Introduction

Studies in animals and humans have brought clear evidence that increasing the operating lung volume, i.e., the end-expiratory lung volume above normal functional residual capacity (FRC) or the tidal volume (VT), reduces airway resistance [1, 2] and can attenuate [3] or reverse [4] the response to bronchoconstrictor stimuli. These effects of breathing at increased lung volume can be explained by either static or dynamic mechanisms. Since airways and lung parenchyma are interdependent, a static increase of lung volume is associated with an increase of airway caliber by the action of tethering forces opposing both the passive elastic recoil of the airway wall and the active contractile forces of airway smooth muscle. On the other hand, studies in-vitro have shown that dynamic swings can blunt the response of airway smooth muscle to contractile stimuli by mechanisms that reduce its force generation capacity [5, 6], though in bronchial segments this effect was observed only when pressure oscillations were raised to twice of those corresponding to normal VT [7]. In vivo, increasing VT [4], or breathing frequency (BF), or both [8] have a bronchodilator effect.

Therefore, it can be expected that increasing FRC or VT, or their combinations, have beneficial effects in counteracting bronchoconstriction in vivo. However, in porcine bronchial segments, static hyper-distension reduced the maximal response to acetylcholine but blunted the relaxant effect of superimposed pressure oscillations of amplitude corresponding to twice the baseline VT [9], raising the possibility that lung hyperinflation may compete with the bronchodilator effects of increasing VT in vivo. In humans, the relative efficacy of physiologically relevant static hyperinflation and increased dynamic swings in countering airway narrowing has not been studied, but it can be hypothesized that they differ, owing to different underlying mechanisms.

To test this hypothesis, we designed the present study to evaluate whether the bronchodilator effect of breathing at increased lung volumes differs depending on whether attained by increasing FRC or VT. Moreover, we investigated whether the bronchodilator effects of increasing FRC and VT were additive.

Methods

Subjects

Nineteen healthy volunteers (13 males/6 females) with no history respiratory/cardiovascular diseases participated in the study. No one was obese. Main anthropometric and respiratory functional data are reported in Table 1. Data were collected at Santa Croce and Carle Hospital (Cuneo, Italy), the protocol was approved by the local Ethical Committee, and each subject gave a written informed consent before participation.

Table 1 Subjects’ anthropometric characteristics and baseline lung functional data
Full size table

Measurements

Spirometry was measured by a mass flowmeter (SensorMedics Inc., CA, USA) following the ATS/ERS recommendations [10]. Respiratory impedance was measured by a forced oscillation technique (FOT) as previously described [11, 12]. Briefly, sinusoidal pressure oscillations (5 and 19 Hz; ~ 2 cmH2O peak-to-peak) were generated by a 16-cm diameter loudspeaker (model CW161N, Ciare, Italy) mounted in a rigid plastic box and connected in parallel to a mesh pneumotachograph and mouthpiece on one side and to a low-resistance high-inertance tube on the other side. Pressure oscillations were applied at the mouth during tidal breathing, while subjects had their cheeks supported by the hands of an investigator to minimize upper airway shunting. The overall load over the tidal breathing frequency range was 0.98 cm H2O•L-1•s. Airway opening pressure and flow were recorded by piezoresistive transducers (DCXL10DS and DCXL01DS Sensortechnics, Germany, respectively) and sampled at 200 Hz. A 15-L/min bias flow of air generated by an air pump (CMP08, 3A Health Care, Italy) was used to reduce dead space to about 35 ml. Pressure and flow signals were processed by a least-square algorithm [13, 14] to calculate respiratory resistance at 5 and 19 Hz (R5 and R19, respectively) and reactance at 5 Hz (X5). Artifacts due to glottis closure or expiratory airflow limitation were avoided by discarding breaths showing any of the following features: i) tidal volume <0.1 L or >2.0 L, ii) difference between measured flow oscillation and ideal sine wave with the same Fourier coefficients >0.2 [15], and iii) ratio of minimum to average X>3.5 [11]. The same breaths were used to measure VT, breathing frequency (BF), inspiratory and total time of each breath (TI and TTot, respectively), and estimate inspiratory drive (VT/TI), inspiratory duty cycle (TI/TTot), and minute ventilation (VE).

