Inspiratory muscle training for enhancing repeated-sprint ability: A pilot study

This pilot study examined the effect of inspiratory muscle training (IMT) on repeated-sprint ability and vastus lateralis reoxygenation. Ten recreationally trained subjects were randomly divided into two groups to complete 4 weeks of IMT or Sham (placebo) training. Pre- and post-intervention, a repeated-sprint ability (RSA) test was performed in both normoxia and hypoxia (FiO2 ≈ 14.5%). Vastus lateralis reoxygenation (VLreoxy), defined as peak to minimum amplitude deoxyhaemoglobin for each sprint/recovery cycle, was assessed during all trials using near-infrared spectroscopy. For total work performed, power analysis revealed that for small, medium and large effects (Cohen’s f), sample sizes of n = 8, 16 and 90 respectively, are required to achieve a power of 80% at an α level of 0.05. Maximal inspiratory mouth pressure increased in IMT by 36.5%, 95% CI [20.9, 61.6] and by 2.7%, 95% CI [−4.46, 8.8] in Sham. No clear difference in the change of work completed during the sprints between groups were observed in normoxia (Sham −0.805 kJ, 95% CI [−3.92, 0.39]; IMT −2.06 kJ, 95% CI [−11.5, 4.96]; P = 0.802), or hypoxia (Sham −3.09 kJ, 95% CI [−7, 0.396]; IMT 0.354 kJ, 95% CI [−1.49, 2.1]; P = 0.802). VLreoxy in IMT increased by 9.34%, 95% CI [5.15, 13.7] in normoxia only. In conclusion, despite a large increase in IMT, this was only associated with a small effect on RSA in our pilot study cohort. Owing to a potentially relevant impact of training the inspiratory musculature, future studies should include a sample size of at least 16-20 to detect moderate to large effects on RSA.


Introduction 19
During whole-body moderate-intensity exercise, the oxygen cost of breathing contributes 3-6% 20 towards total pulmonary oxygen uptake (VȮ 2 ), which increases to 10-15% during high-intensity exercise 21 [1]. Moreover, if a high work of breathing is sustained, respiratory muscle fatigue can develop, resulting 22 in a reflex increase in muscle sympathetic nerve activity [2]. This response, known as the respiratory 23 muscle metaboreflex, attenuates locomotor muscle blood flow in favour of the respiratory musculature, 24 which hastens the development of locomotor skeletal muscle fatigue [3]. Moreover, exercise in hypoxia 25 is associated with a higher ventilatory equivalent for oxygen and peripheral muscle fatigue [4].

26
Therefore, respiratory muscle training may represent an effective strategy to alleviate the detrimental 27 effect of sustained high work of breathing during intense exercise, particularly under hypoxic conditions.

28
Inspiratory muscle training (IMT) has been associated with enhanced exercise performance 29 during the Yo-Yo intermittent recovery test [5,6], time-trials [7][8][9], constant load cycling [10,11], and 30 repeated-sprint exercise (RSE) [12]. By improving the functional capacity of the respiratory muscles, 31 the relative intensity of breathing at a given ventilatory rate decreases. Reducing the relative intensity 32 of hyperpnoea following IMT has been shown to blunt the respiratory muscle metaboreflex [13,14], 33 reduce the O 2 cost of breathing [15], and lessen respiratory muscle fatigue in both normoxia and hypoxia 34 [16]. The application of IMT as a method to enhance repeated-sprint ability (RSA) has only been tested 35 in field-based protocols [12,17], with no work to our knowledge in a controlled laboratory setting under 36 hypoxic conditions.

37
The ability to maintain performance during RSE is underpinned by the capacity to deliver O 2 to 38 the locomotor muscles in the short rest periods between sprints [18]. Thus, when RSE is performed in 39 hypoxia, this capacity is negatively impacted [19]. However, respiratory muscle oxygenation appears to 40 be protected, potentially reflecting preferential blood flow redistribution to the respiratory muscles [20].

41
Interestingly, there is some indication that heightened respiratory muscle work has little consequence 42 on locomotor muscle oxygenation during RSE performed in normoxia [21]. It is possible that when high-43 intensity exercise is interspersed with rest periods, there is enough capacity in the cardiovascular 44 system to maintain O 2 supply to both the locomotor and respiratory muscle. Nevertheless, IMT has 45 been shown to improve RSE in normoxic conditions [12], and thus, could also be beneficial for RSE in 46 hypoxia where a higher work of breathing is typically incurred [4]. By enhancing the capacity of the 47 respiratory musculature, the activation of the respiratory muscle metaboreflex may be delayed, thereby 48 improving RSE [15]. We examined data from a previous investigation [20], and carried out a pilot study 49 to determine the feasibility of IMT to induce an ergogenic effect on RSE performance in normoxia and 50 hypoxia.

