Paper+

Self monitoring of glucose in type 2 diabetes: open parallel group randomised trial

"Impact of self monitoring of blood glucose in the management of patients with non-insulin treated diabetes: open parallel group randomised trial" by Andrew Farmer and colleagues (BMJ 2007;335:132, 25 Jun doi: 10.1136/bmj.39247.447431.BE).

Objective? To determine whether self monitoring, alone or with instruction in incorporating the results into self care, is more effective than usual care in improving glycaemic control in non-insulin treated patients with type 2 diabetes.

Design? Three arm, open, parallel group randomised trial.

Setting? 48 general practices in Oxfordshire and South Yorkshire.

Participants? 453 patients with non-insulin treated type 2 diabetes (mean age 65.7 years) for a median duration of three years and a mean haemoglobin A1c level of 7.5%.

Interventions? Standardised usual care with measurements of HbA1c every three months as the control group (n=152), blood glucose self monitoring with advice for patients to contact their doctor for interpretation of results, in addition to usual care (n=150), and blood glucose self monitoring with additional training of patients in interpretation and application of the results to enhance motivation and maintain adherence to a healthy lifestyle (n=151).

Main outcome measure? HbA1c concentration measured at 12 months.

Results? At 12 months the differences in HbA1c concentration between the three groups (adjusted for baseline HbA1c concentration) were not statistically significant (P=0.12). The difference in unadjusted mean change in HbA1c concentration from baseline to 12 months between the control and less intensive self monitoring groups was −0.14% (95% confidence interval −0.35% to 0.07%) and between the control and more intensive self monitoring groups was −0.17% (−0.37% to 0.03%).

Conclusions? Evidence is not convincing of an effect of self monitoring blood glucose, with or without instruction in incorporating findings into self care, in improving glycaemic control compared with usual care in reasonably well controlled non-insulin treated patients with type 2 diabetes.

Trial registration? Current Controlled Trials ISRCTN47464659.

The researchers wanted to know whether self monitoring of blood glucose concentrations in patients with type 2 diabetes in the community is worth it. They principally wanted to know whether a measure of long term control of blood glucose—glycated haemoglobin (HbA1c)—was reduced in patients who intensively monitored their blood glucose using a meter compared with patients who did not.

Essentially they found no statistically significant effect on the concentrations of HbA1c in patients using a blood glucose meter compared with patients who had no meter, over a one year period. The only variable seen to decrease over the year in patients using a blood glucose meter was total cholesterol.

The question of how we manage the increasing number of patients in the community with type 2 diabetes mellitus, so that their health is maintained to the best possible standards, is a pressing one. Technology allows patients to measure their blood glucose concentrations using a simple finger prick test and an electronic blood glucose meter. Patients with type 1, insulin dependent, diabetes often use blood glucose meters frequently and seem to achieve better control of their diabetes as a consequence. You might assume that it would be equally beneficial for patients with type 2 diabetes, but the evidence is scanty. Most data are from observational studies—that is, studies of cohort and case-control design, which do not involve randomisation. These cannot be considered first rate evidence. Evidence from a randomised controlled trial, usually considered the strongest experimental evidence possible, is, therefore, welcome.

The use of HbA1c as the primary outcome measure—that is, the principal objective of the study—is logical because good evidence shows that this is a valid measure of long term glycaemic control. Compelling evidence shows that good control of blood glucose concentrations correlates with decreased mortality and morbidity from diabetes.

The authors wanted to find out if the use of a blood glucose meter, with or without education about how to use the results of such tests in self care of their condition, would reduce HbA1c concentrations in patients after a year. They compared such patients with patients undergoing "usual care," who do not have a meter, but who had venous blood tests every three months. They also wanted to know whether patients measuring their own blood glucose had any effect on other important variables for patients with diabetes—for example, blood pressure, weight, and cholesterol concentrations (the secondary outcomes).

The authors designed a randomised controlled trial using three parallel arms:

A control group of patients receiving routine care, with blood tests every three months and who did not use a blood glucose meter

A group of patients who used a blood glucose meter but had no education in how to incorporate the results into their own care (the less intensive intervention group)

A group of patients who used blood glucose meters and also received education in how to use the results of their tests to achieve the best possible care of their diabetes (the intensive intervention group).

