Effect of a Heart Failure Program on Hospitalization Frequency and Exercise Tolerance
Background Most patients with heart failure are currently managed by physicians with little specific expertise in heart failure. This management system has been associated with evidence of suboptimal care, such as high rates of hospital readmission and underuse of ACE inhibitors. The current study was undertaken to determine whether hospitalization rates and functional outcomes are improved when patients are managed by physicians with special expertise in heart failure working in a dedicated heart failure program.
Methods and Results All patients with heart failure referred to the Vanderbilt Heart Failure and Heart Transplantation Program between July 1994 and June 1995 were identified. Annual hospitalization rates, medications, and peak exercise capacity before and after referral were compared in patients followed for >30 days. A total of 187 patients were referred during the index time period, of whom 134 (72%) were followed for >30 days. During the year before referral, 94% of the patients were hospitalized (210 cardiovascular hospitalizations) versus 44% of the patients during the year after referral (104 hospitalizations) (53% reduction) (P<.01). Hospitalizations for heart failure decreased from 164 to 60 for all patients regardless of follow-up duration and decreased from 97 to 30 (69% reduction) for patients followed at least 1 year after referral. Eighty-eight of the patients were able to exercise at the time of referral. Peak exercise V̇o2 in this group increased from 12.8±4.7 to 15.7±4.8 mL · min−1 · kg−1 (P<.01) by 6 months after referral. Loop diuretic doses were on averaged doubled during the first 6 months after referral.
Conclusions These findings suggest that patients with heart failure have fewer hospitalizations for heart failure and are significantly more functional when managed by heart failure specialists working in a dedicated heart failure program rather than by physicians with limited expertise in heart failure.
Heart failure is a chronic disease associated with frequent hospitalizations, widespread functional impairment, and a high mortality rate.1 2 3 Despite these serious outcomes, the vast majority of patients with heart failure are currently managed by physicians with limited expertise in heart failure. Most of these physicians are primary care physicians and internists. Approximately 20% of the patients are managed by cardiologists.3 However, few of these cardiologists have special training in heart failure.
Under this system, the cost of caring for patients with heart failure has reached crisis proportions while the quality of care has been suboptimal. Each year, patients with heart failure are hospitalized >2 million times and generate >11 million outpatient visits, resulting in annual costs in excess of $38 billion.2 Readmission rates remain high, while utilization rates of ACE inhibitors, a class of medication clearly indicated in heart failure, remain low.3 4 5 6
Recently, it has been suggested that centers specializing in the management of patients with heart failure may be able to improve these outcomes.7 8 However, to date, there is limited evidence to support this presumption. The present study was undertaken to investigate this hypothesis by comparing hospitalization rates, medication use, and exercise capacity of patients with heart failure before and after referral to the Vanderbilt Heart Failure and Heart Transplantation Program.
All patients with heart failure referred to the Vanderbilt Heart Failure and Heart Transplantation Program between July 1994 and June 1995 were identified (Table 1⇓). A total of 187 patients were referred, 138 as ambulatory outpatients. Cardiologists were involved in the referral of all patients. Forty-nine patients were transferred to Vanderbilt Hospital from other hospitals. Fifty-four patients were women and 133 were men. The average age of the patients was 52±12 years. Heart failure was due to coronary artery disease in 81 patients, nonischemic cardiomyopathy in 97 patients, and other causes in the remaining patients.
A total of 134 of the 187 patients were followed on a long-term basis (>30 days) as outpatients by the Heart Failure and Heart Transplant Program and were the focus of this study. Fifty-three patients were excluded from analysis. Nine patients were listed for heart transplantation, made United Network of Organ Sharing status 1 within 30 days of referral, and subsequently underwent transplantation. Twelve patients died within 30 days of referral. Twenty-nine patients were referred back to their local physician primarily because of minimal or no left ventricular dysfunction. Three patients were transferred to other health care systems due to insurance issues.
