CLINICAL INVESTIGATION
Background
Developing a low-blood-glucose alarm involves the technical development as well as the clinical trials testing the alarm on patients. The investigations we have conducted so far and future investigations making sure that the alarm functions as intended and is safe to use will be described in the following chapter.
Low blood sugar, called hypoglycaemia, is a frequent complication in insulin-treated diabetes. The body has several mechanisms that protect against the development of hypoglycaemia. You can sense when your blood sugar is about to get low as you will shiver, sweat and in general feel unwell, symptoms that all diabetic patients are familiar with. Also the body releases a number of hormones when the blood sugar is falling too low, the so called “counter regulatory hormones”. The most important among these are glucagon. Glucagon causes the liver to convert stored glucagon into glucose. Glucagon is the substance that occurs in GlucaGen® which is used for injection in cases of insulin shock. When you have suffered from diabetes for many years these mechanisms can, however, fail to function so that you are no longer able to feel the looming insulin shock at the same time the body no longer releases counter regulatory hormones to a sufficient extent. Thus there is a large risk of developing insulin shock.
Back in 1988 scientists from Hillerød Hospital proved that the brainwaves, the so called EEG (=electroencephalogram), change when the blood sugar gets low. Figure 1 shows how the EEG looks when the blood sugar is normal (A) and low (B). It is clear that the brainwaves both get larger and slower when the blood sugar is low. We wish to apply this knowledge to the development of a hypoglycaemia alarm.
An ordinary EEG device is, however, quite big and requires that you have electrodes situated several places on your head. For the alarm to be useful in praxis therefore requires that the electrodes do not bother you, that the EEG device must be very small and that you can wear it without any discomfort. In addition the EEG device has to be able to carry out a here-and-know analysis of the EEG and activate the alarm in time i.e. before the blood sugar gets critically low.
The first tests
The purpose of the first tests was to re-establish what was determined earlier. We discovered, exactly like the Hillerød scientists, that the EEG changed significantly during hypoglycaemia. Every second of the EEG was analyzed and these observations were used for the development of a computer programme that automatically analyzes the EEG. For the alarm to be useful the electrodes must be implanted under the skin of the head. This was tested by applying thin electrodes under the skin of the patients concurrently with the traditional electrodes placed on the head. The measuring showed that the signal from the electrodes under the skin was just as good as the signal from the regular electrodes.
Subsequently we completed tests in 15 patients with type 1 diabetes. The investigations were conducted at the department of neurophysiology, Odense University Hospital, where the patients were injected with insulin into a vein and thereby got hypoglycaemia. We continuously checked the well being of the patients by asking them to count and calculate. The purpose was to see if we could find any changes in the EEG already before the patients’ capabilities were affected by the low blood sugar. We ended the test when the patients no longer were able to count and calculate. At that point glucose infusion was given. The EEG was analyzed with the developed computer programme and the result showed changes in the EEG approx. 20 minutes in average prior to the patients being affected by hypoglycaemia. There appeared to be considerable variation, though. The blood glucose varied between 2 and 3.4 mmol/l at the time the EEG showed hypoglycaemia. In order to prove whether or not it was a “random” variation or whether each patient had his/her own blood glucose limit in relation to the changing EEG, several patients were examined one more time and we discovered that the variation from day to day in the same patient was much less than the variation between patients.
Tests involving patients eating their way to a normal blood sugar
One thing is that we were able to discover EEG changes before the patients were affected by the low blood sugar level. Another is whether the patients were able to react if they received an alarm signal when the EEG changed and if they were well enough to be able to eat and drink and thereby adjust the low blood sugar. This was tested in five tests at Sydvestjysk Hospital in Esbjerg.
All patients having problems feeling low blood sugar were given an insulin injection with an ordinary insulin pen at the clinic. In order to obtain low blood sugar the patients afterwards rode on an exercise bike. The EEG was recorded and analyzed during the entire test and when there were signs of low blood sugar the patient was alarmed. In advance they had been told that they should drink a glass of juice and eat a sandwich as soon as they heard the alarm. Four out of five patients were able to eat and drink without any difficulties at the time the alarm was activated. Several of them were not at all affected by the low blood sugar. The last patient did not react to the alarm at all. This patient was awake and able to think straight, though, so perhaps a better instruction would have caused him/her to react to the alarm.
Tests during the night with hypoglycaemia
It is utmost important that the alarm works at night, too, where it can be very uncomfortable for the patients to experience low blood sugar, in particular for the patients living alone. The normal EEG is, however, very different at night than during the day. When we sleep we go through different phases of sleep, NREM (non rapid eye movement) or deep sleep, REM (rapid eye movement) or dream sleep etc. The EEG varies depending on which sleep phases you find yourself in (figure 2) and some phases of sleep are characterized by slow waves like is evident during hypoglycaemia.
