Sage Journals: Discover world-class research

Abstract

Self-monitoring of blood glucose (SMBG) is advocated as a valuable aid in the management of diabetes. The volume and cost of monitoring continues to increase. SMBG has a number of theoretical advantages/disadvantages which might impact on treatment, outcome and wellbeing. Investigating and quantifying the effect of self-monitoring in a condition where self-management plays a central role poses major methodological difficulties because of the need to minimize confounding factors. Despite the absence of definitive evidence, some situations where monitoring is generally accepted to be beneficial include patients on insulin, during pregnancy, in patients with hypoglycaemia unawareness and while driving. An area of controversy is the role of monitoring in non-insulin-requiring type-2 diabetes where observational and controlled studies give conflicting results. The available evidence does not support the general use of monitoring by all patients with type-2 diabetes, although further research is needed to identify specific subgroups of patients or specific situations where monitoring might be useful. The best use of SMBG in patients with type-2 diabetes might be for those receiving insulin and those on sulphonylurea drugs. The impact of monitoring on patient wellbeing must also be considered, with some studies suggesting adverse psychological effects. Given the large increase in the prevalence of type-2 diabetes, it will be important to define the role of SMBG so that resources can be used appropriately. Presently, the widespread use of SMBG (particularly in type-2 diabetes patients) is a good example of self-monitoring that was adopted in advance of robust evidence of its clinical efficacy.

Introduction

Self-monitoring of blood glucose (SMBG) is commonly performed by patients with diabetes and currently costs the United Kingdom National Health Service (NHS) in excess of £130 million per annum.¹ Much of the stimulus for the expansion in self-monitoring has been derived from improved technology on the one hand, which has rendered SMBG simpler and more reliable, and on the other hand, evidence from the large prospective studies such as Diabetes Control and Complications Trial (DCCT) and the United Kingdom Prospective Diabetes Study (UKPDS), which demonstrated the importance of good glycaemic control in type-1 and -2 diabetes patients, respectively.^2,3 However, despite the widespread adoption of SMBG, the evidence for its efficacy in contributing to improved outcomes in patients is unclear. This is particularly the case in non-insulin-requiring patients with type-2 diabetes mellitus, the prevalence of which is increasing rapidly and where expenditure on self-monitoring has the greatest potential to escalate.

This review will consider the evidence for the clinical effectiveness of self-monitoring rather than a comprehensive analytical appraisal of currently available monitoring technology. It will discuss the theoretical benefits and detriments of monitoring, the methodological problems encountered in studying the effects of SMBG on clinical outcomes and will consider the evidence for the role of monitoring in a range of clinical settings.

History

The earliest reference to home glucose-monitoring dates back to 1961 when Knight and Keen⁴ suggested that blood samples collected at home on filter paper and posted to the laboratory for analysis would allow assessment of blood sugar fluctuations and might be of value in improving diabetes control. In 1965 Ames developed and introduced the Dextrostix (Ames Company, Elkhart, ID, USA), the first strip for blood glucose measurement using an enzymatic glucose oxidase–peroxidase technique. A drop of blood was applied to the strip, wiped clean and following a period of incubation the colour was read against a series of standard colours to give a semi-quantitative glucose concentration. The first automated strip reader was the Ames Reflectance Meter (Ames Company, Elkhart, ID, USA, 1971) and shortly afterwards the improved Ames Eyetone was (Ames Company, Elkhart, ID, USA) introduced. Automated strip readers offered an advantage over visually read strips of allowing full quantitation and of eliminating visual reading errors. Such automated readers were initially targeted at health-care workers for use at the bedside or clinic but the potential for use by patients for monitoring glycaemia was soon realized and steadily increased from 1978 onwards.^5–7 In acknowledgement of the expanding interest in self-monitoring, the first consensus statement on SMBG by the American Diabetes Association was produced in 1987.⁸ This expansion in self-monitoring was facilitated by technology advances resulting in smaller, more portable metres that were easier to use and generated a more rapid result.

The role of self-monitoring in both type-1 and -2 diabetes patients was bolstered by the DCCT and UKPDS studies, both of which demonstrated improved microvascular outcomes with better glycaemic control. Furthermore, the design of both studies allowed for self-monitoring in the intensive treatment limbs as a tool for assessing glycaemia and adjusting insulin doses.^2,3 The volume of self-monitoring in the UK continues to increase; in England alone the cost of reagent strips for monitoring was £32.9 million for the fourth quarter of 2006 giving a projected annual cost in excess of £130 million.¹

Somewhat surprisingly for an intervention that is so widely practiced, objective evidence of the benefits of self-monitoring is unclear, particularly for patients with type-2 diabetes. The paucity of evidence reflects in part the availability and incremental adoption of technology in advance of robust studies assessing its clinical efficacy (as is the case with many laboratory tests) but also the considerable methodological difficulties encountered in quantifying the impact of monitoring on clinical outcomes.

Self-monitoring technology

Currently available glucose metres are based on either colorimetric or electrochemical principles.^9,10

First-generation metres used colorimetric detection with glucose oxidase as the recognition material. Hydrogen peroxide formed in the oxidation reaction converts a benzidene derivative to a coloured polymer that is quantified by reflectance spectrophotometry.⁹

Most glucose metres now use electrochemical detection. A drop of blood is applied to a disposable biosensor strip that incorporates glucose oxidase (as the recognition material) and an electron transfer mediator in a carbon paste binder. The mediator, which is usually ferrocene or quinine based, serves to shuttle electrons from the reduced enzyme to the electrode which generates an amperometric signal.⁹

Metres may be either plasma or whole-blood calibrated.^10,11 Important advances have taken place in recent years to improve the ease of use of metres and to reduce the possibility of operator error.^12,13 These include easier calibration through insertion of code or chip, smaller blood volume requirement (0.6–5 μL) and reduced measurement time from 5 to 30 s, improved memory capacity to store several hundred results with a facility for downloading, the generation of basic statistics, e.g. average glucose concentration. Safety features include error message if short sample and lock out if quality control is not performed. All these developments together with a reduction in size have made it easier for patients to generate accurate results.

