- Email: [email protected]
Paravalvular Leak After Mitral Valve Replacement: 20-Year Follow-Up
Paravalvular Leak After Mitral Valve Replacement: 20-Year Follow-Up Ho Young Hwang, MD, PhD, Jae-Woong Choi, MD, Hyung-Kwan Kim, MD, PhD, Kyung-Hwan Kim, MD, PhD, Ki-Bong Kim, MD, PhD, and Hyuk Ahn, MD, PhD Department of Thoracic and Cardiovascular Surgery, and Division of Cardiology, Department of Internal Medicine, Seoul National University Hospital, Seoul, Korea
Background. We evaluated the incidence of paravalvular leak (PVL) after mitral valve replacement (MVR) and analyzed factors associated with early and late PVL during more than 20 years of follow-up. Methods. We studied 1,202 patients (50.4 ± 12.2 years; male/female, 456:746) who underwent MVR between 1992 and 2008. Follow-up duration was 134 months (range, 1 to 272 months). Incidence of early and late PVL was evaluated. The PVL was regarded as major if it caused hemolysis requiring multiple transfusions or regurgitant jet inducing heart failure symptoms. Results. In-hospital mortality rate was 4.5%. Early mitral PVL was found in 23 patients (1.9%), including 7 patients with major PVL. Late mitral PVL without obvious infection occurred in 75 patients (major PVL in 55 patients). Median interval between the index operation and major PVL was 136 months (range, 6 to 250 months). Among 55 patients with major PVL, 50 patients
underwent reoperations and 5 patients were treated medically. Ten- and 20-year late mitral PVL-free rates were 96.2% and 86.9%, respectively. Ten- and 20-year major mitral PVL-free rates were 98.0% and 89.3%, respectively. Cox proportional hazard analysis revealed that age (hazard ratio, 1.052; 95% confidence interval, 1.024 to 1.079), male sex (hazard ratio, 2.804; 95% confidence interval, 1.629 to 4.828), and redo MVR (hazard ratio, 5.193; 95% confidence interval, 2.930 to 9.112) were associated with major mitral PVL during the follow-up. Conclusions. Major PVL without obvious infection occurs even 20 years after MVR with 10- and 20-year major PVL-free rates of 96.2% and 86.9%, respectively. Occurrence of major PVL after MVR is more frequent in elderly, male patients and those who undergo redo MVR.
P
require individual consent based on the institutional guidelines for waiving consent. From January 1992 to December 2008, 1,226 patients underwent MVR at our institution. Twenty patients who underwent index MVR because of PVL after previous operations and 4 patients who had no medical records were excluded. Finally, 1,202 patients (50.4 12.2 years; male:female, 456:746) were enrolled in the present study (Table 1).
aravalvular leak (PVL), which is characterized by eccentric jets originating from outside of the sewing ring, is a nonstructural valve dysfunction that occurs early and late after prosthetic valve replacement. The reported incidence of PVL after mitral valve replacement (MVR) ranges from 5% to 32% [1–4]. The PVL can result from technical failure, which can be identified intraoperatively or by early postoperative echocardiography. It also occurs late after MVR, even several years to a decade after surgery, without any evidence of overt endocarditis [5]. The aims of this study were (1) to evaluate the incidence of early and late PVL, and (2) to analyze factors associated with PVL after MVR during more than 20 years of follow-up.
Material and Methods Study Population The study protocol was reviewed by the institutional review board and approved as a minimal risk retrospective study (approval number H-1405-054-597) that did not Accepted for publication March 18, 2015. Address correspondence to Dr Ahn, Department of Thoracic and Cardiovascular Surgery, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 110-744, Korea; e-mail: [email protected].
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2015;-:-–-) Ó 2015 by The Society of Thoracic Surgeons
Surgical Procedures All operations were performed under aortobicaval cannulation, moderate systemic hypothermia, and cold cardioplegic arrest through median sternotomy. Mitral valve replacement was performed using the everted mattress sutures buttress reinforced with polytetrafluoroethylene as a pledget or a tubule. Use of simple interrupted sutures or continuous suture technique was avoided. Three types of mechanical valves (n ¼ 1,024: Carbomedics; Sulzer Carbomedics Inc, Austin, TX, in 444 patients; On-X valve; On-X Life Technology Inc, Austin, TX, in 392 patients; and St. Jude valve; St. Jude Medical Inc, Minneapolis, MN, in 188 patients) and two types of bioprostheses (n ¼ 165: Carpentier-Edwards Perimount; Edwards Lifesciences LLC, Irvine, CA, in 126 patients; and Hancock II; Medtronic Inc, Minneapolis, MN, in 39 patients) were used in most patients (1,189 of 1,202). Six hundred eighty-four 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.03.104
2
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
Ann Thorac Surg 2015;-:-–-
Table 1. Preoperative Characteristics of the Study Patients n ¼ 1,202
Variable Age (y) Male:female Body surface area (m2) Body mass index (kg/m2) Risk factors, n (%) Smoking Hypertension Diabetes mellitus History of stroke NYHA class 3 Atrial fibrillation LV dysfunction (EF <0.50) Infective endocarditis, n (%) Etiology of mitral valve disease, n (%) Rheumatic Degenerative Other Mitral annular calcification, n (%) History of MVR History of AVR History of TVR
50.4 12.2 456:746 1.57 0.16 21.6 3.1 136 93 74 138 566 885 291 88
(11.3) (7.7) (6.2) (11.5) (47.0) (73.6) (24.2) (6.7)
1,054 87 61 315 330 74 8
(87.7) (7.2) (5.1) (26.2) (27.5) (6.2) (0.7)
AVR ¼ aortic valve replacement; EF ¼ ejection fraction; LV ¼ left ventricle; MVR ¼ mitral valve replacement; NYHA ¼ New York Heart Association; TVR ¼ tricuspid valve replacement.
patients underwent concomitant procedures including aortic valve surgery (n ¼ 420), tricuspid valve operation (n ¼ 335), and arrhythmia surgery (n ¼ 410; Table 2). Mean cardiopulmonary bypass and aortic cross-clamp times were 178 73 min and 117 49 min, respectively. Patients underwent first time (n ¼ 872), redo (n ¼ 322), or second redo MVR (n ¼ 8).
