diabetes associated with statin use?

Dr. Ginsberg: Obviously I am not going to cause diabetes in my patients with diabetes. Will their glucose control get worse? I don’t know. The CARDS study suggests that A1c might have gone up, but they weren’t looking systematically. To date in these trials, data are derived from visits that occur maybe every three months initially, and every six months or every year as the study goes on – one question in the questionnaire might be, do you have diabetes. If the patient says yes, I think it typically means they have A1c or glucose above a certain point, not that they had a hyperosmolar coma or are very symptomatic (though I can’t say this with 100% certainty). This question will be addressed in Hot Topics in Lipids, a session that I think is on Sunday afternoon. 

If you look at people in these trials and classify by number of characteristics of metabolic syndrome, their chance of diabetes goes up during those trials. If you have a patient with fasting glucose 115 mg/dl that you are treating for elevated lipids and you put them on a statin, you need to follow glucose. Would I take them off the stating if they got diabetes or went up a couple tenths on A1c? If I put them on statin because the data say how beneficial statins are, I wouldn’t take them off. As a lipid specialist, I probably deal with a lot more niacin than others. In niacin research, people’s fasting glucose went from 115 mg/dl to 175 mg/dl. I didn’t have much evidence that niacin was beneficial, and I have even less now with AIM HIGH. But with statins I think we have established a good benefit/risk ratio in people with diabetes…. I haven’t seen enough primary data to say this absolutely, but I think that is the case. Even in the Niaspan and short-acting niacin studies about ten years ago – if people with diabetes were in good control, they just modified their treatment regimen and A1c barely went up. 

It is a benefit-risk decision to make. For someone with diabetes that has clinically relevant CV disease, the number needed to treat with statins is probably 20 to save one person. The risk of worsening / developing diabetes might be in the few hundreds. Why is this happening? Hopefully we will find out. There was fairly substantial literature in 1990s and 2000s that looked at whether statins might be insulin sensitizers the thought was that since they are good for CVD they must be good for diabetes, too. Those studies were essentially a wash, considering publication bias.

Q: Could you talk about treating people with triglycerides over 1000 mg/dl?

Dr. Ginsberg: Those people are unusual, and TGs that high are more common among people with diabetes. If you look at the distribution of triglyceride levels, the 90th percentile in the non-diabetic population is probably a little over 200 mg/dl, whereas among people with diabetes, 200 mg/dl that is probably only the 80th percentile or even lower. In ACCORD Lipid, the cutpoint for the upper tertile of TG was 204 mg/dl (though we had enriched the study for high TG). When you look at people with extremely high TG levels, you start to see genetic associations. Robert Hegele has studied roughly 600 people who have had TG over 1,000 at some point; he has come up with roughly 20 alleles, including the gene for angiopoietin, but there is not enough sensitivity to use these diagnostically. 

If someone’s diabetes is out of control – even A1c 9.0% or above – the lipid drugs usually don’t work that well, so you really need to get better glucose control. As for lipid drugs that work relatively well – omega-3 at least 3 or 4 mg / day; I use the concentrated prescription version because the OTC version usually has a lot of inactive ingredients in the pills, so you have to take a lot of them. Fibrates also work in those people. Is one better than the other? I don’t think so – both lower TG around 35-40%. Atorvastatin has been shown to give a 40% reduction in those with TG 600 mg/dl or above. Most statins give similar results but I wouldn’t use them first line. 

Obviously you have to make sure people are not using alcohol; that will clearly make them worse. The immediate cure is a fast – they plummet their TG. Even those not obese will respond to caloric restriction.

The risk with high TG is pancreatitis. Though if you use the evidence-grading criteria, what is evidence for pancreatitis and TG? Probably a B or a C – there has never been a trial and there can’t be; the frequency and incidence is very low. Some people walk around with TG of 2,000-to-3,000 mg/dl and never get pancreatitis, while others get pancreatitis at 800 mg/dl.

To sum up, I think the factors to consider are calories, alcohol, fish oil, fibrates, and maybe statins.

Q: Fibrates are not very potent. Many companies have made efforts to come up with more potent PPAR-alpha agonists. Any thought on this?

Dr. Ginsberg: Some PPAR-alpha candidates were in development that, as seen in a ligand-binding assay, were orders of magnitude more potent than the ones we have. But they didn’t do much more for TG, and sometimes they caused severe myositis. I think the reason we don’t see myositis much with the fibrates we have is that they are low in affinity. If you take potent PPAR-alpha, a little bit of off-target exposure can be bad. Aleglitazar is still in development. One company is targeting creatinine and homocysteine, which go up with fenofibrate and less with gemfibrozil. Estimated GFR went from 90 to 60 in a couple hundred people we followed, because creatinine went up more than 20% in the trial. But within six weeks of cessation, the estimated GFR and creatinine came back to where the placebo group was. The increase in homocysteine seems reversible too; we have funding to study this. If creatinine and homocysteine up in a trial, is that bad? We see no adverse signal among those whose creatinine went up more than 20% vs. everyone else, though it was a negative trial overall.

Q: What is the evidence that triglycerides are an independent risk factor for microvascular and macrovascular complications?

Dr. Ginsberg: Studies going back 40 years have associated TG levels with retinopathy, especially fluffy exudates. Probably the most positive thing out of ACCORD Lipid was that we showed a reduction in retinopathy risk as shown by photos. There is no indication, nor did the companies ever file for an indication, for fibrates in retinopathy. But in both FIELD, of fibrate monotherapy, and our study [ACCORD Lipid], of fibrates on statin background, there was roughly 40% reduction in progression by photographs. We haven’t looked at kidneys in ACCORD Lipid; the use of fenofibrate was associated with decreased micro- and macroalbuminuria in FIELD.

The evidence that TG are independent predictors of macrovascular disease has been controversial for a long time. When you look at plaque in humans, you see cholesterol. But in animal studies, we’ve seen that TG gets into foam cells, and then fatty acids leave but cholesterol remains in the plaques. Why isn’t TG considered more of a target? It tends to be inversely related to HDL, and when you do a study HDL always ‘wins’ [Editor’s note: i.e., has a stronger association with outcomes]. Most recently in JAMA, in a mega- meta-analysis by John Danesh and his group, TG was a predictor of adverse events. The association remained, albeit weakened, even when HDL was put into the equation. Including non-HDL cholesterol wiped out the significance of TG, but non-HDL is essentially TG. Now that we are having second and third thoughts about what HDL means, some people are saying that maybe a lowering of HDL just means apoB or TG increased.

The trials that have been done, where TG was changed, often changed LDL and/or HDL as well. Most major studies of fibrates for triglyceride lowering have been negative, but post-hoc subgroup analyses suggest benefit in those with high TG to start. In ACCORD Lipid we prespecified an analysis, but this was still just a subgroup analysis. If everyone in your study had TG of at least 175 and HDL in the mid-30s or below, I think we’d have a positive trial, but we would have to run it to know for certain. I have never used fibrates for people with TG in the mid-100s, even if they have low HDL, because the efficacy is not that good. I do use it in the upper-TG/lower-HDL tertile group where benefit is suggested.

Q: Given the negative result for niacin in AIM HIGH, what are we left with when patients are on high doses of atorvastatin, but their non-HDL cholesterol is not at goal? What is the next-best drug to add to a statin in type 2 diabetes – Zetia? A bile acid sequestrant?

Dr. Ginsberg: I would say that for someone with LDL of 80 mg/dl, TG of 250 mg/dl, and HDL 32 mg/dl, the evidence base suggests that gemfibrozil is the only drug we have that’s been positive. With this there is risk of myositis and fear of rhabdomyolisis – the absolute risk is still low with gemfibrozil, but is at least tenfold higher than with fenofibrate.

Niacin trial was very disappointing. Why? A lot of speakers that invest greatly in niacin or in HDL say the investigators were stupid because there were not enough subjects – the researchers expected a 25% differential in HDL and never got close, and that’s why the results were negative; the researchers were criticized because it was only a 3,000-person study. But it was not the investigators’ fault that there was no more money; they were constrained in trial design.

A quick aside – in ACCORD; it wasn’t that we thought the A1c target should be <6.0%; no one thought that’s what we were trying to do. Rather, we needed a 1.5% difference from the standard group, and we thought 7.5% was as high as we could put their target while still being ethical.

In niacin’s case – there was a 15%-or-so bigger rise in HDL than with control, but they got a zero-percent benefit. Possibly the study was just underpowered. Note that the LDL differential was only about 5 mg/dl, from around 67-68 mg/dl to around 63 mg/dl. The investigators wanted an HDL trial without an LDL differential, even though niacin also has LDL benefits. So the placebo group had higher rates of ezetimibe and statin use in order to get LDL to similar levels. (There were still a lot of CV events, though, despite the LDL control.) Why was the study negative? I don’t know.

The most recent negative trial was that for dalcetrapib – the study was stopped for futility. You have to wonder whether changing HDL will be valuable no matter how you do it. So if TG is high, HDL is low, and LDL is 80 mg/dl, the best evidence is probably to lower LDL to 60 mg/dl. 60 mg/dl was better than 100 mg/dl in JUPITER, and 70 mg/dl dl was better than 100 mg/dl in several studies. The ongoing IMPROVE IT study is comparing between high-60s and low-50s – if that turns out to be positive, I might even say to take LDL down close to 50. Barring that, you could take a leap of faith and try to lower TG and raise HDL, but we have no evidence. The Heart Protection 2 THRIVE study is looking at 25,000 people – over eightfold more than in AIM HIGH – to compare niacin/simvastatin vs. simvastatin alone. We’ll see if those results, still a couple years away, are positive.


Knut Borch-Johnsen, MD, PhD (University of Southern Denmark, Odese, Denmark)

Dr. Borch-Johnsen began with a question for the audience. Does the treatment for hyperglycemia lead to cardiovascular disease? And does it depend on whether that treatment includes only non-insulin therapies or insulin? The audience’s mixed response set the stage for discussion. Dr. Borch-Johnsen explained that past studies have shown conflicting results regarding the association between insulin level and probability of cardiovascular disease. The association between increasing A1c levels and increased cardiovascular disease risk, however, has been well established. The key, he reminded the audience, was whether treatment of hyperglycemia leads to increased risk for patients. While data from the ACCORD and ADVANCE studies have reported that patients in intensive care groups have greater overall mortality, he did not find this association in recent data from the ADDITION study. Dr. Borch- Johnsen believes that the consistent finding of past studies relating intensive care treatment with increased patient risk can potentially be attributed to confounding by indication – where patients taking insulin do not have the same disease profile as patients treated by diet alone or oral hypoglycemic agents. He concluded with his stated intent to raise concern over whether we are over interpreting the data we have regarding insulin treatment and patient risk.

  • There is conflicting data relating insulin levels to the probability of cardiovascular disease. Dr. Borch-Johnsen explained that studies following patients over only a few years have found an increasing risk with increasing levels of insulin. However, in studies going beyond 10 years of follow up, Dr. Borch-Johnsen noted that the association disappears (this includes populations with diabetes). In contrast, a plethora of studies have confirmed the association between increasing hyperglycemia and increased risk for cardiovascular disease.
  • Dr. Borch-Johnsen reminded the audience that the key question was whether treatment of hyperglycemia leads to increased cardiovascular disease. This is an important question, since the contentious p value (p=o.052) in UKPDS showed a 17% risk reduction for cardiovascular disease in the intensive care group. Dr. Borch-Johnsen shared with the audience what his biostatistics professor once told him: “You have to be medically qualified to tell the difference between a p-value of 0.049 and 0.052.” Regarding this study, Dr. Borch- Johnsen made the important qualification that he did not consider the UKPDS study to be an intensive care study. He described it as an insulin study where the targets reached in the study were far from those we strive for today.
  • Unlike ACCORD and ADVANCE, results from the ADDITION study do not indicate a difference in overall mortality between people with diabetes treated conventionally or intensively. The ADDITION study screened for people with previously undiagnosed type 2 diabetes and randomized patients (n=3,000) to receive either standard care as recommended by local health authorities, or an intensive care regimen that included lowering blood pressure, lipid levels, and an A1c below 7%. In this study, Dr. Borch-Johnsen found no sign of differing overall mortality between the two groups, begging his question of why studies are consistently showing higher mortality in patients on insulin therapy compared to those on diet alone or other hypoglycemic agents.
  • Dr. Borch-Johnsen hypothesized that confounding by indication can explain why studies have consistently shown higher mortality in groups treated with insulin vs. groups treated with diet or oral hypoglycemic agents. He explained what might happen when a patient comes in with diabetes. First, the patient is instructed to change his diet. If the patient reaches his target on diet alone, treatment does not change. If diet does not work, oral hypoglycemic agents are prescribed. Again, if the patient reached his target, treatment would remain as such. If oral hypoglycemic agents do not work, then insulin is prescribed. This means we are not comparing identical groups – the insulin treated group could not reach goal by any other means and are thus likely to have more advanced diabetes than those on diet or orals alone. This, said Dr. Borch-Johnsen, is confounding by indication. He remarked that only an appropriate clinical trial comparing diet, oral hypoglycemic agents, and insulin therapy could assess the comparative risk between insulin therapy and cardiovascular disease or mortality risk. However, such a trial would be unlikely given ethical considerations.
  • Dr. Borch-Johnsen concluded with his own take – the literature is convincing, but he is not personally convinced. While he said he could not give a definite answer that insulin is or is not increasing a patient’s risk, he did want to raise concern that we may be over interpreting some of the data we have.

