Author + information
- Eric D. Peterson, MD, MPH, FACC*,* (, )
- Padma Kaul, PhD*,†,
- Ronald G. Kaczmarek, MD‡,
- Bradley G. Hammill, BA, MS*,
- Paul W. Armstrong, MD, FACC†,
- Charles R. Bridges, MD, ScD, FACC§,
- T.Bruce Ferguson Jr, MD∥,
- Society of Thoracic Surgeons
- ↵*Reprint requests and correspondence:
Dr. Eric D. Peterson, Duke University Medical Center, Box 3236, Durham, North Carolina 27710, USA.
Objectives We sought to examine trends in the use and outcomes of transmyocardial revascularization (TMR) in community practice. We also identified important risk factors for TMR and compared outcomes of TMR combined with coronary artery bypass graft surgery (TMR + CABG) versus bypass alone in patients receiving incomplete revascularization.
Background Although it is approved for use as a stand-alone procedure, there are limited data on the outcomes of (TMR + CABG).
Methods We identified 3,717 patients receiving TMR at 173 U.S. hospitals participating in the Society of Thoracic Surgeons (STS) National Cardiac Database. Baseline characteristics and outcomes in these patients were compared with those from six published randomized TMR trials. Multivariable logistic regression was used to identify clinical risk factors for mortality with TMR. Risk-adjusted mortality was also compared for TMR + CABG relative to CABG only in patients not amenable to complete traditional revascularization.
Results Between January 1998 and December 2001, the number of STS hospitals performing TMR and total procedural counts increased markedly, driven predominately by more TMR + CABG cases. Overall mortality rates for TMR-alone and TMR + CABG were 6.4% and 4.2%, respectively. Operative risks were significantly higher in those patients with recent myocardial infarction, unstable angina, and depressed ventricular function. Among patients receiving incomplete revascularization, TMR + CABG was not associated with decreased mortality risk compared with CABG alone, adjusted odds ratio 1.11 (95% confidence interval 0.74 to 1.67).
Conclusions The use of TMR, and in particular, TMR + CABG, is expanding in community practice. Although procedural risks are high, there is room for optimization through improved patient selection and timing of the procedure. Further studies of TMR + CABG are needed given its growing use and unclear benefits.
Transmyocardial revascularization (TMR) is a recently developed alternative treatment for patients with advanced coronary disease whose anatomy is not amenable to percutaneous coronary intervention or conventional coronary artery bypass graft surgery (CABG). During TMR, small “channels” or holes are created in the heart via a laser device applied to the myocardial surface. This procedure can be performed as sole therapy (TMR-only) or in combination with bypass surgery (TMR + CABG). Although the mechanism(s) by which TMR may work remain unclear (1–7), each of six small randomized studies demonstrated that TMR-only significantly reduced angina symptoms relative to continued medical care, but did not reduce acute or long-term mortality (8–13). An additional randomized trial compared TMR + CABG versus CABG alone in patients not amenable to complete revascularization. This trial could not demonstrate an incremental benefit of TMR on anginal status, yet it did report lower operative mortality with the combined procedure than CABG alone (14).
Although these randomized trials demonstrated the potential benefits of TMR, they also documented the procedure's risks. Overall procedural mortality rates in these trials ranged from 1% to 5%. The TMR risks were particularly high among patients with unstable angina, recent cardiac events, or depressed left ventricular function (9,10). Weighing the risks and benefits of TMR-only, the Food and Drug Administration (FDA) gave market approval for its use in patients with stable angina refractory to medical therapy and lacking traditional revascularization options. This approval for TMR-only, however, came with a requirement for the laser's manufacturers to undertake large post-marketing safety studies. Given a paucity of information, manufacturers have not yet sought an FDA indication for TMR + CABG, and its use in practice remains “off-label.”
In addition to these manufacturers' studies, the FDA, in collaboration with the Society of Thoracic Surgeons (STS) and the Duke Clinical Research Institute (DCRI), initiated an independent evaluation of the use and safety of TMR in community practice. This study reviews trends in the use of TMR-only and TMR + CABG between 1998 and 2002 among more than 400 hospitals participating in the STS National Cardiac Database. We also examined the acute operative mortality and major morbidity following TMR and determined how these risks varied as a function of patient characteristics. Finally, we investigated whether TMR + CABG was associated with fewer procedural complications relative to traditional CABG in patients not amenable to complete revascularization.
