Determinants of maximal oxygen uptake in patients with heart failure.

AIMS: Maximum oxygen uptake (VO2max ) is an essential parameter to assess functional capacity of patients with heart failure (HF). We aimed to identify clinical factors that determine its value, as they have not been well characterized yet.
METHODS: We conducted a retrospective, observational, single-centre study of 362 consecutive patients with HF who underwent cardiopulmonary exercise testing (CPET) as part of standard clinical assessment since 2009-2019. CPET was performed on treadmill, according to Bruce's protocol (n = 360) or Naughton's protocol (n = 2). We performed multivariable linear regression analyses in order to identify independent clinical predictors associated with peak VO2max .
RESULTS: Mean age of study patients was 57.3 ± 10.9 years, mean left ventricular ejection fraction was 32.8 ± 14.2%, and mean VO2max was 19.8 ± 5.2 mL/kg/min. Eighty-nine (24.6%) patients were women, and 114 (31.5%) had ischaemic heart disease. Multivariable linear regression analysis identified six independent clinical predictors of VO2max , including NYHA class (B coefficient = -2.585; P < 0.001), age (B coefficient per 1 year = -0.104; P < 0.001), tricuspid annulus plane systolic excursion (B coefficient per 1 mm = +0.209; P < 0.001), body mass index (B coefficient per 1 kg/m2  = -0.172; P = 0.002), haemoglobin (B coefficient per 1 g/dL = +0.418; P = 0.007) and NT-proBNP (B coefficient per 1000 pg/mL = -0.142; P = 0.019).
CONCLUSIONS: The severity of HF (NYHA class, NT-proBNP) as well as age, body composition and haemoglobin levels influence significantly exercise capacity. In patients with HF, the right ventricular systolic function is of greater importance for the physical capacity than the left ventricular systolic function.

Introduction

Exercise intolerance is one of the main symptoms of heart failure (HF) and a major factor conditioning a patient's quality of life. Moreover, the aetiopathogenesis of exercise intolerance is multifactorial, because it involves central factors related to cardiopulmonary system disorder and also peripheral factors related to both the vascular and musculoskeletal systems. ,

Maximum oxygen uptake (VO2max) or aerobic power is the best single parameter to measure a person's capacity to perform aerobic exercise. This parameter is defined as the maximum amount of oxygen that the body can absorb, transport and consume per unit of time.

In clinical practice, VO2max is determined in a protocolized way using the cardiopulmonary exercise testing (CPET) on a cycle ergometer or on a treadmill. This parameter represents an objective estimation of the functional capacity in patients with cardiovascular disease, and, in addition, it allows prognostic stratification in patients with HF, having been included in some of the most used mortality risk prediction scores such as HFSS or UCLA four‐variable risk score. In this sense, one of the main uses of VO2max is to serve as a guide for the selection of candidates for advanced therapies, including heart transplant. CPET allows the additional determination of other parameters that complement the diagnostic and prognostic value of VO2max such as VO2peak, VO2max/percentage of predicted VO2 ratio (%pp VO2), VO2max/body mass index (BMI) ratio, pulmonary ventilator efficiency slope ratio (VE) /maximum carbon dioxide uptake (VCO2) and oxygen uptake efficiency slope (OUES).

To date, there is little information on the clinical factors that determine the exercise capacity of patients with HF ; the role of some co‐morbidities such as obesity, anaemia or chronic kidney disease (CKD) in this regard has not been completely clarified.

The main objective of this study was to identify clinical factors associated with VO2max in a cohort of consecutive patients with HF who underwent CPET at a Spanish university hospital. As a secondary objective, we set out to evaluate the impact of the main co‐morbidities associated with HF on VO2max.

Methods

Study design and population

We carried out a retrospective observational study that included patients with HF seen in the specialized cardiology unit of the Complexo Hospitalario Universitario de A Coruña (CHUAC) who underwent a CPET between 2009 and 2019 as part of the standard clinical protocol of diagnostic and prognostic evaluation. For the purposes of this study, the follow‐up of the vital status of the patients was extended until 13 March 2020.

The study protocol was approved by A Coruña and Ferrol Ethics Committee in Clinical Research.

Cardiopulmonary exercise testing

All CPET tests of the patients included in the study were carried out in the laboratory of the Cardiac Imaging and Functional Unit of the Cardiology Service of CHUAC using a treadmill and the Bruce or Naughton protocols, depending on the specific characteristics of the patient. , In cases where the patient had more than one CPET during the study period, the first one performed was considered.

