The Value of Cardiopulmonary Exercise Testing in Determining Severity in Patients with both Systolic Heart Failure and COPD

Our aim was to identify optimal cardiopulmonary exercise testing (CPET) threshold values that distinguish disease severity progression in patients with co-existing systolic heart failure (HF) and chronic obstructive pulmonary disease (COPD), and to evaluate the impact of the cut-off determined on the prognosis of hospitalizations. We evaluated 40 patients (30 men and 10 woman) with HF and COPD through pulmonary function testing, doppler echocardiography and maximal incremental CPET on a cycle ergometer. Several significant CPET threshold values were identified in detecting a forced expiratory volume in 1 second (FEV1) < 1.6 L: 1) oxygen uptake efficiency slope (OUES) < 1.3; and 2) circulatory power (CP) < 2383 mmHg.mlO2.kg−1. CPET significant threshold values in identifying a left ventricular ejection fraction (LVEF) < 39% were: 1) OUES: < 1.3; 2) CP < 2116 mmHg.mlO2.kg−1.min−1 and minute ventilation/carbon dioxide production (V̇E/V̇CO2) slope>38. The 15 (38%) patients hospitalized during follow-up (8 ± 2 months). In the hospitalizations analysis, LVEF < 39% and FEV1 < 1.6, OUES < 1.3, CP < 2116 mmHg.mlO2.kg−1.min−1 and V̇E/V̇CO2 > 38 were a strong risk predictor for hospitalization (P ≤ 0.050). The CPET response effectively identified worsening disease severity in patients with a HF-COPD phenotype. LVEF, FEV1, CP, OUES, and the V̇E/V̇CO2 slope may be particularly useful in the clinical assessment and strong risk predictor for hospitalization.

(2013) showed that the peak of V̇O 2 < 10 mL/kg/min and V̇E/V̇CO 2 slope ≥45 were independent predictors of long-term mortality in HF and COPD 14,15 . O 2 pulse, circulatory and ventilatory power (CP and VP, respectively) are related to predictors of mortality in all chronic heart failure patients 16-18 . In addition, the studies that performed CPET focused on the variation between different populations presented as main objective the comparison with tests [10][11][12]19,20 . Thus, there is no information about the impact of HF-COPD on CPET variables, since these patients have marked ventilatory and cardiac limitations and, from a clinical point of view, established cut-off values for these patients based on the disease severity. It would be of utmost importance to determine during the performance of CPET, the clinical condition, diagnosis and best prescription of cut-off training for these patients.
Our aim was to identify optimal CPET threshold values that distinguish disease severity progression in patients with co-existing systolic HF and COPD. Our secondary aim was to evaluate the impact of the cut-off determined on the prognosis of hospitalizations of these patients. We hypothesize that CPET variables are important prognostic determinants for cardiorespiratory worsening in patients with FEV 1 < 1.6 and LVEF < 39%.

