Chapter 4 – Management of Dyslipidemia in the Elderly

Updated 5 January 2011

INTRODUCTION

The treatment of hyperlipidemia is a relatively new concept in medical therapeutics. A relationship between serum cholesterol and the risk of cardiovascular disease has been firmly established in the early 1960's, but only in 1984 the LRC-CPPT trial (1) was the first to confirm the "cholesterol hypothesis", that is that lowering of serum cholesterol results in a reduction in cardiovascular morbidity. In the past 25 years multiple clinical trials have documented the cardiovascular benefit of treatment of hyperlipidemia. Drugs used are usually referred to as "lipid lowering "rugs although their mechanism of action might not be limited to a decrease in the number of atherogenic lipoprotein particles. In spite of the successful use of these drugs, treatment of hyperlipidemia in the "elderly", irrespective of the definition of this group, has remained the subject of controversy. In this chapter we present the evidence of the benefit and risk of lipid lowering in the older age group.

Characteristics of the Aging Process Relevant to Lipid Intervention

 

The US population is aging. In the last 100 years the life expectancy has increased by 30 years. In 2004, 12.4% of US population was in the Medicare age group and this percent is expected to increase progressively in the future (2). In this segment of the population, cardiovascular disease is by far the main cause of mortality in both men and women. This justifies cardiovascular prevention as a subject of academic research and policy debate. Lipid pharmacological intervention is one of the most successful cardiovascular preventative interventions. Concerns about its safety and efficacy in this age group have led different countries to adopt different strategies concerning the use of lipid lowering drugs in the elderly. This is based on physiologic, socio-economic and ethical considerations.

Physiologic characteristics of aging:

Atherosclerosis is a continuous process and its burden increases progressively with age. The pathology includes continuous remodeling of the vascular wall and in final stage, occurrence of calcification. Although lipid lowering is statistically successful, its success might depend on the pathology of the arterial wall. The current belief is that the main mechanism of action of lipid lowering drugs is the stabilization of atherosclerotic plaques. A study using intravascular ultrasound has documented that in young survivors of myocardial infarction the culprit vessel is undergoing constrictive remodeling, usually associated with plaque erosion, while in older subjects expansive remodeling predominates (3). Expansive vascular remodeling results in increased irreversible arterial stiffness, particularly in calcified vessels and in fibrotic plaques (4). This in turn decreases the likelihood for a cardiovascular event to be attributable to a ruptured plaque and increases the likelihood of it being related to increased arterial stiffness.  It is conceivable that in presence of advanced atherosclerosis, lipid intervention might be less successful. On the other hand, the safety of using lipid lowering drugs is diminished in the older age group. The CYP-450 enzyme system and the esterases responsible for drug metabolism may become dysfunctional in the older age group. The kidney function decreases linearly with age further diminishing the excretion of active drugs and their metabolites. All these factors contribute to a modification of the risk/benefit ratio of preventative interventions. In addition, aging increases the number of co-morbidities requiring pharmacologic intervention and this in turn results in polypharmacy. This further decreases the safety of lipid lowering drugs.

Socio-economic characteristics of aging:

The average household net worth increases progressively with age up to age 64, and then it declines slightly (5). In spite of this, almost 10% of people in this age group live in poverty. Gender and ethnicity are important factors since the percent of women living in poverty is double that of men, Asians and Hispanics are twice as likely to live in poverty as Non-Hispanic Whites and African-American three times more likely. The average household net worth is five times lower in Blacks than in Whites. Social isolation is an additional risk factor characteristic for this age group since a third of the women aged between 65 and 74 and half of the women 75 years or older are living alone. In couples, disability of one of the spouses places the other one in the situation of caregiver. This results in an increase in cardiovascular risk for the healthier of the two. The problems are frequently compounded by the occurrence of cognitive impairment affecting understanding of the concept of preventive health care. All these factors contribute to limitations in access to health care and therapeutic modalities.

Ethical considerations in the very old:

There are ethical problems, which need to be addressed when selecting a preventative management program in the elderly. The main question is addressing the value of life versus that of the quality of life. There is a paucity of data concerning the wishes and fears of subjects in this age group. At what point in time is cardiovascular prevention no longer valuable? What is the worst fear of patients in this group, death or disability? If the fear of physical disability prevails, is it the fear of physical disability or of cognitive impairment? At the society level, this is compounded by the cost to society of prolonging a life of doubtful quality to the bearer. All these considerations affect the decisions to treat patients in the older age group with lipid lowering drugs.

LIPOPROTEINS & AGING

 

Biochemistry and Metabolism of Lipoproteins

Lipoproteins are large particles with a hydrophilic surface and a hydrophobic core. The structure of the surface consists of protein referred to as apolipoproteins and phospholipids. The core contains free and esterified cholesterol and triglycerides. The apolipoproteins are responsible for the binding of the lipoproteins to receptors and enzymes and therefore direct lipoprotein metabolism. Table 1 shows the lipoprotein classes:

Table 1: Lipoprotein classes
Lipoprotein	Density (g/mL)	Major Lipid Component	Major Apolipoproteins
Chylomicrons	<0.95	Dietary Triglycerides and Cholesterol Esters	A-I, A-II, A-IV, B48, C1, C2, C3, E
Chylomicron Remnants	<1.006	Dietary Cholesterol Esters	B48, E
VLDL	<1.006	Endogenous Triglycerides	B100, C1, C2, C3, E
IDL	1.006-1.019	Endogenous Cholesterol Esters	B100, E
LDL	1.019-1.063	Endogenous Cholesterol Esters	B100
HDL2	1.063-1.125	Cholesterol Esters and Phospholipids	A-I, A-II, C1, C2, C3, E
HDL3	1.125-1.210	Phospholipids	A-I, A-II, C1, C2, C3, E

Chylomicron metabolism:

Chylomicrons are assembled in the intestinal wall and transported to the general circulation via lymphatics. During the transport they exchange with HDL Apolipoproteins A-I and A-II and triglycerides for Apolipoproteins C, E and cholesterol esters. In the capillary bed, chylomicron particles are exposed to endothelial lipoprotein lipase and their triglyceride content is rapidly depleted. The smaller residual particle is referred to as chylomicron remnant. These particles are removed by the liver through a receptor mediated uptake recognizing Apo E. The entire process of chylomicron metabolism is relatively rapid and the presence of chylomicrons in the serum after an overnight fast is considered an indicator of defective lipoprotein metabolism (6-12).

VLDL metabolism:

Very low density lipoproteins (VLDL) are a spectrum of particles of different sizes and densities. They are assembled in the liver and their secretion is regulated primarily by the flow of fatty acids and their intra-hepatic disposition. The free fatty acids are derived from intrahepatic synthesis, from adipose tissue triglyceride hydrolysis or from the metabolism of chylomicrons. After their secretion, like chylomicrons, they exchange with HDL Apolipoproteins A-I and A-II and triglycerides for Apolipoproteins C, E and cholesterol esters. They share with chylomicrons a common pathway for catabolism involving enzymatic triglyceride hydrolysis and receptor mediated uptake. An abundant production of VLDL particles flooding the removal mechanism may result in an accumulation of chylomicrons. Different particle sizes are removed at different rates with larger VLDL particles being removed faster or being transformed in smaller VLDL particles. The smaller denser fraction resulting from VLDL catabolism is referred to as intermediate density lipoprotein (IDL). The average half-life of a VLDL particle is 12 hours. The regulation of VLDL metabolism is complex and depends on the rate of particle production, enzymatic transformation and receptor mediated uptake (13-18).

LDL metabolism:

Low density lipoproteins (LDL) are the main cholesterol carrier in healthy humans. They are derived mostly from VLDL metabolism and only a small fraction is secreted de novo. LDL particles are also separable in a spectrum of sizes or densities. Larger, more buoyant LDL particles are derived from average size VLDL. Smaller denser LDL particles are derived mostly from very large VLDL particles, which are increased in pathological states. These types of LDL particles are thought to be more atherogenic. The removal of LDL from the circulation is largely a receptor mediated mechanism. A small fraction of the particles is removed by a non receptor mediated mechanism. The apolipoprotein recognized by the receptor is Apolipoprotein B. The receptors are expressed in all cells but the liver uptake accounts for the largest part of particle removal. The half-life of LDL particles is of approximately three days. The variations in LDL concentration are mostly determined by the rate of LDL catabolism. During its catabolism LDL, undergoes a process of oxidations of its cholesterol content which renders it more atherogenic (19-25).

HDL metabolism:

High density lipoprotein (HDL) particles are responsible for reverse cholesterol transport. The first step in the synthesis of HDL is a small, discoid particle containing Apo A1 and phospholipids. Apo A1 is synthesized in the liver or the small intestine and acquires the phospholipids in these organs immediately after synthesis. The native particles undergo a process of lipidation by acquiring cholesterol from all cells. The removal of the cholesterol from the macrophages of the arterial wall acts as a major anti-atherosclerotic mechanism. The process of transfer of cholesterol from the cells to HDL has multiple pathways:

• Active mechanism via ABCA1 cassette receptors
• Transfer of free cholesterol by passive diffusion and subsequent esterification by Lecythin-Cholesterol-Acyl-Transferase (LCAT)
• Transfer via SR-B1 receptors
• Transfer via caveolins
• Transfer via the Cyp 27A1 pathway

Of these the first one is the most important. The pathway is dependent on the activation of a nuclear receptor, liver X receptor (LXR), by oxysterols resulting from LDL catabolism. The mature HDL continues to accept cholesterol from the cellular pool. It also undergoes a process of lipid content transfer. Cholesterol ester transfer protein (CETP) mediates the transfer of cholesterol esters to VLDL particles and the transfer of triglycerides to HDL from VLDL. Triglycerides are rapidly hydrolyzed by Hepatic Lipase (HL) and this process accelerates the removal of HDL particles. The hepatic uptake of HDL particles is essential for determining HDL concentration and therefore the effectiveness of reverse cholesterol transport. Its physiology has not been understood to date (26-50).

Lipoprotein (a):

Lipoprotein (a) is a lipoprotein containing an Apo B100 containing lipoprotein particle indistinguishable from LDL, covalently linked with a unique, carbohydrate rich molecule, Apo (a). Apo (a) has a structural homology with the fibrinolytic proenzyme, plasminogen.
Lipoprotein (a) is highly atherogenic and prothrombotic and might directly promote remodeling of the arterial wall. Its metabolism does not parallel LDL metabolism (51-53).

Besides their structural proteins, lipoproteins transport proteins with functions other than lipoprotein metabolism, conferring special properties to the specific lipoprotein. Table 2 contains a synopsis of apolipoproteins identified to date.

 

Table 2: Apolipoproteins
Apo-lipoprotein	Molecular Weight	Amino acids	Lipoprotein	Function	Associations with  genetic Polymorphism
Apo A1 (54)	28.1 kDa	243	HDL	Main structural apolipoprotein of HDL	Apo A1-C-III-A-IV polymorphism associated with atherosclerosis Apo A1 Milano associated with reduced cardiovascular risk
Apo A2 (54, 55)	18.4 kDa	77	HDL	Structural apolipoprotein of HDL	Associated with waist/hip ratio and triglyceride metabolism
Apo A4  (56, 57)	46 kDa	376	HDL, Chylomicrons	Antioxidant Anti-inflammatory Activator LCAT Regulator food intake	Apo A1-C-III-A-IV polymorphism associated with atherosclerosis
Apo A5 (58, 59)	39 kDa	366	HDL, VLDL	Regulator of triglyceride metabolism	Associated with triglyceride level
Apo B100 (60)	?	4563	LDL, VLDL	Main structural apolipoprotein of LDL & VLDL	Apo B 3500 single point mutation associated with atherosclerosis
Apo B48 (61)	?	2152	Chylomicrons	Main structural apolipoprotein of chylomicrons	Might be associated with chylomicron metabolism
Apo C1 (54)	6.6 kDa	57	VLDL, HDL	Activator LCAT Inhibitor CETP	Associated with cholelithiasis, Alzheimer's disease
Apo C2 (54)	8.8 kDa	79	VLDL, HDL	Main Activator LPL	Apo C2 deficiency is associated with chylomicronemia
Apo C3 (54)	8.8 kDa	79	VLDL, HDL	Main inhibitor ot VLDL catabolism	Associated with hypertriglyceridemia
Apo D (62)	19-32 kDa	189	HDL, VLDL	Associated with neuro-psychiatric pathology and metabolic syndrome	Associated with Alzheimer's disease and increased susceptibility to schizophrenia
Apo E (54, 63)	34.2 kDa	299	VLDL, HDL	Regulator of removal of atherogenic protein remnants	E2 isomorph and point mutations result in type III dyslipoproteinemia. E4 isomorph is associated with atherosclerosis and Alzheimer's disease
Apo F (64)	33 kDa	?	VLDL, HDL	CETP inhibitor	Unknown
Apo H (65)	50 kDa	326	HDL	Anticoagulant	Associated with antiphospolipid antibody, SLE and cerebrovascular accidents
Apo J (66)	80 kDa	449	HDL	Cell-cell interactions Apoptosis	Associated with HDL concentration, preeclampsia and pseudoexfoliation syndrome
Apo L (67, 68)	41 kDa	371	HDL	Apoptosis Associated with hypertriglyceridemia	Associated with hypertriglyceridemia
Apo M (69)	21.3 kDa	188	HDL, VLDL	Antiatherogenic Antioxidant	Associated with atherosclerosis and Alzheimer's disease
SAA4 (70, 71)	12-14 kDa	?	HDL	Acute reactive protein	Unknown

Changes of Lipoproteins with Aging

The most recent average levels for lipoproteins in the US are reported in the NHANES publications (72) }. . Table 3 shows these levels for the older age groups.
Table 3. Average Lipoprotein Levels in the US 2003-4 and Percent at NCEP goal.

	Age: 60-69		Age: >70
	Concentration mg/dl	Percent at Goal	Concentration mg/dl	Percent at Goal
LDLc	126.1	52.8	118.8	45.8
Non HDLc	156.7	53.3	147.8	48.8
HDLc	54.1	72.1	56.2	79.5
Trig.	145.3	60.6	132.7	69.6

There were ethnic differences in the distribution of these concentrations: Hispanic subjects of both gender had higher total cholesterol and triglycerides and lower HDL cholesterol while Nonhispanic black women had higher HDL cholesterol and lower triglyceride level. There was a progressive decrease in the level of total cholesterol, a trend towards increase in the triglyceride level from 1960 to 2002. The trends are particularly striking in older subjects and are probably accounted for by the increased prevalence of treatment with lipid lowering drugs and of obesity. Across all age groups, triglycerides increased with aging and reached a peak in men aged 50-59 and women aged 60-69. Apo B increased progressively with age from the younger group to the elderly groups, and this was associated with an increased prevalence of small dense LDL. HDL cholesterol did not seem to vary with age.

The data show an increase in the number in atherogenic particles with age and a decrease in their number in the oldest population group. As a result, studies show changes occurring with age depending on the population studied. This variable trend with age can be explained by three factors: comorbidity, attrition (survival bias) and frailty.

Comorbidity is at least in part associated with the incidence of metabolic syndrome. Most studies have shown that the prevalence of metabolic syndrome increases with aging (73-77). Aging parallels the changes occurring with metabolic syndrome in both carbohydrate and lipid metabolism. The total cholesterol, Apo B, the prevalence of small dense LDL and triglyceride concentration are increased while HDL cholesterol decreases. Lipoprotein kinetics studies have shown a decrease in fractional clearance rate of VLDL, IDL and LDL Apo B as well as an increase in VLDL Apo B production (78-82). The Apo A1 kinetics shows an increase in both production and removal, and the changes are more pronounced in men (83-85). On the other hand, metabolic syndrome is a strong risk factor for increased mortality through cardiovascular disease and cancer (86-92), and a large number of patients is lost in the transition from younger to an older cohort (survivor bias). Longitudinal studies show the trend for decrease in VLDL and LDL concentration with age to be less striking when compared with cross sectional studies.

The increase in the prevalence of metabolic syndrome with age is shown in national data. In the age group 60-69 it approaches 50% of the population, with the highest prevalence in Hispanic and Nonhispanic black women. In postmenopausal women, the lipids correlate well with BMI, insulin levels and the amount of intra-abdominal fat showing again the strong impact of metabolic syndrome on lipoproteins in older subjects (93-96). The increased prevalence of metabolic syndrome in the old is not associated with changes in energy intake but rather with decreases in energy expenditure and this in turn is associated with decreased functioning (97-100). In addition the trends towards reduction of body weight seen in longitudinal studies of elderly subjects are associated with a disproportionate reduction in lean body mass, further decreasing the ability to function (101).

As related to the increased prevalence of metabolic syndrome, there is an increased prevalence of diabetes in the elderly . Disorders of carbohydrate metabolism do not increase with age after adjustment for visceral obesity (98). The presence of diabetes in suboptimal glycemic control results in abnormalities of lipoproteins of the same type as those of metabolic syndrome but much more severe, in proportion to the degree of hyperglycemia. Other causes of secondary hyperlipidemia, which are more prevalent in the elderly age group, are those of untreated or sub-optimally treated hypothyroidism, chronic kidney disease and liver disease. These have to be ruled out at the time of the assessment of lipoproteins.

Attrition contributes by selecting in the older population groups survivors with lower levels of major risk factors. Consequently the prevalence of severe familial dyslipidemias diminishes in the older age groups.

Frailty is a syndrome associated with aging and increasing in frequency with age. It is usually associated with a lowering of total, LDL and non-HDL cholesterol (102, 103).  While these lipoproteins decrease with age, the average Lp (a) level seems to be unchanged. In centenarians the level of Lp (a) parallels the level of inflammatory markers (104).  It is conceivable that the stable levels of Lp (a) with aging represent the effect of two opposite effects: attrition tending to diminish the levels and frailty increasing the levels and paralleling the rise of inflammatory markers.

Socioeconomic factors, comorbidity and treatment with lipid lowering drugs also contribute to changes in lipoproteins in the elderly. By 2003, 10.2% of the population over age 65 years was living under the nation's poverty level (5). Of these 41.6% reported not having natural teeth. As measured by Healthy Eating Index, only 8.8% of the people over age 65 below the poverty level have a ""good� rating (2). Socioeconomic factors are most of the time associated with comorbidity. Neurological or psychiatric disorders may result in decreases in socioeconomic status and ultimately result in malnutrition. Cancer and frequent hospitalizations contribute to the deterioration of the nutritional status. In some studies, a paradoxical association of high non-HDL cholesterol with survival or recovery from disability in basic activity of daily living was documented (105). An analysis of the data shows a trend towards increased risk of myocardial infarction and stroke and a decreased likelihood of disability and cognitive impairment to be associated with higher non-HDL cholesterol.

 LIPIDS, CARDIOVASCULAR RISK AND AGING

 

Total Cholesterol, LDL Cholesterol and Non-HDL Cholesterol as Risk Factors

The current NCEP guidelines for treatment of adults with hyperlipidemia (106) do not have age limits but specify that "clinical judgment" is necessary for management of patients older than 65 years. The AHA Evidence-based Guidelines for Cardiovascular Disease Prevention in Women: 2007 update (107) specifies that "many studies used to formulate recommendations did not include older women, especially those >80 years of age". Because of this gap in knowledge the need for clinical trials in the very old is being raised frequently. In the Rotterdam study, the use of the Framingham score to predict occurrence of a first coronary event in  elderly subjects resulted in gross underestimation, increasing with increasing age and  so  in men than in women (108). In the Leiden-85 study, in healthy patients over age 85 years, the Framingham score did not predict cardiovascular death (109) The best predictor was homocysteine but the authors did not correct the data for indices of renal function which are usually associated with homocysteine levels..

