Effect of Recurrent Mitral Regurgitation Following Degenerative Mitral Valve Repair

JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY

VOL. 67, NO. 5, 2016

ª 2016 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

ISSN 0735-1097/$36.00

PUBLISHED BY ELSEVIER

http://dx.doi.org/10.1016/j.jacc.2015.10.098

Effect of Recurrent Mitral Regurgitation Following Degenerative Mitral Valve Repair Long-Term Analysis of Competing Outcomes Rakesh M. Suri, MD, DPHIL,a Marie-Annick Clavel, DVM, PHD,b,c Hartzell V. Schaff, MD,a Hector I. Michelena, MD,b Marianne Huebner, PHD,d Rick A. Nishimura, MD,b Maurice Enriquez-Sarano, MDd

ABSTRACT BACKGROUND The risk for and consequences of recurrent mitral regurgitation (MR) following degenerative mitral valve repair are poorly understood. OBJECTIVES This study sought to examine recurrent MR risk along with reoperation and survival rates. METHODS We evaluated patients undergoing primary mitral repair for isolated degenerative MR over 1 decade. Median follow-up was 11.5 years (interquartile range: 9.2 to 13.6 years) and was 99% complete. Multivariate analysis of postrepair MR recurrence employed Cox proportional hazards and multistate modeling. RESULTS A total of 1,218 patients met the study criteria; the mean age was 64
13 years, mean ejection fraction was 63
9%, and 864 (71%) patients were men. Prolapse was posterior in 62%, bileaflet in 26%, and anterior in 12%. The 15-year incidence of recurrent MR (i.e., MR $2) was 13.3%, incidence of mitral reoperation was 6.9%, and overall mortality was 44.0%. Repair before 1996 independently predicted MR recurrence (hazard ratio: 1.52). Additional determinants were: age, mild intraoperative residual MR, anterior leaflet prolapse, bileaflet prolapse, perfusion time >90 min, and lack of annuloplasty. Recurrence of moderate or greater MR was associated with adverse left ventricular remodeling and increased likelihood of death (hazard ratio: 1.72). Among those undergoing repair after 1996, MR recurrence rate was 1.5 per 100 patient-years during the first year post-repair, decreasing markedly to 0.9 thereafter. CONCLUSIONS Our study demonstrated that recurrent MR following degenerative mitral valve repair is associated with adverse left ventricular remodeling and late death. The incidence of MR recurrence decreases markedly following the first year after intervention. A transparent discussion of recurrent MR risk has pressing relevance when referring patients with complex mitral valve prolapse. (J Am Coll Cardiol 2016;67:488–98) © 2016 by the American College of Cardiology Foundation.

E Listen to this manuscript’s

arly surgical correction of severe mitral regur-

Current consensus statements mandate “rescue”

gitation (MR) caused by prolapse due to flail

MR correction in the presence of left ventricular

leaflets improves long-term survival and

(LV) dysfunction or symptoms (8). In the absence

diminishes late heart failure risk (1,2), particularly

of these Class I triggers, early “restorative” surgery

when performed by valve repair specialists within

is

a center of excellence (3–5). Mitral valve (MV) repair

normal in the presence of atrial fibrillation (AF) or

is safe and preferred over replacement in correcting

pulmonary hypertension, or when performed at cen-

MR caused by degenerative valve disease (6,7).

ters where the procedural risk of mortality is <1%

advocated

to

improve

patient

audio summary by JACC Editor-in-Chief Dr. Valentin Fuster.

From the aDivision of Cardiovascular Surgery, Mayo Clinic College of Medicine, Rochester, Minnesota; bDivision of Cardiovascular Diseases, Mayo Clinic College of Medicine, Rochester, Minnesota; cInstitut Universitaire de Cardiologie et de Pneumologie, Université Laval, Québec City, Québec, Canada; and the dDepartment of Statistics and Probability, Michigan State University, East Lansing, Michigan. Dr. Enriquez-Sarano has received a research grant from Edwards. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Suri and Clavel contributed equally to this work. Manuscript received August 6, 2015; revised manuscript received October 16, 2015, accepted October 27, 2015.

prognosis

to

JACC VOL. 67, NO. 5, 2016

Suri et al.

FEBRUARY 9, 2016:488–98

Consequences of Recurrent Regurgitation Following Mitral Repair

and the MV repair rate is >95% (9). Prior to taking

reversion to replacement in the current era (4),

ABBREVIATIONS

the step to generalize the recommendation for early

MR recurrence (not merely reoperation) rates

AND ACRONYMS

MR correction in the presence of Class IIa indica-

from multisurgeon practices must be better

tions, however, it is critically important to under-

understood to formulate guidelines applicable

stand

MV

to community cardiology practices world-

repair, along with the consequences and predictors

wide. Additionally, although mechanisms of

of this occurrence.

and therapeutic approaches to address recur-

recurrent

MR

rates

following

rent MR following prior repair have been

SEE PAGE 499

described (12,13), the effects on LV remodeling

Although prior series have demonstrated that

LV = left ventricular LVEF = left ventricular ejection fraction

MR = mitral regurgitation MV = mitral valve NYHA = New York Heart Association

and long-term life expectancy remain poorly

reoperation rate following degenerative MV repair is

defined. A final important limitation of prior work has

approximately 0.5% to 1% per year (7,10), assessing

been the inability to account for the attrition of pa-

durability on the basis of reintervention alone likely

tients during follow-up due to late death. The true

underestimates the long-term patient risk. While

incidence of recurrent MR and its determinants may

expert single surgeon series (3–5) have suggested that

therefore not be fully appreciated (13–15).

it is possible to repair degenerative mitral prolapse

We hypothesized that the use of multistate

with near 100% certainty (3,11) and infrequent

modeling to account for the competing risk of death

F I G U R E 1 Echocardiographic Follow-Up

Echo performed at Mayo Clinic Echo performed by treating physician Death 0.16% 0.08% 0.08%

Missing Echo End of Follow-up

100

1.23%

2.39%

11.65%

6.88%

2.72% 3.45%

2.46%

7.96%

17.24%

10.36% 4.13%

Percentage of Patients, (%)

80

43.92% 28.15%

29.64% 34.56%

60

39.48% 99.68%

40

4.76% 31.49% 56.51%

33.99% 48.21% 37.11%

20

24.38%

7.81% 9.52%

0 Mitral Valve Repair

30 Days

1 Year

5 Years

10 Years

15 Years

Follow-Up Time Although initially nearly all patients had an echocardiogram (echo) at Mayo Clinic following mitral valve repair, over time, patients more frequently underwent echocardiography at home institutions. The percentages of patients unavailable for an echocardiogram in a given period, had not yet returned, or had died all increased with time.

