Depression is prevalent in patients with heart failure (HF) and is associated with increased mortality. In patients with coronary heart disease (CHD) without HF, exercise training (ET) can effectively decrease depressive symptoms resulting in improved survival. We evaluated 189 patients with American College of Cardiology/American Heart Association stage C HF due to CHD (mean left ventricular ejection fraction 35 ± 10%) enrolled in a structured ET program from January 2000 to December 2008, including a group of 151 who completed the program and 38 patients with HF who dropped out of rehabilitation without ET. Depressive symptoms were assessed by standard questionnaire at baseline and after ET, and mortality was determined at a mean follow-up of 4.6 ± 2.6 years. Prevalence of depressive symptoms decreased by 40% after ET, from 22% to 13% (p <0.0001). Patients initially classified as depressed who remained depressed after ET had nearly a fourfold higher mortality than patients whose depression resolved after ET (43% vs 11%, p = 0.005). Depressed patients who completed ET had a 59% lower mortality (44% vs 18%, p <0.05) compared to depressed dropout subjects not undergoing ET. Survival benefits after ET were concentrated to those patients with depression who improved exercise capacity. In conclusion, depressive symptoms are prevalent in patients with HF and are associated with increased mortality. Structured ET is effective in decreasing depressive symptoms, a factor that correlates with improved long-term survival.
We and others have described the benefits of exercise training (ET) on depressive symptoms in patients with coronary heart disease (CHD) after major CHD events. Specifically, ET decreases depressive symptoms and the prevalence of depression by 50% to 70% and in randomized trials decreased depressive symptoms as effectively as antidepressant medications. We recently reported a significant improvement in overall survival in depressed patients with CHD when depressive symptoms were successfully decreased after ET. However, a beneficial effect of ET on outcomes in patients with heart failure (HF) whose symptoms of depression decreased has not been reported. The present investigation examined the effects of ET on depressive symptoms and subsequent long-term mortality in patients with HF and evaluates the relation between changes in exercise capacity and clinical outcomes.
Methods
We enrolled patients with a left ventricular ejection fraction <45%, a history of HF (American College of Cardiology/American Heart Association stage C) that was considered by their clinical cardiologist to be stable, who had no previous depression or use of antidepressive medications, and whose HF was due to CHD. All patients were referred by their primary cardiologist after a recent CHD event including acute myocardial infarction (MI), coronary bypass, or percutaneous coronary intervention.
From January 2000 to December 2008, 189 patients met the inclusion criteria. All patients entered the program 2 to 6 weeks after a CHD event, including acute MI (38%), coronary bypass surgery (30%), and percutaneous coronary intervention (45% of patients; some patients had >1 clinical event). Thirty-eight patients dropped out of cardiac rehabilitation within 2 weeks of entry (“dropouts”), all of whom attended <5 sessions of exercise; the remaining 151 patients completed phase II cardiac rehabilitation and ET (completing all 36 sessions over a period of 3 to 7 months, mean 3.8) and formed the “treatment group.” All patients completed questionnaires before ET and the treatment group completed questionnaires after ET. Survival status was obtained June 2009, after a mean follow-up of 4.6 ± 2.6 years from the National Death Index. The protocol was approved by the institutional review committee at Ochsner Clinic Foundation (New Orleans, Louisiana).
The Kellner Symptom Questionnaire is a 92-question assessment validated to assess behavioral characteristics including symptoms of depression, anxiety, somatization, and hostility, with a lower score being more favorable for each behavioral symptom (scores can range from 0 to 17 U). The instrument has been validated for its ability to discriminate between psychiatric patients and non-psychiatric patients and for its test–retest and 1/2-split reliabilities. Depressive symptoms were defined to be clinically significant when the depression score exceeded 6, and when present, recovery from categorical depression was defined when the depression score was ≤6. The Medical Outcomes Short Form 36 Survey was used to assess quality of life, with a high score indicating a more favorable quality-of-life trait.
