Although statins significantly decrease the incidence of cardiovascular disease (CVD), residual CVD risk remains high. This may partly be due to uncorrected atherogenic dyslipidaemia. The driving force behind atherogenic dyslipidaemia is hypertriglyceridaemia, which results from hepatic oversecretion and/or hypocatabolism of triglyceride-rich lipoproteins, and is typical of type 2 diabetes and metabolic syndrome. Persistent atherogenic dyslipidaemia in patients treated with a statin according to low-density lipoprotein-cholesterol goals may be corrected with niacin, fibrates or n–3 fatty acids. Clinical trial evidence to inform best practice is limited, but new data support adding fenofibrate to a statin. A consistent feature of fibrate clinical trials is the specific benefit of these agents in dyslipidaemic patients and the improvement in diabetic retinopathy with fenofibrate. Ongoing clinical trials may provide good evidence for adding niacin to a statin. Low-dose n–3 fatty acids could be used routinely after a myocardial infarction, but the value of higher doses of n-3 fatty acids in reducing CVD risk remains to be demonstrated.
Statistics from Altmetric.com
Dyslipidaemia is a powerful predictor of cardiovascular disease (CVD) in high-risk patients,1 such as type 2 diabetics. Low-density lipoprotein (LDL) is the prime target for treatment,2 3 but clinical trials consistently show high residual risk of CVD among statin-treated subjects.4 This may be because statins do not adequately correct atherogenic dyslipidaemia.5 We examine the evidence and give guidance on additional treatments that may deal with this limitation of statins.
We focus mainly on patients with type 2 diabetes and dyslipidaemia and recent clinical trials employing a fibrate, but also refer to other therapeutic agents and non-diabetic subjects at high risk of CVD.
Elevated plasma triglycerides and low high-density lipoprotein (HDL) cholesterol, in both fasting and postprandial states, are characteristic of metabolic syndrome and type 2 diabetes.6 Other abnormalities include accumulation in plasma of small dense LDL particles and triglyceride-rich lipoproteins (TRLs), including chylomicron and very-low-density lipoprotein (VLDL) remnants. This is reflected by elevated plasma concentrations of non-HDL-cholesterol and apolipoprotein B-100 (apoB). Postprandially, there is also accumulation in plasma of TRLs and their remnants, as well as qualitative alterations in LDL and HDL particles. Hence, hypertriglyceridaemia is a marker for a wide spectrum of atherogenic lipoproteins not measured routinely.6 Observational evidence shows that elevated fasting and non-fasting plasma triglyceride concentrations, and by implication triglyceride-rich, apoB-containing remnant lipoproteins, are significant predictors of CVD.7 8 Atherogenic dyslipidaemia is seen in almost all patients with triglycerides >2.2 mmol/l and HDL-cholesterol <1.0 mmol/l, virtually all of whom have type 2 diabetes or central obesity and insulin resistance.6
Triglycerides as a CVD risk factor
The role of triglycerides as a CVD risk factor has been controversial.9 Misunderstanding arises from the apparent weakness of triglycerides as an epidemiological exposure for CVD risk, largely due to its high within-person variability compared with biologically related and more stable factors such as HDL.10 Therefore, allowing for multivariate models in which triglycerides compete with HDL, the latter will often take precedence,11 but this does not necessarily imply biological or pathophysiological relevance and overlooks the inherent atherogenicity of TRLs. The apparent weakness of conventional epidemiology in identifying causality can be overcome by unbiased methods such as Mendelian randomisation, which has recently been harnessed to demonstrate the causal link between elevated triglycerides and CVD. This is not to diminish the importance of measuring HDL-cholesterol, given its value in calculating both LDL and non-HDL-cholesterol and in estimating CVD risk.2 3 We acknowledge indications from recent trials in diabetic patients that low HDL-cholesterol could be particularly atherogenic in the presence of hypetriglyceridaemia,12 13 but perhaps not when identified as an isolated abnormality.14
Centrality of hypertriglyceridaemia in the pathogenesis of atherogenic dyslipidaemia
Hypertriglyceridaemia is central to the pathophysiology of dyslipoproteinaemia in insulin resistance and type 2 diabetes.6 An expanded adipose tissue mass that does not efficiently sequester fatty acids results in oversupply of substrate for triglyceride production in the liver, which in turn leads to hepatic steatosis and oversecretion into plasma of larger triglyceride-rich VLDL particles.15 These particles compete with chylomicrons and its remnants for clearance pathways regulated by lipoprotein lipase, an endothelial-bound enzyme, and by hepatic receptors, thereby exacerbating postprandial dyslipidaemia. Hepatic secretion of apoC-III is increased in insulin resistance and this small protein, which is attached to VLDL, delays the catabolism of TRLs by inhibiting lipoprotein lipase and binding of remnant TRLs to hepatic clearance receptors.15 Expansion of the VLDL triglyceride pool leads to cholesterol depletion and triglyceride enrichment of LDL and HDL facilitated by cholesteryl ester transfer protein, with accumulation in plasma of small dense LDLs and HDLs.6 Hence, expansion of the plasma pool of large TRLs drives a number of atherogenic changes in lipoprotein composition.
