Discussion
The STS adult cardiac surgery database cites the operative mortality of isolated MV replacement at 4.7%. When performing concomitant CABG, the operative mortality increases to 9.8%, with a significantly greater occurrence of post-operative complications and major morbidity occurring in 7.0% to 11.6% (1). Thus, it is hypothesized that the operative risk of combined CABG and MV surgery may be reduced by partitioning the operation into the two lower-risk, less invasive procedures of PCI + MVR. The present study demonstrated a low morbidity and mortality with staged PCI + MVR for significant single coronary artery and MV disease, compared to combined CABG and MVR.
Given the trends toward increases in minimally invasive cardiac surgery, the broad applicability of the hybrid approach described here may be particularly appealing. Gammie and colleagues (2) reported that from 2004 to 2008 the percentage of mitral valve operations that were done via a minimally invasive approach increased from 11.9% to 20.1% (P<.0001). With this progression, it is most likely that a hybrid approach will increase as well. On the other hand, the short-term benefits of a hybrid approach are not without potential long-term hazards.
Fortunately, in-stent restenosis, one of the major limitations of percutaneous revascularization, has decreased with each new generation of coronary stent (3). PCI, however, has yet to match CABG with regard to long-term benefits (4). Multiple studies have consistently demonstrated that for patients with multi-vessel disease and/or left main disease, regardless of the presence of diabetes, CABG yields better outcomes than PCI in terms of mortality, myocardial infarction and need for repeat coronary revascularization (5). Even when comparing CABG versus PCI for patients with proximal LAD disease, Hannan et al., showed that CABG patients had lower rates of repeat revascularization (6). The benefit of CABG over PCI involves the long-term effects of Internal Mammary Artery (IMA) to LAD anastomosis, the potential ability of bypass grafts to ‘treat’ lesions that subsequently develop and resultant downstream effects of cytokines on arterial disease (7). On the other hand, PCI offers lower rates of morbidity and shorter hospital stay.
The primary purpose of a hybrid valve/PCI is to substitute PCI for bypass grafting with saphenous vein grafts (SVGs), particularly for lesions not in the left anterior descending (LAD) coronary artery (8). With the current excellent performance of drug-eluting coronary stents (DES), restenosis and thrombosis rates of DES may be less than the estimated rate of SVG failure of 20 % at 12 months (9). The two most common clinical objectives of hybrid procedures are to reduce overall operative morbidity and mortality by transforming a single, high-risk surgery into two less risky procedures, and to facilitate minimally invasive surgery (8).
Hybrid procedures offer a reasonable alternative to traditional surgery for patients who meet the following basic criteria: non-LAD coronary lesions, not amenable to internal mammary bypass grafting; PCI that is technically feasible and likely durable from a procedural standpoint; and ability to tolerate the required antiplatelet and anticoagulation regimens.
Our study was a prospective comparative review, of 2 years duration, of 80 patients with combined single coronary artery (requiring revascularization in non-LAD territory) and rheumatic MV disease, divided into two groups; Group I consisting of 40 patients who underwent staged PCI, and mitral valve replacement 3 months later, and Group II consisting of 40 patients who underwent combined CABG (using saphenous venous graft) and mitral valve replacement. Our Aim was to compare intraoperative, postoperative, and follow-up outcomes of staged PCI versus CABG with mitral valve replacement in patients with combined single vessel and mitral valve disease.
Reoperative coronary bypass grafting in a patient with valvular disease poses a particular challenge in cardiac surgery. The hybrid approach is of particular benefit in reoperative patients who have had prior CABG with patent grafts. The technical difficulty of accessing lateral wall targets, safely dissecting patent bypass grafts, and obtaining exposure often precludes safe surgery, and these risks are not reflected in traditional scoring systems. Hybrid valve/PCI may be particularly useful in this regard and can dramatically simplify a challenging open valve and CABG surgery by substituting PCI for reoperative bypass grafting in lesions amenable to PCI (9). However, we excluded redo patients from our study, as this cohort is extremely high risk, and would have affected the results in a different way.