Protocol, data collection, and statistical analysis were performed as previously described. Patients were referred to and participated in outpatient phase II cardiac rehabilitation and ET consisting of 36 educational and exercise sessions typically over a 3-month period. Exercise intensity was prescribed within 10 beats of the anaerobic threshold obtained on entry cardiopulmonary exercise testing. Daily lectures and group sessions about CHD risk factors, diet, and HF (symptoms, signs, etc.) were given for patients and spouses. Height, weight, body mass index, and percent body fat were assessed at baseline and 1 week after completing ET. Peak oxygen uptake (V o 2 ) was measured by cardiopulmonary stress testing at baseline and on completion of the program as previously reported. Change in peak V o 2 was calculated as the difference between peak V o 2 after the program and at baseline.
StatView 5.0.1 (SAS Institute, Cary, North Carolina) was used for statistical analysis. Results are reported as mean ± SD or frequencies expressed as percentages. Differences in continuous variables between 2 groups were assessed by paired Student’s t test or nonparametric tests as appropriate. Univariate relations between variables were assessed as partial correlations. Two-tailed p value ≤0.05 was considered statistically significant. Logistic regression analysis was performed to determine independent predictors of mortality. Actuarial survival analysis was used to compute cumulative hazard over time.
Results
Mean age of the 189 patients who met the inclusion criteria was 65 ± 11 years, with men comprising 75% of the cohort. Mean left ventricular ejection fraction was 35 ± 10% and mean peak V o 2 was 15.5 ± 5.0 ml/kg/min. Sixty-one percent of patients were in New York Heart Association class II (23% in class III, 16% in class I). An implantable cardioverter–defibrillator was present in 59 patients (31%). One hundred seventy-four patients (92%) were taking β blockers, 145 (77%) were taking angiotensin-converting enzyme inhibitors, and 21 (11%) were taking angiotensin II receptor blockers. Depressive symptoms were identified in 42 patients (22%) on entry (24% dropouts, 22% treatment; p = NS). Compared to patients in the active treatment group, dropouts had a lower quality-of-life score (88.7 ± 18.9 vs 99.4 ± 18.6; p = 0.05) but were otherwise statistically similar.
Of the 151 patients comprising the treatment group, depressive symptoms were identified in 33 patients (22%) on entry into ET. Table 1 presents baseline differences between treatment subjects based on the presence or absence of significant depressive symptoms. Effects of ET in the depressed cohort are presented in Table 2 . After ET the prevalence of depressive symptoms decreased 40% ( Figure 1 ) from 22% on entry to 13% on completion of the program (p <0.0001).
| Variable | Depressed | Nondepressed | p Value |
|---|---|---|---|
| (n = 33) | (n = 118) | ||
| Age (years) | 62 ± 12 | 67 ± 12 | 0.06 |
| Men | 72% | 78% | NS |
| Body mass index (kg/m 2 ) | 28.6 ± 5.0 | 27.8 ± 4.6 | NS |
| Percent body fat | 28.8 ± 6.7 | 28.5 ± 7.4 | NS |
| Active smoker | 0% | 1% | NS |
| Hypertension | 34% | 32% | NS |
| Diabetes mellitus | 23% | 22% | NS |
| Peak oxygen uptake (ml/kg/min) | 14.6 ± 3.9 | 15.9 ± 4.6 | NS |
| Ejection fraction (%) | 34 ± 10 | 35 ± 10 | NS |
| Depression | 11.0 ± 3.7 | 2.1 ± 1.8 | <0.0001 |
| Anxiety | 12.2 ± 4.8 | 3.1 ± 3.2 | <0.0001 |
| Hostility | 7.8 ± 5.3 | 1.8 ± 2.6 | <0.0001 |
| Somatization | 10.0 ± 4.2 | 6.3 ± 3.5 | <0.0001 |
| Quality of life | 82.4 ± 12.6 | 103.7 ± 17.4 | <0.0001 |
| Variable | Before | After | p Value |
|---|---|---|---|
| Body mass index (kg/m 2 ) | 28.6 ± 5.0 | 28.3 ± 4.7 | NS |
| Percent body fat | 28.8 ± 6.7 | 27.6 ± 5.5 | 0.05 |
| Peak oxygen uptake (ml/kg/min) | 14.6 ± 3.9 | 15.9 ± 4.7 | 0.03 |
| Depression | 11.0 ± 3.7 | 5.1 ± 6.1 | <0.0001 |
| Anxiety | 12.2 ± 4.8 | 7.0 ± 6.4 | <0.0001 |
| Hostility | 7.8 ± 5.3 | 4.3 ± 5.2 | 0.003 |
| Somatization | 10.0 ± 4.2 | 7.5 ± 5.0 | 0.0006 |
| Quality of life | 82.4 ± 12.6 | 102.0 ± 19.9 | <0.0001 |
Of the dropouts, depressed subjects had a threefold higher all-cause mortality compared to nondepressed subjects (44% vs 14%, p <0.05).
