Irl B. Hirsch, MD
Irl B. Hirsch, MD, an internationally-recognized diabetes expert and former ADA Clinician of the Year, is a professor of medicine at the University of Washington School of Medicine and Medical Director of the Diabetes Care Center at the University of Washington Medical Center. Patients trust him because they know they are talking to an expert and a peer -- Dr. Hirsch has had type 1 since he was a child. He discussed with us the importance of "glycemic variability," the risks and rewards of continuous glucose monitors, and the latest on new products.
Kelly: Dr. Hirsch, thanks so much for discussing your current views on diabetes research and treatment. Today we'd like to start by asking you about your focus on glycemic variability. You have said that too much emphasis is placed on A1c control and not enough on keeping all blood glucose values as close to normal as possible. It seems this view is finally starting to move into the mainstream among doctors and nurses. What prompted your initial interest in this area?
Dr. Hirsch: When we started downloading data about 10 years ago, I realized that our gold standard of A1c wasn't enough, and I also realized that looking at simple averages didn't give me the whole picture. I saw that I could get much more meaningful information by evaluating standard deviations [glycemic variability from normal glucose levels]. It all started with an amazing software program from LifeScan that had time-specific standard deviations. I realized that I could look at a download and see who was making smart choices about insulin dosing, who was testing more between meals, and who was in optimal control. I think that a lot of people are embarrassed about the pizza and the beer that they had on Saturday night, and they don't want their doctor to see that 435 blood sugar . . . [but] I see the 400 plus blood sugar when I do the downloads, and that's good data for me. When I see a standard deviation that's way out of whack, when I see certain times of day when the averages are either too high or too low, I can focus on that. When you have such limited time for office visits as we do, you need to learn how to analyze and counsel as optimally as possible.
Kelly: But what made you believe that glycemic variability was important for a patient's overall control?
Dr. Hirsch: It started with a report from the DCCT study group in 1995. As you recall, patients in the intensive therapy arm were mostly taking mealtime regular insulin and basal with NPH or Ultralente. They were doing a better job than the conventional group of matching prandial insulin with their food intake. But at any given level of A1c, the intensive therapy group had 40% to 60% less risk of developing complications.
Kelly: Let's talk about continuous glucose monitoring. Some health care professionals and researchers worry that it provides the patient with too much blood glucose information. Do you feel real-time continuous monitoring is ready for prime time, particularly in light of questions regarding calibration and analysis?
Dr. Hirsch: First of all, nobody really has ever taught health care professionals—much less patients—what to do with the information. We're in the early days. If you have a continuous graph, what do you do with the glycemic profiles, especially if you are a patient? That is, you may see your current blood glucose is 95 and you know that a half hour ago your glucose was 135, and now you're getting ready to eat, and now you would normally give 8 units for what you're eating but your blood sugar is dropping quickly. How much less insulin do you take? How do you change the lag time? You know, usually you take your rapid-acting insulin, you wait a few minutes before eating—eating nothing may be dangerous in this case.
Kelly: How do you think the monitors actually perform?
Dr. Hirsch: We need to all remember they are first generation and they will all improve. Obviously, they are not yet perfect, but for people who understand how to use trends in addition to SMBG, our initial experience has been great. For me, I can't wear the Medtronic as for some reason I get inflammation with their sensor. In my Star1 study, no one had that problem, so I must be unusual in that regard. Clearly, no one has software of the data as good as the Medtronic Carelink, but to be fair, I haven't really had enough experience with the Abbott software to make a good comparison. DexCom does not have all of the bells and whistles literally, but for me that is a good thing. I look at my sensor constantly, so having it alarm all day would just be an annoyance. The arrows one gets with the Abbott and Medtronic are great benefits, and hopefully we will see something similar in the future from DexCom. For me, and as my wife will agree - I won't sleep or leave the house without wearing it, whichever one it is.
Kelly: Both Byetta and Symlin are generating a lot of interest for type 1 and 2 patients respectively. How widely are you prescribing these new drugs and what has your experience been?
