Retrospective versus Prospective Neurologic Assessment

The detection of CNS injury depends critically on the methodology used, and retrospective studies have been deemed insensitive by different authors.^23,24,30,31 As Sotaniemi³¹ demonstrated, a retrospective chart review is inadequate as an assessment of the overall incidence of postoperative neurologic dysfunction. In a study of 100 patients in whom a 37% incidence rate of neurologic dysfunction had been diagnosed by careful neurologic examination, the prevalence rate of cerebral abnormalities detected by retrospective analysis of the same patient pool was only 4%. The reasons for the inability of retrospective chart audit to detect the majority of patients with neurologic dysfunction are readily apparent and include incompleteness of records, a reluctance to document apparently minor complications, and most important, an insensitivity to subtle neurologic dysfunction. Many of the types of neurologic impairment now being documented are subclinical and not readily detectable by a standard “foot-of-the-bed” assessment. The timing, thoroughness, and reproducibility (single examiner) of the neurologic examinations, as well as the incorporation of a preoperative assessment for comparison, all determine the sensitivity and accuracy with which postoperative CNS injury can be detected.^23,24,30,32

Valvular versus Coronary Artery Bypass Graft Surgery

It appears that increasing the complexity or undertaking open chamber–type procedures increases the risk for CNS injury. Ebert et al³³ prospectively studied 42 patients who underwent valve replacement surgery and 42 patients for CABG, with both groups matched post hoc for age, sex, and preoperative cognitive status.

Patients were investigated before surgery, as well as 2 and 7 days after surgery, with a comprehensive neuropsychological and neuropsychiatric assessment. Valve replacement surgery patients exhibited more severe neuropsychological deficits and showed a slower recovery than patients who underwent CABG.³³ In Alexander et al’s¹² study of 64,467 patients who underwent CABG alone and 3297 patients who underwent CABG in conjunction with aortic valve replacement or CABG in conjunction with mitral valve repair or replacement, the incidence rate of type I cerebral injury in patients younger than 80 years was 4.2% for CABG, 9.1% for CABG with aortic valve replacement, and 11.2% for CABG with mitral valve repair or replacement.¹² Notably, the total CPB time was 96 minutes for CABG, 148 minutes for CABG in conjunction with aortic valve replacement, and 161 minutes for CABG with mitral valve repair or replacement. It thus remains unclear whether it is the procedure itself or the prolonged duration of CPB, either acting directly or as a marker of a greater surgical difficulty and thus perioperative hemodynamic instability, that is fundamentally causative.³⁴

Wolman et al²⁰ prospectively studied 273 patients from 24 U.S. medical centers who underwent combined intracardiac surgery and CABG. Included were clinical, historic, specialized testing, neurologic outcome, and autopsy data, and measures of resource utilization. Adverse cerebral outcomes occurred in 16% of patients (43/273), being nearly equally divided between type I cerebral injury (8.4%; 5 cerebral deaths, 16 nonfatal strokes, and 2 new TIAs) and type II cerebral injury (7.3%; 17 new intellectual deteriorations persisting at hospital discharge and 3 newly diagnosed seizures), rates of injury 2- to 3-fold greater than demonstrated after CABG alone by this same group of investigators.²⁰ Associated resource utilization was significantly increased according to type of CNS injury: prolonging median intensive care unit (ICU) stay from 3 days (no adverse cerebral outcome) to 8 days associated with type I injury and from 3 to 6 days in those patients with type II injury. Significant risk factors for type I cerebral injury related primarily to embolic phenomena, including proximal aortic atherosclerosis, intracardiac thrombus, and intermittent clamping of the aorta during surgery. Risk factors for type II cerebral injury included proximal aortic atherosclerosis, as well as a preoperative history of endocarditis, alcohol abuse, perioperative arrhythmia, or poorly controlled hypertension, and the development of a low-output state after CPB.²⁰

Circulatory Arrest

pH Management

The milieu in which HCA is conducted may well also have an important impact on CNS outcomes but has not yet been systematically investigated in adult patients. Although clinical studies and experimental evidence point to a benefit of pH-stat management in infants and children undergoing HCA, it should be noted that neither the clinical studies in pediatric patients⁶¹ nor the experimental models using nonatheromatous animals are necessarily relevant to the adult patient who invariably has substantial atheromatous disease within the ascending aorta, often with concomitant extracranial and intracranial involvement. For adults undergoing moderate hypothermic CPB at least, the weight of evidence from CNS outcomes of at least three separate, prospective, randomized, clinical trials supports alpha-stat pH management over pH-stat.^62–64 In this context, alpha-stat has also been associated with decreased cerebral embolization⁶⁵ and preservation of cerebral autoregulation,^63,66 factors likely of paramount importance in perioperative CNS injury in adult patients undergoing HCA.

In none of these studies were stroke rate, mortality, or other measures of morbidity influenced by treatment mode, though all were underpowered to detect such outcomes. However, Hagl et al⁶⁷ retrospectively analyzed outcomes in 717 survivors of ascending aortic and aortic arch surgery. They determined that the method of cerebral protection did not influence the occurrence of stroke, but that ACP did result in a significant reduction in the incidence of temporary neurologic dysfunction, a result not seen after RCP. Directionally similar results demonstrating that antegrade perfusion was associated with significantly lower incidences of temporary neurological complications, earlier extubation, shorter ICU stay, and shorter hospitalization in comparison with patients managed with RCP has been shown by Apostolakis and colleagues.⁶⁸ Halkos et al⁶⁹ demonstrated that during proximal aortic surgery, selective ACP was associated with lower mortality, as well as improved resource utilization and fewer pulmonary and renal complications.

It is unlikely that pH management will substantially change the results of RCP versus ACP discussed earlier; however, Harrington et al’s⁷⁰ study used alpha-stat management, whereas Reich et al’s⁵⁵ study used pH-stat management, both with similar directional results relatively unfavorable to RCP. Whether pH management will influence the overall incidence of CNS dysfunction after HCA is unknown. A recent meta-analysis has concluded that in the absence of randomized trials, pH-stat management for infants and alpha-stat for adults would appear to be most appropriate strategies for patients undergoing HCA.⁷¹ Based on the impact of pH management on CBF, a strong argument could be made to use pH-stat during the cooling phase before circulatory arrest followed by alpha-stat during rewarming, as practiced in some institutions.⁷²

During HCA, meticulous clinical management with systemic hypothermia combined with topical cooling of the head, avoidance of cerebral hyperthermia during rewarming and in the immediate postoperative interval, maintenance of normal perioperative blood glucose concentrations, careful deairing of graft and arteries, ideally with carbon dioxide flushing before reperfusion, all coupled with an expeditious surgical repair designed to minimize the duration of HCA, should be the goal.

Aortic Atherosclerosis

Atheroembolism from an atheromatous ascending aorta and aortic arch is recognized as a major risk factor in the patient undergoing cardiac surgery and is a widespread problem.^73–79 In a study of 298 asymptomatic members from the Framingham cohort aged 60 ± 9 years and of whom 51% were women, subjects underwent thoracoabdominal aortic cardiovascular MRI and demonstrated aortic plaque of 1-mm radial thickness in 38% of the women and 41% of the men.⁷³ The Stroke Prevention: Assessment of Risk in the Community (SPARC) study used transesophageal echocardiography (TEE) in 581 people older than 44.⁷⁶ Atheroma was identified in 51.3% of patients, of whom 7.6% had severe atheroma (> 4 mm thick, ulcerated or mobile). The prevalence rate of aortic arch atheroma increased with age, such that severe atheroma was seen in more than 20% of patients older than 74.⁷⁹

Atheroembolism in cardiac surgery has a broad spectrum of clinical presentations, including devastating injuries and death, yet its true incidence is probably underestimated.^80–83 Thoracic aorta atheromatosis is associated with coronary artery disease and stroke in the general population.^{74,75,77–79} In Macleod et al’s⁷⁴ review, the evidence shows that the risk for stroke is four times greater in patients with severe arch atheroma. Yahia et al⁷⁷ prospectively studied patients with diagnoses of TIA or stroke using TEE for assessment of aortic atheromatosis. Thoracic aortic atheromas were present in 141 of 237 patients (59%); mild plaque (< 2 mm) was present in 5%, moderate plaque (2 to 4 mm) in 21%, severe plaque (≥ 4 mm) in 33%, and complex plaque in 27%. Plaques were more frequently present in the descending aorta and the arch of the aorta than in the ascending aorta.⁷⁷ Watanabe et al⁷⁸ investigated whether thoracic aorta calcification on computed tomography and coronary risk factors had any correlation with obstructive coronary artery disease on angiography. Two hundred twenty-five consecutive patients underwent both thoracic conventional helical computed tomography and coronary angiography. Thoracic aorta calcification was detected in 185 patients; 141 of 225 patients had significant obstructive coronary artery disease. All of the 13 patients without thoracic aorta calcification and no coronary risk factors had no coronary artery disease.⁷⁸ Overall, it can be seen that atherosclerosis of the ascending aorta is present in 20% to 40% of cardiac surgical patients, the percentage increasing with age (Figure 36-3), and it is an independent risk factor for type I cerebral injury.^{6–8,84–90}

Figure 36-3 Prevalence of severe atherosclerosis of the ascending aorta at autopsy after operations for coronary and valvular heart disease, 1982 through 1989.

(From Blauth CI, Cosgrove DM, Webb BW, et al: Atheroembolism from the ascending aorta. An emerging problem in cardiac surgery. J Thorac Cardiovasc Surg 103:1104, 1992.)

Transesophageal Echocardiography versus Epiaortic Scanning

The detection of ascending aorta atheromatosis is a cornerstone of strategies to decrease cerebral injury in cardiac surgery. Despite its widespread utilization, manual palpation of the aorta has a very low sensitivity for this purpose.^91,92 The association of severe thoracic aortic plaques (defined as 5-mm-thick focal hyperechogenic zones of the aortic intima and/or lumen irregularities with mobile structures or ulcerations) and coronary artery disease is well established.⁸⁹ Identifying severe aortic disease has important clinical implications because surgical technique, including surgical procedure and siting of cannulation and anastomotic sites for proximal grafts, may be altered to avoid producing emboli and stroke. Intraoperative epiaortic ultrasound scanning (EAS) has emerged as a most helpful tool for the diagnosis of ascending aortic atherosclerosis and has revealed major insights into the nature and distribution of this disease.