The fact that the number of atherogenic particles, expressed as total cholesterol or LDL cholesterol or Non-HDL cholesterol or Apo B concentration, is a risk factor for coronary artery disease was known since the 1950's. Numerous cohorts, including elderly enrollees reported a high risk of coronary artery disease for subjects with high but also with low cholesterol concentrations (99, 102, 110, 111). This phenomenon was interpreted as a loss of ability of total cholesterol to predict coronary events, increasing with age (112). This lack of association of coronary disease with total cholesterol levels seems to be more striking in elderly women (113). A recent meta-analysis of the relationship between total cholesterol and coronary events shows a significant association for men aged 65 to 80 years but none for women over 65 years or for men over 80 years (114). Non-HDL cholesterol does not appear to be a better predictor of cardiovascular risk in older subjects. The loss of power of total cholesterol concentration to predict events is shared by other risk factors such as systolic blood pressure, BMI, cigarette smoking and alcohol. This difficulty in identifying subjects at higher cardiovascular risk in older populations was attributed to multiple confounders.

The most important confounder is frailty. With aging the prevalence of frailty increases, and since frailty results in lower cholesterol level and higher risk of death, high total cholesterol in the very old is shown in many studies to be a marker of longevity (115). Low cholesterol is associated with a poor prognosis in the very old and predicts an increased risk of death from infection and cancer (116). Although frailty is associated with weight loss, obesity is not a negative risk factor for frailty, and some authors believe that it might predispose to frailty later in life (117). Since frailty may occur in patients with advanced atherosclerosis, low cholesterol may predict cardiovascular events. This results in a "J" curve of relationship between cholesterol and coronary risk (103). The highest risk of death, including death from cardiovascular disease, is present in subjects with low cholesterol and low BMI or low serum albumin (118). The association between low cholesterol and frailty is also accountable for data showing low cholesterol as a predictor of dementia (119, 120). Because of the fact that there is no consensus of an easy definition of frailty, there are no data showing the relationship between total cholesterol and coronary disease corrected for frailty

Other confounders are "survivor bias" and the "regression dilution" of cholesterol levels when multiple measurements are available Survivor bias eliminates the extreme values of all risk factors, thus decreasing the power of the regression. Regression dilution results from using the mean of multiple readings as a predictor, resulting in the values regressing towards the mean of the population and reducing the variability (121, 122). Another correction was presented from the Framingham Study introducing the concept of "life-time risk of a coronary event" which is higher for subjects with high cholesterol at all ages (123, 124). Multiple studies have shown that elevated cholesterol in mid-life predicts coronary death in old age (125-127). For other forms of atherosclerosis, the predictive effect of serum cholesterol is variable.

 

Cholesterol is not included in engines for stroke prediction and a large meta-analysis including 450,000 patients failed to show a relationship between total cholesterol and stroke (128). More recent studies have documented a relationship between total and LDL cholesterol and thromboembolic, but not hemorrhagic, stroke (129) The relationship between hemorrhagic stroke and low cholesterol has been questioned since this association was noted in the MRFIT cohort in subjects with uncontrolled hypertension (130). This relationship was also observed in more recent studies and attributed to excess alcohol intake (131, 132). Although the average age for stroke is higher than the average age for myocardial infarction, there is no recent analysis of cholesterol and stroke taking into account both type of stroke and age.

Peripheral arterial disease is a form of atherosclerosis in which the average age of the subject is 10-15 years older than that of cohorts of patients with coronary artery disease. As opposed to coronary artery disease, in most recent cohorts, the diagnosis is made by noninvasive tests in asymptomatic individuals. Consequently, its prevalence does not reflect the risk factors for events but reflects the risk factors for progression of atherosclerosis. In most large cohorts with populations screened for peripheral arterial disease, subjects with low ankle-brachial index have a significantly higher LDL and total cholesterol, after adjustment for age (133). High total cholesterol is a risk factor for progression of arterial occlusion and for occurrence of peripheral arterial disease in patients with diabetes (134).

In statin-treated patients aged 65-70 years, LDL cholesterol remains a powerful predictor of coronary heart disease (135). This supports the concept of more aggressive cholesterol lowering therapy in this group. The limitation of these data is the fact that the population was restricted to subjects considered by their physicians or by clinical trialists to be candidates for statin therapy.

Other Lipids and Lipoproteins as Risk Factors

 

Triglycerides are associated with cardiovascular risk in middle age cohorts. A recent meta-analysis including 29 Western countries cohorts and 262,525,participants showed the top quintile of serum triglycerides having 72% higher risk than the lowest quintile, with a very robust level of statistical significance (136). A similar study including 96,224 participants from Asia-Pacific Region reported for the highest triglyceride quintile a 80% increase in the risk of coronary events, a 70% increased risk of coronary death and a 50% increased risk of stroke (137). The investigators reported that the data were similar in all age groups, in both genders. A study including only participants over age 65 years showed triglycerides to be a powerful independent predictor of cardiovascular disease in women but not in men (138).

HDL cholesterol is the most powerful lipid predictor of cardiovascular risk in middle-aged men and women. Subjects with high HDL cholesterol are more likely to achieve high longevity (139, 140). Conversely, healthy subjects over age 80 have high HDL cholesterol levels compared with middle-aged subjects, and their offspring has higher HDL cholesterol compared to the age-matched population (141). These subjects seem to also have larger HDL particles, suggesting an enhanced reversed cholesterol transport (142). Cohort studies have documented the fact that HDL cholesterol loses much less of its predictive power with advancing age than LDL cholesterol. In the Cardiovascular Health Study low HDL cholesterol was the only lipoprotein associated with the risk of myocardial infarction (143). In the Honolulu Heart Study low HDL cholesterol was a powerful predictor of non-hemorrhagic stroke (144). In the Leiden 85-Plus Study subjects in the lower tertile of low HDL cholesterol had 2 times the risk of coronary death and 2.6 times the risk of fatal stroke when compared with those in the upper tertile (145). In this age group HDL cholesterol is also a predictor of total mortality. In nursing home residents low HDL and low albumin were proposed as a mean to diagnose frailty and predicted a 2.5 to 4 fold increase in short term mortality (146). Since frailty is associated with both low HDL and low total cholesterol, the ratio of these parameters was proposed to predict cardiovascular events in old subjects. In an analysis of the data of the PROSPER Study, the investigators reported that low HDL cholesterol in elderly patients predicts the risk of fatal and nonfatal coronary events and stroke as well as the benefit of statin therapy (147).

Apo B and Apo A1 are important predictors of cardiovascular risk and some studies have reported that they might be superior to the measurement of standard lipid parameters (148). In 77-years old men, in the Uppsala Longitudinal Study, low Apo A1 concentration  was the best predictor of coronary death (149) In the Leiden 85-Plus Study, Apo E was documented to be a powerful predictor of cardiovascular death, independent of Apo E genotype and lipid levels (150). In the Cardiovascular Health Study, the risk of stroke was increased 3 times, the risk of coronary death was increased 2.5 times and that of death from all causes twice if the participant was in the higher quintile of Lp (a) (151). The Health, Aging and Body Composition Study has documented the increased risk for coronary heart disease in the upper quintile of oxidized LDL, after adjustment for LDL concentration (152, 153). Serum antioxidants have been associated with a reduced cardiovascular mortality in the elderly (154-156), but, unfortunately, studies of antioxidant supplementation have failed to show benefit in all age groups.

.

Inflammatory Markers as Risk Factors

The advent of inflammatory markers as predictors of atherosclerosis related disorders and events have changed the thinking in cardiovascular prevention. Multiple meta-analyses have documented the powerful ability of high sensitivity C Reactive Protein (CRP) in predicting risk, and recent studies have documented the benefit of targeting CRP when titrating statin therapy in high risk patients. A recent set of engines for predicting cardiovascular risk in women and men included this measurement (157, 158). The engines were recommended for use up to 79 years. The Cardiovascular Health Study has documented that high sensitivity CRP remains a powerful predictor of 10-year coronary disease risk in participants over age 65 years, independently of any other known risk factor (159). The risk of total mortality is even stronger when participants have other elevated inflammatory markers (160). The risk appears to be higher for the immediate rather than for the remote future. Other inflammatory markers such as interleukin 6 (IL-6) and tumor necrosis factor alpha (TNF-alpha), have been also associated with mortality but they are either not standardized or not available commercially (161, 162). In addition, inflammatory markers are associated with procoagulants, and an increased risk of death was shown for high levels of factors such as Factor VIII and D-dimer (163). CRP has been associated with the burden of atherosclerosis documented as decreased ankle brachial index, increased carotid intima-media thickness or vascular calcifications (164). This has been shown to be true in elderly patients.

There are two reasons why there is resistance to the use of CRP in the elderly. Recent studies have shown that CRP testing provides only a modest contribution to reclassification of risk when added to traditional risk factors. The second reason is the progressive loss of specificity of CRP as a cardiovascular risk marker which occurs with aging. This will be discussed below.

 

LIPID INTERVENTION FOR CARDIOVASCULAR PREVENTION

 

Lifestyle Changes and Lipoproteins

Most guidelines for cardiovascular prevention recommend lifestyle changes as an important measure for therapeutic intervention. There are very few randomized clinical trials with clinical endpoints addressing the benefit of different recommendations and none addressing directly the older age group. Most studies of dietary intervention in postmenopausal women and older men have addressed weight reduction as a method of correction of lipoprotein changes associated with the increased prevalence of metabolic syndrome in these groups. Reduction of intra-abdominal fat, irrespective of age, results in a decrease in LDL and Non-HDL cholesterol, triglycerides, hepatic lipase, Apo E and Apo C3 and an increase in HDL cholesterol and in the more buoyant fractions of LDL and HDL. The changes are usually proportional to the percent decrease of body weight and abdominal fat. Other interventions have been tried and are derived from epidemiological studies. Large cohorts followed into the age groups of old and very old have identified use of grains, nuts (particularly walnuts) as predictors of decreased risk of diabetes and coronary artery disease (165). Foods with high glycemic index or containing trans fatty acids have been associated with the risk of coronary artery disease events (166). Mediterranean diet has been recommended for cardiovascular intervention, and some authors have attributed its benefit to monounsaturated fat (olive oil). The rationale is that substitution of carbohydrates with monounsaturated fat increases the HDL cholesterol. Unfortunately HDL cholesterol is not a trans-cultural indicator of coronary artery disease. Vegetarians have lower HDL cholesterol, lower risk of coronary artery disease and higher longevity. More studies are necessary in order to refine our current recommendations for dietary changes. Until then it is reasonable to reduce body weight in subjects with metabolic syndrome and to attempt to reproduce the diet of populations characterized by a lower cardiovascular risk.

The effect of dietary changes on lipoproteins is limited when compared with the effect of lipid lowering drugs. In addition most studies addressing lifestyle changes in order to reduce coronary endpoints have used complex interventions including smoking cessation and exercise. The subjects enrolled have usually been younger survivors of myocardial infarction. Exercise is a powerful predictor of cardiovascular risk two ways: as leisure time exercise and as "time spent in watching television" as an indication of sedentarism (167, 168). Exercise decreases insulin resistance and induces changes in risk factors associated with metabolic syndrome.

Of particular interest has been the use of diet enriched in 3 fatty acids. Vegetables which are sources of 3 fatty acids contain alpha-linolenic acid. Fish containing diets or fish oil supplements are another source of 3 fatty acids. The active ingredients of fish oil are two 3 polyunsaturated fatty acids: docosahexaenoic (DHA) and eicosapentaenoic (EPA) acids. If the diet is low in fish, alpha-linolenic acid is used as a source for the synthesis of DHA and EPA. The risk of coronary events is lower in some studies in subjects using a diet high in fish content. Unfortunately, the use of fish seems to be less beneficial if the products contain mercury (169, 170). This led to the marketing in the US of a prescription product of fish oil supplement (see below). The use of fish oil is poorly understood. The cardiovascular benefit of supplements of these components has been documented in clinical trials but is not dependent on their effect on lipoprotein levels. On the other hand, the administration of fish oil will result in a decrease of triglycerides in hypertriglyceridemic subjects. The dosage of over the counter products is expressed in grams of fish oil, but the content of DHA and EPA is variable. In order to achieve cardiovascular prevention the dose is 1 gram per day. The dose for triglyceride reduction is much higher: 2-4 grams per day. If the capsules used have 120 mg DHA and 180 mg EPA this will mean a dose of 7-13 grams of fish oil per day.

Smoking cessation is strongly advised in patients at increased cardiovascular risk from dyslipidemia. Unfortunately, it usually results in weight gain and worsening of the features of metabolic syndrome other than low HDL cholesterol. Weight gain is associated with decreases in HDL cholesterol, but smoking is also associated with low HDL cholesterol. Consequently, the effect of smoking cessation on HDL cholesterol is unpredictable.

Lipid Lowering Drugs and Cardiovascular Risk

 

Table 4 shows the prescription drugs available in the US for lipid intervention. Different classes of drugs have a different mode of action, and within the class there are differences between the drugs in activation of different metabolic pathways.

 

Table 4: Prescription Lowering Drugs Available in the US
Drug	Dosage (mg)	Half life (hrs)	Time Admin.	 LDLc  (%)	 Trig. (%)	 HDLc (%)
STATINS
Lovastatin	10-20-40-80	2-3	w/dinner	-24- 40	-10.0- 19.0	+6.6- 9.5
Simvastatin	10-20-40-80	2-3	evening	-26- 47	-12.0- 33.0	+8.0- 16.0
Pravastatin	10-20-40-80	2-3	evening	-22- 37	-11.0- 24.0	+2.0- 12.0
Fluvastatin	20-40-80	9	evening	-22- 35	-12.0- 25.0	+3.0- 11.0
Atorvastatin	10-20-40-80	14	any time	-39- 60	-19.0 -37.0	+8.0- 14.0
Rosuvastatin	5-10-20-40	19	any time	-45- 63	-10.0- 35.0	+8.0- 14.0
PPAR Agonists
Gemfibrozil	600 bid	6.4	before meals	-11	-35	+11
Fenofibrate	130-145-160	20	anytime	-20-30	-25-35	+10-15
Pioglitazone	15-20-45	9	anytime	+6-8	-10	+14-19
Rosiglitazone	2-4-8	3.3	anytime	+14-18	NS	+10-18
Bile Acid Binding Resins
Cholestyramine	4-24,000	n/a	anytime	-10-25	+10-50	+0-5
Colestipol	5-30,000	n/a	w/meal	-10-25	+10-50	+0-5
Colesevelam	3,750-4.375	n/a	w/meal	-18	+9	+3
Other Drugs
LA Niacin	500-2,000	?	at bedtime	-5- 15	-20- 30	20- 30
Ezetimibe	10	22	anytime	-18	-8	+1
Lovaza	4,000	n/a	anytime	+20	-45	+9

 

Statins

The primary mechanism of action of statins is the inhibition of HMG CoA Reductase. This enzyme represents the rate limiting step in the synthesis of cholesterol. Decrease of cholesterol synthesis results in a depletion of intracellular cholesterol and a compensatory increase in the number of LDL receptors on the cell surface. Consequently LDL catabolism is increased and the number of circulating LDL particles diminishes. Decrease of cholesterol pool in the liver leads to a change in the bile acid cholesterol ratio of the bile and a decrease of its lithogenicity. Additional properties of the statins are derived from inhibition of other biochemical pathways. The main pathway dependent on HMG Co A Reductase activity, other than cholesterol synthesis, is protein geranylation. The inhibition of this pathway results in anti-inflammatory and anti proliferative properties of this class of drugs. Numerous in vitro and animal studies have documented these properties referred to as "pleiotropic effects" (171-174). Studies have shown this effect to be independent of cholesterol lowering, but since they are dependent on HMG Co A reductase inhibition this relationship is complex. For instance, atorvastatin 10 mg/day has the same effect on CRP as pravastatin 40 mg/day (175) while atorvastatin 80 mg/day is much more powerful in lowering CRP than pravastatin 40 mg/day (176). The clinical benefit of statins in cardiovascular prevention manifests itself earlier than that of other lipid lowering interventions. This and the following observations are considered to represent examples of clinical expression of the pleiotropic effects of statins:

• The earliest benefit of statins is an improvement in endothelial function and a decrease in hospitalization for angina.
• Statins decrease smooth muscle cell proliferation and prevent vascular remodeling manifesting itself in carotid intima-media thickening.
• Statins reduce the level of anti-inflammatory markers such as CRP (175). In clinical trials, lowering CRP concentration does not parallel LDL lowering by the same drug, but the highest benefit is obtained in participants with low levels of LDL cholesterol and CRP. Consequently it has been recommended that both LDL cholesterol and CRP be targeted while using statin therapy (177, 178).

Other hypotheses based on the pleiotropic effect of statins are being tested in clinical trials (see below).

The efficacy of statins is shown in Table 4. different statins have different ranges of effects on lipids and lipoproteins. Two large comparative head to head clinical trials have documented the differences between the statins available in the US market (179, 180). As an example, a decrease of LDL cholesterol of 45% is achieved by 10 mg rosuvastatin, 10 mg atorvastatin or 80 mg simvastatin. Fluvastatin and pravastatin do not achieve LDL cholesterol lowering in this range. The effect of statins on triglycerides in hypertriglyceridemic subjects is of the same magnitude as the effect on LDL cholesterol. There seem to be a small difference between the statins in this respect favoring atorvastatin. The effect of statins on HDL cholesterol is also different, with rosuvastatin being the most effective.

PPAR activators

PPAR alpha activators (fibrates) have been available as commercial drugs before their mechanism of action was known. The activation of these receptors by fibrates results in increased fatty acid oxidation (181, 182). This results in decreased incorporation of fatty acids in VLDL and reduced VLDL production. The decrease in VLDL synthesis parallels a reduction in Apo C3 synthesis and a decreased level of its concentration. VLDL catabolism is increased by a process of activation of lipoprotein lipase for which Apo C3 is a natural inhibitor. The decreases in concentration of large VLDL particles results also in decreases in small, dense LDL particle concentration and a shift of LDL size towards more buoyant and less atherogenic particles. In addition the activation of PPAR alpha receptors results in an activation of the process of reverse cholesterol transport. This is mediated by an increase in Apo A1 production and an activation of the ABC- A1 cassette transporters and is documented in clinical practice by elevation of HDL cholesterol. Fibrates have pleiotropic effects. Administration of fenofibrate resulted in decreases in IL-6 and fibrinogen and cell-adhesion molecules. Both fenofibrate and gemfibrozil result in decreases in CRP.  Fenofibrate decreases the LDL cholesterol slightly more in nonhypergtriglyceridemic subjects (183). It also decreases uric acid as a drug specific pleiotropic effect. Fenofibrate administration results in increases of serum creatinine, cystatin and homocysteine while gemfibrozil does not have these effects (184). Although originally thought that the drug increases creatinine excretion (185), further larger studies have shown that glomerular filtration rate is diminished and expressed concern about it (186). The dose should be decreased to one third if diminished renal function is present.

