Effective diabetes care: a need for realistic targetsBMJ 2002; 324 doi: https://doi.org/10.1136/bmj.324.7353.1577 (Published 29 June 2002) Cite this as: BMJ 2002;324:1577
- Peter H Winocour (), consultant physician
- Department of Diabetes and Endocrinology, Queen Elizabeth II Hospital, Welwyn Garden City, Herts AL7 4HQ
- Accepted 19 December 2001
The research based metabolic and blood pressure targets that the national service framework for diabetes in England and Wales will set in the implementation document will be impractical for use in routine clinical care: targets need to be tailored to the individual patient, according to Peter Winocour
The publication of the national service framework for diabetes in England and Wales will raise public awareness of diabetes throughout the United Kingdom. Its recommendations are influenced by evidence that adverse vascular outcomes may be reduced by tight control of blood sugar and blood pressure1–3 and that secondary prevention of macrovascular disease is feasible through modification of dyslipidaemia and use of antiplatelet agents4–7w1; and by the recognition of an estimated doubling in incidence of diabetes over the next 10 years.8
This has led to the production of guidelines for optimal care that may be incorporated into a nationally recommended standardised approach, with targets for metabolic and vascular variables and recommendations for certain therapies.9–11w2 The effectiveness and quality of healthcare systems are increasingly being measured by how well such standards are adhered to, and it is therefore vital that the targets are appropriate. The following discussion focuses on type 2 diabetes, since this is much more common than type 1 diabetes.
Aggressive treatment of hyperglycaemia, dyslipidaemia, and hypertension and regular use of antiplatelet agents has been advocated in type 2 diabetes
Current targets for glycaemia, lipids, and blood pressure are attainable in only 50-70% of individuals in research studies
The targets are often impractical and involve taking too many drugs, with which patients often will not comply
Individually tailored targets are needed, and their effectiveness will be shown by improvement in the range of metabolic and blood pressure measurements in diabetic clinics
Control of glycaemia
Effective control of glycaemia has been incontrovertibly shown to reduce the development and progression of microvascular disease, particularly retinopathy.1 Subsequently, recommended fasting blood glucose concentrations were set at between 4 mmol/l and 7 mmol/l and glycated haemoglobin (HbA1c) level at or below 7%, albeit with certain provisos about individualising care on the basis of patients' characteristics.9–11w2
The method of measurement of HbA1c may affect the target level. Assay standardisation between laboratories is still some way off, and like will not be compared with like.12 The target HbA1c level of 7% was the mean level achieved in the United Kingdom prospective diabetes study,1 which means that, with the intensive treatment used in that study—which is not currently possible for routine diabetes services in the United Kingdom—50% of patients achieved this target. Even with such an approach, however, 50% were unable to reach this target, and after nine years fewer than 25% of obese subjects maintained this control.13
Furthermore, the type 2 diabetes in patients in the United Kingdom prospective diabetes study was newly diagnosed.1 Biochemical onset invariably precedes clinical diagnosis, and the pancreatic β cell function and insulin secretory reserve in asymptomatic individuals diagnosed through screening and those presenting with symptomatic hyperglycaemia will differ. Glycaemic control deteriorates further with time as complications ensue, despite progressive requirements for higher doses of hypoglycaemic agents.13
Average HbA1c levels in clinics in Britain are 7.4% after five years' estimated duration of type 2 diabetes and at least 8% thereafter, often while on insulin treatment.14w3 It is not yet known whether the introduction of the thiazolidinediones and glitinides will have any major effect on this process, although they can lower HbA1c by 1-1.5% for up to two years.w4 w5
It is thus unrealistic to have a 7% target for HbA1c for all patients with type 2 diabetes. Individualised targeting would be preferable, taking account of estimated duration of diabetes, obesity, concomitant therapy, comorbidity, age, and lifestyle. Even so, observational analysis of data from the United Kingdom prospective diabetes study shows there is no threshold below which there is no microvascular risk for an individual,15 a fact that all clinicians recognise in patients with exemplary glycaemic control but an adverse outcome. The recommended 7% target for HbA1c was a mean for a population. This level may therefore be valid in clinical practice only in as much as we should focus efforts on shifting the mean HbA1c level we try to achieve in our clinics towards this value by a sensible margin which recognises the population being cared for. As metabolic control has been shown to be poorer and morbidity and mortality greater in areas of deprivation,14 additional resources should be preferentially directed to such areas, focusing on “value added” improvements that take account of baseline levels in clinic populations.
