A multi-trial, retrospective analysis of the antihypertensive effects of high-resistance, low-volume inspiratory muscle strength training

Abstract

Above-normal blood pressure (BP) is a primary risk factor for cardiovascular diseases. In a retrospective analysis of five pilot trials, we assessed the BP-lowering effects of high-resistance inspiratory muscle strength training (IMST) in adults aged 18–82 years and the impact of IMST on maximal inspiratory pressure (PI_MAX), a gauge of inspiratory muscle strength and independent disease risk factor. Participants were randomized to high-resistance IMST (75% PI_MAX) or low-resistance sham (15% PI_MAX) training (30 breaths/day, 5–7 days/wk, 6 wk). IMST (n = 67) reduced systolic BP (SBP) by 9 ± 6 mmHg (P < 0.01) and diastolic BP (DBP) by 4 ± 4 mmHg (P < 0.01). IMST-related reductions in SBP and DBP emerged by week 2 of training (−4 ± 8 mmHg and −3 ± 6 mmHg; P ≤ 0.01, respectively) and continued across the 6-wk intervention. SBP and DBP were unchanged with sham training (n = 61, all P > 0.05). Select subject characteristics slightly modified the impact of IMST on BP. Greater reductions in SBP were associated with older age (β = −0.07 ± 0.03; P = 0.04) and greater reductions in DBP associated with medication-naïve BP (β = −3 ± 1; P = 0.02) and higher initial DBP (β = −0.12 ± 0.05; P = 0.04). PI_MAX increased with high-resistance IMST and low-resistance sham training, with a greater increase from high-resistance IMST (+20 ± 17 vs. +6 ± 14 cmH₂O; P < 0.01). Gains in PI_MAX had a modest inverse relation with age (β = −0.20 ± 0.09; P = 0.03) and baseline PI_MAX (β = −0.15 ± 0.07; P = 0.04) but not to reductions in SBP or DBP. These compiled findings from multiple independent trials provide the strongest evidence to date that high-resistance IMST evokes clinically significant reductions in SBP and DBP, and increases in PI_MAX, in adult men and women.

NEW & NOTEWORTHY In young-to-older adult men and women, 6 wk of high-resistance inspiratory muscle strength training lowers casual systolic and diastolic blood pressure by 9 mmHg and 4 mmHg, respectively, with initial reductions observed by week 2 of training. Given blood pressure outcomes with the intervention were only slightly altered by subject baseline characteristics (i.e., age, blood pressure medication, and health status), inspiratory muscle strength training is effective in lowering blood pressure in a broad range of adults.

INTRODUCTION

Cardiovascular diseases (CVD) are the leading cause of death in developed and developing countries. Above-normal blood pressure (BP), defined as having a systolic BP (SBP) ≥ 120 mmHg and/or a diastolic BP (DBP) ≥ 80 mmHg, is the primary modifiable risk factor for CVD (1). Above-normal BP is a highly prevalent health condition that affects more than 50% of adults in the United States (2). The incidence of above-normal BP increases with age, such that 90% of adults who live to 80 years of age will develop unhealthy BP levels, even if they have normal BP at midlife (3). The number of older adults is projected to rapidly rise, predicting a dramatic increase in BP-driven CVD burden (4). Thus, addressing the burden of above-normal BP is an important public health goal.

Interventions that can lower BP to recommended levels (i.e., SBP < 120 mmHg, DBP < 80 mmHg) reduce the risk for both CVD and all-cause mortality by ∼25% (5), and therefore, have a large impact on public health. Lifestyle interventions, such as moderate-intensity aerobic exercise, weight loss, or eating a healthy diet (including restricting sodium intake), are first-line strategies to control BP (6). However, adherence to these interventions among adults in the United States is low, with <40% of adults meeting recommendations for each individual intervention (7–9). Among the most commonly cited barriers to adherence are time availability (10–12), cost, facility access, and transportation issues (12–15).

Originally conceived as a form of physical training intended to improve diaphragm strength and function in ventilator-dependent individuals (16, 17), inspiratory muscle strength training (IMST) entailed repeated inspiratory efforts against a resistance as tolerated for weeks to months. Over the years, a range of IMST protocols, most involving 30 min of training per day, 2–3 times per week (18–21), have been implemented in nonventilator-dependent patient populations (22, 23), healthy adults (24, 25), and athletes (26–28). The reported benefits of IMST in these populations include improved inspiratory muscle strength and athletic performance, as well as reduced perception of effort and dyspnea (22–28).

