13.6.3 Diabetic dyslipidaemia

Management of dyslipidaemia is an integral part of the multifactorial approach to cardiovascular disease (CVD) prevention in people with diabetes. In this chapter the pathogenesis of lipid and lipoprotein disorders in diabetes and their relationship to CVD risk will be discussed together with a practical approach to diagnosis and management.

Quantitative and qualitative lipid and lipoprotein abnormalities characterize diabetic dyslipidaemia (1). Moderately raised triglycerides, low high-density lipoprotein (HDL) cholesterol, and the accumulation of cholesterol-enriched remnant lipoprotein particles are the principal abnormalities. Total and low-density lipoprotein (LDL) cholesterol concentrations generally reflect those of the background population but LDL particle distribution is shifted to smaller, denser particles, which are thought to be more atherogenic Other factors which can affect the phenotype are shown in Table 13.6.3.1. Its pathophysiology is complex and not fully understood but there are strong correlations with insulin resistance. In insulin resistance, increased flux of non-esterified fatty acids (NEFAs) from visceral adipose tissue together with lack of inhibition of very low-density lipoprotein (VLDL) assembly leads to overproduction of VLDL, mainly large VLDL. This, together with chylomicra absorbed from the gut, saturate lipoprotein lipase (LPL) activity, producing prolonged postprandial lipaemia. LPL activity is also reduced by excess NEFAs and increased apoprotein C-III levels, which inhibit LPL activity, and relative insulin deficiency.

Prolonged postprandial lipaemia stimulates lipid exchange by cholesterol ester transfer protein (CETP), which mediates mole for mole cholesterol ester transfer from HDL to lipoproteins of lower density in exchange for triglyceride. This contributes significantly to the qualitative lipoprotein abnormalities. Partially hydrolysed triglyceride-rich lipoproteins or remnant particles become enriched in cholesterol. HDL becomes triglyceride-enriched and a substrate for hepatic lipase which is increased in insulin resistance. Smaller, denser HDL particles are formed which are cleared more rapidly contributing to decreased HDL levels. Low adiponectin concentrations contribute to decreased production of apoprotein A1, the major protein of HDL, and decreased ABC A1, an important peripheral binding site for HDL in reverse cholesterol transport.

Hypertriglyceridaemia is a major factor in determining LDL particle size. Kinetic studies suggest that it is large VLDL which relate strongly to the generation of small, dense LDL. CETP activity is again involved in lipid exchange producing triglyceride-enriched LDL particles which are substrates for hepatic lipase, the resultant triglyceride hydrolysis producing smaller, denser particles.

In type 1 diabetes with good glycaemic control lipid and lipoprotein concentrations are similar to the background population. HDL cholesterol can be higher due to enhanced phospholipid transfer protein activity although there is some evidence that the particle is less effective in protecting against oxidation. Although some qualitative changes have been described, metabolism of apoprotein B containing lipoproteins appears normal in kinetic studies. Albuminuria is associated with increases in LDL (small, dense particles), and triglyceride-rich lipoproteins and a reduction in HDL (I). Excess weight gain with intensive insulin therapy produces a similar lipid and lipoprotein profile to that seen in type 2 diabetes. There are reports of lipid and lipoprotein abnormalities leading to increased risk of microvascular complications but these findings need clarification.

Dyslipidaemia is strongly related to increased CVD risk in diabetes (2). In the United Kingdom Prospective Diabetes Study (UKPDS) LDL cholesterol was the best predictor of myocardial infarction. Based on the observational epidemiology, a 1 mmol/l increase in LDL is associated with a 57% increased risk (3). Small dense LDL particles are less effective ligands for the LDL receptor, a major determinant of LDL clearance from the circulation. More prolonged plasma residence time together with the smaller particle size facilitates penetration to the arterial subintimal space and increased retention due to enhanced binding to glycosaminoglycans. These particles are also more susceptible to oxidation and it is oxidized LDL which is central to many of the processes of atherogenesis (see Fig. 13.6.3.1).

HDL cholesterol concentrations are inversely related to CVD risk. In UKPDS, a 0.1 mmol increase was associated with a 15% decrease in CVD events (3). HDL is likely to protect through its role in reverse cholesterol transport, the removal of cholesterol from the periphery, including the arterial wall, back to the liver for excretion. HDL may also protect through anti-inflammatory, antioxidant, and antithrombotic activity.

