On-X mitral valve replacement: a single-centre experience in 318 patients† Free

Abstract

OBJECTIVES

If mitral valve repair is not feasible, mitral valve replacement remains the only option. Based on our overall experience with the On-X mitral valve, the aim of this study was to investigate early and late outcomes after mitral valve replacement using this latest generation prosthesis.

METHODS

From 1998 to 2016, 600 patients received an On-X prosthesis in the mitral position. Of them, we excluded all patients who had combined aortic procedures and retrospectively analysed 318 consecutive patients who had a mitral valve replacement. Associated procedures (53.5%) were tricuspid valve repair, coronary artery bypass graft and the maze procedure. The mean follow-up time was 5.6 ± 4.0 years.

RESULTS

The overall hospital mortality rate was 4.4%, including acute cases of ischaemic mitral regurgitation (9.4%) and infective endocarditis (9.4%). Survival rates at 1, 3, 5 and 10 years were 97.8 ± 1.0%, 92.4 ± 1.7%, 88.4 ± 2.2% and 70.9 ± 4.0%, respectively. Independent predictors of late mortality were hypertension [hazard ratio (HR) 1.91; P = 0.027], chronic obstructive pulmonary disease (HR 2.91; P = 0.003) and chronic renal failure (HR 5.27; P < 0.001). Freedom from reoperation was 99.3 ± 0.5%, 98.4 ± 0.8%, 97.2 ± 1.2% and 92.5 ± 2.4% at 1, 3, 5 and 10 years, respectively. At follow-up, 8.5% events were recognized as thromboembolic or haemorrhagic events; freedom from events related to anticoagulation therapy at 1, 3, 5 and 10 years was 99.0 ± 0.6%, 96.8 ± 1.1%, 93.7 ± 1.8% and 89.0 ± 2.7%, respectively.

CONCLUSIONS

According to the results of this observational study, the unique design of the On-X valve works well with mitral valve diseases of various aetiologies, especially in cases with an unfavourable anatomy. This prosthesis also guarantees safe long-term durability associated with a low incidence of thromboembolism.

INTRODUCTION

Even though valve repair and transcatheter aortic valve implantation are becoming increasingly popular, for most patients with valvular heart disease, valve replacement with a prosthetic mechanical or biological valve remains the mainstay among the surgical therapeutic options [1, 2]. The On-X bileaflet mechanical valve (On-X Life Technologies Inc., Austin, TX, USA), introduced in 1996 and commercially available in Europe in 1998, uses pure pyrolytic carbon and has a flared inlet orifice designed to reduce flow turbulence and the support of a full annulus to protect the opening of the valve from tissue encroachment [3, 4]. These technological improvements have been hypothesized to limit morbid events in patients.

Reported thromboembolic and haemorrhagic complication rates after valve replacement with the On-X prosthesis are low: 5-year freedom from a major thromboembolism ranges from 96.0% to 97.7% for aortic valve replacement and 93.5% to 99.2% for mitral valve replacement (MVR) [5–8]. Although they are well designed, these studies have been limited by sample size and a short follow-up period. Furthermore, no one seems to be focused on MVR. The end point of this study was to evaluate the early and late adverse clinical event rates with the On-X mechanical prosthesis, particularly with regard to major thromboembolism, haemorrhage and valve-related complications.

MATERIALS AND METHODS

Study population

Between December 1998 and July 2016, 600 consecutive patients received an On-X prosthesis for MVR at S. Orsola Hospital, University of Bologna, Italy. We excluded all the combined aortic procedures and retrospectively analysed only 318 consecutive patients with the following inclusion criteria:

• Isolated MVR

• MVR with tricuspid valve repair, coronary artery bypass graft and/or the maze procedure.

The number of procedures remained stable during the entire study period (Fig. 1).

Management

The On-X prosthesis is available for the mitral position both with a standard sewing ring (valve sizes range from 23 to 31 mm) and a newer conform-X design with a unique valve size of 25/33. In this study, the standard On-X prosthesis was implanted until early 2001; among all the patients, only 3 patients received the 27/29 valve size. The new conform-X sewing ring 25/33 was utilized for all cases since 2001. Operations were performed with the patient on cardiopulmonary bypass with moderate hypothermia and crystalloid cardioplegia. The prostheses were implanted using pledgeted horizontal mattress 2-0 Ethibond (Ethicon, Somerville, NJ, USA) sutures. In the majority of cases, the mitral subvalvular apparatus was preserved to maintain left ventricular geometry and function (Fig. 2).

