Predictors of mortality after discharge in patients hospitalized with heart failure: An analysis from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF)

O'Connor, Christopher M.; Abraham, William T.; Albert, Nancy M.; Clare, Robert; Gattis Stough, Wendy; Gheorghiade, Mihai; Greenberg, Barry H.; Yancy, Clyde W.; Young, James B.; Fonarow, Gregg C.

doi:10.1016/j.ahj.2008.04.030

« Previous Next »American Heart Journal
Volume 156, Issue 4 , Pages 662-673, October 2008

Predictors of mortality after discharge in patients hospitalized with heart failure: An analysis from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF)

Christopher M. O'Connor, MD
- Duke University Medical Center/Duke Clinical Research Institute, Durham, NC
- Reprint requests: Christopher M. O'Connor, MD, Duke University Medical Center, Duke Clinical Research Institute, 2400 N. Pratt Street, Durham, NC 27705.
William T. Abraham, MD
- The Ohio State University, Columbus, OH
Nancy M. Albert, RN, PhD
- George M. and Linda H. Kaufman Center for Heart Failure, Cleveland Clinic Foundation, Cleveland, OH
Robert Clare, MS
- Duke Clinical Research Institute, Durham, NC
Wendy Gattis Stough, PharmD
- Duke University Medical Center, Durham, NC
- Department of Clinical Research, Campbell University School of Pharmacy, Research Triangle Park, NC
Mihai Gheorghiade, MD
- Northwestern University, Chicago, IL
Barry H. Greenberg, MD
- University of California San Diego–Hillcrest Medical Center, San Diego, CA
Clyde W. Yancy, MD
- Baylor Heart and Vascular Institute, Baylor University Medical Center, Dallas, TX
James B. Young, MD
- George M. and Linda H. Kaufman Center for Heart Failure, Cleveland Clinic Foundation, Cleveland, OH
Gregg C. Fonarow, MD
- University of California–Los Angeles Medical Center, Los Angeles, CA

Received 19 November 2007; accepted 16 April 2008.

Background

Patients with heart failure (HF) are at high risk for mortality and rehospitalization in the early period after hospital discharge. We developed clinical models predictive of short-term clinical outcomes in a broad patient population discharged after hospitalization for HF.

Methods

The Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) registry is a comprehensive hospital-based registry and performance-improvement program for patients hospitalized with HF. Follow-up data were scheduled to be prospectively collected at 60 to 90 days postdischarge in a prespecified 10% sample. For the 4,402 patients included in this analysis, 19 prespecified potential predictor variables were used in a stepwise Cox proportional hazards model for all-cause mortality. Logistic regression including 45 potential variables was used to model mortality or rehospitalization.

Results

The 60- to 90-day postdischarge mortality rate was 8.6% (n = 481), and 29.6% (n = 1,715) were rehospitalized. Factors predicting early postdischarge mortality include age, serum creatinine, reactive airway disease, liver disease, lower systolic blood pressure, lower serum sodium, lower admission weight, and depression. Use of statins and β-blockers at discharge was associated with significantly decreased mortality. The C-index of the model was 0.74. The most important predictors for the combined end point of death or rehospitalization were admission serum creatinine, systolic blood pressure, admission hemoglobin, discharge use of angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, and pulmonary disease. From this analysis, 8 factors identified to carry significant risk were selected for use in a point scoring system to predict the risk of mortality within 60 days after discharge, with a C-index of 0.72.

Conclusions

A substantial risk of mortality and mortality or rehospitalization is present in the first 60 to 90 days after discharge from a hospitalization for HF. Several factors were identified that signal high-risk patients. Application of these findings with a simple algorithm can distinguish patients who are low risk from those at high risk who may benefit from closer monitoring and aggressive evidence-based treatment.