Reduction of neurologic complications by intra-aortic filtration in patients undergoing combined intracardiac and CABG procedures

Abstract

Objective: Patients undergoing combined intracardiac and coronary artery bypass grafting (CABG) procedures are at high risk for neurologic events. Previously the Multicenter Study of Perioperative Ischemia (McSPI) Study Group demonstrated that risk factors for neurologic complications in these patients related primarily to embolization. Objective: This study investigated risk factors and outcomes in 185 high-risk patients receiving intra-aortic filtration during combined intracardiac and CABG procedures. Methods: Seventeen European centers comprising the International Council of Emboli Management (ICEM) pooled data on patients receiving intra-aortic filtration (EMBOL-X, Mountain View, CA, USA) into a consecutive enrollment registry. Preoperative, intraoperative, and outcome variables were compared between ICEM patients undergoing combined CABG and intracardiac procedures and the corresponding McSPI subgroup. Results: Type I outcomes (stroke, transient ischemic attack, coma, death due to neurologic causes) occurred in 2.2% (4/185) of ICEM patients compared to 8.4% (23/273) of McSPI patients. No ICEM patient suffered a fatal stroke compared with five in the McSPI study (1.8%). One nonfatal stroke was observed in the ICEM group (0.5%) and 16 were observed in the McSPI group (5.9%). Three patients (1.6%) were reported with transient ischemic attack in the ICEM group, compared to two in the McSPI study (0.7%). One of the reported patients in the ICEM group with transient ischemic attack also suffered a coma that was still evident at discharge. In all, four patients suffered Type I neurologic events in the ICEM group (2.2%), compared to 23 in the McSPI group (8.4%). In the ICEM study group, four patients (2.2%) died due to non-neurologic causes, compared to 18 in the McSPI group (6.6%). Conclusions: Even with a significantly worse preoperative profile, high-risk patients receiving intra-aortic filtration suffered Type I cerebral injury 74% less often than patients in the McSPI group.

Intraaortic filtration, Neurologic outcome, Multicenter Study of Perioperative Ischemia, International Council of Emboli Management

1 Introduction

The persistence of adverse neurologic events after cardiac surgery continues to confound cardiac surgeons seeking to improve procedural outcomes. As a result of the multifactorial dimensions of neurologic damage during cardiac surgery, it is necessary to gather large study populations to design clinical trials capable of yielding definitive data [1]. One of the largest study groups to date, the Multicenter Study of Perioperative Ischemia (McSPI) has addressed this topic in numerous papers, culminating in the article by Roach and collaborators, which reported a 6.1% incidence of overall adverse cerebral outcomes, including a 3.1% incidence of severe Type I outcomes (defined as stroke, transient ischemic attack, coma and death due to neurologic causes) in 2417 patients across 24 United States centers [2].

The bulk of evidence points to aortic atheromatosis as the most important risk factor [3–6]. This suggests that particulate embolization plays a key role in the etiology of adverse neurologic outcomes in cardiac surgery. In 1999 the International Council of Emboli Management (ICEM) was formed to investigate the effect of the use of an intra-aortic filter designed to capture and remove particulate emboli from the ascending aorta [7]. If the embolization hypothesis is correct, removing particulate emboli should have a beneficial effect on neurologic outcome.

Previous publications have shown that filtration successfully removes particulate debris [8,9]. Recently Schmitz, Blackstone and the ICEM Study Group showed that intra-aortic filtration may reduce major perioperative neurologic events in coronary artery bypass grafting (CABG) patients by as much as 50% [7]. Results were obtained by analyzing CABG patients enrolled in the ICEM database and comparing observed outcomes to expected outcomes using a published, validated stroke risk index [7,10]. However, the patients examined in that analysis were limited to selected non-emergent CABG patients in whom the risk for adverse neurologic events was moderate.

