AMBICAN JOURNAL OF EPIDIMIOLOOY
Vol. 109, No. 4
Copyright © 1979 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved
Printed in USA.
PREVENTIVE TRIALS AND TRIBULATIONS WILLIAM B. KANNEL1
1 Heart Disease Epidemiology Study, National Heart Institute, National Institutes of Health, 123 Lincoln St., Framingham, MA 01701.
ministration study of hypertension concluded that there was little efficacy in treating mild hypertension based on the response of the total sample. It is conceivable in the VA study, indeed even likely, that there is a subgroup with a poor coronary risk profile (i.e., with high cholesterol, who smoke, who have impaired glucose tolerance and electrocardiographic (ECG) abnormalities) who were greatly benefited by treatment. In those without a poor risk profile the risk may be so low that the hazards of antihypertensive treatment outweigh any possible benefits. Likewise for secondary prevention it is also possible to identify subgroups whose risk is little more than that of thosefreeof coronary heart disease the same age and less than that of some with only a poor coronary risk profile. Such good risk persons with established coronary disease stand to benefit little from drastic procedures such as cardiac surgery which carries an intraoperative myocardial infarction rate which may be as high as 10 per cent, a 15 per cent graft closure rate, an operative mortality of 4 per cent, a considerable morbidity and expense, and an uncertain (at this point) knowledge of the duration of symptomatic relief and the effect on. survival. Detre and Kuller correctly point out that the chief determinants of survival after the appearance of overt coronary heart disease are the extent of coronary artery involvement and the amount of myocardial damage sustained in the attack. This has been unequivocally and consistently demonstrated. However, these may be chiefly related to proximate mortality in the first six months to one year after the onset of symptoms, particu-
405
Downloaded from https://academic.oup.com/aje/article-abstract/109/4/405/195088 by Lancaster University user on 12 January 2019
Detre and Kuller (1) have called attention to an important aspect of the natural history of coronary heart disease as it applies to the interpretation of clinical preventive trials. What they have to say applies to both primary and secondary prevention. If there is anything we have learned over the past 30 years from epidemiologic studies of the way coronary heart disease arises, evolves and terminates fatally in populations it is that this is a multifactorial problem. Atherogenic traits, risk factors which promote them, signs of pre-clinical compromise of the coronary circulation and host susceptibility to all these influences are involved. The authors correctly assert that patients entered into secondary prevention trials are often a heterogeneous sample subject to a wide range of risks of death and progression of disease. They also accurately point out that erroneous conclusions about the efficacy of therapy or preventive efforts can be drawn from an analysis which deals with findings on the total sample incorrectly treated as a homogeneous entity. They are also correct in their assertion that the usual clinical risk factors have not been shown to be highly correlated with survival following myocardial infarction. To deal with candidates for coronary heart disease or those with established coronary heart disease as a uniform entity is illogical both from the standpoint of prediction and treatment. This applies to primary as well as to secondary prevention. For example, the Veterans Ad-
406
WILLIAM B. KANNEL
well as those following an event may not be characteristic of long-term values. Atherogenic effects are very likely a time-dose product of risk factors. Yet another difficulty is the possibility that factors which promote the underlying atherosclerosis are not necessarily the same as those which precipitate attacks or affect the course of disease once established. It is indeed likely that the non-invasive procedures to assess left ventricular function and coronary artery involvement called for by Detre and Kuller will become available. A battery of tests made up of some combination of the exercise ECG, R-wave loss following myocardial infarction, increase in heart size, decline in vital capacity, rise in heart rate, fall in blood pressure, echocardiographic estimate of ejection fraction or myocardial imaging will probably identify persons with a compromised coronary circulation and poor cardiac function. These should correlate well with the findings on arteriography and ventriculography. The correlation can never be perfect, however. Even angiography does not correlate entirely with clinical status and some persons with severe involvement on coronary cineangiogram have no symptoms at all while some with little involvement have symptoms or even a myocardial infarction. Even an angiogram has a technical error and a stenotic artery is not exactly equivalent to angina, ischemia or infarction. A positive exercise ECG also indicates ischemia, yet this can occur without concurrent ischemic chest pain. Finally, it should be pointed out that clinical trials for secondary prevention should involve preventive measures which really affect the risk attributes. It is really not very relevant to report that diet or drug therapy for hyperlipidemia was ineffective when the change achieved was miniscule. Nor is it convincing to report no effect in elderly subjects with advancedfibrocalcificdisease in a short du-
Downloaded from https://academic.oup.com/aje/article-abstract/109/4/405/195088 by Lancaster University user on 12 January 2019