PPAR gamma activators (thiazolidinediones=TZD) have been marketed as lipid lowering agents. They act through improvement of insulin sensitivity and therefore reverse some of the features of metabolic syndrome. Both TZDs increase HDL cholesterol concentration and reduce the concentration of small dense LDL particles, but pioglitazone is the only TZD on the US market which has a small but statistically significant effect on triglyceride concentration (187). Attempts have been made to synthesize drugs with activation of more than one PPAR receptor. The simultaneous activation of PPAR alpha and gamma will theoretically result in a better effect on lipids. The only drug which received a letter of approval from the FDA was muraglitazar which was expected to be a great cardiovascular preventative drug and had an excellent lipid profile. Unexpectedly a meta-analysis of studies done with this product showed an increase in cardiovascular risk and the drug was not brought to the market (188)

 

Bile Acid Binding Resins

Bile Acid Binding Resins have been used since the late'1960�s. By diminishing bile acid absorption, they increase 7 alpha hydroxylation of intra-hepatic cholesterol and reduce the intracellular cholesterol pool. Consequently, there is an increase in the number of LDL cholesterol receptors on the surface of the hepatocyte and an increase in LDL catabolism. Their use results also in a compensatory increase in cholesterol synthesis and an increased production of VLDL with resulting hypertriglyceridemia. Recently studies have shown a hypoglycemic effect for this class of drugs. The mechanism of action is believed to be the induction of changes in the activation of nuclear receptors FXR and LXR. Interference with the enterohepatic cycle of bile acids results in inhibition of FXR and activation of LXR activity (189). Traditional Bile Acid Binding Resins absorb a large number of compounds besides bile acids and have a large number of drug interactions. Colesevelam has a structure containing polar side chains and has much less potential for drug interactions.

Niacin

Preparations of niacin have been used since the seventies for lipid intervention. The risks of uncontrolled use of this medication have become apparent and a prescription form of long acting niacin was marketed. The mechanism of action of niacin includes an inhibition of triglyceride production by reducing the flow of free fatty acids released by the adipose tissue and by inhibiting the activity of diacylglycerol acyl transferase (190, 191). Niacin also inhibits the uptake of Apo A1 by the hepatocyte without affecting the transfer of cholesterol esters from HDL. The result of these actions is a decrease in triglyceride and total cholesterol concentration and an increase in HDL cholesterol. The drug is also the only product to effectively reduce Lp (a). Niacin releases prostaglandin D2 by binding to specific cutaneous receptors and this effect is responsible for flushing, the main side effect of niacin (192).

Cholesterol absorption inhibitors

Ezetimibe is the only drug available from this class. The drug binds cholesterol absorption receptors on the intestinal brush border cells and inhibits the transfer of cholesterol from the intestinal lumen. Recent studies seem to identify a transport protein, NCP1L1, as the likely target of ezetimibe action (193). There are no pleiotropic effects described for this drug, but it seems to enhance the anti-inflammatory effects of statins reducing further the level of CRP (194).

Omega-3 fatty acids

Recently a preparation of 3 fatty acids has become available by prescription. LovazaR contains 375 mg DHA and 465 mg EPA per capsule. 3 fatty acids decrease triglyceride concentration by reducing VLDL production. Additional pharmacological effects have been described but require further work: increased VLDL catabolism, improvement in endothelial function and reduction in inflammatory biomarkers.                                                                                                                                                               

Combination therapy

Combination therapy between lipid lowering drugs has been increasingly recommended for the management of hyperlipidemia. Numerous studies have shown the effect of simultaneous administration of two lipid lowering drugs to represent the expected additive effect of the pharmacological effects of each drug. The most common combinations have been marketed as single products. Simvastatin and ezetimibe (VytorinR) has been marketed as the single most powerful LDL cholesterol lowering agent. Lovastatin and long-acting niacin (AdvicorR) has been marketed for correction of multiple lipoprotein abnormalities. Other combination medications are being studied. Cost and safety issues are the most likely barriers to the use of multiple lipid lowering drugs simultaneously.

LIPID LOWERING DRUGS AND CARDIOVASCULAR PREVENTION

 

Pharmacologic lipid-lowering therapy has been shown to be highly effective in reducing cardiovascular events in numerous primary and secondary prevention clinical trials over the past 30 years. A few of these large clinical trials have specifically analyzed data from patients older than age 65 years.

Statin Cardiovascular Clinical Trials

 

Much of the evidence regarding cholesterol-lowering and its role in reducing CHD risk has been derived from trials using statins. The statins trials addressing cardiovascular risk reduction in high risk patients with no limitation to the type of disease resulting in atherosclerosis is shown in Table 5. In a recent meta-analysis including most of these trials, total mortality was reduced by 12%, cardiovascular death was reduced by 18%, main coronary events were reduced by 23%, stroke by 17% and coronary revascularizations by 24% (195). Higher differences in LDL cholesterol between the arms have resulted in higher benefit. In another meta-analysis the authors included only trials enrolling subjects with no documented atherosclerosis (196). Statins reduced the risk of main coronary events by 29%, of stroke by 17% and of revascularizations by 34%. Coronary death was reduced by 23%, but the results did not reach statistical significance.

 

Table 5: Successful Statin Studies
Study	Type of patients	Randomization arms	Duration; Primary outcome; Secondary outcomes	Clinical results	Older group
4S  (197, 198)	4,444 patients with TC 213-310 mg/dL, TG </= 220 mg/dL, and CAD	Simvastatin 20-40 mg/day vs. Placebo	5.4 years; all-cause mortality; major cardiovascular event	Significant reductions: LDL: 35% total mortality;  30% coronary events: 34% revascularizations: 37%	See Meta-Analysis
A to Z (199)	4,497 patients with ACS, TC </= 250mg/dL, age 21-80 years	Simvastatin 40mg/day x 1 month, followed by 80mg/day vs. Placebo x 4 months followed by 20mg/day	1.9 years; composite of CV death, nonfatal MI, readmission for ACS, and stroke	Difference in primary outcome NS. CV death occurred in 5.4% vs. 4.1% (p=0.05), but all other components of primary endpoint NS.	Risk reduction of primary outcome: Patients <65 years: 13% Patients 65+ years: 10%
AFCAPS (200)	6,605 patients with average TC and LDL, and below-average HDL	Low saturated fat, low cholesterol diet; Lovastatin 20-40mg/day vs. Placebo	5.2 years; first acute major coronary event	37% risk reduction in primary outcome (p<0.001) in lovastatin group.	1,416 patients age 65+ (no difference in benefit from remainder of study population).
ALLHAT-LLT (201)	10,355 patients age 55+ years, with LDL 120-189 mg/dL (100-129 if known CHD), TG <350 mg/dL, HTN, at least 1 additional CHD risk factor	Pravastatin 40mg/day vs. usual care	4.8 years; all-cause mortality; non-fatal MI or fatal CHD  combined, cause-specific mortality, and cancer.	Difference in primary outcome and CHD events NS.	5707 patients age 65+ (no difference in benefit from remainder of study population).
ALLIANCE (202)	2,442 patients with CHD, hyperlipidemia	Atorvastatin titrated to LDL <80 mg/dL or max 80mg/day vs. usual care	51.5 months; time to first CV event	Atorvastatin vs. usual care: LDL reduction: 34.3% vs. 23.3% (p<0.0001) NCEP goals of LDL<100 met: 72.4% vs 40.0% Primary outcome: 23.7% vs 27.7% (p=0.02) Non-fatal MI: 4.3% vs 7.7% (p=0.0002)	Mean age 61 years. No significant interaction between treatment group and age.
ASCOT-LLA (203)	10,305 patients age 40-79, HTN, and at least 3 other CV risk factors	Atorvastatin 10mg/day vs. Placebo	3.3 years; non-fatal MI and fatal CHD	Non-fatal MI and fatal CHD: 100 vs. 154 events (HR 0.64 [95% CI 0.50�0.83], p=0.0005) TC: lowered by 50mg/dL at 1 year, and by 42.5mg/dL at 3 years	See Meta-Analysis
ASPEN (204)	2,410 patients with type 2 DM	Atorvastatin 10 mg/day vs. Placebo	4 years; composite CV death, non-fatal MI, non-fatal stroke, PCI, CABG, resuscitated cardiac arrest, unstable angina	None of the outcome reductions were significant.	There were 870 participants over age 65 years.
CARDS (205, 206)	2,838 patients with type 2 DM, age 40-75, LDL </= 160 mg/dL	Atorvastatin 10mg/day vs. Placebo;	3.9 years; time to first acute CHD event, coronary revascularization, or stroke	Absolute RR of major CV event: same in both groups All-cause mortality: NS reduced in either group NNT for 4 years to avoid one event: 21 vs. 33	See Meta-Analysis
CARE (207)	4,159 patients with TC <240 mg/dL, h/o MI, LDL 115-174 mg/dL	Pravastatin 40mg/day vs. Placebo	5 years; composite of fatal coronary event or non-fatal MI	Significant Reductions: Fatal coronary event or non-fatal MI: 24% (p=0.003) Need for bypass: 26% (p=0.005) Need for PCI: 23% (p=0.01) Stroke frequency: 31% (p=0.03) Reduction in coronary events was greater in patients with higher pretreatment LDL levels.	See Meta-Analysis
GREACE (208)	1,600 patients age <75, LDL >100 mg/dL, TG <400 mg/dL	Atorvastatin titrated 10-80mg/day for goal LDL <100 vs. usual care	3 years; death, non-fatal MI, unstable angina, CHF, PCI, and stroke; safety, efficacy, and cost-effectiveness of atorvastatin	Significant reductions in: Total mortality (p=0.0021) Coronary mortality (p= 0.0017) Coronary morbidity (p<0.0001) Stroke (p=0.034)	No statistical difference between patients 60-75 years old vs. remainder of population.
HPS (209, 210)	20,536 patients age 40-80, with coronary disease, occlusive arterial disease, or DM	Simvastatin 40mg/day vs. Placebo	5 years; mortality and fatal or non-fatal vascular events	Reductions: All-cause mortality: 12.9% vs. 14.7% (p=0.0003) Coronary death rate: 5.7% vs. 6.9% (p=0.0005) Other vascular deaths: 1.9% vs. 2.2% (p=0.07) Non-vascular deaths: NS First MI, stroke, or revascularization: 24% (p<0.0001) The benefits of simvastatin were additional to those of other cardioprotective treatments.	See Meta-Analysis
IDEAL (211)	8,888 patients age 80 or younger and h/o AMI	Atorvastatin 80mg/day vs. Simvastatin 20mg/day	4.8 years; major coronary event; major CV events (any primary event plus stroke), any CHD event (any primary event, any coronary revascularization procedure, or unstable angina), any CV events (any of the former plus hospitalization with CHF or PAD)	Atorvastatin vs. Simvastatin: Primary outcome: 9.3% vs. 10.4% (p=0.07) Non-fatal AMI: 6.0% vs. 7.2% (p=0.02)	The average age was 61 years. There was no significant interaction between treatment group and age.
LIPID (212, 213)	9,014 patients with previous MI or UA, and TC 155-271 mg/dL	Pravastatin 40mg/day vs. Placebo	6 years; major CV disease events	Significant Reductions: All-cause mortality: 22% (p<0.001) Death from CHD: 24% (p<0.001) CHD death or nonfatal MI: 24% (p<0.001) Stroke: 19% (p=0.05)	See Meta-Analysis
JUPITER (214, 215)	17,802 healthy subjects with CRP >2.0 mg/L and LDL cholesterol < 130 mg/dL	Rosuvastatin 20mg/day vs. Placebo	Stopped after 1.9 years; Myocardial infarction, stroke, arterial revascularization, hospitalization for unstable angina, or death from cardiovascular causes, Secondary endpoints were each one separately	44% reduction in primary endpoint, 54% reduction in myocardial infarction, 48% in stroke, 47% in revascularization or unstable angina and 20% in death of any cause	9,261 patients over age 65 had the same benefit as younger patients
LIPS (216)	1,677 patients age 18-80 years, with stable or unstable angina or silent ischemia after first PCI, with TC 135-270 mg/dL, TG <400 mg/dL	Fluvastatin 80mg/day vs. Placebo	3.9 years; survival time free of major adverse cardiac event (MACE)	21% vs. 26.7% had at least one MACE (p=0.01). Time to first MACE was longer in fluvastatin group (p=0.01).	A Cox Regression analysis was significant after use of age over versus under 65 years as covariate.
MEGA (217)	7,832 Japanese patients with TC 220-270 mg/dL, no h/o CHD or stroke	Diet vs Diet + pravastatin 10-20mg/day	5.3 years; first occurrence of CHD	Primary outcome was significantly lower in the diet plus pravastatin group than in the diet alone group (p=0.01).	There reduction in CHD in patients over 60 years was 41% (22-60) 95%CI
MIRACL (218, 219)	3,086 patients age 18+ years, with UA or non Q-wave AMI	Atorvastatin 80mg/day vs. Placebo (24-96 hours after hospital admission)	4 months; death, non-fatal AMI, cardiac arrest with resuscitation, or recurrent symptomatic myocardial ischemia with objective evidence and requiring emergency rehospitalization	Primary outcome occurred in 14.8% vs. 17.4% (p=0.048). Atorvastatin group had a lower risk of symptomatic ischemia with objective evidence and requiring emergency rehospitalization (p=0.02). There were fewer strokes in the atorvastatin group (p=0.045). Differences in risk of death, non-fatal MI, cardiac arrest coronary revascularization procedures, worsening HF, or worsening angina: NS.	See Meta-Analysis
MUSASHI-AMI (220)	486 patients with Acute MI and  normal TC levels	Any available statin within 96 hours of AMI vs. no statin	2 years: composite of CV death, non-fatal AMI, recurrent symptomatic myocardial ischemia, CHF, and stroke	Primary outcome rates 6.1% vs. 11.4% (p=0.04). Statin group had lower risk of CHF (p=0.015) and symptomatic myocardial ischemia (p=0.026).	Mean age of all patients was 64 years
Post-CABG (221)	1,351 patients from post-CABG trial (1-11 years post-CABG, with LDL 130-175 mg/dL, TG <300 mg/dL, patent grafts, and ejection fraction >/= 30%)	Lovastatin 40-80 mg/day (and 8g cholestyramine/day if needed) to achieve LDL 60 to 85 mg/dL vs. Lovastatin 2.5 or 5mg/day (and 8g cholestyramine/ day if needed) to achieve LDL 130 to 140 mg/dL; patients also randomized to warfarin vs. placebo.	7.5 years; occurrence of CV events and procedures	Significant reductions: Revascularization procedures: 30% (p=0.0006) Composite primary outcome: 24% (p=0.001)	Mean age of all patients was 62 years
PROSPER (147, 222-225)	5,804 patients aged 70-82 with h/o or risk factors for vascular disease	Pravastatin 40mg/day vs. Placebo	3.2 years; composite of coronary death, non-fatal MI, and fatal or non-fatal stroke	Significant Reductions: LDL: 34% Primary outcome: 15% (p=0.014) CHD mortality: 24% (p=0.043) Stroke: NS	See Meta-Analysis
PROVE-IT (226)	4,162 patients hospitalized for ACS in past 10 days	Pravastatin 40mg/day vs. Atorvastatin 80mg/day	2 years; composite of death from any cause, MI, unstable angina requiring rehospitalization, revascularization, and stroke	Primary outcome: 26.3% vs. 22.4% (16% reduction in hazard ratio in favor of atorvastatin, p=0.005)	1230 patients age 65+ (no statistical difference from remainder of study population)
SAGE (227)	893 patients age 65-85 years with CAD and >/= 1 episode of myocardial ischemia with a total duration >/= 3 minutes in 48-hours.	Atorvastatin 80 mg/day vs. Pravastatin 40 mg/day	1 year; absolute change from baseline in total duration of ischemia	MACE: NS All-cause mortality: 1.3% vs. 4.0% (p=0.014) Primary outcome: Significantly reduced in both groups; difference between groups NS	See Meta-Analysis
TNT (228-230)	10,001 patients with CHD, LDL <130 mg/dL	Atorvastatin 10mg/day vs. 80mg/day	4.9 years; first major CV event	Reductions of primary outcome: Absolute risk: 2.2% (p<0.001) Relative risk: 22% (p<0.001) No difference in overall mortality.	37% of patients were age 65+ years (no statistical difference from remainder of study population)
WOSCOPS (231-234)	6,595 men aged 45�64 years with no history of MI and TC 251-309 mg/dL	Pravastatin 40mg/day vs. Placebo	4.9 years; CHD morbidity and mortality	RR of coronary events: 31% (p<0.001) RR of definite non-fatal MI: 31% (p<0.001) RR of death from CHD: NS RR of definite plus suspected cases: 33% (p=0.04) RR of death from all CV causes: 32% (p=0.03) RR if all-cause mortality: 22% (p=0.05)	Results were not significantly influenced by age

 

Very few studies have reported separately subjects in the older age group. A majority of the trials have reported that age is not a factor determining clinical benefit.

The PROspective study of Pravastatin in Elderly individuals at Risk of vascular disease (PROSPER) and the Study Assessing Goals in the Elderly (SAGE) trials are the only randomized, controlled trials to date that exclusively enrolled older subjects (222). In the PROSPER trial, 5,804 subjects aged 70-82 years, with history of, or risk factors for, vascular disease, were randomized to either placebo or 40 mg of pravastatin daily. After 3 years, the pravastatin group had significantly fewer CHD-related deaths, nonfatal MI, and stroke, but showed no impact on stroke or dementia. In the SAGE trial, intensive versus moderate statin therapy was examined in subjects aged 65-85 years with CHD and at least one episode of myocardial ischemia (227). Compared with moderate pravastatin therapy, intensive therapy using atorvastatin was associated with reductions in major acute cardiovascular events and all-cause mortality in addition to the reductions in ischemia observed with both therapies, illustrating the benefit of intensive statin therapy in older men and women.

Subgroup analyses of older subjects from three large secondary prevention clinical trials have demonstrated similar results. Cholesterol And Recurrent Events (CARE) (207), Scandinavian Simvastatin Survival Study (4S) (197), and Long-term Intervention with Pravastatin in Ischemic Disease (LIPID) (212) included large numbers of elderly patients. The CARE trial enrolled 4,159 patients with average cholesterol and history of myocardial infarction and included 2,129 patients age 60 years or older. The 4S trial randomized 4,444 subjects with CHD and high cholesterol, 1,010 above the age of 63 years. The LIPID trial randomized subjects with prior myocardial infarction or unstable angina to receive pravastatin 40 mg/day versus placebo. A subgroup of 3,514 subjects aged 65-75 years was compared to younger subjects. Older patients were at significantly greater risk than younger patients for death, myocardial infarction, unstable angina, and stroke. Pravastatin reduced the risk for all cardiovascular disease events, and similar relative effects were observed in older and younger patients. In both 4S and LIPID subgroup analyses; because older patients were at greater risk than younger patients for cardiovascular events, the absolute benefit of treatment was significantly greater in older patients.

The Heart Protection Study (HPS), the largest statin trial conducted to date, enrolled 20,536 patients with coronary disease, occlusive arterial disease, or diabetes mellitus (209, 210). Of these, 4,891 patients were aged 65-69 years, and 5,806 were aged 70 years or older. The proportional reduction in event rate was similar in all subgroups, including patients 70 years or older, and in patients who presented with LDL <116 mg/dL or total cholesterol <193 mg/dL. This was the first trial to demonstrate the benefit of lipid lowering therapy in patients with CHD risk-equivalents.

Two randomized primary prevention clinical trials reported separately the participants in the older age group. The Anglo-Scandinavian Cardiac Outcomes Trial-Lipid Lowering Arm (ASCOT-LLA) enrolled patients with hypertension and at least three other cardiovascular risk factors. A total of 10,305 patients were randomized and the study included 6,570 patients over age 60 years (203). In the Collaborative Atorvastatin Diabetes Study (CARDS), a post-hoc analysis of 2,838 patients with type 2 diabetes was done, comparing the efficacy and safety of atorvastatin among patients aged 65-75 years versus younger patients (206). After 4 years, atorvastatin treatment resulted in the same proportional reduction in the incidence of major cardiovascular events in older and younger patients. Numbers needed to treat for 4 years to avoid one event were 21 and 33, respectively.

