Orsense NON-INVASIVE meter wins CE approval

[rickst29]

This device is capable of continuous measurement, although the current version is a fairly large "box" connected to a device which compresses your finger while doing the measurement.

I'm GUESSING that this device has no usage of "disposable consumables". If that guess is right, then this is huge. But their accuracy data doesn't look all that great. The pretty graph on their website appears to be for only one person with exceptionally good results, it's got only about 100 data points. ) I've seen "pretty graphs" on websites for non-invasive bG measurement technology before, and the technology has never worked, the devices have never been approved or sold. But this one has CE approval, that's impressive and, frankly, implies that it actually works!

"The trial was conducted at the Sheba Medical center and in an outpatient clinic representing a home-like setting. The trial served to demonstrate the NBM-200G's reliability and patient compliance. Twenty eight subjects with diabetes type I or II participated in the trail, with 1 to 8 sessions of glucose measurements for up to 24 hours. In the 150 sessions collected in the trial, reference blood glucose ranged from 40 to 496 mg/dL. The Median Relative Absolute Error (MRAE) of the evaluated glucose was 10.1%, and the correlation was r = 0.84. A Clarke grid analysis yielded 95.3% of all points in the A+B zones (69.0% A, 26.4% B) and no points in the E zone.

An abstract containing data from clinical trails demonstrating the reliability of NBM-200G was accepted for a poster presentation (number 442-P) entitled "A Non-Invasive Prospective Continuous Glucose Monitoring System" by Amir et. al. at the 67th Scientific Session of the American Diabetes Association (ADA), taking place on June 22-26 in Chicago. In addition, clinical trails results have been accepted for publication at the 4th issue of Diabetes Science and Technology (July, 2007) in a paper entitled "Continuous Non-Invasive Glucose Monitoring Technology Based on "Occlusion Spectroscopy".

Without raw data, I can't exactly compare this MRAE value against a Coefficient of Variance. (The latter places much more penalty on readings which are far form the "true value", by using the SQUARE of the difference, and places much less penalty on readings which are "close".) But I'll guess that since the square root of 10.1 is 3.17, this probably isn't as good as high-quality finger-stick device (such as my One-Touch Ultra). And since you calibrate it using a finger-stick meter, not a YSI, the Orsense errors compound on top of whatever error the finger-stick reading had.

But if you can get CGMS without paying and arm and a leg for Sensors, or making your Insurance pay an arm and a leg for Sensors, this would be really great-- even if you do have to confirm the Orsense reading with a "regular" bG meter before proceeding with any actual treatment. (Heck, all of us CGMS users are SUPPOSED to do that already )

I have no idea how uncomfortable the "squeezing" action is, and whether it would wake you up every time it happens. This technology, unlike many failed non-invasive technologies, stops blood flow while subjecting your finger to the measurements. And, because you wearing the compression ring device in the same place for every test (that's a guess), it avoides a lot of the tissue variability which those other "technologies" could never handle.

I'm hoping to see more about this after the Chicago ADA presentation, and hope that one of our 'big shots' can attend the Session. (Jeff, Judith, Ellen....)

[Ellen]

You can write to Orsense and ask if they have published data besides this that was posted online at the ADA Scientific Sessions page:

The First Non-Invasive CGMS Profiler

Year: 2007
Abstract Number: 0442-P


Authors:
ORNA AMIR, OHAD COHEN, SILVIU ZILBERSTEIN, EFI GABIS, AVRAHAM KARASIK, Nes Ziona, Israel, Ramat Gan, Israel



Results:
There is a mounting body of work supporting the benefit of continuous glucose monitoring in improving the treatment and health of diabetic patients. Currently only invasive subcutaneous devices are used to perform such monitoring, limiting its use in the wider population. In this study, a fully non-invasive (NI) device for continuous blood glucose (BG) monitoring, the NBM (OrSense Ltd.) is examined. The NBM device, based on red near-infrared occlusion spectroscopy technology, is operated as a continuous glucose monitoring system (CGMS) profiler. Glucose predicting parameters were calibrated using four BG reference points from the first three hours of each session and an additional point taken after 8 hours.
A two phase trial was conducted in order to asses the performance of the new profiler calibration scheme. An initial learning phase was conducted at the Sheba Medical center and at a clinic, where subjects participated in 8 sessions lasting 24hr (5M, 2F, aged 22-41, 7DM1). Reference BG readings were collected every 15-30min. This phase served to improve the NBM software and calibration technique. In the second phase, the new software and calibration scheme were tested in an outpatient clinic representing a homelike setting. The second phase trial consisted of sessions lasting up to 14 hours for 3-7 days (14M, 10F, aged 21-75, 14DM1, 10DM2). Reference BG readings were collected every 30 minutes (Freestyle, Abbott). In both phases the non-invasive data was measured every 10 minutes and user compliance was positive.
The overall reference BG range from the 111 sessions, 1924 paired points, collected in the second phase trial was 40-496 mg/dL. The Mean Relative Absolute Error (MRAE) of the evaluated glucose was 15.9%, and the correlation was r = 0.9. A Clarke grid analysis yielded 96.9% of all points in the A+B zones (74.2% A, 22.7% B) and no points in the E zone.
This study shows a significant improvement in the NBM performance as a profiler, and indicates great promise for its use as a non-invasive CGMS. The NBM is currently under evaluation in broader clinical settings.