Patients were followed on a long-term basis by three physicians who work exclusively with heart failure and heart transplantation patients. Two nurse coordinators assisted with the management of patients both during hospitalizations and as outpatients. Home health agencies were involved in the care of ≈10% of the patients.
All patients referred to the program undergo echocardiographic evaluation and, whenever possible, cardiopulmonary exercise testing. Exercise testing is typically repeated at 3 to 6 months to monitor patient status, if the patient is able to exercise. Exercise testing with concurrent hemodynamic monitoring is used extensively in the evaluation and management of patients, as described previously.9 10
Program offices are based in a single area in the Vanderbilt outpatient office building. Cardiopulmonary exercise testing and exercise hemodynamic testing are performed by program staff in an outpatient laboratory adjacent to the office area. Patient information and outcomes are maintained in a Microsoft Access database. Periodic meetings are held with Vanderbilt Home Health Agency and local hospice care programs to integrate patient care.
Hospitalization frequency in the 134 ambulatory patients followed for >30 days was compared before and after referral. The number of hospitalizations during the year before referral was determined by patient history and review of previous medical records. Hospitalizations after referral both at Vanderbilt Medical Center and at other hospitals was compiled in an Access database.
Not all patients were followed a full year after referral due to, for example, deaths and heart transplantation; therefore, annualized hospitalization rates were calculated using life-table analysis. Patients who were lost to follow-up, underwent heart transplantation, or died were treated as censored events.
The use of digoxin, ACE inhibitors, and diuretics at the time of patient referral was compared with utilization rates at 6 months after referral. Loop diuretic doses were calculated as furosemide equivalent doses by multiplying the bumetanide dose by 40 and the torsemide dose by 2. Doses of ACE inhibitors were converted to captopril equivalent doses by multiplying the total daily dose by the following correction factors: enalapril (3.75), captopril (1.0), lisinopril (3.75), quinapril (1.88), and ramipril (7.5).
Maximal cardiopulmonary exercise testing was routinely performed on all stable ambulatory patients at the time of referral and at 3 to 6 months after referral. Eighty-eight patients underwent such testing. The exercise test performed at the time of referral was considered a reflection of the patient’s exercise performance before referral and was compared with the test performed at 3 to 6 months. Patients who were initially hospitalized, stabilized, and then followed as outpatients also underwent exercise testing at 3 to 6 months. These patients were excluded from exercise analysis because a baseline measurement was not available.
Cardiopulmonary exercise testing was performed after a 3-hour fast with the use of a Marquette treadmill and a MedGraphics CardioO2 System. After 3 minutes of resting data acquisition, symptom-limited exercise testing was performed using a 3 minute/stage Naughton protocol. All patients discontinued exercise due to a combination of dyspnea and fatigue. Arterial hemoglobin O2 saturation was monitored continuously via a pulse oximeter.
Peak exercise V̇o2 was defined as the highest V̇o2 achieved at peak exercise. The anaerobic threshold was defined using three criteria: the point after which the respiratory gas exchange ratio (V̇co2/V̇o2) consistently exceeded the resting ratio, the point at which the ventilatory equivalent for oxygen (Ve/V̇o2) was minimal followed by a progressive increase in Ve/V̇o2, and the V̇o2 after which a nonlinear rise in Ve occurred relative to V̇o2.11 12 If these three criteria identified different anaerobic thresholds, threshold values were averaged.
At the time of exercise testing, a subgroup of patients completed the Minnesota Living with Heart Failure Questionnaire.13 14 The Minnesota Living with Heart Failure Questionnaire consists of 21 questions, each of which assesses the patient’s perception of how his or her emotional and physical state is impaired by heart failure. The answer to each question ranges from a score of 0 (No impairment) to 5 (Very much impaired), so the total score can range from 0 to 105. The higher the score, the more severe is the impairment.
