Shortness of breath (dyspnea) is an important symptom during physical exertion in patients with chronic obstructive pulmonary disease (COPD) and is related to respiratory muscle weakness. Dyspnea is a multidimensional sensation. The sensory perceptual domain (perceived dyspnea intensity) has been study extensively. The perception of respiratory distress (unpleasantness of dyspnea) has not received as much attention. Inspiratory muscle training (IMT) has been shown to improve inspiratory muscle function and reduce dyspnea intensity. Balance impairments increasing the risk of falling is another recognized problem in patients with COPD. Postural balance has been shown to be especially impaired in patients with COPD who have pronounced respiratory muscle weakness. Improvements in respiratory muscle function might improve balance control in patients. Respiratory Muscle Metaboreflex is known as respiratory muscle work during exercise reflexively induces sympathetically mediated vasoconstrictor activity, there by compromising blood flow and oxygen delivery to active limb and respiratory muscles.

Eight weeks of controlled IMT is hypothesized to reduce both intensity as well as unpleasntness domain of dyspnea perception, improve postural control and improves blood flow and oxygen delivery to limb muscles in patients with COPD who have pronounced respiratory muscle weakness.

The purpose of this clinical trial is to elucidate mechanisms of dyspnea relief and improvements in postural control after inspiratory muscle training (IMT) in patients with COPD.

An endurance cycle exercise test (Constant work rate (CWR) test) will be used to evaluate dyspnea intensity and unpleasantness at comparable breathing efforts, at comparable work rates on the cycle ergometer, before and after IMT. Patients will be performing a CWR test at 75% of the peak work rate achieved during a maximal incremental cardiopulmonary exercise test (CPET). Before, during and after CWR cycling tests patients will rate their intensity of dyspnea, unpleasantness of dyspnea, breathing related-anxiety and leg discomfort using a modified 10- point Borg scale. Patients will be asked to report qualitative descriptors of dyspnea at the end of the CWR cycling test. Maximum duration of the CWR test will be recorded and levels of minute ventilation will be registered continuously throughout the exercise protocol. With this endurance exercise test the investigators will be able to assess changes in the onset of dyspnea (intensity and unpleasantness) and the performance in endurance exercise before and after IMT.

Surface electromyography (EMG), and a multipair esophageal electrode catheter system will be used during CWR exercise to evaluate respiratory muscle recruitment, respiratory effort and neural drive to the different respiratory muscles. The catheter will be inserted to continuously record Pes (esophageal pressure), Pgas (gastric pressure) and EMGdi (electromyogram of the diaphragm muscle). PesMax (maximal esophageal pressure), PgasMax (maximal gastric pressure) and PdiMax (maximal transdiaphragmatic pressure) will be obtained during maximal sniff and cough maneuvers. Transcutaneous surface electromyography (sEMG) techniques will be applied to scalene, sternocleidomastoid and parasternal intercostal muscles to register neural drive to these respiratory muscles. With this measurements, the investigators will be able to assess whether there are any changes in respiratory muscle recruitment patterns and respiratory neural drive to the different respiratory muscles after IMT.

During exercise 'Respiratory Muscle Metaboreflex' will lead to sympathetically mediated vasoconstriction of limb locomotor muscle, less blood and oxygen supply to active limbs muscle there by locomotor muscle fatigue occurred. Quadriceps twitch forces evoked by magnetic stimulation of femoral nerve will be assessed to measure locomotor muscle fatigue at identical exercise time points during CWR test before and after the intervention. With this technique, the investigators will be able to assess whether the onset of locomotor muscle fatigue is delayed after IMT.

An isocapnic hyperpnea trial will be performed to assess respiratory muscle perfusion, dyspnea intensity and unpleasantness, respiratory muscle recruitment pattern, respiratory effort and neural respiratory drive without the work of locomotor muscle before and after IMT. Patients will be asked to maintain a targeted minute ventilation pattern equal to their breathing frequency, tidal volume and minute ventilation recorded at rest and during the final minute of constant load exercise test (at ~75% WRpeak). Experimenters will provide verbal guidance to patients to adjust the rate and depth of their breathing such that the target ventilation will be obtained and maintained constant within ±5%. Isocapnia will be maintained by having subjects inspire from a Douglas bag containing 5% CO2, 21% O2, balance N2 that will be connected to a two-way non-rebreathing valve (model 2700, Hans Rudolph) by a piece of tubing.

Loaded breathing tests will be performed to assess dyspnea intensity and unpleasantness, respiratory muscle recruitment patterns, respiratory effort and neural respiratory drive. Patients will be asked to breathe through a tapered flow resistive loading (TFRL) device (powerbreathe KH1) as long as they can (3-7 minutes), the resistance will be set at 50% of patients' PImax. Heart rate, oxygen saturation, and number of breaths will be monitored. Before and after the test Borg dyspnea, inspiratory effort and unpleasantness will be recorded. The test will be repeated after 8 weeks of IMT, at the same resistance. Borg dyspnea, inspiratory effort and unpleasantness will be recorded at time limit of pre-training and at the symptom limit. The longest time for post training can be up to 15 minutes.

