Effect of Respiratory Muscle Training on Exercise Performance in Healthy Individuals: A Systematic Review and Meta-Analysis

  • 36 min read

Written and conducted by Sabine K. Illi, Ulrike Held, Irene Frank, and Christina M. Spengler

Article in Sports Medicine

Published July 2012



Objectives: Two distinct types of specific respiratory muscle training (RMT), i.e. respiratory muscle strength (resistive/threshold) and endurance (hyperpnoea) training, have been established to improve the endurance performance of healthy individuals. We performed a systematic review and meta-analysis in order to determine the factors that affect the change in endurance performance after RMT in healthy subjects.

Data sources: A computerized search was performed without language restriction in MEDLINE, EMBASE and CINAHL and references of original studies and reviews were searched for further relevant studies.

Review methods: RMT studies with healthy individuals assessing changes in endurance exercise performance by maximal tests (constant load, time trial, intermittent incremental, conventional [non-intermittent] incremental) were screened and abstracted by two independent investigators. A multiple linear regression model was used to identify effects of subjects’ fitness, type of RMT (inspiratory or combined inspiratory/expiratory muscle strength training, respiratory muscle endurance training), type of exercise test, test duration and type of sport (rowing, running, swimming, cycling) on changes in performance after RMT. In addition, a meta-analysis was performed to determine the effect of RMT on endurance performance in those studies providing the necessary data. 

Results: The multiple linear regression analysis including 46 original studies revealed that less fit subjects benefit more from RMT than highly trained athletes (6.0% per 10 mLkg-1 min-1 decrease in maximal oxygen uptake, 95% confidence interval [CI] 1.8, 10.2%; p = 0.005) and that improvements do not differ significantly between inspiratory muscle strength and respiratory muscle endurance training (p = 0.208), while combined inspiratory and expiratory muscle strength training seems to be superior in improving performance, although based on only 6 studies (+12.8% compared with inspiratory muscle strength training, 95% CI 3.6, 22.0%; p = 0.006). Furthermore, constant load tests (+16%, 95% CI 10.2, 22.9%) and intermittent incremental tests (+18.5%, 95% CI 10.8, 26.3%) detect changes in endurance performance better than conventional incremental tests (both p < 0.001) with no difference between time trials and conventional incremental tests (p = 0.286). With increasing test duration, improvements in performance are greater (+0.4% per minute test duration, 95% CI 0.1, 0.6%; p = 0.011) and the type of sport does not influence the magnitude of improvements (all p > 0.05). The metaanalysis, performed on eight  controlled trials revealed a significant improvement in performance after RMT, which detected by constant load tests, time trials and intermittent incremental tests, but not by conventional incremental tests. 

Conclusion: RMT improves endurance exercise performance in healthy individuals with greater improvements in less fit individuals and in sports of longer durations. The two most common types of RMT (inspiratory muscle strength and respiratory muscle endurance training) do not differ significantly in their effect, while combined inspiratory/expiratory strength training might be superior.  Improvements are similar between different types of sports. Changes in performance can be constant load teats only. Thus, all types of RMT can be used to improve exercise performance in healthy subjects but care must be taken regarding the test used to investigate the improvements.

1. Introduction

Respiratory muscle fatigue is known to compromise exercise performance in healthy subjects.[1,2] Evidence is emerging that fatiguing respiratory muscles may affect exercise performance via the so-called metaboreflex,[3] i.e. accumulation of metabolites, such as lactic acid, in the respiratory muscles activates group III and especially group IV nerve afferents[4-6] that then trigger an increase in sympathetic outflow from the brain causing vasoconstriction in the (exercising) limbs.[7-11] This consequently increases limb muscle fatigue during exercise[12,13] and results in earlier exercise termination compared with conditions where respiratory muscle fatigue is prevented.[14,15] Respiratory muscle training (RMT) has been shown to reduce the development of respiratory muscle fatigue,[16-18] blood lactate concentration during exercise[18-21] and sympathetic activation.[12,22] Therefore, a reduction or delay of the metaboreflex[3] described earlier might be an important mechanism for improving exercise performance by RMT. Interestingly, however, of those studies addressing the effects of specific RMT on exercise performance in healthy subjects, only about twothirds report significant improvements. Therefore, a detailed analysis of potential factors that may contribute to the success or failure of RMT is urgently needed. A brief overview of these factors is given below. First, study outcome may be related to study design, considering that only about half of the RMT studies included a sham-training group to account for a possible placebo effect of RMT. Second, subject selection might influence study outcome, since the extent to which respiratory muscles fatigue may differ, for example, with subjects’ fitness level. Indeed, several studies showed increased respiratory muscle endurance in physically trained compared with sedentary subjects.[23-25] However, when comparing subjects’ physical performance relative to their maximal performance, trained subjects worked at a higher percentage of their maximum and performed more respiratory muscle work,[26] which may theoretically neutralize the effect of increased respiratory muscle endurance on fatigue development. Only two studies[27,28] investigated the difference in development of respiratory muscle fatigue depending on subjects’ fitness. These suggested that respiratory muscles indeed fatigue less in endurance trained subjects compared with sedentary subjects during exhaustive physical exercise.[27,28] This indicates that less fit subjects would generally benefit more from RMT than highly trained athletes. Third, it is unclear whether the type of RMT might influence the degree of improvement in exercise performance. Currently, two distinct forms of RMT are used in healthy subjects: respiratory muscle strength training (RMST; also known as inspiratory muscle [strength] training [IM(S)T], inspiratory [flow] resistive loading [I(F)RL], resistive/ resistance respiratory muscle training [RRMT], concurrent inspiratory and expiratory muscle training [CRMT], or expiratory muscle training [EMT]) and respiratory muscle endurance training (RMET; also referred to as ventilatory muscle training [VMT], voluntary isocapnic hyperpnoea [VIH] or endurance respiratory muscle training [ERMT]). RMST is performed by breathing against an external inspiratory and/or expiratory load. This load consists either of a flow-dependent resistance or of a pressure threshold that needs to be overcome and sustained to generate flow. RMST includes high-force, low-velocity contractions and was shown to specifically increase respiratory muscle strength, i.e. maximal pressure generation capacity of the inspiratory and/or expiratory muscles against a closed airway.[29] In contrast, RMET is performed using normocapnic hyperpnoea. This training consists of low-force, highvelocity contractions of inspiratory and expiratory muscles, and results in improved respiratory endurance.[29] Whether strength or endurance training of the respiratory muscles is more effective in terms of improving exercise performance, remains unclear. From a physiological point of view, it seems that training both inspiratory and expiratory muscles would be most effective, since with elevated breathing, inspiratory as well as expiratory muscles are increasingly recruited. In fact, it has been shown by objectively assessing changes in transdiaphragmatic and abdominal muscle contractility after exercise that not only inspiratory[30-33] but also expiratory muscles[16,34-36] fatigue during exhaustive high-intensity endurance exercise. Therefore, a closer look at the effects of different training regimens is needed. Fourth, different studies use different types of exercise testing, e.g. incremental tests (IT), constant load tests (CLT) or time trials (TT) of different intensities, to assess the effect of RMT on exercise performance. Whether RMT is more likely to result in positive effects during some types of performance compared with others remains to be verified. Considering the degree to which respiratory muscles fatigue after these different types of tests, it could be argued that the effects of RMT are less likely to be detected in ITs than in the other types of tests. This assumption is based on the results of Romer et al.,[37] who demonstrated that the diaphragm of moderately fit subjects did not fatigue during an incremental cycling test, despite subjects reaching maximal exercise intensity. This is surprising, since Johnson et al.[30] showed that higher exercise intensities (oxygen consumption [ . VO2] at >85% maximal . VO2 [ . VO2max]) increase the likelihood for diaphragmatic fatigue to develop. It seems, therefore, that the duration for which a given intensity is sustained is as important as the intensity itself, with respect to both development of respiratory muscle fatigue and a possible benefit from RMT. Finally, the effect of RMT on performance might differ depending on the exercise modality used, e.g. rowing, running, swimming or cycling, since respiratory muscles are well known to realize more than just respiratory tasks, and these tasks differ between exercise modalities. In rowing, for instance, respiratory muscles need to combine the motion of the thorax expanding and contracting with the – sometimes opposing – rowing stroke movement.[38-40] In running, intra-abdominal pressure is increased, which has been attributed to a protecting function of the spine by the abdominal muscles.[41] Furthermore, the diaphragm has been shown to be activated to increase intra-abdominal pressure during movements of upper limbs, such as when running.[42,43] Thus, when running, respiratory muscles of the trunk also serve postural tasks. During swimming and diving, the work of breathing is increased due to the hydrostatic pressure against which the thorax expands causing an increase in end-expiratory lung volume, which in turn leads to suboptimal length for tension development of respiratory muscles.[44] In addition, respiratory muscles are involved in propulsion. Thus, subjects performing exercise modalities that require additional work from respiratory muscles might be more susceptible to respiratory muscle fatigue.[45] Consequently, subjects performing these exercise modalities might benefit the most from RMT. The aim of the present work was, therefore, to assess the importance of the above factors on the effect of RMT to improve exercise performance. For this purpose, a systematic review was performed in MEDLINE, EMBASE and CINAHL (up to October 2011), without language restriction, on all studies including an RMT intervention and assessment of endurance performance as an outcome variable, independent of the presence or absence of a sham-training and/or no-training control group. To specifically analyse the evidence of a positive effect of RMT on exercise performance, a meta-analysis including only controlled studies was performed.