Currently available metres have a typical working range of 0.6–22 mmol/L, with some reading as high as 33.3 mmol/L.¹⁴ The operating haematocrit range varies between metres and is typically in the range of 20–60%, although extreme haematocrits are rarely an issue in home monitoring. Individual metres may experience analytical interference with maltose, icodextrin (a component of peritoneal dialysis fluids) or oligosaccharides (a component of human immunoglobulin preparations).¹⁵

The analytical performance goals for glucose metres remain unclear in large part because readings are used for different purposes each of which may have different analytical requirements, e.g. the titration of insulin doses, detection of hypoglycaemia. Imprecision remains a particular problem. The total and within strip lot coefficients of variation (CV) for commonly used metres may range from 3.1% to 11.3 and 2.1% to 8.5%, respectively, with imprecision being generally higher at the lower end of the range, i.e. at the clinically important interface between euglycaemia and hypoglycaemia.¹³ When used for the selection of insulin doses, simulation modelling suggests that to administer the intended dose >95% of the time, the measurement bias and CV needs to be <2%.¹⁶ The American Diabetes Association has proposed a performance goal of <5% but this is met by few if any metres.^13,17

Alternative site monitoring

SMBG is performed on capillary blood samples generally obtained from a finger tip puncture, although alternative sites may be used. The finger tip offers an easily accessible source of capillary blood due its rich arteriovenous anastomoses. However, as it is highly innervated, lancing may be painful with fingertips becoming sore with frequent testing. This may be a particular problem for patients who perform manual work. Consequently, there has been interest in the use of alternative sampling sites (earlobe, forearm, abdomen, thigh and calf) since these sites are less innervated and therefore sampling may be less painful.^18,19 The major drawbacks are that alternative sampling sites may be less well vascularized (reduced network of arteriovenous anastomoses), yielding a smaller blood sample volume and resulting in a lag time of up to 30 min with respect to the finger capillary blood glucose and venous blood glucose fluctuations, most apparent when the blood glucose concentration is changing rapidly.²⁰ The smaller blood volume requirements of currently available metres make alternative site testing more feasible but the lag time problem renders it unsuitable and precludes its use for the detection of hypoglycaemia and therefore in patients with hypoglycaemia unawareness. It may, however, have a role as an adjunct to fingertip sampling in measuring glycaemia at fixed time points, e.g. fasting or postprandial.¹⁸

Continuous glucose sensors

Blood glucose levels may fluctuate markedly in diabetes. Patients who self-monitor typically measure a capillary glucose at one or more set time points during the day (e.g. fasting, preprandial, postprandial) or when they experience symptoms of hypo- or hyper-glycaemia. Even if performed very frequently it is recognized that clinically significant fluctuations in glucose concentration may be missed.²¹ Accordingly, there has been considerable interest in the development of continuous glucose monitoring (CGM) systems. It is beyond the scope of this article to review CGM in detail but we comment here briefly on the potential of this technology to supplement and eventually replace SMBG using intermittent capillary sample testing.

CGM systems may be invasive or non-invasive. Most invasive systems use a glucose oxidase-impregnated needle-type electrochemical sensor that is placed subcutaneously (often in the anterior abdominal wall) and measures glucose concentration in the interstitial fluid, e.g. Guardian RTS (Medtronic Minimed, Northridge, CA, USA).²² (Other electrochemical needle-type CGM systems that are currently marketed are the Dexcom (DexCom Inc., San Diego, CA, USA) and the Abbott Freestyle Navigator (Abbott Diabetes Care, Alameda, CA, USA), and others are under development). This type of sensor is calibrated using capillary blood glucose measurements that are used to convert the sensor signal into ‘blood glucose’ values. Calibration is performed immediately after sensor insertion with recalibration necessary at certain time intervals afterwards. The dynamic relationship between capillary and interstitial blood glucose is complex and a lag time, particularly if blood glucose levels are changing rapidly, may introduce systematic calibration errors.²³

Microdialysis is an alternative CGM approach in current use (GlucoDay, A Menarini Diagnostics, Florence, Italy) where a probe containing a fine, hollow dialysis fibre is implanted subcutaneously.²⁴ This is perfused with isotonic fluid so that tissue glucose diffuses into the fibre and is then delivered to the outside of the body where the glucose concentration is measured by a conventional glucose oxidase-based biosensor. Microdialysis has not yet enjoyed as much clinical use as enzyme electrodes.

In contrast to these minimally invasive systems for CGM, non-invasive systems are still at the stage of research and development. A variety of spectrophotometric (near-infrared [NIR] light, fluorescence-bases), electrical impedance, photoacoustic, light scattering or iontophoretic techniques have been or continue to be researched, and present technically much greater challenges than for existing CGM.²⁵ For example, the GlucoWatch (Cygnis Inc., Redwood City, CA, USA) device was an early non-invasive technology which reached market and used iontophoresis to transport interstitial fluid glucose across the skin where it was measured on the surface by electrochemical detection at a glucose oxidase-based sensor.²⁶ Inconvenience to patients, largely due to problems with skin irritation, low glucose flux across the skin and consequent inaccuracies limited its use, and it has now been withdrawn.²⁶

NIR spectroscopy is the most researched non-invasive technology, and is based on the tissues being transparent to light at this wavelength range. With multivariate statistical techniques it is possible to derive models that relate NIR absorption of the tissues to blood glucose levels, but the precision is so far too poor for clinical use. Possible interferences that cause unpredictable responses are changing amounts of light scattering (which alters the optical path length), tissue hydration and overlapping absorption by non-glucose metabolites and by water and fat.²⁷

The role of CGM in standard clinical management is under evaluation. A particular application might be the detection of episodes of asymptomatic hypoglycaemia, particularly overnight.²⁸ A recent multicentre study has added evidence that glycaemic control might be improved by CGM, at least in some patients. In this report, 322 adults and children with type-1 diabetes were randomly allocated to CGM or conventional SMBG for 26 weeks.²⁹ At the end of the study, HbA1c was 0.5% less in the CGM group, but this applied only to those ≥25 years of age. Severe hypoglycaemia did not differ between the CGM and SMBG groups, though this complication was already at a low level.²⁹

Theoretical advantages and disadvantages of monitoring

SMBG might contribute to the better management of diabetes in a number of ways:

For insulin-treated patients, it might allow the rational adjustment of insulin doses in response to changing glucose concentrations. This might have particular importance for patients on variable multiple dose insulin injection regimens who can titrate the premeal insulin bolus dose against the capillary glucose concentration;

It offers the possibility of immediate and ongoing feedback to patients on their glycaemic control. This might serve to reinforce concordance with beneficial lifestyle behaviours by increasing patients’ understanding of the effect of diet and exercise on capillary blood sugars;

It might assist in the detection, correction and avoidance of hypoglycaemia. This would be particularly valuable for patients about to embark on activities where hypoglycaemia might be hazardous (e.g. driving) or for patients who have reduced hypoglycaemia warning signs. Even though hypoglycaemia is more common in insulin-treated patients, it is also a problem with sulphonylurea agents and is reported to occur in up to 25% of patients per year although severe hypoglycaemia is much less frequent;^30,31

The feedback provided by SMBG might contribute to increased patient empowerment and an improved sense of wellbeing and adaptation to a serious and lifelong condition.

