To date, the role of CPET (cardiopulmonary exercise testing) for risk stratification in elderly patients with HF (heart failure) with depressed or preserved ventricular function has not been evaluated. In the present study, we analysed whether CPET is useful in predicting outcome in this population. A total of 220 NYHA (New York Heart Association) class I–III patients with HF ≥70 years of age [median age, 75 years; 23% had NYHA class III; and 59% had preserved ventricular systolic function (left ventricular ejection fraction ≥40%)] performed maximal CPET (peak expiratory exchange ratio >1.00). Median peak oxygen uptake was 11.9 ml·kg−1 of body weight·min−1, median V̇E/V̇CO2 slope (slope of the minute ventilation/carbon dioxide production ratio) was 33.2 and 45% had an EVR (enhanced ventilatory response) to exercise (V̇E/V̇CO2 slope ≥34). During 19 months of follow-up, 94 patients (43%) met the combined end point of death and hospital admission for worsening HF, arrhythmias or acute coronary syndromes. By Cox multivariable analysis, a creatinine clearance of <50 ml/min {HR (hazard ratio), 1.657 [95% CI (confidence interval), 1.055–2.602]} and EVR [HR, 1.965 (95% CI, 1.195–3.231)] were the best predictors of outcome, while ventricular function had no influence on prognosis. In conclusion, in elderly patients with HF, a steeper V̇E/V̇CO2 slope provides additional information for risk stratification across the spectrum of ventricular function and identifies a high-risk population, commonly not considered in exercise testing guidelines.

INTRODUCTION

HF (heart failure) is a common chronic disease in the elderly. Although relief of symptoms and acceptable quality of life are the main goals of treatment in frail patients with multiple disabling co-morbidities, ‘robust’ elderly subjects, with no relevant strength or mobility problems, may have an outlook similar to younger patients, and should be carefully assessed for up-to-date treatment.

When assessing prognosis in elderly patients with HF, certain predictors are particularly relevant in view of their prevalence in older age. Renal dysfunction, common in the elderly and frequent in HF, increases mortality across the whole spectrum of ventricular impairment [1,2]. In the elderly, PLVEF {preserved LVEF [LV (left ventricular) ejection fraction}-HF is a common and costly clinical entity, with mortality and 1-year re-admission rates [3,4] similar to SHF (HF with depressed systolic function). Increasing evidence links diastolic dysfunction to exercise intolerance [5].

Among laboratory investigations, CPET (cardiopulmonary exercise testing) with gas exchange measurement, when technically feasible, is an accurate tool in routine clinical evaluation and risk stratification of patients with HF [6,7]. Reduced V̇O2 peak [peak V̇O2 (oxygen consumption)] and ventilatory abnormalities, such as an EVR (enhanced ventilatory response) to exercise provide an objective and non-invasive evaluation of functional capacity and have been shown to predict prognosis in a large subset of patients with SHF and PLVEF-HF [811]. Our group has demonstrated previously the feasibility and safety of CPET in elderly patients with HF [12].

BNP (brain natriuretic peptide), a composite marker of systolic and diastolic function, and neurohormonal activation, is a prominent prognostic parameter, even in elderly patients [13,14]. Previous clinical reports have focused on the prognostic usefulness of BNP and V̇O2 peak in patients with HF [15,16]; however, CPET-derived parameters of ventilatory efficiency may be superior to BNP in predicting the prognosis of patients with SHF [17].

The aim of our present study was to address the additive prognostic role of ventilatory parameters to clinical, neurohormonal and echocardiographic variables in stable elderly outpatients with mild-to-moderate SHF and PLVEF-HF on optimized medical treatment, followed-up at a dedicated HF clinic.

MATERIALS AND METHODS

Study population

Between October 2001 and January 2006, 331 consecutive outpatients aged ≥70 years with a diagnosis of HF of at least 3 months were screened for CPET evaluation, and 95 were excluded for the reasons detailed in Figure 1. Patients were in NYHA (New York Heart Association) classes I–III on stable drug doses without worsening symptoms or in need of intravenous inotropic support during the last 4 weeks. Among the 236 consenting subjects who were able to perform a CPET test, 220 with a V̇O2 peak<25 mg·kg−1 of body weight·min−1 and a RER (respiratory exchange ratio) >1.00 at CPET, indicating the achievement of anaerobic exercise conditions, were enrolled; 16 subjects with an RER ≤1.00 were excluded.

Flow chart of the present study detailing the exclusion criteria

Figure 1
Flow chart of the present study detailing the exclusion criteria

CPX, CPET.

Figure 1
Flow chart of the present study detailing the exclusion criteria

CPX, CPET.

The study protocol was approved by the Institutional Ethics Committee and complies with the ethical rules for human experimentation that are stated in the Declaration of Helsinki. Written informed consent was obtained from each patient.

Baseline laboratory assessment was performed according to standard laboratory methods. Creatinine clearance, as an estimate of GFR (glomerular filtration rate), was determined using the Cockcroft–Gault equation [17a].

Patients were followed up in the outpatient HF clinic of our hospital. Outcome data of patients who did not attend their scheduled appointments (<10% of those enrolled) were obtained by telephone interview of the patient or his/her family; none of these patients had been admitted to hospital. The primary combined study end point was defined as the occurrence of death and hospital admission for worsening HF, arrhythmias or acute coronary syndromes.

