Peak oxygen uptake in relation to total heart volume discriminates heart failure patients from healthy volunteers and athletes

This study follows the Declaration of Helsinki and was approved by the regional ethics committee. All study subjects gave their informed consent to participate in the study.

Thirty-one patients (5 females) who had been clinically diagnosed with HF with varying etiology were retrospectively enrolled and compared with 60 healthy volunteers (20 females) from the local community and 71 elite athletes (30 soccer players [12 females], 23 European handball players [12 females] and 18 triathletes [6 females]) prospectively enrolled in the study. All underwent both CMR and maximal incremental exercise testing on ergometer cycle, with measurement of respiratory gas exchange, within a month for the patients and within a week for the healthy volunteers and athletes. Hemoglobin level was assessed within 2 weeks of the exercise test in patients and on the day of the exercise test in healthy volunteers and athletes who gave their consent for blood sampling.

CMR was performed on a 1.5 T system (Philips Intera CV; Philips, Best, The Netherlands) with a cardiac synergy coil. All subjects were placed in supine position. Cine short-axis gradient-recalled echo images covering the entire heart from base of the atria to the ventricular apex were acquired using a steady state free precession sequence - slice thickness = 8 mm, field-of-view = 340 mm, repetition time = 3.14 ms, echo time = 1.58 ms.

For viability assessment in the patients, late gadolinium enhancement (LGE) images were acquired 20--30 minutes after intravenous injection of an extracellular gadolinium-based contrast agent (gadoteric acid, Gd-DOTA, 0.2 mmol/kg, Guerbet, Gothia Medical AB, Billdal, Sweden). An inversion-recovery (IR) sequence was applied in the corresponding short-axis planes as for the cine images. Typical IR sequence parameters were: slice thickness = 8 mm, repetition time = 3.9 ms, echo time = 1.2 ms, in-plane resolution = 1.5 × 1.5 mm and flip angle = 15° with acquisition every heartbeat. The inversion time, typically 200--300 ms, was manually adjusted to null the signal from remote myocardium [14].

A maximal incremental exercise test was performed using an electronically braked cycle ergometer (Siemens Ergomed 940, Upplands Väsby, Sweden) with respiratory gas analysis equipment (Oxycon Champion, Jaeger, Hochberg, Germany). The test protocols were chosen to yield exercise duration of approximately 8-12 minutes, yielding starting stress between 10-150 W with 10-30 W increment per minute. Peak oxygen uptake was defined as the highest value reached at the end of exercise and expressed both as ml min^-1 (VO₂peak) and ml min^-1 kg^-1 (VO₂peak_kg).

A 12-lead ECG was acquired throughout the test. Blood pressure was measured by sphygmanometer in the supine position before exercise, every two minutes during exercise on the ergometer cycle and in the supine position at rest after exercise.

A ratio between VO₂peak and THV was derived. Furthermore, the ratio was also normalized to hemoglobin levels to partly compensate for differences in oxygen delivery capacity in peripheral tissue in the presence of anemia. The VO₂peak/THV ratio was also compared to levels of plasma amino terminal pro-brain natriuretic peptide (pro-BNP) for the patients who had a pro-BNP taken within 4 months of the exercise test.

Values are shown as mean ± standard deviation. Univariate linear regression analysis was used to assess the correlation between VO₂peak and THV as well as body weight. Unpaired T-test was used to assess differences between variables where normal distribution was assumed. The Mann-Whitney test was used to assess differences where normal distribution could not be assumed. Forward stepwise multivariate regression analysis was performed to assess the independent predictive values of THV, LVEF, LVEDV, VO₂peak, VO₂peak_kg and the VO₂peak/THV ratio for presence of HF. Receiver operating characteristic (ROC) analysis was performed to test the ability of the VO₂peak/THV ratio to distinguish between HF patients and controls. All statistical analysis was performed using SPSS 16.0 (Chicago, IL, USA). A p-value of <0.05 was considered statistically significant.

There was a strong correlation (r² = 0.74, p < 0.001) between VO₂peak and THV for the control subjects (healthy volunteers and athletes) but no correlation (r² = 0.0002, p = 0.95) for HF patients (Figure 2A). Figure 2B shows only a moderate correlation (r² = 0.37, p < 0.001) between VO₂peak and body weight in the control subjects with a considerable variation in VO₂peak for a given body weight.

