Article ID: CJ-19-0175
Background: Although minimally invasive mitral valve surgery via a right minithoracotomy (MICS-mitral) is widely performed, no tool to evaluate its risk has been reported. We sought to establish MICS-mitral risk calculators using a national clinical database for selection of appropriate patients.
Methods and Results: Between 2008 and 2015, 3,240 patients (mean age 59±14 years, males 1,950) underwent a MICS-mitral procedure in Japan and were registered in a national clinical database. We examined mortality and composite outcome (operative mortality, stroke, reoperation for bleeding) using multivariate analysis, then developed a risk calculator for each using stepwise analysis. Operative mortality was 1.1% and the composite outcome rate was 5%. In multivariate analysis, risk factors for operative mortality were shown to be age, respiratory dysfunction, thoracic aortic disease, myocardial infarction, body mass index >30, NYHA class IV, moderate or severe aortic regurgitation, mitral valve replacement, multiple valve surgery, and annual cases <10. ROC curve analysis of our prediction formulas for mortality and composite outcome revealed an area under the curve for operative mortality of 0.877 (95% confidence interval: 0.82–0.94, P<0.01) and for composite outcome of 0.665 (95% confidence interval: 0.62–0.71, P<0.01).
Conclusions: We developed risk calculator formulas using risk factors associated with both operative mortality and composite outcome. The present risk calculator formula is useful for patient selection and may influence future applications for this procedure.
With advances in minimally invasive cardiac surgery (MICS) over the past 25 years, minimally invasive mitral valve (MV) surgery (MICS-mitral) via a right minithoracotomy has become increasingly utilized worldwide.1–4 Furthermore, paradigms for management of MV regurgitation (MR) have shifted to identify benefits earlier in the disease course, before development of adverse effects of long-standing MR on left ventricular function.5 As a result, there is growing advocacy for referral of asymptomatic patients for surgery, indicating the need for a high level of safety for MV repair, as a nationwide study conducted in Japan found that in-hospital death related to that procedure was approximately 1%.6
Even though MICS-mitral is widely performed, controversy remains regarding its application because of procedure-related complications, such as stroke or reoperation for bleeding.7,8 MICS-mitral should not compromise the clinical outcome; thus some surgeons may be reluctant to use this technique. Furthermore, no tool to evaluate its risk has been developed. Because the indications for MICS-mitral will likely be expanded with advancements in techniques and devices,8,9 an appropriate tool for risk assessment of patients undergoing MV surgery is important to ensure quality control, and would be helpful when deciding whether MICS-mitral or conventional MV surgery with a sternotomy is most appropriate for a patient undergoing MV surgery.
With this background in mind, we investigated early results of MICS-mitral cases to aid selection of appropriate patients for that procedure. Furthermore, we aimed to develop MICS-mitral risk calculators using data from the Japan Adult Cardiovascular Surgery Database (JCVSD) in order to provide useful information for patient selection.
For the present study, we selected patients who had undergone initial MV surgery via a right minithoracotomy from the JCVSD. We included those with concomitant operations, such as coronary artery bypass grafting, arrhythmia surgery, and surgery for other valve pathology. Patients who required circulatory arrest or ventricular fibrillation without cross-clamping of the aorta, or with a history of previous cardiac operation, were excluded. Between 2008 and 2015, 63,801 patients underwent initial MV surgery and were registered in the JCVSD. Of those, we selected 3,240 who underwent MICS-mitral via a right minithoracotomy (mean age 59±14 years, males 1,950, females 1,290), as the database includes a surgical approach category.
JCVSDData complied between January 2008 and December 2015 from cardiac surgery units located throughout Japan were obtained. The data registration project was approved by the institutional review board of each participating hospital. Informed consent was also given by all patients at each participating hospital. A high level of data collection was successfully achieved for more than 300 variables, with missing data representing <2% of all obtained information. The JCVSD variables and their definitions (http://jacvsd.umin.jp) were nearly identical to those in the Society of Thoracic Surgeons (STS) National Adult Cardiac Database (http://sts.org), with some slight differences. The methods used to create these databases and details regarding their contents have been reported.10
The JCVSD project includes software developed for a web-based data collection system used by data managers at participating hospitals. The accuracy of submitted data is verified by random monthly visits to each participating hospital by the Site Visit Working Group, comprising administrative office members and investigators who have previously used the JACVSD for clinical studies. Data validity is further confirmed by independent comparisons of the volume of cardiac surgery procedures at a particular hospital entered in the JCVSD with that reported to the Japanese Association for Thoracic Surgery in their annual survey.
