This Article Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Kreisberg, R. A. Search for Related Content PubMed Articles by Kreisberg, R. A.

—Random Thoughts and Opinions

University of South Alabama Mobile, Alabama 36688

Address correspondence to: Robert A. Kreisberg, M.D., Dean and Vice President for Health Affairs, University of South Alabama, CSAB 170, Mobile, Alabama 36688-0002.

	Introduction

Top Introduction

 
IT IS CLEAR that secondary prevention of coronary heart disease (CHD) events is cost-effective. The emphasis now should be on making sure that all eligible patients are identified and treated; something that we have not done very well, so far. Although recurrent myocardial infarction (MI) has been reduced by 2–3% and CHD death by 4–5% annually, the incidence of first MI has not changed, representing a failure in primary prevention. Half of first MIs occur in patients without previous symptoms, and the sad fact is that 25–50% are fatal. Secondary prevention only works for those who survive their first event. Waiting until patients have symptoms of CHD is waiting too long for some. Primary prevention of CHD has obvious implications, particularly if patients at intermediate and high risk for cardiac events can be identified and treated. However, the prevention of clinical events is not the same as prevention of atherosclerosis. Prevention of atherosclerosis requires early and extensive lifestyle changes to prevent or retard the development of the earliest lesions, fatty streaks, and their progression to unstable lesions. As good as lipid-lowering therapy is, it is still imperfect because all intervention studies show a substantial residual risk of 65–70%. Consequently, prevention of disease is much better than prevention of events. Some of the residual risks in these patients reflect the presence of coexistent, but suboptimally treated, traditional CHD risk factors such as hypertension, smoking, and diabetes mellitus that were present, to varying degree, in patients participating in these studies. Some of the residual risks may also be due other recently identified but less well characterized and understood risk factors that also predispose to clinical events. These "emerging" or "conditional" risk factors include, but are not limited to, triglyceride-rich lipoproteins (TRLs), Lp(a), small dense low-density lipoprotein (LDL) particles, homocysteine, oxidative milieu, a procoagulant state, insulin resistance, and C-reactive protein (CRP). Although there is considerable basic information and epidemiologic data that support an important role for some of these factors in atherosclerosis, there is, as of yet, no clear evidence that specific therapy influences CHD risk. The use of antioxidant vitamins or folate, for example, makes perfect sense but may subsequently be incorrect. It is also clear that markers of inflammation, such as CRP, identify patients who have advanced unstable atherosclerotic lesions that are vulnerable to rupture. CRP may be an excellent test for targeting high-risk patients for primary or secondary prevention who would benefit from aggressive therapy.

In patients with established CHD cholesterol-lowering therapy prevents CHD events by presumably modifying the atherosclerotic plaque and making it less susceptible to rupture. This is due in large part to a reduction in LDL-cholesterol (LDL-C), reduced entry of LDL particles into the artery wall, a subsequent reduction in oxidized LDL, reduction or reversal of the ongoing inflammatory and immunologic responses in the vessel, and improvement in endothelial function and in myocardial perfusion. The relationship of the nonlipid properties of statins (on endothelial function, hemostasis, and the cellular, immune, and inflammatory responses of the vessel) to the atherosclerotic process are of great interest, but it is too soon to know whether these or other properties are important. It is becoming increasingly clear that drugs of the same class may not be completely identical. This may be true of the statins where LDL-C lowering may be the major, but not exclusive, mechanism by which CHD risk is reduced. Statins may have antiatherogenic properties independent of their lipid effects. Rigorous head-to-head studies are necessary to learn more about potential mechanisms of action and whether differences in statins are clinically important.

The issue of whether "lower is better" for LDL-C has been contentious. No intervention study has been designed to answer this question, and most of the statements about this issue come from post-hoc analyses of patients at progressively lower LDL-C levels. The number of patients in these subgroups is a small fraction of the number needed to have meaningful statistics. Indirect evidence from other studies support that "lower is better," and studies are underway to answer this question. In the postcoronary artery bypass graft trial an LDL-C of less than 100mg/dL was better than one of 135 mg/dL. In the Atorvastatin versus Revascularization Trial (AVERT), an LDL-C of 75 mg/dL was better than one of 119 mg/dL plus angioplasty. AVERT is an important study because 70% of patients randomized to angioplasty also received statins and had a mean LDL-C of 119 mg/dL, which is quite respectable and in the range associated with maximum benefit in the Cholesterol and Recurrent Events (CARE) study. Yet, those with much lower LDL-C did better. These data suggest that the major effect of statins is mediated by changes in LDL-C. Post-hoc analyses from the CARE study suggested that there was no additional value of LDL-C levels less than 125 mg/dL, whereas in the 4S post-hoc analyses lower was better. It is clear that the relationship of LDL-C to CHD events is curvilinear and that lower is better, but the reduction in CHD risk from further reduction of LDL-C becomes progressively smaller. The decision to aggressively reduce the LDL-C to less than 100 mg/dL depends on numerous factors, which include patient adherence, additional cost, the complexity of the regimen, and the likelihood that the target can be achieved.