Exercise testing

A standard bicycle exercise ramp protocol was used, with increments of 10 W/min. Details of the test protocol have been reported previously [12,17]. Before each test, oxygen and carbon dioxide analysers and a flow mass sensor were calibrated by use of available precision gas mixtures and a 3-litre syringe respectively. To stabilize gas measurements, patients were asked to remain still for at least 3 min before beginning an upright graded bicycle exercise testing, using a continuous protocol, which was continued until exhaustion.

A 12-lead ECG was recorded continuously (MAX-1; Marquette Electronics). BP (blood pressure) was recorded every minute by a cuff sphygmomanometer. Indications to stop the test were exhaustion, dyspnoea, angina, ST-segment depression ≥3 mm, high-degree AV (arteriovenous) block or ventricular tachycardia ≥five consecutive beats at a rate ≥120/min, new-onset atrial fibrillation or SBP (systolic BP) ≥260 mmHg, as well as a progressive decrease in BP.

V̇O2, V̇CO2 (carbon dioxide production) and V̇E (minute ventilation) were measured by breath-by-breath gas analysis with the use of a computerized metabolic cart (V max 29; Sensormedics).

V̇O2 peak was recorded as the mean value of V̇O2 during the last 30 s of the test and is expressed as ml·kg−1 of body weight·min−1. V̇O2 peak was measured and expressed as an absolute and percentage predicted value from the individual age-, gender- and body-weight-corrected mean normal values (% pV̇O2). The %V̇O2 peak was calculated for each patient using Wasserman's equation [17b].

V̇O2 AT (V̇O2 at the anaerobic threshold) was determined by using the V-slope method [18]. Other considered parameters were: (i) V̇E/V̇O2 ratio, (ii) V̇E/V̇CO2 ratio (V̇E, V̇CO2 and V̇CO2 in l/min); (iii) V̇E/V̇CO2 slope, calculated in every subject as the slope of the regression line relating to V̇E to V̇CO2 during exercise testing, with the exclusion of the final non-linear portion after the onset of acidotic drive to ventilation.

A V̇CO2/V̇O2 ratio >1.00 was selected for the definition of maximal exercise.

The V̇E/V̇CO2 slope was used as the main index of the ventilatory response to exercise. We defined EVR as a V̇E/V̇CO2 slope ≥34 [19]. The ratio of V̇E/V̇CO2 slope to V̇O2 peak, suggested previously as a better predictor of prognosis than either variable alone [20], was also analysed.

Doppler echocardiography

Doppler echocardiography was performed using a Phillips MS Sonos 5.500 D device equipped with a multiband 1.8–3.6 MHz probe. LVEF was measured by two expert operators blinded to the CPET results, according to the modified Simpson's biplane method [20a], in the apical four- and two-chamber views. Biplane volumes were calculated from area tracings using the disc summation method (modified Simpson's rule). PAP (pulmonary systolic pressure) was calculated by means of right ventricular/right atrial gradient in patients with detectable tricuspid regurgitation (peak ventricular/atrial gradient +8–10 mmHg). Transmitral flow velocity was determined using the pulsed-wave Doppler approach with a 2–4 mm sample volume placed at the mitral leaflet tips level in the apical four-chamber view. The transmitral pulsed-wave Doppler blood flow velocities were calculated off-line from five consecutive cardiac cycles to measure E (early diastolic peak velocity; in cm/s), A (late diastolic peak velocity; in cm/s), E/A ratio and E deceleration time (ms). A restrictive filling pattern was defined as the presence of an E/A ratio >2.0 cm/s and an E deceleration time <140 ms. In patients with atrial fibrillation, the restrictive filling pattern was defined by an E deceleration time <140 ms calculated off-line from ten consecutive cardiac cycles. PLVEF was defined as LVEF ≥40% [2123].

Measurement of BNP plasma levels

Before exercise testing, blood was collected by venipuncture from an antecubital vein after 10 min of rest in the sitting position. A point-of-care test (Triage BNP test; Biosite Diagnostics) was used, based on an immunofluorimetric assay for the quantitative determination of BNP in whole blood or plasma. Normality range values are 0–40 pg/ml. BNP concentrations were determined by nurses blinded to the CPET results.

Statistical analysis

Categorical variables are presented as numbers (percentage), and continuous variables as medians (interquartile range). Between-group differences were tested using a χ2 test, Student's t test or non-parametric testing according to normal or non-normal variable distribution. A P value <0.05 was considered statistically significant.

To determine the independent predictors of the primary combined end point of occurrence of death and hospital admission for major cardiovascular events, we entered variables significant by univariable analysis (P<0.004 by Bonferroni correction) into a multivariable Cox proportional hazards model in sequential diagnostic blocks of clinical and laboratory, echocardiographic and CPET information to mimic the sequence of information obtained in clinical practice. Relevant covariates were selected through a forward selection procedure. HRs (hazard ratios) are presented with 95% CIs (confidence intervals). Non-categorical variables were entered as continuous values, and HRs are expressed as risk/10-unit change. In the final model, clinically relevant cut-off values were also used. Differences between Kaplan–Meier event-free survival curves, according to the predefined cut-off values, were assessed using the log rank test.

All statistical tests were two-tailed. Results were analysed using SPSS 10 (Microsoft®).

RESULTS

Clinical features and laboratory findings of our population are described in Table 1. Median age was 75 (72, 79) years, LV function was mildly reduced in this series of patients, with PLVEF found in 131 patients (59%), and an EVR was present in 123 out of 220 patients (56%).