52
Power and sample size estimation

65
Power (1 -β) was calculated as a function of sample size (n) and effect size (Cohen's ƒ). Effect 66 size thresholds were set at small, 0.01; medium, 0.25; large, 0.4. The effect of α error probability was 67 also assessed using at the 0.01, 0.05 and 0.1 level.

68
Pilot study 69 Design 70 Ten males accustomed to high-intensity activity were recruited to participate in this study 71 (Sham: age = 24.8 ± 2.4 years, body mass = 77.0 ± 10.3 kg, height = 77.6 ± 6.8 m; IMT: age = 27.2 ± 72 2.2 years, body mass = 80.2 ± 9.3 kg, height = 179.0 ± 9.0 m). Subjects self-reported to be healthy and 73 with no known neurological, cardiovascular or respiratory diseases. After being fully informed of the 74 requirements, benefits, and risks associated with participation, each subject gave written informed

106
Therefore, each sprint/recovery duty cycle was 40 s in total. Before the experimental sessions, 107 participants completed two familiarisation trials within one week of commencing the study.

Near-infrared spectroscopy 109
Locomotor muscle oxygenation was measured using NIRS (Oxymon MKIII, Artinis, The 110 Netherlands). The optical sensor was fixed over the distal part of the vastus lateralis muscle belly 111 approximately 15 cm above the proximal border of the patella. Source-detector optode spacing was set 112 to 4.5 cm, and a differential pathlength factor of 4.95 was used [20]. Data were acquired at 10 Hz. A 113 10 th order zero-lag low-pass Butterworth filter was applied to the data to remove movement artefact and 114 signal oscillation due to pedalling [24]. The filtered signal was used for all data analysis thereafter.

115
Vastus lateralis deoxyhaemoglobin (HHb VL ) was normalised to femoral artery occlusion so that 0% 116 represented a 5 s average immediately prior the occlusion and 100% represented the maximum 5 s 117 average. Arterial occlusion was achieved by placing a cuff around the root of the thigh, which was 118 inflated to 300-350 mmHg until HHb VL plateaued (3-7 min). Peaks and nadirs were identified for each 119 40 s sprint recovery period, and VL reoxy was calculated as the difference between the peak to nadir of 120 the HHb VL signal.

Power and sample size estimation 131
Adopting the conventionally accepted power of 80% and a 5% α error probability (Fig 2 B)

149
Outcomes from the ANOVA's are presented in Table 1 and Table 2

165
The primary objective of this pilot study was to assess the feasibility of using IMT as a tool for 166 enhancing RSA and improving locomotor muscle tissue oxygenation. Based on the power analysis and 167 sample size estimation of total work, we estimated a sample size between 8 to 90 participants would 168 be required to detect large and small effects, respectively. Considering our sample size of n = 10 (two 169 groups of 5), we should have been able to detect a change in total work if the true effect size was at 170 least large, well beyond the effect sizes (Choen's ƒ) observed in the present study of 0.092 and 0.052 171 Table 1 and Table 2). Given that trained individuals are already well adapted to the demands of high-172 intensity exercise, it may be that IMT only yields small to moderate effects on performance. Recruiting 173 90 participants for a training study is not feasible for many exercise science research programs; thus 174 we suggest a total sample size of 16 to 20 participants (two groups of 8 to 10) to provide appropriate 175 statistical power if the effects of IMT on RSE performance is at least moderate.

176
In the present study, we did not observe any clear performance benefit of IMT in either normoxic 177 or hypoxic conditions (Fig 4). protocol, or one with shorter rest periods relative to the sprint, the benefits of IMT may be more obvious.

193
We observed a 36.5% increase in inspiratory muscle strength after the 4-week training 194 intervention (Fig 3) (Fig 4 D). Therefore, it may be that respiratory muscle 206 work (oxygen utilisation) has little effect on locomotor muscle oxygenation in RSE. Previously we have 207 demonstrated that despite an increased inspiratory muscle force development, intercostal muscle tissue 208 oxygenation (ration of oxyhaemoglobin to total haemoglobin) can be maintained relative to breathing 9 209 freely during RSE [21]. The intermittent nature of RSE likely protects against any meaningful 210 competition between the locomotor and respiratory muscles for available oxygen supply.

212
These pilot data showed that IMT readily increases the strength of the inspiratory muscles; 213 however, no effect on RSA was found. Based on our sample size calculations, we estimate that we only 214 had the sensitivity to detect a large effect at 80% statistical power. Moreover, to detect medium and 215 small effects, at least 16 and 90 subjects would need to be recruited, respectively. Based on the 216 resources of exercise science laboratories, recruiting 90 subjects may not be feasible. Therefore, we 217 recommend a total sample size of 16-20 is recruited for at least moderate effect sizes to be detected,

218
thus minimising the chances of type II error. Lastly, a double baseline should be utilised to establish the 219 smallest worthwhile change for which the magnitude to the training effect can be judged against.

221
We would like to thank Mario Popovic for his assistance during data collection. We would also 222 like to thank the laboratory technical staff, Samantha Cassar, Jessica Meilak, and Collene Steward.