Intended for healthcare professionals

Practice Therapeutics

Continuous glucose monitoring

BMJ 2023; 380 doi: (Published 03 March 2023) Cite this as: BMJ 2023;380:e072420
  1. Dana M Lewis, independent researcher1,
  2. Tamara K Oser, associate professor2,
  3. Benjamin J Wheeler, paediatric endocrinologist/head of child health3
  1. 1#OpenAPS, Seattle, WA, USA
  2. 2Department of Family Medicine, University of Colorado School of Medicine, Aurora, CO, USA
  3. 3Department of Women’s and Children’s Health, Dunedin School of Medicine, University of Otago, New Zealand
  1. Correspondence to DM Lewis Dana{at}

What you need to know

  • Continuous glucose monitoring (CGM) is the preferred glucose monitoring method in most recent guidelines (ADA/EASD/ISPAD) for people with insulin-requiring diabetes, but it is also suitable for all people with diabetes

  • Accuracy of CGM is close to that of finger stick blood glucose testing

  • Often, conscious or unconscious bias and concerns about technical or data literacy prevent providers from discussing CGM/intermittently scanned CGM with people with diabetes

  • CGM is costly, which leads to inequity in terms of patient access and research around the topic

A person in their 60s with type 2 diabetes presents to their primary care provider for a diabetes follow-up visit. They express frustration with an above-target glycated haemoglobin (HbA1c), their current glucose management, and the associated burden, which includes finger stick blood glucose testing and multiple daily injections of insulin. They ask if any additional tools are available to help them manage their glucose levels more effectively on a day-to-day basis.

Continuous glucose monitoring (CGM) can be an area of uncertainty for many primary care doctors, which can lead to lower prescribing rates. In this paper, we review the available evidence and guidance on CGM. Currently, this is mostly relevant to people with diabetes in higher resource settings, as use of diabetes devices is lower in resource limited settings, and these populations are poorly represented in previous CGM studies. However, in writing this article, we also advocate for increased worldwide effort to improve access to CGM for these populations, and to improve research in the field.

What is CGM?

Continuous glucose monitors are devices that offer an alternative to finger stick blood glucose testing in adults and children with any type of diabetes (age suitability may vary by product). Monitoring with CGM typically requires the person to wear a small, disposable subcutaneous sensor that measures interstitial glucose levels for 7-14 days or more. Glucose levels are recorded at frequent intervals (typically 1-5 minutes) transmitted to a receiver or smart device. Some systems have a reusable transmitter while others have a single-use disposable transmitter. CGM is increasingly being used also in automated insulin delivery systems, where an algorithm uses CGM data to adjust insulin dosing via a connected insulin pump.

CGM technology has improved in recent years and now includes:

  • Real-time CGM, which provides real-time alerts and transmits glucose data continuously

  • Intermittently scanned CGM (isCGM, previously known as “flash” glucose monitoring), which typically provides glucose data “on demand,” when scanned by a receiver or smart device—some earlier versions have no alarm capabilities

  • A longer term implantable CGM.1

Box 1 compares real-time CGM and isCGM.

Box 1

Comparison between real-time CGM and isCGM

The key difference is that real-time CGM makes data available continuously, compared with isCGM, which enables access only when scanned.4 Real-time data enable glucose threshold alarms (ie, alerts for impending or current hypoglycaemia and hyperglycaemia), following real-time glucose data, and integration with automated insulin delivery.

Systematic review and meta-analysis data show that real-time CGM can improve glycaemic control, and can lead to longer periods in the target glucose range (known as “time in range”), reducing HbA1c and reducing severe hypoglycaemia.67 Real world evidence suggests reduction in A1c also occurs with isCGM.8

Real-time CGM is often more expensive than isCGM.

  • (all comparisons to date have been against first generation isCGM)23


Historically, CGM was used as an adjunct to finger stick blood glucose testing. Now, most systems (intermittent and real-time) are more accurate and “non-adjunctive,” enabling treatment decisions without finger stick blood glucose confirmation if symptoms match glucose levels. However, patients still need access to blood glucose meters and test strips during CGM start-up (the first 30-120 minutes when glucose data are not available) and for when symptoms do not match CGM-reported glucose levels, so access to finger stick test strips and a blood glucose meter is still advised.