Djaiani et al⁹⁰ performed TEE and EAS to assess the severity of aortic atherosclerosis in the ascending aorta and the aortic arch. Patients were allocated to either low- or high-risk groups according to aorta intimal thickness. Transcranial Doppler (TCD) was used to monitor the middle cerebral artery. Diffusion-weighted MRI was performed 3 to 7 days after surgery. The NEECHAM Confusion Scale was used for assessment and monitoring patient consciousness level. In the high-risk group (intimal thickness > 2 mm), confusion was present in six (16%) patients versus five (7%) patients in the low-risk group, and there was a threefold increase in median embolic count, 223.5 versus 70.0. Diffusion-weighted MRI-detected brain lesions were present only in patients from the high-risk group, 61.5% versus 0%. There was significant correlation between the NEECHAM scores and embolic count in the high-risk group.⁸⁷ Multiple studies have documented that most of the significant atherosclerotic lesions in the ascending aorta are missed by intraoperative palpation by the surgeon, and intraoperative echocardiographic studies of the aorta have been recommended^{22,86,88,90–96} (Figure 36-4). However, the ability of TEE to reliably detect all ascending aorta and aortic arch lesions is limited.

Figure 36-4 Transverse ultrasonic image of the ascending aorta and the corresponding segment of aorta in a patient with severe atherosclerosis. Note the calcification (arrowhead) and the projection of atheroma (arrow) into the lumen.

(From Wareing TH, Davila-Roman VG, Barzilai B, et al: Management of the severely atherosclerotic ascending aorta during cardiac operations. A strategy for detection and treatment. J Thorac Cardiovasc Surg 103:453, 1992.)

The high acoustic reflectance attributable to the air-tissue interface resulting from overlying right main bronchus and trachea limits TEE assessment of the upper ascending aorta where cannulation is generally undertaken.^91,92,96,97 Intraoperative EAS has emerged as a most helpful tool for the diagnosis of ascending aortic atherosclerosis and has revealed major insights into the nature and distribution of this disease. Konstadt et al^93,97 investigated 81 patients (57 male and 24 female; aged 32 to 88 years, mean age, 64 years) scheduled for elective cardiac surgery. A comprehensive examination of the entire thoracic aorta in both the longitudinal and transverse planes was performed by biplane TEE. In both echocardiographic examinations, the presence and location of protruding plaques and intimal thickening greater than 3 mm were recorded. Fourteen (17%) of the 81 patients had significant atherosclerotic disease of the ascending aorta as diagnosed by EAS echocardiography. The sensitivity of TEE was 100%, the specificity was 60%, the positive predictive value was 34%, and the negative predictive value was 100%. According to the authors, if the complete biplane TEE examination is negative for plaque, it is highly unlikely that there is significant plaque in the ascending aorta. If the TEE examination is positive for plaque, there is a 34% chance that there is significant disease of the ascending aorta, and EAS should be considered. TEE is a sensitive but only mildly specific method of determining whether ascending aortic atherosclerosis is present.^93–97

The standard for aortic assessment before instrumentation continues to be visual inspection and palpation by the surgeon, despite the fact that this has been shown to identify atheromatous disease in only 25% to 50% of patients, and even then to significantly underestimate its severity.^{91,92,98–100} Identification of ascending aorta atheromatous disease would prompt the surgical team for strategies to either modify, decrease, or avoid aortic manipulation. Management strategies for the diseased ascending aorta range from minimally invasive aortic “no-touch” techniques (NTTs) to maximally invasive procedures, including ascending aorta replacement or extensive aortic debridement under deep HCA.¹⁰¹ A recent review has outlined specific techniques for EAS, as well as steps to be used by the surgical team to mitigate aortic atheroemboli.¹⁰² Operative modifications in CABG include avoidance of aortic cross-clamping, alternative sites of aortic cross-clamping, and avoidance of proximal anastomoses by usage of all arterial conduit or Y-grafts. A decreased incidence of stroke and CNS dysfunction has been associated with this approach (Figure 36-5).¹⁰³

Figure 36-5 Influence on neurologic outcome of altering operative technique on the basis of the finding of mobile aortic atheromas on transesophageal echocardiography.

(Reprinted from Katz ES, Tunick PA, Rusinek H, et al: Protruding aortic atheromas predict stroke in elderly patients undergoing cardiopulmonary bypass: Experience with intraoperative transesophageal echocardiography. J Am Coll Cardiol 20:70–77, 1992.)

“No-Touch” Technique

Avoidance of instrumentation of the ascending aorta in patients with severe aortic atheromatosis has been advocated. Leacche et al⁸⁵ retrospectively reviewed data from 640 off-pump CABG (OPCAB) patients and identified 84 patients in whom they adopted an NTT. In these patients, revascularization was performed with single or bilateral internal thoracic arteries and by connecting additional coronary grafts (saphenous vein, radial artery) in a T or Y configuration. The right gastroepiploic artery was used as a conduit in two patients, whereas the brachiocephalic artery was used as an alternative inflow site for arterial cannulation in three reoperations. Age, sex, risk factors, functional class, and history of congestive heart failure were comparable in the two groups. In the NTT group, the frequencies were greater for severe atherosclerosis of the aorta (13% vs. 0%), carotid disease (25% vs. 16%), and history of previous cerebrovascular accidents (17% vs. 8%). In the NTT group, weak trends toward a lower incidence of postoperative delirium (8% vs. 15%; P = 0.12), a lower incidence rate of stroke (0% vs. 1%; P = 0.85), and a shorter ICU stay (P = 0.07) were observed.⁸⁵ In a review of 1993 beating-heart surgery patients, Calafiore et al¹⁰⁴ observed that in patients with evidence of peripheral vascular disease, use of aortic partial occlusion clamp was associated with a similar stroke rate as in patients in whom conventional CPB was used. They concluded that in patients with extracoronary vasculopathy, aortic manipulation must be avoided to reduce the incidence of stroke. Gaspar et al⁹⁶ used EAS and TEE in 22 patients considered to be at high risk for stroke in whom severe aortic atheroma (maximum aortic wall thickness > 5 mm or mobile plaque) was detected, and with the use of aortic NTT and beating-heart surgery, no strokes occurred.

Royse et al⁹¹ performed screening of the aorta for atheroma before aortic manipulation and used an exclusive Y-graft revascularization technique, which has no aortic coronary anastomoses. Aortic atheroma was detected using EAS and TEE. In the control group, aortic atheroma was assessed by manual palpation, whereas TCD of the right middle cerebral artery was used to detect cerebral microemboli. Neuropsychological dysfunction was assessed using a battery of 10 psychometric tests, and they demonstrated that at 60 days after surgery, dysfunction in the control group was 38.1%, whereas in the TEE/Y-graft group, it was reduced to 3.8%. Microemboli detected by TCD during periods of aortic manipulation were greater for those with late dysfunction (5.2 ± 3.0 compared with 0.5 ± 0.2), consistent with an embolic cause for cognitive dysfunction.⁹¹

Neuropsychological Testing

As noted in the studies discussed, neuropsychological testing has been increasingly used in an attempt to discriminate the efficacy of various treatment modalities or as an index of cognitive functioning after CPB. In large measure, the sensitivity of this type of testing is such that very small decrements in performance can be assessed and quantified. It appears that a patient may have a consistent decrease in cognitive performance, with or without evidence of subtle neurologic abnormalities, yet may be apparently oblivious to it, whether because of denial or an absence of awareness. Commonly, such a patient’s family may have noted some nonspecific alteration in mood or behavior, likely a manifestation of the same dysfunction as detected on cognitive testing.

Preoperative Cognitive Function

One of the earliest prospective reports of neurobehavioral sequelae of cardiac surgery appeared in 1954, and it focused on the acute and chronic stress responses manifested as psychobehavioral syndromes in patients undergoing valvular surgery.¹¹¹ As Tufo and others^112,113 noted, this led to a tendency for some investigators to focus on postoperative emotional reactions, regarding them as primarily psychiatric syndromes, rather than systematically investigating cardiac patients for neurologic deficits. Millar et al¹¹⁴ examined the effect of preexisting cognitive impairment on cognitive outcome in 81 patients undergoing CABG. Patients performed the Stroop Neuropsychological Screening Test and other psychometric assessments before and at 6 days and 6 months after CABG. Those with preexisting cognitive deficits were significantly more likely to display impairment at 6-day and 6-month follow-ups than were those without preexisting deficits, possibly reflecting underlying intrinsic pathology rather than specific intraoperative events. Rankin et al¹¹⁵ used a 1-hour neuropsychological battery administered before surgery to 43 patients before prospective randomization to either CABG or OPCAB and again to 34 of those patients 2 to 3 months after surgery by an examiner blind to surgical condition. Neuropsychological status did not change 2.5 months after surgery between OPCAB or CABG groups. However, both groups showed dramatic presurgical cognitive deficits in multiple domains, particularly verbal memory and psychomotor speed. This corroborates previous research suggesting that patients requiring CABG may evidence significant presurgical cognitive deficits as a result of existing vascular disease. In Di Carlo et al’s¹¹⁰ study of 110 patients undergoing cardiac surgery, the previous level of education was protective against cognitive decline (odds ratio per year of increment, 0.53; 95% confidence interval, 0.31 to 0.90). It is speculative whether this suggests that greater education reflects greater intellectual capacity, and thus a better ability to compensate for perioperative stresses.

Neuropsychological Test Selection

Research examining cognitive functioning in patients undergoing CPB has frequently focused on the assessment of cognitive functioning within the domains of attention/concentration, psychomotor speed, motor dexterity, and verbal learning. Under the “best-case scenario,” it might be desirable to use a complete battery of neuropsychological tests assessing the entire spectrum of cognitive functions. However, the cost of such a procedure and the time demands make such an approach unrealistic. This is especially the case if patients are to be assessed in the perioperative period, when practical time limitations or patient fatigue levels are likely to be prohibitive. A more practical approach is to use tests that can be administered relatively quickly and easily, and that assess functions particularly dependent, for successful execution, on brain areas most vulnerable to insult (i.e., “watershed” areas).

An additional criterion in test selection is that tests chosen be sensitive to brain dysfunction, broadly defined. In perioperative cardiac surgical patients, the evaluation is necessarily limited by constraints of time and fatigue; thus, tests used should have good sensitivity to dysfunction, even if at the expense of specificity. Tests that can be administered quickly and reliably and are highly sensitive, particularly to dysfunction within cognitive domains localized to brain regions vulnerable to effects of microemboli or transient hypoxia, should be selected.

