Healthy volunteers >18 Yrs old, non-smokers, with no known respiratory disease and no respiratory symptoms (e.g., dyspnea, cough) were recruited among hospital staff and their relatives. All had normal spirometry, with FEV₁/FVC > 0.70 and FEV₁ and FVC within the expected ranges for age, sex, and height (Table 1).

To exclude possible obstructive subjects, each participant first performed forced spirometry measurements (M. E. C. PFT systems pocket spirometry, meeting ATS standards and calibrated daily for BTPS conditions).

After demonstrating normal spirometry defined by FEV₁/FVC > 0.7, we asked the subjects to breathe at a quiet, natural rate and to perform a resting inspiratory capacity procedure (IC_rest) at least 4 times. Only the highest volume was chosen from the reproducible IC_rest performed. Finally, with clear instructions and active encouragement from our physiotherapist, the subjects started MVV by breathing in and out as deeply as possible relative to their FVC, with controlled BF, inhaling normal air with a low-resistance mouthpiece (MicroGard II filter, inspiratory/expiratory resistance: <0.04 kPa/(L/s) at 1 L/s; <0.4 cmH2O/(L/s) at 1 L/s).

BF was controlled using a metronome, starting at a target BF of 10 breaths·min⁻¹, which was steadily increased in steps of 10 breaths.min⁻¹, until a maximum BF of 100 breaths.min⁻¹ was reached. Each maneuver was separated by a rest period based on each individual’s need to recover from the previous BF session. If the target BF could not be achieved during the MVV, the maneuver was repeated, with no more than two attempts allowed. Each frequency was maintained for 12 s during MVV, with, again, an IC maneuver at the end of the MVV to detect any DH.

We considered that DH appeared for a given BF when the IC observed was reduced by at least 150 mL compared with the IC_rest (6, 9, 10). The ΔIC ≥ 150 mL cutoff was selected for DH because it exceeds short-term IC variability observed under supervised conditions and has been used to denote clinically meaningful DH; in healthy individuals, this threshold is conservative (12). To minimize false positives from a single IC maneuver, DH required confirmation at the next BF step (BF + 10), thereby establishing a sustained IC decline consistent with best practices for serial IC as a surrogate of EELV (13).

When DH was first observed at a given frequency (BF_DH), it was confirmed by an additional MVV test performed at the BF_DH incremented by 10 breaths.min⁻¹ (BF_DH + 10 breaths.min⁻¹) to validate the recurrence of DH. In this situation, we considered that DH was observed at BF_DH and was documented.

As long as no DH was detected, the MVV procedure was performed with 10 breaths.min⁻¹ increments in BF continue until BF_DH is observed or until a maximum of 100 breaths.min⁻¹-BF is reached.

For each BF investigated, mean tidal volume (V_T) was assessed based on MVV and BF performed (MVV/BF).

To determine the BF at which more than 50% of healthy subjects experienced DH, a time-to-event (Kaplan–Meier) analysis was performed with the occurrence of DH as the dependent variable and BF or ventilatory ratio [MVV/(FEV₁*35)] or volume ratio [V_T/FVC] as the independent variables. The event was defined as the first IC decrease ≥150 mL vs. the rest, confirmed at the subsequent BF step (BF + 10), ensuring a sustained IC decline as a surrogate of EELV increase. Participants who did not develop DH by BF = 100 breaths.min⁻¹ were treated as right-censored at their last observed BF. Kaplan–Meier estimates and 95% CIs were computed; between-age subgroup curves (40–49, 50–59, 60–69 yrs) were compared with the log-rank test (Mantel-Cox), and the same procedure was applied to compare curves between women and men. Censoring was assumed to be non-informative.

Plots of time-to-event analysis are displayed for (1) BF, (2) ventilatory ratio, and (3) the complement [1-(V_T/FVC)], since V_T/FVC decreases with BF, its complement provides an increasing variable consistent with Kaplan–Meier analysis. Participants without DH observed by BF = 100 breaths.min⁻¹ were right-censored at that time point on each scale.

In order to explore potential variables to consider in the appearance of DH, correlations were estimated between BF_DH, ventilatory ratio, and volume ratio with demographic or spirometry parameters.

The analysis performed was two-sided with α = 0.05, using IBM SPSS Statistics 20.

The earliest onset of DH was observed at a BF of 10/min in 2 subjects (5% of the sample), and 4 of the 43 participants were right-censored at BF = 100 breaths.min⁻¹ (Figure 1). Time-to-event analysis revealed that, for the overall sample, DH was found in 50% of the sample at a BF_DH of 52 breaths.min⁻¹ [95%CI: 39–64 breaths.min⁻¹]. After excluding the four subjects without DH, the mean BFDH for the 39 subjects with DH was 52 breaths.min⁻¹ (Figure 1). The log-rank test showed no difference in median time-to-event BF_DH between the three age groups (p = 0.515) (Table 2), nor with gender (p = 0.803).