Even though a significant difference in LVEF was found between the control subjects and the HF patients (p < 0.001; Figure 3A), there was a significant overlap with 29% (9/31) of the HF patient examinations showing LVEF > 50%. Figure 3B shows a comparison between control subjects and HF patients with LVEF > 50%, where the HF patients had a slightly higher LVEF (p = 0.04). Despite the higher LVEF in these patients, the VO₂peak/THV ratio was significantly lower compared to the control subjects (p < 0.001; Figure 3C).

Figure 4A shows the VO₂peak/THV ratio for each subject group. The HF patients showed a significantly lower VO₂peak/THV ratio (p < 0.001) than the other subject groups, even after normalizing for hemoglobin levels (p < 0.001, Figure 4B).

When dividing the different subgroups with regard to VO₂peak alone, there was still a significant difference between healthy volunteers and HF patients (3102 ± 696 vs 1309 ± 430 ml/min; p < 0.001). However, there was a considerable overlap between the two groups (Figure 5A). Ten control subjects and 10 HF patients had overlapping VO₂peak (Figure 5B). The VO₂peak/THV ratio, however, differed significantly between these two groups (p < 0.001) (Figure 5C). The LVEF for the overlapping healthy volunteers did not differ significantly from the overlapping HF patients (42 ± 25% vs 61 ± 5%; p = 0.11), whereas THV was significantly larger for the latter (640 ± 66 vs 1092 ± 336 ml/min; p < 0.001).

In 68% (21/31) of the HF patients a pro-BNP was taken 11 ± 25 days from the exercise test. There was a significant correlation between decreasing VO₂peak/THV ratio and increasing pro-BNP (r² = 0.42, p = 0.002).

Furthermore, the VO₂peak/THV ratio was the only independent predictor of HF (p < 0.001; Table 2). The area under the curve (AUC) for the VO₂peak/THV ratio to diagnose HF was 1.0. With a cut-off value of 3.0, the VO₂peak/THV ratio had a sensitivity of 100% and a specificity of 97%. Even though there was a slight decrease in VO₂peak/THV ratio with increasing age, there was a significant difference between patients and healthy volunteers for all age groups (Figure 6).

It is expected that the VO₂peak/THV ratio will drop in the presence of HF, since cardiac output, and consequently VO₂peak, is disproportional to the metabolic requirements of the peripheral tissue in the situation of a failing heart. However, when and in which stage of HF development the VO₂peak/THV ratio decreases is not known. A decreased VO₂peak/THV ratio could potentially be seen in patients with no symptoms at rest and in the absence of significant structural cardiac abnormalities.

Currently, maximal exercise test with analysis of respiratory gases and analysis of VO₂peak is predominantly used in patients with terminal HF under consideration for heart transplantation [7]. The present study suggests that assessment of VO₂peak in relation to THV for determination of a VO₂peak/THV ratio can be of potential use even in non-terminal and early stage HF, maybe even before the patient exhibit symptoms, such as in stage A HF [4]. Another possible application of a VO₂peak/THV ratio is in management of patients with a history of endurance training and a sudden change in exercise capacity. These patients might perform within normal limits during a maximal exercise test with analysis of respiratory gases when compared to age and gender matched controls. However, in relation to THV the working capacity might be decreased which would be detected by a low VO₂peak/THV ratio as an indicator of early stage HF. This is supported by the findings that the healthy volunteers and the HF patients with overlapping VO₂peak showed a significant difference in VO₂peak/THV ratio with almost no overlap, indicating that the ratio has additional diagnostic value. Furthermore, LVEF was not statistically significant different between the two overlapping groups. Thus, VO₂peak/THV provides diagnostic information beyond LVEF.

Even though there was a significant difference between control subjects and HF patients with regard to LVEF in the present study, almost 30% of the patients had normal LVEF. However, these patients had significantly lower VO₂peak/THV ratio than the control subjects with almost no overlap. Thus, VO₂peak/THV ratio may also be used as a marker of HF in patients with preserved systolic LV function.

