The abdominal drawing-in manoeuvre for detecting activity in the deep abdominal muscles: is this clinical tool reliable and valid?

OBJECTIVE The abdominal drawing-in manoeuvre (ADIM) is a common clinical tool for manually assessing whether a preferential activation of the deep abdominal muscles in patients with low back pain (LBP) is 'correct' or not. The validity and reliability of manual assessment of the ADIM are, however, as yet unknown. This study evaluated the concurrent and discriminative validity and reliability of the manually assessed ADIM. DESIGN Single-blinded cross-sectional study. SETTINGS General population in Stockholm County, Sweden. PARTICIPANTS The study sample comprised 38 participants seeking care for LBP, and 15 healthy subjects. MEASURES The manual ADIM was assessed as correct or not following a standard procedure. Ultrasound imaging (USI) was used as the concurrent reference (gold standard) for the manually assessed ADIM by calculating a ratio of the change in muscle thickness between the resting and the contracted states: the correlation between manual test and USI was calculated. Discriminative validity was analysed by calculating sensitivity and specificity. A sample of 24 participants was analysed with κ coefficients for interobserver reliability between two raters. RESULTS The concurrent validity between the manual ADIM and the ADIM-USI ratios showed poor correlations (r=0.13-0.40). The discriminative validity of the manually assessed ADIM to predict LBP showed a sensitivity/specificity of 0.30/0.73, while the ADIM-USI ratio to predict LBP showed 0.19/0.87. The interobserver reliability for the manually assessed ADIM revealed substantial agreement: K=0.71, CI (95%) 0.41 to 1.00. CONCLUSIONS Although the interobserver reliability of the manually assessed ADIM was high, the concurrent and discriminative validity were both low for examining the preferential activity of the deep abdominal muscles. Neither the manually assessed ADIM nor the ultrasound testing discriminated between participants with LBP and healthy subjects regarding preferential activity of the transversus muscle as this ability/inability was also present in healthy subjects.