2. Methods

A systematic review and meta-analysis were performed on original studies that assessed the effect of RMT (RMST or RMET) on endurance performance in healthy humans by use of at least one of the following exercise tests: a CLT with fixed exercise intensity and subjects performing to exhaustion; a TT with either a fixed distance or a fixed duration and with subjects being required to row, run, swim, cycle, etc. as fast as possible or to cover the largest possible distance; an intermittent incremental test (IIT) with a stepwise increase in exercise intensity including active recovery between steps and subjects performing to exhaustion; or a conventional (non-intermittent) IT with a stepwise increase in exercise intensity and subjects performing to exhaustion, a test that is frequently used to determine . VO2max and/or the anaerobic threshold.

2.1 Search

A computerized search without language restriction was performed in MEDLINE, EMBASE and CINAHL (up to 31 October 2011). The search strategy included the following keywords: ‘respiratory muscle training’, ‘inspiratory muscle training’, ‘expiratory muscle training’, ‘inspiratory training’, ‘expiratory training’, ‘hyperpnoea training’, ‘hyperpnea training’, ‘respiratory muscle endurance training’, ‘threshold training’, ‘resistive training’, ‘inspiratory loading’, ‘expiratory loading’ and ‘resistive loading’ combined with ‘human’, ‘healthy’ and ‘not patient’. Only published studies were included in the analysis.

2.2 Selection

All studies performing RMT in healthy subjects and assessing endurance performance as a main outcome were selected. RMT had to consist of either RMST or RMET. One study[46] combined RMST and RMET and was excluded due to the interaction of combined strength and endurance training in skeletal muscles yielding different specific adaptations compared with training in one modality alone.[47] Studies performing unloaded breathing exercises, breathing therapy, or similar, were not considered. First, all titles of the primary search were screened for potentially relevant articles. Of those, abstracts, reviews, short reports, case reports, editorials and letters were excluded. Original studies were excluded when RMT was not performed, endurance performance was not assessed, physical training was included as an additional intervention and when exercise tests were non-exhaustive. References of the included studies and of the excluded reviews were searched for further relevant studies

2.3 Quality Assessment 

The quality of the selected RMT studies was assessed using the following criteria.[48,49] (i) Randomization: random allocation of the subjects to intervention and sham-training or no-training control group. If a trial was called ‘randomized controlled’ but randomization was not described, it was considered to be a randomized trial (0 points if not reported or not randomized, 1 point if reported, 2 points if randomization procedure specified); (ii) Blinding: observer blinding to group allocation of the subjects (0 points if not reported, 1 point if the observer was blinded); (iii) Allocation of concealment: person in charge of subject recruitment was (at that time) unaware of potential group allocation (0 points if not reported, 1 point if specified); (iv) Dropouts: information about missing data (0 points if not reported, 1 point if reported); (v) Intention-to-treat analysis: all subjects initially considered for the study were included and data was assessed (0 points if not performed, 1 point if performed); (vi) Power calculation: statistical power of the study (0 points if not reported, 1 point if reported). Thus, a maximum of 7 points corresponding to 100% could be reached.

2.4 Data Abstraction Two investigators (IF, SKI) independently abstracted the data. Inconsistencies were cross checked, discussed with the third investigator (CMS) and resolved by consensus.

