Using ACCORD to Target HbA1c: Think Ages and Stages

An Expert Interview With Faramarz Ismail-Beigi, MD

Linda Brookes, MSc


January 04, 2011

An Interview With Faramarz Ismail-Beigi, MD

About the Interviewee

Faramarz Ismail-Beigi, MD, PhD, is a Professor of Medicine, Endocrinology, and Physiology and Biophysics at Case Western Reserve University, University Hospitals of Cleveland, and the Cleveland VA from 1993-2008. He served as Chief of Endocrinology at Case from 1993-2007. His interests are in teaching, clinical diabetes research, and basic research in thyroid hormone action, glucose transport biology, and insulin action. He was the Principle Investigator of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial for the Ohio/Michigan Network. Dr. Ismail-Beigi was also the Diabetologist PI of another multicenter National Institutes of Health (NIH)-funded diabetes trial that was conducted at University Hospitals of Cleveland, the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial, which was in some ways complementary to the ACCORD trial. He has authored more than 150 basic and clinical science papers.

Background to the Interview

The ACCORD trial, sponsored by the National Heart, Lung and Blood Institute (NHLBI), was set up to test the efficacy of 3 medical treatment strategies (blood glucose lowering, blood pressure lowering, and lipid altering) in reducing cardiovascular events in middle-aged and older people with type 2 diabetes plus established cardiovascular disease (CVD) or additional (≥ 2) cardiovascular risk factors.[1,2] The multicenter, double 2×2 factorial design ACCORD trial recruited 10,251 patients in the United States and Canada, all of whom participated in the glycemia arm of the trial. This arm compared a strategy of intensive glycemia therapy targeting normal glycosylated hemoglobin (HbA1c) levels (< 6.0%) with a treatment strategy targeting HbA1c 7.0%-7.9%. The primary outcome measure for this as well as the other 2 arms of the trial was a composite of nonfatal myocardial infarction (MI), nonfatal stroke, or cardiovascular death. A total of 5128 patients were randomized to intensive glycemia treatment group and 5123 to standard treatment. Intensive therapy was stopped in February 2008 after a median of 3.5 years, because of a statistically significant 22% relative and 27% per year absolute increase in all-cause mortality seen in that group.[3] The patients in the intensive glycemia arm were transitioned to standard therapy for the remainder of the median 5.0 years of follow-up. At the point of transition there was no significant reduction in major cardiovascular events (primary study outcome) with intensive glycemic therapy compared with standard therapy.

Among the predefined secondary endpoints of ACCORD was a composite microvascular outcome comprised of renal failure or retinopathy (similar to the United Kingdom Prospective Diabetes Study [UKPDS]), and 13 prespecified secondary individual measures of kidney, eye, and peripheral nerve function. The results of the 2 glycemia control strategies on these microvascular outcomes were presented at the 2010 annual scientific sessions of the American Diabetes Association (ADA) and published simultaneously in The Lancet.[4] This analysis showed no significant difference in the primary composite microvascular outcome between the intensive- and standard-therapy arms, either at transition (hazard ration [HR], 1.00; 95% confidence interval [CI], 0·88–1.14; P = 1.00) or at the end of the study ([HR, 0.95; 95% CI, 0.85–1·07; P=.42). A second composite endpoint, which added peripheral neuropathy to the primary composite outcome, also did not differ at transition (HR, 0.96; 0.89–1.02; P = .19) or at study end (HR, 0.95; 95% CI, 0.89–1.01; P = .12). However, intensive therapy led to a significant 21% reduction in development of microalbuminuria at transition compared with standard therapy (HR, 0.79; 95% CI, 0.69-0.90; P = .0005). This reduction was lower at study end (15%) but still significant (HR, 0.85; 95% CI, 0.77-0.94; P = .0012). The incidence of macroalbuminuria was also significantly reduced in the intensive glycemia group compared with the standard group at transition (HR, 0.68; 95% CI, 0.54-0.86; P = .0013) and at study end (HR, 0.71; 95% CI, 0.59-0.86; P = .0003). These were the only 2 (out of 5) microvascular renal outcomes that differed significantly at transition and study end. The number of patients with overt renal failure was similar in both groups at study end. There was a significant difference between the groups at transition in the microvascular renal outcome based on serum creatinine or estimated glomerular filtration rate (eGFR) that favored the standard glycemia group, but this difference was not significant at study end. Dr. Ismail-Beigi and his colleagues noted that a decline in eGFR in the intensive group in the first 24 months might have been due to a decrease in glomerular hyperfiltration associated with improved glycemic control.

