CLINICAL INVESTIGATIONS EIS SYSTEM

Department of Psychiatry and Behavioral Sciences and Department of Medicine University of Miami Miller School of Medicine, Miami,USA

Abstract:

The Electro Interstitial Scan-Body Composition (EIS-BC) device consists of two modules:  (1) the BC module and (2) the EIS module.  The objective of this study was to compare the BC module to a standardized, valid assessment of BC and to compare the EIS module to a standardized assessment of heart rate variability (HRV).  Fifty subjects between 20 and 62 years of age were assessed for body composition by the BC module (total body water, fat-free mass, and fat mass) on one hand and dual x-ray absorptiometry (DXA; total fat mass and fat-free mass).  On the second hand, spectrum analysis of the EIS module and HRV as measured by a standard HRV device (ES Teck PEMS) to estimate sympathetic nervous system activity were assessed.  Height, weight, blood pressure (BP), and pulse were also measured.  The results of the study indicated that the correlation between DXA and EIS-BC body fat percent measurements was very high (r=.92, p < 0.001).  The correlation between the EIS spectrum analysis and HRV variables was also very high (r=.76, p < 0.001), suggesting that the high conductivity ratio has predictive capability on the sympathetic nervous system activity.  The results of the study suggest that the EIS-BC device has a significant level of reliability in estimating body composition and sympathetic nervous system activity.

Botkin 2006
Summary

Clinical investigations were conducted at the S.P. Botkin Hospital from May 20, 2006, to September 1, 2006, in order to evaluate the Bioimpedance parameters provide from a device named Electro Interstitial Scan (E.I.S), we performed drug administration studies.

The EIS system is measuring the living tissue electrical conductivity and electrical dispersion of the human body via tactile electrodes and displays 3 parameters:

• SDC  segments values (electrical conductivity)
• EPA-SPA segments values (electrical dispersion)
• ESG HF/VLF ratio (spectral analysis of the electrical conductivity )

Two hundred fifteen (215) test subjects (Age 54 + 16) were recorded with the EIS System.
These patients presented affections diagnosed by conventional examinations (hypothyroidism, hypertension, atherosclerosis or thrombosis risk, and Major depression) and were undergoing no treatment.
The treatments corresponding to the diseases were decided by the conventional examinations results, and a follow-up being undertaken on one hand with the EIS System and on the other hand by conventional methods.

Hypothesis
1. Could the drugs ‘administration affect the Bioimpedance parameters estimated from the 3 measured parameters of the EIS system and therefore the EIS can be used in adjunct in treatments’ follow up?
2. Could the EIS parameters be used in diagnosis of the diagnosed affections?
The hypothesis 1 was validated according to the raw data analysis:
Thyroid treatment monitoring
The findings show that SDC 11/12 and TSH has a significant negative correlation to each other (r = -0.975, p = 0.005). It shows that, SDC 11/12 shares approximately 95.1% (that is (-0.975)2x100% or 0.951x100%) of its variability with TSH.  Thus, a high value of SDC 11/12 corresponds to low TSH or low value of SDC 11/12 corresponds to high TSH.
The findings show that EPA-SPA11/12 and TSH has a significant positive correlation to each other (r = 0.926, p = 0.024). It shows that, EPA-SPA11/12   shares approximately 85.7% (that is (0.926)2x100% or 0.857x100%) of its variability with TSH.  Thus, a high value of EPA-SPA11/12 corresponds to high TSH or low value of EPA-SPA 11/12 corresponds to low TSH.
Beta blockers treatment monitoring
The findings show that SDC 2/4/15/17 and Diastolic Pressure has a significant positive correlation to each other
(r = 0.975, p = 0.005). It shows that, SDC 2/4/15/17shares approximately 95.1% (that is (0.975)2x100% or 0.951x100%) of its variability with Diastolic Pressure.  Thus, a high value of SDC 2/4/15/17   corresponds to high Diastolic Pressure or low value of SDC 2/4/15/17   corresponds to low Diastolic Pressure.
The findings show that ESGHF/VLF and Diastolic Pressure has a significant positive correlation to each other
(r = 0.977, p = 0.004). It shows that, ESGHF/VLF shares approximately 95.4% (that is (0.977)2x100% or 0.954x100%) of its variability with Diastolic Pressure.  Thus, a high value of ESGHF/VLF corresponds to high Diastolic Pressure or low value of ESGHF/VLF corresponds to low Diastolic Pressure.
ACE inhibitors treatment monitoring
The findings show that EPA-SPA6/8/19/21 and Diastolic Pressure has a significant negative correlation to each other (r = -0.892, p = 0.042). It shows that, EPA-SPA6/8/1921/ shares approximately 79.6% (that is (-0.892)2x100% or 0.796x100%) of its variability with Diastolic Pressure. 
Anticoagulant treatment monitoring
The findings show that SDC 6/13/19 and PI (Prothrombin Index) has a significant positive correlation to each other
(r = 0.998, p < 0.001).
The findings show that EPA-SPA 6/13/19 and PI (Prothrombin Index) has a significant positive correlation to each other (r = 0.961, p = 0.009).
The findings indicate that ESG HF/VLF and PI (Prothrombin Index) has a significant positive correlation to each other (r = 0.994, p = 0.001).
SSRI treatment monitoring
The findings indicate that there were a significant positive correlations between SDC 9/10 and the treatment Response at D+30 (rho = 0.484, p < 0.001) and D+45and D+60 (rho= 0.557, p < 0.001).
The findings indicate that there were a significant positive correlations between EPA-SPA 9/10 and the treatment Response at D+45 (rho = 0.709, p < 0.001) and D+60 (rho= 0.804, p < 0.001).