2.5 Quantitative Data Synthesis The main variable of interest was the change in endurance performance reported or calculated as the relative difference in test duration or – in case of a TT with fixed duration – the relative change in maximal distance covered. Additional variables of interest were (i) fitness level of the subjects (categorized as follows: level 1 if . VO2max was not provided,[50-59] fitness level was estimated from [a] endurance performance of subjects and [b] description of daily activities compared with subjects in the other studies using the same exercise modality); (ii) respiratory muscles that were trained, i.e. inspiratory and/or expiratory muscles and type of training, i.e. RMST or RMET (since specific expiratory muscle training was investigated in one single subgroup only,[60] it was excluded from the analysis and three categories were generated for the remaining types of training: RMST.IN [inspiratory muscle strength training], RMST.INEX [combined inspiratory and expiratory muscle strength training] and RMET); (iii) type of test, i.e. CLT, TT, IIT or IT; (iv) test duration before RMT; and (v) exercise modality, i.e. rowing, running, swimming (including diving) or cycling. Further potentially relevant variables, such as training modalities (e.g. number of training sessions, duration of a single training session, training intensity, etc.), intensity of the physical endurance test or subjective ratings of breathlessness and respiratory effort, were not included in the multiple linear regression model in order to prevent collinearity and/or as a consequence of missing information in too many of the studies. Collinearity means that two or more variables are interchangeable, e.g. test duration and test intensity are interchangeable because test duration becomes shorter with a higher test intensity. If both variables were included in the model at the same time, this would mean a high degree of multicollinearity and would give invalid estimates for individual predictors. Generalized estimating equations (GEE; with exchangeable correlation structure) were fitted to the dependent variable ‘change in endurance performance’ in order to account for clustered data. Independent variables in the multiple linear regression model were fitness, type of training (RMST.IN, RMST.INEX and RMET), type of test (CLT, TT, IIT and IT), test duration and type of sport (rowing, running, swimming and cycling), including all RMT studies independent of the presence or absence of a sham-training or no-training control group. The multiple linear regression model thus accounts for the influence of the above confounders on changes in exercise performance after RMT. The analyses were performed with R 2.13.1 (statistical computing software).[61] Data from two studies were reported in more than one publication.[17,62-64] In this case, the study that provided more details of the relevant data[63,64] was included in the analysis. Data from two tests were reported in two studies.[65,66] This data appears only once in the present analysis.[66] In three studies,[55,67,68] test duration at baseline was not indicated nor could it be calculated by use of the test protocol. Therefore, these studies were excluded from the regression model. Furthermore, a meta-analysis was performed on the main outcome of those studies that included a sham-training or no-training control group, and that reported the relative change and standard deviation in exercise performance, such that the overall difference in exercise performance including 95% confidence intervals (CIs) between RMT and sham/no-training control group could be calculated. Additionally, subgroup analysis for the different exercise tests was performed. Heterogeneity of the studies was assessed by calculating the I2 -statistics, which are known to be independent of the number of studies included in the meta-analysis, and thus preferable compared with the Cochrane’s chi-squared or Q test.[69] A value of I2 >50% was considered as evidence for heterogeneity.[70]A random effects model was chosen for all tests, since substantial heterogeneity was expected due to differences between fitness level, type of RMT, type and duration of test and exercise modality. As relative improvements in TTs are generally much smaller compared with those in CLTs, mean relative differences in exercise performance were standardized based on their standard deviations. A potential publication bias was assessed by use of a funnel plot. These analyses were performed with Review Manager (RevMan, Version 5.1, The Nordic Cochrane Centre, The Cochrane Collaboration, 2011, Copenhagen, Denmark). In both the multiple linear regression model and the meta-analysis, a p-value of 0.05 was considered significant. In those studies that did not report the standard deviations of relative changes in endurance performance, the relative differences between the RMT and sham/no-training control group are presented without 95% CIs. If this difference was not given, it was calculated from the difference in mean absolute values before and after RMT.

3. Results

3.1 Trial Flow and Study Characteristics

7385 citations were identified of which 236 potentially relevant articles remained for further evaluation (figure 1). Finally, 49 studies were selected.[16,20,21,38,39,50-60,63-68,71-97] Of these, 28 (57%) were randomized controlled trials, 6 (12%) were controlled trials, and 15 (31%) were non-controlled trials. Further characteristics of the studies are given in supplemental table I of the online Supplemental Digital Content (SDC) [http://links. adisonline.com/SMZ/A9]. Note that three[55,67,68] of the 49 studies were excluded from the multiple regression analysis due to missing test durations. Of these, the study by Lomax et al.[55] was included in the meta-analysis and the study by Chatham et al.[67] was included in the fourth figure (see section 3.4) only. Methodological quality scored between 14% and 86% (median 29%, i.e. 2 of maximum 7 points; see supplemental table II of the SDC). Study quality did not correlate with the main outcome, i.e. with the relative change in performance.

3.2 Study Design: Presence/Absence and Type of Control Group Thirteen studies (27%) included a no-training control group while 21 studies (43%) had a shamtraining group. Seventy-five percent of the noncontrolled studies showed an improvement in exercise performance after RMT. In studies with sham-training or no-training control groups, improvements for the RMT group were seen in 71% and 54%, respectively. In those studies that compared improvements of RMT and no-training control groups, improvements in the RMT group were significantly greater in 75% of studies with no-training control and in 69% of studies with sham-training control.

3.3 Linear Regression Model

Table I depicts the linear regression model. The model revealed that (i) less fit subjects benefit more from RMT than fitter subjects; (ii) effects of RMET and RMST.IN are similar, while RMST.INEX seems to be superior to RMST.IN and RMET; (iii) improvements in performance are greater in CLTs and IITs than in ITs, with no significant difference between TTs and ITs; (iv) greater improvements are seen with increasing test duration; and (v) improvements are independent of exercise modality

3.4 Meta-Analysis

RMT results in a significant increase in exercise performance (figure 2, standardized mean difference [SMD] 1.11, 95% CI 0.61, 1.61; p < 0.001), although with moderate heterogeneity (I2 = 71%). Subgroup analysis of different tests revealed significant improvements in exercise performance when assessed by a CLT (SMD 0.66, 95% CI 0.20, 1.12; p = 0.005), a TT (SMD 1.85, 95% CI 0.88, 2.82; p < 0.001) or by an IIT (SMD 2.96, 95% CI 1.12, 4.80; p = 0.002), whereas no significant improvement in exercise performance was detected when assessed by an IT (SMD 0.30, 95% CI -0.20, 0.79; p = 0.30). Furthermore, a significant difference between groups was found (p= 0.003) favouring TT and IIT over CLT and IT (individual p-values not shown), although with substantial heterogeneity (I2 = 78.3%). Subgroup analysis showed evidence for low heterogeneity in CLT and IT (I2 = 0% and 18%, respectively) but high heterogeneity in TT (I2 = 77%). Heterogeneity for IIT could not be calculated as only one study was included in the meta-analysis. Figure 3 shows a funnel plot of those studies that were included in the meta-analysis. Figure 4 shows the overall mean difference of the relative change in exercise performance for all controlled studies. The overall improvement for the RMT group over the sham-training or notraining control group was 11%, while subgroup differences were 21% for CLT, 2% for TT, 13% for IIT and 7% for IT (2% without the studies by Enright et al.[50,51]). These results are in agreement with those of the controlled studies included in the meta-analysis, i.e. RMT effects are seen in CLTs, IITs and TTs.


4. Discussion

The key finding of this analysis is that RMT improves performance in healthy subjects, independent of the type of RMT and exercise modality. Less fit individuals seem to benefit more from RMT than highly trained athletes, and improvements are greater with longer exercise durations. Improvements are significant when the effect of RMT is tested in CLTs, TTs and IITs, while none are seen in ITs, commonly used to assess . VO2max or anaerobic threshold.

