Remodelling of the heart after myocardial infarction
H a w e y D . White
Specialist in Cardiovascular Research, Coronary Care Unit, Green Lane Hospital, Auckland, New Zealand.
Abstract: In the first f m hours after the onset of coronary occlusion the infarct zone stretches due to myocyte slippage. Subsequently the noninfarct zone develops volume overload hypertrophy with series addition of new sarcomeres and fibre elongation. Dilatation is detrimental as it increases ventricular wall stress and oxygen demand, and re-entry of electrical impulses may be influenced by stretching ofthe ischaemic scar resulting in ventricular fibrillation. Left ventricular remodelling and dilatation is a progressive process which begins early and continues in the months after infarction. The major determinants of the extent of remodelling are infarct size and patency of the infarct-relared artery. Late reperfusion may revene initial infarct dilatation and decrease lefr ventricular volumes by inducing calcium-activated contracture of the actomyosin complex. Expansion may also be inhibited by acceleration of healing, splinting of the infarct zone by salvage of subepicardial cells, and blood in the coronary arteries and veins supporting rhe infarct zone. End-systolic volume is the strongest predictor of long-term prognosis after infarction. A number of therapies including throm bolysis, angiotensin-converring enzyme (ACE) inhibition and nitrates have been shown to decrease left ventricular dilatation. The optimal time for commencement, dose, duration and the effects of combinations of therapy are yet to be determined. (Aust N Z J Med 1992; 22: 601-606.) Key words: Myocardial injarction, remodelling of the heart, ACE inhibition.
INTRODUCTION ollowing myocardial infarction (MI) the most important long-term prognostic factor is the end-systolic volume of the left ventricle (LV).’ Major efforts should therefore be made to modify the ‘healing phase’ of acute infarction to prevent stretching of the infarct zone and consequent remodelling and dilatation of the ventricle. There are a number of therapies that have been shown to decrease left ventricular dilatation including thrombolytic therapy,* angiotensinconverting enzyme (ACE) inhibitors3-’ and nitrates.6 Recently ACE inhibition has been shown to improve long-term survival.’ After an acute coronary artery occlusion the immediate haemodynamic response includes both
F
systolic and diastolic dysfunction. Following loss of the contractile function of the infarcted myocardium, systolic ejection decreases and endsystolic volume increases secondary to an increase in diastolic pressure. Although dilatation may occur to maintain stroke volume following loss of myocardial function through the Frank Starling mechanism (Table l), not all dilatation is compensatory.*
CELLULAR BASIS OF REMODELLING In the first few hours after infarction the infarct zone stretches9-” due to slippage of myocytes12and subsequently the noninfarct zone begins to undergo remodelling and dilatation associated with increased wall stress. Volume overload hypertrophy of the
Correspondence: Dr H.D. White, Coronary Care Unit, Green Lane Hospital, Auckland 1003, New Zealand. REMODEILING THE HEART POST MI
Aust NZ J Med 1992; 22
601
I
TABLE 1: Frank Starling Mechanism
I
t LVEDP and LV volume
TABLE 3 Results of Increased Wall Stress
4 Frank Starling Mechanism 4 Increased systolic ejection
Volume overload hypertrophy
Infarct expansion
I
I
LVEDP = left ventricular end-diastolic pressure LV = left ventricular
& contractility
ventricle develops with the series addition of new sarcomeres and fibre elongation. Left ventricular dilatation is a progressive process, beginning in the first few hours," and with most of the dilatation occurring in the first few days but also continuing during the months after infarction. T h e magnitude of the remodelling is related to infarct size and occurs most commonly in anterior Q-wave infarctions (Table 2). Dilatation is also related to patency of the infarct-related a ~ t e r y . ~ , ' ~ Development of left ventricular dilatation is detrimental as it increases ventricular wall stress leading to increased myocardial oxygen demand. T h i s may be a factor in causing reentry of electrical impulses between the normal myocardium and fibres in the stretched ischaemic scar, resulting in ventricular fibrillation. Late remodelling also explains the occurrence of the development of late heart failure (Table 3) despite the absence of an intercurrent infarction. Infarct expansion also distorts the shape of the ventricle which distorts the normal contractile pattern, and patients with the most spherical ventricles are 'more likely to develop heart failure and to do poorly.14
