In order to check the proposed hypotheses, a correlation coefficient (r) of 0.5 was expected for the sample size estimation, considering previous studies that examined the relationship between trunk stability tests and whole-body dynamic balance tests [23,26]. The software G*Power 3.1 (v3.1, University of Düsseldorf) was used with the following parameters: r=0.5; α=.006; 1–β=.8. Calculations suggested a sample size of approximately 46 participants. An additional 20% of the target sample was recruited to anticipate the loss of subjects due to the inability to implement some of the tests presented. A total of 53 volunteer participants aged 65 years or older (mean age 70.3, SD 3.8 y; 21 male participants and 32 female participants) were recruited for the study. The main sociodemographic and anthropometric characteristics and the physical activity levels of the participants are shown in Table 1. Moreover, Table S1 in Multimedia Appendix 1 shows the clinical data of the participants. Participants were asked to present a medical certificate signed by their primary care physician to guarantee that they had no contraindications to perform trunk stability, whole-body balance, and gait-related tests.

The University Office for Research Ethics (DCD.FBM.01.23) approved this study. The study was conducted in accordance with the Declaration of Helsinki, informing the participants of the characteristics and risks of the study. Each participant provided written informed consent. Data were handled to ensure privacy and confidentiality and were accessible only to the research team. No financial compensation was provided. Written informed consent was obtained for the publication of the participant’s images.

Each participant’s mass (mean 69.6, SD 10.3 kg); height (mean 1.62, SD 0.10 m); and height of the center of mass position of the head, arms, and trunk segment (estimated at 62.6% [29] of the trunk height, measured as the distance between the greater trochanter and the glenohumeral joint: 0.53, SD 0.05 m) were registered before testing. The participant’s preferred leg was established as the leg that the participant would use to kick a ball. Participants carried out 2 testing sessions with a day’s rest in-between, and these sessions were repeated 1 week later. The first assessment session consisted of a trunk stability assessment through lumbopelvic stability tests (using a randomized block design to avoid fatigue influence) during the back bridge, the preferred side bridge (determined by the side of the preferred leg), and the front bridge and the bird-dog positions [22]. The second assessment session consisted of trunk stability, dynamic balance, static balance, and gait-related tests (carrying out first those tests in which performance is most likely to be affected by fatigue) performed in the following order: (A) the unstable sitting posturographic test, (B) the tandem dynamic balance stance posturographic test, (C) whole-body static balance tests, (D) straight-line gait tests, and (E) the modified Timed Up and Go test. A 5‐ to 7-minute warm-up was performed at the beginning of each testing session to enhance awareness of spine and pelvis alignment and to facilitate the understanding and practice of neutral spine posture. The following exercises were included in this warm-up: (1) isolated scapular retractions and protractions; (2) isolated pelvic retroversions and anteversions; and (3) the cat-camel exercise.

Statistical analyses were performed to address the study hypotheses and evaluate the reliability of the measurements. Data (presented as mean and SD) from those participants who carried out all the testing sessions were used for the statistical analyses. The normality of the data distribution was explored through the Shapiro-Wilk test. Since the data did not follow a normal distribution, nonparametric statistics were applied. The Wilcoxon test was performed for each test score to explore the existence of statistically significant mean differences between sessions, and the Spearman’s correlation coefficient was used to analyze the relationship between the variables.

In order to avoid possible bias due to the low consistency of certain variables, typical error (TE) and intraclass correlation coefficient (ICC_3,1) were calculated (confidence limits set at 95%) to evaluate the absolute and relative test-retest reliability, respectively, and to determine which variables could be used in subsequent correlational analyses. The TE was calculated as the SD of the difference between the 2 testing sessions divided by √2. Furthermore, TE values were also expressed as percentages to facilitate data extrapolation, interpretation, and comparison with the pertinent literature. Percentage TE was calculated as the TE×100 and divided by the average of test-retest means. ICC values were interpreted according to the following criteria: poor (<0.50), moderate (0.50‐0.75), good (0.75‐0.90), and excellent (>0.90) [30]. Both absolute and relative reliability indexes were calculated through the spreadsheet proposed by Hopkins [31].

If some variables obtained a poor level of relative reliability [30] (ie, ICC <0.50), the Tukey outlier classification test was applied to the change between sessions. Outliers are based on the IQR (ie, [25th percentile]–1.5×IQR and [75th percentile]+1.5×IQR). The data obtained in the second assessment session for the variables that obtained a moderate-to-excellent level of relative reliability [30] (ie, ICC >0.50) were used to perform a correlation analysis (r) between them. Correlational analyses were performed with JASP 0.16.2 software (Eric-Jan Wagenmakers, Department of the Psychological Methods, University of Amsterdam, Nieuwe Achtergracht 129B). The correlation coefficient was interpreted as low (0.30‐0.49), moderate (0.50‐0.69), high (0.70‐0.89), and very high (≥0.90) [32]. In order to minimize the probability of obtaining significant correlations by chance, the prespecified significance level (P<.05) was divided by the number of comparisons analyzed (Bonferroni correction).