In its most recent iteration, high-resistance low-volume IMST comprises 30 daily inspiratory efforts against a resistance equivalent to 75% of an individual’s maximal inspiratory pressure (PI_MAX) and requires ∼5 min to complete (29, 30). This IMST format, implemented in a number of small pilot trials, has yielded significant reductions in casual BP (31–35), while also decreasing sympathoadrenal activity and systemic vascular resistance (33–35), and improving vascular endothelial function (31). High user adherence (>90% of prescribed training sessions completed) has been reported in all trials to date (31–35). Thus, current evidence suggests that high-resistance IMST has potential for public health translation (36).

Despite favorable health and adherence outcomes, the small sample sizes arising from individual pilot trials performed to date both limit confidence in prior results and preclude analysis of subject characteristics that may impact effectiveness of the intervention for lowering BP. Accordingly, to address these specific research gaps, we pooled data from five randomized, controlled trials performed in our laboratories at the University of Arizona and the University of Colorado Boulder that used near-identical protocols but with unique participant populations (i.e., normotensive young adults, middle-aged and older adults with above-normal BP, and adults with obstructive sleep apnea and above normal BP). The principal purposes of this investigation were to quantify the BP-lowering effects of IMST in a larger cohort of participants, evaluate the time course of BP reductions, and identify subject characteristics that may predict the magnitude of the BP reduction in response to the intervention. We also evaluated BP outcomes as a function of gains in respiratory muscle strength (PI_MAX). We hypothesized that 1) IMST would reduce SBP and DBP compared with sham training; 2) reductions in BP would be apparent within the first week of training; and 3) higher initial BP levels would be associated with larger BP reductions with IMST, but other clinical factors would not appreciably predict the efficacy of IMST.

METHODS

The Institutional Review Boards at The University of Arizona and the University of Colorado Boulder approved all experimental procedures. Written informed consent was obtained for all participants. Individual participant data from five studies conducted over a 7-year period (2015 through 2021) were included in the analysis. Subjects in the trials comprised young adults (32, 34), patients with obstructive sleep apnea (33, 35), and otherwise healthy adults with above-normal SBP (31). Based on all participants having initial BP levels ≤ 160/100 mmHg and the presence of other CVD risk factors (e.g., obesity) in some participants, a majority of the participants had low-to-moderate CVD risk (37).

Experimental Measurements

Casual (resting) SBP and DBP were measured in accordance with clinical guidelines (6, 37, 38) using either a manual sphygmomanometer and stethoscope (31–34) or automated oscillometric sphygmomanometer (35) placed over the brachial artery. Subjects were seated quietly for at least 5 min before BP measurements with their feet flat on the floor and arm supported at heart level. Measurements were taken in triplicate and averaged, with >1 min between measurements. BP measurements were made at baseline, weekly during the intervention, and at the end of the intervention.

PI_MAX was measured at baseline, weekly during the interventions, and at end of the intervention to determine each individual’s appropriate training level and document changes in inspiratory muscle function. Subjects generated a maximal inspiration against near-infinite resistance. Per American Thoracic Society/European Respiratory Society guidelines, the average of three trials was used to define an individual’s PI_MAX (39). Four of the studies included in this data set measured PI_MAX using a customized pressure transducer (Omega-dyne, Inc.) (31–34) and one study measured PI_MAX using a POWERbreathe K3 device in TEST mode (35).

Inspiratory Muscle Strength Training Protocol

Detailed descriptions of the interventions have been reported (31–35). In brief, volunteers in all studies were randomized to perform either high-resistance IMST (i.e., training against an inspiratory resistance of 75% PI_MAX) or low-resistance sham IMST (i.e., training at 15% PI_MAX, believed to have no therapeutic effect). In subsequent studies, we have shown low resistance IMST does contribute to improvements in PI_MAX in some individuals and therefore, technically cannot be considered a “sham” intervention. However, for the sake of consistency with our previous work we will continue to refer to low-resistance IMST as sham IMST. All included studies followed a 6-wk intervention wherein participants performed 30 inspiratory efforts/day, performed in a single bout and comprising five sets of six inspiratory efforts, 5–7 days/wk (Table 1). For each inspiratory effort, resistance was controlled but tidal volume and inspiratory duration were not. All participants were instructed to perform as deep of an inspiration as possible on each effort with a target inspiration duration of 1–2 s. There were some between-study differences in the training devices and training load progression (Table 1). In all studies, absolute training intensity was adjusted weekly to account for gains in PI_MAX.

Table 1.