All these trials were placebo controlled. Treating to New Targets (TNT) randomized 10,001 subjects with stable coronary artery disease between pravastatin 40 mg/day and atorvastatin 80 mg/day (228). Among them were 3,809 subjects over age 65 years. The study documented the superiority of more intensive lipid lowering irrespective of age group.

An original approach was taken by investigators in JUPITER (214). This trial randomized between rosuvastatin 20 mg and placebo, 17,802 patients with LDL cholesterol less than 130 mg/dL, and a high sensitivity CRP of at least 2.0 mg/L. The study was stopped after 1.9 years because of overwhelming benefit. The primary endpoint was myocardial infarction, stroke, arterial revascularization, hospitalization for unstable angina, or death from cardiovascular causes. There was a highly significant decrease in primary endpoint (44%), myocardial infarction (54%), stroke (48%), revascularization or unstable angina (47%), death from cardiovascular causes (47%) and all cause mortality (20%). The study reported that the benefit was equal in patients over age 70 or below age of 70 years (235)..  Patients who achieved, for both CRP and LDL cholesterol, levels below the median or decreases larger than the median had the best event-free survival (215).

In the 2007 version of this Endotext chapter, we performed a meta-analysis of the trials in which elderly were reported separately, irrespective of age definition of the group, in which statins were used and in which there was a cholesterol difference between the treatment arms. The results are shown in Fig 1. There was a 21% reduction of main coronary events in the elderly in these trials. This is similar to 29% obtained in the CTT meta-analysis and 25% obtained in the younger groups in the same trials (Fig 2).

 

 

 

Subsequently two meta-analyses were published including respectively 51,351 and 19,569 patients with age 60 to 82, enrolled in statin trials of cardiovascular prevention. In the first one, there was a significant reduction in all cause mortality, CHD death, fatal and nonfatal MI and fatal and nonfatal stroke endpoints (236). In the second one, data were analyzed by hierarchical Bayesian modeling. The authors reported a statistically significant reduction in total mortality (22%), coronary death risk (30%), nonfatal myocardial infarction (26%) revascularization procedures (30%) and stroke (25%). The authors estimated that it would take 28 patient treated to save one life (193).

 

Statins for Primary Prevention in the Elderly

Primary prevention with statin therapy in the elderly is still controversial. Four trials included or reported separately the results of primary prevention intervention in the older age group. PROSPER included 3,239 subjects free of vascular disease while all subjects in ASCOT, CARDS and JUPITER were receiving primary prevention intervention by trial enrollment criteria.  A meta-analysis of these trials shows definite benefit for prevention of cardiovascular events. The number of deaths in these relatively healthy subjects is too small to show statistical significance (237). The impact of these studies on the current guidelines has been to recommend primary prevention in all subjects irrespective of age. Moreover the differences between primary prevention and secondary prevention are less clear than they were at the time when the first primary prevention trials were established. Elderly patients with diabetes enrolled in CARDS or PROSPER, or patients with hypertension enrolled in ASCOT or PROSPER would be considered to have non clinically manifest atherosclerosis and treated with statins by most US physicians.  Most of the weight of evidence for what everybody would consider to be "primary prevention" is based on the JUPITER trial.

The publication of the JUPITER trial in November 2008 has started an international debate, which has not been settled by the time of writing this manuscript. The study intended to determine whether long-term treatment with rosuvastatin (20 mg orally per day) will reduce the rate of first major cardiovascular events, defined as the combined end point of cardiovascular death, stroke, myocardial infarction, hospitalization for unstable angina, or arterial revascularization among individuals with LDL-C levels <130 mg/dL (3.36 mmol/L) who are at high vascular risk because of an enhanced inflammatory response as indicated by hsCRP levels <2 mg/L� (238). At the enrollment, subjects were excluded if they had significant comorbidity or "serious medical or psychological conditions that may compromise successful study participation". The low rate of events attests to the good health of the enrollees. The study was based on epidemiologic data showing that elevated CRP indicates increased cardiovascular risk as well as a previous study by the principal investigator showing that in relatively healthy subjects with normal cholesterol and low CRP, benefit of statin therapy could not be demonstrated (239).

Some of the comments addressing the JUPITER study were related to the mechanism of action underlying the results. The study reported that the rosuvastatin treated group had a lower rate of decline of renal function and some readers questioned if this could have contributed to the results (240). Other readers attributed some of the benefit to a potential effect of rosuvastatin on Vitamin D level, shown in previous research (241). Other comments were addressing questions not intended to be asked by the protocol such as "if rosuvastatin was used based on its anti-inflammatory properties, why not use aspirin as an antiinflammatory agent?" (242), There seem to be some confusion concerning the fact that patients achieving lower levels of CRP had better event-free survival irrespective of the level of LDL cholesterol achieved. Some readers were wondering why CRP was not treated to target (243), while others suggested that CRP should not be targeted since it is a marker and not a risk factor (244).  Other readers would have preferred to have the results expressed as continuous variables rather than grouping them in "high and low" categories SnIderman (245). In the study cardiac mortality was adjudecated and the details of this process were never published. This resulted in unusually low number of cardiac deaths and unusually high number of deaths of unknown cause. There was criticism to the study for this reason (246).

One of the caracteristics of the debate was how violent and personal were the attacks. The chairman of the safety-monitoring board was accused of conflict of interest because of his involvement in other industry-sponsored lipid lowering trials (246). Others claimed conflict of interest because there "ere "potentially millions of dollars in royalties for the principal investigator, who coholds the patent for the hsCRP� (246, 247),  The sponsor was under attack for having agressively marketed the drug prior to the publication of the trial (246). Moreover, since the authors of this paper stated that "sponsor's own investigators controlled and managed the raw data" it was suggested that this "does increase the chance of bias seeping into the data set, as the misrepresentation of data about rofecoxib, gabapentin and ENHANCE have shown". The article mentioned that nine of the fourteen authors had financial ties to the sponsor and stopped short of claiming fraud. Other readers stated that the study sends the wrong message because when dealing with individuals at risk of contracting lifestyle-associated diseases, it is unrealistic to imagine that one can prevent or treat the overt clinical complications of an unhealthy lifestyle merely by means of drugs (248). For most clinicians the data of JUPITER are valid and concerns are related only to its application in clinical practice not to its validity.
 The concerns are related to four main issues:

• What is the long-term safety of statin therapy?
• What is the real number of patient needed to treat in order to prevent an event?
• What is the significance of CRP in the elderly?
• Can the selection of elderly patients for primary prevention be improved??

Long-term safety of statin therapy
Everybody will agree that JUPITER was too short for assessment of the safety of statins in these patients (247, 249). The data concerning safety of statins will be addressed further. Specific concern arising from JUPITER data is attributable to the 25% increased incidence of physician reported diabetes (214).The validity of this finding was tested in a meta-analysis of all statin trials showing no significant effect but a high level of heterogeneity (250). In WOSCOPS, investigators reported a decrease in the risk of incident diabetes in statin treated patients (251)  Elimination of this trial from the meta-analysis reestablishes the homogeneity of the data and shows a significant increase in the incidence of diabetes (RR 1.13 [95% CI 1.03�1.23).

The number needed to treat (NNT) to prevent an event in healthy elderly

NNT is defined as the reciprocal of the absolute difference between the two study arms in proprtions of patients with events. Because each trial has a different duration, it is usually adjusted, for five years therapy (NNT-5). Since clinical practice guidelines depend on cost  & risk versus benefit ratios, an acceptable NNT depends on safety of the exposure and cost of the intervention including the cost of the drug and the cost of the monitoring.  The authors  JUPITER presented the NNT-5 calculated from the trial data in a separate paper (252). They estimated that for the primary endpoint the NNT-5 in patients over 70 was 19. and for the primary endpoint or death 15. The NNT-5 of JUPITER compared favorably with NNT-5 for similar endpoints which were 44-63 for other statins, 86 for diuretics in patients with hypertension and 346 to 426 for aspirin.

The concern about the NNT-5 in JUPITER is derived from the fact that the study was discontinued prematurely.  Truncated trials have been shown to be associated with greater effect sizes than nontrucated trials not stopped early. This difference was estimated to average 30% and was independent of the presence of statistical stopping rules and larger in trials with fewer events (253). In a thorough analysis of JUPITER, Kaul et al (254) estimate that a 43% benefit obtained from a truncated trial corresponds to a 20% benefit from a nontruncated trial. Ridker et al (252) have estimated that a statin delivering only 50% of the benefit of rosuvastatin for the endpoint of myocardial infarction, stroke or death, the  NNT-5 increases from  29 to 57. This number relects probably the NNT-5 for JUPITER patients in a more realistic way. To note is also the fact that the NNT-5 varies with the opresence of other risk factors. If the NNT-5 for the primary endpoint for the total cohort was 25, for patients with parental history the NNT-5 was 10, for a smoker it was 13 and for a patients with no hypertension it was 39. If the Framingham 10-year risk was less than 10%, the NNT-5 was 47 and in this prespecified subgroup benefoit from the intervention was not statistically significant. Based on this the FDA required at least 1 additional risk factor besides the JUPITER entry criteria of age and elevated hsCRP level to define eligibility for treatment with rosuvastatin (255).

In spite of the impression that the cost and safety of the intervention presented in JUPITER might have been over-optimistically assessed by the authors, the data are very robust and primary prevention in the elderly is a fact. A recent ACCF/AHA Guideline (256) states: In men 50 years of age or older or women 60 years of age or older with LDL cholesterol less than 130 mg/dL; not on lipid-lowering, hormone replacement, or immunosuppressant therapy; without clinical CHD, diabetes, chronic kidney disease, severe inflammatory conditions, or contraindications to statins, measurement of CRP can be useful in the selection of patients for statin therapy.  In a more conservative move, the European Medicines Comission has approved rosuvastatin for prevention of major cardiovascular events in patients who are estimated to have a high risk for a first cardiovascular event (257).  Obviously all this is based on interpretations of the JUPITER Trial by different government agencies. In order to increase the benefit and decrease the risk derived from these recommendations one needs to understand the significance, of CRP elevation in the elderly 

Significance of CRP elevation in the elderly

JUPITER has been criticized because it did not prove the need to test for CRP (254, 258-260). Specifically there was no arm of patients with low CRP and low cholesterol. In reality CRP was used to select a group of patients who had higher risk of cardiovascular events and therefore to limit the use of statin therapy to higher risk group. Indeed CRP is an indicator of risk of cardiovascular events and all-cause mortality in all age groups (261). The main problem is the relationship between its specificity for cardiovascular risk and age. Data show that the concentration of inflammatory marker increases with age and their ability to predict cardiovascular events diminishes. They remain strong predictors of mortality including cardiovascular but also noncardiovascular death causes.
In the Health and Body Composition Study, IL-6 was shown to modestly improve coronary risk prediction in older adults while CRP was of borderline significance (262). Another analysis concluded that the improvement in risk stratification or reclassification from addition of CRP to models based on established risk factors is small and inconsistent (263). In the Framingham Heart Study, the net reclassification improvement when CRP was added to traditional factors was 5.6% for total CVD (P=0.014) and 11.8% for hard CHD (P=0.009) (264). In a prospective study of 5,360 men with average age 77 years, enrolled in the Whitehall Study, CRP emerged as a valuable predictor of cardiovascular risk (265). In the same publication, however, higher CRP was also associated with significantly raised non-vascular mortality independently of other characteristics. In another Dutch cohort (266), during a follow-up period of 48 months, 75 (19%) men died and the cause of death was CV disease in 31 of the men (8%). The Framingham, PROCAM and a Dutch Risk Function predicted mortality risk poorly. IL-6 and CRP along with carotid plaque burden allowed identification of subjects with low and high mortality risk but did not discriminate between vascular and nonvascular causes of mortality. CRP is considered an index of intravascular inflammation and an expression of plaque instability (267). The studies reviewed suggest that, with aging, inflammation has an increasing role in indicating risk of death from any cause and a decreasing specificity for cardiovascular death related to plaque instability.

The aging process is characterized by a progressive remodelling of  the immune system. This remodeling is strongly related with an increased inflammatory response. With increasing age there is a reduction of the size of the thymus associated with morphologic changes such as accumulation of necrotic and fibrotic areas (268). This results in a decline in the output of newly released T-cells and therefore a progressively diminishing number of na�ve T-cells. There is also a decrease in circulating dendritic cells (269).. While this process occurs, possibly as a compensatory mechanism, there is a process of "filling the immunological space" with memory and effectot T-cells. There is an expansion of increased number of activated T-cells (HLA-DR+)_(270) and an expansion of CD28 negative cells (271), helper/inducer (CD4+) but mostly supressor/cytotoxic (CD8+). These T-cells are capable of producing cytokines. In addition there is a compensatory increase of macrophage activity (272)..

CRP production is regulated by the levels of IL-6. Plasma levels of IL-6 are low or undetectable until age 50-60 years. With aging the ability of circulating lymphocytes to produce IL-6 upon stimulation increases (273) and the levels of IL-6 increase progressively (104). The most striking increase with age in CD8+ memory cells producing type 1 cytokines, IL-4, IL-6 and IL-10 (271). The activation of macrophages also contributes to IL-6 production. The production of CRP correlates well with IL-6 production in the elderly (274). 

CRP levels increase progressively with age (261). The process has been referred to as "inflammaging" (272). It occurrrs progressively with advancing age even in successful aging but predicts the occurrence of frailty and its co-morbidity.   

Frailty and the spectrum of conditions associated with elevated CRP

Frailty is "the Holy Grail of Geriatric Med"cine� (275). An adequate working definition of this syndrome is not available to date. The current definition is referred to as the "Fried phenotype" comprising three or more of: unintentional weight loss, self-reported exhaustion, weakness (reduced grip strength), slow walking speed, and low physical activity(276). Each one of these features is associated with an increased risk of disability and death and the mortality increases with each additional feature (277).
Patient carrying the Fried phenotype have multiple other objective abnormalities, occurring in proportion with the number of features present and each one associated with a higher risk of death.

• High levels of IL-6, CRP, TNF-alpha, interleukin 1 receptor antagonist (IL-1Ra), soluble TNF receptors, and high WBC (278-287). In well-functioning community-dwelling elderly, TNF- alpha is associated with impaired appetite and therefore with subsequent weight loss (288). The level of inflammatory markers is correlated with the number of features of the Fried phenotype (286).
• Nutritional abnormalities: low albumin, low Vit D, low Vit E, Vit C and low carotene levels (289-295).
• Hormonal abnormalities: high level of PTH and low level of IGF-1 and DHEAS (296-298).
• Coagulation abnormalities: high levels of D-Dimer (287, 299-301).
• Diminished renal function (302, 303).
• High homocysteine (294, 304).
• Sarcopenia detected by dual X-ray absorptiometry (305-307).
• Depression and cognitive impairment, detectable through neuro-psychiatric testing (120, 308-312).

Many studies have pointed out the fact that frailty appears to be a chronic inflammatory state and that inflammation precedes frailty.  Investigators reported from the Iowa site of the Established Populations for Epidemiologic Studies of the Elderly on subjects aged 71 years or older, free of disability at baseline. Participants in the highest IL-6 tertile were 1.76 (95% CI, 1.17-2.64) times more likely to develop mobility-disability and 1.62 (95% CI, 1.02-2.60) times more likely to develop mobility plus ADL-disability compared with to the lowest IL-6 tertile (313). In 718 women enrolled in the Women's Health and Aging Study I, the combination of low IGF-I and high IL-6 levels conferred a high risk for progressive disability (314). In the Health, Aging and Body Composition Study, 2,979 men and women, aged 70 to 79, without mobility limitation at baseline the relative risk (RR) of incident mobility limitation per standard deviation (SD) increase was 1.19 (95% CI: 1.10�1.28) for IL-6 and 1.40 (95% CI: = 1.18�1.68) for CRP (315). In the Longitudinal Aging Study Amsterdam, of the nonfrail subjects at baseline, 14�1% became frail at three years follow-up. After adjustment for potential confounders, moderately elevated CRP levels (OR 1�69, 95% CI 1�09�2�63) and low 25(OH)D (OR 2�04, 95% CI 1�01�4�13) were associated with incident frailty (316). Another study from the same cohort reported that, after adjustment for confounders, including socio-demographic, health, and lifestyle factors, high IL-6 (>5 pg/mL) and high CRP (>6.1 ug/mL) were associated with a 2 to 3-fold greater risk of losing greater than 40% of muscle strength in three years (317). In INCHIANTI the rate of decline in grip strength was progressively greater with the increasing number of dysregulated catabolic biomarkers at baseline (CRP, IL6, IL-1 receptor antagonist, tumor necrosis factor-alpha receptor 1 (318). In the Health and Aging Composition Study, 2,895 patients aged 70-79 of which 702 had diabetes, were evaluated for development of persistent functional limitation after 3.5 years.  After adjusting for clinical and demographic covariates and irrespective of the presence of diabetes, persistent functional limitation in the highest tertile of CRP or IL-6 was significantly greater compared with that in the lowest tertile (319).
In studies including patients with specific disorders associated with frailty risk the results were similar. Higher levels of IL-6 and/or CRP were adversely associated with recovery of function after hip fracture (320), functional recovery after hospitalization for hip fracture, heart failure or infection (321) six minutes walking distance in patients with heart failure (322).

As shown, cognitive impairment and depression are associated with frailty. In some studies inflammatory markers were predictors of cognitive or mood deterioration. A meta-analysis showed that high concentrations of CRP were predictive of cognitive decline and dementia (323). A more recent study reported that high serum hsCRP concentration predicts poorer memory 12 years later in elderly women (324). In the Leiden-85 study, high serum hsCRP concentration predicts poorer memory 12 years later in elderly women (325). In patients with cardiovascular disease, high CRP levels were associated with subtle declines in attention-executive-psychomotor performance (326). In 1,113 men aged 70 years or more, high hsCRP predicted a higher incident Short Zung Self Rating Depression Scale score and depressed mood among older men (327). The changes in cognitive function and mood are important in the etiology of reverse epidemiology of frailty syndrome. Annual changes in MMSE Score were associated with cholesterol and systolic blood pressure decline (328). In another study, increases of CRP and decreases in LDL cholesterol had a negative additive effect on general cognition and memory performance at six years follow-up. (329).

 

Frailty is associated with reverse epidemiology. The highest risk of death in the elderly is present in subjects with low cholesterol and low BMI. Some studies show an increase risk of death from cardiovascular disease in elderly with low cholesterol and / or low BMI, resulting in a "J" curve relationship between cholesterol and coronary risk (103, 330). In Finnish nonagenarians, the levels of IL-6 and CRP were negatively correlated with LDL and nonHDL cholesterol levels . This relationship cannot be detected in younger healthy controls. Low cholesterol and high inflammatory markers are associated with frailty in old subjects and low cholesterol precedes the development of disability features characteristic to the frailty phenotype (331-333). In hospitalized patients aged 80 or more, serum cholesterol and LDL cholesterol are negatively associated with CRP and positively associated with prealbumin indicating that both malnutrition and inflammation are determinants of serum cholesterol (334).
With aging the prevalence of frailty increases, and since frailty results in lower cholesterol level and higher risk of death, high total cholesterol in the very old appears in many studies to be a marker of longevity. In this age group, low cholesterol is associated with poor prognosis and predicts an increased risk of death (115, 330, 335, 336). In the Leiden 85-plus Study, 599 participants aged 85 years were annually assessed during a 5-year follow-up period. Participants who died at age 90 years had stronger annual declines in BMI, total cholesterol levels, and diastolic blood pressure and weaker increases in HDL cholesterol levels than participants who survived until the end of follow-up (328). This was interpreted by the authors as indicating that mortality in this group is preceded by a profile suggesting a wasting disease. Frailty is more prevalent in patients with metabolic syndrome (337-340) and increased insulin resistance (341). In patients without diabetes, however, increased insulin sensitivity is a predictor of death (342). This is another example of reverse epidemiology.
Frailty syndrome in the elderly is analogous to similar conditions in wasting diseases afflicting younger age groups. The best studied is the "malnutrition-inflammation-cachexia syndrome" (MICS) in dialysis patients. In these subjects, traditional risk factors are negatively associated with adverse outcomes resulting in a phenomenon referred to as "reverse epidemiology". Low cholesterol in these patients is associated with increased cardiovascular events while in dialysis patients free of MICS higher cholesterol is associated with poorer cardiovascular outcome. Similarly, reversed epidemiology is reported in some dialysis studies for body weight (343), weight loss (344) blood pressure (345) and Hemoglobin A1c (346). Correction for MICS reverses to a certain extend the reverse epidemiology seen in dialysis patients cohorts (347). Reversed epidemiology has been documented in congestive heart failure (348), chronic obstructive pulmonary disease  (349) and rheumatoid arthritis (350). The prevalence of these disorders is increased in the elderly.  Because of the fact that there is no consensus for a definition of frailty, there are no data showing the relationship between total cholesterol and coronary disease corrected for frailty.

