Influence of Diabetes on Characteristics and Outcomes in Patients Hospitalized With Heart Failure: A Report From the Organized Program to Initiate Lifesaving Treatment in Hospitalized HF Patients

Barry H. Greenberg, MD; William T. Abraham, MD; Nancy M. Albert, PhD, RN; Karen Chiswell, MS; Robert Clare, MS; Wendy Gattis Stough, PharmD; Mihai Gheorghiade, MD; Christopher M. O'Connor, MD; Jie Lena Sun, MS; Clyde W. Yancy, MD; James B. Young, MD; Gregg C. Fonarow, MD

Am Heart J.  2007;154(4):647-654.  ?2007 Mosby, Inc.
Posted 10/24/2007

Abstract and Introduction

Abstract

Background: Diabetes, a common comorbidity in patients with heart failure (HF), is associated with worse long-term outcomes in patients with HF due to systolic dysfunction. Whether diabetes mellitus (DM) influences characteristics and outcomes in patients hospitalized with HF has not been well studied.
Methods: The Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure is a patient registry and performance-improvement program for patients hospitalized with HF that included a prespecified 10% subgroup with 60- to 90-day follow-up data. Data were analyzed as DM compared with no DM. Pearson m2 test for categorical variables and t test for continuous variables were used, as was a multivariable analysis that included a stepwise Cox proportional hazard model.
Results: Among 48,612 patients from 259 hospitals, 42% had DM. Heart failure patients with DM tended to be younger, with greater likelihood of ischemic etiology, and higher serum creatinine levels. Heart failure patients with DM received quality care measures to a similar degree, with a few modest exceptions. No differences in in-hospital mortality were observed, but HF patients with DM experienced modestly longer length of stay (5.9 vs 5.5 days for nondiabetic patients; P b.0001). In the 5791 patients in the follow-up cohort, patients with DM (n = 2464) had similar postdischarge mortality but increased all-cause rehospitalization (31.5% vs 28.2% for nondiabetic patients; P = .006). Multivariable analysis showed that DM was not an independent predictor of in-hospital (odds ratio, 1.00; 95% confidence interval, 0.88-1.14; P = .99) or follow-up mortality (hazard ratio, 1.08; 95% confidence interval, 0.87-1.35; P = .48).
Conclusions: The Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure data reveal a high prevalence of DM in patients hospitalized with HF. Heat failure patients with DM received similar quality of care and experienced similar short-term mortality compared with patients without DM but had higher risk of rehospitalization.

Introduction

Diabetes mellitus (DM) is a common comorbid condition of heart failure (HF),[1] and the prevalence of HF among diabetic patients is 2 to 3 times that of agematched controls.[2,3] Among the selected group of patients enrolled in chronic systolic HF clinical trials, generally 20% to 30% have DM, with a similar observed percentage (28%) for patients with preserved systolic HF investigated in the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM)-Preserved trial.[4-9] Although these patients sometimes have worse long-term outcomes than their nondiabetic counterparts, analyses from several major clinical studies have shown that the reduction in HF mortality and morbidity observed in treatment with angiotensin-converting enzyme inhibitors (ACEI), aldosterone antagonists, and β-blockers extends to diabetic patients.[4,5,7,8,10-12]

Currently, little is known about the prevalence of DM and the effect of its presence on the characteristics, treatment, and outcomes in patients hospitalized for HF. In addition, very little is known about how well guideline-recommended therapies are being used in patients with HF with DM. Although β-blockers are recommended by national guidelines for use in diabetic patients with HF,[13,14] they have been underutilized in this population due to perceived negative metabolic effects.[15,16] Analyses of large patient databases from all regions of the country are critical in providing insight into whether optimal evidence-based therapies are used in HF patients with DM.

Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) is a registry and performance-improvement program for patients hospitalized with HF. This analysis of the OPTIMIZE-HF patient registry data aims to document the prevalence of DM in patients hospitalized for HF; assess the associations between DM and the characteristics, quality of care, treatment, and outcomes in these patients; and improve the understanding of the influence of DM on treatment and outcomes in patients hospitalized for worsening HF with reduced or preserved systolic function.

Methods

The OPTIMIZE-HF was a comprehensive hospital-based registry and process-of-case improvement program designed to provide optimal medical care and education to patients hospitalized for HF. The OPTIMIZE-HF program has been described in detail elsewhere[17-19] and will be briefly summarized below.