CGM data can be used (in real time or in reports that are generated and reviewed retrospectively) to assess changes in behaviour or medications and to facilitate shared decision making between people with diabetes and their healthcare providers.910

Most major diabetes organisations (American Diabetes Association, European Association for the Study of Diabetes, International Society for Pediatric and Adolescent Diabetes) recommend real-time CGM or isCGM as the preferred glucose monitoring method for people with diabetes taking insulin, and many national guidelines (including 2022 guidelines from the National Institute for Health and Care Excellence (NICE) in the UK, table 1) have CGM-related recommendations for people with type 1 diabetes11 and type 2 diabetes.5

Table 1

NICE guidelines on CGM for type 1 diabetes and type 2 diabetes (2022)

View this table:

One Cochrane systematic review covers CGM in type 1 diabetes; however, it is from 2012, is out of date with regard to currently available CGM technology, and doesn’t cover use of CGM for all forms of diabetes. More up-to-date evidence reviews and guidelines on real-time CGM and isCGM are available from the ADA’s Standards of Care,12 which are updated annually.

How well does CGM work?

Clinical studies show that CGM can benefit people with all types of diabetes, regardless of whether they are receiving single daily injections,13 multiple daily injections,14 are using insulin pumps,15 or are engaged with other types of diabetes therapies,16 and regardless of their previous glycaemic outcomes.17

Achieving time in range

A 2022 systematic review and meta-analysis of 22 English language randomised controlled trials looking at real-time CGM (n=15 adjunctive, n=5 non-adjunctive) and isCGM (n=2) for type 1 diabetes presents a large body of evidence of safety and efficacy for people with diabetes, resulting in improvements in time spent within target glucose ranges and reduced times spent above and below target ranges with both real-time CGM and isCGM.6

A1c improvement

In the 2022 meta-analysis, real-time CGM alone had a small but statistically significant improvement in A1c (mean difference (95% confidence interval) −0.22% −0.31 to −0.14, I2=79%) whereas isCGM showed no improvement in A1c,5 although real-world evidence suggests reduction in A1c also occurs with isCGM.8

Early initiation of CGM within one year of diagnosis also appears to have benefits—lower A1c levels and fewer diabetes-related emergency visits were seen up to 2.5 years later when compared with later initiation in a retrospective analysis of electronic health records of 396 people who had just received a diagnosis of type 1 diabetes (94% of whom were children).7

Quality of life

Surveys of people with type 118 and type 2 diabetes19 show CGM and isCGM are also associated with improved quality of life.1819

Data from CGM can be shared by the person with diabetes in real time, so that a friend or family member can “follow”20 or access their data, or receive alerts and alarms as permitted by the person with diabetes. Surveys of 330 parents of children with type 1 diabetes and 214 partners of adults with type 1 diabetes show data sharing itself can improve the quality of life for both people with diabetes and their loved ones.9

A multicentre, prospective observational real-world study evaluated data from 1913 adults with type 1 diabetes in Belgium between July 2016 and July 2018 and found that isCGM resulted in less work absenteeism as well as less severe hypoglycaemia.21

What are the harms?

Skin complications

Irritation and complications of the skin (degrees of allergic contact dermatitis) may occur from the use of adhesive on the skin in CGM/isCGM; however, skin barrier wipes, steroid sprays, or hydrocolloid bandages under the sensor can help with this,22 and several studies show that many individuals continue to use CGM even when experiencing skin complications.

In a 2019 systematic review,23 the overall rate of skin problems was one per eight weeks of CGM wear time, and most problems were mild (68.6%), while 19.8% were considered moderate and 1.5% severe. Rates of associated discontinuation were low.23 A randomised controlled trial study of isCGM for six months also reported low rates of discontinuation in response to repeated skin complications (one of 33 participants) and had an overall lower rate of skin problems (one per 18.1 weeks).24

Interfering substances

The subcutaneous wires of some early generation CGM sensors can interact12 with acetaminophen, hydroxyurea medications, dexamethasone, mannitol, and tetracycline antibiotics, resulting in less accurate glucose readings; and >500 mg/day doses of vitamin C may influence CGM readings. Assess interfering substances for each device.25

Other potential harms and challenges

Users may have difficulties inserting the CGM, discomfort wearing it, alarm fatigue, and a mismatch of expectations in terms of accuracy or performance. These challenges were found in a multi-country analysis (evaluating studies with adherence data from multiple European countries, Israel, and the US)26 and a subsequent large survey of 1503 adults with type 1 diabetes regarding barriers to diabetes device uptake.27

Tips for safer prescribing

Ensure patient access to data

CGM in which the glucose data is not viewable (“blinded”) or available to the person with diabetes (historically often used as “professional” CGM, in which the CGM is provided by the practice rather than owned by the patient) still occurs occasionally. Many providers perceive these systems to increase the risk of hypoglycaemia and harm,28 presumably because patient access to CGM data is shown to reduce the risk of hypoglycaemia. We recommend that, even for trial periods or short-term use of professional CGM, data are always made available to the person with diabetes, unless they indicate that they do not want the ability to view their data.