PROGNOSTIC IMPORTANCE OF LEFT VENTRICULAR VOLUMES Previous studies of long-term survival after infarction have evaluated left ventricular ejection fraction as a prognostic factor. In our study angiog r a p h i c end-systolic volumes and end-diastolic volumes were measured together with other possible clinical and angiographic predictors of cardiac
0 Infarct size 0 Infarct site: anterior or inferior 0 Patency of the infarct-related artery ventricular wall stress preload and
Aust
Infarct zone
I
3. Restored stroke volume
602
Noninfarct zone
NZ J Med
I,
Late heart failure
mortality.' The patients were all males younger than 60 years of age who were investigated four- eight weeks after a first (n= 443) or recurrent (n= 162) MI. They comprised 82% of consecutive cases and follow-up was for 78 -t 32 months. Survival data were analysed by multivariate analysis of ejection fraction, end-systolic volume, end-diastolic volume, myocardial score, occlusion score, stenosis score, age, first or recurrent MI, cardiac surgery, continued cigarette smoking, and fi-blocker treatment. There were 101 cardiac deaths. This is the largest study of long-term prognosis after M I in angiographically defined patients. It describes more than 47,000 patient months of follow-up compared with 3000 to 9000 months in the other studies. As patients with important ischaemia had coronary artery bypass grafting, we found only a weak predictive value for the severity of coronary artery disease (CAD). On multivariate analysis end-systolic volume was the major risk factor for cardiac death. The additional predictive power of end-systolic volume was seen when ejection fraction was reduced. For patients with ejection fractions of 4049% and end-systolic volumes > 95 ml the 5-year survival was 20% less than for patients with the same ejection fraction and end-systolic volume < 95 ml. Although end-diastolic volume is important, endsystolic is more so, and the relationship between endsystolic volume and end-diastolicvolume is not linear. If end-diastolic volume increases by 60 ml and the heart maintains the same stroke volume, end-systolic volume will increase by the same amount but percentage-wise by more. Also, following infarction the percentage change in end-systolic volume is greater than end-diastolic volume (Table ,).I5
THROMBOLYTIC THERAPY AND LEFT VENTRICULAR REMODELLING One of the most important factors affecting remodelling is whether the infarct-related artery is 1992; 22
REMODELLING THE HEART POST MI
TABLE 4 Relationship of End-Systolic Volume to End-Diastolic Volume End-systolic volume = end-diastolic volume End-systolic volume = end-diastolic volume
-
stroke volume (ejection fraction x end-diastolic volume)
Example If EDV increases by 60 ml and the stroke volume is maintained O/o
End-diastolic volume (ml) End-systolic volume (ml) Stroke volume (ml) Ejection fraction (ml)
180 120 60
33O/o
patent or n0t3,I3(Table 2). Jeremy et al. showed in a retrospective analysis of patients with infarction who did not receive thrombolytic therapy that patency of the infarct-related artery was an important predictor of increased ventricular volumes one month after infarction independent of infarct size.13 In our study evaluating the effect of thrombolytic therapy on left ventricular function we showed that streptokinase reduced end-systolic volumes by 18 ml when evaluated at three weeks after infarction, i.e. 50% of the dilatation that occurred in the placebo group.' Clearly patency of the infarct-related artery, if achieved early and if sustained, will result in myocardial salvage and preservation of left ventricular function. However, patency, if achieved late beyond the time of likely myocardial salvage, may also result in significant patient benefit by reducing