This review concluded that the use of real-time continuous glucose monitoring effectively lowered HbA1c in people with type 1 diabetes compared with self monitoring of blood glucose. The conclusions of this well-conducted review are limited by the low quality of the evidence and lack of long-term follow-up.

To evaluate the use of a real-time continuous glucose monitoring system compared with self monitoring of blood glucose in patients with type 1 diabetes.

PubMed, EMBASE and Cochrane Central Register of Controlled Trials (CENTRAL) were searched from 1996 to March 2011. Search terms were reported and the reference lists of included papers and reviews were searched for additional studies.

Randomised controlled trials with parallel or crossover designs that compared real-time continuous glucose monitoring systems with self-monitoring of blood glucose in the management of type 1 diabetes were eligible. Studies that used commercially available real-time glucose monitors were included. Only studies with the same insulin regimen or with a similar proportion of patients who used continuous subcutaneous insulin infusion or multiple daily injections in both experimental and control groups were included. Study duration had to be at least three months with a follow-up rate of over 80%. Studies performed in pre-sugical, post-surgical or cardiac care units or in pregnant women were excluded. The primary outcome was the change in haemoglobin a1c (HbA1c), secondary outcomes were major and minor hypoglycaemic episodes (as defined by the investigators), mean daily area under the continuous glucose monitoring curve less than 3.89 mmol/l, mean daily area over the continuous glucose monitoring curve less than 9.99 mmol/l, local adverse effects and quality of life.

In most studies insulin pump therapy was used in both the experimental and control groups; in the rest the number of patients treated with continuous subcutaneous insulin infusion or multiple daily injections was comparable between groups. One study included a paediatric population only, one included adults only and the rest had mixed populations. Where reported the mean age in the study arms ranged from 8.5 to 44.6 years, one study reported an age range of 25 to 70 years.

Two independent reviewers selected the studies

Study quality was assessed using the following criteria: allocation concealment; blinding of participants, investigators, outcome assessors and data analysts; intention-to-treat analysis; and comprehensive follow-up (greater than 80% of participants included in the final analysis). Study quality was assessed by independent reviewers.

The mean difference between groups was calculated for continuous outcomes and the risk ratio for binary outcomes, both with 95% confidence intervals. Data were extracted by two independent reviewers with disagreements resolved by consensus or referral to a third reviewer.

Study results were pooled using a fixed-effect model, a random-effects model was used in cases of substantial heterogeneity. Statistical heterogeneity was assessed with Ι² and substantial heterogeneity was represented by Ι² of 50% or more. Children and adults, and levels of glycaemic control were compared in subgroup analyses.

Seven trials were included (948 patients). Randomisation methods were described and considered adequate in four trials, allocation concealment was adequate in two trials, two trials used intention-to-treat analyses and two trials reported withdrawals and drop-outs. Follow-up ranged from 87 to 98% of participants during a follow-up period ranging from three to 12 months.

HbA1c: Compared with self monitoring, real-time continuous glucose monitoring systems significantly reduced HbA1c both alone (mean difference (MD) -0.25; 95% CI: -0.34 to -0.17; seven trials) and when combined with an insulin pump (MD -0.26; 95% CI: -0.43 to -0.10; four trials). Heterogeneity was minimal in both analyses (Ι² of 0 and 4%). A statistically significant reduction in HbA1c was seen in subgroups with good and poor glycaemic control but there was no difference between adults and children.

Minor and major hypoglycaemic episodes: Minor hypoglycaemia was reported by five trials, one did not find any difference in hypoglycaemic episodes and the other four found no difference in the time spent in hypoglycaemia between real-time continuous glucose monitoring and self monitoring. There was also no difference in major hypoglycaemic events (six trials).

Results for other outcomes were reported in the paper: hyperglycaemia, mean amplitude of glycaemic excursions, ketoacidosis and local adverse events, sensor compliance and quality of life.

The use of real-time continuous glucose monitoring effectively lowered HbA1c in people with type 1 diabetes compared with self monitoring of blood glucose.

This review had a clear research question and specified inclusion criteria in enough detail to enable independent replication. Three relevant databases were searched but it was not reported if any language restrictions applied or if there were efforts made to locate unpublished research so the risk of language and publication bias could not be ruled out. Review methods were performed independently to prevent errors or bias. Trial quality was assessed and reported in full. The statistical methods were appropriate and heterogeneity was low for most analyses.

The conclusions of this well-conducted review are limited by the low quality of the evidence and lack of long-term follow-up.

Practice: The authors did not make any recommendations for practice.

Research: Further studies were needed to assess the effectiveness, safety, and tolerance of real-time continuous glucose monitoring devices in children and their impact on quality of life. Research that evaluated compliance with the devices was also needed.

The Medical University of Warsaw, Poland.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.