Dr. Hirsch: Quite a bit for both. Symlin is tricky as we have learned that only those sophisticated with their insulin use it long-term. The three-times daily dosing with the old fashioned syringe is a turn-off for some, but those who do well with it don't want to come off of it. A lot of these patients need a lot of hand-holding the first month or two, so providers need to know about that. Byetta seems to cause more nausea than Symlin, and most of us have had great success using it with basal insulin . . . but new type 2 patients when talking about what drug to start often decide to take Januvia simply because they don't want to deal with the shot or shots and the hassle of keeping it cool.
Kelly: Exubera made a big media splash when it was introduced, but patient interest and acceptance appear to be mixed at best. What is your view of this drug and on inhaled insulin more broadly?
Dr. Hirsch: Inhaled insulin may be great for some patients, but it obviously doesn't appear the market is ready. My take is that the insulin will need to be easier to administer than the Exubera is, and the insulin will require some type of pharmacokinetic/phamacodynamic advantage. That is why I think both the Lilly preparation and the Mannkind insulin will do better.
Kelly: For all the tools and technology that we have, patient motivation is still critical for effective management. How do you motivate your patients? If you had advice for motivated patients out there who want to do better, what would that be?
Dr. Hirsch: This is very difficult. I feel I spend a lot of my time as a cheerleader. Some patients do better when family members are involved (and some do worse!). Some do better when writing things down, while others seem to be motivated more with weekly downloads. The biggest problem I see (probably worse in Seattle) are those patients who fail due to depression. Mental illness is a huge problem which often gets overlooked by both patients and physicians. If I see a college graduate, or maybe someone with less education but has done well with a pregnancy, and we download the meter and I see less than two checks daily, my first guess is I am dealing with depression, often not yet diagnosed. So this is a huge issue, and the reason we have a psychologist and psychiatrist in our clinic. When the depression is adequately treated, many (but not all) of these patients do better.
Kelly: You've been a big advocate of Lantus. Do you believe it's prompted more type 2 patients to go on insulin?
Dr. Hirsch: I think Lantus has taken away some of the intimidation of insulin from primary care doctors. I do think that the data suggests that it's taken away a lot of other insulin product sales, such as Lilly's 70/30, NPH, etc. But for all the wonderful things about Lantus, the problem is that the average A1c for starting insulin is in double digits for type 2 patients, so Lantus by itself is not going to get them to target. They're going to need some kind of prandial [mealtime] insulin. And where the practical problems start—let's say you have someone with an A1c of 9.5 or 10. Start them on Lantus and the best you can do is get their A1c level to say 8. If they're on Lantus and 8.2, what are you're going to do? Are you going to go for prandial insulin? You know when you started the Lantus that really what should happen is that you should go straight to four shots a day, and that's the last thing type 2 patients want to do. But if your A1c is that high, and you start them on a premixed insulin, rather than Lantus, you can get them down to 8.2 easily on two shots a day and you can probably get them into the 7 range as well. And even though you have all the now-obvious problems with taking those two shots a day with having to eat the meals on time, with higher risks for hypoglycemia and so forth, there are still a lot of patients who would prefer two shots to four - and as an endocrinologist, even though the premixed insulins are less stable than Lantus, if they have some mealtime insulin, I'm much happier, even if it means foregoing Lantus. Best would be Lantus and rapid acting mealtime insulin, or for some potentially an insulin pump.
Kelly: How willing are type 2 patients to take shots?
Dr. Hirsch: If you look at the national data, it's interesting, there are more type 2s taking three and four shots a day than ever before. So there are some trends showing that we're seeing more MDI [multiple daily injections] in the type 2 population, but it's a very slow pick-up. There are still big problems with the type 2 populations and how they are treated. Say an average doctor wants to get patients' A1c down into the 6 range. They can do that with Lantus if they start them on insulin when their A1c is 8. But when they're starting with an A1c of 10, you know that if they start the Lantus without other insulin, they are not going to reach goal. They're probably going to need more shots. And if you're a type 1 and you don't know any different, that's fine. But for a type 2, it may not fly.
