(Circulation. 1997;96:2121-2123.)
© 1997 American Heart Association, Inc.
Articles |
Is It Important How One Dies?
Questions for Planning Future Revascularization Trials
From Andreas Gruentzig Cardiovascular Center, Emory University, Atlanta, Ga.
Correspondence to Spencer B. King III, MD, Andreas Gruentzig Cardiovascular Center, Emory University, 1364 Clifton Rd, Suite F606, Atlanta, GA 30322.
Key Words: Editorials • angioplasty • revascularization • myocardial infarction • mortality
 |
Introduction |
|---|
 Top Introduction References |
|---|
The randomized trials of angioplasty versus coronary bypass surgery conducted to date have not shown a difference in survival, either individually or collectively.1 2 3 Although total mortality in the largest trial, BARI, remains nonsignificant at 5 years, for the first time a difference in cardiac mortality has emerged. In this issue of Circulation, the BARI investigators report excess mortality in the angioplasty randomized cohort (8.0% versus 4.9% for the surgical group, for a relative risk of 1.55; P=.022).4 The authors also report the occurrence of MI and speculate on the impact of MI on cardiac mortality.
Clearly, the most striking finding in this study is the superiority of freedom from cardiac death in the surgery group without excess noncardiac death. Further analysis of this landmark study shows that the difference in cardiac deaths is completely explained by an excess of cardiac deaths in the diabetic patients in the trial. There are more than three times as many cardiac deaths in the diabetic patients treated with angioplasty as in those patients treated with surgery. The present study concentrates on the 81% of BARI patients who did not have treated diabetes in whom no difference in cardiac mortality rate was found (4.6% for PTCA versus 4.2% for CABG; P=.09).
Two questions regarding cardiac mortality come to mind. Should cardiac mortality replace total mortality as the primary end point in comparative revascularization trials? What is the explanation for the high cardiac mortality rate in the diabetic patients? The question of whether it is important how one dies is not one for patients. Their concern is for longevity and quality of life while they are alive. Perhaps patients would choose a cardiac death over a death from cancer. No, the real reason to explore cardiac death as a primary discriminator between the therapies is that it is the only mode of death that is likely to be influenced by the therapy. As yet, there is no evidence that noncardiac death is found in excess in either treatment group, nor is there any plausible reason why it should be. When choosing a therapy to reduce the risk of future death, the therapy should have some potential to impact the mechanisms by which death would occur. In BARI, which excluded most high-surgical-risk patients, there was no difference in early cardiac mortality, but the significant difference found at 5 years was probably also present at 2 or 3 years, judging from Fig 1 of the present report. Much has already been said about the ability of surgery to provide more complete revascularization than angioplasty. Restenosis after angioplasty is, by consensus, more of a problem in the 5-year time window than is graft closure (80% of BARI patients received an internal mammary artery graft).
Why is cardiac death higher in the angioplasty group? The authors have pointed out the fact that the cardiac mortality rates in the patients without treated diabetes are nearly identical between the angioplasty and surgery groups, and therefore all the excess mortality is in diabetic patients. Diabetic patients having angioplasty or surgery face a significantly higher mortality risk than nondiabetic patients.5 6 Total mortality rate among diabetics was shown to be higher in the BARI PTCA patients (34.5% versus 19.4%),1 and this difference (20 excess deaths in the PTCA group) was fully explained in that report by the excess cardiac mortality in the angioplasty group. Since there was no difference in procedural mortality rate, what are the possible explanations for the late cardiac mortality in the diabetic patients? Completeness of revascularization at the initial procedure was lower in the angioplasty patients treated in BARI and in EAST.1 7 There is more restenosis in diabetics.8 9 In addition, diabetic patients continue to have higher cardiac event rates even after the restenosis process has run its course, suggesting continuing progression of disease in these patients.10 11 In addition to large-vessel disease, diabetics have myocardial and microvascular abnormalities that may hinder long-term survival.12
Interestingly, this excess mortality in treated diabetics undergoing angioplasty was also found in the European revascularization study, CABRI,13 but not in EAST. The diabetic patients in EAST differed from those in BARI in ways that may have relevance. The BARI diabetic patients had more prior MIs, worse left ventricular function, and more total occlusions at baseline than the EAST patients. In addition, the follow-up of EAST patients involved protocol thallium scans and angiograms at 1 and 3 years after the index procedure. Whether these follow-up methods used in a single-center, private hospital setting helped to contribute to the equal 90% 5-year survival rates in the EAST diabetic and nondiabetic patients or whether this is a fluke of a small sample size (n=59) is unclear. Even though the EAST diabetic PTCA patients had a favorable survival rate, two thirds underwent subsequent revascularization procedures during the 5-year follow-up. Careful glycemic control, vigorous use of lipid-lowering agents, and careful surveillance for silent ischemia are potential areas for reducing the late cardiac mortality seen in diabetic patients.