Table 3. Early Clinical Results Variable Mortality, n (%) Complications, n (%) Low cardiac output syndrome Respiratory complication Bleeding reoperation Acute renal failure Stroke Mediastinitis
n ¼ 1,202 54 (4.5) 152 86 61 40 35 16
(12.6) (7.2) (5.1) (3.3) (2.9) (1.3)
especially using color Doppler echocardiography. The normal bileaflet prosthetic valve shows a characteristic regurgitant flow pattern along the closure line of the discs. These regurgitant jets usually arise peripherally and converge with an inverted V shape, or originate from the center of the valve and diverge with a V pattern [6]. A flow within the orifice of the bioprosthesis was considered transvalvular or physiologic regurgitation. Deviations from the patterns described above (ie, large, asymmetric, or eccentric jets originating from outside of the sewing ring) were considered PVL. To delineate regurgitation flow in the left atrium, three apical, parasternal, off-axis, and subcostal views were carefully evaluated with the help of color Doppler imaging. Early mitral PVL was defined as PVL detected by intraoperative or early postoperative echocardiograms. Late mitral PVL was defined as PVL newly diagnosed at the follow-up echocardiography after confirming the absence of PVL on early postoperative echocardiograms. Late PVLs associated with prosthetic valve endocarditis were demonstrated but treated as censored in the subsequent analyses. Mitral PVL was regarded as major if it
Evaluation of Mitral Paravalvular Leak Comprehensive echocardiographic evaluation was performed by experienced echocardiographers. Great care was taken to assess the competence of prosthetic valves, Table 2. Operative Data of the Study Patients Variable
n ¼ 1,202
CPB time (min) ACC time (min) Type of mitral valve replacement, n (%) Mechanical valve Bioprosthesis Concomitant procedures, n (%) Aortic valve surgery Tricuspid valve surgery Arrhythmia surgery Coronary artery bypass grafting Aorta replacement
178 73 117 49
ACC ¼ aortic cross clamp;
1,035 167 684 420 335 410 38 17
CPB ¼ cardiopulmonary bypass.
(84.1) (15.9) (56.9) (34.9) (27.9) (34.1) (3.2) (1.4)
Fig 1. Flow diagram showing occurrence of early and late paravalvular leak (PVL) and subsequent reoperations after mitral valve replacement. (PVE ¼ prosthetic valve endocarditis.)
Ann Thorac Surg 2015;-:-–-
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
3
Table 4. Multivariable Logistic Regression Analysis to Identify Factors Associated With Paravalvular Leak Early After Mitral Valve Replacement
Variable
Univariate Analysis p Value
Infective endocarditis History of stroke Atrial fibrillation Type of prosthetic valve (reference: mechanical valve)
<0.001 0.040 0.018 0.031
Multivariable Analysis Hazard Ratio (95% CI) 6.262 2.632 0.529 2.165
(2.449–16.015) (0.961–7.210) (0.214–1.304) (0.837–5.559)
p Value <0.001 0.060 0.167 0.111
CI ¼ confidence interval.
caused (1) hemolysis requiring multiple transfusions or (2) significant regurgitant jet inducing heart failure symptoms. Early postoperative echocardiography was performed before discharge at 8 days (range, 1 to 75 days) after surgery. Routine follow-up echocardiographic evaluation was performed at the discretion of the operating surgeons or referring physicians. If there was any clinical suspicion of PVL such as patients’ symptoms and signs of heart failure, newly developed holosystolic murmur at auscultation, and laboratory findings of anemia, an additional transthoracic echocardiography was performed. If findings of transthoracic echocardiogram were inconclusive, transesophageal echocardiography was performed. At least one follow-up echocardiography after discharge was performed in 94.2% of patients. The last follow-up echocardiography was performed at 111 months (range, 1 to 265 months) after MVR.
Evaluation of Clinical Outcomes Operative mortality was defined as any death within 30 days after surgery or during the same hospital admission. Postoperative follow-up was performed regularly on an outpatient basis, with a 3- or 4-month interval. Clinical follow-up was closed May 31, 2014. If the patient’s visit did not occur at the scheduled time, they were contacted by telephone to confirm their condition. In addition, data
about vital status and death from cardiovascular diseases were obtained through December 31, 2012, from the death certificates available at Statistics Korea. Follow-up was complete in 94.7% (1,138 of 1,202 patients), with a follow-up duration of 134 months (range, 1 to 272 months).
Statistical Analysis Statistical analyses were performed using the IBM SPSS software package (version 19.0; SPSS Inc, Chicago, IL). Values are expressed as mean standard deviation or proportions. Comparisons were performed with the c2 test or Fisher’s exact test for categorical variables and Student’s t test for continuous variables. Multivariable analysis for categorical variables was performed with logistic regression analysis. Survival rates were estimated using Kaplan-Meier method. Risk factors for time-related events were analyzed using the Cox proportional hazard model. A probability value of less than 0.050 was considered statistically significant.