4.1 Study Design: Presence/Absence and Type of Control Group

It could be assumed that study outcome may be related to study design, since only 43% of RMT studies included a sham-training group to account for a possible placebo effect of RMT. However, a closer look reveals that the presence and type of control group do not influence outcome. When considering differences between RMT and control groups, 75% of studies including a no-training control group and 69% of placebo-controlled studies showed a positive outcome for RMT (i.e. performance improvements for the RMT groups significantly exceeded those for the control groups), similar to the 75% positive outcome in studies without any controls. Thus, the presence or absence and type of control group did not affect the outcome regarding performance improvements for RMT studies in healthy subjects. Likely reasons for the lack of improvement in exercise performance after RMT in some studies include the use of only an IT to evaluate the effects of RMT on endurance performance,[73,82,86,94] low power of the studies,[78,89] lack of recovery time for respiratory muscles prior to the endurance exercise test,[38,46,76,78,96,97] very highintensity exercise,[20,38,58,78,89] a highly trained group of subjects[58,78,96,97] or an increased respiratory drive with concomitantly increased work of breathing in some subjects after RMET.[16,96]

4.2 Effect of Subjects’ Fitness on Improvements in Exercise Performance

The multiple linear regression analysis showed that less fit subjects benefit more from RMT than highly trained athletes. This finding is in accordance with the initial hypothesis suggesting that untrained subjects might benefit more from RMT, since respiratory muscles of less fit subjects were shown to fatigue more during exhaustive endurance performance.[27,28] However, although less fit subjects have a higher potential to increase their physical endurance performance compared with highly trained athletes,[98-101] respiratory muscle performance seems to improve to a similar extent with all levels of fitness. Also, when analysing improvements in maximal inspiratory mouth pressure (MIP), maximal expiratory mouth pressure (MEP) or respiratory muscle endurance separately for the different types of training, no effect of fitness could be observed (data not shown). On the other hand, it might be argued that greater improvements in performance are associated with older age rather than lower fitness, since . VO2max is known to decrease with age.[102] However, separate analyses showed that the relative improvement in exercise performance was negatively correlated with the level of fitness (r = -0.440; p < 0.001), while it was not correlated with age (r = -0.018; p = 0.882). Thus, it seems that the level of fitness is more important than age in affecting the amount of improvement in performance after RMT.


4.3 Influence of the Type of RMT on Improvements in Exercise Performance

The multiple regression analysis revealed that RMST.IN and RMET did not differ in their effect on improving exercise performance. This result seems astonishing, since the degree of fatigue developing during exhaustive exercise was shown to be similar in inspiratory and expiratory muscles[16,30-36] and both of these muscle groups[7,10] were shown to elicit the metaboreflex that is known to impair exercise performance.[14] Thus, one would assume that training both muscle groups, as with RMET, would yield a greater effect than training inspiratory muscles alone, such as during RMST.IN. Thus, the question arises: why would increased inspiratory muscle strength be advantageous for exercise hyperpnoea? Despite exercise hyperpnoea being characterized by high flows, it is known that inspiratory rib-cage muscles produce the pressures needed to expand the rib cage and thereby let the diaphragm act as the main flow generator.[103] Consequently, rib-cage muscles fatigue during high-flow tasks,[104] although to a lesser extent than with high resistances.[105] Thus, it seems likely that RMST.IN provides a larger training stimulus to inspiratory muscles than RMET and that more effectively trained inspiratory muscles, as with RMST.IN, may be superior in preventing or delaying the development of inspiratory ribcage muscle fatigue, compared with RMET. This per se would translate into a greater improvement in exercise performance with RMST.IN than with RMET of inspiratory muscles only. It has, however, been shown that RMET also trains expiratory rib-cage and abdominal muscles, in addition to the inspiratory muscles, which is substantiated in a smaller degree of expiratory muscle fatigue during exercise after this type of training.[16] Therefore, an explanation for the similar improvements in performance with RMST.IN and RMET might be that, on the one hand, inspiratory muscles were trained more effectively with RMST.IN than with RMET, and on the other hand, the combination of ‘less effective’ inspiratory muscle training with expiratory muscle training during RMET results in the same net effect with respect to improvements in exercise performance. The need for training the expiratory in addition to the inspiratory muscles on the one hand and the potential superiority of respiratory muscle strength over endurance training to improve exercise performance, on the other hand, would also be supported by the model showing that the combination of both inspiratory as well as expiratory muscle strength training, i.e. RMST. INEX, improved exercise performance more than RMST.IN or RMET. It should, however, be pointed out that so far only three research groups (six studies) used RMST.INEX and – although the model accounted for differences in fitness, type of testing and sports – one might need to consider that subjects in these studies were slightly less fit than those performing RMST.IN and RMET, and evidence for physical improvements came from one single research group (four studies) testing with CLTs only (known to yield greater improvements). Final proof of a potential difference between RMST.INEX versus RMST.IN and RMET can therefore only be provided when all three types of training are tested in the same study having similar groups of subjects and similar performance tests. For example, a direct comparison of RMST.IN and RMET in one single study showed that the effects of RMET were larger than those of RMST.IN, with respect to the reduction in blood lactate concentration and perception of respiratory sensations.[18] Thus, it also remains to be tested whether alternating RMST and RMET yield even greater improvements in performance than one type of RMT alone.

4.4 Effect of the Type of Exercise Test and Test Duration on the Improvement in Exercise Performance

The multiple regression analysis showed that improvements in exercise performance after RMT were significantly greater when tested with CLTs or IITs compared with ITs, with no difference between TTs and ITs. The meta-analysis of controlled studies revealed no significant effect of RMT when tested with ITs, while improvements in CLTs, TTs and in the IIT were all significant. The fact that RMT does not seem to affect IT performance is consistent with the notion that the duration a subject spends exercising above the threshold of 85% . VO2max, the exercise intensity where respiratory muscles are most likely to fatigue,[30] is too short to elicit respiratory muscle fatigue.[37] This is also supported by the finding that improvements in performance after RMT are greater with longer test duration (+0.4% per minute test duration, table I). Furthermore, of 22 studies assessing . VO2max before and after RMT, all but two studies found no change in . VO2max. Leddy et al.[85] observed a significant increase, while Verges et al.[96] observed a significant decrease in . VO2max. Interestingly, half of those tests that reported exercise intensity (n = 40) were performed below the threshold of 85% . VO2max, maximal workload (Wmax) or maximal velocity (vmax) – the average of the forty tests being 80% . VO2max, 81% Wmax or 98% vmax. All but two of the 20 tests that were performed below 85% showed an improvement in exercise performance after RMT. In contrast, only nine of the 20 tests that were performed above 85% showed increased performance. However, subjects performing above 85% . VO2max, Wmax or vmax were fitter than those performing below this threshold (fitness level of 3.1 and 2.1, respectively), which could partly explain this finding and illustrates the importance of using a model that accounts for confounders. Nevertheless, if respiratory muscles do not fatigue below the suggested threshold of 85% . VO2max, this would mean that a reduction in respiratory muscle fatigue could not be the only mechanism to increase endurance performance after RMT. It is known, for example, that in CLTs, psychological factors such as motivation or boredom may play an important role in determining the point of exhaustion.[106] Accordingly, after an extended period of RMT, motivation to withstand task failure in a CLT might be higher. However, in studies using CLTs at intensities below the threshold and including a sham-training group, improvements in the RMT groups exceeded those of the sham-training groups[66,71,79,85,88] with only one exception.[83] Thus, again, motivation cannot be the only reason for improvements in CLT performance after RMT. Another possible explanation for improved endurance performance after RMT is a reduced perception of respiratory exertion and/or breathlessness. Of the 15 studies testing with a CLT below the 85% threshold, only three specified changes in respiratory sensations, two[79,88] of them reported a significant decrease while one[83] did not. While the regression model, which accounted for confounders such as test duration, subjects’ fitness, type of training and type of sports did not find TTs to be more sensitive than ITs in showing improvements after RMT, the meta-analysis showed greater standardized mean differences in TTs and in the IIT compared with CLTs and ITs meaning that TTs would detect changes better than CLTs and ITs. This seems confusing at first; however, it should be considered that the model is based on all RMT studies, while the meta-analysis is based only on those studies that provided the necessary data. A comparison of average changes of the studies included in the regression model and in the meta-analysis shows that the average change in test duration in studies included in the regression model was significantly larger (+15%) than that of the studies included in the metaanalysis (+5%; p = 0.002). However, the average fitness of subjects included in the regression model tended to be lower (2.6) than fitness of subjects included in the meta-analysis (3.1; p = 0.065), which might possibly explain the greater average improvement in performance of subjects included in the regression model. Thus, it seems that the studies included in the meta-analysis – despite a similar number of positive outcomes – are not fully representative of all of the RMT studies included in the regression model. Also, the studies by Enright et al.[50,51] might contribute to the discrepancy between model and meta-analysis. These authors used a protocol resulting in much shorter test durations (4.4–4.5 minutes) than the suggested 8–12 minutes required for . VO2max determination.[107] Improvements after RMT[50,51] were even greater (approximately 25%) than those reported for physical endurance training (approximately 10%[108]), which raises questions regarding the validity of this protocol. Without the two studies by Enright et al.,[50,51] the overall difference between improvements in IT performance after the RMT and sham/no-training period is 2% – equal to that for TTs. Thus, the relatively small improvements generally seen in TTs, although consistent, might be too small to exceed the changes found in ITs. However, it must be noted that these small improvements in TT performance are highly relevant. For example, mean improvements in the 40 TTs would result in 40 m or five skiff lengths in a 2 km rowing regatta, 100 m in a 2 km running race, 1.2 m in a 200 m swimming competition and 1 km in a 30 km cycling race. The fact that improvements in exercise performance after RMT were significantly greater (19%) also in IITs compared with ITs, suggests that amateur and professional athletes performing intermittent sports, such as football, soccer, basketball, team handball, etc. might benefit from RMT similar to subjects performing endurance-type sports. This is further supported by one study that showed a reduction in recovery duration between sprints, which was in part attributed to a decreased perception of respiratory effort.[20]

