1
Trials and tribulations: the ISIS experience
Peter Sleight Field Morshal Alexander Professor o/ Cardiovascular Medicine, Universiry of Oxford, John Radcliffe Ilaspital, Headington, OxJord, UK.
Abstract: The ISIS trials group has grown slowly over the last ten years and now includes about 1000 collaborating hospitals worldwide. The original stimulus came from the thesis of a research student, Salim Yusuh a newly qualified Rhodes Scholar from India. Yusufs thesis (begun in 1976) examined methods of measuring injarct size quantitatively and paved the way to a preliminary trial 0fi.v. atenolol in 477 patients in Manchester and Oxford. Richard Peto was consulted about the statistical aspects of a trial design and pointed out that most clinical trials at that time were hopelessly inadequate in size 10 have the power reliably to ascertain the value of moderate but medically worthwhile treatments in many common conditions. W e thus developed much larger collaboration, which still continues t o grow. Dr Rory Collins joined the Oxford Cardiac Department and the C T S U in 1981 and now is the coordinator of the ISIS trials, and co-director with Richard Pet0 of the CTSU. Simple trials answer the questions rapidly and effectively. The mechanics of the trial do not hinder patient entry by participating doctors and nurses. The trials are mainly funded by industry, with ‘priming’grants from the British Heart Foundation. Once the protocol is agreed the funding companies have no further input. The trials are planned and supervised by an international steering committee. Sir Richard Doll chairs the data monitoring committees. A brief overview of the three completed trials will begiven, together with some peeps behind the scenes! (Aust NZ 3 Med 1992; 22: 583-586.) Key words: ISIS, clinical rrials, myocardial infarction, coronary heart disease, hypertension, thrombolyis.
INTRODUCTION he group at Prince Henry Hospital showed many years ago that left ventricular hypertrophy in young borderline hypertensives could be reversed by lowering blood pressure, particularly so for those treated by weight reduction compared to drug treatment with metoprolol.‘ It was one of the first studies to emphasise the difficulty in defining a division between normal and abnormal left ventricular mass and hence the importance of prospective studies using the subject as hidher own control. It also foreshadowed the importance
T
of hypertrophy as a marker for prognosis. Perhaps its greatest longterm importance was in the training it gave to the PhD student (non medical) whose thesis it formed, Stephen MacMahon. Stephen MacMahon has gone on to make an international reputation in epidemiology, particularly in the clarification of the impact of hypertension on its major lethal complications - stroke and coronary artery disease.2 He and his colleagues showed that in the general population quite small differences in blood pressure lead to surprisingly large increases in these risks. Although most studies emphasise the
Correspondence 10: Peter Sleight, Field Marshall Alexander Professor of Cardiovascular Medicine, Universiry of Oxford, John Radcliffe Hospital, Headingron, Oxford, U K . T H E lSlS EXPERIENCE
Aust N Z J Med 1992; 22
583
risk of stroke, these population studies show that coronary events are about five times more common than stroke.’ These studies also showed no evidence of a threshold effect for blood pressure, even at the lowest diastolic pressures. There was thus no evidence of a J or U shaped curve indicating an increased risk at lower pressure. O n a doubling ordinate scale the relation is linear. This is of great potential importance in the field of clinical trials, since it makes clear that the proportional decrease in risk by moving down from, for example, quintile 2 to quintile 1 of blood pressure gives the same proportional decrease in risk as moving from quintile (Q) 5 to Q 4. Of course, if everything else is balanced the absolute benefit is greater in the latter case than the former. However, if the subject in Q 2 is at high risk because of some other factors, for example, cholesterol, age, past history etc, then the absolute benefit of moving to Q 1 may be as much as Q5 to 4 4 in a subject with no other risk factors present. This analysis emphasises the need not to be too rigid in setting pressure levels at which we advise lowering blood pressure but to institute trials of lowering near ‘normal’ pressures in subjects at higher risk because of other factors. T h e ISIS trials organisation grew as a result of the PhD (DPhil in Oxford) of a young Rhodes Scholar from Kerala in Southern India, Salim Yusuf. After he had completed his thesis on the measurement of infarct size it seemed possible to mount a trial to see if we could influence the course of acute myocardial infarction (AMI). By what was to prove an extremely fortunate choice we involved Richard Peto, in Sir Richard Doll’s department, in preliminary advice. H e had long felt the need for adequate clinical trials in medicine. He saw clearly the confusion that had arisen from inadequately sized trial^.^ At first the thought of doing a trial in a few thousand patients was daunting and depressing. Perhaps wisely he did not at that stage specify the need for trials with 10-20 000 subjects! But he encouraged us to seek cooperation to try to develop a multicentre organisation.
