Effects of diabetes mellitus and ischemic heart disease on the progression from asymptomatic left ventricular dysfunction to symptomatic heart failure: A retrospective analysis from the Studies of Left Ventricular Dysfunction (SOLVD) Prevention Trial

Effects of diabetes mellitus and ischemic heart disease on the progression from asymptomatic left ventricular dysfunction to symptomatic heart failure: A retrospective analysis from the Studies of Left Ventricular Dysfunction (SOLVD) Prevention Trial Sandeep R. Das, MD, MPH,a Mark H. Drazner, MD, MSc,a Clyde W. Yancy, MD,a Lynne W. Stevenson, MD,b Bernard J. Gersh, MB, ChB, DPhil,c and Daniel L. Dries, MD, MPHa Dallas, Tex, Boston, Mass, and Rochester, Minn

Background Emerging data suggest that diabetes mellitus is a risk factor for the progression of established heart failure only in those patients with ischemic cardiomyopathy. Whether diabetes mellitus is a risk factor for the progression from asymptomatic left ventricular systolic dysfunction to symptomatic heart failure in patients with left ventricular dysfunction of an ischemic cause is not known. Methods

We performed a retrospective analysis of 2821 patients with asymptomatic left ventricular systolic dysfunction from the Studies of Left Ventricular Dysfunction (SOLVD) Prevention trial. We used adjusted survival analysis to examine the effects of ischemic heart disease and diabetes mellitus on 3 prespecified study end points: (1) development of heart failure (HF) symptoms, (2) HF hospitalization, and (3) death or development of symptoms.

Results There is a statistically significant interaction between the cause of left ventricular systolic dysfunction and diabetes mellitus on the risk of development of heart failure symptoms (P ⫽ .020). Patients with ischemic cardiomyopathy and diabetes had an increased risk of progression to symptomatic heart failure (HR ⫽ 1.56, P ⬍ .001), hospitalization for heart failure (HR ⫽ 2.16, P ⬍ .001), and death or development of symptoms (HR ⫽ 1.50, P ⬍ .001), compared with patients with ischemic cardiomyopathy without diabetes. In contrast, diabetes was not associated with an increased risk of reaching these end points in patients with nonischemic cardiomyopathy. Conclusions Diabetes mellitus is a risk factor for the progression from asymptomatic left ventricular systolic dysfunction to symptomatic heart failure, but this risk appears to be confined to those patients with ischemic cardiomyopathy. (Am Heart J 2004;148:883– 8.) In an earlier study,1 we showed that the established adverse effect of diabetes mellitus (DM) on mortality in the combined Studies of Left Ventricular Dysfunction (SOLVD) trials2 was confined to the subset of pa-

From the aDonald W. Reynolds Cardiovascular Clinical Research Center, Division of Cardiology, University of Texas Southwestern Medical School, Dallas, Tex; the bDivision of Cardiology, Brigham and Women’s Hospital, Boston, Mass; and the cDivision of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, Rochester, Minn. Submitted November 18, 2003; accepted April 10, 2004. Reprint requests: Daniel L. Dries, MD, MPH, Heart Failure/Transplant Program, Hospital of the University of Pennsylvania, 6 Penn Tower, 3400 Spruce Street, Philadelphia, PA 19104. E-mail: [email protected] 0002-8703/$ - see front matter © 2004, Elsevier Inc. All rights reserved. doi:10.1016/j.ahj.2004.04.019

tients with ischemic heart disease (IHD). Similar results have been reported for patients with established heart failure from the Beta-Blocker Evaluation of Survival Trial (BEST) trial.3 However, whether IHD and DM interact similarly with respect to the development of heart failure has not been examined. This population is of particular interest given the increasing emphasis now being placed on understanding patients with presymptomatic heart failure at risk for disease progression,4 as documented in the most recent American Heart Association/American College of Cardiology (AHA/ACC) guidelines.5 Therefore, we examined the development of heart failure in the subset of patients from the SOLVD Prevention trial that would fall into the modern stage B category, those with asymptomatic

American Heart Journal November 2004

884 Das et al

Table I. Baseline characteristics stratified by cause and presence of diabetes mellitus (n ⫽ 2818) Ischemic cardiomyopathy (n ⴝ 2431)

Age (y) LVEF (%) Serum creatinine (mg/dL) Length of follow-up (y) Sex Male Female Race/ethnicity White Nonwhite Tobacco use Current smoker Not current smoker Prior CABG Prior MI History of HTN Atrial fibrillation Medication Enalapril Antiplatelet agents Digoxin Diuretic ␤-Blocker CCB Warfarin