Postoperative anticoagulation therapy with warfarin was initiated 2 days after surgery to reach the target international normalized ratio (INR) as soon as possible, which was maintained between 2 and 3, in absence of other risk factors, per European guidelines [1]. In the case of coexisting atrial fibrillation or any additional risk factors for thrombosis, we set the target INR between 3 and 3.5 (for medium risk) or 3.5 and 4 (for high risk). Patients with coexisting coronary bypass grafting also received an antiplatelet agent, usually aspirin 75–100 mg.

Informed consent was obtained from all patients. Data were obtained using an institutional database supplemented by chart review. Follow-up was carried out in our outpatient clinic by direct telephone interview and by consulting the civil registry. When available, clinical records were reviewed to confirm adverse events and causes of death.

As anticoagulation therapy-related events, our combined end point included all ischaemic or haemorrhagic neurological events and thromboembolic/haemorrhagic systemic or peripheral complications.

Postoperative neurological events (temporary or permanent – focal or global) were defined as the onset of a new neurological deficit occurring after the patient awakened from anaesthesia and documented by a neurology specialist. Stroke was defined as a prolonged (<72 h) or permanent neurological deficit associated with a new single brain lesion or multiple brain lesions confirmed by a computed tomography scan or magnetic resonance imaging. A transient ischaemic attack was characterized by reversible neurological symptoms of short duration (<72 h).

A bleeding event was defined as any episode of major internal or external bleeding causing death, hospitalization or permanent injury or needing transfusion.

Statistical analysis

Statistical analyses were performed using the Statistical Package for Social Sciences version 22.0 (IBM SPSS Inc., Chicago, IL, USA). Continuous variables are expressed as mean ± standard deviation and categorical variables as numbers and percentages.

Long-term survival (calculated as the time from the procedure to death) was analysed at 1, 3, 5 and 10 years using the Kaplan–Meier method. Competing risk methods were used to calculate the cumulative incidence of reintervention and anticoagulant-related events. Univariable and multivariable logistic regression analyses were performed to determine risk factors of late death.

Multivariable Cox regression analyses were conducted on variables that reached P < 0.1 on univariable analysis. Multicollinearity among variables was detected by means of the Pearson correlation test. Model building followed a backward-stepwise approach; the test of term significance was the Wald χ² test with a cut-off P-value of 0.1 for removal and 0.05 for addition.

RESULTS

Patient characteristics

Tables 1 and 2 describe the preoperative and operative characteristics, respectively. The mean age of the patients at operation was 60.9 ± 12.6 years; 51.9% of patients were women. Chronic renal failure and chronic obstructive pulmonary disease (COPD) affected 9.1% (29 cases) and 11.0% (35 cases) of the patients, respectively. Preoperative atrial fibrillation was present in 52.5% of the patients, and neurological events occurred in 12.3% (39 patients). The most common aetiology of mitral disease was degenerative (40.9%) or rheumatic (26.7%) factors. Thirty patients (9.4%) were diagnosed with infective endocarditis. Furthermore, 24.2% of the patients (77 cases) had previous cardiac surgery procedures and 53.5% (170 patients) had associated procedures.

In-hospital outcomes

Overall in-hospital mortality was 4.4% (14 of 318). The causes of death were multiorgan failure (n = 8), septic shock (n = 3), respiratory failure (n = 2) and neurological events (n = 1). The adverse events that occurred during the in-hospital stay are described in Table 3. During the perioperative period, myocardial infarctions occurred in 10 patients (3.1%). Eighteen patients (5.7%) required haemodynamic support with an intra-aortic balloon pump: Of them, 5 patients were operated on for ischaemic mitral regurgitation, and 16 patients were weaned off haemodynamic support and discharged. Temporary haemodialysis was necessary for 11 patients (3.5%), of whom 3 had a preoperative diagnosis of chronic renal failure. Moreover, during the in-hospital stay, sepsis occurred in 17 patients (5.3%); 7 of them were treated for ischaemic mitral regurgitation and 2 for endocarditis.

Survival and late morbidities

The mean follow-up time was 5.6 ± 4.0 years; 9 patients were lost to follow-up; therefore, follow-up was 97.2% complete. Survival rates were 97.8 ± 1.0%, 92.4 ± 1.7%, 88.4 ± 2.2% and 70.9 ± 4.0% at 1, 3, 5 and 10 years, respectively (Fig. 3). Freedom from reoperation was 99.3 ± 0.5%, 98.4 ± 0.8%, 97.2 ± 1.2% and 92.5 ± 2.4% at 1, 3, 5 and 10 years, respectively.

The overall rate of reoperation was 3.5% (11 of 318); the causes of reoperation were endocarditis (4), non-structural dysfunction (paravalvular leaks) (3), structural valve dysfunction (2) (1 suture line disruption and 1 leaflet malfunction), 1 left ventricle aneurysmectomy and 1 severe aortic valve stenosis. Overall, the postoperative endocarditis rate was 1.6% (5 of 318).