Study DesignFor the present analysis, both operative mortality and composite operative mortality or major complications (composite outcome) were used as the endpoints. Operative mortality has been shown to be exactly the same as 30-day operative mortality noted in the STS National Database. It included any patient who died during the index hospitalization, regardless of length of hospital stay, and any who died within 30 days of the operation after being discharged from the hospital. Major morbidity was defined as stroke or re-exploration for bleeding as a postoperative outcome because these are considered to be procedure-related complications. Stroke was defined as a new neurological deficit persisting for more than 72 h and re-exploration for bleeding as any reoperation for bleeding.
We examined the relationships of various preoperative factors with in-hospital mortality and major morbidity, and also the effect of patient volume at each institution on postoperative outcomes by checking the distribution of MICS-mitral cases in the latest year (2015). According to the number of MICS-mitral cases treated at each hospital, we divided the patients into those treated at institutions that experienced <10 cases per year and at those that had ≥10 per year, and conducted the same analysis.
Finally, we investigated risk factors for the composite outcome (operative mortality, stroke, reoperation for bleeding) using multivariate analysis. A risk calculator of the composite outcome was made by using those factors.
Statistical AnalysisThe statistical model was analyzed using multiple logistic regression, with variables entered in the model selected from among all variables using bivariate tests. A χ2 test was used for categorical covariates, and an unpaired t-test or Wilcoxon rank-sum test for continuous covariates. We developed the risk models using multivariable logistic regression with stepwise selection of the variables. We included all of the variables presented in Table 1 as candidate predictors. For continuous variables, we created multiple variables that dichotomized the distribution at different cutoff points (e.g., <20, <25, <30 for body mass index [BMI]), and included all those variables in the model as candidates. Probabilities to enter and remove were set at 0.20 and 0.25, respectively. The developed models were then applied to the testing cohort, and their discriminatory performance was assessed using the area under curve of the receiver-operating characteristics (ROC) curve. We used SPSS 20.0 software (IBM Corp, Armonk, NY, USA) for data analyses, with P<0.05 considered to indicate significance.
| All (n=3,240) |
Mortality | Composite outcome | |||||
|---|---|---|---|---|---|---|---|
| (+) (n=35) |
(−) (n=3,205) |
P value | (+) (n=166) |
(−) (n=3,074) |
P value | ||
| Age (years) | 59±14 | 70±10 | 58±14 | <0.01 | 64±14 | 58±14 | <0.01 |
| Sex (M/F) | 1,950/1,290 | 22/13 | 1,928/1,277 | NS | 92/74 | 1,858/1,216 | NS |
| BSA (m2) | 1.64±0.20 | 1.62±0.24 | 1.64±0.20 | NS | 1.59±0.21 | 1.64±0.20 | NS |
| BMI >30 | 76 (2%) | 2 (6%) | 74 (2%) | NS | 4 (2.4%) | 72 (2.3%) | NS |
| History of smoking | 1,102 (34%) | 13 (37%) | 1,089 (34%) | NS | 54 (33%) | 1,048 (34%) | NS |
| Diabetes | 248 (8%) | 2 (6%) | 246 (8%) | NS | 17 (10%) | 231 (8%) | NS |
| Renal dysfunction | 203 (6%) | 8 (23%) | 195 (6%) | <0.01 | 23 (14%) | 180 (6%) | <0.01 |
