Postural control alterations after severe COVID-19 in older adults using smartphone-based assessment of trunk responses to mechanical perturbations

Abstract

Background

Older adults recovering from severe COVID-19 may present persistent postural control impairments that increase fall risk, yet anticipatory and compensatory responses to external perturbations remain poorly understood.

Objective

To investigate postural adjustments during predictable trunk perturbations in older adults with a history of severe COVID-19 and examine agreement between smartphone-derived measurements and a three-dimensional motion capture system.

Methods

Twenty-seven older adults were allocated to a Healthy Group (HG, n = 18) and a Post-COVID Severe Group (PCG, n = 9). Predictable anteroposterior trunk perturbations were delivered by a pendulum, with simultaneous L5 acceleration recordings using motion capture (120 Hz) and a smartphone accelerometer (50 Hz). Outcomes included APA_onset, APA_amplitude, CPA_range and CPA_restabtime. Group comparisons were corrected using the Benjamini-Hochberg false discovery rate (FDR). Instrument agreement was assessed using Spearman's coefficient, intraclass correlation coefficient (ICC), and Bland-Altman analysis. Effect sizes (Hedges' g) are reported with 95% confidence intervals (CIs).

Results

The PCG showed earlier APA_onset and longer CPA_restabtime than the HG. No significant differences were found for APA_amplitude or CPA_range, although the 95% CIs of the corresponding effect sizes were wide, precluding firm conclusions about the absence of group differences. Smartphone and motion capture measurements showed moderate-to-strong correlations, whereas ICC values were lower, particularly in the PCG. The significant between-group difference in CPA_restabtime was detected only by the kinematic reference system. Findings should be interpreted considering the modest sample size, particularly in the PCG.

Conclusion

Older adults recovering from severe COVID-19 exhibit altered postural responses. Smartphone inertial sensors showed acceptable agreement for some variables but should be used cautiously in post-COVID assessment.