March 1, 1997

Comments on FDA Panel Meeting on Invasive and Non-invasive Home Glucose Meters On March 20th and 21st, 1997, the FDA convened a two-day meeting of its "Clinical Chemistry and Clinical Toxicology Devices Panel." The purpose of the meeting was stated as "Issues Regarding Self-monitoring Blood Glucose Systems (SMBG)." This Panel includes experts who do not work for the FDA; e.g. seven medical doctors, most of them directly associated with various aspects of diabetes care and management, three clinical chemists responsible for hospital laboratory analysis including blood glucose analysis, an industry representative as well as a consumer representative. At the start of the meeting, the FDA outlined the many objectives it was hoping that the panel could provide guidance on. These objectives included among others: Should the current FDA method of approving self- monitoring blood glucose systems (SMBG's) be changed? Why does the performance of SMBG's when tested by independent outsiders provide accuracy considerably worse than the accuracy the manufacturers claim in their 510(k) submittals to the FDA? Since SMBG's are used in three different settings (i.e. the home, the doctor's office, and in hospitals), should the FDA establish different performance requirements for SMBG's based upon where their intended use is? Should the FDA define specific performance requirements for SMBG's and if the answer is "yes", what should they be? Should non-invasive blood glucose testing systems be required to meet the same accuracy and performance requirements as the current generation of finger stick type of blood glucose meters? Unfortunately, the panel was rather overwhelmed by the number and scope of the FDA questions and did not provide definitive answers to most of the questions. However, during the panel's discussion, as well as during the presentations by manufacturers of glucose meters, a lot of interesting information was developed. This memo provides a summary of that information with particular emphasis on how it would impact on non-invasive blood glucose meters. Current Generation of Finger Stick SMBG's There are two types of patients that current SMBG's are used on: Stable Patients - These are patients primarily in the home or doctor's office situation where the diabetic is not under trauma or emergency conditions. Unstable Patients - These are patients in emergency rooms, and in hospitals. It was pointed out that the current SMBG's may not provide accurate results on unstable patients. The reasons for this were two-fold: The current meters do not provide accurate readings when hematocrit level is low. On stable patients, the hematocrit level is typically between 40 and 45 percent. However, on unstable patients the hematocrit level could be much lower. These low hematocrit levels may cause significant error in finger stick type units. Low Systolic Blood Pressure - The current generation finger stick meters may not measure correctly if systolic blood pressure is below 80 mmHg. Below this level, blood is primarily in the central part of the body and only a limited amount gets to the capillaries. This causes measurement errors. In addition to the above, the current generation of SMBG's also may have the additional accuracy limitations: At high altitude When humidity is high On neonates How Accurate Do SMBG's Need to Be? Considerable discussion was held on this important topic. It was pointed out that: * In 1986 the American Diabetic Association published the performance goals for SMBG's whereby the accuracy should be within 10% from 30 to 400 mg/dL 100% of the time (see Diabetes Care 1987, 10:95-99). * In 1993 the ADA restated the accuracy requirements to be "5% of lab value" on any new meter brought to market. They agreed that for current meters in the market the old ADA (1986) would remain in force. * In 1987, Clarke, Cox et al provided the unique "Error Grid Analysis" (Diabetes Care 1987, 10:622- 628). This Error Grid Analysis located acceptable error zones within which any meter error would not cause any change in the care of people with diabetes. (For example for glucose levels equal or greater than 70 mg/dL, acceptable for accuracy would be 20% of the laboratory glucose). The Error Grid Analysis also defined unacceptable zones where meter error in those zones would cause dangerous unacceptable management of the person with diabetes. * A recent independent test of the most modern home glucose meters (Trajanoski et al, Diabetes Care 1996, 19:1412-1415) showed that 20% of all readings did not provide results within +/- 15% of laboratory values. However, perhaps the most important point was made by Dr. Gutman of the FDA and by Dr. Ratner representing the Juvenile Diabetes Federation. They both pointed out that the Diabetes Complications in Care Trials (DCCT) which was a