| Dialysis | 27 (0.8%) | 1 (3%) | 26 (0.8%) | <0.01 | 2 (1.2%) | 25 (0.8%) | NS |
| Hyperlipidemia | 823 (25%) | 11 (31%) | 812 (25%) | NS | 41 (25%) | 781 (25%) | NS |
| Hypertension | 1,451 (45%) | 24 (69%) | 1,427 (45%) | <0.01 | 82 (49%) | 1,369 (45%) | NS |
| Recent CVD | 14 (0.4%) | 1 (3%) | 13 (0.4%) | <0.05 | 3 (2%) | 11 (0.4%) | <0.01 |
| Moderate to severe respiratory dysfunction | 46 (1.4%) | 4 (11%) | 42 (1.3%) | NS | 9 (5%) | 37 (1%) | <0.01 |
| Infective endocarditis | 82 (2.5%) | 3 (9%) | 79 (2.5%) | <0.01 | 5 (3%) | 77 (2.5%) | NS |
| MI | 49 (1.5%) | 6 (17%) | 43 (1%) | <0.01 | 11 (7%) | 38 (1%) | <0.01 |
| CAD | 20 (0.6%) | 1 (3%) | 19 (0.6%) | NS | 4 (2.4%) | 16 (0.5%) | <0.01 |
| TR (>grade II) | 305 (9%) | 12 (34%) | 293 (9%) | <0.01 | 31 (19%) | 274 (9%) | <0.01 |
| AR (>grade II) | 71 (2.1%) | 6 (17%) | 65 (2%) | <0.01 | 8 (5%) | 63 (20%) | <0.01 |
| Preoperative AF | 530 (16%) | 13 (37%) | 517 (16%) | <0.01 | 41 (25%) | 489 (16%) | <0.01 |
| Thoracic aortic disease | 23 (0.7%) | 3 (9%) | 20 (0.6%) | <0.01 | 4 (2.4%) | 19 (0.6%) | <0.01 |
| PAD | 49 (1.5%) | 3 (9%) | 46 (1.4%) | <0.01 | 6 (3.6%) | 43 (1.4%) | <0.01 |
| NYHA IV | 51 (1.6%) | 5 (17%) | 46 (1.4%) | <0.01 | 7 (4.2%) | 44 (1.4%) | <0.01 |
| NYHA functional class (I/II/III/IV) | 947/1,294/323/51 | 6/14/7/5 | 941/1,280/316/46 | <0.01 | 40/66/28/7 | 907/1,228/295/44 | NS |
| LV function (Good/medium/bad) | 2,697/518/25 | 19/12/4 | 2,678/506/21 | <0.01 | 122/40/4 | 2,575/478/21 | NS |
| Urgent or emergency | 34 (1%) | 4 (14%) | 30 (0.9%) | <0.01 | 4 (2.4%) | 30 (1.0%) | NS |
| Annual cases ≥10 | 1,791 (55%) | 8 (23%) | 1,783 (56%) | <0.01 | 71 (43%) | 1,720 (56%) | <0.01 |
AF, atrial fibrillation; AR, aortic regurgitation; BSA, body surface area; BMI, body mass index; CAD, carotid artery disease; CVD, cerebrovascular disease; LV, left ventricular; MI, myocardial infarction; NYHA, New York Heart Association; NS, not significant; PAD, peripheral artery disease; TR, tricuspid regurgitation.
The number of MICS-mitral patients in the overall cohort gradually increased from 2008 to 2015, as shown in Figure 1. Among all patients, the ratio of MICS-mitral patients was 1.9% in 2008, and 2015 increased to 7.6%.
Number of MICS-mitral procedures in patients with isolated mitral valve repair between 2008 and 2015. The number and the ratio of MICS-mitral procedures within the overall cohort gradually increased from 2008 to 2015. MICS, minimally invasive cardiac surgery.
Table 1 shows preoperative patient background information for all 3,240 selected MICS-mitral cases, as well as those with and without operative mortality or the composite outcome. The groups were homogeneous in terms of sex, body surface area (BSA), BMI >30, and history of smoking, diabetes, and hyperlipidemia, but there were significant differences between patients with and without a composite outcome for age, renal dysfunction, recent cerebrovascular disease, moderate to severe respiratory dysfunction, myocardial infarction, carotid artery disease, tricuspid or aortic regurgitation >grade II, preoperative atrial fibrillation (AF), thoracic aortic disease, peripheral artery disease, New York Heart Association (NYHA status) IV, and annual cases ≥10 cases/year. Furthermore, as with mortality, patients with hypertension, infective endocarditis, poor left ventricular function, and urgent status were more frequent.
