Modest Increase in Peak VO₂ is Related to Better Clinical Outcomes in Chronic Heart Failure Patients: Results from Heart Failure and a Controlled Trial to Investigate Outcomes of Exercise Training (HF-ACTION)

Abstract

Background

The prognostic ability of a single measurement of peak oxygen uptake (VO₂) is well established in patients with chronic heart failure (HF). The relation between a change in peak VO₂ and clinical outcomes is not well defined.

Methods and Results

This investigation determined if an increase in peak VO₂ was associated with a lower risk of the primary endpoint of time to all-cause mortality or all-cause hospitalization and three secondary endpoints. In Heart Failure and a Controlled Trial to Investigate Outcomes of Exercise Training (HF-ACTION), an exercise training trial for patients with systolic HF, cardiopulmonary exercise tests were performed at baseline and approximately three months later in 1620 participants. Median peak VO₂ in the combined sample increased from 15.0 (11.9–18.0 Q1–Q3) to 15.4 (12.3–18.7 Q1–Q3) mL-kg⁻¹-min⁻¹. Every 6% increase in peak VO₂, adjusted for other significant predictors, was associated with a 5% lower risk of the primary endpoint (HR = 0.95; CI = 0.93–0.98; p < 0.001); a 4% lower risk of the secondary endpoint of time to cardiovascular mortality or cardiovascular hospitalization (HR = 0.96; CI = 0.94–0.99; p < 0.001); an 8% lower risk of cardiovascular mortality or HF hospitalization (HR = 0.92; CI = 0.88–0.96; p < 0.001) and a 7% lower all-cause mortality (HR = 0.93; CI = 0.90–0.97; p < 0.001).

Conclusions

Among patients with chronic systolic HF, a modest increase in peak VO₂ over three months was associated with a more favorable outcome. Monitoring the change in peak VO₂ for such patients may have benefit in assessing prognosis.

Peak oxygen uptake (peak VO₂), the standard for assessing cardiovascular fitness, is a strong prognostic indicator in chronic heart failure (HF) and is a criterion variable for consideration of cardiac transplantation in such patients (1, 2). Peak VO₂, when reported relative to body mass, is responsive to a number of interventions, including drug and device therapies, exercise training and changes in body weight (3–6). While the prognostic ability of a single measurement of peak VO₂ is well established, the relation between a change in peak VO₂ and clinical outcomes is not well defined. Prior studies have involved relatively small samples, and the findings are not consistent (7–11). Studying the change in peak VO₂ and its relation with clinical outcomes would be beneficial for assessing the course of HF and to evaluate the result of a therapeutic intervention such as exercise training, medication changes, or device therapy. In addition, the change in peak VO₂ may have prognostic implications for individuals with chronic HF (8).

Heart Failure and A Controlled Trial Investigating Outcomes of Exercise Training (HF-ACTION) was a multicenter clinical trial designed to study the effects of exercise training on mortality and morbidity in 2,331 patients with New York Heart Association (NYHA) class II to class IV chronic systolic HF receiving guidelines-based optimal medical therapy. Eighty-two clinical sites participated in this trial, with 67 from the United States, nine from Canada, and six from France. Patients enrolled in HF-ACTION were followed for a maximum of four years after baseline testing, with a median follow-up time of 30.1 months. After the first three months the exercise training group demonstrated a modest increase in peak VO₂ in comparison to the usual care group [0.6 (Q1 to Q3 = −0.7 to 2.3) versus 0.2 (Q1 to Q3 = −1.2 to 1.4)] mL-kg⁻¹-min⁻¹, p<.001 (12). HF-ACTION was designed to be sufficiently powered to assess the impact of a change in peak VO₂ on clinical outcomes. Thus, the specific aims of the current analysis were two: 1) To determine if an increase in peak VO₂ was associated with a lower risk of the composite primary endpoint of time to all-cause mortality or all-cause hospitalization and 2) To determine if an increase in peak VO₂ was associated with a lower risk of three secondary endpoints, including time to cardiovascular mortality or cardiovascular hospitalization; cardiovascular mortality or HF hospitalization, and all-cause mortality.

