Author + information
- Received April 12, 2016
- Revision received July 14, 2016
- Accepted July 28, 2016
- Published online October 24, 2016.
- S1936879816312195-44b45ed0292784b94702f3991c74e0e5Jochen Reinöhl, MDa,∗ (, )
- S1936879816312195-280167e9d9d10a6a9a30642b3e9f5cb1Klaus Kaier, PhDa,b,
- S1936879816312195-767b18ae0f5ff38db04b5c967070d061Holger Reinecke, MDc,
- S1936879816312195-09fa6615c37f446805814b915867c163Lutz Frankenstein, MDd,
- S1936879816312195-b24998cfda774e7ac7c3c606735bf332Andreas Zirlik, MDa,
- S1936879816312195-5634a2da4687494741066ca4e315e0caManfred Zehender, MDa,
- S1936879816312195-93022f4593011de76671e3456a4db2cbConstantin von zur Mühlen, MDa,
- S1936879816312195-648d1c10d1bb28c123a93490c649089cChristoph Bode, MDa and
- S1936879816312195-62e3d4eaa4f86b8ab3bee4c952f414e8Peter Stachon, MDa
- aDepartment of Cardiology and Angiology I, Heart Center Freiburg University, Freiburg, Germany
- bCenter for Medical Biometry and Medical Informatics — University of Freiburg, Freiburg, Germany
- cDivision of Vascular Medicine, Department of Cardiovascular Medicine, University Hospital Muenster, Muenster, Germany
- dDepartment of Cardiology, Angiology, and Pulmonology, University of Heidelberg, Heidelberg, Germany
- ↵∗Reprint requests and correspondence:
Dr. Jochen Reinöhl, Heart Center Freiburg University, Department of Cardiology and Angiology I, Hugstetter Straße 55, Freiburg 79106, Germany.
Objectives The aim of this study was to assess how the introduction of transcatheter aortic valve replacement (TAVR) has changed clinical practice and outcome in patients who have previously undergone coronary artery bypass grafting (CABG).
Background A significant proportion of patients admitted for aortic valve replacement have previously undergone CABG and are therefore at increased operative risk in case of redo surgery.
Methods In-hospital outcome data were analyzed from patients with or without previous CABG undergoing isolated surgical aortic valve replacement or TAVR in Germany from 2007 to 2013.
Results In total, 32,581 TAVR and 55,992 surgical aortic valve replacement procedures were performed in patients with (n = 6,221) or without (n = 82,352) previous CABG. TAVR increased markedly in patients with previous CABG, from 18 procedures in 2007 to 1,191 in 2013, while surgical aortic valve replacement decreased in these patients from 471 to 179 procedures. In 2013, TAVR was the preferred therapy in almost 90% of patients with previous CABG. In-hospital mortality was increased in patients with previous CABG compared with those without (7.6% vs. 6.3% for TAVR and 7.2% vs. 2.6% for surgical aortic valve replacement). Bleeding and stroke rates were also increased with redo surgical aortic valve replacement procedures (with vs. without previous CABG: stroke, 3.2% vs. 1.8%; relevant bleeding, 29.6% vs. 13.4%; acute kidney injury, 4.2% vs. 2.9%), whereas this was not the case with TAVR (stroke, 2.1% vs. 2.6%; relevant bleeding, 7.3% vs. 8.3%; acute kidney injury, 6.3% vs. 5.4% respectively). A similar influence was seen in resource utilization (discharge destination home: TAVR after CABG, 51%; surgical aortic valve replacement after CABG, 31%).
Conclusions Since its introduction in 2007, TAVR has been increasingly used in Germany in patients with previous CABG, and in-hospital outcome data support the trend away from redo surgery.
- coronary artery bypass graft (CABG)
- in-hospital mortality
- resource utilization
- transcatheter aortic valve replacement (TAVR)
Clinical outcomes of transcatheter aortic valve replacement (TAVR) are equivalent or even superior to those of surgical aortic valve replacement in inoperable or high-risk patients with severe aortic valve stenosis (1–3). Therefore, TAVR is often the preferred choice of treatment in these patients (4). Because similar risk factors promote the formation of aortic valve stenosis and coronary artery disease (5), a significant proportion of patients with aortic valve stenosis have undergone previous coronary arterial bypass grafting (CABG) (6). Because of technical challenges, a redo surgical approach is associated with increased perioperative risk (7,8). The adoption of TAVR could therefore be expected to initiate a marked shift among patients with previous CABG from surgical aortic valve replacement to TAVR. However, as this has yet not been shown, the aim of our analysis was to evaluate how the treatment of aortic valve stenosis and associated outcomes have changed in patients with previous CABG since the introduction of TAVR in Germany.
