- Email: [email protected]
Transmyocardial laser re vascularisation
CORRESPONDENCE 1
2
Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularisation in patients with refractory angina: a randomised controlled trial. Lancet 1999; 353: 519–24. Prêtre R, Turina MI. Laser to the heart: magic but costly, or only costly? Lancet 1999; 353: 512–13.
Sir—My experience of TMLR over 5 years includes 322 patients with the carbon dioxide laser and 70 with the Holmium-YAG.1 With the CO2 laser, at 1 year thallium studies showed increased vascularity in 55% of patients, 40% had no angina, 31% were class I, and 21% class II, all but six being class IV initially. The discrepancy seems to be due to a difference in patient selection and surgical technique. In their selection the Papworth team 2 insist on areas of reversible ischaemia, whereas we know that many fixed areas on thallium study actually represent muscle that is still viable, especially if the patient has severe pain and the area contracts in catheterisation or echocardiographic studies. These patients respond well to TMLR therapy but at Papworth all such patients were excluded, although they provide some striking results. Another difference in selection is that 73% of the Papworth group had only class III angina, whereas our patients were almost entirely class IV (363 of 369). We believe that benefit is more easily shown in patients who have more severe angina, perhaps because of a more highly developed intramyocardial collateral system available for redistribution of blood flow. Other less important variations include their decision that the patient must be able to carry out a treadmill exercise test, whereas most of our patients were unable to do so because of severe angina, debility, or musculoskeletal problems. They also excluded patients with unstable angina whereas our group includes 108 patients who could not be taken off intravenous medication for over a week without recurrence of chest pain. Finally, only 10% of their patients were female whereas our studies included 25% women. The most important difference, however, is that their report seems to indicate that they treated only areas of reversible ischaemia previously identified by nuclear studies. We learned early that you had to treat the entire ventricle and not only the reversible area. When we confined treatment to this area, we found that the patient still had class IV angina a year later, but re-operation during which the entire ventricle was treated
THE LANCET • Vol 353 • May 15, 1999
resulted in conversion to class I within 3 months. If the whole ventricle and the areas of fixed defect are not treated, we would expect the results to be less striking. Thus, the selection process of patients for the Papworth study was much different from that in our studies, particularly because most patients were class III rather than IV, and Schofield and colleagues insisted on reversible ischaemia by nuclear perfusion as a criterion for selection, excluding those with fixed but viable defects. The most important difference, however, could result from applying TMLR only to areas of demonstrated reversible ischaemia, rather than to the entire ventricle. However, the Papworth study has injected a needed word of caution, especially for the future when multiple types of energy sources come to be used and percutaneous laser revascularisation has become widespread. I received no funds for the project reported here, which was supported entirely by the hospital out of its budget for the Heart Institute. I have acted as a consultant for both PLC Medical Systems and Eclipse Surgical Technology, Inc. For each of these companies I have acted as a teacher to train new teams starting the procedure, and I am now receiving a small honorarium for each such effort. I also have a small number of shares in each company, which I purchased myself, and in each case the value has gone down and I have in fact lost money. Therefore, I do not believe that I have any conflict of interest that would affect the accuracy of my observations.
Allan M Lansing Heart Institute, Norton Audubon Hospital, Louisville, Kentucky, 40217, USA 1
2
Lansing AM. Transmyocardial revascularization: mechanism of action with carbon dioxide and holmium-yttriumaluminium-garnet lasers. J Thorac Cardiovasc Surg 1998; 115: 1392. Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularization in patients with refractory angina. Lancet 1999; 353: 519–224.