4.5 Effects of RMT in Different Types of Sports

Although physiological evidence would suggest that RMT might be more effective in sports where respiratory muscles are subjected to increased respiratory work[44] (swimming) and/or increased non-respiratory work, i.e. postural[41-43] (running) or moving[38-40] (rowing) tasks, the model did not reveal any significant difference between improvements in the different types of sports. Thus, one could assume that respiratory muscles involved in additional tasks resulting in higher respiratory muscle work are sufficiently trained, such that the likelihood to fatigue is similar to that during, for example, cycling. Supporting this assumption is the following interesting observation: for the studies included in the present review that give respiratory muscle strength data and include subjects with a fitness level of 3 or 4, baseline values of MIP and especially MEP expressed as a percentage of predicted values[109] are lowest in cycling and increase with rowing, running and swimming (data not shown).

4.6 Limitations

Several variables of interest were not included in the analysis. For example, duration of the training period or training intensity might also influence changes in endurance performance although these variables were quite similar within RMST and within RMET studies. Therefore, only the type of training was included in the multiple linear regression model, while factors describing training regimens were omitted to prevent collinearity. Furthermore, exercise intensity is believed to play a crucial role with respect to the development of respiratory muscle fatigue and, therefore, with respect to a possible benefit from RMT. Intensity of the exercise tests was not, however, included in the model. Since too many studies did not provide detailed information on exercise intensity (n = 24), the inclusion of this variable would have led to the exclusion of too many studies from the regression model. Furthermore, since only exhaustive tests were included in the model, test duration and intensity would have led to collinearity, with the consequence of excluding one or the other variable from the model. The same holds true for ratings of perceived breathlessness or respiratory effort, which have been shown to be lower after RMT in some studies [16,20,39,53,55,59,60,63,67,72,73,76,79,88,96,110] but not in others.[38,56,58,60,83,97,111] As only 21 studies provided this information, the inclusion of these variables in the linear regression model would have led to the exclusion of too many studies. Therefore, the variables ‘intensity’ and ‘respiratory sensations’ were omitted, despite their potential to explain possible changes after RMT. These variables might, however, be included in the intercept. The significance of the intercept indicates that additional factors not included in the model play a role in determining improvements in exercise performance. A further consideration is that the funnel plot shows a potential publication bias. In general, without publication bias, studies would be evenly distributed around the mean, in the form of a triangle. Studies with small standard errors (often those with many subjects included) are found at the top of the triangle close to the mean. Studies with large standard errors (frequently smaller studies) are found at the bottom of the triangle, with some of them having a greater distance to the mean. In the present meta-analysis, studies at the bottom left of the triangle are missing. This could mean that smaller studies with negative outcome were not published in addition to the possibility that no such studies were ever conducted. If those small studies with a negative outcome were present in the funnel plot, this would mean that the effect of RMT would be smaller than shown in the present analysis.



This is the first study to systematically assess the effect of different types of RMT used to improve exercise performance in healthy subjects. It clearly shows that RMT significantly improves endurance performance, independent of the type of RMT or the type of sport. No difference was found between the effects of the two most commonly used respiratory muscle training modalities, RMST.IN and RMET, while RMST.INEX seemed to be superior. Less fit individuals benefit more from RMT than highly trained athletes, and improvements are greater with longer exercise durations even at intensities lower than the postulated threshold for development of respiratory muscle fatigue (85% . VO2max). This emphasizes the importance to report changes in respiratory sensations after RMT so that this variable can be included in future regression models as well. Furthermore, when assessing the effect of RMT, care must be taken regarding the choice of the test, since effects are not seen in ITs that are commonly used to assess . VO2max or anaerobic threshold. Also, more well controlled studies are needed to prove a superiority of RMST.INEX over the commonly used types of RMT to confirm the positive results observed in the few studies using IITs, and to investigate a possible additional benefit from alternating RMET and RMST.



The authors would like to thank Christoph Bra¨ndle and Dr. Alexander Akhmedov for translating the potentially relevant Japanese and Russian papers, respectively, Dr. Ruth Briggs for English editing, as well as the Swiss Federal Office of Sport (grant no. 11-11) and the Swiss National Science Foundation (grant no. 3200B0-116777) for providing financial support. The authors have no conflicts of interest to declare that are directly relevant to the content of this article.



