Abstract
Transcatheter aortic valve implantation (TAVI) has been introduced as an alternative to conventional surgery for high-risk patients with aortic stenosis. A recently published randomized clinical trial demonstrated reduction of mortality in high-risk or inoperable patients when compared to medical treatment or balloon aortic valvuloplasty. Despite this evidence of superiority, the rate of TAVI complications is high, and perhaps the most devastating of the nonfatal complications is cerebral injury. This review will compare the incidence of stroke and “silent” cerebral injury after surgical aortic valve replacement and after TAVI and will discuss mechanisms that can lead to cerebral injury during these procedures and subsequently how to prevent this with new protection devices.
1
Introduction
Aortic stenosis (AS) is the most common valvular disease in adults. Its prevalence increases with age, reaching 2%–4% in patients over 85 years . Once patients with severe AS become symptomatic or develop left ventricular dysfunction, life expectancy is significantly reduced . Because of this dire prognosis, surgical aortic valve replacement (AVR) is recommended for such patients and substantially improves outlook.
Transcatheter aortic valve implantation (TAVI) has been introduced as a less morbid alternative to conventional surgery for high-risk patients with AS . A recently published randomized clinical trial demonstrated improved outcomes in high-risk or inoperable patients when compared to medical treatment or balloon aortic valvuloplasty (BAV) . However, despite this evidence of superiority, the absolute complication rate of TAVI is high, especially in this high-risk population.
Perhaps the most devastating of the nonfatal complications is cerebral injury. Brain injury, an infrequent complication of all cardiac surgeries, may also occur after TAVI. This review will compare the available evidence bearing on the incidence of stroke and “silent” cerebral injury after surgical AVR and after TAVI. It will also discuss mechanisms that can lead to cerebral injury during these procedures.
2
Cerebral injury after surgical AVR
2.1
Clinical stroke
The frequency of overt neurological injury varies with the type of cardiac surgery. Retrospective studies report an incidence of 0.8%–3.2% after coronary artery bypass grafting (CABG) , while prospective studies report 1.5%–5.2% . A similar incidence (0.7%–5.0%) has been reported in a clinical series of isolated AVR ( Table 1 ). The incidence is higher if the AVR is combined with CABG or mitral valve replacement .
| Study | Design of study | Incidence of stroke |
|---|---|---|
| Bucerius et al. | Prospective Apr 1996–Aug 2001 1830 AVR 2506 CABG+valve surgery | AVR: 4.8% CABG+valve surgery: 7.4% |
| Filsoufi et al. | Prospective Jan 1998–Dec 2006 2808 valve surgery±CABG | Overall 2.2% ( n =63, AVR n =17) 1998: n =3.3% 2002: n =1.3% <50 years: 0.6% 50–79: 1.5% >80: 2.5% |
| Gulbins et al. | Prospective 1996–2005 1014 stentless AVR | Perioperative cerebral event: 1.8% |
| Brown et al. | Prospective 108,687 patients 928 hospitals 1997–2006 | 1997: 1.7% 2006: 1.3% In 2006: <70 years: 0.7% 70–80 years: <2% >80 years: <2.5% |
| Bakaeen et al. | Retrospective 1991–2007 7142 patients isolated AVR | Incidence of stroke <80 years: 1.9% >80: 2.4% |
| Wolman et al. | Prospective 273 patients, 24 centers | Cerebral outcomes: 16% 8.5%: 5 cerebral deaths, 16 nonfatal strokes, 2 TIA 7.3%: 17 new intellectual deteriorations, 3 seizures |
Not only can stroke have a devastating effect on quality of life, but its occurrence is also strongly related to mortality. Hospital mortality in patients with stroke can reach 24%, while the mortality rate in similar patients without cerebral complication is 4.6% . Unsurprisingly, median length of hospital stay is also substantially increased in patients with stroke.
Clinical risk factors for perioperative stroke in cardiac valve surgery include older age, female gender, previous cerebrovascular disease, previous peripheral vascular disease, diabetes mellitus, hypertension, previous cardiac surgery, and urgent operation. In addition, certain facets of the procedure also affect the incidence of perioperative stroke, e.g., cardiopulmonary bypass time longer than 2 h, use of hemofiltration, and requirement for more than usual blood transfusion .
The importance of age is especially noteworthy. For example, Brown et al. found a postoperative stroke rate of 0.7% in those younger than 70 years and 2.5% in those older than 80 years. It is therefore surprising that despite the fact that the age of patients undergoing AVR has increased in the past 10 years, the rate of death or stroke has decreased for isolated AVR .
2.2
Asymptomatic brain injury
A new imaging technology, diffusion-weighted magnetic resonance imaging (DWMRI), has proven to be more sensitive than computed tomography and conventional magnetic resonance imaging (MRI) for the detection of brain lesions . This technique is sensitive to changes in the mobility of water molecules; lowered water mobility is an early event in the ischemic tissue change.