Sources of data

The data presented in this manuscript were obtained from the clinical data system of the Heart Failure Unit of the CHUAC (Intelligent Heart Failure Monitoring System, SiMON‐IC), complemented with the individual review of medical records, both in electronic and in paper formats.

Variable definition

For this study, the demographic, clinical, analytical and functional variables related to common functional studies were collected to describe a population of patients with HF, as well as specific variables in relation to CPET. The main variable of the study was VO2max determined by CPET.

Regarding co‐morbidities, anaemia was defined as any haemoglobin value < 13 g/dL in men and <12 g/dL in women, CKD as a glomerular filtration rate (GFR) < 60 mL/min/m2 and obesity as a BMI ≥ 30 kg/m2.

Statistical analysis

In this study, quantitative variables are described by mean ± standard deviation, whereas variable categories are expressed as absolute and relative frequency. The chi‐square test and Student's t‐test were used to compare groups, depending on the variable being analysed.

A multivariable linear regression model through backward steps was used to identify baseline clinical variables that are independently associated with VO2max by selecting a P out value < 0.05. Candidate variables for entering the first step of the backward stepwise analysis were those who showed a statistically significant or near‐significant univariate association with VO2max (univariate P value < 0.15). Variables with >20% missing values were not considered for these analyses.

A comparison of the VO2max value and VO2max/%pp VO2 ratio was proposed in seven subgroups selected according to the presence of clinical characteristics of special interest, such as age (≥60 years vs. <60 years), sex, left ventricular ejection fraction (LVEF) (<40% vs. ≥40%), GFR (<60 mL/min vs. ≥60 mL/min), BMI (<30 vs. ≥30), anaemia and tricuspid annular plane systolic excursion (TAPSE) (<15 mm vs. ≥15 mm).

A significance level P < 0.05 was assumed for all the contrasts proposed. The analysis was performed with SPSS 20.

Results

Study population

Three hundred sixty‐two patients were studied, of which 89 (24.6%) were female. The mean age of the population was 57.3 +/− 10.9 years. Most (n = 203, 56.1%) of the patients were in New York Heart Association (NYHA) functional class II. Anaemia was observed in 117 (32.3%), arterial hypertension in 161 (44.5%), ischaemic heart disease in 114 (31.5%), CKD in 111 (30.7%), and obesity in 118 (32.6%). Table provides baseline clinical characteristics, co‐morbidity and laboratory values in the study patients.

Table 1

Baseline clinical characteristics of 362 patients with heart failure who underwent cardiopulmonary exercise test

Demographics and clinical status a
Age (years), mean ± SD	57.3 ± 10.9
Male sex, n (%)	273 (75.4%)
Body mass index (kg/m2), mean ± SD	28.1 ± 4.8
Systolic blood pressure (mm Hg), mean ± SD	112.1 ± 17.3
Resting heart rate (beats per min), mean ± SD	67.3 ± 11.5
NYHA class, n (%)
I	25 (6.9%)
II	203 (56.1%)
III	125 (34.5%)
IV	9 (2.5%)

ALT, alanine aminotransferase; COPD, chronic obstructive pulmonary disease; GGT, gamma‐glutamyl transferase; LDH, lactate dehydrogenase; NT‐proBNP, N‐terminal pro‐brain natriuretic peptide; NYHA, New York Heart Association; SD, standard Deviation.

Missing values: body mass index (n = 3), resting heart rate (n = 1), systolic blood pressure (n = 1).

Missing values: chronic kidney disease (n = 3), obesity (n = 5).

Missing values: creatinine (n = 1), glucose (n = 2), urea (n = 13), sodium (n = 2), GGT (n = 46), ALT (n = 35), bilirubin (n = 69), uric acid (n = 132), LDH (n = 233), albumin (n = 247), NT‐proBNP (n = 48), ferritin (n = 339).

Table shows the data referring to complementary studies and pharmacological treatment. The majority (n = 240, 66.2%) was in sinus rhythm and almost a quarter in atrial fibrillation (n = 88, 24.3%). A total of 261 (72.1%) patients presented HF with reduced LVEF (<40%), 58 (16%) patients presented HF with mid‐range LVEF (40%–49%), and 43 (11.9%) patients presented HF with preserved LVEF (≥50%).