OPTIMIZE-HF Patient Eligibility

Eligible patients were hospitalized for episodes of newonset or worsening HF as the primary cause of admission or had significant HF symptoms that developed during hospitalization where HF was the primary discharge diagnosis.[17] Left ventricular systolic dysfunction (LVSD) was defined as a documented left ventricular ejection fraction b40% or qualitative assessment of moderate/severe dysfunction. Diabetes and insulin use was assessed by past medical and medication history as documented in the medical record.

From March 2003 to December 31, 2004, 48,612 patients hospitalized at 259 centers in the United States were enrolled in the OPTIMIZE-HF registry. All regions of the United States were represented, and institutions from community hospitals to large tertiary medical centers participated.[17,18] A prespecified subgroup of patients (10%) enrolled from 91 hospitals had 60- to 90-day follow-up data, and this cohort was demographically similar to patients in the overall registry.[18,19] Participants were screened for inclusion before hospital discharge or identified from administrative databases subsequent to discharge. The registry component captured data on important characteristics (demographic, pathophysiologic, and clinical), treatment patterns, and outcomes of patients hospitalized for HF using a web-based case report form.[17] Written informed consent was obtained before enrollment from patients who participated in the follow-up data collection. The protocol was approved by each participating center?s institutional review board or through use of a central institutional review board. The registry coordinating center was Outcome Sciences, Inc (Cambridge, MA).

OPTIMIZE-HF Registry Process-of-Care

Improvement Program. The OPTIMIZE-HF Process-of-Care Improvement Program assisted hospitals in improving the quality of care in hospitalized patients with HF through the use of a hospital tool kit including order sets, treatment algorithms, discharge checklists, and structured educational materials, as previously described.[17] Through the use of these materials and benchmarked feedback data, OPTIMIZE-HF aimed to accelerate the adoption of evidence-based, guideline-recommended therapies.

This analysis of the OPTIMIZE-HF registry assessed the presence or absence of DM as documented in the medical record in patients hospitalized for HF. The association of DM with the characteristics, quality of care, treatment, and outcomes in these patients was analyzed. Specific measures examined in this patient population include the use of evidence-based medicines (ACEI and/or angiotensin receptor blockers [ARB], aldosterone antagonists, aspirin, β-blockers, statins, and warfarin); Joint Commission on Accreditation of Healthcare Organizations (JCAHO) core HF measures (complete discharge instructions, LVEF assessment, discharge ACEI use, smoking-cessation counseling); and clinical outcomes in-hospital (length of stay, in-hospital mortality) and at 60- to 90-day follow-up (all-cause rehospitalization, all-cause mortality) as previously described.[17-19]

Statistical Analysis. All statistical analyses were performed independently at the Duke Clinical Research Institute (Durham, NC). The data are reported as mean and standard deviation for continuous variables or percentages of nonmissing values for categorical variables. Data on HF medication use are reported as the number and frequency of eligible patients, and contraindications/intolerance were assessed at discharge from the index hospitalization and during follow-up. Patient characteristics and evidence-based treatments at hospital discharge were compared using Pearson Χ2 test for categorical variables and t test for continuous variables. Multivariable models of inhospital death, length of hospital stay, postdischarge mortality, and postdischarge death or rehospitalization were developed to be used for consistent covariate adjustment across all studies. The types of models were logistic for in-hospital mortality, general linear modeling for length of stay, Cox proportional hazards for postdischarge mortality, and logistic for postdischarge mortality and rehospitalization (date of rehospitalization was not available for survival modeling). The modeldevelopment process was similar for all 4 outcomes and used stepwise and backward variable selection methods. The linearity assumption for continuous measures was evaluated using restricted cubic spline transformations. When needed, appropriate transformations such as piecewise linear splines were applied. A P value of .05 was used for both entry and remaining in the model. The potential covariates were preselected with 45 for in-hospital mortality, 39 for length of stay, 19 for postdischarge mortality, and 70 for postdischarge mortality or rehospitalization (Supplemental appendix). To test the relative effect of diabetes in the final adjusted model, an indicator for the presence or absence of diabetes was added to the model. The relative effect of insulin-treated diabetes versus non?insulin-treated diabetes in the final model was investigated in a similar manner. SAS version 8.2 (SAS Institute, Cary, NC) was used for all statistical analyses.