Freedom from clinical anticoagulation-related events was 99.0 ± 0.6%, 96.8 ± 1.1%, 93.7 ± 1.8% and 89.0 ± 2.7% at 1, 3, 5 and 10 years, respectively. In particular, we observed 25 events categorized as thromboembolic or haemorrhagic. Of these, 21 were neurological events, divided as follows: 4 transient ischaemic attacks (4 of 25), 14 ischaemic major strokes (14 of 25) and 3 haemorrhagic major strokes (3 of 25). The remaining haemorrhagic events were gastrointestinal in 3 patients (3 of 25) and lower limb haematoma in 1 patient (1 of 25).

For both reoperations and clinical anticoagulation-related events, we assumed death to be a competing risk factor. Therefore, cumulative incidence rates were estimated as shown in Fig. 4A and B.

Risk factors for late mortality

Of the 295 patients available for follow-up, 54 (18.3%) patients died after hospital discharge. Sixteen patients died of cardiac events, 7 of which were valve related. One patient died of sepsis, 3 died of multiorgan failure, 2 died of pulmonary disease, 2 died of cancer and 10 died of other diseases. A precise cause of death was not identified in 20 patients (6.3%).

At univariable analysis, age ≥75 years (P = 0.017), hypertension (P = 0.010), COPD (P = 0.012), coronary artery disease (P = 0.001), progressive EuroSCORE (P < 0.001), chronic renal failure (P < 0.001) and associated procedures (P = 0.034) showed a statistically significant correlation with the occurrence of late death. On multiple logistic regression analysis, only hypertension [hazard ratio (HR) 1.91; 95% confidence interval (CI) 1.07–3.40%], COPD (HR 2.91; 95% CI 1.45–5.84%) and chronic renal failure (HR 5.27; 95% CI 2.64–10.53%) appeared to be independent predictors of late mortality (Table 4).

DISCUSSION

There are several conditions for which MVR remains the best option to guarantee long-term durability. According to recent guidelines [1], replacement would be indicated in patients with a reasonable life expectancy for whom future redo valve surgery would be high risk, in patients with anatomical valve conditions with annular calcification or infective disruption and in young patients already on long-term anticoagulation because they are at high risk for thromboembolism. In these selected cases, the choice among different types of mechanical prostheses is usually a matter of surgeon preference and only occasionally is forced by surgical anatomy. In our centre, the first implantation of an On-X mechanical prosthesis occurred in 1996, when the previous director and founder of our centre, Prof A. Pierangeli, replaced both the mitral and aortic valves in a young patient with rheumatic fever. Since then, the use of this prosthesis has progressively increased, becoming the preferred option in Bologna (in 2016, it was used in 99% of MVRs). The selection of a mechanical prosthesis was based primarily on the patient’s age. In cases of difficult reinterventions, permanent atrial fibrillation or other coexisting comorbidities requiring lifelong anticoagulation therapy, we chose to implant this type of prosthesis unless the patient specifically expressed a preference for a tissue valve. According to this policy, 10.1% of the population, even those who were older than 75 years, still received a mechanical prosthesis.

The aim of the current study was to review our experience with the ON-X, especially in terms of valve-related complications and long-term durability.

In-hospital results (4.4% mortality) confirmed what has already been reported in previous series for isolated MVR, where operative mortality was reported to range between 4% and 7% [9–11]. On the other hand, we can add a positive comment about the high-risk profile of this population: cardiac reinterventions were 24.2%, and common causes of surgical treatment were endocarditis (9.4%) and functional mitral regurgitation (9.4%). This finding emphasizes that surgical MVR is used in patients with extreme conditions and is limited to diseases in which a durable repair is unlikely to be achieved. Nevertheless, in our institution, the number of procedures using the On-X prosthesis remained stable over time (Fig. 1). This result could be justified by our conservative approach: replacing instead of repairing the mitral valve in patients with ischaemic regurgitation, rheumatic disease or a severely damaged mitral apparatus.

Our follow-up results for late survival (88.4 ± 2.2% at 5 years and 70.9 ±4.0% at 10 years) were extremely satisfying. A recent multicentre European study on the same topic [11] reported a linearized late mortality rate of 3.6% per patient-year in the subgroup of those having MVR. In comparison, we observed slightly better long-term survival rates not because of the similar ages and low rates of follow-up complications in both studies but probably related to the intrinsic longevity of our patients.