Table 2

Baseline electrocardiography, echocardiography, implantable electrical devices and treatments (n = 362)

ECG cardiac rhythm at rest
Sinus rhythm, n (%)	241 (66.6%)
Atrial fibrillation, n (%)	88 (24.3%)
Atrial flutter, n (%)	4 (1.1%)
Paced rhythm, n (%)	29 (8%)

ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; CRT‐D, cardiac resynchronization therapy with defibrillator; CRT‐P, cardiac resynchronization therapy with pacemaker; ECG, electrocardiogram; ICD, implantable cardioverter‐defibrillator; LVEDD, left ventricular end‐diastolic diameter; LVEF, left ventricular ejection fraction; LVESD, left ventricular end‐systolic diameter; MRA, mineralocorticoid receptor antagonist; TAPSE, tricuspid annular plane systolic excursion.

Missing values: LVEDD (n = 28), LVESD (n = 56), TAPSE (n = 74).

Beta‐blockers were prescribed to 330 (91.2%) patients, angiotensin‐converting enzyme (ACE) inhibitors or angiotensin II receptor blockers (ARB) to 248 (68.5%), sacubitril‐valsartan to 30 (8.3%), and mineralocorticoid receptor antagonists to 265 (73.2%). Overall, 181 (50%) patients carried a defibrillator, 49 (13.5%) patients carried a cardiac resynchronization therapy‐defibrillator device, and 5 (1.4%) patients carried a cardiac resynchronization therapy‐pacemaker device.

All CPET studies were performed according to Bruce's protocol, except two, which were performed according to Naughton's one. The CPET parameters are shown in Table . The mean VO2max achieved was 19.8 mL/kg/min (±5.2 SD), with a mean VO2max/%pp VO2 ratio of 75.9% (±21, 0 SD) and a mean duration of exercise of 8.4 (±3.9 SD) min. VO2max was <14 mL/kg/min in 45 (12.4%) patients.

Table 3

Cardiopulmonary exercise testing parameters in 362 patients with heart failure

CPET parameters
VO2 peak (mL/kg/min), mean ± SD	19.8 ± 5.2
%pp VO2, mean ± SD	75.9 ± 21.0
VE/VCO2, mean ± SD	37.8 ± 8.1
% maximum HR, mean ± SD	77.7 ± 16.3
RER, mean ± SD	1.09 ± 0.1
Exercise time (min), mean ± SD	8.4 ± 3.9
METs, mean ± SD	5.7 ± 2.2

% maximum HR, percentage of maximum predicted heart rate; % pp VO2, percentage of predicted peak oxygen uptake; CPET, cardiopulmonary exercise testing; METs, metabolic equivalents of task; RER, respiratory exchange ratio; VE/VCO2, ventilatory equivalent ratio for carbon dioxide; VO2peak, peak oxygen uptake.

Determining factors of VO2max

Table shows the results of the linear regression analyses used to identify the clinical factors that determine VO2max value in the study patients. Twelve variables that showed a univariate P value < 0.15 were selected for entering the first step of the backward stepwise process. Among them, six variables remained as independent predictors of VO2max in the final multivariable linear regression model.

Table 4

Clinical factors associated with peak VO2: multivariable linear regression analysis

	Univariate analysis a		Multivariate analysis b
	B coefficient	P	B coefficient	P
Age (years)	−0.259	<0.001	−0.104	<0.001
BMI (kg/m2)	−0.235	0.131	−0.172	0.002
Atrial fibrillation/flutter	−2.781	0.106	–	–
Hypertension	−3.162	0.001	–	–
Diabetes mellitus	−2.582	0.123	–	–
NYHA class	−2.184	0.059	−2.585	<0.001
Creatinine (mg/dL)	−1.228	0.135	–	–
Urea (mg/dL)	−0.045	0.049	–	–
Haemoglobin (g/L)	+0.642	0.137	+0.418	0.007
NT‐proBNP (1000 pg/mL)	−0.328	0.107	−0.142	0.019
TAPSE (mm)	+0.248	0.002	+0.209	<0.001
Ivabradine use	+7.933	0.001	–	–

BMI, body mass index; NT‐proBNP, N‐terminal pro‐brain natriuretic peptide; NYHA; New York Heart Association; TAPSE, tricuspid annular plane systolic excursion.

Variables presented are only those that showed a univariate association with VO2 with a P‐value < 0.15, as assessed by linear regression. These 10 variables entered the first step of multivariable backward stepwise linear regression analysis. The complete relationships of variables assessed, as well as their univariate regression coefficients, are presented as Supporting Information.

Variables presented are only those that were selected in the final step of multivariable backward stepwise linear regression analysis.