The second item explored in the present report is the occurrence of MI and its influence on cardiac mortality. When the BARI study was in its planning phase, it was recognized that CK elevation in the postoperative period would be common in most surgical patients. Because evaluation of MI by enzyme criteria was impossible for the surgical patients during the early postoperative period, the identification of MI was limited to those with Q waves in both groups. The present study clearly provides the most standardized and validated core laboratory analysis of Q-wave MI of all the trials. A two-step change in the Minnesota code was required, and this fairly stringent criterion resulted in a somewhat lower MI rate than in some other trials.
Should MI be used as an end point in revascularization trials? There are two reasons why MI could be a valuable outcome end point. First, if MI results in hospitalization, prolongation of stay, increased resource utilization, or altered quality of life, then this is an outcome that is interesting, as is repeat revascularization. In that way, however, it is not to be equated with death. The use of death or MI as a prominently reported outcome of trials14 can be misleading. The recently reported antiplatelet trials have a very low death rate that is not different between treatment groups but CK elevations that are significantly different, giving the reporting of excess death or MI a more ominous ring than it deserves.
The other reason to use MI as an end point is that it may be a predictor of future cardiac mortality. Although the correlation between MI and cardiac death seems surprisingly strong in the present report (a risk ratio of 5.9 for cardiac death in patients with MI; see Table 3), this is not an indicator of the prognostic value of MI. The reason this is of limited value is that MI, which is the terminal event, is counted as a death and as a MI. MI as a mode of death cannot be used as a predictor of future mortality because for that patient, there is no future risk. On the other hand, 55 of the MIs were silent and were found on routine follow-up ECGs. By definition, these patients survived their MI, and therefore the association with cardiac mortality is not strong in this group.
When death and MI were combined in the present study, the 121 MIs in the surgery group and the 137 in the angioplasty group overwhelmed the cardiac mortality, making the difference in the combined end point of cardiac death or MI nonsignificant. Because cardiac death and MI are closely related, there is not a great amount of clinically relevant information learned from this carefully adjudicated end point.
Another use of MI as an outcome measure could have been more interesting: what is the prognostic value of a nonfatal MI on future cardiac mortality? Q-wave MI identified in the BARI trial should carry a greater risk than less stringently defined MIs. In the EAST trial,15 identification of a postsurgical MI, defined by an expert panel as the appearance of "new pathologic Q waves," did not predict future mortality and in many cases was compatible with a normal left ventriculogram in late follow-up. But what of the impact of non–Q-wave MIs occurring after surgery or PTCA? This question should be explored but will remain largely unanswered because of the exclusion of CK determinations during the 96-hour period after procedures. Although this was perhaps a fair exclusion for the purpose of comparing procedures, it precludes the evaluation of postprocedure CK elevation on future mortality. A recent meta-analysis16 focused on postintervention CK values and concluded that elevations in the range of 3x normal were predictive of future mortality. Other randomized trials, such as the Balloon versus Optimal Atherectomy Trial,17 however, challenge this conclusion. Perhaps in future trials, the opportunity to capture such data will not be missed.
