We showed that impaired insulin sensitivity in patients with chronic heart failure (CHF) significantly predicted impaired survival. Insulin resistance relates to an advanced disease state and higher mortality independent of body composition and established prognosticators, implicating a pathophysiologic role for insulin sensitivity in CHF. Patients with CHF may possibly benefit from early treatment of impaired insulin sensitivity, but further research is needed.
A major cause of death and disability today is chronic heart failure (CHF). With the expansion of the CHF pathophysiologic concepts beyond mere hemodynamic aspects, hormonal and immune derangements and, more recently, metabolic impairment have been observed to contribute to symptomatic status and disease progression. Insulin sensitivity has been found to occur in CHF independently of ischemic etiology.1 A mutual relationship between CHF and impaired glucose metabolism has been observed, with studies indicating that diabetes mellitus is a predisposing factor for CHF2,3 and CHF patients are at higher risk for developing diabetes mellitus.1,4
The clinical significance of insulin resistance in patients with CHF is not fully known. Insulin resistance is, however, the underlying mechanism for type 2 diabetes mellitus, suggesting pathophysiologic relevance in CHF. Recent studies have shown that insulin resistance relates to CHF severity and key clinical symptoms, such as impaired exercise capacity, early muscle fatigue, and dyspnea.1,5 Whether insulin resistance in CHF is also of prognostic significance in CHF has not been studied.
We evaluated whether the presence of insulin resistance in patients with CHF predicted a worse prognosis. The specific role of regional fat and lean tissue distribution in relation to insulin sensitivity was taken into account.
Patients and methods
Patients with ischemic or nonischemic CHF were enrolled in this longitudinal study (n = 105). Thirty-seven percent of patients had nonischemic CHF, and 63% had ischemic CHF. All patients had systolic left ventricular impairment, with clinical evidence and at least a 6-month history of symptomatic heart failure. At the time patients were evaluated for insulin sensitivity, they were in stable, compensated condition. No patient had diabetes or received antidiabetic treatment. All patients enrolled in the study were receiving medical treatment for CHF in individually optimized regimens. Ethics approval and informed consent were obtained appropriately.
Insulin sensitivity was assessed using minimal modeling analysis of glucose and insulin profiles during an intravenous glucose tolerance test.6 Body composition analysis was done by dual-energy x-ray absorptiometry in a subset of patients, and total and regional distribution of lean and fat tissue was analyzed. Data transformation (log and square root) was used as appropriate to permit parametric statistical analysis. To compare differences in median values between groups, unpaired Student t test was used. For survival analysis of continuous variables Cox proportional hazard analysis in univariate and stepwise multivariate analysis was used.
Over the follow-up period (mean ± standard error of the mean [SEM], 44 ± 4 months), 50% of patients (n = 53) died at a median of 527 days (range, 4-3319 days). For the 52 patients who lived, the follow-up period ranged from 396 to 3719 days (median: 1474 days; mean: 1905 ± 153 days). After 1, 2, and 3 years, cumulative mortality was 22% (95% confidence interval [CI], 14%-30%), 28%, and 40%, respectively.
There was no difference in age between survivors and nonsurvivors (60 ± 2 years vs 64 ± 1 years, respectively). There also was no difference between the 2 groups regarding left ventricular ejection fraction (LVEF; 31% ± 3% vs 26% ± 2%, respectively) and etiology distribution (63% vs 62%, respectively, for those with ischemic cardiomyopathy and 37% vs 38%, respectively, for those with dilated cardiomyopathy). Clinical characteristics of CHF were, however, more advanced in non-surviving patients (higher mean New York Heart As­sociation [NYHA] class [+0.5] and lower peak VO2 [-21%, both P < .004], and lower systolic blood pressure, P < .05).
Insulin sensitivity and body composition parameters for both groups are shown in the Table. Nonsurviving patients had lower insulin sensitivity than survivors (-42%, P = .002). This was paralleled by nonsignificant differences in total and regional fat tissue between survivors and nonsurvivors, showing a trend toward lower total and peripheral fat tissue in nonsurvivors. Total and peripheral lean tissue were significantly lower in nonsurviving compared with surviving patients.
Insulin sensitivity for CHF patients was 2.53 ± 0.26 min-1 • µU • mL-1 • 104 compared with 3.5 min-1 • µU • mL-1 • 104 for healthy control subjects of similar age (reference data from our laboratory). Insulin sensitivity was reduced in a stepwise fashion with advancing CHF severity, as indicated by NYHA functional class (Figure 1; analysis of variance, P = .0007). In linear regression analysis, insulin sensitivity correlated with LVEF (R = 0.36), peak VO2 (R = 0.23), BMI (R = -0.22), and total (R = -0.23) and regional fat mass (R = -0.27; all P < .05). No correlation was found for insulin sensitivity versus age, blood pressure, lean tissue mass, serum cholesterol, and creatinine (all P > .25).
Insulin sensitivity was not significantly different between patients treated with beta blockers and those without (P = .13). Similarly, no difference was found for insulin sensitivity between patients with and without ACE inhibitor (P= .38) and diuretic treatment (P = .49).
Higher insulin sensitivity was predictive of lower mortality, as shown by univariate Cox proportional hazards analysis (relative risk [RR], 0.56; 95% CI, 0.35-0.89; P = .015). Low insulin sensitivity predicted impaired survival also after adjustment for parameters of body composition such as BMI, total amount of fat tissue, and regional fat distribution (P < .01). When adjusted for the use of beta blockers, ACE inhibitors, and diuretics, insulin sensitivity remained a significant predictor of mortality (P = .008) with only diuretic treatment significantly contributing to prognosis (P = .018; ACE inhibitor and beta blocker P > .3). Other prognostic markers were age, NYHA class, and peak VO2, serum uric acid (all P < .01), hemoglobin (P = .018), fat tissue mass and lean tissue mass (both P < .05), but not LVEF and BMI.
