Upper trunk flexion versus whole trunk elevation in semi-Fowler’s position: Effects on ventilation and heart rate responses

An a priori power analysis was not performed, as this study was designed as an exploratory physiological investigation rather than a confirmatory trial. The within-subject repeated-measures crossover design, in which each participant served as her own control across three conditions, substantially increases statistical efficiency by eliminating inter-individual variability from the error term, thereby requiring fewer participants than parallel-group designs to achieve comparable power (Wellek & Blettner, 2012).

Fourteen healthy young women were recruited using convenience sampling (age, 20.57 ± 0.73 years; height, 162.07 ± 5.4 cm; weight, 53.13 ± 5.57 kg; BMI, 20.19 ± 1.46 kg/m²). All participants were physically inactive and did not engage in regular sports activities. None of the participants had smoking or drinking habits. This study deliberately recruited healthy young women to establish baseline physiological responses in controlled conditions. Restricting participants to a homogeneous population minimized confounding factors, such as age-related vascular compliance changes and disease-related compensatory mechanisms. Women were specifically selected, given the documented sex differences in respiratory mechanics: females demonstrate a greater reliance on rib cage inspiratory muscles (Molgat-Seon et al., 2018; Sheel et al., 2016), making them an appropriate model for investigating thoracic positioning interventions. The exclusion criteria were respiratory or cardiovascular diseases, low back pain, and obesity. Low back pain was excluded because individuals with chronic low back pain may exhibit altered neuromuscular strategies during postural tasks, potentially confounding respiratory mechanics (Tajik et al., 2026). Of the 15 participants initially measured, one was excluded due to equipment malfunction, resulting in 14 participants for analysis. This study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the Ethics Committee of the authors’ affiliated institution (Approval No. 21-lg-16). Written informed consent was obtained from all participants before their inclusion in the study.

Participants were positioned in three postures: supine (SPIN) and two semi-Fowler’s positions at 30° upper trunk flexion (UT30) and 30° whole trunk elevation (WT30) (Figure 1). This study employed 30° semi-Fowler’s positions rather than Fowler’s positions exceeding 30°, as lower-angle semi-Fowler’s positions (approximately 30°) are commonly used for frail patients (Rauen et al., 2009; Grap et al., 2005; Grap & Munro, 2005; Zhu et al., 2020). Postures were achieved using an electric hospital bed (KA-335, Paramount Bed, Tokyo, Japan), positioning cushions, and a mattress. The 30° angle was verified using a goniometer attached to the bed base. For the UT30, trunk flexion was adjusted at the tenth thoracic vertebra (T10). The T10 level was identified by palpating the spinous processes, and cushions were placed to ensure that the flexion fulcrum was accurately located, following the method described by Kubota et al. (2015). During both the UT30 and WT30 positions, the participants were instructed to maintain their knees in slight flexion.

Figure 1

Bed positions for each condition. SPIN: Supine. UT30: Lower and upper trunk inclined at 0 and 30°. Segments were subdivided based on the spinous process of the 10th thoracic vertebra. WT30: Lower and upper trunk inclined at 30°

A preliminary experiment was conducted on a separate day to allow the participants to become accustomed to breathing while wearing the ventilation measurement mask. Participants were instructed to avoid strenuous exercise, caffeine, alcohol consumption, and food intake after dinner the day before the experiment and to completely refrain from food and drink for 2 h before the experiment. The laboratory temperature was maintained at 28°C, within the thermoneutral zone, based on previous research (Kubota et al., 2017, 2022). Participants were asked to wear sleeveless tops and shorts with properly fitted, non-constrictive bras to avoid chest compression.

All experiments were conducted between 11:00 and 15:00 to minimize the circadian influences. Measurements were taken after a 15-min rest in the supine position. Each position was measured for 5 min with 10-min rest periods between measurements. The measurement order was randomized to eliminate order effects.

