New Treatments in the Complicated Patient with Type 2 Diabetes and Hypertension

I assume most of you have something to do with the kidney. In my regular life I come close to nephrologists only when I need their help to dialyze a patient with poorly treated diabetes. I do hope that by the end of this exercise we learn how to do a better job and reduce the number of patients who land in the dialysis center.

Nobody doubts anymore that there is a diabetes pandemic in the world. In fact, Dr. King of the World Health Organization (WHO) has projected that by the year 2025, there will be at least 300 million persons with diabetes worldwide; that would roughly represent a doubling of the rate from about the year 2000. So, there is an epidemic, and the vast majority of persons who develop diabetes will have type 2 diabetes. We also are seeing increases in the prevalence and incidence rates of type 1 diabetes across the world, but the magnitude of increase is nowhere near what we see with type 2.

The overwhelming explanation for the diabetes pandemic is attributable to type 2 diabetes; the natural history and pathophysiology of this disease has been increasingly understood. We now know that in persons selected by their genetic makeup and stressed by the environment, very early changes begin to occur in glucoregulatory physiology that translate to insulin resistance. Even in that early stage of "dysglycemia" -- where the blood glucose numbers are still, technically, within the normal range; and with regard to diabetes diagnosis, they are nondiabetic -- you begin to see changes in lipids that are unfavorable, specifically small decreases in high-density lipoprotein (HDL), the protective cholesterol, and increases in triglycerides. About that time, people also begin to accrue increases in both systolic and diastolic blood pressure numbers but they still do not qualify for diabetes diagnosis. With insulin resistance comes a demand for greater and greater secretion of insulin. The natural biologic response to a resistant state is overproduction of the signal hormonal ligand; the beta cells of the pancreas are called upon to work harder and harder to produce more and more insulin. This insulin-resistant hyperinsulinemic phase is really a prodromal phase for future diabetes, a prodrome that "crosses the Rubicon" when people begin to wake up with fasting glucose greater than 99 mg/dL. The individual whose fasting glucose reaches 100 mg/dL is no longer considered normal; 100 mg/dL in the fasting state now has a new name, impaired fasting glucose (IFG), and 126 mg/dL, as you all know, is diabetes in the fasting state. We have people in these prediabetic stages of IFG and impaired glucose tolerance (IGT), and they stay there for variable periods, depending on genetic makeup and other stressors. Some may evolve rapidly to overt type 2 diabetes, others may not progress that rapidly, and yet some others might even recover from this intermediate stage. We call that reversion to normal glucose tolerance. In those who are destined to move on to diabetes, we see the glucose numbers creep up into diagnostic ranges, and they are identified as patients with diabetes and treated accordingly. This is a conceptual slide from a review article by LaSalle, who is a primary care physician with great interest in diabetes and has been honing his skills in the area of diabetes research. Notice that eye damage and kidney damage -- the type you deal with in diabetic nephropathy, as well as diabetic neuropathy, the so-called, "microvascular complications" -- do not begin to occur until after the hyperglycemia has occurred. They are glucose-driven, glucocentric complications. However, long before the clock starts ticking for microvascular complications, you notice the beginnings of risk factors of macrovascular disease, such as dyslipidemia and hypertension, which together with other atheroinflammatory mediators, will set the stage for atherosclerosis. So, this occurs at an earlier time point.