left ventricular volumes. Hochman and Chew showed in rats that late reperfusion had no effect on infarct size but significantly decreased infarct expansion and late aneurysm formation.16 There are a number of mechanisms whereby late reperfusion could prevent left ventricular dilatation (Table 5). Following reperfusion there is increased haemorrhage, cell swelling, oedema and contraction band necrosis. These factors may make reperfused infarcts stiffer and less likely to expand. Late reperfusion may reverse initial infarct dilatation and decrease left ventricular volumes by inducing calcium-activated contracture of the actomyosin complex.s Expansion may be inhibited by acceleration of healing following reperfusion and aneurysm formation may be decreased. Late reperfusion may also salvage a rim of subepicardial cells which may form a splint inhibiting expansion and blood in the coronary arteries and veins can support the infarct zone and prevent infarct expansion. There are a number of other reasons why administration of late thrombolytic therapy may improve long-term survival, such as decreased arrhythmogenesis and provision of collaterals to another infarct zone if a subsequent infarction REMODELLING T H E HEART POST MI
change
240 180 60 25O/o
TABLE 5 Possible Mechanisms for Delayed Reperfusion Preventing Left Ventricular Dilatation ~
Increased stiffness of necrotic muscle opposing expansion calcium-activated contraction of actom yosin complex production of oedema Increased rate of healing Preservation of islets of epicardium forming a splint Coronary arteries forming a skeleton Coronary veins forming a skeleton
occurs in a different coronary artery territory.17 The combined data from the late trials show a 12% reduction in 35-day mortality (1.8% absolute, p < 0.05) for patients treated with streptokinase between seven and 12 hours and a 6% reduction for patients treated between 13 and 24 hours (P=NS).l7This therapy should be given to all patients without contraindications presenting within 12 hours with S T elevation or bundle branch block and to some, such as patients with large infarction, haemodynamic instability or continuing pain, up to 24 hours.
THE RENIN-ANGIOTENSIN SYSTEM AFTER MI Following M I there is increased activity of both the sympathetic nervous system and the reninangiotensin system.18 There may also be local activation of the renin-angiotensin system within the heart. This increased neuroendocrine activity is detrimental because vasoconstriction of the coronary arteries and increased afterload promote infarct expansion, increase heart demands, recurrent ischaemia, heart failure and arrhythmias, and the trophic effects of angiotensin on myocardial protein synthesis may lead to myocardial hypertrophy. BENEFICIAL EFFECTS OF ACE INHIBITORS Mark and Janice Pfeffer showed in rats that captopril reduced the amount of cardiac dilatation Aust NZ J Med 1992; 22 603
in the chronic phase of infarction and long-term survival was i m p r ~ v e d . ’ ~T h e most marked improvement in survival was in animals with moderate-sized infarctions. Subsequently, the benefits were shown angiographically in patients with anterior infarctions where captopril decreased the amount of left ventricular dilatation compared to patients treated with p l a ~ e b o . Sharpe ~ also showed echocardiographically that captopril begun seven-10 days after infarction in patients with impaired left ventricular function decreased left ventricular dilatation compared to patients treated with frusemide or p l a ~ e b o . ~ A further study confirmed these findings in patients treated within 24-48 hours after infarction.20Results from a small study ( n = 38) with intravenous captopril (2 m g + 8 mg) followed by oral therapy four hours after rt-PA show a trend for smaller end-systolic volumes.21 However, seven patients were withdrawn from this study and 19 had angioplasty before assessment of volumes. There have recently been preliminary data presented from the CONSENSUS IIZ2and SAVE7 trials. In the CONSENSUS I1 trial enalapril was given intravenously on the first’ day of infarction (1 mg) followed by 2.5 mg BID titrated up to 20 mg per day. This trial was stopped prematurely after 6000 patients were randomised because