489

490

Suri et al.

JACC VOL. 67, NO. 5, 2016

Consequences of Recurrent Regurgitation Following Mitral Repair

FEBRUARY 9, 2016:488–98

METHODS

T A B L E 1 Baseline Pre-Operative Characteristics

Whole Cohort (n ¼ 1,218)

MR Recurrence (n ¼ 133)

No MR Recurrence (n ¼ 1,085)

p Value

Age, yrs

64
13

66
12

64
13

0.07

nesota. Eligible patients were those who underwent

Female

354 (29)

39 (29)

315 (29)

0.94

primary, isolated MV repair for pure MR (no stenosis),

Atrial fibrillation

308 (25)

37 (28)

271 (25)

0.48

and who had degenerative disease with surgically

Hypertension

421 (35)

60 (45)

361 (33)

0.008

59 (5)

12 (9)

47 (4)

0.03

verified MV prolapse as cause of regurgitation.

CAD

319 (26)

40 (30)

269 (26)

0.28

NYHA functional class III–IV

366 (30)

44 (33)

322 (30)

0.42

LVEDD, mm

59.9
7.4

60.2
8.0

59.8
7.3

0.69

repair; tricuspid valve replacement; previous mitral,

LVESD, mm

36.8
6.8

36.7
7.2

36.8
6.8

0.87

aortic, or tricuspid valve repair or replacement;

63
9

63
8

63
9

0.73

concomitant congenital (other than closure of patent

LV mass, g

265
68

272
65

264
68

0.40

foramen ovale), pericardial, or myocardial (particu-

Ruptured chordae

768 (64)

73 (55)

695 (65)

0.04

We studied consecutive patients who underwent degenerative MV repair between January 1, 1990, and December 31, 2000, at Mayo Clinic in Rochester, Min-

Clinical data

Diabetes

We excluded patients who had: mitral stenosis by hemodynamic assessment or surgical evaluation of the lesions; concomitant aortic valve replacement or

Echocardiographic data

LVEF, %

<0.0001

Prolapse localization

larly with dilated or hypertrophic cardiomyopathy) disease; ischemic MR with or without papillary mus-

Posterior leaflet only

748 (62)

55 (42)

693 (65)

Anterior leaflet only

137 (12)

29 (23)

108 (10)

cle rupture; or organic, nondegenerative MR, such as

Both leaflets

315 (26)

47 (35)

268 (25)

rheumatic heart disease, endocarditis, or miscellaneous causes. We did not exclude patients who

Values are mean
SD or n (%). CAD ¼ coronary artery disease; LV ¼ left ventricular; LVEDD ¼ left ventricular end-diastolic diameter; LVEF ¼ left ventricular ejection fraction; LVESD ¼ left ventricular end-systolic diameter; MR ¼ mitral regurgitation; NYHA ¼ New York Heart Association.

required

tricuspid

valve

repair

for

functional

tricuspid regurgitation, Cox-Maze operation for AF, or coronary bypass for obstructive coronary disease. The study was approved by our institutional re-

would alter prior findings of the long-term outcomes

view board, and informed consent was obtained from

of degenerative MV repair. We thus analyzed out-

study participants.

comes from a multisurgeon, “center of excellence”

ECHOCARDIOGRAPHY. Transthoracic

heart valve practice to determine mortality-adjusted,

grams were performed within routine clinical practice

post–degenerative valve repair recurrent MR rates,

using standard methods. LV assessment employed

determinants, and consequences.

parasternal long-axis views by 2-dimensional direct

echocardio-

measurements or guided M-mode at end-diastole and -systole and measurement of left ventricular

T A B L E 2 Operative and Early Post-Operative Characteristics

ejection fraction (LVEF) as well as LV mass (16,17). Whole Cohort (n ¼ 1,218)

MR Recurrence (n ¼ 133)

No MR Recurrence (n ¼ 1,085)

Surgery year <1996

459 (38)

68 (51)

391 (36)

0.0008

by Doppler echocardiography using comprehensive

Bypass time, min

71
39

86
50

69
37

<0.0001

MR assessment on the basis of systematic collection

Bypass time >90 min

294 (24)

51 (38)

245 (23)

0.0001

of specific and supportive signs with MR quantitation

CABG

914 (25)

43 (32)

261 (24)

0.06

747 (61)

60 (45)

687 (63)

<0.0001

(18) as judged feasible by the responsible physician.

Mitral resection Mitral plication

324 (27)

43 (32)

281 (26)

0.11

Chordal transfer

50 (4)

3 (2)

47 (4)

0.22

Artificial chordae

13 (1)

23 (17)

12 (1)

p Value

views. MR severity was assessed on a scale from 1 to 4

Operative

Annuloplasty No annuloplasty Annuloplasty

45 (4)

12 (9)

1,173 (96)

121 (91)

IABP Inotropes >1 day Residual mild MR in the OR

Immediate post-repair echocardiographic findings were confirmed by a second post-operative echocar-

0.0006

diography performed within 1 month post-repair in

0.007

1,215 (>99%) patients. MR that was mild or greater in hospital was considered “residual MR.” Subsequent

33 (3)

echocardiographic follow-up was performed in >85%

1,052 (97)

Early post-operative Post-operative MI

Left atrial diameter was measured using parasternal

of alive patients for each time point between 0 to 4 (0.3)

0 (0)

4 (0.4)

0.34

17 (1.4)

0 (0)

17 (1.6)

0.05

138 (11)

22 (17)

116 (11)

0.06

27 (2)

8 (6)

19 (2)

0.007

10 years and 80% between 10 to 15 years (Figure 1). Echocardiographic

data

were

used

as

collected

without subsequent modification. FOLLOW-UP. Patients were followed by their per-

Values are n (%) or mean
SD.

sonal physicians at a Mayo Clinic facility or at the

CABG ¼ coronary artery bypass grafting; IABP ¼ intra-aortic balloon pump; MI ¼ myocardial infarction; OR ¼ operating room; other abbreviations as in Table 1.

patient’s home institution. Information on follow-up events was obtained from medical examination or

JACC VOL. 67, NO. 5, 2016

Suri et al.