Kelly: The fact that diabetes care delivered via the telephone or Internet is not routinely—or ever—reimbursed seems an artifact of another age—very inefficient. I think a lot of us would do more to access care if we could receive it via the Internet. The reality is that the hassle factor in visiting a clinic, sad to say, often outweighs motivation. If patients could spend a few minutes writing an update, and CDEs and doctors could write back when time permitted, wouldn't that in many cases be more efficient for everyone? Obviously regular visits are key, but sadly, today, follow-up is really lacking.
Dr. Hirsch: There's a real reimbursement problem in this area, as you say. Basically, we're in the office at 8 o'clock at night working for free. That's the problem. It's ultimately just not affordable - that's why so many endocrinologists have left the field.
Kelly: And because there's no reimbursement, that reduces further the number of CDEs, which makes office visits even less productive.
Dr. Hirsch: The reimbursement for nurses and nutritionists and behavioral scientists, if anything, has gone down in the years since the DCCT was published!
Kelly: It seems clear that patient outcomes would improve with better and easier access. We all know that most of the costs of this disease are associated with chronic complications, not with treatments. How can access be increased?
Dr. Hirsch: Industry probably has some clout. Every state I talk to, everyone is fighting for their lives to try to keep the little bit of coverage they have now. One can't even pay salaries with the kind of coverage that is available in most states.
Kelly: There is great concern about the shortage of endocrinologists. Where do you see this heading?
Dr. Hirsch: If you look at who is going into them, the endocrine programs today are much different than 30 years ago. Then, they got the smartest and brightest people from internal medicine classes. Today, those people are going into cardiology, GI, and pulmonary. There, you get reimbursed for procedures. Right now in pediatric endocrinology, there's truly a crisis. It's very discouraging overall and, actually, as passionate as you are about diabetes, where are you going to get your care 20 or 30 years from now? I worry that 20 years from now, we are not going to have endocrinologists. Nobody is coming in to take the place of the people who are retiring. There is no financial incentive to do it right now.
Kelly: On the bright side, diabetes care is arguably more exciting now than ever before, with the introduction of new classes of drugs and the publication of recent landmark studies. Can you share some of your thoughts on the ADOPT trial?
Dr. Hirsch: ADOPT compared three different drugs in monotherapy: rosiglitazone, metformin, and glyburide. As I recently said in DOC news, the most relevant way to think about the results is that A1c was maintained below 7% for 60 months with rosiglitazone (Avandia), 45 months with metformin, and 33 months with glyburide. So rosiglitazone clearly was more effective and durable, particularly against metformin, at least in a well-performed statistical analysis.
Kelly: And sulfonylureas performed the worst. It's interesting that right now the ADA/EASD's drug therapy algorithm for type 2 diabetes recommends initiating all patients on metformin, not rosiglitazone. We assume this is because it's the least expensive drug and it's weight neutral - although it does have side effects. Does ADOPT change the way we should be initiating drug therapy?
Dr. Hirsch: I think not. Let's look at the drug costs. For the agents used in ADOPT, a month's supply of drug cost $211 for rosiglitazone, $56 for metformin, and $18 for glyburide. Comparing rosiglitazone with metformin, do 15 additional months prior to drug failure justify this added cost, particularly since no added cardiovascular benefits have been proven for this population? Besides, the 6.9 kilogram weight gain [15.2 pounds] with rosiglitazone, compared with metformin, is the fundamental reason I can't recommend this therapy as first-line for this often obese population. We need to look at type 2 diabetes differently now than we did a decade ago, before so many drug therapy options existed. Now there are drugs that are generally weight-neutral, such as metformin, DPP-IV inhibitors, and alpha-glucosidase inhibitors, or even weight-losing like exenatide. These agents must at least be considered in anyone's algorithm.
Kelly: We so appreciate your time. Thank you again and all the very best to you from me and the readers of diaTribe.