Implications of frailty for primary prevention in the elderly.

. Can the benefit demonstrated in JUPITER be extended to frail patients? We believe that there are reasons to reject this hypothesis.

• Frail older subjects are less likely to participate in clinical trials because of limited mobility, feeling of exhaustion and cognitive impairment preventing the investigators from obtained informed consent. No randomized trial of statin therapy including JUPITER has attempted to verify if frail patients were included
• Frail patients have reverse epidemiology. In younger patients, cholesterol is a predictor of risk. In addition, in most statin randomized clinical trials, LDL cholesterol or Apolipoprotein B are predictors of cardiovascular risk after adjustment for treatment arm allocation. In the very old both BMI and cholesterol are negatively associated with risk. This is very likely due to an increase of prevalence of prefrailty and frailty with age. Therefore, assuming that frail patients benefit from cholesterol lowering is contra intuitive.
• Elderly patients have powerful risk factors for cardiovascular disease, different from general population: anemia (351), low albumin (352), low 25-hydroxy Vitamin D and high PTH (353). These are associated with frailty and also risk factors in younger groups of patients with wasting disorders such as patients with end-stage renal disease (ESRD) (354-356) and/or congestive heart failure (CHF) (357). We will present the data of statin trials in patients with ESRD or CHF.
• There is a high prevalence of frailty among patients with advanced atherosclerosis. Conversely, there must be a high prevalence of severe atherosclerosis among frail patients. Frailty occurs more frequently in patients with metabolic syndrome, which is also increasing their risk for atherosclerosis. Advanced atherosclerosis is associated with arterial stiffness and vascular calcification. Sarcopenia, the imaging correlate of frailty is associated with arterial and valvular calcification (358). We will present the results of statin therapy in patients with vascular calcification.

Conversely, if patients with frailty were adequately represented in JUPITER, they had increased survival from all cause death from randomization to rosuvastatin. Therefore statin therapy may be a way to reverse deterioration in aging patients. This hypothesis was already suggested (359) and will require testing in additional trials. An argument against trying it is the observation that statins may exacerbate muscle performance decline and risk of falls associated with aging and frailty and this effect might be reversible with cessation of statin therapy (360).

Can the selection of elderly patients for primary prevention be improved?

In our opinion, the reaction to this trial is due to the concern of expanding the clinical use of statin to patients who are relatively healthy and particularly vulnerable to drug induced adverse events because of their age.  An estimate from most recent NHANES data shows that an additional 3.9 million men age >50 years and 2.6 million women >60 years should be added to existing 6.7 million older adults as candidates for statin therapy (361). Using the same data, other authors estimated that 80% of the elderly should be placed on statin therapy (362). Another group pointed out that among enrollees in JUPITER, 41% required treatment with statins under current guidelines (363). This created a high level of discomfort among the readers and a feeling that the number should be somehow limited.
At the present time, evidence based decisions can only be made based on JUPITER. In addition, traditional risk engines have been shown to lose their value in the very old and there are no new engines predicting cardiovascular events. CRP is an independent predictor of cardiovascular risk but its ability to reclassify patients in different risk categories is limited, as discussed above. It is estimated that more than half of the patients presenting with ST elevation myocardial infarction would not be considered candidates for intensive preventive therapy by current clinical algorithms using traditional risk factors and CRP (364). Imaging studies have been proposed as screening tools in additional to existing clinical predictors in patients of all ages. The only procedures that could be applied to large number of patients are the determinantion of carotid intima-media thickness (CIMT) and electron beam computer tomography (EBCT). The latter has been extensively studied.
The 2010 ACCF/AHA Guideline for Assessment of Cardiovascular Risk in Asymptomatic Adults recommended the use of Computed Tomography for Coronary Calcium Scoring in cardiovascular risk assessment in persons at low to intermediate risk (6% to 10% 10-year risk) and intermediate risk (10% to 20% 10-year risk). CT calcium scoring produces the same amount of radiation as 1 to 2 mammograms performed on each breast (365). In a study including two cohorts of asymptomatic individuals totaling over 35,000 subjects, after adjustment for age, gender, ethnicity, hypertension, hyperlipidemia, diabetes, smoking and family history,coronary calcium score remained a strong predictor of mortality at 5 years and 12 years follow-up (366). In this population, the authors estimated that the cost of finding a high risk patient in patients over 60 years with this methodology varies with the age and gender between $247 and $752 (367).

Coronary calcium score si a powerful tool to reclassify the level of risk. In the Multiethnic Study of Atherosclerosis (MESA), investigators found that the addition of coronary artery calcium (CAC) scores to standard risk factors improved the area under the ROC curve from 0.77 to 0.82 (p_0.001) (368). In another publication the investigators of MESA estimated that the information provided by coronary calcium reclassified an additional 23% patients who experienced a cardiovascular event in the next 5.8 years (median) in the high risk group and an additional 13% patient free of events in the low risk group (369).In the Rotterdam Study, 2,028 asymptomatic subjects aged 69.6 + 6.2 years were evaluated at baseline using the Framingham Score and Coronary calcium score (370). During a median follow-up of 9.2 years, 135 hard coronary events occurred. The authors estimated that 52% of the patient in the Framigham intermediate riisk group were reclassified in more appropriate risk categories.

In older patients CAC continues to be a powerful predfictor of mortality. In Honululu Heart Program in 221 men aged 84 to 96 years observed for 3 years 17 deaths occurred, none opf them if CACS was below 10 (371). Compared to patients with a CAC score (CACS) of 0 to 10, subjects aged over 80 years have 6.3 times higher risk of death for a CACS of 11 to 100, 5.3 times hiogher for a CACS of 100 to 400 and 11.7 times higher for a CACS over 400 (367). Same as CRP, vascular calcification is not a specific marker for coronary events in this age group. In the Study of Osteoporotic Fractures, of 9,704 Caucasian Women aged 65 years or more, 2,056 had aortic calcifications (372). Over 13 years follow up there was an increase in cardiovascular death but also in cancer death (after multivariate adjustment and in noncardiovascular, noncancer death. In elderly women enrolled the Cardiovascular Health Study,CACS was associated with gait speed (373).

The correlation between CACS and CRP is modest. In the Prospective Army Coronary Calcium Study there was no correlation between CRP and CACS (374). In MESA, inflammatory markers were weakly associated with CAC presence and burden. The authors beleved that their data support the hypothesis that inflammatory biomarkers and CAC reflect distinct pathophysiology (375). Using intravascular ultrasound, Kubo et al showed that CRP is associated with the size of the necrotic core of the plaque but not with the the percentage of dense calcium, fibrofatty tissue, and fibrous tissue (267).

In appropriately selected asymptomatic patients, the absence of CAC predicts excellent survival with 10-year event rates of approximately 1% (376). In MESA, 59% of the JUPITER-like population had CAC=0 and CHD event rates in this group were less than 1 event per 1000 person-years (377). The authors estimated that the NNT-5 for CACS of 0, 0-100 and over 100 were respectively 177, 57 and 20. The authors proposed to select patients for cardiovascular prevention studies in the elderly based on selection according to CACS.

Cholesterol lowering in patients with vascular calcification

 

CRP has been used for screening and selection of patients at risk but also as a target for statin intervention. CACS is an unlikely candidate for evaluation of results of therapy.  Nicholls et al sought to determine the relationship between coronary calcification and plaque progression in response to established medical therapies using intravascular ultrasound (378). The interventions were either antihypertensive or lipid lowering drugs. The continuous rate of change in percent atheroma volume (PAV) and total atheroma volume was similar in patients with a lower and higher calcium index and a lower calcium index was associated with a higher rate of patients showing substantial change in atheroma burden (at least 5% change in PAV, 70% vs. 53%, p < 0.001). Clinical trials using vascular calcifications as endpoint were generally unsuccessful.

The St. Francis Heart Study (ST-FRANCIS) enrolled 1,005 asymptomatic healthy men and women age 50 to 70 years with coronary calcium scores at or above the 80th percentile for age and gender (379). They were randomized to atorvastatin 20 mg daily, vitamin C 1 g daily, and vitamin E 1,000 U daily, versus matching placebos. Mean duration of treatment was 4.3 years. Treatment reduced TC from 26.5% to 30.4% (p < 0.0001), LDL from 39.1% to 43.4% (p < 0.0001), and triglycerides from 11.2% to 17.0% (p < 0.02) but had no effect on progression of coronary calcium score. Treatment also failed to significantly reduce the primary end point of a composite of all atherosclerotic cardiovascular disease events.

The Scottish Aortic Stenosis Lipid Lowering Trial, Impact on Regression (SALTIRE) enrolled 155 patients with calcific aortic stenosis and randomized them between atorvastatin 80 mg/day and placebo. At 2 years, atorvastatin reduced LDL (53%, p < 0.001) and CRP (49%, p < 0.001) whereas there was no significant change with placebo (380). At the end of 25 months there was no clinical benefit in terms of either progression of aortic stenosis or of valve calcification. Houslay et al. further investigated the effects of statin on coronary artery calcification (381). They followed 102 patients, mean age 70 years, with coronary artery calcification. The rate of change in coronary artery calcification was 26% log arbitrary units (AU) per year in the atorvastatin group and 18% log AU per year in the placebo group, with a geometric mean difference of 7%/year (95% confidence interval 3% to 18%, p = 0.18). LDL did not correlate with the rate of progression of coronary calcification.

Schmermund et al. evaluated the effect of low versus high-dose atorvastatin on calcified coronary atherosclerosis (382). Four hundred and seventy one patients, mean age 61 years, who had no history of coronary artery disease and no evidence of high-grade coronary stenosis with >/= 2 cardiovascular risk factors and moderate calcified coronary atherosclerosis (CAC score >30) were enrolled and followed for 1 year. After pre-treatment with 10 mg of atorvastatin for 4 weeks, patients were randomized to receive 80 mg or 10 mg of atorvastatin per day. The mean progression of CAC volume scores was not significantly different between the two groups. There was no relationship between on-treatment LDL levels and the progression of calcified coronary atherosclerosis. LDL was reduced from 106 mg/dL to 87 mg/dL in the group randomized to receive 80 mg of atorvastatin (P<0.001), whereas levels remained stable in the group randomized to receive 10 mg.

The Aggressive Versus Moderate Lipid-Lowering Therapy in Hypercholesterolemic Postmenopausal Women: Beyond Endorsed Lipid Lowering With EBT Scanning (BELLES) trial evaluated a subgroup of 615 postmenopausal women and randomized them to high-dose atorvastatin versus moderate-dose pravastatin (383, 384). LDL reductions were 46.6% and 24.5% in the intensive and moderate treatment arms, respectively (P<0.0001); the respective decreases were TC 33.8% and 17.2% (P<0.0001); TG 21% and 19% (P<0.0001); HDL 2.3% and 3.9% (p=0.6), and apo B was 39% and 21.2%. In postmenopausal women, intensive statin therapy for 1 year caused a greater LDL reduction than moderate therapy but did not result in less progression of coronary calcification.

SEAS was a randomized, double-blind trial involving 1873 patients with asymptomatic aortic stenosis (385). The patients received daily either 40 mg of simvastatin plus 10 mg of ezetimibe or placebo. The primary outcome was a composite of major cardiovascular events, including death from cardiovascular causes, aortic-valve replacement, nonfatal myocardial infarction, hospitalization for unstable angina pectoris, heart failure, coronary-artery bypass grafting, percutaneous coronary intervention, and nonhemorrhagic stroke. During a median follow-up of 52.2 months, the primary outcome occurred in 333 patients (35.3%) in the simvastatin�ezetimibe group and in 355 patients (38.2%) in the placebo group (hazard ratio in the simvastatin�ezetimibe group, 0.96; 95% confidence interval [CI: 0.83 to 1.12; P = 0.59). Fewer patients had ischemic cardiovascular events in the simvastatin�ezetimibe group than in the placebo group (hazard ratio, 0.78; 95% CI, 0.63 to 0.97; P = 0.02), mainly because of the smaller number of patients who underwent coronary-artery bypass grafting. Cancer occurred more frequently in the simvastatin�ezetimibe group (105 vs. 70, P = 0.01).

Although vascular calcification is a powerful and independent risk factor for CHD, and is partially related to dyslipidemia, trials have not found an association between the use of statins, control of dyslipidemia and progression of calcifications.

 

Statins in patients with ESRD and CHF

ALERT enrolled 2,102 renal transplant patients and randomly assigned them to either fluvastatin 40 mg / day or placebo for 5.1 years 385, 386). The primary endpoint was the occurrence of a major adverse cardiac event, defined as cardiac death, nonfatal myocardial infarction, or coronary intervention procedure. The primary endpoint was not significantly reduced but there were fewer cardiac death or nonfatal MI in the treated group (risk ratio 0�65 [95% CI 0�48�0�88], p=0�005).
The Deutsche Dialyse Diabetes Studie (4D) enrolled 1,255 patients with type 2 diabetes receiving hemodialysis (387, 388). The subjects were randomized for 4 years to either placebo or atorvastatin 20 mg/day. The endpoint was cardiovascular death, nonfatal myocardial infarction and stroke. There was no reduction in the endpoint or any of its components, and there was a doubling of the rate of fatal stroke, which was statistically significant. The authors concluded that statin intervention might not be effective when atherosclerosis is too advanced. The authors studied in a retrospective analysis the role of CRP in cardiovascular risk and benefit in these patients. The risk of cardiovascular events increased with CRP levels but the benefit of statin therapy did not (386).
In AURORA, 2,776 dialysis patients aged 50 to 80 were randomized to received rosuvastatin 10 mg per day or placebo (387). The combined primary endpoint was death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. Secondary endpoints included death from all causes and individual cardiac and vascular events. After a median of 3.8 years the primary endpoint, its components and all secondary endpoints were not significantly different in the two arms of the study.
The effect of statins in elderly patients with congestive heart failure was tested in CORONA (388). In this study 5,011 patients over 60 years with ischemic cardiomyopathy and systolic heart failure with class II, II and IV New York Heart Association symptoms, were randomized to be assigned to received 10 mg rosuvastatin or placebo. After a median of 32.8 months, there was no significant decrease in the primary outcome, death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. There was, however, a significant 14% decrease in hospitalizations for cardiovascular events. The drug was well tolerated.
GISSI-HF enrolled 5,584 patients aged 18 years or older with chronic heart failure of New York Heart Association class II�IV, irrespective of cause and left ventricular ejection fraction, and randomly assigned them to rosuvastatin 10 mg daily or placebo (389). After 3.9 years, there was no significant reduction of the primary endpoints: time to death and time to death or admission to hospital for cardiovascular reasons.
So far, cholesterol lowering trials in patients with biological characteristics similar to elderly frail subjects do not support a statin benefit in these patients.

Statin clinical trials with outcomes other than cardiovascular acute events

 

Peripheral Arterial Disease

Statins also increase concentrations of nitric oxide in the endothelium, which has vasodilator properties. These properties may also benefit people with known peripheral arterial disease (PAD) and therapy with statins was hypothesized to improve symptoms of PAD. Five small clinical trials showed a statistically (and arguably clinically) significant effect of simvastatin or atorvastatin on pain-free walking distance in patients with claudication (390-394).

Preoperative use

Statins have been shown in some studies, but not in others, to decrease perioperative morbidity, particularly in vascular interventions (395-401). It is believed that this is also a result of the effect of statins on endothelial function (395). Larger randomized trials are necessary to verify this property of this class of drugs.

Use prior to percutaneous revascularization

Statins have been shown to reduce the risk of periprocedural myocardial infarction when administered for a week prior to an elective percutaneous revascularization procedure (402). High dose statins were found to be more efficient than low dose statins (403). In patients with acute coronary syndrome, single dose atorvastatin (404) or rosuvastatin (408, 409) decreased periprocedural additional myocardial injury. The single loading dose was tested also in patients undergoing elective coronary stenting and found to decrease periprocedural myocardial infarction (405). In patients already on statins, reloading with supramaximal doses of the same statin has been found to be beneficial (406). The endpoint of these studies was the elevation of CK-MB and troponins the day after the procedure. These rapid effects of statins are most likely attributable to their pleiotropic effect.

Renal endpoints

Clinical trials of statins (including HPS) have documented a significant beneficial effect of statins on albumin excretion rate. (407-411). There is also a statistically significant benefit in the rate of decrease of glomerular filtration rate in patients with atherosclerosis but its clinical significance is questionable (409). Ongoing large clinical studies will ascertain this effect and compare the effect of different statins, particularly in subjects with nephrotic syndrome.

Anti-inflammatory effects

In animal experiments, pretreatment with statins reduces the effect of induced acute inflammation (412). In a small clinical trial atorvastatin acted as a nonspecific anti-inflammatory agent in patients with rheumatoid arthritis (413). This use of statins requires additional confirmation.

Anti-arrhythmic effects

Observations have been published documenting that patients treated with statins are less likely to have atrial fibrillation or ventricular arrhythmias (414-418). The mechanism of this drug action is unknown and must be confirmed in larger clinical trials,.

Prevention and treatment of Alzheimer�s disease.

PROSPER and HPS have enrolled large numbers of elderly subjects and specifically looked for an effect of statins on cognitive impairment, but no effect was seen. In spite of this, the relationship between cholesterol metabolism and Alzheimer�s disease continues to fascinate researchers and the lay public. While the internet has numerous reports of cases of alleged cognitive impairment induced by statins, small studies seem to document some benefit in small subsets of patients with mild Alzheimer�s disease or vascular dementia(419-421). The majority of researchers remain skeptical that a clinically significant benefit is present. This conclusion is supported by a recent Cochrane Collaboration analysis (422).

Cardiovascular Trials with Other Lipid Lowering Drugs

 

Fibrates

Although in recent years, new data has emerged in the statin trials regarding elderly patients, there is a paucity of data involving other hypolipidemic agents. Fibrates have shown some promise as lipid lowering agents as they increase high-density lipoprotein cholesterol (HDL) and decrease triglycerides and LDL (see Table 6).

