A brief bout of exercise in hypoxia reduces ventricular filling rate and stroke volume response during muscle metaboreflex activation

Eleven healthy and recreationally active Caucasian males aged 22–46 years were recruited to participate in the study. All subjects were regularly involved in leisure-time sports activities at least three times/week. Their average values ± standard deviation (SD) of age, body mass, and height were 32.7 ± 7.2 years, 71.8 ± 10.3 kg, and 174.5 ± 5.0 cm, respectively. All participants underwent a preliminary medical examination to assess their health status. None of them suffered from cardiovascular or respiratory diseases or were on medication at the time of the experiment. All the subjects were non-smokers and abstained from drinking alcohol or coffee for at least 24 h before scheduled tests.

The study was conducted according to the Declaration of Helsinki and was approved by the ethics committee of the University of Cagliari. All the participants signed written informed consent before the beginning of the study.

Preliminary test Each participant underwent a preliminary cardiopulmonary exercise stress test (CPET) on an electromagnetically braked cycle-ergometer (CUSTO Med, Ottobrunn, Germany). Oxygen uptake (\(\dot{V}\)O₂), carbon dioxide production (\(\dot{V}\)CO₂), and V_E were assessed with a gas analyser (VO2000, MedGraphics St. Paul, MN, USA) calibrated immediately before each test. The exercise protocol was incremental and it consisted of a linear increase of workload (30 W min⁻¹), starting at 30 W, keeping a pedalling frequency of 60 rpm until exhaustion, which was considered as the point at which the subject was unable to maintain a pedalling rate of at least 50 rpm. Anaerobic threshold (AT), maximum workload (W_max), and maximum oxygen uptake (\(\dot{V}\)O_2max) were gathered. The following criteria were employed to consider \(\dot{V}\)O_2max achievement: a plateau in \(\dot{V}\)O₂ despite increasing workload (< 80 ml min⁻¹); (2) respiratory exchange ratio (RER) above 1.10; and (3) HR ± 10 beats min⁻¹ of predicted maximum HR calculated as 220-age (Howley et al. 1995). AT was calculated using the V-slope method, which detects AT using a computerised regression analysis of the slope of \(\dot{V}\)CO₂ plotted as a function of \(\dot{V}\)O₂ (Beaver et al. 1986). During the preliminary test, participants familiarised with the laboratory members and equipment, allowing habituation to the environment and the ergometer that was employed in the successive experimental sessions.

Test to study the metaboreflex during normoxia and hypoxia After the preliminary test (interval 4–7 days), subjects underwent randomly assigned two tests to study the metaboreflex in two different conditions: one test was conducted in normoxia (NORMO) and one in normobaric hypoxia (HYPO). Test NORMO and HYPO were separated by at least 7 days (interval 7–10 days). In detail, during both NORMO and HYPO sessions, participants were connected by a mask to a hypoxic gas generator (Everest Summit II Generator, Hypoxico, New York, USA). This device separates nitrogen from oxygen thanks to a molecular sieve system that uses zeolites and provides a gas mixture with a reduced oxygen content that can be regulated to reach a minimum FiO₂ of 12.5%. The latter corresponds approximately to the partial pressure of oxygen at 4000 m of altitude. A gas mixture with a FiO₂ of 13.5% (corresponding to an altitude of about 3500 m) and of 21% were delivered during the HYPO and the NORMO test, respectively. Subjects were blinded about the actual content of oxygen they were breathing, which was constantly checked by an operator by means of oxygen analyser provided with the device (Maxtec, Handi+, Salt Lake City, UT, USA). A similar experimental setting was already employed in our lab in a recent investigation (Mulliri et al. 2019).