Sir—P M Schofield and colleagues1 present the first European randomised trial of TMLR separating the two study groups for the whole observation period without crossover. However, I have some comments. Most patients were in angina class CCS III (73%), which might account for the difference in improvement compared with other studies, with patients mainly in angina class IV. 2,3 Our own experience has also shown that these patients with angina at rest or even unstable angina do benefit most from this procedure and should be the main target group, whereas patients with fewer symptoms might
as well be treated by an improved drug regimen without surgery.4 The evaluation of patients includes few objective data. Two main endpoints are improvement in treadmill test and maximum exercise time, both being symptom limited. More objective data such as stress electrocardiography (ECG) and echocardiography, or even maximum oxygen uptake under exercise conditions, are lacking. The fact that most patients after TMLR reported dyspnoea or fatigue as reasons for stopping these tests might be interpreted as an angina equivalent without chest pain, as assumed by your commentators.5 Most of these patients with endstage coronary artery disease are known to be in poor physical condition and show a limited ventricular performance because of previous and often multiple infarctions, each of which would also be a sufficient explanation for dyspnoea and fatigue. Without ECG recordings to establish the true ischaemic threshold, an interpretation of these data is hardly possible. The only objective data are those obtained by nuclear scans and they do not show any significant difference between the two groups. However, the investigators themselves cast doubt on whether this method provides sufficiently high resolution. Our own data for Sestamibi scans, obtained by an independent investigator, have shown an improvement in ischaemic segments in around 40% of patients after TMLR. 4 Finally the KaplanMeier survival curves, as presented in figure 2, do not correspond to the data in the text. The role of TMLR as a treatment of endstage coronary artery disease remains to be established. Lasers are not magic, but simple physics. They may be helpful to physicians and to the providing industry. More clinical as well as experimental studies are needed, and hasty judgments should be avoided. R Moosdorf Klinikum der Philipps Universität, Klinik für Herzchirurgie, 35044 Marburg, Germany 1
Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularisation in patients with refractory angina: a randomised controlled trial. Lancet 1999; 353: 519–24. 2 Horvath KA, Cohn LH, Cooley DA, et al. Transmyocardial laser revascularization results of a multicenter trial with transmyocardial laser revascularisation used as sole therapy for end-stage coronary artery disease. J Thorac Cardiovasc Surg 1997; 113: 645–53. 3 March RJ. Transmyocardial laser revascularisation with the CO2 laser: one
1705
CORRESPONDENCE
4
5
year results of a randomized, controlled trial. Semin Thorac Cardiovasc Surg 1999; 11: 12–18. Moosdorf R. Transmyocardial laser revascularisation. Z Kardiol 1997; 8 6 (suppl 1): 115–24. Prêtre R, Turina MI. Laser to the heart: magic but costly, or only costly? Lancet 1999; 353: 512–13.
Sir—P M Schofield and colleagues’ 1 study design and statistical analysis of the data make this a valuable contribution to our knowledge on the effectiveness of TMLR. Despite the small number of TMLR patients in this trial (n=94), the results are in broad accordance with those from other large TMLR groups. 2–5 However, what is difficult to comprehend is their view that although they found a highly significant decrease in angina at rest and during treadmill testing 3, 6, and 12 months after TMLR, they concluded that “the adoption of TMLR cannot be advocated”. Like several other study groups, these investigators were unable to demonstrate any proof of an improvement in myocardial perfusion. However, when faced with the poor clinical status of most TMLR patients before the laser procedure (who experience angina at rest despite maximum medical therapy, which results in a severe depression), we as surgeons have to be grateful to be able to alleviate the patient’s complaint with an 80% success rate. Schofield and colleagues, however, are concerned about their perioperative mortality rate of 5%, which indeed is quite high, when compared with other groups. On the other hand, the difference in 1-year mortality between the TMLR group and the medical management group was not significant. Most TMLR users find that with increasing practice and experience the perioperative mortality rate falls continuously. For example, our perioperative mortality rate dropped from 8% in our first 50 TMLR patients to 2% in the last 50 from almost 200 patients we operated on with TMLR at our institution.5 The exclusion criteria for this trial, in our opinion, led to the selection of the more healthy TMLR patients—all patients on intravenous angina therapy and patients who were unable to do treadmill exercise testing because of severe angina were excluded. Thus, exactly the patient group that presumably would have had the greatest benefit from the operation was probably excluded. TMLR is the only surgical treatment option for patients with diffuse, inoperable coronary artery disease. It can reduce the patient’s
1706
complaints substantially for up to 3 years, and can be repeated. I believe that the published study results support these theses, rather than contradicting them. *R Hetzer, T Krabatsch Deutsches Herzzentrum Berlin, D-13353 Berlin, Germany 1
Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularisation in patients with refractory angina: a randomised controlled trial. Lancet 1999; 353: 519–24. 2 Horvath KA, Cohn LH, Cooley DA, et al. Transmyocardial laser revascularization: results of a multicenter trial with transmyocardial laser revascularization used as sole therapy for end-stage coronary artery disease. J Thorac Cardiovasc Surg 1997; 113: 645–54. 3 Burns SM, Sharples LD, Tait S, Caine N, Wallwork J, Schofield PM. The transmyocardial laser revascularization international registry report. Eur Heart J 1999; 20: 31–37. 4 Vincent JG, Bardos P, Kruse J, Maass D. End stage coronary disease treated with the transmyocardial CO2 laser revascularization: a chance for the ‘inoperable’ patient. Eur J Cardiothorac Surg 1997; 11: 888–94. 5 Krabatsch T, Tambeur L, Lieback E, Schäper F, Hetzer R. Transmyocardial laser revascularization in the treatment of end-stage coronary artery disease. Ann Thorac Cardiovasc Surg 1998; 4: 64–71.