Table 1
Clinical characteristics and main parameters in the 220 study patients

*Values are medians (interquartile range). ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocking; CRT, cardiac resynchronization therapy; ICD, implantable cardioverter defibrillator; LVEDV, LV end-diastolic volume; LVESV, LV end-systolic volume. BPCO, broncho-pneumopathie chronic obstructive (chronic obstructive pulmonary disease).

Parameter Value 
Clinical characteristic (n 
 Female gender 70 (32%) 
 Ischaemic aetiology of HF 101 (46%) 
 NYHA functional class III 51 (23%) 
 Left bundle branch block 62 (28%) 
 Chronic atrial fibrillation 33 (15%) 
 Diabetes 68 (31%) 
 Anamnestic mild BPCO 51 (23%) 
 CRT 5 (2%) 
 ICD 5 (2%) 
Medication  
  ACE-inhibitors 196 (89%) 
  ARB 18 (8%) 
  β-Blockers 123 (56%) 
  Non-K+-sparing diuretics 194 (88%) 
  K+-sparing diuretics 53 (24%) 
  Digitalis 13 (6%) 
  Nitrates 123 (56%) 
  Oral anticoagulants 31 (14%) 
Laboratory finding*  
 Haemoglobin (g/dl) 12.7 (11.7, 13.7) 
 Serum sodium (mmol/l) 140 (138, 142) 
 Creatinine clearance (ml/min) 50 (38, 63) 
 BNP (pg/ml) 141 (56, 365) 
Echocardiographic variable*  
 LVEDV (ml) 129 (103, 160) 
 LVESV (ml) 70 (50, 105) 
 LVEF (%) 42 (34, 51) 
E deceleration time (ms) 200 (150, 225) 
 PAP (mmHg) 31 (25, 44) 
 Restrictive filling pattern 40 (22%) 
CPET parameter*  
V̇O2 peak (ml·kg−1 of body weight·min−111.9 (9.76, 14) 
V̇O2 peak (%) 63 (51, 74) 
V̇O2 AT (n=185) (ml·kg−1 of body weight·min−19.20 (7.8, 10.5) 
V̇E/V̇O2 ratio 42 (37, 50) 
V̇E/V̇CO2 ratio 37 (34, 43) 
V̇E/V̇CO2 slope 33.2 (29, 38.1) 
V̇CO2 slope/V̇O2 peak ratio 2.78 (2.24, 3.72) 
 Workload (W) 61 (48, 77) 
Parameter Value 
Clinical characteristic (n 
 Female gender 70 (32%) 
 Ischaemic aetiology of HF 101 (46%) 
 NYHA functional class III 51 (23%) 
 Left bundle branch block 62 (28%) 
 Chronic atrial fibrillation 33 (15%) 
 Diabetes 68 (31%) 
 Anamnestic mild BPCO 51 (23%) 
 CRT 5 (2%) 
 ICD 5 (2%) 
Medication  
  ACE-inhibitors 196 (89%) 
  ARB 18 (8%) 
  β-Blockers 123 (56%) 
  Non-K+-sparing diuretics 194 (88%) 
  K+-sparing diuretics 53 (24%) 
  Digitalis 13 (6%) 
  Nitrates 123 (56%) 
  Oral anticoagulants 31 (14%) 
Laboratory finding*  
 Haemoglobin (g/dl) 12.7 (11.7, 13.7) 
 Serum sodium (mmol/l) 140 (138, 142) 
 Creatinine clearance (ml/min) 50 (38, 63) 
 BNP (pg/ml) 141 (56, 365) 
Echocardiographic variable*  
 LVEDV (ml) 129 (103, 160) 
 LVESV (ml) 70 (50, 105) 
 LVEF (%) 42 (34, 51) 
E deceleration time (ms) 200 (150, 225) 
 PAP (mmHg) 31 (25, 44) 
 Restrictive filling pattern 40 (22%) 
CPET parameter*  
V̇O2 peak (ml·kg−1 of body weight·min−111.9 (9.76, 14) 
V̇O2 peak (%) 63 (51, 74) 
V̇O2 AT (n=185) (ml·kg−1 of body weight·min−19.20 (7.8, 10.5) 
V̇E/V̇O2 ratio 42 (37, 50) 
V̇E/V̇CO2 ratio 37 (34, 43) 
V̇E/V̇CO2 slope 33.2 (29, 38.1) 
V̇CO2 slope/V̇O2 peak ratio 2.78 (2.24, 3.72) 
 Workload (W) 61 (48, 77) 

During a median follow-up of 19 (10, 40) months, 68 patients were admitted at least once for worsening HF (n=45), arrhythmias (n=12) or acute coronary syndromes (n=11), and 32 died (17 from progressive HF, nine from sudden death, four from other cardiovascular causes and two from cancer). Among patients with left bundle branch, 12 (5.4%) underwent re-synchronization therapy. Using a time-to-first-event approach, overall 94 patients (43%) achieved the study end point of death or hospital admission. The significant predictors of major events by univariable analysis (P value <0.004 using Bonferroni correction; Table 2) were entered into multiple multivariable Cox proportional hazard models in a sequential approach. HRs and related 95% CIs per 10-unit increments and according to clinically relevant cut-off values are shown in Table 3.