Operative OutcomesOperative mortality was 1.1% and the composite outcome rate was 5%. Operative, cardiopulmonary bypass, and aortic cross-clamp times were significantly longer in for operative mortality or composite outcome as compared the group without these outcomes, and the incidence of transfusion was significantly higher in the group with mortality or composite outcome. MV replacement and concomitant valve surgery such as aortic or tricuspid valve surgery were more frequently performed in patients with operative mortality or composite outcome (Table 2). Table 3 presents the postoperative characteristics. Among the postoperative morbidities, cardiac tamponade, sepsis, prolonged ventilation, and renal dysfunction showed a tendency to be related to both operative mortality and composite outcome.
| All (n=3,240) |
Mortality | Composite outcome | |||||
|---|---|---|---|---|---|---|---|
| (+) (n=35) |
(−) (n=3,205) |
P value | (+) (n=166) |
(−) (n=3,074) |
P value | ||
| Operation time (min) | 313±96 | 507±198 | 312±92 | <0.01 | 385±152 | 310±91 | <0.01 |
| CPB time (min) | 191±68 | 279±104 | 190±66 | <0.01 | 222±84 | 189±66 | <0.01 |
| Cross-clamp time (min) | 131±54 | 169±59 | 131±54 | <0.01 | 142±53 | 131±54 | <0.01 |
| Transfusion | 1,308 (40%) | 33 (94%) | 1,275 (40%) | <0.01 | 128 (77%) | 1,180 (38%) | <0.01 |
| MV repair | 2,868 (89%) | 20 (63%) | 2,848 (89%) | <0.01 | 135 (81%) | 2,733 (89%) | <0.01 |
| MV replacement | 372 (11%) | 15 (43%) | 357 (11%) | <0.01 | 31 (19%) | 341 (11%) | <0.01 |
| Maze procedure | 234 (7%) | 2 (6%) | 232 (7%) | NS | 15 (9%) | 219 (7%) | NS |
| Concomitant tricuspid valve surgery | 437 (13%) | 11 (31%) | 426 (13%) | <0.01 | 32 (19%) | 405 (13%) | <0.05 |
| Concomitant aortic valve surgery | 13 (0.4%) | 6 (17%) | 64 (2%) | <0.01 | 7 (4%) | 63 (2%) | NS |
| Multiple valve surgery | 498 (15%) | 17 (49%) | 481(15%) | <0.01 | 38 (23%) | 460 (15%) | <0.01 |
| ICU stay (days) | 2.8±4.8 | 21.9±26.7 | 2.6±3.4 | <0.01 | 9.2±16.4 | 2.4±2.8 | <0.01 |
CPB, cardiopulmonary bypass; ICU; intensive care unit; MV, mitral valve; NS, not significant.
| All (n=3,240) |
Mortality (n=35) |
Composite outcome (n=166) |
|
|---|---|---|---|
| 30-day mortality | 25 (0.8%) | – | – |
| In-hospital mortality | 35 (1.1%) | – | – |
| Reoperation for bleeding | 90 (2.8%) | 10 (29%) | – |
| Cardiac tamponade | 13 (0.4%) | 2 (6%) | 9 (5%) |
| Stroke | 34 (1.0%) | 3 (9%) | – |
| Leg infection | 10 (0.3%) | 1 (3%) | 1 (0.6%) |
| Sepsis | 17 (0.5%) | 10 (29%) | 12 (7%) |
| Prolonged ventilation | 101 (3.1%) | 15 (43%) | 35 (21%) |
| Renal failure | 37 (1.1%) | 13 (37%) | 15 (9%) |
| New onset of AF | 463 (14.3%) | 3 (9%) | 34 (20%) |
| PMI | 18 (5.6%) | 2 (6%) | 5 (3%) |
| Aortic dissection | 2 (0.06%) | 0 | 0 |
| Limb ischemia | 6 (0.2%) | 0 | 0 |
AF, atrial fibrillation; PMI, perioperative myocardial infarction.
The distribution of MICS-mitral cases is shown in Figure 2. Comparison of outcome between institutions with small (<10 cases/year) and large (≥10 cases/year) numbers revealed significant differences with regard to postoperative complications (Table 4). Operative mortality rate and rate of composite outcome in the large volume group were 0.4% (8/1,791) and 1.9% (27/1,449), respectively, while those in the small-volume group were 1.9% and 6.6%, respectively. The rates of incidence of prolonged ventilation and renal failure were also higher and the duration of intensive care unit (ICU) stay was longer in the small-volume group (Table 4).