Questions and Answers

Q: The initiation of insulin therapy is occurring earlier in some settings. A patient may go straight from metformin to insulin for example. Does that put another spin on what you were explaining?

A: It may give us some additional insight on whether insulin puts individuals at a greater risk or not. Going back to data from ADDITION, if we start patients earlier on insulin, we are on a much safer ground primarily because patients have a lower risk of cardiovascular disease at this point. We still have concern about patients with past cardiovascular disease or arrhythmias and lowering glucose levels too far.

Q: Can you separate insulin resistance from hyperinsulinemia?

A: It’s tough, because it is rarely done. Insulin is measured at baseline and in a fasting state, and that is used as a proxy for insulin resistance. It doesn’t say if it is peripheral or hepatic or to what extent it is true insulin resistance. To be frank, we don’t know that much about insulin resistance per se because the studies we are doing only give limited insight.

Q: If you had two patients – one patient with type 2 diabetes receiving 120 insulin units and an A1c of 7%, and one patient receiving just 50 units of insulin and an A1c of 8% - which would you prefer to treat from a cardiovascular perspective?

A: I would not be able to say without the patient histories. If I knew of a strong family history of cardiovascular disease or current cardiovascular disease risk in the patient, I would be reluctant to increase insulin levels. But if the patient is young and stands to live with the disease for 30-40 yrs – as most people with type 2 diabetes in the future will – I would carefully weigh the risk of microvascular complications. An increase in A1c from 7% to 8% corresponds to 40% increase risk in retinopathy. It would do the patient a lot of good to get his A1c down.

Q: In intensively insulin treated patients, what do you think the role of hypoglycemia is as far as cardiovascular incidents are concerned?

A: Hypoglycemia may explain cardiovascular disease deaths on intensive treatment. But again, we are in deep trouble here. Once we find the patient, and the patient has died, there is no way to find out what glucose level the patient had when he died. We assume in the absence of other explanations, that the patient could have died from a hypoglycemic event. If you try to lower glucose levels in patients with cardiovascular disease, you introduce electrophysiological changes that could lead to cardiovascular events. You have to be careful.

Q: Do you believe new drugs like GLP-1 analogs will reduce the amount of insulin we use to treat patients with obesity?

A: Five to 10 years ago, I would have said we would use insulin to bring their blood glucose level down as far as possible. Today we have alternatives that can lead to a much lower risk of inducing hypoglycemia. So for severely obese patients, some other kind of treatment would be used. My only concern is that when we introduce a new drug, we may be running a risk because we don’t know the long-term consequences. It was only after 90 years of experience with insulin that we began to question whether it was associated with increased cancer risk.

Q: What is the role of SFU’s in the future given their limited sustainability in their effect and possible association with cardiovascular disease according to ADVANCE?

A: Most clinicians and most writing guidelines are increasingly having concerns about the role of SFU’s in the future. Our problem has been that we had a limited number of other possibilities, but that is rapidly changing. We don’t have sufficient long-term data on new drugs. But if they end up proving true to their promise, my clear assumption would be a change in the partition of the drugs in the future. Newer drugs would take a higher position.


Timothy S. Kern, PhD (Case Western Reserve University, Cleveland, OH) and George L. King, MD (Joslin Diabetes Center, Boston, MA)

Drs. Kern and King eschewed a presentation entirely, offering a two-sentence talk – which really must be some sort of record – that simply defined microvascular complications of diabetes: neuropathy in the nerves, retinopathy in the eyes, and nephropathy in the kidneys. He then threw the entire remainder of the session open to a freewheeling question and answer session. The most important questions pertaining to each of the three major types of complication are included below.

Questions and Answers

Q: What is the current evidence for drugs to reverse retinopathy and is there any evidence for drugs to reverse neuropathy?

Dr. Kern: To the best of my knowledge, only one study claimed reversal of retinopathy. Dr. Mauer did a kidney-pancreas transplant and over 15 years of normal glycemia claimed reversal of retinopathy. In terms of other clinical studies and certainly all the animal studies, you can inhibit progression, but not really say there is a real reversal. No one has claimed that with the exception of this paper. With neuropathy there is considerably more data but it’s more ambiguous.

Dr. King: For reversibility, the best data is in the kidney. Certainly Mike Mauer’s pancreatic transplant as well as Andre Koleski and Bruce Perkins’s New England Journal paper in 2002, if you follow 600- something type 1 diabetic patients with microalbuminuria, you check results six years later and a third of them don’t have microalbuminuria anymore – eGFR hasn’t really decreased. This suggests there is reversibility.

Q: Are there differences in microvascular pathology and different medicines that are affected in kidneys vs. the nerves?

Dr. Kern: Are there structurally different, morphological differences? I guess the one that’s most common in the retina is the capillary degeneration that occurs. That’s been found in skeletal muscle but not identified anywhere else. What seems to account for most capillary degeneration, or at least obliteration, is retinal vessels are surrounded by glial cells unlike any other tissues. If you degeneral stain, they actually maintain normal diameter. That seems strange for a blood vessel that’s no longer functioning. If you stain for glial cells, it’s got processes running right down the blood vessel. That seems to stop degeneration of the blood vessel. Whether that causes or is the result of degeneration, to the best of my knowledge nobody knows. For the kidney, there doesn’t seem to be as much. There’s some thickening that seems to happen anywhere, but the degeneration of the gliaral musculature doesn’t seem to happen anywhere like the retina. With regard to nerve, there have been papers reporting vascular density in the nerve but nobody has ever shown whether that’s a mathematical thing or real structural damage going on.

Dr. King: But I think it’s a very important question that you asked. I think structurally Tim has said, but responses of the microvascular to diabetes or hypoglycemia seem to be quite different. So in the kidney, for example, you have increased VEGF production, while in the retina you have decreased production. You could say there’s hypoxia in all these, why is there a difference? So that makes treatment quite difficult. Because in the retina you’re using an anti-VEGF inhibitor for late stage diabetic retinopathy, but how are you going to do that systemically? In the kidney now there’s controversy whether the VEGF increase is bad or good. There’s 50/50 split over whether VEGF is important for kidney or glmerular structure. VEGF treatment may be bad for nephropathy.

Q: I was just wondering about the kidney. Is there any conclusive evidence suggesting hyperfiltration has implications for late stage retina damage?

Dr. King: Hyperfiltration clearly occurs when there’s poor glycemic control whether it’s in humans or animals. That happens in probably 100% of patients. Only 30 to 40 percent of people have decreased GFR. So the argument then is do you have to go through hyperfiltration stage to develop decreased GFR? Obviously, not everyone will have hyperfiltration-related complications. So the one side is, “All decreased GFR have hyperfiltration.” The other camp says these aren’t related because everyone has hyperfiltration. So here, some of the new thinking is to break this logjam – it’s possible that hyperfiltration itself could cause issues, but it’s also possible that that will induce the body to have counterregulatory defenses. So it’s possible that in those people who reverse, those people could have their body’s own mechanism that kicks in and reduces damage once hyperglycemia has caused the initial effects. You have toxic effects, dodging effects, defensive effects. That’s the reason we’ve been looking at this, and I’ve been looking at diabetes complication methods for at least 25 years. We have to admit it’s a terrible failure. We don’t have a single drug on the toxic side that’s clinically approved. Even ACE-B inhibitor doesn’t change pathology in the kidney all that much. There’s nothing other than anti-VEGF, but that isn’t anti-hyperglycemia’s toxic effects.

Q: Coming back to retinopathy, is there a molecular mechanism involved in vascular regression?

Dr. Kern: So early on in the background of non-clinically important stages of diabetic retinopathy, individual blood vessels die here or there. It’s not arterial disease like hypertension where you have everything downstream of arterial occlusion, it’s just blood vessels here or there. The simple way to look at this is that these are individual insults to vasculature, and over time it can’t repair itself. Some thought that white blood cells along with capillaries might contribute to this subclinical event – 24 hours it goes way and doesn’t come back, but it’s never been proven. But in animals, if you block white blood cells’ interaction with endocapillary cells, the capillary doesn’t die. So there’s a lot of evidence showing if you block this pathway, you can stop degeneration. In my opinion, this concept of degeneration is what best explains loss of some capillaries. Over time as they become larger and larger they begin to group together into a clinical event, and hypoxia starts. There are probably 15 or 20 animal studies showing these can be blocked using a variety of different pathological approaches. The simplest way to group them together is thinking in terms of inflammation.


David G. Armstrong, MD, PhD (University of Arizona, Tucson, AZ)

Dr. Armstrong gave an enthusiastic talk about general treatment of the diabetic foot. The talk began by referencing the comprehensive information about the diabetic foot available at and @dgarmstrong. He emphasized the importance of screening patients with diabetes for neuropathy, or, as he calls it, loss of protective sensation (LOPS). He noted that his favorite tool for this is the monofilament, but also highlighted the ease and aptitude of the Ipswich “touch test” (a simplified approach where the first, third, and fifth toe are touched – this has been lauded for its simplicity and low cost, even perhaps above monofilaments). Dr. Armstrong explained the dichotomous branches of the ADA’s risk score index for podiatric diabetes complications (0-3), noting that previous history of amputations and/or ulcers easily place a patient into the highest risk category (3). A number of questions were asked about the logistics of maintaining effective facilities for treating the diabetic foot, the most insightful of the answers pointed to the importance of individual HCPs aiming to avoid amputation. Dr. Armstrong explained that the mindset of the HCPs at any given hospital was a very important factor in the ratio of above-the-foot to below-the-foot amputations (a higher ratio is better). This is because some HCPs view, for example, a diabetic foot ulcer as only the beginning of a number of problems and jump to amputate, while others strive to save it in the hope that the foot will survive to be effectively utilized by its appreciative owner.


Lawrence Fisher, PhD (University of California San Francisco, San Francisco, CA)

This presentation covered diabetes distress and its negative implications. Feelings like stress and anxiety resulting from diabetes management contribute to diabetes distress. These feelings are not to be confused with those associated with major depression, which can carry a negative connotation. Healthcare providers often feel as if they do not have adequate time or training to deal with diabetes distress. However, Dr. Fisher asserted that it is crucial to acknowledge, normalize, and accept diabetes distress because improved glycemic control and diabetes management will result.

  • Preventing diabetes distress is critical for improved glycemic control and diabetes management. The REDEEM Trial conducted by Dr. Fisher (whose results were presented in another session) tracked stress levels and mood in conjunction with blood glucose levels. The trial found that people who indicated increased levels of stress during the day would experience heightened fasting blood glucose levels. These levels remained abnormally high for most of the day. The trial also found that A1c could not predict distress, and distress could not predict A1c. However, A1c and distress co-varied. That is, an increase in one would result in an increase in the other.

