Lung function was assessed with a portable spirometer by Jaeger (VIASYS® Healthcare GmbBH, Höchberg, Germany). The tests were simple spirometry (volume-time) and forced spirometry (flow-volume curve), which provided data for the following variables: vital capacity (VC), forced vital capacity (FVC), forced expiratory volume in one second (FEV1), peak inspiratory flow (PIF), and peak expiratory flow (PEF). For both tests, we followed the standards required by the American Thoracic Society and the European Respiratory Society (ATS/ERS).17 In the case of forced spirometry, this was repeated until it obtained a minimum of 3 technically satisfactory maneuvers, with a minute of rest in between, and the highest and most reproducible value was recorded, as long as the difference of the 2 highest values, FEV1 and FVC, were not >0.150 l (ATS/ERS, 2005: p. 325).17

PImax and PEmax were measured with an aneroid manometer (Series 2000 Magnehelic® Pressure Gauge, Dwyer Instruments, Michigan City, Indiana, United States) with a sensitivity range between 0 and 300cm H2O of pressure. The measurements were repeated until 3 readings were obtained with a variance ≤10%,18,19 and the highest value was used. The reference values used to obtain the percentages were those chosen by Enright et al.,19 taking into account the following circumstances: (a) the values for normalcy included the population ≥85, with a sample size higher than that of other studies (n=193), and excluded the institutionalized population and/or those who were unable to walk; and (b) the present study reproduces the methodology developed by these authors, following the indications of Black and Hyatt.20

MVV (maximum volume of air that a person breathes voluntarily during an interval of 12s) was measured a minimum of 2 times with a variation ≤20%.17,21 MVV diminished progressively with weak and fatigable RM. In accordance with the ATS/ERS, MVV is an index that is representative of overall RM resistance: “The most important advantage to measuring MVV as an indicator of respiratory muscle endurance is its close resemblance to the task performed during exercise.” (ATS/ERS, 2002, p. 563).18 The reference values used to obtain the percentages were those published by Neder et al.21 All spirometry measurements, as well as PImax, PEmax and MVV, were done by specialized personnel, and with the participants sitting and wearing nasal clips.17,18

The research team and several experts (physicians specialized in geriatrics and pulmonology, physiotherapists and psychologists) elaborated a questionnaire (Appendix A) that allowed us to know: (a) the presence of dyspnea and functional limitation for ADL; (b) the experience during the training sessions; and (c) observations about the training protocol. The members of the research team that carried out the survey were not involved in the development of the training protocol. During the initial assessment, both groups responded to questions 1–4. Later, in the post-training assessment (week 7), only the TG answered questions 5–15. It is an ad hoc questionnaire created for the purpose of this study alone, with descriptive purpose and content validity, because the Spanish versions of the validated scales did not adjust to the characteristics of the population studied.

Table 1 shows the baseline characteristics and the homogeneity tests between groups. As for the anthropometric data, the results only show that the groups differ in height.

No statistically significant differences were observed in the variables regarding inclusion criteria: the Lobo Mini Cognitive Test (23.0±3.5) and Barthel index (68.1±22.7). On an average, the participants had been institutionalized 4.1±5.2 years and only 12.5% were ex-smokers (CG, n=1; TG, n=4; P=.120).

As for questions 1–4 (Appendix A, block A) of the ad hoc questionnaire, there were no significant differences between the 2 groups (P>.05). In 37.5% of the cases, breathing affected in some measure (“a little” or “a lot”) the normal development of activities of daily living (ADL). Among these, walking and/or transference were mostly associated with the presence of dyspnea (47.5%), followed by self-care activities (27.5%). The main causes that impede walking were: amputation of a lower limb or limbs, the effects of a cerebrovascular accident, osteoarthritis and a history of fractures caused by falls.