A more focused way of investigating intra-aortic filtration is to isolate a high-risk group for treatment and analysis, since the effect of intra-aortic filtration should be more pronounced in such a targeted group. However, there is some debate as to what constitutes an appropriate high-risk group. While much attention has been paid to the preoperative risk factors for adverse neurologic outcomes, the type of the procedure may be an equally important variable. In a McSPI study by Wolman and colleagues, patients undergoing open-heart procedures in combination with CABG were found to suffer Type I outcomes at a rate of 8.4% as opposed to the 3.1% in the entire patient population [11]. They conclude that this subgroup of cardiac surgery patients is at extraordinary risk for neurologic injury.

This study was designed to look at the effect of particulate removal by intra-aortic filtration on a defined group of extraordinarily high-risk patients undergoing combined CABG and intracardiac procedures.

2 Patients and methods

2.1 Patients

Between February 1999 and August 2001 a total of 1796 patients had been enrolled in a prospective consecutive observational study of patients receiving intra-aortic filtration at 17 European centers participating in the ICEM study group (see  Appendix A). From this database, the subset including every patient undergoing an intracardiac procedure combined with CABG surgery was extracted. The number of patients analyzed in this subgroup was 185. An additional 13 patients were excluded from the group due to incomplete outcomes data. (Missing outcomes data were attributed to the normal lagtime between enrolling a patient in the study and gathering outcomes data at discharge.) The mean age of patients was 70.5 years (range 35–84). These patients were compared to the corresponding subgroup (273 patients) of the McSPI study. Patient demographics including specific risk factors are listed in Table 1 .

Table 1

Patient history and risk variables

2.2 Methods

The 120-μm intra-aortic filter (EMBOL-X Inc., Mountain View, CA, USA) was placed in the aorta immediately prior to cross-clamp release, and removed when the heart was fully ejecting. The filter was integrated in the arterial return cannula and was designed to extract particulate debris from the ascending aorta proximal to the cannula tip [12]. Full details of the methods of filter usage have been previously described [7–9]. Intra-aortic filtration was used in all ICEM combined procedures described in Table 2 . Atherosclerosis of the ascending aorta was assessed by one or more of the following methods: palpation (79%), transesophageal echocardiography (32%) or epiaortic scanning (17%).

Table 2

Procedures performed

2.3 Definition of outcomes

Type I neurologic outcomes included stroke, transient ischemic attack (TIA), coma at discharge, or death due to stroke or hypoxic encephalopathy as determined by a site investigator at each institution [2]. Type II neurologic deficits such as deterioration in intellectual function, confusion, agitation, disorientation, memory deficit, or seizure were not evaluated in this study. The reporting period included the time from the operation to discharge.

2.4 Statistical methodology

Demographic data, procedural data and neurologic outcomes were analyzed and compared between the ICEM study group and the McSPI study group to evaluate the effect of intra-aortic filtration. Variables that matched across studies were compared using the chi-square test, or Fisher's exact test if a category contained less than 10 data points. The alpha level for significance was considered to be 0.05.

3 Results

3.1 Demographic risk variables

Occurrence of demographic risk factors are summarized and compared to Wolman's findings in Table 1[11]. A significantly higher proportion of ICEM patients were over 70 years of age compared with McSPI patients. Evidence of atherosclerosis of the ascending aorta was significantly higher in the ICEM group as well. However, more McSPI patients underwent redo cardiac procedures than ICEM patients. In the remaining demographic risk factors, the ICEM and McSPI patient subsets were not significantly different.

3.2 Operative procedures

Operative procedures are summarized in Table 2. ICEM procedures were grouped into the same three categories that McSPI had reported in order to enable cross-study comparison. This comparison showed no significant difference across the two studies (P=0.10).

3.3 Observed neurologic events

As shown in Table 3 , no ICEM patient suffered a fatal stroke compared with five in the McSPI study (1.8%). One nonfatal stroke was observed in the ICEM group (0.5%) and 16 were observed in the McSPI group (5.9%). Three patients (1.6%) were reported with TIA in the ICEM group, compared to two in the McSPI study (0.7%). One of the reported TIA patients in the ICEM group also suffered a coma that was still evident at discharge. In all, four patients suffered Type I neurologic events in the ICEM group (2.2%), compared to 23 in the McSPI group (8.4%). In the ICEM study group, four patients (2.2%) died due to non-neurologic causes, compared to 18 (6.6%) in the McSPI group.