Targeting defective lipoprotein transport
Treatment should target the foregoing kinetic defects in lipoprotein particles, aiming to decrease hepatic secretion of VLDL apoB and -TG and transfer of apoB from VLDL to LDL, as well as accelerate clearance of all apoB-containing lipoproteins.15 Although lifestyle modifications, including low-fat diets, weight loss and exercise, are essential, pharmacotherapy is often required.15 16 While statins are first-line drugs, they do not universally correct the metabolic abnormalities.
Statin monotherapy: value and limitations
Statins are the most potent agents for lowering the plasma concentrations of LDL-cholesterol and apoB. They have less potent, but nevertheless significant, effects on plasma triglycerides, reductions being on average directly proportional to the efficacy in lowering LDL-cholesterol. More potent statins, such as rosuvastatin, may also reduce the production of LDL-apoB-100 and decrease the catabolism of HDL-apoA-I particles.15 Statins have not been shown, however, to consistently decrease the hepatic output of VLDL apoB, indicating that limiting de novo cholesterol synthesis is insufficient for reducing VLDL production. Hence, targeting the partitioning of triglycerides in the liver using additional measures, such as lifestyle or other drugs, is required to optimise treatment of dyslipidaemia.15 16
However, several clinical trials show that statins decrease cardiovascular events in subjects with type 2 diabetes, with an average 20% reduction in CVD events per 1 mmol/l reduction in LDL-cholesterol.4 In the Treat-to-New-Target (TNT) trial, a higher-dose statin (atorvastatin 80 mg daily) decreased the relative risk of CVD by 25–30% compared with a lower statin dose (atorvastatin 10 mg daily) in coronary patients with type 2 diabetes or with the metabolic syndrome alone.17 The Collaborative Atorvastatin Diabetes Study (CARDS)18 and Anglo-Scandinavian Cardiac Outcomes Trial-Lipid-Lowering Arm (ASCOT-LLA)19 trials also support the use of statins in diabetic patients without CVD, but with at least one other cardiovascular risk factor, including hypertension. These findings concur with the Heart Protection Study (HPS)20 and a meta-analysis of statin trials.4 The cardiovascular benefits of statins relate principally to the lowering of plasma LDL-cholesterol and lipoprotein remnants, but residual cardiovascular risk in statin-treated patients remains high.4 5 In diabetes, this notion of residual risk extends to both macroangiopathy and microangiopathy.5
Addressing the residual cardiovascular risk in statin-treated patients
The high residual risk may be due to persistent abnormalities in TRLs and HDLs which are not fully corrected by statins.21 22 Higher doses of statins may partially deal with this, but at risk of musculoskeletal (MSS) side effects,23 to which diabetic patients are susceptible given older age, polypharmacy and coexistent chronic kidney disease. Ezetimibe, fibrates, niacin and n-3 fatty acid ethyl esters are adjunctive treatments for residual hypertriglyceridaemia and risk of CVD. But how are these four choices supported by evidence from clinical outcome trials?