Research suggests that among the tests most appropriate under these circumstances are tests of attention/concentration, psychomotor speed, motor dexterity, and verbal learning. Research as to the behavioral consequences of hypoxia (and other conditions associated with more diffuse brain damage) suggests these domains are likely to be compromised.^116,117 This presumption has also been supported by research to date examining behavioral consequences of CPB. Frequently, the Grooved Pegboard Test (motor dexterity), various subtests of the Wechsler Adult Intelligence Scale-Revised (Digit Symbol [psychomotor speed]), some of the seven subtests of the Wechsler Memory Scale (Mental Control [Attention], Digit Span [concentration], Paired Associates Verbal Learning [verbal learning]), as well as the Halstead Reitan Trail Making Test (Trails A and B), have been used in whole or in part for the assessment of cognitive impairment after CPB.^{28,33,40,52,62,91,105,109,110,114,118–123}

Methodologic Issues in Neurobehavioral Assessment

The Statement of Consensus on Assessment of Neurobehavioral Outcomes after Cardiac Surgery encouraged a more standardized and comparable methodology in assessment of cognitive injury, identifying several key issues of concern in perioperative cognitive testing.³²

Based on the proceedings of these consensus conferences, the following cognitive tests were recommended as necessary but not sufficient components of any neuropsychological test battery based on their availability in multiple languages, and availability of paper and pencil versions for use in cardiac surgical patients:

Neuropathologic Studies

In an early series from 1962 to 1970 that examined 206 patients dying after cardiac surgery or CABG and a group of 110 patients dying after non-CPB vascular surgery, Aguilar et al¹⁴⁶ reported that there was a high correlation between the use of CPB and the incidence of brain lesions. They reported that the most significant abnormalities found, in both severity and frequency of occurrence, were emboli in small cerebral vessels; acute petechial, perivascular, and focal subarachnoid hemorrhages; and acute ischemic neuronal damage (Box 36-3). They noted the virtual disappearance from the brain of nonfat emboli such as fibrin, platelet aggregates, polarizable crystalline material, xanthomatous debris, striated muscle, and calcium in cases examined after the introduction of arterial line filtration, whereas they reported that cerebral embolization of such debris was commonly observed in patients dying after CPB before the introduction of arterial line filtration.¹⁴⁶ Other early studies showed that measures taken to decrease the duration of CPB, as well as the introduction of arterial line filtration and filtration of the cardiotomy suction return, decreased overt neurologic dysfunction.^124,125,147

A review of autopsy findings from 221 patients dying after CABG or valve surgery between 1982 and 1989 reported a direct correlation among age, severe atherosclerosis of the ascending aorta, and presence of atheroemboli. Atheroemboli were significantly more common in patients who underwent CABG versus valvular surgery, and there was a high correlation of atheroemboli with severe atherosclerosis of the ascending aorta, being present in 37.4% of patients with severe disease of the aorta versus only 2% of those without. Of all patients who had evidence of atheroemboli, 95.8% had severe atherosclerosis of the ascending aorta.⁸⁰ Doty et al⁸¹ reviewed the records of 49,377 autopsy cases and surgical specimens from the Johns Hopkins Hospital between 1973 and 1995. Three hundred twenty-seven patients (0.7%) had an identifiable atheroembolism on histologic examination. Of these patients, 29 (0.2%) had undergone a cardiac surgical procedure within 30 days of autopsy or surgical resection. Six of the 29 patients (21%) had atheroembolism to the heart, 7 patients (24%) had embolism to the CNS, 19 patients (66%) had embolism to the gastrointestinal tract, 14 patients (48%) had embolism to one or both kidneys, and 5 patients (17%) had embolism to a lower extremity. Sixteen patients (55%) had atheroembolism in two or more areas. In six patients (21%), death was directly attributable to atheroembolism, including intraoperative cardiac failure from coronary embolism (three), massive stroke (two), and extensive gastrointestinal embolization (one).⁸¹

In a neuropathologic study of brains from 262 patients dying after having undergone CABG, valve replacement, or heart transplantation surgery, 49% of cases demonstrated evidence of circulatory disturbances identified as macrohemorrhages and microhemorrhages, infarcts, subarachnoid hemorrhages, or hypoxemic brain damage.¹⁴⁸ The infarcts were caused by local arteriosclerosis of cerebral arteries, fat emboli, arterial emboli from operative sites, or foreign body emboli. These authors concluded that histologically overt microemboli did not play a major role in their findings, and that nonfatal white matter microhemorrhages were found with varying frequency, especially after valve operations. These observations are not inconsistent with the apparent lack of correlation seen in several beating-heart surgery studies between differing incidences of TCD-detected cerebral emboli and cognitive dysfunction.^44,45

Watershed Infarctions

Watershed, or boundary zone, infarcts are ischemic lesions that are situated along borderzones between the territories of two major cerebral arteries (e.g., the middle and posterior, or the anterior and middle cerebral arteries) where terminal arteriolar anastomoses exist^149–152 (Figure 36-6). In a series reported by Malone et al,¹⁵² a correlation was made between the presence of intraoperative electroencephalographic (EEG) abnormalities (virtual or complete electrical silence) usually seen in conjunction with sustained hypotensive episodes and neuropathologic lesions found at necropsy. In all nine patients with clinical evidence of brain damage, cortical boundary zone (watershed) lesions were observed in the parieto-occipital areas. They suggest that this location is the most sensitive area for placement of recording EEG electrodes because it is where minimal boundary zone ischemic lesions occurred in the absence of other lesions, and it is also where ischemic lesions were found in their maximal severity and extent.

Figure 36-6 Hatched areas showing the most frequent locations of boundary area, or watershed zone infarcts in the brain, situated between the territories of major cerebral or cerebellar arteries.

(From Torvik A: The pathogenesis of watershed infarcts in the brain. Stroke 2:221, 1984.)

A profound reduction in systemic blood pressure is the most frequent cause of watershed infarcts. These areas are thought to be more susceptible to ischemia resulting from hypotension because of their critical dependence on a single blood supply. Wityk et al¹⁵³ studied the pattern of ischemic changes on diffusion- and perfusion-weighted MRI in 14 patients with neurologic complications after cardiac surgery, of whom 8 patients presented with encephalopathy, which was associated with focal neurologic deficits in 4, 4 with focal deficits alone, and 2 with either fluctuating symptoms or TIAs. Acute ischemic lesions were classified as having a territorial, watershed, or lacunar pattern of infarction. Patients with multiple territorial infarcts in differing vascular distributions that were not explained by occlusive vascular lesions were classified as having multiple emboli. Acute infarcts were found in 10 of 14 patients by diffusion-weighted imaging. Among patients with encephalopathy, seven of eight had patterns of infarction suggestive of multiple emboli, including three of four patients with no focal neurologic deficits. Several patients had combined watershed and multiple embolic patterns of ischemia. Diffusion-weighted MRI findings were abnormal in two of four patients, showing diffusion–perfusion mismatch; both patients had either fluctuating deficits or TIAs, and their conditions improved with blood pressure increase.¹⁵³

By the same rationale, however, these areas are also highly susceptible to ischemia because of end-artery embolization, and it is also recognized that although severe hypotension is the most common cause, showers of microemboli may lodge preferentially in these areas and cause infarcts in the underlying brain.^153–156 As such, although they are commonly due to profoundly hypotensive episodes, watershed lesions are not pathognomonic of a hypotensive episode and may be the result of cerebral emboli. Embolization and hypoperfusion acting together play a synergistic role and either cause or magnify the brain damage of cardiac surgical patients. The negative influence of hemodynamic instability and hypoxia has been demonstrated by several authors, showing improved outcomes by an early and aggressive recognition and correction of hypoperfusion.^{5,27,40,157–159}

Cerebral Perfusion Pressure

Intraoperative hypotension during cardiac surgery has been related to postoperative neurologic dysfunction.^{4,5,21,27,42,160} Ridderstolpe et al²⁷ published a retrospective study of 3282 patients of mean age of 65.6 years who underwent cardiac surgery in the period from July 1996 through June 2000. Cerebral complications occurred in 107 patients (3.3%). Of these, 60 (1.8%) were early, 33 (1.0%) were delayed, and in 14 (0.4%) patients the onset was unknown. Predictors of early cerebral complications were older age, preoperative hypertension, aortic aneurysm surgery, prolonged CPB time, hypotension at CPB completion and soon after CPB, and postoperative arrhythmia and supraventricular tachyarrhythmia. Predictors of delayed cerebral complications were female sex, diabetes, previous cerebrovascular disease, combined valve and CABG, postoperative supraventricular tachyarrhythmia, and prolonged ventilator support. Early cerebral complications seemed to be more serious, with more permanent deficits and a greater overall mortality (35.0% vs. 18.2%). The results of this study suggest that aggressive antiarrhythmic treatment and blood pressure control may improve the cerebral outcome after cardiac surgery.²⁷

EEG patterns consistent with ischemia—increased slow wave activity, diffuse slowing of EEG activity—have been reported to occur during CPB episodes thought to be associated with cerebral hypoperfusion.^161–163 Episodes of flow reduction during normothermia frequently produced ischemic changes, whereas similar decreases during stable hypothermia were not associated with EEG changes.¹⁶² Ischemic EEG changes are also frequently seen in association with reductions in perfusion flow rate during the initiation of CPB^161,162 (see Chapters 16, 28, and 29). Using computerized EEG to quantitate episodes of low-frequency power as an index of cerebrocortical ischemia in 96 patients undergoing CABG, a correlation was made among episodes of hypotension, focal increases in low-frequency EEG power, and the occurrence of postoperative disorientation.¹⁶⁴

During the transition to CPB, the brain is particularly vulnerable to ischemia, inasmuch as cerebral metabolic rate for oxygen (CMRO₂) is apparently unchanged, yet the brain is initially perfused with an asanguineous prime, and even after equilibration during established CPB, hematocrit is generally maintained at a range between 20% and 30%. As a result, any further decreases in cerebral perfusion, in the absence of concomitant decreases in CMRO₂, are poorly tolerated. During hypothermic conditions, there is a profound decrease in CMRO₂, exceeding 50% for a 10° C reduction in temperature.⁶⁶ Without need to postulate an extension of the lower limit of cerebral autoregulation, it is clear that under anesthesia, and particularly during hypothermic CPB, CBF is maintained at very low levels of cerebral perfusion pressure. As discussed later, this was initially reported by Govier et al¹⁶⁵ and further explored by Murkin et al⁶⁶ and Prough et al.¹⁶⁶ As the average age and extent of disease in patients presenting for CABG continue to increase, however, the number of patients with concomitant cerebrovascular disease, and thus potentially deranged cerebral autoregulation, presents an increasingly important group.