Methods

The eligibility characteristics for subject enrollment and methodology for the exercise testing and training protocols for HF-ACTION have been presented previously (13). Subjects had left ventricular ejection fractions (LVEF) of 35% or less by echocardiography and NYHA class II to class IV HF symptoms despite optimal HF therapy for at least 6 weeks. Cardiopulmonary exercise (CPX) testing was performed pre-randomization and at approximately three months following randomization using an extended, modified Naughton treadmill protocol or ramp (10W/min) stationary cycle protocol. There were 1620 subjects who underwent CPX testing at the pre-randomization and 3-month time points and who did not experience the primary endpoint prior to the 3-month CPX test. The data from these 1620 subjects, (combined exercise training and usual care groups of HF-ACTION) are presented in this paper. Figure 1 shows the flow of these 1620 subjects through this trial. Of these 1620 subjects, 92% were tested on a treadmill while 8% used a cycle ergometer. During CPX testing, subjects were encouraged to achieve a Rating of Perceived Exertion of 17 (hard to very hard) or greater on the 6 to 20 Borg scale and a respiratory exchange ratio of greater than 1.10. Gas exchange results were forwarded to the CPX core laboratory where two independent experts who were blinded to the subject’s treatment assignments, determined peak VO₂. Peak VO₂ was operationally defined as the highest VO₂ for a given 15 or 20 second interval within the last 90 seconds of exercise or the first 30 seconds of recovery (13). Following the initial CPX test, subjects were randomized to either usual care alone or usual care and exercise training. Due to the nature of the intervention, blinding for subjects and site investigators was not possible. However, deaths and cardiovascular hospitalizations for each patient were adjudicated by a clinical end point committee blinded to treatment assignment. Once a patient had an adjudicated HF hospitalization, no further hospitalizations for that patient were reviewed (12).

Figure 1.

Flow of patients through HF-ACTION clinical trial

Statistical analysis

Subject characteristics were summarized by percentages for categorical variables and median with inter-quartile range for continuous variables. Model development for the primary endpoint of all-cause mortality or all-cause hospitalization and secondary endpoints of cardiovascular mortality or cardiovascular hospitalization; cardiovascular mortality or HF hospitalization and for the all-cause mortality endpoint proceeded similarly. In each case, the procedure began with imputation of missing values of potential predictors identified in the main outcomes paper (12) as well as additional predictors related to quality of life questionnaire responses and oxygen pulse. The imputation algorithm used the values of the all-cause mortality or all-cause hospitalization endpoint to impute the values of missing potential predictors in the imputation process. While 78% of the potential predictors had <1% missing data, an additional 18% had 1–5% missing data. Only 2 potential predictors, VO₂ at ventilatory threshold (17%), and hemoglobin (24%) had >15% missing data.

The univariate relationship of each continuous variable with outcome was checked for linearity of the log hazard ratio, and piecewise linear splines were used as transformations when appropriate (14). A bootstrapped backwards selection algorithm of a Cox proportional hazards model was used for model selection, and the C-index was used to choose the final model (15). A similar process was followed in the modeling of the change in peak VO₂, using linear regression rather than Cox proportional hazard and the adjusted model R square statistic rather than the C-index. All Cox analyses used the date of the subject’s 3-month CPX test for time zero, which provided the basis for days to endpoint or censoring. Based on prior studies showing that a 6% change in peak VO2 is clinically meaningful in HF patients (7), the effect of a 6% increase in peak VO2 on primary and secondary outcomes was assessed in separate multivariate models for each clinical outcome. In addition, tertiles of change in peak VO₂ were calculated in the overall sample. Trend tests across the tertiles of percent change in peak VO₂ for each clinical outcome were done using models in which the tertiles of peak VO₂ change were fit as a continuous variable. Survival curves adjusted for significant covariates were constructed by calculating the average survival across all of the patients (16). A 2-tailed p value of < 0.05 was required to reject the null hypothesis.

Results

Table 1 presents relevant subject characteristics at baseline. Median age for the group was 59 years and 28% of patients were women. Baseline median LVEF was 25% and HF was of an ischemic etiology in 52% of the population studied. Ninety five percent of subjects were taking an ACE inhibitor or angiotensin receptor blocker (ARB). Ninety five percent of subjects were also on beta-blocker therapy. Sixty-seven percent of the subjects were classified as NYHA class II and 32% as NYHA class III. The median baseline peak VO₂ was 15.0 ml-kg⁻¹-min⁻¹, and median baseline VO₂ at ventilatory threshold was 10.8 ml-kg⁻¹-min⁻¹.

Table 1.