In addition to the renal outcomes, 4 secondary measures of eye complications and neuropathy favored intensive glycemia therapy (P < .05) at study end. For diabetes-related eye events, cataract extraction was significantly reduced (by 21%) in the intensive group compared with standard group. Other diabetes-related eye outcomes did not differ significantly between the 2 groups. Peripheral neuropathy, defined as a Michigan Neuropathy Screening Instrument (MNSI) score of ≥ 2.0, was less common in the intensive group than in the standard therapy group. Loss of ankle jerk reflex and light touch (10 g monofilament) perception were both less frequent in the intensive than in the standard therapy at study end; light touch was also reduced at transition. Loss of vibratory sensation did not differ significantly between the 2 groups.

In The Lancet report, Dr. Ismail-Beigi and co-authors noted that the microvascular benefits of intensive therapy should be weighed against the observed risks: increase in total and CVD-related mortality, increased weight gain, and increased risk for severe hypoglycemia. They advised caution in pursuing a strategy of intensive glycemic control for prevention of microvascular complications in diabetic patients similar to those in the ACCORD trial. "An HbA1c target of ≤ 6.0% with present strategies seems imprudent," they concluded.

In an accompanying commentary published in The Lancet,[5] Ronald Klein, MD, (University of Wisconsin School of Medicine and Public Health, Madison) pointed out that the reduction of incident microalbuminuria seen with intensive glycemia control in ACCORD was consistent with findings of the Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) trial[6] and the Veterans Affairs Diabetes Trial (VADT).[7,8] The ACCORD investigators' conclusion to exercise caution was "appropriate," Dr Klein said, but he also stressed that importance of not misinterpreting the ACCORD results as a justification to return to the poor glycemic control that was common before the findings of the UKPDS were reported.

Dr. Ismail-Beigi spoke with Linda Brookes, MSc, for Medscape, to discuss the results of the ACCORD analysis and its implications for clinicians.

The Interview

Medscape: When ACCORD was designed there was an expectation, based on previous trial evidence, that as well as cardiovascular benefit, there might be a favorable effect of blood glucose lowering on microvascular outcomes. What was the evidence for this at that time? There was some evidence from the UKPDS, but surely that study population was younger and treated longer than the ACCORD patients?

Faramarz Ismail-Beigi, MD: Yes, the patients in the UKPDS were not only younger, they were newly diagnosed. Of course, one never knows exactly how long a person has had type 2 diabetes before diagnosis (it may be years), but you often do know when the diagnosis is made. UKPDS was designed in a newly diagnosed patient population and the trial lasted close to 11 years.[9] With the tools and medications that were available, the investigators showed very good separation between the standard and the intensive therapy arms (average A1C of 7.9% vs 7.0%, respectively).[10] Of course, the medications that were available at that time were different and much less numerous than what is available now. I was not present when the NIH was conducting initial discussions about the kinds of trials that should be done and the ACCORD trial was formally proposed. However, from the protocol development phase, I think it was clear that the thought was that the UKPDS was getting close to being significant in terms of macrovascular outcome and they felt that there was no sense in repeating that trial, because it was conducted as well as it could be at that time and in that population. So the NIH (and its consultants) must have concluded that a trial in a population apt to have more events would be worthwhile. It was probably surmised the P value of .052 seen for the difference in macrovascular outcomes in UKPDS[11] might become statistically significant in people with more advanced stage of disease. Others might have thought that the disease process, once advanced, might not be easily reversible. Nevertheless, given that cardiovascular event rates are increased 2- to 3-fold among patients with type 2 diabetes (especially with prior CVD events) compared with controls, it must have seemed plausible that over a 5-year period there would be a separation between treatment arms and that the trial would show a significant positive CVD outcome. It is also worth noting that there are millions of people who fit the profile of patients enrolled in ACCORD, and knowledge of how best to manage these patients was (and is) of paramount importance. Concerning microvascular disease, both the UKPDS and the Kumamoto study[12] had shown that intensive glycemic treatment was associated with reduced rate of progression of microvasular complications.