Ezetimibe decreases LDL-cholesterol by 10–20%, but effects on triglycerides and HDL-cholesterol are minor. It increases the catabolism of LDL and possibly LDL apoB by upregulation of hepatic receptors.15 This is due to reduction of the hepatic cholesterol content secondary to inhibition of intestinal cholesterol absorption. ‘Dual inhibition’ of cholesterol absorption and synthesis has complementary effects on the catabolism of apoB-containing lipoproteins. Accordingly, in type 2 diabetic patients, ezetimibe plus lower-dose statin achieves LDL and non-HDL-cholesterol targets more effectively than higher-dose statins alone.24 The dose-sparing effect of ezetimibe decreases the risk of statin-related MSS side effects. Although adding ezetimibe to a statin results in an incremental fall in LDL-cholesterol, this does not consistently lead to reduction in CVD events, at least in patients with aortic stenosis.25 In type 2 diabetic subjects ezetimibe plus a statin can achieve regression of carotid intima-medial thickness (CIMT), a surrogate marker of atherosclerosis, but this relates to the fall in LDL-cholesterol and not to the minimal changes in triglycerides or HDL-cholesterol.26 Hence, the outcome data provide only tentative support for the role of ezetimibe in dealing with residual CVD risk in statin-treated patients.
Fibrates are synthetic agonists of peroxisome proliferator-activated receptor-α, and thereby transcriptionally regulate several genes that control lipid metabolism27 and vascular biology.28 Fibrates can decrease plasma triglyceride and LDL-cholesterol concentrations by up to 50% and 20%, respectively, as well as increase HDL-cholesterol by up to 20% and enhance the formation of large, less dense LDL particles.13 In the metabolic syndrome, fenofibrate significantly increases the catabolism of VLDL, LDL and LDL apoB, which relates to reduction in the production of apoC-III and activation of lipoprotein lipase and hepatic receptors.15 16 As with statins, fenofibrate does not decrease hepatic secretion of VLDL apoB. However, fenofibrate increases the plasma concentration of HDL-cholesterol by increasing the secretion of LpA-I:A-II particles.15
The cardiovascular benefits of fibrates have been confirmed by several clinical trials,29 two of which studied diabetic subjects alone.13 30 Overall, improvements in major cardiovascular events related to coronary events rather than strokes29 and were largely confined to volunteers with dyslipidaemia12 13 31–33 (see table 1).
It is noteworthy that none of these trials selected patients for having atherogenic dyslipidaemia, and particularly plasma triglycerides >2.2 mmol/l, a group in whom the efficacy of fibrates is likely to be greatest.26 In the Fenofibrate Intervention in Event Lowering in Diabetes (FIELD) study reduction in total cardiovascular events was significantly greater in patients with diabetic dyslipidaemia (triglyceride ≥ 2.3 mmol/l, HDL-C <1 mmol/l) than in those without dyslipidaemia.12 There was also improvement in diabetic microangiopathy,30 reflected by reduction in albuminuria, laser photocoagulation34 and minor amputations.35 The impact of an increased serum creatinine and homocysteine on the cardiovascular benefits of fenofibrate remains unclear. An increase in non-CHD deaths in early clofibrate trials has not been confirmed with fenofibrate,13 30 nor in the fibrate meta-analysis.29
The ACCORD-Lipid trial
The Action to Control Cardiovascular Risk in Diabetes (ACCORD)-Lipid trial recently reported on the efficacious effect of adding fenofibrate to ongoing simvastatin treatment on CVD events in type 2 diabetic patients who had CVD or at least two additional CV risk factors, in the setting of good glycaemic and blood pressure control.13 The trial randomised 5518 patients already at LDL-cholesterol target (∼2.6 mmol/l on a mean dose of 22 mg/day of simvastatin) to fenofibrate (54–145 mg/day) plus simvastatin (20–40 mg/day) or simvastatin (20–40 mg/day) alone; it is noteworthy that the parent trial targeted glycaemic control so that the lipid trial was not limited to atherogenic dyslipidaemia. Overall, after a mean of 4.5 years, the combination of fenofibrate and simvastatin did not significantly decrease the primary outcome (fatal cardiovascular events, non-fatal myocardial infarction, or non-fatal stroke) compared with simvastatin alone. Plasma triglycerides fell significantly with fenofibrate (−22.2% vs −8.7% with simvastatin alone) with a more modest but significant increase in HDL-cholesterol (+8.4% vs +6.0% with simvastatin alone), but there were no significant differences in LDL-cholesterol, which averaged 2.0 mmol/l in both groups. Fenofibrate significantly reduced the incidence of both microalbuminuria and macroalbuminuria. However, in a prespecified analysis, there was a 31% reduction (p<0.05 for within-group comparison) in cardiovascular risk in diabetic patients with triglycerides >2.3 mmol/l and HDL <0.9 mmol/l, who comprised approximately 17% of the trial population. The CV benefit in this subgroup is consistent with other data12 31–33 summarised in table 1.