Using radioisotope techniques for measurement of CBF, and incorporating a jugular venous catheter for calculation of CMRO₂, it was determined that there is a profound reduction in CMRO₂ during hypothermic CPB, and that CBF is decreased proportionately and will autoregulate down to a cerebral perfusion pressure of 20 mm Hg, in the presence of alpha-stat pH management.⁶⁶ Low arterial pressure during the hypothermic phase of CPB is thus unlikely to result in cerebral ischemia in the absence of cerebrovascular disease. In a study of high versus low arterial pressure management during CPB in 248 patients undergoing CABG, however, an apparently lower rate of postoperative complications was reported by Gold et al¹⁶⁷ for those patients in the high-pressure group. Although specific CNS morbidity, cognitive and functional status outcomes, and mortality did not differ significantly between groups, the overall complication rate for combined cardiac and neurologic complications was significantly lower in the high-pressure group. This is not inconsistent with data demonstrating a high incidence of cerebrovascular disease in coronary revascularization patients.^108,109

Cerebral Venous Obstruction

It should also be appreciated that during CPB, cerebral venous hypertension can result from partial obstruction of the superior vena cava (Figure 36-7), particularly in the presence of a single two-stage venous cannula, and may cause cerebral edema, as well as produce a disproportionate decline in cerebral perfusion pressure relative to arterial pressure.¹⁶⁸ In a study by Avraamides,¹⁶⁹ surgical dislocation of the heart during CPB produced increases in proximal superior vena cava pressure and resulted in significant decreases in CBF velocity as measured with TCD, despite stable arterial pressure and pump flow rates. This strongly suggests that cerebral venous hypertension, as can occur during CPB with myocardial dislocation and impaired drainage of superior vena cava, may result in cerebral ischemia if unrecognized and untreated. It is feasible that such unrecognized cerebral venous hypertension has resulted in some of the postoperative neurologic syndromes that have been reported.¹⁷⁰

Figure 36-7 A, Systolic, mean, and diastolic arterial blood pressures, with commencement of cardiopulmonary bypass (CPB) indicated at 3:15 pm, after which mean arterial pressure (MAP) is shown. B, Pulmonary artery systolic, mean, and diastolic pressures with proximal jugular venous pressure (JVP) recorded at 3:15 pm, with commencement of CPB. A single two-stage venous cannula was used for CPB and, with rotation of the heart, venous return to the oxygenator decreased and JVP approached MAP values. SVC, superior vena cava.

(Modified from Murkin JM: Intraoperative management. In Estaphanous FG, Barash PG, Reves JG [eds]: Cardiac Anesthesia Principles and Clinical Practice. Philadelphia, Lippincott, 1994, p 326.)

Although the association between hypotension during CPB and cerebral dysfunction remains contentious, there is some evidence that certain subsets of patients may be at particular risk. Newman et al¹⁷¹ used preoperative and postoperative cognitive testing to assess the effects of MAP and rate of rewarming on cognitive decline in 237 patients. They demonstrated significant interactions between cognitive decline and MAP less than 50 mm Hg on one measure of cognitive performance and between rate of rewarming and age on another. They concluded that although MAP and rewarming were not primary determinants of cognitive decline, hypotension and rapid rewarming contributed significantly to cognitive dysfunction in the elderly. Again, because elderly patients comprise an increasing segment of the cardiac surgical population, these aspects are becoming increasingly important clinical management issues.

Hemodynamic Instability

The interaction of emboli, perfusion pressure, and the particular conditions of the regional cerebral circulation (e.g., preexisting cerebral intravascular lesions) will determine the final expression of brain damage in the cardiac surgical patient. In a retrospective study, Ganushchak et al¹⁷² tested the hypothesis that combinations of hemodynamic events from apparently normal CPB procedures are related to the development of postoperative neurologic complications and affect the impact of common clinical risk factors. A multivariate statistical procedure (cluster analysis) was applied to a data set of automatically recorded perfusions from 1395 patients who underwent CABG. The following five parameters emerged for cluster analysis: MAP, dispersion of MAP, dispersion of systemic vascular resistance, dispersion of arterial pulse pressure, and the maximum value of mixed venous saturation. Using these parameters, they found four clusters that were significantly different by CPB performance (first cluster, 389 patients; second cluster, 431 patients; third cluster and fourth cluster, 229 patients each). Patients in the fourth cluster had the greatest dispersion of MAP, that is, greatest instability on CPB. The frequency of postoperative neurologic complications was 0.3% in the first cluster and increased to 3.9% in the fourth cluster. Importantly, the impact of common clinical risk factors for postoperative neurologic complications was affected by the performance of the CPB procedure. For example, the frequency of neurologic complications among patients with cerebrovascular disease in their medical history was 22% in the fourth cluster, whereas it was zero in the second cluster. Patients who underwent CPB procedures with large fluctuations in hemodynamic parameters showed a particularly increased risk for the development of postoperative neurologic complications.¹⁷²

Cerebral Emboli and Outcome

Two different types of cerebral emboli appear to occur during CPB composed of solid or gaseous matter, such as macroemboli (e.g., atherosclerotic debris) and microemboli (e.g., microgaseous bubbles, microparticulate matter). Overt and focal neurologic damage likely reflects the occurrence of cerebral macroemboli (e.g., calcific and atheromatous debris generated during valve tissue removal or instrumentation of an atheromatous aorta), whereas less focal neurologic dysfunction has been ascribed to cerebral microemboli.¹⁴ Microemboli appear to have some role in diffuse, subtle neurologic and cognitive disturbances, whereas macroemboli likely produce clinically apparent catastrophic strokes. Whatever the nature of the cerebral insult, however, it seems that coexistent inflammatory processes can exacerbate the magnitude of injury.

In a study to assess the impact of surgical manipulation of the aorta and correlations with postoperative stroke, Ura et al¹⁷³ performed EAS before aortic cannulation and after decannulation in 472 patients undergoing cardiac surgery with CPB. A new lesion in the ascending aortal intima was identified in 16 patients (3.4%) after aortic decannulation, of whom 10 patients sustained neurologic complications. New lesions were severe, with mobile lesions or disruption of the intima in 10 patients. Six of the severe lesions were related to aortic clamping and the other four to aortic cannulation. Three patients in this group had postoperative stroke. The incidence rate of new lesions was directly related to extent of aortic atheroma, being 11.8% if the atheroma was approximately 3 to 4 mm thick and as high as 33.3% if the atheroma was greater than 4 mm, but only 0.8% when it was less than 3 mm.¹⁷³ Again, this underscores the need to reliably detect and ultimately avoid disruption of aortic atherosclerotic plaque.

Sylivris et al¹⁷⁴ studied 41 consecutive patients undergoing CABG with TCD monitoring and preoperative and postoperative MRI brain scans. A subgroup of 32 patients underwent neuropsychological testing the day before and 5 to 6 days after the operation, of whom 27 had TCD data. Among the subgroup of patients with reliable TCD data and neuropsychological studies, early neuropsychological deficit after CABG was found in 17 (63%) of the 27 patients. On univariate analysis, the time duration on CPB, total microembolic load during bypass, and microembolic rates during bypass were all significantly greater in the group with neuropsychological decline. Actual rates of emboli detected per minute were greatest during release of the aortic cross-clamp. Five patients had strokes, of which four had a significant decline in neuropsychological functioning. Unlike the association between microembolic signals (MESs) during bypass and neuropsychological deficits, there was no relation between these factors and radiologic evidence of cerebral infarction. Not inconsistent with the findings of Ura et al¹⁷³ described earlier, there was a significantly greater microembolic load at cannulation in patients with cerebral infarction, temporally suggestive of particulate emboli, which was not apparent in comparison with patients with neuropsychological deficits alone.¹⁷⁴

A study with a newer generation of TCD, which uses two different frequencies for insonation and purportedly discriminates between gaseous and particulate emboli, compared the number and nature of intraoperative microemboli in patients undergoing on-pump and off-pump cardiac surgery procedures in 45 patients (15 OPCAB, 15 on-pump CABG, and 15 open cardiac procedures).¹²⁸ They demonstrated significantly fewer emboli in the OPCAB versus on-pump CABG and open procedure groups, averaging 40 (28 to 80), 275 (199 to 472), and 860 (393 to 1321) emboli, respectively (P < 0.01). Twelve percent of microemboli in the OPCAB group were defined as solid compared with 28% and 22% in the on-pump CABG and open procedure groups, respectively. In the on-pump groups, 24% of microemboli occurred during CPB, and 56% occurred during aortic manipulation for cannulation, decannulation, application, and removal of cross clamp or side clamp, again underscoring the importance of minimizing aortic instrumentation.⁴⁴

Gaseous emboli are not innocuous, however. It has been demonstrated that the effects of air emboli on the cerebral vasculature not only are due to bubble entrapment with direct blockage of cerebral vessels but represent the effects that such bubbles have on vascular endothelial cells.¹⁷⁵ Ultrastructural examinations of pial vessels in rats exposed to cerebral air emboli demonstrated severe injury to endothelial plasmalemma, leading to loss of cellular integrity and endothelial cell swelling.¹⁷⁶ Such endothelial damage produces disruptions of vasoreactivity, as has been observed in cat pial vessels exposed to air emboli. In these capillary beds, the endothelial layer demonstrated ultrastructural abnormalities that included degradation of intercellular junctions, flattening of nuclei, and crenation of the plasmalemma. Air embolism also produces changes in blood elements leading to formation of a proteinaceous capsule around the bubbles, marked dilation of pial vessels, platelet sequestration, and damage to endothelial cells.^177–179 Air-induced mechanical trauma to the endothelium causes basement membrane disruption, thrombin production, release of P-selectin from intracellular vesicles, synthesis of platelet-activating factor, and a reperfusion-like injury with perturbations in inflammation and thrombotic processes. These phenomena likely impair nitric oxide production, causing alterations in cerebral microvascular regulation.^180–182

In Moody et al’s¹²⁶ study, four of five patients dying after CPB, two patients dying after proximal aortography, and six dogs placed on CPB were all found to have small cerebral capillary and arteriolar dilations (SCADs), consistent with sites of gas bubbles or fat emboli (Figure 36-8). These microvascular anomalies were only found in conjunction with utilization of CPB or proximal aortic instrumentation. In a subsequent series of elegant studies by this same group, use of colored microspheres was able to “time-lock” the development of SCADs to the period associated with CPB¹⁸³ (Figure 36-9). In further studies from this group, Challa et al¹⁸⁴ identified SCADs in thick colloidin sections of the brains of eight patients who died after cardiac surgery supported with a membrane oxygenator and in two dogs that underwent CPB with a bubble oxygenator. In SCADs of the eight patients who had cardiac surgery, both aluminum and silicone values were higher, and silicone values were also high in the two dogs in which a bubble oxygenator was used. Their results indicated that contamination with aluminum and silicone occurred during cardiac surgery assisted by CPB, and that switching to membrane oxygenators from bubble oxygenators for CPB may have reduced silicone contamination of blood.¹⁸⁴ Kincaid et al¹⁸⁵ used a cell saver to process the cardiotomy blood in dogs that underwent hypothermic CPB. The brain tissue from two groups of dogs (group I, cardiotomy suction blood reinfused through arterial line filter; group II, cardiotomy suction blood collected and processed in a cell saver) was examined for the presence of SCADs. Mean SCAD density in the cell-saver group was less than the arterial filter group (11 ± 3 vs. 24 ± 5; P = 0.02). The authors concluded that using a cell saver to scavenge shed blood during CPB decreases cerebral lipid microembolization.¹⁸⁵

Figure 36-8 Microemboli or small capillary and arteriolar dilatations (SCADs) in arterioles of human brain 1 day after cardiopulmonary bypass. The afferent microvessels are black. The SCADs are dilated clear areas (arrows); the largest one here measures 25 μm in diameter. It is believed these represent the “footprints” of emboli that were almost completely removed by the reagents used in this histochemical staining method (alkaline phosphatase–stained 100-μm-thick celloidin section; magnification ×300 before 50% reduction).