Table 2
Univariable predictors of outcome

P<0.004 (Bonferroni correction) for inclusion in the multivariable model. See Table 1 for other abbreviations.

Predictor P value HR (95% CI) 
Age 0.353 1.020 (0.978–1.063) 
Female gender 0.718 0.923 (0.597–1.426) 
Ischaemic aetiology of HF 0.102 1.404 (0.935–2.109) 
NYHA III functional class 0.118 1.452 (0.909–2.318) 
Left bundle branch block 0.018 1.656 (1.092–2.510) 
Chronic atrial fibrillation 0.178 1.440 (0.847–2.446) 
β-Blockers 0.327 0.817 (0.545–1.224) 
Haemoglobin (g/dl) 0.061 0.873 (0.757–1.006) 
Serum sodium (mmol/l) 0.680 0.989 (0.939–1.042) 
Creatinine clearance (ml/min) 0.003 0.982 (0.969–0.995) 
BNP (pg/ml) 0.001 1.001 (1.000–1.002) 
LVEDV (ml) 0.220 1.002 (0.999–1.006) 
LVESV (ml) 0.094 1.004 (0.999–1.008) 
LVEF (%) 0.019 0.979 (0.962–0.997) 
PAP (mmHg) 0.001 1.020 (1.008–1.033) 
E deceleration time (ms) 0.003 0.995 (0.992–0.998) 
Restrictive filling pattern 0.001 2.079 (1.330–3.249) 
V̇O2 peak (ml·kg−1 of body weight·min−10.006 0.915 (0.859–0.974) 
%V̇O2 peak 0.099 0.991 (0.980–1.002) 
V̇O2 AT 0.010 0.867 (0.777–0.967) 
V̇E/V̇O2 ratio 0.000 1.039 (1.018–1.060) 
V̇E/V̇CO2 ratio 0.000 1.040 (1.019–1.061) 
V̇E/V̇CO2 slope 0.000 1.043 (1.020–1.067) 
V̇E/V̇CO2 slope/V̇O2 peak ratio 0.000 1.218 (1.096–1.354) 
Workload (Watts) 0.006 0.987 (0.978–0.966) 
Predictor P value HR (95% CI) 
Age 0.353 1.020 (0.978–1.063) 
Female gender 0.718 0.923 (0.597–1.426) 
Ischaemic aetiology of HF 0.102 1.404 (0.935–2.109) 
NYHA III functional class 0.118 1.452 (0.909–2.318) 
Left bundle branch block 0.018 1.656 (1.092–2.510) 
Chronic atrial fibrillation 0.178 1.440 (0.847–2.446) 
β-Blockers 0.327 0.817 (0.545–1.224) 
Haemoglobin (g/dl) 0.061 0.873 (0.757–1.006) 
Serum sodium (mmol/l) 0.680 0.989 (0.939–1.042) 
Creatinine clearance (ml/min) 0.003 0.982 (0.969–0.995) 
BNP (pg/ml) 0.001 1.001 (1.000–1.002) 
LVEDV (ml) 0.220 1.002 (0.999–1.006) 
LVESV (ml) 0.094 1.004 (0.999–1.008) 
LVEF (%) 0.019 0.979 (0.962–0.997) 
PAP (mmHg) 0.001 1.020 (1.008–1.033) 
E deceleration time (ms) 0.003 0.995 (0.992–0.998) 
Restrictive filling pattern 0.001 2.079 (1.330–3.249) 
V̇O2 peak (ml·kg−1 of body weight·min−10.006 0.915 (0.859–0.974) 
%V̇O2 peak 0.099 0.991 (0.980–1.002) 
V̇O2 AT 0.010 0.867 (0.777–0.967) 
V̇E/V̇O2 ratio 0.000 1.039 (1.018–1.060) 
V̇E/V̇CO2 ratio 0.000 1.040 (1.019–1.061) 
V̇E/V̇CO2 slope 0.000 1.043 (1.020–1.067) 
V̇E/V̇CO2 slope/V̇O2 peak ratio 0.000 1.218 (1.096–1.354) 
Workload (Watts) 0.006 0.987 (0.978–0.966) 
Table 3
Multivariable predictors of death or hospitalization
Predictor P value HR (95% CI) 
Clinical, BNP and echocardiographic variables   
 Per 10-unit increment   
  Creatinine clearance (ml/min) 0.020 0.852 (0.795–0.914) 
  BNP (pg/ml) 0.000 1.010 (1.009–1.012) 
 Cut-off value   
  Creatinine clearance <50  compared with ≥50 ml/min 0.014 1.748 (1.122–2.724) 
  BNP >140 compared with  ≤140 pg/ml 0.001 2.171 (1.388–3.396) 
Clinical, BNP and echocardiographic and CPET variables   
 Per 10-unit increment   
  Creatinine clearance (ml/min) 0.024 0.852 (0.795–0.914) 
  V̇E/V̇CO2 slope 0.001 1.492 (1.323–1.682) 
 Cut-off value   
  Creatinine clearance <50 compared with ≥50 ml/min 0.028 1.657 (1.055–2.602) 
  V̇E/V̇CO2 slope ≥34 (EVR) 0.008 1.965 (1.195–3.231) 
Predictor P value HR (95% CI) 
Clinical, BNP and echocardiographic variables   
 Per 10-unit increment   
  Creatinine clearance (ml/min) 0.020 0.852 (0.795–0.914) 
  BNP (pg/ml) 0.000 1.010 (1.009–1.012) 
 Cut-off value   
  Creatinine clearance <50  compared with ≥50 ml/min 0.014 1.748 (1.122–2.724) 
  BNP >140 compared with  ≤140 pg/ml 0.001 2.171 (1.388–3.396) 
Clinical, BNP and echocardiographic and CPET variables   
 Per 10-unit increment   
  Creatinine clearance (ml/min) 0.024 0.852 (0.795–0.914) 
  V̇E/V̇CO2 slope 0.001 1.492 (1.323–1.682) 
 Cut-off value   
  Creatinine clearance <50 compared with ≥50 ml/min 0.028 1.657 (1.055–2.602) 
  V̇E/V̇CO2 slope ≥34 (EVR) 0.008 1.965 (1.195–3.231) 