Protocol

Pre-study day

Subjects attended the laboratory for spirometry and determination of the dose of methacholine (MCh) to be used for the study day. For this purpose, after baseline FOT measurements, MCh chloride dry-powder (Laboratorio Farmaceutico Lofarma, Milan, Italy) was dissolved in distilled water and administered by an ampoule-dosimeter system (MB3 MEFAR, Brescia, Italy) delivering aerosol particles with a median mass diameter of 1.53-1.61μm, while subjects breathed quietly in a sitting position. The starting dose was of 300 μg followed by doubling doses until R5 increased by at least 100% from baseline.

Study day

Baseline FOT measurements were taken during 2 min of spontaneous tidal breathing. Then, the subjects were trained to breathe, by using visual feed-back of spirometry tracing, for 2 min with imposed combinations of FRC or VT. Thereafter, each subject inhaled a single dose of MCh equal to the last dose given on the pre-study day and R5 was measured 2 min later during spontaneous tidal breathing to confirm the persistence of bronchoconstriction. Then, FOT measurements were taken while subjects maintained for 2 min each of the following imposed breathing patterns in randomized order (Fig. 1): A) spontaneous VT from spontaneous FRC, B) near double VT from spontaneous FRC, C) near triple VT from spontaneous FRC, D) spontaneous VT from FRC increased by 1 VT, E) near double VT from FRC increased by 1 VT, and F) spontaneous VT from FRC increased by 2 VT. For each VT increase the subjects were asked to adjust BF to prevent large increments of VE. Before each change of breathing pattern, R5 was measured during spontaneous tidal breathing to check for the stability of bronchoconstriction. If R5 was 10% or more lower than initial post-MCh value an additional half dose of MCh was given to restore bronchoconstriction. This happened occasionally in 6 subjects, with no relation to any specific breathing pattern. At the end of the study, aerosol albuterol was administered to relieve symptoms if any.

Fig. 1
figure 1

Patterns of breathing before after methacholine (MCh) with tidal volume (VT) initiated from spontaneous or increased functional residual capacity (FRC). For each condition, respiratory impedance measures were calculated over the 3 mid-quintiles of the whole inspiratory phase (upper panel) or over the 3 mid-quintiles of iso-volume inspiratory portions (lower panel) as shown by the thick lines

Full size image

Data analysis

For each breathing pattern, R5, R19, R5-19, and X5 were calculated over the 3 mid-quintiles of the whole inspiratory phase (Fig. 1, upper panel) or over the 3 mid-quintiles of iso-volume inspiratory portions (Fig. 1, lower panel).

Differences in R5, R19, R5-19, X5, VT, BF, VT/TI, TI/TTot, and VE between conditions were tested for statistical significance by a one-way repeated-measure analysis of variance (ANOVA) with Holm-Sidak post-hoc test for multiple-comparisons. Values of p<0.05 were considered statistically significant. Data are presented as mean ± standard deviation (SD).

Results

Breathing patterns during the experimental conditions

The spontaneous breathing pattern after MCh (A) did not differ significantly from the spontaneous pattern before methacholine (Table 2). VT and BF changed with the imposed patterns (B-F) as per protocol. Even though great attention was paid to maintain VE as constant as possible among the imposed breathing patterns, it was with patterns C, E, and F that VE slightly but significantly increased than with patterns than A and B. These differences were associated with significant differences in mean inspiratory, VT/TI. Neither VE nor VT/TI were significantly different among breathing patters C, D, E, and F. There were no significant differences in TI/TTOT among all breathing patterns.

Table 2 Patterns of breathing during experimental conditions
Full size table

Mid-inspiration measures

In general, breathing at increased FRC, increased VT, or both attenuated the changes induced by MCh inhalation on R5, R19, R5-19, and X5 (Fig. 2 and Supplemental Table 1).

Fig. 2
figure 2

Effects of increasing tidal volume from spontaneous functional residual capacity (patterns A, B, C) (A), increasing functional residual capacity with spontaneous (patterns A, D, F) (B), or both (patterns B, E) (C) on mid-inspiration impedance measures. Effects of patterns achieving the same peak volume (C vs. E and vs. F) on mean-inspiratory impedance measurements (D). VT, tidal volume; FRC, functional residual capacity. R5, respiratory resistance at 5 Hz, R19, respiratory resistance at 19 Hz; R5-19, difference in respiratory resistance between 5 and 19 Hz; X5, respiratory reactance at 5 Hz. Columns heights indicate means and error bars standard deviations. *, p<0.005; **, p<0.01; p<0.001

Full size image

Increasing VT from spontaneous FRC was associated with significant reductions of R5, R19, R5-19 and less negative X5 when VT was tripled (pattern C) but not doubled (pattern B) compared to spontaneous breathing (pattern A) VT. Yet, the attenuating effects of pattern C were significantly greater than those of pattern B.