(Reprinted from Moody DM, Brown WR, Challa VR, et al: Brain microemboli associated with cardiopulmonary bypass: A histologic and magnetic resonance imaging study. Ann Thorac Surg 59:1304, 1995.)

Figure 36-9 Microemboli (white arrows) “bracketed in time” by sequentially injected microspheres of different colors. Clear microspheres (small black arrows) and black microspheres (large black arrows) can be seen distal to proximal order in a single arteriolar complex. In this experiment, clear spheres were injected into the carotid artery of a dog, followed in succession by injection of corn oil and then black spheres. Direction of blood flow in the arteriole is from top to bottom (alkaline phosphatase–stained 100-μm-thick celloidin section; microspheres = 15 μm).

(Reprinted from Moody DM, Brown WR, Challa VR, et al: Brain microemboli associated with cardiopulmonary bypass: A histologic and magnetic resonance imaging study. Ann Thorac Surg 59:1304, 1995.)

More recently, two separate randomized, prospective studies in cardiac surgical patients have assessed the impact of cell-saver usage on cognitive dysfunction after cardiac surgery.^186,187 A series of 226 patients older than 60 years undergoing CABG surgery were randomly allocated to either cell-saver or control groups.¹⁸⁶ Anesthesia and surgical management were standardized. Epiaortic scanning of the proximal thoracic aorta was performed in all patients, and TCD was used to measure cerebral embolic rates. Standardized neuropsychological testing was conducted 1 week before and 6 weeks after surgery. Cognitive dysfunction was present in 6% of patients in the cell-saver group and 15% of patients in the control group 6 weeks after surgery (P = 0.038). However, significantly (P = 0.018) more patients in the cell-saver group required transfusion of fresh frozen plasma (25%) versus the control group (12%). In a remarkably similar study from Rubens et al,¹⁸⁷ patients undergoing coronary and/or aortic valve surgery using CPB were randomized to receive unprocessed blood (control, n = 134) or cardiotomy blood that had been processed by centrifugal washing and lipid filtration (treatment, n = 132). The treatment group received more intraoperative red blood cell transfusions (0.23 ± 0.69 vs. 0.08 ± 0.34 units; P = 0.004), and both red blood cell and non–red blood cell blood product use was greater in the treatment group. Postoperative bleeding was greater in the treatment group. Patients also underwent neuropsychometric testing before surgery and at 5 days and 3 months after surgery. There was no difference in the incidence of postoperative cognitive dysfunction in the two groups (relative risk: 1.16, 95% CI: 0.86 to 1.57 at 5 days after surgery; relative risk: 1.05, 95% CI: 0.58 to 1.90 at 3 months). Similarly, there was no difference in the quality of life, nor was there a difference in the number of emboli detected in the two groups. These authors concluded that processing of cardiotomy blood before reinfusion results in greater blood product use with greater postoperative bleeding in patients undergoing cardiac surgery, and that there was no clinical evidence of any neurologic benefit with this approach in terms of postoperative cognitive function. In summary, both these studies showed an increase in utilization of allogeneic blood products and perioperative blood loss as a consequence of routine cell-saver usage, with either no or minor improvements in incidence of postoperative cognitive decline. In view of the variable impact on NCD demonstrated in these studies and the detrimental impact of perioperative allogeneic transfusion,¹⁸⁸ routine usage of cell saver for processing of cardiotomy suction blood is probably unwarranted.

pH Management and Cerebral Blood Flow

Relatively little new information regarding the cerebral circulation in human beings during CPB appeared until 1983, when Henriksen et al¹⁹² reported evidence of cerebral hyperemia occurring during CPB. This report was followed in 1984 by a seminal paper from Govier et al,¹⁶⁵ who not only incited controversy with their observations of ischemic threshold levels of CBF during CPB, in direct contrast with the hyperperfusion reported by Henriksen, but also made preliminary observations on many of the other critical variables thought to influence CBF during CPB.

Without the measurement of concomitant cerebral metabolism, these apparently discordant observations of CBF could not be reconciled. Murkin et al⁶⁶ subsequently reported their observations of both CBF and CMRO₂ during hypothermic CPB in humans, using a xenon-133 clearance technique for measurement of CBF, similar to techniques used by Kubota, Govier et al, and Henriksen et al, but with the addition of a jugular bulb catheter for sampling effluent cerebral venous blood for measurement of cerebral metabolic activity. It was hypothesized that differences in pH management accounted for the divergent values previously reported for CBF during hypothermic CPB. Accordingly, patients were managed with either alpha-stat or pH-stat pH management during hypothermic CPB. A similar and pronounced reduction in CMRO₂ was observed in both groups during hypothermia (Figure 36-10), and in the alpha-stat group, global cerebral flow/metabolism coupling was preserved in comparison with the group managed with pH-stat (Figure 36-11). Decreases in CBF and CMRO₂, significantly lower than similar measures before and after CPB, were still evident after rewarming during normothermic nonpulsatile CPB. These low values for CBF and CMRO₂ were restored to control levels shortly after separation from CPB. Alpha-stat management preserved autoregulation and the relation between CBF and metabolism.⁶⁶

Figure 36-10 Cerebral blood flow (CBF) and cerebral metabolic rate for oxygen (CMRO₂) in the alpha-stat (non–temperature-corrected) and pH-stat (temperature-corrected) groups. Note the convergence of CMRO₂ and divergence of CBF between groups during the hypothermic phase of cardiopulmonary bypass (CPB).

(From Murkin JM: Cerebral hyperperfusion during cardiopulmonary bypass: The influence of PaCO₂. In Mark Hilberman [ed]: Brain Injury and Protection During Heart Surgery. Boston, Martinus Nijhoff, 1988, p 57.)

Figure 36-11 Simple linear regression of cerebral blood flow (CBF) versus cerebral perfusion pressure or cerebral oxygen consumption for temperature-corrected and non–temperature-corrected groups. There is no significant correlation between CBF and cerebral metabolic rate for oxygen (CMRO₂) in the temperature-corrected group (A1), whereas CBF significantly correlates with CMRO₂ in the non–temperature-corrected group (B1). CBF is significantly correlated with cerebral perfusion pressure (CPP) in the temperature-corrected group (A2), whereas CBF is independent of CPP in the non–temperature-corrected group (B2).

(From Murkin JM, Farrar JK, Tweed A, et al: Cerebral autoregulation and flow/metabolism coupling during cardiopulmonary bypass: The influence of PaCO₂. Anesth Analg 66:825, 1987.)

Temperature and Coronary Artery Bypass Grafting

How much does cerebral hyperthermia superimposed on a brain suffused with focal ischemic lesions contribute to postoperative CNS dysfunction? It is known that the vulnerability of the normothermic brain to focal ischemic insult demonstrates a surprising variability in the presence of small gradations in temperature. Busto et al¹⁹⁴ demonstrated that at 33° C, expression of cerebral ischemia was virtually eliminated compared with controls maintained at 36° C. In fact, a large measure of the apparent cerebroprotective efficacy of the glutamate-receptor antagonist MK-801 in global ischemia was demonstrated by Buchan and Pulsinelli¹⁹⁵ to be mediated by just such a small secondary decrease in brain temperature. Conversely, small increases in brain temperature, such as to 39° C increments as may occur during CABG (Figure 36-12), have been shown to profoundly enhance the susceptibility of the brain to focal ischemic insult and result in ischemic lesions of much greater extent in comparison with controls at 37° C.¹⁹⁶

Figure 36-12 Tympanic and bladder temperatures during rewarming.

Time 0 = maximal temperature achieved. Mean ± standard deviation values are shown.

(From Nathan JH, Lavallee G: The management of temperature during hypothermic cardiopulmonary bypass: I. Canadian survey. Can J Anaesth 42:669, 1995.)

Normothermic Cardiopulmonary Bypass

The demonstration of apparently improved myocardial performance and shorter CPB and operating room times after normothermic CPB has prompted several outcome studies to assess the efficacy of this therapy, with particular focus now centered on CNS outcomes. Hypothermia reduces cerebral metabolic rate; thus, mild hypothermia might protect the brain by preferentially suppressing energy utilization to maintain cellular integrity.¹⁹⁷ In support of this are results from a subset of 138 patients randomized to normothermic or hypothermic CPB, in whom a detailed prospective neurologic examination and a series of cognitive tests were performed before surgery, at postoperative days 1 to 3 and 7 to 10, and again at 1 month after surgery.¹⁹⁸ Seven of 68 patients in the normothermic group were found to have a central neurologic deficit compared with none of the patients in the hypothermic group, a significantly greater incidence. In a separate study of 96 patients undergoing CABG and randomized to CPB at either 28° C, 32° C, or 37° C, patients managed at 37° C had a significantly greater incidence of deterioration on cognitive test scores than did those managed at either 28° C or 32° C. No additional benefit in terms of cognitive function was conferred by cooling to 28° C versus 32° C.¹⁹⁹ Nathan et al²⁰⁰ reported on CABG patients operated under hypothermic (32° C) CPB and then randomly assigned to rewarming to 37° C (control) or 34° C (hypothermic), with no further intraoperative warming. Neurocognitive testing was performed 1 week and 3 months after surgery. Eleven tests were combined into three cognitive domains: memory, attention, and psychomotor speed and dexterity. The incidence of cognitive deficits 1 week after surgery was 62% in the control group and 48% in the hypothermic group (relative risk, 0.77; P = 0.048). In the hypothermic group, the magnitude of deterioration in attention and in speed and dexterity was reduced by 55.6% (P = 0.038) and 41.3% (P = 0.042), respectively. At 3 months, the hypothermic group still performed better on one test of speed and dexterity.²⁰¹