But, getting back to the diabetes epidemic, and specifically to the pathophysiology of type 2 diabetes. We now know that there are multiple contributors to the eventual occurrence of hyperglycemia, chief among these is a change in tissue response to insulin. Insulin is the ultimate anabolic hormone that has major influences in the metabolism of glucose, protein, fatty acids, salt and water, and even memory facilitation in the central nervous system, among others. Yet, in persons who have inherited diabetes -- a set of diabetogens from their blood line -- who then also are confronted by environmental challenges, such as obesity or physical inactivity, these risk factors collude to uncover a state that, for lack of a better description, is called insulin resistance. That state is all-pervasive; every cell is involved in that perturbation. The simplest construct for understanding insulin resistance, almost a simplistic approach, is to measure insulin-stimulated glucose uptake by the tissues. If you did that years, and sometimes decades, before the occurrence of type 2 diabetes, individuals at risk might begin to fail that test -- the insulin sensitivity test, where you measure impairment in insulin-stimulated glucose transport. But, that is only one way of identifying insulin resistance. There are many other pathways of insulin signaling for which we, currently, do not have routine measurements or convenient tools to probe at a clinical level. This insulin resistance precedes the occurrence of type 2 diabetes, but it does not go away when the diabetes has occurred. It hangs in there and continues to negate the effect of any treatments we bring to bear on the diabetes, unless such treatments specifically address the resistance; and that is a very important therapeutic point to make. Now, in addition to insulin resistance, the beta cells of the pancreas play a leading role in allowing diabetes to occur. After all, you can induce all the insulin resistance in the world in a given subject; but, if their pancreas responds appropriately, by secreting tons and tons of insulin, the blood glucose will stay in the normal range. So, it is the failure of the beta cells of the pancreas to adequately compensate for progressive insulin resistance that becomes the denouement, the ultimate straw that breaks the proverbial camel's back and allows hyperglycemia to occur. So, insulin resistance is necessary, but not sufficient, to cause diabetes. A second hit, a second defect, residing within the beta cells of the pancreas allows for the diabetes to manifest; and that defect in the beta cell itself is now believed to be largely inherited. So, people in diabetes families would have a set of genes that predicts the insulin resistance and, perhaps, another set that predicts apoptosis, or increased cell death, within the Islets of Langerhans that prevents them from mounting a sufficiently robust insulin secretory response. Many will not know this intuitively, but it is true that gastric emptying rates are accelerated in people with diabetes. When you encounter diabetes in the context of gastroparesis, you tend to think that the bowels are sluggish in people with diabetes; but, actually, the primary motility defect in diabetes is accelerated gastric emptying. Some of the newer agents that have been introduced have actually worked on diabetes by slowing down the gastric emptying rate. Some of the incretin agents work on that. The other defect that we are beginning to understand in diabetes comes out of the adipocytes; lipolytic products like glycerol, free fatty acids can be used by the liver to produce more glucose molecules. Other adipocytokines that come out of the fat cells also are turning out to be very bad for glucose metabolism, by inducing insulin resistance and other processes. Finally, the liver plays a key role in the final pathogenesis of type 2 diabetes. Remember that the liver is a glucose factory, producing on average 1 mg glucose/lb of body weight/min. When people have type 2 diabetes, their average glucose production rate increases from the normal value of 1 mg/lb/min to 2 to 3 mg/lb/min. There are 1440 minutes a day; when you do the math, you can readily appreciate how much glucose is coming out of the liver. That number can be doubled or tripled if you are dealing with a diabetic patient. Interestingly, all of these pathophysiologic defects in type 2 diabetes also are veritable targets for drug development. We have at least 1 drug on the market that will be addressing each of these targets.

Beta Cell Defects and Diabetes Outcomes

Let me spend a moment on the beta cell defect, because it is really the Achilles' heel of type 2 diabetes pathophysiology. If you take away somebody's pancreas, you convert them immediately to diabetes; if you do a total pancreatectomy, if you deprive a system of insulin, that system immediately becomes diabetic. But, if you increase insulin resistance inordinately highly, you still do not have 100% straightforward diabetes, as long as the beta cells of the pancreas are rising to the challenge. The failure of the beta cells to rescue nearly 20 million Americans who already have developed diabetes and 300 million people in the world who will develop diabetes by the year 2025 is an important research matter. We are not quite sure why the beta cells are failing. We do know that the cell mass decreases and that the mechanism of cell mass loss is through programmed cell death, or apoptosis. These triggers for the apoptosis, however, are not well known, but we do have candidates. Time will not permit me to really flesh out this slide, but suffice it to say that current research has accorded priority to elucidation of the mechanism for beta cell death in type 2 diabetes. Numerous targets, including immunologic targets, lipotoxicity, and hyperglycemia itself, have been identified as a signal for cell dysfunction. Above certain thresholds of ambient blood glucose levels, the beta cells begin to take a hit; we call it glucose toxicity. They begin to be affected down to the genomic level, where the transcription and translation of the insulin gene becomes impaired by ambient hyperglycemia. That is a reversible cause of beta cell dysfunction.