True primary prevention means prevention of atherosclerosis. Because the lesions that ultimately cause CHD events later in life begin in childhood and can be well expressed in young adults, early institution of healthy lifestyle practices is required. With the exception of persons with lipid/lipoprotein abnormalities that predispose to early atherogenesis, such as familial hypercholesterolemia or familial defective apolipoprotein B, most primary prevention will be prevention of clinical events in adults who have preclinical disease. In other words, therapy is now directed at prevention of events more than the prevention of disease. In this regard, much of primary prevention is actually secondary prevention because these patients often have extensive disease that has not yet become symptomatic. There is a continuum of risk among asymptomatic patients, some being at very low risk, others at intermediate risk, and some at very high risk. The presence of two or more CHD risk factors identifies the intermediate and high-risk patients. For example, a young menstruating woman with increased LDL-C but no other risk factors is at such low risk of developing CHD in the near future that pharmacotherapy is not cost-effective. In contrast, a middle-aged smoking man with hypertension, central obesity, impaired fasting glucose, and low high-density lipoprotein cholesterol (HDL-C) is a walking "time-bomb." Most patients fall between these two extremes. It is crucial to be able to identify patients without known heart disease who are at high risk for cardiovascular events in the near future. In the United States, the National Cholesterol Education Program guidelines or the Framingham equation can be used to predict risk. More recently, the American Heart Association has recommended the additional use of conditional (emerging) risk factors, the ankle-brachial systolic pressure ratio, exercise testing, and perhaps carotid ultrasound for further stratification of risk. The paradigm has changed from diagnosis to prognosis. The American Heart Association has also emphasized the adverse cumulative impact of multiple minor abnormalities in prediction of risk. There is general agreement that a 2–3% annual risk of clinical CHD events represents high risk.

The intervention studies with statins are impressive and reveal that therapy reduces CHD risk by 25–35%. Baseline risk varied from 5% [Air Force Coronary Artery Prevention Study/Texas Coronary Artery Prevention Study (AFCAPS/TexCAPS), a primary prevention study] to 27% [Scandinavian Simvastatin Survival Study (4S), a secondary prevention study], so that absolute risk reduction varied from 1.5–9%. These numbers reflect the responses of "average" patients within each study and are only applicable to our own patients who have similar characteristics. CHD risk factors are variably present in these studies, putting the patients at higher or lower risk of a CHD event than for the study as a whole. All subgroups of patients within each of the studies benefited from cholesterol lowering (hypertensives, smokers, diabetics, men and women, and the elderly). However, those at the highest risk had greater absolute benefit (e.g. diabetics in the CARE study had an absolute risk reduction of 8% vs. 5% for the group as a whole for combined end points). These data indicate that the higher the baseline risk, the greater the benefit during therapy even if relative risk reduction is similar. The number of patients needed to treat (the reciprocal of the absolute risk reduction) for 5 yr to prevent a single event in the statin studies varies from 12 to 50 depending on the absolute risk reduction. The treatment of 12 patients for 5 yr to prevent an event is considered cost-effective, whereas the treatment of 50 patients for 5 yr is probably not. Physicians as well as patients must understand that if 12 patients must be treated for 5 yr to prevent one event, then the outcome of 11 of the 12 patients will not be changed by therapy; in other words, they have a 1 in 12 chance of benefiting. In contrast, for AFCAPS/TexCAPS the chance of benefit is 1 in 50 and may not be very attractive. Thus, clinical judgement and patient preferences must be taken into consideration. A family history of CHD is crucial in making these decisions; not all of the family predisposition to CHD can be attributed to hypertension, diabetes, dyslipidemia, and so forth. It has been stated that Winston Churchill abused his body and died in his 90s, whereas Jim Fixx was a health nut and died in his 50s. Genetics may well influence the vascular response to known or emerging CHD risk factors. After all, we all have seen patients with lifelong multiple CHD risk factors and no clinical disease and other patients with devastating disease and no obvious risk factors.

Whereas LDL-C is the most important atherogenic lipoprotein, other lipoprotein abnormalities are important and must be considered as management is "fine-tuned." It has been difficult to determine the relationship of triglyceride to CHD because of greater biologic variability in the triglyceride level and larger coefficients of variation when triglyceride is measured. Triglyceride is not directly atherogenic but indicates the presense of TRLs, some of which are atherogenic. The heterogeneity of TRLs plus the other potentially atherogenic lipoprotein abnormalities that accompany hypertriglyceridemia (small dense LDL and reduced levels of HDL-C) make it difficult to "tease out" the precise relationship and mechanisms. In addition, this form of dyslipidemia is characteristic of insulin resistance and the atherogenic metabolic syndrome. The calculated LDL-C level can be misleading because it includes Lp(a) as well as the cholesterol transported in small VLDL, intermediate density lipoprotein, and VLDL remnants. Consequently, the reduction in LDL-C produced by statins is due, in part, to a reduction in the cholesterol transported by these lipoproteins. When triglyceride levels are normal the calculated LDL-C is accurate; but when elevated, it is not. This is important because combined abnormalities of cholesterol and triglyceride predominate in patients with CHD, particularly if triglyceride levels of more than 200 mg/dL are considered abnormal. As described, the non-HDL-C, provides a convenient way to assess apoprotein B-containing lipoproteins (LDL and VLDL) and have a single parameter to express CHD risk in patients with combined hyperlipidemia. Large VLDLs carry little cholesterol relative to triglyceride and will have a smaller influence on this value than small VLDLs or LDLs where a greater percentage of the molecule is composed of cholesterol. Also, the blood sample need not be fasting, making the measurement of triglyceride at that time unnecessary. The non-HDL-C includes all potentially atherogenic lipoproteins, and most hypolipidemic drugs affect multiple lipoproteins. This concept is appealing and simplifies decision making for patients with combined hyperlipidemia.