Results Early Outcomes Operative mortality rate was 4.5% (54 of 1,202 patients). Postoperative complications included low cardiac output syndrome (n ¼ 152), respiratory complications (n ¼ 86),
Fig 2. Kaplan-Meier curves for (A) freedom from late paravalvular leak (PVL) and (B) freedom from major paravalvular leak.
4
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
Ann Thorac Surg 2015;-:-–-
Table 5. Cox Proportional Hazard Model to Identify Factors Associated With Major Paravalvular Leak Late After Mitral Valve Replacement
Variable Age (y) Sex (reference: female) Redo MVR ACC time (min) CPB time (min) ACC ¼ aortic cross-clamp;
Univariate Analysis p Value
Multivariable Analysis Hazard Ratio (95% CI)
0.003 <0.001 <0.001 0.009 0.001 CI ¼ confidence interval;
1.052 2.804 5.193 1.001 1.004
CPB ¼ cardiopulmonary bypass;
bleeding reoperation (n ¼ 61), and acute renal failure requiring hemofiltration (n ¼ 40; Table 3). Early mitral PVL was detected intraoperatively in 7 patients (0.7%). Three of these patients underwent immediate correction of PVL after rerstarting the cardiopulmonary bypass. In the remaining 4 patients, it was left untreated because the leak was less than mild degree. Early mitral PVL was diagnosed in another 16 patients (1.3%) on initial postoperative echocardiography (9 6 days after surgery). Four of those 16 patients had major PVL; 3 underwent reoperation and survived, but the other patient died of pneumonia before reoperation. Overall incidence of early PVL and early major PVL was 1.9% (n ¼ 23) and 0.6% (n ¼ 7), respectively. In 16 patients who had nonmajor PVL, the PVL disappeared in 13 patients during the follow-up, remained unchanged in 1 patient during 6.3 years of follow-up, and was aggravated and became a major PVL in the other 2 patients. Both patients underwent reoperation at 12.5 and 14 years after their index MVR (Fig 1). Multivariable logistic regression analysis revealed that preoperative diagnosis of infective endocarditis was the only factor associated with early PVL after MVR (odds
(1.024–1.079) (1.629–4.828) (2.930–9.112) (0.990–1.011) (0.997–1.011)
p Value <0.001 <0.001 <0.001 0.877 0.268
MVR ¼ mitral valve replacement.
ratio, 6.262; 95% confidence interval [CI], 2.449 to 16.015; Table 4).
Incidence of Late Mitral Paravalvular Leak Late mitral PVL occurred without obvious infective endocarditis in 75 patients including 55 patients with major PVL. The interval between the index MVR and PVL occurrence was 136 months (range, 6 to 250 months). Mitral PVL associated with prosthetic valve endocarditis was found in another 8 patients. Among 55 patients who had major PVL during the follow-up, 29 patients had hemolytic anemia and 26 patients experienced symptoms of heart failure. Fifty patients underwent reoperations, and 5 were treated medically (Fig 1). Ten- and 20-year mitral PVL-free rates were 96.2% and 86.9%, respectively. Ten- and 20-year major mitral PVL-free rates were 98.0% and 89.3%, respectively (Fig 2). Cox proportional hazard analysis revealed that age (hazard ratio [HR], 1.052; 95% CI, 1.024 to 1.079), male sex (HR, 2.804; 95% CI, 1.629 to 4.828), and redo MVR (HR, 5.193; 95% CI, 2.930 to 9.112) were associated with major mitral PVL during the follow-up (Table 5, Fig 3). Other factors including mitral
Fig 3. Differences in freedom from major paravalvular leak (PVL) between (A) male (green) and female patients (blue) and (B) first-time mitral valve replacement (MVR; blue) and redo mitral valve replacement (green).
Ann Thorac Surg 2015;-:-–-
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
Table 6. Clinical Summary of Patients Who Underwent Reoperation for Late Paravalvular Leak After Mitral Valve Replacement Variable Symptoms of paravalvular leak Hemolytic anemia Dyspnea or pitting edema Reoperative procedures, n (%) Isolated mitral valve surgery Concomitant procedure Tricuspid valve replacement Tricuspid annuloplasty Aortic valve replacement Coronary artery bypass grafting Early mortality after reoperation, n (%) Late death during follow-up, n (%)
n ¼ 50 28 (56) 22 (44) 37 13 6 5 3 1 3 12
(74) (26) (12) (10) (6) (2) (6) (24)
annular calcification, small body surface area, and types of prosthetic valves were not statistically significant factors associated with major mitral PVL. At reoperative findings, the most frequent site of PVL was the posterior annulus (n ¼ 25), followed by the anterior (n ¼ 14) and medial portions of the annulus (n ¼ 12). The lateral portion of the annulus was least affected (n ¼ 1). In 6 patients, PVL was found in multiple separate areas.
Long-Term Survival and Results of Reoperations During the follow-up, late deaths occurred in 256 patients, including 141 cardiac deaths. Overall survival rates at 10 and 20 years were 80.8% and 66.1%, respectively. Freedom from cardiac death at 10 and 20 years was 87.7% and 76.9%, respectively. Among 50 patients who underwent reoperation for correction of late mitral PVL, 28 patients had hemolytic anemia and 22 patients experienced symptoms of heart failure. At reoperation, operative mortality occurred in 3 patients (6%). During the follow-up after reoperation, late death occurred in another 12 patients, including 10 cardiac deaths (Table 6).
