Home / Resources / Clinical Gems / Handbook of Diabetes, 4th Edition, Excerpt #8: Diabetes Control and Its Measurement

Handbook of Diabetes, 4th Edition, Excerpt #8: Diabetes Control and Its Measurement

Aug 24, 2014
Rudy Bilous, MD, FRCP
Richard Donnelly, MD, PHD, FRCP, FRACP


Diabetic control ‘defines the extent to which the metabolism in the person with diabetes differs from that in the person without diabetes. Measurement usually focuses on blood glucose: ‘good ‘ control implies maintenance of near-normal blood glucose concentrations throughout the day. However, many other metabolites are disordered in diabetes and some, such as ketone bodies, are now more easily measurable and clinically useful, particularly during acute illness or periods of poor blood glucose control (Figure 9.1)….

In addition to blood and urine glucose concentrations, there are indicators of longer term glycemic control over the preceding weeks using plasma glycated hemoglobin (HbA1c) or fructosamine concentrations (Table 9.1).

Capillary blood glucose monitoring

Single blood glucose measurements are of little use as an assessment of overall control in type 1 diabetes because of unpredictable variations throughout the day and from day to day, although they are important in order to detect hypoglycemia. In order to assess control more meaningfully, serial, timed blood glucose samples are usually needed. In diet-or stable tablet-controlled type 2 diabetes, although blood glucose levels are elevated they tend not to vary widely throughout the day. In these patients a fasting or random blood glucose relates reasonably well to mean blood glucose concentration and to glycated hemoglobin and is probably adequate.

Self-monitoring of capillary blood glucose by patients at home using special enzyme-impregnated reagent strips and a meter is now an integral part of modern diabetes management, especially for those who are on insulin therapy (Figure 9.2). Strips usually contain a combination of glucose oxidase and peroxidase. Colorimetric tests have now largely been superceded by newer electrochemically based strips which generate a current rather than a color change. Meters vary in their need for standardization, their memory and their ability to generate blood glucose profiles when connected to a computer. Some contain algorithms that can give advice on insulin dosage prior to a meal, depending on its carbohydrate content. Some meters require more blood than others. Meters are often made freely available by the manufacturers.

It is worth remembering that all meters tend to be less accurate at lower blood glucose values and usually have an upper limit of detection following which they read ‘high’.





There are many devices which contain a spring-loaded lancet in order to obtain a capillary blood sample (Figure 9.2). This is usually obtained from the fingers; the sides of the fingertip are less sensitive than the pulp. A major reason for poor compliance and low frequency of testing is finger discomfort. In recent years strips require less blood and many devices have a depth adjustment. Some offer the option of testing at alternative sites to the fingers such as the forearm, abdomen, calf and thigh. However, there can be discrepancies in values measured at the finger and these sites, particularly during times of rapid change in blood glucose such as after meals or exercise.
Frequency of testing

Initial trials of home blood glucose monitoring were nothing short of revelatory for patients used to urine tests. Latterly, as part of clinical trials (e.g. Diabetes Control and Complications Trial (DCCT) in type 1 and UK Prospective Diabetes Study (UKPDS) in type 2) and structured educational programs (e.g. Diet Adjustment For Normal Eating (DAFNE) in the UK), they have been shown to help patients achieve sustained long-term improvements in glycemic control. However, systematic reviews have failed to confirm that home blood glucose monitoring alone results in significant glycemic improvement. Many patients, though, prefer it to urinalysis, and it is hard to see how multiple daily insulin injection regimens can be used without it. NICE guidance uses it as an essential component of care management in type 1 diabetes with a frequency dependent upon the clinical circumstances, whereas for type 2 diabetes home blood glucose monitoring should be available for the indications listed in Box 9.1. Both NICE type 1 and type 2 guidelines suggest that knowledge and skills of interpretation, and action based upon home blood glucose monitoring results should be assessed annually. The American Diabetes Association guidelines suggest three or more tests per day in type 1 diabetes on multiple daily injections or pump therapy and for pregnant women. Otherwise their advice is concordant with that from NICE.

Urine glucose monitoring

Glycosuria occurs when blood glucose levels exceed the renal threshold for glucose (usually 10 mmol/L-180 mg/dL). However, urine glucose testing is unreliable in the assessment of blood glucose control because renal threshold varies between and within patients (Box 9). Fluid intake can affect urine glucose concentrations and importantly, the result does not reflect blood glucose at the time of the test but over the duration that the urine has accumulated in the bladder. A negative urine test cannot distinguish between hypoglycemia, normoglycemia and modest hyperglycemia.