SMBG provides information on glycaemia at specific time points, e.g. postprandial hyperglycaemia, and thus is a measure of glycaemic variability. It has been postulated that postprandial hyperglycaemia is an independent risk factor for cardiovascular diseases and that glycaemic variability contributes to oxidative stress.³² Furthermore, glycaemic variability might also be a predictor of microvascular disease.^33,34 However, this remains an area of investigation, conflicting evidence and controversy, and the specific clinical value of measuring postprandial hyperglycaemia or variability is unclear as yet.^34,35

Conversely, there are a number of potential disadvantages associated with monitoring:

SMBG is an invasive technique and some patients find finger pricking painful and uncomfortable;

Some patients may find the requirement to monitor regularly oppressive and a constant reminder of their condition. Furthermore the finding of a high capillary glucose concentration may cause anxiety and contribute to a negative experience of diabetes and its management;

SMBG is expensive and without clear evidence base for its clinical value it may not be cost-effective for health services.

If SMBG is to be valuable in the management of diabetes, it should either contribute to improved glycaemic control (and therefore improved long-term clinical outcomes with a reduced risk of diabetes-related complications) or improve patient wellbeing and experience of diabetes or both.

Methodological difficulties in studying the impact of SMBG

The treatment of diabetes is complex requiring multifactorial interventions that include lifestyle (diet, exercise and weight management) and pharmacological therapy (oral hypoglycaemic agents, incretin or insulin injection) often occurring against background of additional treatments for diabetes-related complications or other co-morbidities. Such interventions are demanding of patients and require a high level of motivation and self-management for successful implementation. Isolating the specific effect of SMBG against these background variables poses major challenges for the investigator and requires careful consideration of study design and the endpoints to be measured.

Study design

Observational studies compare endpoints in patients who have monitored with those who have not (controls), or compare outcomes in the same patients before and after monitoring. These designs introduce potential confounding factors in that patients who actively monitor may differ from control patients in displaying generally more beneficial self-management behaviour, e.g. more active concordance with lifestyle and pharmacological regimens. Before/after studies may suffer from time-related bias;

Randomized controlled trials (RCTs) are generally considered to provide the most robust assessment of a clinical intervention but are not without limitations when applied to the investigation of self-management interventions. For SMBG studies the control intervention might be urine dipstick testing or simply no monitoring. Patient blinding is impossible as there is no control intervention that is indistinguishable from monitoring. Allocation by randomization is potentially problematical if the outcome for individual patients differs from their preference, e.g. patients who are less enthusiastic about self-monitoring but who are randomized to a monitoring group might engage less energetically and this would tend to attenuate any measured effect of monitoring. One method of overcoming this is to allocate intervention on the basis of patient preference. However, as for observational studies such an allocation protocol would be confounded by the fact that patients who are generally active in self-management are more likely to opt for monitoring. Randomized controlled trials may be confounded if participants (monitoring and control) alter their behaviour because they are under observation (the so-called Hawthorne effect).³⁶

Patient inclusion/exclusion criteria: The putative benefits of monitoring may vary in different patient groups, e.g. patients on oral hypoglycaemic regimens, patients on fixed dose insulin regimens, patients on variable dose insulin regimens. It is important therefore for the validity of results that homogeneous groups are studied. Some studies have investigated heterogeneous groups of patients (including patients on diet-only, oral agents and insulin) yielding results that are difficult to interpret. Another important consideration for randomized clinical trials is whether to include patients who currently monitor or have monitored previously. Patients actively monitoring at the time of study recruitment have generally been excluded from study entry. It is likely that patients with established diabetes who are actively monitoring continue to do so because they find it useful while those who are not monitoring have stopped (or never started) because they did not consider it helpful. This introduces a selection bias by excluding those patients on whom monitoring might have had the greatest impact. Such problems can be overcome by recruiting only newly diagnosed patients all of whom are therefore monitoring naive.

Intervention: The intervention employed must be clearly defined and it is important that the study groups are closely matched in all aspects apart from the use of monitoring. In particular, the study design should ensure that control and monitoring groups undergo an identical education programme (other than training in monitoring) and that both groups have equal access to members of the diabetes health-care team for follow-up and support. It is, however, essential that monitoring patients are given advice on how to interpret and act on the glucose readings that have been obtained. This might include the need for dietary review, better to understand the effects of individual dietary components on glucose levels, altered exercise patterns in modifying glucose levels, the identification and treatment of hypoglycaemia and the alteration of insulin dosages and timings.

Endpoints

Studies have considered a number of endpoints:

1. Diabetes-related morbidity and mortality

These include the incidence of microvascular complications (retinopathy, nephropathy), macrovascular complications (myocardial infarction, revascularization procedures, amputation, stroke) and total mortality. These represent the most definitive endpoints that are of the greatest interest to diabetes professionals and patients. However, in practice they have been employed in trials of SMBG only in observational studies and are likely to be unfeasible in prospective controlled trials because of the requirement for a long follow-up period and the need for a sufficiently large number of participants to provide adequate statistical power. Given the well-defined relationship between glycaemic control and microvascular outcomes, HbA1c is a more accessible surrogate outcome marker.

2. Glycaemic control

HbA1c is the primary marker of long-term glycaemic control and constitutes a treatment target that relates to the validated risk of diabetes-specific microvascular complications.^2,3 Other indices of glycaemic control such as postprandial hyperglycaemia may be important both as treatment targets and as indices of risk of complications (as discussed above) but this is less well validated and has not yet been incorporated into standard clinical guidelines.

3. Body mass index (BMI)

Type-2 diabetes is associated with obesity/overweight and many of the antidiabetic treatments may be associated with weight gain (sulphonylureas, glitazones, insulin). Obesity/overweight is associated with worse clinical outcomes and in type-2 diabetes leads to a more rapid progression to insulin requirement. Any effect of SMBG in fostering greater compliance with diet and exercise regimens and improved glycaemic control might therefore have an impact on BMI.

4. Pharmacological therapy

An effect of SMBG on concordance with diet and lifestyle regimens might reduce the requirement for pharmacological intervention (oral hypoglycaemic drugs, insulin, incretins), i.e. the number or dose of agents required to attain acceptable glycaemic control.

