• Home
• Polls
• News
• Forum
• HDL for Beginners
• Archive

• WHATS NEW?

• Suppression of chemokines by HDL
• HDL cholesterol linked to oestrogen levels
• New gene loci regulating HDL cholesterol identified
• Abnormal lipids when young predicts atherosclerosis in later life
• JUPITER: Is HDL cholesterol relevant to residual risk in low-risk patients on potent statin therapy?
• EAS Workshop: How does HDL protect against atherosclerosis?

A balanced perspective of ACCORD Lipid - Philip Barter

 Despite best treatment efforts, type 2 diabetes patients remain at increased risk of cardiovascular events. To address this unmet need, the overall rationale of the Action to Reduce Cardiovascular Risk (ACCORD) research programme was to investigate treatment strategies aimed at reducing this risk. These included intensified treatment (in the case of glycaemia and blood pressure management), as well as extending treatment beyond current options (in the case of ACCORD Lipid).

The aim of ACCORD Lipid was to evaluate whether the addition of fenofibrate to statin treatment, thereby comprehensively targeting high-density lipoprotein cholesterol (HDL-C), triglycerides and low-density lipoprotein cholesterol (LDL-C), would provide additional clinical benefits.

 It was found that this approach is not appropriate for most type 2 diabetes patients.¹ ACCORD Lipid failed to demonstrate any statistically significant benefit for the primary outcome (a composite of nonfatal myocardial infarction [MI], stroke or cardiovascular death) or any of the secondary cardiovascular outcomes in the study. These results are not surprising. Median triglycerides in the study population were 162 mg/dL (1.8 mmol/L), lower than levels that would normally justify treatment with a fibrate in clinical practice. As the effect of fibrates on plasma lipids is influenced largely by baseline lipid levels,² it is probable that the ability of fenofibrate to raise HDL-C and lower triglycerides was not optimal.

 While ACCORD Lipid found that most type 2 diabetes patients do not gain additional clinical benefit from the combination of a fibrate and statin over that achieved by a statin alone, there was a suggestion of benefit in a subgroup of patients who had elevated triglyceride and low HDL-C.1

 17% of patients in ACCORD Lipid had low baseline levels of HDL-C (≤ 34 mg/dL or 0.88 mol/L) and high plasma triglyceride (≥204 mg/dL or ≥2.3 mmol/L).¹ In this subgroup, treatment with fenofibrate reduced the primary event rate from 17.3% to 12.4%. In contrast, in patients without this lipid profile, the event rate was similar with or without fenofibrate treatment (10.1%).

 These findings are consistent with post hoc analyses from other fibrate studies which showed that people with high plasma triglycerides and low levels of HDL-C derive a disproportionately greater benefit from fibrate therapy (see slide 1).

 This is especially noticeable in overweight people, as shown by subgroup analyses from the Helsinki Heart Study⁷ (in which the risk reduction was 78% in patients with BMI >26 kg/m², low HDL-C and elevated triglycerides). Similarly, patients in the Bezafibrate Infarction Prevention study³ with 4-5 components of the metabolic syndrome (as defined by ATPIII) had a 56% reduction in risk of cardiac death (p=0.005) with bezafibrate. Taken together, the balance of evidence therefore supports treatment with a fibrate as a strategy to reduce cardiovascular risk in patients with type 2 diabetes who have high triglycerides and low HDL-C; this conclusion is consistent with current guideline recommendations.^8,9

 However, we do need to be cautious. First, we need to take into account the provisos associated with subgroup analyses. Second, we should be aware that the ACCORD Lipid Investigators have not yet investigated whether these findings can be explained by other characteristics of this subgroup. For example, were there racial disparities, compared with the overall study population, or was there a higher prevalence of pre-existing cardiovascular disease in this subgroup? Third, we need to take into account that there has been no adjustment for the multiplicity of subgroup analyses performed. At least this should be done for analyses performed across different tertiles of HDL-C or triglycerides.

 As regards the safety of adding fenofibrate to a statin, it appears from ACCORD Lipid that there is no excess risk of myopathy, a finding consistent with data from the FIELD study in which about 900 patients  received the combination of fenofibrate plus a statin. It is also reassuring that there were fewer deaths (either all-cause or cardiovascular) in the fenofibrate group than those receiving simvastatin alone in ACCORD Lipid. This is important, as it contrasts with evidence suggestive of the reverse effect in FIELD.¹⁰  Although serum creatinine did increase with fenofibrate treatment (within the first year and stable thereafter), this effect appears to be reversible, as shown in the FIELD study.¹⁰  There was also a lower incidence of albuminuria, both microalbuminuria (p=0.01) and macroalbuminuria (p=0.03), in the fenofibrate group.¹

 However, another subgroup analysis from ACCORD Lipid does raise concern. There was a significant gender interaction (p=0.01), suggestive of possible harm for type 2 diabetic women treated with fenofibrate.1 Whether this effect can be attributed to other baseline or on-study differences between women and men in the study is not known. This is clearly a priority for further investigation. 

 Overall conclusions

• Although ACCORD Lipid does provide evidence suggestive of additional benefit with fenofibrate in simvastatin-treated patients with high triglycerides and low HDL-C we need to cautious in translating these findings.
•  In particular, we need further information on the subgroup analysis suggesting potential harm in type 2 diabetic women.
•  HDL Forum believes that the hypothesis that adding a PPAR-alpha agonist to a statin will reduce cardiovascular risk in diabetic patients with elevated triglycerides and low HDL-C has still not been tested. There is a clear need to conduct a large cardiovascular outcome study specifically in diabetic patients with this lipid phenotype