Ventilation volumes were measured using a ventilation measurement mask and pneumotachograph (BIOPAC Systems, TSD117, USA). Electrocardiogram (ECG) data were collected using a three-lead ECG module (BIOPAC Systems, ECG100C, USA) with disposable Ag/AgCl electrodes (Vitrode M, Nihon Kohden, Tokyo, Japan) placed in the standard Lead II configuration. Data acquisition was performed using a BIOPAC MP150 data acquisition system (BIOPAC Systems, CA, USA), sampled at 1,000 Hz.

Tidal Volume (TV), Minute Ventilation (MV), Respiratory Rate (RespiR), and RR intervals (RRi) were calculated. RRi represents the time required for one cardiac cycle and corresponds to the reciprocal of the heart rate. Subjective breathing difficulty was assessed using a Numerical Rating Scale (0 = easy, 10 = difficult). A Linear Mixed Model (LMM) was used to analyze the effects of postures on these parameters, with postural differences as fixed effects and individual differences as random intercepts, and p-values were calculated using the Kenward–Roger method. The estimated marginal means and confidence intervals were calculated, and comparisons between postures were performed using Tukey’s method. Cohen’s d was calculated as the effect size. The normality of residuals from the LMM was visually confirmed using Q–Q plots. Effect sizes were interpreted as small (d = 0.2), medium (d = 0.5), and large (d = 0.8) according to Cohen (1988). Statistical analyses were conducted using R version 4.3.3 (R Core Team) and the lme4 package. The significance level was set at less than 5%.

RR intervals were used as the primary variable for statistical hypothesis testing of cardiac chronotropy for the following reasons. First, under resting conditions, cardiac chronotropy is predominantly modulated by vagal tone, and vagal activity linearly modulates the cardiac cycle length (RR interval), whereas the relationship between vagal activity and heart rate is hyperbolic (Task Force, 1996; Lee et al., 2018; Heathers, 2014). Analyzing RR intervals, therefore, more accurately reflects the underlying autonomic control. Second, at the relatively low heart rates observed in the present cohort of healthy young women under resting conditions (approximately 65–67 bpm), the non-linear transformation compresses numerical variability, potentially reducing statistical power to detect posture-related changes (Sacha, 2013; Quigley et al., 2024; Berntson et al., 1995). Third, residual analysis confirmed that RR interval data better satisfied the normality assumptions required for the Linear Mixed Model compared to heart rate data. Heart rate values derived from RR intervals are additionally reported in Table 1 to facilitate clinical interpretation.

The significantly higher TV in WT30 than in SPIN was consistent with previous studies comparing the sitting and supine positions (Romei et al., 2010; Agostoni & Hyatt, 2011; Takahashi et al., 1998). This can be attributed to the gravitational effect, which causes the abdominal organs to move downward when the trunk is raised, resulting in a lowered diaphragm and increased thoracic volume (Agostoni & Hyatt, 2011; Mezidi & Guérin, 2018).

In contrast, while UT30 involves raising the thoracic region and head, the abdomen remains unraised. Therefore, the gravitational downward movement of the abdominal organs that occurs in WT30 is less likely to occur. Nevertheless, UT30 showed an increased ventilation volume compared to SPIN. This pattern may reflect an altered thoracoabdominal configuration in UT30, which could reduce the cranial pressure of the abdominal contents on the diaphragm compared with the completely supine position. We further hypothesize that passive support of the upper trunk may modify abdominal wall tension; however, because abdominal muscle activity and intra-abdominal pressure were not measured, this possibility remains untested and should be interpreted cautiously.

Regarding MV, which is the product of TV and RR, considering that RR remained consistent at approximately 14 breaths/min across all postures with no significant differences, we conclude that the postural differences observed in TV influenced MV, resulting in differences between SPIN and the two Semi-Fowler’s positions.