Increasing FRC by either one (pattern D) or two (pattern F) VT with constant spontaneous VT was associated with significant reductions of R5 and R19 than pattern A, while R5-19 was significantly reduced and X5 less negative with pattern F but not pattern D.

Increasing both VT and FRC (pattern E) was associated with significantly lower R5, R19, R5-19 and less negative X5 than increasing VT alone (pattern B) and significantly lower R19 than increasing FRC alone (pattern D).

Breathing patterns with the same peak volume, no matter whether achieved by increasing VT or FRC or both (patterns B vs. D and C vs. E and vs. F) showed insignificantly different effects on airway narrowing.

Notably, R5 (cmH2O•L-1•s) was reduced by 0.57±1.18 when VT was doubled (pattern B vs pattern A), by 1.19±0.70 when FRC was increased by 1 VT (pattern D vs pattern A), and by 1.84±0.88 when both VT and FRC were increased (pattern E vs pattern A). Similarly, R19 (cmH2O•L-1•s) was reduced by 0.29±0.35 when VT was doubled (pattern B vs pattern A), by 0.48±0.46 when FRC was increased by 1 VT (pattern D vs pattern A), and by 0.91±0.42 when both VT and FRC were increased (pattern E vs pattern A). These results suggest simply additive effects, but the increase of FRC was more potent to mitigate airway narrowing than the increase in VT.

Iso-volume measures

In general, R5, R19, and R5-19 were inversely related to the lung volume at which they were measured (Fig. 3 and Supplemental Table 2), while the X5 values were inconsistently related to lung volumes.

Fig. 3
figure 3

Effects of increasing tidal volume from spontaneous (patterns A, B, C) or increased (patterns D, E, F) functional residual capacity on iso-volume inspiratory impedance measures. Other abbreviations as in Fig. 2. Columns heights indicate means and error bars standard deviations. *, p<0.005; **, p<0.01

Full size image

At low iso-volume, R5 and R19, were significantly lower and X5 was less negative than during spontaneous breathing (pattern A) when VT was tripled (pattern C) but not doubled (pattern B). Yet, the attenuating effects of pattern C on R5 and X5 were significantly greater than those of pattern B.

At mid iso-volume, R5, R19, and R5-19 did not differ significantly with increments of VT (patterns B and C), or FRC (pattern D), or both (pattern E). However, X5 was significantly less negative when both FRC and VT were increased (pattern E) than when VT (pattern B) or FRC (pattern D) were increased alone.

At high iso-volume, there were no significant differences with increments of VT (pattern C), or FRC (panel F), or both (panel E).

Discussion

The main findings of the present study in healthy volunteers were that 1) the changes of respiratory impedance induced by inhaled MCh were significantly attenuated by increasing FRC, or VT, or both, 2) increasing FRC had more consistent effects than increasing VT, 3) the effects of increasing FRC and VT were additive, and ) volume-independent effects attributable to tidal stretching were observed only when VT was three times larger than during spontaneous breathing.

Comments on methodology

We used oscillometry because it is the only available method enabling intra-breath measurements of respiratory mechanics over specific portions of lung volume during tidal breathing, but it has two major limitations. First, oscillometry does not directly measure airway resistance but also lung tissue and chest wall resistances. Airway resistance is inversely related to VT whereas lung tissue resistance is inversely related to BF [2]. Therefore, it is possible that the effects of increasing VT on airway caliber were counteracted by the effects of decreasing BF on tissue resistance. We think this had no major effect on our results because the attenuation of R5, which reflects in large part tissue resistance, was not less than the attenuation of R19, which mainly reflect airway resistance. Second, breathing at increased lung volumes requires activation of inspiratory muscles, which increases chest wall elastance [16]. Therefore, we cannot exclude that changes in X5 with different breathing patterns were counteracted by changes in chest wall stiffness.

