13.4.9.1 Glucose monitoring

Blood glucose concentrations are measured in diabetes to detect hyper- and hypo-glycaemia. Health care professionals need this information to diagnose diabetes, or states of impaired glucose tolerance, to adjust therapy and correct hyper- and hypo-glycaemia in established diabetes, to interpret signs and symptoms in patients (e.g. is confusion due to hypoglycaemia or another cause?), and to assess the risk of tissue complications developing in the future (the severity and duration of hyperglycaemia is clearly related to microvascular disease). The patient with diabetes measures blood glucose concentrations to take corrective action with food and insulin, to maintain good control, to check the safety of everyday activities (e.g. not driving when hypoglycaemic), to assess the impact of events and lifestyle and on control (exercise, diet, illness, psychological stress), and to ensure a good quality of life and the ‘peace of mind’ that knowledge of the blood glucose concentration gives.

Glucose monitoring has traditionally been performed by intermittent sampling of blood glucose concentrations, either in hospital or by the patient testing their own blood glucose concentrations at home using finger-prick capillary blood samples applied to reagent strips and inserted into portable glucose meters – self-monitoring of blood glucose (SMBG). In addition, in the last decade or so, continuous glucose monitoring (CGM) has entered clinical practice as a supplement to SMBG, albeit with limited uptake at present. CGM is based on the implantation of needle-type glucose sensors, or microdialysis probes, into the subcutaneous tissue for measurement of interstitial glucose concentrations.

The first strip for blood glucose testing was introduced in 1965; application to a strip containing immobilized glucose oxidase and peroxidase produced a colour change in a dye, which was compared semi-quantitatively with a colour chart in order to estimate the glucose concentration (1). Automated and more quantitative reading of the strips using a portable reflectance meter was introduced in the early 1970s, intended for bedside and clinic use by health care professionals. SMBG by patients themselves at home was introduced into clinical practice in 1978 (2, 3), and has now become an integral part of the modern management of type 1 diabetes and also, more controversially, is used by many with type 2 diabetes.

Most strips and meters are now based not on colour development but on the electrochemical detection of glucose and a current response, which is converted into a digital read out of the glucose concentration (4). A small molecular weight mediator (e.g. ferrocene, hexacyanoferrate, or a quinone) is used to shuttle electrons from glucose to an electrode, thereby producing a current, catalysed by an enzyme, such as glucose oxidase or glucose dehydrogenase (Fig. 13.4.9.1.1). Very small volumes of blood are now required, typically less than 1 µL, with the blood taken up into the strip by capillary fill. A reading can be obtained in about 5 seconds. Although a capillary whole blood sample is measured in SMBG, most modern meters are calibrated to produce a reading that is a ‘plasma equivalent’, based on the assumption of a normal haematocrit. Variations in haematocrit, however, do affect the result and can produce inaccuracies, e.g. in critically ill patients.

Traditional log book recording of SMBG tests is still in use by most patients, and is accepted and understood by most patients and health care professionals, although accuracy and completeness are variable. However, modern meters also have the capability to store between 150 and 500 test results for later display, and have onboard software that can perform simple statistics, such as average daily blood glucose levels and average at a certain time of day. Data can be downloaded to a personal computer for graphical display and more complete statistical analysis, including blood glucose averages, modal days, standard deviations, and percentage of values within, over, or under target.

A major barrier to compliance, and, thus, a high frequency of testing in SMBG, is the discomfort of obtaining finger-prick samples. Certain ‘alternative sites’, such as the forearm, upper arm, calf, thigh, and abdomen, though less well innervated and therefore less painful to obtain a blood sample from, are generally less vascular than the finger. Now that modern devices can operate with very small blood volumes, alternative site testing has become a more realistic option. In the steady state, there is a good correlation between capillary blood glucose concentrations measured at the fingers and at, say, the forearm (5), but when the blood glucose level is rapidly increasing or decreasing, delays of up to 30 min have been noted between the alternative site and the finger value (6). Though rubbing the skin at the alternative site may improve blood flow and reduce delays, it is probably not advisable to rely on testing at alternative sites to assess glycaemic fluctuations and detect developing hypoglycaemia.

The everyday measurement of serial blood glucose concentrations via SMBG offers the opportunity to assess control in a more detailed and quantitative way using numerical indices of the various elements of control: average, within-day and between-day variability; meal-related increases; number of hypoglycaemic episodes, etc. (see Box 13.4.9.1.1) (7, 8). Some of these measures are calculated by software after computer download of SMBG data. So far, most indices have found more research and physician interest than clinical use.