it was judged by the safety monitoring committee that continuation to 9000 patients would still not result in a significant improvement in survival. T h e primary endpoint was 6-month mortality and the mortality was 10.9% in the enalapril group and 10.2% in the placebo group, 95% confidence intervals 0.91-1.26. However, there was a significant reduction in nonfatal heart failure in patients treated with enalapril. Early stopping of the trial may have led to significant biases in interpretation of the results. It is to be noted that in this trial there was no good evidence of hazard of early enalapril treatment. In particular there were no significant differences noted in patients who were hypotensive on entry. T h e safety monitoring committee had been concerned by a higher mortality rate among patients who became hypotensive after the start of treatment. However, subsequent analysis showed that these patients in the enalapril group were already at especially high risk and comparison with placebo patients was biased, and there was no evidence of hypotension induced by enalapril being harmful. I n particular there were no excessive deaths during the first ten days when one would expect any early hazard to be seen if it existed. Furthermore this trial was stopped at six months of follow-up and the benefit of modification of the 604 Aust NZ J Med 1992;
remodelling process is a chronic benefit, and one may not expect a benefit to be seen as early as six months. This trial in fact does not show ‘no benefit’; it merely shows ‘lack of evidence of benefit’, and the result is consistent with a 9% reduction in mortality which could be even twice as great because of the early stopping of the trial which could substantially bias the results. It is to be noted that there are several examples of trials with small numbers o f randomised patients showing unimpressive or nonsignificant trends, and yet when the trials were completed the results were compelling and have had a major impact on clinical practice, eg GISSI- 1 with streptokinase after about 6000 patientsYZ3 ISIS-2 with aspirin after about 6000 patient^.^^ There are three other ongoing trials where ACE inhibition is being administered early: the Chinese captopril study, GISSI-3 and ISIS-4. It may well be (as in other areas of medicine) that the patients who may benefit most are those who are the sickest, i.e. hypotensive patients could benefit most. Preliminary results of the SAVE trial show that oral captopril (test dose of 6.25 mg increasing to 50 mg TID) given to patients with ejection fractions
H a w e y D . White
Specialist in Cardiovascular Research, Coronary Care Unit, Green Lane Hospital, Auckland, New Zealand.
Abstract: In the first f m hours after the onset of coronary occlusion the infarct zone stretches due to myocyte slippage. Subsequently the noninfarct zone develops volume overload hypertrophy with series addition of new sarcomeres and fibre elongation. Dilatation is detrimental as it increases ventricular wall stress and oxygen demand, and re-entry of electrical impulses may be influenced by stretching ofthe ischaemic scar resulting in ventricular fibrillation. Left ventricular remodelling and dilatation is a progressive process which begins early and continues in the months after infarction. The major determinants of the extent of remodelling are infarct size and patency of the infarct-relared artery. Late reperfusion may revene initial infarct dilatation and decrease lefr ventricular volumes by inducing calcium-activated contracture of the actomyosin complex. Expansion may also be inhibited by acceleration of healing, splinting of the infarct zone by salvage of subepicardial cells, and blood in the coronary arteries and veins supporting rhe infarct zone. End-systolic volume is the strongest predictor of long-term prognosis after infarction. A number of therapies including throm bolysis, angiotensin-converring enzyme (ACE) inhibition and nitrates have been shown to decrease left ventricular dilatation. The optimal time for commencement, dose, duration and the effects of combinations of therapy are yet to be determined. (Aust N Z J Med 1992; 22: 601-606.) Key words: Myocardial injarction, remodelling of the heart, ACE inhibition.