If we go back to the diabetes epidemic, one of the most worrisome signs from the clinical research field is that one does not need to have big scale increases in blood glucose to begin to suffer dire consequences. There is a study coming out of the United Kingdom called the European Prospective Investigation Into Cancer (EPIC). It is a cancer-focused study, and in one of the study sites, in Norfolk County in England, they had enough measurements of biochemical values, glucose values, over several years to allow analysis. What you see on the screen is cardiovascular mortality, outcomes data, on the basis of shifts in blood glucose levels captured by the hemoglobin A1c within reasonably normal ranges of life. If you look at physiologic ranges of hemoglobin A1c, people without diabetes score from about 3% to at most 6% on that test. Within up to 5.5% A1c, you are seeing a doubling of mortality from coronary heart disease, compared with people whose A1c is less than 5%. By the time you get into the early diabetic range, A1c of 7% or higher, the increase in the risk of dying from coronary heart disease is 7-fold. If you go to all-cause mortality, you see the same "staircase effect," where every increase within the normal range, an average glucose, is predicting a measurable increase in the risk of mortality. This is a very powerful study that draws attention to early dysglycemia, early escape from normal glucose, being cardiotoxic in ways that were not previously fully appreciated.

When somebody already has the diabetes and then suffers a heart attack, we know that the consequences are disastrous and the prognosis is horrendous. If you do the 200-day mortality analysis for people with or without diabetes who have suffered a heart attack, those who do not have type 2 diabetes (or any diabetes for that matter) and whose in-hospital blood glucose average is lower than 200 mg/dL have the best survival. Persons who have type 2 diabetes and whose glucose exceeds 200 mg/dL have at least a 30% decrease in survival within 200 days. Look at these data. These are folks who did not know they had diabetes. The worst category were people with no known history of diabetes. You get a heart attack, and in the course of that hospitalization, your glucose runs in the 200 mg/dL range -- most likely the 5.5 million Americans walking around with diabetes, but do not know it yet -- those undiagnosed diabetic folks get a heart attack and get in there. Their glucose is not addressed; because they do not have the level of diabetes in the entry history and physical (H and P), nothing is done to address glucose. Look at their survival -- 50% dead in 200 days. So, there is a gradation of glucotoxic effects manifesting in the cardiovascular system that precedes the clinical diagnosis of diabetes, and it then amplifies the risk of dying from diabetes if an event has occurred.

Monodrug vs Combination TherapyMonodrug vs Combination Therapy

Let's switch gears and address the better part of my talk -- what are we going to do about it, because so far I have been painting gory pictures of horror stories happening to our patients. The good part is that we have things to do about it. I mentioned earlier that most of the pathophysiologic contributors to type 2 diabetes themselves present veritable targets for intervention; so, you look at the level of myocyte where insulin resistance is most richly expressed. We do have agents that can reduce insulin resistance and improve insulin sensitivity. The thiazolidinediones (TZDs), and to some extent metformin, can improve peripheral insulin sensitivity. If you look at the excessive glucose production in the liver -- which you recall is 2- to 3-fold increased in type 2 diabetes compared with nondiabetic controls -- we have agents (metformin primarily, but all the TZDs in full dosages) that can be shown to reduce hepatic glucose production. The adipocyte issue is most responsive to TZDs, where the average fall in free fatty acid recorded can approach 47%, 50% reduction from baseline; that is a major target for TZD action. There is insulin secretion, of course; we have had 70-plus years of familiarity with using secretagogues to treat type 2 diabetes. They operate on diabetes by stimulating insulin secretion. Of course, the caveat is that there must be residual beta cells for them to work. That is why we do not prescribe them routinely to kids with type 1 diabetes, who are bereft of all beta cell function. But, if there are some cells left, the secretagogues can stimulate them to release some more insulin, at least for some time.