Methods
Study design. This cross-sectional study was designed following the recommendations of the STROBE statement. All patients were recruited from the Cardiology and Pulmonology outpatient clinic of São Carlos. The study followed the Declaration of Helsinki and was approved by the local ethics committees (Federal University of Sao Carlos) (protocol number: 91088318.7.1001.5504). All volunteers signed a written informed consent statement prior to participation.
Subjects. 51 patients with clinical diagnosis of COPD by pulmonary function test [FEV 1 /forced vital capacity (FVC) ratio of 0.7; FEV 1 60% of predicted] without previous COPD exacerbation (3 months before the study) and clinical diagnosis of HF from a cardiologist with ejection fraction (<50% by echocardiogram), without cognitive impairment or comprehension deficiencies, older than 50 years of age and with HF class I, II or III according to the New York Heart Association Functional Classification (NYHA) were included in the study 16 .
Some HF-COPD patients were excluded from this study, as follows: patients with musculoskeletal disorders or neurological conditions affecting the locomotor system in a way that precluded them from protocol participation, patients recently hospitalized with clinical diagnoses of lung cancer, heavy alcohol drinkers, any patient with observed complex cardiac arrhythmias or electrocardiogram alterations, and patients with uncontrolled metabolic diseases, such as diabetes mellitus.
According to GOLD (2016) 1 the mean FEV 1 value of COPD patients is 1.7 L. Thus, we emphasize that our results are based on the average of our study, which is similar to the recommended by GOLD for FEV 1 1.6 ± 0.1 L and LVEF < 39%.
protocol. All patients underwent an echocardiogram administered by a cardiologist, a pulmonary function exam performed by a pulmonologist, and a clinical assessment. Every patient completed the comprehensive evaluation process in two days: (1) clinical evaluation by a physician and a physical therapist, followed by lung function test and Doppler echocardiography; (2) CPET; (3) follow-up.
Measurements. Doppler echocardiography. Initially for the clinical and diagnostic stratification, the HF-COPD patients were submitted to a 2D-echocardiogram using an iE33 system (Philips, Andover, MA, USA) with a 2-5 MHz matrix transducer and tissue Doppler imaging software. Quantification of the cardiac chambers was performed according to the American Society of Echocardiography. In our study we only included patients with reduced LVEF (<50%). Patients who presented preserved LVEF (>50%) or diastolic HF were excluded from study 16 . Cardiopulmonary exercise testing. In accordance with the American College of Cardiology and American Heart Association Guidelines 15 , a physician and physical therapist supervised the CPETs and the subjects were asked to maintain routine medications on the day of testing. The test was performed on an electronically braked cycle ergometer (Corival Recumbent, Medical Graphics Corporation, St. Paul, Mo, USA) and respiratory gas analysis was measured breath-by-breath with Oxycon Mobile (Mijnhardt/Jäger, Würzburg, German).
The protocol consisted of the following: (I) 5-minute rest period while sitting on the cycle ergometer; (II) 1-minute exercise at free-wheel and 60 rotations per minute (rpm); (III) incremental phase with an increase of 5-10 W/min (ramp protocol); (IV) 1-minute of active recovery at free-wheel; and (V) 5-minute passive recovery resting in sitting position. A twelve-lead Electrocardiogram (ECG) was continuously monitored throughout the test (WinCardio, Micromed, Brasilia, Brazil). The test was finished when subjects were pedaling at their maximum possible effort level (physical exhaustion) and reported at least 2 of the following criteria: (I) age predicted maximal HR (220 j [age]); (II) general/leg fatigue or dyspnea; (III) angina or electrocardiographic evidence of ischemia or malignance arrhythmia (ventricular tachyarrhythmia, ventricular fibrillation, bigeminism); or (IV) inability to maintain a pedaling rate of 60 rpm for 30 seconds.
Follow-up. Patients were included and accompanied during 12 months by telephone calls to their home or family physician. The hospitalizations were determined when the patient was hospitalized for more than one day, our patients had the following hospitalizations: COPD exacerbations (type II and III) (n = 6), HF decompensation (n = 5) and myocardial revascularization (n = 4). COPD exacerbations are classified as: type II -increased medication and medical intervention; and type III -worsening of the clinical condition requiring hospitalization 1 .

Statistical analysis.
The Shapiro-Wilk test was used to verify the data distribution. Descriptive data was shown as a mean, standard deviation and frequency. The parametric Student's t-test was used for normally distributed data. Pearson correlation analysis were performed to investigate the relationship between variables. All tests were made in Statistical Package for the Social Sciences (SPSS) and values were accepted as P ≤ 0.05. ROC curve. First, receiver operating characteristic (ROC) curve analyses selected the optimal threshold values to differentiate the severity of COPD considering FEV 1 (L) and the severity of HF considering LVEF (%). Cut-off points discriminated the precision of CPET variables: VP, CP, O 2 pulse, OUES, V̇E/V̇CO 2 slope and V̇O 2 peak in determining points of predictive cut-off in HF-COPD. The confidence interval (95% CI) was used to determine the ability of the clinical variables, with the lower limit being greater than 0.50. Subsequently, the cut-off points of the variables that obtained significant areas under the ROC curve were identified, with the respective values of sensitivity and specificity.
Kaplan-meier. We examined all hospitalization that occurred during the 12-months follow-up. Hospitalization curves were analyzed according to the Kaplan-Meier method to explore the impact of FEV 1 < 1.6 and LVEF < 39%, CP < 2338, VE/VCO 2 >38 and OUES>1.3. Differences between curves were evaluated using the log-rank test.