Symposium: Occurrence of Chronic Complications of Childhood Diabetes in the 21st Century


David Dunger, MD, PhD (University of Cambridge, Cambridge, United Kingdom)

Dr. Dunger began the pediatric-focused symposium with a discussion of microalbuminuria, the leading indicator of diabetic nephropathy and a cardiovascular risk factor. Dr. Dunger highlighted the impact of renal dysfunction, noting that 10-19 years after diagnosis, renal causes of mortality are the leading cause of mortality, and 20 years after diagnosis, cardiovascular disease is the major cause. He reviewed the Oxford Regional Prospective Study (n=560), which showed 12.8% of the adolescents with type 1 diabetes developed microalbuminuria, and both persistent and intermittent microalbuminuria were predictors of developing macroalbuminuria. Utilizing this study, Dr. Dunger demonstrated that puberty has a significant effect on the development of microalbuminuria, as there is a low incidence in those diagnosed young (<11) until puberty, when all the cases emerge rapidly, compared to the generally increasing incidence with longer diabetes duration seen in those diagnosed after 11 years old. Finally, Dr. Dunger described the setup of the Adolescent Type 1 Diabetes Cardio-Renal Intervention Trial (AdDIT), a large, multi-national intervention study currently underway examining the effect of ACE inhibitors and statins in adolescents with type 1 diabetes and high albumin excretion. Although pediatricians rarely deal with renal complications, they begin developing during adolescence and therefore, Dr. Dunger believes the results of AdDIT will help elucidate whether statins and ACE inhibitors will mediate the risk of developing persistent microalbuminuria.

  • The Oxford Regional Prospective Study followed children who developed type 1 diabetes before the age of 16 to assess albumin excretion. After a mean follow up of 10.3 years, about 12.8% had developed microalbuminuria, separated into persistent (48%), intermittent (13%), and transient (39%). Both persistent and intermittent microalbuminuria were predictors of developing macroalbuminuria (22% and 24%, respectively), and Dr. Dunger emphasized that practitioners should keep this in mind when addressing any indicators of microalbuminuria in youth.
  • The incidence of microalbuminuria depends on glycemic control, diabetes duration, sex, and puberty. In addition to the rapid emergence of microalbuminuria in pubescent adolescents with type 1 diabetes diagnosed as children, deterioration in glomerular filtration deterioration and raised cholesterol and triglyceride levels escalate at 16 years of age. Dr. Dunger also noted that an undetermined peculiarity raises the cumulative prevalence of microalbuminuria in females compared to males – a switch from later onset type 1 diabetes patients.
  • In AdDIT, a double-blind, placebo-controlled, parallel group study (n=900), adolescents (ages 11-15 years) with type 1 diabetes are separated into tertiles of albumin excretion phenotypes based on two sets of three early morning urine samples. The 500 participants chosen from the highest tertile are further separated in a two by two factorial design: statin and ACE inhibitor, statin and placebo, ACE inhibitor and placebo, or placebo and placebo. Participants have follow ups every six months over three to four years, followed by a six-month run out, and a long-term follow-up. In the parallel group, 400 participants from the lower tertile will be assessed at four years. The study aims to determine whether ACE inhibitors, statins, or a combination will reduce albumin excretion, reduce CVD risk, and be well tolerated in participants with high risk for renal decline.

Questions and Answers

Q: How are you going to be monitoring adherence to study medications?

A: We use the track caps to tell when the participants opened the bottle, including time and date. We are doing it very openly, really to help the kids by showing them the records and working with them to figure out why they are not taking it a certain day, if we see a pattern.

Q: Similarly, are you keeping track of adherence to diabetes management?

A: Yes, we will be keeping track of A1c, and they all practice intensive therapy.

Q: The excess of microalbuminuria in females struck me, and I have also, in more recent years, seen exactly the same analyses. We no longer see the male excess of microalbuminuria seen in later onset type 1 diabetes patients. Do you have any ideas why?

A: It doesn’t explain the difference in microalbuminuria between the sexes, but we know there are differences in free testosterone levels between the girls who develop microalbuminuria and those who don’t. It could be hormonal changes related to puberty.


Paul Wadwa, MD (University of Colorado, Aurora, CO)

Dr. Wadwa reviewed several non-invasive measures of cardiac and vascular changes, beginning with flow mediated dilation (FMD) and peripheral arterial tonometry (PAT), measures of endothelial change. Dr. Wadwa highlighted studies showing decreased FMD in adolescents with type 2 diabetes and reduced endothelial function (measured using PAT) in male adolescents with type 1 diabetes compared to control. He then reviewed carotid intimal-medial thickness testing, which utilizes established ultrasound methodology, and noted that preliminary data from a subset of the SEARCH CVD study shows a significant difference in internal carotid intimal-medial thickness between control adolescents (0.5 mm, n=35) and adolescents with type 1 diabetes (0.56 mm, n=127). Dr. Wadwa then noted increased arterial stiffness is associated with cardiovascular disease, and comparing data from the SEARCH CVD study (mean age of 19) and Barbara Davis cohort (mean age of 15), he demonstrated a significant difference in arterial stiffness between type 1 diabetes and control adolescents at 19 years of age but not 15 years of age, despite similar diabetes duration. Moving to coronary artery calcification (CAC), he emphasized that the CACTI study (n=1,420) showed a 3.5 fold increased risk for presence of CAC in participants with type 1 diabetes (mean age of 37) versus control, though the proportion of type 1 diabetes participants with CAC dropped in the 20-29 year old range (25% males, 12% females) and 17- 20 year old range (11%). This indicates that this measure may be less important in pediatric care. Although he emphasized the limited longitudinal data, he noted that these non-invasive measures have provided evidence of early atherosclerotic changes, indicating a need for screening and treatment of modifiable risk factors in youth with diabetes to decrease lifetime risk for cardiovascular disease. Dr. Wadwa closed with a reminder that further work is needed to standardize the surrogate measures to allow clinical use.

Questions and Answers

Q: You talked about the SEARCH CV study and the Barbara Davis cohort, I’m curious – is there a strong interaction with age to manifest these complications? Also, can you go back into that pre-pubertal young group and see what glycemic control was like?

A: We have looked at pubertal staging, and it would be interesting to look at pre-pubertal duration and the correlation between that and cardiovascular risk.

Q: Do you have any data on the effects of treatment with statins or AEC inhibitors after utilizing these indicators?

A: Although there are some studies, the rate of treatment is so low in the pediatric cohort that it is unlikely the data can be drawn together to say anything significant at this point.

Q: Is there any variability in the glucose levels at the time of the measurement, and would that effect these non-invasive measurements?

A: You’re measuring structural changes over a long period of time, so you’re less likely to get an effect from a transient change in glucose level.


Jennifer Sun, MD (Joslin Diabetes Center, Boston, MA)

Dr. Sun delivered a comprehensive presentation regarding the development, diagnosis, and treatment of diabetic retinopathy in young people. She stated that the increasing rates of diabetes worldwide, especially among children, may lead to correspondingly higher rates of young people at risk for vision loss from diabetic retinopathy (DR) and diabetic macular edema (DME). Routine follow-up according to established guidelines is crucial given the potential lack of visual symptoms. Modifiable risk factors in young people include glycemic control, hypertension, lipid status, and potentially BMI. Although effective treatments are available for DR and DME, Dr. Sun was careful to note that safety for many of those options has not yet been established in the pediatric population.

  • Each year, more than 15,000 young people are diagnosed with diabetes, and diabetic retinopathy is the most commonly observed complication. The SEARCH for Diabetes in Youth study found that in 222 youth with type 1 diabetes and 43 with type 2 diabetes, the risk-adjusted prevalence of diabetic retinopathy was 17% for type 1 diabetes and 42% for type 2 diabetes (p=0.40), statistically insignificant yet clinically significant values.
  • There is an increasing spectrum of diabetic retinopathy (DR) with early non- proliferative stages, but diabetic macular edema (DME) can develop with any level of DR. DME, a common cause of moderate vision loss, is characterized by lipid deposits that form around areas of retinal thickening. Though various risk factors have been assessed, Dr. Sun believes that glycemic control remains the best method of DR prevention. Tight blood pressure and lipid control have also been associated with reductions in DR severity. According to a study conducted by the Joslin Diabetes Center titled, “Retrospective effect of BMI Percentile on Risk of Diabetic Retinopathy in a Pediatric Population,” lower BMI percentile and lower A1c values were associated with less severe DR (n=461, <19 years of age, baseline mean A1c: 8.5%).
  • Various surgical and non-surgical treatment options for diabetic eye disease in patients exist, but have not yet been evaluated to a great extent in children. Dr. Sun reviewed laser photocoagulation, intravitreal anti-VEGF agents, intavitreal steroids and invasive surgery, all of which were associated with improved vision. Dr. Sun was particularly concerned about popular anti-VEGF agents, as VEGF is important in blood vessel development and maintenance in children.

Questions and Answers

Q: You made a list of treatable risk factors, but could you comment on the contribution of smoking?

A: This is an interesting question, because in general the studies that have been done haven’t shown a huge association of smoking with diabetic retinopathy progression. Early studies suggest that smoking protects slightly against diabetic retinopathy, but of course we wouldn’t emphasize that with our patients. So, as health care providers, we encourage people not to smoke because it’s better for systemic health overall, but there is not a large correlation of smoking with diabetic retinopathy.


Andrew Paterson, MD, ChB (Hospital for Sick Children, Toronto, Canada)

Dr. Paterson investigated a genetic model of diabetes suggesting that particular single nucleotide polymorphisms (SNPs) in genes affect established risk factors for diabetes (e.g. A1c, blood pressure, BMI), which in turn, contribute to complications. By investigating 1,304 individuals under the age of 19 in the DCCT/EDIC trial, 57,400 A1c measures over 3-9 years and 841,000 SNPs, Dr. Paterson identified four SNPs (most notably rs1358030) as statistically significantly associated with higher A1c values. Though Dr. Paterson identified a potential loci for glycemia in type 1 diabetes, he admitted that his study had no immediate clinical implications, though it may provide insight into understanding the mechanisms that allow different individuals to respond differently to different medications.

Questions and Answers

Q: I know you said that there are no clinical implications, but how soon might we see some people being screened for needing to control their blood sugar, based on genetic information?

A: I doubt providing information of a single loci would provide any information. Perhaps if there is an interaction between multiple loci, we might have things to screen.

Q: The speculation might be that if SNPs had an effect on metabolism, those might not affect glucose, but instead those areas that interact with glucose control. Any thoughts?

A: It’s completely up in the air where these genetic loci are having their effects on glycemia – maybe they are influencing behavior through effects on the brain. There’s no data to address that, I’m afraid.

Q: Could this SNP be having an effect on the rate of glycosylation of hemoglobin and other proteins that might contribute to complications?

A: If that was the case, would SNPs be associated with glucose levels?

Q: That, I don’t know. The association with glucose levels seems different from the relationship with A1c.

A: We have to remember that there is a lot of missing data, and it’s not missing at random, it’s seven time points, one day out of three months, so it’s no surprise that the association with glucose is much less significant than with A1c.

Symposium: Clinical Aspects of Hypoglycemia in Diabetes – Consequences and Prevention


Simon Heller, MB, DM (University of Sheffield, Sheffield, United Kingdom)

Dr. Heller began his presentation by noting hypoglycemia as the main barrier to preventing patients from reaching desirable glucose targets. He reminded the audience of the consistent finding across intensive therapy trials (VADT, ACCORD, and ADVACE) that hypoglycemia was a strong predictor of mortality. Dr. Heller proposed four mechanisms by which hypoglycemia could be contributing to cardiac arrhythmias and sudden cardiac death: 1) increased thrombotic tendency and decreased thrombolysis; 2) cardiac arrhythmias due to abnormal cardiac repolarization; 3) increase in low grade inflammation; and 4) cardiovascular changes induced by catecholamines. He pointed to new data from the laboratory of Chow et al. and ongoing data collection from ambulatory studies indicating a relationship between hypoglycemia and both vagal tone (as measured by RMSSD) and incidence of cardiac arrhythmias, respectively. In his view, the evidence suggests hypoglycemia can increase both morbidity and mortality, most notably by causing sudden death during hypoglycemic events. Dr. Heller remained hopeful that with better understanding, better treatment will follow suit.