There is increasing interest in not only the mean blood glucose levels in patients, but also the glycaemic variability and its clinical importance. While there is continued debate and conflicting evidence on whether glycaemic variability is a risk factor for micro- and/or macro-vascular disease in diabetes (9–12), the clearest consequence of excessive variability is that it restricts the glycated haemoglobin (HbA_1c; see below) that can be achieved by patients. In subjects with type 1 diabetes receiving multiple daily injections (MDIs) of insulin, both within- and between-day blood glucose variability calculated from SMBG tests are significantly related to the HbA_1c on MDI (13) (Fig. 13.4.9.1.2a). Patients with wide and unpredictable swings in blood glucose during MDI are likely to maintain an elevated HbA_1c in order to avoid hypoglycaemia occurring as attempts are made to tighten control. This risk of hypoglycaemia with excessive glycaemic variability is confirmed by the significant relationship between the coefficient of variation of SMBG results, i.e. the standard deviation expressed as a percentage of the mean, and hypoglycaemia frequency, assessed by the percentage of SMBG tests less than 3.5 mmol/l (Fig. 13.4.9.1.2b).

The best indices of glycaemic variability, or ‘lability’, have not been agreed, and amongst the many that are being evaluated as measures of within-day lability are standard deviation (SD), range, coefficient of variation (CV), mean amplitude of glycaemic excursions (MAGE), average daily risk range (ADRR), and the lability index (8, 14, 15); and, for between-day variability, mean of the daily differences (MODD) and SD or interquartile range of fasting blood glucose values (8).

The popularity of SMBG was particularly accelerated after 1993 and the positive results of the Diabetes Control and Complications Trial (DCCT) (16), where SMBG was used as a component of intensified therapy by both MDI and continuous subcutaneous insulin infusion (CSII). However, the use of SMBG in most patients with diabetes is controversial because the study evidence for its effectiveness is so far inconclusive, the best use of SMBG is not agreed by health care professionals, and it is very expensive (17). More is spent on SMBG than on oral hypoglycaemic agents in many countries, with an annual cost of more than £130 million in the UK.

In type 2 diabetes, the majority of trials indicate either marginal or no effectiveness of SMBG (18, 19) over no monitoring, although trials have been notoriously difficult to interpret and incorporate into meta-analyses for many reasons. Studies have often included patients on different treatments, say oral agents, insulin, and diet (a determinant of testing frequency and appropriate use, see below), or failed to differentiate between types of oral agent (i.e. hypoglycaemia-causing agents versus those not associated with significant hypoglycaemia); many have low statistical power; many did not assess hypoglycaemia (a major reason for testing); and many did not give patients instructions on how often to perform SMBG or how to act on the results.

Evidence for the effectiveness of SMBG in type 2 diabetes includes the Kaiser Permante Diabetes Registry Study (20), a cross-sectional survey of more than 23 000 adults with diabetes who were part of a managed care programme. Automated pharmacy records and the redemption of prescriptions were used to calculate the average blood glucose strip usage. HbA_1c was lowest in those patients with type 2 diabetes who used the most strips, whether they were on diet alone, oral agents, or insulin, although it is uncertain whether frequent SMBG use is merely a marker for more intense management, compliance, and healthy lifestyle. The Rosso study (21) retrospectively reviewed more than 3000 people with type 2 diabetes in general practice over a 6.5-year period. The 45% who used SMBG had significantly lower mortality and micro- and macrovascular events, though their HbA_1c was slightly higher than that of the non-SMBG group. The same difficulties of interpretation apply.

Two well-designed, recent, randomized controlled trials show no benefit of monitoring in type 2 diabetes. In the DiGEM (Diabetes Glycaemic Education and Monitoring) study, people with established type 2 diabetes were allocated to non-monitoring, monitoring, or intensive monitoring, with training in interpretation and application of results (22). There was no difference in HbA_1c amongst the three groups after 12 months follow-up, though patients were relatively well controlled at start (HbA_1c, 7.4–7.5% (57–58 mmol/mol)). In the ESMON (Efficiency of Self-Monitoring) study, newly diagnosed subjects with type 2 diabetes were randomized to no monitoring or SMBG (23). At 12 months, there was no difference in HbA_1c, hypoglycaemia, body mass index, or oral hypoglycaemic drug usage between the groups, and those in the monitoring group had a small increase in depression index scores and a tendency towards more anxiety.