Results

Patient Comparison

This analysis included 48,612 patients at 259 hospitals, with follow-up data from 5791 patients. Among patients hospitalized for HF, 42% had diabetes (n = 20,162). Of these patients with DM, 40% were treated with insulin (n = 8058). The overall registry population was elderly, with a mean age of 73.2 years, and patients with DM were slightly younger than their nondiabetic counterparts (71.5 ? 12.2 vs 74.4 ? 15.0 years; P < .0001). Diabetic patients were more likely to have an ischemic etiology, hyperlipidemia, and a hypertension history than the nondiabetic cohort. In addition, these diabetic patients were less likely to be white, have atrial arrhythmia, or be cigarette smokers. They were also more likely to have edema on admission and discharge, to weigh more on admission and discharge, and to have higher systolic blood pressure on admission and discharge. Patients with DM were more likely to have a history of renal insufficiency, and admission serum creatinine was significantly higher than in patients without DM (1.93 vs 1.65 mg/dL; P < .0001). Furthermore, this cohort demonstrated lower B-type natriuretic peptide levels compared with patients without DM. Table 1 summarizes the patient characteristics at admission and discharge in patients with and without DM.

Quality of Care

Quality of care for patients with HF within the OPTIMIZE-HF program was assessed using performance measures from JCAHO as well as additional quality measures, including discharge use of β-blockers, ACEI/ARB, and warfarin for patients with atrial fibrillation, as previously described.[17-19] Analysis of performance on specific JCAHO quality measures within this registry demonstrated that patients with DM were less likely to receive ACEI in the absence of contraindications and intolerance (72.9% vs 76.8%; P < .0001). LVEF assessment was also slightly less in diabetic patients (85.9% vs 87.0%; P = .0013). Diabetic patients participating in OPTIMIZEHF were just as likely as nondiabetic patients to receive smoking-cessation counseling and complete discharge instructions (Figure 1). Additional OPTIMIZE-HF quality measures showed that eligible patients with DM were slightly less likely to be treated at discharge with ACEI/ARB (81.0% vs 83.6%; P < .0001), slightly more likely to be treated with β-blockers (83.9% vs 82.6%; P = .0246), and just as likely to be treated with warfarin and aldosterone antagonists (Figure 2). The use of statins was significantly greater in diabetics than in nondiabetics (41.3% vs 36.7%; P < .0001), a finding that may be related to the higher likelihood of ischemic disease in these enrolled diabetic patients. Further analysis of patients based on insulin treatment showed no significant differences between insulin-treated and non?insulin-treated patients except that insulin-treated patients were slightly less likely to have left ventricle function assessment (84.5% vs 86.8%; P < .0001).

Figure 1. 

Percent of patients receiving HF measures at hospital discharge bydiabetes status.

     

Figure 2 

Eligible patients receiving other medications at hospital dischargeby diabetes status. *ACEI/ARB use in patients with LVSD, excludingpatients with contraindications. ?β-Blocker use in patients withLVSD, excluding patients with contraindications. ?Statin use inpatients with medical history of coronary artery disease, cerebrovascularaccident/transient ischemic attack, diabetes, hyperlipidemia,or peripheral vascular disease. ?Aldosterone antagonist usein patients with LVSD. ||Warfarin use in patients with chronic orparoxysmal atrial fibrillation.

     

Clinical Outcomes

In-hospital mortality was similar for diabetic and nondiabetic patients with HF (3.6% vs 3.9%; P = .08) ( Table 2 ). Patients with DM did experience a modestly increased length of stay (5.9 vs 5.5 days; P < .0001). In addition, among the follow-up cohort, HF patients with DM had a higher incidence of rehospitalization compared with HF patients without DM (31.5% compared to 28.2%; P = .006).

On multivariable analysis, diabetes status was not an independent predictor of in-hospital or postdischarge mortality (in-hospital mortality: odds ratio [OR], 1.00; 95% confidence interval [CI], 0.88-1.14; P = .99; postdischarge mortality: hazard ratio, 1.08; 95% CI, 0.87-1.35; P = .48). In addition, diabetes was not an independent predictor of 60- to 90-day mortality and/or rehospitalization (OR, 1.06; 95% CI, 0.92-1.22; P = .43). An additional multivariable analysis demonstrated that insulin-treated diabetes status was not an independent predictor of in-hospital or postdischarge mortality (inhospital mortality: OR, 1.11; 95% CI, 0.92-1.34; P = .29; postdischarge mortality: hazard ratio, 1.24; 95% CI, 0.89-1.73; P = .21) or 60- to 90-day mortality and/or rehospitalization (OR, 1.15; 95% CI, 0.94-1.42; P = .17).

