International Task Force for Prevention of
Coronary Heart Disease
CORONARY HEART DISEASE: REDUCING THE RISK
1.3 Risk factors as causes of CHD
Several criteria have been used as guidelines in judging whether an epidemiological association reflects a causal role for a particular risk factor:
Hypercholesterolaemia and hypertension fulfil all of these criteria, while cigarette smoking, obesity, diabetes mellitus and physical inactivity meet the first four. These factors influence risk powerfully, are common in populations, and are amenable to prevention or treatment. Most authorities today also regard low HDL-cholesterol or low levels of one of its subclasses as a risk marker. More recently, evidence has accumulated that elevated levels of circulating triglyceride, lipoprotein(a) (Lp(a)) and fibrinogen are associated with increased risk of coronary heart disease. Increased risk is also conferred by physical inactivity, obesity and diabetes mellitus.
Other risk factors under investigation include increased levels of coagulation factor VIIc, homocysteine and plasminogen activator inhibitor I. A further area of research into candidate risk factors concerns polymorphisms and abnormalities in several genes affecting the regulation of lipoprotein and glucose metabolism and blood pressure, including those for apolipoprotein E, angiotensinogen and the enzymes lipoprotein lipase, glucokinase and angiotensin converting enzyme. Inadequate protection of lipoproteins against oxidation is under consideration as a further risk factor. Much of this latter research is still in a preliminary stage with results which are in part inconsistent. Its future impact on clinical practice can therefore not be assessed at present
A relationship between nutrition and CHD has long been recognised based
on three main sources of information. These establish the centrally
important causal role of population-wide adverse eating patterns,
which act through their influence on several risk factors, notably
hyperlipidemia, hypertension, and overweight. The epidemiological evidence
suggests that inviromental factors strongly affect CHD risk. Two classical
studies may be cited. The Seven Countries Study of 14 communities by Ancel
Keys showed a very close positive correlation between CHD mortality rates
and consumption of saturated fat (both also directly related to plasma
cholesterol level). And in the Geographical Pathology of Atherosclerosis
Study, HC McGill found a close relation between the extent of atherosclerosis
in autopsy samples and the saturated fat consumption in each geographical
region studied. Within populations, the satureted fat-CHD relationship is
less evident, presumably because there is less dietary variance within
communities than between them; however, dietary cholesterol intake, and
an index reflecting the combined intakes of saturated fat, unsaturated
fat and cholesterol, have been found to be predictive of CHD mortality
in long-term follow-up.
Within populations, intake of trans fatty acids tends to be associated
with higher risk and intake of polyunsaturated fatty acids (mainly linoleic
acid) with lower risk of CHD.
Nutritional intervention studies have been of several types. In a number of experimental animal models it has been demonstrated that a diet high in cholesterol and saturated fat causes atherosclerosis which (in studies of relatively moderate diets) has been shown to be very similar to the human disease. When atherosclerosis has been induced, and such diets are then replaced by the habitual low-fat diet of the species, the arterial lesions regress partially but substantially. Clinical trials of diet have also been performed, though they are difficult to conduct due to problems of compliance, while blinding is not feasible. In one primary prevention trial using a diet low in saturated fat and moderately supplemented with polyunsaturated fat together with anti-smoking counselling, CHD incidence was reduced by 47% and total mortality by 30%. The smoking intervention was relatively unsuccessful, while plasma cholesterol was decreased consistently during the trial by 10%; it was inferred that the favourable outcome was chiefly attributable to dietary change 2. The available meta-analyses of dietary trials also point to a favourable outcome. Lastly, one trial of diet as a single inter-vention, which used quantitative coronary angiography as the end-point, found a highly significantly favourable effect of dietary intervention on the natural history of coronary athero-sclerosis and on the CHD event rate 3.
That nutritional factors should influence CHD risk is made plausible
by the existence of a number of mechanisms by which this relation
could be predicted. Most evident is the fact that the plasma LDL-cholesterol
level is increased by saturated fats (principally those containing the
widely present palmitic and myristic acids) and by trans-unsaturated
fatty acids (present in partially hydrogenated margarines and cooking
fats and in dairy fats). Trans-fatty acids also lower HDL-cholesterol
and increase lipoprotein (a) and may also increase CHD risk by these
means. Unsaturated fats lower plasma LDL-cholesterol. The effect of
polyunsaturated fatty acids is a little greater than that of monounsaturated
fatty acids, but both have significant favourable actions. A very small
effect of mono- and polyunsaturated fatty acids in lowering HDL-cholesterol
has also been found in some but not all trials and is unlikely to affect
CHD risk. However, the ratio of total to HDL-cholesterol improves markedly
as unsaturated replace saturated or trans-fatty acids.