The present report should be viewed in several ways. First, it challenges the concept that angioplasty is as safe as surgery for patients with multivessel disease suitable for either PTCA or CABG who do not have markers for high surgical risk. Even though the total mortality remained not significantly different at 5 years, the excess cardiac mortality in the angioplasty group suggests that this procedure may not be as protective as surgery. It would be inappropriate, however, not to mention that this difference was found only in patients with diabetes, a variable that although not defined as a subanalysis target in the protocol, was added early in the trial and followed up prospectively. Therefore, patients without treated diabetes have a cardiac mortality rate and total mortality rate that are almost identical between revascularization strategies, and thus selection of therapy in these patients should be guided by patient wishes and physician judgment. Next, the present report and the more in-depth report of diabetic patients18 do not establish that all diabetic patients who need revascularization should have surgery. Registry diabetic patients in whom the revascularization strategy was selected by physician and patient judgment did not have the same excess risk as the randomized patients.
Are there clues that would suggest methods for improving the outcome for diabetic patients undergoing interventions? We have mentioned the value of added surveillance for ongoing ischemia in potentially underrevascularized patients. In addition, advances in medical therapy of diabetes and lipid abnormalities, as well as newer interventional approaches such as stenting, may improve the outcomes in diabetic patients. Currently, two studies of stenting versus surgery are being conducted in Europe, and discussions of targeted studies in diabetic patients are ongoing in this country. Until the outcome of these new approaches is understood, a cautious approach toward recommending interventions in treated diabetic patients with multivessel disease should be pursued.
Could future trials better target the difference in outcomes to reflect those most likely to be impacted by the therapy? Should the primary end point be the difference in cardiac mortality rate? For all the reasons enumerated, I think so.
|
Selected Abbreviations and Acronyms |
|---|
|
|
Footnotes |
|---|
The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association.
|
References |
|---|
|
TopIntroductionReferences |
|---|
1. The Bypass Angioplasty Revascularization Investigation (BARI) Investigators. Comparison of coronary bypass surgery with angioplasty in patients with multivessel disease. N Engl J Med. 1996;335:217-225.
2.
King SB III, Lembo NJ, Weintraub WS, Kosinski AS,
Barnhart HX, Kutner MH, Alazraki NP, Guyton RA, Zhao XQ. A
randomized trial comparing coronary angioplasty with
coronary bypass surgery: Emory Angioplasty versus Surgery Trial
(EAST). N Engl J Med. 1994;331:1044-1050.
3. Pocock SJ, Henderson RA, Rickards AF, Hampton JR, King SB III, Hamm CW, Puel J, Hueb W, Goy J-J, Rodrigues A. Meta-analysis of randomised trials comparing coronary angioplasty with bypass surgery. Lancet. 1995;346:1184-1189.[Medline] [Order article via Infotrieve]
4.
Chaitman BR, Rosen AD, Williams DO, Bourassa MG,
Aguirre FV, Pitt B, Rautaharju PM, Rogers WJ, Sharaf B, Attubato M,
Hardison RM, Srivasta S, Kouchoukos NT, Stocke K, Sopko G, Detre K,
Frye R, for the BARI Investigators. Cardiac mortality and myocardial
infarction in the Bypass Angioplasty
Revascularization Investigation (BARI) randomized
trial. Circulation. 1997;96:2162-2170.
5.
Stein B, Weintraub WS, Gebhart SP, Cohen-Bernstein CL,
Grosswald R, Liberman HA, Douglas JS Jr, Morris DC, King SB III.
Influence of diabetes mellitus on early and late outcome after
percutaneous transluminal coronary
angioplasty. Circulation. 1995;91:979-989.