The descriptive analysis corresponding with questions 5–9 (Appendix A, block B) of the ad hoc questionnaire for the TG shows that close to 42.7% of the participants reported difficulty in breathing during the session, especially for the first few days (47.4%) and when the workload was increased (15.8%). They also considered the 1-min rest intervals (84.2%) necessary. After the conclusion of the training protocol (questions 10–15, Appendix A, block C), 63% of the participants perceived an improvement over their previous state, and 57.9% considered it necessary to make some type of change in the protocol (for example, 47.4% period of one year; 21.1%, weekly sessions; 22%, total number of weeks). As for the number of colds, there were no significant differences (CG 33.3% and TG 52.6%; P=.218). 84.2% of the people with training answered that they would participate again, while 5.3% did not know what to answer. In addition, they scored an average perceived satisfaction of 8.4±2.1 on a scale from 1 (“not at all satisfied”) to 10 (“very satisfied”). Finally, the subjective perception of exertion (Borg CR10 scale)23 for the TG showed no significant differences between the weekly averages (Table 2), and a score of 3 was “moderate” exertion.

The estimated ANOVA evaluating the effect of the intervention on the absolute PImax values did not find statistically significant differences among the groups (F1,38=0.22, P=.637; R2=0.006). Time did show differences and in both groups there was an increase over time (F3,114=12.51, P<.001, R2=0.248). However, the interaction effect, which shows if the treatment is effective, was not statistically significant (F3,114=1.04, P=.368; R2=0.027) (Fig. 2). Although the interaction was not statistically significant, simple effects analyses of the interaction were used to better understand within each group what moments in time differed. The CG showed differences between the baseline and the final measurement (time period 3 or post-training, P<.05). Contrarily, the TG presented differences at all time periods with baseline (P<.05). There were no differences between time periods 3 and 4 (follow-up) (P>.05). This result indicates that the change over time in the TG was more stable, as CG is lower in the follow-up, which does not happen in the TG.

Fig. 2.

Mean maximum inspiratory pressure (PImax, cm H2O) for the control and training groups over the course of the 4 time periods measured: baseline, intermediate, post-training and follow-up.

(0.09MB).

The estimated ANOVA assessing the effect of the intervention on the absolute PEmax values also found no significant differences between the groups (F1,38=2.57, P=.117; R2=0.063). Time did show significant differences and in both groups there was an increase over time (F3,114=15.55, P<.001, R2=0.29). However, the interaction effect was not statistically significant in the case of PEmax (F3,114=1.86, P=.14; R2=0.047) (Fig. 3). Although the interaction was not statistically significant, simple effects were done as in the former variable. In the case of CG, there was only a gain between the baseline and the post-training and, however the follow-up did not differ significantly from baseline (P>.05). In the case of TG, baseline differed from all the other time periods (P>.05), which indicated a stable change, even in follow-up.

Fig. 3.

Mean of maximum expiratory pressure (PEmax, cm H2O) for the control and training groups over the course of the four time periods measured: baseline, intermediate, post-training and follow-up.

(0.08MB).

A third ANOVA evaluated the effect of the intervention on the absolute MVV values and, once again, did not show significant differences between the groups (F1,38=0.47, P=.49; R2=0.012). There was a significant increase in MVV over time (F3,114=3.39, P=.02; R2=0.082), despite the fact that the interaction was not significant (F3,114=1.74, P=.162; R2=0.044) (Fig. 4). Once again, and although the interaction was not statistically significant, simple effects were done. In the case of CG, there were no differences between any of the time periods (P>.05), which indicate stability in the resistance of RM. In the case of TG, the baseline differed only from the follow-up (P<.05), which indicates that, if there were potential effects of training with Threshold® IT on muscle resistance, these tended to appear at later time periods.

Fig. 4.

Means of maximum voluntary ventilation (MVV, l/min) for the control and training groups over the course of the four time periods measured: baseline, intermediate, post-training and follow-up.

(0.08MB).

The workload that the TG group was subjected to significantly improved both absolute values (F5,100=72.031, P<.001; R2=0.791) as well as percentage of baseline PImax (F5,100=67.225, P<.001; R2=0.810). In absolute values, the load increased from −9.7 to −20.7cm H2O over the course of the 6-week training period. This increase was considerably less during the first half of the RM training (difference of means between the first and third weeks, 3.39cm H2O) than during the second half (difference of means between the fourth and sixth weeks, 5.84cm H2O), a result that may be due to the period of familiarization of the first week. Likewise, the relative workload rose from 36% to 78% PImax measured at baseline. The difference of percentages for the first half of the training period was 11%, while for the second half it was 23.3%, an aspect possibly related with familiarization, as previously commented. Last of all, the subjective perception of exertion (Borg scale) did not change (P>.05) during the training period (Table 2).