To examine whether the observed correlations were influenced by the choice of data reduction strategy (session 2 peak-performance vs mean performance averaged across sessions 1 and 2), we conducted complementary sensitivity analyses as described in the previous section. First, we repeated the reliability analyses using mean performance across attempts within each session. Second, we repeated the correlational analyses using the average of session 1 and session 2 mean performance as a more representative estimate of typical performance and to reduce potential measurement noise. The pattern of results across these analyses was then compared with the original session 2 peak-performance analyses.

Given that the lumbopelvic stability tests represent the primary outcome of the study and that they are also the tests most affected by feasibility constraints at higher difficulty levels, an additional focused sensitivity analysis was performed to determine whether test-retest reliability metrics and the key correlations observed in the main analyses based on the session 2 peak-performance criterion were robust enough when analyses were restricted to a more homogeneous group of participants with higher physical fitness who successfully completed all 3 test levels.

The primary findings of this study highlight a significant association between superior lumbopelvic control during specific bridge and bird-dog exercises and enhanced whole-body static balance. Conversely, no such associations were observed between trunk stability and whole-body dynamic balance or gait-related outcomes. Importantly, consistent with existing literature [5,23], the lack of correlation between trunk stability and whole-body balance in laboratory and field tests underscores a critical point for sport science professionals and practitioners: the direct extrapolation of data obtained with the tools available in field settings to laboratory findings remains unfeasible.

Our main results confirmed the first hypothesis, so maintaining a static position in floor-based lumbopelvic stability exercises was associated with the capability to maintain a static upright posture [33]. Thus, incorporating these trunk stability exercises, which require trunk, shoulder, and leg strength to reach the posture, short-term muscle endurance, and muscle coactivation for postural control (via feedforward and/or feedback mechanisms) [34], into a workout routine may enhance whole-body static balance in older adults. However, not all stability exercises showed the same degree of association. Back bridge and bird-dog variations showed more and stronger correlations, emphasizing the role of the trunk-hip extensor (ie, some of the antigravity muscles of the posterior kinetic chain) to maintain static posture. These associations were more evident when balance was increasingly challenged by narrowing the support base through changes in foot position (tandem vs hip-width position), suggesting a greater involvement of the lumbopelvic complex. The asymmetrical tandem stance tends to induce body rotation and lateral adjustments, which may align with the demands of the single-leg back bridge and conventional bird-dog, both involving asymmetry, rotation, and multiplanar control. Conversely, frontal and lateral bridge variations showed fewer and lower correlations. The lack of relevant correlations for the side bridge variations was particularly unexpected as tandem stance tasks seem to have high frontal plane demands and previous studies have linked mediolateral balance to trunk control [34,35]. These results may be due to the fact that control during lateral bridges is heavily mediated by the physical condition of the shoulder complex [36], which could be more affected in older adults by age-related issues such as sarcopenia, tendinopathies, and arthritis [37].

While a significant positive correlation was observed between the well-established trunk stability and whole-body dynamic balance laboratory tests, our second hypothesis was partially rejected, as the strength of these correlations was unexpectedly low (r=0.436). This suggests that, although a statistical link exists, trunk stability as measured by the unstable sitting posturographic test may only have a limited practical impact on whole-body dynamic balance in physically active older adults, particularly compared to populations with greater balance deficits in which stronger associations have been reported [26]. In this sense, tandem stance performance showed stronger associations with trunk stability in people with multiple sclerosis [26] (who present neuromuscular deficits) when balance was greatly challenged by placing the weaker leg behind (r=0.540; P<.05), compared to when the stronger leg was behind (r=0.433; P<.05), suggesting a greater trunk involvement when postural demands increase. In our sample of physically active older adults, the demands in the unstable sitting posturographic test may not be as determinant for whole-body dynamic balance compared with more disabled populations. In this sense, once a minimum level of trunk stability is reached in a seated position, it no longer contributes to improving whole-body dynamic balance in tandem stance unless a greater balance challenge is imposed on this population [23]. This interpretation should be taken with caution, as higher associations were found (r>0.629; P<.05) in other samples of older adults of similar characteristics [38], suggesting that the role of trunk stability may have a different impact on whole-body dynamic balance in other scenarios.

Finally, as expected and given the lack of significant associations, maintaining static positions during bridging and bird-dog exercises seems to play a limited role in whole-body dynamic balance during voluntary weight-shifting tasks and also voluntary control during the unstable sitting posturographic test seems less relevant for whole-body static balance. The complex scenario presented in this section aligns with a previous study [39] that confirmed the involvement of task-specific cortical and subcortical pathways after analyzing the excitability of corticospinal inputs to trunk muscles during both voluntary and involuntary tasks. The study [39] also highlighted that this involvement could not be easily predicted based solely on whether the task was classified as voluntary or more “automatic” in nature. This highlights the need for personalized trunk stability assessments that take into account individual characteristics and the underlying neurophysiological mechanisms related to task demands, which may functionally impact whole-body balance in older adults.

