Propensity Score-Matched Analysis of Minimally Invasive Aortic Valve Replacement

Abstract

Background: Right mini-thoracotomy and partial sternotomy are widely recognized as effective approaches in minimally invasive aortic valve replacement (AVR). The aim of this study was to evaluate the objective benefits of the respective approaches compared to the conventional approach.

Methods and Results: A retrospective analysis was performed in 282 consecutive patients who underwent isolated and initial AVR at a single cardiovascular institute between May 2007 and December 2012. Mini-thoracotomy and partial sternotomy were performed in 62 (22%) and in 26 patients (9%), respectively. Propensity score matching produced 36 (mini-thoracotomy vs. full sternotomy) and 24 (partial sternotomy vs. full sternotomy) well-matched pairs. Compared to the conventional approach, mini-thoracotomy was associated with significantly shorter operative time (235±35 min vs. 272±73 min; P=0.009), lower prevalence of blood transfusion (42%, 15/36 vs. 67%, 24/36; P=0.025), and significantly shorter intensive care unit and postoperative hospital stay (1.4±0.8 days vs. 2.2±1.1 days, P=0.001; and 13.3±6.5 days vs. 21.5±10.3 days, P=0.001; respectively). There were no significant differences in operative and postoperative data between the partial sternotomy and full sternotomy groups.

Conclusions: The objective benefits of right mini-thoracotomy included early rehabilitation and lower prevalence of blood transfusion. Significant advantages of partial sternotomy were not found. (Circ J 2014; 78: 2876–2881)

Techniques for minimally invasive cardiac surgery have been evolving to avoid conventional full median sternotomy in aortic valve replacement (AVR). As minimally invasive AVR (MIAVR), the partial sternotomy and right mini-thoracotomy approaches are widely known, and various advantages of MIAVR have been reported compared to conventional AVR (CAVR) with full sternotomy.^1–3 Different indications for MIAVR, however, can be a cause of preoperative selection bias and it is difficult to prove the real benefits of MIAVR. Propensity score matching, therefore, has been used for comparison between CAVR and right mini-thoracotomy AVR (RTAVR), indicating significant efficacy of RTAVR.^4,5 But there are still few reports on the real benefits of MIAVR including partial sternotomy AVR (PSAVR) and RTAVR. Thus, we compared CAVR with RTAVR and PSAVR using propensity score matching to evaluate the objective benefits of RTAVR and PSAVR at a single center.

Methods

A retrospective analysis was performed in 282 consecutive patients who underwent isolated and initial AVR at a single cardiovascular institute between May 2007 and December 2012. We introduced RTAVR in May 2007, and 62 patients (22%) underwent RTAVR in this cohort. We did not select RTAVR for patients with marginal respiratory function, occlusive disease in the aorta, iliac and femoral artery, and extremely low activities of daily living. PSAVR was performed in 26 patients (9%) who were not candidates for RTAVR, but who still had good indications for early rehabilitation. Because there were significant differences in preoperative patient background, we used propensity score matching to reduce the influences of selection bias. This study was approved by the local institutional review board.

Surgical Technique

CAVR was performed utilizing the standard technique. In PSAVR, we primarily used the lower partial sternotomy approach from the xiphoid to the 2nd right intercostal space for cosmetic reasons. The RTAVR technique has been reported previously, and will be described briefly here.⁶ After intubation with a double lumen endotracheal tube for 1 lung ventilation, surgical access to the aortic valve is through the 3rd anterior intercostal space with a 5–6-cm skin incision. Ribs are not divided. Right femoral artery and vein are cannulated to establish cardiopulmonary bypass (CPB). Left atrial venting is initiated through the right upper pulmonary vein. After direct aortic cross-clamping by modified Chitwood sliding clamp with a greater curvature, cardiac arrest is easily achieved by selective deliveries of cardioplegic solution into both coronary orifices through transverse aortotomy. In patients without aortic insufficiency (AI), antegrade cardioplegic solution is given through aortic root cannula. After a prosthetic valve is sewn into place in standard fashion, the aortotomy is closed. De-airing procedure through the aortic root and left atrial vents is completed, and CPB is terminated.

