A portable bioimpedance instrument for monitoring residual limb fluid volume in people with transtibial limb loss: A technical note

Abstract

People with transtibial limb loss experience daily changes in volume of their residual limb that affect the fit of their prosthetic socket. A portable instrument was developed to monitor fluid volume changes outside of the laboratory setting. The bioimpedance system applied 26 current bursts per second at frequencies between 3 kHz and 1 MHz, and sensed voltage at up to six channels. Among six voltage-controlled current source circuits and five receive-channel amplifier topologies considered, a differential Howland current pump and a single receive-channel instrumentation amplifier proved the best combination of low noise and low power consumption. Mean RMS errors were 0.07% for extracellular fluid resistance, 2.23% for intracellular fluid resistance, and 1.15% for membrane capacitance.

Keywords: Amputee; Bioimpedance; Fluid volume; Prosthetic fit; Residual limb; Volume accommodation.

Copyright © 2019 IPEM. Published by Elsevier Ltd. All rights reserved.

Figures

FIG. 1a,b.. Measurement system.

(a) Electronic components of the portable limb fluid volume monitoring instrument are shown on the left. Up to six channels of impedance (Z1… Z6) are monitored. Thigh and distal residual limb impedances (Zthigh and Zdistal, respectively) are included because they must be accounted for in data processing. Between the DAC and VCCS there are two filters. The first filter is an active third order low-pass filter integrated into a fully differential amplifier with a cutoff frequency of 15 MHz. The differential output of this is fed into the second filter, which is a passive low pass filter (balanced Chebyshev Type II with the filter zero at 24 MHz). (b) System in use on a person with transtibial limb amputation.

FIG. 1a,b.. Measurement system.

(a) Electronic components of the portable limb fluid volume monitoring instrument are shown on the left. Up to six channels of impedance (Z1… Z6) are monitored. Thigh and distal residual limb impedances (Zthigh and Zdistal, respectively) are included because they must be accounted for in data processing. Between the DAC and VCCS there are two filters. The first filter is an active third order low-pass filter integrated into a fully differential amplifier with a cutoff frequency of 15 MHz. The differential output of this is fed into the second filter, which is a passive low pass filter (balanced Chebyshev Type II with the filter zero at 24 MHz). (b) System in use on a person with transtibial limb amputation.

FIG. 2a,b.. Current stimulation and voltage sensing circuits.

a: Current stimulation components. b: Voltage sensing components.

FIG. 2a,b.. Current stimulation and voltage sensing circuits.

a: Current stimulation components. b: Voltage sensing components.

FIG. 3.. Determining impedance from collected current and voltage data.

I= resistive component. Q= reactive component.

FIG. 4a,b.. Performance.

a: Cole–De Lorenzo bioimpedance model fitted to impedance data (phase and magnitude). Dots represent the per frequency impedance that is output from calibration. Lines represent the Cole–Delorenzo model fit. b: RMS errors from check circuit testing. Means and standard deviations from 60 sets of data (five circuit boards, four channels on each board, 3 test circuits) are shown. Mean RMS errors were 0.07% for Re, 2.23% for Ri, and 1.15% for Cm.

FIG. 4a,b.. Performance.

a: Cole–De Lorenzo bioimpedance model fitted to impedance data (phase and magnitude). Dots represent the per frequency impedance that is output from calibration. Lines represent the Cole–Delorenzo model fit. b: RMS errors from check circuit testing. Means and standard deviations from 60 sets of data (five circuit boards, four channels on each board, 3 test circuits) are shown. Mean RMS errors were 0.07% for Re, 2.23% for Ri, and 1.15% for Cm.

FIG. 5a,b.. Participant test data.

Results from a test session of >5 h duration conducted in and around a clinical office. a: Participant #1. b: Participant #2. Percent fluid volume changes for the AM, PM, and between-AM-and-PM sections of the test session are shown. An asterisk (*) indicates a doff. A=anterior. P=posterior.

FIG. 5a,b.. Participant test data.

Results from a test session of >5 h duration conducted in and around a clinical office. a: Participant #1. b: Participant #2. Percent fluid volume changes for the AM, PM, and between-AM-and-PM sections of the test session are shown. An asterisk (*) indicates a doff. A=anterior. P=posterior.