The challenge in diabetes treatment used to be getting patients' blood glucose low enough. But recent evidence has shown that, at least for elderly diabetics, hypoglycemia may be as much or more of a problem as hyperglycemia.
An analysis of Medicare beneficiaries hospitalized between 1999 and 2011 found that hyperglycemia admissions dropped by 38.6% while admissions for hypoglycemia increased by 11.7%. Hypoglycemia surpassed hyperglycemia as a cause of hospitalization in 2001-2002 and continued to do so through the rest of the study, which was published by JAMA Internal Medicine on May 17.
Another study of high-risk, elderly veterans taking sulfonylureas and/or insulin in 2009 found that a significant percentage were at risk for hypoglycemia due to overtreatment: More than 10% had a last HbA1c under 6% and about half had one under 7%, according to results published in the Feb. 1 JAMA Internal Medicine. Such evidence has led experts to make major changes in their thinking about diabetes treatment in the elderly.
“There's been a pendulum shift from ‘We have to focus on poor control’ to ‘Optimal control for everyone,’ and now there's a swing back to the middle, that is, individualized targets,” said Leonard Pogach, MD, MBA, FACP, national director of medicine at the Veterans Health Administration and an author of the veterans study.
This shift has included changes in national recommendations, with the American Diabetes Association (ADA) and American Geriatrics Society (AGS) proposing a new framework for treating glycemia in older adults in a 2012 consensus report and the VA and Department of Defense addressing the issue in 2010. The new framework suggests different HbA1c goals based on whether patients are in good, intermediate, or poor health.
But with a number of other factors potentially affecting treatment choices, from patient preferences to finances to living situations, the changes have been easier said than done, according to experts, who offered their thoughts on reducing glycemic overtreatment of elderly patients.
A bigger trend
The move to less intensive diabetes treatment is part of a larger trend. “It's not just glycemic control. If you look at the new cholesterol guidelines, if you look at the new hypertension guidelines, [prostate-specific antigen] screening, there's been a lot of revision of what was considered dogma,” said Dr. Pogach.
The changes are a response to increasing evidence that many older patients may not have time to benefit from these long-term treatments and screens. “The broader issue would be to incorporate life expectancy into all decision making regarding any intervention,” said Patrick O’Malley, MD, MPH, FACP, professor of medicine at the Uniformed Services University of the Health Sciences in Bethesda, Md.
This concept contradicts the traditional culture of medical practice. “Anything that's abnormal we have a desire to normalize. We think that normal is better,” said Dr. O’Malley. “Unfortunately, we've discovered that logic doesn't always hold up when tested in clinical trials.”
The typical population of a clinical trial also helped delay the realization that intensive treatment wasn't benefiting the elderly. “Older populations don't generally get included in clinical trials, especially older complicated patients, so we're misapplying our evidence and overgeneralizing it in populations that have not been included in clinical trials,” Dr. O’Malley said.
Like trials, quality measurements of physicians and health care systems have been generalized without regard for age or health. “I think some of the [primary care physicians] are getting paid according to what these [HbA1c] numbers are. There's a grading system. They get fliers, reminders around [achieving an HbA1c of] 7%,” said Faramarz Ismail-Beigi, MD, PhD, a professor of medicine at Case Western Reserve University in Cleveland.
“When I've gone out to talk about this, the kinds of questions I've gotten are ‘Won't we be in trouble with performance measurement? Isn't that going to be a problem?’” said Deborah Wexler, MD, an assistant professor of medicine at Massachusetts General Hospital and Harvard Medical School in Boston. Generally, today, it won't, she explained. “One of the reasons the performance measures don't set 100% targets for A1c at a certain threshold is because they recognize that there are clearly exceptions to an A1c target of less than 7%, and many of the big exceptions occur in older patients.”
However, clinicians and guidelines have not always been clear on exactly which older patients should be exempt. “The ADA guidelines a few years ago kept talking about individualizing. What does that mean? What they say is if somebody's really old or dying, then we shouldn't be trying to overtreat them ... But when you read that material, it suggested that they were talking about a minority of patients, not a large number,” said Dr. Ismail-Beigi.
In fact, patients can be quite a long way from death and still not benefit from tight glucose control. “It takes probably about 15 years to start seeing differences in outcomes compared to those not treated aggressively,” said Dr. O’Malley. “An HbA1c goal less than 7% is reasonable for anyone with at least a 15-year life expectancy.”
Even once physicians were aware of this evidence, they may have had difficulty talking to patients about it, because it is both difficult and unpleasant to estimate life expectancy. “There is a general denial of death, denial of frailty,” said Dr. O’Malley. “It is a noxious topic.”
Talking to the patient
The ADA/AGS framework offered an entry to this difficult conversation and some specific criteria on which to base elderly patients' glycemic goals.
According to the consensus paper, healthy patients (those with few or no comorbidities and intact cognitive and functional status) should have an HbA1c goal below 7.5%. Intermediate patients (multiple comorbidities, 2 or more impairments in activities of daily living [ADLs], or mild to moderate cognitive impairment) should aim for under 8%. And finally, patients in long-term care or those with end-stage illness, moderate to severe cognitive impairment, or 2 or more ADL dependencies should have a goal below 8.5%.