Ultimately, efforts to achieve the 7% target for HbA1c will require at least 50% of patients with type 2 diabetes to be treated with insulin, with resource implications as well as practical challenges in inducing obese patients to accept more energetic treatment. Even individualised targets are certain to change with increasing duration of diabetes, and perhaps the real challenge is to assess and delay the point when an individual's previous HbA1c standard needs to be changed.
Control of blood pressure
Similar challenges are faced in managing hypertension. Several studies of hypertension in type 2 diabetes 2 3 w6 have led to recommendations that blood pressure targets should be 130-140 mm Hg for systolic and 80-90 mm Hg for diastolic pressure. 16 17 Tight control of blood pressure in the United Kingdom prospective diabetes study achieved more dramatic results than did better glycaemic control, with not just microvascular benefits but also a substantial reduction in the incidence of strokes and congestive cardiac failure.2 However, as with glycaemia, considerable effort was required to produce a clear separation in blood pressure between the intensive and conventional treatment groups; almost 30% of patients required at least three antihypertensive agents to achieve a mean blood pressure of 144/82 mm Hg. Only 56% reached the preset target of 150/85 mm Hg, and in 4% mean blood pressure did not fall below 180/105 mm Hg.2
In the hypertension optimal treatment (HOT) study of tight control of blood pressure, intensive control required maximum dosage of three antihypertensive agents in over 20% of subjects.3 The incidence of major cardiovascular events was 51% lower in patients with diabetes whose target diastolic blood pressure was set at 80 mm Hg than in those with a target of 90 mm Hg, although only 50% actually reached the 80 mm Hg target.3
Drop-out rates in research studies 2 3 w6 need to be reflected in clinical guidelines, as does recognition of poor long term compliance with polypharmacy of patients with type 2 diabetes.w7 The health survey for England and Wales recently found that blood pressure was controlled below 160/95 mm Hg in less than 50% of hypertensive patients with diabetes.18 Given all these factors, perhaps tight targets for blood pressure (140/90 mm Hg) are unrealistic for more than 30-40% of hypertensive patients in routine diabetes practice, where it is more realistic to aim for 75% of hypertensive patients to be below 160/95 mm Hg and a reduction in mean clinic blood pressure.
Microalbuminuria indicates increased risk of both microvascular and macrovascular disease, and this should also be borne in mind in individualising care.19 The most intensive efforts to control hypertension, hyperglycaemia, dyslipidaemia, and hyperviscosity should be focused on patients with microalbuminuria.
With the exception of small numbers in secondary prevention studies,4–6 there have been no published intervention studies specifically looking at the effects of statins (3-hydroxy-3methyl-glutaryl coenzyme A reductase inhibitors) in prevention of cardiovascular disease in patients with diabetes. The largest diabetic cohort in any secondary prevention study focused on use of the fibric acid derivative gemfibrozil in individuals with low concentrations of high density lipoprotein cholesterol, not with raised total and low density lipoprotein cholesterol.6 This form of dyslipidaemia (with modest hypertriglyceridaemia) is more typical of people with type 2 diabetes and coronary heart disease, and the benefits of gemfibrozil were at least as impressive as the risk reduction shown in the smaller diabetic cohorts treated with statins. 4 5
What then are the concentrations of lipids and lipoproteins at which treatment of dyslipidaemia be recommended, and what target levels could be achievable? In the United Kingdom the concept of separating primary from secondary prevention of coronary heart disease has been promoted, on the grounds that both absolute and relative risks are greater once a myocardial infarction has occurred. This is undoubtedly true in the absence of diabetes, and among people with type 2 diabetes the prognosis of coronary heart disease is also worse once it becomes overt.20w8 However, the risk of coronary heart disease in patients with type 2 diabetes without overt coronary disease may be as high as in patients with established coronary disease but without diabetes,20 in part a reflection of the prevalence of “silent” coronary disease in type 2 diabetes. Although this has recently been contested,21 separation of primary and secondary prevention in type 2 diabetes may still not remove the need to consider treatment of hyperlipidaemia. Thus, at present, in the absence of definitive studies of primary prevention, recommendations and targets are based on observational evidence. When a diagnosis of coronary disease has been established, aggressive treatment of dyslipidaemia is mandatory, and both fibrates and statins should be considered. For patients who have type 2 diabetes but are not known to have coronary disease, alternative recommendations have been made. The most favoured is the use of risk tables,22w9 w10 which have been developed on the basis of observational studies for both diabetic and non-diabetic groups, although there were very few people with diabetes in these studies, providing a woefully small database from which to estimate risk.