Authors' reply Sir—We are delighted with the lively debate generated by the UK TMLR trial results and with the general support we have received. Advocates of TMLR have interpreted the minor improvements in function (40 s additional exercise tolerance from a baseline of 7·5 min) and moderate improvements in angina (25% of survivors with improvement at 1 year), as sufficient to support widespread use of the technology. Others described our interpretation as dispassionate. 1 In a trial of an unmasked surgical procedure in which both patient and clinician have vested interests results must be unequivocal. We chose exercise capacity as the primary outcome since it was important to have an objective method of assessing functional capacity: this included a stress ECG and we preferred stress MIBI to echocardiography, although the two methods of assessing perfusion are similar.2 Maximum oxygen uptake is a measure of heart failure rather than angina and was indeed less appropriate for these patients. Canadian Cardiovascular Society (CCS) angina score was recorded for comparison with other studies and was not regarded as sufficiently reproducible to be a primary endpoint.
John McMurray suggests that the improvements recorded would be acceptable after angioplasty or drug therapy, but these are likely to be less invasive, less risky, and less costly than TMLR. Similarly, the percutaneous approach to myocardial revascularisation may not provide superior results to TMLR, but the lower risk involved may make improvements more acceptable. Compared with TMLR, early results from the PACIFIC trial 3 of PMR suggest slightly greater functional improvement, similar changes in angina score, no deaths in hospital and low morbidity; channels were 6 mm deep rather than the 3–4 mm suggested by Keith Horvath. Well conducted trial evidence is needed and we await the 12 month results of this trial. Horvath should beware of undertaking crude analysis of presented data. Our analysis adjusted for baseline exercise tests and we reported baseline-adjusted means and confidence intervals. The number of sites with irreversible ischaemia was adjusted for baseline, and for repeated within-patient, between-site measures. Horvath and McMurray question hospital readmissions. Although the TMLR group had fewer readmissions for unstable angina (44 v s 68) there were 15 procedure-related readmissions and an excess of cardiacrelated readmissions (19 v s 1 1 ) . Overall the procedure did not reduce the number or length of readmissions in the first year, and the rate of noncardiac related readmissions was very similar for the two groups. There are therefore no savings in readmissions to balance against the costs of the procedure, an issue that will be fully addressed in a future report on costeffectiveness of TMLR. Allan Lansing is correct that patients were required to have areas of reversible ischaemia to enter the trial, but all areas of ischaemia in the ventricle were targeted in the TMLR procedure. Lansing provides the proportion of patients with improvements in angina score in his experience but does not tell us the number of patients and time of followup after the procedure. Four of your correspondents suggest that the greater improvements in CCS noted in other reports may be attributable to patient selection. Sicker patients may have more scope for improvement but will be at greater risk of early mortality and morbidity (US trial perioperative mortality 22% for TMLR within 14 days of unstable angina). 4 It is regrettable that the US trial allowed
THE LANCET • Vol 353 • May 15, 1999
2
Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularisation in patients with refractory angina: a randomised controlled trial. Lancet 1999; 353: 519–24. Prêtre R, Turina MI. Laser to the heart: magic but costly, or only costly? Lancet 1999; 353: 512–13.