ten year long- term study on whether intense testing and treatment of diabetes would lower complications of diabetes, proved that intensive treatment significantly reduced the incidence of blindness, kidney loss and other terrible complications by approximately 70%. The important point made by both of these speakers was: * The DCCT used SMBG's that were one or two generations older than the current generation of finger stick meters. For example, these older meters required blotting the blood off the strips after 90 seconds. Those older meters provided less accuracy than current available meters. However, even with these limitations, the DCCT results clearly showed that major improvements in the health of diabetics would occur if they test often. The above comments by Dr. Gutman and Dr. Ratner clearly point out that you don't really need accuracy of 5% as recently recommended by the ADA in order to avoid the severe complications of diabetes. Clinical Versus Laboratory Considerable discussion was held on the use of SMBG's in hospitals and doctors' offices. However, for most people with diabetes, this discussion was not relevant. The laboratory setting where the desire for high accuracy exists is not the setting that most SMBG's are used. It is estimated that over 95% (perhaps 99%) of all SMBG's are used in the home. These are used in what is called the clinical setting; i.e. it is used to assist the person with diabetes in managing his disease. Thus, to force unreasonable stringent laboratory accuracy requirements on SMBG's that are designed for home use because some people also use them in a hospital laboratory setting, has serious economic and other implications. Non-invasive Blood Glucose Meters Biocontrols Corporation discussed what should be the accuracy requirement for first generation non-invasive blood glucose meters. They pointed out that it is unrealistic to believe that first generation non-invasive blood glucose meters must meet the same accuracy requirements as either the current finger stick units or the more stringent 1995 ADA recommendation. They pointed out that if such stringent requirements are demanded by the FDA, then non-invasive will not become a reality, at least in the foreseeable future. They feel this would do great harm to the diabetic community. Comment on the Meeting by Futrex President, Bob Rosenthal The panel discussed all these matters and many others in an intensive two-day session. Unfortunately, the FDA questions were so many and so broad, and of such significance that the panel was rather overwhelmed by them. Moreover, the panel was predominantly populated by hospital doctors as well as hospital laboratory experts. Their background tended to influence the discussions by always aiming at better accuracy. However, even with such difficulties, the panel discussions proved to be quite fruitful. In my opinion, perhaps the most important points brought out at the meeting were: 1.The DCCT showed the significant reductions in the severe complications of diabetes, and 2.These significant reductions occurred through use of older generation glucose meters that had poorer accuracy than the current generation of finger stick units. It seems to me (and of course I'm prejudiced) that first generation non-invasive meters would be an enormous benefit to the diabetic community if they could provide accuracy equivalent to the older generation finger stick units that were used during the ten year DCCT trials. Moreover, it is my opinion that the Clark, Cox "Error Grid Analysis", which identifies meter accuracy errors that have either significant or non-significant impact on the actual treatment of diabetes, should be intelligently used in the evaluation of non-invasive blood glucose meters. My last comment is that the ADA, which is indeed a most influential and important organization in the diabetic community, should recognize that by establishing very stringent accuracy goal ("5% of laboratory"), they may be inadvertently doing two major harms to those of us who have diabetes? The new 5% ADA accuracy goal may be stopping the possibility of low-cost finger stick units entering the market. Low-cost finger stick units would be of great benefit to the diabetic community provided they had the accuracy that was equivalent to what was obtainable during the DCCT test. All of us with diabetes primarily want to avoid the severe complications that result from our disease, and that is the accuracy that we should demand from our meters. If the "5%" accuracy becomes a rigid requirement for all types of new SMBG's, this may eliminate any non-invasive type of technique. Is it really in the best interest of the diabetic community to stop and strangle this potential major advance in the treatment of diabetes?

Home

E*Mail

Next...

---

©1997 Futrex, Inc. All Rights Reserved.