FEBRUARY 9, 2016:488–98

Consequences of Recurrent Regurgitation Following Mitral Repair

direct patient interview by the research team or use of repeated

follow-up

letters

and

questionnaires.

491

T A B L E 3 Echocardiographic LV Characteristics at Latest

Follow-Up

Follow-up echocardiographic reports from all sources Recurrent MR

No Recurrent MR

p Value

Interventricular septum thickness, mm

12.6
3.0

11.0
1.9

<0.0001

and validated by the investigators. In the present

LVEDD, mm

53.7
7.9

51.3
7.0

0.004

series, 99% of patients were followed from diagnosis

LVESD, mm

37.7
9.0

35.0
7.8

0.008

until death or at least 5 years post-operatively. In

LV mass, g

250
94

210
63

<0.0001

were obtained after authorization. Documentation of

testing

and

surgical

reports

were

reviewed

patients who died during follow-up, cause of death was adjudicated by review of death certificates,

Values are mean
SD. Abbreviations as in Table 1.

physician/hospital notes, and autopsy certificate if available. All patients underwent post-operative echocardiography at our Institution or at the facility of their personal physician (Figure 1). These results were all included in the follow-up; the first Doppler echocardiographic study demonstrating MR of moderate or severe degree was considered “MR recurrence.” No patients left the operating room with moderate or severe MR, but mild MR was noted in a subset of patients and considered insignificant by the surgeon

during

intraoperative

echocardiography.

These patients were considered to have “residual MR,” which was a variable incorporated within the operative characteristics during statistical analysis. STATISTICAL ANALYSIS. Continuous variables were

summarized as mean
SD or median and quartiles as appropriate. Categorical variables were described as frequencies and percentages. Groups were compared

state i to state j is given by l ij(t) ¼ l ij,0(t)exp(Xij b), where Xij are transition-specific covariates and lij,0 is the baseline hazard for the transition from i to j. We investigated whether the time a patient experienced a recurrence was associated with time to subsequent death. Cox proportional hazard models included all clinically relevant variables as well as statistically significant variables in univariate analysis. Proportional hazard assumptions were evaluated with scaled Schoenfeld residuals. The predictive ability for each Cox model was described using a concordance index (c-index). The c-index ranged from 0.5 (no predictive ability) to 1.0 (perfect prediction). Incidences of MR recurrence were calculated (95% confidence intervals [CIs]) after surgery using personyears. Nonoverlapping time windows of 0 to 1 year,

using the Student t test or chi-square test. The main endpoint of the study was first diagnosis of recurrent

T A B L E 4 Determinants of Long-Term Mortality

MR post-repair with secondary endpoints of reopera-

Univariate Analysis

tion and post-operative death. Rates of endpoints were calculated using the Kaplan-Meier method and

HR (95% CI)

expressed as mean
SE. Recurrence and death are not

Age

1.10 (1.09–1.11)

Male

1.12 (0.92–1.37)

prevent later MR recurrence. Hence, multivariate

Atrial fibrillation

1.98 (1.63–1.37)

analysis of post-repair MR recurrence was conducted

NYHA functional class III–IV

not only using Cox proportional hazards but also as a

Ever smoked

to estimate the hazard ratio (HR) of baseline characteristics and operative variables for time (since surgery) to recurrence or death. In these models, surgery

<0.0001 1.09 (1.07–1.11) 0.25

<0.0001 0.34

<0.0001

1.25 (0.98–1.59)

0.07

2.79 (2.32–3.51)

<0.0001

1.75 (1.38–2.23)

<0.001

1.49 (1.24–1.80)

<0.0001 1.28 (1.01–1.62)

0.04

Hypertension

1.60 (1.33–1.93)

<0.0001 1.04 (0.83–1.32)

0.70

Diabetes

2.29 (1.61–3.17)

<0.0001 1.24 (0.62–1.86)

0.28

CAD

2.67 (2.20–3.23)

<0.0001

1.52 (1.02–2.62)

0.04

Renal failure

2.01 (1.50–2.65)

<0.0001 1.08 (0.73–1.58)

0.75

Echocardiographic data LVEF (5%)

recurrence a time-dependent intermediate state.

Operative data

characterize the transition time between 2 states (alive

p Value

1.15 (0.87–1.51)

was the initial state, death the terminal event, and Although survival analysis models estimate HRs to

Multivariate Analysis HR (95% CI)

Clinical data

independent endpoints, as post-operative death may

competing risk to death using a multistate model

p Value

0.97 (0.96–0.98) <0.0001 0.93 (0.88–0.98)

0.006

Surgery before 1996

1.28 (1.05–1.56)

0.02

1.10 (0.73–1.67)

0.14

CABG

2.16 (1.78–2.62)

<0.0001

1.10 (0.98–1.01)

0.53

Bypass time (by 5-min increase) 1.04 (1.03–1.05)

0.64

<0.0001

1.16 (0.91–1.47)

Resection

0.73 (0.61–0.88)

0.001

1.10 (0.85–1.44)

0.37

Plication

1.36 (1.12–1.65)

0.003

1.08 (0.69–1.72)

0.91

to recurrence, surgery to death, and time to death after

Annuloplasty

0.51 (0.36–0.76)

0.001

0.51 (0.32–0.80)

0.004

recurrence. A clock-forward approach was used for the

Follow-up data 2.06 (1.57–2.69)

<0.0001

1.67 (1.20–2.33)

0.002

to death), this approach enabled us to estimate the HRs for risk factors separately for each transition; surgery

time scale. A Markov proportional hazards model was used to estimate HRs for predictors at transitions from surgery to death, surgery to recurrence, or recurrence to death (19). The hazard function for transition from

MR recurrence*

*MR recurrence was analyzed as a time-dependent variable. CI ¼ confidence interval; HR ¼ hazard ratio; other abbreviations as in Tables 1 and 2.

492

Suri et al.