I think that the bulk of the community in diabetes believed that ACCORD would show a positive result of intensive glycemia treatment on the macrovascular outcomes, and they were also expecting a positive result in microvascular outcomes. Of course, the answer was not known, which necessitated the trial. They believed that they were not going to be surprised, because of the results of the UKPDS study,[11] the Kumamoto study,[12,13] and the Diabetes Control and Complications Trial (DCCT).[14] So the macrovascular and microvascular outcomes were predicted to be positive in ACCORD, although there were also some feelings that that might not be the case, in part because of a study done in the 1960s, the University Group Diabetes Program, which really did not show an effect.[15,16] probably because it might have had problems in its design.[17]

It was also not clear what the relative importance of glycemia, blood pressure, and lipid control were in type 2 diabetes, a complex metabolic disease we knew to be more than just hyperglycemia. So in the ACCORD trial, and to some extent in the ADVANCE trial, the investigators looked at other factors. Both ADVANCE[18] and ACCORD[1,2] examined the effects of glycemia treatment and blood pressure control, and ACCORD also tested treatment of dyslipidemia. The VADT trial on the other hand was very pure; it addressed only the glycemic question, with implementation of 2 different glycemia strategies, intensive vs standard,[19] so in a certain way it was more pristine in terms of answering the question of whether glycemia alters cardiovascular outcomes, while actively controlling all other factors. All 3 trials enrolled older patients (~ 60 years old) who had had diabetes for about 10 years. In addition, all patients had at least 1 additional risk factor; in the ACCORD and VADT they had to have more than 1 risk factor and a large percentage, about 40%, in the VADT trial and about one third in the ACCORD and ADVANCE trials had a previous cardiovascular event. So these were people with relatively advanced disease. For example, in the ACCORD cohort, close to 27% of patients entered the trial with microalbuminuria and close to 7% with macroalbuminuria. Of course, serum creatinine had to be < 1.5 mg/dL, because the thought was that metformin would be used heavily in both arms; so one of the exclusion criteria was having creatinine > 1.5 mg/dL. A significant proportion (8%-9%) also had a history of photocoagulation or vitrectomy.

The only difference between the patients in the ACCORD and the ADVANCE trials was that although the duration of the known diabetes was a little longer in ACCORD (10 vs 8 years), only 1% of patients in ADVANCE were on insulin at enrollment, whereas about one third of patients in ACCORD were on insulin at entry into the trial, signifying more "advanced" or "aggressive" disease. In addition, the entry criterion for ACCORD was HbA1c≥ 7.5%, whereas the ADVANCE trial took all-comers; at enrollment, the median A1cs in ACCORD and ADVANCE were 8.1% and 7.1%, respectively. So the trials were a little different, but ACCORD, ADVANCE, and VADT all basically showed the same thing, ie, that in this group of patients, which is actually a sizable group and not a small subset of patients with type 2 diabetes, intensive glycemia treatment does not prevent all-cause mortality or death from cardiovascular events. ACCORD actually found that there was higher all-cause mortality and CVD-related mortality associated with intensive glycemia treatment, but all 3 agreed that there is no benefit in terms of decreasing the above outcomes.