In predicting CVD risk an established approach has been to use the total cholesterol (or LDL cholesterol) to HDL cholesterol ratio. Recently it has become clear that the major apoproteins of LDL and HDL—apoprotein B and apoprotein A-I, respectively—are important risk predictors. In the Collaborative Atorvastatin Diabetes Study (CARDS) the best risk predictor was the apoprotein B/apoprotein A-I ratio (4).

The triglycerides and CVD risk relationship is debated. Triglyceride accumulation is not a feature of the atheroma plaque but the cholesterol contained in some triglyceride-rich lipoproteins such as remnant particles contributes to plaque cholesterol. Postprandial triglycerides are intimately related to LDL and HDL metabolism. As a consequence it is probably unhelpful to use mathematical modelling to determine the independence of the triglyceride CVD relationship. Recently, a meta-analysis involving over a quarter of a million subjects has demonstrated an adjusted odds ratio of 1.72 (95% CI 1.56 to 1.9) between the top third and the bottom third of the triglyceride distribution (5). Nonfasting triglyceride concentrations appear to predict CVD risk better than fasting levels.

Statins are indicated for in the majority of patients. They reduce LDL and other apoprotein B-containing lipoproteins effectively with minimal side effects. In addition, large, placebo-controlled CVD endpoint trials provide an extensive database to inform clinical practice (6).

The Heart Protection Study (HPS) included 2912 diabetic patients (mainly type 2) without CVD, and nonfasting cholesterol >3.5 mmol/l. Simvastatin reduced CVD events by 33% independent of baseline lipids, diabetes duration, glycaemic control, and age (7). In CARDS, 2838 type 2 patients, LDL cholesterol ≤ 4.14 mmo/l and one other CVD risk factor, received atorvastatin (10 mg/day) or placebo. The trial was stopped early as the prespecified early stopping rule for efficacy was met. Atorvastatin reduced CVD events by 37% independent of baseline lipids, age, diabetes duration, glycaemic control, systolic blood pressure, smoking, and albuminuria. Nonhaemorrhagic stroke was reduced by 50% (8).

The trials including diabetic patients are shown in Table 13.6.3.2. There were 202 known diabetic patients in the Scandinavian Simvastatin Survival Study (4S). Half of those on placebo experienced an event during the 5.4 years of follow-up (Fig. 13.6.3.2). Simvastatin reduced CVD events by 55%. 4S and the other trials show that diabetic patients benefit from statins in a similar way to those without diabetes but a high residual risk remains. In HPS, residual risk of a major CVD event in diabetic patients with coronary disease receiving simvastatin remained higher than in nondiabetic patients with coronary disease on placebo (7).

More intensive LDL-lowering with higher statin dose or a more potent statin results in additional benefit. In a meta-analysis of 28 000 patients in four trials of intensive treatment the authors calculated that for every million patients with chronic or acute coronary disease treated more intensively for 5 years, 35 000 CVD events would be saved with a number needed to treat of 29 over 2 years for ACS and 5 years for stable disease (9). Diabetic patients were not analysed separately but it is likely that they would show similar benefit. The Cholesterol Treatment Trialists’ (CTT) collaborators identified 18 686 diabetic patients in 14 major endpoint trials providing 3247 CVD events over a mean follow-up period of 4.3 years. Most patients had type 2 diabetes but 1466 individuals with type 1 disease were identified. Major CVD events were reduced by 21% for each mmol/l reduction in LDL cholesterol. Those with type 1 diabetes showed no heterogeneity of effect of statin therapy. The large number of events permitted useful assessment of many different patient groups by baseline characteristics (Table 13.6.3.3). Absolute size of benefit was determined mainly by the absolute reduction in LDL cholesterol achieved and a statin regimen leading to a substantial LDL cholesterol reduction was recommended (10).

Diet and lifestyle should be optimized as far as is practical for the individual with special focus on smoking cessation, weight reduction, prudent diet, and increasing physical activity.