5. Hypoglycaemia frequency

Hypoglycaemia is an unpleasant and dangerous side-effect of antidiabetic agents and its avoidance is an integral element of good glycaemic control. Any role of SMBG in reducing hypoglycaemia incidence or severity would be relevant.

6. Psychological indices

These include measures of quality of life, wellbeing, treatment satisfaction, psychological state (anxiety/depression) and attitudes to diabetes and its treatment, which might be altered by monitoring.

Evidence of the clinical impact of monitoring

Overall definitive evidence of any beneficial effect of monitoring in terms of either improved glycaemic control or patient wellbeing has proved elusive. For some clinical applications it is impossible or unethical to undertake clinical trials to assess formally the effect of monitoring. There is particular interest in defining the role of monitoring in type-2 diabetes due to the huge increase in its prevalence and the associated expansion in monitoring. A range of specific clinical situations will now be considered.

Insulin-requiring diabetes

Insulin treatment is required for patients with type-1 diabetes or in type-2 diabetes where the use of oral hypoglycaemic agents has not resulted in adequate glycaemic control. The particular insulin regimen chosen will be tailored towards the requirements of individual patients. Commonly used regimens include once-daily injections of long-acting insulin, twice-daily premixed insulins (a mixture of short- and intermediate-acting insulin) injections, basal/bolus regimens (a once- or twice-daily injection of long-acting insulin with 3 or more injections of short-acting insulin premeals) or continuous subcutaneous insulin infusion (CSII).

An important element of more intensive insulin regimens is the opportunity it offers patients to adjust insulin doses taking into account the blood glucose concentration and anticipated food intake or energy expenditure. The DCCT study, which compared intensive versus conventional treatment of type-1 diabetes, incorporated monitoring into the intensive treatment limb (multiple insulin injections or CSII). Intensive treatment was associated with a 50–76% reduction in microvascular endpoints.² The study design did not allow assessment of the specific effect of monitoring. Indeed, given that monitoring (and therefore dose adjustment) is an integral component of intensive insulin regimens it is virtually impossible to design an ethically acceptable prospective study that would isolate and quantify the specific effect of monitoring. Instead, we must look to evidence from observational studies.

A number of observational studies have shown an association between monitoring and improved glycaemic control. In a subset of 1159 patients with type-1 diabetes in the North California Kaiser Permanente Diabetes Registry, a monitoring regimen of >3 times per day as compared with a lesser frequency was associated with a lower mean HbA1c (7.7% versus 8.7%) even after the correction for a range of possible confounding factors including age, gender, educational attainment, number of daily insulin injections, clinic non-attendance rates, self-reported exercise and diet.³⁷ These findings of a beneficial effect of SMBG were supported in a study on a large German/Austrian cohort of patients where each additional SMBG reading per day was associated with an additional 0.26% reduction in HbA1c for patients with type-1 diabetes.³⁸ For the subset of patients on more intensive insulin regimens (CSII or four or more insulin injections daily) each additional SMBG reading was associated with a further 0.32% reduction in HbA1c. The effect was smaller for those patients on less intensive regimens.

The Diabetes Outcomes in Veterans Study (DOVES) investigated the impact of monitoring in a cohort of 201 patients with insulin-treated type-2 diabetes who did not self-titrate their insulin dose.³⁹ This was an open label observational study (i.e. no control group). At eight weeks, the HbA1c had reduced from 8.1 to 7.7% (P < 0.001) and the authors concluded that monitoring provided a strong stimulus for improved self-care.

Coster et al. ⁴⁰ performed a meta-analysis of five RCTs comparing SMBG with urine monitoring in type-1 diabetes patients and found an overall difference in HbA1c of 0.57% favouring SMBG (95% confidence interval −1.07 to −0.06). But there are many problems in interpreting this analysis and combining the results, including the low statistical power in most of the studies, poor reporting in several, the wide variation in the timing and frequency of the testing (from twice weekly to three times daily) and the wide variation in the advice given on action according to the SMBG results (no advice in one study).

What can we conclude from these studies on insulin-treated patients? The finding of an association between SMBG and improved glycaemic control appears consistent. However, in the observational studies it is difficult to eliminate confounding factors such as individual patient motivation and the possibility that a high frequency of SMBG is merely a surrogate marker for better compliance with intensive management and adoption of a healthier lifestyle (and thus related to better HbA1c). There are also insufficient RCTs of SMBG in type-1 diabetes of good quality to perform meaningful meta-analysis, and we must therefore come to the surprising conclusion that although SMBG is rightly regarded as an integral component of modern management of type-1 diabetes and common sense indicates that it should be continued (in our view), the evidence base for this position is, as yet, weak.

Hypoglycaemia and driving

The incidence of severe hypoglycaemia (i.e. requiring third party intervention) in insulin-treated patients is 0.2–1.7 episodes per year.⁴¹ For type-2 diabetes patients on sulphonylureas the self-reported rates of mild and severe hypoglycaemia are 39% and 7%, respectively.⁴² Hypoglycaemia is one of the side effects of insulin treatment most feared by patients and may limit the aggressiveness of treatment regimens.⁴³ Furthermore, up to 25% of type-1 diabetes patients have lost hypoglycaemic warning signs.⁴⁴ This puts patients at significant risk of severe hypoglycaemia and poses particular problems in situations where hypoglycaemia might be dangerous, e.g. driving. It has been demonstrated that driving skills are impaired even at modest hypoglycaemia (blood glucose 3.3–4 mmol/L) and deteriorate markedly at lower levels.^45,46 Drivers are advised to monitor glucose before driving and at two-hourly intervals during long journeys. Since cognitive impairment may persist for up to 30 min following the treatment of hypoglycaemia, driving should not be resumed until 40 min postattainment of euglycaemia.⁴⁵ It seems self-evident that safe driving requires regular monitoring. It would be unjustifiable from an ethical perspective to formally test this in a clinical trial.

Pregnancy

Pregnancy may occur against a background of pre-existing type-1 or -2 diabetes or may develop de novo in pregnancy (gestational diabetes). Clinical trial evidence points to the importance of good glycaemic control at the time of conception and throughout pregnancy in improving clinical outcomes for mother and baby.⁴⁷ Patients on prior treatment with oral hypoglycaemic agents will generally be transferred onto insulin therapy for the duration of the pregnancy while patients with gestational diabetes may be managed initially with dietary treatment but switched to insulin as necessary to attain optimal glycaemic control. HbA1c is too unresponsive a marker for guiding decisions on treatment modality or insulin dose titration. Instead women are given pre and postprandial glucose targets that may be used both by the diabetes team and women to guide decisions on insulin introduction and insulin dose titration.^48,49 Such information can only be obtained by self-monitoring (the alternative being serial venepuncture with laboratory analysis). To test this formally in a clinical trial of monitoring would appear unnecessary and unethical.

