Poor glycaemic control in type 2 diabetes: a conspiracy of disease, suboptimal therapy and attitude

Abstract

Glycaemia in type 2 diabetes is difficult to manage long‐term, and despite a wealth of epidemiological evidence, there continued to be doubts, until recently, as to whether intensive glucose control was beneficial. The publication of robust prospective evidence from the United Kingdom Prospective Diabetes Study1 in September 1998 marked a seminal change. In type 2 diabetes, there was extensive epidemiological data suggesting that complications were linked to glycaemic exposure,2 but the UGDP (University Group Diabetes Program)3 trial had raised doubts about the safety of sulphonylureas in reducing plasma glucose. The DCCT4 showed in 1993 that tight glycaemic control reduced microvascular complications in type 1 diabetes. The UKPDS provided evidence that tight control was beneficial in type 2 diabetes: patients in an intensively treated group achieved a median HbA1c of 7.0% at 10 years compared to 7.9% in those in a conventionally treated group. This improvement in glycaemic control was associated with a 12% reduction in any diabetes end‐points (p=0.029) and a 25% reduction in microvascular end‐points (p=0.0099).1 However the achieved HbA1c values of the trial were neither the aim of the trial, nor the best glycaemic control that could be achieved, because one of the aims was to address the question of the efficacy and safety of monotherapies. Thus the protocol required, in those randomized to sulphonylurea therapy, that, once the maximum dose was reached, the fasting plasma glucose could rise to 15 mmol/l or to the advent of symptoms before additional pharmacological agents were introduced. The epidemiology analysis of the UKPDS suggested that there was no discernible threshold for the improved outcome with lower glycaemia. However, in clinical practice optimal glycaemic control is difficult to obtain. Prospective randomized studies have achieved median HbA1c levels of 7.1%5 and 7.0% in intensively treated patients,1 but, by definition, 50% of the patients must have had values above this level. The reasons for poor glycaemic control are complex, and relate to the disease process itself, the inadequacy of therapeutic regimens and the attitudes of both doctors and patients. We discuss here some of the factors related to poor control, and propose some views about the solutions to the problem. One finding of the UKPDS was that fasting plasma glucose deteriorated with time. The deterioration in the diet‐only policy group strongly suggests that there is progressive deterioration of β‐cell function or an increase in insulin resistance over many years. A parallel deterioration was shown in those on sulphonylurea or metformin, suggesting that neither of these agents either accelerated or slowed the rate of decline. A more detailed analysis of the problem can be undertaken using modelling techniques. β‐Cell function can be measured by the homeostasis model (HOMA),6 which showed a reduction from a mean of 50% function at diagnosis to 25% at 5 years.7 These HOMA estimates of failure rate are concordant with other published estimates.6,8 There is no current identified therapy that stops the decline in β‐cell function and as β‐cell function declines, glycaemia increases. ‘Sulphonylurea failure’ is thus a misnomer: what is observed in the increasing glycaemia of type 2 diabetes over the course of many years as a result of failure of the β cell, not the therapeutic agent. It is also apparent that functional deterioration occurred in those patients in the UKPDS allocated to insulin, and is the explanation why, even when insulin was prescribed in appropriate doses, the HbA1c rose towards that achieved in the DCCT intensive group. This value, about 7%, is probably the best median achievable HbA1c in those who have total β‐cell loss. The rate of failure has been described in detail in the UKPDS patients: failure of sulphonylurea therapy is significantly greater in those who have a higher fasting plasma glucose at diagnosis: in the UKPDS, 61% of those who had fasting glucose >10 mmol/l at randomization required additional therapy at 6 years, compared to 23% of those with a fasting glucose <7.8 mmol/l (p=0.00001).9 In view of this progressive decline in β‐cell function, it is essential to monitor diabetes on a regular basis in order to increase therapy appropriately. In the UKPDS, the fasting plasma glucose rose at a rate of approximately 0.2 mmol/l/year and HbA1c at 0.2% per year (Figure 1). However in the Belfast study,10 this relatively slow rate of increase occurred only in those patients with the lowest mean fasting plasma glucose at diagnosis (7.5 mmol/l) but half of the patients deteriorated at the significantly faster rate of rate of 1.5 mmol/l/year. Patients and physicians need to be aware that patients are likely to need the addition of other therapeutic agents at regular intervals1 and those with diabetes should be made aware at an early stage that the need for such augmentation does not represent ‘failure’ on their p