After cardiac surgery, this modality often demonstrates small, multiple subcortical lesions. Since these lesions appear in patients with no overt signs of brain injury, they can be regarded as “silent brain injury” . The appearance of new DWMRI lesions following surgery has been tentatively associated with the presence and severity of preexisting white matter lesions as well as with age, preexisting T2 lesion volume, and postoperative S100β protein .
New, silent DWMRI lesions appear so frequently after a variety of invasive cardiovascular procedures that Bendszus and Stoll have referred to them as “fingerprints of invasive medical procedures.” Several studies ( Table 2 ) performed DWMRI before and up to 4 months after a cardiac surgical procedure. The incidence of new lesions postoperatively was surprisingly high at 38%–71%.
| Study | Study design | Incidence of clinical cerebral event | Incidence of new MRI lesions |
|---|---|---|---|
| Stolz et al. | Prospective 45 patients, 37 pre- and postoperative (AVR) DWMRI | 3 focal neurological deficits | New postoperative DWMRI lesions in 38% |
| Wityk et al. | Retrospective 14 patients with postoperative (cardiac surgery) DWMRI | 14 patients with clinical cerebral events | New cerebral lesions in 71% |
| Knipp et al. | Prospective 30 patients cardiac valve replacement DWMRI before and after | No focal neurological deficit Impaired cognitive function in 5 of 13 tests (all resolved within 4 months) | New focal brain lesions in 47% |
| Floyd et al. | Prospective 34 patients DWMRI Cardiac surgery | 2 clinical strokes in the AVR group | New infarction in 6 of 15 (40%) in procedures involving aortic valve replacement |
2.3
Neurocognitive decline after AVR
Most patients with new lesions have no overt neurological signs, but there has been considerable interest in a possible association between their appearance and postprocedure cognitive dysfunction. Supporters of this proposition point to the apparent connection between DWMRI lesions and cognitive dysfunction in population-based studies of healthy subjects. While DWMRI lesions are detected in up to 20% of putatively normal elderly participants, they are more frequent and larger in size in those with evidence of dementia.
By means of careful psychological testing, neurocognitive deficits have been observed in 33%–83% of patients in the first weeks after cardiac surgery . As a rule, these deficits resolve within 4 months , but a fear exists among patients, cardiologists, and surgeons that persistent or progressive problems may develop. Evidence points to older age and preoperative vascular disease as major risk factors for postoperative cognitive dysfunction .
An association between the postoperative appearance of new small, “silent,” subcortical brain injury and postoperative cognitive dysfunction is intuitively attractive, but confirmation of this relationship has been elusive . Major practical constraints hinder a careful examination of this association. It is expensive and can be awkward to schedule serial DWMRI studies in the setting of pre- and postoperative clinical care. Moreover, expertise in conducting the required neuropsychological testing can also be expensive and difficult to schedule.
Thus, surgical AVR is associated with a clinical stroke rate of 0.8%–5%. “Silent” brain injury may be detected by sophisticated imaging in about half of surgical AVR patients. At first glance, such small “silent” lesions result in more concern for physicians and patients than their permanent clinical impact deserves, but their possible association with subtle, permanent cognitive impairment appropriately heightens interest in this phenomenon.
2
Cerebral injury after surgical AVR
2.1
Clinical stroke
The frequency of overt neurological injury varies with the type of cardiac surgery. Retrospective studies report an incidence of 0.8%–3.2% after coronary artery bypass grafting (CABG) , while prospective studies report 1.5%–5.2% . A similar incidence (0.7%–5.0%) has been reported in a clinical series of isolated AVR ( Table 1 ). The incidence is higher if the AVR is combined with CABG or mitral valve replacement .
| Study | Design of study | Incidence of stroke |
|---|---|---|
| Bucerius et al. | Prospective Apr 1996–Aug 2001 1830 AVR 2506 CABG+valve surgery | AVR: 4.8% CABG+valve surgery: 7.4% |
| Filsoufi et al. | Prospective Jan 1998–Dec 2006 2808 valve surgery±CABG | Overall 2.2% ( n =63, AVR n =17) 1998: n =3.3% 2002: n =1.3% <50 years: 0.6% 50–79: 1.5% >80: 2.5% |
| Gulbins et al. | Prospective 1996–2005 1014 stentless AVR | Perioperative cerebral event: 1.8% |
| Brown et al. | Prospective 108,687 patients 928 hospitals 1997–2006 | 1997: 1.7% 2006: 1.3% In 2006: <70 years: 0.7% 70–80 years: <2% >80 years: <2.5% |
| Bakaeen et al. | Retrospective 1991–2007 7142 patients isolated AVR | Incidence of stroke <80 years: 1.9% >80: 2.4% |
| Wolman et al. | Prospective 273 patients, 24 centers | Cerebral outcomes: 16% 8.5%: 5 cerebral deaths, 16 nonfatal strokes, 2 TIA 7.3%: 17 new intellectual deteriorations, 3 seizures |
Not only can stroke have a devastating effect on quality of life, but its occurrence is also strongly related to mortality. Hospital mortality in patients with stroke can reach 24%, while the mortality rate in similar patients without cerebral complication is 4.6% . Unsurprisingly, median length of hospital stay is also substantially increased in patients with stroke.