INTRODUCTION ollowing myocardial infarction (MI) the most important long-term prognostic factor is the end-systolic volume of the left ventricle (LV).’ Major efforts should therefore be made to modify the ‘healing phase’ of acute infarction to prevent stretching of the infarct zone and consequent remodelling and dilatation of the ventricle. There are a number of therapies that have been shown to decrease left ventricular dilatation including thrombolytic therapy,* angiotensinconverting enzyme (ACE) inhibitors3-’ and nitrates.6 Recently ACE inhibition has been shown to improve long-term survival.’ After an acute coronary artery occlusion the immediate haemodynamic response includes both
F
systolic and diastolic dysfunction. Following loss of the contractile function of the infarcted myocardium, systolic ejection decreases and endsystolic volume increases secondary to an increase in diastolic pressure. Although dilatation may occur to maintain stroke volume following loss of myocardial function through the Frank Starling mechanism (Table l), not all dilatation is compensatory.*
CELLULAR BASIS OF REMODELLING In the first few hours after infarction the infarct zone stretches9-” due to slippage of myocytes12and subsequently the noninfarct zone begins to undergo remodelling and dilatation associated with increased wall stress. Volume overload hypertrophy of the
Correspondence: Dr H.D. White, Coronary Care Unit, Green Lane Hospital, Auckland 1003, New Zealand. REMODEILING THE HEART POST MI
Aust NZ J Med 1992; 22
601
I
TABLE 1: Frank Starling Mechanism
I
t LVEDP and LV volume
TABLE 3 Results of Increased Wall Stress
4 Frank Starling Mechanism 4 Increased systolic ejection
Volume overload hypertrophy
Infarct expansion
I
I
LVEDP = left ventricular end-diastolic pressure LV = left ventricular
& contractility
ventricle develops with the series addition of new sarcomeres and fibre elongation. Left ventricular dilatation is a progressive process, beginning in the first few hours," and with most of the dilatation occurring in the first few days but also continuing during the months after infarction. T h e magnitude of the remodelling is related to infarct size and occurs most commonly in anterior Q-wave infarctions (Table 2). Dilatation is also related to patency of the infarct-related a ~ t e r y . ~ , ' ~ Development of left ventricular dilatation is detrimental as it increases ventricular wall stress leading to increased myocardial oxygen demand. T h i s may be a factor in causing reentry of electrical impulses between the normal myocardium and fibres in the stretched ischaemic scar, resulting in ventricular fibrillation. Late remodelling also explains the occurrence of the development of late heart failure (Table 3) despite the absence of an intercurrent infarction. Infarct expansion also distorts the shape of the ventricle which distorts the normal contractile pattern, and patients with the most spherical ventricles are 'more likely to develop heart failure and to do poorly.14
PROGNOSTIC IMPORTANCE OF LEFT VENTRICULAR VOLUMES Previous studies of long-term survival after infarction have evaluated left ventricular ejection fraction as a prognostic factor. In our study angiog r a p h i c end-systolic volumes and end-diastolic volumes were measured together with other possible clinical and angiographic predictors of cardiac
0 Infarct size 0 Infarct site: anterior or inferior 0 Patency of the infarct-related artery ventricular wall stress preload and
Aust
Infarct zone
I
3. Restored stroke volume
602
Noninfarct zone
NZ J Med
I,
Late heart failure
mortality.' The patients were all males younger than 60 years of age who were investigated four- eight weeks after a first (n= 443) or recurrent (n= 162) MI. They comprised 82% of consecutive cases and follow-up was for 78 -t 32 months. Survival data were analysed by multivariate analysis of ejection fraction, end-systolic volume, end-diastolic volume, myocardial score, occlusion score, stenosis score, age, first or recurrent MI, cardiac surgery, continued cigarette smoking, and fi-blocker treatment. There were 101 cardiac deaths. This is the largest study of long-term prognosis after M I in angiographically defined patients. It describes more than 47,000 patient months of follow-up compared with 3000 to 9000 months in the other studies. As patients with important ischaemia had coronary artery bypass grafting, we found only a weak predictive value for the severity of coronary artery disease (CAD). On multivariate analysis end-systolic volume was the major risk factor for cardiac death. The additional predictive power of end-systolic volume was seen when ejection fraction was reduced. For patients with ejection fractions of 4049% and end-systolic volumes > 95 ml the 5-year survival was 20% less than for patients with the same ejection fraction and end-systolic volume < 95 ml. Although end-diastolic volume is important, endsystolic is more so, and the relationship between endsystolic volume and end-diastolicvolume is not linear. If end-diastolic volume increases by 60 ml and the heart maintains the same stroke volume, end-systolic volume will increase by the same amount but percentage-wise by more. Also, following infarction the percentage change in end-systolic volume is greater than end-diastolic volume (Table ,).I5
THROMBOLYTIC THERAPY AND LEFT VENTRICULAR REMODELLING One of the most important factors affecting remodelling is whether the infarct-related artery is 1992; 22
REMODELLING THE HEART POST MI
TABLE 4 Relationship of End-Systolic Volume to End-Diastolic Volume End-systolic volume = end-diastolic volume End-systolic volume = end-diastolic volume
-
stroke volume (ejection fraction x end-diastolic volume)
Example If EDV increases by 60 ml and the stroke volume is maintained O/o
End-diastolic volume (ml) End-systolic volume (ml) Stroke volume (ml) Ejection fraction (ml)
180 120 60
33O/o
patent or n0t3,I3(Table 2). Jeremy et al. showed in a retrospective analysis of patients with infarction who did not receive thrombolytic therapy that patency of the infarct-related artery was an important predictor of increased ventricular volumes one month after infarction independent of infarct size.13 In our study evaluating the effect of thrombolytic therapy on left ventricular function we showed that streptokinase reduced end-systolic volumes by 18 ml when evaluated at three weeks after infarction, i.e. 50% of the dilatation that occurred in the placebo group.' Clearly patency of the infarct-related artery, if achieved early and if sustained, will result in myocardial salvage and preservation of left ventricular function. However, patency, if achieved late beyond the time of likely myocardial salvage, may also result in significant patient benefit by reducing left ventricular volumes. Hochman and Chew showed in rats that late reperfusion had no effect on infarct size but significantly decreased infarct expansion and late aneurysm formation.16 There are a number of mechanisms whereby late reperfusion could prevent left ventricular dilatation (Table 5). Following reperfusion there is increased haemorrhage, cell swelling, oedema and contraction band necrosis. These factors may make reperfused infarcts stiffer and less likely to expand. Late reperfusion may reverse initial infarct dilatation and decrease left ventricular volumes by inducing calcium-activated contracture of the actomyosin complex.s Expansion may be inhibited by acceleration of healing following reperfusion and aneurysm formation may be decreased. Late reperfusion may also salvage a rim of subepicardial cells which may form a splint inhibiting expansion and blood in the coronary arteries and veins can support the infarct zone and prevent infarct expansion. There are a number of other reasons why administration of late thrombolytic therapy may improve long-term survival, such as decreased arrhythmogenesis and provision of collaterals to another infarct zone if a subsequent infarction REMODELLING T H E HEART POST MI
change
240 180 60 25O/o
TABLE 5 Possible Mechanisms for Delayed Reperfusion Preventing Left Ventricular Dilatation ~
Increased stiffness of necrotic muscle opposing expansion calcium-activated contraction of actom yosin complex production of oedema Increased rate of healing Preservation of islets of epicardium forming a splint Coronary arteries forming a skeleton Coronary veins forming a skeleton
occurs in a different coronary artery territory.17 The combined data from the late trials show a 12% reduction in 35-day mortality (1.8% absolute, p < 0.05) for patients treated with streptokinase between seven and 12 hours and a 6% reduction for patients treated between 13 and 24 hours (P=NS).l7This therapy should be given to all patients without contraindications presenting within 12 hours with S T elevation or bundle branch block and to some, such as patients with large infarction, haemodynamic instability or continuing pain, up to 24 hours.