Inspiratory muscle strength training trial intervention designs

Author	Training Device	Device Resistance Type	Training Protocol	High Training Load (% PIMAX)	Subject Number (n)	Subject Population
Vranish et al. (32)	Hans Rudolph 2600 series two-way nonrebreathing valve with flow limitation cap	Constant	30 breaths/day; 5 days/wk; 6 wk	75%	IMST: 10 Sham: 10	Healthy young adults
Vranish et al. (33)	POWERbreathe K3	Tapered loading	30 breaths/day; 7 days/wk; 6 wk	75%	IMST: 12 Sham: 10	Adults with obstructive sleep apnea
DeLucia et al. (34)	Hans Rudolph 2600 series two-way nonrebreathing valve with flow limitation cap	Constant	30 breaths/day; 5 days/wk; 6 wk	75%	IMST: 12 Sham: 13	Healthy young adults
Ramos-Barrera et al. (35)	POWERbreathe K3	Tapered loading	30 breaths/day; 7 days/wk; 6 wk	75%	IMST: 15 Sham: 10	Adults with obstructive sleep apnea
Craighead et al. (31)	POWERbreathe K3	Tapered loading	30 breaths/day; 6 days/wk; 6 wk	55% week 1; 65% week 2; 75% weeks 3–6	IMST: 18 Sham: 18	Midlife and older adults with SBP ≥120 mmHg

IMST, inspiratory muscle strength training; PI_MAX, maximal inspiratory pressure; SBP, systolic blood pressure.

Data and Statistical Analysis

Casual SBP, casual DBP, and PI_MAX values from baseline, weekly during the intervention, and at the end of the 6-wk intervention were included for analysis. One study (32) reported SBP and DBP after 5 wk of training. In this instance, SBP and DBP values were imputed with an intention-to-treat approach, wherein the week 5 BP values were carried forward and imputed as the week 6 BP values. A second study (35) reported SBP and DBP for a subset of participants (n = 15) in whom the primary outcomes of muscle sympathetic nerve activity and 24-h BP were obtained; however, the current analysis includes values for SBP, DBP, and PI_MAX from (n = 25) participants who completed that trial’s 6-wk intervention.

Study-level data for SBP, DBP, and PI_MAX were compared using forest plots created with a random-effects model. Standardized mean differences (SMD) with 95% confidence intervals between IMST and sham groups were calculated based on group mean change (end-intervention value – baseline value) as described previously (40). SMD, also known as effect sizes, were defined as very small/no effect (<0.20), small (0.20–0.49), moderate (0.50–0.79), or large (≥0.80) (41). Heterogeneity (I²), study weight, and prediction intervals were calculated as described previously (40). The I² statistic is a measure of relative variance between the studies, with a low variance (<50%) suggesting between-study differences are largely due to random error, while moderate or high variance (>50% and >75%, respectively) suggests a substantial portion of the observed variance is due to real differences between studies that could potentially be explained by covariates (40). Study weight was calculated according to each individual study’s contribution to the pooled estimate (i.e., inverse of the variance of the treatment effect, largely impacted by sample size) (42). Prediction intervals indicate that with 95% confidence the true effect (i.e., SMD) of a future study will fall within the specified range (40).

Participant-level data from the five studies were combined and statistical comparisons made using SPSS version 27 and GraphPad Prism version 9.4. A two-way (Training Group × Time) repeated-measures ANOVA was used to determine significant group and time main effects or interaction effects of SBP, DBP, and PI_MAX at baseline and at end of the intervention. When a significant effect was observed, Sidak’s multiple comparison post hoc test was used to assess within- and/or between-group statistical differences. Unpaired t tests also were used to compare the difference in the change in SBP, DBP, and PI_MAX between the IMST and sham groups.

A one-way repeated-measures ANOVA with Dunnett’s multiple comparisons test was used to assess the onset for changes in SBP, DBP, and PI_MAX compared with baseline in response to the intervention in the high-resistance IMST group only. Simple linear regression of the week-to-week change in BP or PI_MAX was used to assess trends over time.

We then assessed the extent to which baseline subject characteristics influence the responsiveness (e.g., magnitude of change in BP) to high-resistance IMST. Simple linear regression was used to examine the independent effects of age and baseline measures of SBP, DBP, PI_MAX, and body mass index (BMI) to predict changes in SBP, DBP, and PI_MAX during 6 wk of high-resistance IMST. Simple linear regression also was used to evaluate the association between change in PI_MAX and SBP and between PI_MAX and DBP. Unpaired t tests were used to compare the changes in SBP, DBP, and PI_MAX between men and women and between BP-medicated and medication-naive participants. A sensitivity analysis was also performed to assess the impact of BP medication when young adults (aged ≤30 yr) were removed from the analysis to more closely match BP-medicated and medication-naïve participants. Stepwise linear regression including all the aforementioned variables was then used to identify variables that independently influence responsiveness to high-resistance IMST.