THE EARLY DAYS We began in a small way and naturally turned to our seed corn - the young consultants in other hospitals who had trained in Oxford. The first was David Bennett who, with David Ramsdale, helped us with our first attempt - a study of intravenous (IN) atenolol early in AMI. Using Yusufs techniques we showed that IIV atenolol (in patients not at risk of shock or heart failure) reduced enzyme and E C G measures of infarct size.4 We also found that arrhythmias were significantly reduced (principal investigator Dr Paulo Rossi from Curitiba, 584 Aust NZ J Med 1992; 22
Brazil’) and that IIV betablockade led to very rapid pain relief in many subjects (principally the work of David Ramsdale in Manchester).’j These studies, in nearly 500 patients in Oxford and Manchester, formed the pilot studies for the main ISIS-I study of IIV atenolol in more than 16 000 patients, carried out in over 200 hospitals.’ Here I must pay tribute to two other people. We were soon joined by Dr Rory Collins, a young physician from St Thomas Hospital, who had decided to make medical epidemiology his career after his house jobs had been completed. Rory and Salim carried through the organisation of ISIS-1, with Richard Peto’s guidance. Both were to grow and achieve their own independence. T h e second person was ICI’s Dr John Cruickshank who backed ISIS-I enthusiastically and, with no great company support but a lot of personal guts, found the resources in ICI to get the trial going, and completed over a long three and a half years. There have been more pleasures than tribulations with the first ten years of ISIS, but the first tribulation has been the rather poor translation of the 15% mortality benefit with IIV betablockade into clinical practice. There may be several reasons for this. One undoubtedly was that in the then current climate of clinical cardiology the treatment was perceived by many as controversial, potentially hazardous, and only applicable to patients at low risk of death. Another reason was that ISIS was then young with no track record or charisma to help the bandwagon start rolling. A third, perhaps more important, was that it was overshadowed by the advent of the era of thrombolysis with GISSIu and ISIS-29 published in the two years after publication of ISIS- 1. However, the most important lesson from ISIS-1 was that such trials were not only feasible, but that there was huge enthusiasm to join freely in such projects from the nursing and medical staff of ordinary routine coronary care units. The word ‘freely’ is important, for it was one of the initial decisions that if we were to be able to mount such trials they should not be overexpensive. We therefore could not afford to pay our collaborators for their work, nor should we personally reward ourselves, or have any financial links with the companies who fund the trials. Another very important procedure for the credibility of the results is that the sponsoring companies are not involved in the conduct or analysis of the trial, which is run independently by the clinical trials service unit of Oxford University (co-directors R. Peto and R. Collins). This benefits both sides - the triallists and the companies, since this ‘arm’s length’ relationship greatly helps acceptance of the results, not only by the medical community but also THE I S I S EXPERIENCE
by the regulatory authorities such as the Medicines Commission in the UK and the FDA in the US. Regulatory problems are also eased by the multinational nature of the trials. We also involved the C C U senior nursing staff as an integral part of the trial. They were enthusiastic and in general gave greater continuity (and experience!) to the trial than a succession of junior medical staff in six month rotational posts.