larly in the case of myocardial infarction. For men mortality in the first year is about 20 per cent compared to the 4—5 per cent per annum thereafter. In this critical early interval the standard risk factors can be expected to play little role because of the overwhelming influence of the extent to which the coronary circulation is compromised and the amount of myocardial damage sustained. It would appear that preventive trials should take into account this feature of the natural history of coronary heart disease. Short trials can demonstrate only the efficacy of measures which affect cardiac performance, cardiac rhythmicity and the coronary circulation rather directly and rapidly. Thus trials involving correction of atherogenic traits should probably begin after this period of early attrition and continue for long periods thereafter. Another point of confusion about the relation of standard risk factors to prognosis after onset of overt coronary heart disease is failure to take into account the fact that the occurrence of the coronary event can alter the level of the risk factors. Thus, blood pressures, serum cholesterol and weights all tend to fall for a time after interim myocardial infarction. In the case of blood pressure it has been shown that a substantial fall in blood pressure carries an ominous prognosis. This obscures the strong positive relation of post-myocardial infarction blood pressure to survival which is present when those whose pressures have fallen are excluded from consideration. Thus, by considering them all as a homogeneous group it would seem that there is little justification for blood pressure control after myocardial infarction. However, consideration of the subgroup who have not sustained the fall, suggests great potential for blood pressure control in secondary prevention after myocardial infarction or angina pectoris. Another aspect of the problem which is often overlooked is that the level of risk factors measured just before an event as
407
PREVENTIVE TRIALS AND TRIBULATIONS
platelet adhesion to intimal lesions, to reduce the irritability of the myocardium, to reduce the cardiac workload and to promote collateral circulation would seem more rational. REFERENCE
1. Detre K, Kuller LH: Hazards in the evaluation of secondary prevention trials. Am J Epidemiol 109:400-404, 1979 ;
Downloaded from https://academic.oup.com/aje/article-abstract/109/4/405/195088 by Lancaster University user on 12 January 2019
ration trial which attempts to undo 40 years of hypertension, hyperlipidemia, and cigarette smoking in a few years, particularly after the myocardium has been damaged. At that point it is a race against impending cardiovascular catastrophes and death and the subject must survive long enough for the modifications to have an effect. Most will not have enough years left. At that stage measures to reduce
Vol. 109, No. 4
Copyright © 1979 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved
Printed in USA.
PREVENTIVE TRIALS AND TRIBULATIONS WILLIAM B. KANNEL1
1 Heart Disease Epidemiology Study, National Heart Institute, National Institutes of Health, 123 Lincoln St., Framingham, MA 01701.
ministration study of hypertension concluded that there was little efficacy in treating mild hypertension based on the response of the total sample. It is conceivable in the VA study, indeed even likely, that there is a subgroup with a poor coronary risk profile (i.e., with high cholesterol, who smoke, who have impaired glucose tolerance and electrocardiographic (ECG) abnormalities) who were greatly benefited by treatment. In those without a poor risk profile the risk may be so low that the hazards of antihypertensive treatment outweigh any possible benefits. Likewise for secondary prevention it is also possible to identify subgroups whose risk is little more than that of thosefreeof coronary heart disease the same age and less than that of some with only a poor coronary risk profile. Such good risk persons with established coronary disease stand to benefit little from drastic procedures such as cardiac surgery which carries an intraoperative myocardial infarction rate which may be as high as 10 per cent, a 15 per cent graft closure rate, an operative mortality of 4 per cent, a considerable morbidity and expense, and an uncertain (at this point) knowledge of the duration of symptomatic relief and the effect on. survival. Detre and Kuller correctly point out that the chief determinants of survival after the appearance of overt coronary heart disease are the extent of coronary artery involvement and the amount of myocardial damage sustained in the attack. This has been unequivocally and consistently demonstrated. However, these may be chiefly related to proximate mortality in the first six months to one year after the onset of symptoms, particu-
405
Downloaded from https://academic.oup.com/aje/article-abstract/109/4/405/195088 by Lancaster University user on 12 January 2019