Clinical data for the STS Database were collected using methods described more fully elsewhere (15). Briefly, the database was developed in 1990 as a multicenter clinical repository for quality improvement and clinical research. The STS Database collects data from approximately two-thirds of all U.S. cardiothoracic hospitals and contains detailed data on patient demographics, clinical profile, and acute outcomes on more than 2.1 million procedures. Data definitions are standardized and site data coordinators receive initial and ongoing training in these definitions. Sites voluntarily submit data to the data coordinating center (DCRI) on a semi-annual basis and receive site-specific feedback, bench-marked against regional and national results. Accuracy of individual data elements have been validated in regional analyses, and overall completeness of mortality event reporting in patients age 65 years or older has been validated against national Medicare claims files (16).
The study population consisted of all patients in the STS Database receiving TMR between January 1998 and December 2001. Procedures were further subdivided as TMR-only, TMR + CABG, or TMR-other (TMR used in conjunction with other surgical procedures such as valve surgery or ventricular-septal defect repair). The TMR-other cases were included in the trend analysis, but excluded from the outcomes analyses given their higher procedural risks.
A literature search was conducted to identify TMR clinical trials published between 1995 and 2001. This search was supplemented by reviewing study bibliographies and review articles, and direct discussions with all study lead authors. Studies were included if they randomized patients to receive TMR (alone or combined with other procedures). Percutaneous myocardial revascularization studies were not included.
Categorical variables were presented as percentages whereas continuous variables are presented as median and interquartile ranges unless otherwise stated. Chi-square and Kruskall-Wallis tests were used to compare characteristics across groups of patients for categorical and continuous variables, respectively.
The primary outcome of interest was operative mortality, defined as any in-hospital death or death within 30 days of the procedure. In addition, morbidity end points examined included reoperation for any reason, deep sternal wound infection, renal failure, prolonged ventilation, or permanent stroke. Outcomes of TMR in the STS database were compared by procedure type to those seen in the randomized TMR trials. Five of the six TMR trials (8–12)provided complete 30-day mortality results. These outcomes were summated as a random effect estimate using FASTPRO.
Logistic regression was used to identify univariate and multivariate preoperative risk factors for mortality and mortality/morbidity outcomes following TMR. Candidate variables for this analysis came from known predictors of bypass mortality (17)and from those identified from TMR clinical trials. In later analyses, we added procedure type (TMR-only vs. TMR + CABG), temporal trends in outcomes (year of procedure), and TMR hospital volume beyond our clinical factors to our multivariable models.
Finally, we compared the crude and risk-adjusted outcomes of patients with triple-vessel coronary disease who received one to two bypass grafts (that is, incomplete revascularization) with those of similar patients receiving TMR + CABG. These observational treatment comparisons were adjusted for 28 preoperative risk factors associated with CABG mortality or mortality + morbidity (17).
Between January 1998 and December 2001, a total of 3,717 TMR procedures were performed at 173 (31%) of STS National Cardiac Database sites. Of these procedures, 661 (17%) were TMR-only procedures, 2,475 (67%) were TMR + CABG, and 581 (16%) were TMR in combination with some other cardiac procedure. The overall number of sites performing TMR increased from 33 (7% of total STS sites) in 1998 to 131 (36% of sites) in 2001. Similarly, the annual volume of TMR procedures rose from 59 procedures in the first half of 1998 to 572 in the second half of 2001, driven predominately by the expansion in the use of TMR + CABG. The volume of TMR procedures per site ranged from 1 to 150 procedures during the study period, with a median of 12 and an interquartile range of 4 to 27.
Comparison of patient characteristics and outcomes
Table 1provides a comparison of composite baseline patient characteristics from patients enrolled in the randomized TMR trials and those from the STS TMR population. Overall, patients enrolled in the TMR trials had similar demographic and clinical characteristics as those receiving TMR in community practice. Patients receiving TMR had a mean age ranging from 61 to 65 years. Women accounted for one-third of all patients. Comparing the procedure types, patients who underwent TMR + CABG in either the trials or in STS tended to be slightly older, but were markedly less likely to have prior bypass surgery relative to those receiving TMR-only.
Table 2provides STS TMR-only and TMR + CABG mortality, morbidity and mortality or major morbidity end points. The overall STS operative mortality for TMR-only patients was 6.4% (95% confidence interval [CI] 4.8 to 8.7). This tended to be higher than the 30-day mortality rates seen in the individual randomized clinical trials (RCTs), ranging from 1.1% (95% CI 0.0 to 5.9) to 5.3% (95% CI 2.2 to 10.6) or composite average mortality rate for the RCTs 3.5% (95% CI 1.8 to 5.2). Similarly, mortality for the STS TMR + CABG patients was 4.2% (95% CI 3.5% to 5.1%) versus a mortality rate of 1.5% (95% CI 0 to 3) in the single RCT of TMR + CABG (Fig. 1).