Statistical significance was set a priori as α = 0.05 for all comparisons, while P values between 0.05 and 0.10 were noted as trends. Values are presented as means ± standard deviation (SD) unless otherwise specified.

RESULTS

Participants

Data from a total of 128 adults were included for retrospective analysis (Table 2). Sixty-seven subjects (29 women/38 men) had been randomized to receive the high-resistance IMST intervention, whereas 61 (24 women/37 men) were randomized to receive sham training. There were no differences in subject characteristics between groups at baseline (all P > 0.05).

Table 2.

Subject characteristics under baseline conditions

	Sham	IMST	P Values
n (M/F)	61 (37/24)	67 (38/29)
Age, yr	50 ± 25	50 ± 22	0.94
Height, cm	170 ± 9	172 ± 10	0.20
Mass, kg	76.3 ± 16.9	78.7 ± 17.3	0.44
BMI, kg/m2	26.5 ± 5.5	26.5 ± 4.9	0.93
PIMAX, cmH2O	82 ± 22	87 ± 28	0.30
SBP, mmHg	124 ± 14	127 ± 15	0.32
DBP, mmHg	75 ± 9	76 ± 10	0.49
Training session (n)	35 ± 6	37 ± 6	0.30

Data are means ± SD. BMI, body mass index; DBP, diastolic blood pressure; PI_MAX, maximal inspiratory pressure; SBP, systolic blood pressure.

Systolic Blood Pressure

All studies showed a large effect for high-resistance IMST on SBP relative to sham training, with a combined SMD of 1.82 (Fig. 1A). Heterogeneity of the SBP response between studies was moderate (I² = 71.3%).

Figure 1.

Systolic blood pressure (SBP) data in the sham (n = 61; 37 men, 24 women) and inspiratory muscle strength training (IMST; n = 67; 38 men, 29 women) groups. A: forest plot showing standardized mean difference (SMD) between IMST and sham groups. Boxes and whiskers indicate SMD ± 95% confidence intervals. The diamond indicates the combined SMD ± 95% confidence interval, whereas the green line indicates the 95% prediction interval. Positive values indicate favoring IMST and negative values indicate favoring Sham. B: two-way (Training group × Time) repeated-measures ANOVA revealed a group × time interaction (P < 0.01), with SBP in the IMST group significantly lower from baseline and end-intervention (P < 0.01) and lower than sham at end-intervention (P = 0.03). C: independent t tests revealed a significant between-group difference for the change in SBP over the 6-wk intervention (P < 0.01). D: one-way repeated-measures ANOVA with Dunnett’s multiple comparisons test revealed a significant reduction in SBP compared with baseline after 2 wk of training (all P < 0.05). Data (B and D) are means ± SD. *P < 0.05 vs. pre.

When individual participant data from the five studies were collapsed into IMST or sham groups, there was a significant group by time interaction for SBP (P < 0.01). Casual SBP was unchanged in the sham group (baseline: 124 ± 15 mmHg, end-intervention: 124 ± 14; P = 0.98), but in the high-resistance IMST group, SBP decreased from 127 ± 15 mmHg at baseline to 118 ± 15 mmHg at the end of the intervention (P < 0.01), such that end-intervention SBP was significantly lower in the IMST group compared with sham (P = 0.03) (Fig. 1B). Accordingly, the mean reduction in SBP following IMST (−9 ± 6 mmHg) was greater in magnitude than the mean change with sham training (0 ± 6 mmHg; P < 0.01, Fig. 1C).

Casual SBP declined by 1.5 ± 7.9 mmHg/wk on average during the 6-wk high-resistance IMST intervention, such that casual SBP was lower than baseline from the end of week 2 of training each week until end of the intervention (all P < 0.05). The week-to-week rate of reduction in SBP did not change across the 6-wk intervention [β = 0.17 ± 0.23 (standard error of the mean; SEM), P = 0.46].

Diastolic Blood Pressure

Study-level data for all five trials showed large effects of high-resistance IMST on casual DBP relative to sham training (Fig. 2A). In addition, heterogeneity between studies was low (I² = 6.8%).

Figure 2.

