Reproducibility of the transvaginal sonographic assessment of bladder wall thickness

To assess whether transvaginal sonographic measurements of bladder wall thickness (BWT) have adequate reproducibility to detect differences in BWT potentially indicative of detrusor overactivity in women with overactive bladder.

Three reproducibility studies were undertaken to assess (A) total measurement error, (B) intra-observer variability and (C) inter-observer measurement in the interpretation of scans. Women recruited to the Bladder Ultrasound Study underwent a transvaginal ultrasound scan to obtain a measurement of BWT. When a second observer was available, women who agreed to have two transvaginal scans by different operators were recruited into study A. For study B the first observer reassessed a sample of the recorded images at a later date whilst for study C, a random selection of BWT images were read by a second assessor. Analytical variability, percentage of variability attributable to measurement error, within-person variation and the smallest real difference detectable were estimated.

One hundred and twenty-one women took part: 27 had repeat scans, 37 had scans re-read by the same observer, and 57 had scans read by two observers. In study A, 39% of the total variability in measurements was explained by measurement error (the remainder to within person change); the standard deviation (SD) of measurement error was 0.76mm and the smallest detectable clinical difference was 2.1mm. The SD of measurement error from scan interpretation was 0.42mm within observers (study B) and 0.35mm between observers (studyC).

The high levels of measurement error for a small measurement of BWT means it is unlikely Transvaginal ultrasound measurements have insufficient reliability and reproducibility to be an accurate diagnostic test.