The most frequent TMR operative complication was prolonged ventilatory support >24 h required by 7.7% and 9.1% of STS TMR-only and TMR + CABG patients, respectively (Table 2). Renal failure also occurred in nearly 5% of both TMR-only and TMR + CABG patients.
The frequency of TMR postoperative mortality or complications was significantly higher among older patients, in diabetes, and in those patients with recent myocardial infarction (MI) or unstable angina. In contrast, patient gender and degree of underlying coronary disease severity was not associated with increased operative risk. When used in patients without recent MI, unstable angina, and an ejection fraction of >30%, TMR-only (n = 243 patients) mortality rates declined to 3.7% and TMR + CABG (n = 1,067) fell to 2.6%.
Multivariable predictors of TMR risk
Figure 2displays the independent patient risk factors associated with TMR operative mortality. Increasing age, diabetes, unstable angina, peripheral vascular disease, prior CABG, lower ejection fraction, and MI within 21 days of procedure were independently associated with an increased risk of operative mortality. The C-index for this multivariable model was 0.75.
After adjusting for baseline clinical risk, neither patient gender nor procedure type (TMR-only versus TMR + CABG) was associated with increased procedural mortality. Additionally, there were no significant temporal trends in risk-adjusted TMR outcomes over the four years of study (p = 0.31). In contrast, there was a trend towards better TMR outcomes with increasing procedural volume. Low TMR volume sites (1 to 6 procedures) had a mortality rate of 6.2%, versus 4.6% among intermediate sites (7 to 18 procedures) and 4.3% among high-volume sites (>18 procedures) (p = 0.09 for trend).
Comparison of TMR + CABG versus CABG-only
A total of 39,454 STS CABG patients had triple-vessel coronary disease, but received <3 bypass grafts, presumably because of technical inability to graft those regions of the myocardium. Of these patients, 390 underwent a combined TMR + CABG procedure. Compared with CABG-only patients, patients who underwent TMR + CABG had similar to higher unadjusted incidence of acute adverse events (mortality: 4.9% vs. 4.1%, p = 0.37; mortality/morbidity: 18.1% vs. 15.5%, p = 0.09). TMR + CABG was also not associated with lower operative risk after adjusting baseline clinical factors; adjusted TMR + CABG OR for mortality = 1.11 (95% CI 0.74 to 1.67) and for mortality and morbidity 1.13 (95% CI 0.89 to 1.42).
To date, this study is the largest to evaluate TMR use and outcomes. Using the STS National Cardiac Database, we found that the adoption of TMR into U.S. clinical practice has been rapid, with nearly a one-third of centers adopting the procedure within a four-year period. The rise in the procedure's popularity has been driven predominately by increased use of TMR + CABG, an off-label application of the technology. Finally, the operative risks of TMR-only and TMR + CABG in community practice remain higher than those seen with traditional revascularization procedures.
The application of TMR into clinical practice has been rapid and potentially driven by multiple factors. First, TMR represents one of the few existing treatment options for patients with medically refractory angina who are not candidates for traditional revascularization. In these patients, randomized (but unblinded) clinical trials have demonstrated that TMR provided angina relief relative to medical therapy. Finally, TMR requires little additional technical training and commands professional fees similar to those provided for traditional bypass surgery.
The current large series provides clinicians and patients with estimates of mortality and morbidity associated with TMR procedures in contemporary community practice. We found that the operative mortality and morbidity risk of TMR tended to be higher than those seen in prior randomized trial populations. These results are similar to those seen in another international TMR registry (18). These higher risks in community practice may be due to the use of TMR in a less select patient population outside of clinical trials and/or the differential provider effects outside of specialized medical centers. The impact of the latter is suggested by a trend towards lower risks at centers performing higher annual TMR caseloads.
Our study also clearly demonstrates that operative risks for TMR vary markedly as a function of specific patient subgroups (Table 3, Fig. 2). These identified risk factors as similar to those found in prior studies (19,20). However, despite these prior works, our study found that many received TMR in community practice under less than ideal conditions. For example, up to 50% of those receiving TMR were classified as having unstable angina and up to a one-quarter had an MI within 21 days of their TMR procedure. Patients with these unstable symptoms had mortality rates nearly twice as high as those without. On the other hand, among patients with a stable clinical course and preserved ejection fraction (>30%), the procedural risk of TMR-only and TMR + CABG mortality risks was nearly halved. These findings emphasize the need to improve algorithms for TMR patient selection and timing of procedures.