The categories sound fairly definitive, but by calling their advice “a framework for considering treatment goals,” the experts made it clear that even they are not absolutely certain of the optimal treatment course.
“There is not yet a clearly defined target for glucose control that you set a group of experts down around the table and they'll say, ‘Oh, for this patient, it should absolutely be under 6.5% or under 7.5%.’ There's going to be debate,” said Jeffrey B. Halter, MD, professor of internal medicine, director of the Geriatrics Center at the University of Michigan in Ann Arbor, and an author of the consensus paper.
“Therefore, this is an area where involving the patient and family is important to do,” he continued. However, patients vary in both their interest and ability to participate in shared decision making about treatment goals, Dr. O’Malley noted. “In the elderly, there's a much higher prevalence of cognitive impairment, so it's even more challenging,” he said.
Cognitively impaired patients will often come in with a caregiver, though, who can be engaged in the decision making, Dr. Halter said, offering an example of his conversational strategy: “I often say, ‘Do you understand what hemoglobin A1c is? What do you think is the right target?’ I get all sorts of answers to that one. I say, ‘There's national discussions and guidelines, and here's what some of the options are.’”
He offers his analysis of where the patient fits in the ADA/AGS framework and then solicits the patient's and/or caregiver's perspective on the choices. “Sometimes they just say, ‘Doc, what do you think I should do?’” Dr. Halter said, but the majority will have some opinion.
Other factors to consider
Even if the decision is left to the physician, he or she should solicit some other information in conversation and use it in choosing a goal and treatment, according to Dr. Ismail-Beigi. “I need to know how much support they have at home. Do they live alone or have a wife or a husband who is healthy or bedridden?” he said. “Can they treat hypoglycemia if they get it? And can they afford [their medication]?”
If a treatment regimen doesn't work with a patient's lifestyle, it's better to know that up front. “I say to the patient, ‘What can you do? What are you willing to do?’” said Dr. Ismail-Beigi. “No sense asking them to do things they're not going to do.”
These psychosocial issues may actually be more important to choosing a glycemic goal than the medical issues, he added, noting that requirements like frequent blood glucose checks can be a significant burden. “What's their life going to be like? We're trying to improve their life,” Dr. Ismail-Beigi said.
There are also medical issues that didn't make the ADA/AGS framework but should be considered in goal setting. Duration of diabetes is a major one. “If you have a 70-year-old person who's already had diabetes for 15 years and doesn't have cardiovascular disease yet, it may not be very effective to lower that person's glucose intensively. Whereas, a 70-year-old who's been normal until they were 69and a half, maybe it's very effective to keep their glucose levels very close to normal,” said Dr. Halter.
Relatedly, clinicians should consider how difficult it will be to get a patient to a given goal. “A low A1c that can be achieved with diet, exercise, and metformin is a good A1c,” said Dr. Wexler.
Physicians often underutilize lifestyle changes in the treatment of elderly patients, Dr. O’Malley noted. “Exercise is harder in the elderly, where there's less functional capability, but where there is functional capability, we should be emphasizing it,” he said.
When the first-line treatments of lifestyle changes and metformin aren't enough to get elderly patients to their individualized goal, clinicians face the toughest dilemmas of how intensively to treat.
“We need to think long and hard before starting insulin or a sulfonylurea in an elderly diabetic, and we need to be much better aware of the risk of hypoglycemia,” said Dr. O’Malley. The risk of hypoglycemia is the major reason for the move toward less intensive treatment in the elderly, after all. It's not only more common but potentially more risky for them than younger patients, Dr. Ismail-Beigi noted.
“When younger people get severe hypoglycemia, it's damaging but probably not as significant. The younger brain probably recovers, but older people don't seem to recover so well,” said Dr. Ismail-Beigi.
The alternatives pose problems of their own, however. “I personally don't use thiazolidinedione drugs hardly at all. Maybe that's an overreaction, but benefits and risks have been debated,” said Dr. Halter. “The category of thiazolidinediones scares me in this population at high risk for cardiovascular disease.”
There's also the problem that alternatives to insulin are individually less effective. “How many different classes of oral agents should one try before going to insulin? There's no clear answer to that question, but I'm very sensitive to polypharmacy issues,” said Dr. Halter. “I definitely do not like using more than 3 classes of drugs.”
How many medications a patient is already taking for various comorbidities should certainly be a consideration in selection of a glycemic target and treatment, the experts said. And if a patient is newly diagnosed with diabetes, the process of adding medications should be very slow.
“We start conservative and safe, and as the person learns more and they can do better and better, then I can tighten it a bit,” said Dr. Ismail-Beigi. “It doesn't bother me greatly if their A1c is 8% for a while. Their world won't come to an end. But it will be a major problem if an older person develops severe hypoglycemia and has a heart attack.”
Scary as that possibility sounds, the overall message for clinicians, who have been constantly cautioned about the risks of hyperglycemia for the past couple decades, is one of reassurance. “This disease is very, very slow, so everybody could take a chill pill and slowly get people to the target range,” Dr. Ismail-Beigi concluded.