The earliest occurrence of DH was noted in the first subject at a ventilatory ratio of 27% of the maximal ventilation estimated by the ratio [MVV/(FEV₁*35)].

Time-to-event analysis revealed that, across the whole sample, DH occurred in 50% at a ventilatory ratio of 92% [95%CI: 78–106%] (Table 3; Figure 2).

The log-rank test showed no difference in median ventilatory ratio [MVV/(FEV₁*35)] of DH occurrence between the three age groups (p = 0.626) (Table 3), but we found a significant difference between genders, with a DH that appears at lower values in women (p = 0.022).

The earliest occurrence of DH was noted in the first subject at a volume ratio of 77% of the measured FVC estimated by the ratio [V_T/FVC].

Time-to-event analysis revealed that, for the overall sample, DH occurred in 50% of the sample at a volume ratio reduced to 52% [95%CI: 46–58%] (Table 4; Figure 3). For the 39 subjects with DH occurrence, a median reduction of volume ratio of 52% [95%CI: 46–58%] was found (Table 4).

The log-rank test showed no difference in median volume ratio of DH occurrence between the three age groups (p = 0.350) (Table 4), nor with gender (p = 0.166).

In summary, among the 39 subjects (91% of the sample) who experienced a DH when they performed an MVV at a frequency equal to or less than 100 breaths.min⁻¹, the diagram below shows that 50% of these will develop a DH (1) when the imposed BF exceeds a frequency of 52 breaths.min⁻¹, (2) when MVV exceeds 92% of the maximum theoretical ventilation, and (3) when the engaged V_T no longer exceeds 52% of the FVC (Figure 4).

Despite a BF of 100 breaths.min⁻¹ being achieved, the four subjects (9% of the sample) who never experienced DH reached a maximum ventilation ratio of 87 to 131% and a minimal volume ratio between 21 and 34%.

Minimal volume ratios and maximal ventilation ratios were not observed at the precise maximal BF tested of 100 breaths.min⁻¹. While minimal volume ratios were observed at a BF close to the maximum (ranged from [80–100 breaths. Min⁻¹]), peak ventilation ratios appeared when BF ranged between [50–90 breaths.min⁻¹].

We observed that the BMI of the subjects without DH (n = 4) was significantly lower than the BMI of the subjects with DH (n = 39) (20.6 ± 2.9 vs. 25.0 ± 3.3 Kg.m⁻²; p = 0.018). When we compared other demographic or spirometric parameters, we did not observe any differences between these two groups.

Correlations revealed that the lower the resting functional residual capacity (i.e., lower resting EELV), and consequently the higher the [IC_rest/FVC] ratio, the earlier DH appeared at lower BF (r = −0.824; p < 0.0001) and at higher relative tidal volume ratios [V_T/FVC] (r = 0.847; p < 0.0001), as shown in Table 5 and Figure 5. Also, the higher a subject’s BMI, the more DH appears to be associated with high [V_T/FVC] (r = 0.446; p = 0.004) (Table 5).

According to our results, a very wide range of BF (with an average of 52 breaths.min⁻¹) is inducing DH, even in healthy subjects, and even if they are stimulated to deploy as much volume as possible. Conversely, we observed that when the BF prescribed does not allow the healthy subject to perform more than an average of 54% of his FVC, then a DH appears.

In our cohort, during MVV, DH emerged when V_T was ~52% of FVC, consistent with healthy-exercise reports that V_T plateaus ~55–63% of vital capacity (VC) as inspiratory expansion nears thoraco-pulmonary limits: EILV approaches total lung capacity (TLC), and EELV can rise when expiratory time is shortened (13).

Extending this framework, Chuang et al. showed that V_T/TLC discriminates healthy vs. obstructive/restrictive patterns better than V_T/FVC or V_T/IC and proposed practical V_T/TLC bounds linked to DH vs. normal expansion (21). Our metronome-imposed breathing frequency during MVV shortens expiratory time, a constraint under which prior MPT work has shown EELV shifts depend on breathing pattern (20).

Although we did not measure TLC, two features support a mechanical reading: (1) by definition, a fall in IC at constant TLC denotes a rise in EELV (DH), and a ≥ 150 mL threshold with confirmation at BF + 10 limits false positives; (2) the median DH at V_T/FVC ≈ 52% matches the expected V_T “plateau behavior” from Guenette et al. (12, 13).

In line with these mechanics, in our study, DH typically occurred at ~91% of predicted VE (FEV₁*35) and a median BF ≈ 52 breaths.min⁻¹, and lower resting IC_rest/FVC predicted earlier DH, cautioning against reliance on MVV = FEV₁*35 substitutions given their poor agreement with measured MVV (14).

This ventilatory threshold is usually well above the peak VE observed during a CPET, which is consistent with a DH rarely observed in healthy subjects at maximal effort. Moreover, we observed a significant inverse correlation between the BF_DH and the ratio of resting IC to FVC [IC_rest/FVC], which represents the resting static constraints of the respiratory system (p < 0.0001). It reveals that the lower the FRC, the lower the BF_DH occurs.