Category:
Clinical Therapeutics/New Technology - Glucose Monitoring and Sensing

[Ellen]

http://www.glucolight.com/

From Forbes.com

New Devices Shine a Light on Blood Sugar Control
06.23.07, 12:00 AM ET

SATURDAY, June 23 (HealthDay News) --Two new devices that use light to measure diabetics' blood sugar levels show promise, according to studies to be presented Saturday in Chicago at the annual meeting of the American Diabetes Association.
Dr. Robert Gabbay of the Penn State University Medical Center in Hershey, Pa., and colleagues tested an "optical coherence tomography device" -- GlucoLight -- on 27 type 1 and type 2 diabetics. The device, which focuses a beam of light on the patient's skin, was accurate in measuring blood sugar levels, the team found.
"The device maintained calibration over four days with no adverse side effects," Gabbay and his colleagues concluded. "Future efforts will evaluate the accuracy of the system in the hypoglycemic range."
Also at the meeting, a team led by Orna Amir, of OrSense Ltd. In Rehovot, Israel, was slated to present data on another device, the NBM-100, that uses red near-infrared light to measure blood sugar. Their study, involving seven diabetics, found the device also delivered accurate readings.

The abstract on Glucolight:

Optical Coherence Tomography-based Continuous Non-invasive Glucose Monitoring in Patients with Diabetes

Year: 2007
Abstract Number: 0405-P


Authors:
ROBERT GABBAY, SURENDRA SIVARAJAH, ROBERT WILKINS, MATTHEW SCHURMAN, Hershey, PA, Bethlehem, PA



Results:
The objective of this feasibility study is to test a second generation optical coherence tomography (OCT)-based system for non-invasive, sensitive, accurate and continuous monitoring of capillary (as opposed to interstitial) blood glucose concentration in subjects with diabetes, and to evaluate the maintenance of calibration over a four day period. OCT uses low coherence light with precise depth focusing ability to measure changes in scanned images in the microvasculature for glucose detection.
Methods: Thirty three subjects with diabetes had blood glucose concentrations evaluated using simultaneous capillary and OCT-based device (GlucoLight, Bethlehem, PA) measurements. Subjects received a 50 gram carbohydrate load on day 1 and were reevaluated on four subsequent days. Two point calibration was used on day 1 and one point calibration on subsequent days. Six subjects were excluded (two due to hardware and/or software failures at study initiation, two due to sensor connection problems, and two others due to problems with disposable device positioning on the skin). Of the 27 subjects included for analysis (12 type 1; 15 type 2), 13 received an additional 25-30 gram glucose load on Days 2 through 5 to create larger glucose excursions for analysis. Statistical analysis was performed using Clarke error grid, ISO standards, and Pearson correlation coefficient.
Results: Based on a total of 446 matched points, Clarke error grid analysis indicated 81% in category A, 19% in category B and [lt]1% in category C. ISO standards results showed 80% were within 20% for BG[gt]75mg/dL. No conclusions can be made about blood glucoses in the [lt] 75 mg/dL range since only 6 values occurred. The Pearson correlation coefficient comparing OCT vs. capillary blood glucose was 0.88 (95% CI 0.85-0.90, p [lt]0.0001). There were no significant changes in Pearson coefficient or ISO standards for BG[gt]75mg/dL from days one to five. There were 2 self-limited skin reactions but no adverse side effects at study conclusion.
Conclusion: This study demonstrated that the OCT sensor has a high degree of accuracy in the prediction of blood glucose levels in subjects with diabetes and was comparable to the commercially available Dexcom and Guardian devices in the normal and hyperglycemic range. The device maintained calibration over four days with no adverse side effects. Future efforts will evaluate the accuracy of the system in the hypoglycemic range.



Category:
Clinical Therapeutics/New Technology - Glucose Monitoring and Sensing

[rickst29]

Originally Posted by Ellen

The second phase trial consisted of sessions lasting up to 14 hours for 3-7 days (14M, 10F, aged 21-75, 14DM1, 10DM2). Reference BG readings were collected every 30 minutes (Freestyle, Abbott).

Freestyle isn't YSI, obviously. (This "reference" has significant noise of it's own.)

Originally Posted by Ellen

The overall reference BG range from the 111 sessions, 1924 paired points, collected in the second phase trial was 40-496 mg/dL. The Mean Relative Absolute Error (MRAE) of the evaluated glucose was 15.9%, and the correlation was r = 0.9. A Clarke grid analysis yielded 96.9% of all points in the A+B zones (74.2% A, 22.7% B) and no points in the E zone.
This study shows a significant improvement in the NBM performance as a profiler, and indicates great promise for its use as a non-invasive CGMS. The NBM is currently under evaluation in broader clinical settings.

Interesting that the version with "improved" software and calibration got readings with 60% HIGHER MRAE than the older data shown on their site And they're carefully avoiding mention of "C" versus "D" zone readings.

The other one, apparently a company-internal "feasibility study", impresses me no more than Glucoband. (That is to say, not at all, I'll wait for independent verification.)