JACC VOL. 67, NO. 5, 2016

Consequences of Recurrent Regurgitation Following Mitral Repair

FEBRUARY 9, 2016:488–98

RESULTS

T A B L E 5 Determinants of MR Recurrence for the Entire Cohort

Univariate Analysis

Multivariate Analysis

Pre-operative baseline characteristics are shown in

p Value

Table 1. The mean age of patients was 64
13 years,

1.02 (1.01–1.03)

0.01

and 864 (71%) were men. Frequent comorbidities

0.98

1.13 (0.76–1.66)

0.67

included AF (25%), hypertension (35%), coronary ar-

1.70 (1.19–2.40)

0.003

1.52 (1.06–2.19)

0.02

1.74 (1.23–2.45)

0.002

1.57 (1.09–2.24)

0.02

HR (95% CI)

p Value

HR (95% CI)

Age

1.02 (1.01–1.04)

0.001

Male

1.01 (0.69–1.47)

Surgery before 1996 Hypertension

tery disease (26%), and New York Heart Association (NYHA) functional class III/IV status (30%). Mean LV

Bypass time >90 min

2.15 (1.50–3.04)

<0.0001

1.73 (1.19–2.50)

0.004

Residual mild MR in the OR

3.99 (1.79–7.65)

0.002

4.23 (1.86–8.32)

0.001

end-diastolic diameter was 59.9
7.4 mm, LV end-

Resection

0.48 (0.34–0.68)

<0.0001

0.70 (0.48–1.04)

0.07

systolic dimension was 36.8
6.8 mm, and LVEF

Annuloplasty

0.31 (0.18–0.59)

0.0008

0.33 (0.18–0.63)

0.002

was 63
9%. The posterior leaflet of the MV was

0.0002

singularly affected in 62%, bileaflet prolapse was

0.0005

present in 26%, and ruptured chordae tendineae were

0.001

found in 64%. MV repair was performed between 1996

<0.0001

Localization Anterior vs. posterior

2.87 (1.81–4.48)

<0.0001

2.57 (1.54–4.22)

Bileaflet vs. posterior

2.03 (1.37–2.99)

0.0003

2.00 (1.33–2.99)

and 2000 in 755 (62%) patients in the current analysis.

Abbreviations as in Tables 1, 2, and 4.

During surgery (Table 2), mitral leaflet resection was 1 to 2 years, 2 to 5 years, and 5 years or more were selected to distinguish short-, mid-, and long-term outcomes. Rate ratios were calculated in these time intervals to compare MR occurrence between anterior and posterior or bileaflet and posterior. A level of p ¼ 0.05 was considered statistically significant.

All

R

2.15.3

version

analyses (R

were

performed

Foundation

for

using

Statistical

Computing, Vienna, Austria) and the multistate analysis package mstate version 0.2.6 (20).

performed in 61%, leaflet plication in 27%, chordal transfer in 4%, artificial Gore-Tex neochordae (W.L. Gore & Associates, Inc., Flagstaff, Arizona) insertion in 13%, and annuloplasty in 96%. Roughly onequarter (24%) underwent coronary artery bypass graft surgery, and in these patients, the etiology of MR was documented as nonischemic by the operating surgeon. Post-operative intra-aortic balloon counterpulsation was required in 1.4%, and 2% had mild “residual” MR documented in the operating room. RECURRENCE AND SIGNIFICANCE OF MR. Among the 1,218 patients studied, follow-up duration was

T A B L E 6 Transition-Specific Determinants of MR Recurrence

11.0
4.5 years (median 11.5 years; interquartile

Transition Examined From Degenerative MR Repair to

Age (by 5-yr increase) p Value Male p Value Residual mild MR in the OR

The events noted during overall follow-up were as

MR Recurrence (n ¼ 133)

Death Without MR Recurrence (n ¼ 384)

Death After MR Recurrence (n ¼ 62)

1.09 (1.01–1.18)

1.54 (1.44–1.65)

1.67 (1.38–2.00)

0.02

<0.001

<0.001

1.04 (0.70–1.54)

1.12 (0.89–1.40)

2.41 (1.25–4.66)

15-year overall incidence of recurrent MR was 13.3
1.2%. There were 87 cardiac reoperations during

follows: there were 133 patients with a diagnosis of recurrent MR after repair occurring at a median of 3.7 years (IQR: 1.1 to 7.6 years) post-operatively. The

0.86

0.33

0.009

2.96 (2.05–4.30) <0.001

1.21 (0.92–1.59) 0.162

0.81 (0.42–1.56) 0.52

2.68 (1.6–4.43)

1.10 (0.77–1.56)

1.18 (0.54–2.56)

<0.001

0.60

0.67

reoperation was 6.9
1.0%. A total of 452 patients

1.84 (1.23–2.77)

1.14 (0.88–1.47)

0.44 (0.47–1.59)

died post-operatively at a median of 8.2 years (IQR:

0.003

0.32

0.59

4.3 to 11.5 years) following MV repair. The 15-year

0.89 (0.60–1.31)

1.87 (1.50–2.32)

1.31 (0.74–2.33)

0.55

<0.001

0.36

1.09 (0.71–1.66)

1.26 (1.00–1.58)

1.52 (0.84–2.76)

Localization Anterior vs. posterior p Value Bileaflet vs. posterior p Value NYHA functional class III or IV

range [IQR]: 9.2 to 13.6 years) and was 99% complete.

p Value CABG p Value Bypass time >90 min p Value Resection

0.70

0.05

0.02

1.74 (1.18–2.56)

1.00 (0.79–1.28)

1.26 (0.72–2.22)

0.005

0.98

0.42

0.72 (0.49–1.07)

1.01 (0.81–1.26)

1.10 (0.53–1.90)

p Value

0.10

0.95

0.98

Annuloplasty

0.35 (0.19–0.65)

0.60 (0.38–0.96)

0.64 (0.27–1.54)

p Value

<0.001

0.03

0.32

follow-up (64 were related to mitral dysfunction), which occurred at a median time of 4.5 years (IQR: 0.9 to 9.1 years). The 15-year incidence of mitral

overall mortality was 44.0
1.9% (Online Figure 1). To assess the clinical significance of recurrent MR following repair of degenerative MV disease, we first analyzed the influence of this event upon cardiac reverse remodeling at the latest follow-up echocardiogram. Those with recurrent MR had significantly larger residual LV chamber size and myocardial hypertrophy (Table 3). We next studied the association between recurrent MR and death in a Cox proportional hazard model. The

Values are HR (95% CI). Bold text indicates statistical significance. Abbreviations as in Tables 1, 2, and 4.

predictors of overall mortality are presented in Table 4. After controlling for potentially influential

JACC VOL. 67, NO. 5, 2016

Suri et al.