Sir—My experience of TMLR over 5 years includes 322 patients with the carbon dioxide laser and 70 with the Holmium-YAG.1 With the CO2 laser, at 1 year thallium studies showed increased vascularity in 55% of patients, 40% had no angina, 31% were class I, and 21% class II, all but six being class IV initially. The discrepancy seems to be due to a difference in patient selection and surgical technique. In their selection the Papworth team 2 insist on areas of reversible ischaemia, whereas we know that many fixed areas on thallium study actually represent muscle that is still viable, especially if the patient has severe pain and the area contracts in catheterisation or echocardiographic studies. These patients respond well to TMLR therapy but at Papworth all such patients were excluded, although they provide some striking results. Another difference in selection is that 73% of the Papworth group had only class III angina, whereas our patients were almost entirely class IV (363 of 369). We believe that benefit is more easily shown in patients who have more severe angina, perhaps because of a more highly developed intramyocardial collateral system available for redistribution of blood flow. Other less important variations include their decision that the patient must be able to carry out a treadmill exercise test, whereas most of our patients were unable to do so because of severe angina, debility, or musculoskeletal problems. They also excluded patients with unstable angina whereas our group includes 108 patients who could not be taken off intravenous medication for over a week without recurrence of chest pain. Finally, only 10% of their patients were female whereas our studies included 25% women. The most important difference, however, is that their report seems to indicate that they treated only areas of reversible ischaemia previously identified by nuclear studies. We learned early that you had to treat the entire ventricle and not only the reversible area. When we confined treatment to this area, we found that the patient still had class IV angina a year later, but re-operation during which the entire ventricle was treated
THE LANCET • Vol 353 • May 15, 1999
resulted in conversion to class I within 3 months. If the whole ventricle and the areas of fixed defect are not treated, we would expect the results to be less striking. Thus, the selection process of patients for the Papworth study was much different from that in our studies, particularly because most patients were class III rather than IV, and Schofield and colleagues insisted on reversible ischaemia by nuclear perfusion as a criterion for selection, excluding those with fixed but viable defects. The most important difference, however, could result from applying TMLR only to areas of demonstrated reversible ischaemia, rather than to the entire ventricle. However, the Papworth study has injected a needed word of caution, especially for the future when multiple types of energy sources come to be used and percutaneous laser revascularisation has become widespread. I received no funds for the project reported here, which was supported entirely by the hospital out of its budget for the Heart Institute. I have acted as a consultant for both PLC Medical Systems and Eclipse Surgical Technology, Inc. For each of these companies I have acted as a teacher to train new teams starting the procedure, and I am now receiving a small honorarium for each such effort. I also have a small number of shares in each company, which I purchased myself, and in each case the value has gone down and I have in fact lost money. Therefore, I do not believe that I have any conflict of interest that would affect the accuracy of my observations.
Allan M Lansing Heart Institute, Norton Audubon Hospital, Louisville, Kentucky, 40217, USA 1
2
Lansing AM. Transmyocardial revascularization: mechanism of action with carbon dioxide and holmium-yttriumaluminium-garnet lasers. J Thorac Cardiovasc Surg 1998; 115: 1392. Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularization in patients with refractory angina. Lancet 1999; 353: 519–224.