Although our subjects were asked to maintain VE as constant as possible by decreasing BF when VT was increased, there was a tendency for VE to increase (Table 2), thus likely resulting in an increased alveolar ventilation and airway hypocapnia, mainly when achieved by increasing VT. Hypocapnia has a bronchoconstrictor effect [17], thus possibly counteracting the bronchodilator effects of imposed breathing patterns. We did not measure end-tidal CO2, but we believe this had no major impact on our results for two reasons. First, assuming normal anatomical plus instrumental dead space and CO2 production, we estimated a mean difference in alveolar PCO2 between patterns C and A to be approximately 7 mmHg, which was reported to have insignificant effects on the respiratory impedance of healthy subjects [18]. Second, the differences in VE between any imposed patters were insignificant and differences in alveolar PCO2 presumably minimal.

Finally, for changes in VT were associated with changes in BF and the ratio TI/TTOT remained constant, the effects of tissue viscoelasticity could not be evaluated. Nevertheless, breathing patterns with low BF would have increased the time for airway smooth muscle relaxation during the inspiratory phase but also for re-shortening during the expiratory phase.

Interpretation of results

The present study was designed on the premises that both lung hyperinflation and increased breathing depth are mechanisms protecting against airway narrowing, but their relative efficacies are unknown.

That increasing lung volume is associated with a proportional increase of airway conductance, i.e., the reciprocal of airway resistance, was first reported in 1958 by Briscoe and Dubois [1] and subsequently confirmed in excised animal [19] and human [20] lungs with relaxed airways. This effect was simply attributed to a geometric change of airways being distended by the static radial traction of the surrounding lung parenchyma. Studies in contracted airway smooth muscle strips have consistently shown that sustained step-changes of length can rapidly attenuate active tension, possibly due to disassembly of the contractile apparatus, followed by a gradual recovery due to length adaptation [20, 21]. By contrast, in whole bronchial segments a sustained inflationary increase of transmural pressure also caused an immediate reduction in tension, but this was followed by a continuous gradual decrease [22]. Airway wall stiffening was proposed to explain the difference between intact bronchi and muscle strips [22, 23]. In our study, R5 was stable or decreased between the different breathing patterns, but never increased, which makes the occurrence of length adaptation unlikely. Thus, it is possible that the attenuations of airway narrowing we observed after 2 min of breathing at increased FRC reflected not only geometric changes in airway caliber but also mechanisms opposing both the passive elastic recoil of the airway wall and the active contractile forces of airway smooth muscle.

The inhibitory effect of cycling stretching on airway smooth muscle active force generation has been reported consistently in both isolated muscle strips [5, 6] and isolated bronchial segments [7]. It is well-established in animals [7] and humans [4, 24] that the magnitude of the bronchodilator effects of tidal breathing increases with increasing frequency of breathing and with increasing tidal volume. Two independent lines of evidence suggest, further, that the attenuation of smooth muscle contractile force is attributable to changes of VE, which is the product VT x BF, independently of changes of either VT or BF taken individually [24, 25]. Equivalently, neither the amplitude of tissue cyclic strain nor the cyclic frequency is as important as their product, namely, the amplitude of the tissue strain rate. To assess this phenomenon still further, in this report we used an experimental design in which we held the product VT x BF fixed so as to better isolate the effects of changes in VT alone. This is an important issue in our study, as we see that when VE could not be kept constant (pattern C vs A) the impedance values at low iso-volume were significantly attenuated presumably because of the higher mean inspiratory flow (VT/TI ) causing a faster lung stretching rate rather than the increase in VT itself.

Three theories can be invoked to explain the above findings [26], namely, that stretching of airway smooth muscle causes a plastic rearrangement of the contractile apparatus [6, 27, 28], or modifies the crossbridge cycling rate and latch bridges formation [5] or causes temporary detachment of attached cross bridges [29].

In an attempt to examine the relative bronchodilator effects of static hyperinflation and dynamic stretching, we measured inspiratory impedance in healthy subjects with MCh-induced bronchoconstriction breathing with different combinations of FRC and VT. As expected, increasing either VT or FRC significantly attenuated the changes induced by MCh on R5 and R19, R5-19, suggestive of a generalized increase of airway caliber, but also decreased R5-19 and made X5 less negative. To the extent that an increase in R5-19 and a decrease in X5 reflect heterogeneous distribution of time constants within the lung periphery [30], the significant improvement of these variables with the increase in FRC and VT (Figs. 2 and 3) suggests that increasing lung volumes no matter how it was achieved made ventilation more homogeneous. While the effects of increasing VT on R5 and R19 were significant only when it was threefold the spontaneous VT, the effects of increasing FRC where already significant when it was increased by one VT, suggesting a more consistent effect of increasing static than dynamic tidal stretching.