FEBRUARY 9, 2016:488–98

Consequences of Recurrent Regurgitation Following Mitral Repair

the

following

were

associated

with

increased mortality risk: age, NYHA functional class, smoking history, coronary disease, LVEF, no annuloplasty, and recurrent MR expressed as a timedependent variable (HR: 1.72; 95% CI: 1.24 to 2.39). Thus, echocardiogram-documented recurrence of moderate or greater MR resulted in both adverse cardiac remodeling and an increased likelihood of death. DETERMINANTS OF RECURRENT MR. Recognizing that

the recurrence of even moderate MR was a serious event, we next sought to understand the predictors of this problem. In comparing baseline characteristics of patients with and without MR (Table 1), we identified

F I G U R E 2 MR Recurrence (1996 to 2000)

A Recurrence of Mitral Regurgitation After Mitral Valve Repair (%)

variables,

a trend for post-operative recurrent MR patients to

50

Patients at risk: 20

15

14

11

10

8

40

735

700

661

629

581

474

30

20

10

No or trace Residual MR in the OR

0 0

2

rior leaflet prolapse (Table 1). Direct comparison of operative characteristics (Table 2) demonstrated that those with recurrent MR were more likely to have been operated on prior to 1996 and to have spent more time on cardiopulmonary bypass. Intraoperatively, they were also less likely to have undergone posterior leaflet resection but more frequently required artificial neochord placement and had residual (mild) MR

B Recurrence of Mitral Regurgitation After Mitral Valve Repair (%)

likely to have ruptured chordae tendineae and poste-

50

40

83 186 486

76 176 463

C Recurrence of Mitral Regurgitation After Mitral Valve Repair (%)

immediately

post-repair

(Figures 2A and 3A), bileaflet or anterior leaflet prolapse (Figures 2B and 3B), absence of leaflet resection, and

absence

of

prosthetic

69 157 414

annuloplasty

band

(Figure 2C) were all independently associated with an increased likelihood of degenerative MR recurrence following primary MV repair. Interestingly, rates of MR recurrence decreased over time (1-year rate of MR recurrence: 2.7 [95% confidence

63 145 383

51 116 315

p = 0.0002 Anterior Leaflet 10

Bileaflet Posterior Leaflet 2

4

6

8

10

Follow-Up (Years)

These 2 methods of analysis demonstrated almost

identified

72 167 436

20

rence are shown in Table 5 (Cox proportional hazards

before 1996, hypertension, longer bypass time, mild

10

Patients at risk:

0

models) and Table 6 (multistate transitional models). identical results whereby increasing age, surgery

8

30

Multivariate predictors of post-repair MR recur-

MR

6

0

identified immediately post-procedure.

residual

4

Follow-Up (Years)

betes. Although echocardiographically assessed LV between groups, patients with recurrent MR were less

Residual mild MR in the OR

p = 0.006

be older and more likely to have hypertension or diasize, systolic function, and mass were not different

493

50

Patients at risk:

40

19

17

13

12

12

12

736

698

662

628

579

470

8

10

No Ring

30

p < 0.0001

20

10

Ring

0 0

2

4

6

Follow-Up (Years)

interval: 1.9 to 3.9] vs. a subsequent rate of MR recurrence: 0.8 [95% confidence interval: 0.6 to 1.0] per 100

Cumulative incidence of mitral regurgitation (MR) recurrence stratified according to the

person-years; p < 0.001), but the HR related to these

presence of mild intraoperative residual MR (A), localization of prolapse (B), and perfor-

predictors remained constant (all p > 0.14).

mance of annuloplasty (C) during the second half of the study period. Note the increased

The multistate transition-specific method was used to determine predictors of MR recurrence after MV repair (Table 6) identified using the Cox proportional hazards model. We also analyzed the predictors of mortality with and without MR recurrence. Interestingly, the predictors were different in patients who developed MR recurrence. The risk of death without

rate of significant recurrent MR related to the presence of at least mild intraoperative residual MR early following repair, bileaflet or anterior leaflet prolapse, and the absence of ring annuloplasty. OR ¼ operating room.

494

Suri et al.

JACC VOL. 67, NO. 5, 2016

Consequences of Recurrent Regurgitation Following Mitral Repair

FEBRUARY 9, 2016:488–98

Among those undergoing MV repair during the

F I G U R E 3 Rates of MR Recurrence According to Post-Operative Period

A

latter portion of the study (1996 to 2000, n¼755), a total of 59 patients developed moderate or greater

10

recurrent MR and 25 required mitral reoperation

Mitral Regurgitation Recurrence Rates (per 100 Patient-Years)

RR: 3.9(1.0-14.0); p=0.03

following hospital dismissal. In this group of patients, Patients without mild intra-operative residual MR

8

Patients with mild intra-operative residual MR

6

the rate of MR recurrence was 1.5 per 100 patient-years (95% CI: 0.6 to 2.5 per 100 patient-years) in the first year after repair and 0.9 per 100 patient-years (95% CI: 0.6 to 1.1 per 100 patient-years) thereafter. These rates

RR: 4.0(2.3-6.9); p<0.0001

were higher (all p # 0.03) in patients with mild intraoperative residual MR immediately post-operatively

4

(Figure 3A) compared with patients without residual MR. Indeed, mild residual MR immediately following

2

repair was associated with a higher likelihood of progression to moderate or greater MR recurrence at all

0

First year

time points thereafter.

After first year

Further analyzing the subset of patients who un-

Time After Repair Surgery

Mitral Regurgitation Recurrence Rates (per 100 Patient-Years)

B

derwent repair during or after 1996 by leaflet prolapse 10

Anterior leaflet prolapse RR: 8.6(1.3-74.4); p=0.004

8

Posterior leaflet prolapse Bileaflet prolapse

categories,

several

interesting

trends

emerged.

In comparison to patients with posterior leaflet prolapse, those with anterior leaflet disease had higher rates of MR recurrence at all time points after surgery (all p < 0.03) (Figure 3B). In contrast, those with

RR: 5.0(0.9-40.8); p=0.03

bileaflet prolapse only had a higher likelihood of MR

6

recurrence (compared with posterior leaflet) during the first year after surgery (p ¼ 0.001), and there was

RR: 3.2(1.6-6.3); p=0.0003

4

only a nonsignificant trend toward a higher rate of MR recurrence thereafter (p ¼ 0.06) (Figure 3B).

