Study design and participants
This is a cross-sectional observational study in which the assessors were not blinded for the different subgroups and following the STROBE recommendations [29]. The study was performed in accordance with the Declaration of Helsinki (2013) (approved by the ethical committee of the University of Brasília, CAAE 81309417.7.0000.8093). After a careful explanation of the nature and risks of the experimental procedures, all participating patients provided informed consent before starting the measurements. The study was realized between June 2018 and September 2019 at the University of Brasília.
Male and female individuals from a convenient sample, diagnosed with HFpEF or HFrEF, stable and under optimal medical treatment, were recruited and allocated by phenotype. The inclusion criteria were: 1) minimal age of 35 years; 2) at least six months of HFrEF or HFpEF diagnosis [5]; 3) HF with ischemic, hypertensive, or idiopathic etiology; 4) clinically stable for at least three months; and 5) a sedentary lifestyle (in the last six months). The exclusion criteria were: 1) clinically diagnosed pulmonary, inflammatory, musculoskeletal, or orthopedic diseases precluding exercise performance; and 2) functional New York Heart Association (NYHA) [30] class IV.
All participants were assessed during four experimental visits. The first visit was directed to clinical assessment, body composition, and pulmonary function; the second to echocardiogram assessment; and the third for muscle ultrasound and cardiopulmonary exercise testing. Finally, a fourth visit was planned to assess the isokinetic muscle strength and local oxygen with near-infrared spectroscopy (NIRS).
Baseline clinical characteristics
Patients were evaluated by a cardiologist who collected detailed information about the clinical history, diagnosis, and current symptoms. The NYHA [30] and Weber [31] functional classification were included to provide complementary clinical information regarding HF severity. The whole-body composition was assessed using dual-energy X-ray absorptiometry (DXA; Supplementary Material), cardiac function using echocardiography (Supplementary Material), pulmonary function via spirometry (Supplementary Material), and cardiorespiratory fitness via cardiopulmonary exercise test (CPX; Supplementary Material).
Isokinetic muscle strength test
Isokinetic muscle strength tests were performed using the Biodex system III Isokinetic Dynamometer (Biodex Medical, Inc., Shirley, NY). The axis of rotation of the dynamometer arm was adjusted to the right knee, and velcro belts were used to secure the thigh, pelvis, and trunk to the chair to prevent compensatory body movement. The lateral femoral epicondyle was used as the bony landmark for matching the knee joint with the axis of rotation of the dynamometer resistance adapter. Gravity correction was obtained by measuring the torque exerted on the dynamometer resistance adapter with the knee in a relaxed state at full extension. Patients were instructed to fully extend and flex the knee and work maximally during each exercise set. Verbal encouragement was given throughout the test session.
Isokinetic muscle strength assessment protocol comprised 20 repetitions, requiring maximum concentric effort required at an angular velocity of 180°/s. Patients performed six initial submaximal repetitions for familiarization purposes. After three minutes of rest, the isokinetic muscle strength test was performed [32–37]. Variables analyzed were peak torque (Nm) and adjusted per body weight ratio (Nm.kg), total repetition maximum work (J) and adjusted per body weight ratio (%), total work (J), work fatigue (%), and average power (W).
Near-infrared spectroscopy (NIRS)
During isokinetic muscle strength testing, a near-infrared spectroscopy (NIRS) device with a dual-wavelength (760 and 850 nm), continuous-wave system type, containing three pairs of LEDs configured for spatially resolved spectroscopy (SRS) with a source-detector spacing of 30, 35, and 40 mm were utilized to assess local oxygen extraction response (Portamon for OxySoft 3.0.95, Artinis Medical Systems, Amsterdam, Netherlands). Changes in absorbance were recorded using the oxyhemoglobin (O2Hb, µM) and deoxyhemoglobin (HHb, µM) values to assess the oxygenation status of the muscle [38]. In addition, the tissue saturation index (TSI, %) was calculated from the absorption of coefficients derived from the attenuation of light at different source-detector distances and wavelengths as a relative value (%), being feasible for comparing and evaluating the achievement of critical limits during exercise. For this, the equipment was positioned on the right leg vastus lateralis (approximately 5 cm from the lateral patellar border), covered with a dark blue elastic band to avoid interference from ambient light and adhesive tape without pressing the equipment. The data were sampled at 10Hz and stored for offline analysis using the LabChart Pro v8 software (ADInstruments, Sidney, Australia).