Older age (B coefficient for 1 year = −0.104; P < 0.001), worse NYHA functional class (B coefficient for one functional class −2.585; P < 0.001), higher plasma N‐terminal pro‐brain natriuretic peptide (NT‐proBNP) value (B coefficient per 1000 pg/mL = −0.142; P = 0.019) and a higher BMI (B coefficient −0.172; P = 0.002) were statistically significantly and independently associated with a lower VO2max.

On the other hand, a higher TAPSE (B coefficient per 1 mm = +0.209; P < 0.001) and a higher haemoglobin plasma value (B coefficient per 1 g/dL = +0.418; P = 0.007) were associated, also statistically significant and independently, with a higher VO2max value.

Univariate linear regression coefficients for all explored variables are presented in Table .

Subgroup analysis

Table shows the comparison of the VO2max values and the VO2max/%pp VO2 ratio in seven subgroups of patients defined based on clinical variables of interest.

Table 5

Comparison of peak VO2 and % predicted VO2 in selected clinical subgroups

Subgroups	VO2 peak (mL/kg/min) Mean ± SD	P	%pp VO2	P
Gender		0.027		<0.001
Female (n = 89)	18.1 ± 4.2		85.4 ± 18.9
Male (n = 273)	20.4 ± 5.3		72.8 ± 20.7
Anaemia		<0.001		0.215
Without anaemia (n = 51)	20.3 ± 5.4		76.5 ± 21.9
With anaemia (n = 311)	16.8 ± 3.9		72.7 ± 19.9
Left ventricular ejection fraction (%)		0.690		0.454
LVEF < 40 (n = 261)	20.1 ± 5.2		75.6 ± 21.6
LVEF ≥ 40 (n = 101)	19.6 ± 5.5		77.4 ± 19.6
Chronic kidney disease a		<0.001		0.006
GFR ≥ 60 mL/min/m2 (n = 246)	20.1 ± 5.4		77.7 ± 21.9
GFR < 60 mL/min/m2 (n = 111)	17.2 ± 4.2		71.5 ± 17.9
Age (years)		<0.001		0.087
<60 (n = 186)	21.6 ± 5.2		74.1 ± 20.1
≥60 (n = 176)	17.9 ± 4.7		77.9 ± 20.1
Obesity b		0.073		<0.001
BMI < 30 kg/m2 (n = 241)	20.5 ± 5.6		72.4 ± 18.9
BMI ≥ 30 kg/m2 (n = 118)	18.3 ± 5.4		82.8 ± 23.3
Right ventricular function c		0.001		<0.001
TAPSE < 15 mm (n = 100)	17.5 ± 4.2		69.1 ± 18.0
TAPSE ≥ 15 mm (n = 188)	19.9 ± 5.1		79.4 ± 19.9

%pp VO2, percentage of predicted peak VO2; BMI, body mass index; GFR, glomerular filtration rate; peak VO2, peak oxygen uptake; SD, standard deviation; TAPSE, tricuspid annular plane systolic excursion.

Missing data in 3 (0.8%) patients.

Missing data in 5 (1.4%) patients.

Missing data in 74 (20.4%) patients.

Male patients presented in absolute terms a higher VO2max than female patients (20.4 ± 5.3 mL/kg/min vs. 18.1 ± 4.2 mL/kg/min; P = 0.027); however, the VO2max/%pp VO2 ratio was significantly higher in women (72.8% ± 18.9 vs. 85.4% ± 18.9; P < 0.001).

VO2max in patients with anaemia (16.8 ± 3.9 mL/kg/min) was significantly lower than that of the rest of the population (20.3 ± 5.4 mL/kg/min) (P < 0.001); however, VO2max/%pp VO2 ratio showed no significant differences between these two groups.

VO2max value was higher in patients under 60 (21.6 ± 5.2 mL/kg/min) compared with those over 60 (17.9 ± 4.7 mL/kg/min; P < 0.001). However, the difference between both groups in relation to the VO2max/%pp VO2 ratio did not reach statistical significance (P = 0.087).

Patients without CKD showed a superior response to exercise than those with CKD, both in VO2max values (20.1 ± 5.4 mL/kg/min vs. 17.2 ± 4.2 mL/kg/min; P < 0.001) as in the VO2max/%pp VO2 ratio (77.7 ± 21.9 vs. 71.5 ± 17.9; P = 0.001).