During both NORMO and HYPO tests, the subject performed randomly assigned two exercise tests pedalling on the same cycle-ergometer utilised for the CPET. The two tests were almost identical, with the exception of the first 3 minutes of the recovery phase. In detail, the subject in study sat on the cycle-ergometer for 3 min of rest; then, he pedalled for 3 min against a workload corresponding to the 30% of the W_max reached during the CPET. The recovery phase could be either PEMI or control exercise recovery (CER). In detail, during the PEMI test, immediately after the cessation of exercise, a cuff previously applied to the subject' thigh was inflated (in less than 3 s) to a pressure of 50 mmHg above the blood pressure measured at the third minute of exercise. The cuff was kept inflated for 3 min and then removed. Then, a further period of 3-min recovery was allowed. This procedure has been several times employed to elicit the metaboreflex (Crisafulli 2017; Crisafulli et al. 2008; Scott et al. 2002). Indeed, the compression applied after exercise induces a temporary arterial and venous occlusion that traps muscle metabolites produced during exercise thereby stimulating the metaboreflex and, at the same time, excluding the influence of mechanoreflex and central command (Crisafulli et al. 2008; Scott et al. 2002). The PEMI manoeuvre has been demonstrated to be effective in evoking changes in cardiac preload, afterload, and inotropism in healthy humans as well as in patients suffering from various cardiovascular diseases (Crisafulli 2017; Magnani et al. 2018; Marongiu et al. 2013; Milia et al. 2015; Mulliri et al. 2016; Roberto et al. 2017).

The whole PEMI session lasted 12 min in total (i.e., 3 min of rest, 3 min of exercise, 3 min of PEMI, and 3 min of further recovery). The CER session, which had the same duration (i.e., 12 min), was similar to the PEMI session, but the recovery period after exercise was conducted for 6 min without applying any occlusion on the exercised limb. Similar protocols (i.e., 3 min of rest, 3 min of exercise, 3 min of PEMI, and 3 min of further recovery) were employed in previous investigations conducted in our lab to elicit the metaboreflex by PEMI with different muscle groups (Crisafulli 2017). Furthermore, the same protocol was recently employed to study the metaboreflex in normoxia after bouts of exercise in hypoxia (Mulliri et al. 2019). PEMI and CER sessions were separated by a recovery of at least 30 min. Recovery was considered complete when HR return to values not higher than 5 bpm with respect to the pre-exercise level. The design of the study is summarised in Fig. 1. As previously pointed out, PEMI and CER tests were completed in normoxia and hypoxia.

All experiments were conducted in a room at controlled temperature and humidity (22 °C, relative humidity 50%). Tests’ randomisation was obtained using an online random sequence generator (https://​www.​random.​org/​sequences/​).

Throughout all sessions, hemodynamics were collected by mean of impedance cardiography (NCCOM 3, BoMed Inc., Irvine, CA). This method has been previously used in similar experimental settings (Crisafulli et al. 2008; Mulliri et al. 2019). The impedance method allows to detect SV from changes in the thoracic impedance (Z₀) measured while a low-amplitude alternate electrical current is applied to the thorax. Since electricity follows the ways of less impedance, the electrical current mainly flows along the great vessels in the mediastinum (aorta, superior, and inferior vena cava) so that the volume of blood inside the aorta is the major determinant of Z₀. It follows that changes in Z₀ reflect changes in blood volume inside the aorta, which in turn depends on SV. Standard formulas can be employed to derive SV from Z₀ changes.

In detail, analog impedance and ECG traces were converted in digital signal and stored with a dedicated digital recorder (ADInstruments, PowerLab 8sp, Castle Hill, Australia) at a sampling rate of 500 Hz. ECG, Z₀ and its first derivative (dZ/dt) were analysed offline and used to compute the pre-ejection period (PEP), the ventricular ejection time (VET), and the HR as the reciprocal of R–R intervals. For each heart beat analysed, PEP corresponded to the time interval between the onset of the QRS wave on the ECG and the beginning of the systolic deflection on the dZ/dt trace. VET was calculated as the time interval between the systolic deflection of dZ/dt and the local minimum of dZ/dt measured in the same cardiac cycle. SV was indirectly calculated from Z₀, the local maximum of dZ/dt, and VET according to Bernstein's formula (1986). Diastolic time (DT) was measured as the difference between the R‐R interval and the sum of PEP and VET. VFR, an index of cardiac preload and diastolic function, was obtained as the ratio of SV and DT (Gledhill et al. 1994; Marongiu et al. 2013; Milia et al. 2015). Ventricular emptying rate (VER; a measure of cardiac systolic performance) was calculated as the ratio of SV and VET (Gledhill et al. 1994; Sanna et al. 2017).