Study population and data source
A detailed description of methods can be found elsewhere (4). Briefly, diagnoses coded according to the German modification of the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, as well as procedural codes (according to the German procedure classification system) and German diagnosis-related groups of all patients undergoing isolated surgical aortic valve replacement or TAVR from 2007 to 2013 were provided by the Research Data Centers of the Federal Bureau of Statistics (Statistisches Bundesamt, DESTATIS, Wiesbaden, Germany). We excluded admissions for prosthetic valve replacement or with concomitant interventions such as percutaneous coronary intervention or CABG, as well as admissions without the diagnosis of aortic stenosis. Because the datasets are kept anonymous by the Research Data Centers of the Federal Bureau of Statistics in groups with n < 3, procedures, baseline characteristics, post-procedural complications, and in-hospital mortality were not submitted to us. These values, shown as “NA” in the tables, were not included in summary statistics.
In the baseline characteristics, previous coronary bypass or valve surgery was described as “previous cardiac surgery.” Regarding in-hospital complications, bleeding was defined as requiring more than 5 U of red blood cells. For all other comorbidities and complications, the existing anamnestic or acute distinctive codes were used (related German procedure classification system and International Classification of Diseases codes were presented previously ). The logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) score was calculated with the available information (age, sex, admission status, and the International Classification of Diseases codes of comorbidities obtained from the Federal Bureau of Statistics), but critical preoperative state and left ventricular function were not available and were assumed to be insignificant (i.e., no critical preoperative state and no left ventricular dysfunction). Thus, we calculated a best-case scenario for the EuroSCORE.
Differences in baseline parameters were calculated using the chi-square test for categorical variables and the Student t test for continuous variables. All analyses were carried out using Stata version 13.0 (StataCorp LP, College Station, Texas).
Changes in aortic valve replacement in patients with previous CABG since the launch of TAVR
Between 2007 and 2013, a total of 88,573 hospitalizations for isolated aortic valve replacement took place in Germany. About 7% of these patients (n = 6,221) had undergone previous CABG (Table 1). Because of the roughly linear increase in TAVR procedures, without a corresponding decrease in surgical aortic valve replacement procedures, the annual volume of both procedures increased over time by 1.8 times (from 8,766 procedures in 2007 to 16,195 procedures in 2013 in patients without previous CABG). During the same period, the increase in procedures in patients with previous CABG was even more pronounced by about 3 times (from 489 procedures in 2007 to 1,370 procedures in 2013). Thus, the annual proportion of patients with previous CABG rose from 6% to 9%. Whereas TAVR numbers evolved in these patients (from 18 patients in 2007 to 1,191 patients in 2013), the number of surgical aortic valve replacement procedures more than halved over time (from 471 patients in 2007 to 161 and 179 patients in 2012 and 2013, respectively). Accordingly, in patients with previous CABG, TAVR was the preferred therapy in nearly 90% of cases in 2013 (Figure 1).
Patients with previous CABG undergoing surgical aortic valve replacement were older than patients without previous CABG (73 ± 7 years vs. 70 ± 10 years). TAVR was generally performed in older patients, but patients with previous CABG undergoing TAVR were significantly younger compared with those undergoing TAVR without previous CABG (79 ± 6 years vs. 81 ± 6 years). In both groups, patients with previous CABG were more likely to be men (surgical aortic valve replacement with previous CABG, 83%; TAVR with previous CABG, 72%). As expected, the predicted operative mortality by the EuroSCORE and New York Heart Association class was elevated in patients undergoing surgical aortic valve replacement and those undergoing TAVR with previous CABG. The elevated EuroSCORE values in patients with previous CABG were due to a higher prevalence of comorbidities such as recent myocardial infarctions, extracardiac arteriopathy, and diabetes (Table 1).