Diastolic blood pressure (DBP) data in the sham (n = 61, 37 men, 24 women) and inspiratory muscle strength training (IMST; n = 67; 38 men, 29 women) groups. A: forest plot showing standardized mean difference (SMD) between IMST and sham groups. Boxes and whiskers indicate SMD ± 95% confidence intervals. The diamond indicates the combined SMD ± 95% confidence interval, whereas the green line indicates the 95% prediction interval. Positive values indicate favoring IMST and negative values indicate favoring Sham. B: two-way (Training group × Time) repeated-measures ANOVA revealed a group × time interaction (P < 0.01), with DBP in the IMST group significantly lower from baseline and end-intervention (P < 0.01) and lower than sham at end-intervention (P = 0.03). C: independent t tests revealed a significant between-group difference for the change in DBP over the 6-wk intervention (P < 0.01). D: one-way repeated-measures ANOVA with Dunnett’s multiple comparisons test revealed a significant reduction in DBP compared with baseline after 1 wk of training (all P < 0.05). Data (B and D) are means ± SD. *P < 0.05 vs. pre.

For participant-level data, there was a significant group by time interaction for DBP (P < 0.01). Casual DBP was unchanged in the sham group (baseline: 74 ± 9 mmHg, end-intervention: 76 ± 9 mmHg; P = 0.10), but in the high-resistance IMST group, DBP decreased from 76 ± 10 mmHg at baseline to 72 ± 10 mmHg at the end of the intervention (P < 0.01), such that end-intervention DBP was significantly lower in the IMST group compared with sham (P = 0.03) (Fig. 2B). The mean reduction in DBP following IMST (−4 ± 4 mmHg) also was significantly greater than the mean change with sham training (1 ± 5 mmHg; P < 0.01, Fig. 2C).

Casual DBP was reduced by 0.6 ± 5.8 mmHg/wk on average during the high-resistance IMST intervention, such that casual DBP was significantly lower than baseline from the end of week 1 of training until end of the intervention (all P < 0.05). There was no evidence that the rate of reduction in DBP changed during the 6-wk intervention (β = 0.15 ± 0.17 SEM, P = 0.37).

Maximal Inspiratory Pressure

PI_MAX was used to establish the training levels for the high-resistance IMST and sham groups and to document changes in inspiratory muscle strength. Study-level data showed moderate or large effects of high-resistance IMST on PI_MAX relative to sham, with the exception of the study by Ramos-Barrera et al. (35) which indicated a greater effect of sham on PI_MAX (Fig. 3A). Heterogeneity among the studies was high (I² = 80.3%).

Figure 3.

Maximal inspiratory pressure (PI_MAX) data in the sham (n = 61; 37 men, 24 women) and inspiratory muscle strength training (IMST; n = 67; 38 men, 29 women) groups. A: forest plot showing standardized mean difference (SMD) between IMST and sham groups. Boxes and whiskers indicate SMD ± 95% confidence intervals. The diamond indicates the combined SMD ± 95% confidence interval, whereas the green line indicates the 95% prediction interval. Positive values indicate favoring IMST and negative values indicate favoring Sham. B: two-way (Training group × Time) repeated-measures ANOVA revealed a group × time interaction (P < 0.01), with PI_MAX in the IMST group significantly higher from baseline at end-intervention (P < 0.01) and higher than sham at end-intervention (P < 0.01). C: independent t tests revealed a significant between-group difference for the change in PI_MAX over the 6-wk intervention (P < 0.01). D: one-way repeated-measures ANOVA with Dunnett’s multiple comparisons test revealed a significant increase in PI_MAX compared with baseline after 1 wk of training (all P < 0.05). Data (B and D) are means ± SD. *P < 0.05 vs. pre.

With participant-level data there was a significant group by time interaction for PI_MAX (P < 0.01). PI_MAX increased in the sham group (baseline: 82 ± 22 cmH₂O, end-intervention: 88 ± 23 cmH₂O; P = 0.01) and in the high-resistance IMST group (baseline: 87 ± 28 cmH₂O, end-intervention: 107 ± 28 cmH₂O; P < 0.01), such that end-intervention PI_MAX was significantly greater in the IMST group compared with sham (P < 0.01) (Fig. 3B). The average increase in PI_MAX following IMST (+20 ± 17 cmH₂O) also was significantly greater relative to sham training (+6 ± 14 cmH₂O; P < 0.01, Fig. 3C).

PI_MAX increased by 3.4 ± 7.8 cmH₂O/wk on average during the high-resistance intervention, such that PI_MAX was greater than baseline from training week 1 until end of the intervention (all P < 0.05). The magnitude of the week-to-week increase in PI_MAX declined over 6 wk (β = −1.6 ± 0.2 SEM, P < 0.01), indicative of diminishing gains as training progressed.

Predictors of Responsiveness

Baseline characteristics from participants in the high-resistance IMST group were examined to identify factors that may predict changes in BP or PI_MAX. In addition, the relation between the change in BP and PI_MAX was evaluated to determine whether the magnitude of improvement in inspiratory muscle strength bore any correspondence to the magnitude of the BP reduction.