All data are expressed as mean±SD. Differences between groups were evaluated by ANOVA or χ2 analysis. Differences within groups were compared using Student’s t test. Probability of survival and probability of survival without hospital admission were analyzed by the life-table method. Curves were compared using the log-rank test and Wilcoxon analysis. Outcomes were defined as sudden death if the event occurred out of hospital without symptomatic deterioration within the preceding 24 hours or as progressive heart failure in patients with progressive symptomatic or hemodynamic deterioration. Cardiac transplantation was treated as a censored observation (ie, withdrawn at the time of transplantation). Patients lost to follow-up were also treated as a censored event. A value of P<.05 was considered statistically significant.
A total of 187 patients were referred during the index time period, of whom 134 (72%) were followed for >30 days. During the year before referral, 88% of the patients were hospitalized. The calculated annualized rate of first hospitalization for these patients was 94% at 1 year (Fig 1⇓). The incidence of first hospitalization increased during the 2 months before referral; often, such hospitalizations precipitated referral.
During the year after referral, the annual rate of hospitalization was reduced to 44% (53% reduction) (P<.001) (Fig 1⇑). Total cardiovascular hospitalizations decreased from 210 to 104 (P<.01), primarily due to a decrease in hospitalizations for decompensated heart failure from 164 to 60 (Table 2⇓).
Ninety patients were followed for an entire year after referral. In these patients, total cardiovascular hospitalizations decreased from 131 during the year before referral (1.46 hospitalizations/year) to 62 (0.69 hospitalizations/year) during the year after referral. This reduction was primarily due to a decrease in hospitalizations for heart failure from 97 to 30 occurrences (both P<.01). There were 34 non–heart failure hospitalizations in the year before referral versus 32 after referral (P=NS). Reductions in hospitalization for heart failure was observed in these patients regardless of how many hospitalizations were noted during the year before referral (Fig 2⇓).
To determine whether the use of heart transplantation influenced this reduction in hospitalization rates, hospitalization frequency was analyzed with all patients listed for heart transplantation during the year after referral excluded from the analysis. Removal of patients listed for transplantation had little affect on the results. Ninety-four percent of patients were hospitalized during the year before referral versus 41% during the year after referral (P<.015) (Fig 3⇓).
Fifteen of the 134 patients followed for >30 days died during the subsequent year: 12 from sudden death and 3 from low output heart failure. Eight patients underwent heart transplantation, 9 returned to their local medical physician due to improvement, and 9 were lost to follow-up. One-year survival for the entire group was 87% (Fig 4⇓).
Maximal Exercise Capacity
Changes in maximal exercise capacity were compared in 88 patients who were ambulatory at the time of referral and underwent maximal exercise testing at that time (Fig 5⇓). On repeat exercise testing at 3 to 6 months, peak exercise V̇o2 had increased from 12.8±4.7 to 15.7±4.8 mL · min−1 · kg−1 (P<.001). The anaerobic threshold increased from 10.4±2.4 to 11.4±2.7 mL · min−1 · kg−1 (P<.002). Thirty-four of these patients also completed the Living with Heart Failure Questionnaire at baseline and at 3 to 6 months; composite scores improved from 57±27 to 35±26 (P<.01).
At the time of referral, 76% of the patients were receiving ACE inhibitors. Forty-three percent were receiving captopril, 56% enalapril, and 1% other agents. By 6 months, 74% were receiving ACE inhibitors: 37% captopril, 61% enalapril, and 2% other agents. Eighty-eight percent were receiving digoxin on referral; 7 additional patients (5%) were placed on digoxin after referral. Average doses were not changed (Table 3⇓).
One-hundred twenty-four patients (93%) were receiving loop diuretics, 4 thiazide diuretics, and 6 no diuretics. Five were also receiving metolazone on referral. By 6 months, the average loop diuretic dose had been increased from 94±84 to 181±165 mg (P<.001), and 6 additional patients had been placed on metolazone therapy (Table 3⇑).
At baseline, 14 patients were receiving antiarrhythmic agents. Six patients were receiving amiodarone and 1 sotalol for atrial fibrillation or sustained ventricular tachycardia. These medications were maintained. Seven patients were receiving agents for nonsustained, asymptomatic ventricular tachycardia (mexiletine, 5; quinidine, 1; tocainide, 1). All of these agents were stopped except for mexiletine in 1 patient.