To assess the risks of patients with and without previous CABG, typical in-hospital complications were analyzed. As shown in Table 2, bleeding doubled in patients treated with surgical aortic valve replacement and previous CABG: the incidence of relevant bleeding, defined as need for transfusion of more than 5 U of red blood cells, was 13% in patients without previous CABG and 30% in patients with previous CABG. In patients treated with TAVR, no relevant difference in bleeding was observed in patients with previous CABG (TAVR without previous CABG, 8%; TAVR with previous CABG, 7%). Furthermore, the incidence of surgical aortic valve replacement–related stroke increased in patients who had previously undergone CABG (2% without previous CABG, 3% with previous CABG). However, the TAVR group had a slight decrease in the incidence of TAVR-related stroke in patients previously treated with CABG (3% without previous CABG, 2% with previous CAGB). Time on ventilator was roughly 22 h for TAVR patients irrespective of previous CABG but rose significantly in patients with previous CABG treated with surgical aortic valve replacement (25 h without CABG, 45 h with CABG). As expected, need for permanent pacemaker implantation was increased in all patients treated with TAVR but independent of previous CABG (surgical aortic valve replacement without CABG, 4%; surgical aortic valve replacement with CABG, 4%; TAVR without CABG, 18%; TAVR with previous CABG, 15%). Acute kidney injuries occurred more often in both groups with previous CABG. If any of these perioperative complications occurred, related mortality increased (Table 2).
Prolonged time on ventilator is a predictor of hospital resource utilization (9). The length of hospital stay in the primary hospital was comparable in patients treated with either surgical aortic valve replacement or TAVR with previous CABG (18 days vs. 17 days). Previous CABG had only a minor influence on discharge destination: patients treated with surgical aortic valve replacement were discharged mostly to rehabilitation or to secondary hospitals (discharge home without previous CABG, 29%; discharge home with previous CABG, 27%). About one-half of patients after TAVR were discharged home (discharge home without previous CABG, 47%; discharge home with previous CABG, 49%) (Table 3).
Surgical aortic valve replacement after previous CABG was accompanied by higher operative risk and more in-hospital complications. Accordingly, the in-hospital mortality of surgical aortic valve replacement was more than 2-fold higher for patients with previous CABG (3% without previous CABG, 7% with previous CABG) (Figure 2A). Despite the marked increase in operative risk assessment in these patients as shown by the respective EuroSCORE values (TAVR without previous CABG, 21%; TAVR with previous CABG, 33%), the real in-hospital mortality rate rose only slightly from 6% in patients without previous CABG to 8% in those with previous CABG. However, the in-hospital mortality rate among patients with previous CABG treated with TAVR declined significantly over time from 9% (2007 to 2009) to 7% (2012 to 2013), indicating further improvements in the TAVR procedure itself (Figure 2B).
TAVR is the preferred therapy for patients with previous CABG
In the present study we examined a nationwide dataset of 88,573 patients admitted with aortic valve stenosis for treatment with either isolated surgical aortic valve replacement or TAVR in Germany between 2007 and 2013. About 7% of these patients had undergone previous CABG. At the launch of TAVR in 2007, only 489 patients with previous CABG underwent either surgical aortic valve replacement or TAVR, indicating that only a few patients with severe aortic valve stenosis and previous CABG could be treated adequately. Over time, the overall number of TAVR procedures has risen, without a concomitant reduction in surgical aortic valve replacement, but patients with severe aortic valve stenosis and previous CABG have been increasingly treated with TAVR. In contrast, the use of surgical aortic valve replacement in these patients has almost disappeared. This suggests that local heart teams prefer TAVR as the more appropriate treatment for these patients.
Patients with previous CABG are at high operative risk
These decisions of local heart teams are reasonable given that patients with previous CABG are at higher operative risk as assessed by the EuroSCORE. On one hand, this is caused by the previous cardiac surgery, which is 1 element of the EuroSCORE, whereby redo surgery is technically challenging because of problems with scarring, increased calcification of the aortic root, or the risk of damaging the bypasses (7,8). On the other hand, patients with previous CABG more often had comorbidities, which increase operative and post-operative mortality (10,11).
Complications of TAVR are similar in patients with and those without previous CABG
In agreement with the high preoperative risk assessment as indicated by the estimated EuroSCORE, the present study shows increased complications in patients with previous CABG undergoing surgical aortic valve replacement. Substantial bleeding was 2 times more frequent in these patients compared with those without previous CABG. As expected, patients with previous CABG undergoing TAVR did not bleed more than patients without previous CABG. Furthermore, the risk for perioperative stroke rose dramatically in patients with previous CABG undergoing surgical aortic valve replacement. In contrast, patients receiving TAVR after previous CABG experienced even fewer perioperative strokes than those without previous CABG. This surprising finding is in line with data from the FRANCE-2 registry and demonstrates that previous CABG has no influence on the safety of TAVR (12). Time on ventilator is an indicator of the need for intensive care and a severe post-operative course. Patients undergoing surgical aortic valve replacement after previous CABG were mechanically ventilated on average for 20 h longer than those without previous CABG. In patients treated with TAVR, no difference in time on ventilator occurred between patients with or without previous CABG. Prolonged time on ventilator is associated with increased risk for pneumonia (13). This fact could partly explain the worse postoperative courses of patients with previous CABG. As expected, need for permanent pacemaker implantation, a typical complication of TAVR, was greater after TAVR, but this was mostly irrespective of previous CABG. However, acute kidney failure occurred more often in patients undergoing TAVR with previous CABG. In conclusion, apart from acute kidney failure, in-hospital complications in patients with previous CABG undergoing TAVR did not increase, in contrast to patients treated with surgical aortic valve replacement and previous CABG.