CO was obtained as the product of SV and HR, while systolic (SBP) and diastolic blood pressure (DBP) were measured by means of manual sphygmomanometer applied on the non-dominant arm. To avoid any operator-dependent bias, the same physician took care of the blood pressure measurement throughout all the experimental sessions. MAP was derived from SBP and DBP according to Moran's formula that corrects MAP measure considering changes in DT and systolic time during tachycardia (Sainas et al. 2016). Finally, SVR was indirectly obtained from the ratio of MAP and CO multiplied by 80, a conversion factor applied to correctly represent SVR as standard resistance units.

To confirm that the hypoxic stimulus was effective, peripheral blood O₂ saturation (SO₂) was continuously measured through finger pulse oxymetry (Nonin, SenSmart X-100, Plymouth, MN, USA). Moreover, cerebral tissue oxygenation (Cox) was assessed using near-infrared spectroscopy (NIRS) (Nonin, SenSmart X-100, Plymouth, MN, USA). One NIRS probe was positioned on the left side of the forehead over the ipsilateral eyebrow. The sensor was taped and covered with a headband to keep the probe in a fixed position and prevent outer light from interfering. The operator that instrumented the subjects for SO₂ and Cox verified that the headband was comfortable and did not cause any blood flow occlusion.

Data are presented as mean ± SD. Changes in Cox are reported as percent variation against the baseline. All recorded data were averaged over 1 min. Differences in SO₂ and Cox were assessed using a two-way analysis of variance (ANOVA) (factors of time and condition: NORMO and HYPO) followed by Bonferroni post hoc when appropriate. Hemodynamic values were analysed at the third minute of rest, exercise, and PEMI (when metaboreflex activity was expected to be in a steady-state). Two-way ANOVA was employed to confront hemodynamic data for the effects of test (PEMI and CER) and condition (NORMO and HYPO) followed by Bonferroni post hoc when appropriate.

To further analyse the effect of metaboreflex activity on hemodynamic variables, the PEMI minus CER difference at the third minute after exercise was evaluated for all the measurements analysed. This method allowed metaboreflex response to be assessed, i.e., the response due to the metaboreflex activity (Crisafulli et al. 2013; Mulliri et al. 2016). Paired sample t test was used to detect changes in metaboreflex response between NORMO and HYPO. Statistical analysis was performed using commercially available software (GraphPad Prism). A p value < 0.05 was considered to determine statistical significance.

Some limitations of the present study should be honestly acknowledged.

Specifically, no blood samples were collected during experiments, which makes our hypothesis of a metabolite-induced venous dilating effect during exercise in hypoxia speculative at this time. Further studies comprehending blood analysis of NO by-products (nitrites, nitrates) are required to confirm if NO exerts an active role in reducing venous return.

Another limit was the lack of cardiac volumes assessment. In this regard, the addition of echocardiographic and Doppler measurements would have been useful, as this imaging technique could allow the evaluation of end diastolic volume and the indirect measure of pulmonary artery pressure. Thus, echocardiography may clarify the possible mechanisms behind the reduced SV and VFR responses that occur during the metaboreflex in hypoxia. This limitation clearly contributes to cautious interpretation of our results.

Overall, the results of the present investigation supported the hypothesis that brief exercise bouts in hypoxia was capable to blunt the stroke volume response during the metaboreflex activation obtained with the post-exercise muscle ischemia method under hypoxic condition. This was the consequence of a reduced ventricular filling rate probably due to a metabolite-induced venodilation which counteracted the sympathetic-induced venoconstriction that normally occurs during the metaboreflex. Results are in good accordance with those of a recent study conducted during the metaboreflex in normoxia after exercise in hypoxia and collectively support the concept that among hemodynamic modulators, cardiac pre-load is the variable more sensitive to hypoxic stimuli. Considering that exercise in hypoxia has been proposed as a useful tool for training and therapeutic purposes, its cardiovascular effects should be further investigated to better understand physiological and hemodynamic consequences.