At baseline, 3 patients were receiving amlodipine. This agent was stopped in all 3 patients. One patient was receiving vesnarinone as part of a study initiated at another center; this agent was maintained during the study.
Eleven patients were receiving a β-blocker at the time of referral. The β-blocker was continued in 8 of these patients and stopped in 3. Eight patients were begun on a β-blocker during the year after referral, 1 on metoprolol, and 7 on carvedilol.
At present, most patients with heart failure are managed by primary care physicians and internists with little special expertise in heart failure.3 It is becoming clear that this system results in excessive costs and relatively poor outcomes.1 2 3 4 5 6 Consequently, multiple groups have been trying to develop new management systems for heart failure that will provide better outcomes at lower cost.
One approach that has received considerable recent attention has been the development of core resources to help generalists manage patients with heart failure. These core resources were recently summarized in a Cardiology Preeminence Roundtable publication assembled by the Advisory Board Company.5 Proposed core resources included outpatient heart failure clinics staffed by specialized nurses, home health care cardiac specialists, community-based case managers, patient “telemanagement,” hospital-sponsored cardiac rehabilitation, short-stay observation units for patients with decompensated heart failure, and inpatient heart failure subacute care facilities.
To date, most evidence suggests that development of such core resources can reduce hospitalization frequency6 15 16 ; however, there is some evidence that core resources can sometimes increase the cost of care. For example, investigators at Veterans Administration Hospitals recently completed a multicenter trial in which 504 patients hospitalized with heart failure were randomly assigned to either usual care or intensive follow-up and intervention by a nurse and primary care physician.17 The patients subjected to intensive follow-up had a higher readmission rate per month than the control group, possibly because patients were more able to express complaints when seen more frequently.
Another model for managing patients with heart failure is the development of heart failure centers staffed by physicians who exclusively manage patients with heart failure. Theoretically, this system should produce better outcomes than systems managed by generalists. However, to date there is limited evidence to support this conclusion aside from a recent preliminary report by Fonarow et al7 describing a dramatic decrease in hospitalization rates and improvement in exercise capacity in 214 patients listed for heart transplantation at the University of California Los Angeles Medical Center.
The present study was undertaken to examine outcomes in patients managed by the Vanderbilt Heart Failure and Heart Transplantation Program. This program is staffed by physicians and nurses who exclusively manage patients with heart failure. In addition, program physicians routinely use cardiopulmonary exercise testing and hemodynamic monitoring at rest and during exercise to manage patients, based on evidence that such testing provides prognostic and therapeutic information.9 10 18 19 20
Outcomes were examined in all patients referred over a 1-year period and who were subsequently followed as an outpatient for >30 days. Patients who were followed for <30 days due to death, immediate listing for heart transplantation as United Network for Organ Sharing status 1 patients, and referral back to their local physician were not included in the analysis. These patients were excluded based on the premise that outcomes in this groups largely reflected management of the patients before referral. All patients were managed exclusively by physicians in the Vanderbilt Heart Failure and Heart Transplant Program.
After referral, patients exhibited a ≈50% reduction in hospitalization frequency. This reduction was primarily due to a decrease in hospitalizations for heart failure, not to a reduction in hospitalizations for other problems such as arrhythmias, angina, and myocardial infarction. This is not particularly surprising, since it is unlikely that management of the patients over only 1 year would decrease ischemic burdens or propensity for arrhythmias.
Fonarow et al7 reported a much more striking reduction in hospitalization frequency in their preliminary report. The reason for this difference is not entirely clear, but it may be due to patient selection. All of their patients were listed for heart transplantation, whereas most patients in the present study were not listed for transplantation. It is not uncommon for listed patients who decompensate to be hospitalized as status 1 patients until the time of transplantation, thereby avoiding further hospitalizations. Frequent use of this strategy by Fonarow et al could explain the dramatic reduction in hospitalization frequency.