Comment This study revealed two main findings. First, PVL occurs even 20 years after MVR with 10- and 20-year mitral PVLfree rates of 96.2% and 86.9%, respectively. Second, the risk of major PVL after MVR during follow-up increases in elderly, male patients and those who undergo redo MVR. Paravalvular leak, a regurgitant flow between the prosthetic sewing ring and valvular annulus, is a nonstructural valve dysfunction that could occur early and late after prosthetic valve replacement. Reported incidence of PVL after MVR varies from 5% to 32% [1–4]. It could occur early after MVR or newly develop late after MVR, even several years to a decade after surgery, without overt evidence of prosthetic valve endocarditis [5]. In the present study, early mitral PVL was recorded in
5
only 1.9% of the patients, contrary to findings of previous studies that early PVL was detected in up to 32% of MVR patients [1, 3]. This might be owing to the fact that we avoided the continuous suture technique, which has proved to be associated with a high incidence of PVL after MVR [2, 7]. In addition to suture technique, previous studies suggested that the presence of heavy annular calcification, small body surface area, and size and type of prosthetic valve were associated with mitral PVL [1, 2, 7–10]. In the present study, however, we could not demonstrate that such factors are associated with incidence of mitral PVL. Low incidence of early PVL and suture technique as previously described might affect these results. Although only 1.9% of patients showed early PVL after MVR in the present study, new-onset PVL during the follow-up occurred in 75 patients without any evidence of infective endocarditis. A previous study reported that PVL occurred within 5 years after MVR and remained stable thereafter when patients were followed up to 12 years after MVR [11]. On the contrary, another study demonstrated that the longest period between MVR and the diagnosis of PVL was 23 years [8]. Our study results were in agreement with the latter study. Mitral PVL occurred long after surgery, even 250 months after MVR, especially in elderly male patients and those who underwent redo operation. Although the mechanism has not been clearly elucidated yet, there might be some explanations for these findings. Men tend to engage in more physical activity than women. Thus, high pressure loading on the prosthetic mitral valve in men might result in suture dehiscence at the weakest point of the annulus as a wear-and-tear phenomenon. Weakened annular tissues in elderly patients and those who underwent redo MVR might bolster this phenomenon. In addition, residual foreign material after removing a previous prosthesis might interfere with perfect healing of the new prosthesis–annular interface as suggested in our previous study [12]. There are limitations to the present study that must be recognized. First, this study was performed as a retrospective design, although all consecutive patients who underwent MVR were enrolled. Second, not all patients underwent transesophageal echocardiography to confirm the presence of PVL because of the invasive nature of this method. However, this effect might be minimal because transthoracic echocardiography was performed whenever there was any suspicion of mitral PVL, such as holosystolic murmur, heart failure symptoms, and anemia, and additional transesophageal echocardiography was done if the diagnosis was inconclusive with transthoracic echocardiography.
References 1. O’Rourke DJ, Palac RT, Malenka DJ, Marrin CAS, Arbuckle BE, Plehn JF. Outcome of mild periprosthetic regurgitation detected by intraoperative transesophageal echocardiography. J Am Coll Cardiol 2001;38:163–6. 2. Dhasmana JP, Blackstone EH, Kirklin JW, Kouchoukos NT. Factors associated with periprosthetic leakage following
6
3.
4.
5. 6.
7.
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
primary mitral valve replacement with special consideration of the suture technique. Ann Thorac Surg 1983;35:170–8. Ionescu A, Fraser AG, Butchart EG. Prevalence and clinical significance of incidental paraprosthetic valvar regurgitation: a prospective study using transoesophageal echocardiography. Heart 2003;89:1316–21. Movsowitz HD, Shah SI, Ioli A, Kotler MN, Jacobs LE. Longterm follow-up of mitral paraprosthetic regurgitation by transesophageal echocardiography. J Am Soc Echocardiogr 1994;7:488–92. Misawa Y, Saito T, Konishi H, et al. When and how does nonstructural mechanical prosthetic heart valve dysfunction occur? Jpn J Thorac Cardiovasc Surg 2003;51:355–60. Flachskampf FA, O’Shea JP, Griffin BP, Guerrero L, Weyman AE, Thomas JD. Patterns of normal transvalvular regurgitation in mechanical valve prostheses. J Am Coll Cardiol 1991;18:1493–8. Englberger L, Schaff HV, Jamieson WRE, et al, AVERT Investigators. Importance of implant technique on risk of major paravalvular leak (PVL) after St. Jude mechanical heart
Ann Thorac Surg 2015;-:-–-
8.
9. 10. 11.
12.
valve replacement: a report from the Artificial Valve Endocarditis Reduction Trial (AVERT). Eur J Cardiothorac Surg 2005;28:838–44. Genoni M, Franzen D, Vogt P, et al. Paravalvular leakage after mitral valve replacement: improved long-term survival with aggressive surgery? Eur J Cardiothorac Surg 2000;17: 14–9. Schoen FJ. The first step to understanding valve failure: an overview of pathology. Eur J Cardiothorac Surg 1992;6(Suppl 1):S50–3. Shapira Y, Vaturi M, Sagie A. Hemolysis associated with prosthetic heart valves: a review. Cardiol Rev 2009;17:121–4. Nair SK, Bhatnagar G, Valencia O, Chandrasekaran V. Effect of valve suture technique on incidence of paraprosthetic regurgitation and 10-year survival. Ann Thorac Surg 2010;89: 1171–9. Choi JW, Hwang HY, Kim K-H, Kim K-B, Ahn H. Long-term results of surgical correction for mitral paravalvular leak: repair versus re-replacement. J Heart Valve Dis 2013;22: 682–7.