The bad news is that the more medications we get approved in this therapeutic space, diabetes management, the worse the national hemoglobin A1c appears to become. That paradox has not been fully explained to my full understanding. But, look at the report cards. Clearly, well below 50% of our patients with diabetes are getting their hemoglobin A1c controlled to the minimally acceptable target of 7% or lower. The American College of Endocrinology actually asks for 6.5%, and there are studies that are testing 6% as the proper goal. So, we are not really accomplishing a lot by way of control, and that is not an isolated lack of achievement in diabetes, even in blood pressure management. If you combine diabetic folks who also have hypertension and high cholesterol, in whom all 3 conditions are treated to excellent degrees, you are talking about less than 10%. It has been a challenge.

The challenge can be addressed, in my opinion, by adopting a deliberate policy of combination drug therapy. One of the weaknesses of chronic disease management is the patient factor. The patient adherence over prolonged periods of time is challenging. We can "sneak beneath radar detection," if you will, by loading on more than 1 chemical to the pill, so that, when patients think they are taking 1 pill, they actually are taking multiple medications. That may be one way to address the major problem of lack of excellence in glucose control. Whenever dual combination therapies have been pitched against monotherapy, the outcome has always been predictable. Here you have the glyburide/metformin combination pill tested against metformin alone in higher dosages (more than twice as much as was loaded here); or glyburide alone at least twice the dose. "Half and half" of these drugs will get nearly 80% of your subjects at the American Diabetes Association (ADA) target of A1c; whereas, individual drugs titrated to much, much higher milligram dosages failed to match the combination product.

That is an approach that we can increasingly use; but are we using it? Far from it. This study came out of the Kaiser Permanente group, a group model health maintenance organization (HMO), in fact, the granddaddy of all HMOs, where everybody is insured and has access to all kinds of specialties or interventions. In that study, Dr. Brown looked at the charts of diabetic patients whose doctors have registered an A1c of 8% or greater on 1 occasion, and he looked forward in the chart to when an appropriate action was taken. An appropriate action being defined as switching to another drug or adding another diabetes drug onto the one that did not seem to be working that well, because the patient's A1c was high. It turned out that it took too long for any change, any response to high A1c, to be recorded or to be documented -- more than 1 year, whether patients were initially on Glucophage (metformin) or on a sulfonylurea. The physicians were waiting too long. To the extent that Kaiser Permanente doctors are usually well trained, often multiply board certified, and drawn from a cross-section of the US demography, they do not mean ill, they are not incompetent -- to that extent these numbers reflect national habits. I have seen data, not presented today, where a different set of practitioners was audited and the numbers were not much better than you see. People are waiting longer than a year to address high A1c levels. We need tools also. When we advocate combination therapy, we need tools that can get patients close to the goal. The national average A1c is still hovering uncomfortably closer to 9% and 10%, rather than to the 6.5% that we desire.

Effectiveness of Combination Therapies

One of the newer products, a drug that combines rosiglitazone and metformin in a fixed-dose combination pill, recently has been shown to give us as much as a 4 percentage point reduction in hemoglobin A1c when used as first-line therapy for individuals selected for their initially high hemoglobin A1c levels. Now, I must add that that product (rosiglitazone and metformin, marketed as Avandamet) has not yet been approved for initial therapy, but it is approved for later therapy and combination therapy. But this study is so convincing that I would be surprised if that indication does not become available soon, because 4 percentage points mean we can restore many of our patients close to the 7% A1c goal.

Similar data exist, proving the point that combination therapy always delivers better quality care than single agents. Again, I do not want to belabor the point, but, the percentages reaching both the American Association of Clinical Endocrinologists (AACE) goal and the ADA goals are higher when the combination product was administered.

Similar studies have used a different thiazolidinedione, a drug called pioglitazone (marketed as Actos). When combined with metformin, it can give you a nice decrease in hemoglobin A1c. Metformin is combined with a sulfonylurea drug called gliclazide (not available in this country but widely used in Britain and continental Europe). Again, the faster, more rapid action of sulfonylurea is evident, but the long-term impact is similar between the 2 drugs. Combination therapy works; that is the bottom-line message.

In a more recent study, Rosiglitazone Early vs Sulfonylurea Titration (RESULT), we really have the point crystallized for us. In this study, individuals were divided into 2 camps. One-half received titration of a single sulfonylurea drug, glipizide (Glucotrol), to a maximum dose of up to 40 mg/day. The other group had half of the maximum dose of sulfonylurea with rosiglitazone added. Again, you have seen the hemoglobin A1c response to the combination product being exceedingly greater.