Detre and Kuller (1) have called attention to an important aspect of the natural history of coronary heart disease as it applies to the interpretation of clinical preventive trials. What they have to say applies to both primary and secondary prevention. If there is anything we have learned over the past 30 years from epidemiologic studies of the way coronary heart disease arises, evolves and terminates fatally in populations it is that this is a multifactorial problem. Atherogenic traits, risk factors which promote them, signs of pre-clinical compromise of the coronary circulation and host susceptibility to all these influences are involved. The authors correctly assert that patients entered into secondary prevention trials are often a heterogeneous sample subject to a wide range of risks of death and progression of disease. They also accurately point out that erroneous conclusions about the efficacy of therapy or preventive efforts can be drawn from an analysis which deals with findings on the total sample incorrectly treated as a homogeneous entity. They are also correct in their assertion that the usual clinical risk factors have not been shown to be highly correlated with survival following myocardial infarction. To deal with candidates for coronary heart disease or those with established coronary heart disease as a uniform entity is illogical both from the standpoint of prediction and treatment. This applies to primary as well as to secondary prevention. For example, the Veterans Ad-
406
WILLIAM B. KANNEL
well as those following an event may not be characteristic of long-term values. Atherogenic effects are very likely a time-dose product of risk factors. Yet another difficulty is the possibility that factors which promote the underlying atherosclerosis are not necessarily the same as those which precipitate attacks or affect the course of disease once established. It is indeed likely that the non-invasive procedures to assess left ventricular function and coronary artery involvement called for by Detre and Kuller will become available. A battery of tests made up of some combination of the exercise ECG, R-wave loss following myocardial infarction, increase in heart size, decline in vital capacity, rise in heart rate, fall in blood pressure, echocardiographic estimate of ejection fraction or myocardial imaging will probably identify persons with a compromised coronary circulation and poor cardiac function. These should correlate well with the findings on arteriography and ventriculography. The correlation can never be perfect, however. Even angiography does not correlate entirely with clinical status and some persons with severe involvement on coronary cineangiogram have no symptoms at all while some with little involvement have symptoms or even a myocardial infarction. Even an angiogram has a technical error and a stenotic artery is not exactly equivalent to angina, ischemia or infarction. A positive exercise ECG also indicates ischemia, yet this can occur without concurrent ischemic chest pain. Finally, it should be pointed out that clinical trials for secondary prevention should involve preventive measures which really affect the risk attributes. It is really not very relevant to report that diet or drug therapy for hyperlipidemia was ineffective when the change achieved was miniscule. Nor is it convincing to report no effect in elderly subjects with advancedfibrocalcificdisease in a short du-
Downloaded from https://academic.oup.com/aje/article-abstract/109/4/405/195088 by Lancaster University user on 12 January 2019
larly in the case of myocardial infarction. For men mortality in the first year is about 20 per cent compared to the 4—5 per cent per annum thereafter. In this critical early interval the standard risk factors can be expected to play little role because of the overwhelming influence of the extent to which the coronary circulation is compromised and the amount of myocardial damage sustained. It would appear that preventive trials should take into account this feature of the natural history of coronary heart disease. Short trials can demonstrate only the efficacy of measures which affect cardiac performance, cardiac rhythmicity and the coronary circulation rather directly and rapidly. Thus trials involving correction of atherogenic traits should probably begin after this period of early attrition and continue for long periods thereafter. Another point of confusion about the relation of standard risk factors to prognosis after onset of overt coronary heart disease is failure to take into account the fact that the occurrence of the coronary event can alter the level of the risk factors. Thus, blood pressures, serum cholesterol and weights all tend to fall for a time after interim myocardial infarction. In the case of blood pressure it has been shown that a substantial fall in blood pressure carries an ominous prognosis. This obscures the strong positive relation of post-myocardial infarction blood pressure to survival which is present when those whose pressures have fallen are excluded from consideration. Thus, by considering them all as a homogeneous group it would seem that there is little justification for blood pressure control after myocardial infarction. However, consideration of the subgroup who have not sustained the fall, suggests great potential for blood pressure control in secondary prevention after myocardial infarction or angina pectoris. Another aspect of the problem which is often overlooked is that the level of risk factors measured just before an event as
407
PREVENTIVE TRIALS AND TRIBULATIONS
platelet adhesion to intimal lesions, to reduce the irritability of the myocardium, to reduce the cardiac workload and to promote collateral circulation would seem more rational. REFERENCE
1. Detre K, Kuller LH: Hazards in the evaluation of secondary prevention trials. Am J Epidemiol 109:400-404, 1979 ;
Downloaded from https://academic.oup.com/aje/article-abstract/109/4/405/195088 by Lancaster University user on 12 January 2019
ration trial which attempts to undo 40 years of hypertension, hyperlipidemia, and cigarette smoking in a few years, particularly after the myocardium has been damaged. At that point it is a race against impending cardiovascular catastrophes and death and the subject must survive long enough for the modifications to have an effect. Most will not have enough years left. At that stage measures to reduce