Background. We evaluated the incidence of paravalvular leak (PVL) after mitral valve replacement (MVR) and analyzed factors associated with early and late PVL during more than 20 years of follow-up. Methods. We studied 1,202 patients (50.4 ± 12.2 years; male/female, 456:746) who underwent MVR between 1992 and 2008. Follow-up duration was 134 months (range, 1 to 272 months). Incidence of early and late PVL was evaluated. The PVL was regarded as major if it caused hemolysis requiring multiple transfusions or regurgitant jet inducing heart failure symptoms. Results. In-hospital mortality rate was 4.5%. Early mitral PVL was found in 23 patients (1.9%), including 7 patients with major PVL. Late mitral PVL without obvious infection occurred in 75 patients (major PVL in 55 patients). Median interval between the index operation and major PVL was 136 months (range, 6 to 250 months). Among 55 patients with major PVL, 50 patients
underwent reoperations and 5 patients were treated medically. Ten- and 20-year late mitral PVL-free rates were 96.2% and 86.9%, respectively. Ten- and 20-year major mitral PVL-free rates were 98.0% and 89.3%, respectively. Cox proportional hazard analysis revealed that age (hazard ratio, 1.052; 95% confidence interval, 1.024 to 1.079), male sex (hazard ratio, 2.804; 95% confidence interval, 1.629 to 4.828), and redo MVR (hazard ratio, 5.193; 95% confidence interval, 2.930 to 9.112) were associated with major mitral PVL during the follow-up. Conclusions. Major PVL without obvious infection occurs even 20 years after MVR with 10- and 20-year major PVL-free rates of 96.2% and 86.9%, respectively. Occurrence of major PVL after MVR is more frequent in elderly, male patients and those who undergo redo MVR.
P
require individual consent based on the institutional guidelines for waiving consent. From January 1992 to December 2008, 1,226 patients underwent MVR at our institution. Twenty patients who underwent index MVR because of PVL after previous operations and 4 patients who had no medical records were excluded. Finally, 1,202 patients (50.4 12.2 years; male:female, 456:746) were enrolled in the present study (Table 1).
aravalvular leak (PVL), which is characterized by eccentric jets originating from outside of the sewing ring, is a nonstructural valve dysfunction that occurs early and late after prosthetic valve replacement. The reported incidence of PVL after mitral valve replacement (MVR) ranges from 5% to 32% [1–4]. The PVL can result from technical failure, which can be identified intraoperatively or by early postoperative echocardiography. It also occurs late after MVR, even several years to a decade after surgery, without any evidence of overt endocarditis [5]. The aims of this study were (1) to evaluate the incidence of early and late PVL, and (2) to analyze factors associated with PVL after MVR during more than 20 years of follow-up.
Material and Methods Study Population The study protocol was reviewed by the institutional review board and approved as a minimal risk retrospective study (approval number H-1405-054-597) that did not Accepted for publication March 18, 2015. Address correspondence to Dr Ahn, Department of Thoracic and Cardiovascular Surgery, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 110-744, Korea; e-mail: [email protected].
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2015;-:-–-) Ó 2015 by The Society of Thoracic Surgeons
Surgical Procedures All operations were performed under aortobicaval cannulation, moderate systemic hypothermia, and cold cardioplegic arrest through median sternotomy. Mitral valve replacement was performed using the everted mattress sutures buttress reinforced with polytetrafluoroethylene as a pledget or a tubule. Use of simple interrupted sutures or continuous suture technique was avoided. Three types of mechanical valves (n ¼ 1,024: Carbomedics; Sulzer Carbomedics Inc, Austin, TX, in 444 patients; On-X valve; On-X Life Technology Inc, Austin, TX, in 392 patients; and St. Jude valve; St. Jude Medical Inc, Minneapolis, MN, in 188 patients) and two types of bioprostheses (n ¼ 165: Carpentier-Edwards Perimount; Edwards Lifesciences LLC, Irvine, CA, in 126 patients; and Hancock II; Medtronic Inc, Minneapolis, MN, in 39 patients) were used in most patients (1,189 of 1,202). Six hundred eighty-four 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.03.104
2
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
Ann Thorac Surg 2015;-:-–-
Table 1. Preoperative Characteristics of the Study Patients n ¼ 1,202
Variable Age (y) Male:female Body surface area (m2) Body mass index (kg/m2) Risk factors, n (%) Smoking Hypertension Diabetes mellitus History of stroke NYHA class 3 Atrial fibrillation LV dysfunction (EF <0.50) Infective endocarditis, n (%) Etiology of mitral valve disease, n (%) Rheumatic Degenerative Other Mitral annular calcification, n (%) History of MVR History of AVR History of TVR
50.4 12.2 456:746 1.57 0.16 21.6 3.1 136 93 74 138 566 885 291 88
(11.3) (7.7) (6.2) (11.5) (47.0) (73.6) (24.2) (6.7)
1,054 87 61 315 330 74 8
(87.7) (7.2) (5.1) (26.2) (27.5) (6.2) (0.7)
AVR ¼ aortic valve replacement; EF ¼ ejection fraction; LV ¼ left ventricle; MVR ¼ mitral valve replacement; NYHA ¼ New York Heart Association; TVR ¼ tricuspid valve replacement.
patients underwent concomitant procedures including aortic valve surgery (n ¼ 420), tricuspid valve operation (n ¼ 335), and arrhythmia surgery (n ¼ 410; Table 2). Mean cardiopulmonary bypass and aortic cross-clamp times were 178 73 min and 117 49 min, respectively. Patients underwent first time (n ¼ 872), redo (n ¼ 322), or second redo MVR (n ¼ 8).