Non-insulin-requiring diabetes

The use of SMBG in non-insulin-requiring diabetes is the most controversial and difficult area but is of great importance given the huge increase in type-2 diabetes and the parallel escalation in the cost of monitoring in this group of patients. A particular problem has been the conflicting evidence generated by studies of varying quality and design.

Observational studies have generally reported an association between monitoring and improved glycaemic control or clinical outcomes. The German Retroelective Study ‘Self-monitoring of blood glucose and outcome in patients with type-2 diabetes (ROSSO) followed up 3268 patients (mean follow-up period 6.5 years) for a range of clinical endpoints including diabetes-related morbidity and all-cause mortality.⁵⁰ Of these, 1479 patients were designated as ‘monitoring’ (i.e. had performed SMBG for a minimum of 12 months). Monitoring was associated with reduced diabetes-related morbidity (7.2% versus 10.4%, P = 0.002) and all-cause mortality (2.7% versus 4.6%, P = 0.004), but interestingly the fasting blood glucose and HbA1c were higher throughout the observation period in the SMBG group compared with the non-SMBG group.

Karter et al. ⁵¹ undertook a longitudinal analysis of 24,162 patients treated with diet-only or oral agents identified from the Kaiser Permanente Northern California diabetes registry.⁵¹ Initiation of once daily monitoring was associated with 0.35% and 0.42% reduction in HbA1c in diet-only and tablet-treated patients, respectively. There was a dose–response relationship between monitoring frequency and HbA1c reduction up to approximately three readings per day. For those patients established on monitoring, a change in monitoring frequency resulted in improved glycaemic control only in tablet-treated patients (0.16% HbA1c reduction per additional strip per day).

In a cohort of 3567 Italian patients with type-2 diabetes, monitoring was not associated with improved glycaemic control in the 79% treated with diet or tablet therapy.⁵² Interestingly, the study also found increased distress, worry and depressive symptoms in patients who monitored more than once per day.

The Freemantle Diabetes Study reviewed longitudinal observational data on 1280 type-2 diabetes patients and found that SMBG was not associated with improved survival, although monitoring patients treated with diet and or tablets had a 79% increased risk of cardiovascular death.⁵³ The authors were uncertain whether the latter finding related to confounding, incomplete covariate adjustment, chance or was a true association.

As discussed above it is impossible to assign causation in observational studies because of the absence of randomization and the difficulty in eliminating confounding factors: monitoring is more likely to be initiated or intensified in patients with poor glycaemic control, patients who monitor frequently may exhibit greater concordance with other aspects of self-management, and patients who monitor may have greater input from the diabetes team.

While recognizing their limitations, many consider that randomized controlled trials may offer more robust evidence for the efficacy of SMBG in type-2 diabetes patients.^54–72 The Auto Surveillance Intervention Active (ASIA) study was a prospective randomized trial of monitoring versus non-monitoring in patients with established type-2 diabetes over a period of six months.⁶¹ It showed a 0.3% HbA1c benefit in favour of monitoring, with 57.8% of the patients in the SMBG group showing an improvement in HbA1c compared with 46.8% in the control group (P = 0.012). However, the dropout rate was high and there were significant improvements in glycaemic control in both monitoring and control groups, suggesting that pre-existing treatment had been suboptimal and therefore the generalizability of the results is uncertain.

The DiGEM study (Diabetes Glycaemic Education and Monitoring) was a well-designed, three-arm parallel group randomized trial.⁵⁹ Patients with established type-2 diabetes were randomized to non-monitoring (control) or to one of two monitoring groups (a less intensive monitoring group or a more intensive group with specific training in interpretation and application of results). At the end of the 12-month follow-up period, there was no difference in HbA1c (primary outcome measure) between the three groups, suggesting that monitoring even in association with specific education did not result in improved glycaemic control. However, patients in the DiGEM study were well controlled at study entry (mean HbA1c 7.41–7.53%), which would have reduced the sensitivity for detecting an effect of monitoring.

The ESMON (Efficacy of Self-monitoring) study was a randomized control trial of monitoring versus non-monitoring in a cohort of patients with newly diagnosed type-2 diabetes with the primary endpoints of glycaemic control (HbA1c) and psychological indices.⁷⁰ At 12-month follow-up, there was no effect of monitoring on glycaemic control, oral hypoglycaemic drug usage, hypoglycaemia or BMI. The patients in the monitoring group had a small (6%) increase in depression index scores with a trend towards increased anxiety.

A meta-analysis of six RCTs (which predated DiGEM and ESMON) concluded that monitoring was associated with a 0.39% reduction in HbA1c in favour of monitoring, which would translate into a 14% reduced risk of microvascular complications.⁷³ However, this study suffers from the same problems as all meta-analyses in this field: limited quality in four of the six trials, hypoglycaemia was reported in only one study, the subjects were clinically heterogeneous (baseline HbA1c, treatment, etc.) and the effect of co-interventions was not controlled in some (e.g. education only in the SMBG group). A previous meta-analysis of SMBG in type-2 diabetes patients found no effect of SMBG on HbA1c, but with similar problems of interpretation.⁷⁴

What are we to conclude from this conflicting evidence for non-insulin-requiring type-2 diabetes? Certainly, the discrepancies in findings underline the methodological difficulties in gathering robust evidence. Two well-designed and executed randomized control trials (DiGEM and ESMON) have failed to show any benefit for glycaemic control in patients with established and newly diagnosed type-2 diabetes, respectively. This suggests that the indiscriminate use of monitoring in type-2 diabetes patients is unjustified. Indeed, there is evidence that monitoring may be associated with patient detriment by increasing depression and anxiety. However, further research is needed to define whether SMBG has a role in particular situations or particular groups of patients with type-2 diabetes, e.g. in patients with poor control, younger patients or those on sulphonylurea drugs, who are susceptible to hypoglycaemia.