Insulin resistance is the bane of type 2 diabetes, and decreasing that resistance is a target. Rosiglitazone/sulfonylurea combination group did that, whereas sulfonylurea alone did not have an impact on the insulin resistance.

If you look at the same data using pioglitazone, again, TZDs do improve insulin resistance. Sulfonylureas do not work on diabetes by altering insulin sensitivity; they are secretagogues. This is a very well-known point, but from time to time you may hear presentations that do not fully accord with the reality. So, no sulfonylurea on the market has an insulin-sensitization property.

Summarizing Combination Therapy

Advantages of TZD-containing preparations have been written about in the literature, but we need studies

The maximization of a single drug often is not a winning strategy in diabetes management. That has been repeatedly proven for us. Early recourse to combination therapy is a smart move. The use of fixed-dose combination strategies might ease patients into the treatment by optimizing compliance, as compared with their having to open multiple pill bottles and take them. Aggressive treatment of diabetes is really the way to go.

Cardiovascular Implications of Thiazolidinedione Therapy

I want to show you some quick clinical trials, one that recently has concluded, others that are coming. The TZDs, when you look at them, offer so much theoretical benefit that one often jumps to the conclusion that they ought to be part and parcel of every diabetes regimen. Well, before we do that, we need evidence, clinical evidence.

The study here is the PROspective Pioglitazone Clinical Trial in MacroVascular Events Study (PROactive study) of patients with type 2 diabetes who have already had cardiovascular or microvascular events or problems and have multiple risk factors. They were treated to "standards of care" and then one-half were given pioglitazone; the other half were not given pioglitazone, but a matching placebo. They were followed for 36 months. You see a 10% decrease in the composite primary end points, which included myocardial infarction (MI), dying from any cause, cardiovascular accident (CVA), and acute coronary syndrome, among others. This benefit accruing to the TZD barely missed statistical significance, and for that reason we say it was not significant. However, 10% fewer events is clinically meaningful, 1 in 10 fewer events. Why was the primary end point not significant? The point being made here is that perhaps the group of patients enrolled for the study were too far gone. Remember, macrovascular complications begin before clinical diabetes. The presence of diabetes is equivalent to 1 heart attack. You saw from the data of Haffner and colleagues that diabetes is a coronary risk equivalent, not a risk factor. Diabetes, plus events, plus additional risk factors; therefore, it must be a tertiary prevention study. So, it is even highly impressive that, despite that late advanced stage, a TZD study is still able to show some benefit. In fact, when you look at the secondary end points, the prespecified secondary end points, you get a bigger event reduction and statistically significant numbers.

This is time to acute coronary syndrome.

This is time to fatal or nonfatal MI. So, I think that the story continues to look rosy for TZDs. Other studies in the works have learned from this study and this design. They are targeting earlier-stage individuals before any heart attack, sometimes even before diabetes. I will show you some of those studies.

Management Issues With the Thiazolidinediones

People gain weight on TZDs. Part of the weight gain is fluid; the other is actual fat cells being accumulated. That has discouraged many practitioners from considering these agents. It also has demoralized some patients from staying on them.

For that reason, a panel was put together by cardiologists and diabetologists to look into fluid retention issues with regard to thiazolidinediones. Here is a summary of the recommendations. I commend the full article to you to look up, but it is really reassuring. We are quite sanguine with the condition that less than 1% of congestive heart failure is triggered by TZD monotherapy. Even for those who are edematous, the edema will not be due to heart failure in the vast majority of subjects. However, if cardiac status deteriorates and crackles begin to be felt in the lung bases, or shortness of breath occurs, or a cardiologist confirms that there is a worsening of heart failure, the drug should be stopped. Indeed, no TZD should be commenced in a patient who is New York Heart Association class III or IV at baseline. What happens when somebody gains some fluid and has demonstrable edema on a TZD? What do we do? Right now we have been ad-libbing and empirically trying one diuretic or the other, or a drug holiday.