In the first model, among five clinical, laboratory and echocardiographic variables, only renal dysfunction and higher BNP levels were predictive of the end point. In the final model, including four CPET parameters, only renal dysfunction and a steeper V̇E/V̇CO2 slope were significantly associated with outcome.

Survival rates free from HF admission by the Kaplan–Meier method according to the V̇E/V̇CO2 slope, BNP levels and renal dysfunction are shown in Figures 2–4 respectively.

Event-free survival rates in patients without or with an EVR to exercise

Figure 2
Event-free survival rates in patients without or with an EVR to exercise

Log rank P value <0.001.

Figure 2
Event-free survival rates in patients without or with an EVR to exercise

Log rank P value <0.001.

Event-free survival rates in patients with BNP <140 or >140 pg/ml

Figure 3
Event-free survival rates in patients with BNP <140 or >140 pg/ml

Log rank P value <0.0001.

Figure 3
Event-free survival rates in patients with BNP <140 or >140 pg/ml

Log rank P value <0.0001.

Event-free survival rates in patients with creatine clearance≥50 or <50 ml/min

Figure 4
Event-free survival rates in patients with creatine clearance≥50 or <50 ml/min

Creatinine clearance (eGFR) was calculated using the Cockcroft–Gault equation. Log rank P value <0.005.

Figure 4
Event-free survival rates in patients with creatine clearance≥50 or <50 ml/min

Creatinine clearance (eGFR) was calculated using the Cockcroft–Gault equation. Log rank P value <0.005.

DISCUSSION

The present prospective study is among the largest carried out to define, in a multimodal assessment, the additive prognostic role of CPET in an elderly population with mild-to-moderate HF and depressed or preserved LV function. Our results demonstrate that, in ‘robust’ elderly patients with HF able to perform maximal exercise, CPET is a crucial step in prognostic assessment and ventilatory parameters, which, together with evidence of renal dysfunction, has a key role in risk stratification over and beyond the level of LV function.

In advanced age, morbidity, with its attendant impact on quality of life, is a relevant target for HF management beyond extension of life, so we chose to assess outcome in terms of hospital admissions for major cardiac events besides mortality: 43% of the patients met, over a median period of 19 months, the combined end point.

Renal dysfunction, an established risk factor for adverse outcome in HF [1,2], was a main independent prognostic predictor across the spectrum of ventricular impairment in our elderly population. This finding has important clinical implications: pharmacological management, focused on kidney protection to prevent or slow progression of renal impairment, may also prove valuable in improving the outcome in elderly patients with HF.

Our present results underline further the relevance of PLVEF-HF in the elderly: although LV function predicted outcome by univariate analysis, both renal dysfunction and BNP, a composite marker of systolic and diastolic function and neurohormonal activation [13], had a stronger independent association with prognosis. Actually, the role of LVEF is far less important when morbidity, with its attending health care costs and detrimental impact on quality of life, is considered: in overt HF, event-free survival does not differ among patients with depressed or preserved LV function [3,2123]. Moreover, a recent clinical trial, which enrolled 556 patients with HF with both preserved (LFEF >50%) or reduced (LVEF <50%) systolic function [24], demonstrated that PLVEF-HF had a higher 6-months mortality rate than patients with HF with depressed LVEF. These findings are also supported by another recent retrospective study [25] that showed a similar post-discharge mortality and hospitalization rate in patients with preserved (LVEF >40%) and reduced (LVEF <40%) LVEF. A heterogeneous aetiology, the high prevalence in the older population and the underuse of optimal pharmacological treatment are probably the main reasons accounting for the high morbidity and mortality rate in patients with PLVEF-HF and in the reduced prognostic significance of LVEF in this group of patients.

Our main aim in the present study was to assess the additive prognostic role of CPET, an established test for risk stratification in other clinical HF settings, in elderly patients who so far have been substantially neglected in the literature. Controversy still exists on the best ventilatory prognostic predictor in general [11]. Peak exercise V̇O2 has been found to be superior in large series [8] and recent studies [26], and the anaerobic threshold in other studies of younger patients with HF [27]. Other studies have stressed the optimal value of ventilatory efficiency as a marker of adverse outcome [1012,28]. Old age itself might well determine the differences in the definition of the most useful CPET parameters for outcome assessment. Peak exercise V̇O2 is known to be influenced, besides by age and gender, by co-morbidities and noncardiac factors, such as muscle deconditioning, motivation and anaemia, all highly prevalent in the elderly, and might, therefore, be less useful as a prognostic marker in this population, as highlighted by our present findings. V̇O2 AT, an objective parameter of cardiopulmonary exercise capacity that can be derived from submaximal exercise testing found previously to be useful in predicting 6-month mortality [27], could not be determined in approx. 15% of our present study patients and was not independently associated with prognosis in this group.