The psychological impact of monitoring

There is some evidence to suggest that monitoring is associated with adverse psychological consequences. This includes qualitative observational research, which reported that patients may see readings as a proxy measure of good and bad behaviour with women in particular chastising themselves over high readings.^75,76 The ESMON study in patients with newly diagnosed diabetes found that monitoring was associated with a small (6%) increase in depression index score with a trend towards increased anxiety.⁷⁰

Current guidelines

Expert clinical guidelines in general support the use of self-monitoring in both insulin-requiring and non-insulin-requiring patients.⁷⁷ It is interesting that the International Diabetes Federation Clinical Guidelines Taskforce specifically stipulates that for patients with type-2 diabetes ‘comprehensive care’ should include the option of self-monitoring while for patients treated with diet alone monitoring should be used to assess glucose excursions due to lifestyle changes and to monitor changes during intercurrent illness.⁷⁸ Diabetes UK takes a more fluid view and advocates dialogue between patients and the diabetes health-care team to ensure that individual patient needs are addressed and that patients who wish to monitor should be supported in so doing.⁷⁹

What then should the practising clinician, assimilating the evidence and the doubts outlined above, do about SMBG – which patients with diabetes would benefit and how often should it be done? A Multidisciplinary Consensus on SMBG was recently reported that tries to offer some guidance.⁸⁰ It is based on the notion that the frequency of SMBG is determined by three factors: the stability or predictability of the blood glucose levels (which is related to the type of diabetes and the treatment), the use patients make of the results and personal preference.

Patients with type-1 diabetes have very variable control within and between days and logically need frequent blood glucose testing to detect hypo- and hyperglycaemia (e.g. 4–6 times daily). Extra testing is necessary under some circumstances, including when feeling hypoglycaemic, for the patient with hypoglycaemic unawareness, when ill, when exercising, after consumption of alcohol, when driving (see above), during pregnancy and prepregnancy, when there is a change of insulin regimen and for intensive insulin regimens such as CSII. The ‘unempowered’ type-1 diabetes patient does not make any use of SMBG for changing his regimen himself, but the data are used by the physician to alter management, so patients may do tests for the two weeks prior to a clinic visit or at a low frequency at different times of the day each day.

Control in type-2 diabetes is more predictable on a day-to-day basis and therefore testing needs to be less frequent. In patients with type-2 diabetes, those on diet, metformin or glitazones very rarely suffer from hypoglycaemia and therefore SMBG is not needed; fasting and casual blood glucose and HbA1c are reasonable indicators of control in this group. Those on sulphonylureas are at risk of hypoglycaemia and need SMBG at infrequent intervals (e.g. once daily at different times during the week) and those on insulin have hypoglycaemia relatively frequently and therefore need SMBG testing the most often in patients with type-2 diabetes (e.g. 1–4 times/day depending on stability). Extra testing might be necessary in patients with type-2 diabetes for illness, when drugs such as steroids are introduced and when regular HbA1c measurement is not available.

It is uncertain to what extent the current practice follows guidelines of the type described, although there is evidence of variability in the use of SMBG.⁸¹ It is therefore unclear whether stricter adherence to current guidelines would be associated with a greater or lesser volume of monitoring.

Conclusions

The volume of self-monitoring in patients with diabetes is increasing with a parallel rise in cost. SMBG has a number of potential beneficial and adverse effects for patients with diabetes. Investigating and quantifying the clinical effect of monitoring in a condition where self-management plays a central role is very difficult and requires careful attention to the study design to minimize confounding factors. There are certain situations where monitoring is generally accepted to benefit patient care, although definitive evidence is difficult if not impossible to obtain: pregnancy, patients with hypoglycaemia unawareness, driving and patients on variable dose insulin regimens. The role of monitoring in patients with type-2 diabetes is more controversial with conflicting results from observational studies and randomized controlled trials. The currently available evidence does not support the indiscriminate use of monitoring in all patients with type-2 diabetes, although further research is needed to identify subgroups of patients or situations where monitoring might be of value. The psychological effects of monitoring on patient wellbeing require further exploration and must be considered for individual patients. The widespread use of SMBG (particularly in type-2 diabetes mellitus) is a good example of a monitoring test that was adopted in advance of robust evidence of its clinical efficacy.

Declarations

Conflicts of interest: MOK is a member of the Self-monitoring of Blood Glucose International Working Group and in the past has received speaker fees from Abbott, manufacturer of self-monitoring devices and support for clinical evaluation of self-monitoring from Johnson and Johnson, manufacturer of self-monitoring devices. JCP was a member of a Multidisciplinary Consensus on Blood Glucose Self-Monitoring, and has received speaker fees and/or consultancy payments from Roche, LifeScan, Medtronic and Bayer, manufacturers of blood glucose monitoring devices.

Funding: MOK is grateful to the N Ireland Research and Development Office for generous grant support. JCP is grateful to EPSRC, Diabetes Foundation and the Wellcome Trust for generous grant support.

Ethical approval: Not applicable.

Guarantor: MOK.

Contributorship: Both authors contributed equally to the review.

References

Prescribing Support Unit 2007: Report on Prescribing for Diabetes in Primary and Secondary Care in England (online). See: http://www.ic.nhs.uk/webfiles/Services/PSU/diabetes.pdf . (accessed 4 August 2008)

The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long term complications of insulin dependent diabetes mellitus. N Engl J Med 1993;329:977–86

Stratton

, Adler

, Neil

, Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. Br Med J 2000;321:405–12

Knight

, Keen

. Blood sugar by post. Br Med J 1961;1:1168

Sonksen

, Judd

, Lowy

. Home monitoring of blood glucose. Method for improving control. Lancet 1978;1:729–32

Walford

, Gale

, Allison SP: Self monitoring of blood glucose. Improvement of diabetic control. Lancet 1978;1:732–35

Danowski

, Ohlsen

, Fisher

, Sunder

. Parameters of good control in diabetes mellitus. Diabetes Care 1980;3:88–93

American Diabetes Association. Consensus Statement Self-monitoring of blood glucose. Diabetes Care 1987;10:95–9

Turner

APF

, Chen

, Piletsky

. In vitro diagnostics in diabetes: meeting the challenge. Clin Chem 1999;45:1596–1601

10.

Ervin

, Kiser

. Issues and implications in the selection of blood glucose monitoring technologies. Diabetes Technol Ther 1999;1:3–11

11.

Arabdijef

, Nichols

. Assessing glucose meter accuracy. Curr Med Res Opin 2006;22:2167–74

12.

Chan

JCN

, Wong

RYM

, Cheung

C-K

, Accuracy, precision and user-acceptability of self blood glucose monitoring machines. Diabetes Res Clin Prac 1997;36:91–104

13.

Kimberly

, Vesper

, Caudill

, Variability among five over-the-counter blood glucose monitors. Clin Chim Acta 2006;364:292–7

14.

Department of Health Blood Glucose Monitoring: know your systems. See: www.mhra.gov.uk

15.