Results
General characteristics. We initially included 51 HF-COPD patients, however due to our exclusion criteria, 11 patients did not participate in the study protocol: 1 was excluded for brain cancer, 7 did not agree to participate in the study, 1 due to aortic aneurysm, and 2 due to encephalic vascular accident. Therefore the protocol was carried out with 40 HF-COPD patients. Table 1 presents the clinical, echocardiogram and spirometry's characteristics in HF-COPD patients, 40 adult men and 10 woman with beta blocker and beta-agonists use (Table 1).
During follow-up (10 ± 2 months), 15 patients were hospitalized (7 ± 2 months) and 25 were not hospitalized (10 ± 1 months). Table 2  We found significant correlations between: OUES vs LVEF (p = 0.018; r = 0.371) and FEV 1 vs V̇O 2 /WR (p < 0.001; r = −0.735), suggesting that the worse the airway obstruction and LVEF, the worse the behavior of these variables in CPET, thus compromising the performance of these HF-COPD patients.  www.nature.com/scientificreports www.nature.com/scientificreports/ Cut-off points for FEV 1 < 1.6 and LVEF < 39%. From this point of the study, our results will be divided by the mean found in our patients FEV 1 < 1.6 L [n = 21 (52%)] and LVEF < 39% [19 (n = 48%)], our aim in performing this stratification is to demonstrate the impact of diseases on CPET variables and may assist in clinical practice.
The cut-off points, the areas under the ROC curve and 95% CI, as well as the sensitivity and specificity of the clinical variables are shown in Table 3. According to the prediction, the best models of sensitivity and specificity are: OUES < 1.3 and CP < 2383 mmHg.mlO 2 .min −1 identified as cut-off points for HF-COPD patients with FEV 1 < 1.6 L.
The cut-off points, the areas under the ROC curve and 95% CI, as well as the sensitivity and specificity of the clinical variables are shown in Table 4. According to the prediction, the best models of sensitivity and specificity are: OUES: < 1.3, CP < 2116 mmHg.mlO 2 .kg −1 .min −1 and V̇E/V̇CO 2 slope > 38 identified as cut-off points for HF-COPD patients with LVEF < 39% (Fig. 1).

