Continuous Glucose Monitoring for the iPhone Generation?
17 July 2017
Research reviews & expert opinions
Early CGM devices were expensive and had problems with accuracy and reliability, and patients often needed to provide finger-prick data to keep them. Modern CGMs are smaller and cheaper, but are they reliable?
Diabetes is a major contributor to the burden of global chronic disease. In the UK, over 3 million people have type 2 diabetes, and the prevalence is predicted to rise. Some patients can be treated with a combination of diet and exercise, others with oral medications, such as metformin. However, some patients suffer from type 1 or more severe forms of type 2 diabetes and have to inject insulin to maintain good glycaemic control. This can have mixed effects on patients’ quality of life. In one recent survey, 68% of patients reported that insulin treatment had a positive or neutral effect on their quality of life, although 60% of patients agreed that insulin treatment was “restrictive”. Not only do patients have to inject insulin, they should also perform regular finger pricks to check their blood glucose. NICE recommends that all patients with type 1 diabetes should test 4 times per day and in some cases up to 10 times per day.
Blood glucose testing is unpopular, and some patients report frustration and anxiety when trying to meet the requirements. Furthermore, the results are not always as helpful as they might be. A single test cannot distinguish between hypoglycaemia that is resolving and hypoglycaemia that is getting worse and requires treatment. Self-monitoring also fails to detect night-time hypoglycaemia, which is the cause of 6% of all deaths in young people (under 40) with type 1 diabetes.
One potential solution is continuous glucose monitoring. First marketed in the early 2000s, continuous glucose monitors (CGMs) measure the glucose concentration of interstitial fluid with a subcutaneous probe. This lags behind the blood glucose concentration, but only by around 7 minutes. Modern CGMs provide real-time data and can also produce graphs showing glucose concentrations over days, weeks and months, which can be used to guide treatment.
Many devices can download data to smartphones. The possible stated benefits of such CGMs are improved acceptability to patients and improved glycaemic control. In the future, they may even act as a first step in creating an “artificial pancreas” by automatically adjusting the dose of insulin supplied by an insulin pump in response to glucose concentrations. However, in order be more widely adopted, high-quality evidence of a clear benefit for CGMs versus finger prick testing is needed.
A recent systematic review and meta-analysis showed that in adults with type 1 diabetes CGMs provide a modest reduction in HbA1C – a marker of long-term glycaemic control. This reduction, if applied to the baseline HbA1c in adults in the trial would have led to a change from 66.1 to 63.1 mmol/mol (the UK target for type 1 diabetes is 48 mmol/mol)*. This conclusion has been supported by the results of two other trials (DIAMOND and GOLD) published since the review. As well as confirming a benefit in reducing HbA1c, the GOLD trialists investigated the quality of life using three different questionnaires. CGMs were associated with improved quality of life. DIAMOND showed that the CGM group spent less time in hypoglycaemia (blood glucose <3.9 mmol/L**) than the control group at 43 vs 80 minutes per day.
There are still problems. One is the lack of evidence that CGMs reduce HbA1C in children under 16. This may be because the devices are being used to adjust insulin doses, in order to reduce the frequency of severe attacks of hypoglycaemia. Severe hypoglycaemia provides a falsely reassuring HbA1c by lowering the average blood glucose concentration, but it is in fact very dangerous. Unfortunately, since most trials are designed with the statistical power to look for a reduction in HbA1C, there are no data to support this theory in children, even though DIAMOND showed that CGM reduced time spent in hypoglycaemia in adults.
Data are also lacking for elderly people. The mean age in the most recent systematic review was 42. Given the large number of elderly people with diabetes, this is not representative of the target population. Furthermore, low confidence in using new technologies and a lack of trust in their usefulness often act as a barrier to uptake in this group.
While the current evidence does show multiple benefits from CGMs, many problems remain unresolved. A reduction in HbA1c of only 3 mmol/mol compared with finger-pricks may not be clinically significant. Data are lacking for several key groups, including children and elderly people, and no systematic review has shown that these devices reduce the risk of hypoglycaemia. These difficulties are compounded by the rapid pace of change in the field. New, updated devices are released every year, and this affects the relevance of data obtained with earlier devices. In the meantime, it is vital that researchers and healthcare providers continue to monitor new developments and incorporate new evidence into guidelines and reviews as it becomes available
Philip Oddie is a sixth-year Medical Student who has recently completed a Special Study Module (SSM) with the CEBM.
*Values have been converted from the DCCT units used in the paper to the UK standard IFCC units
** Values have been converted from mg/dL to mmol/L