Pavlicek

, Garzoni

, Urech

, Brandle

. Inaccurate self-monitoring of blood glucose readings in patients on chronic ambulatory peritoneal dialysis with icodextrin. Exp Clin End Diabetes 2006;114:124–6

16.

Boyd

, Bruns

. Quality specifications for glucose meters: assessment by simulation modelling errors in insulin dose. Clin Chem 2001;47:209–14

17.

American Diabetes Association. Self monitoring of blood glucose. Diabetes Care 1996;19(Suppl 1):S62–6

18.

Dufaitre-Patouraux

, Djemli

, Vague

. How and when to use an alternative site in self-monitoring of blood glucose. Diabetes Metab 2004;30:471–7

19.

Greenhalgh

, Bradshaw

, Hall

, Price

. Forearm blood glucose tesing in diabetes mellitus. Arch Dis Child 2002;89:516–8

20.

Jungenheim

, Koshinsky

. Glucose monitoring at the arm: risky delay of hypoglycaemia and hyperglycaemia detection. Diabetes Care 2002;25:956–60

21.

Diabetes Research in Children Network (DirecNet) Study Group. Continuous glucose monitoring in children with type 1 diabetes. J Paediatr 2007;151:388–93

22.

Mastrototaro

. The MiniMed Continuous Glucose Monitoring System (CGMS). J Paediatr Endocrinol Metab 1992;12(Suppl 3):751–8

23.

Lodwig

, Heinemann

. Continuous glucose monitoring with glucose sensors: calibration and assessment criteria. Diabetes Technol Ther 2003;5:572–86

24.

Maran

, Crepaldi

, Tiengo

, Continuous subcutaneous glucose monitoring in diabetic patients: a multicentre study. Diabetes Care 2002;25:347–52

25.

Tura

, Maran

, Pacini

. Non-invasive glucose monitoring: assessment of technologies and devices according to quantitative data. Diabetes Res Clin Prac 2006;77:16–40

26.

Tierney

, Tamada

, Potts

, The GlucoWatch biographer: a frequent automatic and non-invasive glucose monitor. Ann Med 2000;32:632–41

27.

Arnold

, Small

. Noninvasive glucose sensing. Anal Chem 2005;77:5429–39

28.

Tarvis

, Shoabi

. The public health impact of the MiniMed Continuous Glucose Monitoring System (CGMS) – an assessment of the literature. Diabetes Technol Ther 2004;6:518–22

29.

Juvenile Diabetes Foundation Continuous Glucose Monitoring Study Group. Continuous glucose monitoring and intensive treatment of type 1 diabetes. N Engl J Med 2008;359:1–13

30.

Burge

, Sood

, Sobhy

, Rassam

, Schade

. Sulphonylurea induced hypoglycaemia in type 2 diabetes: a review. Diabetes Obes Metab 1999;1:199–206

31.

United Kingdom Prospective Diabetes Study (UKPDS) Group. Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352:837–53

32.

Monnier

, Mas

, Ginet

, Activation of oxidative stress by acute fluctuations compared with sustained chronic hyperglycaemia in patients with type 2 diabetes. JAMA 2006;295:1681–7

33.

Hirsch

, Brownlee

. Should minimal blood glucose variability become the gold standard of glycemic control? J Diabetes Complications 2005;19:178–81

34.

Kilpatrick

, Rigby

, Atkin

. The effect of glucose variability on the risk of microvascular complications in type 1 diabetes. Diabetes Care 2006;29:1486–90

35.

Kilpatrick

, Rigby

, Atkin

. Variability in the relationship between mean plasma glucose and HbA1c: implications for the assessment of glycemic control. Clin Chem 2007;53:897–901

36.

Adair

. The Hawthorne effect: a reconsideration of the methodological artefact. J Appl Psychol 1984;69:334–5

37.

Karter

, Ackerson

, Darbinian

, Self monitoring of blood glucose levels and glycemic control: the Northern California Kaiser Permanente Diabetes registry. Am J Med 1996;100:157–63

38.

Schütt

, Kern

, Krause

, Is the frequency of self-monitoring of blood glucose related to long-term metabolic control? Multicentre analysis including 24500 patients from 191 centres in Germany and Austria. Exp Clin Endocrinol Diabetes 2006;114:384–8

39.

Murata

, Adam

, Shah

, Intensified blood glucose monitoring improves glycaemic control in stable, insulin-treated veterans with type 2 diabetes. The Diabetes Outcomes in Veterans Study (DOVES). Diabetes Care 2003;26:1759–63

40.

Coster

, Gulliford

, Seed

, Powrie

, Swaminathan

. Monitoring blood glucose in diabetes mellitus: a systematic review. Health Technol Assess 2000;4:1–93

41.

McLeod

, Hepburn

, Frier

. Frequency and severity of hypoglycaemia in insulin treated diabetic patients. Diabetes Med 1993;10:238–45

42.

Amiel

, Dixon

, Mann

, Jameson

. Hypoglycaemia in type 2 diabetes. Diabetes Med 2008;25:245–54

43.

Cryer

, Davis

, Shamoon

. Hypoglycaemia in diabetes. Diabetes Care 2003;26:1902–12

44.

Gerich

, Mokan

, Veneman

, Hypoglycaemia unawareness. Endocr Rev 1991;12:356–71

45.

Frier

. Hypoglycaemia and driving performance. Diabetes Care 2000;23:148–9

46.

Cox

, Gonder-Frederick

, Kovatchev

, Julian

, Clarke

. Progressive hypoglycaemia's impact on driving simulation performance: occurrence, awareness and correction. Diabetes Care 2000;23:163–70

47.

Jovanovic

, Druzin

, Peterson

. Effect of euglycaemia on the outcome of pregnancy in insulin dependent diabetic women as compared with normal control subjects. Am J Med 1981;71:921–7

48.

Homko

, Sivan

, Reece

. Is self monitoring of blood glucose necessary in the management of gestational diabetes mellitus? Diabetes Care 1998;21(Suppl 2):B118–22

49.

Jovanovic

, Peterson

, Saxena

, Sawood

, Saudek

. Feasibility of maintaining normal glucose profiles in insulin-dependent pregnant diabetic women. Am J Med 1980;68:105–12

50.

Martin

, Schneider

, Heineman

, for the ROSSO study group. Self-monitoring of blood glucose in type 2 diabetes and long term outcome: an epidemiological study. Diabetologia 2005;49:271–8

51.

Karter

, Parker

, Moffett

, Longitudinal study of new and prevalent use of self-monitoring of blood glucose. Diabetes Care 2006;29:1757–63

52.