Although DES have shown excellent results in clinical trials, their effectiveness in clinical practice with more complex patients and complex lesions (high Syntax score, totally occluded coronary vessels, bifurcated lesions, small vessels, long lesions requiring multiple stents, ostial stenosis, calcified vessels) remains to be seen. Patients with diabetes, who comprise 30% of the surgical population and 37.5% of Group I in our study, have higher restenosis rates with DES (10). Late stent restenosis and thrombosis is another concern.
The most recent data from STS (Society of Thoracic Surgeons) demonstrate that in those undergoing isolated MV surgery, the rate of MV repair was 57.4%, and MV replacement was 42.6% (1). In the cases where MV pathology and etiology were documented, 56.6% were identified as having mitral regurgitation due to annular or degenerative disease, without stenosis, of which repair was performed in 75.0% of patients. In the present study, regarding the mitral valve lesions, mitral stenosis was found in most of the patients; 24 patients (60%) of Group I versus 16 patients (40%) of Group II (P=0.074), mixed mitral lesions (stenosis + regurgitation) come in the second place with 8 patients (20%) of Group I versus 20 patients (50%) of Group II (P=0.005) (statistically significant association between groups regarding mixed mitral valve lesions only), and mitral regurgitation in the minority of study groups with 8 patients (20%) of Group I versus 4 patients (10%) of Group II (P=0.210). This means that mitral repair could have been feasible in a minority of patients with pure mitral regurgitation, as the reparability of other pathologies varies markedly. That encouraged us to exclude mitral repair and standardize mitral replacement as the uniform approach for mitral valve surgery in this study.
Although one needs to be cautious when making direct comparison with other studies, reductions in the parameters of morbidity were noted when compared with data from the most recent STS adult cardiac surgery database outcomes. In patients undergoing CABG plus MV replacement, the most common complication is new-onset atrial fibrillation, which occurs in 44.2%, and increases peri-operative morbidity and hospital length of stay (11). This figure is higher than the 16.1% noted in present cohort of PCI + MIMVS (Minimally Invasive Mitral Valve Surgery) and is consistent with prior studies suggesting a reduced incidence of post-operative atrial fibrillation when utilizing a minimally invasive approach for valve surgery (12). In our study, we compared staged PCI + conventional MVR (Group I) to combined CABG + MVR (Group II). We noticed no ECG changes in the form of arrhythmias in both groups, the AF patients preoperatively in both groups went through several changes in the rhythm and rate intraoperatively before returning to baseline AF again. This may be due to the less morbid group of patients in our study, with good EF, low median PAP, relatively good clinical status, and nearly no comorbidities.
The staged strategy ensures optimal myocardial protection during the mitral correction. It is our belief that myocardial protection is greatly enhanced with hybrid procedures. By achieving 100% completeness of revascularization before cross-clamping in all of our valve-PCI patients, cardioplegia administration to all regions of the heart was possible. In contrast, if surgeons are reluctant to attempt revascularization on high-risk or technically difficult to reach lesions, cardioprotection may be compromised, leading to low cardiac output postoperatively, and worsened outcomes. The reduction in cross-clamp time for a hybrid procedure also provides significant myocardial protective benefit, as the heart is faced with a lower overall ischemic time, and potentially less dysfunction upon reperfusion.
By performing PCI to treat the coronary artery disease, one obviates the necessity of performing concomitant CABG at the time of surgery, significantly reducing the complexity of the surgery and shortening the operative times, which was noted in our study when compared with conventional combined CABG and MVR. In our study, the median aortic cross-clamp and cardiopulmonary bypass times were 44 (IQR, 39-48) and 62 minutes (IQR, 59-68) for Group I, versus 60.5 (IQR, 55-65) and 82 minutes (IQR, 75-88) for Group II (P=0.001), that difference between groups is statistically significant.