Statistical Analysis

Continuous data are presented as mean±SD, and were analyzed using 2-tailed t-test, or Mann-Whitney test for independent data as appropriate. Categorical variables are given as count and percentage and were compared using chi-squared or Fisher’s exact test. Multiple comparisons between 3 groups (RTAVR, PSAVR and CAVR) were done using 2-tailed t-test or Mann-Whitney U-test for continuous variables, or chi-squared test for categorical variables with the Holm-Sidak correction. Multivariate stepwise linear regression analysis, including significant parameters identified on univariate analysis (at P<0.1), was used to evaluate the most significant determinants of postoperative length of intensive care unit (ICU) and postoperative hospital stay in MIAVR. To analyze the correlation between operation date and patient age, operative time, perfusion time, and aortic clamp time, and days from date when each procedure was initially introduced, were used. P-value after correction is used, and P<0.05 was considered significant. All data were analyzed using JMP 9.0 (SAS Institute, Cary, NC, USA).

In addition, we performed adjustment for significant differences in patient baseline characteristics with propensity score matching using a 1:1 nearest-neighbor-matching algorithm with a ±0.05 caliper and no replacement, yielding 36 (RTAVR vs. CAVR) and 24 (PSAVR vs. CAVR) propensity score-matched observations, respectively. The propensity score was estimated using a multivariate logistic regression model with indication for MIAVR as the variable, and the 20 baseline characteristics as covariates. The model fit and predictive power were measured with the C-statistic (0.88 and 0.74, respectively). Paired comparison of intraoperative and postoperative data were done using conditional logistic regression analysis for categorical variables and paired t-test for continuous variables.⁷

Results

RTAVR vs. CAVR

Preoperative Patient Characteristics RTAVR was performed in 62 patients (22%). Compared to the CAVR group (194 patients, 69%), significantly younger age, lower prevalence of female gender, smaller body surface area (BSA), lower prevalence of diabetes mellitus and higher creatinine clearance (CCr) were found in the RTAVR group. Additionally, significantly lower prevalence of aortic stenosis (AS; 39%, 24/62 vs. 75%, 146/194; P<0.001), higher grade of AI (2.8±1.4 vs. 2.2±1.2; P=0.003) and better New York Heart Association (NYHA) class (1.5±0.8 vs. 2.0±1.0; P=0.003) were observed in the RTAVR group. The Society of Thoracic Surgeons (STS) score was significantly lower in the RTAVR group compared to the CAVR group (mortality: 0.87±0.72 vs. 2.27±1.69; P<0.001; mortality or morbidity: 9.1±5.0 vs. 14.5±7.6; P<0.001). Table 1 lists the unmatched comparisons. Propensity score matching produced 36 (RTAVR vs. CAVR) matched pairs. Table 1 lists comparisons of preoperative patient background after propensity score matching. There were no significant differences in all covariates including STS risk score in the pairs, and the models fitted well.

Table 1. Preoperative Characteristics: RTAVR vs. CAVR

Variables	Unmatched comparison			Matched comparison
	RTAVR (n=62)	CAVR (n=194)	P-value	RTAVR (n=36)	CAVR (n=36)	P-value
Age (years)	56.3±13.8	73.6±11.0	<0.001	62.6±10.7	62.5±14.0	0.963
Female	22 (35)	107 (55)	0.021	20 (56)	19 (53)	1.000
BSA (m2)	1.72±0.20	1.53±0.19	<0.001	1.66±0.17	1.61±0.19	0.223
Hypertension	30 (48)	126 (65)	0.060	20 (56)	23 (64)	0.631
Hyperlipidemia	12 (19)	58 (30)	0.315	8 (22)	8 (22)	1.000
DM	1 (2)	41 (21)	0.001	1 (3)	1 (3)	1.000
COPD	0 (0)	9 (5)	0.168	0 (0)	0 (0)	–
CAD	2 (3)	12 (6)	0.744	2 (6)	0 (0)	0.493
PVD	0 (0)	7 (4)	0.258	0 (0)	2 (6)	0.493
CAS	0 (0)	6 (3)	0.483	0 (0)	0 (0)	–
HD	2 (3)	13 (7)	0.621	1 (3)	2 (6)	1.000
eCCr (ml/min)	88.3±37.1	54.9±29.8	<0.001	77.1±31.5	78.6±39.0	0.859
AF	3 (5)	7 (4)	0.663	2 (6)	4 (11)	0.674
IE	3 (5)	7 (4)	0.663	2 (6)	4 (11)	0.674
Urgent+emergency	0 (0)	5 (3)	0.403	0 (0)	1 (3)	1.000
AS	24 (39)	146 (75)	<0.001	16 (44)	18 (50)	0.814
AI (0–4)	2.8±1.4	2.2±1.2	0.003	2.6±1.4	2.3±1.4	0.381
MR (0–4)	1.3±1.0	1.3±0.8	0.999	1.4±1.1	1.0±0.9	0.146
LVEF (%)	65.4±8.3	63.2±13.8	0.700	64.3±8.8	66.7±6.1	0.184
NYHA class	1.5±0.8	2.0±1.0	0.003	1.6±0.9	1.9±0.9	0.250
STS score (mortality)	0.87±0.72	2.46±2.33	<0.001	1.08±0.84	1.30±0.98	0.346
STS score (morbidity or mortality)	9.1±5.0	14.9±7.5	<0.001	10.3±5.9	11.3±5.2	0.485