Statins are first line therapy for the majority of patients. Given the benefits of intensive statin therapy, a new goal of therapy, LDL cholesterol <1.8 mmol/l has been proposed in the USA for those at highest risk, regardless of baseline LDL cholesterol (11). The Joint British Societies’ (JBS) guidelines in the UK have lowered the LDL cholesterol goal to 2 mmol/l (12).

Absolute risk rather than LDL cholesterol concentration should determine decisions on therapy. Most patients aged over 40 years will fulfil the accepted risk threshold for statin therapy, 20% CVD risk over 10 years. Some younger patients will be at high risk and JBS2 has suggested that some patients aged 18–39 years should be considered for therapy (Table 13.6.3.4). Lifetime CVD risk in type 1 diabetes is high and current guidelines do not distinguish between type 1 and type 2.

Atherogenic cholesterol is carried on particles (e.g. remnants) other than LDL in diabetes. When LDL cholesterol is to goal and plasma triglycerides remain ≥2.0 mmol/l, a secondary goal is non-HDL cholesterol (total cholesterol- HDL cholesterol) set at 0.8 mmol/l above the LDL goal (12). The first approach is to increase the statin dose and switch to a more potent statin if necessary. If the treatment goal is not reached then combination therapy is considered.

Statin choice depends on factors such as efficacy, safety, other medication, and medical conditions, Randomized controlled trial (RCT) evidence of CVD reduction, baseline lipids and, increasingly, cost, given the availability of generics. Statins differ in potency, the most potent being atorvastatin and rosuvastatin. The drugs are safe if used appropriately and drug interactions avoided. Simvastatin, lovastatin, and atorvastatin are metabolized through cytochrome p 450 (CYP) 3A4 so are best avoided in patients taking drugs that inhibit this pathway. Large amounts of grapefruit juice can also inhibit CYP 3A4. Atorvastatin appears to be less susceptible to this interaction than simvastatin. Fluvastatin is metabolized through CYP 2C9 and rosuvastatin through CYP 2C9 and 2C19. Pravastatin is not metabolized through the cytochrome system. Gastrointestinal disturbances, weakness, headache and general aches and pains are common side effects. However, in RCTs which provide the best unbiased data, there is little difference in these symptoms between active and placebo groups. Several possible adverse reactions including sleep and mood disorders, dementia and peripheral neuropathy have been reported spontaneously but not seen in RCTs (13).

Myopathy characterized by painful, tender muscles often with flu-like symptoms is vanishingly rare. Creatine phosphokinase (CPK) is at least 10-fold increased. Patients are warned to stop the drug if these symptoms develop. Myopathy resolves on stopping the drug. Very rarely acute tubular necrosis can occur following rhabdomyolysis. I do not measure CPK routinely except in complex patients at risk of drug interactions. CPK levels vary enormously and the normal range is higher in black patients. Other causes of raised CPK are exercise and hypothyroidism. Although advocated in some guidelines, I do not use simvastatin at a dose of 8omg/day because of an increased risk of myopathy. Liver abnormalities are rare even with high doses. If transaminases are greater than threefold increased (c. 1 in 400) dosage is reduced. Many patients have abnormal liver function due to fatty liver and these patients should not be denied a statin. Women of childbearing potential should use effective contraception and stop the statin at least 6 weeks prior to conception.

Diabetic patients may develop severe hypertriglyceridaemia, fasting triglycerides ≥11 mmol/l, from a combination of exogenous and endogenous particles, namely chylomicra and VLDL. Diabetes alone does result in such high levels and there is usually an underlying lipid disorder such as familial combined hyperlipidaemia. Hypothyroidism, high alcohol intake and renal disease should be excluded. Recurrent abdominal pain and pancreatitis can occur. Hepatosplenomegaly due to accumulation of lipid-laden macrophages may be present. Rarely there may be memory disturbances and lack of concentration. Some patients develop eruptive xanthomata The measurement of other analytes such as haemoglobin, bilirubin, and liver transaminases may be affected and, because of decreased water volume in plasma, sodium levels may appear low. The assay for amylase may also be affected.

Treatment is urgent given the risk of pancreatitis. A low total fat diet together with reductions in alcohol and refined carbohydrate are important. High-dose omega-3 fish oils are useful together with a fibrate or nicotinic acid. As diet and lifestyle measures progress it is often possible to stop these medications. If significant mixed lipaemia persists a statin is indicated.

