Glycemia Levels Prognostic in Nondiabetics With Heart Failure: A New HF Treatment Target?

November 21, 2006

November 21, 2006 (Chicago IL) – Two common measures of insulin resistance are associated with poor outcomes in a "dose-response" fashion in nondiabetic patients with heart failure, suggested two studies presented at last week's American Heart Association 2006 Scientific Sessions [1,2].

In one of them, an analysis based on the combined interim results from two prospective randomized trials, rates of HF hospitalization over more than two years rose in tandem with baseline fasting plasma glucose levels in a "high-risk" population of >30 000 diabetic and nondiabetic patients with vascular disease. The association was significant even after adjustment for medical therapies, diabetic status, and other components of the metabolic syndrome, reported Dr Claes Held (Karolinska University Hospital, Stockholm, Sweden).

Similarly, the two-year rate of cardiac death went up significantly and independently with baseline levels of glycated hemoglobin (HbA1 C) in a prospective study of several hundred patients with systolic HF. As reported by Dr Nicolas Lamblin (Centre Hospitalier Regional Universitaire de Lille, France), baseline HbA1 C levels also seemed to predict severity of HF and other morbidity measures.

"These studies provide further evidence that insulin resistance plays an important role in the risk of developing heart failure and the risk of mortality once heart failure develops," Dr Gregg C Fonarow (University of California, Los Angeles Medical Center), who wasn't a coauthor of either report, told heartwire . "These studies have important implications for how physicians recognize and manage patients with or at risk for heart failure."

Definitions of diabetes are based on degree of glycemia, but patients with glycemic indices below the arbitrary diagnostic thresholds can still be at increased risk for worsening heart failure, according to Held. "I look at glucose levels as I do cholesterol or blood pressure. It seems to be a continuous risk variable," he told heartwire . "The lower the better, to a certain limit, of course."

No one is claiming, based on current evidence, that either fasting glucose or HbA1 C is a viable target for therapy of heart failure specifically; that would have to be established in prospective, randomized trials, all three researchers emphasized. But both new studies are consistent with research in other populations suggesting that insulin resistance is closely tied to HF progression [3,4].

For example, both studies are consistent with an analysis from the randomized Heart Outcomes Prevention Evaluation (HOPE) trial in which the risks of CV events, HF, death, and clinically evident nephropathy were each independently associated with rising levels of HbA1 C among diabetic patients. In the same study, a mixed cohort of diabetics and nondiabetics showed similar relationships between fasting plasma glucose and the same set of outcomes. Observational studies have long suggested that diabetes is common among patients with HF and may contribute to the syndrome's progression.

Glucose levels and risk of new-onset HF

Held's report was based on a prospectively planned, blinded interim analysis of patients initially free of HF enrolled in two ongoing "parallel" randomized trials, one ( ONTARGET) comparing telmisartan, ramipril, and placebo and the other ( TRANSCEND) looking at telmisartan vs placebo in patients intolerant of ACE inhibitors. Entry to the trials required presence of CAD, peripheral artery disease, cerebrovascular disease, or diabetes with end-organ damage.

Over a mean follow-up of 2.4 years, the hazard ratio for HF hospitalization climbed 5% (95% CI 1.02-1.08, p<0.001) for each 1-mmol/L increment in the baseline level of fasting plasma glucose, independent of diabetic status. When outcomes were analyzed by glucose quartiles, the risk went up significantly even at levels within the "normal" range and was more pronounced among patients with established diabetes compared with those with diabetes diagnosed at baseline.

Relationship between baseline fasting plasma glucose (by quartiles) and risk of HF hospitalization (mean follow-up, 2.4 years)
<5.0 mmol/L, n=5136
5.0-5.5 mmol/L, n=6298
5.6-6.9 mmol/L, n=6650
>7.0 mmol/L, n=1006 (new DM)
>7.0 mmol/L, n=11 708 (known DM)
HF hospitalization (%)
HR (95% CI)*
1.07 (0.78–1.46)
1.24 (0.92–1.68)
1.74 (1.09–2.76)
2.59 (1.99–3.36)
*Adjusted for age, sex, smoking, prior MI, hypertension, diabetes, creatinine, waist-to-hip ratio, and use of aspirin, beta blockers, or statins. p for trend across quartiles, levels, <0.0001. DM=diabetes mellitus.

The findings suggest that fasting plasma glucose independently predicts HF hospitalization and that "the degree of dysglycemia is the key determinant of this relationship, although they do not prove causality," Held said during his presentation.

Interviewed, Held said other data from the same patients suggest a similar link between baseline fasting plasma glucose levels and the composite rate of death, MI, or stroke. "So it's not just heart failure, but the strongest relationship is with heart failure."

HbA1C levels and mortality in the HF population

Baseline levels of HbA1 C were inversely related to LV systolic function and rose with increasing ventricular dimensions and greater HF morbidity in a cohort of consecutive nondiabetic patients with heart failure and an LVEF <45%. The group's two-year rate of cardiac death also went up with increasing HbA1 C levels, Lamblin reported. In this analysis, cardiac death included urgent cardiac transplantation.

Baseline clinical and echocardiographic parameters* plus clinical outcome (median follow-up, 2.1 years) by HbA1C tertiles
HbA1C <5.5%, n=123
HbA1C 5.5-5.9%, n=133
HbA1C >5.9%, n=121
p for trend
NYHA class 3-4* (%)
LVEF* (%)
LVEDD* (mm)
Atrial fibrillation* (%)
Cardiac deaths (%)
*Across tertiles, patients were similar with respect to distribution of age, sex, body-mass index, smoking, hypertension, hypercholesterolemia, triglycerides, hemoglobin, creatinine clearance, and prevalence of ischemic HF. LVEDD=left ventricular end-diastolic diameter.

"Our study shows for the first time that HbA1 C levels are associated with the severity and prognosis of heart failure in nondiabetic persons," according to Lamblin. He said that HbA1 C may not be very useful as a prognostic marker in ongoing heart failure, noting that his group's results--while controlled for other components of the metabolic syndrome--were not adjusted for brain-type natriuretic peptide (BNP) levels or other established HF prognostic markers.

Fonarow noted that the study's "minor elevations" in HbA1 C were "strongly linked to increased heart-failure severity. So whether this is merely another indicator of heart-failure severity or providing a truly independent mediator needs further study. Furthermore, knowing how best to lower this risk requires further studies, as trials of glycemic-control medications have generally excluded patients with preexisting heart failure."

  1. Held C, Gerstein HC, Zhao F, et al. Fasting plasma glucose is an independent predictor of hospitalization for congestive heart failure in high-risk patients. American Heart Association 2006 Scientific Sessions; November 13, 2006. Abstract 2562.

  2. Lamblin N, Bauters C. Hemoglobin A1c levels are associated with severity and prognosis of systolic chronic heart failure in nondiabetic patients. American Heart Association 2006 Scientific Sessions; November 13, 2006. Abstract 2372.

  3. Gerstein HC, Pogue J, Mann JF, et al. The relationship between dysglycemia and cardiovascular and renal risk in diabetic and non-diabetic participants in the HOPE study: Aprospective epidemiological analysis. Diabetologia 2005; 48:1749-1755.

  4. Barsheshet A, Garty M, Grossman E, et al. Admission blood glucose level and mortality among hospitalized nondiabetic patients with heart failure. Arch Intern Med 2006; 166:1613-1619.

The complete contents of Heartwire , a professional news service of WebMD, can be found at, a Web site for cardiovascular healthcare professionals.


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.