Data given as n (%) or mean±SD. AF, atrial fibrillation; AI, aortic insufficiency; AS, aortic stenosis; BSA, body surface area; CAD, coronary artery disease; CAS, carotid artery stenosis; CAVR, conventional aortic valve replacement; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; eCCr, estimated creatinine clearance; HD, hemodialysis; IE, infectious endocarditis; LVEF, left ventricular ejection fraction; MR, mitral regurgitation; NYHA, New York Heart Association; PVD, peripheral vascular disease; RTAVR, right mini-thoracotomy aortic valve replacement; STS, Society of Thoracic Surgeons.

Operative and Postoperative Data On unmatched comparison, significantly prolonged CPB and aortic clamping time (136±31 min vs. 110±32 min, P<0.001; and 95±23 min vs. 80±22 min, P<0.001; respectively), and larger size of prosthetic valve (22.1±2.1 mm vs. 20.8±1.9 mm; P<0.001) were seen in the RTAVR group, compared to the CAVR group. The prevalence of blood transfusion was significantly lower in the RTAVR group (P<0.001). Overall 30-day mortality was 1.1% (3/282), and there was 1 death as a result of bleeding from the intercostal artery in the RTAVR group. There were no significant differences in postoperative major complications between all groups. Significantly shorter length of ICU and postoperative hospital stays were observed in the RTAVR group compared to the CAVR group (1.3±0.7 days vs. 2.4±1.5 days, P<0.001; and 13.3±5.6 days vs. 20.5±9.1 days, P<0.001; respectively), and CCr significantly decreased in the CAVR group.

Regarding the matched comparison between RTAVR and CAVR, significantly shorter operative time (235±35 min vs. 272±73 min; P=0.009) and lower prevalence of blood transfusion (42%, 15/36 vs. 67%, 24/36; P=0.025) were found in the RTAVR group. Additionally, significantly shorter ICU and postoperative hospital stays were observed in the RTAVR group (1.4±0.8 days vs. 2.2±1.1 days, P=0.001; and 13.3±6.5 days vs. 21.5±10.3 days, P=0.001; respectively). Table 2 lists the unmatched and matched comparisons of operative and postoperative data.

Table 2. Operative and Postoperative Data: RTAVR vs. CAVR

Variables	Unmatched comparison			Matched comparison
	RTAVR (n=62)	CAVR (n=194)	P-value	RTAVR (n=36)	CAVR (n=36)	P-value
Operative data
Operative time (min)	244±43	239±60	0.999	235±35	272±73	0.009
CPB time (min)	136±31	110±32	<0.001	129±28	125±47	0.808
Aortic clamping time (min)	95±23	80±22	<0.001	91±22	90±30	0.945
Blood transfusion	20 (32)	155 (80)	<0.001	15 (42)	24 (67)	0.025
Prosthetic valve size (mm)	22.1±2.1	20.8±1.9	<0.001	22.1±2.1	21.6±2.4	0.448
Postoperative data
30-day death	1 (2)	2 (1)	0.711	0 (0)	0 (0)	–
Stroke	1 (2)	1 (1)	1.000	1 (3)	0 (0)	0.368
Re-exploration for bleeding	2 (3)	2 (1)	0.675	1 (3)	0 (0)	0.368
Reintubation	2 (3)	4 (2)	1.000	1 (3)	0 (0)	0.368
Initial ventilation time (h)	4.7±5.5	8.2±16.2	0.264	5.4±7.0	8.0±13.6	0.282
Complete AV block	1 (2)	4 (2)	0.824	1 (3)	1 (3)	1.000
Serious surgical site infection	1 (2)	2 (1)	1.000	1 (3)	0 (0)	0.368
ICU stay (days)	1.3±0.7	2.4±1.5	<0.001	1.4±0.8	2.2±1.1	0.001
Hospital stay (days)	13.3±5.6	20.5±9.1	<0.001	13.3±6.5	21.5±10.3	0.001
Maximum eCCr (ml/min)	82.1±30.5	46.5±24.2	<0.001	72.6±25,8	63.8±30.7	0.214
Change in eCCr (%)	2.1±64.2	−12.9±20.6	0.017	8.2±80.3	−17.0±18.2	0.089