Given the physiological loss of elastic release of the thoraco-pulmonary system with age, which affects maximal expiratory flow, we expected to observe a lower BF threshold for inducing DH in the oldest participants in our experiment. Our observations could not confirm this hypothesis since no differences were found between younger and older participants within our 40–70 Yrs age range (p > 0.05).

Our sample was selected from a healthy population showing expected values for the spirometry parameters, such as FEV₁ or forced expiratory flow (FEF) at different FVC levels, that could not be related to the onset of DH, whether considering the BF, the ventilation ratio, or the volume ratio.

Conversely, we found a difference for the ventilation ratio according to gender (p = 0.022). This last observation could be explained by the fact that for authors, predicted maximum ventilation may vary according to gender, estimated for women at [FEV₁*35] (what we have chosen for our entire population) and for men at [FEV₁*38] (22).

Although we did not directly assess inspiratory muscle fatigue, several features of our protocol and data argue against fatigue as the primary driver of the IC fall at DH onset. Ventilation during the 12-s bouts increased from 10 to 70 breaths.min⁻¹ (mean MVV: 33.6 → 98.4 L.min⁻¹), showed a modest inflection at 80 (96.2 L.min⁻¹), then rose again to a peak at 90 (115.3 L.min⁻¹) and remained near-plateau at 100 (113.3 L.min⁻¹). Among participants with a BF_DH of around 70–80–90 breaths.min⁻¹, 60% increased from 70 → 80 breaths.min⁻¹ and 60% from 80 → 90 breaths.min⁻¹ and in those with BF_DH = 90, 33% increased from 90 → 100. These observations are consistent with performance maintenance under the imposed pacing. Together with short bout duration and individualized rest, this pattern suggests that the observed IC decrease reflects a real decrease in operating lung volume rather than a progressive decrement in ventilatory performance.

Obesity shifts operating lung volumes by reducing Expiratory Reserve Volume (ERV) and Functional Residual Capacity (FRC) while generally preserving TLC, thereby lowering resting IC and bringing EILV closer to TLC during high ventilatory demand, a mechanism that can advance DH onset (23).

Consistent with an added thoracic load, obesity blunts the ventilatory response to exercise and is associated with higher EELV from rest to peak, with EILV approaching TLC in older adults, supporting greater mechanical constraints on V_T expansion (24).

In our cohort, resting IC_rest/FVC predicted the BF at DH (higher IC_rest/FVC → earlier DH), and BMI correlated with the V_T/FVC at DH, suggesting that higher BMI (and possibly central adiposity) may depress IC and may precipitate DH under MVV.

These observations argue for stratifying future analyses by BMI and fat distribution and for incorporating direct operating-volume measures (EELV/EILV or TLC) to quantify BMI’s modulation of DH risk.

Also, BMI, which significantly correlated with IC_rest (r = 0.333; p = 0.029) and may therefore play a role in CRF levels, appears to be an important parameter to consider in the onset of DH. Firstly, we observe that BMI differentiates between subjects with and without DH (p = 0.018), with overweight (25.0 ± 3.3 Kg.m^-2) and standard (20.6 ± 2.9 Kg.m⁻²) BMIs, respectively. Furthermore, BMI correlates with the volume ratio at which DH appears (p = 0.004).

In light of these observations, it would be interesting to investigate whether, beyond obesity, being overweight and/or gender could be risk factors for DH.

Convenience and professional-network recruitment may have introduced selection and clustering biases, combined with age and sex imbalances, further limiting generalizability.

Participants were classified as healthy based on a fixed FEV₁/FVC > 0.70, which we acknowledge is insufficient on its own to exclude subclinical disease; ideally, static lung volumes (TLC, RV, DLCO) would be included, but these were not measured here. Nevertheless, baseline spirometry was broadly normal (FEV₁ averaged ~106% predicted with nearly all participants ≥80% predicted), supporting a healthy classification. Despite identical instruction and encouragement from the same physiotherapist, V_T and therefore VE during the MVV could still have been influenced by individual motivation, while BF was the only parameter monitored.

Respiratory muscle fatigue, whether acute during repeated maneuvers or reflecting underlying strength and endurance, was not formally assessed. DH and fatigue may coexist, and serial IC (a valid proxy for EELV) cannot, by itself, exclude that some inspirations at the end of a bout were not truly maximal (13). In line with ERS perspectives that MVV is influenced by chest-wall/lung mechanics as well as “fatigue,” causal attribution from MVV alone is non-specific (14). Future studies combining IC with direct fatigue indices (e.g., MIP/MEP or SNIP, and plethysmographic EELV) could help clarify the relative contributions of these mechanisms.

The four subjects who did not develop a DH at the imposed 100 breaths.min⁻¹ BF were included in the Kaplan Meier estimates, which may therefore have slightly overestimated (for the mean BF_DH and the mean ventilatory ratio) or underestimated (for the mean volume ratio). Future studies should specifically address the role of respiratory muscle function in relation to DH.