Beyond these risk factors, the size of the STS registry allowed us a unique opportunity to explore the outcomes of TMR in women. Fewer than 200 women had been enrolled in the six prior clinical trials combined. In our registry of nearly 1,000 women receiving TMR, we did not find evidence for gender-specific risks for mortality or major morbidity following TMR.
Importantly, our study revealed that TMR + CABG accounted for the majority of procedures currently performed, yet there is limited information regarding the safety and efficacy of TMR in this setting. In a single randomized study of 266 patients not amenable to complete revascularization, Allen and colleagues reported that TMR + CABG significantly reduced in-hospital mortality rates relative to CABG only (1.5% vs. 7.6%, p = 0.02) (14). Although these results were promising, observational comparison of TMR + CABG in a single institution study (n = 169) and in our larger STS analysis failed to confirm these acute survival benefits (Table 4).
Although we used this study to address national trends in TMR use and outcomes, our data were limited to those available in the STS database. However, this data source represents nearly two-thirds of U.S. operative centers. Second, there are currently two laser devices approved by the FDA for the TMR procedure: the CO2laser and the Ho:YAG laser. As the STS Database does not collect device-specific results, we cannot comment on whether results differ by TMR device type. Third, participation in the STS Database is voluntary without full external clinical event validation. Any under-reporting of events would make our risk estimates conservative. Finally, our database was observational and comparisons of specific treatments, even when risk-adjusted, represent hypothesis generation rather than hypothesis testing.
The STS National Database offered a prime opportunity to track use and results of a new technology in the “real world.” Our finding of the widespread use of TMR + CABG indicates a strong need for further studies of this procedure. Additionally, our study allows for the estimation of patient-specific risks with TMR. Understanding these risks is key to the informed decision process. Although prior randomized trials have demonstrated the symptomatic benefits from TMR, these must be weighted against the upfront procedural risks.
Finally, the results of this study point out the potential for a reduction in TMR morbidity and mortality through optimization of the timing of the procedure. Up to one in two patients currently receiving TMR have the procedure following a recent cardiac event or had unstable symptoms. As previously published studies and the current study document the relatively adverse events associated with TMR in these settings, we believe that these refinements in timing of TMR can lower the procedure's risks.
☆ This study was supported by a grant from the Food and Drug Administration to the Society of Thoracic Surgeons. This project was supported in part by an Agency for Healthcare Research and Quality Centers for Education and Research on Therapeutics cooperative agreement (grant #U18HS10548). Dr. Peterson is a recipient of a Paul Beeson Career Development Award. Dr. Kaul was supported by a fellowship award from the Canadian Institute of Health Research, Heart and Stroke Foundation of Canada, and the Alberta Heritage Foundation for Medical Research.
- coronary artery bypass graft surgery
- confidence interval
- Duke Clinical Research Institute
- Food and Drug Administration
- myocardial infarction
- odds ratio
- randomized clinical trial
- Society of Thoracic Surgeons
- transmyocardial revascularization
- American College of Cardiology Foundation
- Krabatsch T.,
- Schaper F.,
- Leder C.,
- Tulsner J.,
- Thalmann U.,
- Hetzer R.
- Gassler N.,
- Wintzer H.O.,
- Stubbe H.M.,
- Wullbrand A.,
- Helmchen U.
- Mack C.A.,
- Magovern C.J.,
- Hahn R.T.,
- et al.
- Al-Sheikh T.,
- Allen K.B.,
- Straka S.P.,
- et al.
- Aaberge L.,
- Nordstrand K.,
- Dragsund M.,
- et al.
- ↵Society of Thoracic Surgeons. Available at: http://www.sts.org. Accessed in August 2002
- ↵Ferguson TB, Peterson ED, Eiken MC, et al. CQI in CABG: a quality improvement national randomized trial. Presented at the Late Breaking Clinical Trials, American Heart Association 2002 Annual Scientific Session, Chicago, Illinois
- ↵Burns SM, Sharples LD, Tait S, Caine N, Wallwork, Schofield PM. The transmyocardial laser revascularization international registry report. Eur Heart J 1999;20:31–7
- Burkhoff D.,
- Wesley M.N.,
- Resar J.R.,
- Lansing A.M.
- Hughes G.C.,
- Landolfo K.P.,
- Lowe J.E.,
- Coleman R.B.,
- Donovan C.L.