One of the most important risk factors for CHD is diabetes mellitus. Diabetes mellitus and impaired glucose tolerance are disproportionately represented among patients with CHD. Hyperglycemia is a late event in the natural history of the insulin resistance syndrome. Because insulin resistance exists from birth the vascular endothelium is exposed to a constellation of proatherogenic metabolic abnormalities for decades. This often leads to development of advanced macrovascular disease before the appearance of hyperglycemia and explains why there is a poor correlation between the duration of type 2 diabetes mellitus and the development of macrovascular disease. This has led to the concept that "the clock starts ticking" for atherosclerosis in type 2 diabetes long before the appearance of hyperglycemia. However, it is also clear that hyperglycemia worsens the biology of atherosclerosis. Acute, intermediate, and long-term morbidity and mortality following MI is worse in diabetic than in nondiabetic patients even when adjusted for the severity of CHD, the size of the infarct, the use of thrombolytic agents, or the interval between the time of the infarct and receiving medical care. Adaptation of the myocardium to the infarct is less efficient in patients with diabetes. If myocardial perfusion and adaptation following an infarct are influenced by the microvasculature, then glycemic control may be an important factor independent of its effect on the atherosclerotic lesion. Diabetics without clinical CHD are at the same risk of a CHD event as nondiabetics with CHD, implying that CHD prevention in asymptomatic patients with type 2 diabetes is secondary prevention. In the 4S, diabetics with CHD and elevated LDL-C were at approximately twice the risk of an event as nondiabetics. In the CARE study, diabetics with CHD and average American LDL-C levels had a 50% higher risk of CHD events than nondiabetics. Under virtually all conditions diabetics are at much higher risk than nondiabetics for CHD events and adverse outcomes. In the 4S and CARE study, diabetics had a greater reduction in absolute CHD risk than nondiabetics and, therefore, experienced greater cardioprotection. In the Long-Term Intervention with Pravastatin in Ischemic Disease (LIPID), the differences between diabetics and nondiabetics with LDL-C lowering was not significant. In the recent Veterans Administration High Density Lipoprotein Cholesterol Intervention Trial (VA-HIT) (which included 625 diabetics with HDL less than 40 mg/dL, LDL-C less than 140 mg/dL, and TG less than 300 mg/dL), gemfibrozil reduced the relative risk of CHD events by 22% in diabetics (from 36% to 28% compared with nondiabetics, from 23% to 18%). The absolute risk reductions were 8% and 5%, respectively. The number of diabetics that need to be treated for 5 yr to prevent an event is 12, comparable to what was observed in the 4S and CARE study. In contrast, glycemic control in the United Kingdom Prospective Diabetes Study (UKPDS) study was associated with only a 16% reduction in relative CHD risk over a period of 12 years, which was of marginal statistical significance. Perhaps better glycemic control than that achieved in the UKPDS would have been more effective in reducing CHD events; the mean Hgb A1_C in the intensively treated group was only 0.9% lower than in the conventional group, and glycemic control in the intensively treated group actually worsened during the study. It is possible that better glycemic control than was achieved is required to prevent CHD. These results suggest that aggressive management of hypercholesterolemia and dyslipidemia, as well as other standard CHD risk factors in patients with type 2 diabetes, is more likely to influence CHD events than treatment of hyperglycemia. However, achieving good glycemic control has other obvious benefits.

Medicine and CHD, in particular, have never been more exciting. Impressive progress has been made in understanding atherosclerosis and preventing (CHD) events in the past decade. More impressive discoveries can be expected in the next 5–10 yr that may dramatically change our approach to the prevention and treatment of CHD. Lipids are only one part of this complex disorder, and endocrinologists must understand more than just the diagnosis and treatment of hyperlipidemia/dyslipidemia if we are to play an important future role in the prevention of CHD.

Somehow, most of the truly exciting work in this field is being done by lipid specialists and cardiologists and not by endocrinologists. After all, molecular biology is one of the things we do best and is a field in which we have been a leader. Is this yet another area that we have abdicated?

	Footnotes

 
"Clinical Perspectives" are an occasional feature of The Journal of Clinical Endocrinology & Metabolism. They present the opposing views of invited contributors on a topic. All reprints must include the complete Clinical Perspective, so that each section can be read in context.

Accepted March 6, 2000.

This Article Full Text (PDF) Submit a related Letter to the Editor Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Kreisberg, R. A. Search for Related Content PubMed Articles by Kreisberg, R. A.