OTHER TRIBULATIONS Perhaps the biggest challenge is persuading pharmaceutical companies, who are naturally intensely jealous of their product, to subject their drug to a rigorous trial. Behringwerke, a subsidiary of Hoechst, were particularly unusual in funding ISIS-2, since their product Streptase (SK) was out of patent, had a viable competitor, and had been little used in myocardial infarction. Another challenge was to persuade them to allow us to ‘factor’ in aspirin; at that time the streptase package insert warned of the dangers of combining SK and aspirin! Other tribulations occur when something serious goes wrong. Computers arc the key to ISIS or any large trial. Normally they are a boon. Their most obvious application is for data handling and analysis. However, they are equally useful for the innumerable ‘house keeping’ tasks in a trial. For example, the computer logs and charts the number of times the randomisation phone rings before it is answered by the randomisation staff. This enables the trial coordinator to adjust the number of staff needed to handle calls quickly. At one time in ISIS-3 the randomisation rate rose to a figure five or six times the highest rate achieved in any Oxford based trial! One potentially serious computer ‘glitch’occurred in ISIS-29 when one of the programmers modified the algorithm designed to randomise patients in a way which ensured even balance for such entry variables as age, blood pressure etc. Unfortunately the programme change also unintentionally resulted in many successive patients being allocated SK and aspirin placebo instead of a balance between active and placebo. Fortunately this was discovered relatively quickly and for the next period the computer ‘caught up’ by allocating more patients to active SK. These unbalanced blocks had to be analysed and reported separately; luckily the results during this period were the same as the rest of the trial, so in the end no harm was done.9 COLLABORATION WITH OTHER GROUPS We have had a long and symbiotic relationship with the GISSI trialists. ISIS is represented on the GISSI steering committee and vice versa. So, although the trials such as GISSI-2” and ISIS-3” have T H E ISIS EXPERIENCE
differences, for example in the drugs used, they also have much in common. This enables the trials to be combined in an overview of say tPA versus SK with regard to risk benefit ratio. The result has been that the impact on medical practice is much larger when two very large trials give a similar message. This was seen in the wide and very rapid acceptance of the use of thrombolytics which followed the publication of ISIS-2 and GISSI- 1. One of the tribulations common to all triallists is posthoc criticism of trial design or execution. Sometimes this is well deserved on a scientific basis; more often it is greatly fuelled by commercial rivalry, leading to downright misrcpresentation of the results of ISIS-3, for example.
THE FUTURE It is frequently said that megatrials are too large and clumsy to be affordable or desirable. This is a mistake. We are now at a stage where at least three groups - GISSI, ISIS and G U S T O - can randomise 20-50 000 patients and get results within 18 months or so. T h e first two organisations are very low cost compared with previous trials. Unfortunately, myocardial infarction is still common and the most promising of the newer strategies and compounds will need proper evaluation of their (increasing) cost against potential benefit. References I . MacMahon SW, MacDonald GJ, Bernstein L, Andrews G , Blacket RB. Comparison of weight reduction with metoprolol in the treatment of hypertension in young overweight patients. Lancet 1985; i: 1233-6. 2. MacMahon S, Peto R, Cutler J er al. Blood pressure, stroke, and coronary heart disease. Part 1. Prolonged differences in blood pressure: prospective observational studies corrected for the regression dilutioh bias. Lancet 1990; 335: 765-74. 3 . Yusuf S, Collins R, Peto R. Why do we need some large, simple randomized trials? Statistics in Medicine 1984; 3: 409-20. 4. Yusuf S, Sleight P, Rossi P e t a/. Reduction in infarct size, arrhythmias and chest pain by early intravenous beta blockade in suspected acute myocardial infarction. Circulation 1983; 67: 132-41. 5. Rossi PR, Yusuf S, Ramsdale D, Furze L, Sleight P. Reduction of ventricular arrhvthmias bv earlv inrravenous --atenolol in suspected acute miocardial infarction. Br Med J 1983; 286: 506-10. 6. Ramsdale DR, Faragher EB er ul. Ischaemic pain relief in patients with acute myocardial infarction by intravenous atenolol. Am Heart J 1982; 103: 459-67. 7. ISIS-I Collaborative Group. Randomised trial of intravenous atenolol among 16 027 cases ofsuspectrd acute myocardial infarction. Lancet 1986; ii: 57-66. 8. Gruppo Italian0 per lo Studio Della Streptochinasi nell‘Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986; i: 397-402. 9. ISIS-2 (Second International Study of Infarct Survival Collaborative Group). Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17 187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988; ii: 349-60.
Aust NZ J Med 1992; 22
2
i
585
10. Gruppo Italiano per lo Studio della Sopravvivenza nell’lnfarto Miocardico. GISSI-2. A factorial randomised trial of alteplase vs streptokinase and heparin vs no heparin among 12 490 patients with acute myocardial infarction. Lancet 1990; 336: 65-71.
586
Aust NZ J Med 1992; 22
11. ISIS-3 (Third International Study of Infarct Survival) Collaborative Group. ISIS3 a randomised comparison of strep tokinase vs tissue plasminogen activator vs anistreplase and of aspirin plus heparin vs aspirin alone among 41 299 cases of suspected acute myocardial infarction. Lancet 1992; i: 1-18.