RR: 1.7(0.9-3.1); p=0.11 2

DISCUSSION 0

After first year First year Time After Repair Surgery

On the basis of data from a large multisurgeon heart valve center of excellence over 1 decade with 99% follow-up, this study detailed the long-term recur-

The rate of mitral regurgitation (MR) recurrence for the first year after mitral valve repair for degenerative MR is compared with that for subsequent years in the second one-half of the study period (1996 to 2000) according to the presence of mild intraoperative residual

rence of MR in patients undergoing degenerative MV repair. The results reflected an important evolution in

MR (A) and localization of prolapse (B). Patients with no or trivial intraoperative residual

surgical technique, repair volumes, and repair expe-

MR and posterior or bileaflet prolapse exhibited very low rates of recurrence. Bars ¼ 95%

rience over time. We uniquely accounted for the

confidence intervals. RR ¼ relative risk.

competing risk of death, which otherwise masks the true long-term recurrence rate for MR. We were thus able to more accurately study risk factors predispos-

MR recurrence was predicted by age, NYHA functional

ing to repair failure, along with predictors of death in

class, associated coronary artery bypass grafting,

those with and without MR recurrence following

and absence of annuloplasty, whereas death following

repair. We found that the 15-year incidence of recur-

MR recurrence was only associated with age and

rent MR was 13.3% and the mitral reoperation rate

male sex.

was 6.9%. Importantly, recurrent MR following MV

MR RECURRENCE IN SUBGROUPS. In view of the

repair was associated with adverse LV remodeling as

observation of both diminished MR

recurrence

well as increased likelihood of late death (Central

(Table 5) and mitral reoperation rates over time

Illustration). Results of mitral repair in a more

(Figure 4), rates were analyzed among subsets of pa-

contemporary era (after 1996) revealed a lower rate of

tients operated on after 1996. Importantly, there were

MR recurrence: 1) after the first post-surgical year (0.9

no statistically significant interactions identified be-

per 100 patient-years); 2) in those without mild intra-

tween era of operation and studied risk factors (i.e.,

operative residual MR; and 3) in patients without iso-

residual MR or prolapse localization; all p > 0.43).

lated anterior leaflet prolapse.

JACC VOL. 67, NO. 5, 2016

Suri et al.

FEBRUARY 9, 2016:488–98

Consequences of Recurrent Regurgitation Following Mitral Repair

COMPETING RISK OF DEATH. Accounting for patient

attrition by death is an important concept that has

F I G U R E 4 Incidence of Mitral Valve Reoperation According to Study Period

not been prioritized in prior multisurgeon analyses 6

of long-term outcomes following MV repair. The current multistate analysis allowed us to understand patients for follow-up due to death. The results we presented are timely in that they reflected the excellent long-term durability of low-risk and effective MV repair operations in a specialized heart valve center. It is important to remember that the aim of valve repair is to completely, or almost entirely, eliminate degenerative MR burden. Although surgical mitral repair techniques have further evolved over the past 2 decades (3,11,21), we offered a contemporary reappraisal

5 Occurrence of Mitral Mitral Reoperation (%)

MR recurrence after controlling for unavailability of

495

Patients at risk: 463 755

415 713

385 685

4

361 651

326 503

312 493

Surgery before 1996 Overall

3

p=0.003

2 Surgery after 1996 1 0 0

2

of the predictors of MR recurrence, providing crucial

4

6

8

10

Follow-Up (Years)

guidance to clinicians, imagers, and surgeons caring for patients with degenerative MV disease. Finally, the

The incidence of mitral valve reoperation decreased considerably in the more recent study

importance of eliminating MR at the time of correc-

period (i.e., after 1996; p ¼ 0.003).

tion of degenerative valve disease was eminently apparent. Patients who undergo incomplete reduction of MR have elevated risks of adverse left ventricular,

MV repair have been noted. The evolution of posterior

remodeling, heart failure, and late death.

leaflet prolapse repair from McGoon plication (25) to

PRIOR STUDIES. Previous studies have analyzed

mitral reoperation and MR recurrence in populations of

patients

following

degenerative

MV

repair

(13–15,22). David et al. (4) recently published results of a single-surgeon series of 840 patients undergoing MV repair for MR due to degenerative disease between 1985 and 2004 with a median 10.4 years of follow-up. The authors found that age, LVEF, and functional class were independent predictors of late cardiac- and valve-related death. Recurrent severe MR occurred in 37 patients, and moderate leakage was found in 61 patients. The 20-year risk of mitral reoperation was thus 5.9%, and the freedom from recurrent moderate or severe MR was 69.2% as predicted by age, anterior leaflet prolapse, extent of myxomatous disease, lack of mitral annuloplasty, and duration of cardiopulmonary bypass. As the authors reported, “This is equivalent to almost one-third of all patients developing significant recurrent MR by 20 years.” Despite this, the excellent results of this world-class master technician’s repair experience have been difficult to match by others to date (14,23,24). The predictors of MR recurrence in our large multisurgeon series were strikingly similar. Uniquely, however, our current report established for the first time that MR risk falls following the first year after repair and that MR recurrence is prognostically influential upon long-term survival.

the Schaff triangular resection (26) occurred early in the current repair series at Mayo Clinic. As such, the learning curve for posterior leaflet repair was mastered earlier than for nonposterior leaflet subsets. The superb stability of posterior leaflet repair was reflected in the negligible very long-term rates of MR recurrence and reoperation in this disease subset. Involvement of the anterior leaflet in bileaflet disease often represents “bystander pathology.” In a significant proportion of such cases, the anterior leaflet may be only mildly myxomatous and can thus be addressed by the annuloplasty as an adjunct to posterior leaflet repair, without any anterior leaflet manipulation at all. It is also possible that some of these patients subsequently undergo anterior leaflet disease progression resulting in slightly more recurrence than that seen in posterior leaflet prolapse alone. Despite this possibility, a limited number of options exist to treat anterior leaflet prolapse aside from chordal-based maneuvers; as anterior leaflet resection may cause leaflet restriction that impedes the normal anterior leaflet excursion so critically important to long-term MV competence. It is interesting to note, however, that in our most recent experience presented (1996 to 2000), long-term MR recurrence rate past the first year after bileaflet MV repair became indistinguishable from that seen following posterior leaflet correction alone. Isolated anterior leaflet prolapse, however, was quite different. The historical use of chordal shortening and/or trans-

EVOLUTION IN SURGICAL TECHNIQUE. Over the

fer, which was quite prevalent initially, subsequently

period of our study, several technical improvements in

diminished with time in our practice. Instead,

Suri et al.