The less traumatic nature of isolated MVR and reduced operative times in Group I likely conferred lower bleeding and transfusion requirements, the median bleeding loss (ml) postoperatively was 925 (IQR, 650-1200) in group I versus 1075 (IQR, 900-1400) in group II (P= 0.021), which is statistically significant, with a median of 2 units of packed red blood cells (RBCs) transfused (IQR, 1–2) in group I compared to a median of 2 packed RBCs units transfused (IQR, 2-3) in group II (P= 0.002), which is statistically significant also. However, the rate of re-exploration for bleeding did not differ significantly between both groups, with 1 case only (2.5%) in group I versus 2 cases (5%) in group II (P= 0.556). All reoperation for bleeding cases were related to sternal wire placement, and no postoperative delayed cardiac tamponade was noted in any of the two groups.
Although, our study was not powered to detect a statistically significant difference, shorter operative times and less blood product use during cardiac surgery are associated with fewer infections, and a lower morbidity and mortality (13, 14). However, composite post-operative complications (prolonged mechanical ventilation (>24 h), respiratory complications, wound infection, cerebrovascular accidents, re-operation for bleeding, acute kidney injury, and 30-day mortality) occurred less frequently in group I than in group II.
With the goal being to optimize stent patency while minimizing the risks of bleeding, there is concern regarding the risk of bleeding if the surgery is performed after the PCI and the possibility of stent thrombosis with protamine reversal. Of particular concern is the risk of bleeding with dual antiplatelet therapy, largely based on known higher rates of bleeding after CABG in patients receiving clopidogrel (15). In the current study, median time of PCI to mitral valve surgery was 93 days, the patients had their antiplatelet agents stopped 5 days before surgery and resumed their antiplatelet regimen within 24 to 48 hours after surgery.
In the previously mentioned study by Byrne and colleagues (16), because of the use of dual antiplatelet therapy, a high incidence of bleeding occurred, with 22 (85%) of the 26 patients requiring blood transfusions. In an attempt to reduce the incidence of bleeding, Brinster and colleagues (17) performed the PCI the day of, or evening before, the scheduled minimally invasive aortic valve replacement in 18 patients. There were no reoperations for bleeding, and only 8 (44%) patients required blood transfusions.
Santana O, et al. compared the outcomes of patients taking clopidogrel with those who were not taking clopidogrel. In the intraoperative period, there were no differences in the requirement of blood products, whereas in the postoperative period, there was a significantly higher number of patients taking clopidogrel who required blood products compared with those not taking clopidogrel. Out of concern for the possible development of stent thrombosis, they prefer the continuation of antiplatelet therapy at the time of valve operation (18).
In the study of Mihos CG, et al. (19), even though there was a higher use of pre-operative clopidogrel in those undergoing PCI+MIMVS, there were fewer intraoperative transfusions required, when compared with CABG+MVS. The lower need for blood products in the PCI+MIVS group is most likely due to the fact that, by its less traumatic nature, minimally invasive valve surgery is associated with less blood loss. Also, by virtue of the fact that there was no need to place bypass grafts, the operative times were much shorter in this group, thereby having less bleeding (20). Importantly, there were no cases of acute stent thrombosis peri-operatively. In their previous work, they evaluated 222 patients who had PCI+MIMVS, 183 of which were on clopidogrel and were compared with 38 who were not (18). In the intra-operative period, there were no differences in the requirement of blood products between the two groups. Post-operatively, there was a higher proportion of patients on clopidogrel requiring blood products compared with those who did not take it (50.5% versus 26.3%, P=0.005); however, there was no significant difference in the need for re-operation for bleeding. Because clopidogrel use perioperatively appears to be safe (21), their clinical practice has been to continue anti-platelet therapy at the time of valve surgery to minimize the risk of acute stent thrombosis.
Ideally, these patients would be best managed by either a longer staging duration so that the clopidogrel can be stopped (three to six months with drug-eluting stents), or by a very short staging duration (under 6 h), so that clopidogrel’s actions are just beginning to take effect once the surgery has been completed. Our study differs from the previously mentioned studies in that we had a significant variation on the use of antiplatelet agents, our Group I patients had their antiplatelet agents stopped 5 days before surgery. The median time of PCI to mitral valve surgery in our study was 93 days, so we feel it is safe to withhold the antiplatelet therapy.