To date, only three studies have analysed the prognostic value of CPET in patients of an age similar to ours [2931]. Both Davies et al. [29] and Mejhert et al. [30] demonstrated in small series the prominent predictive role of the V̇E/V̇CO2 slope, but their analyses were restricted to patients with SHF. In a larger study by Cicoira et al. [31], among 102 patients with SHF or PLVEF-HF with a maximal CPET (RER >1), both V̇O2 peak and the V̇E/V̇CO2 slope were predictors of mortality. The inclusion of subjects with hypertrophic or restrictive cardiomyopathy and dynamic flow obstruction, the different end point (death compared with mortality and morbidity) and the underuse of β-blockers (15 compared with 56%) may explain the different prognostic weight of LVEF in the experiences obtained by Cicoira et al. [31] and ourselves in the present study. In none of the above studies was CPET compared with BNP as a risk stratification tool.

In a group of patients with SHF and a median age of 71 years, we have demonstrated previously the prognostic superiority of ventilatory and diastolic function parameters in comparison with BNP [17]. In the present study, although independently predictive in the clinical model as anticipated by its significant relationship with ventilatory parameters, BNP lost its power when ventilatory parameters, namely the V̇E/V̇CO2 slope, were included in the multivariable model. On the basis of these findings, BNP may be used as a reliable surrogate predictor of prognosis when CPET cannot be adequately performed in the elderly.

The prognosis of elderly patients with HF is poor, and NYHA classification, a simple and widely used approach to the assessment of their functional status, appears unreliable for risk stratification.

CPET, an inexpensive and reliable tool for testing cardiopulmonary function, can be safely performed in the vast majority of older patients, adding important information on disease severity and, thus, also providing a guide for tailored HF treatment in this population normally neglected by exercise testing guidelines.

Study limitations

Our present study population was relatively small and included only clinically stable ‘robust’ outpatients with mild-to-moderate HF who did not have relevant co-morbidities and were, furthermore, able to perform CPET. Therefore our conclusions cannot be extended to all elderly patients with HF.

β-Blockers reduced mortality and hospital admissions in landmark HF trials, but did not predict outcome in the present study. One possible explanation is the relatively low proportion of sudden deaths with respect to worsening HF in our patients. Furthermore, stratification of ventilatory parameters by β-blocker treatment did not show any significant differences.

We did not measure plasma NT-pro-BNP (N-terminal pro-BNP), reported recently [32,33] as an accurate prognostic predictor in HF, as point-of-care testing was not available for NT-pro-BNP at the time the study was performed. Patients with advanced renal dysfunction were excluded from the present study because of the non-specific elevation of plasma BNP in this clinical condition; however, even in this relatively healthy population, mild degrees of renal dysfunction were one of the strongest predictors of outcome, independent of BNP levels.

We did not assess the value of other emerging ventilatory parameters, such as the V̇O2 efficiency slope [34]. Indeed, others [28] have recently reasserted the superior prognostic characteristics of the V̇E/V̇CO2 slope over V̇O2 peak, both at maximal and submaximal testing, in younger patients with HF treated with β-blockers in a proportion similar to our present study.

Conclusions

In ‘robust’ elderly patients with mild-to-moderate HF and impaired function or PLVEF, reduced eGFR (estimated GFR) and a steeper V̇E/V̇CO2 slope provide additional independent prognostic information across the spectrum of ventricular function and identify a high-risk population commonly not considered in exercise testing guidelines. These findings should extend some of the clinical applications of CPET validated in middle-aged patients to older populations with HF not eligible for heart transplantation, suggesting a wider use of this technique.