Dietary cholesterol increases plasma cholesterol. However, there is
wide variation between individuals in the extent of this effect. This
individual responsiveness is consistent on repeated testing and is
greater in subjects with higher plasma cholesterol levels. A genetic
basis is therefore assumed. In hyper-responders dietary cholesterol
has a greater than usual effect in stimulating LDL production which
is accompanied by a lower activity of LDL receptors and a lower LDL
fractional catabolic rate. When given a lipid-lowering diet, women show
a slightly lesser degree of lowering of plasma- and LDL-cholesterol than
men.
Saturated fats may confer risk by mechanisms other than their effect on plasma lipoproteins. Platelet aggregability is enhanced by a wide range of saturated fats, and other coagulation variables, e.g. Factor VII coagulant, are increased. Long chain saturated fatty acids including stearic acid confer these untoward thrombogenic effects. It may prove to be of great importance that dietary fats modify the proneness of LDL to oxidation, which in turn profoundly alters its atherogenicity. Oleic acid renders plasma LDL less prone to oxidation. In a trial of the effect of diet on coronary atherosclerosis, consumption of saturated fatty acid was strongly related to disease progression as measured by quantitative coronary angiography. Also, the intakes of stearic acid and of trans-unsaturated fatty acids were correlated with progression of disease through mechanisms other than effects on plasma lipids. Polyunsaturated fatty acids may affect platelet function; high amounts of n-6 polyunsaturated fatty acids cause increased aggregability, while n-3 polyunsaturated fatty acids decrease aggregation. One of the effects of omega-3 fatty acids (mainly present in fatty fish) is to promote the vasodilation mediated by endothelium-derived nitric oxide. This mechanism is impaired in atherosclerotic coronary arteries and is restored by large doses of fish oil. Intake of alpha-linolenic acid is reported to be inversely related to the risk of CHD.
Many, though not all, epidemiological studies have shown an inverse relation between intake of fish and CHD mortality; in particular mortality is higher among men who do not eat fish than among those who do.
Intake of cereal fibre has been reported to be inversely related to the risk of CHD. Soluble fibre, such as occurs in fruit and vegetables and in some cereals, including oats, significantly lowers plasma cholesterol levels.
One of the most important, though often under-acknowledged determinants of CHD risk, is obesity, notably truncal obesity, and this in turn has powerful nutritional implications. The long term data from the Framingham survey have established an increase in CHD in overweight patients. Thus excess dietary energy (calorie surplus) may have a fundamental role as a potentially avoidable and/or treatable cause of CHD.
The importance of certain nutrients in blood pressure control is discussed in section 3.5.2 (Non-pharmacological treatment of hypertension).
Dietary intake of anti-oxidant substances has been extensively
studied: intependent epidemiological studies have shown an inverse
relationship between CHD risk and the intakes of alpha-tocopherol
and beta-carotene. Intervention studies have provided some evidence
for a beneficial effect of supplemental vitamin E, however they have
not unequivocally proved that anti-oxidants reduce the incidence of
CHD end-points; protection of LDL from oxidative change is established,
and there is an impressive body of evidence from animal studies. Many
foods of vegetable origin contribute anti-oxi-dants to the diet,
including native olive oil, tomatoes, carrots and other vegetables,
whole wheat, onions and tea.
Elevated levels of homocysteine are suspected to be a CHD risk
factor, and vitamin intake has been shown to influence plasma
levels of this amino-acid. Higher homocysteine levels are seen in
subjects with low intakes of folic acid, vitamin B12 and pyridoxal
phosphate. In most studies the association with folate is the
strongest. Dietary supplementation with these vitamins, or with
folate alone, lowers plasma homocysteine effectively.
These nutritional effects on lipid and other CHD risk factors are not only of central importance in the aetiology of cardiovascular disease; they are also highly significant in risk factor manage-ment. But the potential benefits of improved nutrition can only be realised when patients at risk and populations with high CHD prevalence can be induced to adopt healthier eating habits. Failure to recognise this issue has led some to doubt the role of diet as a means of lowering lipid-mediated risk. Where efforts have been made to maximise dietary compliance - as numerous studies some of 3 years or longer in duration have reaffirmed - the effectiveness of lipid-lowering diets both in free-living persons and under controlled conditions has been established. This places a responsibility on health professionals to acquire skills in nutritional counselling. Necessary to such skills are enthusiasm, a working knowledge of the nutrient content of common foods, and an ability to adapt the principles of lipid-lowering and calorie-restricted diets to the food preferences and lifestyle of each individual. Initial success in lipid lowering is a powerful source of motivation to the patient. Effective use of dietary treatment lessens dependence on drug therapy and offers economic advantages.