After completion of ET, treatment group patients initially classified as depressed who remained depressed had a fourfold higher mortality (43% vs 11%, p = 0.005) than depressed patients who converted to a nondepressed status ( Figure 2 ). Depressed patients in the treatment group demonstrated an early and incremental increased mortality compared to nondepressed patients (p = 0.0003) by time-dependent actuarial hazard survival ( Figure 3 ).
By multivariate analysis ( Table 3 ), decreased peak V o 2 and advancing age were independent predictors of death in the treatment cohort, whereas decreased ejection fraction and presence of depression were of borderline significance.
| Variable | Chi-square | p Value |
|---|---|---|
| Decreased peak oxygen uptake | 7.7 | 0.006 |
| Advancing age | 6.3 | 0.01 |
| Decreased ejection fraction | 3.0 | 0.08 |
| Presence of depression | 2.9 | 0.09 |
To evaluate the impact of ET on survival in depressed subjects, we assessed mortality in all patients (treatment and dropouts) who were identified as being depressed at study entry. Depressed dropouts had a greater than twofold higher mortality than depressed patients who completed ET (44% vs 18%; p <0.05).
Treatment patients were divided into 2 groups based on change in exercise capacity after ET. Patients exhibiting a training effect, demonstrated by an increase in peak V o 2 after ET, were labeled “V o 2 gain” and were compared to those with no improvement or a loss in exercise capacity, labeled “V o 2 loss.” Baseline differences between these 2 groups are presented in Table 4 . Patients with V o 2 gain (n = 115) had higher total quality-of-life scores and trended toward having fewer smokers and lower depression scores than patients with V o 2 loss (n = 36).
| Variable | V o 2 Gain (n = 115) | V o 2 Loss (n = 36) | p Value |
|---|---|---|---|
| Age (years) | 66 ± 12 | 67 ± 12 | NS |
| Men | 75% | 80% | NS |
| Body mass index (kg/m 2 ) | 27.7 ± 4.8 | 28.1 ± 4.8 | NS |
| Active smoker | 0% | 3% | 0.08 |
| Hypertension | 25% | 33% | NS |
| Diabetes mellitus | 21% | 22% | NS |
| Peak oxygen uptake (ml/kg/min) | 15.7 ± 4.9 | 15.7 ± 4.9 | NS |
| Ejection fraction (%) | 34 ± 10 | 35 ± 11 | NS |
| Depression | 3.5 ± 4.0 | 4.9 ± 5.0 | 0.11 |
| Anxiety | 4.6 ± 4.9 | 5.7 ± 5.7 | NS |
| Hostility | 2.7 ± 3.7 | 3.8 ± 4.9 | NS |
| Somatization | 6.8 ± 3.9 | 7.4 ± 4.1 | NS |
| Quality of life | 102.1 ± 17.7 | 89.8 ± 18.9 | 0.0009 |
At follow-up, patients with V o 2 gain had a 61% lower mortality than patients with V o 2 loss (13% vs 33%; p = 0.006). Time-dependent actuarial cumulative hazard for survival was assessed in patients with V o 2 gain and V o 2 loss ( Figure 4 ). Patients with V o 2 loss showed an early increase in mortality compared to patients with V o 2 gain (p = 0.001).