JACC VOL. 67, NO. 5, 2016

Consequences of Recurrent Regurgitation Following Mitral Repair

FEBRUARY 9, 2016:488–98

CENTR AL I LLU ST RAT ION

Degenerative Mitral Valve Disease Presentation and Consequences of Recurrent MR Following

Mitral Valve Repair

Posterior Leaflet

Anterior Leaflet

100

Bileaflet

No MR Recurrence

80

Overall Survival (%)

496

76 ± 1%

MR Recurrence

57 ± 2% 60

44 ± 5%

40

29 ± 6% 20 HR: 2.06 (1.57 - 2.69); p <0.0001 *HR: 1.67 (1.20 - 2.33); p = 0.002 0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Follow-up (Years) Patients at Risk: 1,085 1,004 131 100

935 67

831 37

477 15

169 10

Suri, R.M. et al. J Am Coll Cardiol. 2016; 67(5):488–98.

Anatomic presentation of posterior, anterior, and bileaflet mitral valve prolapse. As shown in the Kaplan-Meier curves, late survival is diminished in patients with significant recurrent mitral regurgitation (MR) (expressed as a time-dependent variable) in comparison to patients without recurrent MR. Hazard ratios (HRs) are shown using univariate analysis and after adjustment (*) for age, New York Heart Association functional class, smoking history, coronary artery disease, left ventricular ejection fraction, and presence of annuloplasty.

Gore-Tex neochord replacement has become the pri-

familiarity with the technical nuances of artificial

mary tool utilized to treat anterior leaflet prolapse. As

neochord placement and length adjustment (7). We

previously concluded by others, freedom from recur-

have further noted that some patients who return for

rent MR improved as surgeons gained greater

reoperation

following

failed

Gore-Tex

neochord

JACC VOL. 67, NO. 5, 2016

Suri et al.

FEBRUARY 9, 2016:488–98

Consequences of Recurrent Regurgitation Following Mitral Repair

resuspension often bear evidence of the cord itself

the current era. As such, although it is possible that

tearing through the leaflet tissue. Experienced mitral

tissue quality of these individuals was less robust than

repair surgeons have noted, not infrequently, that the

it might have been in younger patients, the incidence

second loop of these neochords can erode through the

of recurrent MR with time was nonetheless low.

diseased anterior leaflet surface when tissue quality is

STUDY LIMITATIONS. The present study was obser-

poor. If not immediately recognized and corrected by

vational and retrospective, bearing associated biases.

adding an additional loop and/or protecting the knot

However, follow-up was comprehensive and included

with a pledget, this may lead to catastrophic early

assessments of clinical visits and echocardiograms

failure following cardiac reanimation. The nearly 2

performed at our Institution and elsewhere. After

decades of subsequent experience with the Gore-Tex

the initial echocardiogram (i.e., within 30 days),

neochords have led to a greater familiarity with this

one-third to one-half of subsequent studies were not

technique, and the analysis of future long-term out-

performed nor reviewed at Mayo Clinic. Because the

comes will be important to test our hypothesis.

quality of echocardiograms performed elsewhere

CLINICAL

IMPLICATIONS. Prior

studies

have

an-

could not be individually confirmed, it is theoretically

alyzed either mitral reoperation alone, which un-

possible that some patients who were not available for

derestimates the senescence rate of mitral repair, or

follow-up at Mayo Clinic had MR recurrence and

echocardiographic studies, which generally have not

escaped detection. It was reassuring, however, that

accounted for the competing risks of death (13–15). The

99% of patients were followed from diagnosis until

current report conveyed several important messages

death or at least 5 years post-operatively. The fact that

that differentiate it from existing series. First, our

residual/recurrent MR was highly significant in pre-

findings demonstrated for the first time in a contem-

dicting outcomes such as persistent LV and left atrial

porary multisurgeon practice very low MR recurrence

enlargement, as well as excess long-term mortality

following repair of isolated posterior leaflet prolapse,

when analyzed in a time-dependent manner, supports

providing an important contemporary comparator

the importance of high-quality echocardiographic

against which forthcoming percutaneous technology

surveillance for MR recurrence following MV repair.

will be compared. Second, we showed that recurrent

Finally, because the primary endpoint of our study was

MR was associated with poor subsequent outcome,

residual MR following MV repair, patients who un-

including poor reverse remodeling, along with an

derwent attempted repair and had a replacement were

increased risk of MV reoperation and death in

not included in the present analysis.

a competing risk analysis accounting for lack of reoperation in the oldest patients with recurrent MR.

CONCLUSIONS

These findings caution surgeons to consider the not insignificant consequences of residual mild intra-

The overall risk of recurrent MR is very low: 0.9 per

operative and subsequent recurrent moderate MR

100 patient-years following the first year post-MV

following mitral repair.

repair. Despite this reassurance, recurrent moderate-

Third, the potential technical complexity of repair

or-greater MR is a serious problem associated with

for anterior and bileaflet prolapse subsets indicate that

adverse consequences, including left heart enlarge-

referral to high-volume MV repair specialists may be

ment and death. Contemporary results indicate that

necessary to optimize quality and diminish both short-

those with posterior-leaflet or bileaflet disease who

and long-term recurrence risks (27). Due to the fact that

undergo

these subsets are technically challenging and associ-

continue to be free of mild or greater MR early after

ated

with

greater

recurrence

overall,

repair

supported by annuloplasty and

current

surgery have the lowest long-term recurrence rates.

consensus statements also indicate that the perfor-

Early referral for MV repair should thus be predicated

mance of early MV repair in asymptomatic or mini-

upon the assurance of near complete elimination of

mally symptomatic patients should occur at centers

regurgitation to reduce the long-term risk associated

where high-quality repair expertise exists (8). Fourth,

with MR recurrence. Further work will be necessary

these results largely confirm those detailed in the

to understand how recent technical improvements in

single-center experiences of experts such as David

repair quality will affect long-term outcomes.

et al. (4), Castillo et al. (3), and Okada et al. (5) and likely allow generalizability to multisurgeon valve repair

REPRINT REQUESTS AND CORRESPONDENCE: Dr.

practices in certain centers of excellence. Finally, the

Rakesh M. Suri, Department of Thoracic and Cardio-

average age of patients in our series was older than

vascular Surgery, Cleveland Clinic Foundation, 9500

those reported previously, reflecting an evolution in

Euclid Avenue, J4-1, Cleveland, Ohio 44195. E-mail:

the demographics of patients undergoing MV repair in

[email protected].