Accordingly we have run a number of tests at Sydvestjysk Hospital in Esbjerg where the patients spent the night at the hospital. Electrodes were applied under the patients’ skin and throughout the night the EEG was monitored. During insulin injection a low blood glucose level was induced and the alarm was activated when the EEG showed signs of hypoglycaemia. The patients should preferably wake up and be able to eat and drink what was at their disposal and thereby avoid the blood glucose to drop any further. The preliminary results show that all patients go through EEG changes, also during the night, like the changes we discovered during the day. Not all patients were able to wake up when the alarm set off with the current setting of the alarm. We do realize, however, that a more favourable setting of the computer programme, making the alarm go off earlier, presumably would have woken up the patients in due time. These studies are not complete yet.
The clinical pilot project
To determine whether or not the alarm works under day-to-day circumstances it of course has to be tested over a longer period of time with the target group of patients, namely, the patients who do not counter regulate to a sufficient extent and who are not able to sense low blood glucose. We will test this in a large study described below. Before we initiate the large study we will, however, complete a smaller pilot project where the alarm will be tested.
Where?
The study will take place at Sydvestjysk Hospital in Esbjerg.
Who can participate and who cannot participate?
In this study six patients with type 1 diabetes and with the inability to react to low blood glucose participate. This is either recorded by the patient scoring low on a scale where the ability to sense low blood glucose is measured or by receiving help in connection with low blood glucose within the past 12 months at least twice. Neither patients suffering from epilepsy or having had a stroke can participate, nor can patients with a severe heart or kidney disease.
What is going to happen?
The patients show up at the hospital where they will get examined to see if they are suitable for participating in the study. Afterwards, the patient is asked to go to a Plastic Surgical Clinic in Odense, where the alarm will be implanted. This is described below.
After approx. a week the stitches will be removed and the patient can start using the alarm. In order to make sure that the alarm functions as intended, and that the patient is able to react to the alarm, it will be tested at the laboratory. This takes place in terms of the patient getting an insulin injection that triggers an incidence of hypoglycaemia. During the study the patient has to measure his/her blood sugar levels on a daily basis and will also use a blood glucose monitor that measures the blood glucose all the time without showing the result. This leaves an opportunity to compare the different methods to demonstrate hypoglycaemia.
The next two months the patient has to attend the laboratory regularly in order to get the insulin dosage adjusted and where the case of hypoglycaemia will be registered.
After two months the alarm will be removed.
Further information about the study is described in the patient information sheet.
The clinical test
The clinical test has several purposes: We wish to test if patients, who do not react to hypoglycaemia, will experience fewer cases of severe hypoglycaemia or insulin shocks by using the alarm. We will test how well the alarm detects instances of hypoglycaemia, i.e. if there are any instances of hypoglycaemia that the alarm do not detect and situations of false alarm. Finally we will test if there are any problems involved with using the alarm.
Where?
This study will include 120 patients at five examination places in Denmark, namely Hillerød Hospital, Steno Diabetes Centre in Gentofte, Odense University Hospital, the Sygehus Lillebælt in Fredericia and Sydvestjysk Hospital in Esbjerg.
Who can participate and who cannot?
In this study 120 patients with type 1 diabetes and with the inability to react to low blood glucose will participate. This is either recorded by the patient scoring low on a scale where the ability to sense low blood glucose is measured or by receiving help in connection with low blood glucose within the past 12 months at least twice. Neither patients suffering from epilepsy or having had a stroke can participate, nor can patients with a severe heart or kidney disease.
What is going to happen?
The patients show up at the hospital where they will get examined to see if they are suitable for participating in the study. Afterwards, the patient is asked to go to a Plastic Surgical Clinic in Odense, where the alarm will be implanted. This is described below.
After approx. a week the stitches will be removed and the patient can start using the alarm. In order to make sure that the alarm functions as intended, and that the patient is able to react to the alarm, it will be tested at the laboratory. This takes place in terms of the patient getting an insulin injection that triggers an incidence of hypoglycaemia. During the study the patient has to measure his/her blood sugar levels on a daily basis.
The next six months the patient has to attend the laboratory regularly in order to get the insulin dosage adjusted and the case of hypoglycaemia will be registered. All instances of discomfort and potential side effects will also be registered.
After six months the alarm will be removed. This will also take place in Odense.
Further information about the study is described in the patient information sheet.
The operation
As it appears from the description [internal link] the alarm consists of an inner device that will be implanted under the skin on the head and an outer device that is attached like a “behind-the-ear-aid” that detects the signals from the inner device. The inner device is placed under the skin by a minor operation. The operation will be performed under local anaesthesia, where the anaesthesia will be injected into the skin. The alarm will be placed by the hair limit, but it will not be necessary to cut any hair off. During the operation the surgeon cuts an approx. 2 cm incision, forms a small “pocket” under the skin at the implantation site and subsequently lifts the electrode in place under the skin and then sutures the incision. The entire operation lasts for about 12 minutes. Right after the operation you will be able to drive a car or take the train home. We do not expect any significant discomforts to occur, but the patient is encouraged to keep an eye on the operation wound in order to prevent infection.
Relevant links for people participating in a clinical test
Rules and regulations regarding participation in the clinical tests are available on the website of The Committee on Research Ethics where it is possible to read the brochure ”Før du beslutter dig” (before you make up your mind). and where it is possible to read about the patients’ rights in connection with participation in clinical tests.