Sir—P M Schofield and colleagues1 present the first European randomised trial of TMLR separating the two study groups for the whole observation period without crossover. However, I have some comments. Most patients were in angina class CCS III (73%), which might account for the difference in improvement compared with other studies, with patients mainly in angina class IV. 2,3 Our own experience has also shown that these patients with angina at rest or even unstable angina do benefit most from this procedure and should be the main target group, whereas patients with fewer symptoms might
as well be treated by an improved drug regimen without surgery.4 The evaluation of patients includes few objective data. Two main endpoints are improvement in treadmill test and maximum exercise time, both being symptom limited. More objective data such as stress electrocardiography (ECG) and echocardiography, or even maximum oxygen uptake under exercise conditions, are lacking. The fact that most patients after TMLR reported dyspnoea or fatigue as reasons for stopping these tests might be interpreted as an angina equivalent without chest pain, as assumed by your commentators.5 Most of these patients with endstage coronary artery disease are known to be in poor physical condition and show a limited ventricular performance because of previous and often multiple infarctions, each of which would also be a sufficient explanation for dyspnoea and fatigue. Without ECG recordings to establish the true ischaemic threshold, an interpretation of these data is hardly possible. The only objective data are those obtained by nuclear scans and they do not show any significant difference between the two groups. However, the investigators themselves cast doubt on whether this method provides sufficiently high resolution. Our own data for Sestamibi scans, obtained by an independent investigator, have shown an improvement in ischaemic segments in around 40% of patients after TMLR. 4 Finally the KaplanMeier survival curves, as presented in figure 2, do not correspond to the data in the text. The role of TMLR as a treatment of endstage coronary artery disease remains to be established. Lasers are not magic, but simple physics. They may be helpful to physicians and to the providing industry. More clinical as well as experimental studies are needed, and hasty judgments should be avoided. R Moosdorf Klinikum der Philipps Universität, Klinik für Herzchirurgie, 35044 Marburg, Germany 1
Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularisation in patients with refractory angina: a randomised controlled trial. Lancet 1999; 353: 519–24. 2 Horvath KA, Cohn LH, Cooley DA, et al. Transmyocardial laser revascularization results of a multicenter trial with transmyocardial laser revascularisation used as sole therapy for end-stage coronary artery disease. J Thorac Cardiovasc Surg 1997; 113: 645–53. 3 March RJ. Transmyocardial laser revascularisation with the CO2 laser: one
1705
CORRESPONDENCE
4
5
year results of a randomized, controlled trial. Semin Thorac Cardiovasc Surg 1999; 11: 12–18. Moosdorf R. Transmyocardial laser revascularisation. Z Kardiol 1997; 8 6 (suppl 1): 115–24. Prêtre R, Turina MI. Laser to the heart: magic but costly, or only costly? Lancet 1999; 353: 512–13.
Sir—P M Schofield and colleagues’ 1 study design and statistical analysis of the data make this a valuable contribution to our knowledge on the effectiveness of TMLR. Despite the small number of TMLR patients in this trial (n=94), the results are in broad accordance with those from other large TMLR groups. 2–5 However, what is difficult to comprehend is their view that although they found a highly significant decrease in angina at rest and during treadmill testing 3, 6, and 12 months after TMLR, they concluded that “the adoption of TMLR cannot be advocated”. Like several other study groups, these investigators were unable to demonstrate any proof of an improvement in myocardial perfusion. However, when faced with the poor clinical status of most TMLR patients before the laser procedure (who experience angina at rest despite maximum medical therapy, which results in a severe depression), we as surgeons have to be grateful to be able to alleviate the patient’s complaint with an 80% success rate. Schofield and colleagues, however, are concerned about their perioperative mortality rate of 5%, which indeed is quite high, when compared with other groups. On the other hand, the difference in 1-year mortality between the TMLR group and the medical management group was not significant. Most TMLR users find that with increasing practice and experience the perioperative mortality rate falls continuously. For example, our perioperative mortality rate dropped from 8% in our first 50 TMLR patients to 2% in the last 50 from almost 200 patients we operated on with TMLR at our institution.5 The exclusion criteria for this trial, in our opinion, led to the selection of the more healthy TMLR patients—all patients on intravenous angina therapy and patients who were unable to do treadmill exercise testing because of severe angina were excluded. Thus, exactly the patient group that presumably would have had the greatest benefit from the operation was probably excluded. TMLR is the only surgical treatment option for patients with diffuse, inoperable coronary artery disease. It can reduce the patient’s
1706
complaints substantially for up to 3 years, and can be repeated. I believe that the published study results support these theses, rather than contradicting them. *R Hetzer, T Krabatsch Deutsches Herzzentrum Berlin, D-13353 Berlin, Germany 1
Schofield PM, Sharples LD, Caine N, et al. Transmyocardial laser revascularisation in patients with refractory angina: a randomised controlled trial. Lancet 1999; 353: 519–24. 2 Horvath KA, Cohn LH, Cooley DA, et al. Transmyocardial laser revascularization: results of a multicenter trial with transmyocardial laser revascularization used as sole therapy for end-stage coronary artery disease. J Thorac Cardiovasc Surg 1997; 113: 645–54. 3 Burns SM, Sharples LD, Tait S, Caine N, Wallwork J, Schofield PM. The transmyocardial laser revascularization international registry report. Eur Heart J 1999; 20: 31–37. 4 Vincent JG, Bardos P, Kruse J, Maass D. End stage coronary disease treated with the transmyocardial CO2 laser revascularization: a chance for the ‘inoperable’ patient. Eur J Cardiothorac Surg 1997; 11: 888–94. 5 Krabatsch T, Tambeur L, Lieback E, Schäper F, Hetzer R. Transmyocardial laser revascularization in the treatment of end-stage coronary artery disease. Ann Thorac Cardiovasc Surg 1998; 4: 64–71.