Reductions in casual SBP with high-resistance IMST were not influenced by sex (women: −9 ± 7 mmHg, men: −9 ± 6 mmHg; P = 0.55) nor BP-lowering medication (medication-naïve BP: n = 54, mean change −9 ± 6 mmHg, medicated: n = 13, mean change −8 ± 8 mmHg; P = 0.48); the impact of BP-lowering medication on SBP was similar when young adult were removed from the analysis (medication-naïve BP: n = 31, mean change −11 ± 6 mmHg, medicated: n = 13, mean change −8 ± 8 mmHg; P = 0.11). Reductions in SBP were influenced by older age and higher initial SBP (Table 3). Stepwise linear regression analysis identified only subject age for inclusion in the model (β = −0.07 ± 0.03 SEM, R² = 0.07, P = 0.04). The decline in SBP was not associated with the change in PI_MAX (β = 0.02 ± 0.05 SEM, R² < 0.01, P = 0.69).

Table 3.

Predictors of responsiveness to 6 weeks of high-resistance inspiratory muscle strength training

	ΔPIMAX			ΔCasual SBP			ΔCasual DBP
Baseline Variable	β	R 2	P Value	β	R2	P Value	β	R 2	P Value
Age	−0.22 ± 0.09	0.09	0.02*	−0.07 ± 0.03	0.07	0.04*	0.01 ± 0.02	<0.01	0.67
BMI	−0.15 ± 0.43	<0.01	0.73	0.10 ± 0.16	0.01	0.56	−0.03 ± 0.11	<0.01	0.75
PIMAX	−0.17 ± 0.07	0.08	0.02*	0.01 ± 0.03	<0.01	0.66	−0.03 ± 0.02	0.04	0.13
SBP	−0.16 ± 0.14	0.02	0.25	−0.10 ± 0.05	0.06	0.04*	−0.01 ± 0.03	<0.01	0.69
DBP	−0.16 ± 0.22	0.01	0.46	−0.11 ± 0.08	0.03	0.16	−0.10 ± 0.05	0.06	0.06

Results are from simple linear regression. β values are means ± SE. BMI, body mass index; DBP, diastolic blood pressure; PI_MAX, maximal inspiratory pressure; SBP, systolic blood pressure.

^*P < 0.05.

Reductions in casual DBP with high-resistance IMST were slightly greater in men than in women (men: −5 ± 5 mmHg, women: −3 ± 3 mmHg; P = 0.07) and greater in participants naïve to BP-lowering medications (medication-naïve BP: −5 ± 4 mmHg; medicated: −2 ± 3 mmHg; P = 0.04); the impact of BP-lowering medication on DBP persisted when young adults were removed from the analysis (medication-naïve BP: mean change −5 ± 4 mmHg, medicated: mean change −2 ± 3 mmHg; P = 0.01). A higher initial DBP also was associated with a greater reduction in DBP (P = 0.06); no other subject characteristics were associated with the change in DBP (Table 3). Stepwise linear regression identified medication-naïve BP (β = −3.12 ± 1.3 SEM, R² = 0.09, P = 0.02) and above-normal baseline DBP (β = −0.12 ± 0.05 SEM, R² = 0.08, P = 0.04) as the strongest predictors of IMST-related reductions in DBP. Consistent with SBP, the decline in DBP was not associated with change in PI_MAX (β = −0.02 ± 0.05 SEM, R² = 0.01, P = 0.45).

Gains in inspiratory muscle strength (i.e., PI_MAX) with high-resistance IMST were not influenced by sex (women: +20 ± 16 cmH₂O, men: +20 ± 18 cmH₂O; P = 0.95) but were influenced by BP medication status (medication-naive: +23 ± 17 cmH₂O, medicated: +11 ± 11 cmH₂O; P = 0.03); however, there was no impact of BP-lowering medications on PI_MAX when young adults were removed from the analysis (medication-naïve BP: mean change +20 ± 17 cmH₂O, medicated: mean change +11 ± 11 cmH₂O; P = 0.10). Older age and higher baseline PI_MAX were negatively associated with training-related improvements in PI_MAX; no other variables were associated with change in PI_MAX (Table 3). Stepwise linear regression identified age (β = −0.20 ± 0.09 SEM, R² = 0.07, P = 0.03) and baseline PI_MAX (β = −0.15 ± 0.07 SEM, R² = 0.07, P = 0.04) as the strongest predictors of change in PI_MAX.