In-hospital mortality of TAVR is comparable in patients with and those without previous CABG
According to the increased EuroSCORE values in patients with previous CABG undergoing surgical aortic valve replacement, in-hospital mortality was more than doubled compared with patients without previous CABG. However, the increase in mortality was less distinct in patients treated with TAVR after previous CABG, despite increased predicted operative mortality. Because of the disparate groups, a direct comparison of surgically and transcatheter-treated patients is not possible. However, in-hospital outcomes improved during the past 8 years in all groups, in particular in the group undergoing TAVR after CABG, demonstrating a learning curve and technical improvements for the TAVR procedure, which may continue even further (1).
Resource utilization increases after surgical aortic valve replacement in patients with previous CABG
Analyses of the resource utilization of TAVR and surgical aortic valve replacement show that time on ventilator and length of hospital stay were longer in patients undergoing surgical aortic valve replacement after previous CABG. Furthermore, patients were more often discharged to second hospitals after surgery, indicating an increased utilization of resources.
It was a retrospective analysis for the in-hospital period only, so we cannot provide mid- or long-term outcomes. It was also based on administrative data, which ensures high quality and completeness, but coding errors cannot be ruled out. A comparison between the groups is difficult, because clinical details of the patients are not completely known and reasons for individual decisions are not traceable. Furthermore, previous cardiac surgery other than CABG could not be included in the analysis, because of small numbers of procedures, resulting in the information’s not being releasable by the Research Data Centers of the Federal Bureau of Statistics because of privacy concerns.
Finally, our analysis reports data from Germany only, and our conclusions may not be applicable in other systems with different clinical experiences and economic situations. However, we present a dataset with clinical outcomes of more than 80,000 patients, which is highly relevant and may be useful in estimating achievable outcomes for patients.
We conclude that TAVR is a safe and appropriate procedure for patients with previous CABG, very comparable in terms of perioperative complications apart from acute kidney failure, and in-hospital mortality. Further studies are required to determine long-term outcomes for patients with previous CABG and treatment with either TAVR or surgical aortic valve replacement.
WHAT IS KNOWN? A significant proportion of patients with aortic valve stenosis have histories of CABG. Because of technical challenges, a redo surgical approach is associated, with associated increased perioperative risk. The adoption of TAVR could improve medical care of these patients by making available a new treatment option with safe clinical outcomes.
WHAT IS NEW? Previous CABG has only minor influence on in-hospital outcomes after TAVR, in contrast to surgical aortic valve replacement. Consequently, there is an obvious trend toward treating patients with severe aortic valve stenosis and previous CABG with TAVR.
WHAT IS NEXT? The present study presents in-hospital outcomes. Further prospective studies are required to determine mid- and long-term outcomes in patients undergoing TAVR or surgical aortic valve replacement after previous CABG.
The authors are indebted to the excellent assistance of Angelika Gerlach and Marion Rüdiger. The authors would also like to thank the staff of the Research Data Centers of the Federal Bureau of Statistics and the statistical offices of the federal states for providing the data and technical support during the analysis. The authors thank Philip Hehn (Center for Medical Biometry and Medical Informatics, Medical Center, University of Freiburg) for assistance with English language and grammar.
Internal funding from Heart Center Freiburg University supported the present study, without any conflicts of interest. Dr. Reinöhl works as proctor for Edwards Lifesciences and Direct Flow Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery bypass grafting
- European System for Cardiac Operative Risk Evaluation
- transcatheter aortic valve replacement
- Received April 12, 2016.
- Revision received July 14, 2016.
- Accepted July 28, 2016.
- American College of Cardiology Foundation
- Stewart B.F.,
- Siscovick D.,
- Lind B.K.,
- et al.
- Castellant P.,
- Didier R.,
- Bezon E.,
- et al.