There have been reports that some subsets do benefit from intensive glycemia therapy. In 2009, Peter Reaven, MD (Phoenix Veterans Affairs Health Care System, Phoenix, Arizona), and colleagues reported that people who entered the VADT trial with relatively high coronary artery calcium (CAC) scores (a marker of advanced atherosclerosis) did not benefit from intensive glucose-lowering therapy, whereas people who entered the trial with very low CAC scores had a reduction in CVD events.[20] Again, this is consistent with the idea that it depends on the age of the patient and the duration of disease.

The entire VADT cohort had relatively more advanced disease (compared with ACCORD and ADVANCE) with probable significant atherosclerosis, most with at least 2 risk factors for CVD, and a large proportion with a previous event. So, even in this group of patients with relatively advanced disease, one can see positive effects of intensive glycemia treatment on cardiovascular outcomes in certain patients, and positive effects on microvascular outcomes.

Medscape: The incidence of microalbuminuria and macroalbuminuria were reduced in all the trials?

Dr. Ismail-Beigi: The VADT investigators first reported that intensive glucose control in patients with poorly controlled type 2 diabetes had no significant effect on progression of macroalbuminuria or microalbuminuria[7] then corrected the results.[21] So all 3 trials are in agreement that you can decrease development of microalbuminuria and macroalbuminuria with intensive glucose therapy. ACCORD also showed some positive effects on eye outcomes, and a separate eye substudy (ACCORD-EYE) that used fundus photography, showed a significant reduction in the rate of progression of diabetic retinopathy with intensive glycemic management.[22]But, as noted above, the prespecified microvascular outcome, which was a composite of renal failure and retinopathy, did not show any benefits from intensive glucose treatment. One cannot say why, but there was no effect on this composite outcome of advanced microvascular disease designed to mimic the UKPDS microvascular outcome. Some people have said that maybe 3.7 years was not long enough to show any benefit, but there is no evidence for any improvement with intensive glycemic control.

Medscape: Was ACCORD really long enough?People who have carried out similar trials with renal endpoints always stress how long it takes to reach renal endpoints like this.

Dr. Ismail-Beigi: It was expected in the 5 years there might be a difference, but I continue to believe that what we see is that there is no positive signal for renal failure. Outcome number 3, development of renal failure, was the same. Now you could say that you have to wait another 10 years or so; after all, UKPDS was 11 years in duration, so you have to wait a long time before the groups separate. That is what we hope to try to do in a follow-on study of the ACCORD trial. But we should remember that even in the UKPDS (and in the DCCT), there was no benefit from intensive glycemic control on end-stage renal disease. Perhaps it takes a very long time for intensive glycemic control to result in a benefit on this advanced complication.

Medscape: Have the ACCORD patients been approached to enroll for the follow-up study?

Dr. Ismail-Beigi: They are still in the study, because we asked for and received an extension of the study from the NHLBI. The extension phase started July 2009 and runs through December 2010. Most of the participants have remained in the trial. By the end of 2010 they will have had 3 telephone visits, 1 every 6 months. So we have remained in contact with most of the participants. There has been some data collection, but at the moment it is rudimentary. Patients tell us when they have been hospitalized or have had a stroke or a heart attack or whether they have had a severe hypoglycemic episode, or have had other serious medical conditions. We are not managing the patients any longer, and we are not doing any laboratory tests. So we are not getting all the information that we need to do any analysis of value, other than capturing mortality and major CVD events. We are also not getting hospital records. We are currently waiting to see whether NIH will approve the follow-on study. It would run for 4 years of follow-up with clinic visits for examinations, data collection, and some laboratory tests, with interim phone contacts to obtain information about clinical events. We expect to hear soon whether this extension study will be approved, and if it is, it will begin in January 2011.

Medscape: You mentioned benefits seen in some subgroups. Was there any difference in microvascular outcomes according to glucose-lowering drugs or regimen used?

