Author + information
- Received October 1, 1998
- Revision received January 22, 1999
- Accepted February 15, 1999
- Published online June 1, 1999.
- ↵*Reprint requests and correspondence: Dr. John R. Wilson, Cardiology Division, 315 MRB II, Vanderbilt University Medical Center, Nashville, Tennessee 37232-2170
The present study was undertaken to investigate the relationship over time between exertional symptoms in heart failure and functional capacity.
Most clinicians rely on exertional symptoms rather than on exercise testing to assess functional capacity in heart failure. However, it remains uncertain whether the subjective symptoms reported by patients provide a reliable index of functional capacity.
Fifty patients with heart failure underwent serial cardiopulmonary exercise testing and evaluation of exertional fatigue and dyspnea over a period of one to four years. Exercise testing was performed using the Naughton treadmill protocol and a MedGraphics metabolic cart. Fatigue and dyspnea were each scored from 0 to 3 (p = none, 1 = mild, 2 = moderate, 3 = severe). A composite symptom score was determined by adding together the fatigue and dyspnea scores.
Patients underwent a total of 185 tests at an average interval of 4.3 months (average tests/patient = 3.7). Composite symptom scores noted at the time of exercise testing correlated significantly with peak exercise minute oxygen consumption (VO2) (r = 0.47, p < 0.01). In addition, the change in symptoms scores and change in peak VO2noted between the baseline and final exercise test correlated significantly (r = 0.50, p < 0.01). However, patients reported few or no symptoms (symptom score ≤2) 45% of the time when peak VO2was <14 ml/min/kg, consistent with a severe functional disability, and 72% of the time when peak VO2was 14 to 18 ml/min/kg, consistent with moderate functional disability.
Exertional symptoms reported by patients with heart failure generally correlate with maximal exercise capacity. However, exertional symptoms frequently underestimate the severity of functional disability. Cardiopulmonary exercise testing rather than symptoms should be used to assess functional capacity in heart failure.
Functional disability, or an inability to perform activities normally, is one of the most common problems experienced by patients with heart failure (1–3). Therefore, one of the major goals of heart failure management is to improve functional capacity. To accomplish this, it is important that clinicians utilize reliable methods for detecting and monitoring the functional capacity of this population.
Most clinicians currently use the clinical history rather than formal exercise testing to assess and monitor the functional capacity of patients with heart failure. This strategy has also recently been recommended by the American Heart Association and American College of Cardiology guidelines for the management of heart failure (4). These guidelines recommend the use of exercise testing only in potential candidates for heart transplantation and specifically discourage the use of exercise testing as a method of following clinically stable patients with heart failure.
The relationship between symptoms reported by patients and functional capacity remains uncertain, however. Functional capacity indicates the ability of a patient to perform work and is traditionally measured by comparing a patient’s peak exercise capacity with the exercise capacity of age- and gender-matched normal control subjects. In contrast, the symptoms reported by patients reflect their subjective response to daily activities. Patients with heart failure often learn to avoid activities that produce exertional symptoms or may change the pace of activities to minimize symptoms. Therefore, the clinical history may provide unreliable information about the functional capacity of patients.
The present study was undertaken to further clarify the relationship over time between exertional symptoms reported by patients with heart failure and their functional capacity as measured by maximal cardiopulmonary exercise testing.
Studies were performed on 50 sequential ambulatory patients followed by the Heart Failure and Heart Transplantation Program at Vanderbilt University Medical Center. All patients had a history of chronic heart failure for over six months, and all patients were ambulatory and receiving digoxin, an angiotensin-converting enzyme inhibitor and diuretics. Forty-one patients were male and nine were female. The average age was 51 ± 9 years (range 39 to 68 years). The left ventricular ejection fraction averaged 26 ± 10%. Heart failure was attributed to coronary artery disease in 34 patients and to a cardiomyopathy in 16 patients. During exercise, all patients were limited by dyspnea or fatigue; no patient was limited by angina or intermittent claudication. Ten patients were New York Heart Association functional class II, 37 were class III and 3 were class IV.
As part of clinical care, patients routinely undergo cardiopulmonary exercise testing on initial evaluation by the Vanderbilt Heart Failure Program. Exercise testing is then repeated three to six months after initial evaluation and annually thereafter. Additional testing is performed if the patient reports major changes in exertional symptoms.