  • Dr. Heller marked hypoglycemia as the main barrier to preventing patients from reaching desirable glucose targets. He explained that hypoglycemia causes problems for patients in a number of ways: 1) seizures; 2) cognitive dysfunction; 3) accidents; 4) fear; and 5) quality of life.
  • In all three trials of intensive therapy – VADT, ACCORD, and ADVANCE – severe hypoglycemia predicted mortality. In VADT, history of severe hypoglycemia within 90 days was the strongest independent predictor of mortality. In ACCORD, severe hypoglycemia was associated with increased mortality to a greater extent than in the standard care group. In ADVANCE, severe hypoglycemia was associated with increased mortality at a mean of 1.05 years. Dr. Heller noted that severe hypoglycemia was strongly associated with increased risk for a range of adverse clinical outcomes, but that severe hypoglycemia could be either contributing to these outcomes or serving as a marker for these outcomes.
  • Dr. Heller highlighted that the important question was whether hypoglycemia can stimulate the cascade from cardiac risk factors to arrhythmia and sudden cardiac death. He proposed four highly plausible mechanisms: 1) increased thrombotic tendency and decreased thrombolysis; 2) cardiac arrhythmias due to abnormal cardiac repolarization; 3) increase in low grade inflammation; and 4) cardiovascular changed induced by catecholamines including increased heart rate, silent myocardial ischemia, and angina and myocardial infarction.
  • According to unpublished data by Chow et al., hypoglycemia confers clot resistance to lysis when compared to patients with euglycemia. In patients with type 2 diabetes, while platelet aggregation is increased compared to baseline in patients 150 minutes after a hypoglycemic event, seven days after a hypoglycemic event there appeared to be no difference between the hypoglycemic group vs. euglycemic group. Increased resistance to clot lysis only persisted in patients following the hypoglycemic event, despite both the hypoglycemic and euglycemic group showing increased resistance to clot lysis at day one.
  • Dr. Heller hypothesized that abnormal cardiac repolarization could explain the considerable evidence (nine studies so far) linking sudden death to hypoglycemia in people with type 1 diabetes. Dr. Heller recapped data from Marques et al. study (Diabetic Medicine 1997) showing that a markedly prolonged QT interval is associated with sudden death, and that the QT interval rose during hypoglycemic events. Dr. Heller explained that hypoglycemia can cause a lengthened QT interval through multiple mechanisms: 1) insulin lowers serum potassium, which prolongs cardiac repolarization by reducing K+ efflux; 2) hypoglycemia reduces hERG channel conduction that mediates K+ efflux (has been shown in vitro and in animalstudies); and 3) increased sympathetic activity reduces serum K+ and promotes Ca2+ influx. However, he noted that the relative contribution of these factors is still under debate.
  • According to ongoing work by Chow et al., vagal tone (measured via RMSSD [vagus- mediated heart rate variability]) is diminished seven days after a hypoglycemic event. Dr. Heller explained that the current data (n=4) shows that for euglycemic patients, there was no activity in RMSSD. However, in the hypoglycemic group, vagal tone diminished 30 minutes after an event compared to baseline. While vagal tone returned to normal 90 minutes from baseline, vagal tone falls again by seven days and seems to remain there.
  • Dr. Heller connected hypoglycemia events to cardiac arrhythmias. Dr. Heller presented evidence from the ongoing data collection from ambulatory studies in which patients were monitored simultaneously with CGM and a 12 lead holter monitor for five days. He pointed to two cases’ electrocardiograms collected so far as prime examples of cardiac arrhythmias during hypoglycemic events. The first showed sinus bradycardia with trigeminy at a glucose of 48.6 mg/dl (2.7 mmol/l) and the second showed non-sustained VT at 57.6 mg/dl (3.2 mmol). He particularly emphasized the increased incidence of bradycardia during hypoglycemia compared to euglycemia – at a confidence interval of 95%, bradycardia had an 8.98 IRR (7.16-11.26) during hypoglycemia.
  • Dr. Heller sees adverse consequences of hypoglycemia manifesting in two ways: 1) when repeated episodes and tight control impair the counter regulatory responses and 2) when occasional episodes affect the sympathoadrenal response. According to Dr. Heller, evidence suggests that hypoglycemia can increase both morbidity and mortality and it may be causing sudden death during hypoglycemic events. He hopes with better understanding, better treatment will follow.

Symposium: Diabetes to Diabetic Kidney Disease – Easing the Nutritional, Medical, and Behavioral Transition


Susan Guzman, PhD (Behavioral Diabetes Institute, San Diego, CA)

Heralded by many audience members as one of the most valuable talks they attended at ADA 2012, Dr. Guzman’s presentation provided valuable lessons for health care professionals in making the transition to chronic kidney disease more manageable. She offered five pillars of advice: (1) realize that the transition to CKD comes with many tough feelings; (2) be aware of your own attitudes; (3) understand that there are good reasons for why people struggle with diabetes, so developing the right strategies is key; (4) realize that obstacles are abundant; (5) be on alert for depression; and (6) be a partner against the disease. Dr. Guzman advises that care providers focus on concrete, achievable, actions that will get the patient the “biggest bang for their buck.” The success, hope, and confidence that result can make a difference and can go a long way when treating patients with diabetes, CKD, or other chronic diseases.

  • Patients often do not understand the consequences of poorly managing their diabetes, and given that there is no tangible reward for managing it well, patients have little incentive to do so. What patients notice are the obstacles they face daily that impede their ability to achieve their desired lifestyle. Other obstacles that further challenge proper diabetes management include lack of knowledge about treatment options, healthcare professional and patient communication problems, depression, medication side effects, unachievable goals, and cultural norms that interfere (especially with regards to diet).
  • While it is well known that people with diabetes are at a greater risk for depression, many HCPs claim to have neither time nor responsibility to assess the mental status of their patients. Regardless, depression is associated with poor self-care, a 2.5-fold increased mortality rate, and a general apathetic attitude that can seriously impede successful therapy. Thus, Dr. Guzman emphasizes that it is imperative that healthcare professionals realize and act on the signs and symptoms of depression in their patients.

Questions and Answers

Q: Can you chat for a minute about some of the workshops you do for people who are dealing with CKD or chronic conditions in general?

A: I do a one-day workshop for people living with complications. They find us locally on our website. We also do a depression series that is a seven-week series, all offered in San Diego , but $10!

Q: I feel like this was one of the most valuable sessions I went to in the entire conference. I just wanted to say thank you.

A: Thank you

Q: I wholeheartedly agree. I’ve had type 1 diabetes for 48 years, and so I have a lot of empathy for my patients. I think one of the most valuable things I do is meet with my patient for an hour each time, often “knee to knee.” Thank you

A: Thank you. Hope is a big thing – if you give them that they can start to see that they can make a difference with their disease.


Rudy Bilous, MD (Newcastle University, Middlesbrough, United Kingdom)

Dr. Bilous presented updates to the National Kidney Society’s Kidney Disease Outcomes Quality Initiatives (KDOQI) recommendations for diabetes and chronic kidney disease (CKD), which will be released later this year. He specifically reviewed guidelines two, four, and six and discussed the quality of evidence grades (A through D) and recommendation levels (one for recommend, two for suggest).

  • Revisions to guideline two focus on the management of hyperglycemia and general diabetes care in chronic kidney disease (CKD). First, they establish a target A1c of 7%, further explaining not to treat to a level under 7% in those at risk of hypoglycemia. Individual needs will determine if a target should be above 7%. Dr. Bilous briefly explained that the guidelines were based on UKPDS, DCCT, ACCORD, and ADVANCE results, as poor glycemic control is correlated with decreased glomerular filtration rates (GFR) and increases in diabetic nephropathy.
  • Guideline 4.1 recommends using LDL lowering medications, like statins, to reduce the risk of major atherosclerotic events in patients with diabetes and CKD, including those with kidney transplants (1B). However, in Guideline 4.2 the NKS recommends not initiating statin therapy in patients with diabetes treated by dialysis (1B). Dr. Bilous explained that evidence for these guidelines comes from the significant improvement in GFR in participants treated with statins in the CARDS and HPS studies, and the 17% reduction in people suffering major atherosclerotic events when treated with statins in the SHARP study.
  • Dr. Bilous concluded with a review of guideline six, which recommends not using ACE inhibitors or ARBs in primary prevention of diabetic kidney disease innormotensive, normoalbuminuric patients with diabetes (1A). However, Guideline 6.2 suggests not using these medications in patients with albuminuria levels ≥30 mg/l who are at risk of diabetic kidney disease (2C). A 2011 meta-analysis showed ACE inhibitors in combination with ARBs are actually less beneficial than using ACE inhibitors or ARB alone, but the NKS cannot conclusively recommend against using combination therapy

Corporate Symposium: Protecting Vision in Your Diabetic Patients (Sponsored by Genentech)


Sunir Garg, MD (Thomas Jefferson University, Philadelphia, PA)

Dr. Garg began his talk with the surprising statistic that one-third of people with diabetes have never had an eye exam. He provided a comprehensive overview of the classifications of diabetic retinopathy, and explained how diabetic macular edema (DME) can occur at any point along the spectrum. DME is the leading cause of vision loss in patients with diabetes and is diagnosed with a dilated fundus exam. Next, Dr. Garg discussed the most common treatment: laser therapy. The ETDRS study not only defined DME but also set laser therapy as the standard of treatment, as it was shown to reduce the risk of DME by 50%. Triamcinolone (Kenalog), an innovative steroid introduced around 2002, was discussed next. Although it initially had impressive results, a Diabetic Retinopathy Clinical Research Network (DRCR) trial illustrated that after three years, the beneficial effects of triamcinolone regressed. Yet, laser therapy still proved beneficial. Anti-VEGF agents were the last class of therapeutics he discussed, which help reduce vascular permeability. Interestingly, all four Anti-VEGF agents – pegaptanib, bevacizumab, ranibizumab, and afilbercept – are all used off-label. Bevacizumab showed superior vision gains compared to laser therapy in the bevacizumab or Laser Therapy (BOLT) trial. Similar superior vision gains were observed with ranibizumab in RISE/RIDE, a multicenter, randomized trial (n=759). Overall, Dr. Garg believes anti-VEGF agents are a very promising, low-cost alternative to laser treatment.


Carol Wysham, MD (University of Washington School of Medicine, Seattle, WA) and Doron Schneider, MD, FACP (Drexel University, Abington, PA)

Dr. Wysham, Dr. Garg, and Dr. Schneider focused on the control of diabetic retinopathy in patients with type 1 and type 2 diabetes. Dr. Wysham emphasized the importance of ophthalmology screening for diabetes patients, citing both ADA recommendations and practical reasons for doing so. The ADA recommends screening type 1 diabetes patients within five years of diagnosis after age 10 and type 2 diabetes patients at the time of diagnosis. Such screening is important because diabetic retinopathy is highly prevalent, associated with increases in morbidity and decreases in quality if life, and can be controlled with systematic therapy targeting A1c or blood pressure. Dr. Schneider talked about strategies for increasing ophthalmology referrals for diabetes patients. He promoted a systematic approach embedded within a Chronic Care Model that encourages patients, front desk personnel, and medical assistants (MAs) to participate as fully in the healthcare process as their knowledge and license would allow. For example, the patient could be asked to review a customized checklist on a website for prevention, the front staff could remind patients to pull out or complete a visit checklist, and the MA could review the date of the last ophthalmology appointment.

  • Dr. Wysham emphasized that ophthalmology screening for diabetes patients is important because diabetic retinopathy is highly prevalent, associated with significant morbidity increases, and can be controlled with systematic therapy. She noted that diabetic retinopathy is the most common cause of blindness among adults 20-74 years of age. During the first two decades of disease, nearly all patients with type 1 diabetes and over 60% of patients with type 2 diabetes have retinopathy. Dr. Wysham then talked about the impairment of quality of life in patients with diabetic retinopathy, including social isolation, stressful family dynamics, unemployment, psychological distress, and difficulty in the management of diabetes (due to difficulty reading labels, exercising, reading the glucose meter, and administering insulin). She then emphasized that diabetic retinopathy can be minimized through glycemic control (UKPDS trials), fenofibrates (FIELD and ACCORD trials), and statins. Laser photocoagulation therapy is also an option to minimize visual loss in those with more advanced retinopathy. The American Diabetes Association (ADA) recommends screening type 1 diabetes patients within five years of diagnosis after age 10 and type 2 diabetes patients at the time of diagnosis. Despite this, Dr. Wysham noted that only 46% of type 1 diabetes patients in Germany had their retinal status documented (Hammes 2011), and that a significant proportion of diabetic patients in the US are not getting their ophthalmology screening.
  • Dr. Schneider talked about developing strategies to increase ophthalmological referrals given to diabetes patients. Based on National Committee for Quality Assurance (NCQA) data, he noted that the best practices in the United States (at the 90th percentile) have an ophthalmological referral rate of 79.3% for diabetes patients, compared to practices at the 10th percentile with a rate of 49.7%. Dr. Schneider recommended that private practitioners evaluate their own performance in this aspect by looking at a sample of diabetes patients’ files. He promoted the use of the Chronic Care Model as a tool to facilitate brainstorming around the reasons and the methods to close the gap. In this model, the patient would be asked to review a customized checklist on a website for prevention, the front staff would remind the patient to pull out or complete a visit checklist, the medical assistant (MA) would confirm adherence to medications and review the date of the last ophthalmology appointment, and the doctor would handle aspects of the patient’s chronic problems. Information systems can also be used to facilitate referral and can help ophthalmology practices identify patients who may require ophthalmology visits (e.g., those with no record of eye exams in the past year). He emphasized that doctors need not handle every aspect of disease management, and that practices should promote team-based care where each staff works up to the limits of their license and ability. That way, doctors can focus on more complex decision-making. According to Dr. Schneider, if it came to the doctor to remember whether the patients require an ophthalmology referral, “shame on that practice.”