Other studies have demonstrated apparently different results, however. Engelman et al²⁰¹ randomized a series of 291 patients undergoing coronary revascularization to either hypothermic or tepid/normothermic perfusion. Twelve intraoperative ischemic strokes occurred; six of these were in the group receiving hypothermic perfusion, and six were in the group receiving the tepid/normothermic perfusion. Measuring the infarct volume documented that three of the strokes in each group resulted in minor or small infarcts, and that three in each group were significant, major strokes. The volume of infarction, whether including all six patients in each group or only those with major strokes, was no different between the hypothermic and the tepid/normothermic groups. The authors observed no relation between the size of a cerebral ischemic infarct and the perfusate temperature during coronary revascularization.²⁰¹ Similarly, Dworschak et al²⁰² found no difference in brain isoenzyme S-100β release pattern or in clinical neurologic complications between two groups of CABG patients randomly assigned to normothermic or hypothermic (32° C) CPB. In a large, prospective study, Grigore et al¹¹⁸ randomly assigned 300 patients undergoing elective CABG to tepid/normothermic (35.5° C to 36.5° C) or hypothermic (28° C to 30° C) CPB and used a battery of neurocognitive tests evaluating four distinct cognitive domains administered before surgery and at 6 weeks after surgery. Again, there were no differences in neurologic or neurocognitive outcomes between normothermic and hypothermic groups in multivariable models. In a separate study, Grimm et al²⁰³ actually found subclinical impairment of cognitive brain function to be more pronounced in CABG patients undergoing mildly hypothermic CPB compared with normothermic CPB. Accordingly, a protective effect of hypothermia on neurologic outcome in CABG has not been verified in most clinical trials, possibly because of too fast and/or excessive rewarming of patients or, more likely, because the impact of only transient cooling during CPB does not extend into the postoperative interval when cerebral hyperthermia has also been demonstrated and correlated with impaired cognitive performance at 6 weeks after surgery.^21,204–208

Cerebral Hyperthermia

Cerebral hyperthermia during the rewarming phase of CPB can exacerbate a preexisting injury before rewarming and may be detrimental in itself. Hyperthermia can have a strong impact on cerebral oxygen transfer and neurologic outcome. Glutamate levels can increase during cerebral hyperthermia, leading to eventual cell death. Rapid rewarming decreases jugular venous hemoglobin saturation, creating a mismatch between cerebral oxygen consumption and delivery.^209,210 Okano et al²¹¹ assessed the effects of normothermia and mild hypothermia (32° C) during CPB on jugular oxygen saturation (Sjvo₂) in 20 patients scheduled for elective CABG. The Sjvo₂ in the normothermic group was decreased significantly at 20 and 40 minutes after the onset of CPB compared with pre-CPB, whereas there was no change in Sjvo₂ in the mild hypothermic group during the study. The authors concluded that cerebral oxygenation, as assessed by Sjvo₂, was increased during mild hypothermic CPB compared with normothermic CPB. Kawahara et al²¹⁰ examined the effect of rewarming rates on Sjvo₂ in 100 patients scheduled for elective CABG and randomly divided into two groups: a control group and a slow rewarming group. Cerebral desaturation (defined as an Sjvo₂ value < 50%) during rewarming was more frequent in the control group than in the slow group. Cerebral desaturation time was defined as duration when Sjvo₂ was less than 50% and the ratio of the cerebral desaturation time to the total CPB time in the control group differed significantly from those in the slow group (control group: 17 ± 11 minutes, 12% ± 4%; slow group: 10 ± 8 minutes, 7% ± 4%, respectively; P < 0.05).²¹⁰ Consistent with this, in a study of the impact of rate of rewarming on cognitive outcomes in 165 CABG patients randomized to two differing rewarming strategies, Grigore et al¹¹⁸ demonstrated a significant association between change in cognitive function and rate of rewarming.

Grocott et al²⁰⁸ recorded hourly postoperative temperatures in 300 patients undergoing CABG on CPB and determined the degree of postoperative hyperthermia using the maximum temperature within the first 24 hours, as well as by calculating the area under the curve for temperatures greater than 37° C. Patients underwent a battery of cognitive testing both before surgery and 6 weeks after surgery. The maximum temperature within the first 24 hours after CABG ranged from 37.2° C to 39.3° C, and these investigators demonstrated that the maximum postoperative temperature was independently associated with cognitive dysfunction at 6 weeks.¹⁷⁶ Accordingly, slower rewarming rate with lower peak temperatures during CPB may be an important factor in the prevention of neurocognitive decline after hypothermic CPB, and interventions to avoid postoperative hyperthermia may be warranted to improve cerebral outcome after cardiac surgery.

Cerebrovascular Disease

Relatively few studies have examined the cerebrovascular responses to CPB in patients with known cerebrovascular disease. Because of the vasodilatory effects of increased carbon dioxide in patients with cerebrovascular disease, pH-stat management could theoretically induce redistribution of regional CBF from marginally perfused to well-perfused regions (i.e., an intracerebral steal). Gravlee et al²¹² investigated patients with cerebrovascular disease undergoing CABG and assessed the CBF responses to varying pH management, between alpha-stat and pH-stat. They confirmed the responsiveness of the cerebral vasculature to changes in Paco₂ during hypothermic CPB but did not demonstrate evidence of intracerebral steal at greater Paco₂ levels in any of these patients. In all patients, however, arterial perfusion pressure was greater than 65 mm Hg during CBF measurements, which may have offset any tendency for regional CBF inhomogeneities.

Using TCD monitoring of CBF velocity, 18 patients with severe carotid stenosis and 37 with no or mild stenosis were monitored during CPB.²¹³ Although not specified, it appears as though pH-stat management was used, because flow velocities correlated with Paco₂ and arterial pressure. There were no significant differences detectable in flow velocity between patients with or without significant carotid stenosis. In a case report of a single patient with bilateral carotid stenoses, alpha-stat pH management was used and arterial pressure was varied from 35 to 85 mm Hg, whereas CBF was measured during hypothermic CPB.²¹⁴ The CBF values obtained from contralateral hemispheres were essentially equal and remained so throughout the range of different perfusion pressures used. These studies suggest that the cerebrovascular responses to CPB in patients with cerebrovascular disease do not differ significantly from those of normal patients with respect to gross measures of cerebrovascular responsiveness. This suggests that other factors, including associated aortic atherosclerosis, may be a more important factor.

Hogue et al⁷ examined demographic and perioperative data prospectively collected from 2972 patients undergoing cardiac surgery. Carotid artery ultrasound examination was performed before surgery for patients aged 65 years or older or when there was a history of TIAs or prior stroke. Epiaortic ultrasound was performed at the time of surgery in all patients to assess for atherosclerosis of the ascending aorta. Strokes occurred after surgery in 30 women and 18 men (P < 0.0001). A history of a stroke was the strongest predictor of new stroke for both women and men. A prior cerebrovascular event was a more important predictor of stroke for men than women.⁷

Delirium

In Bucerius et al’s study⁵ assessing CNS outcomes from 16,184 patients undergoing cardiac operations, the overall prevalence rate of postoperative delirium was 8.4%. Stepwise logistic regression revealed history of cerebrovascular disease, peripheral vascular disease, atrial fibrillation, diabetes mellitus, left ventricular ejection fraction of 30% or less, preoperative cardiogenic shock, urgent operation, intraoperative hemofiltration, operation time of 3 hours or more, and a high perioperative transfusion requirement as being independent predictors of delirium. In a prospective follow-up study of 112 cardiac surgical patients, the incidence rate of postoperative delirium was 21% and was associated with significantly increased mortality and readmission to hospital, as well as significantly greater incidences of cognitive and sleep disturbances.²¹⁵

Carotid Endarterectomy

In the current cardiac surgical population, 17% to 22% of patients have a moderate carotid artery stenosis of 50% or more, and 6% to 12% have a severe stenosis of 80% or more.²¹ The risk for postoperative stroke is 10% in patients with moderate and 11% to 19% in patients with severe stenosis, whereas it remains 2% or less in patients with a stenosis of less than 50%. Although in patients presenting for cardiac surgery severe bilateral carotid artery disease is rare, the risk for perioperative stroke is as high as 20%.^21,216–218 It is not clear that carotid endarterectomy decreases this rate, however, because in a meta-analysis, pooled data for stroke or death did not support carotid endarterectomy for risk reduction from asymptomatic carotid stenosis during CABG (relative risk, 0.9; P = 0.5).²¹⁹ In a review, it was estimated that only about 40% of perioperative strokes (at most) could be directly attributable to ipsilateral carotid artery disease.²²⁰ Accordingly, in a patient with asymptomatic carotid stenosis, combined surgery should not be undertaken unless the surgical team is very experienced in combined carotid endarterectomy/CABG procedures. Concomitant carotid endarterectomy is unlikely to decrease a patient’s stroke risk. Rather, carotid stenosis should be regarded as indicating a high likelihood of aortic and/or concomitant intracerebral disease, and there is increasing evidence that use of EAS with appropriate modification of surgical approach and, potentially, applied neuromonitoring can be of particular benefit in this high-risk group.

Diabetes Mellitus and Hyperglycemia

The presence of diabetes is recognized as a factor related to increased morbidity and mortality in cardiac surgical patients.^221–225 The incidence of diabetes mellitus increases with age, and its presence is known to accelerate the damage caused by atherosclerosis; thus, an increasingly greater percentage of patients coming for CABG have concomitant diabetes, currently estimated as a comorbidity in approximately 30% to 40% of CABG patients. Bucerius et al²²⁶ found diabetes to be associated with increased incidence of stroke and delirium, and prolonged ICU and hospital stay. Mortasawi et al²²⁷ and Nussmeier²⁶ describe diabetes mellitus as associated with increased incidences of stroke and mortality. In a large study, McKhann et al¹⁹ prospectively collected data on 2711 CABG patients and identified diabetes mellitus as an independent risk factor for both strok and encephalopathy. Part of the risk may involve cerebral hypoperfusion because cerebral oxygen desaturation during CPB has been documented in diabetic patients, with patients with insulin-dependent diabetes demonstrating the lowest values as measured via jugular oximetry and the poorest response to increases in MAP.²²⁸

Studies identify normoglycemia as a desirable perioperative goal in cardiac surgical patients regardless of whether they are diabetic.^223,225 There is both experimental and clinical evidence that hyperglycemia is associated with exacerbation of neurologic injury.²²⁹ Approaches to maintain serum glucose values less than 150 mg/dL have shown favorable results. Furnary et al²²¹ reported on 3554 patients who underwent CABG from 1987 through 2001, demonstrating that the observed mortality in the group managed with tighter glucose values was lower and concluded that continuous intravenous insulin infusion added a protective effect against death. Carvalho et al²²⁴ used an aggressive approach to maintain serum glucose values between 80 and 110 mg/dL and reported that this is a safely attainable goal. One study positively correlated average blood glucose on the first postoperative day with a variety of adverse outcomes (stroke, myocardial infarction, septic complication, or death). For each 1-mmol/L increase greater than 6.1 mmol/L (1 mmol = 18 mg/dL), risk increased by 17%.²³⁰ The ideal value of serum glucose in cardiac surgical patients remains unknown, but the evidence available suggests that maintenance of euglycemia is related to a better prognosis.