FUNDING

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Abbreviations

     
  • A

    late diastolic peak velocity

  •  
  • BNP

    brain natriuretic peptide

  •  
  • BP

    blood pressure

  •  
  • CI

    confidence interval

  •  
  • CPET

    cardiopulmonary exercise testing

  •  
  • E

    early diastolic peak velocity

  •  
  • eGFR

    estimated GFR

  •  
  • EVR

    enhanced ventilatory response

  •  
  • GFR

    glomerular filtration rate

  •  
  • HF

    heart failure

  •  
  • HR

    hazard ratio

  •  
  • LV

    left ventricular

  •  
  • LVEF

    LV ejection fraction

  •  
  • NT-pro-BNP

    N-terminal pro-BNP

  •  
  • NYHA

    New York Heart Association

  •  
  • PAP

    pulmonary systolic pressure

  •  
  • PLVEF

    preserved LVEF

  •  
  • PLVEF-HF

    HF with PLVEF

  •  
  • RER

    respiratory exchange ratio

  •  
  • SHF

    HF with depressed systolic function

  •  
  • V̇CO2

    carbon dioxide production

  •  
  • V̇E

    minute ventilation

  •  
  • V̇O2

    oxygen consumption

  •  
  • V̇O2 peak

    peak V̇O2

  •  
  • V̇O2AT

    V̇O2 at the anaerobic threshold

References

References
1
McAlister
F. A.
Ezekowitz
J.
Tonelli
M.
Armstrong
P. W.
Renal insufficiency and heart failure: prognostic and therapeutic implications from a prospective cohort study
Circulation
2004
, vol. 
109
 (pg. 
1004
-
1009
)
2
Hillege
H. L.
Nitsch
D.
Pfeffer
M. A.
, et al. 
Renal function as a predictor of outcome in a broad spectrum of patients with heart failure
Circulation
2006
, vol. 
113
 (pg. 
67
-
68
)
3
Hogg
K.
Swedberg
K.
McMurray
J.
Heart failure with preserved left ventricular systolic function; epidemiology, clinical characteristics and prognosis
J. Am. Coll. Cardiol.
2004
, vol. 
43
 (pg. 
317
-
327
)
4
O'Conner
C. M.
Gattis
W. A.
Shaw
L.
Cuffe
M. S.
Califf
R. M.
Clinical characteristics and longterm outcomes of patients with heart failure and preserved systolic function
Am. J. Cardiol.
2000
, vol. 
86
 (pg. 
863
-
867
)
5
Packer
M.
Abnormalities of diastolic function as a potential cause of exercise intolerance in chronic heart failure
Circulation
1990
, vol. 
81
 (pg. 
III78
-
III86
)
6
Working Group on Cardiac Rehabilitation Exercise Physiology, Working Group on Heart Failure of European Society of Cardiology
Recommendations for exercise testing in chronic heart failure patients
Eur. Heart J.
2001
, vol. 
22
 (pg. 
37
-
45
)
7
Lang
C.
Agostoni
P. G.
Mancini
D. M.
Prognostic significance and measurement of exercise-derived hemodynamic variables in patients with heart failure
J. Card. Failure
2007
, vol. 
13
 (pg. 
672
-
679
)
8
O'Neill
J. O.
Young
J. B.
Pothier
C. E.
Lauer
M. S.
Peak oxygen consumption as a predictor of death in patients with heart failure receiving β-blockers
Circulation
2005
, vol. 
111
 (pg. 
2313
-
2318
)
9
Corra
U.
Mezzani
A.
Bosimini
E.
Scapellato
F.
Imparato
A.
Giannuzzi
P.
Ventilatory response to exercise improves risk stratification in patients with chronic heart failure and intermediate functional capacity
Am. Heart J.
2002
, vol. 
143
 (pg. 
418
-
426
)
10
Arena
R.
Myers
J.
Aslam
S. S.
Varughese
E. B.
Peberdy
M. A.
Peak V̇O2 and V̇E/V̇CO2 slope in patients with heart failure: a prognostic comparison
Am. Heart J.
2004
, vol. 
147
 (pg. 
354
-
360
)
11
Guazzi
M.
Myers
J.
Arena
R.
Cardiopulmonary exercise testing in the clinical and prognostic assessment of diastolic heart failure
J. Am. Coll. Cardiol.
2005
, vol. 
46
 (pg. 
1883
-
1890
)
12
Scardovi
A. B.
Coletta
C.
De Maria
R.
, et al. 
Cardiopulmonary exercise test is safe and reliable in elderly patients with chronic heart failure
J. Cardiovasc. Med.
2007
, vol. 
8
 (pg. 
608
-
612
)
13
Silver
M. A.
Maisel
A.
Yancy
C. W.
, et al. 
BNP Consensus Panel 2004: a clinical approach for the diagnostic, prognostic, screening, treatment monitoring, and therapeutic roles of natriuretic peptides in cardiovascular diseases
Congest. Heart Failure
2004
, vol. 
10
 (pg. 
S1
-
S30
)
14
Feola
M.
Aspromonte
N.
Canali
C.
, et al. 
Prognostic value of plasma brain natriuretic peptide, urea nitrogen, and creatinine in outpatients >70 years of age with heart failure
Am. J. Cardiol.
2005
, vol. 
96
 (pg. 
705
-
709
)
15
Isnard
R.
Pousset
F.
Chafirovskaia
O.
, et al. 
Combination of B-type natriuretic peptide and peak oxygen consumption improves risk stratification in outpatients with chronic heart failure
Am. Heart J.
2003
, vol. 
146
 (pg. 
729
-
735
)
16
de Groote
P.
Dagorn
J. I.
Soudan
B.
Lamblin
N.
McFadden
E.
Bauters
C.
B-type natriuretic peptide and peak exercise oxygen consumption provide independent information for risk stratification in patients with stable congestive heart failure
J. Am. Coll. Cardiol.
2004
, vol. 
43
 (pg. 
1584
-
1589
)
17
Scardovi
A. B.
De Maria
R.
Coletta
C.
, et al. 
Multiparametric risk stratification in patients with mild to moderate chronic heart failure