Carol Wysham, MD (University of Washington School of Medicine, Seattle, WA); Sunir Garg, MD, FACS (Thomas Jefferson University, Philadelphia, PA); Doron Schneider, MD, FACP (Drexel University, Abington, PA)

Q: Why do you think that a rise was observed in the sham arm of the RISE trial?

Dr. Garg: the sham group still got laser therapy, which has been the standard of care for diabetic macular edema. Since the sham group didn’t have injections but had the option of laser treatment, many opted for the laser treatment. The patients looked at in this trial were also selected for Macular Edema, and these patients respond better to laser treatment.

Q: Our practice performed a program where we tried to have independent eye care providers give information back to us, and said that as a private practice, they could provide nothing in return to us. What strategies do you suggest with a practice like this?

Dr. Schneider: Really, competition is key. If there is a practice in your community that is more amenable to working together, that practice should be supported. The national trend is towards Accountable Care Organizations and healthcare teams, so practices that “don’t play ball” would be under pressure to change over the long term. Right now, practices that do not fill out forms must still document their treatments, and you can try to get that information through their records.

Dr. Wysham: My first reaction is to look for other eye doctors to refer patients to. The eye doctors I refer patients to are all fairly open to responding, and this is becoming a standard.

Dr. Garg: As ophthalmologists, we don’t necessarily appreciate how important the forms are to primary care doctors. So, education would be very helpful. Instead of looking for a new eye doctor right away, if you like their care, you can try to educate them on the importance of the information and tell them that you would have to look for other options if they cannot facilitate the transfer of that information. Hopefully this will incentivize the ophthalmologists to do this.

Q: What patients would you not use anti-VEGF drugs with?

Dr. Garg: There are no other systemic medications we are aware of that will cross-react with anti-VEGF drugs, with the exception of intravenous immunoglobulins (IVIG). The concern is with patients that have had a recent stroke or heart attack. We really don’t know what the risk is – by and large the risk seems small if at all. We will often reduce the dose for patients with recent stroke or heart attack. We don’t know if anti-VEGF agents increase the risk of stroke or heart attack, but there is little data to support that anti- VEGF increase the risk.

Q: Are there any drugs available for patients with prediabetes if they potentially have retinopathy?

Dr. Wysham: There is no specific treatment that would be indicated to lower the risk for retinopathy. It is best to focus on lifestyle changes and establishing control of their CV risk factors. When you have retinopathy, it is a global vascular risk so focusing on CV risk control is very important.

Symposium: Changing the Course of Diabetic Retinopathy


Thomas Gardner, MD (University of Michigan, Ann Arbor, MI)

Dr. Gardner transitioned the discussion of diabetic retinopathy from the global perspective presented by Dr. Barbara Klein to a more local approach. Dr. Gardner presented the eye, body, and soul parts to treating diabetic retinopathy. He began his talk specifically focusing on the eyes. Dr. Gardner reviewed current treatments (notably the success of ranibizumab in improving visual acuity in patients with diabetic macular edema [, Ophthalmology 2012]) for treating diabetic macular edema. Importantly, he noted that different patients respond differently to treatments, with some patients having limited or no success. Consequently, he concluded that while we’ve made progress in treating diabetic retinopathy and macular edema, there is a lot more to be gained. To that effect, he noted that many new therapies are currently being investigated, including fenofibrate, doxycycline, minocycline, and tissue plasminogen activators. Next, he presented the challenge of treating diabetic retinopathy from a whole-body standpoint. He reminded us that diabetic retinopathy is not an isolated complication. A1c, diabetic nephropathy, and ulcers are all associated with an increased risk of diabetic retinopathy. Most notably, a 1% increase in A1c is associated with a 14% increased risk for developing proliferative retinopathy. Dr. Gardner stressed that diabetic retinopathy is a complex picture and multiple factors need to be addressed to ameliorate it. Then, Dr. Gardner moved on to the soul. If the eye is the window of the soul, he said, we should be able to get insight into what motivates patients. He emphasized that in practice it is important to 1) help patients get expert help; 2) encourage patients that they can control their diabetes; 3) remind patients that small differences can make big differences; 4) talk to patients in terms they can understand; and 5) reassure patients that they will not go blind if they take care of themselves. Dr. Gardner concluded with an outline for future strategies in patient treatment to prevent diabetic retinopathy and macular edema: 1) quantify interactions of multiple risk factors; 2) identify a distinct risk profile for each patient; and 3) treat the profile to prevent the complications. Dr. Gardner explained that distinct profiles mean more specific treatments and a better chance of treating complications.


Barbara Klein, MD (University of Wisconsin, Madison, WI)

Dr. Klein opened the symposium with a comprehensive overview of diabetic retinopathy on the global level. First, she reminded the audience of the burden and dynamic nature of diabetic retinopathy. A 2010 WHO estimate rated diabetic retinopathy as the fifth leading cause of blindness. (Dr. Klein did note that a number of factors introduce error and variability into WHO estimates including heterogeneity of survey methods, extrapolation from specific areas to provide national estimates, and differing definitions of eye disease.) Yet, care seems to be improving. She pointed to data from the Wisconsin epidemiologic study of diabetic retinopathy that showed the percentage of eyes with proliferative diabetic retinopathy and clinically significant macular edema in people with type 1 diabetes receiving treatment is decreasing. However, she commented that the growing burden of diabetes (a 2010 study in China of 46,239 adults estimated 4.5 times more people have diabetes than in 2002) and the increasing duration of diabetes suggests we may see increased prevalence of diabetic retinopathy. Dr. Klein went on to discuss the resources necessary to implement effective screening and treatment for diabetic retinopathy. She highlighted five requirements: 1) regional centers or mobile exam units; 2) computer and Internet access; 3) resources for follow up; 4) availability of personnel and technical equipment for treatment; and 4) communication systems for practitioners and patients to enable education. Furthermore, Dr. Klein suggested that digital photography combined with telemedicine links could be used to screen remote populations in a cost-effective way. She emphasized the need for testing strategies to balance sensitivity and specificity in order to get the most bang for our buck.

Questions and Answers

Q: There are higher rates of prevalence in Southeast Asian countries. Why is that?

A: It is really hard to know. There are some of the problems in data that plague WHO studies. It is probably true that things we use in screening for diabetes like BMI need to be revised. Standards in more Western countries probably don’t capture what is important in the way of obesity. This is really true in India. Public health efforts are often geared off of things that are simple. Some of it is that. I think the increase in economic stability in some of these areas and greater availability of food adds to the burden of obesity.

Q: Is there any geographic data that you can normalize according to A1c and find that at an equivalent A1c, rates are much lower in some regions than others? If you have a 9% in France and a 9% in Wisconsin, is the prevalence of any parameter lower?

A: It is uniform in Caucasian populations. We can’t define a group in particular. It looks to be pretty linear with glycated hemoglobin, but it is dangerous to extrapolate world wide from experience in Caucasian populations.


Renu Kowluru, PhD (Kresge Eye Institute at Wayne State University, Detroit, MI)

Dr. Kowluru has conducted a number of animal studies investigating the effect of antioxidant micronutrients on the progression of diabetic retinopathy in her prolific career. She asserts that there is adequate evidence to initiate well-designed human clinical trials to assess the value of micronutrients as an inexpensive adjunct therapy to inhibit the development of diabetic retinopathy.

  • Dr. Kowluru first described the physiologic progression of diabetic retinopathy (DR) in order to illustrate how micronutrients have the potential to inhibit microvascular disease. Diabetes, which causes sustained hyperglycemia, can cause biochemical changes and oxidative stress. These lead to growth factor alterations and vascular occlusion, culminating in neovascularization and proliferative retinopathy. The step between hyperglycemia and oxidative stress has been targeted for inhibition.
  • A series of animal studies found that multi-antioxidants, lipoic acids, and AREDS- based micronutrients (age-related eye disease study) each inhibit retinal oxidative stress. Multi-antioxidant therapy included a variety of widely known antioxidants. Normal levels of CSH, HdG and acellular capillaries (each established measures of retinal oxidative stress) were restored in diabetic mice on multi-antioxidants, lipoic acid, and AREDS-based nutrients respectively. Each regimen was shown to reduce hyperglycemia severity as well. However, Dr. Kowluru was careful to note that micronutrient regimens began within three to four days of diabetic induction in mice.
  • However, studies of the effect of micronutrients on diabetic retinopathy in humans leave much to be desired, and Dr. Kowluru believes it is time for a new beginning. Dr. Kowluru reviewed several commercially available supplements (e. g. bilberry, Nepretin, and superoxide dismutase) that have fared well in non-clinical trials in slowing DR, but acknowledged that some clinical studies have indicated no link between many antioxidants and DR progression. However, Dr. Kowluru noted that unlike in her studies, these studies initiated antioxidants subsequent to the development of background retinopathy. Ultimately, Dr. Kowluru finds that animal studies provide convincing data to initiate human clinical trials with micronutrients.

Questions and Answers

Q: Patients come to us all of the time asking us if nutritional supplements work for their diabetes and diabetic eye disease. There are many corporations that make many claims, but is there any convincing evidence that any of these have a beneficial effect on diabetic retinopathy?

A: Antioxidants will have some beneficial effect depending on the state of diabetic retinopathy these patients are in. I would say that there’s no harm in taking them, but until we have a good clinical trial, we don’t know the true benefit.

Q: The data you showed was for animals that were started on antioxidants at the onset of diabetes. Have you done studies where you’ve delayed antioxidant supplements to later? And is there a stage of DR at which antioxidant supplementation is no longer beneficial and even harmful?

A: We’ve done some work where we’ve started animals on antioxidants at a much later stage and we’ve seen some beneficial effect, but we’ve only done one study where we started them at six months of diabetes, which for animals is a long time. We haven’t started in the middle of the game, we’ve started at six months – and we should see some beneficial effect but not really great.

Q: In your feeding studies, when do you provide antibiotics?

A: They are mixed into the diet.

Q: That might explain some of the distinction between animal and human studies; we are getting the nutrients when we need them if we mix them into the diet. When we eat, the mitochondria produce superoxides, whereas when most humans take antioxidants they take them in the evening or when they wake up when they don’t need them. My guess is that you wouldn’t see much of an improvement as with people. So a better approach would be to take antioxidants when they wake up.

A: Yes, sure.

Oral Sessions: Predictors and Consequences of Diabetic Kidney Disease


Paul Beisswenger, MD (Dartmouth University, Hanover, NH)

Dr. Beisswenger presented what he considered a groundbreaking study looking at novel biomarkers for prediction and identification of diabetic nephropathy. Five advanced glycation products were studied: carboxymethyl lysine (CML), carboxyethyl lysine (CEL), Glyoxal hydroimidazolone (GHI), methylglyoxal hydroimidazolone (MGHI), and 3-deoxyglucosone hydroimidazolone (3DG-H). While Dr. Beisswenger believes that the current testing method (microalbuminuria) is good, he thinks it could be better as there is a poor correlation between the level of microalbumin and the degree of biopsy-proven diabetic nephropathy. The novel biomarkers presented have strong scientific basis to justify their use, as they are long-term indicators of chemical pathways that lead to complications. The current study examined the early development of diabetic nephropathy in 103 patients with type 1 diabetes, as determined by electron-microscopic quantification of two research renal biopsies five years apart. Three biomarker levels were higher in nephropathy progressors than non-progressors – CML, CEL, and MGHI. All three biomarkers had higher odds ratios for nephropathy progression relative to A1c, and enhanced A1c’s predictive value. Overall, this study has made impressive progress on improving the prediction and treatment of diabetic nephropathy.