In these hybrid procedures, the optimal timing of the valve operation once PCI has been performed is not known. At our institution, the time delay between PCI and the valve operation is mainly driven by a desire to reduce the incidence of acute kidney injury counterbalanced with the urgency of the operation. It has been noted that the closer the 2 procedures are in time, the higher the incidence of acute kidney injury. Data from 4,440 patients undergoing coronary angiography and cardiac operation on the same day demonstrated this approach to be an independent predictor for the development of acute kidney injury (22). Another study evaluated the incidence of acute kidney injury in patients who had cardiac catheterization and cardiac operations during the same admission and compared it with a group of patients who had cardiac catheterization followed by operation at a later admission (23). The incidence of acute kidney injury in the patients who had same-admission cardiac catheterization and operations was 50.2% compared with 33.7% in those who had operations at a later date (p = 0.009). To reduce the incidence of acute kidney injury, several institutes prefer to wait at least 3 weeks after PCI to perform valve operations (22, 23).
We decided to wait 3 months after PCI to perform MVR, to avoid renal failure, and to be able to stop the antiplatelet therapy safely. We have got 2 patients (5%) in group I versus 1 (2.5%) in group II of acute kidney injury (P= 0.556), in the form of elevated creatinine levels, which resolved medically, without the need for dialysis. This difference was not statistically significant. We attribute this to the period of 3 months between PCI & MVR, which provide us with the protection window against acute kidney injury.
Santana O, et al (24) compared 65 patients who had a hybrid approach with 52 matched control patients who underwent conventional bypass grafting and valve operation. The results demonstrated a significant reduction in composite complications and hospital lengths of stay in the hybrid group when compared with conventional group. This is similar to the results per our study which showed fewer composite complications in Group I, as well as statistically significant lower median intensive care unit (ICU) length of stay (hours) and hospital length of stay (days); 39 hours (IQR, 32-45) and 5.5 days (IQR, 5-6), versus 56.5 hours (IQR, 49-69) and 8.5 days (IQR, 7-13) for Group II (P=0.001), respectively.
The postoperative complications were comparable with no statistically significant difference for groups I and II, with less prolonged mechanical ventilation (>24 h) 0 (0%) versus 3 (7.5%), and less respiratory complications 0 (0%) versus 1 (2.5%) for group I. This may be due to less aggressive and less time consuming is the isolated MVR than the combined CABG + MVR, which paves the way for faster extubation with less respiratory complications. The 3 patients of Group II eventually got extubated, one of them got chest infection which has been resolved using appropriate antibiotics. However, both approaches did include a sternotomy.
By virtue of avoiding a sternotomy, minimally invasive surgery results in less thoracic surgical trauma and alterations in pulmonary physiology and biomechanics, which contributes to an enhanced post-operative recovery and faster extubation, with a reported significantly lower incidence of prolonged mechanical ventilation occurring in 18.3% of the PCI + MIMVS cohort, and 29% in CABG plus MV replacement, leading to shorter intensive care unit length of stay with the PCI + MIMVS approach, when compared with sternotomy. We don’t see this significant difference in our study, because both groups were approached through a median sternotomy, with low incidence of prolonged mechanical ventilation, most probably due to exclusion of any lung disease or other comorbidities from our study groups.
Our present study showed slightly less wound infection in group I; 0 (0%) versus 2 (5%), most probably due to less operative time, ICU & hospital stay, blood transfusion requirements, & other postoperative complications which usually encourage infection. However, this difference is not significant, probably due to the similar approach used in both groups; sternotomy, and similar baseline characteristics between both groups.
The cerebrovascular accidents encountered in our study was one stroke patient in group II without any residual deficit. No other cases were reported in both groups, probably due to relatively good baseline characteristics of the patients included in our study, with good carotid duplex preoperatively, and exclusion of old patients above 60 years old with any central or peripheral vascular disease.
