Romberg ratio in quiet stance posturography-Test to retest reliability

- Repeated Romberg ratio tests has often been used to evaluate postural control.
- We investigated the validity of repeated Romberg ratio in quiet stance posturography.
- Although consistent quiet stance data were found, calculated ratios were inconsistent.
- Repeated Romberg testing in posturography seem to be unreliable in healthy subjects.
- Validity of Romberg ratio in evaluating single sensory components is questionable.

We investigated test to retest reliability and intraindividual variability of Romberg ratios in quiet stance posturography. Thirty-six healthy young adults (17 males, 19 females aged 15-38 years) were divided into 3 groups with different time-intervals between consecutive trials (20 min, 3 h and 24 h respectively). Each group performed 5 posturography recordings in a randomized order of eyes open (EO) or closed (EC) + once after 3 months. We measured the torque variance in posturography and calculated Romberg ratios. Total postural sway as well as sway above and below 0.1 Hz was analyzed.ResultsTest to retest reliability was found to be poor for Romberg ratios (intraclass correlation coefficients (ICC) <0.4) despite that the individual EO and EC posturography recordings were consistent. For sway >0.1 Hz the Romberg ratios were found to be more consistent (fair to good, ICC 0.49-0.71). The variation between two consecutive tests (absolute difference (%)) was high when using the traditional Romberg ratio (EC/EO), but became less varied if an alternate formula that includes the total postural sway was used ((EC − EO)/(EC + EO) × 100).ConclusionIn healthy young adults the evaluation of ratios from repeated quiet stance posturography show great intraindividual inconsistency. This questions the Romberg ratio as being a reliable tool for evaluation of postural performance and determination of sensory preference in postural control, at least in healthy controls. Whether test-retest reliability is acceptable in patient cohorts needs to be evaluated for proper validity of intervention and outcome studies and for detection of clinical relevance.