Table 3. Early Clinical Results Variable Mortality, n (%) Complications, n (%) Low cardiac output syndrome Respiratory complication Bleeding reoperation Acute renal failure Stroke Mediastinitis
n ¼ 1,202 54 (4.5) 152 86 61 40 35 16
(12.6) (7.2) (5.1) (3.3) (2.9) (1.3)
especially using color Doppler echocardiography. The normal bileaflet prosthetic valve shows a characteristic regurgitant flow pattern along the closure line of the discs. These regurgitant jets usually arise peripherally and converge with an inverted V shape, or originate from the center of the valve and diverge with a V pattern [6]. A flow within the orifice of the bioprosthesis was considered transvalvular or physiologic regurgitation. Deviations from the patterns described above (ie, large, asymmetric, or eccentric jets originating from outside of the sewing ring) were considered PVL. To delineate regurgitation flow in the left atrium, three apical, parasternal, off-axis, and subcostal views were carefully evaluated with the help of color Doppler imaging. Early mitral PVL was defined as PVL detected by intraoperative or early postoperative echocardiograms. Late mitral PVL was defined as PVL newly diagnosed at the follow-up echocardiography after confirming the absence of PVL on early postoperative echocardiograms. Late PVLs associated with prosthetic valve endocarditis were demonstrated but treated as censored in the subsequent analyses. Mitral PVL was regarded as major if it
Evaluation of Mitral Paravalvular Leak Comprehensive echocardiographic evaluation was performed by experienced echocardiographers. Great care was taken to assess the competence of prosthetic valves, Table 2. Operative Data of the Study Patients Variable
n ¼ 1,202
CPB time (min) ACC time (min) Type of mitral valve replacement, n (%) Mechanical valve Bioprosthesis Concomitant procedures, n (%) Aortic valve surgery Tricuspid valve surgery Arrhythmia surgery Coronary artery bypass grafting Aorta replacement
178 73 117 49
ACC ¼ aortic cross clamp;
1,035 167 684 420 335 410 38 17
CPB ¼ cardiopulmonary bypass.
(84.1) (15.9) (56.9) (34.9) (27.9) (34.1) (3.2) (1.4)
Fig 1. Flow diagram showing occurrence of early and late paravalvular leak (PVL) and subsequent reoperations after mitral valve replacement. (PVE ¼ prosthetic valve endocarditis.)
Ann Thorac Surg 2015;-:-–-
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
3
Table 4. Multivariable Logistic Regression Analysis to Identify Factors Associated With Paravalvular Leak Early After Mitral Valve Replacement
Variable
Univariate Analysis p Value
Infective endocarditis History of stroke Atrial fibrillation Type of prosthetic valve (reference: mechanical valve)
<0.001 0.040 0.018 0.031
Multivariable Analysis Hazard Ratio (95% CI) 6.262 2.632 0.529 2.165
(2.449–16.015) (0.961–7.210) (0.214–1.304) (0.837–5.559)
p Value <0.001 0.060 0.167 0.111
CI ¼ confidence interval.
caused (1) hemolysis requiring multiple transfusions or (2) significant regurgitant jet inducing heart failure symptoms. Early postoperative echocardiography was performed before discharge at 8 days (range, 1 to 75 days) after surgery. Routine follow-up echocardiographic evaluation was performed at the discretion of the operating surgeons or referring physicians. If there was any clinical suspicion of PVL such as patients’ symptoms and signs of heart failure, newly developed holosystolic murmur at auscultation, and laboratory findings of anemia, an additional transthoracic echocardiography was performed. If findings of transthoracic echocardiogram were inconclusive, transesophageal echocardiography was performed. At least one follow-up echocardiography after discharge was performed in 94.2% of patients. The last follow-up echocardiography was performed at 111 months (range, 1 to 265 months) after MVR.
Evaluation of Clinical Outcomes Operative mortality was defined as any death within 30 days after surgery or during the same hospital admission. Postoperative follow-up was performed regularly on an outpatient basis, with a 3- or 4-month interval. Clinical follow-up was closed May 31, 2014. If the patient’s visit did not occur at the scheduled time, they were contacted by telephone to confirm their condition. In addition, data
about vital status and death from cardiovascular diseases were obtained through December 31, 2012, from the death certificates available at Statistics Korea. Follow-up was complete in 94.7% (1,138 of 1,202 patients), with a follow-up duration of 134 months (range, 1 to 272 months).
Statistical Analysis Statistical analyses were performed using the IBM SPSS software package (version 19.0; SPSS Inc, Chicago, IL). Values are expressed as mean standard deviation or proportions. Comparisons were performed with the c2 test or Fisher’s exact test for categorical variables and Student’s t test for continuous variables. Multivariable analysis for categorical variables was performed with logistic regression analysis. Survival rates were estimated using Kaplan-Meier method. Risk factors for time-related events were analyzed using the Cox proportional hazard model. A probability value of less than 0.050 was considered statistically significant.