6. Salomon NW, Page US, Okies EJ, Stephens J, Krause AH, Bigelow JC. Diabetes mellitus and coronary artery bypass. J Thorac Cardiovasc Surg. 1983;85:264-271.[Abstract]
7.
Zhao X-Q, Brown BG, Stewart DK, Hillger LA, Barnhart
HX, Kosinski AS, Weintraub WS, King SB III. Effectiveness of
revascularization in the Emory Angioplasty versus
Surgery Trial: a randomized comparison of coronary angioplasty
with bypass surgery. Circulation. 1996;93:1954-1962.
8.
Detre K, Holubkov R, Kelsey S, Bourassa M, Williams D,
Holmes D Jr, Dorros G, Faxon D, Myler R, Kent K, Cowley M, Cannon R,
Robertson T. One-year follow-up results of the 1985-1986 National
Heart, Lung, and Blood Institute's Percutaneous
Transluminal Coronary Angioplasty Registry.
Circulation. 1989;80:421-428.
9. Weintraub WS, Kosinski AS, Brown CL III, King SB III. Can restenosis after coronary angioplasty be predicted from clinical variables? J Am Coll Cardiol. 1993;21:6-14.[Abstract]
10.
Weintraub WS, Ghazzal ZMB, Cohen CL, Douglas JS Jr,
Liberman HA, Morris DC, King SB III. Clinical implications of
late proven patency after successful coronary
angioplasty. Circulation. 1991;84:572-582.
11.
Ellis SG, Cowley MJ, DiSciascio G, Deligonul U, Topol
EJ, Bulle TM, Vandormael MG. Determinants of 2-year outcome
after coronary angioplasty in patients with multivessel disease
on the basis of comprehensive preprocedural evaluation: implications
for patient selection. Circulation. 1991;83:1905-1914.
12.
Nahser PJ, Borwn RE, Oskarsson H, Winniford MD, Rossen
JD. Maximal coronary flow reserve and
metabolic coronary vasodilation in patients with
diabetes mellitus. Circulation. 1995;91:635-640.
13. CABRI Investigators. Long-term follow-up of European revascularization trials. Presented at the 68th Scientific Sessions of the American Heart Association (Plenary Session XII); November 16, 1995; Anaheim, Calif.
14. Lincoff AM, Tcheng JB, Califf RM, Cabot CF, Miller DP, Booth JE, Montague EA, Anderson KM, Topol EJ. Abciximab (c7E3 GP IIB/IIIA Fab, ReoPro) with reduced heparin dosing during coronary intervention: final results of the EPILOG trial. J Am Coll Cardiol. 1997;29:187A. Abstract.
15.
Hodakowski GT, Craver JM, Jones EL, King SB, Guyton
RA. Clinical significance of perioperative Q
wave myocardial infarction in the Emory Angioplasty versus Surgery
Trial. J Thorac Cardiovasc Surg. 1997;112:1447-1453.
16. Califf RM, Abdelmeguid A, Kuntz R, Popma JJ, Tardiff BE, Crenshaw B, Bauman RP, Davidson CJ, Cohen EA, Kleiman NS, Mahaffey KW, Topol EJ, Pepine CJ, Lipicky R, Granger CB, Harrington RA, Zuckerman B, Chaitman BR, Bittl JA, Ohman EM. Myonecrosis after revascularization procedures. J Am Coll Cardiol. In press.
17. Cutlip DE, Baim DS, Senerchia C, DeFeo-Fraulini T, Ho KKL, Chauhan N-N, Kuntz RE. Clinical consequences of myocardial infarction following balloon angioplasty or directional coronary atherectomy: acute and one year results of the Balloon vs. Optimal Atherectomy Trial (BOAT). J Am Coll Cardiol. 1997;29:187A-188A. Abstract.
18. BARI Investigators. The influence of diabetes on 5-year mortality and morbidity in a randomized trial comparing CABG and PTCA in patients with multivessel disease: the Bypass Angioplasty Revascularization Investigation (BARI). Circulation. In press.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||