The totality of evidence therefore suggests that fibrates are effective when triglycerides are elevated. Notwithstanding methodological limitations, the recent meta-analysis of fibrates suggested a 5% reduction in CVD events per 0.1 mmol/l improvement in plasma triglyceride concentration.29 In a diabetic patient a statin alone could on average achieve a proportional risk reduction in CVD events of 20% per 1 mmol/l fall in LDL-cholesterol,4 so that a decrease in triglycerides of at least 15% from adding a fibrate in a patient with pretreatment triglycerides of 2.3 mmol/l could theoretically achieve a further 17% reduction in CVD events.
The ACCORD-EYE trial
In a subgroup of 2856 patients, the effect of intensive treatment for hyperglycaemia (glycated haemoglobin <6% vs 7-7.9%), systolic blood-pressure (<120 vs <140 mm Hg) and dyslipidaemia (fenofibrate plus simvastatin vs simvastatin alone) on retinopathy was assessed using fundal photography.36 At 4 years, progression of retinopathy decreased significantly with improved control of glycaemia (10.4% vs 7.3%, p=0.003) and dyslipidaemia (10.2% vs 6.5%, p=0.006), but not with blood pressure control. Visual loss was not affected by any intervention. Progression of retinopathy with fenofibrate was less in men and in patients with previous retinopathy but the effect was apparently global and not confined to the subgroup with atherogenic dyslipidaemia. The benefits of intensive glycaemic control were offset by an increase in total mortality, weight gain and hypoglycaemia.37 That blood pressure control did not improve retinopathy may reflect the narrow range of blood pressure, small treatment effect and short duration of the intervention.36
What are the implications of ACCORD?
Targeting atherogenic dyslipidaemia
The lipid subgroup findings support recommendations that fibrates, and specifically fenofibrate, may be used to treat residual dyslipidaemia in diabetic patients receiving a statin.2 3 38 39 The number-needed-to-treat of 20 patients over 5 years to prevent one cardiovascular event implies cost-effectiveness. Because the study tended to recruit patients with low HDL-cholesterol, a more conservative estimate of the proportion requiring fenofibrate is ≤10%, especially in those treated with more potent statins.
ACCORD-EYE confirms that fenofibrate delays progression of diabetic retinopathy irrespective of overt dyslipidaemia, the number-needed-to-treat of 27 over 4 years also suggesting cost effectiveness. The findings may reflect improvement in generalised microvascular function, but this needs verification. Dyslipidaemia is a risk factor for diabetic retinopathy40 and the presence of both indicates the value of fenofibrate.
Safety of statin–fenofibrate combination
Concerns relate chiefly to risk of myopathy and rhabdomyolysis.41 This is most common with gemfibrozil, which has different pharmacokinetics from fenofibrate and is more likely to interact with statins, especially simvastatin. Importantly, rhabdomyolysis was not reported with fenofibrate–statin combinations in the recent trials. In ACCORD-Lipid there were also no significant increases in venous thromboembolic disease, transaminases, pancreatitis or non-CVD mortality with combination therapy.13 30 Elevation in serum creatinine with fenofibrate was moderate, rapidly reversible and associated with reduction in albuminuria, excluding significant nephrotoxicity. The exact significance of transient elevations in plasma homocysteine with fenofibrate remains unclear.
In pharmacological doses, niacin exerts a distinctive, global improvement in lipid and lipoprotein metabolism.16 Niacin and its derivatives can decrease plasma triglycerides and LDL-cholesterol by up to 35% and 15%, respectively, and increase HDL-cholesterol by 25%. Importantly, niacin also decreases plasma concentration of lipoprotein(a).16 The impact of niacin on plasma lipids and lipoproteins is dose dependent, the most efficacious and tolerable dose being 1.5 g/day. The mechanism of action of niacin remains unclear. However, it may decrease adipose tissue lipolysis and hence the flux of free fatty acids to the liver, which together with direct inhibition of hepatic triglyceride synthesis, decreases the secretion of VLDL and hence the production of LDL.15 16 The HDL-raising effect may be due to increased secretion of HDL-apoA-I and delayed catabolism of larger HDL particles.15 Like fibrates, niacin induces qualitative changes in LDL, with a shift of small dense LDL to larger buoyant LDL particles.