Exertional symptoms are recorded at each visit by quantifying the level of fatigue and dyspnea from 0 to 3 (0 = none; 1 = mild; 2 = moderate; 3 = severe). A composite symptom score was determined by adding the fatigue and dyspnea scores.
For the present study, all maximal exercise tests and symptom assessments performed on the patients were analyzed. Patients performed an average of 3.7 exercise tests at average intervals of 4.3 months. Exercise tests were excluded from analysis if the respiratory gas exchange ratio did not exceed 1.0 at peak exercise, a finding that suggests an inadequate patient effort (2,5).
Maximal exercise testing was performed using a 3-min Naughton protocol and a Marquette treadmill (Milwaukee, Wisconsin). During exercise, patients were connected to a CardioO2 System (MedGraphics, Minneapolis, Minnesota) via a disposable pneumotach. The patient was also attached to a Pulse Oximeter via a device on the finger to noninvasively measure arterial oxygen saturation. Exercise loads were increased every 3 min. The electrocardiogram and heart rate were monitored continuously. Arterial blood pressure was measured at rest and every 3 min during exercise using a cuff sphygmomanometer.
Peak exercise minute oxygen consumption (VO2) was defined as the maximal oxygen consumption level noted during exercise and was expressed as ml/min/kg body weight. The anaerobic threshold was defined using three criteria: the point after which the respiratory gas exchange ratio (minute carbon dioxide consumption [VCO2]/CO2) consistently exceeded the resting ratio, the point at which the ventilatory equivalent for oxygen (minute ventilation [VE]/VO2) was minimal followed by a progressive increase in VE/VO2and the VO2after which a nonlinear rise in VE occurred relative to VO2(2,6,7). When these three criteria identified different anaerobic thresholds, results of the different threshold were averaged. The anaerobic threshold was expressed as ml/min/kg of oxygen uptake.
During the course of the study, most patients exhibited improvements in symptoms and peak exercise VO2. These improvements were primarily related to adjustments in medications, particularly diuretic doses, and to efforts to increase patient activity levels.
All data are expressed as mean ± standard deviation. Correlations between variables were assessed using least square regression analysis. Comparison between group data was performed using Student paired or unpaired ttesting, as appropriate. A p value < 0.05 was considered statistically significant.
The 50 study patients underwent a total of 185 cardiopulmonary tests at an average interval of 4.3 months (average tests/patient = 3.7). At baseline exercise testing, peak exercise VO2averaged 14.7 ± 4.7, the anaerobic threshold averaged 11.2 ± 1.9 ml/min/kg and composite symptom scores averaged 2.9 ± 1.8. During the study period, peak exercise VO2improved to 16.1 ± 4.3 (p < 0.02 vs. baseline), whereas symptom scores improved to 1.4 ± 1.8 (p < 0.01 vs. baseline).
Composite symptom scores correlated significantly with peak exercise VO2both at the time of baseline exercise testing (r = 0.45) and throughout the study period (r = 0.47) (both p < 0.001) (Fig. 1and Fig. 2). Composite symptom scores also correlated significantly with peak exercise minute ventilation (r = 0.24), peak VCO2/kg (r = 0.47), the ratio of peak minute ventilation to peak VCO2(r = 0.48) and the anaerobic threshold (r = 0.378) (all p < 0.001). There was no significant correlation between composite symptom score and the peak exercise respiratory gas exchange ratio.
Changes in symptom scores from the baseline to final exercise test correlated significantly with changes in peak exercise VO2 (r = 0.50, p < 0.01) (Fig. 3). Regression equations suggested that exertional symptoms on average developed when peak VO2decreased below 18 ml/min/kg, with a decrease in peak VO2of 1 ml/min/kg being associated with an increase in symptom score of approximately 1.
However, symptom scores often underestimated the severity of functional disability (Fig. 4). Peak exercise VO2was <14 ml/min/kg, consistent with severe functional limitation, during 65 exercise tests. During these tests, patients reported minimal or no clinical symptoms (composite scores ≤2) 45% of the time. Peak exercise VO2was 14–18 ml/min/kg, consistent with moderate functional limitation, during 74 exercise tests. At the time of these tests, 72% of the patients reported minimal or no clinical symptoms. Peak exercise VO2was >18 ml/min/kg during 54 tests. At the time of these tests, 87% of patients reported minimal or no clinical symptoms. The anaerobic threshold occurred at a similar percentage of peak exercise VO2in patients reporting minimal or no clinical symptoms (73 ± 12%) versus patients with composite scores >2 (74 ± 13%) (p = NS).
