Management outcomes of patients with type 2 diabetes: targeting the 10-year absolute risk of coronary heart disease

The macro- and microvascular complications of type 2 diabetes significantly increase patient morbidity and mortality.1 However, there is also evidence from randomised controlled trials that treating hyperglycaemia, dyslipidaemia and hypertension will decrease the severity and/or delay the onset of diabetes-related complications, particularly coronary heart disease (CHD).2-5 Accordingly, the Royal Australian College of General Practitioners (RACGP) has proposed target threshold values for glycated haemoglobin (HbA1c), blood pressure and total cholesterol of < 7.0%, < 130/80 mmHg, and < 4 mmol/L, respectively.6 Smoking cessation, low-dose aspirin (50–150 mg/day), statins, and angiotensin-converting enzyme (ACE) inhibitors are also an integral part of this management strategy. Given that, for patients with diabetes, preventing CHD remains the primary goal of risk modification, estimating the 10-year absolute risk of CHD events provides a unique summary of key risk factors for an individual patient.7 Additionally, an absolute risk > 15% has been chosen to define patients within the high-risk category for CHD events, and therefore most likely to benefit from aspirin and statin treatment.8-10 Estimating the 10-year absolute risk also provides a single focus with which to review the current efficacy or assess the likely impact of different risk-factor modification strategies in the primary care setting.

We assessed the management of patients with type 2 diabetes mellitus in the primary care setting, with respect to risk factors associated with CHD.

Our study cohort comprised 328 patients with type 2 diabetes (52 patients with existing CHD were excluded) seen at the hospital’s Diabetes Assessment Clinic during the period March 2004 – February 2005. Patients were referred by their general practitioners for a comprehensive diabetes assessment and to receive diabetes education.

At the clinic, details of all current medications were recorded, blood pressure was measured, and a venous blood sample drawn for fasting lipids and HbA1c measurement. The 10-year absolute risk of CHD was calculated using the United Kingdom Prospective Diabetes Study (UKPDS) risk engine equation,7 which incorporates nine variables: age at diagnosis, duration of diabetes, sex, ethnicity, smoking status, systolic blood pressure, HbA1c level, and total cholesterol and high-density lipoprotein cholesterol levels. The UKPDS equation is not valid for patients with existing CHD. A summary of the assessment, highlighting non-attainment of RACGP target thresholds,6 as well as suggestions for treatment change, was mailed to the GPs.

Statistical analyses were performed using SPSS, version 14.0 (SPSS Inc, Chicago, Ill, USA).

The final cohort comprised 52% women and had a mean age of 58.3 years (95% CI, 57.5–59.1); body mass index (BMI) of 31.1 kg/m2 (95% CI, 30.4–31.9) (52% were obese); and duration of diabetes of 5.5 years (95% CI, 5.0–6.1). The respective group means for HbA1c, systolic blood pressure, and total cholesterol level were 7.4% (95% CI, 7.3%–7.4%), 139 mmHg (95% CI, 137–142), and 5.0 mmol/L (95% CI, 4.9–5.1). Among the cohort patients, 42%, 61% and 43%, respectively, were receiving medication to control glycaemia, blood pressure and cholesterolaemia. Ten per cent of all patients were using insulin. The proportions of patients achieving RACGP target values6 for HbA1c, blood pressure and total cholesterol were 46%, 29% and 15%, respectively, while 78% were non-smokers. Only 2% of patients attained all these four targets, while 6% attained none. Patients following a diet and/or lifestyle schedule were 3.9 times more likely to have an HbA1c level < 7.0% than those taking hypoglycaemic medication (95% CI, 2.5–6.3; P < 0.001). In contrast, patients treated with cholesterol-lowering drugs were 5.2 times more likely than untreated patients to have a total cholesterol level < 4 mmol/L (95% CI, 2.6–10.4; P < 0.001). There was no significant difference between those taking or not taking antihypertensive medication with respect to achieving a blood pressure level < 130/80 mmHg (odds ratio [OR], 1.3; 95% CI, 0.8–2.2; P > 0.1).