Several groups have investigated hybrid approaches of PCI combined with valve operations. In 2014, Santana et al., published the results of over 200 patients who underwent PCI for coronary revascularization followed by a minimally invasive valve procedure. They found a mortality rate of 3.6% and an all-cause mortality rate of 12% at 4.5 years. They also demonstrated a decreased complication rate and length of stay for the hybrid group compared to those undergoing conventional sternotomy (24).
In 2015, George et al. described a series of 26 patients who underwent a single-stage hybrid procedure involving PCI of a non-LAD vessel followed by a valve operation (21). Recalculating the STS risk after the PCI was performed, they found a 35% risk reduction in the re-operative group and a 17% risk reduction in the non-reoperative group. In addition, they had no in-hospital mortalities and very few complications. No coronary-stent thromboses were noted during a follow-up period of two years.
Specific to mitral valve, Umakanthan et al. described the Vanderbilt experience with 32 consecutive patients who underwent a hybrid procedure, including PCI and mitral valve surgery. Of these procedures, 28 (89 %) were performed as a single-stage procedure in a hybrid operating room (25). The observed in-hospital mortality rate was 3 % (1/32) and survival at one and two years was 96 % and 89 %, respectively. The series was expanded to 39 patients and reported by Solenkova et al., noting a predicted mortality for conventional CABG/ mitral of 14.1 % versus an observed in-hospital mortality of only 2.6 % (1/39). (26)
This is different from our study which showed no operative/30-day mortality in both groups. This may be attributed to the baseline characteristics of our group of elective patients who are low risk patients with few or no comorbidities.
As demonstrated by 5-year outcomes from the Synergy Between Percutaneous Coronary Intervention With TAXUS and Cardiac Surgery (SYNTAX) trial (27), a strong argument can now be made for PCI in patients with left main or multivessel disease with low SYNTAX scores (less than 23). However, the SYNTAX trial also clearly demonstrates a survival benefit of CABG for patients with a higher burden of disease, as reflected by a high SYNTAX score (greater than 33), and in specific patient subsets, such as patients with diabetes mellitus (28). The benefit of CABG is primarily attributable to left IMA grafting to the LAD, and the patency of IMA grafting consistently exceeds 95% at 10 years, setting the gold standard with which other revascularization strategies should be compared. Yet, significant limitations of both PCI and CABG persist. Whereas PCI is burdened by the need for repeat target lesion interventions, saphenous vein graft failure for non-LAD targets in CABG can reach 30% at 1 year, and at 10 to 15 years, only 50% to 60% of the SVGs have been reported to be patent (29). Conversely, the early restenosis and thrombosis rate of the drug-eluting stents (DES) in non-LAD vessels is lower than that reported for SVG failure (30).
This is to a great extent in line with the results per our study. Postoperative routine trans-thoracic echo (TTE) follow-up was done upon discharge, as well as after 3 months, 6 months, and 1 year for all patients in both groups. Follow-up was completed after one year. Regional wall motion abnormalities were noted in 15 patients (37.5%) of group I versus 17 patients (42.5%) of group II (P=0.648), who all undergone stress ECG, of whom 9 patients (22.5%) in group I versus 11 patients (27.5%) in group II showed positive results (P=0.606), and qualified for diagnostic coronary angiography, which confirmed the need for reoperation for myocardial ischemia/infarction within the first year of follow up post-operatively in 4 patients (10%) of group I versus 8 patients (20%) of group II (P=0.210). However, all these follow up outcomes showed no significant difference between both groups within the first year of follow up. None of the patients in both groups had major cardiac events or CCU admission. We attribute this to the nature of our patients in both groups, who have single non-LAD vessel disease supplying limited heart territories with good functional reserve. Long term data are needed for more informative conclusion.
As more hybrid PCI/valve procedures are being performed, many questions remain unanswered, including the optimal order for the procedures, their timing, the management of dual antiplatelet therapy, and the optimal costs and logistics of the procedures (8).