The reasons for, and likely benefits of, SMBG in type 1 diabetes seem more obvious than in type 2 diabetes: hypoglycaemia is more common and control more erratic so frequent monitoring will allow adjustments in insulin dosage, the avoidance of hypoglycaemia, and improved control. Surprisingly, though few doubt the central part that SMBG should and does play in the management of type 1 diabetes, the evidence for its effectiveness from formal randomized studies is weak, mostly because of the poor design and reporting. Cross-sectional studies generally favour lower HbA_1c levels in those testing frequently in type 1 diabetes, e.g. the Kaiser Permanente study mentioned above (20) and a similar study in Tayside, Scotland, of the number of prescribed strips redeemed (24) showed a high correlation in type 1 (but not type 2) diabetes between (presumed) strip usage and HbA_1c (r = 0.61, p <0.001).

With conflicting evidence for the effectiveness of SMBG, how should the practicing physician decide how often to test in a particular patient? Major determinants of blood glucose testing frequency are the predictability of the blood glucose concentration (the more unpredictable the control, the more frequent should be the SMBG); the use made of the data (patients who make no use of SMBG do not need to do perform it frequently) and personal preference (patients who dislike finger-prick testing will not do frequent SMBG) (25).

The predictability of blood glucose levels depends on the type of diabetes and the treatment. Because glycaemia usually varies markedly within and between days in type 1 diabetes, a single blood test gives little or no idea of the control, and blood tests need to done at several times during the day (8). In type 2 diabetes, glucose control is mostly stable and, although concentrations are elevated, they are fairly predictable from day to day. A single blood glucose measurement in type 2 diabetes, therefore, relates reasonably well to overall control and, indeed, correlates well with HbA_1c (26).

The treatment of patients with diabetes also determines the frequency of SMBG because those type 2 diabetic patients on diet alone, metformin, glitazones, or DPP-IV inhibitors have little or no hypoglycaemia (and, therefore, need no testing to detect it); those with type 1 and 2 diabetes on insulin suffer hypoglycaemia relatively frequently (and need frequent testing to detect it); and those on sulphonylureas have some risk of hypoglycaemia and, therefore, need some testing.

In type 1 diabetes, frequent SMBG is therefore desirable (Fig. 13.4.9.1.3 top panel), usually before, and sometimes after, meals and at bedtime, with extra tests at certain times, including when there are symptoms of hypoglycaemia, when driving, with illness, after exercise, with consumption of alcohol, with pregnancy and pre-pregnancy and after a change of insulin regimen. Some simple rules should be given to patients concerning driving and SMBG that will ensure that there is no hypoglycaemia before driving and that any hypoglycaemia on a prolonged journey will be detected and corrected (Box 13.4.9.1.2). In the ‘unempowered’ patient with type 1 diabetes who does not make any use of everyday frequent monitoring data, recent SMBG results are nevertheless used by the health care professional to adjust therapy; testing, say, regularly for 2 weeks prior to a clinic visit or at a low frequency at different times of the day is valuable (before and after breakfast on Monday, before and after lunch on Tuesday, before after dinner on Wednesday, etc.).

In type 2 diabetes (Fig. 13.4.9.1.3 lower panel), most patients are stable and those on diet, metformin, glitazones, and gliptins do not need any monitoring, except, perhaps, when severely ill, during steroid treatment and when regular HbA_1c testing is not available. Those with type 2 diabetes receiving sulphonylureas, and, therefore, susceptible to hypoglycaemia, need a low frequency of SMBG, e.g. once daily at different times on each day and when feeling ill or hypoglycaemic. Pre-breakfast and pre-evening meal testing are in common use in patients with type 2 diabetes for assessment of control and the detection of hypoglycaemia, respectively. Empowered patients with insulin-requiring diabetes should test about 1–4 times daily, depending on stability (plus additional testing when ill/hypoglycaemic), and unempowered patients need test only in the week or two before a clinic visit (Fig. 13.4.9.1.3).

SMBG has notable limitations, particularly, as mentioned, patients’ willingness to perform frequent testing being restricted by the discomfort of blood sampling. In addition, monitoring cannot be done at times when patients are at especial risk of hypoglycaemia, e.g. at night and when driving a motor vehicle. Furthermore, because of the marked and unpredictable glycaemic variability characteristic of type 1 diabetes, intermittent testing can miss episodes of hypo- or hyperglycaemia, and single tests give no indication of the rate or direction of glycaemic change. The ideal glucose monitoring technology would, therefore, be either noninvasive (not involving insertion into the body) or at least minimally invasive, and continuous (27). Several devices for CGM are now available for clinical use.