Results Early Outcomes Operative mortality rate was 4.5% (54 of 1,202 patients). Postoperative complications included low cardiac output syndrome (n ¼ 152), respiratory complications (n ¼ 86),
Fig 2. Kaplan-Meier curves for (A) freedom from late paravalvular leak (PVL) and (B) freedom from major paravalvular leak.
4
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
Ann Thorac Surg 2015;-:-–-
Table 5. Cox Proportional Hazard Model to Identify Factors Associated With Major Paravalvular Leak Late After Mitral Valve Replacement
Variable Age (y) Sex (reference: female) Redo MVR ACC time (min) CPB time (min) ACC ¼ aortic cross-clamp;
Univariate Analysis p Value
Multivariable Analysis Hazard Ratio (95% CI)
0.003 <0.001 <0.001 0.009 0.001 CI ¼ confidence interval;
1.052 2.804 5.193 1.001 1.004
CPB ¼ cardiopulmonary bypass;
bleeding reoperation (n ¼ 61), and acute renal failure requiring hemofiltration (n ¼ 40; Table 3). Early mitral PVL was detected intraoperatively in 7 patients (0.7%). Three of these patients underwent immediate correction of PVL after rerstarting the cardiopulmonary bypass. In the remaining 4 patients, it was left untreated because the leak was less than mild degree. Early mitral PVL was diagnosed in another 16 patients (1.3%) on initial postoperative echocardiography (9 6 days after surgery). Four of those 16 patients had major PVL; 3 underwent reoperation and survived, but the other patient died of pneumonia before reoperation. Overall incidence of early PVL and early major PVL was 1.9% (n ¼ 23) and 0.6% (n ¼ 7), respectively. In 16 patients who had nonmajor PVL, the PVL disappeared in 13 patients during the follow-up, remained unchanged in 1 patient during 6.3 years of follow-up, and was aggravated and became a major PVL in the other 2 patients. Both patients underwent reoperation at 12.5 and 14 years after their index MVR (Fig 1). Multivariable logistic regression analysis revealed that preoperative diagnosis of infective endocarditis was the only factor associated with early PVL after MVR (odds
(1.024–1.079) (1.629–4.828) (2.930–9.112) (0.990–1.011) (0.997–1.011)
p Value <0.001 <0.001 <0.001 0.877 0.268
MVR ¼ mitral valve replacement.
ratio, 6.262; 95% confidence interval [CI], 2.449 to 16.015; Table 4).
Incidence of Late Mitral Paravalvular Leak Late mitral PVL occurred without obvious infective endocarditis in 75 patients including 55 patients with major PVL. The interval between the index MVR and PVL occurrence was 136 months (range, 6 to 250 months). Mitral PVL associated with prosthetic valve endocarditis was found in another 8 patients. Among 55 patients who had major PVL during the follow-up, 29 patients had hemolytic anemia and 26 patients experienced symptoms of heart failure. Fifty patients underwent reoperations, and 5 were treated medically (Fig 1). Ten- and 20-year mitral PVL-free rates were 96.2% and 86.9%, respectively. Ten- and 20-year major mitral PVL-free rates were 98.0% and 89.3%, respectively (Fig 2). Cox proportional hazard analysis revealed that age (hazard ratio [HR], 1.052; 95% CI, 1.024 to 1.079), male sex (HR, 2.804; 95% CI, 1.629 to 4.828), and redo MVR (HR, 5.193; 95% CI, 2.930 to 9.112) were associated with major mitral PVL during the follow-up (Table 5, Fig 3). Other factors including mitral
Fig 3. Differences in freedom from major paravalvular leak (PVL) between (A) male (green) and female patients (blue) and (B) first-time mitral valve replacement (MVR; blue) and redo mitral valve replacement (green).
Ann Thorac Surg 2015;-:-–-
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
Table 6. Clinical Summary of Patients Who Underwent Reoperation for Late Paravalvular Leak After Mitral Valve Replacement Variable Symptoms of paravalvular leak Hemolytic anemia Dyspnea or pitting edema Reoperative procedures, n (%) Isolated mitral valve surgery Concomitant procedure Tricuspid valve replacement Tricuspid annuloplasty Aortic valve replacement Coronary artery bypass grafting Early mortality after reoperation, n (%) Late death during follow-up, n (%)
n ¼ 50 28 (56) 22 (44) 37 13 6 5 3 1 3 12
(74) (26) (12) (10) (6) (2) (6) (24)
annular calcification, small body surface area, and types of prosthetic valves were not statistically significant factors associated with major mitral PVL. At reoperative findings, the most frequent site of PVL was the posterior annulus (n ¼ 25), followed by the anterior (n ¼ 14) and medial portions of the annulus (n ¼ 12). The lateral portion of the annulus was least affected (n ¼ 1). In 6 patients, PVL was found in multiple separate areas.
Long-Term Survival and Results of Reoperations During the follow-up, late deaths occurred in 256 patients, including 141 cardiac deaths. Overall survival rates at 10 and 20 years were 80.8% and 66.1%, respectively. Freedom from cardiac death at 10 and 20 years was 87.7% and 76.9%, respectively. Among 50 patients who underwent reoperation for correction of late mitral PVL, 28 patients had hemolytic anemia and 22 patients experienced symptoms of heart failure. At reoperation, operative mortality occurred in 3 patients (6%). During the follow-up after reoperation, late death occurred in another 12 patients, including 10 cardiac deaths (Table 6).