With DM (n ⴝ 367)

Without DM (n ⴝ 2064)

62.1 ⫾ 8.6 28.5 ⫾ 5.5 1.15 ⫾ 0.29 2.72 ⫾ 1.20

58.6 ⫾ 10.3 28.7 ⫾ 5.5 1.14 ⫾ 0.25 2.91 ⫾ 1.15

326 (88.8) 41 (11.2)

1882 (91.8) 182 (8.8)

290 (79.2) 76 (20.8)

1869 (90.6) 193 (9.4)

59 (16.1) 308 (83.9) 131 (35.9) 319 (86.9) 187 (51.0) 10 (2.9)

465 (22.5) 1599 (77.5) 718 (34.9) 1916 (92.8) 669 (32.4) 35 (1.8)

180 (49.1) 215 (58.7) 42 (11.4) 72 (19.6) 98 (26.7) 124 (33.8) 37 (10.1)

1028 (49.8) 1245 (60.4) 191 (9.3) 255 (12.4) 583 (28.3) 647 (31.4) 240 (11.6)

Nonischemic cardiomyopathy (n ⴝ 387) With DM (n ⴝ 50)

Without DM (n ⴝ 337)

59.7 ⫾ 9.8 27.8 ⫾ 5.8 1.16 ⫾ 0.42 2.89 ⫾ 1.22

55.6 ⫾ 12.9 27.0 ⫾ 6.1 1.12 ⫾ 0.25 2.77 ⫾ 1.20

41 (82.0) 9 (18.0)

283 (84.0) 54 (16.0)

27 (54.0) 23 (46.0)

239 (70.9) 98 (29.1)

.716 ⬍.001 ⬍.001 .181

9 (18.0) 41 (82.0) ... ... 35 (70.0) 6 (12.0)

94 (27.9) 243 (72.1) ... ... 123 (36.5) 43 (13.4)

... ... ⬍.001 .859

.788 .563 .189 ⬍.001 .541 .358 .390

23 (46.0) 9 (18.0) 11 (22.0) 22 (44.0) 12 (24.0) 9 (18.0) 3 (6.0)

167 (49.6) 54 (16.0) 72 (21.4) 84 (24.9) 17 (5.0) 26 (7.7) 58 (17.2)

.639 .724 .919 .005 ⬍.001 .018 .042

P value ⬍.001 .610 .473 .005 .150

⬍.001

P value .032 .400 .324 .533 .724

.016

.006

.140

ALVD, Asymptomatic left ventricular dysfunction; CCB, calcium channel blockers; CABG, coronary artery bypass graft; DM, diabetes mellitus; LVEF, left ventricular ejection fraction; HF, heart failure; HTN, hypertension; IHD, ischemic heart disease; MI, myocardial infarction. Data are presented as mean value ⫾ SD or number (percentage) of patients. P values represent results of Student t test for continuous variables or Pearson ␹2 for dichotomous variables, comparing patients with DM with those without DM in each group, according to cause.

left ventricular systolic dysfunction (ALVD). Specifically, we tested three hypotheses: (1) that IHD is associated with faster progression from ALVD to symptomatic HF, (2) that DM is associated with faster disease progression, and (3) that there is a qualitative interaction between DM and IHD with respect to disease progression.

Methods Patient population The current study uses data from the SOLVD Prevention trial, which has been described previously.6,7 Briefly, SOLVD Prevention was a randomized, double-blinded, placebocontrolled trial of the ACE inhibitor enalapril in a cohort of 4228 patients with left ventricular systolic dysfunction without diagnosed heart failure. Patients in SOLVD Prevention were required to have a left ventricular ejection fraction (LVEF) of ⬍35% and to have little to no limitation of exercise tolerance caused by dyspnea or fatigue. Patients could be taking cardiovascular medications, as long as the indication for use was not HF. We restricted our analysis to the 2821 patients from SOLVD Prevention who were classified as New

York Heart Association (NYHA) class I at the time of random assignment. The Institutional Review Board at the University of Texas Southwestern Medical School approved this study.