The patient population in the OPTIMIZE-HF registry was also stratified based on the presence or absence of LVSD ( Table 2 ). For patients with LVSD, those with DM had increased incidence of rehospitalization (33.7% vs 27.2%; P = .0003) compared to nondiabetic patients with LVSD. For patients with LVSD, the presence of diabetes was associated with an increased risk for 60- to 90-day mortality and/or rehospitalization (OR, 1.33; 95% CI, 1.12-1.59; P = .001).

In those patients without LVSD, DM was not related to in-hospital mortality or follow-up rehospitalization ( Table 2 ). Diabetes was associated with a lower likelihood of follow-up mortality (6.2% vs 9.2%; P = .008) in patients with preserved systolic function ( Table 2 ). In contrast to what was observed for patients with LVSD, in patients with preserved systolic function, DM was not associated with an increased risk for 60- to 90-day mortality and/or rehospitalization (OR, 0.98; 95% CI, 0.81-1.19; P = .87). The association of the use of ACEI/ARB and h-blocker in eligible patients with and without diabetes with 60- to 90-day death and death/rehospitalization was similar to the findings for the overall cohort.

Discussion

This analysis from OPTIMIZE-HF demonstrates that a very high prevalence of DM exists among hospitalized patients with HF, with 42% of those enrolled having DM documented. A prevalence similar to that of the current study has also been observed in the Acute Decompensated Heart Failure National Registry, which showed that 44% of hospitalized patients with HF have DM.[20] In the Enhanced Feedback for Effective Cardiac Treatment study of 4031 community-based patients presenting with new onset HF at multiple hospitals in Ontario, Canada, from 1997 to 2001, the prevalence of diabetes was 34%.[21] Interestingly, clinical trials investigating the use of HF therapies have generally enrolled populations with a substantially lower prevalence of DM.[4-9]

One possible explanation for the underrepresentation of the diabetic cohort in clinical trials is a common presumption that patients with DM are not treated in the same manner as their nondiabetic counterparts due to potential problems associated with tolerance and adherence to medication strategies, particularly β-blockade. Results from OPTIMIZE-HF demonstrate that this assumption may not be founded. Participation in the OPTIMIZE-HF performance-improvement program was associated with remarkably high rates of β-blocker treatment at hospital discharge among eligible HF patients with and without DM. This is a significant finding, considering conflicting views on β-blocker use being contraindicated in diabetic patients with HF due to the possible worsening of glycemic and metabolic parameters.[15,16] In fact, as noted earlier, the actual percentage of HF patients with DM using β-blocker therapy at the 60- to 90-day follow-up period was higher than that among HF patients without DM. In addition, patients with DM were more likely to be retained on ACEI/ARB at the 60- to 90-day follow-up compared with HF patients without DM. These results describing the use of evidence-based medication use shed new light on managing medication strategies in HF patients with DM, and, based on them, one may argue that HF patients with DM need not be excluded and that aggressive evidence-based therapies are warranted in both diabetic and nondiabetic cohorts of patients with HF.

The presence of DM in patients with HF has been shown in previous studies to be associated with an increased risk for poor outcomes.[21-26] An analysis from the Studies of Left Ventricular Dysfunction trials and registry program demonstrated that the presence of DM in patients with HF resulted in an increased risk for mortality.[23] However, this increased mortality risk with DM was confined to patients with coronary artery disease, with no increase in mortality risk in patients with nonischemic HF.[26] The Beta-Blocker Evaluation of Survival Trial reported a DM prevalence of 36% among patients with HF enrolled, and on multivariable analysis showed that DM was independently associated with a 22% increased risk of allcause mortality ( P = .007).[25] Interestingly, this increased risk was also found to be confined to those patients with an ischemic cardiomyopathy etiology of HF. Results from the Danish Investigations of Arrhythmia and Mortality on Dofetilide study demonstrated an independent risk for mortality associated with DM in patients hospitalized with HF.[26] The increased risk of diabetes was similar in those patients with depressed and normal left ventricular systolic function.[24] Recently, an analysis from the CHARM program showed that the predicted risk for mortality associated with DM in this 2-year study was found to be 2 times greater for patients treated with insulin and 58% greater for non?insulintreated patients compared to those HF patients without DM.[22]

