ALGOS - An algorithm to guide patients with Type 1 Diabetes treated with continuous subcutaneous insulin on how to respond to real-time continuous glucose levels©
Long-term complications such as kidney, eye and nerve damage can be minimized in people with type 1 diabetes through tight blood sugar (glucose) control, as well as non-smoking and good control of blood pressure and blood fats (cholesterol). In previous research studies tight blood glucose control has been facilitated through self-monitored home finger-prick glucose readings with frequent, weekly input by health care professionals. However, such intensive input is not always possible for health-care professionals or convenient for people living with diabetes. A disadvantage of tight glucose control is an increase in low blood glucose levels, also known as hypoglycaemia or ‘hypos’.
Currently insulin doses in most people with type 1 diabetes are determined by the person with diabetes and/or their carer in consultation (usually several times a year) with their diabetes care team, largely based on home finger-prick blood glucose readings. Such readings provide only single “snapshots”, of glucose readings at selected times, usually before meals and bed during the day, with post-meal and overnight readings being less frequent. It has been shown that the more frequent the readings, the better diabetes contol.
Continuous subcutaneous insulin infusion (CSII) pumps (insulin pumps) are becoming increasingly widely used by people with type 1 diabetes. Some insulin pumps can be linked to a ‘real time continuous glucose monitor (RT-CGM)’, which measures glucose levels in tissue fluid, which is usually a bit lower and later than the finger prick blood glucose level. This RT-CGM device includes a line which is inserted under the skin by the user or their carer and it measures glucose levels in tissue fluid utilizing the same chemistry as the finger-prick glucose sticks, and transmits a glucose level to the pump screen every five minutes. Audible and vibration alarms can be set to alert to high or low glucose levels, and arrows on the pump screen show if the tissue fluid glucose level is rapidly rising or falling. The glucose data is stored and can be uploaded to a computer or website and reviewed retrospectively. This technology is available in Australia, currently only via Medtronic, but there is currently no subsidy from the NDSS, other Government sources, or from private health insurance companies for the sensors, which last three days and cost about $75 each, and the transmitter, which costs about $1200.
At present insulin pumps, even if linked to a RT-CGM device, cannot directly determine rates of insulin delivery. Changes in insulin dosing must be initiated by the individual or their carer who interprets the glucose information provided by the RT-CGM. To facilitate appropriate use of the insulin pump and RT-CGM device we have designed an algorithm to guide pump and RT-CGM user responses to the observed tissue fluid glucose levels detected by the RT-CGM device. Following algorithm education, which takes about 90 minutes, pump users are provided with a written form of the algorithm along with a wallet card (and in the research study on a mobile phone handset) with guidelines on how to respond to glucose levels shown on the pump, as well as a wall chart which assists in making changes based on retrospective review of glucose level uploads.
The algorithm was developed and evaluated in a major investigator-initiated study, supported by Medtronic Australasia and the NHMRC funded Diabetes CCRE. Diabetes teams and sixty (adult and teenage) patients with type 1 diabetes from The University of Melbourne (St. Vincent’s), Monash Medical Centre, the Royal Childrens Hospital, Royal Melbourne Hospital and the Western Hospital were involved. All patients were experienced insulin pump users, and all were provided with the RT-CGM system for three or for six months.
Results of the recently completed study shows that for adults using the algorithm in addition to their insulin pump and the real-time glucose sensor data, their HbA1c improved significantly, compared with no change in HbA1c levels in non-algorithm users. In addition, after 16 weeks of RT-CGM use, the number of adults with the desired HbA1c of 7% or less was significantly greater in algorithm users than non-users. Even with the better glucose control there was no increased low glucose time. Adolescents did not significantly improve their glucose control with use of the glucose sensor.
These results were recently presented at major diabetes meetings in Europe, the USA and will be presented at the forthcoming Australian Diabetes Society meeting.
Investigators are making the algorithm freely available and will be available to discuss its use with diabetes care teams and patients who wish to use it. Insulin pump using patients should discuss their potential use of this algorithm with their diabetes care team so that treatment can be individualized.
If you wish to find out more about the algorithm, or to volunteer for evaluation studies, please contact us at enquiries@diabetesccre.unimelb.edu.au or (03) 9288 3646.