Dr. Ismail-Beigi: That is not known at present, but some of this analysis is ongoing.

Medscape: Because of the factorial design of the study, might there have been an additive influence of blood pressure lowering on microvascular outcomes in addition to glycemia treatment, as recently reported in ADVANCE?[23]

Dr. Ismail-Beigi: We are in the process of analyzing that question in the ACCORD trial, because it is of interest to know whether there was any interaction or whether there was additive effects between these 2 arms.

Medscape: In ADVANCE there was no interaction, but there was an additive effect on microvascular but not macrovascular outcomes. So do you think that patients who were on optimal blood pressure lowering, such as a renin-angiotensin system (RAS) inhibitor, and optimal glycemia therapy, etc, show a beneficial effect on vascular outcomes?

Dr Ismail-Beigi: In the ACCORD blood pressure trial there was no positive effect of intensive blood pressure lowering on the primary cardiovascular outcome.[24] The intensive arm of ACCORD-BP trial, targeted a systolic blood pressure of < 120 mm Hg, which is a target that no trial has ever attempted or achieved before. The ACCORD-BP trial achieved an average systolic blood pressure of 119.3 mm Hg on the intensive-therapy arm and 133.5 mm Hg on the standard-therapy arm. So there was a good separation, but no effect on the primary outcome, but a reduction in strokes. So I think that in this population we have reached a lower limit of what does not help any more. The UKPDS intensive blood pressure-lowering arm achieved 144/82 mm Hg,[25] and in ADVANCE blood pressure lowering was not target-specific, but achieved an average of 135/75 mm Hg, which was associated with a significant positive effect on cardiovascular outcomes.[26] So it may be that systolic blood pressure somewhere around 135-140 mm Hg is the lower limit and is a reasonable target in patients similar to those in ACCORD.

Medscape: I think no trial in type 2 diabetes mellitus has demonstrated a benefit in lowering blood pressure to below 130 mm Hg?

Dr. Ismail-Beigi: That is correct. Current ADA guidelines recommend an systolic blood pressure target of <130 mm Hg,[27] but there is no strong type A evidence for that target. It is possible that the authors of the guidelines may decide to reconsider that goal in certain patients.

Medscape: What about the recommended HbA1c level in the guidelines (< 7%)? This is are already being reconsidered on the basis of the macrovascular outcomes, but since microvascular outcomes showed some benefit with intensive lowering, what do you think the recommendation should be?

Dr. Ismail-Beigi: This is a point that is going to come up a lot, because the way the ADA guidelines are written, even in the 2010 version, they give 1 A1C target number "in general" and then state that there are exceptions. Of course there are 3 or 4 exceptions that are obvious, such as someone who is dying or has another serious disease. The guidelines list some exceptions and then recommend individualizing, but they do not specify in necessary detail what they mean by individualization.

Medscape: The guidelines appeared to be a big topic of discussion at the 2010 ADA meeting.

Dr. Ismail-Beigi: At the end of our session about ACCORD, all the discussion was about guidelines, what the results mean, and what we should do. We are in the process of writing an article (opinion piece) about what it might mean to individualize treatment. It is possible that the ADA will issue a new, expanded statement about the topic of individualization. The following is my personal view, based on what we have learned from the long-term follow-up of the UKPDS,[28] the ACCORD, VADT, and ADVANCE trials, and the DCCT follow-on study.[29] If I see a patient who is similar to an ACCORD participant, who has fairly advanced disease or is in the middle stage of their disease, and have had an event or have CVD risk factors, I pay a lot of attention to treatment of all the risk factors. The Steno-2 trial showed the cardiovascular benefits of controlling blood pressure and lipids in addition to glycemic control, use of aspirin, lifestyle changes, and smoking cessation in high-risk diabetic patients,[30] and I would pay attention to all of them. I would probably target HBA1c around 7.5% unless there were reasons that I thought the patient was going to develop advancing microvascular disease. So in 60-year-old patients who have had the disease for a long time and have very little microvascular disease, I am not going to push hard to get HbA1cway below 7%. It depends on other comorbid issues, how many other disease states they have, what their finances are, whether they have some back-up at home, what their support system is, how able they are to do whatever they are trying to do, but I would aim the glucose somewhere around 7.5%. I would aim to control the blood pressure to around 135 mm Hg, and low-density lipoprotein cholesterol, to well below 100 mg/dL, or to below 70 mg/dL if they have had a CVD event. I would get them to stop smoking, and if there are secondary conditions such as mentioned, I would make sure that they are on aspirin. The A1c target would be lower if there were early microvascular complications.