In contrast, the present analysis of short-term clinical outcomes in the OPTIMIZE-HF registry demonstrates that in-hospital and 60- to 90-day follow-up mortality rates were similar in patients with DM and those without, despite the increased frequency of baseline characteristics associated with increased mortality risk. Surprisingly, among HF patients without LVSD, those with DM actually had significantly lower all-cause postdischarge mortality than patients without DM (6.2% vs 9.2%; P = .008). The outcome data within OPTIMIZEHF could have been influenced by the high use of evidence-based therapies among patients with DM in this study. Use of these proven treatments may have minimized the outcomes differences between the diabetic and nondiabetic patients enrolled, although patients with DM and HF were shown to experience a longer length of hospital stay and higher rehospitalization rates compared with HF patients without DM. Alternatively, in the first 60- to 90-day follow-up period, DM may have little adverse effect on mortality, but with longer term follow-up, an adverse effect on prognosis becomes evident. Results of previous studies demonstrating an increased risk for mortality were derived from patients over longer periods of time.[22-26] In the Enhanced Feedback for Effective Cardiac Treatment study, diabetes was a univariate predictor of a 25% lower 30-day and a 19% lower 1-year post hospital discharge mortality; on multivariate analysis, diabetes was not independently predictive of either higher or lower mortality risk.[21]

There are a number of important limitations to this analysis. All clinically relevant medical histories, including diabetes status, were obtained as documented in the medical record. There may have been patients with DM who had not yet been diagnosed at study entry, which would further increase the already high proportion of patients with HF and DM. Information regarding the degree of glycemic control, duration of diabetes, and diabetes medications other than insulin was not captured. Medication use in OPTIMIZE-HF is as documented in the medical record, and it is therefore possible that actual adherence rates may have been lower than reported. In addition, contraindications and intolerance were as documented in the medical record, and a proportion of patients reported to be eligible but not treated may have had contraindications that went undocumented. The OPTIMIZE-HF was not designed as a prospective randomized trial, and unmeasured confounders may have influenced clinical outcomes. These findings may not apply to hospitals that differ in patient characteristics or care patterns from OPTIMIZE-HF hospitals. Given the overall large number of patients observed, some differences, although statistically significant, may not be clinically relevant. Follow-up data were obtained in a subset of patients and were limited to 60 to 90 days. Despite these limitations, this analysis provides new insights into the characteristics, treatments, and outcomes from a large representative data set of patients hospitalized with HF, including patients with preserved systolic function and multiple comorbidities.

Conclusion

This report from the OPTIMIZE-HF registry revealed a high prevalence of DM (42%) among patients hospitalized with HF. In addition, this registry demonstrated that HF patients with DM differed from those without DM in a number of important characteristics, including younger age, greater ischemic etiology, and increased renal dysfunction. Treatment and clinical outcomes in HF patients with DM were relatively similar to those without DM. Importantly, the presence of DM was not associated with increased mortality, either during hospitalization or during the 60 to 90 days of postdischarge follow-up. These results provide additional support for the use of aggressive evidence-based therapies in HF patients with DM, and they challenge traditional lines of thought on the difficulties in treating this at-risk patient population.


Table 1. Patient Characteristics at Hospital Admission and Discharge


Table 1

 

Table 2. In-Hospital and 60- to 90-day Follow-up Outcomes by Diabetes and Left Ventricular Function


Table 2

 