Subject Characteristics at Baseline

Characteristic	Statistic	Patients with a 3- month CPX*
Age at randomization	N, median, (Q1, Q3)	1620 59.4 (51.6 – 67.9)
Female	n/N (%)	457/1620 (28)
BMI	N, median, (Q1, Q3)	1618 29.7 (26.0 – 34.4)
Weight (kg)	N, median, (Q1, Q3)	1620 89.8 (76.0 – 105.0)
Race
Black or African American	n/N (%)	485/1620 (30)
White	n/N (%)	1038/1620 (64)
Other	n/N (%)	75/1620 (5)
Unknown	n/N (%)	22/1620 (1)
Geographical region
West U.S.A.	n/N (%)	199/1620 (12)
Midwest U.S.A.	n/N (%)	495/1620 (31)
Northeast U.S.A.	n/N (%)	172/1620 (11)
South U.S.A.	n/N (%)	567/1620 (35)
Canada	n/N (%)	150/1620 (9)
France	n/N (%)	37/1620 (2)
NYHA class at randomization
II	n/N (%)	1091/1620 (67)
III	n/N (%)	518/1620 (32)
IV	n/N (%)	11/1620 (1)
Clinical history at baseline
History of diabetes	n/N (%)	503/1620 (31)
History of hypertension	n/N (%)	956/1612 (59)
History of myocardial infarction	n/N (%)	691/1620 (43)
Prior CABG	n/N (%)	418/1620 (26)
Prior valve surgery	n/N (%)	77/1620 (5)
Prior PCI	n/N (%)	367/1620 (23)
History of COPD	n/N (%)	175/1604 (11)
History of peripheral vascular disease	n/N (%)	96/1611 (6)
History of depression	n/N (%)	322/1620 (20)
History of atrial fib. or atrial flutter	n/N (%)	338/1620 (21)
Renal dysfunction	n/N (%)	18/1447 (1)
Medications and Devices at baseline
On a bi-ventricular pacemaker at baseline	n/N (%)	292/1620 (18)
On pacemaker at baseline	n/N (%)	284/1620 (18)
On an ACE inhibitor	n/N (%)	1207/1620 (75)
Ace inhibitor or ARB	n/N (%)	1546/1620 (95)
On a beta blocker	n/N (%)	1537/1620 (95)
Ischemic etiology of heart failure	n/N (%)	838/1620 (52)
Left ventricular ejection fraction	N, median, (Q1, Q3)	1619 24.8 (20.3 – 30.3)
CPX baseline parameters
Rest ECG rhythm, CPX test	n/N (%)	1291/1597 (81)
Sinus	n/N (%)	92/1597 (6)
Atrial fibrillation and other	n/N (%)	214/1597 (13)
Resting ventricular conduction
Normal	n/N (%)	689/1580 (44)
LBBB	n/N (%)	267/1580 (17)
RBBB	n/N (%)	61/1580 (4)
IVCD	n/N (%)	200/1580 (13)
Paced	n/N (%)	363/1580 (23)
Resting heart rate/blood pressure
Heart rate (bpm)	N, median, (Q1, Q3)	1618 69 (62 – 76)
Systolic blood pressure (mmHg)	N, median, (Q1, Q3)	1619 112 (102 – 126)
Diastolic blood pressure (mmHg)	N, median, (Q1, Q3)	1619 70 (62 – 80)
CPX maximal parameters
HR at peak exercise (bpm), CPX test	N, median, (Q1, Q3)	1620 120.0 (105.0 – 134.0)
Peak respiratory exchange ratio, CPX test	N, median, (Q1, Q3)	1606 1.09 (1.02 – 1.16)
Peak VO2 (ml-kg−1-min−1), CPX test	N, median, (Q1, Q3)	1620 15.0 (11.9 – 18.0)
VO2 at ventilatory threshold, CPX test	N, median, (Q1, Q3)	1401 10.8 (9.1 – 12.7)
VE-VCO2 slope, CPX test	N, median, (Q1, Q3)	1603 32.6 (28.1 – 38.3)

Abbreviations: ACE = angiotensin converting enzyme; ARB = angiotensin receptor blocker; BMI = body mass index; CABG = coronary artery bypass graft; COPD = chronic obstructive pulmonary disease; CPX = cardiopulmonary exercise test; IVCD = intra-ventricular conduction defect; LBBB = left bundle branch block; NYHA = New York Heart Association Class; PCI = percutaneous coronary intervention; PVD = peripheral vascular disease; RBBB = right bundle branch block; VO₂ = oxygen uptake; VE = ventilation; VCO₂ = volume of carbon dioxide produced

^*Subset of patients from HF-ACTION cohort who were not missing baseline or 3-month peak VO₂, and free of primary end point events through the 3-month measurement