Insulin sensitivity was shown to be a significant predictor of impaired survival on multivariate analysis independent of clinical variables, which included age, NYHA class, peak VO2, mean arterial pressure, uric acid, hemoglobin, cholesterol, sodium, creatinine, total fat tissue mass, lean tissue mass, and diuretic treatment (RR, 0.30; 95% CI, 0.14-0.63; P = .002). Insulin sensitivity was also shown to be a predictor independent of clinical, biochemical, body composition, and treatment parameters using varying stepwise Cox proportional hazards models.
Patients with insulin sensitivity be­low the median had worse survival than patients with insulin sensitivity above the median (1.82 min-1 • µU • mL-1 • 104), as shown by analysis of insulin sensitivity as a dichotomized variable (RR, 0.38; 95% CI, 0.21-0.67; P = .001). For patients with insulin sensitivity above the median, survival was 83%, 76%, and 73% at 2, 3, and 4 years, respectively. Corresponding data for patients with insulin sensitivity below the median were 61%, 44%, and 37% at 2, 3, and 4 years, respectively (Figure 2).
Our results indicate that for patients with CHF, impaired insulin sensitivity is an independent marker of impaired prognosis. In a prior study, insulin resistance had been shown to occur in CHF patients independently of ischemic heart disease.1 The current study extends previous data showing that insulin resistance increases in proportion to the severity of CHF and relates to major clinical symptoms, such as LVEF and peak VO2. Insulin sensitivity occurs in CHF independently of body composition measures and, specifically, of fat tissue distribution.
Because insulin resistance is a characteristic of type 2 diabetes as well as a precursor to it, data on CHF and type 2 diabetes may be helpful because it may indirectly supply information about the clinical importance of insulin resistance in patients with CHF. There is a high prevalence of diabetes mellitus in patients with CHF (20%-25%). Compared with patients without diabetes, men and women with diabetes have a 2.4-fold to 5.1-fold greater risk of developing CHF, as was shown in the Framingham study.7 Large trials have corroborated this relationship.2,3 This reciprocal relationship indicates that impaired insulin sensitivity as a feature of metabolic imbalance is a major contributor to the pathophysiology of heart failure syndrome.
It is important to note that type 2 diabetes mellitus is only the late consequence of enduring insulin resistance, with the latter preceding frank diabetes for years if not decades. Therefore, the prevalence of insulin resistance in CHF may be considerably higher than the prevalence of type 2 diabetes. Accordingly, an analysis from the RESOLVD study has shown that, based on fasting blood glucose criteria from a single point assessment, 43% of the patients had abnormal glucose metabolism.5
Because obesity is a common confounding factor in type 2 diabetes, impaired insulin sensitivity has been recognized to be associated with in­creased BMI. Not surprisingly, our results showed that insulin sensitivity was associated with fat tissue mass and BMI among the overall study population. But when subgrouping patients, nonsurviving patients were similar to surviving patients with regard to fat tissue mass and distribution, despite significantly lower insulin sensitivity (-42%). In fact, the nonsurviving patients (those with greater insulin resistance) tended to have lower fat tissue mass. This corroborates earlier studies showing that CHF patients with higher BMI had lower mortality.8,9 Impaired insulin sensitivity predicted increased mortality independently of parameters of body composition such as weight, BMI, and total and regional fat and lean tissue mass in our study. These results seem to be contrary to the anticipated correlation between body composition and insulin metabolism. It may be concluded that in CHF, impaired insulin sensitivity is not merely a function of adiposity but rather develops secondary to the underlying chronic disease itself and may therefore indeed have implications in the pathophysiology of CHF. These results further indicate that disease-related factors causing impaired hormonal and metabolic balance might supersede the physiologic feedback regulation between metabolic composition and body composition.
The principal mechanisms of in­sulin resistance in CHF are not entirely clear. A combination of elements, including increased immune and neuroendocrine activation, lower GLUT4 transport protein amount in skeletal muscle,10 impaired oxidative metabolism, and decreased peripheral tissue perfusion, probably plays a role. A decrease in exercise and changes in the diet may also be factors. Metabolic balance may be affected by treatment with ACE inhibitors and beta-blocking agents. Negative effects on metabolic control have been reported for some beta-blockers, whereas for others, such as carvedilol, neutral11 or beneficial12 effects on insulin sensitivity were observed. Insulin sensitivity was independent of treatment with beta blockers, ACE inhibitors, or diuretics. In our study, and in survival analysis, the significant predictive power of insulin sensitivity remained unchanged when adjusted for treatment.
Patients with CHF are usually not tested for insulin sensitivity, and our goal in performing this study was not to establish insulin sensitivity as yet another marker for prognostic evaluation in patients with CHF. This study does highlight, however, the significance of hormonal imbalance and metabolic derangement in CHF, which has only recently been receiving increased notice as an important as­pect of CHF pathophysiology.
Based on the results of this study, additional research may help to de­termine whether early detection and treatment of in­sulin resistance in patients with CHF will improve their clinical status and potential outcome. Long-term studies showing that exercise decreased mortality indirectly support this approach.13 The benefits of exercise for patients with CHF may be related to the well-recognized favorable effects of exercise on insulin sensitivity. It is not clear, however, whether drug treatment directed at improving in­sulin sensitivity will result in advantages for CHF patients. Be­cause insulin sensitizers, such as thiazolidinediones, intensify retention of fluids and possibly may increase edema, their use in CHF patients is controversial. Intervention studies are required to test whether targeting impaired insulin sensitivity may be of benefit in patients with CHF, but need to be done with great care and attention to potential adverse effects.