Nutrition plays a significant role in blood pressure regulation, and the following are of therapeutic importance:
1.3.3 Contribution of socioeconomic and psychosocial factors to CHD risk
A growing literature has documented associations between several psychosocial factors and increased incidence of CHD. The most prominent among these are stress, lack of social support, depression, and socioeconomic status. All of these have been associated either with clinical outcomes in CHD patients or poor prognosis through mechanisms of action that involve neuroendocrine or sympathetic nervous system pathways, as well as possible failure to adopt such aspects of a healthy lifestyles as sticking to a recommended diet, taking exercise, losing weight and stopping. In many cases, these psychosocial factors have also been shown to have adverse effects on adherence to medical care, thereby decreasing its effectiveness.
Few intervention studies have been conducted, and some of these have had low statistical power or other methodological flaws. The content of various treatments for diagnosed psychosocial conditions shows some overlap. For example, group therapy for depressive symptoms or stress management also provides a socially supportive environment, making it difficult to attribute observed improvement to a specific psychosocial factor. Nonetheless, considered together, the studies support the concept of providing treatment for psychosocial factors. At a minimum, such treatments result in better adherence to medical care and improved quality of life.
Evidence for some of the psychosocial influences may be sufficiently good to justify testing whether they should be included in algorithms for computing cardiovascular risk in the context of primary or secondary incidence of CHD events.
In most countries lower socioeconomic status is associated with higher total and CHD mortality rates, a reversal of the position in early post-war years. The reasons for this are not fully understood but it is clear that many aspects of health related behaviour are less favorable among the less privileged, partly because of a low standard of education. For example, the prevalence of smoking is higher and the frequency of quitting is lower among those of lower socioeconomic status, and the prevalence of obesity is higher. However, the gradient is not restricted to the extremes of the socioeconomic status distribution at which disparities in education would be expected to have substantial effects on health knowledge. Health status also is more favorable for those at the highest levels of socioeconomic well-being compared to those in adjacent strata, in spite of the fact that they can be considered, by any standard, to be among the well-educated and economically advantaged. Furthermore, in most studies, the relationship between socioeconomic status and CHD risk remains significant after adjusting for traditionally accepted risk factors, recognizing that health behaviours such as diet and physical activity are not measured precisely. No studies have been conducted to examine the effects of socioeconomic mobility per se on health.
The molecular and physiological processes which govern an organism's responses to stress are complex, and are mediated by both neuroendocrine and neural pathways some of which overlap extensively with systems involved in cardiovascular regulation. Good evidence exists that acute stress acts as a trigger of cardiac events such as myocardial ischemia and arrhythmias.
Although stress has been shown in numerous studies to influence cardiovascular function, a clear consensus of the relationship between chronic stress and the development of cardiovascular diseases has not emerged. To a large extent, this is because stress has been difficult to quantify. In contrast to biological attributes such as serum cholesterol levels which can be measured under standardized conditions, such as fasting, standard methods of preparing individuals for measurement of basal stress levels have not been developed. Additionally, it has been difficult to measure the stress burden imposed by various life events, such as family illness, job loss, or chronically adverse social conditions in a way that can be used to quantify their contribution as a risk factor for CHD. Furthermore, the effects of stressors on individuals are affected by endogenous variables, such as genetic factors, which influence the response of the central nervous system, and exogenous variables such as the availability of social support and previous experience.
Nonetheless, good evidence exists from research in animals and humans that stress promotes the progression of CHD. For example, diet-induced atherosclerosis is accelerated in cynomolgus macaques who are exposed to the chronic stress of competition brought about by periodic disruption of hierarchical relationships. In humans, the most consistent evidence of a relationship between stress and CHD shows that CHD progresses faster in individuals who have a large cardiovascular response under mental stress testing. Myocardial ischemia provoked during mental stress testing has being found to be prognostic of future cardiac events.