As expected, no association was observed between trunk stability and gait-related outcomes. However, the third hypothesis was partially rejected, as the anticipated correlation between the unstable sitting posturographic test and the modified Timed Up and Go test was not observed. As a previous study on physically active older adults showed [38], trunk stability does not seem to play a significant role in gait performance or even in a gait-related task, such as the Timed Up and Go test, which requires standing, sitting, and turning actions in this population. A previous study in minimally and moderately impaired people with multiple sclerosis [26] found an association with the Timed Up and Go test but not with a straight-line gait test, which is justified by the greater relevance of trunk control needed for its performance. Therefore, in populations with neuromuscular deficits, voluntary trunk control may have a more relevant role during the sit-to-stand, turn, and stand-to-sit phases of the Timed Up and Go test. Again, this interpretation should be taken with caution, as another study involving individuals with greater disability, such as those with hemiparesis following a stroke [40], found no association between an unstable sitting posturographic test and the Timed Up and Go test.

The seemingly counterintuitive inverse association between whole-body static balance (higher acceleration indicating worse control) and walking speed (straight-line gait and Timed Up and Go) might be influenced by anthropometric factors, such as participant height. Taller individuals, while potentially achieving higher walking speeds due to longer strides, often face greater challenges in maintaining static postural control due to a higher center of mass and increased pendulum-like dynamics [41]. This suggests a complex interplay wherein static balance demands differ from dynamic mobility requirements, particularly during phases like sit-to-stand, turning, and stand-to-sit in the Timed Up and Go test [42].

The lack of correlation between tests grouped under the same capability (ie, trunk stability and whole-body balance) reinforces the notion of the multidimensional nature of postural control in older adults, as previously documented in studies involving other populations [5,23]. Therefore, sport science professionals and practitioners should be aware that the lack of correlations between the novel inertial-sensor–based field tools (CoreMaker and G&B apps) and the well-established force-platform-based assessments makes it unfeasible to extrapolate research results found through laboratory tests to the field. Despite the lack of relationship between laboratory and field tools, the technologically based trunk stability and balance assessment through field tools such as CoreMaker and G&B apps will enhance experimental studies, allowing postural control during training exercises to be assessed in larger, more representative samples, leading to more meaningful clinical findings [21]. In addition, measuring trunk stability during the exercises used for training may allow us to determine to what extent the potential benefits caused by trunk-focused exercise programs (bridges and bird-dogs [24,25]) on balance and gait may be due to an improvement in trunk stability.

As with any research, it is important to highlight some limitations. A much larger sample would be desirable in the most difficult variations of bridges (Table S2 in Multimedia Appendix 1) to minimize the probability of a type II error due to undersampling. Second, our results are based on a sample of physically active older adults and of both sexes, so future studies would need to examine the relationships of these tests in more fragile populations and also potential sex differences. Although our study focused on well-established force platform and inertial sensor protocols to assess trunk stability, balance, and gait, it is important to admit that other assessment tools exist and that using different tests might have led to different results. Future studies should analyze and develop field tests to evaluate the reactive control of the trunk due to its apparent potential relationship with the mechanisms required during a fall situation in older people. Moreover, trunk-focused stability outcomes should also be analyzed during specific tests designed to quantify whole-body balance, gait, and functional mobility, particularly under conditions imposing high demands on trunk control, to better understand how performance in these tasks relates to direct measurements of trunk involvement [42,43]. Although for the accelerometry-based field tests postural control was summarized using the average of the acceleration magnitude data series obtained from the 3-axis as a proxy for performance, future studies with more focused designs should incorporate axis-specific and advanced signal analyses to further elucidate control mechanisms during the studied tasks. Although the main conclusions were robust across multiple analytical strategies and the qualitative interpretation of the results remained similar, as with any complex analytical framework, the choice of data reduction strategy may have had a modest influence on the magnitude of some associations. Importantly, the convergence of findings across these contrasting analytical strategies strengthens confidence in the robustness, task-specificity, and interpretability of the observed relationships, indicating that the primary analysis represents a conservative estimate rather than a biased one.

This study provides novel evidence linking lumbopelvic control in floor-based exercises (mainly back bridge and bird-dog) to whole-body static balance in physically active older adults, supporting trunk stability as a crucial factor for maintaining steady standing positions in physically active older adults. Crucially, this research highlights a notable disconnection between novel inertial sensor–based field assessments (CoreMaker and G&B apps) and traditional force platform–based laboratory protocols for trunk stability and balance. While direct extrapolation between these methods is not advisable, the accessibility of field tools like CoreMaker offers a promising avenue [21] for standardized lumbopelvic stability assessment and the monitoring of training load in broader populations. Future interventions targeting static balance in older adults should consider incorporating specific bridging and bird-dog positions [24,25] to enhance trunk control.