CGM used in clinical practice at the moment is based on needle-type electrodes (28), or microdialysis probes (29), that are inserted subcutaneously. It is therefore still minimally invasive. Amperometric (current-measuring) enzyme electrodes consist of glucose oxidase immobilized at a charged (usually about +700 mV) base electrode; oxidation of glucose to gluconic acid and hydrogen peroxide is monitored by the electrochemical detection of the hydrogen peroxide (Medtronic, DexCom).

One glucose sensor on the market (FreeStyle Navigator) uses a modification of this system, whereby an osmium mediator covalently bound to a polymer matrix ‘wires’ the enzyme to the electrode.

Microdialysis-based CGM (GlucoDay), involves insertion of a probe containing a fine, hollow dialysis fibre into the subcutaneous tissue, with perfusion of the fibre with isotonic fluid. Glucose from the tissues diffuses into the fibre and is pumped to the outside of the body where glucose is measured by an electrochemical biosensor based on glucose oxidase, as above (29). So far, microdialysis has been used less than enzyme electrodes.

Glucose sensors are implanted in the subcutaneous tissue and record interstitial fluid glucose concentrations every 1–5 min over (for current devices) an operating period of 3–7 days. A transmitter attached to the sensor relays the data wirelessly to a storage and display device, either a portable meter (Medtronic Guardian RT, DexCom STS, FreeStyle Navigator) or a modified insulin pump (Medtronic Paradigm RT). The sensor must be calibrated using conventional finger-prick capillary blood glucose tests performed when the blood glucose is not changing rapidly. This is because, although the blood glucose concentration is similar to that of the interstitial fluid in the steady state, when the blood glucose is changing rapidly, there is a variable lag time between the two compartments that will affect the sensitivity and calibration (30).

The restriction on the long-term use of sensors and the need for frequent calibration are due to impaired responses of sensors in vivo, probably caused by a variety of biocompatibility issues, including protein and cellular coating of the sensor, changing blood flow and oxygen tension at the implantation site, and electroactive interfering substances in the tissues.

CGM data can be either downloaded to a computer and reviewed by the health care professional and patient retrospectively in order to identify patterns and aid treatment changes (Fig. 13.4.9.1.4), or the patient can view the glucose information on the meter or pump in real time. CGM meters have alerts/alarms for high and low blood glucose levels and some have predictive alarms that warn of developing hypoglycaemia or hyperglycaemia. Trend arrows give an indication of the direction and rate of change of blood glucose levels.

There is continuing debate on how best to assess the accuracy of CGM systems. Traditional error grid analysis (scatter plots of the test results versus reference blood glucose measurements) was devised for intermittent SMBG, and does not take into account the rate and direction of change of glucose. A modified error grid for use with continuous data has been reported, but has so far not been widely employed (31). It is certainly true that accuracy of CGM depends on calibration (proper performance of capillary blood glucose testing in the steady state) and is confused by the subcutaneous tissue fluid–blood lag time for glucose, which is generally about 10 min, but can vary from about 3 min to 25 min. When the blood glucose is rising, e.g. after a meal, CGM readings lag behind the blood readings, but, when blood glucose is falling, CGM data can either precede the blood (if the sensor is sited next to tissue consuming glucose) or can follow after glycaemic changes. Several groups have also reported that current CGM devices tend to overestimate the degree of hypoglycaemia in subjects, e.g. a high frequency of nocturnal hypoglycaemia is often recorded with CGM though is not confirmed by SMBG at this time (30).

Recent randomized controlled trials show that CGM use in type 1 diabetes can improve HbA_1c by about 0.5% (5 mmol/mol) on average, but this seems to be dependent on frequent use of the technology. In a multicentre trial sponsored by the Juvenile Diabetes Research Foundation, 322 adults and children treated by MDI or CSII were randomised to SMBG or CGM (32). After 26 weeks, HbA_1c was 0.53% (5–8 mmol/mol) (CI−0.71 to −0.35%; p <0.001) less in the CGM group who were over 25 years old, but there was no difference in those who were 15–24 or 8–14 years old. Compliance with CGM was reasonably good in the adults (83% used the sensor for 6 or more days per week), but less so for the younger subjects (30% for the 15–25 and 50% for the 8–14 year olds, respectively).