1. Mador MJ, Acevedo FA. Effect of respiratory muscle fatigue on subsequent exercise performance. J Appl Physiol 1991 May; 70 (5): 2059-65
2. Verges S, Sager Y, Erni C, et al. Expiratory muscle fatigue impairs exercise performance. Eur J Appl Physiol 2007 Sep; 101 (2): 225-32
3. Dempsey JA, Romer L, Rodman J, et al. Consequences of exercise-induced respiratory muscle work. Respir Physiol Neurobiol 2006 Apr 28; 151 (2-3): 242-50
4. Hill JM. Discharge of group IV phrenic afferent fibers increases during diaphragmatic fatigue. Brain Res 2000 Feb 21; 856 (1-2): 240-4
5. Graham R, Jammes Y, Delpierre S, et al. The effects of ischemia, lactic acid and hypertonic sodium chloride on phrenic afferent discharge during spontaneous diaphragmatic contraction. Neurosci Lett 1986 Jun 30; 67 (3): 257-62
6. Jammes Y, Balzamo E. Changes in afferent and efferent phrenic activities with electrically induced diaphragmatic fatigue. J Appl Physiol 1992 Sep; 73 (3): 894-902
7. Derchak PA, Sheel AW, Morgan BJ, et al. Effects of expiratory muscle work on muscle sympathetic nerve activity. J Appl Physiol 2002 Apr; 92 (4): 1539-52
8. Rodman JR, Henderson KS, Smith CA, et al. Cardiovascular effects of the respiratory muscle metaboreflexes in dogs: rest and exercise. J Appl Physiol 2003 Sep; 95 (3): 1159-69
9. Sheel AW, Derchak PA, Morgan BJ, et al. Fatiguing inspiratory muscle work causes reflex reduction in resting leg blood flow in humans. J Physiol 2001 Nov 15; 537 (Pt 1): 277-89
10. St Croix CM, Morgan BJ, Wetter TJ, et al. Fatiguing inspiratory muscle work causes reflex sympathetic activation in humans. J Physiol 2000 Dec 1; 529 Pt 2: 493-504
11. Harms CA, Babcock MA, McClaran SR, et al. Respiratory muscle work compromises leg blood flow during maximal exercise. J Appl Physiol 1997 May; 82 (5): 1573-83
12. McConnell AK, Lomax M. The influence of inspiratory muscle work history and specific inspiratory muscle training upon human limb muscle fatigue. J Physiol 2006 Nov 15; 577 (Pt 1): 445-57
13. Romer LM, Lovering AT, Haverkamp HC, et al. Effect of inspiratory muscle work on peripheral fatigue of locomotor muscles in healthy humans. J Physiol 2006 Mar 1; 571 (Pt 2): 425-39
14. Harms CA, Wetter TJ, St Croix CM, et al. Effects of respiratory muscle work on exercise performance. J Appl Physiol 2000 Jul; 89 (1): 131-8
15. Babcock MA, Pegelow DF, Harms CA, et al. Effects of respiratory muscle unloading on exercise-induced diaphragm fatigue. J Appl Physiol 2002 Jul; 93 (1): 201-6
16. Verges S, Lenherr O, Haner AC, et al. Increased fatigue resistance of respiratory muscles during exercise after respiratory muscle endurance training. Am J Physiol Regul Integr Comp Physiol 2007 Mar; 292 (3): R1246-53
17. Romer LM, McConnell AK, Jones DA. Inspiratory muscle fatigue in trained cyclists: effects of inspiratory muscle training. Med Sci Sports Exerc 2002 May; 34 (5): 785-92
18. Verges S, Renggli AS, Notter DA, et al. Effects of different respiratory muscle training regimes on fatigue-related variables during volitional hyperpnoea. Respir Physiol Neurobiol 2009 Dec 31; 169 (3): 282-90
19. McConnell AK, Sharpe GR. The effect of inspiratory muscle training upon maximum lactate steady-state and blood lactate concentration. Eur J Appl Physiol 2005 Jun; 94 (3): 277-84
20. Romer LM, McConnell AK, Jones DA. Effects of inspiratory muscle training upon recovery time during high intensity, repetitive sprint activity. Int J Sports Med 2002 Jul; 23 (5): 353-60
21. Spengler CM, Roos M, Laube SM, et al. Decreased exercise blood lactate concentrations after respiratory endurance training in humans. Eur J Appl Physiol Occup Physiol 1999 Mar; 79 (4): 299-305
22. Witt JD, Guenette JA, Rupert JL, et al. Inspiratory muscle training attenuates the human respiratory muscle metaboreflex. J Physiol 2007 Nov 1; 584 (Pt 3): 1019-28
23. Clanton TL, Dixon GF, Drake J, et al. Effects of swim training on lung volumes and inspiratory muscle conditioning. J Appl Physiol 1987 Jan; 62 (1): 39-46
24. Martin BJ, Stager JM. Ventilatory endurance in athletes and non-athletes. Med Sci Sports Exerc 1981; 13 (1): 21-6
25. Robinson EP, Kjeldgaard JM. Improvement in ventilatory muscle function with running. J Appl Physiol 1982 Jun; 52 (6): 1400-6
26. Aaron EA, Seow KC, Johnson BD, et al. Oxygen cost of exercise hyperpnea: implications for performance. J Appl Physiol 1992 May; 72 (5): 1818-25
27. Choukroun ML, Kays C, Gioux M, et al. Respiratory muscle function in trained and untrained adolescents during short-term high intensity exercise. Eur J Appl Physiol Occup Physiol 1993; 67 (1): 14-9
28. Coast JR, Clifford PS, Henrich TW, et al. Maximal inspiratory pressure following maximal exercise in trained and untrained subjects. Med Sci Sports Exerc 1990 Dec; 22 (6): 811-5
29. Leith DE, Bradley M. Ventilatory muscle strength and endurance training. J Appl Physiol 1976 Oct; 41 (4): 508-16
30. Johnson BD, Babcock MA, Suman OE, et al. Exerciseinduced diaphragmatic fatigue in healthy humans. J Physiol 1993 Jan; 460: 385-405
31. Mador MJ, Magalang UJ, Rodis A, et al. Diaphragmatic fatigue after exercise in healthy human subjects. Am Rev Respir Dis 1993 Dec; 148 (6 Pt 1): 1571-5
32. Babcock MA, Johnson BD, Pegelow DF, et al. Hypoxic effects on exercise-induced diaphragmatic fatigue in normal healthy humans. J Appl Physiol 1995 Jan; 78 (1): 82-92
33. Babcock MA, Pegelow DF, McClaran SR, et al. Contribution of diaphragmatic power output to exercise-induced diaphragm fatigue. J Appl Physiol 1995 May; 78 (5): 1710-9
34. Taylor BJ, How SC, Romer LM. Exercise-induced abdominal muscle fatigue in healthy humans. J Appl Physiol 2006 May; 100 (5): 1554-62
35. Taylor BJ, Romer LM. Effect of expiratory muscle fatigue on exercise tolerance and locomotor muscle fatigue in healthy humans. J Appl Physiol 2008 May; 104 (5): 1442-51
36. Verges S, Schulz C, Perret C, et al. Impaired abdominal muscle contractility after high-intensity exhaustive exercise assessed by magnetic stimulation. Muscle Nerve 2006 Oct; 34 (4): 423-30
37. Romer LM, Miller JD, Haverkamp HC, et al. Inspiratory muscles do not limit maximal incremental exercise performance in healthy subjects. Respir Physiol Neurobiol 2007 Jun 15; 156 (3): 353-61
38. Riganas CS, Vrabas IS, Christoulas K, et al. Specific inspiratory muscle training does not improve performance or VO2max levels in well trained rowers. J Sports Med Phys Fitness 2008 Sep; 48 (3): 285-92