In accordance with these data, recent Society of Thoracic Surgeons guidelines have been published outlining recommendations for glucose control in diabetic and nondiabetic patients undergoing cardiac surgery with an overall recommendation that in both diabetic and nondiabetic patients, blood glucose levels should be maintained at less than or equal to 180 mg/dL with intravenous insulin as required.²³¹ However, because of concerns regarding potential adverse effects associated with hypoglycemia, including both increased risk for mortality associated with even a single episode of severe hypoglycemia as seen in medical/surgical intensive care patients,²³² and because in a randomized, prospective study of 400 cardiac surgical patients managed either with tight glucose control (intravenous insulin to maintain intraoperative glucose between 80 and 100 mg/dL) or conventional management (glucose level < 200 mg/dL) a significantly greater incidence of stroke was found in the treatment group,²³³ avoidance of hypoglycemia should be paramount. Accordingly, an important caveat recommending preservation of lower limit of glucose level greater than 100 mg/dL should be appended to the guidelines.²³⁴ Overall, it would appear that maintenance of perioperative serum glucose between 100 and 180 mg/dL in both diabetic and nondiabetic patients is desirable.

Hemodynamic Instability

Hemodynamic complications, either before, during, or after surgery, have been found to increase cerebral injury in cardiac surgical patients. In Bucerius et al’s⁵ report, ejection fraction less than 30%, urgent operations, and preoperative cardiogenic shock were related to increased postoperative delirium. Ridderstolpe et al²⁷ found that hypotension and postoperative arrhythmias were related to cerebral complications, whereas Stanley et al¹³² reported that postoperative atrial fibrillation was related to increased cognitive decline. Ganushchak et al¹⁷² reported in a retrospective analysis of 1395 patients that the frequency of neurologic complications was 3.9% in the group of patients who experienced large fluctuations in hemodynamic parameters while on CPB, whereas in the group of patients with more stable values on CPB, the incidence rate of neurologic complications was 0.3%. These studies indicate an increased susceptibility of the brain in cardiac surgical patients to apparently “benign” hemodynamic alterations that either produce or enhance cerebral injury, probably through hypoperfusion of the brain tissue, particularly because it has been estimated that more than 50% of CABG patients have coexisting cerebrovascular disease.^137,138

Cardiopulmonary Bypass Equipment

Early studies demonstrated increased microemboli in patients undergoing CPB using bubble oxygenators, with a reduction in cerebral embolization with the use of membrane oxygenators and arterial line filtration^{124,127,235–237} (Box 36-4). Using intraoperative fluorescein retinal angiography, Blauth et al¹²⁴ reported retinal microembolizations occurring during CPB with much greater frequency in patients in whom bubble versus membrane oxygenators were used. Using TCD for detection of cerebral emboli, Padayachee et al²³⁶ demonstrated continuous generation of cerebral emboli in all patients managed using bubble oxygenators and none in patients in whom a membrane oxygenator was used. They also demonstrated the efficacy of arterial line filtration to significantly decrease cerebral embolic load.²³⁷ It is apparent that emboli may be generated continuously during CPB and that equipment modification (e.g., arterial line microfiltration and preferential usage of membrane oxygenators) can decrease the generation of such emboli.²³⁸ Membrane oxygenators are currently recommended for CPB.¹³¹ (See Chapters 28 and 29.)

It is equally evident that equipment modifications, although decreasing the embolic load, cannot completely eliminate it.^94,236 Georgiadis et al²³⁹ used Doppler ultrasound and evaluated the percentage of MES reduction caused by the arterial filter and the proportion of MES actually reaching the brain by comparing the MES counts detected before the arterial filter, after the arterial filter, and in both middle cerebral arteries. Eleven patients underwent surgery using normothermic CPB, alpha-stat, a membrane oxygenator, and a 40-μm arterial filter. Evaluation of MES was only performed during extracorporeal circulation, was initiated after cannulation and clamping of the ascending aorta, and was terminated shortly before the aortic clamp was removed. The arterial filter resulted in a 58.9% reduction of MES, with only 4.4% (2624/59,132) of the MES detected after the arterial filter. The proportion of MES detected in the middle cerebral artery corresponded to the total cerebral perfusion under CPB, estimated as 5% to 10% of the total perfusion volume.²³⁹

Schoenburg et al²⁴⁰ used a dynamic bubble trap, incorporated in the arterial line after a 40-μm filter, to reduce the number of gaseous microemboli in 50 patients undergoing CABG. In 26 patients, a dynamic bubble trap was placed between the arterial filter and the aortic cannula (group 1), and in 24 patients, a placebo dynamic bubble trap was used (group 2) with TCD continuously measured on both sides during bypass, which was separated into four periods: phase 1, start of bypass until aortic clamping; phase 2, aortic clamping until rewarming; phase 3, rewarming until clamp removal; and phase 4, clamp removal until end of bypass. The bubble elimination rate during bypass was 77% in group 1 and 28% in group 2. The number of high-intensity signals was lower in group 1 during phase 1 (5.8 ± 7.3 vs. 16 ± 15.4; P < 0.05 vs. group 2) and phase 2 (6.9 ± 7.3 vs. 24.2 ± 27.3; P < 0.05 vs. group 2) but not during phases 3 and 4.²⁴⁰ The authors found that the dynamic bubble trap can remove gaseous microemboli. Unfortunately, no psychometric studies were performed on either group of patients, but future research in this area might yield positive influence on neurologic outcome by decreasing gas microemboli. Other investigators have demonstrated that air within the venous line of the CPB circuit, resulting from air entrainment at the venous cannulation site, injection of drugs into the venous line, or use of cardiotomy suction can pass through the oxygenator and appear as microemboli within the arterial line, even in the presence of a venous line defoamer and with use of a membrane oxygenator.^239,240 Georgiadis et al²³⁹ used tubing systems that included an arterial line 40-μm filter and demonstrated that the arterial filter resulted in a 58.9% reduction of microemboli signals with only 4.4% of the MESs detected after the arterial filter being actually detected in the middle cerebral artery.

Modification of the inflammatory response to the CPB using modified surface CPB circuits and leukocyte-depleting filters has been explored. Hamada et al¹³⁴ examined the combined use of heparin coating of the CPB circuit and a leukocyte-depleting arterial line filter in 30 patients allocated randomly to equal groups with a conventional circuit and arterial line filter, a heparin-coated circuit with a conventional filter, or a heparin-coated circuit with a leukocyte-depleting arterial line filter. Plasma interleukin-6 and -8 concentrations in the heparin bonded with leukocyte-depleting filter group were lower than in the conventional circuit group. Although a decrease in inflammatory mediator release has been observed by using leukocyte-depleting filters, the impact of these types of filters on neurologic outcome is less clear. In a meta-analysis of 28 relevant clinical studies, Whitaker et al²⁴³ concluded that conventional arterial line filtration had a definite effect in reducing neuropsychological deficit post-CPB. The results of studies using the leucocyte-depleting filter were less clear-cut.

In an attempt to decrease emboli originating from the surgical field, cell savers have been used for processing cardiotomy suction blood before returning it to the CPB circuit. Jewell et al²⁴⁴ reported on 20 patients prospectively randomized to either cell saver or cardiotomy suction and demonstrated that compared with cardiotomy suction, cell saver removed significantly more fat from shed blood, such that the percentage reduction in fat weight achieved by cell saver or cardiotomy suction was 87% compared with 45%. de Vries et al²⁴⁵ published a study on patients randomly assigned to have a fat removal filter for the cardiotomy suction. The fat filter removed 40% fat, leukocytes, and platelets from cardiotomy suction blood during cardiac surgery compared with the control group without the filter.

In addition, various intraoperative manipulations, particularly instrumentation of the atherosclerotic aorta, are independent risks for the generation of cerebral emboli and likely produce particulate or macrogaseous emboli, rather than oxygenator-generated microgaseous and microaggregate emboli.^{80,84,246,247} Avoidance of manipulation of a diseased aorta seems to decrease embolization and cerebral injury.^91,97,248 An alternative approach, emboli reduction by capture using an intra-aortic filter inserted through a side chamber of a modified aortic cannula, has also been assessed (Figure 36-13).

Figure 36-13 The “Brain Milieu” concept. Cerebral injury in cardiac surgery is the product of concurrent risk factors related to the surgical procedure and the patient’s previous condition. This unique combination of embolic, perfusion, inflammatory, coagulation, and metabolic events yields a broad spectrum of neurologic manifestations. The graphic shows factors and potential interventions to decrease neurologic morbidity in cardiac surgery. These are numbered clockwise: 1, tight perioperative control of glycemia; 2, epiaortic scanning to detect severe atheromatosis of ascending aorta; modified surgical technique to decrease or avoid aortic manipulation; 3, early pharmacologic and/or mechanical support to attain stable circulatory conditions; 4, cardiopulmonary bypass (CPB) circuitry of reduced and modified surface, use of arterial line 40-μm filters, alpha-stat CO₂ management, mild/moderate hypothermia, use of cell savers to process shed blood before reinfusing to CPB circuitry, arterial inflow temperature less than 37°C; and 5 and 6, intraoperative neurophysiologic monitoring (central venous pressure on CPB, brain oximetry [near-infrared spectrophotometry], retinal and transcranial Doppler, evoked potentials), and maintenance of cerebral perfusion pressure greater than 60 mm Hg.

In a nonrandomized study, Schmitz et al¹³³ examined the impact of intra-aortic filtration during CPB.^134,249 Three hundred four cardiac surgical patients had intra-aortic filtration using a 150-μm net deployed through the aortic cannula, whereas a further 278 patients formed the control group. Patients in the filter group experienced a lower incidence of adverse neurologic outcomes than patients in the control group (4.3% vs. 11.9%), with significantly fewer TIAs (0% vs. 1.4%), delirium (3.0% vs. 6.5%), and memory deficit (1.3% vs. 6.2%). There were also fewer strokes in the filter group compared with the control group (0.7% vs. 2.2%). Although the sample size was relatively small and the patients were not randomly assigned, the study findings suggest a protective effect of intra-aortic filters on CNS injury.¹³³ In a large, multi-institutional, prospective, randomized trial of 1289 patients in whom intra-aortic filtration or conventional CPB cannula was used, emboli were identified in 598 (96.8%) of 618 filters successfully deployed, and their post hoc analysis indicated a reduction in postoperative renal complications.²⁴⁹ In other studies, intra-aortic emboli trapping devices have been used with varied results.^97,239 In a small study by Eifert et al⁹⁴ in 24 patients, the use of the intra-aortic filter device did not show any difference in neurologic, neuroradiographic, or neuropsychological outcomes, yet the intra-aortic filter was effective in capturing particulate material. Similar efficacy in capturing intraoperative material using intraaortic filters during CPB was reported by Reichenspurner et al.²⁵⁰ It does appear as though intra-aortic filters can be safely deployed, and that they do capture particulate emboli, the predominant origin of which is atheromatous.

