Intensive Glycemic Control May Harm Some Diabetes Patients

Miriam E Tucker

June 08, 2016

One in five elderly and "clinically complex" patients with type 2 diabetes may be receiving unnecessarily intensive glucose-lowering treatment, leading to a dramatically increased risk for severe hypoglycemia, a new study finds.

The findings were published online June 6 in JAMA Internal Medicine by Rozalina G McCoy, MD, assistant professor of medicine at the Mayo Clinic, Rochester, MN, and colleagues.

"It is time that we, as physicians and patients, recognize that high-quality diabetes care should emphasize not only avoiding hyperglycemia but also preventing hypoglycemia," Dr McCoy told Medscape Medical News.

She added, "We should recognize the harms of intensive treatment and hypoglycemia, particularly when there is little likely benefit of keeping HbA1c low in the setting of limited life expectancy or multiple comorbidities."

The retrospective database analysis included over 30,000 adults with type 2 diabetes who had HbA1c levels less than 7% without using insulin and had no episodes of severe hypoglycemia or hyperglycemia in the prior 12 months.

Nearly 4000 of the subjects were aged 75 years or older and/or had serious comorbidities. During the 2-year study period, 19% of that "high clinical complexity" group received intensive glucose-lowering therapy. And among those patients, the risk for severe hypoglycemia necessitating medical care was 3%, compared with just 1.7% for the high-complexity patients not receiving intensive treatment.

"Severe hypoglycemia is not rare, even among patients who are not treated with insulin and who have no prior history of severe hypoglycemia," Dr McCoy commented.

More Research Needed on How to Deescalate Treatment

The findings highlight the lack of clinical guidance on when to stop or "deescalate" treatment in people with chronic conditions, and in fact clinicians are often incentivized not to, say Eve A Kerr, MD, of the Veterans Affairs Center for Clinical Management Research, University of Michigan Medical School, and Timothy P Hofer, MD, professor in the division of general medicine, University of Michigan, Ann Arbor, in an accompanying editorial.

"Balancing the medical profession's focus on aggressively treating patients who are likely to benefit with an explicit consideration of when to deintensify treatments when they are no longer useful or are potentially harmful, and doing so in a manner that is respectful to the patient-physician relationship and promotes shared decision making, is the next frontier for improving care quality," Drs Kerr and Hofer write.

Dr McCoy agrees. "The culture of healthcare has been that 'more is better,' and doing less is automatically an indicator of low-quality care. We need studies looking at how and when treatment should be deescalated…without causing patient harm."

Also in the same issue of JAMA Internal Medicine is a short research letter describing a small study by Medha N Munshi, MD, of Joslin Diabetes Center, Boston, Massachusetts and colleagues, in which the insulin regimen in older adults with type 2 diabetes was simplified. This was achieved by switching multiple daily premeal dosing of insulin to once-daily insulin glargine, resulting in decreased hypoglycemia and disease-related distress without compromising glycemic control.

Dr McCoy commented, "It was reassuring to see that their HbA1c did not rise, which I think is a common fear of physicians and patients when faced with the prospect of deintensifying glucose-lowering therapy."

Clinical Complexity and Hypoglycemia

In Dr McCoy et al's study, the researchers obtained 2001–2011 figures for adults with diabetes who had HbA1c levels below 7% over 2 years from an administrative claims database of more than 100 million individuals enrolled in private and Medicare Advantage plans across the United States.

After exclusion of patients with severe hypoglycemia in the prior 12 months, those prescribed insulin in the prior 120 days, and those with type 1 or gestational diabetes or missing data, a total of 31,542 patients were included.

High clinical complexity, defined as a composite of age 75 years or older or high comorbidity burden defined by the presence of end-stage renal disease, dementia, or three or more serious chronic conditions, was identified in 12.4% of the total (3910 patients).

"Intensive" treatment was defined by the addition of any medication in those with baseline HbA1c levels of 5.6% or below, the use of two or more drugs at baseline or the addition of one more over the study period in those with initial HbA1c 5.7% to 6.4%, or the addition of two or more drugs or insulin in those with baseline HbA1c 6.5% to 6.9%.

Regimens that did not meet intensive-treatment criteria were considered standard treatment.

In all, 8048 patients (25.5%) were treated intensively, including 7317 (26.5%) with low clinical complexity and 731 (18.7%) with high clinical complexity.

Of note, 76% of the intensively treated patients did not have their treatment deescalated after the low HbA1c test result was obtained. Approximately three-quarters of both the low and high clinical complexity groups continued with their baseline intensive regimen.

The risk-adjusted probability of intensive treatment was 25.7% for patients with low clinical complexity and 20.8% for those with high clinical complexity (P < .001 for the absolute difference).