J. Card. Failure
2007
, vol. 
13
 (pg. 
445
-
451
)
17a
Cockcroft
D. W.
Gault
M. H.
Prediction of creatinine clearance from serum creatinine
Nephron
1976
, vol. 
16
 (pg. 
31
-
41
)
17b
Stringer
W. W.
Hansen
J. E.
Wasserman
K.
Cardiac output estimated non-invasively from oxygen uptake during exercise
J. Appl. Physiol.
1997
, vol. 
82
 (pg. 
908
-
912
)
18
Cohen-Solal
A.
Zannad
F.
Kayanakis
J. G.
Gueret
P.
Aupetit
J. F.
Kolsky
H.
Multicentre study of the determination of peak oxygen uptake and ventilatory threshold during bicycle in chronic heart failure
Eur. Heart J.
1991
, vol. 
12
 (pg. 
1055
-
1063
)
19
Ponikowski
P.
Francis
D. P.
Piepoli
M. F.
, et al. 
Enhanced ventilatory response to exercise in patients with chronic heart failure and preserved exercise tolerance. Marker of abnormal cardiorespiratory reflex control and predictor of poor prognosis
Circulation
2001
, vol. 
103
 (pg. 
967
-
972
)
20
Guazzi
M.
De Vita
S.
Cardano
P.
Barlera
S.
Guazzi
M. D.
Normalization for peak oxygen uptake increases the prognostic power of the ventilatory response to exercise in patients with chronic heart failure
Am. Heart J.
2003
, vol. 
146
 (pg. 
542
-
548
)
20a
Otto
C. M.
Schwaegler
R. G.
Echocardiography Review Guide
2007
New York
Saunders
(pg. 
22
-
23
)
21
Permenkil
R.
Vinson
J. M.
Shah
A. S
Beckham
V.
Wittenberg
C.
Rich
M. W.
Course and prognosis in patients ≥70 years of age with congestive heart failure and normal versus abnormal left ventricular ejection fraction
Am. J. Cardiol.
1997
, vol. 
79
 (pg. 
216
-
219
)
22
Smith
G. L.
Masoudi
F. A.
Vaccarino Radford
M. J.
Krumholz
H. M.
Outcomes in heart failure patients with preserved ejection fraction. Mortality, readmission and functional decline
J. Am. Coll. Cardiol.
2003
, vol. 
41
 (pg. 
1510
-
1518
)
23
Ahmed
A.
Perry
G. J.
Fleg
J. L.
Love
T. E.
Goff
D. C.
Jr
Kitzman
D. W.
Outcomes in ambulatory chronic systolic and diastolic heart failure: a propensity score analysis
Am. Heart J.
2006
, vol. 
152
 (pg. 
956
-
966
)
24
Ezekowitz
J. A.
Lee
D. S.
Tu
J. V.
Newman
A. M.
McAlister
F. A.
Comparison of one-year outcome (death and rehospitalization) in hospitalized heart failure patients with left ventricular ejection fraction >50% versus those with ejection fraction <50%
Am. J. Cardiol.
2008
, vol. 
102
 (pg. 
79
-
83
)
25
Fonarow
G. C.
Stough
W. G.
Abraham
W. T.
, et al. 
OPTIMIZE-HF Investigators and Hospitals Characteristics, treatments, and outcomes of patients with preserved systolic function hospitalized for heart failure: a report from the OPTIMIZE-HF Registry
J. Am. Coll. Cardiol.
2007
, vol. 
50
 (pg. 
768
-
777
)
26
Ingle
L.
Witte
K. K.
Cleland
J. G. F.
Clark
A. L.
Combining the ventilatory response to exercise and peak oxygen consumption is no better than peak oxygen consumption alone in predicting mortality in chronic heart failure
Eur. J. Heart Failure
2008
, vol. 
10
 (pg. 
85
-
88
)
27
Gitt
A. L.
Wasserman
K.
Kilkowski
C.
, et al. 
Exercise anaerobic threshold and ventilatory efficiency identify heart failure patients for high risk of early death
Circulation
2002
, vol. 
106
 (pg. 
3079
-
3084
)
28
Arena
R.
Myers
J.
Hsu
L.
, et al. 
The minute ventilation/carbon dioxide production slope is prognostically superior to the oxygen uptake efficiency slope
J. Card. Failure
2007
, vol. 
13
 (pg. 
462
-
469
)
29
Davies
L. C.
Francis
D. P.
Piepoli
M.
Scott
A. C.
Ponikowski
P.
Coats
A. J.
Chronic heart failure in the elderly: value of cardiopulmonary exercise testing in risk stratification
Heart
2000
, vol. 
83
 (pg. 
147
-
151
)
30
Mejhert
M.
Linder-Klingsell
E.
Edner
M.
Kahan
T.
Persson
H.
Ventilatory variables are strong prognostic markers in elderly patients with heart failure
Heart
2002
, vol. 
88
 (pg. 
239
-
243
)
31
Cicoira
M.
Davos
C. H.
Florea
V.
, et al. 
Chronic heart failure in the very elderly: clinical status, survival, and prognostic factors in 188 patients more than 70 years old
Am. Heart J.
2001
, vol. 
142
 (pg. 
174
-
180
)
32
Grewal
J.
McKelvie
R. S.
Persson
H.
, et al. 
Usefulness of N-terminal pro-brain natriuretic peptide and brain natriuretic peptide to predict cardiovascular outcomes in patients with heart failure and preserved left ventricular ejection fraction
Am. J. Cardiol.
2008
, vol. 
102
 (pg. 
733
-
737
)
33
Dini
F. L.
Conti
U.
Fontanive
P.
Andreini
D.
Panicucci
E.
De Tommasi
S. M.
Prognostic value of N-terminal pro-type-B natriuretic peptide and Doppler left ventricular diastolic variables in patients with chronic systolic heart failure stabilized by therapy
Am. J. Cardiol.
2008
, vol. 
102
 (pg. 
463
-
468
)
34
Davies
L. C.
Wensel
R.
Georgiadou
P.
, et al. 
Enhanced prognostic value from cardiopulmonary exercise testing in chronic heart failure by non-linear analysis: oxygen uptake efficiency slope
Eur. Heart J.
2006
, vol. 
27
 (pg. 
684
-
690
)