A parallel arm randomised controlled trial is one in which the control and intervention groups receive their treatments concurrently, after random allocation to any of the arms, so that baseline risk factors (known and unknown) are evenly distributed. Such a randomised controlled trial produces strong evidence, usually considered to be as good as it can be, provided the trial is well run. This design is well suited to the aims of this study. The authors used three arms to differentiate whether the effects of the intervention (the use of blood glucose meters) were related to more knowledge of the concentration of blood glucose or whether it was because patients were also educated as to how to use this knowledge in their own care.

Patients were recruited from 48 general practices in Oxfordshire and South Yorkshire over four years. The trial was strictly restricted to patients aged over 25 who were not taking insulin or using blood glucose meters at the time of recruitment.

To ensure that all patients entered the trial with a similar knowledge about diabetes and the effects of high blood glucose concentrations, they were all given a standardised education session designed to help changes in behaviour, and they were encouraged to review their diet and exercise levels, and to set goals for their self care. In this way the authors ensured that any effect on HbA1c that they measured should not have been related to differences in the groups of patients when they entered the trial, regarding basic knowledge of the influences of lifestyle on diabetes. All patients also had baseline measurements taken about their state of health—for example, blood tests, blood pressure, and so on. The matching of the groups seems to have been good at baseline, implying successful randomisation—an essential aspect of any study design such as this.

Patients were randomised to one of the three groups by minimisation, which is a system by which allocation is forced to achieve a balance according to known baseline factors that may influence the outcome. Obviously, the results of the randomisation should not be known by the investigators. This is known as single blinding and was achieved in this trial. The ideal—double blinding, in which the patients involved, and the investigators, do not know which intervention they are receiving—was not possible because of the overtly different nature of the interventions.

Patients in the control group simply had HbA1c measurements taken at three monthly intervals after the start of the trial. The results of these tests were normal.

Patients in the less intensive intervention were given a blood glucose meter and told to record readings for a minimum of two days each week. They were told to aim for a certain concentration of blood glucose and to contact their general practitioner if the readings fell outside a specified range. They were not told how to interpret the readings.

The patients in the intensive intervention group were given meters and training on how to interpret of the results and how to best use the results in their own care.

Any difference between the two intervention groups will indicate the effect of educating or not educating patients as to best use of the meters.

Both intervention groups kept diaries. General practitioners looking after the patients were told to change their diabetic treatment as necessary, in line with guidelines from the UK National Institute for Health and Clinical Excellence and according to their results.

The study includes a power calculation to estimate the numbers of patients needed to test the hypothesis, and is unlikely to be producing a false negative result. Also the statistical analysis of the results is clearly defined and seems appropriate. Importantly, the analysis was by intention to treat—that is, it included all patients who were randomised to one of the groups (whether they completed the trial or not) This ensures that the results bear a close resemblance to the results that might ensue from patients given this intervention in ordinary clinical settings (where many patients never follow treatment through to completion). Patients who did not complete a year of follow-up, for whatever reason, had HbA1c values carried forward from the last properly done measurement to fill in the missing values. Fortunately only 12.6% of patients were lost to follow-up, so this artificial technique is unlikely to have unduly influenced the result.

The study did not show a statistically significant difference of the HbA1c concentration between the groups after a year of follow-up. And most of the secondary outcomes examined showed no difference, with the exception of the total cholesterol concentration. In the group of patients with the intensive intervention the reduction in cholesterol was −0.40, whereas in the control group it was −0.16. This is a statistically significant difference (P=0.01), but whether this is of clinical importance is questionable. No difference was seen in changes to treatment (of drugs or insulin) during the study.

The results are robust and believable and seem to conclusively answer the main research question and point to the fact that using a blood glucose meter is of little value in this group of patients. Also the use of the meters did not seem to be popular: for example 72 of the 151 (48%) in the intensive intervention groups did not persist with home readings for the whole study period. But this is a selected group of patients who have relatively well controlled diabetes (HbA1c average at baseline about 7.5%), and it may well be that the real potential for improvement lies in the group of patients with poor control of their diabetes. However, such patients might not agree to test their blood glucose frequently at home; only 32% of the patients with diabetes approached in the practices involved in this study agreed to take part.

This is an important study, even though it shows a negative result, as it will be relevant to a large number of patients, who have reasonably controlled type 2 diabetes. Such patients should not be actively encouraged to use a blood glucose meter.

Competing interests: None declared.

1. Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. New Engl J Med 1993;329:977-86.
2. The UK Prospective Diabetes Study Group. Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes. Lancet 1998;352:837-53.

   

 Printable version    Download PDF    E-mail this to a friend    Respond to this article