Table 2 presents the relation between the change in peak VO₂ at three months and the primary and secondary endpoints after accounting for significant predictors. The significant predictors for each clinical outcome are also presented in Table 2. An increase in peak VO₂ from baseline to three months was related to a significantly lower risk for the primary outcome of time to all-cause mortality or all-cause hospitalization (HR = 0.95; CI = 0.93–0.98; p < 0.001). For each 6% increase in peak VO₂, there was a 5% lower risk for this endpoint. A 6% increase in peak VO₂ from baseline to three months was related to a 4% lower risk of the secondary endpoint of cardiovascular mortality or cardiovascular hospitalization (HR = 0.96; CI = 0.94–0.99; p < 0.001). A 6% increase in peak VO₂ from baseline to three months was related to an 8% lower risk of the secondary endpoint of cardiovascular mortality or heart failure hospitalization (HR = 0.92; CI = 0.88–0.96; p < 0.001). A 6% increase in peak VO₂ from baseline to three months was related to a 7% lower risk of the secondary endpoint of all-cause mortality (HR = 0.93; CI = 0.90–0.97; p < 0.001). Similarly significant though smaller HR reductions were seen for a 4% increase in peak VO2, the mean change in the training arm (data not shown).

Table 2.

Relation of 6% increase in peak VO₂ at 3 months to clinical outcomes from multivariable models

Outcome	p-value	Hazard ratio (95% CI)	C-Index
All-cause mortality or all-cause hospitalization	<.0001	0.95 (0.93–0.98)	0.64
Cardiovascular mortality or hospitalization	0.001	0.96 (0.924–0.99)	0.74
Cardiovascular mortality or HF hospitalization	0.001	0.92 (0.88–0.96)	0.65
Mortality	0.001	0.93 (0.90–0.97)	0.74

Abbreviations: VO₂ = oxygen uptake

Covariates in the final model for each clinical end point with treatment arm forced into all models.

All-cause mortality or all-cause hospitalization: change in peak VO₂ at 3 months; baseline peak VO₂; KCC total symptom score; ventricular conduction on CPX; LVEF; mitral regurgitation grade; on nitrate; BUN; gender; region; beta-blocker dose

Cardiovascular mortality or cardiovascular hospitalization: change in peak VO₂ at 3 months; ventricular conduction on CPX; LVEF; mitral regurgitation grade; on nitrate; BUN; gender; region; beta-blocker dose

Cardiovascular mortality or heart failure hospitalization; change in peak VO₂ at 3 months; ventricular conduction on CPX; LVEF; mitral regurgitation grade; BUN; loop diuretic dose; gender; peak RER on CPX; region; beta-blocker dose

Mortality: change in peak VO₂ at 3 months; ventricular conduction on CPX; LVEF; mitral regurgitation grade; BUN; loop diuretic dose; gender; marital status; baseline peak VO₂; region and beta-blocker dose

Abbreviations: CPX = cardiopulmonary exercise test; LVEF = left ventricular ejection fraction; BUN = blood urea nitrogen; RER = respiratory exchange ratio; KCC = Kansas City Cardiomyopathy scale

Table 3 presents trend tests for each clinical outcome. With the exception of cardiovascular mortality or heart failure hospitalization, a significant trend for better outcomes was observed from lowest to highest tertiles of peak VO₂ change.

Table 3.

Trend Tests for Clinical Outcomes Across Tertiles of Peak VO₂ Change

Clinical Outcome	Estimate	Chi-Square	p-value
All-cause mortality or all-cause hospitalization	−0.15	13.08	<0.001
Cardiovascular mortality or hospitalization	−0.12	7.35	0.007
Cardiovascular mortality or HF hospitalization	−0.15	5.88	0.061
Mortality	−0.18	4.62	0.032

Figure 2 presents adjusted survival curves for the 1620 subjects for each of the clinical endpoints stratified by change in peak VO₂. Only the primary end point of time to all-cause mortality or all-cause hospitalization demonstrated a significant difference in event rates. Specifically, individuals whose peak VO₂ decreased by greater than 6% experienced a 9% lower event rate at 3 years compared to those individuals whose peak VO₂ changed less than this amount. Figure 3 presents an algorithm for using the change in peak VO₂ as assessed in the current study with repeat CPX testing and evaluating HF prognosis for the clinician. This algorithm allows the clinician some opportunity to consider alternative therapies based on the measured change in peak VO₂

Figure 2.