The most common reason for statin discontinuation, in my experience, is muscle ache generally without significant CPK elevation. Benefits of therapy need to be re-emphasized together with an explanation that the drug is unlikely to be at fault. However, dose reduction is sometimes necessary. If a small dose is tolerated and LDL cholesterol is not to goal, the addition of ezetimibe, a potent specific inhibitor of cholesterol absorption, may provide a further 20–25% LDL cholesterol lowering. Side effects of the combination are largely similar to the statin alone. Some patients refuse to take statins despite best efforts by the physician. In these cases it is worth trying other drugs. As a sole agent, ezetimibe reduces LDL cholesterol by 15% although response varies depending on individuals’ ability to absorb cholesterol.

Fibrates, which are agonists for peroxisome proliferator activator receptor α (PPARα), have modest effects in reducing LDL cholesterol. Their main effects are to reduce triglycerides and increase HDL cholesterol. Outcome data with the fibrate class is mixed. Gemfibrozil has the best RCT evidence but the drug should not be combined with a statin as it can increase plasma levels through interaction at glucuronidation sites involved in drug metabolism. Bezafibrate has a better impact on LDL cholesterol but the RCT evidence relates only to post hoc analyses. In a long-term RCT, fenofibrate increased HDL cholesterol by just 2% and the primary CVD endpoint was not reached. Information on the potential benefits or otherwise of the statin/fenofibrate combination will become available from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study lipid arm (www.accordtrial.org).

Bile acid sequestrants, reduce LDL cholesterol by up to 30% and their use is supported by RCTs but they are not well tolerated because of the inconvenience of their administration and gastrointestinal effects. In patients on multiple drug therapies resin administration is difficult because of the potential to decrease absorption of other drugs.

Nicotinic acid, which has a major effect on hepatic VLDL synthesis and output and HDL metabolism, could be the ideal agent but it is poorly tolerated due to cutaneous flushing. There is no definitive RCT to guide therapy but there are studies showing benefit on surrogate endpoints. An extended-release (ER) preparation is better tolerated in terms of flushing. Most physicians use ER nicotinic acid at doses of 2 g/day to lower triglycerides and increase HDL in combination with a statin. At this dose adverse effects on insulin sensitivity and glucose tolerance are less marked. ER nicotinic acid alone will reduce LDL cholesterol by about 18%, increase HDL cholesterol by 19% and reduce triglyceride by 21%.

In a recent consensus from American Diabetes Association and American College of Cardiology (14), the advantage of using apoprotein B as a therapeutic target in patients with cardiometabolic risk has been raised as a better indicator of the level of atherogenic lipoproteins.

An approach currently in trial is to raise HDL cholesterol in addition to intensive LDL cholesterol lowering. HDL cholesterol remains a risk predictor even in patients on intensive statin therapy. Fibrates have traditionally been used to increase HDL cholesterol but their effects are modest and good RCT evidence is limited to gemfibrozil, which should not be used concomitantly with statins. Consistent effects on HDL cholesterol are seen with the PPARγ agonists particularly pioglitazone. In a large study of the effects of this agent on carotid intima/media thickening (IMT) the benefit of the drug in stopping IMT progression has been attributed to this effect (15).

Nicotinic acid at a dose of 2 g/day increases HDL cholesterol approximately 20%. Trials with ER nicotinic acid added to statin therapy have been encouraging in surrogate endpoint trials but large robust CVD event point trials are required. In AIM-HIGH (Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High triglyceride and Impact on Global Health outcomes), a secondary prevention trial, the combination of simvastatin with ER nicotinic acid is being compared to simvastatin alone at comparable LDL cholesterol levels (www.aimhigh-heart.com). ER nicotinic acid and the prostaglandin D2 receptor blocker, laropiprant which reduces flushing is being tested in a major RCT. HPS2-THRIVE (Treatment of HDL to Reduce the Incidence of Vascular Events) will recruit 25 000 men and women aged 50–80 years with CVD including 7000 diabetic patients. The trial will test the potential benefit of adding nicotinic acid/laropiprant to optimal statin therapy (www.ctsu.ox.ac.uk). These trials should provide robust evidence as to the potential benefit of a more global approach to lipid management.