Hospitalization analysis.
In the Kaplan-Meier analysis, when we associated patients with LVEF < 39% and FEV 1 < 1.6 [n = 14] with a higher probability of hospitalizations over a 12-month period. The hospitalization curve differed significantly in the log-rank test (p = 0.018). In the analysis for the CPET variables, we found that only V̇E/V̇CO 2 > 38, OUES < 1.3 and CP < 2383 mmHg.mlO 2 .min −1 were prognostic indicators for hospitalization (P ≤ 0.05) in COPD-HF patients (Fig. 2).    Table 4. Cut-off values, sensitivity and specificity of hemodynamic response in CPET in HF-COPD with LVEF < 39%. HF: heart failure; COPD: chronic obstructive pulmonary disease; V E/V CO 2 slope: Linear relation between minute ventilation and carbon dioxide production; CP: Circulatory power and OUES: Oxygen uptake efficiency slope.
We found that airway obstruction and LVEF directly influence aerobic function and OUES. Our study corroborates with Hansen et al. 24 , who found a lower V̇O 2 /WR in cardiorespiratory disease and suggesting that cardiorespiratory limitation may decrease the load ratio, even with optimal exercise duration.
The main result of our study was to find predictive CPET values for worsening of cardiorespiratory capacity based on airway obstruction in patients with COPD and LVEF in patients with HF and prognostic indicators for hospitalization. We can highlight that through these cutoff points it will be possible to determine diagnosis, training prescriptions, as well as the general state of these patients. De Miguel et al. 25 highlighted that HF-COPD patients have a mixed pulmonary dysfunction. COPD is characterized by airway obstruction and emphysema. In HF, the enlargement of the heart, venous congestion, and interstitial fibrosis compresses the lungs, leading to a restrictive pulmonary disorder 26 . The coexistence of diseases may lead to further increases in disability and mortality, perhaps even impairing the CPET results since variables of the test are poorer in these patients [17][18][19][20][21][22][23][24][25][26][27] .
We emphasize the importance in determining other prognostic indices of cardiorespiratory worsening of CPET in these patients. Another cut-off values in our study were CP < 2383 mmHg.mlO 2 .min −1 and CP < 2116 mmHg.mlO 2 .kg −1 .min −1 for FEV 1 < 1.6 and LVEF < 39% respectively, a new cardiac index recently studied, and an original result for this population. A previous study demonstrated that CP is a surrogate index of cardiac power, shown to have a better prognostic value than peak V̇O 2 , V̇O 2 /HR and peak SBP in HF 20,22,27 . Physiologically, CP represents the volume of O 2 added to the mixed venous blood by the lungs and transferred to systemic arterial circulation, against a pressure gradient produced by the heart 22 . CP is related to the central and peripheral components of the cardiac work 16,22 . We emphasize that CP < 2383 mmHg.mlO 2 .min −1 was a predictor of hospitalization in these patients.
In COPD, an emphysema burden has been associated with increased V̇E/V̇CO 2 , as a consequence of increased ventilatory drive and greater neuromechanical dissociation 28,29 . In HF, disease progression is associated with higher V̇E/V̇CO 2 because of an increased ventilatory drive leading to hypocapnia in highly variable combinations 30,31 . Aposto et al. 20 , found that increased V̇E of the linear V̇E/V̇CO 2 relationship during ramp-incremental exercise should raise the suspicion of coexistent COPD in patients with HF. The new result of our study is that the V̇E/V̇CO 2 slope > 38 for patients with LVEF < 39% that was identified as cut-off points and predictor of hospitalization for HF-COPD patients.
Lin et al. 32 found that HF patients with OUES < 1.3 had a higher risk of cardiac events. OUES is considered a method of evaluating cardiopulmonary endurance and it is easily determined when using the breath-by-breath respiratory analysis method 32 . Compared with the values already described in the literature, our result of OUES < 1.3 for patients with FEV 1 < 1.6 and LVEF < 39%, emphasizes that our study is the first to find a cut-off value for a HF-COPD population.
Thus, our results highlight the importance of identifying predictive values of disease severity-based exercise testing (COPD and HF), as well as the prognosis value of cardiopulmonary exercise testing to help healthcare professionals to identify potential patients for hospitalization. It is possible with cut-off values to obtain a response about the patient's general health, as we know that hospitalizations limit or even preclude rehabilitation and may aggravate the exercise intolerance of this population.

Strength and Limitation of the Study
The main limitation in our study is the small number of COPD-HF coexistence, however, despite the high prevalence of COPD and HF, such coexistence has been poorly studied, especially considering the impact of disease severity on exercise capacity. In this sense, future studies should be conducted with a larger sample to confirm our findings 33 .
The strengths of our study include the novelty of finding cut-offs for new rates (CP, OUES and V̇E/V̇CO 2 slope) and predictors of hospitalization for HF-COPD patients with FEV 1 < 1.6 L and LVEF 39%. Thus, this is the first study to investigate these indices in HF-COPD patients based on the severity of the diseases. We emphasize as a limitation the sample size, however it is known the difficulty in screening COPD-HF coexistence and that so far no follow-up studies and prognosis of hospitalizations in this population have been conducted, thus highlighting the importance of this work.

conclusion
In conclusion, heart failure with reduced left ventricular ejection fraction as well as the severity of COPD impacted negatively selected physiological responses obtained by CPET. CP, OUES and V̇E/V̇CO 2 slope provided to be useful in practice to clinical interpretation of cardiopulmonary responses during exercise and are a strong risk predictor for hospitalization in HF-COPD coexistence patients.