The effects of increasing both VT and FRC were additive, i.e., the effect of dynamic stretching was not blunted by an increased static stretch. This finding is in apparent contradiction with a study showing that in isolated bronchial segments hyperinflation blunted the effect of pressure oscillations corresponding to twice a normal VT [9] In that study, bronchi were hyperinflated at a transmural pressure of 20 cmH2O, where airway compliance is reduced [7] and so are the amplitude of volume oscillation and airway smooth muscle strain. Examining our data in the light of a previous study [31], (Fig. 3), we estimate that the largest end-tidal inspiratory volumes achieved as with patterns C, E and F would have not exceed the values associated with transpulmonary pressures in excess of 20 cm H2O. Since bronchial transmural pressure might differ from transpulmonary pressure in the presence of bronchoconstriction [32], we cannot exclude that stress on airway walls increased with the increase of end-inspiratory volume. Therefore, the increments of VT in our study were likely to reflect increments of airway smooth muscle strain but not stress. The latter, however, does not seem to be the major determinant of the decrease in airway smooth muscle contractility with breathing maneuvers [33, 34].

The fact that the effects of FRC and VT were simply additive does suggest that lung hyperinflation and tidal swings operated via a similar mechanism, viz. increase of operational lung volume. This interpretation is supported by the lack of differences at iso-volumes among most breathing patterns. The only exceptions were the lower R5, R19, R5-19, and less negative X5 at low lung volume after triple VT and the less negative X5 at mid lung volume with breathing patterns with the highest end-inspiratory lung volume, i.e., tripling VT (pattern C) and doubling VT from increased FRC (pattern E). These findings are consistent with a study in airway segments showing modest dilator effects with peak-to-peak pressure oscillations of 10 but not 5 cmH2O [7]. As FOT measurement were taken during the inspiratory phase, these findings possibly reflect volume-independent dynamic effects on airway smooth muscle persisting after the expiratory phase, even when BF and, in turn, expiratory time for re-narrowing was the largest (pattern C).

Why was hyperinflation more potent than tidal swings against airway narrowing in the present study is a matter of speculation. Increasing either FRC or VT results in increased mean operational lung volume, which is associates with an increase of airway caliber owing to the tethering force of lung parenchyma opposing the passive elastic recoil of airway walls. However, the mechanisms of static and dynamic stretching on airway smooth muscle active force may be different. One possibility is that in our study the sustained increments of operational lung volume maintained the airway smooth muscle in a condition of reduced force generation capacity by disassembling the contractile apparatus before the occurrence of length adaptation [20, 21] or substantial reduction of tethering force due to stress relation of lung parenchyma [35]. By contrast, additional time-dependent effects of tidal stretching, e.g., on cross-bridge cycling rate, were possibly obscured by the re-constriction during expiratory phase unless started from very high end-inspiratory volume. Another possible mechanism explaining the larger bronchodilator effects yielded by the increase in FRC rather than VT could be the larger amount of nitric oxide penetrating the airway lumen when narrowing is relieved by distending lung parenchyma [36].

The results of the present study in healthy subjects cannot be directly extrapolated to asthma because the mechanisms regulating airway smooth muscle contractility and heterogeneity of ventilation may differ in disease. Yet, it is known that FRC increases in asthma with the occurrence of expiratory flow limitation [37] and decreases after bronchodilator treatments [38]. Moreover, some beneficial effects of continuous positive airway pressure against airway responsiveness have been reported. To what extent hyperinflation can alleviate asthma symptoms remains to be elucidated, considering that above a given threshold it may cause an increase of inspiratory work of breathing [39] and limit the increase in VT [21].

In conclusion, this study provides evidence that both lung hyperinflation and increased tidal stretching yield substantial bronchodilatation in human lungs exposed to a constrictor agent, though the former seems more effective than the latter presumably because of additive effects on airway smooth muscle contractile force and non-contractile airway tissues.