Table 3

International Council of Emboli Management cerebral outcomes

4 Discussion

Adverse neurologic outcomes are still the most relevant perioperative morbidity in cardiac surgery. The severity of these events is far reaching, especially for the patient who must endure a life altering and likely permanent deficit. The strain on already overburdened budgets is enormous as hospitals struggle in an increasingly competitive environment. Although pathogenicity of cardiopulmonary bypass has been reduced to a minimum through advances in technology, adverse neurologic events continue to plague cardiac surgeons worldwide [13,14]. The emergence of consensus surrounding the importance of aortic atherosclerosis as one of the most predictive risk factors for these devastating events has focused the quest for improved outcomes to the question of embolic generation [4–6]. The development of an intra-aortic filter has allowed us to begin to examine the relevance of particulate removal as a means of improving neurologic outcomes.

Previous studies by Barbut demonstrate the embolic activity associated with surgical manipulation of the heart and the aorta during surgery [4]. The filter is intended to be deployed during the surgical period most associated with embolization, including partial clamping, sewing anastomoses, and removal of the aortic cross-clamp. In previous intra-aortic filtration studies it has been shown that particulate material is recovered in over 90% of cardiac surgery procedures [8,9]. Future versions of the filter may increase the deployment period to protect even more surgical manipulations of the aorta.

While the benefit of particulate removal may seem to be intuitive, the clinical relevance has yet to be fully defined. Efforts to study this phenomenon are complicated due to the multifactoral causes of adverse neurologic events. In addition to particulate and gaseous emboli, it has been shown that the etiology of adverse neurologic events can be attributed to reduced cerebral blood flow, hyperthermia on rewarming, and systemic inflammatory response [14–17].

Since the actual incidence of severe neurologic events is relatively low, ranging from 1 to 6%, large populations are required for study [18]. Even larger populations are needed to demonstrate improvement in outcomes. The formation of large multicenter study groups such as the McSPI and ICEM groups are necessary in order to gather enough patients for meaningful analysis.

Recently the ICEM study group showed evidence that particulate removal in CABG patients reduced the incidence of severe neurologic events [7]. In the present study, we selected clearly defined high-risk patients to further confirm the clinical benefit and to begin to identify the specific patient populations likely to achieve the greatest benefit. Combined CABG and intracardiac procedures pose a high risk of embolization due to excision of heavily degenerated valves, resulting in surgical debris.

The McSPI study is particularly well suited for co-analysis with the ICEM study due to its size, design, and structure. Both studies look at similar numbers of patients (2107 for McSPI, 1796 for ICEM) drawn from similar numbers of centers (24 for McSPI, 17 for ICEM). The large numbers and multicenter approach help to minimize bias inherent in smaller studies at fewer centers. In addition, the Wolman analysis of the McSPI data clearly identifies a group highly likely to benefit from intra-aortic filtration [11].

We found several statistically significant differences in the comparison of preoperative risk factors between the ICEM and McSPI data. The proportion of patients older than 70 years was significantly higher in the ICEM group (66 vs. 52%, P<0.002) and the incidence of proximal aortic atherosclerosis was also significantly higher. (52 vs. 30%, P<0.0001). This result may partially be due to some use of more sensitive detection techniques in the ICEM group, such as epiaortic scanning (17%) and transesophageal echo (TEE) (32%). Since palpation was the dominant method for detecting aortic atherosclerosis in both groups (and the only method used in McSPI), the highly significant difference between the groups is assumed to be valid.

These data reflect the changing characteristics of patient populations throughout the last decade. Age and atherosclerosis have been identified as strong predictors for neurologic events. In the McSPI study published by Roach and colleagues, atheromatosis of the proximal aorta was shown to be the most significant predictor of neurologic events [2]. The incidence of neurologic events following cardiac surgery has been shown to dramatically increase in patients over the age of 70 [3].