The cardiovascular responses differed markedly in their pattern relative to baseline. The significant shortening of RR intervals in WT30 reflects the expected orthostatic stress response, in which gravitational pooling reduces venous return, triggering a baroreceptor-mediated heart rate increase to maintain cardiac output. In contrast, UT30 preserved the supine RR interval values. This pattern is consistent with the possibility that UT30 imposes less hydrostatic challenge to venous return than WT30, but the direct statistical comparison between UT30 and WT30 did not reach significance (p > 0.05). These findings are consistent with previous reports that upper trunk flexion may lessen cardiovascular perturbation during trunk elevation (Kubota et al., 2015, 2017). However, given the limited sample size and the possibility of a Type II error, the present data should not be interpreted as demonstrating a significant between-posture advantage. Rather, the current results show that WT30 differed significantly from SPIN, whereas UT30 did not.

Although no significant differences were observed in the NRS, this experiment was conducted under resting conditions and represented a low physical burden for healthy individuals. Therefore, we believe that the participants had difficulty perceiving the differences in breathing between the different postures.

Upper trunk flexion in the semi-Fowler’s position represents a simple, non-pharmacological approach without specialized equipment, requiring only strategic pillow placement or adjustable bed positioning. These findings also suggest that upper trunk flexion may serve as a comfortable and physiologically efficient resting posture in daily life, providing adequate ventilation without imposing additional cardiovascular strain. Although the present findings are derived exclusively from healthy young women, this favorable physiological profile raises the possibility that UT30 could offer advantages for populations with limited cardiovascular reserve; however, this extrapolation remains speculative and requires direct validation in clinical samples. Notably, healthy hearts readily compensate for orthostatic changes in venous return, whereas a failing heart may respond differently to the hemodynamic alterations induced by trunk positioning; therefore, direct clinical application should be approached with caution. For patients with conditions such as heart failure, recent myocardial infarction, or hemodynamic instability, avoiding additional cardiovascular stress is a primary therapeutic goal. Supporting the clinical relevance of positioning interventions, Zhu et al. (2020) demonstrated that the semi-Fowler’s position can improve comfort and reduce adverse symptoms during extubation after abdominal surgery. If the ventilatory enhancement without concurrent tachycardia observed in healthy participants is replicated in clinical populations, UT30 could represent a useful positioning strategy for individuals requiring respiratory support or for whom circulatory stability is a concern. Future studies in patients with compromised cardiovascular reserve are necessary to test this hypothesis.

This study has several limitations. First, the sample consisted exclusively of healthy young women with intact autonomic function and cardiovascular reserve, limiting generalizability to men, older adults, or clinical populations. Patients with impaired autonomic reflexes, reduced cardiac reserve, or altered vascular compliance may exhibit qualitatively different hemodynamic responses to postural changes; therefore, the clinical implications discussed above should be regarded as hypothesis-generating. Second, while we hypothesized that reduced abdominal wall tension drives ventilatory improvement in UT30, we did not directly measure abdominal muscle activity (EMG) or intra-abdominal/gastric pressures. Future mechanistic studies should incorporate these measures to confirm the proposed pathways. Third, we measured RR intervals as a surrogate for cardiac chronotropy but did not directly assess stroke volume or cardiac output. Finally, the lack of a statistically significant difference in the direct comparison of the RRi between UT30 and WT30 warrants caution; larger studies are needed to determine whether a between-posture hemodynamic difference exists. Furthermore, the Numerical Rating Scale for breathing difficulty may lack sufficient sensitivity to detect subtle differences under resting conditions in healthy individuals, potentially contributing to the absence of significant NRS differences between postures. Additionally, an a priori power analysis was not conducted. While the sample size of 14 participants is consistent with recommendations for exploratory investigations (Julious, 2005; Kunselman, 2024) and is supported by the statistical efficiency of the within-subject crossover design (Wellek & Blettner, 2012), it may have been insufficient to detect a significant difference in the direct UT30–WT30 comparison for RR intervals. Larger confirmatory studies with prospective sample size calculations are needed to evaluate between-posture cardiovascular differences.