However, urine testing still remains a reasonable option in stable type 2 diabetic patients treated with diet or oral agents, particularly in those who are unable or unwilling to perform blood glucose monitoring. It should be supplemented by glycated hemoglobin tests once or twice a year. It is worth noting that there is no benefit in terms of assessment of control in using freshly voided urine samples.


Box 9.1 Indications for capillary blood glucose monitoring in type 2 diabetes

  • Insulin therapy
  • On oral therapy with a risk of hypoglycemia (e.g. sulphonylureas, glitinides)
  • Assessment of response of glycemia to changes in management or lifestyle
  • Monitoring of glycaemia during intercurrent illness
  • Avoidance of hypoglycemia during driving, employment or physical activity

Box 9.2 Limitations of urine testing for glucose

  • Variations in renal threshold, especially in pregnancy
  • Variable result depending on urinary output/concentration
  • No immediate relationship to current blood glucose
  • Negative test unhelpful for detection of hypoglycemia
  • Visual reading of colour required
  • Accuracy may not be as precise at urine concentrations around 5.5 mmol/L
  • Some drugs may interfere with the test
Glycated hemoglobin

Hemoglobin A comprises over 90% of most adult hemoglobin and is variably glycated by the non-enzymatic attachment of sugars. HbA1c comprises the major glycated component and has been shown in numerous studies to correlate with average blood glucose (Figure 9.3).

As the average life span of the red cell is 90-120 days, the percentage glycated hemoglobin is a reflection of glycemic control over the 8-12 weeks preceding the test. However, the level of glycation is not linear with time-50% of the value reflects the 30 days prior to the test, and only 10% the initial 30 days of red cell life.

It is important to remember this because if the life span of a patient’s red cell is less than 90 days then theoretically the HbA 1c could be 50% of the expected value. Although high-pressure liquid chromatography (HPLC) methodology and the new International Federation for Clinical Chemistry and Laboratory Medicine (IFCC) standard have largely eliminated the confounding problems of aberrant hemoglobins, these can still cause falsely high values in some populations where they exist in high prevalence. It is important to check local assays in order to be aware of potential confounding factors.

The more common causes of misleading HbA1c values are shown in Box 9.3. Carbamylation due to uremia increases HbA1c by 0.063% for every 1 mmol/L increase in plasma urea concentration so is of relatively minor consequence.


Of more importance is the observation that there is considerable interindividual variation in the correlation between average blood glucose and HbA1c. Analysis of the DCCT cohort showed that for a mean blood glucose of 10 mmol/L based upon 7-point home blood glucose monitoring profiles over 24 hours, the HbA1c can range from 6% to 10%. A concept of rapid and slow glycators has been proposed to explain this phenomenon, but it is more likely to reflect variable red blood cell membrane transport of glucose. Recent research has shown a range of approximately 0.7-1.0 for this property between individuals and could explain HbA1c differences of 1.5-2.3% for any given mean blood glucose value.

These observations question whether a single target HbA1c should be used and perhaps explains some of the often observed discrepancy between recorded home blood glucose tests and glycated haemoglobin concentrations.

The recommended frequency of testing of HbA1c is twice per year in stable patients and 4-6 times for those undergoing treatment changes.


Box 9.3 Potential reasons for a misleading HbA1c Altered red blood cell turnover
  • Blood loss
  • Hemolysis
  • Hemoglobinopathies and red cell disorders
  • Myelodysplasia
  • Pregnancy
  • Iron deficiency

Interference with assay

  • Persistent foetal hemoglobin
  • Hemoglobin variant
  • Carbamylation
  • Too frequent testing
  • Differences of approximately 0.4% reflect + 2 SD for most modern assays

Variability in red blood cell membrane transport (slow/rapid glycators)


Estimated average glucose (eAG)

Because many patients have difficulty relating HbA1c to their results of home blood glucose monitoring, glycated hemoglobin levels are now often reported together with an estimated average blood glucose (eAG). The initial equations were based upon the DCCT cohort but the more recently used conversion has come from the A1c-Derived Average Glucose (ADAG) Trial (Table 9.2) utilizing frequent capillary blood glucose measurements and continuous subcutaneous blood glucose monitoring. The strongly positive correlation r = 0.92) has not been reproduced in children and there may also be differences in African Americans. There is ongoing debate about the utility of eAG and there are as yet no data to suggest it has clinical benefit over and above HbA1c. It is also important to remember that eAG relates to plasma, not whole blood, so there will inevitably be some discrepancy with meter-based home blood glucose measurements. However, eAG could provide a more accessible estimate of control for patients and create the basis for more meaningful discussions about management.