Patients with a TAPSE ≥ 15 presented a higher VO2max value (19.9 ± 5.1 vs. 17.5 ± 4.2; P < 0.001) and a higher VO2max/%pp VO2 ratio (79.4 ± 18.0 vs. 69.1 ± 18.0; P < 0.001) than patients with TAPSE < 15.

We did not observe statistically significant differences in VO2max based on the presence or absence of obesity or the presence or absence of an LVEF > 40%.

Discussion

VO2max is an essential parameter for the functional and prognostic assessment of patients with HF that depends on central and peripheral factors. In our research, in which we have studied a cohort of patients with HF followed in a specialized cardiology unit, we have identified six independent clinical predictors of VO2max, including age, NYHA functional class and the presence of obesity, haemoglobin and NT‐proBNP plasma levels and TAPSE.

The relationship between ageing and a progressive reduction in exercise capacity is well known. As in our study, the literature has referenced that younger patients present higher absolute VO2max values in CPET. This trend towards a decrease in the theoretical VO2 as age increases may result in a less prognostic value of VO2max in patients with HF and older age. Therefore, the use of VO2max/%pp VO2 ratio is recommended in this subgroup with the intention of, for example, selecting candidates for advanced therapies.

The association observed between a higher NYHA functional class and a lower VO2max value in patients with HF is evident. The NYHA functional class, although conditioned by a subjective evaluation, is a good initial approach to the global symptom burden of patients with HF and has shown a significant prognostic value in these individuals. The findings of lower VO2max values in patients with higher plasma levels of NT‐proBNP are consistent with the known diagnostic and prognostic value of this biomarker as an indicator of the degree of congestion and severity of HF.

An interesting finding of our study is the independent association observed between an indicator of right ventricular function, TAPSE and VO2max. This relationship is more interesting, if possible, taking into account that in our series we have not been able to evidence an impact of LVEF on this parameter. These results seem to suggest that in patients with HF, the preservation of the contractile function of the right ventricle would be a more determining factor for keeping the ability to exercise than the systolic function of the left ventricle itself. The prognostic importance of the parameters that assess right ventricular function in patients with HF, even in those with preserved LVEF, has been previously demonstrated. It should be noted, however, that a study with a slightly larger sample size than ours, Nadruz et al., did observe a significant impact of LVEF on VO2max, so patients with LVEF > 40% had higher values than patients with LVEF < 40%.

Obesity and anaemia are two frequent co‐morbidities in patients with HF that significantly affect their exercise capacity. Other authors have previously described an association between a higher BMI or a lower haemoglobin plasma value and a lower VO2max value in CPET. Patients with anaemia experience a decline in mitochondrial oxidative function, reduced adenosine triphosphate production and a decline of the capacity of muscular contraction that affects notably its exercise tolerance.

In our study, the univariate analysis showed that patients with CKD had significantly lower VO2max and VO2max/%pp VO2 values than patients with preservation of renal function. CKD is a known factor of poor prognosis and poor functional capacity in patients with HF, which can affect exercise capacity through inflammatory mechanisms, loss of muscle mass and increased oxidative stress. , However, CKD was not founded as an independent predictor of VO2max in the multivariable analysis, a result that we attribute, as the most likely explanation, to its tendency to coexist in the same patient with other independent risk markers such as age and advanced NYHA functional class, anaemia or elevated NT‐proBNP, as well as the limited sample size of our study.

The present study has several limitations. Firstly, it is an observational and retrospective study, so it is exposed to the typical biases of this type of research. Secondly, CPETs were performed within the framework of routine clinical care, and not according to a predefined research protocol. Finally, it should be noted that the sample size of the study, even being relevant, may have been insufficient to detect as statistically significant other risk associations that do may be clinically relevant

Conclusions

The severity HF (NYHA class, NT‐proBNP) and age, body composition and haemoglobin levels influence significantly exercise capacity. In patients with HF, the right ventricular systolic function is of greater importance for the physical capacity than the left ventricular systolic function.

Our work highlights the importance of associated clinical conditions as determinants of exercise tolerance in patients with HF and reinforces the importance of appropriate handling of co‐morbidities to improve functional capacity in these individuals.

Conflict of interest

MGCL reports to have received research support from Novartis, Vifor Pharma and FEDER Funds and personal fees (travel grants, lecture fees and/or advisory boards) from Novartis, Abbott Vascular, Astellas, MSD and Servier. None of them related to this work. The rest of the authors declare no conflict of interest.

Funding

This work was supported by FEDER Funds, Instituto de Salud Carlos III, Madrid.

Table S1. Supporting information.

Click here for additional data file.