On the other hand, I can tell you that if tomorrow a 35-year old patient with recently diagnosed type 2 diabetes with no complications walks into the clinic – and unfortunately we see this type of patient almost every day – I tend to be as aggressive as I can be, given all the safety issues that I mentioned. I try to get them to be normal in terms of their glucose, blood pressure, and lipids. I will try hard to get close to normal, as long as I don't cause harm, as long as they can follow the protocol, and as long as it is not exorbitantly expensive and they can afford it. People with relatively newly recognized type 2 diabetes mellitus are easier to control. They usually do well with a couple of medications. If they have almost no beta cells or they have nonfunctioning beta cells, they need to be on insulin; that is a different disease.

In presentations that I give about diabetes, I often mention "ages and stages" of the disease. What is important right now is to classify patients as much as possible in that broad spectrum and by individualizing them. I think of the age of the person, the duration of their disease, whether they have microvascular disease, macrovascular disease, or history of severe hypoglycemia. Then I look at what other diseases they have. Then I look at whether they are depressed or not, at their cognitive function, what their financial situation is. What is it that they want to do? After all, the patient has got to be in charge of their health and therapeutic program: it is a lot of work and they have to understand it and be motivated. Their target can change with time. Patients who are newly diagnosed may have difficulty in absorbing all this information, because they are shocked by their diagnosis. So we may have 1 target and later, as they get better and they see success and they lose weight and they feel better, the target may change and we may make it more stringent. I regard this as a chronic disease; you cannot have 1 idea and 1 target all the time. So I believe that it would be nice if the guidelines changed to accommodate the ideas of individualization, and biologically they accommodated the age and the stage of the patient's disease. The disease is not the same early on in its course, and what needs to be done is not the same over time.

Medscape: When you mentioned the guidelines, were you referring mainly to the ADA guidelines?

Dr Ismail-Beigi: The guidelines of the American Association of Clinical Endocrinologists (AACE) recommends an HbA1c target of ≤ 6.5%[31] and the European Society of Cardiology (ESC) and the European Association for the Study of Diabetes (EASD) recommend 6.5%,[32] and the ADA has said < 7%.[27] Again, I think that an A1C of < 7% or < 6.5% may make a lot of sense in a younger individual who is going to live 40 years, hopefully. If you and they do not manage their disease, they might be on dialysis and may become blind and have amputations over the decades if not properly controlled. But in someone who is 69 years old and has other diseases, diabetes is a problem, but it is not their preeminent problem. That is what I mean by individualization. Primary care physicians ask me all the time what this all means and what they should do; there is confusion out there. The global message is that in relatively advanced disease and older age, you have to have different goals. You cannot just have 1 goal and push that; 1 goal for everyone does not work anymore. At the 2010 ADA meeting, we presented 5 papers on ACCORD, and then we had a half hour discussion. As best as I recall, not a single question from the audience was addressed to anything we said about the data. All that was discussed was what does this all mean and how is this going to alter our practice? Should the guidelines be changed? So that is what is causing anxiety and what healthcare providers want to know; that is the need right now. It is interpretation of all this information that is needed, and I have given you my own personal version, but someone else may have a slightly different version, so I think that we need to have discussion and hopefully reach a consensus.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.