Diagnostic accuracy of bioimpedance spectroscopy in patients with lymphedema: A retrospective cohort analysis

Publication date: Available online 9 April 2018
Source:Journal of Plastic, Reconstructive & Aesthetic Surgery
Author(s): Evelyn S. Qin, Mindy J. Bowen, Wei F. Chen
BackgroundBioimpedance spectroscopy (BIS) is used by healthcare specialists to diagnose lymphedema. BIS measures limb fluid content by assessing tissue resistance to the flow of electric current. However, there is debate regarding the validity of BIS in diagnosing early lymphedema. Indocyanine green (ICG) lymphography has been established as the most accurate diagnostic modality to date for lymphedema diagnosis. In this retrospective study, we test the sensitivity, specificity, and diagnostic accuracy of BIS in diagnosing lymphedema by referencing its results with ICG lymphography.MethodsPatients presented to the University of Iowa Lymphedema Center from 2015 to 2017 were evaluated with a standardized protocol that included history and physical examination, a validated lymphedema-specific quality-of-life assessment (LYMQOL), circumference –measurement-based index, BIS, and ICG lymphography. Diagnostic accuracy of BIS was assessed using ICG lymphography as a reference test.ResultsFifty-eight patients had positive ICG lymphography results, which confirmed the diagnosis of lymphedema. ICG lymphographic findings consistently correlated with clinical examination, LYMQOL evaluation, and lymphedema indices. By contrast, BIS demonstrated a false-negative rate of 36% – 21 out of 58 patients had normal BIS readings, but a positive ICG lymphography result. The 21 false-negative results occurred in patients with early-stage disease. Sensitivity and specificity for BIS were 0.64 and 1, respectively.ConclusionBIS carries an excessively high rate of false-negative results to be dependably used as a diagnostic modality for lymphedema. ICG lymphography highly correlates with other tracking modalities, and it remains the most reliable tool for diagnosing lymphedema.