For statistical analysis and graph signal processing analysis of the NIRS curve, baseline (mean obtained value for the 30s of the resting phase), exercise (lowest obtained value for TSI, %, and, O2Hb, µM and highest for the HHb, µM) with a maximum interval variation acceptance of 4 seconds (20 to 24s, depending on manual NIRS mark) and recovery (highest obtained value for TSI, % and O2Hb, µM and lowest for the HHb, µM) were considered as time points for comparison [25;39]. An individual visual inspection of the curves was made to exclude possible failures/noise from the graph signal. Lately, eligible individuals were analyzed and presented on graphs that included the individual mean values from each variable (representative cases).
Ultrasound-derived measures: echo intensity and muscle thickness
The ultrasound images were captured by using an ultrasound device (HD11XE, Phillips, Amsterdam, Netherlands) with a 7.5MHz linear matrix transducer. The individuals were evaluated in a supine position with the knee in passive flexion with a 15cm under-knee support and neutral rotation. The images were always acquired on the right leg with the transducer placed transverse and perpendicular to the long axis of the anterior thigh, rectus femoris (RF), and vastus lateralis (VL) muscles (50% of the distance between the iliac spine anterior superior to the superior edge of the patella) to assess muscle thickness, using appropriate transmission gel [40]. The images were analyzed using the ImageJ software (1.52q version, Bethesda, EUA)[41]. The quadriceps femoris was analyzed between the uppermost part of the femur and the superficial fascia of the rectus femoris (which includes the rectus femoris and vastus intermedius) and the isolated rectus femoris [42,43]. The measurement of echo intensity was determined by a gray-scale analysis using ImageJ software. The region of interest was selected for each assessed muscle, including all muscle areas and removing bone or surrounding fascia from the selected area [42]. The mean was calculated using an 8-bit resolution measure, resulting in a number between 0 = black and 255 = white. An average of the three measurements per muscle was calculated. In the quadriceps femoris, only the rectus femoris muscle was used for analysis [42,43]. Patients were instructed not to perform any physical activities 24 hours before testing.
Statistical analysis
Data are expressed as mean ± standard deviation (SD), absolute (n), or relative frequencies (%). Shapiro-Wilk test was used to indicate sample data distribution. Parametric or non-parametric tests were applied accordingly. Group differences for continuous outcome variables were compared using unpaired t (mean difference and 95% confidence interval) or Mann-Whitney U test (Hodges-Lehmann's difference). Categoric variables were compared using Fisher's exact test.
We performed a bivariate correlation (Spearman's or Pearson's) analysis to investigate the associations between muscle microcirculation dynamics (TSI, %; O2Hb, µM and HHb, µM) and echo intensity (EI, 0-255) with isokinetic muscle strength (PT, Nm), cardiorespiratory fitness (peak V̇O2, mL.kg−1.min−1) and peak power output (Watts) among HF phenotypes (HFpEF and HFrEF) and severity of functional impairment classification (Weber A + B and Weber C).
Moreover, to verify possible association among muscular thickness (cm) with muscle strength (peak torque, PT, Nm), cardiorespiratory fitness (peak V̇O2, mL.kg−1.min−1), and peak power output (W). Association levels was defined according to correlation coefficient (r) (0.00 no association; 0.20 weakly; 0.50 moderately; 0.8 strongly and 1.00 perfectly) [44] or (rho) (0.00 to 0.20 negligible; 021 to 0.40 weak; 0.41 to 0.60 moderate; 0.61 to 0.80 strong and 0.81 to 1.00 very strong) [45].
Considering the absence of similar studies involving microcirculatory dynamics withing resistance exercise in heart failure, the calculation of sample size was not possible. Because of this, we decided to run the post-hoc analysis in order to detect the power calculation of the study (effect size). By considering an alpha error of 0.05, and the power (1-beta) 0.80 for the parameter O2Hb during the exercise period on Weber C in between phenotypes, we have an effect size result of 2.47 (high); for the parameter O2Hb during the recovery period on Weber C in between phenotypes, we have an effect size result of 2.38 (high); for the parameter HHb during the recovery period on Weber A + B in between phenotypes, we have an effect size result of 2.31 (high); for the parameter O2Hb during the exercise period on Weber A + B in HFpEF group, we have an effect size result of 2.07 (high); and for the parameter HHb during the recovery period on Weber A + B in HFrEF group, we have an effect size result of 2.68 (high). These parameters was chosen because its statistical difference is < 0.05. The effect size for group comparisons was estimated using G*Power Software 3.1.
Statistical software GraphPad Prism (8.4.0, California, San Diego) was used for statistical analyses and figure production. All analyses considered 95% confidence interval (CI) and statistically significant was set at p-value ≤ 0.05 (two-tailed).