  • The study was designed to examine the early development of diabetic nephropathy in 103 subjects with type 1 diabetes, as determined by electron-microscopic quantification of two research renal biopsies fives years apart. The present study used data from the Natural History of Diabetic Nephropathy Study (NHS), which was a younger population (mean age of 16.9 years), had a short diabetes duration (mean of 7.9 years), and a mean baseline A1c of 8.5%. The primary indicator used to assess nephropathy progression was glomerular basement (GBM) width.
  • Three biomarker levels were higher in nephropathy progressors than non- progressors – CML, CEL, and MGHI. All three biomarkers had higher odds ratios for nephropathy progression relative to A1c, and enhanced A1c’s predictive value. The respective odds ratios (with 95% CI) for A1c, CML, CEL, and MGHI are as follows: 1.29 (0.82-2.02), 1.95 (1.14-3.35), 1.72 (1.06-2.77), and 1.63 (1.004-2.31). More importantly, when CML, CEL, and MGHI were combined with A1c, the prediction of variation in GBM width increased nearly 2.5 times from 4.7% to 11.6%.
  • Overall, this study has made impressive progress on improving the prediction and treatment of diabetic nephropathy. Follow-up outcome studies should be run to validate the role of these biomarkers. Ultimately, the clinical use of these new biomarkers would allow early, individualized treatment to prevent complications.

Q: When have you observed renal damage following renal albuminuria?

A: You can already have significant renal damage without renal albuminuria. The other thing we are interested in is whether these new biomarkers are aiding in protection or susceptibility. There were only two progressors in the lower tertile – we need to understand whether they are less susceptible or protected.

Q: Can you clarify the timing of when you measured the biomarkers studied?

A: The biomarkers were measured at year five. My initial plan was to look at serum samples throughout the study. Ultimately we had to use plasma samples, as the serum samples were not stable.


Hillary Keenan, PhD (Joslin Diabetes Center, Boston, MA)

Dr. Keenan described the results of the 50-year Medalist Study, which identified protective factors for diabetic retinopathy and nephropathy. The study focused on patients with type 1 diabetes who had been insulin dependent for 50+ years. Patients were classified into five groups based on the severity of chronic kidney disease (CKD) and assessed for estimated glomerular filtration rate (eGFR) and albumin/creatinine ratio (ACR). More severe chronic kidney disease was associated with increased cardiovascular disease, higher C-reactive peptide levels, and lower eGFR. Interestingly, ACR was not correlated with renal glomerular pathology in the cohort. Based on these findings, Dr. Keenan believes that further validation of the presence of diabetic nephropathy by histological study may be necessary for therapeutic plans in patients without proliferative diabetic retinopathy and low eGFR.

  • Patients were classified into five groups based on the severity of chronic kidney disease (CKD) and assessed for estimated glomerular filtration rate (eGFR) and the albumin/creatinine ratio. The following are clinical characteristics of the 50-year medalist cohort –average age: 70 years; mean diabetes duration: 59 years; mean BMI: 25 kg/m2; mean A1c: 7.2%; neuropathy: 34.5%; proliferative diabetic retinopathy: 46.8%.
  • More severe chronic kidney disease was found to be associated with increased cardiovascular disease, higher C-reactive peptide levels, and lower eGFR. The two patient groups with the most severe CKD were classified as severe, while the other three patients groups were classified as mild, collectively. Between the severe and mild CKD groups,cardiovascular disease prevalence was 55.5% vs. 35.4% (p <0.01), C-reactive peptide levels were 1.3 vs. 0.9 mg/l (p <0.01), and eGFR levels were 48.5 vs. 77.5 ml/min/1.73m3 (p <0.01).

Q: Did you have any opportunity to measure pre and post removal of anti-hypertensive drugs?

A: No we don’t have pre and post data on when they were started.

Q: Clearly there are people with long duration diabetes that don’t develop kidney disease. Do these data change your perspective?

A: When I was first introduced to diabetic nephropathy, I was taught that as the duration progresses in diabetes, your glomeruli increases. However, some of these data defies this dogma.


Amanda Adler, MD, PhD (Addenbrooke’s Hospital, Cambridge, United Kingdom)

Diabetes significantly increases the risk of death in dialysis. Given the increased prevalence of cardiovascular disease and hypoglycemic unawareness in patients with diabetes, aggressive glucose- lowering could be associated with risks. With this in mind, Dr. Adler outlined an observational study using the UK Renal Registry to determine whether an association exists between glycemic control (measured by A1c) and mortality in individuals with diabetes on dialysis. The study included 3,157 individuals with information on covariates including age, sex, cause of renal failure, and a measure of social deprivation. After accounting for confounders, the study found that for patients under 60 years, as A1c increases there is a significant increase in the incidence of mortality. We found these results very interesting because if this risk is assumed reversible, the study suggests that improving glycemic control to an A1c between 6.5-7.4% prior to dialysis could improve survival by approximately one year.

  • Dr. Adler described an observational study using the UK Renal Registry to determine whether there is an association between glycemic control (measured by A1c) and mortality in individuals with diabetes on dialysis. Currently, the National Kidney Foundation advocates an A1c < 7.0% irrespective of the presence or absence of chronic kidney disease (CKD), while the UK Renal Association advises an A1c between 6.5-7.5% in dialysis patients. The current study sought to determine whether there was an association independent of factors related to both glycemia and mortality and to quantify the potential survival gain in months with better glycemic control.
  • The study included 3,157 individuals with information on covariates including age, sex, cause of renal failure, and a measure of social deprivation. The cohort had the following characteristic: median A1c of 7.0%, median age of 60 years, 38.6% women, deprivation of 0.2 based on the townsend score of material deprivation, and 85.2% of renal failure caused by diabetes.
  • After accounting for confounders, the study found that for patients under 60 years, as A1c increases there is a significant increase in the incidence of mortality (p<0.0001). These results were considered generalizable given ‘free’ healthcare in the UK.
  • The study suggests that improving glycemic control to an A1c between 6.5-7.4% prior to dialysis could improve survival by approximately one year.

Q: Is it possible that the A1c is a risk factor for earlier death?

A: Yes, but we did try to control for residual confounding.

Q: Your data relates to A1c starting at dialysis, do you have A1c over time? Also, what were the causes of death?

A: We do have cause of death, it is a good suggestion to look at A1c as a time variant

Symposium: How to Measure Success in the Treatment of Diabetic Neuropathy


Aaron Vinik, MD, PhD, FCP, MACP (Eastern Virginia Medical School, Norfolk, VA)

In this talk, the highly-regarded complications expert Dr. Vinik discussed pain alleviation associated with diabetic neuropathy, focusing on distal symmetric polyneuropathy, which affects both small and large nerve fibers. According to Dr. Vinik, treating small nerve fiber neuropathy is far more challenging. He provided a review of the efficacy of various treatments being developed. Ultimately, Dr. Vinik emphasized that although there are many promising agents in the wings that can address the core pathogenic mechanisms of neuropathy, metabolic (glucose and lipid) control and treatment of macrovascular risk factors are key to prevention.

  • The most common form of diabetic neuropathy is distal symmetric polyneuropathy, which involves small and large fibers, with damage to each fiber producing its own constellation of features. Large fiber neuropathy is associated with loss of balance, which places individuals at a greater risk of falling and causing injury. Small fiber neuropathy is associated with loss of intra-epidermal nerve fibers, which induces C-fiber type pain (e.g., burning sensations and allodynia), but strength and reflexes are normal. Though the symptoms of large fiber neuropathy have been shown to improve significantly with exercise training to improve balance and reduce reaction time to avoid falls, there is little consensus about how to effectively treat small fiber neuropathy.
  • A meta-analysis of a number of clinical trials identified SNRIs, gabapentinoids, and the opioids as the most effective treatments for alleviating pain associated with small nerve fiber neuropathy, which is difficult to address. Difficulty sleeping, lack of energy, and drowsiness were found to be the most common small-fiber neuropathy-related symptoms. Dr. Vinik cited data from three double-blind trials with 652 patients with diabetic peripheral neuropathy receiving pregabalin for 5-8 weeks. Pain improved significantly, with greater improvements observed as baseline pain severity increased.
  • Topiramate and Metanx, along with glycemic control, were found to be effective in reversing symptoms as well as alleviating pain associated with diabetic neuropathy. Dr. Vinik cited a study he helped conduct titled the Norfolk Quality of Life – Diabetic Neuropathy, finding that individuals with diabetic neuropathy and anxiety or depression reported significantly higher pain scores and had less pain acceptance than those with only diabetic neuropathy. The study also found that topiramate, which stimulates nerve regeneration, significantly reduced weight (228 to 220 lbs), BMI, blood pressure, A1c levels (7.4 to 6.8%) as well as neuropathy scores (31.1 to 21.0). Metanx was also found to significantly reduce pain associated with diabetic peripheral neuropathy. Dr. Vinik then reviewed the famous DCCT/EDIC trial, which found that glycemic control prevents neuropathy both in the short and the long-term, even when glycemic control is not maintained in the long run.



Sebhat Erqou, MD, PhD (University of Pittsburgh, Pittsburgh, PA)

The association between hyperglycemia and the risk of congestive heart failure (CHF) has been suggested but has yet to be quantified through clinical trials. Dr. Erqou performed an analysis of prospective epidemiological data on the association between glycemia measured by A1c and the risk of CHF in patients with diabetes. Through an analysis of ten studies, the overall adjusted risk ratio (RR) for CHF was 1.15 (1.10-1.21, 95% CI) – for every one percentage point increase in A1c, the risk of CHF increased by 15%. Interesting, the RR for the single study comprising patients with type 1 diabetes was 1.30  (1.21-1.40, 95% CI). Despite the absence of conclusive evidence from clinical trials, this prospectivestudy provides epidemiological support for the potential benefit of glucose lowering on the risk of CHF.

  • Dr. Erqou performed an analysis of prospective epidemiological data on the association between glycemia measured in A1c and the risk of CHF in patients with diabetes. Electronic databases and relevant studies were analyzed that included baseline and endpoint A1c measurements, as well as clinically significant CHF. Relative risk estimates were calculated for each study, and a cumulative risk was derived from these. 178,929 participants with diabetes were studies, with 14,176 incident CHF cases.
  • Through an analysis of ten studies, the overall adjusted risk ratio (RR) for CHF was1.15 (1.10-1.21, 95% CI) – for every one percentage point increase in A1c, the risk of CHF increased by 15%. Within the ten studies, there was substantial heterogeneity across the studies. After removing one study (which was later confirmed by the author to have been flawed), the pooled adjusted RR for CHF was again 1.15 (1.11-1.19, 95% CI).

Q: What about the Nichols study?

A: Maybe by chance they got it like that. This is the source of your heterogeneity by and large. However, without this study the association is still the same.

Q: Does type 1 make a difference?

A: Type 1s were followed for a longer period of time.



Lisa Olsson, MS (Karolinska Institute, Stockholm, Sweden)

Ms. Olsson’s study concluded that the risk of mortality in autoimmune diabetes and latent autoimmune diabetes (LADA) patients may be higher than that of type 2 diabetes patients. Her study used data from the Nord-Trondelag Health Study (HUNT, 1995-1997), which had a sample size of 46,303 patients and included 2,699 cases of diabetes, 208 cases of autoimmune diabetes, and 150 cases of LADA. Patients were classified as LADA if they were anti-GAD positive. Patients in the study were tracked using the Norwegian National Cause of Death Registry (1995-2009), and during the 14 years of follow up, mortality was 33% in individuals with diabetes and 15% in individuals without diabetes. Those with autoimmune diabetes had hazard ratios of 1.56 for all cause mortality, 1.89 for cardiovascular mortality, and 2.44 for mortality due to ischemic heart disease. All three of those numbers are higher than the hazard ratios for type 2 diabetes patients (1.48, 1.8, and 2.61, respectively). Constricting the analysis to LADA patients led to similar results (see chart below). The differences between the groups are statistically significant and cannot be eliminated by controlling for A1c levels.