Data given as n (%) or mean±SD. AV, atrioventricular; CPB, cardiopulmonary bypass; ICU, intensive care unit. Other abbreviations as in Table 1.

PSAVR vs. CAVR

PSAVR was performed in 26 patients (9%). There was a significantly higher prevalence of atrial fibrillation (15%, 4/26 vs. 4%, 7/194; P=0.029) and lower grade of AI (1.7±1.4 vs. 2.2±1.2; P=0.049) in the PSAVR group compared to CAVR. There were no significant differences in STS score between PSAVR and CAVR. Propensity score matching produced 24 (PSAVR vs. CAVR) matched pairs, and there were no significant differences in all covariates (Table 3).

Table 3. Preoperative Characteristics: PSAVR vs. CAVR

Variables	Unmatched comparison			Matched comparison
	PSAVR (n=26)	CAVR (n=194)	P-value	PSAVR (n=24)	CAVR (n=24)	P-value
Age (years)	71.8±14.8	73.6±11.0	0.495	74.6±9.5	70.8±13.2	0.253
Female	12 (46)	107 (55)	0.387	12 (50)	8 (33)	0.380
BSA (m2)	1.51±0.24	1.53±0.19	0.669	1.49±0.24	1.49±0.16	0.995
Hypertension	17 (65)	126 (65)	0.965	16 (67)	18 (75)	0.752
Hyperlipidemia	6 (23)	58 (30)	0.944	6 (25)	7 (29)	1.000
DM	3 (12)	41 (21)	0.251	3 (13)	2 (8)	1.000
COPD	2 (8)	9 (5)	0.502	2 (8)	2 (8)	1.000
CAD	2 (8)	12 (6)	0.768	2 (8)	2 (8)	1.000
PVD	1 (4)	7 (4)	0.952	1 (4)	0 (0)	1.000
CAS	0 (0)	6 (3)	0.726	0 (0)	0 (0)	–
HD	3 (12)	13 (7)	0.372	2 (8)	1 (4)	1.000
eCCr (ml/min)	60.8±47.4	54.9±29.8	0.401	55.0±26.9	63.3±30.6	0.322
AF	4 (15)	7 (4)	0.029	3 (13)	2 (8)	1.000
IE	0 (0)	7 (4)	0.649	0 (0)	0 (0)	–
Urgent+emergency	1 (4)	5 (3)	0.709	0 (0)	1 (4)	1.000
AS	19 (73)	146 (75)	0.809	18 (75)	14 (58)	0.359
AI (0–4)	1.7±1.4	2.2±1.2	0.049	1.8±1.4	2.3±1.4	0.171
MR (0–4)	1.2±0.9	1.3±0.8	0.999	1.2±0.9	1.3±0.9	0.887
LVEF (%)	64.7±11.0	63.2±13.8	0.999	65.4±10.6	65.4±11.7	0.990
NYHA class	2.2±1.1	2.0±1.0	0.360	2.2±1.0	2.2±0.8	0.801
STS score (mortality)	2.27±1.69	2.46±2.33	0.596	2.44±2.33	1.81±1.19	0.261
STS score (morbidity or mortality)	14.5±7.6	14.9±7.5	0.834	14.2±6.8	12.6±5.0	0.370

Data given as n (%) or mean±SD. PSAVR, partial sternotomy aortic valve replacement. Other abbreviations as in Table 1.

In the unmatched comparison, the prevalence of blood transfusion was significantly lower in the PSAVR group (P=0.035). There were no significant differences, however, in operative and postoperative data in the matched comparison between the PSAVR and CAVR groups (Table 4).