39. Volianitis S, McConnell AK, Koutedakis Y, et al. Inspiratory muscle training improves rowing performance. Med Sci Sports Exerc 2001 May; 33 (5): 803-9
40. Cunningham DA, Goode PB, Critz JB. Cardiorespiratory response to exercise on a rowing and bicycle ergometer. Med Sci Sports 1975 Spring; 7 (1): 37-43
41. Grillner S, Nilsson J, Thorstensson A. Intra-abdominal pressure changes during natural movements in man. Acta Physiol Scand. 1978 Jul; 103 (3): 275-83 42. Hodges PW, Butler JE, McKenzie DK, et al. Contraction of the human diaphragm during rapid postural adjustments. J Physiol 1997 Dec 1; 505 (Pt 2): 539-48
43. Hodges PW, Gandevia SC. Changes in intra-abdominal pressure during postural and respiratory activation of the human diaphragm. J Appl Physiol 2000 Sep; 89 (3): 967-76
44. Pendergast DR, Lundgren CE. The underwater environment: cardiopulmonary, thermal, and energetic demands. J Appl Physiol 2009 Jan; 106 (1): 276-83
45. Guenette JA, Sheel AW. Physiological consequences of a high work of breathing during heavy exercise in humans. J Sci Med Sport 2007 Dec; 10 (6): 341-50
46. Sonetti DA, Wetter TJ, Pegelow DF, et al. Effects of respiratory muscle training versus placebo on endurance exercise performance. Respir Physiol 2001 Sep; 127 (2-3): 185-99
47. Leveritt M, Abernethy PJ, Barry BK, et al. Concurrent strength and endurance training: a review. Sports Med 1999 Dec; 28 (6): 413-27 
48. Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health 1998 Jun; 52 (6): 377-84
49. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996 Feb; 17 (1): 1-12
50. Enright SJ, Unnithan VB. Effect of inspiratory muscle training intensities on pulmonary function and work capacity in people who are healthy: a randomized controlled trial. Phys Ther 2011 Jun; 91 (6): 894-905
51. Enright SJ, Unnithan VB, Heward C, et al. Effect of highintensity inspiratory muscle training on lung volumes, diaphragm thickness, and exercise capacity in subjects who are healthy. Phys Ther 2006 Mar; 86 (3): 345-54
52. Johnson MA, Sharpe GR, Brown PI. Inspiratory muscle training improves cycling time-trial performance and anaerobic work capacity but not critical power. Eur J Appl Physiol 2007 Dec; 101 (6): 761-70
53. Kilding AE, Brown S, McConnell AK. Inspiratory muscle training improves 100 and 200 m swimming performance. Eur J Appl Physiol 2010 Feb; 108 (3): 505-11 54. Kohl J, Koller EA, Brandenberger M, et al. Effect of exercise-induced hyperventilation on airway resistance and cycling endurance. Eur J Appl Physiol Occup Physiol 1997; 75 (4): 305-11
55. Lomax M, Grant I, Corbett J. Inspiratory muscle warm-up and inspiratory muscle training: separate and combined effects on intermittent running to exhaustion. J Sports Sci 2011 Mar; 29 (6): 563-9
56. Nicks CR, Morgan DW, Fuller DK, et al. The influence of respiratory muscle training upon intermittent exercise performance. Int J Sports Med 2009 Jan; 30 (1): 16-21
57. Tong TK, Fu FH, Chung PK, et al. The effect of inspiratory muscle training on high-intensity, intermittent running performance to exhaustion. Appl Physiol Nutr Metab 2008 Aug; 33 (4): 671-81
58. Wells GD, Plyley M, Thomas S, et al. Effects of concurrent inspiratory and expiratory muscle training on respiratory and exercise performance in competitive swimmers. Eur J Appl Physiol 2005 Aug; 94 (5-6): 527-40
59. Huang CH, Yang GG, Wu YT, et al. Comparison of inspiratory muscle strength training effects between older subjects with and without chronic obstructive pulmonary disease. J Formos Med Assoc 2011 Aug; 110 (8): 518-26
60. Griffiths LA, McConnell AK. The influence of inspiratory and expiratory muscle training upon rowing performance. Eur J Appl Physiol 2007 Mar; 99 (5): 457-66 61. Team RDC. R: A language and environment for statistical computing. 2011 [online]. Available from URL: http:// www.R-project.org/ [Accessed 2011 Nov 21] 62. Markov G, Spengler CM, Kno¨pfli-Lenzin C, et al. Respiratory muscle training increases cycling endurance without affecting cardiovascular responses to exercise. Eur J Appl Physiol 2001 Aug; 85 (3-4): 233-9
63. Romer LM, McConnell AK, Jones DA. Effects of inspiratory muscle training on time-trial performance in trained cyclists. J Sports Sci 2002 Jul; 20 (7): 547-62
64. Stuessi C, Spengler CM, Kno¨pfli-Lenzin C, et al. Respiratory muscle endurance training in humans increases cycling endurance without affecting blood gas concentrations. Eur J Appl Physiol 2001 Jun; 84 (6): 582-6
65. Lindholm P, Wylegala J, Pendergast DR, et al. Resistive respiratory muscle training improves and maintains endurance swimming performance in divers. Undersea Hyperb Med 2007 May-Jun; 34 (3): 169-80
66. Wylegala JA, Pendergast DR, Gosselin LE, et al. Respiratory muscle training improves swimming endurance in divers. Eur J Appl Physiol 2007 Mar; 99 (4): 393-404
67. Chatham K, Baldwin J, Griffiths H, et al. Inspiratory muscle training improves shuttle run performance in healthy subjects. Phys Ther 1999; 85 (12): 676-83
68. Esposito F, Limonta E, Alberti G, et al. Effect of respiratory muscle training on maximum aerobic power in normoxia and hypoxia. Respir Physiol Neurobiol 2010 Mar 31; 170 (3): 268-72
69. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002 Jun 15; 21 (11): 1539-58
70. Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003 Sep 6; 327 (7414): 557-60
71. Aznar-Lain S, Webster AL, Can˜ete S, et al. Effects of inspiratory muscle training on exercise capacity and spontaneous physical activity in elderly subjects: a randomized controlled pilot trial. Int J Sports Med 2007 Dec; 28 (12): 1025-9
72. Bailey SJ, Romer LM, Kelly J, et al. Inspiratory muscle training enhances pulmonary O(2) uptake kinetics and high-intensity exercise tolerance in humans. J Appl Physiol 2010 Aug; 109 (2): 457-68
73. Belman MJ, Gaesser GA. Ventilatory muscle training in the elderly. J Appl Physiol 1988 Mar; 64 (3): 899-905
74. Boutellier U, Bu¨chel R, Kundert A, et al. The respiratory system as an exercise limiting factor in normal trained subjects. Eur J Appl Physiol Occup Physiol 1992; 65 (4): 347-53
75. Boutellier U, Piwko P. The respiratory system as an exercise limiting factor in normal sedentary subjects. Eur J Appl Physiol Occup Physiol 1992; 64 (2): 145-52