That niacin alone may decrease cardiovascular events and mortality in patients with coronary disease was suggested by the Coronary Drug Project (CDP) and, in particular, by its 9 year post-trial follow-up data,16 42 which also showed that benefit was independent of glycaemic status.43 This is important because it has been suggested that niacin impairs glucose/insulin homeostasis.44 Niacin significantly improves diabetic dyslipidaemia,45 and deleterious effects on glycaemia can be counteracted by adjusting antidiabetic drugs.46
The combination of a statin and niacin may diminish progression of atherosclerosis in high-risk patients, including diabetic subjects.47 The Arterial Biology for the Investigation of the Treatment Effects of Reducing Cholesterol-2 (ARBITER-2) study reported significant inhibition in progression of ultrasonographic CIMT over 12 months with a combination of a statin plus niacin compared with statin plus placebo in a group of patients with coronary disease, of whom 51% had the metabolic syndrome and 28% diabetes.48 The ARBITER-6—HDL and LDL Treatment Strategies in Atherosclerosis (HALTS) study—recruited patients with coronary heart disease, or risk equivalents, receiving long-term statin treatment with near-optimal LDL-cholesterol and low HDL-cholesterol, and showed that the addition of niacin-ER achieved significant regression of CIMT thickness compared with progression with addition of ezetimibe.49 This was seen despite significantly greater reduction in LDL-cholesterol with ezetimibe, and signifies benefit related to reduction in triglycerides and increase in HDL-cholesterol with niacin. Two large clinical trials, Atherothrombosis Intervention in Metabolic Syndrome with Low HDL-cholesterol/High Triglyceride and Impact on Global Health Outcomes (AIM HIGH)50 and Heart Protection Study-2: Treatment of HDL to Reduce the Incidence of Vascular Events (HPS-2 THRIVE)51 studies, which will recruit a significant proportion of patients with metabolic syndrome and diabetes, are currently testing whether adding niacin to a statin has an impact on CVD events.
Flushing is a dose-dependent and common side effect of niacin preparations,44 which is reduced with niacin-ER, the co-administration of niacin and aspirin, and the combination of niacin with laropiprant, a prostaglandin D2 inhibitor. The safety of extended-release niacin preparations in combination with statins has been confirmed in short-term studies. The significance of deterioration in glycaemic control with niacin remains a concern for non-diabetic and diabetic patients45 but this, as well as the long-term risk of myotoxicity and hepatoxicity, is being addressed by AIM HIGH50 and HPS-2 THRIVE.51
N-3 Fatty acid ethyl esters
Supplemental n-3 fatty acids (eicosapentaenoic acid and docosahexaenoic acid) dose dependently lower plasma triglycerides,15 particularly in hypertriglyceridaemic subjects. Doses of 3–4 g of eicosapentaenoic acid plus docosahexaenoic acid are required, which involves use of commercially available concentrates of n-3 fatty acid ethyl esters, such as Omacor. These are FDA approved as an adjunct to diet for lowering plasma triglycerides >5.5 mmol/l to mitigate risk of acute pancreatitis. n-3 Fatty acids improve triglyceride metabolism and hepatic steatosis by several mechanisms, including the transcriptional regulation of hepatic genes involved in lipogenesis and fatty acid oxidation.52 In the metabolic syndrome n-3 fatty acids decrease the hepatic secretion of large size VLDL,15 explaining the fall in triglycerides in those receiving a statin.53 In hypertriglyceridaemic patients, n-3 fatty acids increase the conversion of VLDL to LDL, but the elevation in LDL-cholesterol can be offset by a statin.15
The Gruppo Italiano per lo Studio della Sopravivenza nell'Infarto Miocardio GISSI-Prevenzione (GISSI-P) trial showed that 1 g/day of n-3 fatty acid ethyl esters, taken as one capsule of Omacor, reduced all-cause mortality and sudden death in patients with previous myocardial infarction.54 Reduction in major coronary events was also shown in both a primary and secondary prevention setting by the Japan Eicosapentaenoic Acid Lipid Intervention Study (JELIS) trial, in which subjects also received low-dose statin.55 These benefits of low-dose n-3 fatty acids are partially related to their antiarrhythmic effects and are independent of minor but statistically significant changes in plasma triglycerides. The CVD effects of n-3 fatty acid ethyl esters have recently been reviewed.56 Importantly, there are no published outcome trials of the efficacy of treating residual hypertriglyceridaemia with high-dose n-3 fatty acid ethyl esters in type 2 diabetic subjects receiving a statin.