Table 4. Operative and Postoperative Data: PSAVR vs. CAVR

Variables	Unmatched comparison			Matched comparison
	PSAVR (n=26)	CAVR (n=194)	P-value	PSAVR (n=24)	CAVR (n=24)	P-value
Operative data
Operative time (min)	245±41	239±60	0.999	239±30	222±48	0.172
CPB time (min)	115±27	110±32	0.419	110±18	103±27	0.298
Aortic clamping time (min)	87±22	80±22	0.167	83±19	77±24	0.302
Blood transfusion	16 (62)	155 (80)	0.035	15 (63)	17 (71)	0.480
Prosthetic valve size (mm)	21.0±2.0	20.8±1.9	0.573	21.0±2.0	20.5±2.0	0.475
Postoperative data
30-day death	0 (0)	2 (1)	1.000	0 (0)	0 (0)	–
Stroke	0 (0)	1 (1)	0.719	0 (0)	0 (0)	–
Re-exploration for bleeding	0 (0)	2 (1)	0.603	0 (0)	0 (0)	–
Reintubation	1 (4)	4 (2)	1.000	1 (4)	0 (0)	0.368
Initial ventilation time (h)	7.3±13.1	8.2±16.2	0.744	4.6±2.7	4.7±2.4	0.920
Complete AV block	1 (4)	4 (2)	1.000	1 (4)	0 (0)	0.368
Serious surgical site infection	0 (0)	2 (1)	0.603	0 (0)	0 (0)	–
ICU stay (days)	2.4±1.6	2.4±1.5	0.893	2.3±1.5	1.9±0.7	0.187
Hospital stay (days)	18.3±6.1	20.5±9.1	0.199	17.8±6.1	17.3±5.5	0.722
Maximum eCCr (ml/min)	53.5±36.5	46.5±24.2	0.215	49.8±23.0	55.7±27.0	0.470
Change in eCCr (%)	−8.1±35.0	−12.9±20.6	0.519	−6.4±35.9	−10.5±19.4	0.635

Data given as n (%) or mean±SD. Abbreviations as in Tables 1–3.

Risk Factors for Length of ICU and Hospital Stay in MIAVR

Uni- and multivariate analysis was done to identify risk factors for prolonged ICU and postoperative hospital stay in MIAVR, including PSAVR and RTAVR. Univariate linear regression test isolated age, BSA, CCr, NYHA class, prosthetic valve size, operative time, lowest core temperature, hypertension, dialysis, AS, PSAVR, and blood transfusion as correlated factors with postoperative ICU stay. On multivariate analysis, smaller BSA, dialysis, PSAVR and extended operative time were identified as risk factors for prolonged ICU stay in MIAVR (P=0.023, 0.023, 0.025 and 0.022, respectively), whereas age, female gender, BSA, CCr, atrial fibrillation, AS, PSAVR, prosthetic valve size and blood transfusion were isolated as correlated factors with length of postoperative hospital stay on univariate testing. On multivariate linear regression analysis smaller BSA and AS were risk factors for extended postoperative hospital stay in MIAVR (P=0.003 and 0.010, respectively; Table 5).

Table 5. Risk Factors for ICU and Hospital Stay in MIAVR

Variables	Univariate analysis		Multivariate analysis
	R2	P-value	β	P-value
ICU stay
Age	0.117	0.001
BSA	0.125	<0.001	−0.230	0.023
Hypertension (+)	0.046	0.044
HD (+)	0.088	0.005	0.210	0.023
eCCr	0.119	0.001
NYHA class	0.117	0.001
AS (+)	0.043	0.054
PSAVR (vs. RTAVR)	0.182	<0.001	0.235	0.025
Prosthetic valve size	0.108	0.002
Operative time	0.046	0.044	0.212	0.022
Lowest core temperature	0.049	0.043
Blood transfusion (+)	0.069	0.014
Postoperative hospital stay
Age	0.082	0.007
Female (vs. Male)	0.039	0.067
BSA	0.159	<0.001	−0.399	0.003
eCCr	0.041	0.064
AF	0.041	0.059
AS (+)	0.139	<0.001	0.276	0.010
PSAVR (vs. RTAVR)	0.139	<0.001
Prosthetic valve size	0.139	<0.001
RBC transfusion (+)	0.080	0.008

RBC, red blood cellt. Other abbreviations as in Tables 1–3.