THE ISIS EXPERIENCE
Trials and tribulations: the ISIS experience
Peter Sleight Field Morshal Alexander Professor o/ Cardiovascular Medicine, Universiry of Oxford, John Radcliffe Ilaspital, Headington, OxJord, UK.
Abstract: The ISIS trials group has grown slowly over the last ten years and now includes about 1000 collaborating hospitals worldwide. The original stimulus came from the thesis of a research student, Salim Yusuh a newly qualified Rhodes Scholar from India. Yusufs thesis (begun in 1976) examined methods of measuring injarct size quantitatively and paved the way to a preliminary trial 0fi.v. atenolol in 477 patients in Manchester and Oxford. Richard Peto was consulted about the statistical aspects of a trial design and pointed out that most clinical trials at that time were hopelessly inadequate in size 10 have the power reliably to ascertain the value of moderate but medically worthwhile treatments in many common conditions. W e thus developed much larger collaboration, which still continues t o grow. Dr Rory Collins joined the Oxford Cardiac Department and the C T S U in 1981 and now is the coordinator of the ISIS trials, and co-director with Richard Pet0 of the CTSU. Simple trials answer the questions rapidly and effectively. The mechanics of the trial do not hinder patient entry by participating doctors and nurses. The trials are mainly funded by industry, with ‘priming’grants from the British Heart Foundation. Once the protocol is agreed the funding companies have no further input. The trials are planned and supervised by an international steering committee. Sir Richard Doll chairs the data monitoring committees. A brief overview of the three completed trials will begiven, together with some peeps behind the scenes! (Aust NZ 3 Med 1992; 22: 583-586.) Key words: ISIS, clinical rrials, myocardial infarction, coronary heart disease, hypertension, thrombolyis.
INTRODUCTION he group at Prince Henry Hospital showed many years ago that left ventricular hypertrophy in young borderline hypertensives could be reversed by lowering blood pressure, particularly so for those treated by weight reduction compared to drug treatment with metoprolol.‘ It was one of the first studies to emphasise the difficulty in defining a division between normal and abnormal left ventricular mass and hence the importance of prospective studies using the subject as hidher own control. It also foreshadowed the importance
T
of hypertrophy as a marker for prognosis. Perhaps its greatest longterm importance was in the training it gave to the PhD student (non medical) whose thesis it formed, Stephen MacMahon. Stephen MacMahon has gone on to make an international reputation in epidemiology, particularly in the clarification of the impact of hypertension on its major lethal complications - stroke and coronary artery disease.2 He and his colleagues showed that in the general population quite small differences in blood pressure lead to surprisingly large increases in these risks. Although most studies emphasise the
Correspondence 10: Peter Sleight, Field Marshall Alexander Professor of Cardiovascular Medicine, Universiry of Oxford, John Radcliffe Hospital, Headingron, Oxford, U K . T H E lSlS EXPERIENCE
Aust N Z J Med 1992; 22
583
risk of stroke, these population studies show that coronary events are about five times more common than stroke.’ These studies also showed no evidence of a threshold effect for blood pressure, even at the lowest diastolic pressures. There was thus no evidence of a J or U shaped curve indicating an increased risk at lower pressure. O n a doubling ordinate scale the relation is linear. This is of great potential importance in the field of clinical trials, since it makes clear that the proportional decrease in risk by moving down from, for example, quintile 2 to quintile 1 of blood pressure gives the same proportional decrease in risk as moving from quintile (Q) 5 to Q 4. Of course, if everything else is balanced the absolute benefit is greater in the latter case than the former. However, if the subject in Q 2 is at high risk because of some other factors, for example, cholesterol, age, past history etc, then the absolute benefit of moving to Q 1 may be as much as Q5 to 4 4 in a subject with no other risk factors present. This analysis emphasises the need not to be too rigid in setting pressure levels at which we advise lowering blood pressure but to institute trials of lowering near ‘normal’ pressures in subjects at higher risk because of other factors. T h e ISIS trials organisation grew as a result of the PhD (DPhil in Oxford) of a young Rhodes Scholar from Kerala in Southern India, Salim Yusuf. After he had completed his thesis on the measurement of infarct size it seemed possible to mount a trial to see if we could influence the course of acute myocardial infarction (AMI). By what was to prove an extremely fortunate choice we involved Richard Peto, in Sir Richard Doll’s department, in preliminary advice. H e had long felt the need for adequate clinical trials in medicine. He saw clearly the confusion that had arisen from inadequately sized trial^.^ At first the thought of doing a trial in a few thousand patients was daunting and depressing. Perhaps wisely he did not at that stage specify the need for trials with 10-20 000 subjects! But he encouraged us to seek cooperation to try to develop a multicentre organisation.