Classification of principal exposures and outcomes In SOLVD, patients were classified as having a primary cause of HF, based on the opinion of the individual site investigator into one of 3 categories: ischemic, nonischemic, and other. For the current study, a patient was also classified as having ischemic cardiomyopathy (ICM) if he or she had a prior myocardial infarction or coronary revascularization procedure. The presence of DM was established by patient selfreport at the SOLVD baseline interview. Three prespecified outcomes from the SOLVD Prevention study are examined in this report: (1) development of clinical HF, (2) hospitalization due to HF, and (3) death or development of HF. Three patients were missing data for either DM or IHD and were therefore excluded, leaving a cohort of 2818 patients for analysis.

Statistical analysis Bivariate comparisons were performed by means of 1-way analysis of variance and Pearson ␹2, as appropriate. Two-

American Heart Journal Volume 148, Number 5

Figure 1

Das et al 885

Figure 2

A, Kaplan-Meier estimates show cumulative incidence of heart failure symptoms over time for patients with nonischemic cardiomyopathy (NICM, solid line) and patients with ischemic cardiomyopathy (ICM, dashed line). B, Kaplan-Meier estimates show cumulative incidence of heart failure symptoms over time for patients with NICM (thick solid line), patients with ICM with comorbid diabetes mellitus (dashed line), and patients with ICM without diabetes mellitus (thin solid line).

A, Kaplan-Meier estimates show cumulative incidence of heart failure hospitalization over time for patients with nonischemic cardiomyopathy (NICM, solid line) and patients with ischemic cardiomyopathy (ICM, dashed line). B, Unadjusted Kaplan-Meier estimates show cumulative incidence of heart failure hospitalization over time for patients with NICM (thick solid line), patients with ICM with comorbid diabetes mellitus (dashed line), and patients with ICM without diabetes mellitus (thin solid line).

sided probability values ⬍.05 were considered statistically significant. Cumulative survival curves were constructed by Kaplan-Meier methods with differences between curves tested for statistical significance by the log-rank statistic.8,9 Adjusted Cox proportional hazards estimates were used to determine the effect of diabetes stratified by underlying cause for each of the three prespecified end points.10 The following characteristics, based on clinical significance and on the baseline evaluation, were considered as potential confounders: age, LVEF, serum creatinine, sex, ethnicity, tobacco use, history of hypertension, history of atrial fibrillation, and baseline medication use. Medications considered in multivariate modeling were enalapril, antiplatelet agents, digoxin, diuretics, ␤-adrenergic antagonists, calcium channel blockers, and warfarin. Age, EF, and creatinine were modeled as continuous variables; the rest were modeled as dichotomous variables. Ethnicity was classified on the basis of patient self-report at the SOLVD baseline interview as white or nonwhite, the lat-

ter group consisting primarily of patients who described themselves as African American (black). For the purposes of this study, ethnicity is considered solely as a covariate and probably conflates several potential confounders.11 Prior myocardial infarction and prior revascularization were not included in the model because they were part of our definition of IHD. The final Cox model contained the following covariates: age, EF, serum creatinine, ethnicity, use of antiplatelet agents, use of ␤-adrenergic antagonists, and random assignment to enalapril. None of these covariates were deemed to violate the proportional hazards assumption. The Stata 7.0 statistical software package (Stata Corp, College Station, Tex, 2001) was used for all analyses.

Results Baseline characteristics As shown in Table I, our study population was generally white and male. There was a high prevalence of

American Heart Journal November 2004

886 Das et al

Table II. Relative risk associated with diabetes mellitus, stratified by cause of heart failure ICM (n ⴝ 2431)

NICM (n ⴝ 387)

Outcome

HR

(95% CI)

P

HR

(95% CI)

P

Development of HF symptoms First hospitalization for HF Death or symptomatic HF

1.52 2.10 1.47

(1.22–1.89) (1.53–2.89) (1.22–1.77)

⬍.001 ⬍.001 ⬍.001

0.70 0.70 0.68

(0.39–1.28) (0.31–1.61) (0.40–1.16)

.247 .401 .156

CI, Confidence interval; HF, heart failure; HR, hazard ratio; ICM, ischemic cardiomyopathy; NICM, nonischemic cardiomyopathy. Hazard ratios are derived from Cox proportional hazards estimates adjusted for age, ejection fraction, serum creatinine, ethnicity, and medication use at baseline (antiplatelet agents; ␤-adrenergic antagonists, enalapril). P values are for log-rank test comparing patients with and without diabetes mellitus in each group, according to cause.

Figure 3

SOLVD. Roughly one fourth of these patients were taking a ␤-adrenergic antagonist and, by design, half were randomly assigned to take enalapril.