Discussion

The partial sternotomy approach has been recognized as an effective option for AVR to reduce blood transfusion and improve postoperative course.^1,2,8 Brown et al, however, in their meta-analysis of 4 randomized trials, noted that there were no significant advantages of PSAVR in early mortality, or operative and postoperative data.^1,9–12 Thus, the advantages of partial sternotomy for AVR are still controversial. Although the efficacy of RTAVR has been reported, a lack of prospective randomized trials comparing CAVR and preoperative significant selection bias make it difficult to confirm the real benefits of RTAVR.^13,14 Therefore, propensity score was used for matched comparison between RTAVR and CAVR, and shorter ventilation time, ICU and hospital stay, lower incidence of AF and blood transfusion have been reported as advantages of RTAVR.^4,5 In these reports, PSAVR and RTAVR were evaluated separately, and the outcomes of PSAVR and RTAVR were not simultaneously compared with the conventional approach within a single institute.

In the present study, comparison of postoperative outcome between RTAVR, PSAVR and CAVR was done at a single center. On unmatched comparison, there were significant differences in preoperative characteristics between RTAVR and CAVR, given that we initially introduced RTAVR for younger patients suffering from AI. Therefore, we used propensity score to adjust for significant differences in preoperative patient background, and significantly shorter operative time, lower prevalence of blood transfusion, significantly shorter ICU and postoperative hospital stay were observed in the RTAVR group compared to the CAVR group. Additionally, significantly longer CPB and aortic clamping time seen on unmatched comparison were not observed in the matched comparison, and operative time was significantly shorter in the RTAVR group. The faster closing time in mini-thoracotomy may be responsible for these results, but there were no significant differences in operative and postoperative data in the PSAVR group compared with the CAVR group. Additionally, PSAVR can be a significant cause of prolonged ICU stay compared to RTAVR. Regarding these findings, while objective advantages were not found for partial sternotomy, actual benefits for right mini-thoracotomy approach were identified. And smaller BSA, dialysis, AS and extended operative time were identified as correlated indices with length of ICU and postoperative hospital stay in MIAVR. Considering that 1 of the most important objectives of MIAVR is an early rehabilitation, these factors should be considered as crucial in the indications for MIAVR.

Sutureless AVR has been evolving, and satisfactory results including excellent hemodynamic performance, shortened surgical time and improvement of outcome have been reported in high-risk patients or minimally invasive procedures.^15,16 Sutureless AVR is reported to be an effective option for high-risk patients and can reduce the occurrence of paravalvular leakage compared to transcatheter AVR (TAVR).^17,18 Considering the significant advantages of RTAVR, sutureless RTAVR can be an effective alternative with broad indications. Although RTAVR was initially introduced for patients without preoperative risk factors, the indications have been expanding and it can be an effective option for high-risk patients, as well as TAVR.^19,20

There were several limitations in this study. First, this was a non-randomized retrospective observational study. Second, propensity score matching was used to reduce selection bias, but the matching is limited and arbitrariness was not fully denied. The greater prevalence of AS in the RTAVR group and surgeon bias may be influential factors. Additionally, a small sample of propensity score-matched pairs remained due to preoperative significant differences in characteristics, particularly in the results of partial sternotomy. And we could not evaluate the effects of upper partial sternotomy AVR as opposed to a lower partial sternotomy. Finally, propensity score-matched comparison between the RTAVR and PSAVR groups was not done directly because only a limited number of pairs can be matched on propensity score.

Conclusions

On propensity score matching to reduce selection bias, objective benefits of RTAVR including early rehabilitation and lower prevalence of blood transfusion were identified. Significant advantages of PSAVR over CAVR, however, were not found, and PSAVR was identified as a correlated factor with length of ICU stay in comparison to RTAVR.