THE RENIN-ANGIOTENSIN SYSTEM AFTER MI Following M I there is increased activity of both the sympathetic nervous system and the reninangiotensin system.18 There may also be local activation of the renin-angiotensin system within the heart. This increased neuroendocrine activity is detrimental because vasoconstriction of the coronary arteries and increased afterload promote infarct expansion, increase heart demands, recurrent ischaemia, heart failure and arrhythmias, and the trophic effects of angiotensin on myocardial protein synthesis may lead to myocardial hypertrophy. BENEFICIAL EFFECTS OF ACE INHIBITORS Mark and Janice Pfeffer showed in rats that captopril reduced the amount of cardiac dilatation Aust NZ J Med 1992; 22 603
in the chronic phase of infarction and long-term survival was i m p r ~ v e d . ’ ~T h e most marked improvement in survival was in animals with moderate-sized infarctions. Subsequently, the benefits were shown angiographically in patients with anterior infarctions where captopril decreased the amount of left ventricular dilatation compared to patients treated with p l a ~ e b o . Sharpe ~ also showed echocardiographically that captopril begun seven-10 days after infarction in patients with impaired left ventricular function decreased left ventricular dilatation compared to patients treated with frusemide or p l a ~ e b o . ~ A further study confirmed these findings in patients treated within 24-48 hours after infarction.20Results from a small study ( n = 38) with intravenous captopril (2 m g + 8 mg) followed by oral therapy four hours after rt-PA show a trend for smaller end-systolic volumes.21 However, seven patients were withdrawn from this study and 19 had angioplasty before assessment of volumes. There have recently been preliminary data presented from the CONSENSUS IIZ2and SAVE7 trials. In the CONSENSUS I1 trial enalapril was given intravenously on the first’ day of infarction (1 mg) followed by 2.5 mg BID titrated up to 20 mg per day. This trial was stopped prematurely after 6000 patients were randomised because it was judged by the safety monitoring committee that continuation to 9000 patients would still not result in a significant improvement in survival. T h e primary endpoint was 6-month mortality and the mortality was 10.9% in the enalapril group and 10.2% in the placebo group, 95% confidence intervals 0.91-1.26. However, there was a significant reduction in nonfatal heart failure in patients treated with enalapril. Early stopping of the trial may have led to significant biases in interpretation of the results. It is to be noted that in this trial there was no good evidence of hazard of early enalapril treatment. In particular there were no significant differences noted in patients who were hypotensive on entry. T h e safety monitoring committee had been concerned by a higher mortality rate among patients who became hypotensive after the start of treatment. However, subsequent analysis showed that these patients in the enalapril group were already at especially high risk and comparison with placebo patients was biased, and there was no evidence of hypotension induced by enalapril being harmful. I n particular there were no excessive deaths during the first ten days when one would expect any early hazard to be seen if it existed. Furthermore this trial was stopped at six months of follow-up and the benefit of modification of the 604 Aust NZ J Med 1992;
remodelling process is a chronic benefit, and one may not expect a benefit to be seen as early as six months. This trial in fact does not show ‘no benefit’; it merely shows ‘lack of evidence of benefit’, and the result is consistent with a 9% reduction in mortality which could be even twice as great because of the early stopping of the trial which could substantially bias the results. It is to be noted that there are several examples of trials with small numbers o f randomised patients showing unimpressive or nonsignificant trends, and yet when the trials were completed the results were compelling and have had a major impact on clinical practice, eg GISSI- 1 with streptokinase after about 6000 patientsYZ3 ISIS-2 with aspirin after about 6000 patient^.^^ There are three other ongoing trials where ACE inhibition is being administered early: the Chinese captopril study, GISSI-3 and ISIS-4. It may well be (as in other areas of medicine) that the patients who may benefit most are those who are the sickest, i.e. hypotensive patients could benefit most. Preliminary results of the SAVE trial show that oral captopril (test dose of 6.25 mg increasing to 50 mg TID) given to patients with ejection fractions