Franciosoi

, Pellegrini

, De Beradis

, The impact of blood glucose self-monitoring on metabolic control and quality of life in type 2 diabetic patients: an urgent need for better educational strategies. Diabetes Care 2001;24:1870–7

53.

Davis

, Bruce

, Davis

TME

. Does self-monitoring of blood glucose improve outcome in type 2 diabetes? The Freemantle Diabetes Study. Diabetologia 2007;50:510–15

54.

Allen

, DeLong

, Feussner

. Impact of glucose self-monitoring on non-insulin treated patients with type 2 diabetes mellitus. Diabetes Care 1990;13:1044–50

55.

Brown

, Garcia

, Kouzekanani

, Hanis

. Culturally competent diabetes self-management education for Mexican Americans: the Starr County border health initiative. Diabetes Care 2002;25:259–68

56.

Cho

, Chang

, Kwon

, Long term effect of the Internet-based glucose monitoring system on HbA1c reduction and glucose stability: a 30 month follow-up study for diabetes management with a ubiquitous medical care system. Diabetes Care 2005;29:2625–31

57.

Davidson

, Castellanos

, Kain

, Duran

. The effect of self-monitoring of blood glucose concentrations on glycated haemoglobin levels in diabetic patients not taking insulin. Am J Med 2005;118:422–5

58.

Estey

, Tan

, Mann

. Follow-up intervention: its effect on compliance behaviour to a diabetes regimen. Diabetes Educ 1990;16:291–5

59.

Farmer

, Wade

, Goyder

, Impact of self monitoring of blood glucose in the management of patients with non-insulin treated diabetes: open parallel group randomised trial. Br Med J 2007;335:132–6

60.

Fontbonne

, Billault

, Acosta

, Is glucose self-monitoring beneficial in non-insulin treated patients? Results of a randomized comparative trial. Diabetes Metab 1989;15:255–60

61.

Guerci

, Drouin

, Grange

, for the ASIA group. Self-monitoring of blood glucose significantly improves metabolic control in patients with type 2 diabetes mellitus: the Auto-Surveillance Intervention Active (ASIA) study. Diabetes Metab 2003;29:587–94

62.

Jaber

, Halapy

, Fernet

, Tummalapalli

, Diwakaran

. Evaluation of a pharmaceutical care model on diabetes management. Ann Pharmacother 1996;30:238–43

63.

Johnson

, Majumdar

, Bowker

, Toth

, Edwards

. Self-monitoring in Type 2 diabetes: a randomised trial of reimbursement policy. Diabetes Med 2006;23:1427–51

64.

Jones

, Edwards

, Vallis

, Changes in diabetes self-care behaviors make a difference in glycemic control: the Diabetes Stages of Change (DiSC) study. Diabetes Care 2003;26:732–7

65.

Kibriya

, Ali

, Banik

, Khan

. Home monitoring of blood glucose (HMBG) in Type-2 diabetes mellitus in a developing country. Diabetes Res Clin Pract 1999;46:253–7

66.

Kwon

, Cho

, Kim

, Establishment of blood glucose monitoring system using the internet. Diabetes Care 2004;27:478–83

67.

Miles

, Everett

, Murphy

, Kerr

. Comparison of blood or urine testing by patients with newly diagnosed type diabetes: patient survey after randomised crossover trial. Br Med J 1997;315:348–9

68.

Moreland

, Volkening

, Lawlor

. Use of a blood glucose monitoring manual to enhance monitoring adherence in adults with diabetes: a randomized controlled trial. Arch Intern Med 2006;166:689–95

69.

Muchmore

, Springer

, Miller

. Self-monitoring of blood glucose in overweight type 2 diabetic patients. Acta Diabetol 1994;31:215–9

70.

O'Kane

, Bunting

, Copeland

, Coates

. Efficacy of self monitoring of blood glucose in patients with newly diagnosed type 2 diabetes (ESMON study): randomised controlled trial. Br Med J 2008;336:1174–7

71.

Schwedes

, Siebolds

, Mertes

. Meal-related structured self-monitoring of blood glucose: effect on diabetes control in non-insulin-treated type 2 diabetic patients. Diabetes Care 2002;25:1928–32

72.

Wing

, Epstein

, Nowalk

. Does self-monitoring of blood glucose levels improve dietary compliance of obese patients with type II diabetes? Am J Med 1986;81:830–6

73.

Welschen

LMC

, Bloemendal

, Nijpels

, Self-monitoring of blood glucose in patients with type 2 diabetes who are not using insulin. Diabetes Care 2005;28:1510–17

74.

Coster

, Gulliford

, Seed

, Powrie

, Swaminathan

. Self monitoring in Type 2 diabetes mellitus: a meta-analysis. Diab Med 2000;17:755–61

75.

Peel

, Douglas

, Lawton

. Self monitoring of blood glucose in type 2 diabetes: longitudinal qualitative study of patients’ perspectives. Br Med J 2007;335:493

76.

Peel

, Parry

, Douglas

, Lawton

. Blood glucose self-monitoring in non-insulin-treated type 2 diabetes: a qualitative study of patients’ perspectives. Br J Gen Pract 2004;54:183–8

77.

American Diabetes Association. Standards of Medical Care in Diabetes-2007. Diabetes Care 2007;30(Suppl 1):S4–41

78.

IDF Guidelines Taskforce. Global guidelines for Type 2 Diabetes: recommendations for standard, comprehensive and minimal care. Diabetes Med 2006;23:579–93

79.

Diabetes UK, 2006 Care recommendations. Self-monitoring of blood glucose (SMBG) (online). See: http://www.diabetes.org.uk/About_us/Our_Views/Care_recommendations/Self-monitoring_of_blood_glucose (accessed 4 August 2008)

80.

Owens

, Barnett

, Pickup

, Blood glucose self-monitoring in type 1 and type 2 diabetes: reaching a multidisciplinary consensus. Diabetes Primary Care 2004;6:8–16

81.

Farmer

, Neil

. Variations in glucose self-monitoring during oral hypoglycaemic therapy in primary care (letter). Diabetes Med 2004;22:511

Self-monitoring of blood glucose in diabetes: is it worth it?

Abstract

Introduction

History

Self-monitoring technology

Alternative site monitoring

Continuous glucose sensors

Theoretical advantages and disadvantages of monitoring

Methodological difficulties in studying the impact of SMBG

Study design

Endpoints

Evidence of the clinical impact of monitoring

Insulin-requiring diabetes

Hypoglycaemia and driving

Pregnancy

Non-insulin-requiring diabetes

The psychological impact of monitoring

Current guidelines

Conclusions

Declarations

References