Adjusted survival curves for the 1620 subjects who completed pre-randomization and 3 month CPX tests categorized by direction of changes in peak VO₂ over time for the primary outcome of all-cause mortality or all-cause hospitalization and three secondary endpoints, including time to cardiovascular mortality or cardiovascular hospitalization; cardiovascular mortality or HF hospitalization, and all-cause mortality.

Figure 3.

Clinical algorithm relating change in peak VO₂ from repeat CPX testing to HF outcomes and therapy recommendations.

Discussion

HF-ACTION is the largest study to address whether a change in peak VO₂ is related to mortality and morbidity clinical endpoints for individuals with chronic HF. The results confirm the hypotheses that an increase in peak VO₂ was significantly related to a lower event rate for the primary and each of the three secondary clinical endpoints. Specifically, for each 6% increase in peak VO₂ over three months there was a 5% lower risk for the primary outcome of all-cause mortality and all-cause hospitalization; a 4% lower risk for the secondary endpoint of time to cardiovascular mortality or cardiovascular hospitalizations; an 8% lower risk for cardiovascular mortality or HF hospitalizations, and a 7% decreased risk for all-cause mortality after accounting for other significant predictors. A 6% increase in peak VO₂ was achieved by 33% of the patients randomized to usual care and 44% of the patients randomized to exercise training.

Other studies (7–11) that have evaluated the change in peak VO₂ and related the change to clinical outcomes in HF patients have had relatively small sample sizes, and their findings have not been consistent. Corra et al (7) studied 231 stable chronic HF patients, of which 200 were men, with a peak VO₂ of 14.3 ± 8 ml-kg⁻¹-min⁻¹ and LVEF of 24 ± 8%, who underwent two CPX tests separated by 258 ± 42 days. They observed that the change in peak VO₂ had significant prognostic ability. Survival analysis indicated that 51% of subjects having a decrease in peak VO₂ did not survive for the follow-up period of 1167 ± 562 days compared to 14% of subjects with an increase in peak VO₂ (3). Stevenson et al, (10) evaluated 83 HF patients who underwent CPX testing between 3 and 12 months after their initial evaluation. Patients who demonstrated an increase in peak VO₂ of greater than 2 ml-kg⁻¹-min⁻¹ or a peak VO₂ greater than 12 ml-kg⁻¹-min⁻¹ showed a better survival rate. Florea et al, (8) studied 62 chronic HF patients, of which 58 were men, with a mean peak VO₂ of 18.2 ± 5.9 ml-kg⁻¹-min⁻¹ and a mean LVEF of 38.9 ± 15.8% and found that the change in peak VO₂ monitored over 24 months, predicted non-transplant survival. In contrast, Gullestad et al, (9) reported that the change in peak VO₂ assessed by two separate CPX tests performed 7.8 months apart did not add any prognostic information in 283 HF patients with a mean peak VO2 of 16.7 ± 0.3 ml-kg⁻¹-min⁻¹ being evaluated for heart transplantation. In 155 stable chronic HF patients studied before and after 20 supervised exercise training sessions, Tabet et al (11) found that the change in peak VO₂ and B-type natriuretic peptide level were the only significant independent predictors of clinical outcomes. Similar to our own findings, the investigators found that the lack of improvement in exercise capacity had prognostic value for adverse clinical events. Finally, in 417 coronary patients without evident heart failure, Vanhees et al (17) observed that every 1% increase in peak VO₂ after a 3-month training program was associated with a 2% lower risk of cardiovascular death.

Several possible factors might contribute to a change in peak VO₂ over time, including random variation, a familiarization effect with repeated testing, the natural history of the disease, a change in physical activity patterns, a change in drug regimen, or cardiac resynchronization (biventricular) pacing therapy (3–6). In this study, exercise training is a potential reason for the observed changes in peak VO₂ because it represents the intervention under investigation. Despite exercise group assignment not appearing as an independent covariate in predicting clinical outcomes, a significantly greater increase in mean peak VO₂ (0.6 ml-kg⁻¹-min⁻¹ for exercise training and usual care in comparison to an increase of 0.2 ml-kg⁻¹-min⁻¹ for usual care alone, respectively) was observed after 3 months (12). The majority of single site randomized controlled trials have observed a significant increase in peak VO₂ with exercise training for individuals with chronic HF (18–27), although not all studies have observed significant changes (28–32). Single site studies however, are typically not designed to address the relation between the change in peak VO₂ and clinical outcomes, nor are they powered to do so (27).