2. Could the EIS parameters be used in diagnosis of the affections diagnosed?
The hypothesis 2 was NOT validated according to the raw data analysis:
Only the following correlations had been showed:
The Hypertension group 2A data shows the significant positive correlation between EPA-SPA and Diastolic pressure before the treatment (D).
The major depression group 4 data shows the significant positive correlations between SDC 9/10 and the treatment response (R) for D+ 30 and it also provide evidence of significant positive correlation between ESG HF/VLF and the treatment response (R) for D+45 and D+60 measurements.

Caudal 2007
Summary

Clinical trials were conducted at the office of Dr. Frederique Caudal, pediatrician and specialist in Attention-Deficit/Hyperactivity Disorder (ADHD) in children.
Symptoms of this disorder are related, in the current literature, to a low level of cerebral neurotransmitters.
The diagnosis of ADHD children is almost symptomatic, which leads to the dramatic possibility of error and treatment (Ritalin®, or SSRI or catecholamine’s) with medications associated with numerous side effects in particular for the age of the population.
For this reason, a new, measurable, and therefore objective marker was proposed using the Electro Interstitial Scanner (EIS) Bioimpedance measurement device, in adjunct to the conventional diagnoses and treatment monitoring of the ADHD children.
From 10.04.2006 to 05.16.2007, data from 59 children (age 12 + 5) presenting ADHD diagnostic and not undergoing treatment were recorded with the EIS System. This database was compared with another control group database (age 14 + 6) of non-hyperactive children also recorded with the same EIS System.
Hypothesis tested
The hypothesis tested was:

• Can the EIS device with reference to its ESG (Electro Scan Gram) graph be used as a marker for ADHD children and therefore as adjunct to the conventional diagnosis of ADHD children?

This hypothesis was validated by statistical analysis.
Independent Sample T-test
In order to determine the differences between ADHD and the control group in the values of v9/v10, v2/v4/v15/v17 and v1/v3/v16/v18 comparison of means via independent samples t-test was conducted. Table 1.1 presented the t-test results for v9/v10 scores at ADHD and the control group. It shows the number of cases per group used in the analysis, the means, the standard deviation, degrees of freedom, t value and the significance of the test (p-value). The findings in Table 1.1 indicate that the mean of v9/v10 at ADHD (M=78.38) was significantly (p<0.001) higher than the mean of v9/v10 at the control group (M=21.75). Thus, the score of v9/v10 at ADHD was expected to be higher than the scores of v9v10 at control group.
                                                                         
Table 1.1
Independent Sample T-test for v9/v10 between ADHD and Control

Table 1.2 shows the t-test results for v2/v4/v15/v17 between ADHD and the control group. Accordingly, the mean of v2/v4/v15/v17 at ADHD (M=52.96) was significantly (p=0.003) higher than the mean of v2/v4/v15/v17 at the control group (M=47.96). Thus, the score of v2/v4/v15/v17 at ADHD was expected to be higher than the scores of v2/v4/v15/v17 at the control group.

Table 1.2

Independent Sample T-test for v2/v4/v15/v17between ADHD and Control

T-test results for v2/v4/v15/v17 between ADHD and the control group were presented in Table 1.3. the findings shows that the mean of v1/v3/v16/v18 at ADHD (M=30.34) was significantly (p<0.001) higher than the mean of v1/v3/v16/v18 at the control group (M=16.75). Thus, the score of v1/v3/v16/v18at ADHD was expected to be higher than the scores of v1/v3/v16/v18 at the control group.

Table 1.3
Independent Sample T-test for v1/v3/v16/v18between ADHD and Control

ADHD Sensitivity and Specificity Measures
Table 1
Classification table

                Based from table 1 we can see that the correct percentage of classification of No ADHD is 88.5% which is good since we it is greater than 50%. Moreover the classification of person with ADHD is 69.2% which is also fair. Based from table 1 we can generate table 2 to determine the specificity and sensitivity in our classification.
Table 2
Sensitivity and Specificity Analysis

                Based from table 2 we can see that of all those who have ADHD we correctly predict that 69% has ADHD and of all those who doesn’t have ADHD we correctly predict that 89% doesn’t have ADHD.

Graph 1
ROC  Curve

 

Area Under the Curve
Area	Std. Errora	Asymptotic Sig.b	Asymptotic 95% Confidence Interval
			Lower Bound	Upper Bound
.500	.052	1.000	.398	.602
The test result variable(s): VAR00006 has at least one tie between the positive actual state group and the negative actual state group. Statistics may be biased.
a. Under the nonparametric assumption
b. Null hypothesis: true area = 0.5

 

 

 

 

 

Based from graph 1 we can see that the area below the ROC curve is 0.5 which is fair enough. We can also see that the significance value is 1.000 which means that the true area is really 0.5.