Reproducibility and Concurrent Validity of Manual Palpation with Rehabilitative Ultrasound Imaging for Assessing Deep Abdominal Muscle Activity: Analysis with Preferential Ratios.

The abdominal drawing-in maneuver (ADIM) is a clinical tool used for identifying preferential activity of deep abdominal muscles. However, concurrent validity and reproducibility of palpation during the ADIM has not been formally investigated. The aims of this study were (1) to assess intra- and interrater reliability of manual palpation during the ADIM, and (2) to determine the concurrent validity of manual palpation during the ADIM by calculating preferential activation ratio cut-off as assessed with ultrasound imaging (RUSI). Thirty-two subjects ( = 16 patients with nonspecific low back pain and 16 comparable healthy individuals) performed the ADIM in a supine hook-lying position. Two experienced assessors evaluated the presence or absence of preferential contraction of the deep abdominal muscles by palpation during the ADIM on 2 different days. Intrarater (test-retest) and interrater reliability of palpation were calculated using Cohen's kappa coefficients. Muscle thickness of the transverse abdominis (TrA), internal oblique (IO), and external oblique (EO) muscles at rest and during the ADIM were also measured. TrA-Contraction Ratio (TrA-CR), TrA-Preferential Activation Ratio (TrA-PAR), and Modified-TrA-PR (M-TrA-PAR) were calculated. The concurrent validity of manual palpation was determined using the correlation between manual palpation and imaging and by calculating ROC curve (operating characteristics curve), Youden index, and sensitivity and specificity. Intra- and interrater reliability of manual palpation during the ADIM was excellent (: 0.82-1.00) and good to excellent (: 0.71-1.00), respectively. Interrater reliability for muscle thickness ranged from good to excellent (ICC3,1 0.79-0.91). Manual palpation and TrA ratio showed low to moderate correlations (: 0.36-0.60). When evaluating the diagnostic accuracy of manual palpation, the best predictive model (ROC value: 0.89; < 0.001) for correct a preferential contraction of TrA was obtained when the M-TrA-PAR was ≥0.08 (sensitivity: 0.95-1.00; specificity: 0.62). Good to excellent intra- and interrater reliability of manual palpation was found during the ADIM in both patients and healthy groups. Manual palpation showed concurrent validity for identifying the preferential activity of the TrA muscle supporting its use in clinical practice.