Glucose variability

Attempts have been made to obtain an estimate of blood glucose variability based upon the ranges or standard deviations of the mean of profiles, or continuous subcutaneous monitoring. So far, these analyses using the DCCT and other data sets have not been shown to provide advantages over HbA1c alone.

IFCC standard


Most HbA1c assays have been standardized to that used in the DCCT as part of work carried out by the National Glycohemoglobin Standardization Program (NGSP) in the USA. However, Sweden and Japan have each had their own standard. The IFCC has developed a new reference method that specifically measures only one molecular species of HbA1c and relates this to total hemoglobin. This method is expensive and laborious and can only be used to standardizes local assays. It reports in units of mmol/mol and the absolute values will be quite different from the current familiar percentage. However, it has been decided internationally that there should be a gradual switch to the IFCC standard with its new units. An international consensus agreed the following.

  • HbA 1c results would be standardized worldwide to the new IFCC standard.
  • The IFCC method is currently the only valid anchor that permits such standardization.
  • HbA1c would be reported in both new and old units for the time being (probably until 2011) together with eAG.
  • Glycemic goals should be expressed in IFCC units, NGSP percent and eAG mmol/L or mg/dL.



A 24 – year – old white Europid man developed classic symptoms of type 1 diabetes and was commenced upon insulin as a basal-bolus regimen. His initial HbA1c was 9.7%. Six months later at

regular review his home blood glucose monitoring showed excellent control with readings between 3.8 and 8.9 mmol/L (68 – 160 mg/dL) and he only reported occasional, mild, effort – related hypoglycemia. However, his HbA1c value came back the next day at 8.3%. He was contacted at home and an increase in his insulin of 2 units per dose was recommended and a repeat HbA c ordered for 6 weeks  time. This was 8.1% and further insulin increase was instituted. Five days later he was admitted to hospital following a profound nocturnal hypoglycemic episode during which he was found fitting.

Hemoglobin electrophoresis revealed the presence of HbS. The laboratory used an HbA1c assay that was sensitive to HbS, particularly at lower HbA1c concentrations. On direct questioning, it transpired that his parents were from the Mediterranean area.

Comment: Several learning points emerge. HbS can occur in non – African populations, so a family history in all people with diabetes is important. Secondly, it is important to know the limitations of the assays used by local laboratories. Lastly, in the presence of discrepancies between home monitoring and laboratory, do not always assume the patient’s tests are incorrect.


Serum fructosamine is a measure of glycated serum protein, mostly albumin, and is an indicator of glycemic control over the preceding 2-3 weeks (the lifetime of albumin). Colorimetric assays for fructosamine, which are now adapted for automated analyzers, give a normal reference range of 205-285 μ mol/L. Fructosamine generally correlates well with HbA1c, except when control has changed recently.

It has potential advantages over HbA1c, particularly in situations such as hemoglobinopathies or pregnancy when the glycated hemoglobin is hard to interpret. However, standardization is difficult: uremia, lipemia, hyperbilirubinemia and vitamin C use can affect the assay, and there may be an effect of high or low circulating blood proteins.



Urine and blood ketone measurements

Ketones can be measured in urine using a colorimetric test or in capillary blood using an electrochemical sensor similar to those now used for glucose (Figure 9.4).

Acetoacetate and acetone are detected by the urine test, β hydroxybutyrate by the blood sensors. As the ratio of β hydroxybutyrate to acetoacetate is around 6:1 in human ketoacidosis, the blood sensor offers a convenient way to monitor diabetes control during intercurrent illness or in situations that may predispose to ketoacidosis, such as pregnancy, or where it can occur relatively quickly, such as in patients using continuous subcutaneous insulin infusion pump therapy. As yet there is little evidence on which to form a consensus but blood ketone testing should be available in acute medical and obstetric assessment units as well as for inpatients with diabetes with intercurrent illnesses and perhaps as a means of monitoring response to treatment for diabetic ketoacidosis. Many units also provide their insulin pump users with blood ketone monitoring.


Koenig RJ, Peterson CM, Jones RL, Saudek C, Lehrman M, Cerami A. Correlation of glucose regulation and hemoglobin A1c in diabetes mellitus. N Engl J Med 1976; 295: 417-420.