The mechanism for an increase in peak VO₂ impacting clinical outcomes related to mortality and morbidity likely involves a number of factors. The determinants of peak VO₂ include central (heart rate, stroke volume, cardiac output) and peripheral (muscle oxygen extraction) components. Each of these factors tends to respond favorably (positively) to exercise training for most individuals with chronic HF. Exercise training induces significant increases in maximal heart rate (25); stroke volume (23), cardiac output (33, 34), arterial-venous O₂ difference (35), skeletal muscle blood flow, likely related to reversal of endothelial dysfunction (22, 24); skeletal muscle oxidative function (19, 21, 29, 34, 36); attenuation of sympathetic activation (18, 28, 33, 37); and increases in chronotropic responsiveness in such patients (18, 26). Exercise training also attenuates the production of pro-inflammatory cytokines and natriuretic peptides (38, 39). In addition to the responses already noted, increased blood vessel formation, a higher threshold for arrhythmias and sudden death, and healthier lifestyle decisions that accompany a change in fitness level may also help explain the positive relation between an increase in peak VO₂ and clinical outcomes (20).

This investigation had the following limitations. The HF patients enrolled in this study were younger and with fewer co-morbid conditions compared with the general population with HF, and this study only included patients with reduced LVEF (systolic HF). The exercise intervention was not blinded to either patient or site investigator; however, all CPX tests were reviewed by experts from a core laboratory who were blinded to treatment assignment. Despite extensive efforts, adherence to exercise training was less than optimal, with about 40% of patients at 3 months and about 35% of patients in year 2 training at or above the target number of minutes per week. The improvements in peak VO₂ with training were less than in many other studies and limit the ability to determine whether greater improvements in peak VO₂ would have resulted in larger differences in clinical outcomes. Although the change in peak VO₂ observed in the current study is less than the within-subject absolute variation of 1.3 ml/kg/min measured with repeated CPX testing in HF patients observed by Bensimhon et al (40), chance variation in peak VO₂ would not likely be associated with differences in clinical outcomes as observed here. Due to the fact that about 30% of the randomized subjects were excluded from our analyses because they either did not have a baseline or 3-month CPX test or experienced a primary outcome prior to the 3-month test, these findings should be interpreted with caution since they involve a more select group of patients than the overall HF-ACTION trial cohort. Unmeasured confounding variables may account for some or all of these findings. Imputation of missing data, particularly variables with >15% missing may have added some bias to the selection of the predictor variables for the final model.

Notwithstanding these limitations, several strengths of the current analysis are noteworthy. The large sample, which is more than twice that of prior similar studies (7–11) combined, the inclusion of large numbers of women and minorities, and the high use of evidence-based background therapy make our findings relevant to clinical practice. The protocol-based CPX testing three months after baseline assessment, with analysis by blinded core laboratory personnel provided further quality assurance.

In summary, the results of this secondary analysis of the largest randomized trial of exercise training in HF patients indicate that a modest increase in peak VO₂ over three months is associated with a lower composite rate of all-cause mortality and all-cause hospitalization, even in the setting of optimal medical and device therapy for chronic systolic HF. Since exercise training and other interventions can be effective therapy in improving aerobic exercise tolerance, a key limiting factor in patients with HF, these findings suggests that monitoring the change in peak VO₂ over time may be useful in this setting.

Clinical Impact Commentary for the Clinician.

In this sub-study of the landmark clinical trial entitled Heart Failure and a Controlled Trial to Investigate Outcomes of Exercise Training (HF-ACTION), Swank and colleagues report that small increments in peak VO₂ are associated with clinical outcome benefits in patients with heart failure. Specifically, every 6% increase in peak VO₂, adjusted for other significant predictors, was associated with a 5% lower risk of the primary composite endpoint of all-cause mortality and morbidity; a 4% lower risk of the secondary endpoint of time to cardiovascular mortality or cardiovascular hospitalization; an 8% lower risk of cardiovascular mortality or HF hospitalization and a 7% lower all-cause mortality. While it was difficult to attribute the results to exercise training per se, the fact that a relatively small increase in fitness improved clinical outcomes supports the concept that improving VO₂ may be an important target of therapy, however it is achieved. The study also has implications for the often-asked question not only in rehabilitation studies but also for drug and device interventions: "what constitutes a significant change in peak VO₂". The information provided from this study that an incremental change in peak VO₂ as low as 6% has morbidity and mortality benefit provides a potential benchmark for clinicians and trialists.