Authors' reply Sir—We are delighted with the lively debate generated by the UK TMLR trial results and with the general support we have received. Advocates of TMLR have interpreted the minor improvements in function (40 s additional exercise tolerance from a baseline of 7·5 min) and moderate improvements in angina (25% of survivors with improvement at 1 year), as sufficient to support widespread use of the technology. Others described our interpretation as dispassionate. 1 In a trial of an unmasked surgical procedure in which both patient and clinician have vested interests results must be unequivocal. We chose exercise capacity as the primary outcome since it was important to have an objective method of assessing functional capacity: this included a stress ECG and we preferred stress MIBI to echocardiography, although the two methods of assessing perfusion are similar.2 Maximum oxygen uptake is a measure of heart failure rather than angina and was indeed less appropriate for these patients. Canadian Cardiovascular Society (CCS) angina score was recorded for comparison with other studies and was not regarded as sufficiently reproducible to be a primary endpoint.
John McMurray suggests that the improvements recorded would be acceptable after angioplasty or drug therapy, but these are likely to be less invasive, less risky, and less costly than TMLR. Similarly, the percutaneous approach to myocardial revascularisation may not provide superior results to TMLR, but the lower risk involved may make improvements more acceptable. Compared with TMLR, early results from the PACIFIC trial 3 of PMR suggest slightly greater functional improvement, similar changes in angina score, no deaths in hospital and low morbidity; channels were 6 mm deep rather than the 3–4 mm suggested by Keith Horvath. Well conducted trial evidence is needed and we await the 12 month results of this trial. Horvath should beware of undertaking crude analysis of presented data. Our analysis adjusted for baseline exercise tests and we reported baseline-adjusted means and confidence intervals. The number of sites with irreversible ischaemia was adjusted for baseline, and for repeated within-patient, between-site measures. Horvath and McMurray question hospital readmissions. Although the TMLR group had fewer readmissions for unstable angina (44 v s 68) there were 15 procedure-related readmissions and an excess of cardiacrelated readmissions (19 v s 1 1 ) . Overall the procedure did not reduce the number or length of readmissions in the first year, and the rate of noncardiac related readmissions was very similar for the two groups. There are therefore no savings in readmissions to balance against the costs of the procedure, an issue that will be fully addressed in a future report on costeffectiveness of TMLR. Allan Lansing is correct that patients were required to have areas of reversible ischaemia to enter the trial, but all areas of ischaemia in the ventricle were targeted in the TMLR procedure. Lansing provides the proportion of patients with improvements in angina score in his experience but does not tell us the number of patients and time of followup after the procedure. Four of your correspondents suggest that the greater improvements in CCS noted in other reports may be attributable to patient selection. Sicker patients may have more scope for improvement but will be at greater risk of early mortality and morbidity (US trial perioperative mortality 22% for TMLR within 14 days of unstable angina). 4 It is regrettable that the US trial allowed
THE LANCET • Vol 353 • May 15, 1999