Table 6: Fibrates
Study	Type of patients	Randomization arms	Duration; Primary outcome; Secondary outcomes	Clinical results	Older group
ACCORD (423)	Type 2 diabetes with cardiovascular disease or two additional risk factors	Simvastatin in all patients + fenofibrate or Placebo	Primary outcome: Nonfatal MI, nonfatal stroke or cardiovascular death	No significant difference in any outcomes. Possible benefit to men and harm to women. Possible benefit if triglycerides high and HDLc low	Average age 62 years. No benefit in any age group.
BECAIT (424)	92 survivors of MI younger than 45 years	Bezafibrate 400mg/day vs. Placebo	Primary outcome: Angiographic stenosis Secondary outcome: coronary events	There were 3 coronary events in the bezafibrate group and 11 in the placebo group (p=0.02)	n/a
BIP (425)	3,090 patients with previous MI or stable angina, TC 180-250 mg/dL, HDL </=45 mg/dL, TG </=300 mg/dL, and LDL </=180 mg/dL	Bezafibrate 400mg/day vs. Placebo	6.2 years; fatal or non-fatal MI or sudden death	Primary outcome:  13.6% vs. 15.0% (NS). Reduction in the cumulative probability of the primary outcome by bezafibrate was 39.5% (p=0.02) when TG >/= 200 mg/dL. Total and noncardiac mortality rates were similar, and adverse events and cancer were equally distributed.	Not announced
FIELD (426)	9,795 patients age 50-75 years with type 2 DM, not on statin therapy	After a placebo and a fenofibrate run-in phase,  patients randomized to micronized fenofibrate 200 mg/day vs. placebo	5 years; coronary events	Fenofibrate did not significantly reduce the risk of the primary outcome. There was a 24% reduction in non-fatal MI (p=0.01) and an increase in CHD mortality (NS). There was a  21% reduction in revascularizations (p=0.003).	No benefit over age 65 years
Helsinki (427)	4,081 men aged 40-55 years with hyperlipidemia	Gemfibrozil 600 mg twice daily vs. Placebo	5 years; fatal or non-fatal MI and cardiac death	Primary outcome NS. 34% reduction in coronary events with gemfibrozil (p<0.02). Each 1% increase in HDL was associated with a 3% reduction in coronary events, independent of changes in LDL and TG.	n/a
MEADE (428)	1,568 men with lower extremity arterial disease	Bezafibrate 400mg/day vs. Placebo	4.6 years; combination of CHD and stroke; all coronary events, fatal and non�fatal coronary events separately, and strokes alone	Reduction in primary outcome, fatal and non-fatal CHD, strokes NS. In men age <65 years, non-fatal coronary events reduced (p=0.02), and all events combined were 62% lower in bezafibrate group (p=0.01).	Benefit limited to enrollees under 65 years
VAHIT (429)	2,531 men age <74 years with documented CAD and HDL </=40 mg/dL, LDL </=140 mg/dL, TG </=300 mg/dL	Gemfibrozil 1200mg/day vs. Placebo	5.1 years; MI or death from CAD	Primary event: 17.3% vs. 21.7% (p=0.006). Overall reduction in risk of an event was 4.4%, and the reduction in relative risk was 22% (p=0.006). There was 24% reduction in the combined outcome of death from CHD, nonfatal MI, and stroke (p<0.001). Differences in the rates of coronary revascularization, hospitalization for UA, death from any cause, and cancer all NS.	In patients over age 66 years. Gemfibrozil was associated with 26% reduction in primary outcome (p=0.0070

• A meta-analysis of these studies including all age groups shows significant benefit (Fig 4).Unfortunately some of the studies did not include elderly and in some that included subjects over age 65 years, there was no benefit for them Fibrate studies are controversial.
• Both gemfibrozil studies were successful. Both were done in men. Studies using other available fibrates were at best borderline successful.
• In the FIELD trial there was a trend for more benefit to women. The opposite is true for the ACCORD trial. 
• In the BIP trial and in the ACCORD trial benefit was present in patients with high triglycerides and low HDL cholesterol.  In the VA-HIT most of the benefit was attributable to patients with insulin resistance. In the Helsinki Heart Study patients with high triglycerides and low HDL cholesterol benefitted more. None of the studies has addressed directly the issue that, in clinical practice, the patient who receives fibrate in addition to statin therapy is the patient with high triglycerides and low HDL cholesterol.

Figure 4: Metaanalysis of reduction of risk of main coronary events in patients of all ages treated with fibrates.

                                  

In summary, the reduction in major coronary events was significant in the two studies using gemfibrozil, VAHIT, and Helsinki Heart Study, and only trended toward significant in the other fibrate trials. Fibrates may have benefit in select patients in whom statins are not an option and when HDL or TG modification is indicated.

Niacin

 

 Two clinical endpoint, placebo-controlled trials were performed with niacin. Niacin monotherapy was used in the coronary drug trial, which enrolled 2,789 myocardial infarction survivors in the placebo group and 1,119 in the niacin group (430). At the end of 5 years, there was a significant difference in recurrent MI (27%), revascularizations (47%) and stroke (26%). After 15 years the niacin treated subjects had a significantly lower (11%) mortality (431). The HDL-Atherosclerosis Treatment study (HATS) enrolled 160 patients with stable coronary artery disease in order the study the effect of simvastatin-niacin combination therapy and / or the effect of antioxidants on the progression of coronary atherosclerosis (432). At the end of two years there were 12 cardiovascular events in the placebo group and 1 in the simvastatin-niacin group (p=0.003). Treatment effect was diminished by co-administration of antioxidants. There are no data showing the clinical efficacy of niacin in the elderly.

Lovaza

Gissi-Prevenzione enrolled 11,324 survivors of myocardial infarction in a study randomizing them between Lovaza 1 g/day and placebo (433, 434). After 3.5 years there was a significant 15% reduction in the combined endpoint of death, nonfatal myocardial infarction and stroke. Co-administration of vitamin E was neither effective nor detrimental. Sixteen percent of the patients were over age 70 years but their benefit was not reported separately. .

Recently, OMEGA included 3,827 MI survivors mean age 64 years in a similar trial and failed to confirm the benefit (435)  The authors speculate that the difference is attributable to the more aggressive risk factor intervention in their patients.

ALPHA-OMEGA assigned 4837 patients, 60 through 80 years of age (78% men), who had had a myocardial infarction and were receiving state-of-the-art antihypertensive, antithrombotic, and lipidmodifying therapy to receive for 40 months one of four trial margarines: a margarine supplemented with a combination of EPA and DHA (with a targeted additional daily intake of 400 mg of EPA�DHA), a margarine supplemented with ALA (with a targeted additional daily intake of 2 g of ALA), a margarine supplemented with EPA�DHA and ALA, or a placebo margarine. Neither EPA�DHA nor ALA reduced the primary end point of major cardiovascular events, which comprised fatal and nonfatal cardiovascular events and cardiac interventions. In patients with diabetes, there was a borderline decrease of coronary death in patients taking EPA-DHA, 0.51 (0.27�0.97) and of ventricular arrhtythmia related events in patients taking ALA, 0.39 (0.17�0.88) (436).

SU.FOL.OM3 enrolled 2501 patients with a history of myocardial infarction, unstable angina, or ischaemic stroke, aged 45-80 years in a randomized trial a 2-2 factorial design, The investigators examined the effects of a daily dietary supplement containing 5-methyltetrahydrofolate (560 μg), vitamin B-6 (3 mg) and B-12 (20 μg) versus placebo, and a supplement containing omega 3 polyunsaturated fatty acids (400 mg of EPA 200 mg DHA) versus placebo. There was no benefit from either intervention (437).

In GISSI-HF there was a 9% statistically significant reduction in death or hospitalization for heart failure and 8% reduction in mortality (438). The authors concluded that treatment of 56 patients for 3.9 years will prevent one death. Although this means a NNT-5 of 72, the cost of the medication is low compared with other drugs used to prolong life in these patients.
The investigators of JELIS randomized 18,645 patients with or without preexisting coronary artery disease and a serum cholesterol higher than 250 mg/dL to statin + a preparation containing 1800 mg EPA or statin + placebo (439). After 4.6 years, there was a significant reduction in the risk of major coronary events and nonfatal coronary events, each of 19%. Most of the events occurred in the secondary prevention subcohort. there was a significant 19% decrease in risk of major coronary events, a 28% decrease in the risk of unstable angina and a 20% reduction in stroke (440), There was a higher reduction in events in survivors of myocardial infarction (441) and patients with peripheral arterial disease (442). EPA treatment reduced significant the risk of coronary events in patients with high triglycerides and low HDLc (443), patients with abnormal glucose metabolism (444).
The role of omega 3 fatty acid supplementation in cardiovascular prevention is not yet clarified.

Probucol

Probucol was withdrawn from the US market in the 1990's after a clinical trial showed no efficacy of improvement in femoral atherosclerosis. It acts as an antioxidant and lowers moderately LDL cholesterol. It may improve reverse cholesterol transport, in spite of decreasing HDL cholesterol levels. It is used in Japan where it has been shown to cause marked regression of cutaneous and tendon xanthomata. Multiple recent studies have documented an effect of probucol in prevention of restenosis after angioplasty. Two studies performed in older subjects have documented that this effect is clinically relevant.

• The Probucol Angioplasty Restenosis Trial (PART) enrolled 90 subjects with coronary artery disease requiring angioplasty, average age 60 years (445). They were randomized between probucol 500 mg/day and placebo for one year. Probucol therapy was a negative predictor of repeat revascularization (p=0.034).
• The Fukuoka Atherosclerosis Trial (FAST) enrolled 246 asymptomatic hypercholesterolemic patients, average age 66 years, to receive probucol 500 mg/day or pravastatin 10 mg/day or placebo (446, 447). The primary endpoint was carotid IMT progression and the secondary endpoint cardiovascular events. Over two years probucol decreased significantly the rate of IMT increase and reduced significantly cardiovascular events (2.4% vs. 13.6%; p=0.0136).

Treatment of Hyperlipidemia-induced Pancreatitis

Hyperlipidemia is a rare, but well-established cause of acute and recurrent pancreatitis, and it is estimated that hypertriglyceridemia is the cause of acute pancreatitis in about 1-4% of cases although biliary tract disease and alcohol use cause the majority of cases (448, 449). The overall mortality rate of patients with acute pancreatitis is 10-15%, but pancreatitis caused by hypertriglyceridemia tends to be more severe than other forms (450). The median age of onset of acute pancreatitis is dependent on type, with elderly individuals affected most by pancreatitis caused by biliary tract disease and trauma. Idiopathic pancreatitis, a type including hypertriglyceridemia-induced pancreatitis, increases in frequency with age. Although genetic disorders resulting in massive increase of triglyceride concentration manifest themselves at a younger age, age increases the incidence of metabolic syndrome, worsening genetic hypertriglyceridemia.
While the exact mechanism of pancreatic injury is unknown, it has been proposed that free fatty acid release is primarily responsible. When triglycerides and phospholipids (lecithin) comprising chylomicrons or very-low-density lipoprotein particles (VLDL) are hydrolyzed by pancreatic lipase, the resulting free fatty acids and lysolecithin cause chemical injury to pancreatic acinar cells and capillaries. This leads to further release of lipase, thereby perpetuating the inflammatory cycle. Clinically significant pancreatitis usually does not occur until triglycerides reach at least 1000 mg/dL (451-454). The activity of lipoprotein lipase (LPL) is crucial in removing triglycerides from the plasma. LPL gene mutations and dyslipoproteinemia (seen in Type I, IV, or V hyperlipidemia) combined with secondary causes of lipoprotein abnormalities (such as poorly controlled diabetes, proteinuric kidney disease, hypothyroidism, pregnancy, alcohol use, obesity, and certain medications) can cause hypertriglyceridemia-induced pancreatitis (454-456).
Acute pancreatitis due to hyperlipidemia should be diagnostically and therapeutically managed in much the same way as pancreatitis from any other cause. This should be supplemented by other therapeutic options aimed specifically at treatment and prevention of hypertriglyceridemia.

Diet

In addition to resting the bowel in the setting of acute inflammation, acute fasting serves to cut off the exogenous supply of triglycerides. This also causes reduced hepatic VLDL secretion. Long-term dietary modification in patients with hypertriglyceridemia should include restriction of simple carbohydrates and alcohol. Dietary fat should be limited to 10-15% of total daily caloric intake. Medium-chain triglycerides (found in coconut and palm oils) have also been used, as their absorption does not require chylomicron formation (454).

Insulin and Heparin

Insulin and heparin are both potent stimulators of LPL activity, and therefore reduce hypertriglyceridemia. Several case reports have been published describing the safety and efficacy of these therapies in the treatment of hypertriglyceridemia-induced pancreatitis in diabetic and non-diabetic patients (457-462). However, no randomized controlled trials comparing these therapies to standard treatment of pancreatitis have been published.
Berger et al. reported five patients with hypertriglyceridemia-induced pancreatitis, all with initial triglyceride levels > 1000 mg/dL, who received heparin and/or insulin administered intravenously by continuous infusion (in addition to standard treatment), initiated on day of presentation. Heparin dose was adjusted to maintain normal coagulation parameters, and insulin dose was adjusted to maintain blood glucose 120-160 mg/dL. Triglyceride levels decreased to < 500 mg/dL by day 3 in all cases (457). Case reports of other dosing schedules of insulin, such as lispro insulin sliding scale administered every 4 hours, or 20 units of regular insulin administered intravenously with dextrose infusion over 24 hours, have also been shown to decrease serum triglyceride levels, though somewhat less effectively (457-459). Various effective dosing schedules of heparin have also been reported, including the continuous infusion described above, as well as heparin 5000 units twice daily and 10,000 units daily (both intravenously) (457, 459, 460).
The duration of effective heparin use is also not clearly established. Nasstrom et al. (461) measured plasma LPL and hepatic lipase prior to, during, and serially after an 8 hour heparin infusion in 10 healthy subjects. Peak activity of LPL was seen at 1 hour, decreased to 20% by 2 hours, and 15% at 8 hours. A second bolus of heparin administered after 4 hours caused LPL levels to increase to only 35% of the original peak level. These results suggest that LPL released into plasma by heparin is susceptible to hepatic degradation, and after 1 hour, the stores of LPL are dependent on release of newly synthesized molecules. The use of heparin in the setting of acute hypertriglyceridemia-induced pancreatitis must be considered carefully, as a paradoxical increase in triglycerides due to depletion of LPL is a possibility (461, 462).

Plasmapheresis

Several reports have described the benefits of plasmapheresis in the setting of acute pancreatitis caused by hyperlipidemia (463-468). Plasmapheresis serves to directly remove circulating chylomicrons, triglycerides, and excessive proteases in the plasma. Additionally, introduction of donor plasma can replenish the recipient plasma with fresh LPL and apolipoprotein C-II (a required cofactor for LPL).
Yeh et al. examined 17 patients with hypertriglyceridemia-induced pancreatitis who all received one or two consecutive treatments with plasmapheresis. The mean removal rates of triglycerides, amylase, and lipase were 66.3%, 70.0%, and 84.8%, respectively, after one treatment. A second treatment increased the removal rate to 83.3%, 85.5%, and 87.0%, respectively (463). Kyriakidis et al. (452) evaluated 10 patients with hyperlipidemic pancreatitis, all of whom were given standard treatment as well as plasmapheresis within 48 hours of admission. Triglyceride levels were lowered in all cases, and plasmapheresis improved abdominal pain and other signs and symptoms of pancreatitis.
Data from numerous case reports suggest that outcomes are best when plasmapheresis therapy is started as soon as possible, preferably within the first 48 hours of presentation (467-469). In most of these cases, plasmapheresis was used at least two times during the hospital course. In some cases, observance of the appropriate diet and use of lipid-lowering drugs is insufficient to maintain safe triglyceride levels. In these instances, routine plasmapheresis may be of value in preventing recurrent acute pancreatitis (456, 471). Piolot et al. reported two cases of severe hypertriglyceridemia with recurrent acute pancreatitis. Plasmapheresis was then performed every 4 weeks and reduced the incidence of pancreatitis (470).

Fibric Acid Derivatives

Fibrates have been effectively used in both the acute and long-term treatment of hypertriglyceridemia-induced pancreatitis. Fibrates cause an increase in LPL activity, a decrease in apolipoprotein C-III (a potent LPL inhibitor) concentration, and reduced hepatic secretion of VLDL and triglycerides. They can lower triglycerides by approximately 35-50%, and increase high-density lipoprotein (HDL) by about 20%.
In the Helsinki Heart Study, gemfibrozil 600 mg twice daily for 5 years significantly reduced triglyceride levels by 43% (427). The Veterans Affairs High-Density Lipoprotein Intervention Trial (VA-HIT) demonstrated that gemfibrozil 1200 mg daily for 5 years significantly reduced triglycerides by 33%, and increased HDL by 6% (429). In the Bezafibrate Infarction Prevention Study, bezafibrate 400 mg daily for 6 years significantly decreased triglycerides by 21%, and increased HDL by 18% (425). Data from these studies have shown that the benefits of fibrate therapy were greatest in patients with triglyceride ≥ 200 mg/dL.
Elevated plasma viscosity is another proposed mechanism by which hypertriglyceridemia increases pancreatitis risk, and may be decreased with fibrate therapy. Stein et al. (471) prospectively studied 24 adult patients with severe hypertriglyceridemia, and measured plasma lipids and plasma viscosity before and after gemfibrozil 1200 mg daily. Triglyceride levels decreased by 70% (p<0.001) and plasma viscosity decreased by 0.082 mPa/s (p=0.003).

Niacin

Niacin (vitamin B3) is effective alone or in combination with fibrates to decrease triglyceride levels. In the Coronary Drug Project, one of the largest randomized controlled trials involving niacin to date, 8,341 men with history of myocardial infarction were randomized to several treatment arms including niacin 3 grams per day for 6.2 years (430). The niacin group had a 26% reduction in triglyceride levels.
In most cases, niacin reduces triglycerides by at least 25-30%, and increases HDL by about 15-35%. In higher doses (at least 1.5 grams per day), niacin may decrease triglycerides by as much as 50%. However side effects such as elevation of hepatic transaminases, flushing, and rash may limit the use of niacin as a first line agent in the setting of hyperlipidemia-induced pancreatitis.

Fish Oil

Increased intake of -3 fatty acids via diet and/or dietary supplements is also recommended for long-term use, although no studies have been done in the acute setting of pancreatitis. These polyunsaturated fats cause decreased production of hepatic VLDL and increased clearance of plasma chylomicrons, and can significantly reduce plasma triglyceride levels when consumed in adequate amounts (at least 4 grams per day). Fatty fish such as sardines, herring, and mackerel are rich in  -3 fatty acids, but purified capsules of fish oil are usually necessary to achieve sufficient doses for triglyceride reduction.

Anti-oxidants

Anti-oxidant therapy with vitamin C, alpha-tocopherol, selenium, and beta-carotene have been shown to successfully reduce the incidence of recurrent pancreatitis in patients with familial LPL deficiency and hypertriglyceridemia (472). One of the proposed mechanisms for this effect of anti-oxidants is the protective effect against free radical damage of pancreatic acinar cells by ischemic changes seen in pancreatitis.

Acipimox

Acipimox, an orally active xenobiotic nicotinic acid analogue, has anti-lipolytic activity, decreases free fatty acid release into the bloodstream, and causes increased clearance of VLDL and triglycerides. Acipmox can be used as an adjunct to other anti-hyperlipidemic drug therapy in the setting of hyperlipidemia-induced pancreatitis (455). It is not available in the United States.

Pioglitazone

In patients with type 2 diabetes, hyperlipidemia and risk of pancreatitis pioglitazone is indicated on the grounds that it has additional triglyceride lowering efficacy.  In average dose lowering of triglycerides is in the range of 10-15% (187, 478). This is not sufficient for the drug to be a first line one for prevention of pancreatitis but it can add to the triglyceride lowering of a complex medial management plan.
***
After the initial treatment of acute hypertriglyceridemia-induced pancreatitis, therapy should be aimed at preventing recurrent episodes by controlling triglyceride levels. Because triglyceride levels can fluctuate significantly (depending on dietary intake, alcohol, etc.), they should be maintained well below 1000 mg/dL. As most cases develop when a genetically predisposed individual is exposed to a secondary risk factor, prevention of pancreatitis must involve careful treatment of these factors. This may include dietary modification, weight loss, abstinence from alcohol, control of diabetes, and treatment of concurrent hypothyroidism.

