Electrode comparison for impedance-based venous occlusion plethysmography

Abstract

Deep vein thrombosis (DVT) is the third most common cardiovascular disease, immediately after heart attack and stroke. The methods widely used for diagnosing DVT, contrast venography and ultrasound, often provide a reliable diagnosis, but both still have weaknesses, such as invasiveness and ionizing radiation in contrast venography, and the requirement of a skilled operator in ultrasound. Therefore, new methods are continuously needed for early DVT detection.

Impedance-based venous occlusion plethysmography (VOP) has been a widely studied method for DVT detection. This method utilizes a thigh cuff to cause temporary venous occlusion causing the pooling of blood in the lower limb. This pooling is measured using impedance plethysmography (IPG). The deflation of the cuff allows venous return, and a rapid venous outflow is observed in healthy individuals. Slowed venous outflow and decreased venous capacity can indicate DVT. However, there is limited research related to the optimization of the method, including IPG electrodes. Particularly interesting are novel textile-based materials, which could enable the use of VOP method for continuous monitoring and integration into wearable devices.

The purpose of this thesis is to evaluate the capability and performance of different IPG electrodes for VOP measurement. The thesis particularly focuses on band-type textile electrodes that provide comfort for continuous monitoring and easy integration into wearable devices. The current distribution of band electrodes is also more uniform, which could enable more reliable measurement of blood pooling. In addition to the silver-plated textile band electrodes, textile spot electrodes plated with silver and stainless steel, metal plate electrodes made of stainless steel, and traditional adhesive and disposable Ag/AgCl electrodes were tested in this thesis.

The comparison of the electrodes was conducted by collecting IPG-data in three separate experiments, involving a total of ten healthy volunteers. The signal-to-noise ratio (SNR) of the electrodes was calculated for all four electrode types. The time dependency of electrode-skin contact impedance and SNR was evaluated for both textile electrodes and gel electrodes. Textile band electrodes were compared to gel electrodes in VOP measurement, where DVT was simulated using a pneumatic tourniquet.

The median SNR for textile band electrodes was 8.67 dB, for textile spot electrodes 8.07 dB, for metal plate electrodes 8.28 dB and for gel electrodes 7.42 dB. However, there was no statistically significant differences between any electrodes according to Wilcoxon signed-rank test. The decrease in contact impedance of textile electrodes over time was considerably higher compared to gel electrodes, whose contact impedance was relatively stable. Band electrodes provided high signal quality immediately, similar to gel electrodes, while textile spot electrodes required a long settling time to obtain high-quality signal. Gel electrodes provided better classification between the VOP measurements with and without DVT simulation. The diagnostic sensitivity for DVT simulation of gel electrodes was better (72 %) compared to textile band electrodes (40 %). Diagnostic specificity was 100 % for both electrodes. Results still show that textile band electrodes are capable of measuring IPG and VOP signals, even if their performance to detect DVT can be lower compared to gel electrodes. The performance of textile band electrodes can also be improved by developing them further, making them a more desired option for continuous monitoring and wearable device than traditionally used gel electrodes.