Algorithm for Characteristic Parameter Estimation of Gastric Impedance Spectra in Humans

Abstract

Impedance spectroscopy has been proposed as a method of monitoring mucosal injury due to hypoperfusion and ischemia in the critically ill. The present paper presents an algorithm developed to calculate the characteristic electrical values that best describe human gastric impedance measurements and simplify the information obtained with this method. An impedance spectroscopy probe and nasogastric tube (ISP/NGT) was placed into the stomach of healthy volunteers, cardiovascular surgery and critically ill patients, and a database with 16199 spectra was obtained. The gastric spectrum forms two semi circles in the complex domain, divided into low frequency (F < 10 kHz) and high frequency (F > 10 kHz). A fitting algorithm was developed based on the Cole model, and central characteristic parameters were calculated. The parameters were validated using the normalized mean squared error and 0.66% of the spectra were discarded. From the experimental data obtained in humans, the greatest changes observed as the gastric mucosa becomes ischemic occur at low frequencies, which are specific and sensitive to tissue damage, and vary with the degree of hypoperfusion.

Gastrointestinal ischemia monitoring through impedance spectroscopy as a tool for the management of the critically ill

“…Gastric reactance shows high sensitivity to predict outcome, compared with some hemodynamic and regional perfusion variables. 25,26 Low-frequency gastric reactance, X L , had a significant predictive value of postoperative persistent shock requiring more than 48 h of vasopressors and associated complications, before, during, and after surgery (P < 0.05) (see Table 1). Patients with a high reactance (X L > 26) before surgery had a significantly higher incidence of complications, higher mortality, and more days in the intensive care unit (ICU) than patients with a low reactance (X L < 13).…”

“…With this process, we can condense the information from 50 measurements to six characteristic parameters: R L (central resistance at low frequency), R H (central resistance at high frequency), X L (central reactance at low frequency), X H (central reactance at high frequency), f L (central frequency at low frequency), and f H (central frequency at high frequency). 24,26 Of all these measurements, we propose the use of the X L as the primary measurement to assess the health of the gastric mucosa as it is sensitive to mucosal cell edema and changes in intracellular to extracellular volume ratio (IV/EV), but it is insensitive to changes in gastric contents and the electrode to mucosal interface impedance. Experimental results described below confirmed that X L is the most descriptive and reliable index of all the other characteristic parameters.…”

“…Our device ( Figure 10) has been shown to provide reproducible measurements under clinical conditions that contain relevant information on the patient's progress. 25,26,28 The impedance spectra also provide a distinctive shape or fingerprint that will be effective to identify improper placement, excessive movement artifact or noise and device failures, and therefore improve our confidence in the measurement. The challenge was to reduce the dimensionality of the complex measurements to one or two simple variables that were robust and easy to interpret so that this system could be used effectively in clinical practice.…”

“…We are using theoretical models to try to extract from the overall spectral shape simple parameters that are more robust to noise and more importantly have a physiological meaning that will be more easily interpreted. We proposed that low-frequency central reactance (X L ) was the principal parameter that reflects tissue edema caused by prolonged ischemia, 26,28 also it only appears in live tissues. We used that parameter to classify injured mucosa (patients in red in Figure 9) in human gastric tissue with the proposed thresholds, showing a very high incidence of gastric ischemia in the critically ill. 28 We also observed alterations in cellular and tissular integrity by light microscopy analysis that confirmed gastric reactance increments obtained during ischemia in a rat model.…”

Gastric Tissue Damage Analysis Generated by Ischemia: Bioimpedance, Confocal Endomicroscopy, and Light Microscopy

“…Most of the authors in the bioimpedance field use the Cole-Cole equation to describe their experimental results characterizing the tissue bioimpedance with four parameters. For our device, an adaptation of Cole-Cole model was performed to develop an algorithm to calculate the 6 characteristic electrical values that best describe human gastric impedance measurements [17]. The device has been tested and validated during the last 12 years.…”