Symptom changes noted in individual patients over the study also tended to correlate poorly with changes in functional capacity. When individual changes in peak exercise VO2were examined, 18 patients exhibited a >20% increase in peak VO2(11.3 ± 3.6 to 17.2 ± 4.7 ml/min/kg), while 8 patients had a >20% decrease in peak VO2(18.2 ± 2.6 to 12.9 ± 2.4 ml/min/kg) (both p < 0.01). On average, patients with a >20% increase in peak VO2had a highly significant improvement in symptom score, from 3.3 ± 1.9 to 1.2 ± 1.8 (p < 0.0001). However, eight of the patients exhibited minimal or no changes in symptom score (change in symptom score = 0–1). Patients with a >20% decrease in peak VO2actually had no significant change in symptom score (2.8 ± 1.5 to 3.1 ± 2.0).
Coronary artery disease versus cardiomyopathy
To determine if the relationship between symptoms and peak VO2differed in the patients with coronary artery disease versus those with a cardiomyopathy, analysis was performed on these two subgroups. At baseline, both groups had similar symptom levels (coronary disease: 2.7 ± 1.7 vs. cardiomyopathy: 3.0 ± 1.9) and peak VO2(coronary disease: 15.2 ± 4.7 vs. 14.1 ± 4.6 ml/min/kg) (both p = NS). During the course of the study, patients with coronary disease exhibited significantly less change in peak VO2when compared with patients with a cardiomyopathy (0.2 ± 3.3 vs. 2.9 ± 5.4 ml/min/kg [p < 0.04]) but exhibited similar changes in symptom scores (−1.4 ± 1.7 vs. 1.5 ± 2.3 [p = NS]).
Over the past decade, exercise testing has been used widely in research studies to assess the functional capacity of patients with heart failure and to evaluate the effect of therapeutic interventions (1–3,5). In contrast, exercise testing is rarely used by clinicians managing patients with heart failure. Instead, clinicians rely almost exclusively on the clinical history to assess the functional capacity of patients with heart failure and to track responses to therapy.
This approach is based on the assumption that the symptoms reported by a patient with heart failure are related to the patient’s functional ability. This assumption may appear intuitively correct. However, the symptoms reported by a patient potentially could have little or no relationship to functional capacity. Functional capacity is a measure of a person’s ability to perform work. In contrast, the exertional symptoms reported by a patient reflects the patient’s subjective response to activities actually performed and therefore will be influenced by the intensity of activities undertaken and by psychological factors.
The present study was undertaken to investigate the relationship between symptom levels reported by patients with heart failure and an objective index of exercise capacity: peak exercise VO2. To assess exertional symptoms, exertional dyspnea and fatigue were each scored separately and then added together to obtain a composite score. Peak exercise VO2was used to objectively assess maximal exercise capacity based on extensive data, suggesting that this index provides the most accurate and reproducible index of peak exercise capacity available (2,5). Exercise tests were included in the analysis only if the respiratory gas exchange ratio, the ratio of VCO2to VO2, exceeded 1.0 at peak exercise, suggesting adequate patient effort (2,5).
Results of this study indicate that there is a definite relationship between symptoms reported by patients and functional capacity. Symptom scores compiled at the baseline exercise test correlated with peak exercise VO2(Fig. 1). A similar correlation was observed when results from all 185 exercise tests were compared with symptom scores (Fig. 2). Changes in symptom score over the study period also correlated with changes in peak exercise VO2(Fig. 3).
The correlations noted between peak exercise VO2and symptom scores indicated that exertional symptoms were on average minimal or absent when the peak exercise VO2achieved a level of approximately 18 ml/min/kg. A decrease in VO2of 1 ml/min/kg below this level was associated with an increase in symptom score of approximately 1. Therefore, one would predict severe symptoms (composite symptom score = 6) at a peak VO2level of approximately 12. This finding is consistent with prior observations. Weber et al. considered a peak VO2>20 ml/min/kg to be associated with a relatively asymptomatic state, and a peak VO2<10 to be associated with severe functional limitation (2). Solal and Gourgon noted that most patients with few exertional symptoms had a peak VO2level >18 ml/min/kg, and that patients with severe exercise limitation had an average peak VO2level of 12 ml/min/kg (5).