To determine whether the availability of heart transplantation at our center was responsible for the reduction in hospitalization frequency, we examined annual hospitalization rates in patients not listed for heart transplantation (Fig 3⇑). Hospitalization rates were reduced by >50% even when patients listed for transplantation were excluded.
In any event, our results and the preliminary results of Fonarow et al suggest that heart failure programs directed by specialists in heart failure can markedly reduce hospitalization rates in heart failure. In contrast, two recent randomized studies suggest that management systems directed by nurses or primary care physicians may have less of an affect on patient outcomes. Rich et al6 examined the affect of a nurse-directed multidisciplinary approach on readmission rates in 282 elderly patients with heart failure and a high risk of readmission. Patients were randomized to standard treatment by their primary physician or to a treatment regimen consisting of intensive education, social-service consultation, an analysis of medications by a geriatric cardiologist, and intensive follow-up after discharge through home care services. Survival for 90 days without readmission, the primary end point of the study, was not significantly different between the two groups. When the analysis was restricted to survivors of the initial hospitalization, a modest but significant decrease in survival for 90 days without readmission was noted (controls, 67%; treatment, 54%; 19% reduction). The total number of readmissions was 94 in the control group versus 53 in the treatment group, reflecting a decrease of 44% (P<.02). However, this decrease appeared to be primarily due to a marked reduction in the number of readmissions by a small subgroup of patients readmitted at least two times.
As noted previously, investigators at Veterans Administration hospitals compared rehospitalization rates in 504 Veterans Administration patients with heart failure randomized to standard care versus close follow-up by a nurse and primary care physician.17 The patients subjected to intensive follow-up had a higher readmission rate/month and a 70% increase in days of rehospitalization compared with the control group.
It is important to emphasize, however, that patient selection and the methodology used by these randomized trials differed from the approach used both in the present study and by Fonarow et al.7 Our study and that of Fonarow et al compared changes in outcomes over time within a somewhat younger population. To clearly establish whether management by heart failure specialists produces better outcomes than management by nurses or primary care physicians, patients should be randomly assigned to these different systems, and outcomes should then be compared.
The decrease in heart failure admissions observed in the present study may have been due at least in part to more aggressive use of diuretic therapy. On average, loop diuretic doses were doubled after referral to treat obvious fluid overloading detected at the time of referral or documented at the time of Swan-Ganz catheterization. We and others have noted that a large number of patients with heart failure appear by physical examination to be in optimal fluid balance but, when evaluated with Swan-Ganz catheters, are found to have markedly elevated intracardiac pressures.18 19 20 Recently, we reported that these patients respond to aggressive diuresis with improved exertional symptoms.18 Stevenson et al19 20 also noted the utility of hemodynamic monitoring in managing patients with heart failure. At the time of referral, most patients were already receiving ACE inhibitors and digoxin.
The reduction in hospitalization frequency was accompanied by an increase in peak exercise V̇o2 and the anaerobic threshold. In a subgroup of unselected patients, Living with Heart Failure Questionnaire composite scores also improved. These findings suggest that patients were symptomatically improved after referral. This improvement may have been due to increased diuretic doses. In addition, patients were frequently encouraged to be inactive by referring physicians. After referral, patients were usually encouraged to become more active, an intervention that likely also improved exercise performance.
Survival at 1 year in our population was surprisingly good: 83%. A number of prior studies have demonstrated a relationship between peak exercise V̇o2 and survival21 22 23 and have suggested annual mortality rates of >25% in patients with peak V̇o2 levels similar to levels noted in our study population. The majority of deaths was due to sudden death rather than to progressive heart failure, whereas in large studies such as the SOLVD trial, the majority of deaths are usually due to heart failure.24 Stevenson et al23 also noted more sudden deaths than deaths due to heart failure in their population. The aggressive use of hemodynamic monitoring and correction of high intracardiac pressures both in our program and in that of Stevenson et al may decrease the frequency of death due to heart failure. However, it should be noted that the aggressive use of diuretics can increase activation of the renin-angiotensin system and produce electrolyte imbalances, changes that could decrease long-term survival and increase the incidence of sudden death.