Comment This study revealed two main findings. First, PVL occurs even 20 years after MVR with 10- and 20-year mitral PVLfree rates of 96.2% and 86.9%, respectively. Second, the risk of major PVL after MVR during follow-up increases in elderly, male patients and those who undergo redo MVR. Paravalvular leak, a regurgitant flow between the prosthetic sewing ring and valvular annulus, is a nonstructural valve dysfunction that could occur early and late after prosthetic valve replacement. Reported incidence of PVL after MVR varies from 5% to 32% [1–4]. It could occur early after MVR or newly develop late after MVR, even several years to a decade after surgery, without overt evidence of prosthetic valve endocarditis [5]. In the present study, early mitral PVL was recorded in
5
only 1.9% of the patients, contrary to findings of previous studies that early PVL was detected in up to 32% of MVR patients [1, 3]. This might be owing to the fact that we avoided the continuous suture technique, which has proved to be associated with a high incidence of PVL after MVR [2, 7]. In addition to suture technique, previous studies suggested that the presence of heavy annular calcification, small body surface area, and size and type of prosthetic valve were associated with mitral PVL [1, 2, 7–10]. In the present study, however, we could not demonstrate that such factors are associated with incidence of mitral PVL. Low incidence of early PVL and suture technique as previously described might affect these results. Although only 1.9% of patients showed early PVL after MVR in the present study, new-onset PVL during the follow-up occurred in 75 patients without any evidence of infective endocarditis. A previous study reported that PVL occurred within 5 years after MVR and remained stable thereafter when patients were followed up to 12 years after MVR [11]. On the contrary, another study demonstrated that the longest period between MVR and the diagnosis of PVL was 23 years [8]. Our study results were in agreement with the latter study. Mitral PVL occurred long after surgery, even 250 months after MVR, especially in elderly male patients and those who underwent redo operation. Although the mechanism has not been clearly elucidated yet, there might be some explanations for these findings. Men tend to engage in more physical activity than women. Thus, high pressure loading on the prosthetic mitral valve in men might result in suture dehiscence at the weakest point of the annulus as a wear-and-tear phenomenon. Weakened annular tissues in elderly patients and those who underwent redo MVR might bolster this phenomenon. In addition, residual foreign material after removing a previous prosthesis might interfere with perfect healing of the new prosthesis–annular interface as suggested in our previous study [12]. There are limitations to the present study that must be recognized. First, this study was performed as a retrospective design, although all consecutive patients who underwent MVR were enrolled. Second, not all patients underwent transesophageal echocardiography to confirm the presence of PVL because of the invasive nature of this method. However, this effect might be minimal because transthoracic echocardiography was performed whenever there was any suspicion of mitral PVL, such as holosystolic murmur, heart failure symptoms, and anemia, and additional transesophageal echocardiography was done if the diagnosis was inconclusive with transthoracic echocardiography.
References 1. O’Rourke DJ, Palac RT, Malenka DJ, Marrin CAS, Arbuckle BE, Plehn JF. Outcome of mild periprosthetic regurgitation detected by intraoperative transesophageal echocardiography. J Am Coll Cardiol 2001;38:163–6. 2. Dhasmana JP, Blackstone EH, Kirklin JW, Kouchoukos NT. Factors associated with periprosthetic leakage following
6
3.
4.
5. 6.
7.
HWANG ET AL PARAVALVULAR LEAK AFTER MVR
primary mitral valve replacement with special consideration of the suture technique. Ann Thorac Surg 1983;35:170–8. Ionescu A, Fraser AG, Butchart EG. Prevalence and clinical significance of incidental paraprosthetic valvar regurgitation: a prospective study using transoesophageal echocardiography. Heart 2003;89:1316–21. Movsowitz HD, Shah SI, Ioli A, Kotler MN, Jacobs LE. Longterm follow-up of mitral paraprosthetic regurgitation by transesophageal echocardiography. J Am Soc Echocardiogr 1994;7:488–92. Misawa Y, Saito T, Konishi H, et al. When and how does nonstructural mechanical prosthetic heart valve dysfunction occur? Jpn J Thorac Cardiovasc Surg 2003;51:355–60. Flachskampf FA, O’Shea JP, Griffin BP, Guerrero L, Weyman AE, Thomas JD. Patterns of normal transvalvular regurgitation in mechanical valve prostheses. J Am Coll Cardiol 1991;18:1493–8. Englberger L, Schaff HV, Jamieson WRE, et al, AVERT Investigators. Importance of implant technique on risk of major paravalvular leak (PVL) after St. Jude mechanical heart
Ann Thorac Surg 2015;-:-–-
8.
9. 10. 11.
12.
valve replacement: a report from the Artificial Valve Endocarditis Reduction Trial (AVERT). Eur J Cardiothorac Surg 2005;28:838–44. Genoni M, Franzen D, Vogt P, et al. Paravalvular leakage after mitral valve replacement: improved long-term survival with aggressive surgery? Eur J Cardiothorac Surg 2000;17: 14–9. Schoen FJ. The first step to understanding valve failure: an overview of pathology. Eur J Cardiothorac Surg 1992;6(Suppl 1):S50–3. Shapira Y, Vaturi M, Sagie A. Hemolysis associated with prosthetic heart valves: a review. Cardiol Rev 2009;17:121–4. Nair SK, Bhatnagar G, Valencia O, Chandrasekaran V. Effect of valve suture technique on incidence of paraprosthetic regurgitation and 10-year survival. Ann Thorac Surg 2010;89: 1171–9. Choi JW, Hwang HY, Kim K-H, Kim K-B, Ahn H. Long-term results of surgical correction for mitral paravalvular leak: repair versus re-replacement. J Heart Valve Dis 2013;22: 682–7.