76. Downey AE, Chenoweth LM, Townsend DK, et al. Effects of inspiratory muscle training on exercise responses in normoxia and hypoxia. Respir Physiol Neurobiol 2007 May 14; 156 (2): 137-46
77. Edwards AM, Cooke CB. Oxygen uptake kinetics and maximal aerobic power are unaffected by inspiratory muscle training in healthy subjects where time to exhaustion is extended. Eur J Appl Physiol 2004 Oct; 93 (1-2): 139-44
78. Fairbarn MS, Coutts KC, Pardy RL, et al. Improved respiratory muscle endurance of highly trained cyclists and the effects on maximal exercise performance. Int J Sports Med 1991 Feb; 12 (1): 66-70
79. Gething AD, Williams M, Davies B. Inspiratory resistive loading improves cycling capacity: a placebo controlled trial. Br J Sports Med 2004 Dec; 38 (6): 730-6
80. Guenette JA, Martens AM, Lee AL, et al. Variable effects of respiratory muscle training on cycle exercise performance in men and women. Appl Physiol Nutr Metab 2006 Apr; 31 (2): 159-66
81. Hanel B, Secher NH. Maximal oxygen uptake and work capacity after inspiratory muscle training: a controlled study. J Sports Sci 1991 Spring; 9 (1): 43-52
82. Hart N, Sylvester K, Ward S, et al. Evaluation of an inspiratory muscle trainer in healthy humans. Respir Med 2001 Jun; 95 (6): 526-31
83. Holm P, Sattler A, Fregosi RF. Endurance training of respiratory muscles improves cycling performance in fit young cyclists. BMC Physiol 2004 May 6; 4: 9 84. Kwok TMK, Jones AYM. Target-flow inspiratory muscle training improves running performance in recreational runners: a randomized controlled trial. Hong Kong Physiother J 2009; 27: 48-54
85. Leddy JJ, Limprasertkul A, Patel S, et al. Isocapnic hyperpnea training improves performance in competitive male runners. Eur J Appl Physiol 2007 Apr; 99 (6): 665-76
86. Markov G, Orler R, Boutellier U. Respiratory training, hypoxic ventilatory response and acute mountain sickness. Respir Physiol 1996 Sep; 105 (3): 179-86 87. McMahon ME, Boutellier U, Smith RM, et al. Hyperpnea training attenuates peripheral chemosensitivity and improves cycling endurance. J Exp Biol 2002 Dec; 205 (Pt 24): 3937-43
88. Mickleborough TD, Nichols T, Lindley MR, et al. Inspiratory flow resistive loading improves respiratory muscle function and endurance capacity in recreational runners. Scand J Med Sci Sports 2010 Jun; 20 (3): 458-68
89. Morgan DW, Kohrt WM, Bates BJ, et al. Effects of respiratory muscle endurance training on ventilatory and endurance performance of moderately trained cyclists. Int J Sports Med 1987 Apr; 8 (2): 88-93
90. Mucci P, Lesaignoux Y. Inspiratory-muscle training and critical velocity [in French]. Sci Sports 2008; 23: 255-7
91. Ray AD, Pendergast DR, Lundgren CEG. Respiratory muscle training improves swimming endurance at depth. Undersea Hyperb Med 2008 May-Jun; 35 (3): 185-96
92. Ray AD, Pendergast DR, Lundgren CEG. Respiratory muscle training reduces the work of breathing at depth. Eur J Appl Physiol 2010 Mar; 108 (4): 811-20
93. Spengler CM, Laube SM, Roos M, et al. The effect of breathing pattern during respiratory training on cycling endurance. In: Steinacker JM, Ward SA, editors. The physiology and pathophysiology of exercise tolerance. New York: Plenum Press; 1996; 315-9
94. Sperlich B, Fricke H, de Marees M, et al. Does respiratory muscle training increase physical performance? Mil Med 2009 Sep; 174 (9): 977-82
95. Swanson GD. Pulmonary training may alter exertional dyspnea and fatigue via an exercise-like training effect of a lowered heart rate. Adv Exp Med Biol 1998; 450: 231-6
96. Verges S, Kruttli U, Stahl B, et al. Respiratory control, respiratory sensations and cycling endurance after respiratory muscle endurance training. Adv Exp Med Biol 2008; 605: 239-44
97. Williams JS, Wongsathikun J, Boon SM, et al. Inspiratory muscle training fails to improve endurance capacity in athletes. Med Sci Sports Exerc 2002 Jul; 34 (7): 1194-8
98. Pollock ML. The quantification of endurance training programs. Exerc Sport Sci Rev 1973; 1: 155-88
99. Saltin B, Hartley LH, Kilbom A, et al. Physical training in sedentary middle-aged and older men. II: oxygen uptake, heart rate, and blood lactate concentration at submaximal and maximal exercise. Scand J Clin Lab Invest 1969 Dec; 24 (4): 323-34
100. Sharkey BJ. Intensity and duration of training and the development of cardiorespiratory endurance. Med Sci Sports 1970 Winter; 2 (4): 197-202
101. Wenger HA, Bell GJ. The interactions of intensity, frequency and duration of exercise training in altering cardiorespiratory fitness. Sports Med 1986 Sep-Oct; 3 (5): 346-56
102. Wilson TM, Tanaka H. Meta-analysis of the ageassociated decline in maximal aerobic capacity in men: relation to training status. Am J Physiol Heart Circ Physiol 2000 Mar; 278 (3): H829-34
103. Aliverti A, Cala SJ, Duranti R, et al. Human respiratory muscle actions and control during exercise. J Appl Physiol 1997 Oct; 83 (4): 1256-69
104. Illi SK, Hostettler S, Mohler E, et al. Compartmental chest wall volume changes during volitional normocapnic hyperpnoea. Respir Physiol Neurobiol 2011 Aug 15; 177 (3): 294-300
105. McCool FD, Hershenson MB, Tzelepis GE, et al. Effect of fatigue on maximal inspiratory pressure-flow capacity. J Appl Physiol 1992 Jul; 73 (1): 36-43 106. Jeukendrup A, Saris WHM, Brouns F, et al. A new validated endurance performance test. Med Sci Sports Exerc 1996 Feb; 28 (2): 266-70
107. Buchfuhrer MJ, Hansen JE, Robinson TE, et al. Optimizing the exercise protocol for cardiopulmonary assessment. J Appl Physiol 1983 Nov; 55 (5): 1558-64
108. Jones AM, Carter H. The effect of endurance training on parameters of aerobic fitness. Sports Med 2000 Jun; 29 (6): 373-86
109. Wilson SH, Cooke NT, Edwards RH, et al. Predicted normal values for maximal respiratory pressures in caucasian adults and children. Thorax 1984 Jul; 39 (7): 535-8
110. Suzuki S, Sato M, Okubo T. Expiratory muscle training and sensation of respiratory effort during exercise in normal subjects. Thorax 1995 Apr; 50 (4): 366-70
111. Suzuki S, Yoshiike Y, Suzuki M, et al. Inspiratory muscle training and respiratory sensation during treadmill exercise. Chest 1993 Jul; 104 (1): 197-202

Correspondence: Prof. Dr. Christina M. Spengler, Exercise Physiology, Institute of Human Movement Sciences, University and ETH Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland. E-mail: christina.spengler@hest.ethz.ch