DISCUSSION

We combined data from five randomized, sham-controlled clinical trials to evaluate the efficacy and time course of high-resistance IMST for lowering BP and to identify subject characteristics that may impact the efficacy of the intervention. Whereas individual studies included in the data set represent distinct populations [i.e., normotensive young adults (32, 34), middle-aged, and older adults with obstructive sleep apnea and above-normal BP (33, 35), and generally healthy older adults with above-normal BP (31)], the combined data set included adults across a broad age range (18–80+ years) with low to moderate CVD risk. Overall, this analysis confirmed our primary hypothesis, as IMST significantly reduced both SBP and DBP relative to sham training. We also hypothesized that significant reductions in BP would be apparent after 1 wk of IMST. Significant reductions in DBP were apparent after 1 wk, and for SBP after 2 wk—both of which highlight a relatively rapid time course for BP reduction with IMST. Finally, our hypothesis that initial BP levels would be the primary clinical factor impacting the effectiveness of IMST was partially confirmed, as initial BP levels and other select clinical factors were only modestly related to changes in BP. Collectively, this large heterogeneous sample supports high-resistance IMST as an effective nonpharmacological therapy to reduce BP in all age groups relatively quickly (1–2 wk).

In the analysis of this large and diverse subject cohort, we found that 6 wk high-resistance IMST lowers average casual SBP by 9 mmHg and casual DBP by 4 mmHg, reductions consistent with a 30% lower risk for CVD (43). This clinically significant reduction in BP is comparable with that observed with pharmacotherapy (44–46) but larger than those seen on average with established lifestyle interventions, such as aerobic exercise (47, 48), dietary sodium restriction (49, 50), and weight loss (51). Importantly, the BP reductions reported here were attained with a weekly training commitment of ∼25–35 min, equivalent to 20% of the weekly time required to meet aerobic exercise guidelines (52).

Clinically significant IMST-related reductions in SBP (−4 mmHg) and DBP (−3 mmHg) emerged in the second and first training weeks, respectively. This time course of BP reduction is similar to that observed with dietary sodium restriction (53), but more rapid than the response to aerobic exercise training; the latter has been reported to take ∼5 wk of training to lower BP in hypertensive adults (54) and up to ∼6 mo in normotensive young adults (55). Importantly, IMST-related declines in SBP and DBP continued throughout the 6-wk intervention, with no evidence of a plateau. We do not expect linear reductions in BP to be sustained indefinitely. Although we do not yet know the precise timepoint at which the response diminishes, one of the individual trials that contributed to the current dataset (31) documented reductions in SBP that persisted (∼75%) after abstaining from IMST for 6 wk. On this basis, a reduced training frequency (perhaps 3 days/wk) may be sufficient to preserve the BP reduction over the longer term. The trajectory and potential magnitude of the BP reduction that can be attained with high-resistance IMST currently is the focus of additional ongoing clinical trials with longer treatment durations (56, 57).

Although this retrospective analysis does not provide insight on the mechanisms through which IMST lowers BP, results from the individual trials that comprise this analysis suggest that point to reductions in sympathetic nerve activity (33, 35), improvements in peripheral resistance and vascular endothelial function (31, 34), and changes in circulating factors related to oxidative stress and inflammation (31) as potential mechanisms mediating the BP-lowering effects of IMST.

Six weeks of high-resistance IMST increased PI_MAX by 20 cmH₂O on average, with improvements evident after the first training week. However, unlike BP, the magnitude of the weekly gain in PI_MAX declined over the course of the intervention and reached a plateau by week 6. Similar plateauing has been reported in other muscle groups (e.g., quadriceps) in response to resistance training, wherein the greatest rates of gains in strength occur at the outset of training and diminish over time, despite progressive increases in resistance (58). Sham training also improved PI_MAX, although the magnitude of improvement was less than one-third of that observed in the high-resistance IMST groups. In view of these outcomes, and given that age-related declines in PI_MAX may increase chronic disease risk (59–67), high-resistance IMST may be an appropriate and accessible training format for use by older adults to improve respiratory muscle strength.

The dissociation of IMST-related changes in PI_MAX and BP is a key finding and consistent with previous observations. First, the time course of improvement for BP and PI_MAX differed. That is, gains in PI_MAX tapered and plateaued by week 6 whereas SBP and DBP showed steady declines throughout the intervention without evidence of plateau. Second and more notable, whereas both sham training and high-resistance IMST groups made gains in PI_MAX there was no relation between gains in PI_MAX and reductions in SBP or DBP. Taken together, high-resistance IMST has concomitant effects on PI_MAX and BP, however, improvements in BP are not PI_MAX-dependent. This suggests it is the repeated generation of large, negative intrathoracic pressure during IMST, and not the concomitant change in inspiratory muscle strength, that leads to reductions in BP.