497

498

Suri et al.

JACC VOL. 67, NO. 5, 2016

Consequences of Recurrent Regurgitation Following Mitral Repair

FEBRUARY 9, 2016:488–98

PERSPECTIVES COMPETENCY IN PATIENT CARE AND PROCE-

TRANSLATIONAL OUTLOOK: Further studies are

DURAL SKILLS: In patients with degenerative MR due to

needed to understand the impact of technical advances

posterior-leaflet or bileaflet disease undergoing surgical

that improve the quality of MV repair on long-term out-

repair with annuloplasty, recurrence risks are low. Mild or

comes in patients undergoing surgery to correct severe

greater early post-operative MR is associated with adverse

degenerative MR.

ventricular remodeling and higher rates of recurrent MR, need for reoperation, and death than in patients with noneto-trivial early post-operative MR.

REFERENCES 1. Suri RM, Vanoverschelde JL, Grigioni F, et al. Association between early surgical intervention vs watchful waiting and outcomes for mitral regurgitation due to flail mitral valve leaflets. JAMA 2013;310:609–16. 2. Kang DH, Park SJ, Sun BJ, et al. Early surgery versus conventional treatment for asymptomatic severe mitral regurgitation: a propensity analysis. J Am Coll Cardiol 2014;63: 2398–407. 3. Castillo JG, Anyanwu AC, Fuster V, Adams DH. A near 100% repair rate for mitral valve prolapse is achievable in a reference center: implications for future guidelines. J Thorac Cardiovasc Surg 2012; 144:308–12. 4. David TE, Armstrong S, McCrindle BW, Manlhiot C. Late outcomes of mitral valve repair for mitral regurgitation due to degenerative disease. Circulation 2013;127:1485–92. 5. Okada Y, Nasu M, Takahashi Y, et al. Late results of mitral valve repair for mitral regurgitation. Jpn J Thorac Cardiovasc Surg 2003;51:282–8. 6. Enriquez-Sarano M, Schaff HV, Orszulak TA, et al. Valve repair improves the outcome of surgery for mitral regurgitation. A multivariate analysis. Circulation 1995;91:1022–8. 7. Suri RM, Schaff HV, Dearani JA, et al. Survival advantage and improved durability of mitral repair for leaflet prolapse subsets in the current era. Ann Thorac Surg 2006;82:819–26. 8. Nishimura RA, Otto C. 2014 ACC/AHA valve guidelines: earlier intervention for chronic mitral regurgitation. Heart 2014;100:905–7. 9. Enriquez-Sarano M, Sundt TM 3rd Early surgery is recommended for mitral regurgitation. Circulation 2010;121:804–11. 10. Mohty D, Orszulak TA, Schaff HV, et al. Very long-term survival and durability of mitral valve repair for mitral valve prolapse. Circulation 2001; 104:I1–7.

11. Suri RM, Burkhart HM. Optimizing outcomes of robotic mitral valve repair for all prolapse anatomy: the Suri-Burkhart technique. Ann Cardiothorac Surg 2013;2:841–5. 12. Cerfolio RJ, Orzulak TA, Pluth JR, et al. Reoperation after valve repair for mitral regurgitation: early and intermediate results. J Thorac Cardiovasc Surg 1996;111:1177–83. 13. Suri RM, Schaff HV, Dearani JA, et al. Recurrent mitral regurgitation after repair: should the mitral valve be re-repaired? J Thorac Cardiovasc Surg 2006;132:1390–7. 14. Flameng WJ, Herijgers P, Szecsi J, et al. Determinants of early and late results of combined valve operations and coronary artery bypass grafting. Ann Thorac Surg 1996;61:621–8. 15. Murashita T, Okada Y, Fujiwara H, et al. Mechanism of and risk factors for reoperation after mitral valve repair for degenerative mitral regurgitation. Circ J 2013;77:2050–5. 16. Lang RM, Badano LP, Mor-Avi V, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2015;28:1–39. e14. 17. Quinones MA, Waggoner AD, Reduto LA, et al. A new, simplified and accurate method for determining ejection fraction with twodimensional echocardiography. Circulation 1981; 64:744–53. 18. Zoghbi WA, Enriquez-Sarano M, Foster E, et al. Recommendations for evaluation of the severity of native valvular regurgitation with twodimensional and Doppler echocardiography. J Am Soc Echocardiogr 2003;16:777–802. 19. Putter H, Fiocco M, Geskus RB. Tutorial in biostatistics: competing risks and multi-state models. Stat Med 2007;26:2389–430.

20. De Wreede LC, Fiocco M, Putter H. mstate: An R Package for the Analysis of Competing Risks and Multi-State Models. J Stat Softw 2011;38. 21. Suri RM, mitral valve techniques establishing

Burkhart HM, Daly RC, et al. Robotic repair for all prolapse subsets using identical to open valvuloplasty: the benchmark against which percu-

taneous interventions should be judged. J Thorac Cardiovasc Surg 2011;142:970–9. 22. Gillinov AM, Blackstone EH, Nowicki ER, et al. Valve repair versus valve replacement for degenerative mitral valve disease. J Thorac Cardiovasc Surg 2008;135:885–93.e1–2. 23. Shimokawa T, Kasegawa H, Matsuyama S, et al. Long-term outcome of mitral valve repair for infective endocarditis. Ann Thorac Surg 2009;88:733–9. 24. Stevens LM, Basmadjian AJ, Bouchard D, et al. Late echocardiographic and clinical outcomes after mitral valve repair for degenerative disease. J Card Surg 2010;25:9–15. 25. McGoon D. Repair of mitral insufficiency due to ruptured chordae tendineae. J Thorac Cardiovasc Surg 1960;39:357–62. 26. Suri RM, Orszulak TA. Triangular resection for repair of mitral regurgitation due to degenerative disease. Oper Tech Thorac Cardiovasc Surg 2005; 8:194–9. 27. Suri RM, Taggarse A, Burkhart HM, et al. Robotic mitral valve repair for simple and complex degenerative disease: midterm clinical and echocardiographic quality outcomes. Circulation 2015; 132:1961–8.

KEY WORDS leaflet, left ventricular, recurrence, reoperation

A PPE NDI X For supplemental methods and a supplemental figure, please see the online version of this article.