However, our ability to draw conclusions from observational studies is limited, and only one small, controlled clinical trial of 107 patients has been conducted to assess the influence of a stress management intervention for CHD patients specifically on the occurrence of cardiac events, defined as revascularization and reinfarction. In that study, only 9% of the patients in the stress management group experienced an event during 38 months of follow-up, compared with 21% of those enrolled in an exercise rehabilitation programme 3b. Stress management has also been studied as adjunctive therapy for hypertension, with mixed results, and was a prominent component in the Lifestyle Heart Trial to assess a multifactor intervention for improving myocardial ischemia (see footnote 25.). In general, stress management interventions seem sufficiently promising to warrant further evaluation with sufficient power for disease endpoints, including morbidity and mortality.
Depression is common among persons with CHD, with some studies reporting major depression in as many as 20% of myocardial infarction (MI) patients. Depression following an MI is associated with increased risk of reinfarction and mortality in several studies, independent of disease severity. In one study, major depression was a significant predictor of cardiac mortality in post-MI patients after controlling for left ventricular dysfunction and previous MI (adjusted hazard ratio, 4.29) 3c.
Recently, depression has been reported to be a risk factor for CHD morbidity and mortality in healthy individuals or those with risk factors, but no evidence of existing disease. Several large, prospective cohort studies have found depression and depressive symptoms to be associated with fatal and non-fatal MI in persons free of diagnosed disease. Participants enrolled in the Systolic Hypertension in the Elderly Program (SHEP) 3d. who also had an increase in depressive symptoms during the 4.5 year follow-up period were more likely to die and to experience stroke or MI than those who did not have an increase in depressive symptoms. These studies show that depression may be a predictor of initial, as well as recurrent cardiac events. Due to increasing recognition that depression may influence cardiac outcomes, more CHD patients with depressive symptoms are being treated.
While treatment of depression improves patients' ability to function and quality of life, it is not known whether such treatment will improve cardiac outcomes. Whether depression is merely a marker for or a consequence of the symptoms of the underlying, perhaps even subclinical disease, or a risk factor that influences disease progression, is not understood. A large randomized, controlled clinical trial, the Enhancing Recovery in Coronary Heart Disease (ENRICHD) 3e. Trial conducted in the United States by the National Heart, Lung, and Blood Institute, is underway to determine whether intervention for depression and low social support affects risk of subsequent events and death in MI patients. ENRICHD will randomize 3,000 eligible post-MI patients to either therapy or usual care, with an average follow up of three years for morbidity and mortality.
A growing number of observational studies have linked an absence of social support with all-cause and cardiovascular disease mortality. The Alameda County study was one of the first to demonstrate that a lock of social integration, or participation in intimate social relationships, predicts mortality risk 3f. These findings have been replicated in at least eight population-based prospective studies, showing that individuals who are socially isolated or lack social networks are at greatest risk for death from all causes as well as from heart disease.
The presence of social support is associated with decreased risk of future nonfatal events as well as cardiac mortality in persons with existing heart disease. Men who are socially isolated and report high levels of stress experience a greater than four-fold increase in post-MI mortality. For patients randomized to placebo in the Cardiac Arrhythmia Suppression Trial-1 (CAST-1) 3g., the level of perceived social support was a significant predictor of mortality after adjusting for disease severity.
In two recent studies of CHD, patients who met criteria for social isolation had a two- to three-fold increased risk of death compared to patients not meeting the social isolation criterion. This difference in survival was independent of all underlying cardiac anatomic and functional factors documented at baseline. Finally, while few studies have examined the effects of social support available prior to the onset of MI, a longitudinal study of elderly persons showed that presence of emotional support was the best and most consistent predictor of survival after MI, with lack of emotional support related to early in-hospital as well as out-of-hospital mortality over a six-month period. After 12 months, 55% of those with no sources of emotional support had died in contrast to only 27% of those with 2 or more sources of support, adjusted for disease severity, age, and sex.
Although the epidemiological evidence linking lack of social support to all-cause and cardiovascular death is strong, few studies have investigated whether providing social support to healthy individuals or CHD patients decreases risk of future disease. In two studies which provided supportive group therapy to post- MI patients, cardiac morbidity and mortality were reduced or showed a favourable trend. Several other studies that involved interventions for MI patients, including the Recurrent Coronary Prevention Project and a nursing intervention designed to provide supportive contacts have also shown promising results, but the outcomes have not been consistent.
Patients screened for the ENRICHD study, mentioned above, are evaluated for lack of social support and if eligible, are enrolled in a specific therapeutic programme designed to develop its availability. The study should help to define whether such an approach decreases morbidity and mortality in post-MI patients.