In a recent study of 150 CABG patients randomized to either intra-aortic filtration, dynamic bubble trap or conventional management, no difference was found in the overall incidence of postoperative MRI-detected small ischemic brain lesions (17/143) between groups, whereas dynamic bubble trap was associated with significantly fewer TCD-detected cerebral emboli and improved cognitive performance at 3 months after surgery in comparison with control and intra-aortic filtration groups.²⁵¹

Applied Neuromonitoring

Intraoperative neurophysiologic monitoring may be of benefit to decrease CNS injury.²⁵² Intraoperative TCD has been demonstrated to detect embolic events in real time and allows modification of perfusion and surgical techniques. It has been shown that the numbers of emboli generated by perfusionist interventions (e.g., drug injection, blood return), as well as episodes of entrainment of air from the surgical field, are rapidly identified and corrected by TCD detection of intraoperative emboli.²⁵³ (See Chapter16.)

Brain oximetry studies using noninvasive near-infrared spectrophotometry (NIRS) have shown promising results.^254–260 In Goldman’s²⁵⁹ large, nonrandomized cohort study, NIRS was used to monitor cerebral oxygen saturation in 1034 cardiac surgical patients and was compared with outcomes in 1245 patients who underwent cardiac surgery immediately before cerebral oximetry was incorporated. The study group had significantly more patients in New York Heart Association (NYHA) Classes III and IV than the control group, but the study group overall had fewer permanent strokes (10 [0.97%] vs. 25 [2.5%]; P < 0.044) and the proportion of patients requiring prolonged ventilation was significantly smaller in the study group, as was the length of hospital stay.²⁵⁹ Murkin et al²⁶⁰ reported a prospective, blinded, randomized study of NIRS cerebral oximetry in 200 cardiac surgical patients and demonstrated significantly fewer adverse clinical outcomes in NIRS versus control groups (P = 0.027).

Even during beating-heart procedures, compromised cerebral perfusion can occur relatively frequently, and if unrecognized, may account for the relative lack of difference in CNS outcomes between CABG and OPCAB surgery.²⁶¹ Combined EEG and cerebral oximetry identified episodes of cerebral ischemia in 15% of a series of 550 beating-heart patients; all were treated successfully by a combination of pharmacologically improved cardiac output, increased perfusion pressure, and cardiac repositioning.²⁶¹

In a study utilizing cerebral oximetry in 265 patients undergoing primary CAB surgery and randomized to active monitoring and a series of interventions designed to improve rSo₂ or to a control group in which blinded monitoring was used, a significant association was found between prolonged cerebral desaturation and early cognitive decline, as well as a threefold increased risk for prolonged hospital stay.⁵⁴ However, cerebral desaturation rates were similar between groups and ascribed to poor compliance with the treatment protocol, resulting in no difference in the incidence of cognitive dysfunction between groups. In a study of 103 patients undergoing valvular heart surgery in whom blinded cerebral NIRS monitoring was used, cerebral oxygen desaturation was again found to be associated with significantly longer duration of postoperative hospitalization.²⁶² In a prospective, randomized, blinded study in 200 patients undergoing coronary artery grafting, Murkin et al²⁶⁰ demonstrated that active treatment of declining cerebral rSo₂ values prevented prolonged cerebral desaturations and was associated with a shorter ICU length of stay and a significantly reduced incidence of major organ morbidity or mortality in comparison with a similar control group. In this study, the intervention protocol undertaken to return rSo₂ to baseline resulted in a rapid improvement in rSo₂ in 84% of cases and did not add undue risk to the patient, including no increase in allogeneic blood transfusions.^260,263 There were also numerically fewer clinical cerebrovascular accidents in the monitored patients directionally consistent with previous studies.²⁵⁹ As such, a physiologically derived treatment algorithm for management of perioperative cerebral oxygen desaturation has been proposed and is shown in Figure 36-14.²⁶⁴ An important confounder in evaluating the role of cerebral NIRS devices is the efficacy of treatment for cerebral desaturation.

Figure 36-14 Algorithm for the use of brain oximetry.

CT, computed tomography; ICHT, intracranial hypertension; MAP, mean arterial pressure; MRI, magnetic resonance imaging.

(Reprinted from Denault A, Deschamps A, Murkin JM: A proposed algorithm for the intraoperative use of cerebral near-infrared spectroscopy.Semin Cardiothorac Vasc Anesth 11:274–281, 2007.)

Pharmacologic Cerebral Protection

In general, pharmacologic protection from cerebral ischemia remains an elusive goal. Although there had been one clinical study in which a significant reduction in persistent neurologic defects after open-chamber cardiac surgery was reported after administration of high-dose thiopental, this was not confirmed in closed-chamber CABG procedures.^282,283 Other data have suggested that if there is any such thiopental-derived cerebroprotective effect, the mechanism may be caused by a metabolically driven decrease in CBF, with a concomitant reduction in the delivery of emboli into the brain, rather than occurring primarily as a result of a decrease in CMRO₂.²⁸⁴ However, a three-center study in 225 patients undergoing mitral or aortic valve surgery and randomized to high-dose propofol with induction of burst-suppression on EEG or a sufentanil control group was unable to detect any significant differences in neurologic or neuropsychological outcomes between groups.²⁸⁵ These authors concluded that neither cerebral metabolic suppression nor reduction in CBF reliably provides neuroprotection in open-chamber cardiac surgery. A large trial of perioperative nimodipine in valve replacement surgery had to be terminated prematurely after enrollment of only 150 of 400 patients because of a significant increase in major bleeding and an increased mortality in the 6-month follow-up period.²⁸⁶

There are some interesting associations between certain drug therapies having anti-inflammatory properties and lowered incidences of stroke and adverse CNS events. In a retrospective review of 2575 CABG patients, Amory et al²⁸⁷ reported that patients who received perioperative β-blockers had a significantly lower incidence of severe neurologic outcomes versus those who did not receive these drugs, demonstrating a 1.9% incidence rate of stroke and coma versus 4.3%.

However, concerns regarding perioperative β-blocker therapy has been raised by the results of the PeriOperative ISchemic Evaluation (POISE) trial in which 8351 patients with, or at risk for, atherosclerotic disease who were undergoing noncardiac surgery were randomized to receive extended-release metoprolol (n = 4174) or placebo (n = 4177).²⁸⁸ Although significantly fewer patients in the metoprolol group than in the placebo group had a myocardial infarction, there were more deaths in the metoprolol group than in the placebo group (3.1% vs. 2.3%), and more patients in the metoprolol group than in the placebo group had a stroke (1.0% vs. 0.5%).²⁸⁵ The implications of this study for cardiac surgical patients remains unclear, but there does not appear to be any increased risk for perioperative stroke in noncardiac surgery patients chronically maintained on β-blocker therapy.²⁸⁹

The serine protease inhibitor aprotinin has been shown to positively impact coagulation and inflammatory alterations triggered by CPB and has also been associated with decreased incidences of stroke and major CNS injury in cardiac surgical patients.^135,136 Aprotinin has also been demonstrated to have anti-inflammatory and antithrombotic effects.²⁹⁰ Frumento et al⁸ retrospectively analyzed stroke outcomes in 1524 patients undergoing cardiac surgery and identified a subset of 149 patients older than 70, with history of hypertension, diabetes mellitus, stroke or TIA, and presence of aortic atheroma who were deemed to be at similarly high risk for stroke. The authors reported that in this high-risk subset of patients, intraoperative administration of full-dose aprotinin, but not half-dose, was associated with a lower incidence of stroke. However, because the clinical usage of aprotinin has been suspended indefinitely because of several reports of increased mortality and adverse events associated with aprotinin therapy in cardiac surgical patients,^291–293 the future of this drug remains controversial.^294,295

In a prospective study of 5065 patients undergoing CABG conducted at 70 centers in 17 countries, the relation between early aspirin use and fatal and nonfatal outcomes was investigated.²⁹⁶ Among patients who received aspirin within 48 hours after revascularization, subsequent mortality rate was 1.3% compared with 4% among those who did not receive aspirin during this period. Aspirin therapy was associated with a 48% reduction in the incidence of myocardial infarction (2.8% vs. 5.4%; P < 0.001), a 50% reduction in the incidence of stroke (1.3% vs. 2.6%; P = 0.01), a 74% reduction in the incidence of renal failure (0.9% vs. 3.4%; P < 0.001), and a 62% reduction in the incidence of bowel infarction (0.3% vs. 0.8%; P = 0.01). Taken together, these studies seem to indicate that therapies associated with decreased inflammatory and sympathetic responses appear to be associated with decreased incidence of stroke and adverse CNS events.

Several preliminary studies had suggested that intraoperative administration of lidocaine infusion during cardiac surgery was associated with a decreased incidence of postoperative cognitive dysfunction.^297,298 However, in larger, randomized, prospective trials, this was not demonstrated.²⁹⁹ In view of these divergent results, whether lidocaine infusion will have any further role as a cerebroprotectant remains uncertain.³⁰⁰ An increasingly promising line of investigation for cerebral protection is the role of perioperative statin therapy, with evidence accruing that increased statin dosages and combination with other antiinflammatory agents such as angiotensin converting enzyme inhibitors can significantly decrease inflammatory markers and associated stroke risk.^301,302

Although these results suggest that there is as yet no pharmacologic “magic bullet” that can be used to reduce neurologic injury in patients undergoing cardiac surgery, a combination of technical and pharmacologic measures is currently available that might positively affect the CNS outcomes of these patients.¹⁶⁰ In patients identified as being at risk for perioperative cerebral injury, preventive measures as outlined in Box 36-4 should be instituted with organ-targeted management to guide the whole intraoperative and postoperative period. The Best Practice Cardiopulmonary Bypass Group summarized these measures as the avoidance of embolic injury by means of intraoperative EAS before aortic instrumentation, the employment of arterial line filters, modified-surface and reduced-area CPB circuits and membrane oxygenators, minimization of transfusion of unprocessed cardiotomy suction blood, the use of alpha-stat pH management during moderate hypothermic CPB, monitoring cerebral venous outflow pressure via proximal jugular venous pressure, avoidance of hypotension, the avoidance of cerebral hyperthermia during rewarming, maintenance of euglycemia, and the use of “tepid” rather than normothermic perfusion during CPB. As the age and incidence of comorbid disease in the cardiac surgical population continue to increase, the importance of these issues becomes ever more acute. In summary, primary prevention continues to be the only effective measure to decrease cerebral injury in cardiac surgical patients.