Distribution of numbers of MICS-mitral cases. There were many small-volume institutions performing fewer than 10 cases per year. MICS, minimally invasive cardiac surgery.
| Large volume (n=1,791) |
Small volume (n=1,449) |
P value | |
|---|---|---|---|
| Operation time (min) | 293±87 | 340±100 | <0.01 |
| CPB time (min) | 185±68 | 197±67 | <0.01 |
| Cross-clamp time (min) | 128±50 | 135±58 | <0.01 |
| Transfusion | 617 (34%) | 691 (48%) | <0.05 |
| 30-day mortality | 7 (0.4%) | 18 (1.2%) | <0.01 |
| In-hospital mortality | 8 (0.4%) | 27 (1.9%) | <0.01 |
| Reoperation for bleeding | 44 (2.4%) | 48 (3.3%) | NS |
| Cardiac tamponade | 6 (0.3%) | 7 (0.5%) | NS |
| Stroke | 20 (1.1%) | 14 (1.0%) | NS |
| Leg infection | 4 (0.2%) | 6 (0.4%) | NS |
| Sepsis | 6 (0.3%) | 11 (0.8%) | NS |
| Prolonged ventilation | 28 (1.6%) | 73 (5%) | <0.01 |
| Renal failure | 11 (0.6%) | 26 (1.8%) | <0.01 |
| New onset of AF | 246 (14%) | 217 (15%) | NS |
| PMI | 9 (0.5%) | 9 (0.6%) | NS |
| ICU stay | 2.0±2.2 | 3.7±6.7 | <0.01 |
MICS, minimally invasive cardiac surgery. Other abbreviations as in Tables 1–3.
Multiple regression analysis for all patients identified several risk factors affecting operative mortality and composite outcome (Table 5). Four risk factors were found to be associated with both operative mortality and composite outcome: age, respiratory dysfunction, myocardial infarction, and the number of annual cases. In the multivariate analysis, risk factors associated with operative mortality were age, respiratory dysfunction, thoracic aortic disease, myocardial infarction, BMI >30, NYHA class IV, moderate or severe aortic regurgitation, MV replacement, multiple valve surgery, and the number of annual cases, while those associated with a composite outcome were age, recent cerebrovascular disease, carotid artery disease, respiratory dysfunction, myocardial infarction, moderate or severe tricuspid regurgitation, and annual cases.
| Mortality | Composite outcome | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| Age (years) | 1.350 | 1.079–1.689 | <0.05 | 1.251 | 1.135–1.378 | <0.01 |
| Recent CVD | 5.167 | 1.352–19.756 | <0.05 | |||
| Carotid artery disease | 3.122 | 0.976–9.980 | <0.05 | |||
| Respiratory dysfunction | 3.606 | 1.050–12.381 | <0.05 | 3.178 | 1.438–7.024 | <0.01 |
| Thoracic aortic disease | 5.452 | 1.243–23.910 | <0.01 | |||
| MI | 4.384 | 1.495–12.858 | <0.05 | 2.876 | 1.364–6.065 | <0.01 |
| BMI >30 | 6.451 | 1.404–29.630 | <0.05 | |||
| NYHA class IV | 3.277 | 1.049–10.238 | <0.05 | |||
| TR >2 | 1.651 | 1.073–2.541 | <0.05 | |||
| AR >2 | 5.584 | 1.866–16.709 | <0.01 | |||
| MV replacement | 2.738 | 1.267–5.917 | <0.05 | |||
| Multiple valve surgery | 2.275 | 1.048–4.939 | <0.05 | |||
| Annual cases ≥10 | 0.290 | 0.125–0.672 | <0.01 | 0.653 | 0.473–0.901 | <0.01 |
CI, confidence interval; OR, odds ratio; PMI, perioperative myocardial infarction. Other abbreviations as in Tables 1–3.
Our prediction formulas for mortality and composite outcome are shown in Figure 3. Model performance was evaluated using the C-index (area under the ROC curve) as a measure of model discrimination. ROC curve analysis revealed an area under the curve for operative mortality of 0.877 (95% confidence interval (CI): 0.77–0.92, P<0.001) and for composite outcome of 0.665 (95% CI: 0.62–0.71, P<0.001).