We also investigated the perioperative variables influencing survival. The analysis revealed that they were mostly represented by the common cardiovascular risk factors in developed countries (or their direct consequences). In this analysis, hypertension (HR 1.9; P = 0.027), COPD (HR 2.9; P = 0.003), coronary artery disease (HR 1.9; P = 0.06) and chronic renal failure (HR 5.3; P < 0.001) all appeared to be independent risk factors for late mortality. The surgical procedure itself, along with the type of prosthesis and the underlying mitral valve disease, did not influence the outcomes of these patients.

This finding fits well with the low occurrence of valve-related complications such as endocarditis (1.35%), non-structural dysfunction such as paravalvular leaks (1.02%), structural prosthesis dysfunction (0.68%) and the need for cardiac reinterventions (92.5 ± 2.4% at 10 years) observed at follow-up.

A major concern of the mechanical prosthesis is the lifelong need for anticoagulation and the occurrence of thromboembolic complications. Thromboembolisms usually occur at a rate of 1.5–2.0% per patient-year; this rate increases significantly in patients with chronic atrial fibrillation and large left atrial size [12, 13]. In this cohort, the incidence of thromboembolic events at follow-up was 6.1% (18 of 295), and freedom from anticoagulation-related events was 93.7% at 5 years. In the multicentre study of Chan et al. [5], focusing on thromboembolic and haemorrhagic events using the On-X valve, they reported a 5-year freedom from major thromboembolic events for MVR of 97.7%. Our rate was similar, but we also included the haemorrhagic events that in the Canadian study were accounted for separately. Also, we had a higher percentage of patients with preoperative atrial fibrillation (53% vs 28%).

One strength of our study was our ability to consolidate the results on thromboembolic and haemorrhagic events over the long term. At 10 years, freedom from anticoagulation-related events for the ON-X mitral prosthesis was 89.0%. Several researchers reported acceptable rates for thromboembolisms for other bileaflet mitral valves of 1.1–2.8%/patient-years [14–17]. However, it is difficult to compare our results with those of these studies because there is a substantial heterogeneity in valve designs, patient profiles, INR regimens, the combined end point for thromboembolic complications and methods of reporting data at follow-up.

Japanese researchers recently [18] investigated 78 patients who had an On-X bileaflet mechanical aortic valve and who were treated with a lower dose of warfarin (INR 1.5–2.0) according to the new evidence from the Prospective Randomized On-X Anticoagulation Clinical Trial. Freedom from major valve-related stroke and cerebral bleeding events was 93.3% and 98.6%, respectively. We obtained almost identical results (at 5 years, the number was 93.7 ± 1.8%) after MVR in a population of patients, 52% of whom had preoperative atrial fibrillation. Therefore, the high regimen of anticoagulation therapy required for patients with the mechanical mitral valve was not a strong limitation to the use of this prosthesis. Indeed, the US Food and Drug Administration recommendation for a lower anticoagulation regimen in the aortic position is proof of the low thrombogenicity of the On-X prosthesis because of its pure pyrolytic carbon structure and improved flow dynamics (actuated pivot, flared inlet, 2-point leaflet contact and 180° opening) [3, 19, 20].

Another relevant point of discussion is the incidence of prosthetic valve obstruction after hospital discharge. The most common causes usually include valve thrombosis or pannus formation. Thrombosis is more likely to occur early after valve implantation, in the setting of an inadequate anticoagulation, whereas pannus is a more chronic process associated with ingrowth of tissue that causes an obstruction [21]. According to the data collected at follow-up, this condition has been eradicated with the use of this prosthesis (0 events in the overall population). This result is directly related to the valve design: a main body that fully extends through the annulus and provides protection from contact with tissue. This feature allows us also to safely preserve the entire valvular apparatus in most cases, excluding cases of infective endocarditis, where we perform a more extensive resection of the mitral valve.

The major findings of the study are as follows: (i) in the era of transcatheter technology, conventional MVR with a mechanical prosthesis is still needed. (ii) The reliability of the latest generation mechanical heart valve is confirmed by the rare prosthesis-related complications observed at follow-up. (iii) Thromboembolism is a major concern with mitral valve prostheses, requiring postoperative anticoagulation, but the incidence of this complication after implantation of an On-X mitral valve is limited.

Limitations and strengths

The limitations of our study are related to its retrospective and single-institutional nature. No definite conclusions may be drawn regarding the safety of this prosthesis.

The consistent follow-up combined with the large number of patients may contribute to an increase in the interest and significance of this study regardless of the observational design.

The strengths of the present analysis include the long-term experience with this mechanical prosthesis in our centre during the last 20 years and the unique population in which all patients were operated on using well-defined techniques and protocols, ensuring uniform patient care.

This study has also the advantage of comprising one of the largest series published to date regarding patients undergoing MVR using the On-X prosthesis.

Conflict of interest: none declared.

Footnotes

REFERENCES