Clinical risk factors for perioperative stroke in cardiac valve surgery include older age, female gender, previous cerebrovascular disease, previous peripheral vascular disease, diabetes mellitus, hypertension, previous cardiac surgery, and urgent operation. In addition, certain facets of the procedure also affect the incidence of perioperative stroke, e.g., cardiopulmonary bypass time longer than 2 h, use of hemofiltration, and requirement for more than usual blood transfusion .
The importance of age is especially noteworthy. For example, Brown et al. found a postoperative stroke rate of 0.7% in those younger than 70 years and 2.5% in those older than 80 years. It is therefore surprising that despite the fact that the age of patients undergoing AVR has increased in the past 10 years, the rate of death or stroke has decreased for isolated AVR .
2.2
Asymptomatic brain injury
A new imaging technology, diffusion-weighted magnetic resonance imaging (DWMRI), has proven to be more sensitive than computed tomography and conventional magnetic resonance imaging (MRI) for the detection of brain lesions . This technique is sensitive to changes in the mobility of water molecules; lowered water mobility is an early event in the ischemic tissue change.
After cardiac surgery, this modality often demonstrates small, multiple subcortical lesions. Since these lesions appear in patients with no overt signs of brain injury, they can be regarded as “silent brain injury” . The appearance of new DWMRI lesions following surgery has been tentatively associated with the presence and severity of preexisting white matter lesions as well as with age, preexisting T2 lesion volume, and postoperative S100β protein .
New, silent DWMRI lesions appear so frequently after a variety of invasive cardiovascular procedures that Bendszus and Stoll have referred to them as “fingerprints of invasive medical procedures.” Several studies ( Table 2 ) performed DWMRI before and up to 4 months after a cardiac surgical procedure. The incidence of new lesions postoperatively was surprisingly high at 38%–71%.
| Study | Study design | Incidence of clinical cerebral event | Incidence of new MRI lesions |
|---|---|---|---|
| Stolz et al. | Prospective 45 patients, 37 pre- and postoperative (AVR) DWMRI | 3 focal neurological deficits | New postoperative DWMRI lesions in 38% |
| Wityk et al. | Retrospective 14 patients with postoperative (cardiac surgery) DWMRI | 14 patients with clinical cerebral events | New cerebral lesions in 71% |
| Knipp et al. | Prospective 30 patients cardiac valve replacement DWMRI before and after | No focal neurological deficit Impaired cognitive function in 5 of 13 tests (all resolved within 4 months) | New focal brain lesions in 47% |
| Floyd et al. | Prospective 34 patients DWMRI Cardiac surgery | 2 clinical strokes in the AVR group | New infarction in 6 of 15 (40%) in procedures involving aortic valve replacement |
2.3
Neurocognitive decline after AVR
Most patients with new lesions have no overt neurological signs, but there has been considerable interest in a possible association between their appearance and postprocedure cognitive dysfunction. Supporters of this proposition point to the apparent connection between DWMRI lesions and cognitive dysfunction in population-based studies of healthy subjects. While DWMRI lesions are detected in up to 20% of putatively normal elderly participants, they are more frequent and larger in size in those with evidence of dementia.
By means of careful psychological testing, neurocognitive deficits have been observed in 33%–83% of patients in the first weeks after cardiac surgery . As a rule, these deficits resolve within 4 months , but a fear exists among patients, cardiologists, and surgeons that persistent or progressive problems may develop. Evidence points to older age and preoperative vascular disease as major risk factors for postoperative cognitive dysfunction .
An association between the postoperative appearance of new small, “silent,” subcortical brain injury and postoperative cognitive dysfunction is intuitively attractive, but confirmation of this relationship has been elusive . Major practical constraints hinder a careful examination of this association. It is expensive and can be awkward to schedule serial DWMRI studies in the setting of pre- and postoperative clinical care. Moreover, expertise in conducting the required neuropsychological testing can also be expensive and difficult to schedule.
Thus, surgical AVR is associated with a clinical stroke rate of 0.8%–5%. “Silent” brain injury may be detected by sophisticated imaging in about half of surgical AVR patients. At first glance, such small “silent” lesions result in more concern for physicians and patients than their permanent clinical impact deserves, but their possible association with subtle, permanent cognitive impairment appropriately heightens interest in this phenomenon.
3
Brain injury and TAVI
3.1
Clinical strokes
To this point, patients are referred to TAVI specifically because they are at high risk for surgical AVR and are, therefore, typically older and sicker than those who undergo surgical AVR. As a consequence, they arguably also have a higher risk for perioperative brain injury.
The early experience with TAVI indicates that the incidence of clinically apparent neurological events varies widely. A broad review of all reported series reveals a range of such events of from 0% to 10% ( Table 3 ). It is important to note that these reports describe initial experiences with a variety of approaches to TAVI, including transfemoral and transapical procedures, as well as utilization of both the Edwards valve (Edwards LifeSciences LLC, Irvine, CA, USA) and the CoreValve (Medtronic CoreValve LLC, Minneapolis, MN, USA).