THE EARLY DAYS We began in a small way and naturally turned to our seed corn - the young consultants in other hospitals who had trained in Oxford. The first was David Bennett who, with David Ramsdale, helped us with our first attempt - a study of intravenous (IN) atenolol early in AMI. Using Yusufs techniques we showed that IIV atenolol (in patients not at risk of shock or heart failure) reduced enzyme and E C G measures of infarct size.4 We also found that arrhythmias were significantly reduced (principal investigator Dr Paulo Rossi from Curitiba, 584 Aust NZ J Med 1992; 22
Brazil’) and that IIV betablockade led to very rapid pain relief in many subjects (principally the work of David Ramsdale in Manchester).’j These studies, in nearly 500 patients in Oxford and Manchester, formed the pilot studies for the main ISIS-I study of IIV atenolol in more than 16 000 patients, carried out in over 200 hospitals.’ Here I must pay tribute to two other people. We were soon joined by Dr Rory Collins, a young physician from St Thomas Hospital, who had decided to make medical epidemiology his career after his house jobs had been completed. Rory and Salim carried through the organisation of ISIS-1, with Richard Peto’s guidance. Both were to grow and achieve their own independence. T h e second person was ICI’s Dr John Cruickshank who backed ISIS-I enthusiastically and, with no great company support but a lot of personal guts, found the resources in ICI to get the trial going, and completed over a long three and a half years. There have been more pleasures than tribulations with the first ten years of ISIS, but the first tribulation has been the rather poor translation of the 15% mortality benefit with IIV betablockade into clinical practice. There may be several reasons for this. One undoubtedly was that in the then current climate of clinical cardiology the treatment was perceived by many as controversial, potentially hazardous, and only applicable to patients at low risk of death. Another reason was that ISIS was then young with no track record or charisma to help the bandwagon start rolling. A third, perhaps more important, was that it was overshadowed by the advent of the era of thrombolysis with GISSIu and ISIS-29 published in the two years after publication of ISIS- 1. However, the most important lesson from ISIS-1 was that such trials were not only feasible, but that there was huge enthusiasm to join freely in such projects from the nursing and medical staff of ordinary routine coronary care units. The word ‘freely’ is important, for it was one of the initial decisions that if we were to be able to mount such trials they should not be overexpensive. We therefore could not afford to pay our collaborators for their work, nor should we personally reward ourselves, or have any financial links with the companies who fund the trials. Another very important procedure for the credibility of the results is that the sponsoring companies are not involved in the conduct or analysis of the trial, which is run independently by the clinical trials service unit of Oxford University (co-directors R. Peto and R. Collins). This benefits both sides - the triallists and the companies, since this ‘arm’s length’ relationship greatly helps acceptance of the results, not only by the medical community but also THE I S I S EXPERIENCE
by the regulatory authorities such as the Medicines Commission in the UK and the FDA in the US. Regulatory problems are also eased by the multinational nature of the trials. We also involved the C C U senior nursing staff as an integral part of the trial. They were enthusiastic and in general gave greater continuity (and experience!) to the trial than a succession of junior medical staff in six month rotational posts.