References

• 1.    Tabata  M,  Fukui  T,  Takanashi  S. Do minimally invasive approaches improve outcomes of heart valve surgery? Circ J 2013; 77: 2232–2239.
• 2.    Bakir  I,  Casselman  FP,  Wellens  F,  Jeanmart  H,  De Geest  R,  Degrieck  I, et al. Minimally invasive versus standard approach aortic valve replacement: A study in 506 patients. Ann Thorac Surg 2006; 81: 1599–1604.
• 3.    Glauber  M,  Miceli  A,  Bevilacqua  S,  Farneti  PA. Minimally invasive aortic valve replacement via right anterior minithoracotomy: Early outcomes and midterm follow-up. J Thorac Cardiovasc Surg 2011; 142: 1577–1579.
• 4.    Brinkman  WT,  Hoffman  W,  Dewey  TM,  Culica  D,  Prince  SL,  Herbert  MA, et al. Aortic valve replacement surgery: Comparison of outcomes in matched sternotomy and PORT ACCESS groups. Ann Thorac Surg 2010; 90: 131–135.
• 5.    Glauber  M,  Miceli  A,  Gilmanov  D,  Ferrarini  M,  Bevilacqua  S,  Farneti  PA, et al. Right anterior minithoracotomy versus conventional aortic valve replacement: A propensity score matched study. J Thorac Cardiovasc Surg 2013; 145: 1222–1226.
• 6.    Hiraoka  A,  Kuinose  M,  Chikazawa  G,  Katayama  K,  Totsugawa  T,  Yoshitaka  H. Minimally invasive aortic valve replacement surgery: Comparison of port-access and conventional standard approach. Circ J 2011; 75: 1656–1660.
• 7.    Austin  PC. Propensity-score matching in the cardiovascular surgery literature from 2004 to 2006: A systematic review and suggestions for improvement. J Thorac Cardiovasc Surg 2007; 134: 1128–1135.
• 8.    Mihaljevic  T,  Cohn  LH,  Unic  D,  Aranki  SF,  Couper  GS,  Byrne  JG. One thousand minimally invasive valve operations: Early and late results. Ann Surg 2004; 240: 529–534.
• 9.    Brown  ML,  McKellar  SH,  Sundt  TM,  Schaff  HV. Ministernotomy versus conventional sternotomy for aortic valve replacement: A systematic review and meta-analysis. J Thorac Cardiovasc Surg 2009; 137: 670–679.
• 10.    Aris  A,  Camara  ML,  Montiel  J,  Delgado  LJ,  Galan  J,  Litvan  H. Ministernotomy versus median sternotomy for aortic valve replacement: A prospective, randomized study. Ann Thorac Surg 1999; 67: 1583–1587.
• 11.    Machler  HE,  Bergmann  P,  Anelli-Monti  M,  Dacar  D,  Rehak  P,  Knez  I, et al. Minimally invasive versus conventional aortic valve operations: A prospective study in 120 patients. Ann Thorac Surg 1999; 67: 1001–1005.
• 12.    Bonacchi  M,  Prifti  E,  Giunti  G,  Frati  G,  Sani  G. Does ministernotomy improve postoperative outcome in aortic valve operation?: A prospective randomized study. Ann Thorac Surg 2002; 73: 460–465.
• 13.    Plass  A,  Scheffel  H,  Alkadi  H,  Kaufmann  P,  Genoni  M,  Falk  V, et al. Aortic valve replacement through a minimally invasive approach: Preoperative planning, surgical technique and outcome. Ann Thorac Surg 2009; 88: 1851–1856.
• 14.    Glower  DD,  Teng  L,  Desai  B. Aortic valve replacement through right minithoracotomy in 306 consecutive patients. Innovations 2010; 5: 326–330.
• 15.    Cerillo  AG,  Bevilacqua  S,  Farneti  PA,  Concistrè  G,  Glauber  M. Sutureless aortic valve replacement through a right minithoracotomy. J Heart Valve Dis 2012; 21: 168–171.
• 16.    Santarpino  G,  Pfeiffer  S,  Concistré  G,  Grossmann  I,  Hinzmann  M,  Fischlein  T. The Perceval S aortic valve has the potential of shortening surgical time: Does it also result in improved outcome? Ann Thorac Surg 2013; 96: 77–82.
• 17.    Lorusso  R,  Gelsomino  S,  Renzulli  A. Sutureless aortic valve replacement: An alternative to transcatheter aortic valve implantation? Curr Opin Cardiol 2013; 28: 158–163.
• 18.    D’Onofrio  A,  Messina  A,  Lorusso  R,  Alfieri  OR,  Fusari  M,  Rubino  P, et al. Sutureless aortic valve replacement as an alternative treatment for patients belonging to the “gray zone” between transcatheter aortic valve implantation and conventional surgery: A propensity-matched, multicenter analysis. J Thorac Cardiovasc Surg 2012; 144: 1010–1016.
• 19.    Kobayashi  J. Changing strategy for aortic stenosis by transcatheter valve treatment in Japan. Circ J 2013; 77: 309–310.
• 20.    Sawa  Y,  Saito  S,  Kobayashi  J,  Niinami  H,  Kuratani  T,  Maeda  K, et al; MDT-2111 Japan Investigators. First clinical trial of a self-expandable transcatheter heart valve in Japan in patients with symptomatic severe aortic stenosis. Circ J 2014; 78: 1083–1090.