In another study, adults and children were allocated to SMBG or either continuous or two 3-day CGM periods (Guardian RT) every 2 weeks (33). Though HbA_1c fell in all three groups over 3 months, only in the continuous group was it significantly less than the control. In a further randomized controlled trial over 6 months comparing SMBG versus CGM-augmented CSII, there was no difference in HbA_1c between the groups as a whole, though more than 60% use of CGM was associated with a significant HbA_1c reduction (34).

Hypoglycaemia changes during CGM are more uncertain at present. Though some studies show less time is spent in the hypoglycaemic range on CGM (35), there is little or no information on the clinical impact of this, particularly whether the rate of severe hypoglycaemia is reduced. Indeed, in the sensor-augmented pump study mentioned above (34), severe hypoglycaemia events were increased (compared to the control arm), possibly due to overcorrection of hyperglycaemia by bolus administration. One must also bear in mind that time spent in CGM-recorded hypoglycaemia may not be an accurate estimate of true low blood glucose concentrations.

The best use of CGM is still under discussion and the technology is generally only funded by health services or reimbursed by insurance organizations for selected patients with defined clinical problems. A reasonable consensus would be that those people with type 1 diabetes who have failed to achieve target levels of control using best attempts with intensified insulin therapy (MDI or CSII, and usually MDI followed by CSII), SMBG, structured diabetes education, and frequent contact with health care professionals may benefit for a trial of real-time CGM. This group will include those with continued elevated HbA_1c and/or those with continued frequent hypoglycaemia. On present evidence, CGM is likely to involve frequent use over many weeks, and, possibly, months, and must include full education on the use of CGM and instructions for changes in therapy in accordance with CGM results during the day. CGM should be used to supplement SMBG and is not a replacement.

Short-term CGM use over a few days and retrospective analysis by patient and health care professional is in much more widespread use than long-term, real-time CGM. There are potential and seemingly logical uses for this option, including identification of glycaemic patterns; relating glycaemic excursions to meal times, content, and duration (allowing adjustment of bolus doses in CSII or prandial doses in MDI); relating excursions to exercise intensity, duration, and type; relating hypoglycaemia to the time of day, insulin dose and timing, activity, meal, and alcohol intake, and hypoglycaemia symptoms; and using CGM to educate and motivate patients. However, more evidence is needed that CGM is more effective in this respect than frequent SMBG.

Non-invasive blood glucose monitoring (NIBGM) has not yet become a clinical reality, in spite of decades of research, but it continues to be an area of active investigation (36). A number of technologies are being researched (Box 13.4.9.1.3), and they can be divided into direct methods, where some intrinsic property of the glucose molecule is measured, and indirect methods, where the effect on some secondary process influenced by glucose is measured.

The most studied direct approach is near-infrared (NIR) spectroscopy, which is based on the relative transparency of the tissues to NIR light, i.e. wavelengths between about 700 and 1000 nm. There are glucose-related absorption peaks in the NIR region, but specificity is compromised by spectral overlap from many nonglucose substances in the tissues, including water, fat, and protein. Variable blood flow and changing hydration may also affect scattering of the light so that the optical path length of the NIR beam changes unpredictably. Multivariate models can be built using NIR signals recorded at many wavelengths and glucose concentrations to predict blood glucose concentration. So far, precision from one day to the next has been too poor to allow clinical use (37).

Raman spectroscopy is complementary to NIR and depends on the fact that a proportion of the light scattered by a molecule is shifted in frequency depending on the properties of the molecule. Glucose has sharp Raman spectral features, and, though weak, NIBG approaches using these are being investigated.

In reverse iontophoresis, a small current is passed between two skin-surface electrodes, which draws ions to the surface. Glucose is carried with the electroosmotic flow of water to the skin surface where it can be measured with a glucose oxidase-based biosensor. A glucose sensor using reverse iontophoresis was marketed some years ago (GlucoWatch), but has now been withdrawn because of a number of problems, including skin rash and irritation, a low flux of glucose, and skips in readings due to movement and sweating, with consequent inaccuracies.

Another transdermal technology uses low frequency ultrasound (sonophoresis) to increase skin permeability, and interstitial fluid can be extracted (sometimes with the application of a vacuum to the skin surface), and glucose subsequently measured (36).

Indirect methods include light scattering and the related technique of optical coherence tomography, where depth-resolved scattering measurements are made (36). These methods are based on glucose-related changes in the refractive index of the fluid surrounding the cells, membranes, and fibrils of the tissues, which change the scattering coefficient. Alternatively, NIBGM using impedance or dielectric spectroscopy involves applying an alternating current to the skin to measure impedance as a function of frequency. Acceptable correlations between light scattering or impedance and blood glucose are apparent in many, but not all, subjects, and probably the same influences of blood flow, temperature, hydration, and movement affect the accuracy and precision, and must be compensated if a workable device is to result.