Performance of each prediction model: The areas under the receiver-operating characteristic curve (ROC) for (A) operative mortality and (B) composite outcome. AR, aortic regurgitation; BMI, body mass index; CVD, cerebrovascular disease; MI, myocardial infarction; MV, mitral valve; TR, tricuspid regurgitation.
Recently, MICS-mitral has become the preferred technique at many centers because of lower rates of postoperative bleeding and AF, reduced incidence of wound infection, shorter hospital stay, quicker recovery, and improved outcomes.7,8,11–13 Nevertheless, the benefits of MICS-mitral are not universally accepted and many surgeons remain apprehensive about developing a minimally invasive MV repair program, because of the learning curve14 and potential complications. The present investigation of the early results of MICS-mitral procedures using data included in a national database of cases performed in Japan demonstrated risk factors for operative mortality and composite outcome, which have not been previously analyzed. Although many studies from leading institutions have shown favorable results, those reports lack significant statistical power to investigate the outcomes of MICS-mitral and do not reflect real-world outcomes. Furthermore, there is currently no risk model for MICS-mitral. Therefore, data from a national database with a large cohort, such as in the present study, are meaningful and necessary. We consider that the present findings provide useful information regarding MICS-mitral procedures and can help with appropriate patient selection.
Mortality and MorbidityAlthough the advantages of MICS-mitral have been highlighted in recent years, adaption of the procedure should not compromise the outcome of patients undergoing MV surgery. Previous studies revealed that the rate of operative mortality of conventional MV repair was 1.1–1.7%.8,14 In the present cohort, the rate of operative mortality following a MICS-mitral procedure was 1.1%. In Japan, 30-day and in-hospital mortality rates following such a procedure are similar to those in Western countries: 1.1% (isolated MV repair) in a report of the STS National Database15 and 1.1% in the present investigation using data from JCVSD. Other studies have also shown excellent results, with low 30-day mortality rates of approximately 1.0%,16–18 although the results were from individual experienced centers. The present findings were obtained from multiple centers throughout Japan and show the safety of MICS-mitral for patients undergoing MV surgery.
As for postoperative complications, one concern is the rate of re-exploration for bleeding, as that has been reported to occur more frequently in MICS-mitral patients.7,8,11–13 Another concern is postoperative stroke, as previous reports have shown that the risk of stroke significantly increases in patients who undergo a minimally invasive procedure,7,19 though other studies have found that use of a minimally invasive approach does not affect the rate of postoperative stroke.13 Because stroke is a devastating complication that has serious effects on postoperative quality of life, we have to be careful about this complication. Our investigation revealed that 1.0% of the patients in the study cohort experienced a stroke, which is nearly the same or slightly better than the results in previous review studies.7,19 In the present analysis, the rate of re-exploration for bleeding was 2.8%, which is comparable to previous findings, though some reports have noted lower rates.11,18
We consider that the MICS-mitral procedures conducted for the present patients were performed with acceptable safety. However, some patients experienced complications. Therefore, to achieve the goal of better outcomes following MICS-mitral, it will be important to examine the risks of mortality and serious complications using detailed analyses.
Factors Affecting Perioperative OutcomeFour risk factors were found to be associated with both operative mortality and composite outcome: age, respiratory dysfunction, myocardial infarction, and the number of annual cases. Others, such as thoracic aortic disease, BMI, NYHA, aortic regurgitation, MV replacement, and multiple valve surgery, were significant risk factors for operative mortality, while recent history of cerebrovascular disease, carotid artery disease, and tricuspid regurgitation were significant risk factors for composite outcome.
Age is consistently recognized as a risk for operative mortality in various analyses of cardiovascular surgery, including several large registries and risk calculator findings.10,20,21 In a previous study that compared the results of MICS-mitral with those of a conventional sternotomy using a propensity matched analysis, a higher rate of re-exploration for bleeding was seen in older patients (≥60 years old), which was speculated to be related to tissue fragility in the thoracic cavity.8 Therefore, care should be taken when performing MICS-mitral procedures for older patients. Respiratory dysfunction, recent history of myocardial dysfunction, and NYHA have also been noted as risk factors for mortality and morbidity in studies of EuroSCORE, STS scoring, and JapanSCORE.10,20,21 The present study noted respiratory dysfunction as a risk factor for MICS-mitral patients, which might be related to the requirement of a right thoracotomy. These ordinary factors should also be considered when performing MICS-mitral procedures.