Patients with high clinical complexity were significantly less likely to be treated intensively than patients with low clinical complexity (odds ratio [OR], 0.76).

The overall unadjusted 2-year incidence of severe hypoglycemia — defined as episodes necessitating an outpatient, inpatient, or emergency visit — was 1.4%.

Severe hypoglycemia was significantly more frequent among patients with high vs low complexity (2.9% vs 1.2%, P < .001).

Among patients with low clinical complexity, the risk-adjusted probability of severe hypoglycemia did not increase with intensive treatment (1.02% with standard treatment vs 1.30% with intensive treatment).

In contrast, among patients with high clinical complexity, the risk-adjusted probability of severe hypoglycemia increased significantly from 1.74% with standard treatment to 3.04% with intensive treatment.

Sulfonylurea and glinide therapy significantly raised the risk of severe hypoglycemia (OR, 2.19). However, the overall results remained consistent in a sensitivity analysis even after researchers excluded patients using these agents, although the numbers were small.

Dr McCoy noted that newer diabetes drugs are "thought not to cause hypoglycemia and therefore used with less caution...yet we now see that these medications too can cause hypoglycemia among clinically complex patients…I think that physicians need to ask all patients receiving pharmacotherapy for diabetes about hypoglycemia, not just if they receive insulin or sulfonylureas."

Also of note, patients treated by endocrinologists had significantly higher risk of hypoglycemia (OR, 1.65), even after adjustment for the HbA1c level and medications used.

Reasons for Not Deescalating

In their editorial, Drs Kerr and Hofer summarize four main reasons that treatment deintensification is rare even when clearly indicated.

First, the question is rarely examined in clinical trials, so data are lacking about when to stop most medical services.

Second, physicians and patients may suspect that cost saving is behind recommendations to stop treatments. Third and, "less nobly, physicians may be concerned about their own performance on report cards."

Finally, the editorialists suggest, lack of time and effective decision-support tools may hinder discussion about stopping or scaling back on established treatments and services.

"When guidelines are silent on the limits of generalization, the default in clinical practice and pharmaceutical marketing is often to generalize to the entire population all treatment benefits in the absence of definitive proof of harm," they write.

Dr McCoy added another possible factor: "I think we also need to recognize the discomfort that physicians may feel in telling patients that they may not benefit from intensive treatment. I think that we, as physicians, may worry that this comes across as us saying that patients don't have long to live or do not 'deserve' high-quality care."

And regulatory bodies should not necessarily always equate low HbA1c with high-quality care, she said, adding, "High-quality care is care that is efficient and safe, so quality metrics should incorporate indices of overtreatment and hypoglycemia in them."

"If physicians are 'graded' and reimbursed on the basis of how low their patients' HbA1c is, there is inevitable potential for seeking lower HbA1c at the expense of potential patient harm."

One Way to Simplify

The single-arm intervention study by Dr Munshi and colleagues recruited 65 adults with type 2 diabetes who were 65 years or older and had a mean HbA1c of 7.7%, taking two or more insulin injections a day (mean 3.7) and who had had hypoglycemia in the past 5 days detected by continuous glucose monitoring.

Mealtime insulin was stopped, and all were switched to once-daily insulin glargine, with or without noninsulin agents as needed, over 5 months. Hypoglycemia duration decreased at 5 and 8 months (P < .001) without any change in HbA1c levels after simplification.

Improvements in HbA1c occurred in patients with baseline HbA1c levels 8% to 9% (P < .001) and above 9% (P = .03), whereas there was a small worsening in those with baseline HbA1c levels below 7% (P = .03), and no change in those with HbA1c levels between 7% and 8% (P = .80).

Diabetes-related distress scores improved at 5 months and remained low at 8 months (P < .001).

In a short note, journal editor Deborah Grady, MD, of the Veterans Affairs Medical Center, San Francisco, California, points out that the study by Dr Munshi and colleagues is small and uncontrolled and therefore should be considered preliminary.

However, she wrote, "we decided to publish the study because we believe it should inspire larger trials to investigate optimum insulin regimens that minimize hypoglycemia and patient burden."

Dr McCoy has no relevant financial relationships; disclosures for the coauthors are listed in the article. Dr Munshi received an investigator-initiated grant from Sanofi and works as a consultant for Sanofi and Novo-Nordisk; disclosures for the coauthors are listed in the article. Drs Kerr and Hofer are leading a grant, funded by the Department of Veterans Affairs, on deintensification of medical services. Dr Grady has no relevant financial relationships.

For more diabetes and endocrinology news, follow us on Twitter and on Facebook.

JAMA Intern Med . Published online June 6, 2016. Abstract, Editorial, Research letter, Editor's note


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.