Fluorescence is attracting increasing interest as both a minimally invasive and NIBGM technology, since it is very sensitive, free from electroactive interferences that affect implanted enzyme electrodes, and because both intensity and lifetime can be measured (37). Fluorescence lifetime is relatively independent of light scattering and fluorophore concentration, so potential covering of an implanted sensor by protein or cells will not alter the signal. Both direct and indirect NIBGM fluorescence-based glucose sensing are being researched. The production of cell NAD(P)H due to a number of glucose-dependent metabolic pathways can be detected by its fluorescence, and is related to glucose concentrations in cell culture (38); it is being investigated for skin-surface NIBGM. Micro- or nanosensors based on a fluorescent-labelled glucose receptor might be used as a ‘smart tattoo’ when impregnated in the skin or implanted subcutaneously; a suitable receptor is fluorophore-tagged glucose-binding protein that undergoes a marked change in conformation and fluorescence intensity and lifetime on binding glucose (39).

Various minor components of adult haemoglobin (HbA_o) result from the slow non-enzymatic attachment of glucose and other sugars to amino groups on haemoglobin over the lifetime of the red cell. The component present in largest amount is HbA_1c, formed when glucose links to the N-terminal valine of the β chain of haemoglobin. Glucose initially forms a Schiff base linkage with the N-terminal amino group of the protein; this then rearranges to a more stable ketoamine product (8).

Glycated haemoglobin (previously called glycosylated or glyco-haemoglobin) and HbA_1c have been in common use since the 1970s as a measure of average glucose control over the preceding weeks and months, although there is weighting for glycaemic changes in the one month preceding the sampling. Most major clinical trials relating diabetes control to outcomes such as the frequency of micro- and macro-vascular complications (e.g. the DCCT and the UK Prospective Diabetes Study) have used HbA_1c as an index of glycaemic control and thus HbA_1c is now seen as a risk factor for the development of complications.

Several methods exist for measuring HbA_1c, including ion exchange chromatography, electrophoresis, affinity chromatography (based on the binding of the cis-diol group of glucose to immobilized boronic acid derivatives) and immunoturbimetric methods. The reference range for HbA_1c has been taken to be about 4.5% (26 mmol/mol) to 6.5% (48 mmol/mol) (i.e. the percentage of HbA_o that is glycated), and values in the worst-controlled subjects with diabetes reach levels of about 10–14% (86–129 mmol/mol). However, this reference range and the actual values have not been comparable between laboratories (partly because of differing methods and glycated species measured), and, in recent years, methods have generally been ‘aligned’ to the method used in the DCCT, in an effort to reduce variability between laboratories. The International Federation of Clinical Chemistry (IFCC) has now developed a more definitive standardization system for HbA_1c based on purified HbA_1c and HbA_o. Although highly correlated with the more familiar DCCT-aligned results, the IFFC HbA_1c value is reported in mmol/mol, e.g. 5% (DCCT) being equivalent to 31 mmol/mol (IFCC), 7% to 53 mmol/mol, and 9% to 75 mmol/mol. For the time being, it is recommended that both the IFCC and a DCCT value (derived from the IFCC value using a master equation) be reported.

Serum proteins are also glycated in a manner analogous to haemoglobin, and fructosamine is the generic name for these products, though it is mostly glycated albumin. It can be measured using a colorimetric procedure in an automatic analyser. Because the half-life of albumin is about 17 days, fructosamine is a measure integrated glycaemic control over a much shorter period than HbA_1c—about 2–3 weeks. It is a less popular method for assessing control than HbA_1c, and is in most use when control is changing rapidly, e.g. during pregnancy in the patient with diabetes.

Blood glucose monitoring in diabetes is entering an exciting new phase with CGM. It gives patients information about the direction of change in glycaemic control, and allows alarms when hypo- and hyper-glycaemic thresholds are exceeded. The aim of coupling sensor readings to an insulin pump to provide automatic closed-loop insulin delivery—an artificial pancreas—has been a goal for many decades. Although a moment-to-moment, fully closed-loop system is some way off as a routine treatment, simpler CGM-assisted pumps that suspend the basal insulin rate when the sensor detects hypoglycaemia are already entering clinical practice.

NIBGM has seen slower progress, but new technologies are being explored. Some 45 years after blood glucose reagent strips were introduced, innovative technologies for glucose monitoring continue to be actively researched.