References

  1. Kannel WB, D'Agostino RB, Silbershatz H, et al. Profile for estimating risk of heart failure. Arch Intern Med 1999;159: 1197-204.
  2. Nichols GA, Hillier TA, Erbey JR, et al. Congestive heart failure in type 2 diabetes: prevalence, incidence, and risk factors. Diabetes Care 2001;24:1614-9.
  3. Kannel WB, Hjortland M, Castelli WP. Role of diabetes in congestive heart failure: the Framingham study. Am J Cardiol 1974;34:29-34.
  4. MERIT-HF Investigators. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet 1999; 353:2001-7.
  5. Poole-Wilson PA, Swedberg K, Cleland JG, et al. Comparison of carvedilol and metoprolol on clinical outcomes in patients with chronic heart failure in the Carvedilol or Metoprolol European Trial (COMET): randomised controlled trial. Lancet 2003;362:7-13.
  6. CONSENSUS Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure. Results of the Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS). N Engl J Med 1987;316:1429-35.
  7. Cohn JN, Tognoni G. A randomized trial of the angiotensinreceptor blocker valsartan in chronic heart failure. N Engl J Med 2001;345:1667-75.
  8. Pitt B, Poole-Wilson PA, Segal R, et al. Effect of losartan compared with captopril on mortality in patients with symptomatic heart failure: randomised trial?the Losartan Heart Failure Survival Study ELITE II. Lancet 2000;355:1582-7.
  9. Yusuf S, Pfeffer MA, Swedberg K, et al. Effects of candesartan in patients with chronic heart failure and preserved left-ventricular ejection fraction: the CHARM-Preserved Trial. Lancet 2003;362:777-81.
  10. Yusuf S, Sleight P, Pogue J, et al. Effects of an angiotensinconverting-enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. The Heart Outcomes Prevention Evaluation Study Investigators. N Engl J Med 2000;342:145-53.
  11. Shekelle PG, Rich MW, Morton SC, et al. Efficacy of angiotensinconverting enzyme inhibitors and beta-blockers in the management of left ventricular systolic dysfunction according to race, gender, and diabetic status. A meta-analysis of major clinical trials. J Am Coll Cardiol 2003;41:1529-38.
  12. Deedwania PC, Giles TD, Klibaner M, et al. Efficacy, safety and tolerability of metoprolol CR/XL in patients with diabetes and chronic heart failure: experiences from MERIT-HF. Am Heart J 2005;149:159-67.
  13. Hunt SA, Abraham WT, Chin MH, et al. ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines American College of Cardiology Website. Available at: http://www.acc.org/qualityandscience/clinical
    /guidelines/failure/update/index.pdf, 2005.
  14. Adams KF, Lindenfeld J, Arnold JMO, et al. HFSA 2006 comprehensive heart failure practice guideline. J Card Fail 2006;12:e1-e122.
  15. Giugliano D, Acampora R, Marfella R, et al. Metabolic and cardiovascular effects of carvedilol and atenolol in non-insulindependent diabetes mellitus and hypertension: a randomized, controlled trial. Ann Intern Med 1997;126:955-9.
  16. Jacob S, Rett K, Wicklmayr M, et al. Differential effect of chronic treatment with two beta-blocking agents on insulin sensitivity: the carvedilol-metoprolol study. J Hypertens 1996;14:489-94.
  17. Fonarow GC, Abraham WT, Albert NM, et al. Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF): rationale and design. Am Heart J 2004;148:43-51.
  18. Gheorghiade M, Abraham WT, Albert NM, et al. Systolic blood pressure at admission, clinical characteristics, and outcomes in patients hospitalized with acute heart failure. JAMA 2006;296:2217-26.
  19. Fonarow GC, Abraham WT, Albert NM, et al. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA 2007;297:61-70.
  20. Adams KF, Fonarow GC, Emerman CL, et al. Characteristics and outcomes of patients hospitalized for heart failure in the United States: rationale, design, and preliminary observations from the first 100,000 cases in the Acute Decompensated Heart Failure National Registry (ADHERE). Am Heart J 2005;149:209-16.
  21. Lee DS, Austin PC, Rouleau JL, et al. Predicting mortality among patients hospitalized for heart failure: derivation and validation of a clinical model. JAMA 2003;290:2581-7.
  22. Pocock SJ, Wang D, Pfeffer MA, et al. Predictors of mortality and morbidity in patients with chronic heart failure. Eur Heart J 2006;27:65-75.
  23. 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.
  24. Gustafsson I, Brendorp B, Seibaek M, et al. Influence of diabetes and diabetes-gender interaction on the risk of death in patients hospitalized with congestive heart failure. J Am Coll Cardiol 2004;43:771-7.
  25. 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.
  26. 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.
Funding Information

The OPTIMIZE-HF registry and this study were supported by GlaxoSmithKline, Philadelphia, PA.

Reprint Address

Barry H. Greenberg, MD, University of California-San Diego Medical Center, 200 W. Arbor Drive, #8411, San Diego, CA 92103-8411. E-mail: bgreenberg@ucsd.edu


Barry H. Greenberg, MD,a William T. Abraham, MD,b Nancy M. Albert, PhD, RN,c Karen Chiswell, MS,d Robert Clare, MS,d Wendy Gattis Stough, PharmD,e Mihai Gheorghiade, MD,f Christopher M. O'Connor, MD,e Jie Lena Sun, MS,d Clyde W. Yancy, MD,g James B. Young, MD,c Gregg C. Fonarow, MDh

aUniversity of California-San Diego Medical Center, San Diego, CA,
bOhio State University, Columbus, OH,
cCleveland Clinic Foundation, Cleveland, OH,
dDuke Clinical Research Institute, Durham, NC,
eDuke University Medical Center, Durham, NC,
fNorthwestern Feinberg School of Medicine, Chicago, IL,
gBaylor University Medical Center, Dallas, TX,
hUCLA Medical Center, Los Angeles, CA.