Respiratory (i.e., intercostal, scalene and abdominal) and locomotor muscle blood flow index (i.e., vastus lateralis) will be simultaneously measured during isocapnic hyperpnea trials as well as during CWR test (at 75% of the peak work rate) before and after IMT according to a previously established method using Near-Infrared Spectroscopy (NIRS) and indocyanine green (ICG). Respiratory (i.e., intercostal, scalene, abdominal) and locomotor muscle (i.e., vastus lateralis) oxygen delivery will be calculated by multiplying blood flow index to arterial oxygen content, the latter will be calculated non-invasively by pulse oximetry. Respiratory (i.e., intercostal, scalene, abdominal) and locomotor muscle (i.e., vastus lateralis) oxygen saturation (Stio2) - an index of oxygen availability reflecting the balance between oxygen supply and demand-, will be recorded continuously during the trial by NIRS.

A noninvasive technique for studying the neural processing of respiratory sensations will be used to assess changes in the affective unpleasantness component of dyspnea during a standardized loaded breathing task. Electroencephalography (EEG) will be used to measure respiratory related evoked potentials (RREPs) during loaded and unloaded breathing. The RREP recorded from EEG is a measurement of cerebral cortical activity, which is elicited by the activation of lung and muscle mechanoreceptors due to short inspiratory occlusions. Patients will be wearing EEG sensors [129 channel system, Electrical Geodesics Inc., Eugene, USA] and breathe through a breathing circuit with a non-rebreathing valve via a mouthpiece. The inspiration will be interrupted briefly for 150 milliseconds every two to six breaths by activation of the occlusion valve with pressurized air, which induces the RREP. Patients will be rating their perceived intensity and unpleasantness of dyspnea and intensity of occlusion on a Borg scale while inspiratory load and occlusion will be applied via a breathing valve. This method will be used to assess if there is less unpleasantness associated with a giving level of dyspnea elicited by a resistive breathing task and whether these changes are correlated with changes in the central processing of the dyspnea sensation.

To evaluate postural balance, displacement of the center of pressure (CoP) in will be estimated from the raw force plate data using the equation: CoP = Mx/Fz (medio-lateral) and CoP=My/Fz (anterior-posterior). CoP will be measured during upright standing with and without vision, on stable and unstable (foam pad) support surface. During some conditions, local muscle vibration on ankle and/or back muscles will be applied to evaluate the role of proprioception in postural control. Furthermore, in some conditions a repetitive ballistic arm movement will be asked to evaluate the effect of a internal perturbation on postural control. Root mean square values of the CoP displacements will be used for the analysis of postural stability measures and mean values will be calculated for the vibration trials to analyze the expected directional effect. A ratio of the CoP displacements of the ankle muscles vibration trial regard to the back muscles vibration trial will be calculated to determine the proprioceptive postural control strategy.

A sample size of 16 participants for intervention group and 8 for control group are required to detect a difference of one unit in dyspnea Borg-10 scale, assuming a SD of 1 unit in the change in dyspnea score between pre and post measurements (power 80%, level of significance p<0.05). These estimates are based on previous work on dyspnea during exercise. Therefore 24 clinically stable COPD patients with inspiratory muscle weakness (PiMax<70% predicted or <60cmH2O) and dyspnea symptoms (BDI<7) will be included. Patients who are unable to perform exercise testing will be excluded.

Patients will perform daily training consisting of two training sessions of 30 breaths (intensity ~50% of PiMax; 4-5 minutes per session). One session per week will be performed supervised at the research center. IMT will be performed using an electronic tapered flow resistive loading (TFRL) device [POWERbreath®KH1, HaB International Ltd., Southam, UK] for 8 weeks. The PiMax will be measured every week in order to increase the appropriate training intensity to around 50% of the PiMax at that moment. The sham group will perform IMT at an inspiratory load that is not expected to improve inspiratory muscle function (intensity <10% of baseline PiMax; unchanged throughout the protocol).

Differences in primary and secondary outcomes between groups after 8 weeks of IMT will be compared adjusting for baseline differences in an analysis of covariance (ANCOVA).

Inclusion Criteria:

• Clinical Diagnosis of COPD
• Inspiratory Muscle Weakness (Pi,max <70% predicted or < 60 cmH2O)
• Baseline dyspnea index (BDI) < 7
• Peripheral muscle fatigue present after CPET

Exclusion Criteria:

• Major cardiovascular
• Limiting exercise capacity more than pulmonary function impairment
• Severe orthopedic with major impact on daily activities
• Psychiatric or cognitive disorders
• Progressive neurological or neuromuscular disorders
• Longterm O2 therapy
• Previous inclusion in rehabilitation program (<1 year)