Hazard Ratios (HR)

95% CI

All Cause Mortality


No diabetes






Type 2 diabetes



Cardiovascular mortality


No diabetes






Type 2 diabetes



Mortality of ischemic heart disease


No diabetes






Type 2 diabetes



Questions and Answers

Q: Do you have information on insulin used? Some LADA patients may have switched to insulin after getting correct diagnosis

A: We do have that information but we didn’t do this. Thank you.

Q: How often do you follow-up: once at the end of the 14 years or once every few years?

A: We follow them through the use of the death registry, both continuously and at the end of follow-up.


Nilka Burrows, MPh (Centers for Disease Control and Prevention, Atlanta, GA)

Ms. Burrows examined trends in hospitalization rates for acute myocardial infarction (AMI) and stroke among people with diabetes in the US. She noticed a decline in cardiovascular hospitalization rates within both diabetic and non-diabetic populations, but found faster decline among diabetics and a narrowing gap between diabetics and non-diabetics. From 1998 to 2009, the age-adjusted AMI hospitalization rate declined from 24.7 to 7.7 per 1,000 diabetes patients, and the stroke hospitalization rate declined from 18.9 to 9.2. The decline occurred among all age, sex, and race groups, and disparities by age, sex, and race either narrowed or disappeared. Ms. Burrows chalks up the decline to a number of factors, including but not limited to a reduction in risk factor prevalence and more aggressive treatments of cardiovascular incident risk factors. She considers these findings “encouraging.” The study used data from the National Hospital Discharge Survey (NHDS), which is limited in the sense that the diabetes population was identified by self-report and does not include patients in nursing homes.

Questions and Answers

Q: Is the change in hospitalization rates for AMI significant? Could it be caused by the fact that statins became popular in the early 2000s?

A: We can only speculate as to what is going on, but I suspect it’s a combination of many things. The trend is significant. My boss also thinks that statins are part of the reason for the decline.

Q: In the Kaiser Northwest Diabetes Registry, we injected a lot of data into our registry and suddenly the risk factors all went down. Could this type of sampling error explain happened here also?

A: Continued surveillance could help determine if this trend is real or not.

Q: Did you study the hospitalization rates for other reasons than AMI and stroke in diabetic patients?

A: We limited our study to AMI and stroke. We have to take the sample sizes of the groups into account, especially when doing trend analysis. That’s part of the reason we focused on these two large groups.


Baqiyyah Conway, PhD (Vanderbilt University, Nashville, TN)

Dr. Conway’s study found that all cause mortality and coronary artery disease mortality are lower in black patients with type 2 diabetes than in white patients with type 2 diabetes, after controlling for socioeconomics status. The study was conducted for low-income blacks and whites with and without type 2 diabetes using a sample size of n=12,498 for a mean follow-up period of 5.9 years. Overall, compared to whites with type 2 diabetes, blacks with type 2 diabetes had lower all-cause mortality (HR=0.95, 95% CI = 0.72-0.99) and coronary artery disease mortality (HR=0.64, 95% CI=0.47-0.86), but higher (though statistically insignificant) mortality from heart failure (HR=1.67, CI=1.01-2.76) and respiratory failure (HR=1.09, CI = 0.68-1.74).

Questions and Answers

Q: We saw the same thing in the Britain in the UKPDSL: blacks had lower all cause mortality than whites. Did you look at how BMI relates to all cause mortality?

A: We don’t have that data among whites, but among blacks the relationship is U shaped.

Q: Where is the lowest point of the U?

A: I’m not sure.


Oral Sessions: Diabetes and Macrovascular Disease – From Biomarkers to Therapeutics


Jenica Wong, MBChB, FRACP, PhD (University of Sydney, Sydney, Australia)

Dr. Wong presented data comparing a cohort of people with early onset type 2 diabetes (n=354) to a cohort of 1:1 matched controls with type 1 diabetes (n=354). Early onset was defined as a diagnosis between 15 and 30 years of age. The study examined the prevalence of diabetes complications and case fatality using records from the National Death Index. Ten years after diagnosis, more people with type 2 diabetes presented with microvascular and macrovascular complications than people with type 1 diabetes including 1) albuminuria (49.5% vs. 13.4%; p <0.0001); 2) vibration perception threshold (Z score: 2.5 vs. 1.9; p <0.0001); 3) stroke (6.3% vs. 1.0%; p=0.004); and 4) ischemic heart disease (21.5% vs. 2.4%). People with youth onset type 2 diabetes also showed a less favorable risk factor profile. Notably, the cohort with type 2 diabetes had higher triglyceride levels 1.8 (32.4) vs. 1.0 (18) mmol/L (mg/dl); p <0.0001). While both groups showed similar blood pressure rates and cholesterol levels, the group with type 2 diabetes achieved this with higher antihypertensive and statin treatment. The type 2 diabetes cohort had more deaths (n=39 vs. n=23; p=0.03) and more deaths due to cardiovascular disease (50% vs. 38%; p=0.047). Dr. Wong concluded that this data seen in the context of increasing evidence suggests earlier onset of type 2 diabetes is a “clarion call to further action”.

  • Youth onset of type 2 diabetes is growing. In the youth aged 10-19 with diabetes, the relative proportion of type 2 diabetes growing. In American Indian (86% vs. 14%), African American (58% vs. 42%), and Asian/Pacific Islander (70% vs. 30%) populations type 2 diabetessurpasses type 1.
  • The study examined the prevalence of diabetes complications and case fatality in youth onset type 2 diabetes compared to a similar cohort with type 1 diabetes. Youth onset was defined as a diagnosis between 15 and 30 years of age. The study examined records for 354 people with youth onset type 2 diabetes and 354 matched controls with type 1 diabetes. Controls were matched for age of diagnosis. Mortality data was obtained from the National Death Index. Notable differences in baseline characteristics (given type 2 diabetes cohort vs. type 1 diabetes cohort) include: 1) BMI (32.2 kg/m2 vs. 25.9 kg/m2); 2) diabetes treatment profile within five years of diagnosis; 3) percent of cohort of Anglo-Celtic descent (28.1% vs. 76.3%, p<0.0001);5) number of males (50.6% vs. 61.3%, p=0.004); and 6) age of onset of diabetes (25.6 vs. 24 years, p=0.0001). Dr. Wong noted that she did not believe the difference in age of onset was clinically significant. Baseline A1c was identical (8.1%).
  • More microvascular and macrovascular complications occurred in individuals with type 2 diabetes vs. type 1 diabetes 10 years after diagnosis. Dr. Wong drew attention to differences in albuminuria, vibration perception threshold, stroke, and ischemic heart disease (IHD).

Type 2 diabetes

Type 1 diabetes



Duration of diabetes (yrs)











Retinopathy (%)




Albuminuria (%)




Albumin- creatine ratio







Vibration perception threshold Z score




Stroke 6.3 1 0.004
IHD 21.5 2.4 <0.0001
  • Youth onset type 2 diabetes also showed a less favorable risk factor profile 1o years after diagnosis. Dr. Wong explained that even though blood pressure was similar, individuals with type 2 diabetes achieved the blood pressure level with higher antihypertensive treatment. Similarly, individuals with type 2 diabetes achieved similar LDL profiles to people with type 2 diabetes with higher statin use. She also noted that snapshots of the cohorts two to five years after diagnosis were remarkable similar to the 10 year profile.

Type 2 diabetes

Type 1 diabetes



Antihypertensive treatment (%)






p <0.0001

Systolic BP (mm Hg)






p = 0.06

Diastolic BP (mm Hg)






p = 0.08

Statins (%)



p <0.0001

Cholesterol (mmol/L [mg/dl])



5.0 (90)



5.0 (90)



p = 0.7

Triglycerides (mmol/L [mg/dl])



1.8 (32.4)



1.0 (18)



p <0.0001

HDL (mmol/L [mg/dl])


1.2 (21.6)


1.5 (27)


p <0.0001

LDL (mmol/L [mg/dl])


2.8 (50.4)


2.8 (50.4)


p = 0.9

Ever smoked (%)



p = 0.5

  • There were more deaths (39 vs. 23, p=0.03), and more deaths due to cardiovascular disease (50% vs. 38%, p=0.047), in the type 2 diabetes cohort. Dr. Wong commented that the increase in case fatality in the type 2 diabetes cohort suggested: 1) younger mean age of death; 2) shorter duration of diabetes before death; and 3) increased cardiovascular disease deaths. The first deaths and the first vascular deaths occurred in the third decade of life for individuals with type 2 diabetes, but later for individuals with type 1 diabetes. Furthermore, she noted that people of Anglo-Celtic descent with early onset type 2 diabetes were most at risk and had the most unfavorable metabolic syndrome factors.

Questions and Answers

Q: Are there any associations with socioeconomic status?

A: We don’t have SES data. But SES does have an impact in mortality and development of type 2 diabetes.

Q: Do you think the atherosclerosis process is the same in type 2 diabetes and type 1 diabetes?

A: I don’t know. I imagine that we have evidence for it being more aggressive. In terms of vasculature it looks similar in terms of it being a multivascular disease.


Andrew Althouse, MA (University of Pittsburgh, Pittsburgh, PA)

Mr. Althouse presented the results from the BARI 2D randomized clinical trial comparing insulin- sensitizing (IS) therapy to insulin providing (IP) therapy in peripheral arterial disease (PAD) risk. The BARI 2D design assessed the primary endpoint of incident PAD, defined by a new low ankle-brachial index (ABI) <0.9, and secondary endpoints of lower extremity revascularization and lower extremity amputation in people with type 2 diabetes randomly assigned to IS therapy (n=735) or IP therapy (n=744). Mr. Althouse showed a decreased five-year incidence in PAD composite, a measure including both primary and secondary endpoints, with IS therapy (24.1% in IS arm vs. 16.9% in IP arm). Similarly, he found a lower five-year cumulative incidence of PAD with IS therapy (19.9% vs. 28.6%). Interestingly, Mr. Althouse noted that cumulative incidence in the IS and IP arm did not begin to diverge until after two years. Additionally, Mr. Althouse pointed to a decreased risk for PAD (HR=0.76; p=0.02), incident low ABI (HR=0.73; p=0.002), and lower extremity amputation (HR=0.12, p=0.04) in the IS arm vs. the IP arm after adjustment for in-trial A1c. (HR for lower extremity vascularization was non-significant.) Mr. Althouse concluded that given these results, IS therapy may be preferable to IP therapy as the method of glycemic control in people with type 2 diabetes to reduce the risk of PAD.

  • The BARI 2D trial clinical trial assessed incident PAD, defined by a new low ABI (<0.9), in people with type 2 diabetes randomly assigned to IS therapy (n=735) or IP therapy (n=744). Participants were excluded from PAD analysis for a history amputation, lower extremity revascularization, missing baseline ABI, or abnormal baseline ABI (ABI ≤0.9 or ≥ 1.3). The IS arm did not differ significantly from the IP arm at baseline. This included age at study entry, sex, race, BMI, BP, current smoking, duration of diabetes, percent on insulin, and A1c.
  • The five-year incidence of PAD composite, a measure of both primary and secondary outcomes, was less in the IS arm (24.1%) than in the IP arm (16.9%). This corresponded to a 5-year incidence difference of: 1) 22.7% vs. 16.5% for incident low ABI; 2) 2.6% vs. 1.1% for lower extremity revascularization; and 3) 1.6% vs. 0.1% for lower extremity amputation (IS vs. IP; p-values not given). Similarly, Mr. Althouse presented data showing a decreased 5-year cumulative incidence of PAD in the IS arm compared to the IP arm (19.9% vs. 28.6%). He noted that incidence was consistent between the IS and IP arm through two years, but began to diverge to result in the almost 10% difference after five years.
  • Hazard ratios for PAD, incident low ABI, and lower extremity amputation in the IS arm vs. the IP arm indicated decreased risk with IS therapy even after adjusting for A1c.