A study is now available that directs attention to what may work and what may not work. In this study (using hematocrit dilution as a surrogate for fluid expansion, volume expansion), 381 patients with type 2 diabetes being treated with sulfonylurea alone or sulfonylurea and metformin, were then given a TZD, rosiglitazone. Nearly two-thirds of them developed fluid expansion, as indicated by a dilution of the hematocrit. They were then randomly assigned to take furosemide, a loop diuretic; hydrochlorothiazide, a thiazide diuretic; Aldactone (spironolactone) an aldosterone-inhibiting potassium-sparing diuretic; or stop the TZD and see what happens to the fluid retention.

This is the first of its kind, of therapeutic studies, for TZD fluid retention. What became very clear was that an ordinarily weak diuretic in the grand scheme of things, spironolactone, proved to be the most potent in preventing fluid retention on a TZD, followed by thiazide diuretic;. The loop diuretic, furosemide, was without significant benefit, and compared with continuing the rosiglitazone, giving furosemide (Lasix) did not change things much. Stopping the rosiglitazone was moderately effective in limiting edema; but the best results came when you continued the rosiglitazone and you gave a 25 mg dose of Aldactone. So, that is something we can use.

Safety-wise from the RESULT study, emergency room visits decreased as did hospitalization rate when individuals in the TZD, rosiglitazone/sulfonylurea arm were compared with those whose diabetes treatment was a single drug titrated to the maximum. The mechanism is not very clear, but it is good news and reduced cost.

Studies of Early-Stage Diabetes

In the home stretch, I want to show you what is in the works. You have seen this graph before. I told you that, right up to this point, blood tests will not show diabetes in any of your patients. In this early stage, we begin to see IFG or prediabetes, and we have had a paucity of studies that have intervened in that early stage. The PROactive study went beyond late diabetes, for diabetes plus event, to see if a TZD would change the natural history. What we have are 2 powerful studies that are on the horizon of being completed and one about to be presented. The "A Diabetes Outcome Progression Trial" (ADOPT) trial goes a little later into early diabetes.

The Diabetes Reduction Approaches With Ramipril and Rosiglitazone Medications (DREAM) trial is targeting and has enrolled prediabetics, completely nondiabetic people in the IFG or IGT range. They are being followed for diabetes as an outcome, as well as others (long list of coronary, cardiovascular, and total life outcomes, mortality, etc). These studies are going to be very valuable in the way we approach diabetes and the way we decide to intervene early, and what tools we use for early intervention. In the DREAM trial you have ramipril, an angiotensin-converting enzyme (ACE) inhibitor, being used alone or in combination with rosiglitazone, which is also being used alone or in the combination. Then, you have double placebo; it is a 2 X 2 factorial design. They are being followed after enrollment. They will stay on protocol a maximum of 5 years and then be washed out.

A variety of outcome measurements, including diabetes, coronary events, stroke, and macrovascular disease are being tracked. This study is at an advanced stage of completion, and we will hear results from it before long. It is powered to the extent of 90% power to detect a 22% reduction in the intervention arm, so it is well designed.

In ADOPT, (this is a very busy slide), the crux of this trial is that individuals recently developing diabetes, less than 3 years (on average only 8 months of diabetes history) would immediately be randomly assigned to take either glyburide, metformin, or rosiglitazone, representing the 3 most widely used classes of oral antidiabetic agents.

They will then be followed for an end point that consists of a strictly and rigorously defined failure of monotherapy. We will try to determine which one of these 3 drugs will carry the most patients the farthest along in a state of good control. If that result comes to the finishing line and we have the data, we will be able to share with each other what is the best initial oral agent to introduce shortly after diagnosis of type 2 diabetes on the basis of the evidence.

Finally, I have told you that there is a diabetes epidemic raging. There is no sign of a crest or a plateau at all. It is associated with a galaxy of complications that conveniently can be divided into small-vessel and large-vessel complications. The large-vessel complications are what kill our patients, the heart attacks and strokes. They occur even before the diabetes has become clinical grade, and they are now the target of antidiabetic interventions, not just the glucose. We have been overly focused on glucose outcome measures. We are now expanding and demanding more of our diabetes drugs. What more can we do for our patients with diabetes besides lowering their glucose? Would we prevent the heart attack, etc? Studies are on the horizon that will tell us that.