ABC | Volume 114, Nº4, April 2020

Original Article Marchese et al. Inspiratory exercise in heart failure Arq Bras Cardiol. 2020; 114(4):656-663 would be observed. Therefore, this study aimed to assess CHR in a single session of inspiratory exercise with different loads (placebo, 30 and 60%) in HFREF. Methods Randomized, placebo-controlled clinical trial. The load was placed on the linear load resistor, in a way that participants could not see at which level the marker was positioned and were also not informed about the load used. Inclusion and exclusion criteria To meet the objective of this study, 29 patients with HFREF from the Heart Failure Clinic (CLIC) of Centro Universitário Serra dos Órgãos (UNIFESO) were selected. They all met the following inclusion criteria: clinical diagnosis of heart failure, age over 21 years, Doppler echocardiogram with left ventricular ejection fraction (LVEF) <45% (Simpson method), class II and III by the New York Heart Association (NYHA), stable disease for at least three months, never having undergone or not being treated with IMT. None of the following exclusion criteria were present: clinical (medical) diagnosis of chronic obstructive pulmonary disease, unstable angina, major cardiac arrhythmias, acute myocardial infarction in the last three months, inability to perform the IE session. And yet none of the exclusion criteria for cardiothoracic bioimpedance: massive pleural effusion, anasarca, moderate or severe aortic insufficiency, use of intra-aortic balloon, mean arterial pressure >130mmHg, height <1.20m or >2.30m, weight <30kg or >155kg, and use of pacemakers with sensors to adjust heart rate according to respiratory rate. Assessment methods Collection instruments used were: an analogue manovacuometer (Critical Med®, Brazil), a linear load resistor (PowerBreathe Light®, United States), and a cardiothoracic bioimpedance (CTB) device (Niccomo™ CardioScreen®, Germany). The inspiratory muscle exercise (IME) sessions were performed according to the randomization made by the Randomizer website, using the linear load resistor for 15 minutes with the following loads: 0 (placebo), 30% and 60% of the maximum inspiratory pressure (MIP) value measured previously by manovacuometry, with a 1-hour washout. To monitor the hemodynamic repercussions, the CTB device was used. Inspiratory exercise As this was the first time participants used the linear load resistor, after the initial evaluation they were instructed on how to perform the IE and then remained at rest for 15 minutes before beginning hemodynamic monitoring. Following the load randomization done previously (placebo, 30% or 60%), the IE was performed for 15 minutes, with the patient in supine position on a reclining chair, at 45º of elevation. All participants used the same linear load resistor, but an individual filter from the same manufacturer was used and discarded after the experiment. Throughout the IE, the patient was instructed to perform inspiration and expiration according to the sound signal emitted by a software (Paced Breathing), so all participants performed 15 breaths per minute. 8 The sessions with the other loads were carried out after a one-hour interval between each. For the IE with placebo, the device's spring was removed, only the unidirectional valve was left, therefore no resistance was present to the patient's inspiration. Statistical analysis The appropriate number of participants to be studied was calculated based on previous publications that showed which intervention, such as the effects of exercise, caused significant changes, such as increased heart rate, among others. For this magnitude of effects and to set statistical power at 0.8 and alpha error at 0.05, the sample should be comprised of 20 individuals. All data were subjected to Kolmogorov-Smirnov analysis to determine whether or not there was a normal distribution of the sample and data. Hemodynamic variables during IE, in Placebo, 30% or 60% groups, were compared using the Student's t test for paired variables. For the association of independent variables, Pearson's correlation was used. When p values were significant, paired comparisons were made using the Bonferroni test (post-hoc). The data were transferred to a systematic spreadsheet in Prism GraphPad 5.0 software (GraphPad Software, San Diego, CA). Categorical variables were expressed as absolute numbers. All results were expressed as mean±standard deviation and p values <0.05 were considered statistically significant. Ethical considerations All participants in this study received detailed information about the purpose of the research and the procedures to be performed. The protocol was sent to UNIFESO's Research Ethics Committee and approved under opinion number 420.737, registered at Plataforma Brasil . Before taking part in the study, all participants signed the informed consent form, according to resolution 466/2012 of the National Health Council. Results Among the 29 participants selected for the study, 20 completed the experiment (9 patients refused to participate) (Figure 1). Table 1 describes the demographic, clinical and pharmacological treatment characteristics of the sample. Responses of central hemodynamic variables to IE The central hemodynamic response had a different behavior according to different IE loads in our sample. HR increased with loads of 30% (C30) (64 ± 15 vs 69 ± 15 bpm; p=0.005) and 657