SAFETY OF LIPID LOWERING DRUGS IN THE ELDERLY

Pathophysiology of Drug Interaction and Toxicity

The likelihood of a drug related adverse event to occur is associated with the individual susceptibility for a specific drug effect. In most cases this is either genetically determined or related to accumulation of toxic doses of the drug and/or its metabolites. This accumulation occurs through disease or through drug interaction.
An example of genetically determined susceptibility to drug related adverse events is the high concentrations of rosuvastatin present in the serum of Asian subjects upon administration of average doses (473). Because of this it is recommended that this drug be started at a lower dose and increased with utmost care when administered to an Asian patient.
Disorders resulting in a higher likelihood of drug related adverse events are impairments of major organs such as chronic kidney disease, liver disease, congestive heart failure, or endocrine disorders (diabetes, hypothyroidism).
In situations in which the pathophysiology has been studied, drug related toxicity is attributed to drug accumulation to toxic concentrations. The risk of occurrence of this accumulation increases with age through a series of mechanisms:

• There is an age related decrease in glomerular filtration rate (474)
• There is an age related decrease in hepatic blood flow and decrease in drug clearance (475)
• Aging maybe associated with changes in induction and inhibition of different Cyp-450 enzymes (476).
• Aging might be associated with an increased expression of P-glycoproteins resulting in alterations in the rate of drug transport across cellular membranes (477) .
• Frailty (but not aging per se) has been associated with diminished drug esterase and conjugation activity (478).

Muscle toxicity of lipid lowering drugs

Muscle related adverse events are the most severe and the most feared side effects of lipid lowering drugs. From the least to most severe the entities are:

• Myalgia, defined as muscle ache with no biochemical indication of muscle pathology
• Creatine-kinase (CK) increases with no muscle symptoms
• Myopathy, defined as muscle symptoms of weakness or pain associated with CK elevation.
• Rhabdomyolysis, defined as muscle symptoms, CK elevation more than 10 times the upper limit of normal and presence of myoglobinemia, myoglobinuria or acute renal failure.

Because statins are by far the most prescribed lipid lowering drugs, muscle disorders are automatically associated with their use. In reality all lipid lowering drugs may cause muscle disorders. The mechanism through which statins induce muscle disorders is unknown. The following mechanisms have been proposed.

• Depletion of intracellular cholesterol pool (479).
• Diminished isoprenylation of muscle proteins through inhibition of HMG CoA reductase (480).
• Reduction of muscle co-enzyme Q10 (487-489).
• Reduction of muscle ubiquinone and reduction of respiratory chain enzyme and citrate synthetase (481).
• Disruption of the balance of muscle protein degradation and repair through statin effect on the ubuiquitin-proteasome pathway (482).
• Further inhibition of lipid oxidation by statins in a subgroup of subjects with diminished lipid oxidation (476).

The incidence of drug induced adverse events in the muscle is not known. There are three ways of obtaining information. FDA reports are useful to compare differences between drugs, but only 1% of adverse events are reported to FDA and the reports are subject to reporting bias determined by the belief of the reporter in the specific toxicity of a given drug. Meta-analyses of data from randomized clinical trials are scientifically correct but are subject to enrollee selection bias, particularly if the randomization occurs after a trial of "tolerance to the study drug". Claim databases are subjected to the prescribing bias with patients at higher risk of adverse events being prescribed drugs perceived as safer.

Myalgia

Randomized clinical trials report the frequency of myalgia. A few meta-analyses of these trials have been reported, all showing that the incidence of muscle complaints with normal CK levels is equal in the statin and placebo treated groups (492-494). This is in contrast with occasional reports documenting the presence of abnormal biopsy findings in patients with muscle complaints and normal CK (483). There is no standard recommendation for this clinical situation but most physicians will discontinue the statin and after a few weeks re-challenge the patient with the same statin or with a different statin.

Elevation of CK levels without symptoms

Although clinical trials have collected data, the significance of these biochemical findings is unknown. No recommendation has been made and most authors warn against monitoring of CK levels. Some authors recommend a baseline CK in high risk patients in order to have available a comparison, should muscle complaints emerge. Certain ethnic groups such as Philippinos or black people of Sub-Saharan-African descent have higher levels of tissue and serum CK (484).

Myopathy

There are difficulties in analyzing myopathy due to the fact that there is disagreement concerning how high does the CK level have to be in order to qualify for the diagnosis. Some studies have grouped myopathy with myalgias and muscle weakness. Since muscle symptoms are equal in placebo and statin treated patients, this would result in an underestimate of the relative risk of myopathy. A meta-analysis of 71,108 subjects enrolled in randomized clinical trials (485) has shown a 28.5% increase in myopathy defined this way in statin treated subjects (p<0.001). Another meta-analysis including over 86,000 subjects showed a 2.56 times higher risk of myopathy (486). Some authors believe that the risk of myopathy is dose related although there are no sufficient data to ascertain this hypothesis. Rosuvastatin 80 mg was tested but not marketed because of this perception (487). At high dose atorvastatin seems to be as safe as at low doses (488). Some authors believe that different statins have different potential risk of inducing myopathy with high dose simvastatin constituting a higher risk than atorvastatin and pravastatin a higher risk than fluvastatin (485).

Rhabdomyolysis

Rhabdomyolysis is a potentially fatal adverse event of lipid lowering drug use. Elderly subjects are the most likely subjects because of the alterations of drug metabolism listed above, the increased likelihood of co-morbidity and of polypharmacy. It is a rare but devastating occurrence. The syndrome presents as an acute or subacute disorder with proximal myalgias, "flu-like" symptoms, fatigue and muscle weakness (489). The average length of time on statin therapy before the onset of the clinical syndrome is one year. After the episode, the duration of myalgia is 2-3 months (490). Death and permanent renal dysfunction may occur but are relatively rare (~5% each).

In 2001, the FDA has removed from the market cerivastatin because of an unusually high risk of rhabdomyolysis, approximately 10-fold higher than for the other statins. Subsequently, most meta-analyses have analyzed the risk of each drug separately. In clinical trials, rhabdomyolysis is a rare event, so analysis of the data collected this way might not constitute a very reliable estimate of its incidence. Some meta-analyses reported no increased incidence of rhabdomyolysis in statin users. One study reported a number needed to harm (NNH) of 7,428 (485). A study analyzing the data available to FDA has reported a risk of death associated with fatal rhabdomyolysis of 0.15 per 1 million prescriptions (491). The risk of nonfatal disease is 15-20 times higher, with an NNH of 500,000. Other estimates are 0.0042% or an NNH of 25,000. A large administrative claim database is reporting the risk of "myopathy requiring hospitalization" to vary between 1.58 for fluvastatin to 3.49 for simvastatin per 10,000 patient years exposed (492). There is no report of incidence of adverse events by age group.

Hepatotoxicity of lipid lowering drugs

As opposed to muscle toxicity, liver toxicity is frequent, definitely dose related but not related to LDL lowering (493) and entirely reversible on discontinuation of the statin.  In randomized clinical trials the risk of transaminase elevation is significantly higher in patients receiving statin therapy. Elevation of transaminases up to 3 times the upper limit of normal is not a contraindication to statin therapy. Patients with elevated baseline liver enzymes do not have higher frequency of hepatotoxicity from statins than those with normal liver enzymes (494). The prevalence of metabolic syndrome increases with aging and the prevalence of fatty liver parallels the prevalence of metabolic syndrome. Either statins or fibrates decrease the level of elevation of transaminases in most studies. Thiazolidinediones have been proposed as therapeutic agents for non-alcoholic steatotic hepatitis. Of the 51,741 liver transplants performed in the US between 1990 and 2002, there were three patients in whom statin therapy was a possible contributor (494). Niacin has been associated with severe hepatotoxicity (495). Although severe hepatotoxicity is rare with crystalline niacin therapy, the advent of "non-flush" long-acting niacin was associated with multiple cases of liver transplant. Clinical studies done with NiaspanR have shown that transaminitis is rare if the maximum dose is below 2,500 mg/day. Fibrates, but not statins, have been associated with an increased incidence of cholelithiasis (496).

Other toxicity of lipid lowering drugs

Renal toxicity

The issue of renal toxicity was discussed when Phase III clinical trials of rosuvastatin showed cases of proteinuria and occasional of microhematuria in statin treated patients. In reality, no decrease in renal function is seen in statin treated subjects, and in some trials there is an improvement in GFR or albumin excretion rate induced by statins. Large clinical trials are recently addressing this issue. Fenofibrate, but not gemfibrozil, has been shown to be associated with an increase in creatinine, but its clinical significance is unknown (185, 506, 507).

Central and peripheral nervous system

The concern of the elderly patients of having a more rapid decrease in their cognitive function if treated with a statin has been inflated by the internet and the media. There are no clinical trial data to support this concern. Rare cases of peripheral neuropathy attributed to statins or fibrates have been reported. In a population study, patients diagnosed to have peripheral neuropathy were more likely to be treated with lipid lowering drugs after corrections for diabetes, end-stage renal disease, anemia and hypothyroidism (508, 509).  Another study estimated the NNH for statin induced neuropathy at 14,000 (497). Contrary to beliefs expressed in the late nineties that statin therapy and or low cholesterol is associated with depression and suicide, this is not confirmed by recent studies (498).

Cancer and lipid lowering drugs

A study performed in the 1970's under the auspices of the WHO randomized subjects between a fibrate no longer available, clofibrate, and placebo (499). At the end of the trial there was a decrease in the risk of myocardial infarction but an increase in the risk gastro-intestinal cancer. There is no such signal for the currently used fibrates. Off and on, an increase of cancer incidence is seen in the statin treated group in randomized clinical trials. Subsequent meta-analyses using larger number of events have unequivocally alleviated these concerns. In population studies, where prescribing bias is significant, a protective effect against certain forms of cancer was seen (500) but, again, larger databases have not confirmed this effect (513-519). A recent study attempted to determine the relationship between statin therapy and the three main forms of cancer in the elderly: lung, colo-rectal and breast (501). The authors concluded that it is unlikely that statins have any relationship with cancer incidence. In small studies statins have been used as an adjuvant to cancer chemotherapy (520, 521). More data are necessary to establish such practice.

Risk attributable to drug interaction

 

Different statins have different mechanisms of metabolism and elimination. In general the statins requiring an intact Cyp-450 3A4 system (lovastatin and simvastatin) for their metabolism are more sensitive to drug interaction. The main culprit in the medical literature is gemfibrozil. Since statins have complementary biological effect on lipids and each has superb clinical benefit in clinical trials using monotherapy, it is tempting to treat patients with statin and fibrate combination therapy. Unfortunately, gemfibrozil interferes with the metabolism of all statins (522-525) with the exception of fluvastatin ,and hundreds of cases of severe drug interaction resulting in rhabdomyolysis have been reported to FDA. The mechanism of action for increased toxicity of most statins is an inhibition of glycuronidation by gemfibrozil. Fenofibrate does not have the same potential for drug interaction. In pharmacokinetic studies fenofibrate was tested against each statin, and its presence resulted in no changes in statin kinetic parameters (526-528). Niacin has been also suspected to increase the risk of rhabdomyolysis induced by statins, but recent data do not seem to confirm this effect (491). Other main sources of drug interaction are Cyp-450 3A4 inhibitors: macrolides (except azythromycin) and azole antifungals (502). Nondihydropyridine calcium channel blockers are also moderate inhibitors of Cyp-450 3A4 (503). Statins have variable effect on INR in patients receiving anticoagulation. The statin least likely to require a modification of warfarin dosage is atorvastatin (504). The elderly are more likely to receive multiple drugs and also have an increased likelihood of severe events resulting from drug interaction. It is advisable to consult the package insert when adding a new drug to statin therapy. A large proportion of severe adverse events secondary to drug interaction occur in a patient who was tolerating  statin therapy but another drug was subsequently added and changed the metabolism of the statin.

In summary the risk of adverse events is not negligible in the elderly treated with lipid lowering drugs, but it is exceeded by far by the benefit in qualified subjects.

EPIDEMIOLOGY OF PRESCRIBING LIPID LOWERING DRUGS FOR THE ELDERLY

If elderly benefit as much as younger subjects from lipid lowering drugs, then age should not be a factor in prescribing or complying with the prescribed medications. Studies have shown that this is not the case. Patients with ischemic heart disease aged 65-74 were 36% less likely and patients over age 65 years were 84% less likely to be treated with statins as compared with patients aged less than 65 years (505). Treatment discontinuation did not account for these differences. The prescription of statins decreases progressively as baseline cardiovascular risk and future probability of death increases. A large study including patients 66 years or older estimated that the likelihood of prescribing decreases 6.4% with each year of age or with each 1% increase in the predicted 3-year mortality risk (506). Treatment discontinuation is very high in the older age group. In a study of patients older than 65 years, after two years of treatment, half of the secondary prevention and three quarters of the secondary prevention cohorts have discontinued statin therapy (507).
Since there is no scientific reason to deny treatment for elderly patients, the explanation for this phenomenon is speculative:

• Statins do not benefit subjects with less than two years life expectancy. This is correct but with increasing longevity of our population less and less applicable.
• Adherence is poor in patients with cognitive impairment. This may be so but it applies only to fraction of the population in whom it could be improved through outside intervention.
• Cost is an issue. Medicare reform and the transition of some of the lipid lowering drugs have reduced the burden of cost.
• Polypharmacy is an issue. This can be overcome by careful prescribing and counseling on benefits of taking each medication.
• Physician's compliance. Physician's compliance with the guidelines is affecting the patient's compliance with the prescribed drug.
• Omission, oversight, or organizational problems. They may lead to missed opportunities. This can be overcome by a practice protocol.
• Patient resistance to prescribing. It is usually due to lack of symptomatic relief by the drug, to perceived risks or to fatalistic attitude. It can usually be overcome by the physician.

Short term discontinuation of statin therapy has no serious health consequences for patients with stable coronary artery disease (508). Lack of treatment or long-term discontinuation of treatment can have a marked impact on morbidity and mortality in high risk patients. In a large study including 1.5 million veterans, statin prescribing was a negative factor for mortality after adjustment for age (509). The benefit of statins increased with cardiovascular risk score. In the National Registry of Myocardial Infarction the patients who continued statin therapy and patients who received an early new prescription for statin therapy at the time of their MI had an odds ratio of 0.46 and 0.42, respectively, for secondary prevention when compared with patients who never received statins (510). Patients in whom statin therapy was discontinued had a risk ratio of 1.25. The situation is similar in primary prevention. A study of statin treated patients, two third of them over age 60 years, looked at the relationship between LDL goal achievement in primary prevention and cardiovascular outcomes (511). Compared to patients with complete goal achievement over 3 years, temporary goal achievement and lack of goal achievement resulted in risk ratios of 2.34 and 2.99 respectively.
The patient adherence to the regimen is also very important. In statin user MI survivors, low adherers had a 1.25 risk ratio of death while intermediate adherers had a 1.12 risk ratio when compared to high adherers (512). After adjustment for age and co-morbidity, patients with diabetes non-adhering to prescribed statins have a two-fold increase in mortality and 40% increase in hospitalization (513, 514). In a cohort of over 20,000 patients free of cardiovascular disease and newly treated with statins, high adherence to statin therapy reduced nonfatal cardiovascular events by 20% (515).
Since the benefit of statin therapy in the elderly group seem to exceed the risks, several proposals have been made to improve the prescribing and/or the adherence:

• Initiation of treatment during hospitalization enforced by hospital protocols and audits (516).
• Computerized alerts (517).
• Academic detailing by pharmacists to primary care physicians (518).
• Administrative measures such as decreased co-payments by health plans (519).
• Physician continuity (520).
• Treatment to goal (521).

Of the above measures the first one has been demonstrated to have an impact on the long term adherence and persistence of treatment with statin therapy.

SHOULD LIPID LOWERING DRUGS BE PRESCRIBED FOR THE ELDERLY?

There is overwhelming evidence that treatment with lipid lowering drugs decreases cardiovascular risk and that this benefit is extended at least until age 75. In cardiovascular risk prediction age is a vague indicator of the presence of atherosclerosis. No guideline should use age limits for management of cardiovascular prevention. It is acceptable, however that patients younger than age 75 years have too advanced atherosclerosis for events to be prevented by lipid lowering drug therapy. In addition, it is conceivable that in frail elderly cholesterol predicts negatively cardiovascular risk and its further lowering might be an unwise intervention. In the current status, however, the burden of proof has shifted to the conservative approach. Elderly patients should receive lipid lowering drug therapy until proven that the recommendations should be limited to certain subgroups within this population.

INITIATION OF LIPID INTERVENTION

Initiation of lipid therapy is recommended, in vast majority of situations, for reduction of cardiovascular risk and in a minority of situations for prevention of acute pancreatitis, which is discussed in separate chapter. The decision to initiate an intervention for reduction of cardiovascular risk is based on estimation of global risk, estimation of longevity of the patient and estimation of individual risk of the therapeutic intervention. Cardiovascular prevention through lipid intervention has different distinct stages:

• Evaluation of goals (global risk, longevity and addressing patient's and family's concerns).
• Choice of atherogenic particle to target, and of its target level
• Choice of statin
• Choice of drugs to target additional or alternative targets
• Choice of alternative or additional drug used to reach target

 

Evaluation of goals

Obviously, lipid intervention in the elderly patient is to certain extent an art rather than a science. The guidelines are specifying that the therapeutic intervention should not be denied based on age alone but the decision to intervene should belong to the patient or to the patient's advocate. The global risk is evaluated based of different estimates of risk of a cardiovascular event in the subsequent five to ten years. The most used in the US is the Framingham risk score which is included in the ATP III panel recommendations (106). This score will give the 10 year risk of events of men and women up to age 79 years. A 65-year-old man is expected to survive an additional 17.2 years and a 65-year-old woman is expected to survive an additional 20 years (5). Therefore, most "young elderly" could be treated according to the guidelines addressing younger subjects. For a subject over 79 years the ten-year survival risk chart is not applicable. The decision should not be automatic. The recommendation of the physician should be accompanied by a discussion with the patient or with the decision-maker in which the informed consent process should be documented. In patients who have cognitive impairment or terminal illnesses or poor quality of life the option might be to withhold cardiovascular prevention therapy. Frail patients, who usually fall in this category and for whom no direct evidence of benefit of statin therapy exists, should be informed of their options. In the majority of healthy elderly in this age group, preservation of the function appears to be the main desired purpose of preventive therapy. Some patients or families are capable to verbalize their concerns and the physician should be prepared to address the role of lipid intervention in the prevention of the specific risk. Does the intervention decrease the risk of stroke, of hospitalization, of nursing home admission, etc?  The discussion is similar to the informed consent for an invasive procedure or clinical study. The physician makes a brief recommendation and outlines his opinion about risks and benefits. The patient or his decision-maker will question the decision or accept it. 
Therefore, patients are classified in three groups:

• Group 1
• Patients with age less than 80 years of age for whom evidence of benefit is available and who have good quality of life
• Patients in whom the decision has already been made by a recent revascularization procedure performed or planned
• Patients over 80 years of age who express interest in lipid intervention

In these patients lipoprotein testing should be ordered and intervention initiated.