Outcomes Figure 1, Figure 2, and Figure 3 show the unadjusted Kaplan-Meier survival curves for each of the considered end points. For each end point, patients with ICM were at lower risk of disease progression compared with patients with nonischemic cardiomyopathy (Figures 1A, 2A, and 3A). Stratifying patients with ICM further by DM (Figures 1B, 2B, and 3B), it now becomes apparent that the low-risk group consists of those patients with ICM without DM; the risk of disease progression for patients with ICM and DM was similar to that for patients with nonischemic cardiomyopathy. These results were consistent across all three considered end points. Adjusted Cox proportional hazards estimates showing the interaction between DM and underlying cause are shown in Table II. Among patients with nonischemic cardiomyopathy, differences in outcomes for patients with and without DM were not statistically significant. In contrast, among patients with ICM, DM was associated with an increased risk of disease progression for each outcome.

Discussion A, Kaplan-Meier estimates show cumulative incidence of death or development of heart failure symptoms over time for patients with nonischemic cardiomyopathy (NICM, solid line) and patients with ischemic cardiomyopathy (ICM, dashed line). B, Unadjusted Kaplan-Meier estimates show cumulative incidence of death or development of heart failure symptoms over time for patients with NICM (thick solid line), patients with ICM with comorbid diabetes mellitus (dashed line), and patients with ICM without diabetes mellitus (thin solid line).

IHD; more than two thirds of patients had a history of myocardial infarction. The mean EF at baseline was low, owing to the requirements for inclusion in

Heart failure (HF) has traditionally been a clinical diagnosis, based on signs and symptoms of circulatory insufficiency in the form of volume overload or target organ hypoperfusion.12 Detailed information on the natural history of ALVD is currently lacking, but further inquiry is warranted, as there is evidence that treatment of at least a subset of these patients improves outcomes.4,6,13 Ischemic heart disease has long been associated with deleterious structural changes in the heart and with progression of left ventricular dysfunction.14 IHD is an independent risk factor for incident HF and has become the primary cause of HF in the United States.12,15–17 In patients with existing HF, IHD is asso-

American Heart Journal Volume 148, Number 5

ciated with worse prognosis, in proportion to the angiographic severity of disease.14,18 –21 DM is also associated with adverse cardiac structural effects and with increased risk for incident HF.12,15,17,22–24 The current study extends previous findings1,3 by demonstrating a qualitative interaction between IHD and DM in the risk for progression from ALVD to symptomatic HF. In this population, DM was associated with an increased risk for development of HF only in those patients with IHD. The pathophysiologic basis for this interaction has not been established. However, alterations in cardiac structure and function attributable to DM have been reported.22,23,25 Specifically, insulin is a coronary vasodilator that plays an important role in cardiac signaling; it is possible that direct vascular effects or impaired signaling may potentiate left ventricular dysfunction.23 Further potential for adverse synergy between DM and IHD may result from alterations in myocyte substrate utilization. Ischemia results in a substrate shift toward glycolytic pathways, and ischemic myocardium may therefore be more susceptible to the adverse effects of DM on myocardial glucose utilization.25 Patients with nonischemic cardiomyopathy were at increased risk for the development of HF compared with patients with ICM who did not have DM. The increased risk of progression to HF in patients with ALVD caused nonischemic cardiomyopathy is at odds with data in patients with symptomatic HF.14,17,18 The basis of this observation is not entirely clear and may be due to chance involved in subgroup analysis. Alternatively, if patients were screened for entry into SOLVD because of a recent history of myocardial infarction, patients with ICM may have been identified at an earlier point in their disease progression compared with patients with nonischemic cardiomyopathy. Nevertheless, we recognize that this question needs to be tested in other large data sets of patients with ALVD.

Limitations This is a retrospective study, involving a population that was generally white and male; most had ICM and a history of myocardial infarction. There were a small number of patients with nonischemic cardiomyopathy without DM, so inferences for that subgroup are sharply limited. There were limitations in measuring the exposures; no attempt at invasive validation was performed to validate the classification of ischemic versus nonischemic cause. Misclassification in this case would bias the result toward the null. Presence of DM was determined by patient self-report, and a high rate of undiagnosed or unreported diabetes would also bias the result toward the null. The multivariate model was limited by what data were available for the SOLVD patients. Other potential confounders are likely to exist

Das et al 887

that were not available for inclusion in the model. Finally, the SOLVD Prevention trial was published more than 10 years ago and thus reflects clinical practice of that time. Baseline rates of disease progression in a modern cohort may well differ.