The mean 10-year absolute risk of CHD was 16.8% (95% CI, 15.7%–17.9%), and 48% of patients had a 10-year absolute risk of > 15%. The rates of aspirin, statin and ACE inhibitor use by the low- and high-risk patients (defined by a 10-year absolute risk 15% threshold) were 36% v 39% (OR, 1.1; 95% CI, 0.7–1.8; P = 0.7); 47% v 39% (OR, 0.7; 95% CI, 0.5–1.1; P = 0.15); and 44% v 53% (OR, 1.4; 95% CI, 0.9–2.2; P = 0.13), respectively. If each of the four risk factors were lowered within the patient group to achieve the respective targets, then the mean 10-year absolute risk of CHD would decrease to 12.6% (95% CI, 11.8%–13.4%); that is, 42 CHD events would be prevented per 1000 patients, or there would be a 4.2% absolute risk reduction in CHD events (25% relative risk reduction). Alternatively, focusing on individual risk factor reduction, achieving RACGP targets for total cholesterol, HbA1c, blood pressure and smoking would theoretically decrease the likelihood of 10-year CHD events per 1000 patients by 22.5, 9.2, 4.0 and 9.8, respectively. The increased effect for total cholesterol reduction reflects the low number of patients with a total cholesterol level < 4.0 mmol/L.

Our study highlights that a majority of patients with type 2 diabetes managed in the primary care setting and referred to a diabetes clinic are not achieving RACGP recommended target thresholds for HbA1c, blood pressure, and total cholesterol and may potentially benefit, in terms of a decreased 10-year risk of CHD, from clinically realisable interventions. For example, the UKPDS intensive arm achieved a 0.9% decrease in HbA1c and a 10 mmHg reduction in systolic blood pressure level,2,5 while the Heart Protection Study achieved a 1 mmol/L lowering in total cholesterol level using simvastatin.11 However, it is known that glycaemic control progressively deteriorates with diabetes duration,12,13 requiring increasingly aggressive antihyperglycaemic treatment. Accordingly, and consistent with other studies, we find that patients who are adhering to a diabetes diet have the lowest HbA1c values, and insulin-treated patients the highest.14-17 Intensive antihyperglycaemic treatment, though, has practical limitations, which relate to hypoglycaemia and, to a lesser extent, weight gain. For that reason, extensive debate persists about the appropriateness of strict HbA1c target thresholds for individual patients.18,19 In contrast, meeting total cholesterol target levels directly relates to the intensive use of cholesterol-lowering drugs, with the only limitations being adverse side effects and the economic cost when patients do not meet Pharmaceutical Benefits Scheme eligibility criteria.

We acknowledge that our findings are biased by both the geographical location and the referral characteristics of the GPs. However, recently published comparable Australian surveys14,17,20-22 for patients with type 2 diabetes also contain similar biases, and do not exclude patients with pre-existing CHD. Only the AusDiab Study (1999–2000),17 which identified 439 of 11 247 participants as having previously diagnosed type 2 diabetes, is a population-based survey. While the NEFRON study22 contained the largest cohort (n = 3893) of patients with type 2 diabetes, and involved some 500 GPs throughout Australia, its exclusive focus was dyslipidaemia, and no data are reported concerning either type of treatment or the degree of hyperglycaemia. The Box compares variables between the AusDiab Study,17 our study, and the study by Bryant et al21 (n = 509) (which contained the most important variables for direct comparison with the AusDiab and our studies). If we accept that the AusDiab Study17 has the most representative random diabetes cohort, then the characteristics of our study are more comparable, particularly for diet-only treatment and HbA1c < 7.0%, than those reported by Bryant et al.21 Two previous publications, by Ackermann and Mitchell20 and Thomas and Nestel22 also reported CHD risk estimates using the same UKPDS risk equation. Ackermann and Mitchell, however, only reported the 5-year absolute risk estimates, so any direct comparison is not possible.20 Thomas and Nestel showed that about 56% of patients not receiving lipid-lowering treatment have a 10-year absolute risk of CHD > 15%, quite similar to our finding of 52% of patients.22

In the main, and unaddressed by previous studies, our estimates of the 10-year risk of CHD reveal that primary CHD prevention for patients with type 2 diabetes in the primary care setting, dependent on the use of aspirin, statins and ACE inhibitors, is essentially unstructured. In particular, patients at the highest predicted risk of CHD are not preferentially targeted for aspirin, statin or ACE inhibitor treatment. For statins, there is even a tendency to target patients in the lowest CHD risk category. This represents a significant predicament, as the maximum likelihood of reducing CHD events is to treat high-risk patients with lipid-lowering drugs.

The results of our study and previous studies12-17,20-22 reiterate that target goals, like those recommended by the RACGP6 for patients with type 2 diabetes, are not being achieved. This unsatisfactory management outcome is inevitably framed in terms of patient compliance and adherence. Recent discourses,23,24 however, have highlighted that the nature of the compliance–adherence issue is complex and does not simply reflect on the patient, but also on the diabetic specialist and the GP and their connectivity. Accordingly, for patients with type 2 diabetes, meeting target goals will require new initiatives focusing on patients and their doctors, as well as the clinical and community environment.