Heart failure is often described based on clinical symptoms as being mild, moderate, or severe. Mild HF is used for patients with no significant physical limitations due to dyspnoea or fatigue and severe HF is used for patients who are markedly symptomatic with a need for frequent medical attention. Moderate HF is used for the remaining patients. For clinical management of HF patients and for randomized clinical trials, the New York Heart Association (NYHA) functional classification [22, 23] is often employed. However, the accuracy of diagnosing HF by clinical means alone is often inadequate [2, 3]. Hence, there is a need for objective measures of cardiac performance in order to determine the efficiency of different therapeutic strategies in randomized clinical trials.

The clinical use a VO₂peak/THV ratio could potentially complement commonly used cardiac diagnostic biomarkers such as ECG and biochemical markers. A commonly used biochemical marker for severity of HF is pro-BNP, which has been shown to correlate with the severity of HF [24, 25]. In the present study, the VO₂peak/THV ratio was shown to correlate with pro-BNP, indicating that the ratio can potentially serve as a measure for severity of HF.

Lower limits for VO₂peak are usually determined in relation to body weight and used to determine if a patient is suitable for heart transplantation or not [26, 27]. However, the present study show a considerable variability in VO₂peak for a given body weight (Figure 2B). Total heart volume, however, was shown to be closely correlated to VO₂peak in healthy volunteers and athletes but not in HF patients. Thus, for the non-failing heart, the VO₂peak/THV ratio is preserved in the presence of a physiologically enlarged heart, whereas for the failing heart the VO₂peak/THV ratio is not preserved as THV increases due to pathological enlargement of the heart. This suggests that VO₂peak should be normalized to THV and not to body weight. This is in line with a recent paper by Saltin et al [28] and experimental animal studies [9, 10] showing that VO₂peak is determined primarily by cardiac dimension and not peripheral factors such as the efficiency of oxygen uptake in peripheral tissue [29]. This is also supported by the present study showing that a VO₂peak/THV ratio is significantly lower in HF compared to healthy volunteers and athletes even after normalizing for hemoglobin levels.

In the present study THV was determined by CMR, which is considered the gold standard for assessing cardiac dimensions and function [13]. In situations when CMR is not available or is contraindicated, i.e. in patients with biventricular pacing, THV could potentially also be determined by low dose-cardiac CT, preferably gated. However, non-gated examinations would also be possible if CT thorax is performed for other reasons, since THV varies little during the cardiac cycle [15, 17, 18]. Three dimensional echocardiography might potentially also be used to determine THV, although this technique may suffer from incomplete coverage of the tissue within the pericardium. By using the THV, enlargement of all cardiac chambers including the atria is considered. Enlargement of the left atrium has been shown to provide prognostic value in HF patients with normal LVEF [30].

The results of the present study should be interpreted in the light of some limitations. The number of patients in the present study is limited and the patients were retrospectively included with clinically diagnosed HF and clinically determined etiology. Thus, no specific HF inclusion criteria based on biochemical markers, ECG or echocardiographic variables were applied. Furthermore, the patients had different etiologies and different stages of HF. However, this pilot study can be considered a proof of the concept study for determining a VO₂peak/THV ratio as an objective measure of cardiac capacity, independent of HF etiology.

Due to symptoms during the exercise test, HF patients might not reach VO₂peak representative of the individual's actual maximal oxygen uptake capacity. For this reason VO₂peak is referred to as peak and not maximal oxygen uptake in the present study.

There was a significant difference in mean age between the patient population and the control groups. However, the slight decrease in VO₂peak/THV ratio found with increased age in the healthy volunteers (21-65 years) is negligible compared to the decrease in VO₂peak/THV ratio found in the HF patients.

Since VO₂peak can not be obtained in HF patients with acute decompensation, the VO₂peak/THV ratio can not be used in these patients. The VO₂peak/THV ratio might, however, be of significant benefit when the HF diagnosis is uncertain, such as in the situation of stage A HF or if the progression of the disease is less clear.

Furthermore, the prognostic value of the VO₂peak/THV ratio also remains to be determined in prospective follow-up studies of different patient groups with established HF or at risk for developing HF.