The monitoring of diabetes mellitus

‘There is no reason why a diabetic should not, if he can be taught to do so, lead a long and normal life. True, the diabetic life demands self-control from all its subjects … … … … But he must accept the diabetic creed and follow it faithfully.’

RD Lawrence, in Preface to the First edition of ‘The Diabetic Life. It’s Control by Diet and Insulin’. 1925. 1

Changes in clinical practice are often preceded by the gathering of evidence which can take several years. Diabetes mellitus is a prime example of how its management has changed with a mix of advances in laboratory medicine and technology.

I was first introduced to the world of the Maillard reaction, ² Amadori rearrangement and Schiff bases in the early 1980s while working with Professor Sir George Alberti when he handed me a paper on non-enzymatic glycosylation of proteins and their use in diabetes control. Fast forward to the late 1980s when the University of Manchester Department of Medicine at Manchester Royal Infirmary moved from World War II hutments to a new building. The departmental library moved too, and volumes lay haphazardly on shelves. Professor Peter Adams, turning to me, a research fellow, said ‘You look a bit spare, how about sorting out this library?’ Among the journal volumes, there was a historical treasure trove of publications from famous names, but it was here that I first saw a complete collection of the Annals of Clinical Biochemistry starting from volume one. It belonged to the late Jeff Lumb who produced seminal works on vitamin D and calcium metabolism. It was interesting to see the current Annals evolve from the Proceedings of the Association of Clinical Biochemist meetings around the country.

The monitoring of diabetes was initially based on urine glucose testing. RD Lawrence, the well-known Chemical Pathologist and Physician, in his 1937 edition of his book ¹ gave details of the best times to test for urine glucose and suggested that the Benedict’s test was better than using Fehling’s reagent. For general practitioners, he recommended tests such as the fermentation test for sugars and an iodine test for starch to determine suitability of ‘diabetic’ foods. By 1941, tablets were available to test for urine sugar and over time these morphed into the present day urine dipsticks. In 1962, a semiquantitative method for blood glucose was available. ³ By the 1970s, self-monitoring of blood glucose (SMBG) was being developed and demonstrated to improve glycaemic control. ⁴

In parallel to these developments, a new way to monitor glycaemic control was being developed. It started very much with Samuel Rahbar’s chance discovery of an unidentified band on electrophoresis in blood from a 67-year-old diabetic woman when he was working on a project looking at haemoglobin variants in Iran. He published his findings in a sister clinical biochemistry journal ⁵ and confirmed his results in a larger series in the USA. ⁶ Glycosylated haemoglobin A1c (HbA_1c) was first described in 1958 by Huisman et al. ⁷ and was characterized later in 1966, but the diabetes connection was only confirmed by Rahbar in 1969. Further work by Koenig et al.^8,9 established the clinical utility of HbA_1c through a series of papers in the mid 1970s. In the early 1990s, the Diabetes Control and Complications Trial confirmed the utility of HbA_1c in type 1 diabetes and in the late 1990s, the United Kingdom Prospective Diabetes Study demonstrated the usefulness of HbA_1c in type 2 diabetes. Glycaemic control based on HbA_1c was initially related to data from frequent SMBG and expressed in percentage points. In 2007, the International Federation of Clinical Chemistry (IFCC) put out a consensus statement on HbA_1c standardization based on a mixture of HbA_1c and haemoglobin A0 as calibrators and determined using mass spectroscopy and capillary electrophoresis with a further statement in 2010. ¹⁰ In 2011, in the UK, after two years of dual reporting of HbA_1c in both percentage and mmol/mol units, IFCC-only reported HbA_1c was introduced. In the same year, the World Health Organization produced guidance on using HbA_1c for diagnosis of diabetes mellitus.

A walk through the pages of the Annals reveals some interesting contributions to the literature on monitoring of diabetes. The first reference to diabetes in the Annals appears in August 1961, where urinary excretion of pyruvate in diabetes is described. In the same issue, a glucose oxidase method for measuring glucose is described. By 1968, the need for monitoring blood glucose close to the patient prompts a report of a rapid method for measuring blood glucose in the diabetes clinic. ¹¹ Earlier in 1965, there is a fascinating paper on the relationship between oral and intravenous glucose tolerance and insulin secretion. ¹² This is part of the evidence for the therapeutic basis of the currently used glucagon-like peptide-1 (GLP-1) agonists and dipeptidyl peptidase-4 (DPP-4) inhibitors.

In the 1970s, the emphasis remained on finding more convenient ways of home blood glucose monitoring, but the first reference to HbA₁ measurement was in 1980. ¹³ More papers appear in the 1980s, describing various methods of determination of HbA_1c.^14,15 To warm the hearts of the current proponents of screening a paper in 1982 makes a case for including serum glucose in the standard biochemical profile. ¹⁶ A review of ‘glycosylated haemoglobins’ in 1984 still questions its clinical utility. However, a paper from the Birmingham group in 1985 comparing the oral glucose tolerance test and total HbA₁ suggests that HbA₁ may be useful for the diagnosis of diabetes.^17,18 Geberhiwot et al. ¹⁹ take a subtle approach to diagnosis using HbA_1c. It has only taken 25 years for HbA_1c to be deemed good enough for diagnosis of diabetes!

In 1985, serum fructosamine, a measure of glycation of pooled plasma proteins, is proposed as an alternative to HbA_1c to monitor diabetes. ²⁰ A couple of years later, fructosamine was also being suggested as an alternative to the oral glucose tolerance test for impaired glucose tolerance. Winocour et al. ²¹ also highlighted that fructosamine and HbA_1c reflected glycaemic control to a differing degree in patients with type 1 diabetes. Wiener from Manchester, in the 1990s, asks some probing questions on type of glucose used in the oral glucose tolerance test.^22,23

In the last two decades, notable papers in the Annals include those on the diagnostic utility of HbA_1c and interference in the methods used in HbA_1c measurement. There is a seminal review on SMBG by Maurice O’Kane in 2009 ^{2 4} and an accompanying editorial. ²⁵ Sporadic reports on the utility of glycated albumin continue to be published ²⁶ and there is a useful audit of oral glucose tolerance testing in Australia indicating that nearly 40% of tests were done in the presence of confounding factors. ²⁷

The Annals has continued to publish papers that are of practical clinical utility. This is well illustrated by the literature on diabetes and its monitoring.