Assess available systems

Before recommending a particular CGM, review the latest versions available in your country or area for accuracy and performance. Bear in mind that some systematic reviews may have included studies about older CGM systems, which had different levels of accuracy.1229

Consider also any potentially interfering substances, which vary by CGM system.1229

If the person with diabetes does not like or benefit from alarms, then certain versions of isCGM may be a better fit than real-time CGM, and/or intermittent use of either isCGM or real-time CGM.1

How cost effective is CGM?

Depending on insurance or public health coverage, the short term costs of real-time CGM or isCGM may seem overwhelming to patients. However, CGMs, as well as improving glycaemic control and providing quality of life benefits, are considered a cost-effective strategy long term compared with finger stick blood glucose testing.303132

Cost may be a reason why CGM/isCGM is sometimes used intermittently rather than continuously—while not ideal, patients may still benefit from occasional, intermittent CGM/isCGM use.12

Ongoing efforts are required to increase access to CGM worldwide. Those who are likely to benefit most from CGM may be least likely to have access to the technologies, including people with higher A1c and lower time in target range who may be excluded from CGM for a variety of reasons, including provider biases.333435

How does CGM compare with finger stick blood glucose testing?

The accuracy of newer CGM systems is close to that of finger stick blood glucose testing,36 and most are approved for therapeutic medication dosing (non-adjunctive use), so that people with diabetes can take action without having to use finger stick blood glucose tests.

Additionally, CGM glucose trend arrows provide additional contextual information above and beyond the reliability of any single data point from finger stick blood glucose testing, enabling individuals to see past and recent glucose levels and observe recent trends.373839

Data on comparative effectiveness include:

  • A 12-month observational study in 519 adults with insulin-requiring diabetes (types 1 and 2) observed a 63% reduction in frequency of hypoglycaemia compared with finger stick blood glucose testing.38

  • The recent FLASH-UK study on second generation isCGM enrolled patients with higher A1c at baseline and showed a 0.5% reduction in A1c with isCGM use compared with finger stick blood glucose testing.40

Tips for patients

  • Reports from your CGM can show patterns in your glucose levels and help you identify times when your glucose is going out of range more often, even if it’s not something that happens every day. This can empower you to learn more about your body, how it works, and what diabetes management strategies work best for you.

  • Getting familiar with CGM takes some time.

  • When you start on CGM, prioritise areas you’d like to focus on first—eg, reducing or avoiding hypoglycaemia or post-meal hyperglycaemia.

  • To prevent alarm fatigue with CGM, try starting without alarms or set your alarms to go off at a higher level than the high end of your glucose target range. As you get more comfortable, reduce the alarm levels to lower levels.

  • Others can’t see your data without your permission. Think about who you would like to have remote access to your data—ie, a health practitioner/friend/relative—and why. This can be for ongoing support, help reviewing patterns, or for specific use cases like assisting in the case of hypoglycaemia overnight. Data sharing settings can be amended at any time.

Education into practice

  • What CGM systems are available to your patients, and how do you ensure they are offered to those who might benefit?

  • How might you describe the quality-of-life benefits of CGM and isCGM to a person with diabetes?

  • If a person with diabetes wanted to use CGM and expressed cost concerns, what recommendations would you offer?

  • What actions would you take if a person with diabetes asked you to review their CGM data?

How patients were involved in the creation of this article

The lead author of this guidance is a person with diabetes who has used six different CGM systems; another author is a caregiver to a spouse and a child with diabetes. Additionally, another person with diabetes who was newer to CGM and their caregiver both provided feedback on this article, such that we changed the order of questions in the “Education into practice” box to prioritise quality of life above glycaemic outcome. In the box “Tips for patients,” we framed CGM as a tool for empowering people with diabetes to learn more about their bodies and what works for them, and further emphasised quality of life as what matters to patients in the section “How well does CGM work?” Another person with diabetes—an experienced CGM user who has used eight different systems—provided additional input on substances that may interfere with the functioning of different types of CGM.


  • This is one of a series of occasional articles on therapeutics for common or serious conditions, covering new drugs and old drugs with important new indications or concerns. To suggest a topic, please email us at

  • Advisers to this series are Robin Ferner and Patricia McGettigan.

  • Competing interests: The BMJ has judged that there are no disqualifying financial ties to commercial companies. The authors declare the following other interests: DL reports no relevant interests. TO reports that she has served on a Physician Advisory Board for Dexcom, served as a consultant for Cecilia Health, DiabetesWise, and received an investigator-initiated award from Abbott Diabetes. BW reports that he has received research funding from Medtronic, Dexcom, and iSENS.

  • The case(s) in this article are fictitious and therefore no consent was needed.

  • Provenance and peer review: commissioned, externally peer reviewed.


Log in

Log in through your institution


* For online subscription