American Association for Clinical Chemistry
Better health through laboratory medicine
Clinical Laboratory Strategies: February 25, 2010

Strategies logo

Focusing Glycemic Control with HbA1c
Comparison of Subgroups Underscores Importance of Comorbidities
By Bill Malone

Clinicians have suspected for a long time that older, sicker diabetic patients don’t benefit as much from some interventions, including controversial intensive glucose control that keeps HbA1c below 7%. A new observational study confirms this suspicion and highlights the significance of paying attention to effects on subgroups that can get lost in clinical trials. This issue of Strategies examines its findings.

Clinicians have suspected for a long time that older, sicker diabetic patients don’t benefit as much from some interventions, including controversial intensive glucose control that keeps HbA1c below 7%. A new observational study confirms this suspicion and highlights the significance of paying attention to effects on subgroups that can get lost in clinical trials. This issue of Strategies examines its findings.

In just the last few years, three large, randomized, controlled trials have examined the association between tight glucose control and cardiovascular disease and found no benefit for people with diabetes who kept their HbA1c at 6.5% or lower. The trials include the ADVANCE Collaborative Group, the ACCORD study from the National Heart Lung and Blood Institute, and the Veterans Affairs Diabetes Trial. But despite these findings, many clinicians and researchers were not ready to give up on tight glycemic control. Within a year, two post hoc metaanalyses
that included data from the ADVANCE and ACCORD trials found evidence to suggest that more aggressive glycemic control might in fact be appropriate for younger patients and those without previous cardiovascular disease (CVD), redeeming the CVD benefits of tight glycemic control for at least some patients.

Now a recent observational study has brought new clarity to how researchers can make sense of the seemingly contradictory findings, as well as how clinicians can know when to push patients to achieve an HbA1c below 7% and when to hold back (Ann Intern Med 2009; 151: 854-860). Unlike previous trials and post hoc analysis, the new 5-year observational study took on the issue of comorbidities directly as part of the study design, seeking to discern whether this would be the key to determine which patients could benefit from tight glycemic
control and which could not.

In the study, 2,613 patients completed a questionnaire called the Total Illness Burden Index (TIBI), a novel tool previously developed by the authors that assesses the presence and severity of eight comorbid conditions and has been validated as a predictor of 3.5-year mortality and health-related quality of life. After statistical analysis, including multivariate Cox proportional hazards regression models, patients were divided into two subgroups depending on their level of comorbidity set at a TIBI score of 12. That number was taken from experience with TIBI in prior studies, but the authors also computed hazard ratios for both cardiovascular event risk and mortality risk with the TIBI score as the continuous, independent variable. This helped make sure that study results were not the artifact of the TIBI score cutpoint.

The study found that attaining an HbA1c level of 6.5% or less was associated with lower 5-year incidence of cardiovascular events in the low-comorbidity, healthier subgroup. For the sicker subgroup, achieving 6.5% or 7.0% did not predict fewer cardiovascular events. The message of the paper is straightforward, said the study's lead author, Sheldon Greenfield, MD, Donald
Bren professor of medicine and the executive director of the Health Policy Research Institute at the University of California, Irvine. “The key is that crucial subgroups respond differently to treatment because of their characteristics. This study adds to a body of literature that says, if a person has a lot of comorbidity, ease up on the HbA1c and focus on everything else, like
blood pressure and cholesterol,” he said. “In a person with low comorbidity— usually young people without much disease—they should give it a shot and try to bring their HbA1c down.”

Targeting Treatment

Making sure that only those patients who can benefit from an intervention get that intervention is especially important with a disease like diabetes, Greenfield emphasized. That’s one reason the paper explicitly focused on comorbidities from the outset. “It’s difficult for patients with comorbidities to achieve 6.5% or 7% HbA1c, because meanwhile they have a long list of medical conditions that also need attention,” he said. “Why make people crazy getting their HbA1c down if it doesn’t help much at all? A lot of these people with comorbidities are on
eight to ten medications a day, plus working very hard on diet and exercise.”

 Pushing HbA1c lower means more medication, and every new class of medication for a patient can lead to hypoglycemia, interactions with other drugs, and side effects. “Every time you take a medicine, it adds something good and something bad, and the bad can start to add up,” he said.

On the other hand, achieving lower HbA1c levels is reasonable for the mostly young, healthier group of patients who have a good chance of benefiting from the strategy down the road. According to Greenfield, the reason previous randomized controlled trials failed to show a distinction that tight glycemic control worked for some patients and not others has to do with so-called average effects. “If older patients with substantial comorbidity are less likely to benefit from intensive glycemic control and younger patients with less comorbidity are more likely to benefit, then the ‘average effect’ will be influenced by the proportion of study patients that represent each group,” the authors wrote.

This was the case with the previous three large trials, Greenfield said. While there were people who in the post-hoc analyses appeared to respond—mainly younger patients without heart disease—there were so few in each trial that their number was overwhelmed by the older, sicker participants who could not respond and thereby dragged the average effect to null, he explained.

The way the new study underscored the importance of differences between patients from the beginning is representative of a new wave of thought in research, according to Greenfield.

 “Unlike the olden days, when if somebody had a diagnosis they were pretty much in a homogenous group, now, with people living longer with more disease—with diabetes or any medical condition—two people can have the same diagnosis with one quite healthy and the next person dying,” he said. “The spectrum of patients with a given disease these days is so large, that inevitably there are many, many subgroups at differential risk. So with any disease, there will some who just will not respond no matter what you’re doing.”

Next Steps

Although the preponderance of evidence now leans toward the theory that tight glycemic control does work with younger, healthier people and that it is easier for them to achieve, an observational study such as Greenfield’s can’t draw lines of cause and effect. However, his experience as a co-chair of the Institute of Medicine report on comparative effectiveness research mandated by the American Recovery and Reinvestment Act (ARRA) last year has convinced Greenfield that randomized trials aren’t always necessary to reach good

“I think in the new world of observational studies with good databases and registries, there is now room for accumulating that kind of evidence that’s necessary without the traditional clinical trial,” he said. “You’d need a randomized trial more where you don’t know what the important variables are.

But in diabetes you know what they are and what the risks for confounders are because it’s been such a highly-studied field. So we can balance the groups pretty well.” While Greenfield admits there is still a lot of work to be done in the area of tight glucose control and CVD, it likely won’t take the form of a huge trial of the younger, healthier patients researchers believe can benefit. With data from his paper and others, professional organizations and other authorities may now feel more comfortable moving guidelines more in this direction.

Rate this page:       

Clinical Laboratory Strategies Podcast
Focusing Glycemic Control with HbA1c©