Significantly more McSPI patients underwent redo operations. However, since no study has identified reoperations as a significant risk factor for neurologic injury, this should have little impact on the preoperative risk profile of these patients. Therefore, we can derive that the ICEM patient population in this comparison was at greater risk for adverse neurologic events than the McSPI group. Nevertheless, the incidence of neurologic outcomes was much lower in the ICEM group (2.2%) than in the McSPI group (8.4%), confirming that particulate removal clearly benefits this high-risk group of patients.

These results indicate that particulate filtration has a positive impact on patients who receive concomitant CABG and intracardiac procedures. This is also consistent with previous reports of improved neurologic outcomes in CABG patients who received intra-aortic filtration [7]. However, the existence of gaseous emboli must be given special consideration in the subset of patients with concomitant procedures. It is commonly thought that gaseous emboli pose a greater risk in open-heart procedures and could be an additional cause of adverse neurologic events. Attempts to study the impact of gaseous versus particulate debris have been hampered by the inability of transcranial Doppler to distinguish between gaseous and particulate matter [4]. In 172 of 177 (97%) filters analyzed in this study, particulate matter was recovered, yet the incidence of adverse neurologic outcomes decreased. This result calls into question the role of gaseous microemboli as causative factors in major neurologic adverse events. It may be posited that particulate matter is more likely to cause severe neurologic events than gaseous emboli. These data may also be explained by the impact of the filter on large air bubbles. Perhaps large air bubbles are reduced to smaller ones simply by passing through the filter mesh, and therefore have less of a clinical impact. Further study is needed to understand the relative impact of gaseous and particulate debris on the brain.

4.1 Limitations

This study was based on a prospective registry but results would be even stronger if a randomized protocol to examine adverse neurologic events was feasible. While the relatively low incidence of adverse neurologic events requires a very large study population, the selection of a high-risk group of patients should narrow the potential study population. Due to the infrequent occurrence of concomitant CABG and intracardiac procedures, and the still fairly low incidence of adverse neurologic events, a randomized protocol to study the impact of intra-aortic filtration in these high-risk patients has not been completed to date.

Other limitations of this analysis include the comparison of European and American data, which may include technical and procedural differences, and the comparison of data collected from 1991 to 1993 with data collected from 1999 to 2001. In addition, in this analysis ICEM neurologic outcomes data were entirely self-reported, while the McSPI outcomes were adjudicated by a panel to ensure maximum validity and consistency.

Increased awareness by the surgeon of neurologic damage caused by particulate emboli generated in the ascending aorta during open heart surgery may also lead to more gentle handling. This modification of surgical technique may also contribute to the reduction of adverse neurologic events seen in this data set.

Finally, although a total of 198 patients qualified for this analysis from the ICEM database, complete data were available on 185 patients. We have attempted to address this by indicating the total number of responses for each field analyzed.

4.2 Conclusion

Even with a significantly worse preoperative profile, high-risk patients receiving intra-aortic filtration suffered Type I cerebral injury 74% less often than patients in the McSPI group. The beneficial effect of intra-aortic filtration seems to be augmented in this high-risk group when compared to the 50% reduction in expected neurologic outcomes previously reported by Schmitz, Blackstone and the ICEM Study Group [7].

Appendix A. International Council of Emboli Management Study Group

Appendix B. Conference discussion

Dr R. de Vivie (Cologne, Germany): Did you ever have a complication by cannulating the aorta with this type of cannula?

Dr Wimmer-Greinecker: Not in this subgroup. But in the overall group there was one Type A dissection reported in the 1,700 patients.

Dr J. Pirk (Prague, Czech Republic): I would like to ask you where the big emboli came from? In your oxygenator you have filtration also.

Dr Wimmer-Greinecker: This comes directly either from the ascending aorta or from the heart. This particular one was a patient, who was operated on with an intracavitary thrombus and a ventricular aneurysm, and that is what we caught.