• Group 2
• Frail patients
• Patients over 80 years
• Patients with diminished longevity through comorbidity
• Patients with ESRD (but not patients with chronic kidney disease or kidney transplant)
• Patients with extensive vascular calcifications
• Patients with non-ischemic cardiomyopathy
• Patients with cognitive impairment
• Patients with chronic liver disease

In these patients an informed process is necessary. The patient or its representative should be informed either of the lack of evidence for lipid intervention or that lipid intervention is not expected to modify the course of the main comorbidity. In patients with mild to moderate chronic liver disease statins are not contraindicated but the baseline elevation o transaminases might prevent the detection of statin related toxicity. In these patients it is acceptable to use intermittent treatment with statins with discontinuation if transaminases are increased three times or more over the upper limit of normal range. Patients with end-stage liver disease should not be prescribed statins since the risk exceeds the benefit.

• Group 3
• Patients with terminal illness
• Patients with irreversibly poor quality of live

In these patients, most prescribers would not consider lipid intervention and should notify the caregiver of this recommendation.

Choice of atherogenic particle targeted

Majority of physicians in the country practicing lipid lowering therapy are aware that their main target is LDL cholesterol (522). In elderly, majority of patients will be targeted to LDL cholesterol less than 100 mg/dL. This is the target recommended for patients with prevalent atherosclerosis, diabetes or metabolic syndrome, all disorders with high prevalence in the older age group. In patients with documented atherosclerosis and diabetes or uncontrolled risk factors, the goal of treatment is LDL cholesterol less than 70 mg/dL. The main problem is that, in this age group, LDL should be a target for therapy in a smaller percentage of patients than in the younger age group.

Non-HDL cholesterol as target for lipid intervention

The ATP III Panel guidelines are specifying that Non-HDL cholesterol should be targeted if triglycerides are higher than 200 mg/dL. This is usually forgotten by the practicing physicians (523). Obviously VLDL cholesterol contains atherogenic particles and omitting it from the evaluation of the risk will be an error of the magnitude proportional with the concentration of triglycerides. In addition, ignoring the role of VLDL as a cardiovascular risk factor will be a more significant error in populations in whom triglycerides are a risk factor independent from LDL or total cholesterol. In elderly subjects the presence of hypertriglyceridemia increased the prevalence of cardiovascular disease by 30% (524), was associated with increased arterial stiffness (525) and with large artery stroke (526).
Multiple studies have documented the superiority of Non-HDL cholesterol over LDL cholesterol as a predictor of risk. The arguments are:

• In some studies Non-HDL cholesterol predicted cardiovascular risk while LDL cholesterol did not. This was reported in patients with multivessel coronary artery disease undergoing revascularization (527), in patients with diabetes with or without coronary artery disease (528) or patients with diabetes free of coronary artery disease at baseline (529), or as a predictor of first myocardial infarction in healthy middle aged subjects (530). In adolescents enrolled in NHANES, Non-HDL but not LDL cholesterol predicted clustering of risk factors identifying better the subjects targeted for intervention (531).
• Non-HDL cholesterol but not LDL cholesterol emerged as independent predictor in multivariate analyses including both factors for future coronary risk (532) or for the burden of atherosclerosis expressed as coronary calcium score (533).
• The risk ratio per unit change is higher for Non-HDL cholesterol than for LDL cholesterol. This was shown in the Women Health Study (534) and particularly in patients with diabetes in this study (535), in the Health Professionals Follow Up Study (536), in the Lipid Research Clinics Prevalence Program (537), in the 4S study in the placebo arm (538) and in native Americans enrolled in the STRONG Study (539).
• Using discriminant analysis, Non HDL cholesterol was shown to be a better predictor than LDL cholesterol for myocardial infarction in subjects under age 36 years (540). These unusual patients were mainly smokers with high triglycerides.
• Remnant-like particle cholesterol is a risk factor for coronary artery disease (541) and stroke (542). They are elevated in patients with diabetes (543). Achieving Non-HDL targets in patients with diabetes treated with atorvastatin in different doses, discriminated between patients achieving a reduction in remnant-like particle cholesterol while LDL cholesterol target did not show a difference (544).  

LDL cholesterol is usually calculated by the Friedewald formula in which VLDL cholesterol is estimated as triglycerides / 5. The threshold for initiation of lipid intervention as well as the target for Non-HDL cholesterol is 30 mg / dL higher than the threshold and target for LDL cholesterol. Therefore, for practical purposes, there is little to gain by choosing a Non-HDL cholesterol threshold or target over a LDL cholesterol one in patients with triglycerides in the normal range.
Recently, studies have documented the ability of nonfasting lipid levels to predict cardiovascular outcomes (572-574). Non-HDL cholesterol levels, as opposed to LDL cholesterol levels are minimally modified by normal food intake in general population (545) or specifically in patients in whom Non-HDL has been shown to be a better predictor of risk such as patients with diabetes (546) or patients with ESRD (547). In some patient groups including elderly patients with slight cognitive impairment, there is a diminished certainty that a specimen was obtained in fasting state. This adds to the value of choosing the Non-HDL cholesterol as guide to therapy.

Apolipoprotein B as target for lipid intervention

All atherogenic particles contain a single molecule of Apo B but variable amounts of cholesterol. Apo B determination is a surrogate for determination of particle number. Cholesterol content of a particle is not positively associated with cardiovascular risk. Substituting Non-HDL cholesterol for LDL cholesterol as a target in patients with hypertriglyceridemia, accounts for the additional atherogenicity of VLDL particles. It does not, however, account for the differences in atherogenic particle composition, resulting in wide variations in the number of atherogenic particles for the same value of LDL cholesterol.
Particle number can be determined by nuclear magnetic resonance technology. Its relevance as a predictor of risk has been confirmed for incident coronary artery disease in middle aged (548) and older (549) women, new coronary events in male veterans treated with gemfibrozil (550), progression of coronary atherosclerosis (551), carotid intima-media thickness in general population (552) and in type 1 diabetes (553), as well as incident diabetic nephropathy in men with type 1 diabetes (554). The determination is relatively expensive and has not been recommended for routine use by any guideline. Apo B assay is available commercially and relatively inexpensive. Sniderman (555) and then Walldius (556). proposed Apo B determination for improving the prediction of coronary events. Two large studies have used this concept. In AMORIS the ability of Apo B and Apo A1 to predict cardiovascular events was documented in 149,121 subjects followed for 17 years (557). Interheart studied 15,152 cases of myocardial infarction from 52 countries and 14,820 matched controls (558). Only apolipoproteins B and A1 were directly determined and fractions of cholesterol were calculated from their values.
Apo B is superior to LDL cholesterol as a predictor of risk in the same situations as Non-HDL cholesterol is. In patients with hypertriglyceridemia, Apo B will document the number of atherogenic particles in the VLDL fraction as well as accounting for the prevalence of small dense LDL. Supporters of the use of Apo B argue the superiority of Apo B over Non-HDL cholesterol as follows:

• The overlap between quintiles of Non-HDL cholesterol and Apo B is only 40-75% and patients with above average Apo B and average Non-HDLc are more likely to have features of metabolic syndrome (559).
• The risk ratio for coronary events in the upper quartile or quintile over the lowest one is higher for Apo B or for LDL particle number than for Non-HDL cholesterol (577, 589, 590)..
• In some studies, Apo B is associated with subclinical atherosclerosis while Non-HDL cholesterol is not as strongly (560) or not at all (561) associated.
• In some studies, Apo B is associated with endothelial function while Non-HDL cholesterol is not Lind (562).
• In multivariate selection the program picks up Apo B as a significant independent predictor (563, 594).
• In patients treated with statin therapy Apo B is the best predictor of events (595-598).
• Apo B determination in patients with metabolic syndrome or diabetes picks up more patients at risk and targeting LDL or Non-HDL cholesterol will miss them (599, 600)
• Apo B determination in general population predicts the development of metabolic syndrome in both gender and of hypertension or increased CRP in women (601, 602).
• Apo B determination in patients with diabetes is better associated with abdominal adiposity (563).
• Apo B is an independent predictor of recurrent events in patients with metabolic syndrome (564).
• In patients with diabetes and coronary artery disease the ratio Apo B  / Non-HDL is higher when coronary artery disease is present (605, 606).
• Apo B is a predictor of coronary death while Non-HDL cholesterol is not in subjects over age 70 (148) or 75 years (565).
• I the Women Health Study Apo B was unchanged with estrogen treatment predicting no estrogen benefit while LDL was decreased and HDL was increased, predicting benefit (566).
• The best argument was presented by Stein et al (567). The investigators assessed in over 22,000 samples of blood available at a central research laboratory, if in presence of triglycerides 200-500 mg/dL the goal for LDL cholesterol<100 mg/dL, Non-HDL cholesterol<130 mg/dL or Apo B<90 mg/dL was achieved:
• If LDL cholesterol was at goal, then 84% of the patients were at goal for Non-HDL cholesterol and only 30% at goal for apo B.
• If Non-HDL cholesterol was at goal, then 97% of the patients were at goal for Non-HDL cholesterol and only 34% at goal for apo B.
• If apo B was at goal, then 99% of the patients were at goal for LDL cholesterol and 98% at goal for Non-HDL cholesterol.

Arguments against the use of Apo B are the following:

• Most statin clinical trials have focused on cholesterol but not on Apo B so the clinician will have will have to measure both. This will increase the cost.
• The cost of changing to Apo B technology will also be increased since more subjects will be treated and it will require more aggressive therapy to reach goals.
• Apo B or LDL particle number determination is not superior to cholesterol fractions determinations in all studies (561, 610-613). 
• Changes the targets for management of hyperlipidemia will create confusion among the US physicians (568).
• Area ROC is not improved by adding the lipoproteins to traditional risk factors (569).
• The relationship between Apo B and Non-HDL cholesterol is changing when the patient is treated with statins which complicates the interpretation of Apo B values in terms of traditional lipid intervention (570).

Guidelines and consensus panels for choice of target lipid intervention

In the US, the most influential guidelines are the ATP III Panel � NCEP Guidelines (106). They recommend LDL cholesterol targeting in most patients and Non-HDL cholesterol targeting in patients with triglycerides higher than 200 mg/dL. The targets for Apo B have been established by this organization but the concerns expressed above have prevented its introduction in the recommendations to clinicians. View of the enormous influence this organization has on the education of clinicians in the US, cholesterol management has improved (616-618).  The implementation of the guidelines is, however, still insufficient (619, 620) and interventions were devised to assist with this problem (621-623).
European guidelines on cardiovascular disease prevention are similar to the ATP III panel recommendations (571) but there are variations in guidelines between different European countries (572). These guidelines are recommending goals based on LDL and total, rather than Non-HDL cholesterol. Canadian Cardiovascular Position Statement recommend LDL cholesterol targets similar to those of ATP III panel but recommend Total Cholesterol / HDL cholesterol ratio as second target (573). A consensus panel of the Ministry for Health, Labor and Welfare of Japan has recommended adding Non-HDL cholesterol as a second target rather than waiting to obtain evidence that hypertriglyceridemia is a risk factor in Japanese population (574).
No guideline has specifically recommended targets in term of Apo B. An international consensus panel strongly recommended the use of Apo B and established a goal of 80 mg/dL for very high-risk patients (575). A consensus panel of the American College of Cardiology and American Diabetes Association recommended for patients with cardio-metabolic risk on statin therapy, guiding therapy with measurements of Apo B and treatment to Apo B goals in addition to LDL cholesterol and non-HDL cholesterol assessments (576). The recent American Diabetes Association are acknowledging the conclusion of this panel but are stopping short of recommending its conclusions (577). A panel of the American Association of Clinical Chemists concluded that it appears prudent to consider using Apo B along with LDL-C to assess LDL-related risk for an interim period until the superiority of Apo B is generally recognized (578).

Choice of statin

View of the almost universal success of statin clinical endpoint trials, therapy targeting atherogenic particles is initiated with a statin. Choice of statin depends mostly on the percent difference between the current level of the particle targeted and the goal proposed. Most guidelines recommend that statin therapy produce at least 30% decrease in LDL cholesterol. The target chosen for each patient represents "maximum allowed", therefore therapy with at least 30% decrease in cholesterol should be initiated, even if the target will be exceeded by a large margin. Statin intervention can be divided in two groups: patients requiring LDL cholesterol reduction 30-45% and patients requiring reduction over 45%. In patients with metabolic syndrome, for LDL cholesterol reductions of 30, 40 and 45%, equivalent reductions of Non-HDL cholesterol would be 28, 37 and 42% and for Apo B would be 24, 32 and 36% (579).
LDL cholesterol reductions up to 40 % can be achieved with generic statins in most patients. Reductions of 40 to 45 % occur rarely with pravastatin and fluvastatin and in some patients tolerating and responding well to simvastatin 80 mg. In elderly patients generics should be preferred on cost containment reasons. Pravastatin and fluvastatin will require maximum dosage to achieve LDL cholesterol levels in the desired range. Simvastatin is usually the preferred drug but the prescriber should review the risk of potential drug interaction. Maximum dose of simvastatin, 80 mg , should be prescribed with caution view of the warning issued for FDA concerning increased risk of rhabdomyolysis (580). The main drug interactions for simvastatin are itraconazole, ketoconazole, erythromycin, clarithromycin, telithromycin, HIV protease inhibitors, nefazodone precluding its use, gemfibrozil, cyclosporine and danazol limiting the dose to 20 mg, amiodarone and verapamil limiting the dose to 20 mg and diltiazem limiting the dose to 40 mg.
If the target is not reached with these statins or if LDL cholesterol decreases over 45% are required, high-end statins, atorvastatin or rosuvastatin, should be initiated.  With these statins in maximum dose, LDL cholesterol decreases up to 55% are expected in most patients. Maximum dose of rosuvastatin is 40 mg and atorvastatin 80 mg. In this dose rosuvastatin decreases more LDL cholesterol than atorvastatin while atorvastatin decreases triglycerides more than rosuvastatin. These differences are statistically but not clinically significant. The dose of atorvastatin does not have to be reduced in patients with impaired renal function so this drug is usually preferred in these patients. Rosuvastatin starting dose should be 5 mg in Asians or Filipino because of concerns over toxicity in these populations.

Other targets for lipid intervention

Lipid intervention is documented to be beneficial with two additional targets: HDL cholesterol and C Reactive Protein.

HDL cholesterol as target for lipid intervention

Trying to increase HDL cholesterol is usually considered upon achieving the goals for intervention on atherogenic particles. The drug of choice is niacin. Niacin has some clinical trials to support its use but large clinical endpoint trials are still ongoing. The main problem with initiation of niacin is usually the risk of discontinuation because of side effects. Patient needs to be explained that the expected benefit of the drug is higher than the transient discomfort associated with its use. The starting dose is 500 mg of long acting niacin. The dose is increase progressively at one to two week intervals. Adequately coached patients are able to tolerate the drug and achieve the desired dose. In sensitive patients such as postmenopausal women it is recommended to use low dose aspirin therapy 30 to 45 minutes before the administration of the dose. Increase in serum glucose is the second feared side effect and in patients with diabetes the possibility of a step-up in management of glycemia should be discussed. This step-up will occur in at least 30% of the patients (581).   

C Reactive Protein as target for lipid intervention

Statin intervention trials in high risk patients have documented the importance of bringing CRP to goal using statin therapy. In PROVE-IT patients with acute coronary syndrome who did not achieve LDL cholesterol <70 mg/dL or CRP<2 mg/L had an event rate of 4.6% per year. This was decreased to 3.1 or 3.2% if one of the targets was achieved and was 2.4% per year if both were achieved. In patients achieving a CRP of <1 mg/L the event rate was 1.9% per year an almost 60% reduction (176). Similar conclusions were reached by the investigators of A to Z trial (582). ACS survivors enrolled in this trial had an end of trial mortality of 1.65 if CRP reached after days was less than 1 mg/L, 3.7% if CRP was between 1 and 3 mg/L and 6.1% if CRP was over 3 mg/L.. JUPITER has also documented the role of CRP as a target rather than a threshold for initiating lipid intervention (215). In patients treated with rosuvastatin having a LDL cholesterol over 70 mg/dL and a CRP over 2 mg/L there was no benefit. In patients having one of these two goals reached the benefit was 47 to 58%. In patients who reached goal for both targets the benefit was 65% reduction in primary endpoint. If LDL cholesterol was less than 70 mg/dL and CRP was less than 1 mg/L the benefit was 79%. Targeting CRP does not imply that CRP contributes to the process of atherosclerosis but rather that achieving lower targets for this indicator of risk provides better protection against atherosclerotic events.

Any role for triglyceride targeting?

Triglycerides are a risk factor in certain populations but it is not clear if they are an independent risk factor or part of the metabolic syndrome. There is insufficient evidence that specific targeting of triglycerides in patients reaching the primary goals of lipid intervention is beneficial. Since in most of patients treatment is initiated with statins the main question is of adding a fibrate to statin therapy. In patients with low HDL cholesterol the second drug to be added is niacin. Patients with normal HDL cholesterol and high triglycerides are uncommon. In patients with low HDL cholesterol and high triglycerides treated with statins but intolerant to niacin fenofibrate could be added to statin therapy. Gemfibrozil, the only available fibrate with positive clinical outcomes should not be used in patients treated with statins. The combination statin-fibrate has been extensively studied for fenofibrate cholinate (TrilipixR) and has been approved by FDA. Fibrate therapy has been shown to be clinically beneficial in patients with high triglycerides and low HDL cholesterol in post-hoc analyses of BIP (636, 637) and ACCORD (423) for bezafibrate or fenofibrate, respectively.

Choice of additional or alternative drugs

 

Targeting atherogenic particles

Some patients will not tolerate any statin in any dose. In others, the goal cannot be reached with maximum dose statin therapy. For these, treatment is usually initiated with bile acid binding resins or with ezetimibe with the goal of eventually combined these drugs. The combination will produce LDL cholesterol lowering in the range of 30-35% (583). The combination is relatively expensive since elderly tolerate poorly cholestyramine or colestipol and the prescriber has to recommend non-generic colesevelam. Addition of niacin to any of these two drugs or both is feasible but its effect on atherogenic particle number is minimal. View of the difficulties of achieving adequate cardiovascular prevention in absence of statin therapy, prescriber should not give up on statin therapy when the patient is complaining of side effects, usually myalgia. Patient should be made aware repeatedly of the benefits of statin therapy and of the fact that myalgia with statins is a common complaint but myopathy is rare and rhabdomyolysis is not preceded by warning myalgias. In some patients alternative day statin therapy using atorvastatin or rosuvastatin diminishes myalgia. Alternative day therapy with these statins has been shown to have acceptable effectiveness view of their long half-life (639, 640).  Some authors gave proposed the use of red yeast rice (monascus purpureus) as an alternative to statin therapy. Reductions in LDL cholesterol of 20-30% were reported (641-643). The problem with this approach is that the active ingredients in the yeast are monacolins, which are 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors similar to statins (584).

Is gemfibrozil monotherapy an alternative to statin therapy?

In patients who cannot tolerate statins gemfibrozil monotherapy is a viable alternative. The approach has been known to work in high-risk elderly men (429) or in healthy middle age men with high Non-HDL cholesterol (427). No information is available concerning the benefit of this approach in women. The benefit of gemfibrozil is not depending on achieving targets in terms of atherogenic particles. There are data showing that on-study HDL cholesterol (585) or the number of small LDL and HDL particles predict the results, but Apo B was not a predictor (550). HDL cholesterol monitoring and use of niacin would be logical add-ons to gemfibrozil therapy. There is no information concerning adding niacin to gemfibrozil for cardiovascular prevention. In small trial, 143 military retirees with coronary artery disease were randomized to either placebo or gemfibrozil, niacin and cholestyramine. A composite cardiovascular event end point was reached in 26% of patients in the placebo group and 13% of those in the drug group and the difference was statistically significant (586).

INITIATION OF LIPID PREVENTION