Although abnormal haemoglobin electrophoresis had been described in diabetes since the 1950s, this was the first correlation between change in glycaemia and change in HbA1c. Five patients with a fasting blood glucose ranging from 280 to 450 mg/dL (15.6-25.0 mmol/L) were hospitalized and their values corrected to 70-100 mg/dL (3.9-5.6 mmol/L). HbA1c was 6.8-12.1% initially, falling to 4.2 – 7.6% after glycemic improvement. Later, much larger studies confimed the linear relationship but the author’s conclusion was spot on: "Periodic monitoring of hemoglobin A1c levels provides a useful way of documenting the degree of control of glucose metabolism in diabetic patients."



  • National Glycohemoglobin Support Program. Excellent information on HbA1c, eAG and the new IFCC standards: www.ngsp.org
  • National Institute for Health and Clinical Excellence (NICE). All UK guidelines available on this site (Type 1 Clinical Guidance CG15, Type 2 CG 66): www.nice.org.uk
  • Diabetes UK. Guidance on monitoring: www.diabetes.org.uk
  • American Diabetes Association. Standards of care published in Diabetes Care as a supplement each January: http://professional.diabetes.org/
  • SIGN Guidelines: www.SIGN.ac.uk

Continuous glucose monitoring systems

A major objective of diabetes research has been to provide continuous real-time monitoring of blood glucose so that insulin therapy can be matched to glycemia. Ideally such a system would be linked to an insulin delivery device automatically, thus ‘closing the loop’. In the last decade, huge strides have been made to achieve this goal but the current systems based upon capillary blood glucose monitoring technology using electron transfer do have their drawbacks.

Firstly, they are based upon measures of interstitial fluid, not blood glucose (Figure 9.5). This inevitably means that there is a delay or lag between detecting changes in blood glucose (mean delay 6.7 minutes, range 2-45 minutes). This lag can be affected by level of blood glucose, exercise, food intake and blood flow to the interstitial sampling site. Accuracy of the current devices tends to be less good at lower blood glucose levels. Secondly, such systems are by definition invasive as they require subcutaneous sensor insertion, usually on the abdominal wall. Thirdly, linkage to subcutaneous insulin infusion pumps introduces a further time lag in responsiveness (that of insulin absorption from the subcutaneous site). Finally, they need intermittent calibration with capillary blood glucose tests. The technology is also expensive and requires replacement every 5 days or so.


Clinical trials have, however, shown modest improvements of around 0.5% HbA1c at 6 months in young adults > 25 years of age; children and adolescents in the same study showed no significant benefit. In a separate study, adults with an HbA1c < 7% had a significant reduction in the time spent in biochemical hypoglycemia when randomized to continuous glucose monitoring compared to intermittent capillary blood tests. However, there was no difference in HbA1c in this study. These trials have used open loop algorithms and results were highly dependent upon patient motivation, training and education. Those who used the devices more consistently and made regular adjustments to their insulin dose obtained most benefit.


Closed loop devices and truly non-invasive glucose monitoring systems are under intense research and there are certain to be rapid developments in the near future. Meanwhile, the existing systems based upon interstitial glucose sensing probably have a role for patients struggling with glycemic control (particularly those with unpredictable and severe hypoglycemia), who are either on an insulin pump or multiple daily injections, and who are looked after by specialist teams who are well versed in the technology (Figure 9.6). 

Next Week: Metabolic control and complications

For more information and to purchase this book, just follow this link:

Rudy Bilous MD, FRCP, Professor of Clinical Medicine, Newcastle University, Honorary Consultant Endocrinologist, South Tees Foundation Trust, Middlesbrough, UK
Richard Donnelly MD, PHD, FRCP, FRACP, Head, School of Graduate Entry Medicine and Health, University of Nottingham, Honorary Consultant Physician, Derby Hospitals NHS Foundation Trust, Derby, UK 
A John Wiley & Sons, Ltd., Publication

This edition first published 2010, © 2010 by Rudy Bilous and Richard Donnelly. Previous editions: 1992, 1999, 2004

Note: The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting a specific method, diagnosis, or treatment by physicians for any particular patient. The publisher and the authors make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. Readers should consult with a specialist where appropriate. The fact that an organisation or website is referred to in this work as a citation and/or a potential source of further information does not mean that the authors or the publisher endorse the information the organisation or website may provide or recommendations it may make. Further, readers should be aware that Internet websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the authors shall be liable for any damages arising herefrom.