Conclusions These findings demonstrate a qualitative interaction between DM and the cause of ALVD on the risk for progression to symptomatic HF. In patients with IHD, DM had an adverse impact on progression from ALVD to symptomatic HF. In patients with nonischemic cardiomyopathy, however, DM had no statistically significant effect on disease progression. The pathophysiologic basis for this interaction between diabetes and underlying cause warrants further investigation.

References 1. Dries DL, Sweitzer NK, Drazner MH, et al. Prognostic impact of diabetes mellitus in patients with heart failure according to the etiology of left ventricular systolic dysfunction. J Am Coll Cardiol 2001;38:421– 8. 2. Shindler DM, Kostis JB, Yusuf S, et al. Diabetes mellitus, a predictor of morbidity and mortality in the studies of left ventricular dysfunction (SOLVD) trials and registry. Am J Cardiol 1996;77:1017– 20. 3. Domanski M, Krause-Steinrauf H, Deedwania P, et al. The effect of diabetes on outcomes of patients with advanced heart failure in the BEST trial. J Am Coll Cardiol 2003;42:914 –22. 4. Wang TJ, Levy D, Benjamin EJ, Vasan RS. The epidemiology of “asymptomatic” left ventricular systolic dysfunction: implications for screening. Ann Intern Med 2003;138:907–16. 5. Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2001;104:2996 –3007. 6. SOLVD Investigators. Effect of enalapril on mortality and the development of heart failure in asymptomatic patients with reduced left ventricular ejection fractions. N Engl J Med 1992;327:685–91. 7. Johnstone D, Limacher M, Rousseau M, et al. Clinical characteristics of patients in Studies of Left Ventricular Dysfunction (SOLVD). Am J Cardiol 1992;70:894 –900. 8. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457– 81. 9. Lee TE. Nonparametric methods of estimating survival functions: statistical methods for survival data analysis. Belmont, Calif: Lifetime Learning Publications; 1980. 75– 87. 10. Cox DR. Regression models and life-tables. J R Stat Soc [B] 1972; 34:187–220. 11. Kaplan JB, Bennett T. Use of race and ethnicity in biomedical publication. JAMA 2003;289:2709 –16. 12. Adams KF. New epidemiologic perspectives concerning mild-tomoderate heart failure. Am J Med 2001;110:6S–12S. 13. Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction. N Engl J Med 1992;327:669 –77.

888 Das et al

14. Bart BA, Shaw LK, McCants CB, et al. Clinical determinants of mortality in patients with angiographically diagnosed ischemic or nonischemic cardiomyopathy. J Am Coll Cardiol 1997;30:1002– 8. 15. Nichols GA, Erbey JR, Hillier TA, Brown JB. Congestive heart failure in type 2 diabetes. Diabetes Care 2001;24:1614 –9. 16. Wood DA. Preventing clinical heart failure: the rationale and scientific evidence. Heart 2002;88(Suppl II):II-15–22. 17. He J, Ogden LG, Bazzano LA, et al. Risk factors for congestive heart failure in US men and women. Arch Intern Med 2001;161: 996 –1002. 18. Gheorghiade M, Bonow RO. Chronic heart failure in the United States: a manifestation of coronary artery disease. Circulation 1998;97:282–9. 19. Likoff MJ, Chandler SL, Kay HR. Clinical determinants of mortality in chronic congestive heart failure secondary to idiopathic dilated or to ischemic cardiomyopathy. Am J Cardiol 1987;59:634 – 8.

American Heart Journal November 2004

20. Dei Cas L, Metra M, Nodari S, et al. Prevention and management of chronic heart failure in patients at risk. Am J Cardiol 2003; 91(Suppl):10F–7F. 21. Follath F, Cleland JGF, Klein W, et al. Etiology and response to drug treatment in heart failure. J Am Coll Cardiol 1998;32:1167–72. 22. Devereux RB, Roman MJ, Paranicas M, et al. Impact of diabetes on cardiac structure and function: the Strong Heart Study. Circulation 2000;101:2271– 6. 23. Taegtmeyer H, McNulty P, Young ME. Adaptation and maladaptation of the heart in diabetes, I: general concepts. Circulation 2002; 105:1727–33. 24. McKelvie RS, Benedict CR, Yusuf S. Prevention of congestive heart failure and management of asymptomatic left ventricular dysfunction. BMJ 1999;318:1400 –2. 25. Young ME, McNulty P, Taegtmeyer H. Adaptation and maladaptation of the heart in diabetes, II: potential mechanisms. Circulation 2002;105:1861–70.