Risk tables make no estimation of the impact of varying degrees of hyperglycaemia, hypertriglyceridaemia, and microalbuminuria on vascular risk. Current risk stratification charts estimate risk over a 5-10 year period21w9 w10 or use the same database to annualise risk, presenting a uniform equivalent risk each year over 10 years. Both approaches are artificial. The risk of coronary disease over 10 years in the United Kingdom prospective diabetes study is not the same each year but increases with age over this period.w8 However, from its larger representative database the evolution of cardiovascular disease at different time points in the clinical course of type 2 diabetes could be examined, allowing calculation of real and not estimated annual risk. Also, the effects of glycaemia and albuminuria, as well as of high density lipoprotein cholesterol, triglycerides, and electrocardiographic findings, could be taken into account, providing a more valid risk table. A preliminary report has produced a predictive model with this objective.23
The alternative approach has been to acknowledge that there is an enhanced risk of coronary disease in diabetes and aim to treat patients whose lipid and lipoprotein concentrations suggest they are at high risk. Thresholds for treatment vary from 4.8 or 6.0 mmol/l for total cholesterol, and from 3.0 or 4.0 mmol/l for low density lipoprotein cholesterol. 10 11 24 This fraction is often not directly measured and cannot be estimated in the presence of the severe hypertriglyceridaemia found in over 5% of patients with type 2 diabetes.10 The European and American recommendations 10 11 24 correctly state that levels of other lipid and lipoprotein fractions need to be taken into account. Triglyceride concentrations recommended for treatment range from 1.7 to 4.5 mmol/l, whereas recommendations for high density lipoprotein cholesterol levels needing treatment range from less than 0.9 mmol/l to less than 1.2 mmol/l. 10 24 Although women have higher concentrations of high density lipoprotein cholesterol than men there is no reference to the need to take sex differences into account.
The pattern of dyslipidaemia may make the choice of drug critical, but this is often not emphasised. Statins may not be desirable in dyslipidaemia with low concentrations of high density lipoprotein cholesterol and modest hypertriglyceridaemia (>4.5 mmol/l): patients with this pattern were excluded from the secondary prevention statin studies, 4 5 and statins have been found unsatisfactory in such patients.25
In current guidelines, target lipid concentrations have been set at 2.6 mmol/l for low density lipoprotein cholesterol, below 1.7 mmol/l for triglycerides, and for high density lipopotein cholesterol above 0.9 mmol/l for men and above 1.1 mmol/l for women. 10 24 In practice, these targets are attained in only 71% of patients, and over 80% require treatment with a combination of statin and fibrate (and nicotinate).26 For routine clinical practice, these recommendations seem fanciful, and perhaps no more than 50% of patients would be expected to attain them; individualisation of care and shifting of clinic means is a more realistic strategy.
Aspirin and clopidogrel have been evaluated for secondary prevention in diabetes. Aspirin gave a 15% reduction in cardiovascular events7 and clopidogrel may be even more effective.27 Routine use of aspirin has therefore been recommended for patients with type 2 diabetes and coronary disease. 7 10 w2 w11 Given the high absolute cardiovascular risk in type 2 diabetes, aspirin has also been recommended when additional risk factors, such as treated hypertension, albuminuria, smoking, dyslipidaemia, or retinopathy, are present,10 w11 which means, effectively, in at least 50% of all patients with type 2 diabetes.
Polypharmacy may be unavoidable. Given the cardiovascular risk profile of type 2 diabetes, up to 10% of patients could require two or three hypoglycaemic agents (ultimately including insulin), at least three antihypertensive agents, two hypolipidaemic agents, and aspirin. A high proportion will also require treatment for coexistent cardiovascular disease and coincidental unrelated chronic disease. It is difficult to see how we can realistically expect patients to comply for long with such a draconian regimen requiring so many separate drugs. The development of combination tablets by the pharmaceutical industry seems of the utmost priority for effective prevention of complications of type 2 diabetes. Until then we must be pragmatic. We should set targets that are reasonable for our clinics and our individual patients (table). The targets will be evidence based in so far as they will be derived from the findings of research studies, but they cannot in reality seek to achieve the results attained in such artificial circumstances.
I am grateful to my colleagues Miles Fisher and Jonathan Katz for reading and commenting on an earlier draft of this manuscript.
Competing interests PHW has served on professional advisory boards for Novo Nordisk and Takeda Pharmaceuticals and has received support for clinical research, presentations, and attendance at scientific meetings from these and other pharmaceutical companies.
Additional references for this article appear on the web