Another factor that may have influenced survival and hospitalization outcomes was more appropriate use of antiarrhythmic agents and β-blockers. Antiarrhythmic agents were discontinued in patients being treated for nonsustained, asymptomatic ventricular tachycardia based on the lack of evidence that this therapy improves survival. Amlodipine was discontinued in 3 patients based on the lack of clear evidence that this intervention is beneficial. Finally, 8 patients were placed on β-blocker therapy, an intervention that may have improved survival.25 These changes are consistent with the suggestion of Ayanian et al26 that specialist physicians aware of key advances in a field generally use medications more appropriately than do generalists. Unfortunately, only a small number of patients had changes in their medication, so multivariate analysis could not be reliably used to determine the impact of such changes on outcomes.
In summary, our findings suggest that patients with heart failure have fewer hospitalizations and demonstrate functional improvement when their care is transferred from generalists to cardiologists and a support staff who specialize in the management of such patients.
This work was supported in part by a Grant-in-Aid from the National American Heart Association. We would like to thank Lori Miller, RN, and Ann Bishop, RN, for their assistance in data acquisition and analysis.
- Received April 7, 1997.
- Revision received June 5, 1997.
- Accepted June 19, 1997.
- Copyright © 1997 by American Heart Association
Bourassa MG, Gurne O, Bangdiwala SI, Ghali JK, Young JB, Rousseau M, Johnstone, DE, Yusuf, S. Natural history and patterns of current practice in heart failure. J Am Coll Cardiol. 1993;22(suppl A):14A-19A.
Beyond Four Walls: Cost-Effective Management of Chronic Congestive Heart Failure. Washington, DC: Advisory Board Company; 1994.
Fonarow GC, Stevenson LW, Walden JA, Stevenson LW, Walden JA, Hamilton MA, Natterson BJ, Steimle AE, Moriguchi Jaime D, Middlekauff, HR. Impact of a comprehensive management program on the hospitalization rate for patients with advanced heart failure. J Am Coll Cardiol. 1995(suppl):25:264A. Abstract.
Wilson JR, Groves J, Rayos G. Circulatory status and response to cardiac rehabilitation in patients with heart failure. Circulation. 1996;94:1567-1572.
Chomsky DB, Lang CC, Rayos GH, Shyr Y, Yeoh TK, Pierson RN, Davis SF, Wilson JR. Hemodynamic exercise testing: a valuable tool in the selection of cardiac transplantation candidates. Circulation. 1996;94:3176-3183.
Wasserman K, Whipp BJ, Koyal SN, Beaver WL. Anaerobic threshold and respiratory gas exchange during exercise. J Appl Physiol. 1973;35:236-243.
Wilson JR, Fink LI, Ferraro N, Dunkman WB, Jones RA. Use of maximal bicycle exercise testing with respiratory gas analysis to assess exercise performance in patients with congestive heart failure secondary to coronary artery disease or to idiopathic dilated cardiomyopathy. Am J Cardiol. 1986;58:601-606.
Rector TS, Kubo SH, Cohn JN. Patients self-assessment of their congestive heart failure: content, reliability and validity of a new measure, the Minnesota Living with Heart Failure questionnaire. Heart Failure. 1987;3:198-209.
Singh P. Managing chronic congestive heart failure in the home. Home Healthcare Nurse. 1995;13:11-13.
Cohn JN, Johnson G, Shabetai R, Loeb H, Tristani F, Rector T, Smith R, Fletcher R, for the V-HeFT VA Cooperative Studies Group. Ejection fraction, peak exercise oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation. 1993;87(suppl VI):VI-5-VI-16.
Mancini DM, Eisen H, Kussmaul W, Mull R, Edmunds Jr. LH, Wilson JR. Value of peak exercise oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation. 1991;83:778-786.