Cardiac surgery for patients with recent cerebrovascular disease is obviously challenging, because of the risk of recurrent stroke or intracranial hemorrhage after performing cardiopulmonary bypass.22 Carotid artery disease is also well recognized as a risk factor for stroke,23 so it might be better to avoid performing MICS-mitral in affected patients. Additionally, tricuspid regurgitation has been noted as a risk factor of the composite outcome, which might reflect longer operative, cardiopulmonary bypass, and cardiac arrest times needed for tricuspid surgery, which in turn affects the outcome.
Interesting findings obtained in the present study include BMI >30 and aortic regurgitation as risk factors for death. Ailawadi et al24 reported their outline of current best practice regarding patient evaluation and selection for MICS-mitral as an expert opinion, which includes important disciplines for MICS-mitral complete myocardial protection and optimal exposure. Notably, surgery for obese patients with thick chest walls can be challenging because of the added distance to the MV. Furthermore, the approach in patients with more than mild aortic valve insufficiency should be done with caution through the right chest, as arresting, protecting, and decompressing the heart may be challenging. The present analysis found these factors to be significantly related to operative mortality, thus care should be taken when considering MICS-mitral in such patients.
Effects of Hospital VolumeThe volume of operations performed at a single institution is generally considered to affect operative results in cases of cardiac surgery,25 and most of the previously reported excellent outcomes for MICS-mitral have come from institutions with a large number of patients.8,11,16,18 Therefore, to achieve better results of MICS-mitral, this factor must be considered. As shown in the present study, most institutions in Japan perform this procedure relatively few times each year. Hospital volume was detected as a risk factor of the composite outcome, which confirms a previous investigation that found that hospital volume correlated with both operative mortality and morbidity. Based on our results showing an apparent relationship between hospital volume and outcome, it may be necessary to consider the effects of volume and adjust the number of institutions performing MICS-mitral procedures. This issue requires further investigation, because there are many MICS-mitral training programs and opportunities to be exposed to the procedure as a resident physician are increasing. We consider that careful patient selection should be implemented during the initial phase of the program, especially in small-volume institutions.
Risk ModelIn our analysis, the C-index value for operative mortality was 0.877, which indicates an acceptable risk (C-index ≥0.7). Although the C-index for the composite outcome was relatively low (0.665), it is considered to have value as a reference. The discriminatory ability of predicted operative mortality was comparable with that reported in a recent meta-analysis of EuroSCORE II, a multicenter database, regarding isolated and combined valve surgery, which showed C-index values ranging from 0.70 to 0.81.26
Study LimitationsFirst, because of its retrospective nature, the results are weaker compared with those obtained from studies with a randomized prospective design. Second, our investigation was primarily focused on detailed short-term outcomes and did not include long-term results. Third, precise information regarding the surgical procedures utilized was not available, including cannulation strategy and method used for MV surgery. Fourth, the JCVSD does not provide any data regarding postoperative status of MV regurgitation, which is a major limitation in this large national database. Unfortunately, the JCVSD also does not include detailed echocardiographic data or long-term results, thus it is impossible to evaluate those factors. On the other hand, it includes data from all cardiovascular surgical institutions in Japan and the results reflect real-world data obtained from each treated patient. As a result, we consider that the shortcomings of the available data are outweighed by the all-encompassing nature of this national database. Finally, because the study population was derived from a single reference health system, external validity is partially limited at the expense of enhancing internal validity.
MICS-mitral procedures were safely performed with acceptable outcomes in the present study cohort. Several risk factors, including age, respiratory dysfunction, myocardial infarction, and the number of annual cases, were found to be associated with both operative mortality and composite outcome. We were able to develop a risk calculator formula using the results obtained by this comprehensive multicenter analysis. In addition to its usefulness for patient selection, we consider that the present risk calculator formulas may influence future application of this procedure.
The authors thank Chieko Fujimura (Department of Healthcare Quality Assessment, The University of Tokyo, Graduate School of Medicine, Tokyo, Japan) for her assistance with data collection.
All authors declare no conflicts of interest.