Clinical factors had a modest impact on BP outcomes. Both older age and higher initial SBP were associated with larger reductions in casual SBP with IMST. However, as advancing age correlates with increased SBP (68), these two variables were considered potential confounding variables in predicting the efficacy of IMST. Further analysis suggested that age was the primary predictor for SBP reductions with IMST, with older adults experiencing larger reductions in SBP. However, although statistically significant, the impact of advancing age appeared modest (i.e., 15 years of aging associated with a 1.0 mmHg greater reduction in SBP with IMST). On this basis, high-resistance IMST may be considered similarly effective at lowering SBP across a broad age range.

Other factors were found to influence the effectiveness of IMST for lowering casual DBP. Adults naive to antihypertensive medications and with higher initial DBP were more likely to exhibit greater reductions in casual DBP (i.e., 8 mmHg higher baseline DBP corresponding to a 1.0 mmHg reduction in DBP). Conversely, those taking BP-lowering medication exhibited smaller reductions in DBP with IMST (−2 mmHg vs. −5 mmHg for those not on BP medication). Importantly, even a 2 mmHg reduction in DBP in those on BP medication confers a significant reduction in CVD risk (69).

Older age was associated with more modest improvements in PI_MAX, whereas lower initial PI_MAX values were modestly associated with larger improvements in PI_MAX. Taking BP-lowering medications initially appeared to attenuate the increase in PI_MAX with IMST; however, a sensitivity analysis in which young adults were excluded from consideration nullified this effect. Including age in the regression model also mitigated the impact of BP-lowering medications on PI_MAX; therefore, age appears to be the factor driving this effect. In general, aging is associated with changes in respiratory function (59–61), whereas those with a lower initial PI_MAX may have greater capacity for training-induced improvements in PI_MAX. Nevertheless, these associations were quite modest, indicating broad efficacy.

Experimental Considerations

All trials that comprised this data set implemented 6 wk of high-resistance (75% PI_MAX), low-volume (30 breaths) IMST, 5–7 days/wk. The similarity in the designs of the included studies is a strength of this analysis and allowed for direct comparisons between trials. This study included men and women across a broad age range, making the findings generally applicable across sex and age groups. However, as most participants were non-Hispanic white adults, additional studies that include participants from other racial and ethnic groups are needed. Finally, as the most common health conditions in our cohort were above-normal SBP and obstructive sleep apnea, the findings cannot be extrapolated to other patient populations, such as those with established CVD, at this time.

Conclusions

Using compiled data from several independent small trials, here we provide the strongest evidence to date that 6 wk of high-resistance IMST consistently induces clinically meaningful reductions in casual SBP and DBP, while improving inspiratory muscle function in adult men and women across a wide age range. In addition, we show that improvements in BP and PI_MAX are apparent after only 1–2 wk of the intervention. We also demonstrate that the efficacy of high-resistance IMST is only modestly impacted by select subject characteristics, indicating it is likely to be an effective BP-lowering lifestyle intervention for a broad range of users. Finally, we show that the change in BP is not related to changes in PI_MAX, suggesting it is the generation of large, negative intrathoracic pressures during IMST, and not the associated change in inspiratory muscle strength, that drive the vascular conditioning effects (e.g., BP-reduction, improved endothelial function) of the intervention. These results provide support for time-efficient, high-resistance IMST as a promising lifestyle intervention for lowering BP and possibly decreasing the risk for CVD and other BP-related health conditions.

GRANTS

This work was supported by grants by National Institutes of Health (NIH) R01AG065346, American Heart Association (AHA) 16GRNT26700007, NIH R21AG061677, NIH K01HL153326, AHA 18POST33990034, NIH 5T32HL007249-45, and NIH F31HL154782.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

E. Fiona Bailey is an editor of Journal of Applied Physiology and was not involved and did not have access to information regarding the peer-review process or final disposition of this article. An alternate editor oversaw the peer-review and decision-making process for this article.

AUTHOR CONTRIBUTIONS

D.H.C. and D.T. conceived and designed research; D.H.C., K.A.F., J.R.V., and C.M.D. performed experiments; D.H.C., D.T., K.A.F., and J.L.M. analyzed data; D.H.C., D.T., K.A.F., J.L.M., D.R.S., and E.F.B. interpreted results of experiments; D.H.C. and D.T. prepared figures; D.H.C. and D.T. drafted manuscript; D.H.C., D.T., K.A.F., J.L.M., J.R.V., C.M.D., D.R.S., and E.F.B. edited and revised manuscript; D.H.C., D.T., K.A.F., J.L.M., J.R.V., C.M.D., D.R.S., and E.F.B. approved final version of manuscript.