Practical guidance: treating dyslipidaemia in a high-risk patient who is already receiving a statin
The recommended steps that could be followed for managing patients with diabetes, established CVD or an estimated moderate-to-high risk of CVD are summarised in the algorithm (see figure 1). LDL-cholesterol is the primary therapeutic target for statins. HDL-cholesterol is required to evaluate risk and calculate LDL and non-HDL-cholesterol. In patients with triglycerides >4.5 mmol/l, LDL-cholesterol should be measured directly or apoB used as a target. Elevated plasma triglycerides, particularly with low HDL-cholesterol, should be a trigger for considering further treatment with fibrates, niacin or n-3 fatty acids. Non-HDL-cholesterol or apoB could be used as secondary targets for managing hypertriglyceridaemia. When therapeutic targets are not achieved or maintained, compliance must be carefully evaluated and corrected before introducing new agents. Fasting or non-fasting blood samples after light meals may be used to assess patients, but care should be taken when interpreting plasma triglycerides after consumption of high-fat meals.
Statins and lifestyle modifications
Type 2 diabetes is a coronary-risk equivalent.2 Irrespective of pretreatment LDL-cholesterol levels all type 2 diabetic patients and non-diabetic subjects with established CVD should receive a statin with preferably low acquisition cost.39 Statins should also be used in primary prevention in moderate-risk or higher-risk patients without diabetes (ie, estimated 10-year risk of CVD of ≥20%) if the LDL-cholesterol level remains >2.6 mmol/l after 3–6 months of lifestyle changes.2 3 Secondary causes of dyslipidaemia, especially hyperglycaemia, must be dealt with and a healthy lifestyle emphasised. We estimate that at least 60% of all high-risk patients receiving a statin will require optimisation of treatment to lower LDL-cholesterol further and a smaller proportion, 10–15%, will need additional treatment for atherogenic dyslipidaemia.
Increasing the statin dose and adding ezetimibe
The statin dose should be increased to achieve an LDL-cholesterol target <2.6 mmol/l and lower (<1.8 mmol/l) if there is established CVD or diabetes plus other CVD risk factors (hypertension, albuminuria, smoking, family history of early CVD).3 If targets are not achieved with higher doses of weaker statins, ezetimibe may be added, or a more potent statin, such as atorvastatin or rosuvastatin, introduced. Ezetimibe has a dose-sparing advantage in patients intolerant of higher-dose statins, but there is no outcome evidence supporting its use. Colesevelam is an alternative to ezetimibe if incremental reduction in glycated haemoglobin in a diabetic is required,57 but gastrointestinal side effects and triglycerides must be monitored. These additional strategies may be superseded by the use of niacin in a statin-treated patient with atherogenic dyslipidaemia not at LDL-cholesterol goal, but this option needs confirmation by clinical outcome data.50 51
If there is residual hypertriglyceridaemia >2.2 mmol/l (especially with HDL-cholesterol <1.0 mmol/l) in patients receiving statin treatment,12 13 fenofibrate could be considered after dealing with secondary causes, including poor glycaemic control, obesity, diets high in refined carbohydrate, lack of exercise and smoking. Hypertriglyceridaemia with elevated HDL-cholesterol commonly points to alcohol excess and use of exogenous oestrogen. Because of adverse pharmacokinetic interaction with statins, we do not recommend gemfibrozil over fenofibrate in a patient already taking a statin.23 41 Bezafibrate is also more likely to interact with statins and the clinical trial evidence is limited.32 In our opinion, serious consideration should also be given to adding fenofibrate to a statin in type 2 diabetic subjects aged 40–70 years who have evidence of mild-to-moderate retinopathy,36 especially with triglycerides >1.5 mmol/l, and no contraindication to combination therapy. Given that at least 40% of type 2 diabetic subjects have retinopathy, the number of patients requiring fenofibrate for this indication will be greater than for overt, atherogenic dyslipidaemia alone. Fibrates should be avoided in patients with a history of cholelithiasis.