OTHER TRIBULATIONS Perhaps the biggest challenge is persuading pharmaceutical companies, who are naturally intensely jealous of their product, to subject their drug to a rigorous trial. Behringwerke, a subsidiary of Hoechst, were particularly unusual in funding ISIS-2, since their product Streptase (SK) was out of patent, had a viable competitor, and had been little used in myocardial infarction. Another challenge was to persuade them to allow us to ‘factor’ in aspirin; at that time the streptase package insert warned of the dangers of combining SK and aspirin! Other tribulations occur when something serious goes wrong. Computers arc the key to ISIS or any large trial. Normally they are a boon. Their most obvious application is for data handling and analysis. However, they are equally useful for the innumerable ‘house keeping’ tasks in a trial. For example, the computer logs and charts the number of times the randomisation phone rings before it is answered by the randomisation staff. This enables the trial coordinator to adjust the number of staff needed to handle calls quickly. At one time in ISIS-3 the randomisation rate rose to a figure five or six times the highest rate achieved in any Oxford based trial! One potentially serious computer ‘glitch’occurred in ISIS-29 when one of the programmers modified the algorithm designed to randomise patients in a way which ensured even balance for such entry variables as age, blood pressure etc. Unfortunately the programme change also unintentionally resulted in many successive patients being allocated SK and aspirin placebo instead of a balance between active and placebo. Fortunately this was discovered relatively quickly and for the next period the computer ‘caught up’ by allocating more patients to active SK. These unbalanced blocks had to be analysed and reported separately; luckily the results during this period were the same as the rest of the trial, so in the end no harm was done.9 COLLABORATION WITH OTHER GROUPS We have had a long and symbiotic relationship with the GISSI trialists. ISIS is represented on the GISSI steering committee and vice versa. So, although the trials such as GISSI-2” and ISIS-3” have T H E ISIS EXPERIENCE
differences, for example in the drugs used, they also have much in common. This enables the trials to be combined in an overview of say tPA versus SK with regard to risk benefit ratio. The result has been that the impact on medical practice is much larger when two very large trials give a similar message. This was seen in the wide and very rapid acceptance of the use of thrombolytics which followed the publication of ISIS-2 and GISSI- 1. One of the tribulations common to all triallists is posthoc criticism of trial design or execution. Sometimes this is well deserved on a scientific basis; more often it is greatly fuelled by commercial rivalry, leading to downright misrcpresentation of the results of ISIS-3, for example.
THE FUTURE It is frequently said that megatrials are too large and clumsy to be affordable or desirable. This is a mistake. We are now at a stage where at least three groups - GISSI, ISIS and G U S T O - can randomise 20-50 000 patients and get results within 18 months or so. T h e first two organisations are very low cost compared with previous trials. Unfortunately, myocardial infarction is still common and the most promising of the newer strategies and compounds will need proper evaluation of their (increasing) cost against potential benefit. References I . MacMahon SW, MacDonald GJ, Bernstein L, Andrews G , Blacket RB. Comparison of weight reduction with metoprolol in the treatment of hypertension in young overweight patients. Lancet 1985; i: 1233-6. 2. MacMahon S, Peto R, Cutler J er al. Blood pressure, stroke, and coronary heart disease. Part 1. Prolonged differences in blood pressure: prospective observational studies corrected for the regression dilutioh bias. Lancet 1990; 335: 765-74. 3 . Yusuf S, Collins R, Peto R. Why do we need some large, simple randomized trials? Statistics in Medicine 1984; 3: 409-20. 4. Yusuf S, Sleight P, Rossi P e t a/. Reduction in infarct size, arrhythmias and chest pain by early intravenous beta blockade in suspected acute myocardial infarction. Circulation 1983; 67: 132-41. 5. Rossi PR, Yusuf S, Ramsdale D, Furze L, Sleight P. Reduction of ventricular arrhvthmias bv earlv inrravenous --atenolol in suspected acute miocardial infarction. Br Med J 1983; 286: 506-10. 6. Ramsdale DR, Faragher EB er ul. Ischaemic pain relief in patients with acute myocardial infarction by intravenous atenolol. Am Heart J 1982; 103: 459-67. 7. ISIS-I Collaborative Group. Randomised trial of intravenous atenolol among 16 027 cases ofsuspectrd acute myocardial infarction. Lancet 1986; ii: 57-66. 8. Gruppo Italian0 per lo Studio Della Streptochinasi nell‘Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986; i: 397-402. 9. ISIS-2 (Second International Study of Infarct Survival Collaborative Group). Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17 187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988; ii: 349-60.
Aust NZ J Med 1992; 22
2
i
585
10. Gruppo Italiano per lo Studio della Sopravvivenza nell’lnfarto Miocardico. GISSI-2. A factorial randomised trial of alteplase vs streptokinase and heparin vs no heparin among 12 490 patients with acute myocardial infarction. Lancet 1990; 336: 65-71.
586
Aust NZ J Med 1992; 22
11. ISIS-3 (Third International Study of Infarct Survival) Collaborative Group. ISIS3 a randomised comparison of strep tokinase vs tissue plasminogen activator vs anistreplase and of aspirin plus heparin vs aspirin alone among 41 299 cases of suspected acute myocardial infarction. Lancet 1992; i: 1-18.
THE ISIS EXPERIENCE