Such observations suggest that exertional symptoms parallel functional capacity to some extent. However, the degree of correlation between symptoms and peak exercise VO2noted in this study was extremely weak. For any given symptom level, patients exhibited widely varying levels of peak exercise VO2. Of particular note, symptom scores tended to underestimate the level of functional disability. Patients who had peak exercise VO2levels <14 ml/min/kg, consistent with moderate to severe functional disability, reported minimal or no symptoms 45% of the time. When peak exercise levels were between 14 and 18 ml/min/kg, consistent with mild to moderate functional disability, patients reported minimal or no symptoms 72% of the time. The relationship between respiratory gas data and symptoms was not improved by substituting the anaerobic threshold, peak exercise ventilation or the ratio of minute ventilation to carbon dixoide production for peak exercise VO2.
Changes in symptoms over time also did not track closely with changes in functional capacity. Although patients who exhibited a >20% increase in peak exercise VO2during the study exhibited a highly significant improvement in symptoms as a group, 44% of these patients reported minimal or no change in symptom scores. Even a larger percentage of patients who exhibited a >20% decrease in peak VO2reported minimal or no change in symptom scores.
These observations strongly suggest that clinical symptoms do not provide a reliable index of functional capacity in heart failure. In particular, reliance on clinical symptoms is likely to underestimate the extent of functional disability experienced by patients, because many functionally disabled patients appear to report few if any clinical symptoms.
Why try to identify patients with reduced functional capacity if they report few exertional symptoms? It is very likely that such patients are functionally limited during normal daily activities, but have altered their lifestyle to adapt to their disability. Such patients should be made aware of their disability and efforts made to improve their functional capacity. Many of these patients are probably deconditioned and may improve with an exercise program (8–10). Some patients may have associated lung dysfunction and benefit from bronchodilators. Other patients may have unsuspected fluid retention and benefit from additional diuresis (11). Still other patients may benefit from medication adjustments.
This management strategy is likely to substantially improve the overall functional ability of patients with heart failure. In addition, correcting functional disability may potentially improve prognosis. A number of groups have demonstrated that peak exercise VO2correlates with survival in heart failure, suggesting that exercise capacity influences prognosis, possibly via an impact on neurohomones (12–14). Stevenson et al. (15)have also demonstrated that improvements in peak exercise VO2are associated with improved clinical stability and with fewer hospitalizations. However, it should be emphasized that this study did not address the impact of therapies on functional capacity. Further studies are needed to determine if cardiopulmonary exercise testing can be used to guide therapeutic interventions.
In summary, results of this study suggest that the clinical history does not provide a reliable method of detecting and monitoring functional capacity in heart failure. Formal exercise testing should be used more frequently in this population to identify patients with substantial functional disability.
☆ This work was supported in part by Grant RO-1 HL53059 from the National Institutes of Health, Bethesda, Maryland and by a Grant-in-Aid from the National American Heart Association, Dallas, Texas.
- minute ventilation
- minute oxygen consumption
- minute carbon dioxide production
- Received October 1, 1998.
- Revision received January 22, 1999.
- Accepted February 15, 1999.
- American College of Cardiology
- Wilson J.R.,
- Martin J.L.,
- Schwartz D.,
- Ferraro N.
- Weber K.,
- Kinasewitz G.,
- Janicki J.,
- Fishman A.
- Williams J.F.,
- Bristow M.R.,
- Fowler M.B.,
- et al.
- Wasserman K.,
- Whipp B.J.,
- Koyal S.N.,
- Beaver W.L.
- Wilson J.R.,
- Fink L.I.,
- Ferraro N.,
- Dunkman W.B.,
- Jones R.A.
- Coats A.J.S.,
- Adamopoulos S.,
- Radaelli A.,
- et al.
- Sullivan M.J.,
- Higginbotham M.B.,
- Cobb F.R.
- Wilson J.R.,
- Groves J.,
- Rayos G.
- Mancini D.M.,
- Eisen H.,
- Kussmaul W.,
- Mull R.,
- Edmunds L.H. Jr.,
- Wilson J.R.
- Cohn J.N.,
- Johnson G.R.,
- Shabetai R.,
- Loeb H.,
- et al.
- Stevenson L.W.,
- Steimle A.E.,
- Fonarow G.,
- et al.