Niacin and n-3 fatty acid ethyl esters
By contrast to fenofibrate, the clinical outcome evidence supporting adding niacin or n-3 fatty acids to a statin is in our view more tenuous,16 but these agents are suitable alternatives if fenofibrate cannot be tolerated or is contraindicated. On the basis of surrogate end-point data48 49 and the Coronary Drug Project findings,42 the case for using niacin for residual dyslipidaemia is stronger than for n-3 fatty acids. However, there is good evidence for using low-dose n-3 fatty acid ethyl esters after a myocardial infarction.56
Very high triglycerides
Patients with triglycerides >5.5 mmol/l, and particularly >10 mmol/l, some of whom will have familial hypertriglyceridaemia and chylomicronaemia, are special cases not shown in the algorithm in whom a fibrate, niacin or high-dose n-3 fatty acids, together with a very-low-fat diet (<10% of calorie intake) and exercise are recommended as first-line treatment to prevent acute pancreatitis.2 3 39 Either one of these agents may also be added to statin treatment. Alcohol and dietary excess and poor glycaemic control in a diabetic must be addressed.
Secondary targets for hypertriglyceridaemia
Some authorities advise that with triglycerides >2.2 mmol/l, non-HDL-cholesterol (total HDL-cholesterol, a measure of all atherogenic lipoproteins including remnant particles) should be estimated and a therapeutic target set at 0.7 mmol/l higher than the corresponding LDL-cholesterol goal2 3; apoB is an alternative target at <0.8 g/l for patients with established CVD or diabetes plus other risk factors.3 In statin-treated patients, both non-HDL-cholesterol and apoB predict CVD outcomes better than LDL-cholesterol,58 consistent with other studies.21 The precise clinical yield of these secondary targets remains to be demonstrated. However, as indicated in the algorithm, they could be valuable for deciding whether to alter an existing statin regimen or to add niacin in patients with persistent hypertriglyceridaemia who are already receiving statin–fenofibrate combination therapy. Because of lack of clinical outcome evidence, this strategy could be reserved for the highest-risk patients.
Monitoring safety of combination therapy
Plasma transaminases, creatine kinase and creatinine should be measured at baseline. Creatine kinase should only be repeated if MSS symptoms are reported and treatment discontinued if levels exceed five times the upper limit of normal and/or symptoms are severe. Alanine and aspartate transaminases should be monitored 3 months after starting treatment and annually thereafter, or more frequently if uptitrating the dose of statin; hepatotoxicity is a potential problem with all drugs (except n-3 fatty acids) combined with a statin. Plasma creatinine should also be monitored with a statin–fenofibrate combination, although the increases reported in clinical trials are transient and not related to impaired glomerular filtration. All patients receiving niacin should have plasma glucose and urate monitored regularly because of possible hyperglycaemia and hyperuricaemia, respectively.
Concluding opinion: quo vadis after a statin?
The extant evidence provides a compelling argument for adding fenofibrate to a statin in type 2 diabetic patients who have residual dyslipidaemia and retinopathy (recommendation level A, evidence level 1b).12 13 34 36 There is less evidence for doing so in people who do not have diabetes but are at high risk of CVD, however, and no data in diabetic subjects who are young, elderly and have chronic kidney disease. Other options are niacin and n-3 fatty acid ethyl esters, with steadily growing but incomplete evidence supporting niacin as the preferred agent.48–51 In practice, drug compliance is a key issue that needs addressing with combination therapy,59 and this will require safer prescribing, improvement in doctor–patient alliance and reduction in drug costs.
This is a reprint of a paper that first appeared in Heart, March 2011, Volume 97, pages 350–356.
Competing interests GFW has received honoraria for educational activities and scientific advisory boards from Pfizer, AstraZeneca, Merck-Schering Plough, Abbott, Sanofi-Aventis, Glaxo-Welcome, Novartis and Genfit.
Provenance and peer review Commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.