Adjusted for in-trial A1c


HR (95% CI)


HR (95% CI)


Peripheral arterial disease


0.66 (0.51- 0.83)




0.76 (0.59-0.96)



Incident low ABI

0.68 (0.54-0.86)


0.73 (0.57-0.93)


Lower extremity revascularization

0.47 (0.20-1.09)


0.58 (0.24-1.37)


Lower extremity amputation

0.08 (0.01-0.63)


0.12 (0.02-0.91)


Questions and Answers

Q: Why do outcomes differ for PAD versus CAD?

A: I am not well positioned to answer that question. The risk factors for PAD versus CAD are different.

Q: Did you adjust for smoking?

A: Smoking rates at baseline were equivalent in the IS and IP arm, so we did not adjust for that.


Michael Lehrke, MD (University of Aachen, Aachen, Germany)

Dr. Lehrke presented results showing increased atherosclerosis in mice with overexpressed GLP-1 metabolite. He began his presentation distinguishing active GLP-1 from GLP-1 metabolite (GLP-1 metabolite is DPP-4-inactivated GLP-1). In a human study comparing total GLP-1 levels (approximately 10% GLP-1 and 90% GLP-1 metabolite) with coronary plaque in individuals undergoing CT- angiography for chest pain (n=303), increased GLP-1 levels were associated with increased plaque burden (OR=1.22 [1.03-1.48]; p=0.03). “We know this is very provocative,” he said and “needs to be confirmed in other studies.” Dr. Lehrke turned to a mouse model to compare active GLP-1 to GLP-1 metabolite. Eight-week-old LDL-receptor deficient mice were injected with an adeno-associated viral vector to overexpress GLP-1, GLP-1 metabolite,,or β-galactosidase. After 20 weeks on a high fat diet, the study found GLP-1 metabolite increased atherosclerosis. Dr. Lehrke proposed that the elevated non-HDL serum cholesterol levels observed in the study were at least partially responsible. Additionally, both GLP-1 and GLP-1 metabolite showed anti-inflammatory properties in the mice.

  • Dr. Lehrke noted the important distinction between active GLP-1 and GLP-1 metabolite. DPP-4 inactivates GLP-1 to form GLP-1 metabolite. While GLP-1 metabolites cannot activate GLP-1 receptors, Dr. Lehrke hypothesized GLP-1 metabolites may still affect the cardiovascular system.
  • Increased total GLP-1 levels were associated with increasing atherosclerotic plaque burden in humans (n=303). Coronary plaque was assessed by CT-angiography in 303 individuals (BMI: 23.2 kg/m2, 60% males, 63 years old, 7% with diabetes) with chest pain. The study found a weak, but significant relationship (p=0.03) between GLP-1 and total plaque (OR=1.22 [1.03-1.48]). The association persisted (p=0.02) even after adjusting for gender, BMI, C-reactive protein, medicines, TNF, IL6, and glomerular filtration rate (OR=1.30 [1.06-1.65]). Dr. Lehrke found this result surprising, expecting to find the opposite. He noted that while the study only measured total GLP-1, we have low levels of active GLP-1, begging the question: Is it possible that GLP-1 metabolites have a positive association with atherosclerosis?
  • GLP-1 metabolite, not GLP-1, increased atherosclerotic formation in mice. The study overexpressed GLP-1, GLP-1 metabolite,, or β-galactosidase by injecting adeno-associated viral vectors in eight-week-old LDL-receptor deficient mice. Mice were fed a high fat diet and analyzed at 28 weeks old at which point the peptide had reached pharmacological levels. As expected, GLP- 1 improved glucose tolerance. There was no appreciable difference in body weight. GLP-1metabolite significantly increased atherosclerosis (p value not given) compared to controls, whereas GLP-1 did not. Dr. Lehrke proposed that this was at least partially due to increased non- HDL serum cholesterol levels. Compared to controls, GLP-1 metabolites significantly increased total cholesterol, while GLP-1 did not. Further analysis revealed the increase was in non-HDL cholesterol. In accord with this pathogenesis, GLP-1 metabolites significantly decreased Cyp7A1 mRNA expression (Cyp7A1 is critical in synthesizing bile acid from cholesterol) compared to controls.
  • Both GLP-1 metabolite and GLP-1 had anti-inflammatory properties in mice. Notably, GLP-1 metabolite and GLP-1 significantly reduced mRNA expression of MCP-1 – implicated in monocyte recruitment.


Symposium: Macrovascular Disease in Type 1 Diabetes Mellitus


Maria Lopes-Virella, MD, PhD (Medical University of South Carolina, Charleston, SC)

Dr. Lopes-Virella presented data accumulated from more than 20 years of the Diabetes Control and Complications Trial (DCCT) and Epidemiologic Diabetes Interventions and Complications (EDIC) follow-up. Dr. Lopes-Virella described the data analysis of 880 individuals from the DCCT and EDIC to determine whether modified LDL in circulating immunocomplexes (mLDL-IC), inflammation biomarkers, or conventional risk factors could predict later coronary artery disease (CAD). High internal carotid artery intima-media wall thickness (ICA IMT) (> 1.7 mm) was used as a marker for CAD. Odds ratio for potential predictors were assessed at DCCT baseline (1983-1989), DCCT conclusion (1994), EDIC Pass 3/4 (2003/4) and 12 year EDIC (2006). At baseline, Dr. Lopes-Virella noted that mLDL-IC and A1c were good predictors of later CAD, conventional risk factors were less good, and novel inflammatory biomarkers were not good. However, she noted the caveat that DCCT participant inclusion had required normal blood pressure and lipids. At DCCT closeout, predictors did not differ tremendously from their power at baseline. Finally, at EDIC pass 3/4, inflammation biomarkers emerged as potential predictors – most notably, active PAI-1 and C-reactive protein (CRP). Dr. Lopes- Virella then asked whether inflammatory biomarkers and mLDL-IC could predict acute cardiovascular disease events. She presented data from a population of people with type 2 diabetes followed for a period of 40 years for acute MI and pointed to MDA levels in isolated LDL-IC as the only appreciable predictor of myocardial infarction.

  • Dr. Lopes-Virella described the analysis of data from the longitudinal, observational DCCT and EDIC study to determine whether mLDL, inflammatory, or conventional risk factors could predict atherosclerosis in people with type 1 diabetes. High ICA IMT at EDIC year 12 (2006) was used as the marker for CAD. Odds ratios for possible predictors of high ICA IMT were calculated at baseline (DCCT year of entry ranged from 1983-1989), DCCT conclusion (1994), and at EDIC pass 3/4 (2003/4). Dr. Lopes-Virella commented that DCCT baseline characteristics did not differ too much between those included in the analysis (n=880) and those not included (n=561). The mean age for included participants(27.1 years), diabetes duration (66 months), and albumin excretion rate (mg/ 24 H) did not differ significantly. However, A1c (8.8% for included vs. 9.0 for non included; p <0.01) and percent male (55.3% for included vs. 48.8% for not included; p=0.02) did differ.
  • At baseline, high levels of mLDL in circulating IC and high A1c were predictive of CAD in people with type 1 diabetes. Dr. Lopes-Virella drew attention to the odds ratio for theupper quartile of: 1) oxLDL (2.56 [1.17-5.74]); 2) AGE-LDL (3.18 [1.42-7.13]); and 3) MDA-LDL(3.18 [1.42-7.13]). Additionally, she pointed to the high odds ratio for the upper quartile of A1c(2.93 [1.39-6.17]). Interestingly, conventional risk factors were less useful predictors. She noted however that there was some bias in the study. Patients recruited for DCCT had to have normal lipids and blood pressure to be included. Inflammatory markers were not good predictors for the development of CAD. Dr. Lopes-Virella commented that predictive capabilities did not differ tremendously at DCCT closeout.
  • At EDIC pass 3/4, PAI-1 biomarkers were predictive of later CAD. Interestingly, mLDL was no longer predictive in people with type 1 diabetes. Dr. Lopes-Virella highlighted odds ratio for total PAI-1 (1.37 [1.06-1.77]) and active PAI-1 (1.52 [1.20-1.92]). She also pointed to CRP as a potential emerging predictor (OR=1.24 [1.01-1.52]).
  • Dr. Lopes-Virella suggested that MDA-levels in isolated LDL-IC were predictive of late acute myocardial infarction in people with type 2 diabetes in a population of people with type 2 diabetes followed for a period of 40 years. While the exact number was not given, from the graph shown, we estimated the proportional hazard of the fourth quartile relative to the first quartile was a little over 2.


Tina Costacou, PhD (University of Pittsburgh, Pittsburgh, PA)

Even though life expectancy and risk of kidney disease has markedly improved for patients with type 1 diabetes in the last 60 years, a similar decline has not been observed for coronary artery disease (CAD). While most standard risk factors for CAD are still predictive of CAD in people with type 1 diabetes, the protective effect of being female is lost, and there is conflicting evidence for the association of CAD risk and A1c. People with type 1 diabetes exhibit a disproportionately increased risk of CAD compared to the general population, and Dr. Costacou presented data indicating that haptoglobin (Hp) genotype may help predict risk for CAD for people with diabetes, allowing for preventative measures to be taken. Two common Hp alleles exist: Hp1 and Hp2. People with the Hp 2-2 genotype are twice as likely to develop CAD than people with the Hp 1-1 genotype. Encouragingly, evidence from the ICARE clinical trial indicated that people with diabetes and the Hp 2-2 genotype could take a vitamin E supplement to lower risk of CAD by 55%.

  • Great strides have been made in improving life expectancy and kidney disease for people with type 1 diabetes, yet the same improvement has not been seen with respect to macrovascular complications, namely coronary artery disease (CAD). Even though patients with type 1 diabetes present with a more favorable risk factor profile (e.g., total cholesterol for people with type 1 diabetes was 35% lower than that of the general population according to the CACTI study), they still exhibit a disproportionate risk of developing CAD. One explanation that has been proposed to explain the lack of improvement in CAD for people with type 1 diabetes is that prevalence of intensive therapy has been accompanied by an increase in the prevalence of obesity for people with type 1 diabetes.
  • Standard risk factors for CAD do predict CAD in people with type 1 diabetes, but there are discrepancies when it comes to gender and glycemic control; additionally, diabetes specific risk factors (especially kidney disease) are also strong determinants of CAD risk. There is a loss of the female protection in CAD for people with type1 diabetes. This may be due to less aggressive risk factor management (e.g., women are less likely to be prescribed preventative medication) or a loss of the relationship between HDL and gender (men’s risk for CAD is inversely related to HDL levels, but this relationship is not true for women). With respect to glycemic control, there is great variability of CAD risk compared to A1c. Some studies have shown no association between A1c and the development of CAD in people with type 1 diabetes, whereas a larger meta-analysis found that those with an A1c of 10.8% had a 2.8- fold higher risk for CAD.
  • Variants in the haptoglobin (Hp) gene may help predict risk for CAD in people with diabetes – risk of CAD is two times higher for diabetes patients that have the Hp 2-2 genotype compared to those with the 1-1 genotype. Hp is a plasma protein that binds to hemoglobin (Hb). Hp-Hb binding is crucial for normal hemoglobin function; it reduces hemoglobin’s oxidative potential for tissue damage, reduces glomerular filtration and loss of hemoglobin and iron following erythrocyte destruction, and reduces kidney damage resulting from iron deposition in renal proximal tubule cells. For hemoglobin that is highly glycosylated, as it is in people with diabetes, Hp binding to Hb does not confer as much protection from oxidative damage by Hb (Hp binding in a low glycosylation state can inhibit 90% of Hb’s oxidative potential, whereas Hp binding in a high glycosylation state might only inhibit 10% of Hb’s oxidative potential). People with diabetes and the 2-2 genotype have a two times greater risk for CAD compared to those with the 1-1 genotype.
  • Prediction is only useful if it can lead to prevention, and clinical trials have shown that cardiovascular disease risk for those with the 2-2 genotype can be significantly reduced with a vitamin E supplementation. In the ICARE trial, there was a 55% reduction in the risk for developing cardiovascular disease for people with diabetes that took a vitamin E supplement.

-- by Adam Brown, Eric Chang, Cati Crawford, Jessica Dong, Dana Lewis, Kira Maker, Lisa Rotenstein, Joseph Shivers, Kyle Rudolph, Mark Sorrentino, Josh Tennefoss, Tanayott Thaweethai, Katrina Verbrugge, Alasdair Wilkins, Nick Wilkie, Vincent Wu, David Zhang, and Kelly Close